V2 Materials Technology 2

Volume 2
Materials Technology II
1. Clinker Burning: Material Technological Aspects....................................................... 2

2. Portland Cement Clinker.............................................................................................44
3. Refractory Materials and Coating Formation............................................................ 92
4. Rings, Balls and Build-Ups.......................................................................................136
5. Fuels for Cement Kilns..............................................................................................160
6. Clinker Burning Behaviour of Heavy Metals............................................................ 182
7. Kiln Dust..................................................................................................................... 223
8. Quality Assurance.....................................................................................................262
9. Quality Control: Concept, Control Plan, Procedures.............................................. 291
10. Basic Principles of Sampling and Testing................................................................ 323
11. Chemical and Physical Methods of the Laboratories.............................................. 368
12. Automation in Quality Control..................................................................................391
13. Statistics..................................................................................................................... 414
14. Case Study.................................................................................................................. 521
Cement Manufacturing Course Volume 2 - Page 1

Chapter 1
Clinker Burning: Material Technological Aspects

Clinker Burning - Material Technological Aspects
1. introduction.................................................................................................................... 4
1.1 Important Features...................................................................................................... 4
1.2 Control of the burning process.................................................................................... 5
1.3 Basic Aspects of the Material Changes Involved in Clinker Formation........................5
2. Reaction pathways encountered during clinker formation......................................... 5
2.1 General Considerations.............................................................................................. 5
2.2 Basic Sequence of Reactions..................................................................................... 6
2.3 Mineralogical and Chemical Characteristics of Raw Mixes......................................... 7
2.4 Intermediate Products................................................................................................. 9
2.5 Liquid Phase............................................................................................................... 9
2.6 The Overall Reaction Sequence............................................................................... 13
3. Reaction mechanisms.................................................................................................15
3.1 Definitions................................................................................................................. 15
3.2 Examples.................................................................................................................. 17
4. Kinetics of clinker burning.......................................................................................... 23
4.1 General Aspects........................................................................................................ 23
4.2 Practical Considerations........................................................................................... 25
4.3 Assessment of Raw Meal Burnability........................................................................ 30
5. Thermochemistry of clinker formation.......................................................................38

凾嚙 ࿿࿿࿿࿿࿿࿿࿿࿿࿿È䱔 ࿿࿿࿿࿿࿿࿿࿿࿿࿿É붚‫ވ‬࿿࿿࿿࿿࿿࿿࿿࿿࿿Êꕦ䟊 ࿿࿿࿿࿿࿿࿿࿿࿿࿿Ë鶾䠤 ࿿࿿࿿࿿࿿࿿࿿࿿࿿Ì崌Ἆ࿿࿿࿿࿿࿿࿿࿿࿿࿿
࿿࿿࿿࿿࿿࿿࿿࿿࿿Ò햼报࿿࿿࿿࿿࿿࿿࿿࿿࿿Ó࿿熏 ࿿࿿࿿࿿࿿࿿࿿࿿࿿Ô祼睵 ࿿࿿࿿࿿࿿࿿࿿࿿࿿Õ‫ﮌ‬ᘦ ࿿࿿࿿࿿࿿࿿࿿࿿࿿Ö番㥗 ࿿࿿࿿࿿࿿࿿࿿࿿࿿
࿿࿿ÛဂᏢ࿿࿿࿿࿿࿿࿿࿿࿿࿿Ü援╽࿿࿿࿿࿿࿿࿿࿿࿿࿿Ýᆲ烬࿿࿿࿿࿿࿿࿿࿿࿿࿿Þì⻚ ࿿࿿࿿࿿࿿࿿࿿࿿࿿ß遲猠 ࿿࿿࿿࿿࿿࿿࿿࿿࿿à頖徲 ࿿࿿࿿࿿࿿࿿࿿࿿࿿
࿿å釰滎 ࿿࿿࿿࿿࿿࿿࿿࿿࿿æ혀俄࿿࿿࿿࿿࿿࿿࿿࿿࿿ç䈐寽 ࿿࿿࿿࿿࿿࿿࿿࿿࿿è侊癗 ࿿࿿࿿࿿࿿࿿࿿࿿࿿é筦 ត࿿࿿࿿࿿࿿࿿࿿࿿࿿ê INTRO
DUCTION
In the production of Portland cement a mixture of natural minerals is transformed into an
intimate mixture of synthetic minerals having the required hydraulic properties.
Natural minerals Temperature (T) → Synthetichydraulic minerals
time (t),pressure (p)
This process is similar to that occurring during the transformation of sedimentary and
igneous into metamorphic rocks, but with the exception that the time (t) is much shorter and
that the pressure (p) is only 1 bar. Consequently, a high temperature is required.
During the transformation into clinker, two principal steps may be differentiated, namely the
disintegration of the original rock structures followed by the formation of new ones. Even
these categories can be subdivided as follows:
Disintegration of original structure
0 mechanical crushing and grinding
1 thermal decomposition
2 structural rearrangement on heating(e.g. polymorphism)
3 melting
Formation of new structures
0 occurrence of intermediate products
1 genesis and growth of final clinker minerals
2 crystallization of liquid phase
1.1 Important Features

Several important features characterize the clinker formation process. These are:
5888 the clinker burning system is complex and occurs via a series of diverse
mechanisms
5889 the transformation requires mechanical, thermal and electrical energy
5890 the reaction rate is slow in comparison with homogenous reactions (i.e. gas
or liquid phase reactions). This leads to the necessity of high temperatures, extended
reaction times, and finely dispersed materials
5891 the desired hydraulically active clinker minerals are not all stable at normal
temperature, and so the hot clinker must be rapidly cooled
5892 the quality of the product is determined by the
23 clinker chemistry
24 clinker microstructure
1.2 Control of the burning process
The transition into clinker is influenced by both material and process technological aspects,
namely:
5888 Material technological aspects
5888 raw meal burning behavior including burnability, dust formation, coating
behavior, granulation of clinker, etc.
5889 The quantity and properties of the liquid phase
5889 Process technological aspects
5888 temperature profile
5889 kiln atmosphere
5890 fuel type
5891 flame characteristics
5892 etc.
1.3 Basic Aspects of the Material Changes Involved in Clinker Formation

To properly understand the material changes occurring, it is an advantage to have a good
knowledge of the most basic characteristic aspects associated with clinker formation. They
are listed in the following table:
Table 1 Important aspects of the understanding of chemical reactions

Aspect Characteristic Features
Reaction Pathway indicates the intermediate products occurring
between reactants and products
Reaction Mechanism type(s) and reaction(s) taking place
Reaction Kinetics indicates rate at which the final products are
produced
Reaction Thermodynamics dictates whether reaction will at all be possible, and
what the heat and temperature requirements will be
It is therefore logical, when considering the production of Portland cement clinker, that
consideration should be given to these and earlier mentioned points. In the following pages, a
review will be given of the principal material aspects associated with clinker formation.
23 REACTION PATHWAYS ENCOUNTERED DURING CLINKER FORMATION
2.1 General Considerations

Cement clinker is an intermediate product in the manufacture of the hydraulic binding agent
known as Portland cement. By means of a single thermal treatment, a mixture of the non-
hydraulic, naturally occurring minerals calcite, quartz, clays and feldspars, it is transformed
into the intimate mixture of hydraulically active minerals alite, belite, aluminate and ferrite,
called Portland cement clinker. Like other products such as ceramics, refractories, bricks,
etc., cement clinker can be considered as being synthetically produced rock. The thermal
treatment of heating and cooling, responsible for this “metamorphosis”, is termed
“clinkering” and, with the exception of a small number of shaft kilns, is carried out in one of
the several types of rotary kilns. As has already been discussed, only a short specific range
of mixtures of calcareous and siliceous minerals can be transformed into Portland cement
clinker.

To fully describe the pathway of clinkering, it is necessary to consider the following aspects:
23 the chemical and mineralogical content of the raw mix
24 the overall sequence of reactions
25 the chemical and mineralogical nature of the intermediate products
2.2 Basic Sequence of Reactions

During the heating and cooling of the clinker formation process, a complex series of
reactions takes place within the rotary kiln, exhibiting a wide spectrum of reaction
mechanisms. Concerning first the reactions involved, the following summarizes the general
sequence of events taking place during the formation process.
Reactants + products Alite + Belite

Raw mixture → + → + → Cooled
(20°C) Intermediate products Melt (1450°C) clinker
(450 - 1300°C)
As can be seen, the raw mixture is not transformed into the final cement clinker by rapid
subjection to the required clinkering temperature. The process of clinker formation, as
demanded by the construction of the rotary kiln, takes place over a wide range of
temperatures, during which a complex mixture of reactants, intermediate and final products
often coexist. Even after reaching the maximum temperature (approx. 1450 °C), a certain
time must also elapse before a clinker of acceptable quality is produced.
In somewhat more detail, the principal sequence of events occurring and the temperature
range over which they take place, can be represented as shown in Table 2:
Table 2 Sequence of reactions occurring in a rotary kiln

Temperature range (°C) Type of reaction
Heating:
20 - 100 Evaporation of H2O
100 - 300 Loss of physically adsorbed water
400 - 900 Removal of structural H2O (H2O and OH
groups) from clay minerals
>500 Structural changes in silicate minerals
600 - 900 Dissociation of carbonates
>800 Formation of belite, intermediate products,
aluminate and ferrite
>1250 Formation of liquid phase (aluminate and
ferrite melt)
approx. 1450 Completion of reaction and recrystallization
of alite and belite
Cooling
1300 - 1240 Crystallization of liquid phase into mainly
aluminate and ferrite

From Table 2, several points can be observed, namely:
5888 that the dissociation of carbonate minerals does not take place at any single
temperature but occurs over a temperature range (600 - 900 °C)
5889 that, with respect to the formation of clinker minerals, three important
stages can be distinguished:
5890 formation of belite, aluminate and ferrite over a wide temperature range
starting at approx. 800 °C
5891 start of appearance of liquid phase and commencement of formation of
principal clinker phases (alite) at approx. 1250 °C
5892 that the aluminate and ferrite found in the final clinker are formed by
crystallization of the liquid phase during the cooling cycle. The individual aluminate and
ferrite crystals formed during the heating cycle at >800 °C loose their identity by melting
(> 1260 °C) to form the liquid phase
2.3 Mineralogical and Chemical Characteristics of Raw Mixes

During a detailed study performed by the Materials Division of the raw meals and slurries
from all the plants of the Holcim Group, a mineralogical and chemical analysis of a large
number of mixes was undertaken. The following Tables 3 and 4 list the large variety of
observed minerals and the range of raw mix chemistry.
Table 3 Minerals frequently found in cement raw mixes

Type Example
Carbonates calcite CaCO3, aragonite CaCO3, dolomite
CaCO3 ⋅ MgCO3, ankerite CaCO3 ⋅ (Mg, Fe) CO3,
magnesite MgCO3, siderite FeCO3
Simple Oxides quartz SiO2, cristobalite SiO2, hematite Fe2O3,
magnetite Fe3O4
Feldspars potassium feldspars (Na, K) Si3O8 and
plagioclase series
(Na, Ca) (Si, Al) Al2Si2O8
Sheet silicates* minerals of the mica and chlorite groups (e.g.
biotite, muscovite, chlorite), clay minerals (e.g.
kaolinite, montomorillonite, illite, palygorskite) and
other sheet silicates (e.g. pyrophyllite) *
Hydroxides* Al-hydroxides (e.g. boehmite), Fe-hydroxides
(e.g. goethite, limonite)
Sulfides and sulfates pyrites FeS2, anhydrite CaSO4, gypsum Ca SO4 ⋅
2H2O
Fluorides fluorspar CaF2
*see chapter 4 for chemical formulae

Table 4 Range of chemical compositions (dry basis) from which OPC clinker is
normally produced
From the preceding data it is evident that the various mixes are composed of very different
minerals and in addition vary within a certain range in their chemical composition. In spite of
these differences, Portland cement clinkers composed of the same hydraulically active
minerals are produced from all these mixes. The fact that such a diversity of mixes results in
the same end product, underlines the complexity of the transformation process.

2.4 Intermediate Products
The clinkering reaction does not take place in a single direct step from the natural to the final
minerals, but proceeds via intermediate products. These intermediate products can be
observed by heating up raw mixes in the laboratory or by withdrawing samples from various
points along the kiln. This latter method, although more difficult to carry out, has the
advantage over the laboratory method in that those intermediate minerals, whose presence
is also dependent on the kiln gas atmosphere, are observed. The following table lists the
most commonly encountered intermediate minerals.
Table 5 Intermediate minerals encountered during clinker production

Type Mineral Name Chemical Composition
Simple Sulfates anhydrite CaSO4
arcanite K2SO4
Compound Sulfates “sulfate”-spurrite 2(C2S) ⋅ CaCO3
“calcium”-langbeinite K2Ca2 (SO4)3
Compound Carbonates spurrite 2(C2S) ⋅ CaCO3
Simple Chlorides sylvite KCl
Calcium Aluminates mayenite 12 CaO ⋅ 7 Al2O3
CaO ⋅ Al2O3
Calcium Ferrites 2 CaO ⋅ Fe2O3
Calcium Alumino-Silicates gehlenite 2 CaO ⋅ Al2O3 ⋅ SiO2
That these materials are formed at all, when it is considered that the final equilibrium
products are alite, belite, aluminate, and ferrite, can be explained as follows:
Table 6 Reasons for the formation of the intermediate products
23 the intermediate products are preferentially formed by kinetically faster reaction rates
24 the intermediate products are the reaction products of localized zones in the meal charge,
i.e. local equilibrium but not overall equilibrium was reached (for example gehlenite
formation)
25 the intermediate products are really the equilibrium products at the given temperature and
gas atmosphere, but not at the final clinkering temperature (the formation of spurrite is a
typical example)
The most important intermediate product, however, is not a crystalline product at all, but is
the liquid phase first formed above ca. 1250 °C, and which only forms crystalline products
(mainly aluminate and ferrite) after cooling.
2.5 Liquid Phase

The composition of the raw mix determines
5888 temperature at which liquid will first be formed
5889 amount of liquid formed at any given temperature
5890 the physical properties of the liquid at any particular temperature, especially its
viscosity

2.5.1 Temperature of Liquid Formation
The following are the lowest melting temperatures of various mixtures pertinent to Portland
cement production:
Components Temperature °C of
liquid formation
C-S 2065
C-S-A 1458
C- S-A- F 1338
C- S-A- F- M 1300
C- S-A- F- M-N 1280
Due to the inclusion of additional oxides (e.g. K2O, TiO2, Mn 2O3, etc.) the lowest
formation temperature of the liquid phase in clinker is in reality around 1250°C.
2.5.2 Quantity of Liquid
Although most raw mixes show about the same minimum temperature of liquid formation, the
quantity of liquid formed at this and progressively higher temperatures varies according to
the raw mix chemistry.
In the Portland cement relevant parts of the system C - S - A - F, in which melting begins at
1338 °C, the composition of the liquid is:
CaO - 55 %
SiO2 - 6% Alumina ratio
Al2O3 - 22 % (AR) = 1.38
Fe2O3 - 16 %
The percentage liquid forming at various temperature (1338, 1400, and 1450 °C) is given in
Table 7.
Table 7 Formulae used for estimating the quantity of the liquid phase according
to Lea & Parker
Temperature °C Quantity of liquid phase

1338 AR > 1.38: 6.1 Fe2O3 + MgO + Na2O + K2O
AR < 1.38: 8.5 Al2O3 - 5.22 Fe2O3 + MgO + Na2O + K2 O
1400 2.95 Al2O3 + 2.2 Fe2O3 + MgO + Na2O + K2O
1450 3.0 Al2O3 + 2.25 Fe2O3 + MgO + Na2O + K2O
Because MgO has limited solubility in the liquid, there is a restriction of 2 % maximum MgO
which can be incorporated into these equations.
For the majority of clinkers, the quantity of liquid formed at 1450 °C varies according to the
above formula from approx. 20 to 30 %. Figure 1 shows the quantity of liquid formed at the
given temperature for several Group plant clinker, the influence of MgO and alkalis being
included.

Figure 1 Quantitative change of liquid phase with temperature
(INFLUENCE OF MgO, Na2O AND K2O INCLUDED)
Figure 2 displays graphically the influence of Al2O3 and Fe2O3 alone on the quantity of liquid
formed at 1338 °C, the solid contours representing the percentage of liquid formed. It can be
seen that the effect of adding Fe2O3 or Al2O3 on the proportion of liquid formed at 1338 °C is
influenced strongly by the alumina ratio. If in a clinker composition containing 4.5 % Al2O3
and 2 % Fe2O3, for example, the Fe 2O3 content is held constant and the Al 2O3 content is
increased, the percentage of liquid remains unchanged at 14 %, because the iso-Fe2O3
contour lies parallel to the iso-liquid contour in this region. If, however, the iron content is
increased and the Al 2O3 content remains constant, the proportion of liquid at first increases
rapidly until at 3.3 % Fe2O3 it attains a maximum of 21 %. With further increase in Fe2O3 the
quantity of liquid decreases. Therefore, the most effective use of Al2O3 and Fe2O3 - with
respect to liquid formation at 1338 °C - occurs when the two are used in the weight ratio of
1.38.

Figure 2 Quantity of liquid phase at 1338 °C
AR = Alumina ratio
OB, GM, UN, VN = Different Plants
At 1400 °C and above increasing Al2O3 and / or Fe2O3 always result in a larger quantity of
liquid phase, Al2O3 being the one with the larger influence.
2.5.3 Viscosity of Liquid Phase

The viscosity of the liquid phase diminishes exponentially with increasing temperature and at
1400 °C is reduced by addition of fluxing components in the following order:
Na2O < CaO < MgO < Fe2O3 < MnO
With increasing SiO2 content of the melt and to a lesser extent with increasing Al2O3,
appreciable increases in viscosity occur.

2.6 The Overall Reaction Sequence
Several different approaches have been used in the past to display the sequence of events
taking place during the formation of cement clinker. These differ mainly in the aspect
chosen, the following examples giving emphasis to differing features.
2.6.1 Display Based on Qualitative Change of Minerals from Samples taken from an
Operational Kiln
The minerals existing at various points during the production process in the kiln can be seen
in Figure 3. It is worth noting that the feed contains several minerals (e.g. KCl, spurrite,
C12A7, CaO) that were not present in the freshly milled raw mix. This is due to the return of
dust to the slurry feed.
Figure 3 Minerals identified at different locations (long wet kiln)

2.6.2 Schematic Display according to Chemistry
A recent example of this approach to clinker formation is proposed in Figure 4.
Figure 4 Sequence of compound formation according to chemical composition
CF
2C2S.CaCO3 2
In order to allow an overall picture to be quickly gained, it must be mentioned that the SiO2
content of the clay minerals and the pressure of feldspars have been omitted. Of interest in
the above display is that the presence of spurrite is considered as being a normal
intermediate product and not only a component found solely in kiln deposits.

2.6.3 Quasi-Quantitative Display of Minerals Present at Various Temperatures
An informative schematic display based mainly on laboratory studies depicts in a quasi-
quantitative manner changes of the various minerals with temperature. This is shown in
Figure 5
Figure 5 Quasi-quantitative variation of minerals with temperature
0 REACTION MECHANISMS
3.1 Definitions
For the discussion of reaction mechanisms the following concepts with regard to states of
matter and types of reaction will be employed.
3.1.1 States of Matter
All matter can be assigned at any specific temperature to at least one of the three defined states of matter,
namely to the solid, liquid, or gaseous state. Each of these states has specific properties which allow a
simple distinction between them to be made as follows:
State of Matter Properties
solid definitive volume and definite shape
liquid definitive volume, assumes shape of container
gaseous neither definitive volume nor definite shape
Under specific conditions (of temperature and pressure) it is possible for more than one state
to coexist with another. An example demonstrating this effect is, of course, water which at
various temperatures and pressures can coexist in the liquid/gas, solid/liquid and
solid/liquid/gas states.

3.1.2 Classification of Reactions
Chemical reactions may be classified according to their type, namely:
Type Example from clinker burning

structural change high quartz ↔ low quartz
decomposition CaCO3 ↔ CaO + CO2
combination 2CaO + SiO2 ↔ C2S
A further differentiation of these reaction types may be made:
0 According to the state of matter
reaction type Example from clinker burning

solid - solid quartz and free CaO → belite
solid - liquid liquid phase → crystallization of aluminate + ferrite
solid - gas CaCO3 ↔ CaO + CO2
liquid - liquid -
liquid - gas drying process, volatilization of alkalis, etc.
gas - gas CO + 1/2 O2 ↔ CO2
0 according to rate controlling step (kinetics of reaction)
Type Example from clinker burning

diffusion formation of alite
phase boundary quartz + free CaO → belite (initial reaction)
nucleation liquid phase → crystallization of aluminate + ferrite
alite formation
For many reactions the rate controlling pressure may not be known.
Since clinker burning is a most complex chemical process, it is not surprising that most of
the mentioned reaction types are operative during the process. In the following, examples of
these reaction types taking place in the kiln during production will be considered.

3.2 Examples
3.2.1 Structural Changes

This type of change occurring within single minerals in the solid state is also called a
polymorphic transition (see paper 4). In this process the chemical composition of the
minerals remains constant, but the spatial arrangement of the component atoms alters,
usually upon reaching a specific temperature. Relevant examples of such polymorphic
transition in clinker burning include the following:
a) The low - high quartz inversion (both have composition SiO2)
Figure 6 Arrangement of the atoms in low and high quartz
In this transition the distorted low from, stable under 573 °C, changes into the more ordered
high modification, the stable species above the transition temperature.
b) aragonite Î calcite inversion (both have composition CaCO3)
Aragonite present in the raw meal transforms on heating into calcite. Consequently, kiln dust
always contains calcite and not aragonite. The reason for this transition is, that under normal
atmospheric pressure calcite and not aragonite is the stable from of CaCO3. This can be
easily seen in Figure 7, showing the stability conditions of calcite and aragonite.

Figure 7 Stability conditions of the various forms of CaCO3
c) alite and belite transformations

Even more important than the polymorphic transitions in the raw materials are the structural
changes occurring in the principal clinker minerals alite and belite. These occur during both
formation and cooling of the clinker, and contribute to the complexity of the clinker system.
These transitions will be discussed in more detail in paper 13, pages 11 - 13.
3.2.2 Decomposition Reactions
The principal decomposition reactions encountered during clinker production include the
following:
solid / gas type
0 dehydroxylation of the clay minerals (kaolinite, etc.)
1 decarbonation of the carbonate minerals (magnesite, dolomite, calcite, spurrite)
solid / solid type
2 decomposition of alite
Characteristic of this reaction type is that the single reactant is transformed into two
products.
23 Dehydroxylation of the clay minerals
The removal of structural water from the clay minerals is a complex decomposition, as it
further involves structural changes of the dehydrated product. This may be demonstrated for
one of the simplest of the clay minerals compounds kaolinite, 2SiO2, ⋅ Al 2O3 ⋅ 2 H2O.
There are two alternative theories about the changes that kaolinite undergoes when it looses
water at about 500 - 600 °C. One theory is that the anhydrous alumino silicate 2SiO 2 ⋅ Al2O3 is
formed, the other than an intimate mixture of amorphous silica and alumina results

5888 Decarbonation of the simple carbonate
minerals These reactions are of the type
MCO3 ↔ MO + CO2
(solid) (solid) (gas)
where M = Ca (Calcite)
23 Ca/Mg (Dolomite)
24 Mg (Magnesite)
in which, if the CO2 is not removed, there exists an equilibrium between the reactant and the
reaction products. At any single temperature, both the degree of decarbonisation of the
carbonate and the partial pressure of CO2 (i.e. pCO2) are fixed, providing equilibrium has
been achieved. With increasing temperature the pCO2 value increases until a certain given
temperature the value equals atmosphere pressure as shown in Figure 8, so that the
carbonate will completely decompose.
Figure 8 Equilibrium dissociation pressure of calcite with temperature
Under conditions in which the CO2 is constantly removed (e.g. in a flowing gas atmosphere
at reduced pressure in which the pCO2 value is lower than the equilibrium level)
decomposition of the carbonate will take place at a reduced temperature.
In the simple carbonates found in cement raw mixes, the temperature at which the
equilibrium pCO2 level attains 1 bar is as follows:
Mineral Temp (°C) pCO2 = 1 bar
Calcite 890
Dolomite 745 and 890

Magnesite 650
23 Decomposition of compound carbonates (e.g. spurrite)
The decarbonisation of spurrite 2C2S ⋅ CaCO3 proceeds according to the equation
2C2S ⋅ CaCO3 ↔ 2C2S + CaO + CO2
spurrite (solid) belite (solid) free lime (solid) carbon dioxide (gas)
Just like the simple carbonates, the above equilibrium is reversible, and its extent depends
on the temperature and pCO2 value. Unlike the simple carbonates, two solid reaction
products are formed, belite and free CaO, which at the temperature associated with spurrite
decomposition under normal pressure, cannot react with each other.
The equilibrium pCO2 value for spurrite is given in Figure 9 along with that pertaining
to CaCO3.
Figure 9 Equilibrium dissociation pressure of calcite and spurrite with

temperature
From the figure it is evident that spurrite is thermally more stable than calcite. As a
consequence of the greater stability of spurrite, it may form in cement kilns either by direct
reaction of calcium carbonate and silica, or by reaction of CaO, Ca2SiO4 and gaseous CO2.
The formation of spurrite with the latter mechanism may occur only in a small range of
temperatures above the CaCO3 decomposition temperature, for example, between 890
°C and 912 °C at 1 bar CO2 pressure.
5888 Decomposition of alite
A reaction of the solid-solid decomposition type which can, under certain circumstances, occur
during clinker cooling, is the decomposition of alite. From the phase diagram for the system CaO
- SiO2 it can be deduced that C 3S has a lower limit of stability at 1250 °C, below which it will
decompose into two solids, C2S and CaO. The maximum rate of decomposition occurs at 1175
°C, at which temperature the S-shape decomposition curve (Figure 10)
indicates that the initial decomposition is controller by nucleation of the products on the
C3S surface.
Figure 10 Decomposition of C3S at 1175 °C
As can be deduced from figure 10, the rate of decomposition is so slow that for the cooling rates
encountered during clinker cooling, the decomposition of C3S should not be observed.
In the case of impure C3S, i.e. clinker alite, the rate of decomposition is
appreciably accelerated by:
23 the presence of lime and C2S nuclei
2+
24 the presence of Fe , H2O and K2SO4 / CaSO4
melts This aspect will be further considered in paper 13.

3.2.3 Combination Reactions
The main combination reactions are the formation of the two major phases, the calcium
silicates alite and belite, the only principal phases remaining unmelted at the highest
clinkering temperature. The other major phases, aluminate and ferrite, although formed by
combination reactions during heating, are melted at the highest temperatures and are
present in the final cooled clinker as crystallization products of the melt.
23 Belite formation
The formation of belite starts at ≈ 800 °C and is mainly formed by solid -state reaction.
In many raw mixes it is formed in quantities exceeding those found in the final clinker
(see Figure 5) and then further reacts with free lime (via the liquid phase) to yield alite,
the quantity of belite thus falling back to that finally observed.
Belite formation (see Figure 4) is the result of a combination between the calcite and silica
components of the raw mix.
2 CaCO3 + SiO3 ≈ 800°C 2 C3S CO3
calcite e.g. quartz belite +
Although the pCO2 value of calcite does not reach 1 bar until 890 °C, the initial formation of
belite does take place at lower temperatures, due to the stability of the C2S phase.
During the solid state reaction of silica and calcite (or CaO) the first reaction product is
always C2S. This can bee seen in Figure 11.
Figure 11 Sequence of appearance of phases in the solid state reaction of CaO

and SiO2
CaO SiO2 → CaO C2 S SiO2 → CaO C2S CS SiO2
showing the order of appearance of phases at a CaO / SiO2 interface. Only later does the
most silica rich phase CS appear as a result of reaction s between C2S and SiO2. This
demonstrates the remarkably strong nature of the belite formation reaction. At temperatures
at which solid state reactions occur, the formation of C3S does not take place due to its
thermal instability.
The rate limiting mechanism by which belite is formed (after an initial phase boundary
controlled reaction) depends on the diffusion of ions through the solid state. The rate of this
reaction is thus dependent on:
5888 the path distance that the diffusing species have to travel
5889 defects in the reactant’s crystal lattices.
The former factor is dependent on both the size of the CaO and SiO2 particles, and the
effective contact area between them. It is this easy to understand why finely milled and/or
densely intergrown powder mixtures react faster than large-grained coarse ones, and thus,
why cement raw meals must be finely ground and intimately mixed.
2+
From the results of many studies, the main diffusion species is known to be the Ca ion
which migrates into the solid SiO2, although evidence does exist to indicate that the diffusion
4+
of Si into CaO can occur to a lesser extent.
2+
That the Ca ion is the major diffusing species is confirmed by the well known fact in
clinker formation that the quartz must be finer ground than the limestone if complete
reaction is to take place.
23 Formation of alite
The formation of alite commences only when the temperature is above its lower stability limit
of ≈ 1250 °C. At that temperature, the liquid phase is also starting to form. Consequently the
formation of alite is a liquid - solid reaction, as reaction via a liquid medium is faster than by
solid state processes. In the laboratory, C3S can be formed by the solid state reaction of
CaO and SiO2, but this requires many hours at temperatures above those encountered
during clinker formation. The resulting finely crystalline product must be rapidly quenched, at
much faster rate than can be achieved in a production kiln, to prevent its decomposition at
normal temperatures.
In the industrial production of clinker, the formation of alite and its stabilization is thus wholly
dependent on the presence of the liquid phase.
At the temperature at which alite formation begins (≈ 1250 °C), the material consists mainly of free
lime, belite and liquid phase. The formation reaction is, thus, of the type
C2S + CaO liquid phase C3S
→
> 1250 °C
From the results of many laboratory experiments, the rate controlling step is considered to
be the movement (diffusion) of dissolved CaO through the melt. This is supported by the
following observations:
23 alite particles often contain belite inclusions (see paper 13)
24 when free lime is added to an already formed clinker, and the mixture again heated,
pores are to be found where the lime particles had been, surrounded by a zone
containing only alite.
Again, as in the case of belite formation, the rate of reaction is influenced by
5888 the distance the CaO has to travel (depending on raw mix fineness and
homogeneity), and in addition to the quantity and viscosity of the liquid phase
(controls the gas with which the movement of CaO can occur).
As the viscosity of the raw mix is reduced by
MnO > Fe2O3 > MgO > CaO> Na2O
it can thus be appreciated that the presence of Fe2O3 and MgO in the liquid phase assists
in the alite formation process.
0 KINETICS OF CLINKER BURNING
4.1 General Aspects

The burning of cement consists of a series of reactions taking place between finely divided
solids, and it is only at temperatures above 1250 °C that a liquid is formed and becomes the
medium through which reaction occurs. When liquids or solutions react together, the actions
usually occur rapidly and the products depend only on the composition of the reaction
mixture and the temperature. With reactions between solids the conditions are very different.
Reaction takes place at the surface of grains, and diffusion of fresh solid to the surface
proceeds but slowly. If a particle of clay is imagined surrounded by particles of lime, the
reaction will commence at the surface of the clay particle. This in turn will react with further
clay inside. If complete combination is not attained, the stage to which the reaction has
progressed depends not only on the temperature, but also on the time, or rate of heating and
the general physical and chemical condition of the reacting materials.
This phenomenon of the rate of reaction can be schematically represented as follows:

raw mix kT
→clinker
(1 - α) α
where kT is the rate of reaction at temperature T and α the extent of clinker formation after
a certain time. The rate of reaction kT is dependent on temperature and material
characteristics (mainly mineralogy and fineness as expressed in the fundamental
Arrhenius equation
Ea
− RT
kT =Ae
kT = rate constant at temperature T
T = reaction temperature
R = gas constant
Ea = activation energy
A = frequency factor
From the above equation it can be seen that the rate of reaction
0 increases with them T and contact surface between raw mix components
(frequency factor A)
1 decreases with higher activation energy Ea for raw mix components.
This can be illustrated as follows. Experience demonstrates that a raw mix composed of
non- reactive minerals (high activation energy Ea) such as quartz, forms clinker more slowly
than a raw mix incorporating more reactive components such as clay minerals or amorphous
silica. The lower activation energy required for the reaction with lime results in a high rate of
clinker formation. To compensate for the slow reactivity of the less reactive minerals, a
higher burning temperature and / or longer burning period (longer clinkering zone) is
required.
The quantitative characterization of the kinetics of clinker formation, however, is extremely
complicated, due to:
0 Portland cement having a range of chemical compositions and being produced from
a wide spectrum of different minerals
1 the transformation takes place via a series of sequential and simultaneous reaction steps
of various types
2 the burning process is not isothermal but occurs in a heating process
3 the non-availability of a simple, precise method for determining the extent of reaction (α)
Various attempts have, in spite of the complexity of the system, been undertaken to evaluate
the kinetics of clinker formation. An example is given in the following table.

Table 8 Various models for describing the kinetics of clinker formation
The main feature is that the process has been subdivided into four stages having different
reaction mechanisms and rates.
For further details Report No. MA 75/2408/E may be consulted.
4.2 Practical Considerations

When comparing the overall reaction velocity with the kinetics of the individual stages (Table
0 , it is known that the calcite decomposition, i.e. decarbonation of the raw meal, is not
critical. The combination reaction stages (solid/solid and solid/liquid) are the rate
determining steps in the process. The practical consequence of this is that it has been
possible to replace the long rotary kiln by a suspension preheater kiln, in which the calcite
decomposition may take place to a large extent within minutes. In the clinker formation
stages, however, a rotary part of adequate length so as to allow sufficient reaction time, is
indispensable. This is illustrated in Figure 12.

Figure 12 Temperature distribution in wet and dry process kilns
Consequently, to assess the progress of clinker formation, most consideration is given to

events occurring after decarbonation.

In practice, the most convenient method of following the reaction is by measuring the rate of
decrease of uncombined lime (i.e. free lime) . This technique is illustrated in Figure 13 showing
two raw mixes, I and II, of identical chemistry (LS = 95, SR = 3.2 and AR = 2.2) and
similar fineness (R200µ ≅ 0.5 %, R90 ≅ 7 % and R60 ≅ 15 %). It is evident that the difference in
mineralogy and actual particle size of the individual crystals influence both the mechanism
and rate of reaction, especially at start of the clinker formation.
As can be seen in Phototable 1, mix I is composed of limestone and the fine-grained shale A
which itself consists of finely mixed calcite (40 %) and siliceous (60 %) components, whilst
mix II is composed of the same limestone but with the much coarser grained shale S
containing a high proportion of coarse quartz particles. While raw mix I is considered to have
a good burnability , raw mix II is considerably more difficult to burn.
Figure 13 Temperature change of DTA curves, phases identified and free lime

Figure 13 a Temperature change of DTA curves, phases identified and free lime

Phototable I Texture and Mineralogy of Mix Components
Designation Limestone Shale A Shale S

Calcite 97 % ≈ 40 % ≈ 10 %
Dolomite ≈2% - -
Quartz traces ≈ 25 % ≈ 55 %
Chlorite - ≈ 20 % ≈ 10 %
Illite and Micas - ≈ 10 % ≈ 20 %
Pyrite traces ≈2% traces
Feldspars - ≈2% traces

4.3 Assessment of Raw Meal Burnability
From the foregoing it will be realized that a true determination of reaction kinetics is at best
feasible in fundamental research. In practice, simple methods are mostly applied to asses
the “burnability” of a mix, i.e. the ease of formation of the clinker minerals. Three
distinct methods are practiced at the Holcim Technical Center:
0 Holcim burnability test -
involves the heating of a raw mix under well defined laboratory conditions, and
subsequent laboratory determination of the non-combined lime.
1 Statistical burning model -
in which ten material parameters, proven to be of significance, influence the rate of
clinker formation. The uncombined CaO value, of any raw mix, relative to that of a
standard raw mix prepared from identical components, is calculated.
2 Physicochemical burning model -
requires no standard raw mix. Only 4 parameters need to be considered.
4.3.1 Holcim Burnability Test

The Holcim burnability test is based on the isothermal burning at 1400 °C of raw meal
nodules, details of which can be obtained in paper 5, page 5/18 ff.
By being carried out under standard laboratory conditions, this test allows the determination
of the relative influence of the various material parameters to be ascertained, free
from the influence of process technological disturbances.
Typical results obtained by this test for the raw mixes of all Group plants are listed in Figure
14, arranged in order of decreasing free CaO, i.e. increasingly better burnability.

Figure 14 Range of CaO values determined by the Holcim burnability test on plant
raw mixes
The free lime value of the laboratory clinker extended from a minimum value of 0.94 % to
a maximum value of 9.5 %. The values are distributed rather regularly around the mean
value of 4.7 %. Based on this distribution and the corresponding plant experience, the
burnability of raw mixes is ranked as follows:
Evaluation Free Lime %

very good 0-2
good 2-4
moderate 4-6
poor 6-8
very poor >8
This classification has been shown to be applicable over many years of practical experience.
With the same raw meals the free lime values of the industrial clinker vary within a much
smaller range, from a minimum value of 0.6 % CaOfree to a maximum value of 2.8 % CaOfree.
The mean free lime value of the clinker examined from the Holcim Group is about 1.3 %
CaO with a standard deviation of s = 0.4 %.
The difference between the distribution of free lime for the laboratory produced clinks of the
burnability test and plant clinkers can be seen in Figure 15.

Figure 15 Distribution of free CaO values for plant clinkers and for clinkers
produced by the Holcim burnability test
This narrow range of free lime for plant clinkers can be explained by the quality requirements
in the standards and of the market. The free lime values of industrial clinker, therefore, do
not offer an explanation for the actual burning behavior of the raw mixes. The variability of
raw mix laboratory burnability, as shown in Figure 14, is compensated by technological
measures in the plant.
4.3.2 Statistical Burnability Model
In a quantitative evaluation of the data obtained by the Holcim burnability test, a model has
been developed - Statistical Burnability Model. The initial data for development of the model
were the free lime values from burnability test at 1400 °C on 168 different raw mixes, the
chemical and mineralogical composition, and the overall fineness of these mixes.
With this model, the 1350, 1400 and 1450 free lime values of other raw mixes from the same
raw material components can be determined based on one single burnability test of one mix.
In this way, the burnability model can be used as an instrument for optimization of raw
mixes by allowing for the effect due to changes of the following factors:
Chemical Parameters lime saturation, silica ratio, alumina ratio,
Physical Parameters K2O + Na2O, MgO
residue on 200 and 90 µm sieves,
quantity of mica, quartz and iron
minerals
An impression of the use of the statistical burnability model may be gained from the data listed
in Table 9 and presented graphically in Figures 16 and 17. Using a standard raw mix (no. 2), of
overall fineness R90µ= 11 %, and determined free lime value of 4.9 % that would be observed
by varying the temperature, overall fineness and chemical composition have

been calculated. For a series of mixes from any one set of components it is cheaper,
quicker and equally accurate to use the model rather than burning each mix in the
laboratory.
Table 9 Influence of fineness and chemistry on laboratory burnability of a raw

mix

Figure 16 Change of free lime with raw meal fineness
Figure 17 Change of free lime with raw meal composition
Influence of burning temperature

As can be seen in Table 9 and Figures 16 and 17, increasing the burning temperature
results - as would be expected - in lower free lime values. Thus, for example, mix 2 (4.9 %
free lime at 1400 °C) would contain 6.3 % at 1350 °C and 3.5 % at 1450 °C. The influence of
reducing the temperature 50 °C to 1350 °C is thus roughly the same as that resulting by
increasing the lime saturation from 93 to 96%, as in mix 4.
Influence of raw meal fineness
Increasing the mix 2 residue on 90 µ from 11 to 17 %, results in the free lime at 1400 °C
increasing to 5.6 %, while reducing the residue to 8 % decreases the free lime to 4.3 %.
Increasing the residue to 17 %, thus has the same influence as that caused by raising the
silica ratio from 2.37 to 2.6. Most practical use of being able to calculate the effects due to
fineness is the assistance it gives when evaluating the optimum of raw meal mills.
Influence of LS, SR, and AR
The influence of increasing the LS, SR, and AR of mix 2 (LS 93, SR 2.7, and AR 1.72) can
be seen in Figure 17. Increasing the AR to 1.9 has practically no effect, increasing the SR
to 2.60 does, however, increase the free lime values (mix 5), but to a lesser extent than that
caused by increasing the LS to 96. Increasing both the LS to 96 and the SR to 2.6,
increases the free lime the most.

4.3.3 Physicochemical Burnability Model
Further investigations during recent years have contributed additional knowledge with regard
to burnability. It has long been known that the amount of uncombined lime depends on:
0 Specific reaction area, i.e. the area of contact between the grains
1 local oversaturation, i.e. grain size of individual minerals
2 ambient conditions, i.e. pressure, temperature, and burning time
3 diffusion coefficient of CaO through the liquid phase, i.e. composition of the liquid phase
4 amount of liquid phase formed during burning
5 supply and demand of CaO
It has recently been found that all these influencing factors may be incorporated in four
parameters, namely silica ratio, lime saturation, amount of oversized quartz grains and
amount of oversized calcite grains. Pressure, temperature and burning time are considered
to be constant.
Silica ratio (SR) and lime saturation (LS)
The formation of C3S from C2S and CaO is governed by the diffusion of CaO through the
melt. The silica modules and lime saturation are sufficient to describe this chemical reaction
quantitatively.
The amount of CaO which can be accommodated within the liquid phase and in which it can
diffuse and thus react, is inversely proportional to the silica ratio. A linear relationship exists
between max. lime saturation and silica ratio values at which no free lime can be observed.
Quartz and calcite grains
Whether a grain of material reacts fully under given burning conditions depends on its
diameter, structure and chemical composition. Monomineral particles above a certain critical
size do not react fully, even after a long burning period. They are equivalent to local
oversaturation of one chemical component. The overall raw mix residue is therefore not
always a wholly reliable criterion for assessment of the burnability, unless the character of
the grains is known. Quartz grains which exceed certain dimensions react with adjoining lime
grains to from belite clusters. This prevents further reaction and gives rise to free lime.
Too large calcite grains result in CaO not being completely combined as also results from
grains whose lime saturation is over 100 %.
For the Holcim burnability test conditions the following grain diameters, quartz ≥ 32 µm and
calcite ≥ 90 µm, were found to be critical limits.

Figure 18 Influence of chemistry and coarse grains on burnability of raw mixes
An impression of the advantages offered by the physiochemical model may be drawn from
the data presented in Figure 18 showing the influence of oversize quartz and calcite grains
and of lime saturation (LS) and silica ratio (SR) on the uncombined CaO content of various
industrial raw mixes.
The total free lime values calculated from the four mentioned parameters correlate well with
the values determined experimentally by the Holcim burnability test (r = 0.97), thus
confirming that the four chosen parameters are capable of characterizing to a considerable
extent raw meal burnability.
As can be easily observed (Figure 18), the quantity of free lime attributed to the individual
parameters varies considerably. Whereas in some mixes the free lime is primarily associated
with the chemical parameters (LS/SR), in others the influence of the quartz and calcite
grains is predominant.
For example in raw meal no. 3, despite a high overall residue of 24 % on a 90 µm screen, no
free lime results from the oversize calcite grains. The reason for this lies in the cement stone
composition of this component, in which the individual calcite crystals are fine -grained and
homogeneously distributed. The benefit of distinguishing between the overall fineness and
the actual dimensions of the individual crystals can thus be appreciated.
In contrast to raw meal no. 3, about half the free lime in raw meal no. 10 is due to oversize
calcite, originating from a high-grade limestone component of a titration value of 95 %.
In raw meals no. 46 and 53 the uncombined CaO can be mainly attributed to chemical rather
than physical parameters. In the former it is mainly the high silica ratio (3.5), in the latter the very
high lime saturation (103 % ) which is responsible for the resulting content of free lime. Finer
grinding of these raw meals would thus not appreciably lower the uncombined lime

values of clinker burnt in the burnability test. Only a change in chemical composition would
improve the burnability of these raw meals.
In raw meal no. 48, despite the large overall residue (43 % > 90 µm) and the given 6.5 %
quartz content, the influence of quartz is not unduly great. The reason for this is that the
individual crystals are small and well distributed, even within the individual grains.
It is now obvious that by individually examining the 4 parameters which principally influence
the burnability, it is easier to characterize, then was previously the case, the critical features
of any raw mixture.
It is thus, for instance, easier to decide whether components should be combined or separately
ground, or to decide the optimal chemical and physical properties of any raw mix.
23 THERMOCHEMISTRY OF CLINKER FORMATION

A detailed discussion of the heat balance of rotary kiln lies outside the scope of this paper
and will, in any case, be considered later under the heading of process technology. What
will, however, be mainly discussed are the basic thermochemical data relevant of the
material aspects of clinker production.
During clinker production, heat is both absorbed (endothermic heat changes) and produced
(exothermic heat changes in certain characteristic temperature rings).
On the basis of the reaction sequence already presented in Table 2, the heat changes
pertinent to the individual reactions are known to be as follows:
Table 10 Nature of heat changes occurring during clinkering

Temperature Type of Reaction Heat Change
range (°C)
20 - 100 Evaporation of free H2 O Endothermic
100 - 300 Loss of physically adsorbed H2O Endothermic
400 - 900 Removal of structural H2O (H2O, OH Endothermic
groups from clay minerals)
600 - 900 Dissociation of CO2 from carbonate Endothermic
> 800 Formation of intermediate products, Exothermic
belite, aluminate and ferrite
> 1250 Formation of liquid phase (aluminate Endothermic
and ferrite melt)
Formation of alite Exothermic
1300 - 1240 Crystallization of liquid phase into Exothermic
(cooling cycle) mainly aluminate and ferrite
A most effective way of visualizing the temperature at which the heat changes occur, their
nature and magnitude, can be obtained from differential thermal analysis (DTA) curves of
cement raw mixes, an example of which is given in Figure 19.

Figure 19 DTA curves of typical cement raw meals
It is obvious from these curves that by far the greatest heat requirement occurs between 850
5888 900 °C, i.e. for the decomposition of the carbonate minerals. The total heat
requirements for dehydration, decarbonisation and melting exceed the heat liberated by
the formation of belite and the intermediate and final products.
Not included in the DTA representation, however is the heat involved in the actual heating of
the raw meal and the cooling of the clinker and of the CO 2 and H2O gases obtained from
the calcite and the clay minerals.
The heat of absorption and liberation, however, varies according to the nature of the
individual raw mix, as is clearly demonstrated in Figure 13. The DTA curve, therefore,
allows a differentiation of the reactivities of different raw mixes to be made.
When the quantitative values for the various reactions are considered, the heat
requirements for the material only with respect to clinker formation may be expressed as
follows in approximate terms.

Table 11 Thermal balance of heat changes occurring during clinker formation
Endothermic processes: dehydration, decreasing,
heating
kJ/kg clinker (kcals/kg)
dehydration of clays ≈ 170 ≈ 40
decarbonisation of calcite ≈ 1990 ≈ 475
heat of melting ≈ 105 ≈ 25
heating of raw materials 0 - ≈ 2050 ≈ 490
1450 °C
≈ 4315 ≈ 1030
Exothermic processes: heat of formation, cooling
of clinker, CO2, H2O
kJ/kg clinker (kcals/kg)
crystallization of dehydrated clay ≈ 40 ≈ 10
heat of formation of clinker ≈ 420 ≈ 100
minerals
crystallization of melt ≈ 105 ≈ 25
cooling of clinker ≈ 1400 ≈ 335
cooling of CO2 (ex calcite) ≈ 500 ≈ 120
cooling of H2O (ex clays) ≈ 85 ≈ 20
≈ 2550 ≈ 610
Net theoretical heat of clinker ≈ + 1765 ≈ + 420
formation (endothermic -
exothermic)
Depending on the carbonate content, lime saturation and mineralogical composition, the net
heat requirement varies between approx. 1630 - 1800 kJ/kg clinker (i.e. 390 - 430 kcals /kg).
These values, of course, only apply to a dry raw meal. For a slurry, the heat requirements
will certainly be higher by that necessary of the removal of water. This value will be
dependent on the actual slurry water content, and for the average slurry of 35 % H2O is ≈
2100 kJ/kg clinker. The net theoretical heat of formation of 1 kg clinker from slurry is,
therefore ≈1765 + 2100, i.e. 3865 kJ/kg (920 kcals/kg).
When considering the heat required for actual clinker production, involving heat losses, the
values of approx. 1700 and 3865 kJ/kg clinker are too low. A more realistic heat balance
taking process inefficiency into consideration, is approximately as follows:

Table 12 Heat balance of wet and dry kiln
Dry kiln Wet kiln
kJ/kg clinker
Evaporation of H2O (0.4 % ≈ 20 (≈ 0.6 %) 2100 (38 %)
dry kin and 35 % wet kiln)
Heat of reaction ≈ 1765 (≈ 54 %) ≈ 1765 (≈ 32%)
Heat losses through gas, ≈ 840 (≈ 26 %) ≈ 1250 (≈ 23 %)
clinker, dust, etc.
Heat losses by radiation ≈ 650 (≈ 20 %) ≈ 360 (≈ 7 %)
and convection
≈ 3275 kJ/kg ≈ 5475 kJ/kg
The heat losses due to radiation, exhaust gas, etc. will be explained in more detail under the
heading of process technology.

Literature
F. Becker and W. Build-up of Rings caused by Spurrite Formation. Cement and
Schrämli: Lime Manufacture 42 (1969), pp. 91 - 94.
H. Braun, H. Burnability tests on cement raw mixes of the Holcim Group.
Guntlin and H. Report MA 77/2510/E
Kock:
J. Gebauer and E. Reaktionen zwischen Kalk und Kieselsäure bei hohen
Fitzer: Temperaturen. ZKG 22 (1969), pp. 402 - 406.
J.A. Imlach: Non-Isothermal Investigation of the Kinetics of Reactions
Occurring during Clinker Formation. Cement and Concrete
Research 6 (1976), pp 747 - 755.
H. Kock, G. Rey, F. A statistical Model for Determining the Burnability of Cement
Becker: Raw Mixes. VI. Int. Congr. Chem. Cem. Moscow (1974).
Report MA 74/ 3274/E.
Frankenberger: Wärmeübergang in Zementdrehöfen. Schriftenreihe der
Zementindustrie (1969), H. 36, Beton-Bauverlag GmbH,
Düsseldorf.
F.P. Glasser and Role of Iron and Aluminum Oxides as Fluxes during the
R.M. Herath Burning of Portland Cement. Cement and Concrete Research
Banda: 8 (1978), No 319 - 324.
F.M. Lea: The Chemistry of Cement and Concrete. 3rd Edition, Edward
Arnold (Publishers) Ltd., London 1970.
A.E. Moore: The Sequence of Compound Formation in Portland Cement
Rotary Kilns. Cement Technology 1976, pp. 85 - 91 and pp.
134 - 138.
P. Weber: Heat Transfer in Rotary Kilns. Bauverlag GmbH, Wiesbaden -
Berlin (1963)
H. Guntlin and M. Physikalisch-Chemisches Modell zur Charakterisierung der
Kristmann: Brennbarkeit von Rohmischungen.

Chapter 2
Portland Cement Clinker

Portland Cement Clinker
1. clinker mineralogy......................................................................................................... 46
1.1 Theoretical Clinker Phases.........................................................................................46
1.2 The Real Clinker Minerals...........................................................................................50
1.3 Polymorphic Modifications of the Clinker Minerals......................................................51
1.4 Quantitative Clinker Mineralogy - Potential and Reality..............................................55
2. Clinker Microstructure...................................................................................................62
2.1 Microscopical Appearance of the clinker minerals......................................................62
2.2 Formation of Clinker Microstructure During Burning................................................... 73
2.3 Influence of Production Conditions on Clinker Microstructure..................................... 76
3. Methods of clinker investigation...................................................................................81
3.1 Introduction................................................................................................................. 81
3.2 Microscopy and Microprobe Analysis..........................................................................82
3.3 X-ray Powder Diffraction.............................................................................................87
3.4 Clinker Analysis by Wet Chemical Extraction Method.................................................90

23 CLINKER MINERALOGY
1.1 Theoretical Clinker Phases

The chemical composition (wt.%) of ordinary Portland cement clinker in the Holcim Group is
on the average as follows:
Loss on ignition 0.48
SiO2 22.2
Al2O3 5.0
Fe2O3 2.8
CaO 65.4
MgO 1.8
SO3 0.54
K2O 0.63
Na2O 0.25
TiO2 0.27
Mn2O3 0.09
P2O5 0.14
- 0.01
Cl
-
F 0.08
99.70
CaOfree 1.3
With this knowledge it is possible to predict both:

5888 the phases which should theoretically be present in a cooled clinker of this
composition
5889 the quantity of each phase which should be present.
For the given average chemical composition the potential mineralogy is calculated to be as
follows:
C3S = 54.3 % C3A = 8.5 %
C2S = 22.8 % C4AF = 8.7 %
How was this calculation performed?

To predict the minerals which should potentially be present, the phase equilibrium diagram
for the appropriate oxide system was consulted. Such diagrams are most easily located in
the standard reference series ‘Phase Diagrams for Ceramists’ (Ref. 1).
To a first approximation the average clinker may be considered as consisting of only the
three major oxides CaO, SiO2 and Al 2O3 and hence its composition may be plotted using
the well-known ternary phase diagram, CaO - Al2O3 - SiO2. The given composition, i.e. 65.4
% CaO, 22.2 % SiO2, and 5.0 % Al2O3, when recalculated in terms of these three oxides
alone, yields 70.6 % CaO, 24.0 % SiO2, and 5.4 % Al2O3, totaling altogether 100 %.

Consulting the CaO - Al2O3 - SiO2 phase equilibrium diagram reproduced in a simplified
formation in Figure 1, exhibiting only the sub-solidous equilibrium relationships, it can be
demonstrated that the average composition lies within the C3S - C2S - C3A compatibility
triangle.
Because the chosen clinker composition lies within this triangle, these three phases will be
formed on heating, provided that equilibrium is attained.
Small changes in the chemical composition of the mix are accompanied by changes in the
quantity of the clinker phase C3S - C2S - C3A. As soon as the changes exceed a certain
limit, different minerals are then formed.
Thus, if the SiO2 composition is held constant at 24 % and the Al2O3 reduced to 2 % (i.e. point
1 in Figure 1), the phases formed would be C 3S - C3A- CaO, whereas increasing the Al2O3 to
12.5 % (point 2 in Figure 1) would result in the formation of C2S - C3A - C12A7.
It can thus be realized why maintaining the correct composition is important in the
production of Portland cement, and why it cannot be varied over too wide a range.
Because only the three phases C3S - C2S - C3A possess the desired hydraulic
properties, cement clinker composition should lie within the C3S - C2S - C3A triangle.
Fig. 1: Sub-solidous phase equilibrium relationships in the system CaO - Al2O3 - SiO2

Clinker containing CaO - C3S - C3A would contain undesired free lime (responsible for
expansion on hydration) while clinker containing C2S - C3A - C12A7, would contain no C3S,
and would therefore exhibit unfavorable hydration and hardening characteristics.
A more accurate prediction concerning the theoretical phases present is obtained when the
presence of Fe2O3 is taken into consideration. On recalculation in terms of the 4 principal
oxides, the composition of the so far considered average clinker becomes 68.5 % CaO,
23.2. % SiO2, 5.3 % Al2O3 and 3.0 % Fe2O3, again totaling 100 %.
As in the case for CaO - Al2O3 - SiO2, the above composition when plotted in the CaO -
Al2O3 - Fe2O3 - SiO2 phase equilibrium diagram (Figure 2), shows the potential
composition as lying within the C3S - C2S - C3A - C4AF compatibility volume and hence
consists of these phases.
Fig. 2: Sub-solidus compatibility relations in the system CaO - Al2O3 - Fe2O3- SiO2
A point of special interest here is that three of the phases C3S - C2S - C3A are binary
compounds, i.e. composed of only two oxides. The other phase C4AF is a ternary oxide,
composed of CaO, Al2O3 and Fe2O3. Although C4AF is often considered as being an
individual compound of fixed composition, it is in reality a member of the solid solution series
3+ 3+
2 CaO. Fe2O3 - ‘CaO . Al2O3 ‘ in which 50 % of the Fe atoms are replaced by Al atoms,
hence giving the composition 2 CaO ⋅ 1/2 Al2O3 ⋅ 1/2 Fe2O3, i.e. 4 CaO ⋅ Al2O3 ⋅ Fe2O3.

This solid solution series can be clearly seen in Figure 3, running from composition 2 CaO ⋅
1/2 Al2O3 ⋅ 1/2 Fe2O3 to the end member of approximate composition 6 CaO ⋅ 2 Al2O3 ⋅
3+ 3+
Fe2O3, i.e. C6A2F, in which 2 out of 3 Fe atoms have been replaced by Al . The theoretical
3+ 3+
end member with 100 % replacement of Fe by Al , i.e. ‘C2A’, is not stable at normal
atmospheric pressure, but only at pressures in the order of 15000 bars.
Figure 3 Phase equilibrium diagram of the portion CaO - 2CaO . Fe2O3 - CaO
⋅ Al2O3 of the CaO ⋅ Al2O3 ⋅ Fe2O3 system
In the already mentioned average clinker the oxides CaO, Al2O3, Fe2O3 and SiO2, total 95.6
%, the remaining 4.4 % being principally composed of MgO (1.8 %), the alkalis K2O and
Na2O (0.9 %) and SO3 (0.5 %). These minor oxides not only are located within the lattices
of the C3S, C2S, C3A and C4AF, (as will be demonstrated later) but also exist in the clinker
as free MgO (periclase) and as alkali and alkali/calcium sulfates.

1.2 The Real Clinker Minerals
We are now acquainted with the clinker phase from the theoretical point of view, but what
are the real chemical compositions of the clinker minerals?
Except for periclase (MgO) and free lime (CaO), all clinker minerals are solid solutions, i.e.
they occur as single crystalline phases which vary in composition within finite limits without
the appearance of an additional phase. The ‘pure’ compounds C3S, C2S, C3A, thus, do not
occur in this simple form, but contain a whole series of oxides such as MgO, Na2O, K2O,
Fe2O3, and Al2O3.
Whereas for the pure phases we use the abbreviations C3S, C2S, etc., for the real
minerals only the names alite, belite, aluminate and ferrite should be used.
Until the development of extremely sophisticated analytical techniques (electron microprobe
analyzer and energy disperse analyzer), the determination of the maximum quantities of
these minor oxides that could be incorporated into the 4 principal clinker minerals was based
on laboratory experiments using the pure theoretical phases C3S, C2S, C3A and C4AF. The
quantities actually present, i.e. the true distribution of the oxides, remained unknown. By
using the microprobe technique, however, it is now possible to determine the rage of
composition of even the individual alite, belite, aluminate and ferrite crystals in any factory
produced clinker.
In a recent study (2), based on 28 clinkers from Holcim Group plants, it was found by
electron microprobe analysis that the range of MgO, Fe2O3, Al 2O3, K2O and Na2O present
in the alite and belite crystals varies from plant to plant (see Table 1 and Figure 4).
Moreover, even within one particular clinker grain, the individual alite and belite crystals
show varying minor element incorporation.

Figure 4 Wt % chemical variation ( ± s) of MgO, Fe2O3, Al2O3, K2O, and Na2O in
alites and belites in various cement plants
Plant Plant
Table 1 Real clinker minerals: chemical variation between different plant

clinkers
Alite Pure Belite Pure Aluminate Pure Ferrite Pure
min. max. C3S min. max. C 2S min. max. C3A min. max. C4AF
X X X X
SiO2 24.4 25.9 24.5 26.31 29.8 32.9 31.3 34.88 3.1 4.3 3.9 1.8 3.1 2.5 20.98
Al2O3 1.0 1.8 1.3 - 1.5 2.8 2.2 - 28.7 31.8 30.8 37.74 20.2 24.4 22.5 32.86
Fe2O3 0.4 1.7 0.8 - 0.7 2.3 1.2 - 5.2 8.3 6.9 - 20.8 26.7 23.7 46.16
CaO 69.3 72.4 70.6 73.69 62.2 66.4 63.8 65.12 51.5 58.3 54.6 62.26 47.4 49.9 49.1
MgO 0.5 2.1 1.1 - 0.2 0.8 0.5 - 0.8 1.5 1.1 - 1.9 3.2 2.6
K 2O <0.1 0.3 0.1 - 0.3 1.9 0. - 0.3 4.6 2.0 - 0.1 0.3 0.2
Na2O <0.1 0.3 0.1 - 0.1 0.8 0.3 - 0.2 0.8 0.8 - 0.1 0.2 0.1
1.3 Polymorphic Modifications of the Clinker Minerals

If the clinker minerals are studied in detail, we find out that not only does the chemical
composition vary, but the same clinker phase also exists in different polymorphic forms,

which can exhibit different physical properties. Determining the various polymorphic forms
present in clinker has not been an easy task so that it was first necessary to study the
polymorphism of the pure phases C3S, C2S, C3A and C4AF in order to obtain sufficient
information and experience to be able to advance to alite, belite, aluminate and ferrite.
1.3.1 C3S and Alite
Between room temperature and 1100 °C pure C3S displays 6 polymorphic modifications,
namely:
♦ 3 triclinic TI, TII, TIII
♦ 2 monoclinic MI and MII
♦ 1 rhombohedral RI
From DTA analysis it is known that the transitions take place as follows:
600° 920° 980° 990° 1050°
TI ↔ TII ↔ TIII ↔ MI ↔ MII ↔ R
and because the changes involve only a light displacement within the lattice structure (i.e. no
breakage of the strong chemical bonds), they are therefore easily reversible. These
transformations consequently cannot be prevented, so that in pure C3S, the high
temperature modifications can only be stabilized by quenching.
When passing to alite , the real mineral in industrial clinker, incorporating Al2O3,,
Fe2O3, MgO, K2O, Na2O, and TiO2 within the crystal lattice, it has been found to exist
in the following modifications:
♦ trigonal R
23 monoclinic MI
24 triclinic TII
The incorporation of foreign atoms thus ensures the stabilization of the high temperature
monoclinic and trigonal forms, and in some clinker even both of these modifications can be
simultaneously present. Because the various modifications are crystallographically all very
similar, the various polymorphic forms of alite in cement clinker are considered as having
essentially the same hydraulic activity. Observed differences in activity seem to be related to
structural faults in the crystal as a result of incorporation of elements such as Al, Fe, Mg, etc.
A differentiation between the various polymorphic forms is partially possible by both
microscopy and by x-ray diffraction as shown in Figure 5 using sophisticated apparatus and
analyzing equipment.

Figure 5 X-ray diffraction patterns of the various C3S polymorphs

1.3.2 C2S and Belite
Pure C2S exhibits 5 crystalline modifications between room temperature and 1500°C,
namely the form α, β, α‘L, α‘H and α. The stability range of these is displayed in a simplified
form in Figure 6.
Figure 6 Stability ranges and polymorphic sequences of C2S
The changes β → α and α → α‘ L are irreversible and slow as they involve breaking all the
primary chemical bonds. These structural changes are accompanied by a considerable
volume change (13 %). The β-belite is considered to have good hydraulicity, while the α-form
is hydraulically only little active. Alkalis play an important role: in the presence of K 2O and
Na2O and α‘ and an α-form develop, which have a higher activity than the β-form.
1.3.3 C3A and Aluminate
Pure C3A is a definite compound with a cubic lattice, showing no polymorphic transformations.
When increasing quantities of Na2O, they are incorporated within the lattice, forming a solid
solution series incorporating up to 5.7 % Na2O, orthorhombic (O), monoclinic
5888 and tetragonal (Tg) modifications occur; their stability relations are summarized in
Figure
7.

Figure 7 Phase equilibrium diagram for Na2O - C3A solid solutions
The aluminate phase in clinker exists in the cubic (C), orthorhombic (O) and tetragonal (Tg)
forms. Two forms, e.g. O + C or O + Tg, can coexist. Up to date, the monoclinic (M)
formation has not been observed. There is some evidence to suggest that the various forms
differ in their reactivities.
1.3.4 C4AF and Ferrite
Pure C4AF is not a definite compound but a member of the continuous solid solution series
extending between 2 CaO ⋅ Fe2O3 and 2 CaO ⋅ Al2O3 , the latter end-member being stable
only at high pressure, up to approx. 66 % Fe2O3 in 2CaO ⋅ Fe2O3 can be replaced by Al2O3.
The member of 4 CaO ⋅ Al2O3 ⋅ Fe2O3 (C4AF) corresponds to the 50 % replacement.
The ferrite phase in Portland cement varies chemically between C6A2F and C6AF2 (solid
solution series) and is orthorhombic like the pure C2F. As yet no other modification has been
reported. The ferrite phase is considered to be only slightly hydraulically active.
For a detailed x-ray diffraction investigation of aluminate and ferrite in clinker, the phases
can be concentrated dissolution of alite and belite using salicylic acid.
1.4 Quantitative Clinker Mineralogy - Potential and Reality
1.4.1 Calculation of Potential Clinker Composition

For almost 50 years the potential clinker mineralogy has been calculated by the procedure
first proposed by Bogue. This method assumes that:
23 chemical equilibrium has been attained

5888 Fe2O3 reacts with Al2O3 and CaO to form C4AF
5889 Al2O3 remaining from combination as C4AF reacts with CaO to form C3A
5890 SO3 combines with CaO to formation CaSO4
5891 CaO remaining from the above combinations reacts with the silica. C3S is the first
formed
5892 with the SiO2, and any CaO left over reacts with the C2S to form C3S. If CaO
remains after converting all of the C2S to C3S it remains as uncombined CaO.
5893 MgO remains essentially uncombined.
Figure 7A Mineralogical composition derived from chemical analysis (Graphical

representation of the steps in calculation)
For the purposes of calculation, the ‘CaO’ and ‘SiO2’ values used are given.as
‘CaO’ = CaOtotal - CaOfree
‘SiO2’ = SiO2 total - insoluble residue
The calculation does not take into account the influence of the minor elements:
alkalis, Mg, Ti, Mn, Cr, P, etc., and the solid solution of the principal oxides: CaO, SiO2,
Al2O3 and Fe2O3 within the alite, belite, aluminate and ferrite phases. As we have already seen,
the composition of the 4 principal phases is not C3S, C2S, C3A and C4AF. It is therefore not
surprising that the potential and real clinker compositions never agree. This situation is
further aggravated by the non-attainment of equilibrium during clinker production. In addition it is
now known that SO3 in clinker is mainly combined with the alkalis
and only to a slight extent with CaO. The currently used Bogue formula for clinker does,
therefore, not consider the influence of SO3.
1.4.2 Differences between Potential and Real Phase Composition

The most practiced method at present to determine the real mineralogical composition is by
means of microscopic point counting, complemented by the use of crystallized aluminate
and ferrite. The results obtained by these methods, as is to be expected, differ from those
obtained during the Bogue calculation procedure.

How large are the differences between the potential and real mineralogical composition of
the clinker?
This can best be seen in Figure 8 based on the results of measurements on over 50 different
plant clinkers (3).
It can be seen that:
23 Alites as calculated are always too low, ranging from 3 - 24 %, on average 13 %
24 Belites are mostly calculated too high, on average 6 %
25 Aluminate and ferrite as calculated can either be too high or too low, but in general both
phases are calculated approx. 2 % too high.

Figure 8 Difference between Actual and Potential Phase Composition
(experimental minus Bogue)

1.4.3 Modifications of Bogue’s Formula
A useful improvement in the correlation between potential and real mineralogical
composition can be obtained if slight corrections are made to the Bogue calculated values.
On the basis of data obtained (Figure 8), the following formulae have been proposed:
C3Scorr. = C3SBogue + 4.0 MgOclinker+ 5.5 K2Oclinker
C2Scorr. = C2 SBogue- 1.5 MgOclinker- 2.2 K2Oclinker
C3Acorr. = C3ABogue+ 7.8 Na2O + 1.5 AR - 2.1 SO3 - 5
C4AFcorr. = C4AFBog - 6.5 Na2O - 1-7 AR + 5.0 Mn2O3 + 3
These corrected formulae are valid under the following conditions:

MgO <4% Na2O < 0.5 % Mn2O3 < 0.3 %
K2O <1% SO3 < 1.0 % AR> 1.0
In the case of AR < 1.0 uncorrected Bogue’s formulae are more precise to calculate C3A and
C4AF.
The improvement achieved by using these modified formulae can be seen in Figures 9 and
5888 These corrections improve Bogue’s original concept in that the clinker phases
can be more accurately calculated, taking into consideration the incorporation of minor
elements without having to undertake time consuming mineralogical analyses.
For pure routine quality control purposes in any individual plant, the use of the original
Bogue equations is still acceptable as long as no significant change in raw materials occurs.
For comparison of calculated values between different plants or with directly measured
clinker mineralogy (microscopy, x-ray diffraction) the use of the modified formulae is
recommended.

Figure 9 Alite and Belite

Figure 10 Aluminate and Ferrite
23 CLINKER MICROSTRUCTURE
2.1 Microscopical Appearance of the clinker minerals
2.1.1 Introduction
In the foregoing chapter we have discussed the principles of clinker formation, its chemistry
and mineralogy. These parameters can only account for part of the clinker and of the
resulting cement properties. The microstructure or fabric of a clinker is also decisive when
considering phenomena such as clinker grinding, cement hydration and hardening
characteristics.
In this chapter only an introduction to the microscopical appearance of industrial clinker will
be given. As this is a complex topic, only the basic principles will be discussed. For further
information reference should be made to any of the reference works on this subject (4, 5).
2.1.2 General
The appearance of a crystalline aggregate is generally referred to as fabric, which is defined
as the whole of all geometric features or properties of an aggregate. Three groups of
geometric features can be discerned: phase association, texture and structure.

Phase association is the information group which defines what phases constitute the fabric
and in what quantities they exist. Texture includes the information concerning size, shape
and arrangement of the crystals and also the intergrowth relationship of the individual
phases. Finally, structure indicates to what extent a fabric is anisotropic, whether some
crystal phases are oriented in a certain direction and what type of space-filling, i.e. porosity,
occurs.
With these geometric features, we can describe the fabric of a material. This description
explains how and out of what a piece of clinker is build up. When completely differently
manufactured clinkers are investigated and compared, one realizes to what degree the
‘what’ and ‘how’ differ, thereby resulting in the most varied clinker fabrics.
The fact that clinker does not always have the same appearance leads to the important
question of why one particular fabric and not another is formed. It is overall accepted that the
genesis of a crystalline material influences to a great extent its appearance or fabric. The
genetic factors of the clinker formation process are identical to those conditions on which the
industrial manufacture of clinker is based, this being the interaction of parameters beginning
with the raw material, passing through the sintering process and ending with clinker storage.
2.1.3 Microscopical Appearance of Clinker Phases
2.1.3.1 Alite
External Form
Alite, the principal constituent of cement clinker, occurs most frequently as idiomorphic
crystals, i.e. as well-defined, sharp-edged crystals. The characteristic form, as can be seen
in Plate 1, are compact tabular crystals, often exhibiting a hexagonal section. Alite crystals
can also, however, be found in a form whose outlines are dissected by indentations and
protrusions, and whose hexagonal shape is barely discernible. Such crystals are termed
xenomorphic alite and can be seen in Plate 2. Lath-shaped alite crystals have also been
reported.
Plate 1 Idiomorphic alite (brown). 250 x

Plate 2 Xenomorphic alite. 250 x

Internal Features
A feature commonly observed within alite crystals is the inclusion of small rounded belite
crystals, these having been formed before the alite, as has previously been indicated in
Paper 10. An example of this phenomenon can be seen in Plate 3.
Plate 3 Alite exhibiting belite (blue) inclusions. 250 x
Other features can also be observed within alite crystals and include the presence of single
and polysynthetic twinning, as can be seen in the excellent plates in Ref. 4 (p. 11, Fig. 4).
At the highest sintering temperature, alite always exhibits its highest symmetry, i.e. trigonal
23 form, but on cooling to normal temperature, passes through several polymorphic stages,
and so in industrial clinker appears mostly in the monoclinic (M) an triclinic (T) forms.The
presence of simple twinning is indicative of crystals having passed through the
trigonal (R) → monoclinic (M) polymorphic transition. The presence of polysynthetic
twinning is associated with the change from the monoclinic to the lowest symmetry,
i.e. triclinic (T) form. Differentiation between the individual monoclinic or triclinic forms is
not possible by microscopy but requires the use of x-ray diffraction (see Figure 5) or DTA
techniques. The optical distinction of the various polymorphic changes is made possible by
the different etching behavior of the various crystal modifications.
Another internal feature of alite crystals is the presence of zones having different etching
properties, i.e. a zonar structure. These zones are associated with the selective inclusion of
different ions during the formation process. Examples of the various forms of twinning and
zonar structure are to be found in (4).

Alite decomposition
As has already been mentioned, pure C3S is unstable below 1250 °C and tends do
decompose into belite and sub-microscopic free lime. Although this decomposition is
strongly suppressed in alite due to the incorporation of foreign ions, the presence of
2+
bivalent iron i.e. Fe and CaF2 opposes this stabilization effect. An example of partially
decomposed alite crystals can be seen in Plate 4.
Plate 4 Alite decomposition into belite and free lime. 250x
Scanning electron microscope (SEM) photographs of alite, (see Plates 21 - 23)

also indicate the idiomorphic formation shown in Plate 1.

2.1.3.2 Belite
External Form
The round crystal form (Plate 5) is the most characteristic shape of belite, although in rare
cases crystals showing a tendency to a hexagonal form are also observed (Plate 6).
Plate 5 Typical rounded belite 250x
Plate 6 Belite with complex lamellar structure due to α → α‘ transition. 640x

Internal Features
Just like in the case of alite, the passing through the α → α‘, α‘ → β and β → α polymorphic
transformations can also be microscopically observed. The first mentioned transition realties
in belite crystals exhibiting a cross lamella structure as shown in Plate 6, while after
undergoing the α → β transition, polysynthetic twinning occurs.
Cross lamella and polysynthetic twinning can sometimes be observed in the same crystal.
The transition from β → α belt, accompanied by an appreciable change of volume can be
observed microscopically.
23 - belite is undesirable in clinker because the strong volume expansion breaks up
the clinker granules, producing much dust, and also because it is not hydraulically
active.
Excellent plates showing polymorphic transformation phenomena are to be found in (4).
Secondary Belite
The liquid in contact with alite and belite at clinking temperature contains approximately 6 %
SiO2 as well as CaO, Al2O3 and Fe2O3. During cooling, immediately after the maximum
temperature, this SiO2 comes out of solution in the form of small secondary belite crystals
as sown in Plate 7. Such belite can exhibit a drop-like or dendritic structure.
Plate 7 Secondary crystallization of belite 250 x
Scanning electron microscope photographs of belite seen in Plate 22 also confirm the
rounded shape of belite observed in polished sections.
2.1.3.3 Uncombined Lime
This is more commonly known as free lime, and always appears as rounded spherical
crystals (Plate 8), even when it is not constricted by other crystalline phases. When viewed
at higher magnifications, e.g. 10,000 through scanning electron microscopy as shown in
Plate 24, it can be seen that the rounded free lime grains do have planar surfaces. It is
invariably found along with alite, but not in contact with belite, as in the latter case the two
would react to form alite.

Plate 8 Spherical free lime (off-white) 250 x
Uncombined lime should not be confused with the micro-crystalline secondary free
lime, resulting form the decomposition of alite.

2.1.3.4 Matrix
The principal clinker minerals alite and belite are surrounded at clinking temperature by a
liquid phase (matrix). Upon cooling the latter crystallizes to yield principally aluminate and
ferrite but, depending upon its composition, can also yield periclase and sulfate containing
minerals.
2.1.3.5 Aluminate
The form of the aluminate phase depends both upon the cooling rate and the alkali content
of the liquid phase. When the cooling rate is slow, well- separated idiomorphic aluminate
crystals can be observed (and consequently also ferrite crystals), as can be seen in Plate 9.
This type of aluminate is the normal or cubic form, containing insufficient incorporated
foreign ions (alkalis) to change its symmetry to one of the lower symmetry types.
Plate 9 Differentiated aluminate (gray) and ferrite (white) caused by slow

cooling. 640 x
When the liquid phase is rich on alkalis, an aluminate solid solution containing both K2O and
Na2O crystallizes in the elongated lath from shown in Plate 10, often termed ‘alkali
aluminate’.

Plate 10 Alkali aluminate with lamellar habit. 570 x
This form, as has been reported earlier, may have orthorhombic (O) or tetragonal (Tg)
crystal symmetry, requiring x-ray diffraction to differentiate between the two modifications.
However, when the clinker has been rapidly cooled, the aluminate and ferrite are finely
grained and closely intergrown as shown in Plate 11.
Plate 11 Finely grained aluminate and ferrite due to rapid cooling. 250 x
Although the individual crystals are often so fine that they cannot be optically distinguished,
x-ray diffraction confirms that they are present in a crystalline and not amorphous form.
2.1.3.6 Ferrite
Just like the aluminate, the external form of the ferrite, as can be seen in Plate 9 and 11,
depends upon the cooling rate. The composition of this phase varies, due to solid solution,
around the ratio C4AF. The actual value cannot be determined by microscopy but x-ray
diffraction gives an indication of their composition.

2.1.3.7 Periclase
Periclase occurs only in clinker rich in MgO as an independent phase. In general two types
can be distinguished, one idiomorphic with polygonal sections, and one dentritic. Periclase is
very easily identified under the microscope due to its resistance to grinding and etching.
Plate 12 Idiomorphic periclase (gray). 640 x
2.1.3.8 Sulfate containing phases

The last phases to crystallize, long after the aluminate and ferrite have solidified, are the
alkali sulfates. These have been identified as being of two separate types, namely a simple
alkali sulfate of composition K2SO4, and a calcium potassium sulfate of composition 2
CaSO4, both being able to incorporate small amounts of sodium. They are often found
around the edges of pores. Micrographs showing these two types of sulfates can be seen in
Plates 13 and 14.
Plate 13 Alkali sulfate (black, angular) 250 x

Plate 14 Calcium potassium sulfate (needle shaped). 250 x
2.2 Formation of Clinker Microstructure During Burning

As clinker formation takes place over a range of temperature, it is not surprising that a
gradual change of fabric occurs between the initial raw meal and the finally cooled clinker.
The most common way of ascertaining such changes is to microscopically examine polished
sections prepared from raw materials which have been heated to selected temperatures and
then immediately quenched, to retain their high-temperature microstructure. Suitable
samples may be either obtained via sampling parts located at various positions along a kiln
shell, or by heating raw mixes in the laboratory. Now following, the change of microstructure
of samples obtained at various temperatures during clinkering in a laboratory DTA
investigation are summarized.
As can be deduced from Figure 11 and from other sources (6), the clinker on attaining the
various indicated temperatures, shows the following changes in microstructure:

Figure 11 Changes in Microstructure at various temperatures
Upheating cycle:
1160°C
Individual particles of decarbonized limestone (i.e. free lime) and unreacted hematite can be
seen against the sub-microscopic, non- differentiated mass. Quartz particles are
surrounded by reaction rims, several microns thick. The presence of free lime and belite
can be demonstrated by x-ray diffraction. No large pores can be observed.

1280°C
The reaction rim around quartz particles has now become thicker and developed into
rounded belite crystals, large enough to be optically identified. The outline of large
decarbonized limestone grains can still be recognized. The presence of large pores is now
observed.
1295°C
Quartz grains have mostly disappeared and in their place are now grains of belite. While
new belite crystals are formed, the already existing ones at the same time increase in size.
Small pockets of liquid phase can already be observed. The structure is now decidedly
porous.
1330°C
The presence of liquid phase is very evident at this temperature, as is the presence of
optically identifiable alite and free lime crystals.
1410°C
Alite and belite crystals are still growing and showing their typical morphological forms in
liquid rich zones. Large pores can still be observed.
1450°C
Alite and belite have further increased in size but have still not reached their final
dimensions. Free lime can still be observed in many areas of the clinker in its typical
rounded form.
Cooling Cycle:
1290°C
Alite and belite have reached their final size while porosity has generally decreased. The
liquid phase is still undifferentiated.
1260°C
Crystallization of the ferrite phase from the melt is detectable (see DTA curve).
1200°C
The clinker has its final microstructure. Alite and belite can now be observed in a matrix
consisting of crystallized aluminate and ferrite of sufficient size so as to allow their
identification. Alite often appears to have been resorbed (xenomorphic crystals), while on
account of slow cooling belite often shows finger-like growths. Small areas of exsolved
material can sometimes be seen within aluminate areas, i.e. secondary crystallization of
belite has occurred.

2.3 Influence of Production Conditions on Clinker Microstructure
2.3.1 General
The microstructure of industrially produced clinker is influenced by several process
parameters, the most important of which are given in Table 2.
Table 2 Industrial parameters influencing clinker texture

Influence Parameter Specific factors
Raw mix properties chemical composition
fineness and Mineralogy
homogeneity
Fuel type coal ash
Burning conditions burning time / temperature
kiln atmosphere
Cooling conditions rate of cooling
2.3.2 Chemical Composition

If burning is properly performed, the quantitative mineralogical composition of a clinker
depends only on the chemical characteristics of the raw feed. From theoretical
considerations (phase diagrams and Bogue calculation) and microscopical observations, the
change of clinker mineralogy with varying clinker compositions can be predicted.
Although it may seem trivial, it is considered useful to give the clinker mineralogy associated
with clinker chemical composition (Table 3).
Table 3 Effect of chemical composition on clinker mineralogy

Chemical composition Clinker mineralogy
lime saturation too low alite little
belite abundant
free lime traces
lime saturation too high alite abundant
belite little
free lime excessive
alumina ratio high aluminate > ferrite
alumina ratio low ferrite > aluminate
silica ratio high little liquid phase
silica ratio low liquid phase abundant
alkalis alkali - aluminate present
alkalis together with SO3 alkali - sulfates present
more Mg than can be incorporated in periclase present
the lattice of principal phases

2.3.3 Fineness and Mineralogy
During the clinkering of a raw mix, a migration and exchange of material takes place. The
smaller the grain size, the shorter the transport distance required to achieve complete
reaction.
Once a certain grain size is exceeded, the lime and quartz components can no longer react
completely within the allowed time. Critical particle size for minerals such as calcite and
quartz lie in the range of approx. 50 - 100 µm. The maximum grain size is, therefore, that at
which the raw material particles are still just able to react. When this limit is exceeded,
‘undigested’ relicts remain. If the clusters of free lime and belite are examined, particularly
those with sharp boundaries (Plates 15 and 16), it is evident that they owe their formation to
the presence of oversize raw material grains. They still show the original appearance of a
raw meal grain, the siliceous particles being present as belite clusters and the carbonate
particles as free lime clusters as a result of incomplete reaction.
Plate 15 Belite cluster due to large, SiO2 - rich grain. 220 x
Plate 16 Free lime cluster due to large, lime-rich grain. 64 x

2.3.4 Homogeneity
If the raw mix is insufficiently homogenous, volume units of varying size will exist which do
not have the overall bulk chemical composition. It is easy to deduce form the phase diagram
for the system CaO - Al2O3 - Fe2O3 - SiO2 the phase compositions which can coexist
assuming different volumes to have difference chemical composition. In Figure 12 the
different phase assemblages in the system CaO - Al2O3 - SiO2 can be seen.
Figure 12 Possible mineral assemblages in the system CaO - Al2O3 - SiO2
With insufficiently homogenous raw feeds we should therefore expect a clinker

microstructure of varying mineralogy.
In reality this model exists also in the clinker. Plate 17 shows a clinker with a very
inhomogenous distribution of alite, belite and free lime, reflecting directly raw meal
inhomogeneity.
Plate 17 Inhomogeneous distribution of clinker minerals. 64x
In comparison with the inhomogenous microstructure produced from an inhomogenous

raw mix, Plate 18 sows the clinker fabric when the raw mix is correctly homogenized.

Plate 18 Homogenous distribution of minerals. 250 x
In comparison with the effects caused by oversize particles, the concentrations areas due to
inhomogeneity contain normally more than one clinker component and usually extend over
lager areas. Experience shows, however, that a certain clustering, particularly of belite, is
almost always observed in plate clinker. The even distribution (Plate 18) is rather an
exception.
2.3.5 Coal Ash
Rotary kilns are in ever increasing numbers again being fired using coal.
Unlike gas and oil, this type of fuel imposes traces of its presence on clinker textures. During
the burning of coal, siliceous ash is produced as molten ash drops and these fall as a
sediment of the clinker being formed. There is thus a SiO2 concentration on the surface of
the kiln charge. Depending on how far the ash drops are carried into the kiln, they will be
more or less assimilated with the clinker granules. Due to the deposition of ash drops rich in
SiO2 on the kiln charge, there is a local lack of CaO. At these places only belite can be
formed and thus the deposition leads to typical belite streaks in the clinker as can be seen in
Plate 19. Generally, these streaks are found in the peripheral areas of the pieces of clinker.
Plate 19 Belite layer due to coal ash. 64 x

2.3.6 Burning Time / Temperature
Large-scale industrial process conditions do not allow thermodynamic equilibrium to be
reached. If the kiln charge passes through the kiln too quickly (flushing) or the sintering
temperature is too low (underburning), the same characteristic fabric results as shown in
Plate 20.
Plate 20 Underburnt clinker. 64 x
Whether the temperature, the heating time of both parameters were too low cannot be
distinguished. As can be seen in Plate 20, the alite an belite crystals are unusually small.
Alite is surrounded by large amounts of liquid phase, while belite is concentrated in clusters.
In the areas where alite and liquid phase coexist, accompanying free lime appears as a
result of uncompleted reaction between lime and silica. High porosity with vesicular cavities
in places is also typical.
If all these characteristics are considered, it may be concluded that here the formation of the
clinker phases stopped at an early stage. The time and / or temperature was not sufficient to
form normal clinker from the raw materials.
Another time - temperature anomaly occurs if the clinker remains in the highest temperature
range for too long or if the sintering temperature is too high. Coarsening of the alite and
belite grains takes place. It can be shown that the crystallization of calcium silicate also
continues during cooling as long as the cooling rate is sufficiently slow. Alite and belite can
continue to grow under such conditions - provided that the quantity of the reacting material is
large enough - until the liquid phase crystallizes. Therefore, with an extreme crystal growth
of calcium silicate, it is not clearly distinguishable if the clinker was over burnt or whether the
cooling rate was too slow. In all these events, hard burning results in clinker of reduced
hydraulic activity.

2.3.7 Cooling Rate
Clinker cooling takes place in two stages, the first cooling stage occurring within the
kiln, the second in the clinker cooler.
The rate of cooling within the kiln depends upon the flame length, on its position in the kiln
and on the throughput and speed of the kiln charge. The clinker temperature at the kiln
outlet is probably between 1350 and 1200 °C. If the flame is long, this part of the cooling
process is very slow and alite and belite can grow to an excessive crystal size. In some
cases (when the cooler efficiency is low), alite partially decomposes into belite and free lime
(see Plate 4).
The texture of the solidified liquid phase is also very dependent on the cooling rate. During
slow cooling, the crystals have time to grow. Ferrite and aluminate form a coarsely grained
matrix (Plate 9). Alternatively, if the cooling process proceeds quickly, the opposite takes
place, i.e. the crystals are finely grained (Plate 11).
Cooling can also proceed so quickly that the crystals can only form in the sub-microscopic
range. Distinction between aluminate and ferrite is no longer possible by microscopy but can
be effected by x-ray methods.
23 METHODS OF CLINKER INVESTIGATION
3.1 Introduction
Investigations on clinker are usually restricted to determining the following properties:
5888 identification of the minerals
5889 estimation of the quantity of each species present
5890 determination of the composition of individual crystals of any single species
and hence the composition range of that species
5891 determination of the habit and spatial distribution (texture) of the minerals.
Several groups of methods are available for the determination of these properties:
23 light microscopy, scanning electron and electron microprobe method
24 x-ray diffraction
25 thermo-analytical methods
26 wet-chemical extraction techniques
Often it is necessary to use the combination of these methods if the maximum amount of
information or if the highest degree of accuracy is required. No single technique is capable
of determining all the desired properties. In Table 4 the normal range of application of each
method is listed.
As can be seen, wet chemical methods are certainly capable of being performed at all
plants, whereas only a limited number of plants are equipped to perform x-ray diffraction and
reflected light microscopy. These latter methods are, however, certainly available in the
central laboratory of a group of plants. Investigations using scanning electron microscopy,
and an electron microprobe analyzer can only be undertaken by specialized laboratories
possessing such advanced and expensive instruments.

Table 4 Normal application range of the various methods of clinker
investigations
Investigation Clinker properties Equipment
Method
Mineral identification Mineral Habit / Found in Cost of Available in
composi- texture cement plant equipment TS
tion laboratories
Qualitative Quantitative
Microscopy
Reflected light + + o + sometimes moderate yes
Scanning electron o o (+) + definitely not very high no external
Electron + o + o definitely not very high equipment
microprobe used
Diffraction
X-ray + aluminate / ferrite o in a few cases high* yes
ferrite free solid
CaO periclase solution
Wet chemical
extraction
Salicylic acid alite + belite o o very low yes
extraction No special
Glycol extraction free CaO o o equipment very low yes
Controlled water soluble alkali o o required very low yes
extraction sulfates
* except when purchased as an addition for an existing x-ray fluorescence spectrometer

3.2 Microscopy and Microprobe Analysis
Cement clinker is, almost without exception, investigated by reflected light microscopy on
etched, polished sections. The use of transmitted light microscopy either on thin sections or
on powders is rarely practiced.
Great care must be exercised when selecting representative samples, the quantity required
depending on the topic under investigation. When information on minerals content, porosity,
etc. , is required, the chosen material should be sampled from not only one but from several
clinker lumps.
The selected sample is first embedded into a suitable resin (e.g. epoxy resin) to facilitate
handling and then treated as follows:
5888 ground to produce a flat, matt surface (e.g. using carborundum powder)
5889 polished to obtain a flat, shining surface (e.g. using 0.3 µm diamond paste)
5890 etched to differentiate between the individual clinker minerals (e.g. using
hydrofluoric or acetic acid vapors, or using nitric acid solution).
By using non aqueous grinding and polishing media, the loss of the water soluble alkali
sulfates and free lime during sample preparation can be prevented.

3.2.1.1 Qualitative mineral analysis
Using the thus prepared polished section it is possible by using magnifications of up to 500 x
to determine qualitatively in clinker:
23 the minerals present and their habit
24 their polymorphic modifications (sometimes)
25 clinker fabric (porosity, homogeneity, etc.)
and from this information deduce the influence of such process conditions as:
23 fineness of raw mix grinding
24 homogeneity of raw feed
25 time / temperature of burning
26 rate of cooling
27 type of fuel used
The use of this method has already been demonstrated in Chapters 2.1, 2.2, 2.3 (see also
(4), (5).
3.2.1.2 Quantitative mineral analysis
Clinker properties are quantitatively determined by the use of one or two methods, namely:
5888 Linear point counting i.e. the sample is moved in regular steps and the
mineral under the cross-wire is identified and counted.
5889 Area point counting i.e. a point grind containing 100 points, placed in the
microscope ocular, lies apparently over the clinker, and the mineral at each point is
counted without moving the clinker. Only after all the points have been counted is the
clinker moved to another field of view.
Linear point counting Area point counting
Both methods determine the area percent of the minerals in the polished section, and to
3
convert the results into weight percent, the following mineral densities (g/cm ) must be taken
into account:
alite 3.20 belite 3.28 aluminate 3.04
ferrite 3.77 periclase 3.58 free lime 3.30

To obtain a high accuracy, the TS/MA counts 4000 points (N), giving an absolute statistical
counting error (δ) at the 95 % confidence level of
δ =1.96 P(100− P)
N
P = % points for the counted minerals
The most usual features to be quantitatively determined include:
0 clinker mineralogy
1 size distribution of minerals (is related to hydration reactivity)
2 homogeneity
3 porosity
Whereas the Bogue formula calculates the potential quantities of minerals present,
microscopical analysis determines the quantities actually present . The difference between
both methods has already been demonstrated in Chapter 1.4.2 for alite and belite.
3.2.2 Scanning Electron Microscope Analysis

By using a fine electron beam in place of light in such a way that the beam scans the sample
surface in a raster pattern, it is possible to obtain an optical image of the surface on a TV
screen. This image can have a magnification of up to 10,000 x, combined with a good depth
of focus. By further increasing the magnification (up to 100,000 x) the quality of the image is
reduced.
Using the energy disperse x-ray spectrometry technique (EDAX), it is now also possible to
determine semi-quantitatively the chemical composition of the observed minerals.
Examples of clinker investigated by scanning electron microscopy are displayed in Plates 21
- 24.
Plate 21 Idiomorphic alite (A) embedded in sulfate-rich matrix (S). 2000 x

Plate 22 Alite (A) and belite (B) in sulfate-rich matrix (S). 2000 x

Plate 23 Alite (A) in aluminate / ferrite matrix (M). 2000 x
Plate 24 Free lime (L) in contact with alite (A). 9500 x

3.3 X-ray Powder Diffraction
3.3.1 Methodical
The examination of clinker by x-ray powder diffraction can with much experience be used to
determine:
0 the minerals present - quantitative and qualitative
1 their polymorphic modification (s)
2 the approximate composition of certain minerals
The full potential use of the method as a basic routine technique, however, is made difficult
by the following:
0 variations in the chemical composition of almost all minerals
1 more than one modification of the same mineral almost always present
2 the x-ray intensities of the various modifications are not equal
3 overlapping of various reflections and small concentrations require pre-preparation of the
sample by wet chemical extraction
4 the relatively high cost of the equipment. This can be somewhat reduced when x-ray
diffraction facilities are combined with existing x-ray fluorescence equipment.
3.3.2 Experimental Technique

Samples to be investigated must be finer than 20 µm when determinations are desired. The
prepared sample is then irradiated in a focused x-ray beam (e.g. Cu or Cr radiation) and the
characteristic peaks of each mineral recorded.
For qualitative analysis the 2 θ values of the peaks are compared with the characteristic
peaks of the known clinker minerals and in this way the following can be determined:
0 the presence and polymorphic modifications of alite, belite
1 the presence and solid solutions of aluminate and ferrite

2 the presence of MgO, CaOfree
For quantitative analysis, suitable strong reflections of the required minerals, free of
interference from other mineral peaks, are required. As measurement criterion either the
effective peak height or integrated peak area may be applied, this being obtained by
subtracting the background radiation level form the overall peak height or peak area.
Peak area method

Peak height method
Two types of calibration can be applied in which the measured effective peak area of height
is compared to that of either an internal or external standard (e.g. SiO2, CaF 2, TiO2).
The external standard method, i.e. standard and sample not mixed together, eliminates
errors due to fluctuations in the equipment itself. (e.g. instability of x-ray generator, x-ray
counter, etc.), but makes no allowance for changes in material parameters (e.g. absorption
by matrix) . This latter effect can be compensated for by employing an internal standard,
i.e. by mixing the standard powder into the powder being investigated.
3.3.3 Applications of x-ray diffraction
Determination of ferrite solid solution
As mentioned previously, the ferrite composition is only approximately C4AF. The true solid
solution present in clinker can, however, be determined using x-ray diffraction, employing the
change in the 2 θ value of characteristic peaks as evaluation criterion. The latter are shown
in Figure 13.
Figure 13 Change of peak reflections (002, 200, 141) for ferrites in the C2F, C6A2F
series.

A useful application of knowing the actual ferrite composition is, that it can be used to obtain
a better Bogue formula estimation of aluminate and ferrite than the one obtained by
assuming the composition to be C4AF.

Determination of free lime
By extending x-ray fluorescence facilities to include x-ray diffraction (at a fraction of the cost
for x-ray diffraction alone), it is possible to quickly and accurately measure free lime as a
process control method, avoiding the use of glycol extraction. The main advantage of the x -
ray method is the reduced time required (10 minutes) using the same sample prepared for x
- ray fluorescence. Free lime values in the range 0.5 - 5 % can be determined at an
accuracy of ± 0.3 % CaOfree at the 95 % confidence level.
For the determination of other, more complicated clinker analyses (e.g. quantitative
determination of the clinker minerals), it is highly recommended that contact be first
established with the TS/MA or other suitable qualified laboratories.
3.4 Clinker Analysis by Wet Chemical Extraction Method

When neither microscope nor x-ray facilities are available, certain minerals or groups of
minerals can still be quantitatively determined by selective extraction using wet chemical
methods.
Minerals which can be quantitatively determined to an acceptable degree of accuracy by
such inexpensive but rather limited techniques include:
Alite + belite + free CaO: using ethanolic salicylic acid
Free lime and Ca(OH)2: using ethylene glycol extraction. The proportion of free lime alone
can be ascertained by subtracting the Ca(OH)2 value determined via its H2O loss on
thermal decomposition
Alkali and alkali/calcium sulfates: by controlled extraction with H2O and determination
of Na2O, K2O and SO3 levels in the extracted solution

Literature
23 American Ceramic Society: Phase Diagrams for Ceramists. The American Ceramic
Society, Columbus, Ohio, USA. 1975.
24 M. Kristmann: Portlandzementklinker. Mineralogische und chemische
Untersuchungen.’Holderbank’ Management und Consulting Ltd., 1977 (in German).
Part I: Cement and Concrete Research 7, 649, 1977.
Part II: Cement and Concrete Research 8, 93, 1978.
23 M. Kristmann: Clinker Mineralogy - Actual and Potential Phase Composition. MA Report
77/2511/II/E.
24 F. Hofmänner: Microstructure of Portland cement clinker. ‘Holderbank’ Management und
Consulting Ltd., 1975 ( in English and German).
25 F. Gille et al.: Mikroskopie des Zementklinkers. Bilderatlas. Beton-Verlag GmbH,
Düsseldorf, 1965 (in German).
26 J.A. Imlach and F. Hofmänner: Investigation of Clinker Formation by DTA and Optical
Microscopy. MA Report 74/2373/E.
Additional Literature:
J.D. Dorn: Mictroscopic Methods for Burnability Improvement. Cement and Concrete
Research 8, pp 635 - 646, 1978.
D.H. Campell, Micrascafic Examination and Interpretation of Portland Cement and Clinker,
Construction, Technology Laboratories, Skokie, Illinois, USA.

Chapter 3
Refractory Materials and Coating Formation

Refractory Materials and Coating Formation
1. REFRACTORY MATERIALS AND THEIR PROPERTIES............................................... 94

1.1 General Introduction...................................................................................................94
1.2 Classification of Refractory Products..........................................................................96
1.3 Manufacture of Refractory Products.........................................................................102
1.4 Testing of Unused Refractories.................................................................................108
1.5 ZONING of Cement Kilns..........................................................................................116
1.6 Wear of Kiln Refractories..........................................................................................118
1.7 Chemical Reactions Producing Refractory Damage.................................................123
2. COATING: PROPERTIES AND BEHAVIOR..................................................................127
2.1 General..................................................................................................................... 127
2.2 Coating..................................................................................................................... 127
2.3 Influencing Parameters.............................................................................................128
3. REFERENCES............................................................................................................... 134

5888 REFRACTORY MATERIALS AND THEIR PROPERTIES
1.1 General Introduction

The refractory lining of a modern, high performance Portland cement kiln is a complex
system performing different specific functions at various locations. These include:
0 protection of the kiln shell and tyres from reaching excessive temperatures
1 protection of the kiln shell against chemical attack by aggressive gases and liquids
2 protection of the kiln shell from mechanical abrasion by kiln charge
3 reduction of heat losses to acceptably low values
4 increase of heat exchange between gases and kiln charge
5 control of material flow
Whereas in the rotary kilns in service until the late 1930's these duties were performed by
fireclay products of approximately 40% Al2O3 content, the high performance kilns of today
necessitate the use of a wide spectrum of specially developed high MgO, Al2O3 and SiO2
refractory types. Within each of these main categories, each manufacturer has designed his
product differently to incorporate the required degree of resistance to the combined
chemical, thermal and mechanical attack that will be encountered in service.
From the viewpoint of clinker production, the financial outlay of relining a kiln (including
removal and installation of the worn and new refractories) represents approximately 2-4 % of
total cement manufacturing costs. The effective loss of cement sales resulting from reduced
kiln availability constitutes, however, a much higher loss should the kiln be shut down for
more than 10-15 days.
Within the plants of Holcim Group, an estimated 30'000 tons of refractories are consumed
each year which, when the following sales costs are assumed
magnesite : 1'400 DM/ton
dolomite : 1'100 DM/ton
others : 1'050 DM/ton
results in an annual expenditure for refractory materials alone of $20'000'000. To this
already substantial sum must further be added the man -hours associated with relining,
and lost cement sales resulting from 44'000 hr. of kiln down time.
It is therefore obvious that the life expectancy of the kiln refractory lining is a topic of basic
financial importance. Depending on kiln zone, refractory types and production conditions,
service lives in the critical sintering and transition zones may range from only several
weeks to well in excess of 1 year.
The principal factors determining a linings service life are schematically displayed in Fig. 1,
indicating those topics which lie within the responsibility of the refractory supplier and those
over which the cement producer has most influence.

Fig.1 Factors Influencing Refractory Lining Lifetime

The recent reintroduction of coal-firing, together with the use of dry process kilns has
resulted in refractories being subjected to appreciably increased attack by fused coal ash, by
local reducing atmospheres and by alkali sulfates. To combat these trends, refractory
manufacturers have recently increased their range of cement kiln refractories to include iron
and chrome free periclase-spinel and magnesium silicate (forsterite) types.
The technological complexity of the kiln linings has been developed to such a stage and is of
such financial importance that each plant should have a refractories specialist possessing a
good basic technical/theoretical knowledge and having additionally sound practical
experience. To help achieve this aim, the Holcim Refractories Handbook for the Cement
Industry is intended as a compact source of cement industry specific information and is
intended for both newcomer and refractories specialist alike.
1.2 Classification of Refractory Products
1.2.1 General
Refractory materials are mainly composed of Al2O3, SiO2, MgO, CaO and Cr2O3 existing as
the free oxide or combined together. They are defined according to ISO Standard R 836 as
"non metallic materials or products having a minimum pyrometric cone equivalent of
1500°C. Internationally they are classified, according to PRE recommendations, as
compromising of the following technological categories:
0 dense, shaped refractory products .
1 shaped insulating refractory products .
2 prepared unshaped dense and insulating refractory materials
1.2.2 Chemical Classifications

The different groups of refractory materials, based on their chemical composition, is listed in
Table 1.

Table 1 Chemical classification of dense shaped refractory products according
to ISO 1109 1975 (E)

As can be observed according to ISO Recommendations, there are
0 acidic refractory products
1 basic refractory products
2 special products
each with a defined percentage of the component oxides in the finished products. In a
further subdivision of the fireclay products (i.e. 30% Al2O3 < 45 %), the following
categories are recognized according to PRE/R 29:
Table 2 Groups of Fireclay Products

Group Al2O3 content (%)
30 30-33
33 33-37
37 37-40
40 40-42
42 42-45
The properties of several ranges of refractories are provided in Table 3.
Table 3 Composition and Properties of Refractory materials

Appendix 1 Structures of “Spinel” Bricks

1.2.3 Nomenclature of Manufacturing Process
To provide further information on the various chemically based classes of shaped
refractories, the manufacturer may supply additional characterizing information such as:
0 process of shaping
1 method of bonding the material
With regard to the shaping process, various methods are employed by the refractory industry
in the manufacture of bricks.
Besides the casting of melts of specific compositions (fusion cast products), the methods of
fine ceramics, such as slip casting, continuous extrusion or pressing of finely crushed
materials, have been used increasingly during recent years.
However, the so-called coarse ceramic method is preferred for the manufacture of refractory
bricks. Production by this technique includes the following steps:
0 crushing
1 classification (grading)
2 mixing
3 shaping (pressing)
4 drying
5 firing
1.2.4 Types of Bonding

From the viewpoint of bonding, various methods are recognized which can be classified
according to the temperature at which strength properties are developed, e.g.
0 unfired products (<150°C)
1 fired at moderate temperatures (150-800°C)
2 fired at high temperatures (>800°C)
and also according to the material nature of the bond, e.g.
0 fireclay bond
1 inorganic bond
2 organic bond
3 hydraulic bond
4 ceramic bond
Since the bonding in many refractory materials tends towards ceramic bonding at
operating temperatures, it is necessary to state that the previous classification of refractory
material bonds is based on products coming from the supplier ready to be put into the kiln.
Examples of the bonding types encountered in refractories commonly supplied to the cement
industry are given in Table 4.

Table 4 Examples of the Various Material Bonds
Nature of bond Refractory Type Characteristics
CERAMIC Fireclay bricks The brick is first heated up above
(Eutectic) High-alumina bricks the eutectic temperature of the
Silicate bonded naturally present bonding oxides,
magnesite bricks and the structure formed by the
liquid on cooling bonds the
principal components together
CERAMIC High fired mangesia The principal components are "dry"

(High fired) spinel bricks bonded by the use of very high
temperatures (1700 -
1850°C),which causes the principal
solid components to react directly
with one another
INORGANIC Phosphate-bonded The bricks are bonded by chemical

alumina bricks reaction between the added
bonding agents(e.g. phosphates,
sulfates, chlorides, chromates) and
the bricks components at
temperatures up to 150°C
HYDRAULIC Refractory castables The refractory materials are

Gunning mixes bonded by a hydraulic calcium
aluminate cement, just like normal
concrete
1.2.5 Shaped Insulating Refractory Products

These are shaped products having the same chemical classification as dense shaped
refractories (see Table 1) but having a total porosity of at least 45 %. They are further
characterized (ISO 22451972 E) by the temperature at which they do not display a
permanent linear change of above 2 %, and by their bulk density. Such products are also
known as lightweight insulating refractories and are mainly used as insulating backing in the
heat exchangers of dry process kilns.
1.2.6 Prepared Unshaped Dense and Insulating Refractories
These are prepared mixtures which contain an aggregate and a binder ready for use either
0 as supplied e.g. ramming mixes
1 after addition of suitable liquids
e.g. refractory castables
These may be either dense or insulating, the latter requiring a porosity of ≥ 45 %.
Such materials be used for the following purposes, with further subdivisions being
recognized within each group.

Table 5: Categories of Unshaped Refractories
Ramming mixes
delivered in a pliable non-coherent state
and first harden upon heating
Moldable mixes
delivered in a plastic form and harden upon
MONOLITHIC heating
Refractory castables
CONSTRUCTION/ delivered dry Setting and hardening occurs
REPAIRS without heating after mixing H2O or other
liquids
Gunning mixes
delivered in a non-coherent state, suitable
for placing by pneumatic projection
Refractory cements
these are placed wet, and harden at high
temperature by ceramic bonding
MORTARS/ Refractory cements (hydraulic binder)
as above but harden without heating
JOINTING MATERIALS
Refractory cements (chemical binder)
harden by chemical reaction at lower
temperatures than required for cement
bonding
1.3 Manufacture of Refractory Products
1.3.1 Manufacturing Process

A typical flowsheet for the production of shaped and unshaped magnesite based refractories
is shown in Fig. 2. As can be seen, the process can be split into four principal stages
including
23 preparation of sintered magnesia and dried chromite in defined particle size ranges
24 preparation of mix having specific proportions of the various magnesite and chromite
fractions
25 pressing of required format (shaped refractories)
26 bonding process (direct, eutectic, chemical)

Fig. 2 Flow Diagram for the Production of Shaped and Unshaped Magnesite
Based Refractory Products

As can be seen, the magnesite sinter may originate from either natural magnesite or from
Mg(OH)2 extracted from sea water. To produce the latter, sea water containing about 200 g
MgO/100 liters is reacted with slaked lime slurry to precipitate Mg(OH) 2. The resulting
slurry is filtered to produce a paste which is calcined at 900- 950°C to form "caustic calcined
magnesia". This is then fed through briquetting rollers to press dry, dense tablets, which are
then dead burnt at 1800°C in rotary kilns. During the burning process, the undesired B2O3
content of the MgO is appreciably reduced. In the production from natural magnesite,
impurities such as dolomite are removed by floatation after crushing. From the dead burnt
sinter the coarser fractions are then prepared by crushing and sieving, with the finest being
prepared by subsequent milling.
By using different sizes of sinter and chromite in various proportions, and controlling the total
porosity and pore size distribution, a variation of refractory properties may be achieved.
1.3.2 Source of Materials
The production of refractory materials was originally located in areas near to iron works
where clay deposits were available with a sufficiently high Al2O3 content to allow the
production of fireclay bricks. With the development of refractories to include high alumina,
magnesite, magnesite chrome, and most recently periclase spinel types, the raw
materials are not available in the traditional production areas and have to be purchased
on the international market. The refractory manufacturer therefore cannot always produce
his product with materials from the same source.
Of the various materials of a grade suitable for refractory production, the most well known
sources are indicated in Table 6.

TABLE 6 Sources of Raw Materials for Refractory Production
Raw Material Source
For Basic Refractory Products Country Location
Magnesite
• coarse crystalline(natural) Austria Radenthein
Veitsch
Canada Radium Hot
Spring
• crypto crystalline(natural) Greece Euboea

Turkey
Yugoslavia
India
Australia
South Africa
• sea water (synthetic) Italy Sardinia

England Hartlepool
Ireland Drogheda
Japan
USA
chrome ore (natural) Philippines

Turkey Transvaal
South Africa
Greece
Iran
magnesia spinel(synthetic) Japan
England
Germany
The key factor deciding whether a raw material can be used for refractory production is the
content of the minor oxides, i.e. CaO, SiO2, Fe2O3 and Al2O3 which, when they are present
in excessive quantities, lead to a reduced refractoriness. Synthetic raw materials (e.g. sea
water magnesite) are often too pure and have to have their composition modified by the use
of additives to control their CaO/SiO2 ratio and hence improve their sintering behavior.
Although the critical composition limits at which a manufacturer accepts or rejects a specific
raw material are not made available and remain a trade secret, typical compositions of
acceptable materials are as indicated in Table 7.

Table 7 Typical Compositions of Raw Materials
Material Chemical Composition (wt %)
SiO2 Al2O3 Fe2O3 CaO MgO Cr2O3 TiO2
BAUXITE 6.2 86.9 1.8 0.04 0.08 -- 2.7
(sintered)
CHROME ORE 5.4 20.4 18.4 0.6 18.1 37.0 --
DOLOMITE 2.3 1.0 0.8 58.8 36.2 -- 0.2

(sintered)
MAGNESITE 0.4 0.2 0.6 0.9 98.1 -- --

(sea water-
sintered)
MAGNESITE 2.1 0.6 1.5 2.1 93.6 -- --

(raw)
1.3.3 Production Details

Production details for the manufacture of refractories are seldomly made available. The
following points, however, give an indication of typical layout and performance of a
refractories plant, which receives already sintered raw materials (see Table 8).
Table 8 Production Details
No of silos per component: up to 4 for various size fractions
Weighing and mixing usually automatically controlled by

process: computer. Proportions change with brick
types and material source
Brick presses: 1500 T total pressure, fully automatic

operation. Often up to 5 presses/plant
Tunnel oven (for firing): capacity 30'000 t/y
Firing data: time up to 3 days temperature 1550 -

1850°C heat consumption 1500-2000
kcal/kg

1.4 Testing of Unused Refractories
1.4.1 General
Refractory products are routinely tested at several stages during their manufacture and in
some industries (e.g. steel production) are even tested by the user prior to delivery before
actually being put into service. Testing may be considered as falling into the following
categories and fulfilling the indicated functions.
Table 9 Testing of Unused Refractories

Type of testing Purpose
Manufacturers Process Check on Production Process/Equipment/Operations
and Quality Control
Standard Refractory Tests To ensure that product meets manufacturers description
(e.g. DIN, ASTM, etc.)
Non-Standard Tests To assess whether product is suitable for operations in the
(e.g. at HMC) specific user industry
Users Acceptance Quality Check whether delivered refractories meet manufacturers

Control specifications or are suitable for users specific
requirements
Whereas in the cement industry the used standard ASTM, DIN, BS tests are considered as
giving a not unreliable indication of the performance of the mortar or concrete in service, the
prescribed standard tests for refractories are not specific for clinker production conditions. It
has even been suggested that it would be more useful to test refractories that had already
been in service, and not new unused refractories, and so obtain a practice related indication
of the suitability of refractories under clinker production conditions.
1.4.2 Manufacturers Process Control
Upon delivery, the refractory manufacturer checks the chemical composition of the supplied
chrome ore, sintered magnesite, dolomite, bauxite etc. to ensure that the levels of the minor
oxides, e.g. CaO, SiO2, Fe2O3, Al2O3, do not exceed the requirements for the refractory
type concerned (e.g. direct or indirect bonded basic bricks). In the case of sea water
magnesite the boron content is checked, as is the loss on ignition in the case of sintered
dolomite, which must be used within approximately 48 hours of delivery.
During the mixing process, the granulometry of the various fractions of each component (e.g.
chrome ore, magnesite) is routinely checked as is the mixing ratio of the raw components. These
checks are especially necessary when more than one type of for example magnesite - chrome
brick are being simultaneously produced. In larger production plants, quality control samples are
taken on a statistical basis and are sent to the quality control laboratory. In several companies, all
the quality control and production data are stored in a computer databank, thus allowing the
production history of any specific delivery to be easily recalled.
In the most modern production lines, the whole process is fully automated from component
mixing to packing. After pressing, each unfired brick is automatically checked for its
dimensional accuracy, whilst the weights are checked only on chosen samples. After firing,
the bricks are all individually tested automatically for dimensions, and by a controller who
checks for various subjective factors. The following table lists typical factors evaluated in the
subjective tests:

Table 10 Criteria for “Optical” inspection of Refractory Bricks

1.4.3 Standard Refractory Tests
The product description of refractories is defined in terms of various standard test methods
(ASTM, BS, DIN, ISO) which measure their chemical composition and physical properties.
Testing is performed both at room and elevated temperatures and for shaped refractories
includes the following:
Table 11 Standard Tests for Refractory Materials

Temperature Property Standard
ASTM BS DIN
Ambient Dimensions C-134 1902
Density C-134 1902 51'065
Porosity C-20 1902 51 ‘056

51'058
Cold crushing C-113 1902 51’048
strength 51'067
Chemical C-572, 1902 51'070
composition 573,574
Elevated Pyrometric cone C- 24 1902 51'063

equivalent
Hot rupture C-583 1902 51'048
modulus
Refractoriness C-16 1902 51'053
under load
Creep in 1902 51'053
compression
. Thermal C-832 1902 51'045

expansion
Permanent linear C-113 1902 51'066

change
Shock resistance C38, 1902 51'068

107,122
Thermal C-201, 202 1902 51'046
conductivity
Testing procedures have also been established by ASTM for unshaped refractories and
include the following in addition to those applicable for shaped products:
23 workability index (C-181)
24 dry firing shrinkage (C-179)
25 cold bond strength
It should be appreciated, however, that various test methods supply data which cannot be
numerically come pared with that obtained by another standard test method, e.g. thermal
conductivity. This must be kept in mind when selecting refractories from international
suppliers, who often define the quality of their product according to their national standards.
A higher catalogue value does therefore not always necessarily imply a higher quality
product, and makes a rational selection even more difficult. To avoid such difficulties,
international standard specifications are in the process of being established, but progress in
this direction has been slow. So far only test procedures for Pyrometric Cone Equivalent
(ISO R 528) and refractoriness under load (ISO R 1893) are internationally defined. Test
procedures for high temperature modulus of rupture, bulk and true density, and permanent
linear change have reached a "draft international standard" status. For the majority most
important properties no international standard yet exists.
The previously listed standard tests, although they do to a moderate extent indicate the
general performance of a refractory, are not wholly specific for conditions representative of
those encountered in cement rotary kilns. It is therefore not surprising that in the cement
industry no detailed generally known and accepted guidelines have been established for the
exact properties of the refractory types required in the various kiln zones.
Although qualitative rather than quantitative, a hypothetical relationship can be proposed
(Table 12), indicating which physical and chemical properties exert an influence on the rate
of wear by the common mechanical, thermal and chemical wear factors. From this it can
therefore be deduced that the standard tests are more suited for determining the resistance
to mechanical attack, less for thermal attack, and almost no indication against resistance to
attack by chemical influences.
Table 12 Significance of Characteristic Refractory Properties for In-service Life

Expectancy

1.4.4 Non-Standard Tests
As a means of assessing the performance of refractories under more practice
related conditions, almost all manufacturers have developed their own non-standard
test procedures. These include methods for the determination of
23 coating adhesion thermal
24 shock resistance
25 resistance to alkali attack
all of which can be tested by not too complicated test procedures.
1.4.5 Sinterzone Coating Adhesion

The tendency towards coating adhesion can be determined at HMC/MA by means of its own
recently developed Fig. 3, and fully described in various publications. In developing this
method, all the previously developed procedures were evaluated, their deficiencies
eliminated and their advantages combined in a single large scale laboratory method.
Fig. 3: Coating Test Furnace - Longitudinal section
Examples representative of the type and degree of various observed contact surfaces
obtained by this method are given in Fig. 4.
Fig. 4: Typical Contact Zones in Coating Zones

1.4.6 Thermal Shock Resistance
A simple, but extremely aggressive thermal shock resistance test, employed internally in the
German steel industry, requires a heated prism of the refractory material under test to be
placed on its end on top of a water cooled copper plate (see Fig. 5).
Fig. 5 Arrangement of VDEhL Thermal Shock Test
This test is considered to be more practice relevant than the temperature shock caused by
spraying the complete refractory test piece with water or air, conditions never encountered in
practice. The test, because only one end of the refractory material is cooled, is said to be
much more demanding than either of the two DIN methods and at the same time more
practice-like.
1.4.7 Alkali Attack
To determine the resistance of high alumina refractories to alkali attack, many manufacturers
use a modified form of the DIN 51'069 test procedure originally developed for testing the
resistance of refractories to blast furnace slags. In this method 50 mm diameter x 35 mm
deep holes in the test pieces are filled with K2CO3 and the material heated for 5 hr. at
1100°C. After cooling, the test samples are cut with a diamond saw and attack judged
according to depth of penetration and to whether the bricks show cracking or not. Examples
showing a weak and a strong resistance to attack by this simple test method are shown in
Figs. 6a and 6b.

Fig 6a Strong Alkali Attack

Fig 6b No Alkali Attack
1.4.8 Acceptance Testing Criteria

Although acceptance testing of refractories is not practiced in the cement industry, this is not
the case in steel. Refractories are not, however, tested just for testings sake, but are viewed
from the aspect of their costs and from the demands placed upon them, and quality control
testing carried out accordingly. A principal objective of quality control is to prevent cases of
failure in sensitive aggregates which in addition to immense financial consequences may in
the steel industry also endanger human lives.
For the cheaper refractory grades, for less critical applications, the manufacturers data sheet
specifications and quality control are accepted, with no essential input on the part of the
user. Simple optical inspection of the state of the delivered materials is considered as being
sufficient. Even for moderate cost refractory grades in service under moderately critical
applications, testing is confined to carrying out spot checks, after delivery, on random
samples, determining such conveniently measurable properties as dimensional accuracy,
cold compressive strength. The results are collected to detect trends over long periods and
are also if anomalous communicated to the supplier, for the lather’s “information".

Only for applications for which the refractory properties are critical are quality control
examinations carried out systematically on a statistical basis. In such cases, the refractories
are controlled prior to dispatch from the suppliers plant and checked according to
specifications and tolerance limits previously agreed upon between supplier and client. The
chosen test samples are normally divided into 2 sub- samples, one portion being tested in
the supplier's laboratory, the other in the user's. This class of refractory is ordered well in
advance of the specified delivery date e.g. 6 months, and this long range planning assures
that the material to be supplied lies within the acceptance limits. The costs for statistically
significant quality control are borne equally by supplier and user and often involve 3 - 4 man
days working time for a quantity of 150 - 200 tons, the costs corresponding to 1.5 % of the
delivery value.
1.5 ZONING of Cement Kilns

Because the refractories in the cement kiln have to perform several duties, as already listed
in 1.1, no single refractory type can now fulfill all the requirements. As factors such as
5888 the changing temperature profile in the kiln
5889 the changing physical and chemical nature of the kiln feed ranging from raw
meal to clinker
must be taken into consideration, the zoning of the kiln, with respect to the refractory lining,
is a necessity. Indeed in some kiln zones it is normal to have two refractory types at any one
position, namely a "working lining" and a backup insulating lining.
In a commonly established practice, the kiln zones are defined according to the changes
occurring in the kiln charge as indicated in Table 13.
Table 13 Characterization of the Various Kiln Zones

Kiln Zone Process Material Change Temp.Range
Drying Wet and semi-wet Expulsion of slurry or ≤100°C
nodule moisture
Preheating Wet and dry Expulsion of chemically 100-700°C
bonded OH groups
Calcining “ Expulsion of CO2 from 600-900°C
carbonates
Transition “ Formation of 900-1350°C
intermediate and
clinker minerals
Sintering “ Granulation of clinker ≥1350°C
and formation of
permanent coating
Discharge “ Cooling of clinker 1400-1250°C
within kiln
Variations to the previous scheme include the definition of a chain zone (i.e. a combined
drying and preheating zone) and the division of the transition zone into the safety zone and
the transition zone, often described as the upper and lower transition zones.
In Figs. 7 and 8 a typical zone plan can be seen for wet and dry process operations with
indications of the kiln temperature and various types of attack to be expected.

Fig. 7 Typical Rotary Kiln Data - Dry Process
Fig. 8 Typical Rotary Kiln Data - Wet Process

1.6 Wear of Kiln Refractories
1.6.1 General
Even under the most favorable operating conditions, refractory linings are subject to a
definite measurable rate of wear. Based on plant experience (Fig. 9), the minimum
consumption rate achievable in practice, for basic and acidic types may be considered as
follows:
2
0 basic refractories ≈ 6 cm /t clinker
2
1 acidic refractories ≈ 4 cm /t clinker
It should be considered that these values are average values for the total basic and acidic
linings and that at various positions inside the kiln both higher and lower specific
consumption rates are observed. Whereas some plants can consistently approximate to
these desired low consumption rates, in most cases refractory consumption is higher with
2
average consumption rates of 30 cm /t clinker and above being not uncommon.
Fig. 9 Consumption of Kiln Linings

1.6.2 Causes of Wear
The principal causes of wear may be attributed to thermal, mechanical or chemical factors,
acting either alone or more frequently in combination. Details concerning these various wear
types are given in Fig. 10.
Fig. 10 Influence of chemical, thermal and mechanical factors on kiln refractory

linings
Depending on the period for which observations are based it can be observed that the
frequency of wear, due to any particular mechanism, is not constant but varies. This can be
seen in Fig. 11 showing the frequency of the wear mechanisms from 1964 - 1968 and from
1968 - 1974, i.e. since the introduction of large diameter process kilns. As can be deduced,
the identified causes of wear before 1968 were principally chemical (silicate and alkali
infiltration) in nature with mechanical and thermal factors being of equal but secondary
importance. Between 1968 and 1974 mechanical stresses became by far the principal failure
cause, with chemical and thermal being of equal importance. Whereas before 1968 the
causes could clearly be identified in only 50 % of the investigated cases, the percentage of
those not identified dropped to 25 % between 1968 and 1974.
Although no statistics are presently available, the principal causes of wear may again be
chemical in nature, originating from silicate attack and reducing conditions stemming from
the widespread reintroduction of coal firing.

Fig. 11 Main Causes of Wear In Sintering Zones
1.6.3 Mechanisms by which Damage Occurs

The chemical, thermal and mechanical factors responsible for wear cause changes to occur
in the refractory materials by means of material transport stress and erosion. This results in
either a modification or a destruction of the refractories texture leading to discontinuous wear
by spalling, or to a continuous wear by abrasion. The relationship between these factors are
indicated in Fig.12.

Fig. 12 Characterization of Wear Magnesite-Type Linings
1.6.4 Recognition of the Various Wear Causes

To help identify the cause or causes of refractory wear, a range of characteristics can be
proposed summarizing the changes resulting from the various attack mechanisms. A list of
these is presented in Tables 14a and 14b. As can be seen, many of the changes can be
observed macro or microscopically or by chemical analysis, and so a preliminary conclusion
as to why refractories have failed can often be drawn by plant personnel.

Table 14a CHARACTERISTICS OF THE VARIOUS DAMAGE MECHANISMS FOR
ACIDIC AND BASIC BRICKS
0 Chemical Damage
0 Liquid infiltration (clinker / coal ash
0 lower porosity due to penetration
1 change of colour
2 change of mineralogy
3 increase in CaO, SiO2, (Al2O3, Fe2O3)
1 Alkali condensation
0 increaes in Na2O, K2O, Cl, SO3 levels
1 porosity decrease
2 presence of light coloured layers
3 alkali bursting (Al2O3 refractories)
2 CO2/SO3 attack (on CaO containing compounds)
0 decrease of porosity
1 presence with whie coloured layers
2 increase in CO2, SO3
Table 14b CHARACTERISTICS OF THE VARIOUS DAMAGE

MECHANISMS FOR ACIDIC AND BASIC BRICKS
0 Thermal damage
0 Overheating
0 melting of hot face
1 change (increase / decrease) in type of porosity
2 elongation of periclase crystals
3 reduction of chrome content
4 diffusion of low melting silicates, aluminates
5 change of colour
1 Shock
0 spalling of hot face
1 Mechanical damage
0 Kiln deformation / thermal expansion
0 porosity unchanged
1 chemistry / mineralogy unchanged
2 spalling over wide areas
3 crack formation on micro to macro scale

Appendix 2: A graphical representation of the damage zones
1.7 Chemical Reactions Producing Refractory Damage

The chemical reactions occurring when refractories are attacked by siliceous or alkali sulfate
infiltration’s or by reducing atmospheres have been thoroughly investigated. Examples will
now be given of those reactions occurring at various positions in the kiln.
1.7.1 SO3 Attack on Castables in Heat Exchangers
The bonding agent of many refractory castables is monocalcium aluminate (CA) which exhibits
both hydraulic properties (e.g. high alumina cement) and a high melting point (> 1600°C). At
service temperatures in the range of 800 1000°C, CA reacts with SO 3 from the kiln gases and
on cooling (during shutdowns) a further reaction with moisture can occur.

The foil owing reactions have been demonstrated to take place:
0 In Service
2 CaO.Al2O3 + SO3 → CaO.2Al2O3 + CaSO4 (fast)
(monocalcium aluminate) → (calcium dialuminate) + (anhydrite)
CaO.2 Al2O3 + SO3 → 2 Al2O3 + CaSO4 (slow)
(calcium dialuminate) → (alumina) + (anhydrite)
1 During Shutdowns
CaSO4 + 2 H2O → CaSO4 . 2
H2O (anhydrite) → (gypsum)
As a consequence of all these reactions their occurs
0 a destruction of the bonding and a change of structure
1 a loss of strength
which are claimed as promoting conditions beneficial for the enhanced adhesions of build-
ups.
1.7.2 Alkali Bursting of Alumina Refractories
The most well known method by which high alumina bricks are damaged is the so-called
alkali- "spalling" or "bursting". By the pick up of alkalis, the formation of feldspathoids
takes place which, because of an increase in volume, causes a disruption of the brick at a
given depth below the hot face. The ensuing cracking over a large area causes the heads
of the brick to spall off. For the various components of high alumina bricks different
volume changes occur.
0 Mullite Based Refractories
3Al2O3 2SiO2 + 3 K2O + 4 SiO2 → 3 K2O.Al2O3 . 2
SiO2 (Mullite) → (Kalsilite)
∆Volume = + 29 %
23 Bauxite Based Refractories
Al2O3 + K2O + 2 SiO2 → K2O.Al2O3 . 2
SiO2 (Bauxite) → (Kalsilite)
∆Volume = + 2.5 %
As a consequence, should it be necessary to install high alumina refractories in zones in
which the danger of alkali spelling occurs (e.g. transition zone in the vicinity of the kiln type),
preference should be given to a bauxite rather than mullite based product.
1.7.3 Reducing Atmosphere Attack in Magnesite Refractories
The occurrence of local reducing conditions in the burning and transition zones can, when
no protective coating is available, bring about a dramatic loss of brick strength. Damage by
this mechanism has been more frequent since the reintroduction of coal firing and has been
one of the factors decisive in the necessity to develop new Fe 2O3 and Cr2O3 free magnesia
refractories i.e. magnesia-spinel products.
Typically, this type of chemical attack takes place on the Fe-containing magnesia ferrite
(MgO . Fe2O3) component, present as a precipitate within the periclase crystals, when the
refractories have been in service at approx. 1100°C over an extended period in an oxidizing
atmosphere.

Chemically, the following reaction can be considered to occur:
III II
MgO.Fe 2O3 → (Mg/Fe )O ss
(Magnesia Ferrite) → (Magnesia Wüstite)
23 Volume = -20 %
The reduction in volume leads to an increase in porosity with a resulting decrease in
strength due to the fo rmation of voids.
1.7.4 Gaseous Attack on Dolomite Bricks
Although they give excellent service when protected by a permanent coating, without this
protection, the reactive free CaO component of dolomite refractories is very susceptible to
reaction by various reagents, all of which result in an increase in volume. Examples of these
include:
Reaction Volume Increase

CaO → CaS + 155 %
CaO → Ca(OH)2 + 200 %
CaO → CaCO3 + 221 %
CaO → CaSO4 + 275 %
CaO → CaCl2 + 313 %
The MgO component remains in comparison much less reactive to gaseous attack.
Depending on the temperature conditions and the kiln atmosphere, the following reactions
occur in the various sections of a dolomite lining:
23 Middle Brick Zone
CaO + SO3 → CaSO4
-2
CaO + S → CaS
2 CaO + 3 SO3 + K2O → 2CaSO4.K2SO4
24 Lower Brick Zone
CaO + C02 → CaCO3
Because of the increase in volume at a specific temperature and hence depth the increase
in volume results first in the formation of cracking and then in the spelling of the hot brick
face.
Dolomite bricks are also subject to hydration in air when their temperature is under 450°C.
Because used bricks have different concentrations of newly formed minerals, their hydration
rate is not the same over the whole structure. It has been shown that the middle CaS
containing zones hydrate faster than the hot face, which in turn hydrates faster than the hot
face, which in turn hydrates faster than the cold one. This is shown in Fig. 13 showing the
rate of destruction of a single used dolomite brick exposed to laboratory air at 22°C and 65%
r.h.

Fig. 13 Destruction rates of Dolomite Brick Zones

23 COATING: PROPERTIES AND BEHAVIOR
2.1 General
The formation of durable, protective coating on the refractory lining of a cement kiln is
desirable, in a technical as well as in an economical respect, to guarantee steady kiln
operations.
Basically, it is known that the properties and behavior of coating have to be regarded as the
result of a combination of constructional, process technological and material inherent
parameters. The three partial-problems:
23 Formation of coating
24 Adhesion and strength of coating
25 Prevention of ring formation
are closely related to the material parameters as
23 Refractory type
24 Clinker dust and ash characteristics
25 Gas phase and volatile elements
and to the constructional/operational aspects as:
23 Kiln system
24 Kiln operation
25 Operating personnel
2.2 Coating
2.2.1 Definition
The term "coating" is applied in the cement industry for different phenomena. It is therefore
necessary to distinguish between the following:
23 Coating in the cement kiln is the desired protective layer on the refractory lining, which is
formed during kiln operation under the influence of high temperature from the traveling
material on the refractory lining in the transition and sinter zones.
24 Ring formation is an undesired degree of growth of coating.
25 Incrustations, referred to as build-ups, are "coatings" which are formed below 1100°C by
alkali salts and which occur particularly in pre-heater systems.
26 Refractory lining encompasses all construction materials which are intended to eliminate
or reduce the effects of thermal strain.

2.2.2 Significance of Coating
Coating has the following functions to fulfill:
23 Protection of the refractory brick against overheating and abrupt temperature changes
(improvement of heat distribution).
24 Protection of the refractory lining against the abrasive effect of the moving clinker bed
and the dust.
25 Protection of the refractory lining against chemical attack of clinker constituents, fuel ash
and gas atmosphere.
26 Augmenting the mixing effect, i.e. improvement of the blending effect of the clinker mass
due to unevenness of the coating surface.
27 Reduction of heat radiation and protection of the kiln shell against thermal overloading.
According to experience with large kilns, coating characteristics and behavior were found to
be the most significant parameters associated with steady kiln operations. The operating life
of refractory linings is substantially longer, when stable coating is achieved over the whole
length of the kiln.
Protective coating begins normally in the transition zone which is also the zone, where ring
formation may frequently occur.
2.2.3 Formation of Coating
Two basic hypothesis are plausible as to the mechanism of coating formation:
23 Formation of clinker melts (above cat 1280°C) and their sticking together with the
refractory lining with simultaneous cementing of solid particles.
24 Adhesion of solid crystalline phases on the refractory lining.
In both cases, chemical and/or physical reactions (impregnation/infiltration) with the upper
zone of brick strengthen the adhesion of the coating.
The temperature gradient of coating refractory lining kiln shell effects a further stabilization of
the coating as a result of cooling and crystallization. The formation of coating increases to
that point at which its inherent insulation characteristics prevent further growth.
2.2.4 Thickness of Coating
Usually, the following classification is applied as to the thickness of coating:
30 - 50 cm thick coating
~ 20 cm normal coating
0 - 10 cm thin coating
Coating thicker than 50 cm is classified as ring formation.
2.3 Influencing Parameters

Coating is preponderantly formed from the components of raw material and clinker, whose
characteristics and composition primarily determine the quality and quantity of the coating. In
addition, the type of fuel and diffusion products of the refractory lining also exert a certain
influence.
Industrial experience demonstrates that a correlation exists between the characterizing
moduli (silica ratio, alumina ratio, lime saturation) and the rate and ability of coating

formation. Within the system SiO2 Al2O3 Fe2O3, i.e. S A F, the dependence of coating
formation and chemical composition has been demonstrated as occurring in Fig. 14.
Fig. 14 Correlation of coating formation and chemical parameters
Coating formation is facilitated by decreasing the ratios A/F and S/A.

In a second triangle (right) indicating two series of lines, viz. equal percentage of liquid
phase at 1450°C and equal fluidity of the liquid phase it is demonstrated that, disregarding
the lime saturation, the tendency to coating formation increases with increasing proportion
and fluidity of the liquid phase.
The liquid phase of a clinker is defined as that material present in liquid form, after beginning
of the formation of clinker minerals. This usually first forms at temperatures of approximately
1250°C. This liquid phase content at 1450°C may be up to 35 %, but on average lies around
25 %.
The quantity of liquid phase and its physical properties (viscosity, surface tension, etc.) are
dependent on chemical composition and temperature.

From experience, a correlation is known to exist between the quantity of liquid phase (C3A +
C4AF + MgO + K2O + Na2O + SO3) and the thickness of the coating. If a raw mix with less
than 20 % potential liquid phase is employed, the formation of coating in the burning zone is
expected to be weak, whereas liquid phase contents of more than 25 % lead to thick coating
and frequent ring formation.
Another display divides the field of potential coating formation into five areas (Fig. 15).
Fig. 15 Clinker composition and coating formation as influenced by the ratios of

SiO2 Al2O3 (Fe, Mn)2O3

Coating formation is subdivided as follows:
field 1 strong
field 2 medium strong
field 3 medium
field 4 low
field 5 very low
Generally, the following can be deduced from Fig. 14:
23 Increasing alumina ratio and SiO2/Al2O3 ratio impair coating formation and vice-versa.
24 Favorable coating conditions are achieved with a SiO2/Al2O3 ratio of 2.5 to 3.3 and
an alumina ratio of 1.8 to 2.3.
25 Favorable coating conditions are achieved at constant medium or high SiO2/Al2O3
ratios, by lowering the alumina ratio.
26 Extremely SiO2/Al2O3 high ratios necessitate extremely low alumina ratios.
27 Main influencing parameter is the alumina ratio.
In addition, a number of other measures to improve the tendency to coating formation have
been proposed:
23 Lowering of the CaCO3 content in the raw mix
24 Addition of CaF2 (fluorspar)
25 Addition of pyrites ash
etc.
However, the chemical characteristics of the raw material should not be over-emphasized,
since the characteristics of the refractory lining represent additional, determining parameters
as to coating formation.
2.3.2 Mineralogical Composition
Experience shows that raw mixes containing elevated clay mineral contents and low quartz
contents lead to better coating ability than raw mixes of reversed mineralogical
composition. Since the mineralogical composition influences the ratios and liquid phase
contents, the chemical composition is already a conclusive criterion.
2.3.3 Homogeneity of Raw Mix
Inhomogeneity of the raw mix composition is normally responsible for fluctuations in coating
formation. This applies not only to chemical inhomogeneity, but also to the variations of the
raw mix fineness and to the characteristics of the fuel.
The effects of changes of raw mix properties on the already formed coating layer may be
deduced with the aid of Fig. 16.

Fig. 16 Schematic Temperature profile of kiln clinker as a function of material
properties
Clinker A of good burnability upon up-heating along a - b enters the burning zone at
temperature T (point b). It remains during clinker formation at this temperature and then
cools down, along the temperature profile c - d. During the contact of clinker A with the
refractory lining, a layer of coating is assumed to have been formed.
Any change of the raw mix composition, which impairs its burnability, often leads to a change
of clinker melting properties. For the more difficult to burn clinker b let us assume that an
increased amount of energy will be necessary resulting in a new temperature profile a' - b' -
c' - d'. With this changed profile, the already formed coating layer becomes softer, more fluid,
and is easily scraped off by clinker B. If the refractory linings are already worn out, the kiln
shell will become hotter and perhaps even red spots may occur. In any case, the refractory
lining in the burning zone will be weakened as a result of the new temperature at the brick
surface. Frequent changes and loss of coating (due to fluctuation of composition of raw mix)
particularly in big kilns often results in a considerable decrease in refractory life caused by
spelling.
2.3.4 Circulating Elements
-
Circulating elements are the alkalis (K2O, Na2O), chlorine (Cl ) and sulfur (SO2, SO3,
2-
eventually S ). They evaporate from the raw materials (raw mix and fuel) and condense on
the colder raw mix and penetrate into the refractory lining. They play a significant role with
regard to the stability of coating and refractory lining life, but influence coating formation to a
lesser extent.

Alumina-rich refractory bricks are easily infiltrated by alkalis which react with the brick
constituents forming minerals in this zone, such as leucite and kalsilite. These minerals
crystallize upon cooling and lead to spelling of the brick surface. Such reactions don't occur
between the alkalis and the basic bricks. Corrosion of bricks is also observed, if sulfates are
present in notable amounts. The destructiveness of alkali sulfates (and chlorides) on coating
is caused by the following mechanism:
23 The sulfate-melt penetrates and becomes enriched in the pores and cracks of the brick,
the position of enrichment depending upon the thermal gradient.
24 Periodic changes of the temperature gradient produce both melting and recrystallization,
which results in crack formation vertical to the temperature gradient, due to pressure of
crystallization.
Such an infiltrated zone is the cause for the spelling of 4 - 6 cm thick layers of brick and
adhering coating. As a rule of thumb it can be said that large quantities of alkali and sulfate
are always especially harmful, if a molar excess of sulfate exists.
2.3.5 Crystal phases
As already discussed, coating formation is mainly considered as being controlled by the
properties of the clinker liquid phase.
However, there is also another possibility to be considered: the compaction of powdery
materials at high temperatures by sintering or recrystallization. This effect may be explained
by the preferential grow of bigger crystals at the expense of smaller ones, a diffusion
process which takes place at temperatures below the melting points of the substances.
In nearly all cases in laboratory tests as well as under plant conditions a small layer of C2S,
less than 1/2 mm in thickness can be found in the contact zone between brick surface and
coating, whereby the crystals of C2S in this zone seem to be glued together without
interstitial phases. The C 2S zone extends into coating as well as into the upper parts of the
(porous) brick to about 1 mm in depth, thus connecting both, brick and coating.

23 REFERENCES
H. Kock: Probleme der Bildung und Erhaltung eines schützenden
Ansatzes in der Brennzone von Zementdrehöfen
Literaturstudie, HMCReport MA 75/2414/D.
(This report compiles all the meaningful literature references
on refractory material coating and related phenomena.)
D. Opitz: Steinverschleiss in grosser Zementdrehöfen. Vortrag
XVII.Intern. Feuerfest. Kolloquim, 24/25.10.1974, Aachen

(BRD).
D. Opitz: Die Ansatzringe in Zementdrehöfen. Schriftenreihe der
Zementind. Heft 41 (19~/4 VDZ.

J. Forest: Etude des croûtages de zone dans les fours rotatifs et des
bélites. Silicates Ind. XXXII (1967), No. 11, 373 384.

W. Polesnig, W. Beitrag zur Bewertung des Ansataverhaltens und der
Zednicek: Ansatzbildung in Zementrotieröfen. RadexRdsch. (1973), H. 5,

695 712.
M. Majdi (, H.E. Ueber die Ansatzbildung im Drehofen. ZKG 12 (1959), H. 3,
Schwiete: 89 101.
L.G. Huggett: Rotary cement kiln linings, refractory problems in the burning
zone. Trans Brit. Ceram. Soc. 56 (1957), 87 109.
H. Gygi: Wärmetechnische Untersuchungen des Drehofens zur
Herstellung von Portlandzement klinker. Dissertation ETH,

Zürich (1937).
G. Routschka, A. Feuerfeste Baustoffe für die Zementindustrie im Spiegel der
Majdic: Literatur. ZKG 27 (1974), H. 10,469 485

F. Harders, S. Feuerfestkunde. Herstellung, Eigenschaften und Verwendung
Kienow: feuerfester Baustoffe. Berlin/Göttingen/Heidelberg (1960).

D.S. Buist, J.R. Solid Liquid Reactions and Brick/Clinker Adhesion in Rotary
Gelsthorpe: Cement Kilns. Science of Ceramics 4 (1968), 255 273

Chapter 4
Rings, Balls and Build-Ups

Rings, Balls and Build-Ups
1. INTRODUCTION............................................................................................................138
2. LOCATION OF RINGS...................................................................................................138
2.1 Classification............................................................................................................ 138
3. THEORETICAL ASPECTS OF RING AND DEPOSIT FORMATION.............................140
4. CHARACTERISTICS OF VARIOUS RING AND DEPOSIT TYPES...............................140
4.1 Exhaust Fan Deposits...............................................................................................141
4.2 Slurry Rings (including mud balls).............................................................................141
4.3 Cyclone and Grate Preheater Deposits....................................................................141
4.4 Meal Ring (Calcining Ring) in Long Kilns..................................................................143
4.5 Middle Rings in Large Pre-heater Kilns.....................................................................144
4.6 Sinter Rings (excluding coal-ash rings).....................................................................144
4.7 Coal Ash Sinter Rings...............................................................................................145
4.8 Clinker Rings / Cooler Inlet Deposit (snowman)........................................................145
4.9 Kiln Charge Balls......................................................................................................146
5. METHODS OF REMOVAL/ELIMINATION.....................................................................147

23 INTRODUCTION
Rings and deposits are accumulations of solid materials (from the powdery kiln charge) in
the rotary or static sections of clinker production lines. They have been encountered since
the earliest days of rotary kiln production, with each development in process technology, e.g.
grate and cyclone pre-heaters, grate cooler, bringing with them their own specific type of
deposit.
Rather than being of academic interest, ring and deposit formation has an appreciable
influence on plant operations, frustrating operations personnel by their impairing or even
impeding production, and annoying the company management by lowering production (and
sales) and increasing production costs.
As a direct consequence of rings and deposits, the gas and material flow through the kiln is
restricted, resulting in a reduced kiln output. Especially in the sinter zone, the presence of
rings can interfere with combustion of the fuel and can result in improper combustion. From
time to time unstable rings and deposits can break away leading to blockage or mechanical
damage in the cooler, or in cyclone blockages. The partial shedding of coating from the
exhaust fan blades results in severe vibration which mostly requires a short shutdown for
complete removal. The breaking of a ring almost always causes a flush of material into the
burning zone and a temporary loss of stable kiln operations.
The formation of deposits in cyclones results in extra costs for the labour needed to remove
the deposits by poking. The introduction of air canons (big blasters) provides a method for
their regular automatic removal and has been installed in Group plants with persistent pre-
heater blockages. High pressure water jets may also be employed.
In the worst cases, a complete shutdown is necessary to allow entrance to the affected area
and mechanical removal of the blockage with compressed air drills. This shutdown invariably
weakens the sinter zone refractories, and accelerates the next shutdown for rebricking.
23 LOCATION OF RINGS
2.1 Classification
Unwanted build-ups may be classified with regard to the type of material from which they are
formed, either sintered or unsintered. Within these two groups the various types can be
classified as follows:
unsintered:
5888 exhaust fan deposits
5889 cyclone and grate preheater deposits
5890 slurry or mud rings
5891 meal rings

sintered:
࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿첏
࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿첏첧쳔쳖쳮쳱쳴촌촥촧촿쵂쵅쵝춇춉춡춤춧췀췧췩츂츆츉츢칅칇칠칤칧캀캛캝
ꡃꡜꡠꡣ꣢ ꢧꢩꢩ࿿࿿꣢࿿ ꣢ꤎ ࿿࿿ꤐꤐꤐ ꥣꥥ࿿꣢࿿࿿࿿࿿࿿࿿࿿࿿
ꨌ ཇཋཎཧཧྱ࿐
ၯ࿿࿿࿿࿿࿿
ၒ
ၐ
သ
စ
ဂ
࿿ດ
࿿࿿࿿
ꦂ ༮
༬
༟
༆
༃
ະྨ
ະ
ທ ꨢꨤꩁꩄ꩝
࿿࿿
࿿࿿꣢࿿ ࿿
꣢࿿
ളളച ࿿࿿࿿
࿿࿿࿿࿿꣢࿿ ඝ࿿࿿࿿࿿
ꦂൢꦂංൾඨඪස ꦂ
࿿࿿࿿࿿࿿࿿
ꦂ් ࿿࿿࿿
ข ࿿࿿࿿࿿
ฎฐษอะะ ࿿࿿࿿࿿࿿
࿿ ࿿ ࿿࿿࿿࿿࿿࿿࿿࿿
࿿࿿
࿿࿿࿿࿿
࿿࿿ ࿿࿿࿿࿿࿿ ࿿࿿࿿
࿿ ᎦᎨᏁᏅᏈᏡᐫᐭᑆᑊᑍᑦᑪᑬᒅᒎᒐᒩᒭᒰᓉᓯᓱᔊᔎᔑᔪᔸᔺᕓᕗᕚᕳᖃᖅᖞᖢᖥᖾᗏᗑᗪᗮᗱᘊᘞᘠᘹᘽᙀᙙᙰᙲᚋᚏᚒᚫᚯᚱᛊᛖᛘ ࿿࿿
ꙎꙑꙪꙮꙮꚉ࿿࿿꜑ ꜘꜱꜵꜷꝐ
꜕ ࿿࿿࿿࿿࿿꟮ ꡏꡑꡪꡮꡱꢊꢎꢐꢩꤌꤎꤎꤎ꤮ ꣢ꥦ࿿࿿࿿࿿࿿ꨄ
࿿꣢࿿࿿࿿ ꨈꨊꨣ
࿿࿿࿿࿿࿿࿿࿿࿿࿿ ꭕꭗ࿿࿿࿿࿿࿿࿿࿿ 갪갬걅걉걌걥걩걫겄곻곽괖괚괞괷
middle ring
࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿첏
ꡃꡜꡠꡣ꣢ ꢧꢩꢩ࿿࿿꣢࿿ ꣢ꤎ ࿿࿿ꤐꤐꤐ ꥣꥥ࿿꣢࿿ ཇཋཎཧཧྱ࿐
࿿࿿ ༮
༬
༟
༆
༃
ະྨ
ၯꦂ࿿࿿࿿࿿࿿
ၒ
ၐ
သ
စ
ဂ
࿿࿿࿿࿿࿿࿿࿿ꨌ ࿿࿿ ະ
ທ
ດ ꨢꨤ
࿿࿿ ꩄ꩝ꩁ
࿿࿿꣢࿿
࿿꣢࿿࿿࿿࿿
࿿࿿࿿࿿꣢࿿
ളളച ࿿ꦂൢ
࿿࿿࿿࿿࿿
ꦂං ࿿࿿ ꦂ ꦂ්࿿࿿࿿
ඝ࿿࿿
ൾඨඪස ࿿࿿࿿࿿ ࿿࿿࿿࿿࿿
࿿ฎฐษอะะ
ข ࿿ ࿿࿿࿿࿿࿿࿿࿿࿿
࿿࿿
࿿࿿࿿࿿
࿿࿿ ࿿࿿࿿
࿿࿿ ࿿࿿࿿
꜕ ࿿࿿࿿࿿࿿꟮ ꡏꡑꡪꡮꡱꢊꢎꢐꢩꤌꤎꤎꤎ꤮ ꣢
࿿꣢࿿࿿࿿ ꥦ࿿࿿࿿࿿࿿ꨄ ꨈꨊꨣ
࿿࿿࿿࿿࿿࿿࿿࿿࿿ ꭕꭗ࿿࿿࿿࿿࿿࿿࿿ 갪갬걅걉걌걥걩걫겄곻곽괖
sinter
ring
࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿첏
࿿࿿ ༮
༬
༟
༆
༃
ະྨ
ၯꦂ࿿࿿࿿࿿࿿
ၒ
ၐ
သ
စ
ဂ
࿿࿿࿿࿿࿿࿿࿿ꨌ ࿿࿿ ະ
ທ
ດ ꨢꨤ
࿿࿿ ꩄ꩝ꩁ
࿿࿿꣢࿿
࿿꣢࿿࿿࿿࿿
࿿࿿࿿࿿꣢࿿
ളളച ࿿ꦂൢ
࿿࿿࿿࿿࿿
ꦂං ࿿࿿ ꦂ ꦂ්࿿࿿࿿
ඝ࿿࿿
ൾඨඪස ࿿࿿࿿࿿ ࿿࿿࿿࿿࿿
ข ࿿ ࿿࿿࿿࿿࿿࿿࿿࿿
࿿࿿
࿿࿿࿿࿿
࿿࿿ ࿿࿿࿿
࿿࿿ ࿿࿿࿿
࿿꣢࿿࿿࿿ ꥦ࿿࿿࿿࿿࿿ꨄ ꨈꨊꨣ
clinker
ring
࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿첏
ꡃꡜꡠꡣ꣢ ꢧꢩꢩ࿿࿿꣢࿿ ꣢ꤎ ࿿࿿ꤐꤐꤐ ꥣꥥ࿿꣢࿿ ࿿࿿࿿࿿࿿࿿࿿ꨌ ࿿࿿ ཇཋཎཧཧྱ࿐
࿿࿿ ༮
༬
༟
༆
༃
ະྨ
ၯꦂ࿿࿿࿿࿿࿿
ၒ
ၐ
သ
စ
ဂ ະ
ທ
ດ ꨢꨤ
࿿࿿ ꩄ꩝ꩁ
࿿࿿꣢࿿
࿿꣢࿿࿿࿿࿿
࿿࿿࿿࿿꣢࿿
ളളച ࿿ꦂൢ
࿿࿿࿿࿿࿿
ꦂං ࿿࿿ ꦂ ꦂ්࿿࿿࿿
ඝ࿿࿿
ൾඨඪස ࿿࿿࿿࿿ ࿿࿿࿿࿿࿿
ข ࿿ ࿿࿿࿿࿿࿿࿿࿿࿿
࿿࿿
࿿࿿࿿࿿
࿿࿿ ࿿࿿࿿
࿿࿿ ࿿࿿࿿
࿿꣢࿿࿿࿿ ꥦ࿿࿿࿿࿿࿿ꨄ ꨈꨊꨣ
"snowm
an" in grate cooler
࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿࿿첏
࿿࿿ ༮
༬
༟
༆
༃
ະྨ
ၯꦂ࿿࿿࿿࿿࿿
ၒ
ၐ
သ
စ
ဂ
࿿࿿࿿࿿࿿࿿࿿ꨌ ࿿࿿ ະ
ທ
ດ ꨢꨤ
࿿࿿ ꩄ꩝ꩁ
࿿࿿꣢࿿
࿿꣢࿿࿿࿿࿿
࿿࿿࿿࿿꣢࿿
ളളച ࿿ꦂൢ
࿿࿿࿿࿿࿿
ꦂං ࿿࿿ ꦂ ꦂ්࿿࿿࿿
ඝ࿿࿿
ൾඨඪස ࿿࿿࿿࿿ ࿿࿿࿿࿿࿿
ข ࿿ ࿿࿿࿿࿿࿿࿿࿿࿿
࿿࿿
࿿࿿࿿࿿
࿿࿿ ࿿࿿࿿
࿿࿿ ࿿࿿࿿
࿿꣢࿿࿿࿿ ꥦ࿿࿿࿿࿿࿿ꨄ ꨈꨊꨣ
kiln
charge ball
Process technological characteristics of such build-ups e.g. kiln type, location, temperature
of gas and kiln charge can be seen in Table 1. Material technological characteristics e.g.
state of kiln charge, enrichment in various elements, and type of texture are summarised in
Table 2.
The location of the various types of the above rings and deposits can be seen in Fig. 1.
Fig.1 Rings and Build-ups in Different Kiln Systems

0 THEORETICAL ASPECTS OF RING AND DEPOSIT FORMATION
Although of much practical significance, little quantitatively based, fundamental knowledge is
available on the formation of deposits from solids suspended in gas streams. In a qualitative
way, however, the more important features of such processes are known.
The formation of a deposit is always a dynamic process in which the factors responsible for
formation outweigh the forces of degradation. In general, the stronger the forces of
destruction, the more unlikely the chance of deposit formation, but when this does occur, a
strong, hard to remove agglomeration is the result.
After the transport of material to the area of deposition, a definite force is required to make it
adhere to the wall. This can range in magnitude from that caused by turbulence within the
gas stream, increasing to centrifugal forces when the stream changes direction, to that due
to mechanical pressure. Whereas pre -heater deposits involve the first two, mechanical
pressure certainly plays a part in ring formation within the rotating kiln.
The forces according to Rumpf considered to cause deposit formation can be grouped as
follows:
1057 melting or softening of surface due to friction or collision
1058 melting or freezing due to addition or removal of heat
1059 interlocking of aggregates built up of finer particles held together by surface forces
1060 interlocking of long fibrous particles
1061 electrostatic attraction
The mechanisms b, c and d are the ones encountered in kiln operations. In general, the finer
the powder, the greater the tendency towards agglomeration, and in many cases the
absence of particles under a critical size (e.g. 5 um) ensures freedom from deposition.
0 CHARACTERISTICS OF VARIOUS RING AND DEPOSIT TYPES

Tables 3 and 4 contain a list of typical properties of rings and build- ups encountered within
the Holcim Group plants (with full chemical analysis being provided in Appendices I, II and
III). Included are such factors as volatile element concentrations and moduli of the deposited
materials. As an indication of the texture, the size of the pores and the particles or
aggregates of particles, from which the materials were built up, is given. In many cases the
mineralogical composition is also given.

4.1 Exhaust Fan Deposits
In the case of kilns with pressure filter systems, in which unfiltered dust-laden gas passes
through the exhaust gas fan, deposit formation causes problems. These arise when the
deposit falls off one blade, and brings the rotating fan out of balance. Deposits of up to 3.5
kg/blade are known to occur.
4.1.1 Characteristic properties:
Exhaust fan deposits, composed of the finest raw meal particles are usually red-brown, hard
and quite brittle. They exhibit a compact layered structure and have a very low porosity of
8%.
Their chemical and mineralogical composition is basically that of the raw meal but often the
plate -shaped clay particles are preferentially deposited parallel to the blade surfaces. Due
to their long stay in the system, fan deposits are enriched in the volatile components K2O,
Na2O, SO3. Typical values include the following:
LSF 20 - 100
SR 1.0 - 1.5
K2O + Na2O 2.1 -3.0 %
SO3 4.7 - 6.0 %
The SO3 is usually present as anhydrite (CaSO4).
Binding mechanism:
In this case, the temperature is such that liquid phase involvement - aqueous or molten salts
- can be ruled out.
The dust particles, because of the fan rotation, strike the blade surfaces with a high velocity
and are so compacted. As the texture of the surface, after even a short time in operation,
possesses undulations in the order 0.5 - 20 um, the smaller dust particles can be
mechanically "locked-on". Subsequent development of the deposit follows by an identical
mechanism.
4.2 Slurry Rings (including mud balls)

Characteristic properties:
These occur in long wet kilns and are composed of the partially dried kiln charge somewhat
enriched in alkalis and SO3. They are soft and can usually be broken - and hence prevented by -
heavier chains. The H2O content lies between 20 and 30%, a range in which clay materials
exhibit a sticky, plastic consistency. The content of the alkalis which greatly increases the
tendency to adhesion (influence on rheological properties) can be up to 10% K 2O + Na2O, and
about the same level of SO3. In many cases, balls form (in addition) on the chain links by the
same mechanism. A typical example of a mud ball is plant I in Table 4.
Binding mechanism:
The binding mechanism is the well known ability of clays to form a sticky, plastic mass when
containing the correct quantity of H2O, and to harden on the further water loss. To this
mechanism must also be added the crystallization of K2SO4 solution and the further
strengthening of the structure by formation of CaSO4. Photo 1a gives an example of such a
ring.
4.3 Cyclone and Grate Preheater Deposits

Characteristics:

These deposits form on the roofs, walls, outlet and riser pipes of cyclone preheaters, in the
hot chamber of grate preheaters, and vary considerably in appearance and homogeneity. In
general, they have a light colour varying from cream to brown to pink, indicating that the
component particles had not been heated higher than 1200°C. In some cases, darker zones
of harder burnt material can be observed. Depending on their place of deposition, they range
from a dense, compact, definitely layered structure, hard to break to a porous (30%) material
with only moderate strength with less obvious layering. The former type is typical of cyclone
cones and discharge pipes while the latter is to be found in the transition and swirl
chambers. Soft deposits can, however, also be found in the cyclones.
From a chemical viewpoint, this deposit type usually is characterized by a concentration of
the volatile elements in the following range:
K2O = 1 - 30%
SO3 = 1 - 35%
Cl = 1 - 25%
Na2O = 0 - 2%
In some cases, therefore, deposits can occur with no appreciable increase in concentration.
Typical analyses found for deposits are given in Table 3 and Appendix I.
The mineralogical composition of preheater deposits differs as would be expected from that
of the raw meal in that the clays are essentially decomposed, and a reaction to form
intermediate minerals has taken place. Minerals containing only the volatile elements can
also be found.

Amongst the minerals found in preheater deposits are the following:
raw meal:
calcite CaCO3
quartz SiO2
normal intermediate phases:
free lime CaOf
periclase MgO
mayenite C12A7
belite C2S
typical deposit phases with low melting point:
sylvite KCl
halite NaCl
langheinite 2CaSO4 . K2SO4
arcanite K2SO4
typical deposit phases without melt involvement:
carbonate spurrite 2C2S . CaCO3
sulfate spurrite 2C2S . CaSO4
anhydrite CaS04
Formation mechanism:
The binding substance in this deposit type is the low melting point Na2O, K2O, SO3, Cl based
compounds. These are molten in the kiln gas and are deposited on the cyclone walls and pipes,
or first on dust particles which then themselves are deposited out of the gas stream in these
areas. Cooling on contact or with increasing thickness results in an appreciable strengthening of
the originally sticky deposit. Because of the extensive duration of stay in the kiln system, a
reaction takes place with gaseous CO2 and SO3, resulting in the formation of lath shaped
spurrite and sulfate spurrite which additionally strengthen the texture. Typical textures for
unsintered, pre-heater and kiln inlet deposits can be seen in photos lb - ld.
4.4 Meal Ring (Calcining Ring) in Long Kilns

Characteristics and formation:
The meal rings, often called "calcining rings" in long kilns, are in their properties and
mechanism of formation very similar to those of pre -heater deposits in heat exchanger
kilns. This is perhaps not surprising in that both build-ups occur in the same temperature
zone. Meal rings are mostly less troublesome than preheater deposits because often, due
to their relatively poor strength, thermal fluctuations, kiln deformation and the action of the
material stream, they fall off periodically under their own weight. A typical example of a
calcining ring is plant J in Table 4.

4.5 Middle Rings in Large Pre-heater Kilns
Characteristics:
Unlike meal rings, middle "rings" are dense (fine grained) of low porosity, very hard and seldom
fall off during operations. Although termed as a ring they are rather more elongated, like a band,
being often some 15 - 20 m long extending from 7 to 11 diameters from the outlet, e.g. 35 - 55
m for a 5 m ∅ kiln. Unlike previous types, this deposit is clinker like in colour indicating it being
composed of well burnt kiln charge. Perpendicular to the direction of deposition, the fine layered
structure can be seen showing the curvature of the kiln shell.
The chemical composition of middle rings is very similar to that of clinker. This is surprising
because considering the long duration of the stay in the kiln, no increase in concentration of
the alkalis or SO3 takes place, and often the ring shows lower volatile element values than
for clinker. Typical analyses of a middle ring are given in Table 5.
The minerals found in middle rings are the clinker minerals elite, belite, aluminate, ferrite and
free CaO, the elite having often decomposed into microscopically mixed belite and free CaO,
resulting from the temperature at the site of the ring being under the lower stability
temperature of elite (i.e. 1260°C).
Formation Mechanism:
The mechanism of bonding is the freezing of the clinker alumino-ferrite melt. Due to a long
cool flame, the clinker has a tendency to be fine, and the smallest clinker particles of 150 -
450 um are carried back by the flame and deposited onto the kiln wall in a zone where
temperatures of below 1250°C exist. The particles immediately freeze in place, and because
the kiln charge is still fine, it does not possess sufficient abrasive action to remove the
growing ring. The typical compact structure of a middle ring can be seen in photo 1e.
4.6 Sinter Rings (excluding coal-ash rings)

Characteristics:
These occur at the beginning of the sinter zone some 4 - 5 D from the kiln outlet. They are
greyish-black in colour, strong and are (usually) agglomerations of small clinker pellets and
clinker dust. No layer structure is obvious because of the presence of large pores and voids.
In general, the chemical composition is that of the clinker with no appreciable concentration
of volatile elements.
From a mineralogical viewpoint, the normal clinker minerals elite, belite, aluminate, ferrite
and free CaO are observed, with reactions to form belite and CaO free, spurrite and belite being
found with increasing depth in the ring, i.e. decreasing temperature, similar to the case
of middle rings.
Bonding mechanism:
The bonding is created by the freezing of the alumino ferrite clinker liquid in the case of pure
sinter rings. This phenomenon occurs especially at the beginning of the sinter zone, where
the liquid phase is just starting to form (approx. 1280°C). Due to the rotation of the kiln, the
charge in this zone freezes with each kiln revolution: a new wet layer sticks on, and with time
a thick deposit builds up consisting of particles of less than 1mm.

4.7 Coal Ash Sinter Rings
Characteristics:
In kilns fired with a high ash content coal, sinter/coal ash rings can form at 7 - 8.5 D from the
kiln outlet. They are dense, often layered and sometimes glassy in appearance and built up
from particles some 150 250 um in size. They are rather less dense and have larger pores
and voids than middle rings. Photo 1f gives an example of the microstructure of such
material, showing the coal ash layers.
From the viewpoint of their chemical and mineralogical composition they are essentially
similar to clinker, exhibiting the minerals elite, belite, aluminate, ferrite and free CaO. With
decreasing temperature (increasing ring depth) reactions to form spurrite and calcite take
place, and also the transformation of alite ⇒ belite + CaOf and β belite ⇒ α belite. Details
of the chemistry and mineralogy are given in Table 6 . No enrichment of the volatile phases
can be observed. Because of the enrichment in coal ash, the belite content is higher than
that of the clinker, and tends to be found in layers.
Formation mechanism:
The bonding medium here is the sticky molten coal ash particles and perhaps to a slight
extent, the alumino ferrite clinker liquid phase occurring by a mechanism such as in Fig. 2
showing the typical build up during kiln rotation.
Fig. 2: Mechanism for ring formation
The molten coal ash droplets adhere to the exposed kiln lining at a point and temperature at
which they are still partially fluid and sticky. When this sticky layer passes under the kiln
charge on each rotation, it is assumed that a single layer of the still very fine kiln charge
adheres to it. Because of the presence of fine crystalline elite and the overall occurrence of
liquid phase, it must be assumed that the material temperature at the position of the ring lies
above 1260°C.
The elite crystals are very small and certainly much smaller than those of the clinker.
Because of this, it can definitely be said that the ring is not formed from clinker dust blown
back down the kiln.
4.8 Clinker Rings / Cooler Inlet Deposit (snowman)

Such rings and deposits are formed from normal size clinker granules and have a high
porosity containing many voids. They are usually not troublesome to kiln operations as they
can easily be removed. Their composition and mineralogy is identical to clinker, but in some
cases, rings of up to 3.5% K2O and 3.0% SO3 have been observed.
The mechanism of bonding is the freezing of the clinker liquid phase as the clinker passes
through the cooling zone (ring) or on falling down the chute into a grate cooler, grate kilns
being usually operated so as to have no cooling zone within the kiln itself.
4.9 Kiln Charge Balls

Kiln balls occur in cases where a tendency to meal or sinter ring already exists and can be
up to 1m in diameter. The chemical composition is, thus, an important factor. They are
usually found upstream of meal or sinter rings. They are usually made up of already calcined
material and can have a porosity of up to 55%, consisting of many fine pores. Often they
consist of a hard dense porous core, surrounded by the majority of porous material. The
core usually is a piece of coating from say the lower heat exchangers or the transition
chamber, and often has a come position different from the kiln charge in the area of
formation. Differences in composition can be seen in Table 7.
The mechanism of meal ball formation can be due to either, or a combination of the following
factors:
0 stripping and subsequent "balling" of old, excess coating
1 agglomeration enhanced by available clinker and/or salt melt
2 ring section acts as a dam, retaining "pieces" of material for long periods. Radial growth
of the pieces occurs by compaction and adherence of fresh surface due to continual
rolling action of the pieces/balls over the charge.
In most cases, no liquid phase participation in sufficient quantities is possible so that the
balls behave like a snowball and by their own pressure material sticks to the surface. This
mechanism is similar to that of deposition on the exhaust fan blades.

0 METHODS OF REMOVAL/ELIMINATION
An important prerequisite for minimizing the tendency to form objectionable coatings and
rings, is stable kiln operation. This applies to the composition, fineness and feed rate of the
raw material and fuel, and burning zone heat control.
The tendency to form coatings in the kiln is reduced by lowering the dust load of the kiln gas.
Objectionable coatings and rings which are formed as a consequence of high concentrations
of various circulating elements can be obviated by appropriate reduction of the cycles in
question.
This can be achieved by:
0 Employing different raw materials and/or fuel with lower concentration of the offending
element. This is generally not practicable.
1 Control of the raw meal milling so as to reduce the concentration of the very fine particles
of sizes under 20 µm.
2 Intervention into cyclic process by either discarding dust in which the circulating
elements have become concentrated, or by means of a bypass installation which
extracts a portion of the kiln gas.
The penetration of false air into the preheater and kiln inlet chambers should be avoided, as
such cold areas will act as sites for preferential build-up.
In order to reduce the tendency to form sinter rings, it is in the first place necessary to
reduce the proportion of fusible matter in the clinker, i.e. the lime standard and silica
modulus should be increased. "Coating -inactive" bricks have also proved successful
in certain cases, in reducing the tendency toward sinter ring formation.
In coal-fired kilns a coal with a normal ash content should be employed as coals having ash
contents of 40% are characterized by a very high tendency to ring formation. No general
approach can be given to the effectiveness of other measures, e.g. alteration to firing
conditions, as these represent variables which are peculiar to the particular installation.
Clinker rings can be avoided by shifting the flame further back, thus increasing the clinker
temperature at the kiln outlet. As a result of this, however, the "stickiness" range of the
clinker is shifted towards the cooler inlet. Coatings can then be formed on the cooler inlet
chute. This is particularly problematic with satellite coolers. In instances where this occurs
with grate coolers, these coatings can be eliminated by the use of water cooled plates on the
inlet chute.

Table 1 Time and Location of Rings and Coating
(according to Opitz, 1974)

Table 2 General Characteristics of Various Ring Types

Table 3 Properties of Build-ups Examined
Table 4 Properties of Rings Examined

Table 5 : Typical analyses of a middle ring at various depths
sample ring ring ring clinker
internal middle external
Loss on ign. 0.47 1.82 0.94 0.37
SiO2 23.0 17.8 21.9 21.9
Al2O3 4.9 5.1 5.5 5.6
Fe2O3 3.0 2.9 3.1 3.2
CaO 66.4 70.7 66.8 66.1
MgO 1.1 1.0 0.95 0.95
SO3 0.13 0.07 0.27 0.48
K2O 0.16 0.09 0.30 0.35
Na2O 0.06 0.06 0.12 0.14
TiO2 0.25 0.23 0.25 0.25
Mn2O3 0.02 0.02 0.02 0.02
P2O5 0.24 0.21 0.20 0.20
Cl- 0.01 0.01 0.01 0.01
Total 99.74 100.36 100.01 99.57
LS 92 122 96 93
SR 2.91 2.23 2.55 2.49
AR 1.63 1.76 1.77 1.75

Table 6: Chemistry and Mineralogy of a (coal ash) ring
kiln point 29.5 32.5 33.5
depth in ring under hot middle near shell
surface
chemical analysis:
SiO2 24.65 24.61 21.77
Al2O3 5.68 5.70 5.48
Fe2O3 4.23 4.29 3.53
Cao 61.27 61.92 65.31
MgO 0.49 0.79 0.64
K2O 0.58 0.34 0.17
Na2O 0.36 0.03 0.20
SO3 0.15 0.27 0.34
Loss on ign. 2.61 1.63 2.48
Total 100.0 99.68 100.0
LS 78.1 79.0 93.7
SR 2.49 2.46 2.42
AR 1.34 1.33 1.55
free lime 1.0 4.0 2.7
minerals det.:
(wt. %)
alite ~ 30 ~ 15 ~ 40
α and β belite ~ 35 - 40 ~ 55 - 60 ~ 30
liquid phase ~ 15 ~ 15 ~ 15
(aluminate + ferrite)
carbonate ~ 15 10 ~ 15
spurrite+CaOf
λ belite <5 <5
texture:
average pore size 400 - 800 400 - 800 100 - 300
(estimated) µm

Table 7. Chemical comopsition of core and rim of meal ball and the kiln charge
composition at the corresponding zone
kiln ball kiln charge
rim core calcining zone
Loss on ign. 3.9 4.3 1.9
SiO2 20.9 17.2 21.2
Al203 5.6 8.0 5.7
Fe203 2.8 4.4 2.7
Cao 63.0 62.5 62.1
SO3 1.9 1.7 2.2
K2O 2.1 1.5 2.3
Cl- 0.1 0.1 0.1
LSF 94 104 91
SR 2.4 1.4 2.5
Appendix I: Chemical analysis of preheater cyclone deposits

Appendix II: Chemical analysis of riser pipe kiln inlet deposits

Appendix III Chemical analysis of rings from rotary section of Kiln

Phototable 1 S.E.M Micrographs of Deposits

Chapter 5
Fuels for Cement Kilns

Fuels for Cement Kilns
1. INTRODUCTION............................................................................................................162
1.1 Objectives...........................................................................................................163
1.2 Definitions...........................................................................................................163
2. General parameters to characterize fuels..................................................................163
2.1 Chemical Composition........................................................................................163
2.1.1 Sulfur..................................................................................................................164
2.1.2 Nitrogen............................................................................................................. 164
2.2 Heat of Combustion............................................................................................164
2.2.1 Definition of Gross and Net Calorific value.........................................................164
2.2.2 Conversion from Gross to Net Calorific Value....................................................165
2.3 CO2 Emissions from Fuel Combustion...............................................................167
3. Characteristics of traditional fuels..............................................................................168
3.1 Solid Fuels (Coal, Petcoke).................................................................................168
3.1.1 Formation and Classification of Coals................................................................168
3.1.2 Petroleum Coke.................................................................................................170
3.1.3 Analyses Coal and Coke....................................................................................171
3.1.4 Reporting Coal or Coke Analyses and Conversion to Different Bases................173
3.1.5 Impact of Solid Fuel Characteristics on the Clinker Burning Process.................175
3.2 FUEL OIL............................................................................................................ 177
3.2.1 Compositions and Classification.........................................................................177
3.2.2 Density............................................................................................................... 178
3.2.3 Viscosity............................................................................................................. 178
3.2.4 Pour Point.......................................................................................................... 180
3.2.5 Water Content....................................................................................................180
3.2.6 Other Properties.................................................................................................180
3.3 NATURAL GAS...................................................................................................181

MT 04/13701/E
Henry Salazar April 19, 2004
0 INTRODUCTION
Summary
According to their physical state, fuels are classified into 3 main groups:
0 Solid
1 Liquid
2 Gaseous
The International Panel for Climate Changes (IPCC) and in the Guide for National
Greenhouse Gas Inventories defined the traditional fuels for the cement industry
as follows:
0 Solid: Coal, petcoke, lignite, and lower grades of the same fuel
1 Liquid: Fuel oils (heavy), fuel oils (light, incl. diesel), other petroleum (incl.
petrol)
2 Gas: Natural gas
All fuels that are not traditional are called alternative fuels.
All traditional fuels consist mainly of carbon and hydrogen accompanied by
minor amounts of nitrogen, oxygen, sulfur and sometimes inorganic minerals.
The carbon and hydrogen contents of the fuels determine their most important
property, namely their heat of combustion or calorific value.
The calorific value describes how much energy or heat is released during
combustion. There are two values:
0 Gross calorific value (which includes the heat from water condensation)
1 Net calorific (which excludes the heat from water condensation) is
normally used in the cement industry.
There are three general types of coal and coke analysis:
0 Chemical analysis
1 Physical analysis
2 Microscopic analysis
For practical purposes the chemical composition of coal and coke is given in terms
of its "proximate" and "ultimate" analysis or elemental composition.
The ash content of solid fuels is important for the chemical composition of the
clinker, since the fuel ash is incorporated into the clinker and changes its chemical
composition.
The volatile mater helps to ignite the fuel and has a great impact on the fineness.
A certain fineness is required to guarantee a good combustion.
The grindability, expressed in the Hardgrove index, gives information about how
difficult it is to grind the fuel. The higher the Hardgrove index value, the easier it is
to grind the fuel.

Grinding is generally combined with drying of the solid fuel. Drying should not go
too far because completely dry coal could be difficult to ignite. 1 - 1.5 % residual
moisture should be sufficient for good ignition.
The viscosity of the liquid fuel is important for the transport and atomization
characteristics. For transport purposes, the fuel has to have a viscosity of less
than 600 - 800 cSt. For atomization, the viscosity has to be in the range of 12
- 125 cSt. Therefore, most liquid fuels have to be heated, since the viscosity
decreases with increasing temperature.
6.1 Objectives
The objective of this report is to present the most important and relevant fuel
parameters, to be able to understand the basic characteristics of traditional fuels in
the cement industry.
6.2 Definitions
By common definition, a fuel is any material that liberates heat when it reacts with
oxygen. Practically speaking, however, only those materials that:
0 ignite at moderate temperatures,
1 burn with comparative rapidity, and
2 are available in quantity at reasonable prices
are regarded as fuels.
According to their physical state, fuels are classified into 3 main groups:
0 Solid
1 Liquid
2 Gaseous
The International Panel for Climate Changes (IPCC) and in the Guide for National
Greenhouse Gas Inventories defined the traditional fuels for the cement industry
as follows:
0 Solid: Coal, petcoke, lignite, and lower grades of the same fuel
1 Liquid: Fuel oils (heavy), fuel oils (light, incl. diesel), other petroleum (incl. petrol)
2 Gas: Natural gas
All fuels that are not traditional are called alternative fuels.
The ASTM standard D 121 (standard definitions of terms relating to coal and coke)
defines the technical terms that are used in relation with coal and coke.
0 GENERAL PARAMETERS TO CHARACTERIZE FUELS
7.1 Chemical Composition

All ordinary fuels consist mainly of carbon and hydrogen accompanied by minor
amounts of nitrogen, oxygen, sulfur and sometimes inorganic minerals. For
example, Table 1 shows the analyses of the most used fuels on a dry, mineral-free
basis and without inert components like N2 and CO2.

Table 1: Ultimate analyses of fuels (weight percent)
Bituminous Fuel oil Natural gas
coal
C 88 - 90 83 - 87 75 - 77
H 4-5 11 - 13 23 - 25
O 3-4 0.1-0.4 Traces
S 1-2 1-3 Traces
Appreciable amounts of minerals, which form ashes after combustion, occur only
in solid fuels and will be discussed in chapter 3.1.5.3.
7.1.1 Sulfur
Sulfur, although a minor component of fuel, is of particular importance because of
its effects on clinker quality and the environment. Furthermore, owing to its
volatility, it can accumulate in certain parts of the kiln system and form build-ups
and rings. Therefore, it is absolutely necessary to closely supervise the sulfur
contents of both raw mix and fuel.
Due to the strong alkaline nature of the sintering zone, calcination zone and lower
stage of the preheater, the sulfur in the fuel fed to preheater kilns will not lead to
significant SO2 emissions. The sulfur will be captured in the clinker. In long kilns
the contact between SO2 and the alkaline material is not so good and sulfur in the
fuel can lead to significant emissions.
7.1.2 Nitrogen
Nitrogen in the fuel either combines with other nitrogen atoms to form N 2 gas or
reacts with oxygen to form " fuel NO"x. In a precalciner kiln the prevailing
o
temperature is in the range of 850 - 950 C, which is not high enough to form
significant "thermal NOx" but fuel NOx will occur. Similarly, other types of
secondary firing of fuel in the back end of the kiln system, such as in the kiln riser
pipe of a suspension preheater kiln or calcining chamber of a preheater grate, may
give rise to "fuel NOx".
There are two main sources for production of NOx:
0 Fuel NOx: nitrogen-containing compounds, chemically bound in the fuel, react with
oxygen in the air to form various oxides of nitrogen. In other words fuel NOx is
generated by the combustion of the nitrogen present in the fuel.
1 Thermal NOx: part of the nitrogen in the combustion air reacts with oxygen to form
0
various oxides of nitrogen. Thermal NOx forms at temperatures above 1200 C and
is mainly produced in the burning zone.
7.2 Heat of Combustion
7.2.1 Definition of Gross and Net Calorific value

The carbon and hydrogen contents of the fuels determine their most important
property, namely their heat of combustion or calorific value. Chemically speaking
combustion is nothing else but oxidation. The elements of the fuel, mainly carbon
and hydrogen, combined with oxygen from the air and a certain amount of energy
are set free as heat. The heat is measured in:
0 Joule (J) (International Standard Unit)
1 Calorie (cal) (no longer valid but still widely used)
2 British Thermal Unit (BTU) (United Kingdom, USA, etc.)

The conversion factors are:
0 1 cal = 4.187 J
1 1 BTU = 1055 J
For the requirements of the cement industry, a larger unit is more suitable, i.e. the
Mega-Joule (MJ).
6
0 1 MJ = 10 J
The heat generated per unit weight of fuel is called specific "heat of combustion" or
"calorific value". For example, 1 kg of carbon will combine with 2.67 kg of oxygen
to form 3.67 kg of CO2, releasing 33.9 MJ of heat:
0 1kg C + 2.67kg O2 → 3.67kg CO2 + 33.9 MJ
Accordingly, the calorific value of carbon is 33.9 MJ/kg.
The calorific value is normally calculated for complete combustion to the highest
possible state of oxidation. If sufficient oxygen is not present, toxic CO will be
formed with much less heat development:
0 1kg C + 1.33 kg O2 → 2.33kg CO + 10.0 MJ
This means that the emission of carbon monoxide is not only a pollution of the
atmosphere but also a waste of energy.
For hydrogen, one obtains 119.9 MJ for the formation of water vapor:
0 1kg H2 + 8kg O2 → 9kg H2O + 119.9 MJ
This is called the 'net calorific value'. By condensation of the vapor, another 20.3
MJ would be released, bringing the total heat development of the reaction to 140.2
MJ. This is the 'gross calorific value'.
Of course, for cement kilns only the net values are important because the fuel
gases leave the kiln far above 100°C and, therefore, the heat of condensation of
the water vapor cannot be utilized. Nevertheless, in some countries, including the
USA, the gross calorific value is generally used.
The contribution of sulfur to the calorific value of fuels is very small:
0 1kg S + 1 kg O2 → 2 kg SO2 + 9.3 MJ
7.2.2 Conversion from Gross to Net Calorific Value

The gross calorific value is usually determined with the help of a bomb-calorimeter
(e.g. ASTM 2015, DIN 51900).
The net calorific value is then calculated according to the following procedures:
According to DIN 51900:
0 Hu = Ho - 24.41 x F [J/g]
Hu = net calorific value [J/g
Ho = gross calorific value [J/g]
F = water determined by elementary analysis (ultimate analysis)
According to ASTM:
0 Q net = Q gross - 5.72 (H x 9) [cal/g]
Q net = net calorific value [cal/g]
Q gross = gross calorific value
H = total hydrogen from the ultimate analysis [%]

Other formulas found in the literature:
N = G - (WB + WM) HC
WM = moisture content of the fuel [kg]
WB = water generated by combustion [kg]
HC = heat of condensation of water vapor, which is 2.45 MJ/kg (at 20 o C).
Normally, cement plant laboratories are not equipped for the hydrogen
determination (ultimate or elementary analysis). A possible way to accurately
calculate the net calorific value is by establishing an empirical formula.
At HGRS, the author established such a formula with the help of the fuel analyses
carried out during the years as follows:
0 H net = H gross - Y [J/g]
0 = 9.28 x % volatile + 588 [J/g]
H net = net calorific value [J/g
H gross = gross calorific value [J/g]
This empirical formula is applicable for coal, petcoke and their mixes. It can also
be applied for other fuels, but a higher error (up to 4 % e.g. for oil) should be
expected.
At the Untervaz plant in Switzerland, a similar formula was established when coal
was re-introduced during 1980 and has been used since.
The empirical formula applied at Untervaz is:
0 H net = H gross - Y [J/g]
0 = 14.5 x % volatile + 642 [J/g]
H net = net calorific value [J/g
H gross = gross calorific value [J/g]
Besides MJ/kg, other units for calorific values are still in use such as:
1 J/g (or kJ/kg)
2 Cal/g (or kcal/kg)
3 BTU/lb (British Thermal Units per pound)
For gaseous fuels, the calorific values usually refer to units of volume under
3
standard conditions (0°C, 760 mm Hg) instead of mass (Nm = cubic meter or cf =
3
cubic foot. 1 Nm = 35.31 cf). The conversion factors of these units are listed in
Table 2.

Table 2: Conversion factors for calorific values
a) Units for solid and liquid fuels
J/g cal/g BTU/lb
J/g 1 0.2388 0.4299
Cal/g 4.187 1 1.800
BTU/lb 2.326 0.5555 1
b) Units for gaseous fuels
kJ/Nm3 kcal/Nm3 BTU/cf
3 1 0.2388 0.0268
kJ/Nm
kcal/Nm3 4.187 1 0.1122
BTU/cf 37.25 8.896 1
The typical calorific values of the main fuels employed in the cement industry are
shown in Table 3.
Table 3: Net calorific values of various fuels

a) Solids + liquids MJ/kg Cal/g BTU/lb
Bituminous coal (dry, 33 - 35 7800 - 8400 14100 - 15200
mineral-free)
Heavy fuel oil 40 - 41 9600 - 9900 17300 - 17800
b) Gaseous MJ/Nm3 kcal/Nm3 BTU/cf

Natural gas 34 - 38 8000 - 9000 900 - 1000
From the foregoing, it can be inferred that the calorific value of a fuel may be
calculated from its elementary composition. However, the energy to break up the
chemical compounds of the fuel has to be deducted. In general, this is not known
with sufficient accuracy. Therefore, the direct determination of the heat of
combustion by calorimetric methods is to be preferred. It is simple and quick and
requires no special skill if a modern calorimeter is available. Fuel prices depend
directly on calorific values. Frequent checks of the fuel caloric value are strongly
recommended. Owing to the tremendous heat requirements of a cement kiln, even
small variations of the calorific values may become very costly.
7.3 CO2 Emissions from Fuel Combustion

As mentioned in the previous chapter, the amount of heat generated during fuel
combustion depends on the amounts of carbon, hydrogen and oxygen present in
the fuel.
Hence, the emissions of CO2 from fuel combustion primarily depend on the ratio
between the carbon content and the calorific value of the fuel. The CO2 emissions
resulting from the combustion of the carbon of the fuel are directly proportional to
the specific heat required for clinker burning as well as the ratio of carbon content
to the calorific value of the fuel. Taking into account only the influence of the fuel,
we can say that a high ratio carbon to calorific value means high CO2 emissions.
Natural gas emits approximately half of the amount of CO2 as does coal for the
same heat output (see default CO2 emission factors for fuels).

Table 4 Default CO2 emission factors for fuels
Type Category IPCC default CSI default kg Comments
kg CO2/GJ CO2/GJ
Traditional fossil fuels
1 Coal + anthrazite 96 IPCC defaults are: 94.6 for coking coal

and other bituminous coal, 96,1 for sub-
+ waste coal bituminous coal, and 98.4 for anthracite
2 Petrol coke 92.8

3 (Ultra) heavy fuel 77.4
4 Diesel oil 74.1
5 Natural gas 56.1
6 Shale + lignite 107
7 Gasoline 69.2
Alternative fossil fuels
8 Waste oil 80 Best estimate; water content can have
relevant influence
9 Tyres 85 Best estimate
10 Plastics 75 Best estimate
11 Solvents 75 Best estimate
12 Prepared solid substitute fuels 80 Best estimate
13 Other fossil based wastes 80 Best estimate
0 CHARACTERISTICS OF TRADITIONAL FUELS
01 Solid Fuels (Coal, Petcoke)
8.1.1 Formation and Classification of Coals

The Collier's Encyclopedia defines coal as a "combustible, sedimentary rock
formed of vegetable matter by physical and chemical alteration through geologic
processes". The first intermediate product of these processes is peat, which is not
yet considered as coal. Then follows lignite (or brown coal), several types of
bituminous coal, and finally anthracite. During this transformation, the carbon
content and with it the calorific value increase continuously, while the volatile
matter (mainly hydrocarbons) decreases. Thus coals are classified according to
their contents of volatile matter and/or their calorific value.
Fig. 1: Coal formation

Table 5: Coalification Sequence
Peat Brown Coal Sub-bituminous Bituminous Anthracite

coal coal
% H2O 75 - 80 50 - 70 25 - 30 5 - 10 2- 5
%C 50 - 60 60 - 75 75 - 80 80 - 90 90 - 95
%H 5-6 5-6 5-6 4-5 2-3
%O 35 - 40 20 - 30 15 - 20 10 - 15 2-3
% Volatile 60 - 65 45 - 55 40 - 45 20 - 40 5- 7
C.V. kJ/kg
Dry- basis 25 25 - 30 28 - 32 30 - 35 35 - 38
Net wet basis 5 5 - 15 20 - 27 24 - 33 35 - 38
8.1.2 Petroleum Coke

Petroleum coke (petcoke) is a carbonaceous solid residual by-product of the oil
refining coking process. There are three different types of petcoke coking
processes:
0 Fluid coking (very hard small spheres with a Hardgrove of e.g. 20 - 40) its physical
structure makes it difficult to calcine and grind.
1 Needle coking (premium coke with low sulfur content) specially used in the steel
industry.
2 Delayed coking is the most common and accounts for approximately 90 % of all
petcoke produced.
Petcoke has a high calorific value, but low volatile matter content and normally
higher levels of sulfur and nitrogen than coal, fuel oil and natural gas. This makes
it necessary to take special precautions when changing from other fuel types to
petcoke. The low volatile matter content and high sulfur and nitrogen content may
cause environmental or operational problems like:
0 High SO2, NOx and CO emissions
1 Unstable kiln operation due to build-ups and cyclone blockages.

Origin Carbon Hydrogen Sulfur C.V. gross C.V. net
[% C] [% H] [% S] [kJ/kg] [kJ/kg]
Holcim Trading (E) 89.2 3.5 3.63 34’951 34’188

Holcim Trading (E) 86.6 3.2 4.74 33’921 33’223
Jerez 89.0 3.1 4.86 34’155 33’479
Obourg 89.5 3.2 4.51 34’742 34’044
Obourg 87.7 2.5 5.20 33’352 32’807
Altkirch 91.1 2.7 3.25 35’375 34’786
Dannes 89.2 3.0 2.43 34’536 33’877

Lumbres 87.1 3.4 2.47 34’310 33’569
Table 6 Analyses of Petcoke
Figure 2: Petcoke grindability
20
18
16
spec. energy
[kWh/t] 14
12
10
8
6
4
2
0
0 20 40 60 80 100
Ha rdgrove Inde x
8.1.3 Analyses Coal and Coke

There are three general types of coal and coke analysis:
0 Chemical analysis
1 Physical analysis
2 Microscopic analysis
8.1.3.1 Chemical Analysis

For practical purposes, the chemical composition of coal and coke is given in
terms of its "proximate" analysis and by its "ultimate" or elemental composition.
0 Proximate analysis:
Covers the determination of moisture, volatile matter, fixed carbon, and ash by
determining the weight losses at certain temperatures. For example the ASTM-
Standard D 3172 prescribes:
0 for moisture, 107°C in air
1 for volatile matter, 950°C with exclusion of air (covered crucibles)
2 for fixed carbon, 750°C in air
3 the residue in ash.
The proximate analysis is simple and fairly quick. Together with the calorific value
and sulfur content, it is sufficient for all practical purposes.
0 Ultimate analysis
The ultimate analysis determines the elementary composition. It requires
special equipment and can only be carried out in laboratories with appropriately
trained personnel.
0 Sulfur
Sulfur is analyzed by ignition and subsequent determination of SO2 in the
combustion gases by infrared absorption or colorimetry.
0 Chlorine
Chlorine can be determined in the liquid from the calorimetric bomb by titration
with silver nitrate.
0 Ash composition
The ash composition, which is important for a correct calculation of the raw mix
(see chapter 3.5), is analyzed by the same methods as other raw materials.
0 Heavy metals
Fuels will always contain metals; especially in petcoke, appreciable amounts of
vanadium and nickel are possible (e.g. V: 1500 ppm; Ni 550 ppm; Cr: 7 ppm).
For this reason prior to using a petcoke from a new source, the heavy metals
content should be assessed.
8.1.3.2 Physical Analysis
The test carried out in this analysis include surface area determination, mercury
and helium density and porosity, shrinkage and hardness. For the cement industry,
only the hardness is of importance. The grindability of coal is usually determined
according to the Hardgrove test.

Fig. 3: Hardgrove test
Hardgrove (coal hardness)
Typical values:
Coal: Hi: 40 to 60 (max 90) (bad to good)
Petcoke: Hi: 35 – 45 (max 90)
(Clinker: Hi: around 30)
8.1.3.3 Microscopy Analysis of Coal

The microscopic analysis serves to identify the coal type of a deposit, its formation
and origin. The microscopy distinguishes the basic particles of the coal.
Microscopic analyses of coal are normally not carried out in relation to clinker
burning.
8.1.4 Reporting Coal or Coke Analyses and Conversion to Different Bases
8.1.4.1 Definitions
When reporting coal or coke analyses, the bases of the analyses must be given
together with the analysis results.
The bases most commonly used in the cement industry are:
0 As-determined basis
1 As-received basis
2 Dry basis
3 Dry and ash-free basis
The ASTM D3180 (standard for calculating coal and coke analyses from
as-determined to different bases) defines these bases as follows:
As-determined basis (ad)
Analytical data obtained from the analysis sample of coal or coke after conditioning
and preparation in accordance with the ASTM D 2013. As-determined data
represents the numerical values obtained at the particular moisture level in the
samples at the time of analysis. The term air-dry basis was also used before for
this sample basis.
As-received basis (ar)

Analytical data calculated to the moisture condition of the sample as it arrived at
the laboratory and before any processing or conditioning.
Dry basis (d)
Data calculated to theoretical base of no moisture associated with the sample.
Dry, ash-free basis (daf)
Data calculated to theoretical base of no moisture or ash associated with the
sample.
8.1.4.2 Converting of Coal or Coke Analysis to Different Bases
The results can be converted to different bases. The ASTM D3180 defines the
calculation procedure; this procedure is applicable for the following parameters:
0 Calorific value
1 Ash
2 Volatile matter
3 Fixed carbon
4 Sulfur
5 Chlorine
6 Nitrogen
Converting from as-determined basis to the as-received basis
Moisture:
Mar = Mad x + ADL
Other parameters:
Par = Pad x
The symbols used above:

0 A= ash; weight %
1 M= moisture; weight %
2 P= any analysis parameter (see previous list)
• ADL = air-dry loss, weight % of as-received sample (ASTM method D 2013)
An example of a proximate analysis and calorific value reported on three different

bases is shown in the table below:
Table 7: Proximate analysis of coal

Coal analysis Air-dried basis Dry basis As-received basis
Moisture [%] 8.23 - 23.24
Ash [%] 4.46 4.46 3.73
Volatile [%] 40.05 43.64 33.5
Fixed carbon [%] 47.26 51.5 39.53
Total 100.00 100.00 1000.00
Net Calorific value 4540 4947 3797
[Kcal/kg]
(Air-dry loss in accordance with method D 2013 = 16.36%)
8.1.5 Impact of Solid Fuel Characteristics on the Clinker Burning Process
8.1.5.1 Volatile Matter

Not all types of coal are equally suitable for cement kilns. If the volatile contents
are too low, ignition of the coal particles is delayed and the flame becomes too
long. On the other hand, high volatile contents increase the risk of self-ignition and
explosions. For a long time, about 20 % of volatile matter was considered an
acceptable compromise, and preferably bituminous coal of that composition was
used or, if that was not available, a corresponding mixture of coals with higher and
lower volatile contents. In the meantime, engineers have learned to master the risk
of self-ignition and explosions, and brown coal containing 50 % of volatiles or even
more can be employed.
8.1.5.2 Grinding and Drying
Besides a certain content of volatile matter, the fineness of the coal dust is of
prime importance for the generation of a good flame in the kiln. In fact, finer
grinding of the coal as can be seen in Fig. 4 can, to a certain extent, compensate
for a lack of volatiles:
Fig. 4: Required fineness of coal as a function of volatile content (sieve

numbers indicate mesh openings in mm)
Grinding is generally combined with drying of the coal. This drying should not go
too far because completely dry coal could be difficult to ignite. In the absence of
water, the reaction of carbon with oxygen is very sluggish. 1 - 1.5 % of residual
moisture in the coal will accelerate the reaction sufficiently. The exact drying

temperature for a certain coal can be determined in the laboratory by establishing
the desiccation curve, i.e. residual water as a function of the temperature.
8.1.5.3 Ash Content
It has already been mentioned that coal contains mineral components, which leave
solid remainders after combustion, the ash. Bituminous coal, which is generally
used for clinker burning, gives a siliceous ash in the form of molten drops. These
are carried away by the flame gases and finally settle on the clinker being formed.
There is thus a higher SiO2 - concentration on the surface of the kiln charge.
Depending on how far the ash drops are carried into the kiln, they will be more or
less assimilated with the clinker granules. Due to the deposition of ash drops rich
in silica on the kiln charge, there is a local lack of CaO. Here only belite can be
formed and thus the deposition of ash leads to typical belite streaks in the clinker.
Because the coal ash is incorporated into the clinker, it is absolutely necessary to
take it into account when calculating the raw mix. The following example may
serve to demonstrate this. Suppose the raw materials and ash have the
compositions listed in Table 5 a) and the coal required for clinker burning leaves 2
0 of the raw mix weight as ash. To produce a clinker with 55 % C3S, the raw mix
shown in the first column of Table 5 b) would be required.
If the coal ash is not taken into account, the calculation for 55 % C 3S will give the
raw mix of the second column. After incorporation of the coal ash, the C 3S -
content of the clinker would come down to only 39 %.
Coal ash: Impact on raw mix design

0 Example:
* Coal: C.V. net: 26'000 [kJ/kgcoal]
Ash: 15 %
0 1.0 %
1 Kiln heat cons.: 3'500 [kJ/kgclinker]
2 Coal consumption: 3'500/26'000 = 0.13 [kgcoal/kgclinker]

3 As intake: 0.13 x 15 x = 2.0 % (rel. to clinker)
4 SO3 intake: 0.13 x 1.0 x 2.5 = 0.31 % (rel. to clinker)

Table 8: Composition of raw mix components and coal ash
Limestone Clay Sand Coal Ash

L.o.I. 42.2 12.5 16.7 -
SiO2 3.1 52.2 59.4 51.2
Al2O3 0.9 15.6 1.5 26.2
Fe2O3 0.6 13.2 3.0 9.7
CaO 52.6 2.4 17.2 3.0
Table 9: Mix designs to obtain target clinker composition
Mix designed for Mix designed Resulting

ash-free fuel for coal firing clinker
(2.0 % ash)
Limestone 77.7 78.8
Clay 16.9 15.4
Sand 5.4 5.8
LS 91.5 96.2 91.5
SR 2.30 2.32 2.30
AR 1.20 1.19 1.20
C3S 55
8.2 FUEL OIL
8.2.1 Compositions and Classification

Fuel is one of the products of the refining of petroleum. Petroleum or crude oil is a
liquid found in subterranean deposits. Chemically, it is a complex mixture of
hydrocarbons, which range from molecules containing one carbon atom to more
than ninety. In addition, sulfur, nitrogen, oxygen and traces of some metallic
elements appear in varying amounts in the molecules.
By distillation and other processes ("refining"), the crude oil is split up into a large
number of different products, among them several types of fuel oil. These are
classified according to their density, viscosity, and other properties. For cement kiln
firing, only the heavy quality is used, designated as Bunker C or No. 6 Fuel Oil in
the USA. In Germany and Switzerland it is called Heavy Fuel Oil (Schweres Heizöl
= Heizöl S).
Heavy fuel oil consists primarily of the residue from distillation and other refinery
processes, such a thermal cracking, visbreaking, etc., with the addition of various
diluents to meet various specification limits for the viscosity, density and sulfur
content of the finished product.
The chemical and physical properties of heavy residual fuels are dependent both
on the origin of the crude oil and on the processing conditions in the refinery. The
main properties of fuel oil derived from petroleum are subject to certain standards.
For example, Table 6 shows the German Standard DIN 51603.
Table 8: Requirements for heavy fuel oil as per German Standard DIN 51603
Flashpoint in closed cup: min. °C 65;100
Kinematic viscosity: max. 2 at 50°C: 450
mm /s (cSt)
at 100°C: 40
Pour point: °C no specifications
Conradson carbon residue: max. % mass 15
Sulfur content: max. % by mass 2.8
Water content, not removable: max. % mass 0.5
Sediment content: max. % mass 0.5
Calorific value: min. MJ/kg 39.8
Ash (oxide ash): max. % mass 0.15
8.2.2 Density
The density of fuel oil is strongly dependent on temperature. For 1°C increase in
the temperature of heavy oil, the density decreases, on average, by 0.63 %.
Therefore, the density must always be given with a reference temperature, usually
15°C (d15) . There are statistical relationships between density and other
properties of fuel oil, specially the calorific value. Being relatively easy to measure,
density is suitable for regular and frequent quality control.
8.2.3 Viscosity
Apart from the calorific value, viscosity is the most important property of fuel oils.
Both pumping and atomization require certain viscosities. These can be obtained
by heating the oil because the viscosity depends strongly on temperature. Fig. 5
shows this together with the atomization range and pumping limits. It can be seen
that for optimal atomization, the heavy fuel oil must be heated to about 120°C. The
oil temperature has to be strictly controlled, because even small variations would
cause marked changes of viscosity and thereby also atomization.

Fig. 3: Kinematic Viscosity of fuel oil as a function of temperature
The International Standard Unit of viscosity is the Poise (P):
1P = 10−1 N.s
m2
It is called "dynamic viscosity". The ratio of dynamic viscosity to density is
designated as "kinematic viscosity", usually measured in centistokes (cSt). In
practice, empirical units are often employed such as Redwood-seconds in Great
Britain, or Saybolt seconds ("Universal" and "Furol") in the USA. Table 7 contains a
comparison of the different viscosity scales.

Table 9: Comparison of viscosity units
Saybolt seconds Redwood cSt
Universal Furol Seconds
100 15 89 20.6
160 20 145 34.2
300 32 292 65.0
500 52 438 108
800 81 800 173
1500 150 1400 324
3000 300 2800 648
8.2.4 Pour Point

Some fuel oils show an anomaly in their temperature/viscosity behavior: already
more or less above the upper pumping limit, viscosities rise abruptly and the oil
suddenly becomes a stiff gel. This temperature is called the "pour point".
Obviously, it is most important for storage and transport and, depending on the
climate, upper limits should be fixed between buyer and supplier.
8.2.5 Water Content
Heavy fuel oil as delivered to consumers can contain slight traces of water. In
general, the water content is well below 0.5 % mass. The relatively small quantity
of water found in heavy fuel oil is normally present in a finely dispersed form, which
does not cause corrosion or problems in storage, transport or combustion.
A water content higher than 0.5 % mass can be caused by water condensation in
storage tanks that are not completely full or by leaks from pipes, steam or hot
water heating coils. Excess water gradually collects on the floor of the tank
together with other constituents of the oil or contaminants (e.g. sediments, dirt)
and forms sludge. Regular checks and proper care of the tanks are therefore
necessary in order to prevent build-up of sludge on the bottom of the tank.
Care must be taken to ensure that, when oil is removed from the tank, none of the
sludge is allowed to get into the burner circuit, as it may cause blocking of filters
and other problems. In normal circumstances, heavy fuel oil tanks should be
cleaned every two to three years.
Even if some water in the oil is not likely to cause any trouble in kiln operation,
etc., it still means less calorific value and must not be neglected when comparing
fuel prices.
8.2.6 Other Properties

0 Calorific values 40 - 41 MJ/kg (see chapter 2.2)
1 Sulfur contents 1 - 3 % (see chapter 2.1)
2 Ash contents < 0.05 % (negligible for raw mix calculations)
3 Flash point: >100°C (no danger of inadvertent or spontaneous ignition)
4 Carbon residue *) ~12 %

*) As measured by the Conradson test, i.e. heating of an oil sample with the
exclusion of air until a solid residue of oil coke remains, which is expressed in
percent mass of the oil sample. The procedure in this test, however, does not
correspond to the conditions in a normal burner, where heavy fuel oil is first
atomized into very fine droplets and is then burnt in suspension in an excess of air.
8.3 NATURAL GAS

The term "natural gas" is used to designate gases from below the surface of the
earth. They consist of methane and higher paraffins, nitrogen, carbon dioxide,
hydrogen sulfide, and sometimes helium. In a narrower sense, only combustible
mixtures consisting predominantly of methane are called "natural gas".
Many natural gases, especially those from oil deposits, contain some natural
gasoline, mainly paraffins with 5 - 8 carbon atoms, which are liquid at room
temperature under atmospheric pressure. Depending on their gasoline content,
three types of natural gas are distinguished:
3
0 dry gas with < 10 g gasoline per m
3
1 lean gas with 10 - 50 g gasoline per m
3
2 wet gas with > 50 g gasoline per m .
Before feeding the gas into the supply system, it is purified. Especially the evil-
smelling and poisonous hydrogen sulfide, which may amount to 15 % of the crude
gas, has to be eliminated. The valuable gasoline is also taken out and recovered.
Besides sulfur, many other things can cause trouble when employing liquid or solid
fuels, such as grinding, preheating, atomization etc. These problems do not exist
with natural gas. Supervision of the fuel quality is scarcely necessary. Only the
calorific value may vary sometimes, mainly because of varying concentrations of
3
inert compounds like CO2. Abrupt changes of ±300 kJ/Nm have been reported. In
cases of wide variations, the installation of an on-line calorimeter is to be
recommended, which measures and records continuously the heating value of the
gas.
Another disadvantage of gas is the low emissivity of the flame, but owing to the
dusty kiln atmosphere, the heat transfer to the kiln charge is not as bad as might
be apprehended.
CTS/Materials Technology
Henry Salazar

Chapter 6
Clinker Burning Behaviour of Heavy Metals

Clinker Burning Behaviour of Heavy Metals
1. INTRODUCTION ............................................................................................................. 185

2. ESSENTIAL AND TOXIC ELEMENTS ........................................................................... 187
2.1 Essential Elements................................................................................................... 187
2.2 Toxic Elements......................................................................................................... 188
2.2.1 Toxicity .............................................................................................................. 188
3. ANALYSIS OF TOXIC METALS..................................................................................... 191

3.1 Qualitative Analysis of Toxic Elements .................................................................... 191
3.2 Quantitative Analysis of Toxic Metals....................................................................... 193
3.3 Evaporation of Mercury during Handling in Laboratory ............................................ 194
4. SOURCES OF TOXIC METALLIC SUBSTANCES ........................................................ 194

4.1 Common Sources of Toxic Metallic Substances ...................................................... 194
4.2 Natural Cement Raw Materials Components .......................................................... 194
4.3 Secondary Raw Meal Components.......................................................................... 197
4.4 Toxic Elements in Raw Meal .................................................................................... 198
4.5 Natural Fuels ............................................................................................................ 199
4.6 Secondary Fuels ...................................................................................................... 200
5. BEHAVIOUR OF TOXIC METALS IN CEMENT KILNS................................................. 201

5.1 Volatility of Toxic Element Compounds.................................................................... 201
5.2 Distribution of Toxic Heavy Elements....................................................................... 202
5.3 Distribution of Cadmium ........................................................................................... 203
5.4 Distribution of Mercury ............................................................................................. 203
5.5 Distribution of Thallium............................................................................................. 203
5.6 Typical Balance of Toxic Elements In Kilns.............................................................. 204
5.6.1 Wet Kiln............................................................................................................. 204
5.6.2 Lepol Kiln .......................................................................................................... 204
5.7 Class 1 Elements (Cd and Tl) .................................................................................. 206
5.8 Class 2 Elements (As, Co and Ni)............................................................................ 206
5.9 Class 3 Elements (Cr, Cu, Pb, V)............................................................................. 206
5.10 Non-Classified Elements................................................................................... 207
5.11 Maximum Allowed Levels in Stack Gas Dust.................................................... 207

6. TOXIC ELEMENTS IN CLINKER.................................................................................... 207
6.1 Levels of Toxic Elements ......................................................................................... 207
6.2 Influence of Toxic Metals on Cement Properties...................................................... 208
6.3 Ability of Clinker Compounds to Immobilize Toxic Metals........................................ 208
7. INFLUENCE OF KILN TYPE ON HEAVY METALS EMISSIONS .................................. 209

7.1 Experience from Literature and Plants ..................................................................... 209
7.1.1 Non-Volatile Metals ........................................................................................... 209
7.1.2 Volatile Metals................................................................................................... 210
7.1.3 Long Wet and Long Dry Kilns ........................................................................... 210
7.2 Specification of Missing or Inaccurate Data ............................................................. 210
8. STANDARD TECHNOLOGY FOR REDUCTION OF HEAVY METAL EMISSION ....... 211

8.1 Reduction of Heavy Metals Input into Kiln System................................................... 211
8.2 Modification of Process ............................................................................................ 212
8.3 Secondary Reduction Measures .............................................................................. 212
9. BAT (BEST AVAILABLE TECHNOLOGY) FOR REDUCTION OF HEAVY METAL
EMISSION ....................................................................................................................... 213
9.1 Non-volatile metals................................................................................................... 213
9.2 Volatile metals .......................................................................................................... 213
10. LITERATURE........................................................................................................... 214
11. Appendix ................................................................................................................. 216
11.1Thalium Documentation Dyckerhoff-Zementwerke AG, Lengerich ................... 216

0 INTRODUCTION
For hundreds of millions of years the heavy toxic metals remained buried in the ground in
non-soluble forms, mainly as oxides and sulphides, and as such were not a pollutant of the
biosphere. The use of metals by man started about 6000 B.C. when natural metallic gold,
silver and copper supplemented the use of stone, bone and wood. Between 4500 and 3500
B.C. the discovery that metals such as copper could be extracted from native ores, made
malleable by heat and mixed with other metals to form alloys, formed the basis of the vast
metal industries of today. Accompanying this came potential hazards to health.
The early ores were mined from pits without adequate ventilation and resulted in an
exposure to dusts and fumes whose composition was unknown and whose toxicity was on
the whole not even suspected. Strabo, in describing the arsenic mines in Pontus, however
reported:
"the air in the mines is both deadly and hard to endure on account of the grievous odour
of the ore, so the workmen are doomed to a quick death".
From the early beginnings modern technology has made available metals whose toxic
properties were first realised after incidents of ill-health or death caused by inhalation of dust
and fume or by contact with the skin. Such metals include beryllium, selenium, cadmium,
gallium, indium, etc. That cadmium and selenium have toxic properties was first realised
after these metals had already achieved commercial success.
What significance do heavy toxic metals have today in the cement industry? Since the mid
1980's strict regulations for gaseous emissions and immissions have started to be
implemented in highly developed countries. In Europe, especially in Switzerland and
Germany, very low limits have been set for the emission of various elements such as defined
in the Swiss (LRV) and German (TA- Luft) Clean Air Acts. These regulations, given in Table
1, classify the toxic elements, mainly heavy metals, into various classes depending on their
toxicity.

Table 1: The Limits for the Emission of Heavy Toxic Metals as Defined in the
German (TA-Luft) and Swiss (LRV) Clean Air Acts
Element Class Class Emission Limits

3
(mg/Nm )
Cd 1
Hg 1 0.2
Tl 1
As 2
Co 2
Ni 2 1
Se 2
Te 2
Cr 3
Cu 3
Pb 3
Pd 3
Pt 3 5
Rh 3
Sb 3
Sn 3
V 3
Zn n.c. n. defined
3
Due to the very low emission limit of 0.2 mg/Nm for the sum of the Class 1 elements
(cadmium, mercury and thallium) and the high volatility of the elements mercury and
thallium, Class 1 is of utmost importance to the cement industry. That thallium can be
emitted from a cement plant is well documented (Figure 3) following the "incident" in 1979 at
Dyckerhof'f's Lengerich plant in which flora and fauna around the plant were shown to have
suffered as a result of its emission. The source of this thallium was mainly the iron corrective
used (4.8 ppm) and also the coal (1.2 - 1.8 ppm), both natural materials.
With the now common and ever increasing use of secondary materials, combustible and
non-combustible, increased quantities of the heavy, toxic elements just defined are now
being introduced into cement kilns. Which of these elements are the most common ones
encountered, where do they come from and how do they behave inside the kiln? What
happens to them after the cement is hydrated?

0 ESSENTIAL AND TOXIC ELEMENTS
Whereas some elements are essential, others are decidedly toxic. Which elements fit into
each category?
2.1 Essential Elements

Those elements considered to be essential are given in Table 2.
Table 2: Essential and Toxic Elements (Iyengar and Sansoni, 1976)
Structural Elements: C, Ca, H, N, O, P, S
Electrolyte Elements: Ca, Cl, K, Mg, Na
Trace Elements:
0 Essential
0 biologically important: Co, Cr, Cu, F, Fe, I, Mg, Mn, Mo, Ni, Se, Si. Sn, V, Zn
1 Clinically significant: Co, Cr, Cu, Fe, Mg, Se, Zn
1 Suspected to be Essential: As, Ge, Rb
2 Regularly Found in Tissue: Al, B, Br, Ga, Li, Sc, Sr, Ti
3 Toxic
0 potentially toxic: As, Cd, Hg, Pb, Sb, Se
1 environmental contaminants: Cd, Hg, Pb
2 industrial hazards: Be, Bi, Cr, Mn, Ni, Sb, Th, Te,U, V, W
Others: Ag, Au, Ba, Ce, Cs, Nb, Pt, rare earths, Te, W, Zr
Radioactive Contaminants: Po, Ra, Rn, Th, U, Am, Cm, Np
Not all elements are toxic, indeed many are even essential to life. As can be seen in Table 2,
the elements H, C, N, O, Na, Mg, P, S, Cl, Ca are the major elements necessary, the
additional elements B, F, Si, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Se, Mo, Sn and I being necessary
in only trace quantities. The main body of the essential trace elements is grouped between V
and Zn in the first transition series of the periodic table. With the exception of Mo, Sn and I,
all the essential elements lie within the first four layers of the table. The form in which these
elements is present is also important as some compounds, eg. HCN (prussic acid) contain
essential elements but must definitely be classed as being highly toxic.
Regarding the quantity of the trace elements required this varies from element to element.
As given in Table 3, showing the situation for a 70 kg adult, the daily requirements range
from 0.03 mg for Co up to 15 mg for Fe. These elements are not accumulated within the
body and are excreted and so must be continually replaced. The total burden of the
essential elements is also shown in Table 3 with Fe at 4.1 g (4100 mg) and Zn at 2.3 g
(2300 mg) being the most abundant.

Table 3: Estimated Daily Intake and Human Body Burden of Essential Trace Elements
Element Class* Daily Intake (mg) Burden (mg/70 kg)

Co 2 0.03 1
Ni 2 0.45 10
Se 2 0.06 15
Cr 3 0.06 6
Cu 3 3.2 100
V 3 2.5 30
Fe n.c. 15 4,100
Mn n.c. 5 20
Mo n.c. 0.35 9
Zn n.c. 12 3,200
* Class in German and Swiss Clean Air Acts
2.2 Toxic Elements
2.2.1 Toxicity
Some of the substances which are emitted to the atmosphere during the burning of cement
clinker are annoying and some can be even harmful to the environment. Annoying
substances are those which only affect well-being, whereas harmful substances impair the
biological activity (metabolism, growth) of an organism. In medical terminology the latter are
referred to as "atoxic". While it is usual to speak globally of toxic elements, toxicity itself is by
no means a fixed, well- defined property. Firstly it depends to a decisive extent on the
chemical compound in which the particular element is found. For instance sulphur is very
toxic as dioxide, hydrogen sulphide etc., but many other sulphur compounds are perfectly
harmless e.g. sulphates like gypsum. It is also very important whether a compound is easily
resorbed by the organism or not. Examples of compounds containing well-known "toxic"
elements which on account of their being insoluble cannot be resorbed and are therefore
practically harmless are:
0 arsenic sulphide
1 barium sulphate
2 mercury sulphide
The tolerance limits for toxic elements to be found in official regulations are of course
determined on the basis of the most dangerous compounds of the element concerned.
The toxic efficacy of a substance also depends to a large degree on whether it is absorbed
by inhalation, ingestion or skin exposure. For example, metallic mercury can pass through
the stomach and intestines without serious effects whereas inhaling the vapour can be very
harmful. The toxicity of mercury compounds also depends on their solubility which in turn
determines their absorption and distribution in the organism involved. Thus whilst metallic
mercury and almost all mercury compounds are toxic , mercury sulphide is non-toxic
because of its practical insolubility in the body.

Furthermore, a substance which harms one particular kind of organism can leave others
entirely unaffected or even promote their biological activity. Moreover, the biological effect
depends largely on the dose, as already recognized by the Swiss born physician and
alchemist Paracelsus (1493 - 1541) who established the role of chemistry in medicine:
"the dose alone decides whether a substance is poisonous or not"
"if given in small doses what makes a man ill also cures him"
"miners disease results from inhaling metal vapours and is not a punishment for
sin administered by mountain spirits"
No substance is absolutely harmless as there is always an upper limit to the concentration
which may not be exceeded without serious consequences for the organism. For example,
common salt is a dangerous poison when consumed in too large quantities. On the other
hand, there is no absolutely toxic substance because below a certain minimum
concentration even the strongest poison becomes ineffective.
Figure 1 shows schematically the basic relationship between dose and effect. In order to
assess the toxicity of a substance, all the above factors have to be taken into account.
Figure 1: Troxicity is a Question of the Dose

Toxic Elements
In many cases it has been found that elements which are essential to life can also be toxic
depending on the dose and the form in which the element is present. In general it can be
said that the toxicity of elements roughly parallels their abundance in the Earth's crust and in
sea water. Elements that are present in low levels are more likely to be toxic as shown in
Table 4 in which those metals for which toxic properties have been reported are given. In
this table mainly the toxic metals are included, the toxic non-metals having being omitted.

Table 4: Composition of the Lithosphere indicating Essential and Toxic Elements
Concentration (%) Element
46.6 O
27.7 Si
8.1 Al
5.0 Fe
3.6 Ca
2.8 Na
2.6 K
2.1 Mg
-1 Ti H, Mn, P
10
10-2 F, S, C, Cl, Rb, Sr, Ba, Zr, Cr, V, Zn
-3 Ni, Cu, Li, N, Sn, Co, Pb, Th
10
-4 As, B, Mo, Br, W, U
10 (ppm)
-5 Sb, Bi, Ag
10
-6 Hg, Se
10
-7 Au, Pt, He
10 (ppb)
-8
10 I
essential elements in bold text, toxic elements in italics
Also presented in Table 4 are those elements previously shown in Table 3 which are
essential for life. As can be seen many of the elements reported are considered as being
both essential and toxic. Included in this category are the elements B, Mg, V, Cr, Mn, Co, Ni,
Cu, Zn, Se, Mo and Sn. When, for example, a manganese deficiency occurs in humans this
can be accompanied by anaemia, and bone changes in children. Considering the toxic
effects of manganese these are listed as including metal fume fever, psychological and
neurological disorders, the two latter being contracted after long term exposure to air
containing freshly drilled dust particles rich in manganese ore. When a deficiency of copper
occurs this is accompanied by a severe anaemia. Metallic copper is much less toxic than its
salts and for humans only 1 g of CuSO4 is highly toxic, causing vomiting, diarrhoea,
convulsions and is often fatal.

0 ANALYSIS OF TOXIC METALS
In order for a cement plant to be able to safely use secondary materials containing toxic metals
without the risk of excessive emissions, it is essential that the materials received conform to the
reference samples initially submitted. To do this it is necessary to carry out control analyses for
those elements listed in the applicable Clean Air Act. This is especially the case, because of their
volatility, for the Class 1 elements Hg, Tl and Cd but also for the elements of Classes 2 and 3
should these be present in appreciable concentrations. To avoid the necessity of carrying out a
full quantitative analysis, which would be costly and time consuming and would attempt to
determine elements even when they were not present, it is advisable to first carry out a
qualitative analysis. This is the method used at HMC/TC.
3.1 Qualitative Analysis of Toxic Elements

A qualitative analysis (spectrum) of all the toxic elements with the exception of Be can be
carried out using such methods as X-ray fluorescence (XRF) and inductively coupled plasma
emission (I.C.P.) spectroscopy, provided that the instrument is of the sequential and not of
the multi-channel type. Whereas the I.C.P. method requires the sample being measured to
be in liquid form, both solid and liquid samples can be measured by the XRF, the latter being
possible when the spectrometer is flushed with helium instead of operating under vacuum.
Because of the comfort of the operating software and the general ease of operations the
XRF technique is preferred at HMC/TC.
An example of an XRF spectrum is given in Fig. 2. This was taken using a sample of slightly
contaminated excavated material being considered for disposal by incineration in a Swiss
cement kiln. In this diagramme the wave length of the characteristic radiation (expressed as
two theta in degrees for the measuring crystal) is presented on the x-axis and the intensity of
the radiation (expressed as counts/sec) on the y-axis. As can be seen the presence of Pb
(410 ppm) and Zn (7700 ppm) can easily be determined in this figure. The peak heights
observed are not the same for equal concentrations of different elements. The peak height
for different elements is not proportional to the quantity present. The values for these
elements were found in subsequent quantitative determinations. Attempts to ascertain the
qualitative presence of Cd, Hg and Tl, which were later shown to be present in quantities of
4.4, 0.14 and < 1 ppm, were not successful by this technique and should in any case always
be quantitatively determined.

Figure 2: XRF Specturm of excataed material

3.2 Quantitative Analysis of Toxic Metals
Several methods can be used for the quantitative analysis of toxic metals and include those
shown in Table 5. Of these the most sensitive of the methods that can generally be used is
certainly the AAS technique with a graphite furnace which allows a lower limit of detection
than the flame AAS, XRF and ICP methods. The limits of detection of the cold vapour and
hydride AAS techniques lie within those of the two groups of methods previously mentioned.
With regard to the limits of measurement these are some 4 to 5 times higher than the limits
of detection.
Table 5: Detection Limits (ppm) of Environmentally Sensitive Elements by

Various Measurement Techniques
Element Class AAS AAS AAS AAS ICP XRF
Flame Furnace Cold Hybride
Vapour
Be 1 Car. 1 0.003 0.003
Cd 1 1 0.0002 1.5 1 - 10
Hg 1 200 0.1 0.05 12
Tl 1 30 0.01 27
As 2 100 0.08 1 30
Co 2 5 0.008 3
Ni 2 8 0.05 4 1 - 10
Se 2 100 0.05 1 30
Te 2 70 0.03 0.50 20
Cr 3 3 0.04 3
Cu 3 2 0.005 1
Pb 3 20 0.007 25
Pd 3 20 0.05 13 1 - 10
Pt 3 50 0.20
Rh 3 2 0.01 8
Sb 3 60 0.08 5 40
Sn 3 80 0.03 5 30
V 3 20 0.10 3
Zn n.c. 0.6 0.003 2 1 - 10
In HMC/TC laboratories, where all the above methods are available Hg is measured by the
hydride process and all the others with a few exceptions using the graphite furnace AAS
technique. These methods were chosen because, being wet chemical techniques, there is
no matrix correction and standard calibration samples are easily made by the addition
method.

3.3 Evaporation of Mercury during Handling in Laboratory
It should be pointed out that sampling and sample preparation are most important aspects of
a trace element analysis. Contamination or the loss of a portion an element are an
omnipresent danger. This is actually the reason why there are often doubts about the validity
of values given in literature, particularly at low concentration levels. An evaporation test
demonstrated that varying amounts of mercury can escape during the preparation and
drying of a sample (see Table 6).
Table 6: Hg-Concentrations in Precipitator Dust and Sewage Sludge Samples

after Treatment at Different Temperatures
Temperature (°C) ESP Dust (ppb) Sewage Sludge (ppb)
20 430 6,400
100 90 6,400
200 60 6,300
300 40 700
0 SOURCES OF TOXIC METALLIC SUBSTANCES
4.1 Common Sources of Toxic Metallic Substances

To provide an overview with regard to the normal uses to which the elements of Classes 1, 2
and 3 of the previously mentioned Clean Air Acts are put, Appendix II gathers together some
relevant information. As can be seen all of the elements are extracted from ores in which
they are present as either oxides or sulphides and are thus found in nature in water insoluble
forms. In nature they are often only present in low concentrations but after extraction are
then present in much higher concentrations and may be in a soluble form. The information
given in Appendix II is intended to inform the reader which industries are likely to have by -
products that could contain the listed elements. Although it is very unlikely that any of the
primary products listed in the "uses" table would ever be considered as secondary materials
the same cannot be said for the by- products of these manufacturing industries, which
because of restrictions on dumping materials, are now looking for other means of disposal.
Today some Group plants are in fact disposing of preservatives, pigments, dyes etc. which
enter the kiln via secondary materials such as waste oils or woods.
4.2 Natural Cement Raw Materials Components

As shown earlier in Table 4 the heavy toxic metallic elements make up a small but definite
proportion of the materials in the Earth's crust. As such their introduction into a cement kiln
via the natural raw materials cannot be prevented. This question of interest is therefore how
much of the classified and non-classified elements (at the ppm and ppb levels) can be
present in the various rock types used for raw mix components?
The answer to the above question is given in Table 7 for the main rock types limestone,
limey marl and clays and in Table 8 for the commonly used corrective, iron ore. As can be
seen in Table 7 the limestone and limey marl components tend to have lower contents of
heavy metals than clay components. This does not mean, however, that all limestones can
be assumed to be low in heavy metals. In fact Zn levels of 1900 ppm and V levels of 3000
ppm have been encountered in extreme cases.
For the limey marls reported the values were all low and in the same magnitude or lower
than the average levels in the Earth's crust. The highest level was 55 ppm reported for Zn.

When considering the heavy metals present in iron ore, an often used corrective, as shown
in Table 8 such materials can often contain concentrations of heavy metals appreciably
higher than the average for the crust. Such elements include Tl (400 ppm), As (680 ppm), Cr
(1500 ppm), Pb (8700 ppm) and Zn (9400 ppm). It is fortunate that only small quantities of
iron corrective have to be used in cement kiln raw mixes.
In the case of the clays values very much higher than in the Earth's crust were reported for
Cd (500 ppm), Cr (1500 ppm), V (3600 ppm) and Zn (1300 ppm).

Table 7: Range (ppm) of Heavy Metals Reported for Rock Types Used as Raw
Mix Components
Element Class Limestone Limey Marl Clay Earth's Crust
Cd 1 0.02 - 0.50 0.04 - 0.35 0.02 - 500 0.2
Hg 1 0.005 - 0.10 0.009 - 0.13 0.08
0.02 - 0.15
Tl 1 0.05 - 1.8 0.07 - 0.68 0.20 - 1.6 0.5
As 2 0.1 - 12 3 - 25 1.8
Co 2 25
Ni 2 1.4 - 13 5.9 - 21 11 - 71 75
Se 2 0.05
Te 2 0.01
Cr 3 0.70 - 16 4.6 - 35 20 - 1500 100
Cu 3 55
Pb 3 0.27 - 21 1.3 - 8.5 1.0 - 219 13
Sb 3 0.2
Sn 3 2
V 3 5 - 3000 30 - 3600 135
Zn n.c. 0.10 - 1900 24 - 55 2 - 1300 70
Table 8: Range (ppm) of Heavy Metals Reported for Correctives used in Cement
raw Mixes
Element Class Iron Ore Earth's Crust
Cd 1 0.2 - 15 0.2
Hg 1 0.08
Tl 1 0.07 - 400 0.5
As 2 4 - 680 1.8
Co 2 25
Ni 2 10 - 340 75
Se 2 0.05
Te 2 0.01
Cr 3 9 - 1400 100
Cu 3 55
Pb 3 9 - 8700 13
Sb 3 0.2
Sn 3 2
V 3 135
Zn n.c. 6900 - 9400 70

4.3 Secondary Raw Meal Components
To improve the economics of clinker production or improve cement quality many plants are
today using so called secondary components in their raw mixes. These materials are mainly
waste products from other processes and are often available in large quantities. Examples of
such materials are given in Table 9 with fly ash and foundry sand in a few cases being the
main alumino-siliceous raw mix component. Of these the purest material is probably the man
made FCC catalyst but, depending on the source of the oil being processed in the refinery,
this material can contain up to 1860 ppm of Ni, 1265 ppm of Sb and over 5000 ppm of V. As
can be seen a natural material such as fly ash can contain up to 530 ppm of As, 600 ppm of
Ni, 650 ppm of Cu, 800 ppm of Pb, 500 ppm of V and 910 ppm of Zn. In cases in which coal
is used as the main fuel and fly ash as a major mix component, appreciably levels of heavy
metals may be introduced into the kiln. Foundry sand is a component that is just starting to
be used but as seen in Table 9 may contain appreciable maximum levels of Ni (1520 ppm),
Cu (1500 ppm) and Pb (2300 ppm).
Table 9: Range of Heavy Metals (ppm) in Secondary Components used in

Cement Raw Mixes
Element Class Foundry Sand FCC Catalyst Fly Ash Earth's Crust
Cd 1 300 < 0.6 0.2 - 8 0.2
Hg 1 2 0.04 - 2 0.2 - 0.8 0.08
Tl 1 < 1.5 - 145 2 - 15 0.5
As 2 2 2-6 40 - 530 1.8
Co 2 13 - 93 90 - 250 25
Ni 2 1,520 265- 1860 150 - 600 75
Se 2 1 < 11 - 128 1 - 35 0.05
Te 2 0.01
Cr 3 10 58- 142 60 - 330 100
Cu 3 1,500 6- 93 140 - 650 55
Pb 3 2,300 12- 174 125 - 800 13
Sb 3 <25 - 1265 0.5 - 90 0.2
Sn 3 100 7- 23 (38) 2
V 3 5 1050 - 5515 230 - 500 135
Zn n.c. 130 32- 115 350 - 910 70

4.4 Toxic Elements in Raw Meal
The levels of the heavy toxic elements previously given for the mix components, major,
minor and secondary give the maximum range encountered on a world-wide basis. Of more
importance is the levels of the classified elements that are actually encountered in practice
and these are presented in Table 10. As can be seen the concentrations of many elements
are less than the average for the crust. The only elements which are present in
concentrations above the average for the Earth's crust are limited to Tl (6 ppm) and As (28
ppm). These levels can in no way be considered as high.
Table 10: Heavy Toxic Metal Levels (ppm) in Raw Meal
Element Class Raw Meal Earth's Crust

Cd 1 0.04 - 1 0.2
Hg 1 < 0.02 - 0.6 0.08
Tl 1 0.1 - 6 0.5
As 2 3 - 28 1.8
Co 2 11 25
Ni 2 18 - 30 75
Se 2 < 0.2 0.05
Te 2 0.01
Cr 3 23 - 39 100
Cu 3 17 55
Pb 3 4 - 15 13
Sb 3 0.2
Sn 3 2
V 3 32 - 102 135
Zn n.c. 15 -90 70

4.5 Natural Fuels
Because the normally used fuels coal and oil have been formed over millions of years by
naturally occuring processes in the Earth's crust, it will not now be surprising that these also
contain heavy, toxic elements in trace quantities. The quantities reported by various authors
are given in Table 11.
As can be seen brown coal contains the smallest quantities of classified heavy metals with
only Zn at up to 70 ppm being above the 10 ppm level. For hard coal the maximum
quantities of the major heavy metals present are Pb (270 ppm), Zn (220 ppm), V (180 ppm),
Ni (96 ppm) and Cr (80 ppm). For bunker oil the heavy metals present in the maximum
concentrations are V (117 ppm), Zn (85 ppm) and Ni (43 ppm). For petcoke, a residue of oil
distillation, the levels of the heavy metals are more concentrated with the maximum levls
reported being Ni (355 ppm) and Cr and Pb at both just over 100 ppm.
Table 11: Range of Heavy Metals (ppm ) in Natural Fuels Used in Clinker
Production
Element Class Hard Brown Petroleum Bunker Oil Hard Coal Earth's
Coal Coal Coke Fly Ash Crust
Cd 1 0.1 - 10 0.1 - 2.4 0.1 - 4.0 0.02 - 0.4 0.2 - 8 0.2
Hg 1 0.1 - 1.0 0.01 - 0.09 0.2 - 0.8 0.08
Tl 1 0.1 - 1.2 0.07 - 0.3 0.04 - 3.1 < 0.02 - 0.12 2 -15 0.5
As 2 2 - 50 0.3 - 9 < 0.01 - 0.1 40 -530 1.8
Co 2 5 - 34 90 -250 25
Ni 2 15 - 96 0.6 - 1.9 24 - 355 5 - 43 150 - 600 75
Se 2 0.6 - 3.4 1 -35 0.05
Te 2 0.2 - 0.6 0.01
Cr 3 5 - 80 0.9 - 8 3.6 - 104 2-4 60 -330 100
Cu 3 10 - 60 140 - 650 55
Pb 3 10 - 270 0.8 - 6 6 - 102 1 - 34 125 - 800 13
Sb 3 0.4 - 2.0 0.5 - 90 0.2
Sn 3 (38) 2
V 3 30 - 180 2-7 2 - 117 230 - 500 135
Zn n.c. 6 - 220 1 - 70 (7.6) 5 - 85 350 - 910 70

4.6 Secondary Fuels
Table 12 lists the range of heavy metals used in the most commonly used secondary fuels.
As can be seen, the levels of many elements are below the average for the Earth's crust but,
when these latter values are exceeded, this is by a considerable margin and in elements
specific for the individual fuel concerned.
For example waste automobile oil is enriched in those elements which are introduced by the
wear of bearing materials. Thus elevated values of Pb (21'000 ppm) and Zn (3000 ppm) are
to be found in waste oil. Automobile tyres are especially enriched in Zn for which levels
between 0.93 % and 2.05 % are reported. The impurities in used ion exchange resins will
depend on the material stream which they have been used to clean up, with especially high
levels being reported for Pb (6400 ppm) and Zn (3900 ppm). In contrast the levels of the
heavy metals in used wood are considerably lower but with maximum values of Cd (16
ppm), Se (5ppm), Cu (77 ppm), Pb (70 ppm), Sb (10 ppm) and Zn (560 ppm) which are
definitely above the average level in the crust.
Table 12: Range of Heavy Metals (ppm ) in Secondary Fuels Used in Clinker
Production
Element Class Waste Oil Fullers Ion Automobile Used Earth's
Earth Exchange Tyres Wood Crust
Resins
Cd 1 4 0.01 - 2 9 - 50 5 - 10 1.5 - 16 0.2
Hg 1 < 0.05 0.08
Tl 1 0.02 0.2 0.03 - 0.07 0.2 - 0.3 0.5
As 2 1.8
Co 2 25
Ni 2 3 - 30 0.001 - 30 8 - 87 77 11 - 50 75
Se 2 5 0.05
Te 2 0.01
Cr 3 5 - 50 2 - 11 20 - 330 97 16 - 25 100
Cu 3 40 - 77 55
Pb 3 10 - 21000 2 - 2500 150 - 6400 60 - 760 38 - 70 13
Sb 3 10 0.2
Sn 3 2
V 3 135
Zn n.c. 240 - 3000 10 - 480 56 - 3900 9300 - 20500 130 - 560 70

0 BEHAVIOUR OF TOXIC METALS IN CEMENT KILNS
Now that we know the concentrations of the heavy toxic metals in the raw material
components and fuels let us now turn our attention to the behaviour of these metals in the
cement kiln.
5.1 Volatility of Toxic Element Compounds

To gain an impression of the volatility behaviour of the toxic metals listed in the German and
Swiss Clean Air Acts, the vapour pressures of the most likely compounds to be present in a
cement kiln, i.e. when the elements are not incorporated into the lattice of the clinker
minerals, have been reviewed by Haegermann. Those having a partial pressure of at least 1
0 Hg at 300°C are listed in Table 13. As can be seen in addition to the high volatility of Hg
compounds, which are essentially completely emitted from cement kilns, other elements
listed in the German clean air act (Be, As, Se and V) can all form compounds with vapour
pressures of 1 mm Hg at temperatures of lower than 300°C and are hence capable of being
emitted from cement kilns.
Table 13: Heavy Metal Compounds of Class 1, 2 and 3 Elements which have a
Partial Pressure of over 1 mm Hg at or below 300 °C
Elemen Class Fluoride Chloride Bromide Iodide Oxide Sulphate Hydride Meta
t l
Be 1 * * *
carcin
Cd 1
Hg 1 * * * *
Tl 1
As 2 * * * * * *
Co 2
Ni 2
Se 2 * * *
Te 2
Cr 3
Cu 3
Pb 3
Sb 3
Sn 3
V 3

5.2 Distribution of Toxic Heavy Elements
For a comprehensive assessment of the distribution volatilization and emission of an
element, the exact balance of its intake and output must be known. To this end it is
necessary to analyse all material and gas streams, such as raw mix and fuel, in order to
determine the intake and the amount present in clinker and discarded dust.
Normally these calculations of the balances indicate that the total intake of an element from
raw material (Ir) and from fuel (If) minus discharges via clinker (Dc) and discarded dust (Dd)
equals the emission (Ex) plus that portion (Rk) retained inside the kiln system:
Ir + If - Dc - Dd = Ex + Rk (1)
According to the behaviour of the elements in cement kilns, three separate cases can be
distinguished, one for non-volatile and two for volatile compounds:
0 Case 1: For non-volatile compounds: Ir + If = Dc, i.e. there is no emission and no
accumulation in the kiln system: Ex, Rk = 0
1 Case 2: For volatile compounds: Ir + If > Dc
0 (a): surplus intake remains totally in the "kiln" system with no emission, i.e. Ex = 0
1 (b): surplus intake is wholly emitted with no retention in the kiln: Rk = 0
The transposition of equation (1) results in:
Dd + Rk = Ir + If - Ac - Ex (2)
To a first approximation the terms on the right side of the equation (2) can be considered as
constant:
Dd + Rk = constant (3)
By means of equation (2) it is possible to calculate that a certain portion of a compound will
remain in the kiln system if no dust is rejected and also the amount of dust required to be
rejected in order to keep the retained portion Rk below a certain limit. For this purpose it is
necessary to know the emission Ex. Only limited data are available for the latter as this is
sensitive information which cement plants do not make readily available.
The quantity emitted Ex can, however, be estimated if it is known which of the three cases
best describes the compound in question.
According to HMC/TC's own investigations and the findings of the Research Institute of the
German Cement Industry (VDZ), the Class 2 and 3 elements are not likely to exceed the
3
emission limits of 1 resp. 5 mg/Nm , respectively, in normal operation and with normal raw
materials. Even if sewage sludge is used as combustible, as partial substitute for coal, the
emission of all Class 2 and 3 elements is well below the limits. The elements of Class 1,
especially Hg and Tl show quite a different behaviour. This Class, characterized by its very
3
low emission limit of 0.2 mg/Nm , represents a permanent danger, a fact which also
becomes evident from the concentration of the elements in the Earth's crust shown in Tables
7 to 12. Therefore, it is worthwhile to give a more detailed description of the findings of these
investigations.

5.3 Distribution of Cadmium
According to HMC/TC's measurements, the degree of binding of cadmium in clinker shows
very large variations from 41 to 100 %. This corresponds to the findings of Sprung (VDZ).
The reasons for the variations are not clear. Relatively small temperature fluctuations and
also the burning conditions will have an effect. Due to the fact that the clinker absorbs only
part of the cadmium and the emission is insignificant, the rest will remain in the kiln system
causing recirculations with cadmium enrichment in different materials. The cadmium
concentration increases from raw meal to electrostatic precipitator dust to the hot meal. The
dust shows large differences in concentration between direct and compound operation.
These findings show that cadmium as a partly volatile element belongs to case 2(a) with no
emission (Ex = 0) but with accumulation in the kiln system.
5.4 Distribution of Mercury

The investigation of different materials has shown that the mercury contents vary widely and
are relatively low in raw materials and coal compared to the average concentration in the
lithosphere.
Mercury was not detected in clinker, however, in some cases an accumulation was observed
in kiln dust, although these concentrations (max. 190 mg/t) are insignificant compared to the
other elements, e.g. thallium. Furthermore, it should be mentioned that the mercury emission
in compound operation is appreciably lower than that in the direct mode. In compound
operation the emitted quantities are even lower than the intake. Obviously, much of the
mercury vapour is retained in the raw mill perhaps by adsorption on the surface of the meal
particles. In direct operation, the retained mercury will be released again and is emitted.
Summarising, it can be generally stated that the mercury introduced by raw materials and
fuels is completely emitted in gaseous form (in the dry process mainly in direct operation) so
that its behaviour is that of a volatile compound without retention in the kiln (Rk = 0, i.e. case
2b). These findings were also confirmed by the stack gas analyses available.
5.5 Distribution of Thallium

According to German investigations (Sprung, VDZ) and HMC/TC findings, cement raw
mixes contain about 100 to 800 mg thallium per ton corresponding to 150 to 1200 mg
thallium per t clinker. The concentration can be much higher in kiln feed and coal dust
because large quantities of thallium are accumulated from the kiln gases during drying and
grinding. Furthermore, the kiln feed shows significant differences between direct and
compound operation: The thallium concentration increases considerably during direct
operation and may ultimately be up to 8 times as high as in the compound mode. The
measurements reveal that neither the clinker nor the stack gas discharge the kiln system of
thallium. Only very high exhaust gas temperatures above 200°C (long dry and wet kilns)
may allow some thallium vapour to escape. Nearly the entire intake remains in the system.
Therefore, thallium is to be considered as a volatile compound with no emission (Ex = O)
and no output by the clinker (Ac = O), i.e. all Tl is accumulated in the kiln system. Hence, as
will be appreciated in light of the "Lengerich incident" it is particularly important to relieve the
kiln/grinding system e.g. by creating a "valve" in order to avoid excessive accumulations of
thallium which ultimately lead to kiln dust with extremely high Tl concentrations. Especially in
cases of disturbed operations e.g. insufficient or lacking precipitator function could the
emission of Tl become considerable.

5.6 Typical Balance of Toxic Elements In Kilns
5.6.1 Wet Kiln

A simple balance of the toxic elements in a large wet kiln fired with coal is displayed in Table
14 in which the percentage input as coal and slurry as well as the output via the clinker, dust
and stack gas are given. Here a simplified case is presented in that the totals for the inputs
and outputs are all 100 %. As can be seen the major input always comes from the slurry but
for As and Be almost 25 % is delivered from the coal. At the other end of the scale less than
10 % of the elements Pb, Tl and Zn comes from the fuel.
Table 14: Balance (% total) of Toxic Elements in a Large Wet Kiln
Element class Input Input Output Output Output

slurry coal clinker E.S.P.dust Stack
Be 1 carcin 76 24 96 4 0
Cd 1 82 18 18 81 1
Tl 1 95 5 22 14 42
Ni 2 82 18 97 3 0
As 3 77 23 92 8 0
Pb 3 93 7 22 77 1
Zn n.c. 93 7 88 12 0
Regarding the output of Hg, although not mentioned, this essentially leaves the kiln system
completely via the stack gas. The next most volatile element was found to be Tl of which 42
0 of the total input left the kiln in the stack gas whilst for Cd and Pb only 1 % left by this
way.
Of the elements which did not leave the kiln system, with the exception of Tl these were
mainly found in either the filter dust or in the clinker. Those less volatile elements found
predominantly in the clinker were As (92 %), Be (96 %), Ni (97 %) and Zn (88 %). In
contrast, for the more volatile species, 81 % of the Cd and 77 % of the Pb were found in the
kiln dust.
5.6.2 Lepol Kiln
An example from Krcmar et.al. of a heavy metal balance for a Lepol kiln fired with petroleum
coke as fuel is presented in Table 15. This presentation is more comprehensive than that
previously given in that the actual concentrations of the individual elements are given for the
various types of inputs and outputs. Those elements leaving the kiln in the form of a vapour
were not measured and the fact that the output can be either higher or lower than the input
is acknoweledged, these indicating a gain resp. loss within the kiln system. A missing output
therefore does not automatically mean the emission of the element concerned as a vapour.
This is certainly the case for Tl where additional measurements showed that the missing Tl
was not being emitted.

On the basis of the present balance the volatility of the classified heavy elements can be
classified by the volatility factor f, defined as:
f = concentration in clinker / concentration in raw meal
The following classes have been defined
♦ volatile (f = 0 to 0.35): Cd, Tl, Se, Pb
0 semi-volatile (f = 0.7 to 1.1):
♦ on-volatile (f = 1.2 to 2.7): As, Co, Ni, Cr, Cu, Sb, Zn
Table 15: Heavy Metal Balance for a Lepol Kiln Fired with Petroleum Coke
(Concentrations marked * given in ppb)
TOXIC ELEMENTS IN KILN DUST
The concentrations of toxic elements in cement kiln dusts have been measured for the wet,
semi wet/semi dry and dry types of cement kiln and also for rotary lime kilns. The results
obtained are listed in Table 16 in which it can be seen that no values are given for mercury
(Hg) due to the fact that it is essentially wholly emitted and not retained in the dust or in the
clinker. The values given refer to a period prior to the introduction of the German (TA Luft)
and Swiss (LRV) Clean Air Acts. In some cases the plants were using secondary materials
(combustible and non combustible) and also iron pyrites as a corrective. In those cases
where the filter dust was sampled the values given in Table 16 are indicative of the range of
these toxic elements and, as far as is known, are in no way associated with the burning
conditions. It is not possible to adjust conditions to ensure that the toxic elements are
steered into the clinker minerals and not into the dust. The values given for the stack gas
dust are again indicative of the raw materials used but now for their interpretation the
efficiency of the filter system must also be taken into consideration.

Table 16: Concentration (in ppm) of Toxic Elements in Kiln Dusts
Class Kiln Wet Semi Dry Lime
1 Cd 1-21 3-120 1-15 1-2
1 Tl 1-237 43-586 1-617 1-56-69
2 As 63-225 1-103 37-222 12-33
2 Co 1-2 1-11 1-169 41-765
2 Ni 11-75 20-87 22-4145 61-90
3 Cr 37-65 61-282 31-4107
3 Cu 45-1180 128 44-69 27-77
3 Pb 31-2830 522-7000 12-321
3 V 77-79 12-321 1-33
n.c. Zn 80-4500 27-77 107-147
5.7 Class 1 Elements (Cd and Tl)

As seen in Table 16 the values for these elements can be as low as 1 ppm to as high as 120
ppm in the case of Cd and 617 ppm (0.06 %) for Tl. It is interesting that both the wet and dry
kiln dusts had all Cd values essentially under 20 ppm whereas for the semi processes the
lowest value was 31 ppm and the highest 120 ppm. When Tl is considered, which is mainly
to be found in the dust and not in the clinker, no obvious trend could be seen but again the
lowest value for semi processes was higher than for the wet and dry types. For the two lime
kiln dusts studied, for which limestone was the only component of the kiln charge, the levels
of Cd and Tl were both under 2 ppm, an indication that the limestone is usually low in toxic
metals.
5.8 Class 2 Elements (As, Co and Ni)

Of the five Class 2 elements only As, Co and Ni were determined as the elements Se and Te
are not often encountered. The absence of these elements should not be taken for granted
as they are in the same column of the periodic table as S, and their absence should be first
be confirmed by taking an XRF-spectrum. In one secondary fuel tested at HMC/TC the level
of Se was found to be 12'000 ppm i.e. 1.2 %.
For As the values found in the dusts ranged essentially from 37 to 225 ppm, for Co from 1 to
169 ppm, with all values except one being under 37 ppm, and for Ni from 11 up to 4145 ppm
(0.41 %) in an extreme case. For Ni the values were either under 200 ppm or well above
with the extreme values all being for dry process kilns. With regard to the lime kiln dusts the
maximum values were 69 ppm for As, 33 ppm for Co and 765 ppm for Ni.
5.9 Class 3 Elements (Cr, Cu, Pb, V)

The very exotic elements in this class (Pd, Pt and Rh) as well as Mn, which is always
present in cement raw materials in small percentages, were not analysed. Up to date none
of the very expensive elements have ever been found in XRF-spectra performed at
HMC/CS.
Regarding the toxic metal Cr, whose presence is associated with chrome eczema, this has
been found in quantities between 31 and 4107 ppm (0.41 %) in kiln dusts. Most values lie
under 300 ppm and the appreciably higher values were found in only three cases. With
regard to Cu most values lie between 40 to 70 ppm and only in two cases (128 ppm and
1180 ppm) were higher values determined.
The toxic element present in the highest levels in kiln dust was found to be Pb which ranged
in value between 12 and 7000 ppm (0.7 %). For this element the level in the dusts of the

semi process plants were definitely the highest and with only two exceptions the values lie
between 1274 and 7000 ppm. For the dry process plants no values were higher than 321
ppm and for wet process kiln dusts most were lower than 150 ppm but with two extreme
values of over 2220 ppm. The levels determined for V lie between 1 and 79 ppm and so are
under the average level of this element (110 ppm) present in the Earth's crust.
5.10 Non-Classified Elements

Although Zn is not specified in the German and Swiss Clean Air Acts it can be found in kiln
dusts. The levels found ranged mainly between 80 and 230 ppm with one extreme value of
4500 ppm (0.45 %) also being found. This highest value was determined on a stack gas
dust.
5.11 Maximum Allowed Levels in Stack Gas Dust

Although the levels of the toxic elements present in the emitted stack gas dusts are high,
their significance should be considered by taking not their concentrations but the quantities
actually emitted into consideration. If it is considered that the maximum allowed emissions or
3
Classes 1, 2 and 3 are 0.2, 1.0 and 5.0 mg/Nm respectively, then for a stack gas stream
3
with 50 mg/Nm dust load the maximum allowed concentrations in the dust are 0.4 %, 2.0 %
and 10.0 %. In none of the cases investigated did the emissions come near to these values
and even with filters of poorer efficiency is it unlikely that the limits would be exceeded.
0 TOXIC ELEMENTS IN CLINKER
6.1 Levels of Toxic Elements

The levels of toxic elements in clinkers, reported in the literature are given in Table 17. As
can be seen for the elements of Class 1 i.e. Cd, Hg and Tl, the levels found were at the
highest 1.5 ppm in the case of Cd. For the elements of Classes 2 and 3 the values reported,
with the exception of Se, varied over a larger range. The lowest values could be low for As
(2ppm), Pb (1 ppm) but could be much higher such as in the case of Zn (530 ppm), Ni (400
ppm) and Cr (320 ppm)

Table 17: Range of Heavy Metals (ppm) in Clinker
Element Class Clinker Earth's Crust

Cd 1 0.01 -1.5 0.2
Hg 1 < 0.01 -0.12 0.08
Tl 1 <0.01 -1.2 5
As 2 2-87 1.8
Co 2 10 -21 25
Ni 2 10-397 75
Se 2 <0.2 -<1.0 0.05
Te 2 0.01
Cr 3 10-319 100
Cu 3 5 -136 55
Pb 3 1 -105 13
Sb 3 0.1 -1.5 0.2
Sn 3 (2.9) 2
V 3 10-135 135
Zn n.c. 29-531 70
6.2 Influence of Toxic Metals on Cement Properties

As just seen in Table 17 the highest concentration of a toxic element found in clinker was
531 ppm (0.05 %) in the case of Zn and 397 ppm (0.04 %) for Ni. What effect will these have
on cement properties?
From the studies of C. Tashiro the influence of additions of Cr, Cu, Zn, Cd, Hg and Pb, in
both soluble and poorly soluble forms has been well documented. At an addition level of 0.5
0 on a clinker basis no appreciable changes in hardening properties were detected and thus
with the much lower levels of these metals,and then already incorporated within the clinker,
definitely no change in cement properties is expected. The results of studies in which toxic
metal dosages to cement of appreciably more than 0.5 % and which showed retarding and
accelerating properties are not relevant to the properties of clinker. Statistical studies by
Blaine on normally produced cements also showed that at the levels of the toxic metals
present in clinker no changes are to be expected.
6.3 Ability of Clinker Compounds to Immobilize Toxic Metals

At the present time much research is being carried out into the use of cements to immobilize
the toxic metals present in ashes, slags and other industrial waste materials so that they are
not removed by leaching after dumping in an approved site. From the studies of Bhatty it is
known that the toxic elements can be incorporated into the structure of the CSH-gel during
its formation but can also enter the structure after its formation with the displacement of Ca
ions. In certain cases new unknown compounds containing a toxic metal may also be
formed.

From the studies of the VDZ (Sprung) it is known that even the water soluble toxic elements
will be completely immobilized on hydrating. This has been shown to be the case from
studies in which larger quantities of toxic metals than are possible in clinker as found in
wastes have been completely bonded by the use of cement pastes. The bonding is chemical
and adsorbtive together with the formation of a dense structure.
From the data shown in Table 18 it can be seen that the solubility of the toxic metals in a
1 -6
cement suspension is between 10- and 10 % of the quantities present.
Table 19: (1) Solubilities of Heavy metal Compounds at pH Values of 12.6 to 13.0:
0 Leaching Rates of Compacted Mortars (11 % Cement) after 28
Days Hardening
Element Solubility in Solubility in Degree of
Ca(OH)2 / KOH Cement Suspen- Solubility (%) from
Solution sion (%) Mortar
As -3 -5
8 x 10 4 x 10
Pb 93 -3
7 x 10
Zn -4 -6
1 x 10 2 x 10
Cd -4 -6 -4
1 x 10 < 1 x 10 2 x 10
Tl 90 -1 4
1 x 10 1 x 10-
With regard to the leaching out of the elements from 28- day mortars prepared with 11 % by
-4
weight of cement, the quantity of the elements removed by the eluate was maximum 2 x 10
0 of the quantity added. With regard to the small quantities normally present in clinker these
will be certainly immobilized when the cement is present in hydrated form in concrete.
0 INFLUENCE OF KILN TYPE ON HEAVY METALS EMISSIONS
01 Experience from Literature and Plants
7.1.1 Non-Volatile Metals

Emission of low or non-volatile heavy metals are not critical with regard to actual emission
limits. They can be kept under control by an efficient dedusting.
According to comparatively secured information [1, [18 the emissions of As, Be, Co, Cr, Cu,
Mu, Mo, Ni, P, Pb, Sb, Se, Te, V, Zn are extremely low on suspension preheater kilns with
efficient dedusting. As a general rule the emission of non- or low-volatile elements is always
much below 0.1% of the corresponding metal input into the suspension preheater kiln.
Thallium has a different behavior. It is known to form an outer cycle (not inside kiln) which
can create a high accumulation in the filter dust (a few thousand times) if no
countermeasures are taken. This can also increase the normal solid emission or become a
potential danger as at Lengerich.
On grate preheater kilns not only Tl but also additional elements as Cd, Pb and Zn can
become highly enriched in the filter dust. The emissions of all these elements are still mainly
in solid form and therefore related to the efficiency of the filter. Due to the enrichment effects
the emission may reach the order of 1% (instead of 0.1% as before).

All elements mentioned so far and under the conditions of suspension or grate preheater
kilns are known to have a very low or insignificant vaporous proportion (equilibrium
concentration). This is certainly applicable for exhaust gas temperatures <150°C. It should
be kept in mind that the volatility depends not only on the temperature but also on the other
accompanying components occurring on cement kilns (e.g. Cl).
7.1.2 Volatile Metals
Volatile elements cannot be controlled by efficient dust filters and are also classified as most
toxic (class 1). Therefore, they need special attention.
The most prominent example is the mercury. Hg is suspected to be almost totally emitted
3
[18, 24, 25]. At the same time the emission limit is only 0.2 mg/m , therefore, it may soon
become critical if the Hg input is significant. Hg emissions on cement kilns are almost
completely in form of vaporous Hg. Therefore, many older emission measurements (based
on pure solid sampling) either omitted Hg or did show unrealistically low figures.
Although Hg is very volatile, a certain retention effect in the cement kiln system is also
possible. This may e.g. be the case under the influence of some sulphur compounds or in a
raw mill during compound operation [7]. However, the retention during mill operation creates
a Hg accumulation which can evaporate when switching to direct operation thus creating a
high peak emission and a less favorable long term retention effect. This was e.g. observed
3
in [18] when some 0.2 mg/m were found during direct operation. However, this high
emission value seems to be very exceptional in general because normal raw materials and
fuels do not create critical Hg inputs. The problem of the highly volatile Hg is therefore
limited to very special cases and is not a general problem under the present emission limits.
It is not quite clear whether also other elements could become as volatile as Hg under
certain conditions. According to [14] Selenium on grate preheater kilns would be highly
volatile but this is not confirmed.
7.1.3 Long Wet and Long Dry Kilns
Unlike suspension and grate preheater kilns long wet and long dry kilns are less explored in
their heavy metal emission behavior. From the basic principle and from the higher stack gas
temperatures (high vaporous equilibrium concentrations) one could expect a less favorable
situation but the practical evidence is incomplete.
The most complete and extensive information basis so far are heavy metal balances from
the wet kiln at Obourg [16]. The results are partly surprisingly favorable. The emissions for
all heavy metals investigated such as Be, Cd, Hg, Tl, As, Co, Ni, Se, Cr, Cu, Mn, Pb, Sn, V,
Zn are very low and the calculated relative emissions are below 0.5% even for Cd, Tl, Hg. In
other words: none of the heavy metals seems to be volatile. The elements of class 1 and 2
3 3
are typically < 0.01 mg/m and class 3 elements are < 0.1 mg/m .
7.2 Specification of Missing or Inaccurate Data
When reviewing the corresponding literature one is often impressed by the very low heavy
metal emission or the excellent metal retention effect found for certain elements. When
going more into details it becomes evident that such impressive results are usually based on
pure solid sampling on the clean gas. The vaporous proportions are also verified in some
cases but the measuring methods on cement kiln exhaust gases do not allow for extremely
3
low detection limits (< 1 mg/m ) under practical conditions (method in [20] is not accepted for
practical emission measurements).
Measuring methods for vaporous proportions of metals also appear as possible weak point
in some cases. Measurements do often consider only solid sampling and the assumption is
made that vaporous proportions are negligible. The analytical confirmation is often missing
due to the lack of sufficiently sensitive measuring methods. And if the dedusting performs
very well, very low emissions and impressive retention effects for heavy metals of 99.999%
or even better can be calculated. Although there seems to be no problem with respecting
emission limits there could be a problem with the impressive number of nines in the retention
effect. In view of vision towards a future "zero emission" sampling methods and analytical
procedures need to be improved.
0 STANDARD TECHNOLOGY FOR REDUCTION OF HEAVY METAL EMISSION
8.1 Reduction of Heavy Metals Input into Kiln System

The input of heavy metals into the kiln system from raw materials and fuels can be controlled
and limited in order to avoid undesired effects. The justification can be:
0 Limits set for heavy metals in clinker (usually "political" limits, concentrations are often
too low to produce clear negative effects in clinker, except Cr)
1 Cycles of semi-volatile heavy metals and their influence on (CKD) kiln dust or bypass
dust
2 Emission of volatile heavy metals (e.g. Hg)
With regard to emissions, it is most important that the heavy metal input is limited, especially
if conditions are present which promote emissions such as:
0 high stack gas temperature
1 low metal retention capacity of kiln system
2 presence of compounds which can volatilize the metal, such as chlorine
For example, if sewage sludge is burnt in a cement kiln the measuring input has to be limited
[18, 19].
The emission of a metal only becomes important if the metal cycle can actually reach the
filter and if there is a problem with the function of the filter (EP shut down or poor
performance), e.g.Thallium cycle on SP kiln
Besides these special cases most of the known heavy metals behave as low volatile and are
virtually not emitted. The question is rather which metal concentrations in clinker and kiln
dusts are acceptable from a political and environmental point of view. Therefore, input
reduction will always maintain its importance, even if metal emission values are very low.

8.2 Modification of Process
Break the Cycle of Enrichments
If semi-volatile metals as e.g. Tl cannot be avoided in the raw mix there is a possibility to
break the cycle of Tl enrichment. This will then limit the maximum concentration in the filter
dust and thus reduce the potential risk during any malfunction of the filter.
The principle is often used on systems which originally worked with complete reintroduction
of the kiln dust (closed systems), as e.g. suspension preheater kilns or grate preheater kilns
without any bypass. Enriched filter dust is extracted from the system and usually introduced
into the cement grinding process, e.g. on suspension preheater kilns in combination with a
raw mill the kiln dust during direct operation can be extracted. The costs are very moderate
(< 0.2 Mio CHF). The effects in terms of normal emission is not the main argument, it is
rather a reduction of a potential risk.
8.3 Secondary Reduction Measures

Improve Dedusting
In case of non or semi-volatile heavy metals the emission can be reduced by improving or
replacing the dedusting. The effect of an improved emission is an almost linear decrease of
the heavy metal emission according to the reduction of the dust emission. However, the
heavy metals are usually not the main justification for such modifications.

0 BAT (BEST AVAILABLE TECHNOLOGY) FOR REDUCTION OF HEAVY METAL
EMISSION
9.1 Non-volatile metals
For most of the heavy metals which are non-volatile or semi-volatile the emission
reduction can be achieved by reducing the dust emission (see paper PT 95/4391/E, dust).
9.2 Volatile metals

Special technology in form of secondary reduction measures is only justified for volatile
metals as e.g. Hg. The following technologies are possible:
0 Dry scrubbing by circulating fluidized bed is also effective for Hg reduction [26. The use
of other absorbers (activated coke) is also possible and would improve the effect.
1 Activated coke filter The Polvitec at "HCB Siggenthal removes Hg at an efficiency of
some 99%. As a pure Hg removal system it would be extremely expensive for this
particular purpose (Investment of approx. 30 Mio CHF). Since the activated coke is re-
used as kiln fuel the mercury would form a cycle. The cycle is broken up by extracting
filter dust during direct operation and using it in cement grinding. Treatment of the dust
(thermal Hg-expulsion) has been tested but not realized on a large scale due to low Hg
concentrations.
2 Wet scrubbing This is not yet known on cement kilns but rather on waste incinerators.
Wet scrubbing of mercury is inefficient if Hg is present in metallic form (insoluble) and
may not be automatically the appropriate method [29. Oxidyzing the Hg would be
necessary. Again, this method would be extremely expensive only for the purpose to
remove volatile heavy metals.
3 Removal of critical heavy metal from specific fuels or material components
Example: Hg can be thermally expelled from sewage sludge before feeding it into a
cement kiln [27].

LITERATURE
Sprung S., Technologische Probleme beim Brennen des Zementklinkers, Ursacheund

Lösung, Kap. 7
Schriftenreihe der Zementindustrie, Heft 43/1982, Beton-Verlag, Düsseldorf
Sprung S., Emissionsprognosen beim Einsatz von Abfallbrennstoffen Zement-Kalk-Gips Nr
10/1984
Kirchner G., Verhalten der Schwermetalle beim Brennen von Zementklinker
Zement-Kalk-Gips 10/1986
Welzel K., Winkler H.D., Thalliumemissionen, Ursachen und interner Kreislauf
Kamm K., Schwermetallanreicherung bei der Elektrofilterung
Staub Reinhaltung der Luft Bd 45 1985 Nr 1
Weisweiler W., Krcmar W., Arsen- und Antimon-Bilanzen einer Zementofenanlage mit
Rostvorwärmung
Schütte R., Kupper D., Die Bedeutung von Kreislaufbetrachtungen für Produktqualität und
Umweltverträglichkeit der Zementherstellung
Weisweiler W. Krcmar W., Schwermetall-Bilanzen einer Zementofenanlage mit
Rostvorwärmung
Sprung S., Kirchner G., Rechenberg W., Rektionen schwer verdampfbarer
Spurenelemente beim Brennen von Zementklinker Zement-Kalk-Gips
10/1984
Weisweiler W., Blome K., Kaeding L., Thallium- und Blei-Kreisläufe in
Zementofenanlagen
Staub Reinhaltung der Luft 45/1985 Nr. 10
Weisweiler W., Mallon E., Schwarz B., Anreicherung von Thallium- und Bleihalogeniden
Staub Reinhaltung der Luft 46/1986 Nr.3
Seebach M., Tompkins B., The behavior of heavy metals in cement kilns Rock Products 26th
Int. Cement Seminar, New Orleans Dec. 5, 1990
Seebach M., Tompkins B., Metal emissions are predictable Rock Products April 1991
Krcmar W., Linner B., Weisweiler W.
Untersuchungen über das Verhalten von Spurenelementen beim Klinkerbrand in einer
Drehofenanlage mit Rostvorwärmung
Märki M., Schwermetalluntersuchungen an Roh-, Korrektur-, Brenn- und Zusatzstoffen
sowie an Kreisläufen bei der BCU
BCU Bericht 1.3.1994
Degré J.P., Comportement des métaux lourds dans les fours d'Obourg Obourg - Bericht
JPD/ID/94.471 vom 23.8.94
Imlach J., Sources and Reduction of HM-Emissions

MA 95/3389/E
Obrist A., Klärschlammstudie Teil 5: Bestimmung von Emissionswerten und
Produktqualität bei Klärschlammverbrennung...
VA 95/5170/D
Obrist A., Klärschlammstudie Teil 3: Die Möglichkeiten der CH-Zementindustrie...
VA 86/5300/D S. 136 ff "Rückhaltevermögen des Zementofensystems"
Hagermann B., Dampfförmige Schwermetallverbindungen im Zementofenabgas
Diplomarbeit Techn Univers. Clausthal 1982
Vgl. dazu DB 987: Emission Schwermetalldämpfe
Kirchner G., Reaktionen des Cadmiums beim Klinkerbrennprozess
Wickert H., Schwermetalle der Giftklasse 1 LRV im Abgas RK MA
86/518/D
Wickert H., Elemente der Giftklasse 1 LRV im Abgas UV MA
86/520/E
Obrist A., Iskraut A., Quecksilberemissionen in der Zementindustrie
Ergebnisse bisheriger Untersuchungen / Literaturstudien
VA 85/09/85/D
Weisweiler W., Keller A., Zur Problematik gasförmiger Quecksilber-Emissionen aus
Zementwerken
Zement-Kalk-Gips Nr 10/1992
Obrist A., Orientierende Untersuchung an ZWS-Absorber über Hg-Rückhaltevermögen VA
90/27/D
Stadtentwässerung Zürich, Klärschlamm-Entquickung, Technikumsversuche im
kontinuierlichen Verfahren bei der FA List
MA 88/3542/D
De Quervain B., Umweltfreundliche Klärschlammverbrennung am Beispiel PCW GWA des
Schweiz. Vereins des Gas- und Wasserfachs 3/92
Muster U., Abtrennung von Quecksilber mittels Keramikreaktoren Müll
und Abfall 7/95

APPENDIX
11.1 Thalium Documentation Dyckerhoff-Zementwerke AG, Lengerich

Characteristics of Toxic Elements quoted in TA Luft and LRV
Element Class Critical Dose Uses Source Natural

WHO max. Lethal dose concentrations
dose
Cd I 0.5 mg / Coating for metals (hot dip) Zn / Pb 1 ppm in plant
week Alloying material (wires, bearings, batteries) Ores and animal
Pigment (CaS in glass, paints, etc.) tissues
Insecticide (CdO, Cd(OH)2)
Hg I Scientific instruments HgS Soils, dust,
Mercury vapour lamps (Cinnabar) water, 0.005 -
Solders 0.25 ppm in
Mercury arc rectifiers nearly all foods
Anti-fouling paints
Pharmaceuticals
Seed dressing
Tl I 100 ppm Rodenticide / insecticide (Thallous sulphate) Pyrites Not naturally
poss. toxic Lenses, IR instruments found in plants
to soils Alloying material or soils
Fireworks
Dyes, pigments
12 mg/kg Low temperature thermometer (with Hg)
Catalyst
body weight Hair remover
Be I Cu / Be alloys Beryl ore
(cancer.) X-Rax tubes
Fluorescence light tubes
Nuclear reactor parts
As II MAC of Insecticides, Weed killers, fungicide (as K- By-product In all soils from
0.5 mg/m3 arsenate), Wood preservative Cu, Pb, <10 to 500 ppm
Glass Zn, Sn,
Alloying material Au-ores
Therapeutic (organic As compounds)
Co II 10-20 mg Alloying material In Cu / Co
per 100 g Magnet alloys ores in
body weight Cutting tools Zaire
Binder for tungsten carbide
Blue pigment in china glass
Ni II Ni and soluble Ni-salts are Alloying material S-ores in 0,0016% of the
not considered as being Special steels Ontario Earth’s crust
toxic to humans. The most Electroplating and is widely
important industrial aspect is Catalyst for hydrogenation of oils found in soils
their trend to cause Coinage and plants
dermatitis. Paints and enamels
Ni-Cd batteries
Se II Se rectifiers Anode Il soils from
Anti-corrosive coating on copper, steel etc. slime of trace to 30 ppm,
Decolouriser in glasses Cu- at which level
Paints and dyes refineries animals suffer
Photoelectric cells from alkali
Insecticides disease
Catalyst
Te II In cast iron to improve properties Bi-ore Traces in
In various Pb, Cu and Fe alloys residues seleniferous
In rubber industry soils but not in
As catalyst crops
Glass colouring agent
In bacteriological tests
In thermocouples for low temperature

dose
Sb II Alloying material Sb2S3
Battery grids (Stibnite)
Pewter
Printers type
Lead electrodes
Bearing metals
Rubber compounding
Flameproof clothing
Paint pigment
In ceramic / glass
Pb III Pipes, Roof coverings PbS In soils and plants
Tanks for acids (Galena) in very low levels.
Lead shot, Bullets, Linotype metal Clay soils make
Battery plates Pb insoluble but
Alloying material acid soils
In paints and pigments increase solubility
In glasses and enamels and toxicity.
Litharge rubber
Insecticides
Noise barriers
Cr III Alloying material in steels FeO.Cr2O3 In small quantities
Chromium plating (Chromite) in all plants and
Cutting tools (as carbide) soils. Plants
Paint primer (zinc chromate) grown on soils
Leather tanning agent with 0.2-0.4% are
Wood preservative (zinc, chloride, chromate) infertile
Pigment (paint, inks, dyes, etc.)
In photo industry as sensitiser
CN III
F III
Cu III 780 mg as Alloying material As S-ores
CuSO4 As electrical conductor in several
Gas / water pipes, Roof sheeting areas
Bronze paint
Anti-fouling paint
Insecticide
Mn III Alloying material As oxide up to 20 ppm in
Potassium permanganate sulphide some nuts
Colour in glasses / ceramics and
Paints, inks carbonate
Matches, Fireworks
Enrich manganese deficient soils
Pd III Alloying material In certain
Caralyst coals as
Dental alloys aresenide
Jewellery or selenide
Pt III Contact material in electrical circuits As native
Electrode material metal in
Catalysts Cu / Ni
High temperature electric furnaces ores
Dies for glass fibre production
SiO2 III
Rh III Pt / Rh alloys In wastes
Electroplating from Cu,
Electrical contacts surfaces Ni and Pt
Reflector in high power lamps extraction

dose
V III In special steels and alloys Oil refinery To 470 ppm in
Industrial catalyst residues soil; to 10 ppm
Pigment in paints, dyes, inks (Carnotite) in plants
In insecticides
In photography
Sn III

Chapter 7
Kiln Dust

Disposal of Wastes from Cement Production
1. INTRODUCTION ............................................................................................................. 226

2. REGULATORY ASPECTS OF DUMPING SOLID WASTE MATERIALS...................... 227
2.1 Swiss TVA 95 .......................................................................................................... 227
2.2 E.P.A.'s TCLP Leaching Test .................................................................................. 230
2.3 European DIN 38 405 D24 Leaching Test .............................................................. 231
3. DISPOSAL OF General CEMENT PLANT WASTES .................................................... 231

4. DISPOSAL OF KILN DUSTs .......................................................................................... 234
4.1 Discarded Dust, its Potential and Drawbacks ......................................................... 234
4.2 Methods of In-Plant Disposal of Untreated Kiln Dust .............................................. 236
4.2.1 Addition of Dust to Cement ............................................................................... 236
4.2.2 Returning to Kiln after Granulation or Compaction ........................................... 237
4.3 Methods for In-Plant Use of Kiln Dust after Removal of Alkalis .............................. 240
4.3.1 Removal by Simple Leaching............................................................................ 240
4.3.2 Removal by Passamaquoddy Technology Recovery Scrubber ........................ 243
4.3.3 Removal by Auxiliary Kiln or by Fuller Fluidised Bed Reactor .......................... 244
4.4 Reported Plant External Uses of Kiln Dust.............................................................. 246
4.4.1 As Fertilizer ....................................................................................................... 246
4.4.2 As Raw Material for Artificial (N-Viro) Soil......................................................... 248
4.4.3 As Asphalt Filler ................................................................................................ 250
5. Disposal of Bypass Dust .............................................................................................. 251

5.1 In Plant Use............................................................................................................. 251
6. DISPOSAL OF REFRACTORIES ................................................................................... 253
6.1 Background ............................................................................................................. 253
6.2 Materials Involved and their Chemical Compositions.............................................. 253
6.2.1 Materials Involved ............................................................................................. 253
6.2.2 Chemical Composition ...................................................................................... 254
6.3 Regulatory Aspects / Manufacturers Safety Data Sheets ....................................... 255
6.4 In-plant refractories disposal ................................................................................... 255
6.4.1 Plant "A" ............................................................................................................ 255
6.4.2 Plant "B" ............................................................................................................ 255
7. Recycling of spent cement kiln refractories ............................................................... 256
8. APPENDIXES ................................................................................................................. 258

8.1 Appendix Ia............................................................................................................. 258
8.2 Appendix Ib............................................................................................................. 259
8.3 Appendix Ic............................................................................................................. 260
8.4 Appendix Ib............................................................................................................. 260

INTRODUCTION
Whereas natural processes generally occur in cycles, in which the waste products created
do not harm the biosphere, the same cannot be said for human activities at today's high
population densities. The cement industry is no exception in that it generates wastes whose
properties range from being an unsightly nuisance to being definitely toxic. Table 1 lists
these principal solid, liquid and gaseous wastes. Whereas the gaseous wastes go into the
atmosphere and rapidly lose their identity by dilution, the same is not true for solid types.
The latter remain in possession of the plant, which has the responsibility for their safe
disposal
Table 1: Wastes generated in Cement Plants
Solid Wastes ♦ overburden, earth

kiln dust, bypass dust
mill dust
sulphur absorbent
reject production
used refractories
cleaning dust
scrap iron
rubber wastes (tyres, belts)
test prisms / cubes
Liquid Wastes ♦ rainwater / wash water

sewage water
oils / greases / emulsions
laboratory chemicals
solvents
Gaseous Wastes ♦ CO2, NOx, SO2, Hydrocarbons, etc.

In the present article the main emphasis will be placed on the management of the large
tonnage wastes namely refractory tear out and the various types of kiln dusts. No mention
will be given to that waste generated in the largest quantities, namely CO2 emissions, of
which the cement industry is a significant but not the major contributor.

REGULATORY ASPECTS OF DUMPING SOLID WASTE MATERIALS
In the past most manufacturing and processing industries simply dumped their waste
materials, euphemistically calling them "landfill". They did this because it made economic
sense and an awareness of the accompanying burden to the environment existed in only
rare cases, if at all.
Today the situation has changed. Environmental awareness is only all too present and those
harmlessly dumped wastes of yester year are now becoming the potential environmental
time-bombs of today and tomorrow. Many a cement plant is going to be confronted with a bill
for cleaning up its dumped materials, costs which will certainly be a multiple of the sum
originally "saved". No cement company in a developed country is today going to be allowed
to wind up its operations until it has made certain that all the materials it dumped during its
operational lifetime are in a form which poses no threat to the environment.
Some examples of the regulations applicable to the dumping of materials will now be given.
2.1 Swiss TVA 95

The Swiss TVA regulation of December 1990 is a typical example applicable to the dumping
of all types of material. It defines three categories of dumping operations, namely of the
'inert material", "residual material" and of the "reactor" types. Details of the requirements for
end storage in inert and residual dumps are given in Tables 2, 3 and 4. The use of reactor
dumps is presently reserved for slag from rubbish incinerating plants and for sewage sludge
(< 65 % H2O), inert/building waste materials, etc. when no capacity is available in inert or
residual material dumps. Dumps for highly toxic materials do not exist and such materials
have to be eliminated by other means.

Table 2: Swiss Dumping Regulations (TVA-1990)
Compositional Requirements
MT-Mat. No TVA Requirements
Material for End Storage
Condition Inert Residual
Composition known (%) > 95
Stone like Material (%) > 95
H2O-soluble substances (%) 0,5 5
Organic material (% C) 5
Pb (ppm) 500
Cd (ppm) 10
Cu (ppm) 500
Ni (ppm) 500
Hg (ppm) 2
Zn (ppm) 1000
Lipo. org. Cl compods. (ppm) 10
Gas formation with H2O none
Reaction with self, H2O,air none
Neutral, Capacity (Mol/kg) 1
The following building material can be included in an inert substance deposit:
concrete, tiles, asbestos cement, class masonry and road construction wastes.
Table 3: Elutriation Test (No. 1)

Requirements in Swiss Dumping Regulations (TVA 1990)
(Elutriation in CO2 Saturated Solution)
MT-Mat. No TVA Requirements (mg/l)
Al 1.000 10.00
As 0.001 0.10
Ba 0.50 5.00
Pb 0.10 1.00
Cd 0.01 0.10
Cr-III 0.05 2.00
Cr-VI 0.01 0.10
Co 0.05 0.05
Cu 0.20 0.05
Ni 0.20 2.00
Hg 0.005 0.01
Zn 0.50 5.00
Sn 0.20 2.00
Materials which may be deposited in inert substance dumps can either be unreactive
substances as defined in the mentioned tables or "building wastes" such as concrete, tiles,
asbestos cement products, glass and masonry/road construction wastes. Materials such as
metal, paper, wood, and textiles must be removed from "building wastes" as much as is
economically practical.
As can be seen in the Tables 4 to 6 there are limits set down on the chemical composition
(toxic metals, organic substances) of the materials themselves and on their leaching
behaviour in two elutriation tests. In the first test CO2 is bubbled through the solution to
keep the pH value low and thus ensure that the toxic cations are brought into solution and
not precipitated. In the second elutriation test, in which the anions and organic substances
present are determined, distilled water without CO2 is employed.
Table 4: Elutriation Test (No. 2)

Requirements in Swiss Dumping Regulations (TVA 1990)
(Elutriation in Distilled Water)
MT-Mat. No TVA Requirements (mg/l)
4+ 0.50 5.00
NH / NH3
Tl not specified not specified
Cyanide 0.01 0.10
Fluoride 1.00 10.00
Nitrite 0.10 1.00
Sulphite 0.10 1.00
Sulphide 0.01 0.10
Phosphate 1.00 10.00
DOC * 20.00 50.00
HC’s 0.50 5.00
LOC * 0.01 0.05
Cl Sol * 0.01 0.10
pH Value 6 - 12 6 - 12
DOC: dissolved organic carbon
HC: hydrocarbons
LOC: Iypophilic organic chlorine compounds
Cl Sol: chlorinated solvents
As a result of these regulations the following restrictions prevent (in Switzerland) the
dumping of waste materials such as kiln dust and refractories:
H2O soluble substances < 0.5 %, resp < 5.0 % exceeded due to presence of soluble
alkali chlorides and sulphates
total Pb and Zn levels over 500 ppm due to use of alternative fuels and raw material
components
soluble Pb and Zn limits exceeded due to use of alternative fuels and raw material
components.
6+
soluble Cr limit exceeded in the case of spent magnesia-chrome refractories

2.2 E.P.A.'s TCLP Leaching Test
In the USA the disposal of kiln dust and of spent kiln refractories is mainly by dumping the
refractories, often in plant external sites. This disposal is regulated by the Resources
Conservation and Recovery Act (RCRA) and hence the class of dump to which they can be
sent, is decided by their behaviour according to the Toxicity Concentration Leaching
Procedure (TCLP) test. Details of the latter and the limits defined for the various elements in
the leachate are given in Table 5.
Table 5: Details of USEPA’s TCLP Leaching Test

Leaching Medium Demineralised H2O Element Extract Limit (mg/l)
+ acetic acid (pH=5) Ag 5
Liquid / Solid ratio 20 : 1 As 5
Max. Particle Size (mm) 9.5 B
Nr. of Extractions 1 (agitation) Ba 100
Extraction Time (hr) 18 ± 2 Be 0.007
Cd 1
Cr 5
Cu
CrHg 0.2
Ni 70
Pb 5
Se 1
Sb 1
Tl 7
Zn
As can be seen the elements defined are (with one exception) trace toxic elements and as
such only present in natural raw materials and in new, unused refractories in trace
quantities, if at all. Trace quantities can, however, be picked up during their presence in the
kiln. The main oxides from which kiln refractories are produced i.e. SiO2, Al2O3, CaO, MgO
and the trace oxides Fe2O3, K2O, Na2O etc. are not stipulated and are anyway the same
oxides as found in PC clinker.
The one exception is Cr which, as Cr2O3, is a major oxide component of magnesia chrome
refractories, with some refractories containing up to almost 18 %. Although it is wholly present
as insoluble Cr2O3 in the new, unused bricks this is partially converted to a soluble alkali
6+
sulphate chromate during kiln operations and therefore contains soluble Cr . On the basis of
6+
tests carried out at HMC Corporate Services up to 0.4 % of soluble Cr has been determined
in spent magnesia chrome bricks which originally contained 17 % Cr2O3.
Spent magnesia-chrome bricks are therefore not likely to pass the TCLP test for disposal in
a municipal garbage dump and hence will have to be treated before such a disposal with
ensuing costs. A change to refractories of the chrome-ore free magnesia-spinel or doloma
types avoids this disposal problem completely.

2.3 European DIN 38 405 D24 Leaching Test
6+
According to this regulation the following concentrations of Cr dictate the disposal of all
wastes, but especially those containing chromite ore (magnesia chrome bricks) as follows:
♦ < 0.1mg/l: in household garbage dump
♦ > 0.1 mg/l < 0.5 mg/l: in toxic material dump
♦ >0.5 mg/l: special treatment or must be dumped underground
DISPOSAL OF GENERAL CEMENT PLANT WASTES

Of the general wastes generated in cement plants examples of their management at the
Rekingen and Ciminas plants are listed in Tables 6 and 7. Whereas the procedures for
dealing with wastes are already established in developed countries, plants in developing
countries invariably have to set up their own infrastructure. It is important that clearly
marked collection centres are located near to the places the wastes are generated and that
their use is promoted.

Table 6: Waste Concept at Rekingen
Material Quantity per Month Internal Disposal External Disposal
Office, stores
Paper carton Kiln
Workshop
Ferrous 4t Scap metal
Non-ferrous 2t Scap metal
Rags Garbage incinerator
Waste Wood Plant
Ferrous 40t Scrap metal
Non-ferrous 3t Scrap metal
Masonry 60t Mix bed
Paper, styropor 0.5t Garbage incinerator
Water Mix bed
Refractories
Cuttings 10 Mix bed
Tear out 500 t/a Mid bed
Garage
Tyres 0.5t Siggenthal
Oil / Air filters 50 kg Siggenthal
Waste Oil 1000l Waste oil dealer
Laboratory
Liquid waste 10l Special collection
Solid waste 10l Special collection
Radioactive source Supplier
Various
Glass Local collection
Light tubes 100 Special Collectio

Table 7: Waste to be Disposed at Ciminas Plant
Material Means of Disposal
Internal & External use
Absorbed Diesel Spillage Used as combustable
Filter bags Reused, sold for reuse
Internal use
Asbestos components Crushed, added to raw mill
Cement ex spillage’s Returned ro process
Clinical wastes Incinerated in containers
Glass Crushed, added to raw mill
Lamps Added to raw mill
Rags / cloths Incinerated in containers
Refractories (Cr containing) Crushed, added to raw mill
Wastes from charcoal Added to raw mill via clay
Sax ex sand blasting Added to raw mill
External use
Askarel oil Incinerated externally
Clothes / shoes / helmets Sold for recycling
Concrete cubes Given to personnel
Fibre glass Sold for recycling
Filter elements Pressed and sold
Iron ex slag Sold to foundries
Metal / electrical scrap Sold for recycling
Oils / greases Sold for recycling
Paper ex office Sold for recycling
Plastics Sold for recycling
Refractories Sold for reuse
Rubber / car batteries Sold for recycling
It is the disposal of the larger quantities of cement plant specific wastes kiln dust (up to
300’000 tpa at one Group plant) and spent refractories (up 800 tons at some Group plants)
that must receive the greatest attention. Whereas refractories are consumable items and it is
understandable that there will be remnants to be disposed of, this is not the case with kiln
dusts. In fact most Group plants do not have to discard dust and those actually doing so are
forced to do so for several reasons:
production in wet kilns of low(er) alkali cements from raw materials which do not have
sufficiently low alkali levels to produce such cements without the discarding of dust
use of a bypass in dry process kilns when the raw materials have a too high Cl content
to prevent the formation of rings in wet process kilns caused by a molar excess of SO3
over alkalis.
Most of the dust discarded in the Group is generated at wet plants in the USA and Canada
for the production of low alkali cements and thus in countries whose standard specifications
still do not allow the incorporation of kiln dust into cement.

DISPOSAL OF KILN DUSTS
4.1 Discarded Dust, its Potential and Drawbacks

The total quantity of dust discarded annually in the Holcim Group amounts to some 1.8 mio.
tpa, with the quantity per plant ranging from zero (most cases) to over 200'000 tpa at a few
N. American plants. In these latter cases the kiln feed / clinker factor can be as high as 1.91
which means having to process excessive quantities of kiln feed and a faster depletion of the
reserves.
The chemical composition of the discarded kiln and bypass dust varies with kiln type as well
as from from plant to plant. Details main oxide composition of kiln dust discarded at Group
Plants is given in Table 8, with details of the heavy metal levels in American kiln dusts being
listed in Table 9.
Table 8: Analyses of Dusts from Various Kiln Types

Kiln type Suspension preheater Grate preheater Wet long Dry long
Filter By-pass (Lepol)
Abbreviation DS-F DS-BP DG WL DL
Number of 19 6 13 36 2
samples
from-to average from-to average from-to average from-to average from-to average
L.o.l 31.3-37.9 34.5 4.3-13.3 8.9 5.2-28.6 15.6 3.8-33.8 22.6 21.8-27.0 24.4
SiO2 9.0-16.8 13.1 13.1-16.4 14.4 4.9-21.0 14.0 3.5-18.3 13.3 7.6-11.3 9.5
Al2O3 2.2-6.6 4.1 3.0-4.5 3.7 1.1-9.0 3.7 1.4-6.9 3.9 1.6-3.1 2.9
Fe2O3 1.2-3.3 2.2 1.2-2.5 1.7 0.4-3.7 1.5 0.6-4.3 1.9 1.1-1.7 1.4
CaO 38.6-45.7 42.2 44.4-60.8 52.3 14.9- 36.1 10.6-53.8 41.7 31.7-49.6 40.2
52.7
MgO 0.70-2.4 1.4 0.80-2.7 1.7 0.4-2.6 1.0 0.3-4.5 1.3 1.3-1.3 1.3
SO3 0.04-2.4 0.75 2.6-4.8 3.6 1.3-32.2 121.1 0.68-36.7 6.3 6.0-14.8 10.4
K2O 0.35-2.7 1.07 4.1-12.6 8.8 3.4-35.2 13.2 0.79-38.1 6.0 0.88-15.8 8.3
Na2O 0.07-0.30 0.17 0.17-1.6 0.62 0.18-2.6 1.1 0.12-14.1 1.3 0.07-1.3 0.69
Cl 0.04-0.68 0.25 2.3-7.1 4.2 0.10-2.0 0.81 0.07-10.8 1.3 0.88-1.3 0.69
Free lime 0.00-0.54 0.12 2.6-31.1 4.2 0.71-7.7 6.32 0-17.9 6.33 1.3-8.8 5.1

Table 9: Total (acid soluble) metals in cement and cement kiln dust
Total in Cement, mg/kg Total in CKD, mg/kg Ratio of Avg.
Volatile Min. Avg. Max. No. Min. Avg. Max. No. Cement /CKD
Mercury <0.001 0.014 0.039 21 0.004 0.660 25.500 68 2%
Selenium 0.62 NM 2.23 6 2.68 28.14 307.00 54 NM
Thallium 0.01 1.08 2.68 32 1.40 43.24 776.00 89 2%
Cadmium 0.03 0.34 1.12 42 0.1 10.3 59.6 81 3%
Lead 1 12 75 82 34 434 7390 95 3%
Antimony 0.7 NM 4.0 2 0.3 NM 3.4 9 NM
Silver 6.75 9.20 19.90 93 4.8 10.53 40.70 92 87%
Arsenic 5 19 71 54 2 18 159 50 106%
Nickel 10 31 129 85 1 22 60 65 141%
Barium 91 280 1402 94 35 172 767 95 163%
Berrylium 0.32 1.13 305 93 0.13 0.65 3.54 94 174%
Chromium 25 76 422 94 8 41 293 95 185%
Refractory
Notes:
Exclude non-detectable values
NM indicates average not meaningful because of large number of non-detectable values
No. indicates number of detected values
Kiln dusts, although discarded by a few plants can, in specific cases, possess characteristics
that make them of interest to other uses. The most important properties are:
low price
the very fine particle size / high absorptive capacity
neutralising properties due to the presence of CaCO3 and free CaO
enhanced levels of K2O and SO3
content of hydraulically active components
These properties have lead to proposals being made for using cement kiln dust in the
following applications:
absorber material
filler component in cement and asphalt
neutralizating agent
fertilizer/agricultural lime
stabilisation of mud from oil drilling
substitute for lime in
sewage sludge dewatering
waste stabilization
exhaust gas scrubbing

glass manufacture
rearing cattle
production of soil replacement
production of low grade binder
as cat litter
Here it must be emphatically stated, that only in exceptional cases has cement kiln dust
actually been used for any of the above applications in large quanties and over many years.
No universally acceptable uses have yet been located for kiln dusts outside of the cement
plant and their disposal problem has yet to be solved.
4.2 Methods of In-Plant Disposal of Untreated Kiln Dust
4.2.1 Addition of Dust to Cement

The most cost-effective method of using kiln dust is by incorporating it into cement. This
practice was allowed early on in a few national standard specifications in European
countries, which in 1989 resulted in 7 Group plants employing kiln dust in this manner. The
quantity involved at single plants varied from a minimum of 0.4 wt.% to a maximum value
of 8.5 wt.% of the OPC produced, with a total of 170'000 t being used in this way. In
Germany kiln dust could also be added to blast furnace slag cements, where the addition
resulted in improved early strength characteristics.
This use of kiln dust has since been incorporated into the European cement standard (ENV
197-1) which, as seen in Table 10,allows the addition of to 5 % of “minor additional
constituents” which may include kiln dust. As can be seen this addition is allowed in all
cement classes.
Table 10: ENV 197-1 Cement Types and their Compositions (wt.%)
Cement Clinker B.F. Pozzolana Possolana Fly Ash Fly Ash Burnt Shale Limestone Silica Minor
Type (K) Slag (S) Natural Indust. (Zi) Siliceous Calcareous (B) (L) Fume (D) Additions
(Zn) (Cs) (Cc) (F)
I 95-100 - - - - - - 0-5
II/A 80-100 S+Z+C+B+L+D+F = 6 - 20 -
II/B 65-79 S+Z+C+B+L+D+F =21 - 35 -
III/A 35-64 35-65 - - - - - 0-5
III/B 20-34 60-80 - - - - - 0-5
III/C 5-19 81-95 - - - - - 0-5
IV 60-90 - 0-40 - - - 0-10 0-5
V/A 40-64 18-30 18-30 - - - - 0-5
V/B 20-39 30-50 30-50 - - - - 0-5
1) Values in the table refer to cement nucleus, excluding calcium sulphate and additives
2) Minor additional constituents may be filler or may be one or more of the main constituents unless
inclueded as main constituents
3) Portion of silica dust is limited to 10% in types III/A and III/B
4) Portio of filler is limited to 10% in types III/A and III/B
For those countries in which the local standard specifications do not allow the addition of
modest additions of kiln dust as in ENV 197-1, cement companies should, by active
participation in standards committees, attempt to bring about the necessary changes in the
national standard.

4.2.2 Returning to Kiln after Granulation or Compaction
Essentially this method prepares kiln dust for re-entry into the kiln by simply increasing the
particle size, either by granulation (1-2 mm) or by briquetting it into long thin rods.
As far as is known the granulation method only found one practical application, on kiln 8 at
Obourg, but was not needed for the larger kilns 9 and 10 as the introduction of standard
specifications, which allowed the addition of dust to cement, made its use irrevelant. The
layout of the kiln 8 granulator (Schugi) is shown in Figures 1 and 2.

Figure 1: Granulater for production of fine granulated powder (1 - 2 mm)

Figure 2: Dust circuit, Slurry circuit, Mixture and compressed air circuit

As can be seen, the granulator consists of a vertical cylinder in which a concentric shaft is
placed, rotating at a speed between 1'000 and 3'000 rpm. Attached to the shaft are several
groups of knives whose angle to the shaft is continuously variable. To prevent the granules
sticking permanently to the cylinder wall the latter is lined with a deformable neoprene liner
which can be deformed from the outside by using pneumatically powered rollers, so causing
any deposits to fall off. Kiln dust and a binding liquid are introduced into the granulator,
which by the action of the rotating knives are formed into small granules, a residence time of
0.5 to 1 sec. being sufficient for granules of 0.5 mm to 1.5mm, max. 2 mm in diameter.
Typically for a granulator of 10 t/h capacity only several kg of dust are present in the
granulator at any one time, thus requiring an accurate and reliable dosage of both kiln dust
and binder. Normally the kiln dust is itself sticky enough so that water alone can be used as
granulating fluid.
At the Obourg plant the use of equal quantities of dust and slurry allowed the whole of the dust
produced in kiln 8 to be granulated and fed back to the kiln. The return of the dust in this way
allowed the water content of the slurry fed to the kiln to be reduced by some 2 to 3
and increased the quantity of the dust generated about 2 to 4 %. To prevent the granules
from sticking when being fed into the kiln via the 12 m tube, they were held in suspension in
a carrier gas stream. The velocity of the gas suspension was about 9 to 10 m/s and the extra
quantity of air introduced some 2 %. The tube was held about 40 cm above the level of the
slurry on which floated a compact bed of granules, of which about 97 % were captured by
the slurry. The granulator was shut down when kiln 8 went out of production.
4.3 Methods for In-Plant Use of Kiln Dust after Removal of Alkalis
4.3.1 Removal by Simple Leaching

An early method (1963) for the treatment of wet kiln dust, so as to allow its return to kiln, is
the leaching system developed and used at Holnam’s Dundee plant in the USA. As far as is
known this is the only leaching plant of this type in operation. Here some 350 tpd dust are
collected daily from both of the plant's 1'500 tpd wet kilns. Before the installation of the bag
filter in the 1990’s, the coarser dust taken from the smoke chamber and the first two
compartments of the electrofilter was returned to the kiln by insufflation at a rate of about 10
tph through the burner pipe. The remaining fine dust with higher levels of volatiles (from
chambers 3 and 4) was processed in the leaching plant, (see Fig. 3) in which over 70 % of
the alkalis were removed using quarry water. This leaching plant is still operating today but
the electro filter has since been replaced by a bag type.

Figure 3: Dust Leaching System

The kiln dust is slurried batch-wise in a mixing basin using a water to dust ratio of 9:1 and
then fed to the leaching tank from which the thickened slurry containing some 50 % water is
then returned to the feed end of the kiln. The use of the leaching plant drops the K2O value
of the clinker from 1.1 to 0.5 % and the S03 from 1.0 to 0.4 %.
An important factor for the continuing operation of the leaching plant is ensuring that the
effluent complies with the water pollution regulations. In the situation at Dundee the waste
water disposal problem was two-fold, namely the high pH value of the leaching plant effluent
and the high H2S levels of the quarry seepage water. The two problems were reduced to one
by using the quarry water for the leaching operations.
The effluent from the leaching plant has initially a pH value of 12.5 and contains S, SO4, K,
Na and Ca ions. It is neutralized by using steel mill waste pickling acid and ferrous sulphate
producing CaSO4 and Fe(OH)2, with the H2S being converted into insoluble sulfides. This
treated effluent is then discharged into a system of three lagoons in the mined-out quarry
where the solids settle out and the water is then aerated and clarified. Part of this treated
water is used for clay slip preparation, the rest being discharged into the river under state
regulations. As can be seen in Table 11 the operating permit of 1963 does not mention limits
for the heavy toxic metals. Because of their low solubility at high pH values the concentration
of the heavy toxic metals in the effluent will be anyway low.
Table 11: National Pollutant Discharge Elimination System

(Permit No. MI. 0002020)
Effluent Characteristics Discharge Limitations Monitoring Requirements
Outfall 001
Flow 3 Daily
17’500 m /d
Total suspended solids 20 - 30 ppm Daily - 24 hr. comp.
Total dissolved solids 5’000 - 6’000 ppm Daily - 24 hr. comp.
Oil and grease No visible film Daily visual obs.
Fecal coliform bacteria 200 / 100 m/l Weekly grab
Total Iron-Potassium -- Monthly 24 hr. comp.
Hydrogen sulfide 0.1 ppm Weekly grab.
pH 6.5 - 9.5 Bihourly grab
2. Outfall 001
Flow 3 Daily
4’500 m /d
Total suspended solids -- Monthly grab
Total dissolved solids -- Monthly grab
Oil and grease No visible film Weekly visual obs.
Temperature -- Weekly regarding
pH 6.5 - 9.5 Monthly grab

4.3.2 Removal by Passamaquoddy Technology Recovery Scrubber
This is essentially a modern development of the previously described leaching system and
first went into operation in 1990 at the Thomaston plant in the USA. In spite of world-wide
sales efforts this technology has only been installed at the above mentioned plant.
The layout of the equipment is shown in Fig. 4. An improvement over the Dundee leaching
plant is that there is no discharge of alkali salt effluent into the waterways, these being
converted into solid alkali sulphate/chloride to be sold as fertilizer, plus distilled water. The
system also includes a scrubbing system which removes 90 % of the SO2 from the kiln
exhaust gases as well as reducing particulate emissions. It further allows the reprocessing of
already dumped kiln dust.
Figure 4: Layout of the Passamaquoddy Scrubber

In the Passamaquoddy system kiln dust is mixed with water to form slurry just like at the
Dundee leaching plant. Cooled kiln exhaust gas is bubbled through a slurry bedwhich reacts
with lime and gypsum to form limestone and so avoids setting problems. At the same time
the SO2 anc Cl in the kiln exhaust gas converts the alkalis in the slurry into soluble sulphates
and chlorides. The alkaline leachate is then fed to the crystallizer, which, using waste heat
from the recuperator and reduced pressure, turns it into the desired products, namely alkali
salts (mainly K2SO4) for sale as fertiliser and distilled water. It was also said that the
scrubber would be able to remove dioxins and furans and convert the heavy toxic metals into
insoluble carbonates.
The reasons for the non- preliferation of this system are believed to be that the costs were
very high, in the order of 30 mio USD for a 3000 tpd kiln, and that a revenue from the sale of
alkali sulphate fertilizer and of distilled water was never demonstrated.
4.3.3 Removal by Auxiliary Kiln or by Fuller Fluidised Bed Reactor
In very recent times Holnam Inc.,working with Fuller have carried out initial small, pilot scale
trials to evaluate the possibility of producing low alkali clinker (Na2O eqvt < 0.60%) from kiln
dust using the Fuller’s Fluid Bed CKD Recovery Process. Discarded dust from Holnam’s
Clarksville plant, both fresh and landfilled, was used for the test which demonstrated an
alkali volatilisation rate < 65 % utilising a maximum permissible fluid bed operating
temperature of 1310'C and oxidising conditions. The use, however, of a mildly reducing
reactor atmosphere (0.8 % to 1 % CO) increased the alkali volatilisation rates to > 90 %
utilising a slightly lower fluid bed temperature of 1300-1305 °C. Details of the Fuller reactor
are presented in Fig. 5.
Figure 5: 18’’ Fluid Bed CKD Recovery System

From the results of the trials it can be said that:
a clinker with 0.33 % free CaO was obtained after a 3 hr retention time
the fluid bed system has a low rate of fines generation
this fine dust, with typical composition 51.5% K2O, 1.33% Na2O, 33.29% SO3 and
11.32% Cl, might be suitable as fertiliser
the TCLP heavy metals in the dust appears to meet the limits of the RCRA leaching test.
At one plant now being built in Egypt, Polysius have foreseen the use of an auxiliary rotary
kiln to concentrate the bypass dust prior to its disposal or further use.

4.4 Reported Plant External Uses of Kiln Dust
4.4.1 As Fertilizer
The features of kiln dust which make it potentially useful as a fertilizer are its potassium
content, its lime content and its acid neutralizing capacity.
When compared to potassium-lime fertilizers produced by the fertilizer industry, the
disadvantages of kiln dust include its variable chemical composition, its lower K2O
content, its high fineness and associated nodulizing problems.
In general a potassium fertilizer should fulfill the following requirements which may differ
somewhat from country to country:
should be in nodule form
only slight variations in chemical composition
should have a minimum content of soluble K2O and CaO (In Germany the K2O content of
potassium fertilizers should be between 40 to 60 % for the chloride types and between
25 to 50 % for sulphate formulations)
for some crops a low Cl content is required
Lime fertilizers should have a CaO content of between 40 to 90 % depending on whether
they consist of calcium carbonate, burnt lime or hydrated lime.
Only two plants were known in the past to successfully market their kiln dust. These were
the ENCI plant at Maastricht and Alsen-Breitenburg's Itzehoe plant, which sold their dusts as
fertilizers under the "KENCICA" and "Alsen" brand names. None of these plants today sells
kiln dust as fertilizer; the Itzehoe plant was closed in the early 1980’s and at ENCI the kiln
dust is being legally added to the cement.
With regard to their chemical compositions the dusts from these plants had quite different
characteristics as shown in Table 12. The fertilizer from Alsen was of the potassium type,
that from ENCI of the lime-potassium type, and as can be seen these compositions are
typical of the dusts from many wet and semi-wet plants.
Table 12: Chemical Composition (wt%) of the ENCI and Itzehoe Dusts
ENCI ITZEHOE
L.O.I. (500°C) 1.5 1- 2
CO2 18 2- 6
SiO2 18 2- 6
Al2O3 3 0.5 - 1.5
Fe2O3 3 0.2 - 0.9
CaO 46 5 - 15
MgO - 0.1 - 0.3
SO3 5 32 - 39
K2O 5 32 - 42
Na2O - 2- 3
Cl - 1

For the successful marketing of kiln dust a very large input may not be required. ENCI for
instance did not market directly under their own name but did so through a separate
company having a long experience in the business of providing farmers with fertilizers. They
were able to claim on the basis of crop tests that barley, turnips, potatoes and sugar beet
produced the same yields as when conventional lime or potassium sulphate fertilizers were
employed. It was even claimed that sugar beet contained up to 30 % more sugar when
"KENCICA" was used. ENCI claimed that the CaO to K2O ratio of 9:1 and the presence of
calcium silicates made their product especially valuable as a fertilizer.
In the case of "Alsen" the company did its own marketing, an example of the promotional
literature being shown (in German) in Figures 6 and 7.
Figure 6: Promotional Literature for “Alsen” Fertilizer

4.4.2 As Raw Material for Artificial (N-Viro) Soil
An interesting use of kiln dust, presently practised in the USA is in the production of artificial
soil. The kiln dust and bypass dust from a few US plants is being sold to a specialist
company - N -VIRO - which mixes the dust with partially dried sewage sludge to produce an
artificial soil, marketed as N-VIRO SOIL. The kiln dust, especially bypass dust, behaves as
lime and when mixed with the dried sludge, causes an increase in temperature which kills
most pathogens and greatly reduces the sludge odour, all this happening within only 7 days.
The production of this artificial soil is approved by the EPA as a PFRP (Process to Further
Reduce Pathogens) treatment. Uses to which N-VIRO soil is put include:
daily landfill cover material
land and mining reclamation
land remineralization
highway embankment construction
landscaping and turf material
agricultural lime/soil conditioner

Details from N-VIRO's promotional literature are shown in Fig. 8 and the composition of
the dust in Table 13.
Table 13: Chemical Composition of B

ANALYSIS EPA-LIMIT BYPASS DUST (%)
SiO2 17.55
Al2O3 4.87
Fe2O3 1.84
CaO 61.41
MgO 0.72
SO3 4.23
Na2O 0.51
K2O 1.53
TiO2 0.39
LOI 6.97
Total 100.02
E.P. TOXICITY PROCEDURE
Max. Limit ppm ppm
As 5.0 0.003
Ba 100.0 2.44
Cd 1.0 0.05
Cr 5.0 0.21
Pb 5.0 0.55
Hg 0.2 < 0.001
Se 1.0 0.006
Ag 5.0 0.07
TOXICITY CHARACTERISTIC LEACHING PROCEDURE
Max. Limit ppm ppm
5.0 0.004
Ba 100.0 1.7
Cd 1.0 0.05
Cr 5.0 0.25
Pb 5.0 0.6
Hg 0.2 0.003
Se 1.0 0.002
Ag 5.0 0.04

Figure 8: Promotional Literature for “N-Viro Soil”
4.4.3 As Asphalt Filler

Only one plant in the Holcim Group has, as far as is known, been able to sell some of its kiln
dust (approx. 200 t/a) for use as a filler in asphalt, this amounting to only 10 % of the approx.
2'000 t of dust sold each year.
On the basis of laboratory trials it was ascertained that only the dust from the first chamber of
the electrofilter was promising as a filler. the other dusts having the following drawbacks:
♦ smoke chamber: too coarse
♦ chambers 2, 4 and 4: too low density
too high solubility
Chamber 1 dust could be mixed with a normal filler (Vaulx) and used in the preparation of
asphalt according to the Dutch "Zwacke" class. The production of "Zeer Zwacke" and
"Middlesoort" classes was not possible. The proportion of natural filler that could be replaced
with chamber 1 dust in the "Zwacke" grade is about 50 %. The performance of the various
proportions of filler and chamber 1 dust are given in Table 14.

Table 14: Comparison of Performance of Different Proportions of Filler and Kiln
Dust as Filler Component in Asphalt
Dutch Standard Proportion Filler: Cust (Filed 1)
Characteristics Belgian Z.Zwakke Zwakke Middlesoort Filler 20:80 30:70 40:60 50:50 60:40
Standards (Vaulx)
Residue on:
1 mm 0 0 0 0 0 0 0 0 0
0.09 mm 0 - 15 0 - 15 0 - 15 5 - 10 5 - 10 5 - 10 5 - 10 5 - 10 5 - 10
0.63 mm 5 - 25 5 - 25 5 - 25 13 - 19 14.5 - 20.0 14.5 - 20.0 14.5 - 20.0 14.5 - 20.0 14.5 - 20.0
Apparent density 0.5 - 0.8 0.5 - 0.6 0.43 0.45 0.43 0.46 0.48
in toluene
Porosity 30 - 50 > 29 > 36 > 44 30 - 34 46.5 46.0 43.7 43.9 39.5
Sensitive to < 10 < 10 < 10 < 10 < 10 0 0 0 0 0
water
Swelling <3 <3 <3 <3 0.35 - 1.0 3.0 2.8 2.7 2.7 2.5
L.O.I (150°C) < 1.5 < 1.5 < 1.5 0.1 - 0.21 < 0.5 < 0.5 < 0.5 < 0.5 < 0.5
Solubility in water < 10 < 10 < 10 0.38 - 1.90 12.5 11.6 9.6 8.5 7.1
Bitumen (Getal) 28 - 30 40 - 50 54 - 60 32 - 37 58 56 55 50 44
DISPOSAL OF BYPASS DUST
5.1 In Plant Use

Because the original purpose of a bypass is to reduce the levels of Cl or alkalis in the kiln
system, it is just not possible to return the dust so extracted to the kiln. At one Group plant
(Höver) some 16 tpd of bypass dust (at times up to 30 tpd) is generated, equivalent to some
4'400 t/a. The composition of the dust is given in Tables 15 and 16. As can be seen the
enrichment in Pb, Cd and Hg is high, the value for Pb being higher when tyres are being
added as secondary fuel at the kiln inlet. Although it was possible in the past to dump
bypass dust in the quarry, this is no longer the case and so uses have had to be found for
this material.
Table 15: Bypass dust chemical composition

main components % raw meal clinker bypass dust enrichment factor
raw meal
bypass dust
ignition loss 35.8 0.3 6.5
SiO2 13.5 21.7 12.5
Al2O3 3.6 6.1 2.2
Fe2O3 1.3 2.0 2.5
CaO 43.2 66.5 31.0
MgO 0.6 1.1 0.4
SO3 0.6 1.0 11.0 18.3
Na2O 0.3 0.3 2.6 8.6
K2O 0.6 0.9 18.4 30.1
Cl 0.06 0.007 11.5 191.6

Table 16: Bypass dust chemical composition
trace elements raw meal clinker bypass enrichment factor
-4
mg/kg/10 % (not analyzed) raw meal
bypass dust
Sb 2.1 -- 10.0 5.0
As 10.4 -- 64.0 6.1
Be 1.0 -- 1.6 1.6
Pb 21.0 -- 947.0 (450)* 45.0 (21)*
Cd 1.0 -- 46.0 46.0
Cr 16.0 -- 90.0 5.6
Co 2.0 -- 11.0 5.5
Cu 5.0 -- 28.0 5.6
Mo 2.0 -- 6.0 3.0
Ni 12.0 -- 38.0 3.2
Hg 0.05 -- 2.0 40.0
Se 2.0 -- 56.0 28.0
Tl 4.0 -- 27.0 7.0
V 41.0 -- 80.0 2.0
*without burning tyres
Several plant internal uses have been found for the bypass dust including:
addition at raw material preparation
feeding to the raw meal
addition to OPC and slag cements as allowed by ENV 197-1
The bypass dust is added to the raw meal only when the Cl content in the kiln feed falls to
0.04% to 0.05 % and sufficient dust is added to give a constant level of 0.08 % Cl. This
ensures a smooth bypass operation. When added to cement the quantities are limited by the
requirement that the maximum Cl content should be less than 0.1 %, and is in fact held
constant at 0.05 %. The quantity of dust added on a cement basis varies between 0.8 and
2.25.%. The maximum quantity of bypass dust that can be used within the plant by this
method amounts to some 2'400 t/a. As an instantaneous recycling is not always possible,
some dust must be temporary stored until it can be used as previously described. Other
plant internal potential uses which have been evaluated but did not lead to success were in
the production of:
low quality hydraulic binders
special binding agents for salt mining

DISPOSAL OF REFRACTORIES
6.1 Background
Prior to the 1980's the disposal of spent refractories invariably involved dumping the broken
out material in the quarry. Thus only monetary aspects were taken into consideration and
any potential burden on the environment was ignored. The introduction of strict dumping
regulations has also altered this practice but in some countries an equally undesirable
situation now exists. There, the same refractory materials which are continuously being worn
down inside the kiln during normal operations cannot be disposed of by slow addition to the
raw materials after their removal.
What quantities of spent refractories have to be disposed of? For say an average wet kiln,
with a total consumption of 750 g/t basic refractories and 450 g/t of acidic materials this will
mean that about 190 g/t of basic and 110 g/t of acidic refractories, i.e. 300 g/t in total actually
have to be disposed of. This represents only some 0.03 % of the annual clinker production
(0.019 % on a raw material basis). Assuming the spent refractories are disposed of over a
period of 1 month this would mean the incorporation of 0.23 % spent refractories into the raw
materials. This quantity is very small and is less than the quantity of most waste materials
from other industries being disposed of in cement industry kilns.
HMC / CS is of the opinion that the plants of the Holcim Group should aim at being able to
dispose of all their spent refractories by incorporation into the raw materials.
6.2 Materials Involved and their Chemical Compositions
6.2.1 Materials Involved

In the widest sense the waste materials associated with cement kiln refractories must also
be include the packing materials in which they are delivered. This includes wooden pallets,
paper and plastic packing, steel and plastic drums etc. Such materials are disposed of within
the framework of a plant's general refuse concept (see Chapt. 2) and will not be further
considered.
The spent refractory materials fall mainly into the following categories:
basic bricks of the magnesia and doloma types
high alumina bricks
fireclay and lightweight insulating bricks
dense refractory concrete
insulating boards
steel plates, steel anchors
Most is in the form of bricks.

The chemical compositions of typical cement kiln refractories as supplied are given in Table
Here it can be seen that, apart from the elevated MgO and Cr2O3 contents of some types,
their compositions are similar to the oxide components used in clinker manufacture.
During their operational lifetime the original chemical composition changes due to various
reactions which include:
reaction with non volatile CaO and SiO2 from kiln charge
reaction with CO2 and SO2 kiln atmosphere (doloma types)
6+
reaction with with alkalis and sulphates to form Cr compounds (magnesia chrome
types)
pick up of alkalis, sulphates and chlorides into open pores
Table 17: Typical Cemical Cmposition of Various Classes of Refractories (Unused)

Element Magnesia Magnesia High Dolomite Fireclay Refractory
Spinel Chrome Alumina Concrete
L.o.l. 0.15 0.22 (2.7) 0.29 0.12 0.98
SiO2 63 2.3 11.5 0.61 55 49.8
Al2O3 10.5 8.5 77.5 0.27 39.1 24.2
Fe2O3 0.57 4.9 1.6 0.88 1.8 2.6
CaO 0.91 1.2 0.13 58 0.32 13.9
MgO 87 71.5 0.23 37.6 0.11 1.5
SO3 0.01 0.01 0.02 0.01 0.14
K2O 0.02 0.02 0.32 0.02 1.1 0.94
Na2O < 0.01 < 0.01 0.02 0.01 0.12 3.7
TiO2 0.04 0.11 3.1 0.02 1.5 1.3
Cr2O3 0.15 12.3
Mn2O3 0.01 0.08 0.01 0.08 0.02 0.05
P2O5 0.07 0.04 2.8 0.02 0.15 0.05
Cl < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01

6.3 Regulatory Aspects / Manufacturers Safety Data Sheets
Within the European Union and the USA various regulations are now in operation which
regulate the disposal of spent kiln refractories. Details have already been given in Chapt. 2.
Of importance to cement plant refractories specialists is the safety data sheets that most
European refractory manufacturers are now beginning to draw up. An example of such a
safety sheet, issued to comply with EU Guideline 91/155/EG, is given in Appendix I. As can
be seen in page 3 of this guideline, comments are made regarding the disposal of the
refractory materials (cut pieces and breakout) and also of the packing materials.
These guidelines are for consumers in general and do not take into account the cement kilns
capability of disposing of its worn refractories. Normally mention is made to the presence of
6+
Cr .
6.4 In-plant refractories disposal
In many Group plants, especially in Europe, spent refractories are cost effectively and safely
disposed of via the kiln, into which they are introduced at a controlled dosage rate by:
milling in the raw mill or by
milling in the coal mill
In both cases the spent refractories first require to be crushed and any steel plates, shims or
anchors removed by a magnet (or in some cases by hand) before grinding
The crushing process can be either carried out in:
existing plant crushers (raw materials or clinker) or using
a mobile crusher from an external company (disposal via coal mill)
The route via the coal mill is of special importance for the case of spent doloma refractories
in wet process plants as their inclusion into the raw materials would have an adverse effect
on slurry viscosity and pumpability.
In the following examples are given of several methods of in plant disposal.
6.4.1 Plant "A"

In this semi- wet process plant fitted with a filter press, the spent bricks leave the kiln via
the grate cooler and are crushed via the clinker crusher . They are then transported on a
band, where a magnet removes any metal, to a small silo from which the crushed
refractories are weighed into the raw feed at a slow controlled rate. The waste refractories
are thus kept within the plant, require no additional handling and are effectively out of sight.
6.4.2 Plant "B"
In this large wet process plant the spent refractories are removed from the kiln via a man
hole and are taken by truck to a covered storage. At regular intervals an external contractor
crushes the spent material with a mobile crusher and removes the steel present. The
crushed refractories are fed slowly into the coal mill and introduced into the kiln as an
additional "fuel ash". It is thus homogeneously distributed.

Figure 9: In-Plant disposal of Worn Refractories
RECYCLING OF SPENT CEMENT KILN REFRACTORIES The

philosophy of:
♦ avoiding / reducing waste, ♦
recycling or down cycling it,
♦ generating energy by its disposal and ♦
finally disposing of the "ultimate waste"
is now becoming established in many developed countries. When applied to cement kiln
refractories all these possibilities, with the exception of energy generation, are in principle
thinkable.
Under the motto, "nothing will be recycled until it is economic", the recycling of spent cement
kiln refractories does not appear to offer much potential. The only materials that could in
theory be completely recycled are the small volumes of cuttings from unused bricks bricks
and rest bricks which were never installed. Once used refractories from cement kilns are
normally not suitable for recycling.
When viewed from a purely financial viewpoint low cost firebricks will never be worth
recycling and only high cost qualities such as magnesia spinel might have some potential.
From a technical viewpoint used cement kiln refractories are always contaminated with a
host of substances which substantially lower their melting temperature and so prevent their
re-use in cement kilns. At best recycling means use in a lower or non refractory context and
certainly not in the cement industry.

From a practical point of view compound materials such as monolithics will probably never
be able to be recycled. The recycling of dolomite qualities would necessitate the resintering
of the material because of deterioration of the sinter due to hydration.
In Japan used magnesia chrome bricks from the cement and glass industries are classified
as hazardous wastes and are returned to the supplier. This magnesia chrome waste is
recycled as gunning mixes for application in BOF steel works.
However, even if landfilling costs were to become astronomical the fact that the possibility
remains (or should remain) for a cost effective disposal via the clinkering process, means
that it is very unlikely that cement industry refractories will ever need to be recycled on a
regular basis.
As far as the production of cement is concerned the remnants of kiln refractories cannot be
returned and recycled by the refractory manufacturer. The techniques required for this are
not available and the quantities involved are small. Kiln refractories are and will stay
consumable items. Only in a few developing countries are worn kiln refractories used
outside of the plant as aggregate for the production of low quality refractory concrete. This
form of "disposal" could not be practised in developed countries.

Appendixes
1 Appendix Ia

8.2 Appendix Ib

8.3 Appendix Ic
8.4 Appendix Id

Chapter 8
Quality Assurance
1. Quality Assurance.......................................................................................................263
2. Quality Systems in the Cement Industry according to ISO Standards.....................281

Quality Assurance
1. INTRODUCTION............................................................................................................264
2. DEFINITION OF QUALITY AND ITS CEMENT RELATED INTERPRETATION............265
3. OBJECTIVES OF QUALITY ASSURANCE...................................................................267
4. ACTIVITIES IN QUALITY ASSURANCE.......................................................................268
4.1 Quality Planning........................................................................................................268
4.2 Quality Control.......................................................................................................... 271
4.3 Quality Audit............................................................................................................. 271
5. QUALITY ASSURANCE SYSTEM.................................................................................272
5.1 The Elements of a Quality Assurance System..........................................................272
5.2 Organisation of Quality Assurance............................................................................272
5.3 Quality Assurance Handbook....................................................................................276
6. QUALITY COSTS..........................................................................................................276
6.1 Introduction............................................................................................................... 276
6.2 Elements of Quality Costs.........................................................................................277
6.3 Relationship of Quality Costs to Sales......................................................................278

INTRODUCTION
Every industry producing goods or services is confronted with the subject of quality. Their
products are in the true sense of the word qualified by the customer with respect to the
extent they meet his expectations. Included in his expectations is a conception of the price;
thus, quality cannot be dealt with apart from cost aspects: on one hand, a product able to
fulfill high and special requirements usually fetches a higher price than a "middle of the road"
product; on the other hand, its production costs are as a rule also higher. Furthermore and
that is not of less importance for the producer the market for a highly specialized product
might be limited and the profitability questionable in spite of the high specific selling price.
It is the fundamental task of the quality assurance of a company to deal with the product
quality and its relation to production costs, selling price and market potential, with the
objective to manufacture a profitable product.
The establishment of an efficient quality assurance and its integration within all levels of
company organisation, has for many companies been the basis for their success the brilliant
advance of various sectors of the Japanese industry is an impressioning example.
In view of this experience, a generally applicable concept of quality assurance has been
systematically developed, and has reached now the stage where quality assurance systems
are specified in national or international standards.
Also in the cement industry, the needs for a well conceived quality assurance have
significantly increased over the last twenty years, mainly due to the following developments:
more stringent requirements on concrete and cement
growing competition due to over-capacities
escalating energy costs
more efficient - though also more delicate - production facilities
increasingly tighter environmental regulations
involvement in active environmental protection

DEFINITION OF QUALITY AND ITS CEMENT RELATED INTERPRETATION
Quality is defined as follows:
Quality is the totality of features and characteristics of a product that bear on its ability to
1
satisfy a given need or fulfill certain requirements .
It is obvious that cement producers and consumers have a different attitude towards cement
quality.
The cement manufacturer is above all interested in low investment and production costs.
Quality, therefore, is primarily associated with costs for materials, energy, equipment and
labour. His major concern usually was - and often still is - concentrated on the suitability
and uniformity of raw materials and intermediate products to allow optimum operation of
the expensive installations with respect to quantity and costs.
On the other hand, the consumer expects a product with favourable characteristics for its
many applications. The changed market conditions - these days most countries envisage
a very pronounced buyer's market - put him in a very strong position with regard to
quality requirements. He need not be satisfied by the mere fulfilment of the standard
specifications of the purchased cement, but may well demand more stringent
requirements. Often, these additive requirements mean higher strengths and excellent
uniformity - permitting to minimize this, according to the user, expensive ingredient in
concrete - but they can also refer to other properties such as rheology (concrete
workability) or colour. Furthermore, the consumer may urge the producer for detailed
information not only about the final product, but even on all stages of cement
manufacture. This may ultimately result in periodic inspection visits to the plant by
external officials to so-called protect the consumers interests. Quality assurance by the
cement manufacturer, therefore, is interpreted by the consumer as quality insurance in
the sense of product liability.
Hence, the quality requirements of producers and consumers are by far not congruent; while
a certain limitation of the product variability is generally in line with the manufacturer's
attitude towards quality, his product costs can be seriously affected by the user's request for
pushed quality levels, excessive uniformity and special requirements (Fig. 1) . Depending on
the suitability of the available raw materials, correctives, additives etc., it will be more or less
costly for the manufacturer, if he has to comply with all the specific requirements. Moreover,
the requirements of the two parties change significantly with time, and this change does
usually not follow along the same lines.
1
ISO 8402 - 1986: Quality : Vocabulary
Fig 1 Cost/Quality Relation

OBJECTIVES OF QUALITY ASSURANCE
It is the task of quality assurance to find an appropriate compromise between the divergent
interests of producer and consumer.
More specifically, quality assurance has to ensure (Fig. 2):
the manufacturing of market-conform products
a smooth process by an appropriate design of the raw mix and other intermediates
an economic utilisation of raw materials
the compliance with environmental regulations
Fig 2 Objectives of Quality Assurance

The prevailing trends within these four subjects can be summarised as follows:
cement application:
Product oriented aspects gained in importance over the past years. For many plants it is
no longer sufficient to produce cement meeting the standard specifications or any
nationally accepted quality level, but they are obliged to adapt their products and product
ranges to specific customer requirements, be it for the inland or export market.
The implementation of technical consulting services for cement application provides an
important marketing support for many plants in competitive markets.
process:
Induced by world-wide cement over-capacities and rising energy costs, plants are under
considerable pressure to produce more cost-efficiently. A detailed optimisation of the raw
mix and other intermediates with respect to process requirements can significantly
contribute to this aim. Rationalisation by means of automatisation of process and quality
control are also potential measures. However, they require a careful return-on-
investment analysis.
raw materials:
Optimal utilisation of raw material deposits, both long - and short-term, has become an
economic necessity and is nowadays facilitated by powerful computer programs. Waste
materials - both combustibles and non-combustibles - are successfully utilised in
cement manufacturing. Special measures and investigations are necessary prior to and
during their application.
environmental aspects:
Just as any other industry, the cement industry has to comply with increasingly stringent
regulations on emission and immission. Different to most other industries, it offers,
however, the possibility to actively contribute to a safe disposal of waste materials by
means of their utilisation for cement manufacturing.
ACTIVITIES IN QUALITY ASSURANCE

By definition (ISO 8402), quality assurance comprises all those planned and systematic
actions necessary to provide adequate confidence that a product or service will satisfy given
requirements for quality.
Within the concept of quality assurance, three principle functions are distinguished:
Quality Assurance
Quality Quality Quality
Planning Control Audit
4.1 Quality Planning

The basic objective of quality planning is to determine the company's strategy and tactics in
quality assurance. This begins with defining the relative significance of quality assurance
and its main areas of emphasis, taking the particular situation of a company, e.g. with
respect to market or to raw materials resources, into account.
The individual subjects that have to be dealt with are compiled Fig. 3, and can be
summarised as follows:

Outwards activities have to include the decision making in co-operation with marketing on
number and types of products to be manufactured on the basis of criteria such as: demand
of domestic and export market, raw materials situation, plant characteristics, etc.; evaluation
of the demand for an extended support of the marketing department, e.g. in form of a
technical customer service, and, if required, the realisation of such a service; furthermore, to
represent the company's or cement industry's interest in the design of cement and concrete
standard specifications or in the environmental legislation, by participation in the respective
committees or bodies.
Plant internal planning activities comprise the design of the raw mix(es) as a direct follow-up
to product planning and, in the same context, the establishment of a concept for short-and
long -term raw material exploitation and securing. Of a very practical relevance is the set-up
of a concept for quality control; questions regarding the personnel structure, the equipment,
the degree of automation as well as data-processing and communication are of primary
importance. Last but not least it is also incumbent on quality planning to increase the
knowledge on the quality criteria and to develop a quality consciousness of the personnel by
adequate training. Addressants are on the one hand the management of both the production
and marketing sectors, and, on the other hand, the subordinate staff and unskilled labour
members, since it will ultimately depend on them whether quality assurance is properly
observed in all sectors of the plant activities.

Fig 3 Quality Planning

4.2 Quality Control
The task of the quality control (Fig. 4) is the routine monitoring of material properties
throughout the process and the initiation of necessary adjustments in the process controls.
The pre-requisite are an elaborate control plan and well - established methods and
procedures for sampling and testing. It needs no particular mention that the infrastructure of
quality control can also be used for non-routine purposes.
Fig 4 Quality Control
In view of new technologies, special attention has to be paid nowadays to data storage,
processing, presentation as well as to data communication to the plant management and to
process and sales departments. Data processing and communication not only serve to
initiate process control measures - which can, in certain cases, be done directly by computer
control - but also as a "quality certificate" for produced or dispatched products and as basic
information for quality planning.
4.3 Quality Audit

Quality audit - as an integral part of the QA system - has the objective to critically scrutinise from
time to time the procedures and efficiency of quality planning and quality control and to suggest
adjustments if deemed necessary. Particular reasons for performing a quality audit are certainly
significant changes, e.g. in the market situation, in the raw material situation or

in the plant itself (extension, shutdown of production lines). It might be sensible to have
a quality audit carried out by a third party in order to ensure an unbiased approach.
QUALITY ASSURANCE SYSTEM
5.1 The Elements of a Quality Assurance System

In slight deviation from the official definitions, quality assurance system may be defined as
the organisational structure, responsibilities, procedures, processes and resources for
1
implementing quality assurance.
In very general terms, the elements of a quality assurance system comprise:
Organisation of personnel, dedication of tasks and responsibilities
Lines of communication of information and decisions in routine and non-routine
situations
Quality control proper, i.e. sampling, testing, initiation or direct performance of reactions
(automatic and manual)
Training scheme
Technical customer service
Co-operation with external official bodies
Quality cost control
It is evident that a quality assurance system has to be individually tailored to the specific
needs of a company, taking regard of the complexity of the plant situation, of the prevailing
market situation, and of the relationship with official institutions.
Detailed lists and descriptions of the required elements of a quality assurance system are
contained in national or international standards, particularly in the ISO 9000 standard series.
Companies meeting the standard requirements can have their quality assurance systems
certified by recognised certification organisations. The ISO 9000 standard series or quality
assurance systems find increasing application in the cement industry, in respond to
corresponding demands from the market or from product conformity certification schemes
(see report MA 91/3866/E).
5.2 Organisation of Quality Assurance

It is the responsibility of the plant managers to supervise, on the basis of operating reports
and statistics, the suitability and reserves of the raw materials, the quality of intermediate
and final products as well as the conformance of the latter with valid and relevant standards.
For these tasks, the plant manager requires an organisation for quality assurance to
establish the required information and data, and thus to prepare proposals in the context of
quality assurance on which the plant manager can base his decisions.
There are several ways of organising quality assurance: again, the right choice is strongly
dependent on the significance of quality assurance for the company.
The term quality assurance system is not defined in ISO 8402. The term quality system is
defined in ISO 8402 as the organisational structure, responsibilities, procedure processes
and resources for implementing quality management.
The personnel from the chemical and physical laboratory may be placed in an advisory
position (staff function) to the plant manager and is merely responsible for the technical
execution of quality control (Fig. 5). Where the aspect of quality control is mainly connected
with the process of cement manufacture, a form of organisation is often established in which
the production superintendent is directly responsible for the laboratory. These kinds of
organisation have, however, the consequence that practically no direct decisions can be
taken by the organisation which carried out the actual quality control. All decisions and
instructions have to come from the plant manager and this bears the danger of an
overburdened and inflated plant management. Moreover, it hinders communication and
direct co-operation of the quality control department with other departments which have a
strong relation with quality assurance. Consequently, this kind of organisation may serve the
needs in a situation where material problems in production are not too complicated and
when only a limited product range is supplied to a relatively uncritical market.
Fig 5 “Low Profile” Quality Assurance
Otherwise, the situation implies usually that the scope of the company's quality assurance is
split so that various parts of the organisation are actively involved (Fig. 6). The plant
laboratory is only concerned with the aspects of quality control in production; in addition to
this, a separate group is established to cope with the quality problems directly associated in
the application of cement. Such a group for technical customer services is often helpful when
a company runs several plants - it is usually part of a company's technical planning and
marketing divisions respectively. This form of organisation certainly offers the advantage that
in the individual fields of clinker/cement production and of technical consultation of
customers, a high level of expertise may be achieved with a limited number of specialists.
On the other hand, the following problematic aspects have to be considered:

Fig 6 Quality Assurance Organisation in Multi-Plant Company
An optimum reflection of the entire system of quality assurance from raw materials
through the process to cement application is lost. As a consequence of such a partial
consideration and the exclusive activity as a service organisation, the interest and
comprehension and thus motivation and initiative of the people responsible for actual
quality control may be strongly reduced; it thus remains questionable if a continuous
optimisation of production and product will still take place.
Without an organisation which is in a position to consider all relevant aspects of cement
manufacture and cement application, there is an inherent danger that the producer's
interests are not sufficiently protected against the users' requests on product
specifications (e.g. when testing and product specifications are established).
The aim of delegating responsibilities and competence to those who are fully dedicated to
quality assurance leads to the assignment of a superintendent of quality assurance within
the line organisation of the plant (Fig. 7). Today, this type of organisation is widely applied,
most functions of quality assurance are assigned to the quality superintendent who is also
responsible for the laboratories.

Fig 7 Typical Quality Assurance Organisation
He reports directly to the plant manager and assists the latter in all matters concerning
quality and quantity of raw materials. His further tasks can comprise:
to advise the production superintendent
to examine the final products
to handle technical problems
to advise plant management and production superintendent on environmental aspects
to participate in standardisation committees
Before deciding on the one or the other form of organisation for quality assurance, the
situation should always be thoroughly analysed. A selection of the organisational form purely
on the basis of tradition ("because it always had been this way") is definitely not adequate.
Also preferences and dislikes of available personnel should not become decisive factors
when organising a company's quality assurance. To achieve an organisation with an
optimum efficiency not only in the short but also in the long run, it is essential
to define clearly the company's tasks of quality assurance, considering prevailing and
possible future conditions of raw materials, process and market,
to formulate duties and responsibilities as well as authority for the personnel in charge of
quality assurance (job description, performance standards).
From this, clearly defined requirements as to the qualification of the personnel for quality
assurance and hence to the training scheme according to their tasks, can be derived.
Depending on the level of responsibility for quality assurance, this training scheme should
not only be limited to techniques of quality control (sampling, testing, data processing), but
should also include sound experience in the entire field of cement manufacture and
application. Moreover, in addition to the professional skill, basic knowledge and
understanding of economic principles and relations is required for the superintendent of
quality assurance to achieve the primary goal, namely to make a profit on quality. Being a
leader of an organisational entity, he also requires training in the field of leading personnel.

5.3 Quality Assurance Handbook
It is strongly recommended that for every company the quality assurance system and
generally all features related to quality assurance are compiled and documented in a
"Quality Assurance Handbook". This handbook should be of a binding nature and serves to
ensure transparency and efficiency in quality assurance. Although primarily intended for
plant internal purposes, the awareness of its existence by customers or official bodies can
only be beneficial to the company's quality image.
In order to meet the requirements of national or international standards on quality assurance
systems, a comprehensive quality documentation needs to be established. It usually
comprises two or three document levels, i.e. a top level quality manual, mid-level quality
procedures and lower level work instructions (see Report MA 91/3866/E).
QUALITY COSTS
6.1 Introduction
“Quality assurance: from outlay to profit”
It is increasingly recognised that quality assurance is no longer primarily - though inevitably -
a cost generating nuisance but represents a valuable instrument to increase profit.
Quality assurance no doubt creates costs - as does any other activity within an enterprise -
but, on the other hand, it also allows to reduce unnecessary costs which may arise in case
of inadequate quality of services or goods produced.
To ensure the economic benefits of quality assurance, the costs referred to above should be
assessed and presented in a transparent way. This is ultimately the most effective way to
convince the top management of the benefits of quality assurance, and to obtain their
support in establishing and maintaining an integral company policy for quality assurance.
The purpose of assessing the quality costs of a company can be summarised as follows:
the quality cost assessment allows to recognise and to rank weak points. It makes evident
what kind of deficiencies generate the highest (additional) costs at what stage of
production. Measures to improve quality and quality control can be implemented
accordingly. Also, sectors apparently causing very high control costs can be re-assessed
as to whether the risks justify the costs.
quality costs serve as basis for a cost/benefit calculation in connection with planned
investments for production installations and for the instruments of quality control. Also
the consequences of changes in personnel structure can be estimated on this basis.
quality costs can represent a motivation at all levels to increase the efficiency of quality
assurance
quality costs may permit a comparison with similar enterprises. This, however, requires
due precaution because either the assessment of quality costs or the means of
production or both may be different.

6.2 Elements of Quality Costs
Quality costs are in general terms defined as the difference between the real expenditures
and revenues and those which would arise, if no quality related failures during product
development, manufacture and marketing would and could occur.
It is generally accepted that quality costs should be subdivided into three elements:
QUALITY COSTS
Prevention Costs + Appraisal Costs + Failure Costs

Applied to the cement industry, these cost categories can be characterised in the following
way:
6.2.1 Prevention Costs
These basically comprise the costs arising in connection with quality planning as outlined in
chapter 4.1.
Cost compilation may be unproblematic as far as it concerns the QA-internal personnel
costs. For the investigations involving other departments - which is often the case - the
fraction of "quality costs" is difficult to determine and would have to be estimated.
6.2.2 Appraisal Costs
These comprise all expenditures of quality control:
personnel costs
costs of accommodation and infrastructure
purchase and amortisation of equipment
consumables
These costs can easily be ascertained.
6.2.3 Failure Costs

Failure costs are additional costs or non -realised revenues arising as a result of
inadequate quality (“non-quality") of intermediate or final products. They can have most
different origins as documented by the following list which is by no means exhaustive:
undue consumption of expensive corrective material on account of non-optimal mix
design
elevated heat consumption, reduced production rate, or excessive refractory wear due to
inappropriate raw mix quality
increased cement grinding energy (higher fineness) due to inadequate clinker quality
separate storage and blending of under-burnt clinker - . addition of costly "quality
improvers" upon cement grinding when using inferior (e.g. outdoor stored) clinker
use of off-specification special clinker as ordinary clinker
treatment of customer complaints; payments for damage in cases of failure in cement
application
sales losses owing to manufacture of not market-conform products or due to loss of
confidence after a case of damage.

It is evident that a comprehensive quantitative assessment of these failure costs is very
difficult.
The relation between these three cost elements is shown in figure 8.
Fig 8 Relationship between Quality Cost Elements
6.3 Relationship of Quality Costs to Sales

Reliable figures for quality costs in the cement industry are not available, mainly due to the
lack of a quantitative assessment of failure costs. Reasonable estimates of the expenditure
for personnel, equipment and infrastructure associated with quality assurance, which may be
taken as a rough approximation of the prevention and appraisal costs, reveal figures in the
order of 1 to 2% of the cement net sales price. Sound estimates of the failure costs (“non-
quality” costs) are not available; ad-hoc estimates range between 5 and 20 % of net sales
price

For comparison, some figures related to other industries for quality costs and improvements
achieved after optimisation are given in figure 9.
Fig. 9 Quality costs in different industries before and after optimization, expressed as
a percentage of turnover.
Industry before after
motor cars 6- 9 4-6
chemicals 5- 8 3-5
- raw materials
food 7-12 3-6
chemicals 7-12 4-7
- complex products
ceramics 15 10
overall average 11 6
The different types of quality costs before, during an after successful

optimisation, expressed as a percentage of turnover.
Before during after
prevention 0.5 1.5 1.0
appraisal 3.5 3.0 2.5
consequential 7.0 4.0 2.5
total 11.0 8.5 6.0

Quality Systems in the Cement Industry according to
ISO Standards
1. INTRODUCTION............................................................................................................283
2. REASONS FOR IMPLEMENTATION OF DOCUMENTED QUALITY SYSTEMS..........283
2.1 Internal reason..........................................................................................................283
2.2 External reasons.......................................................................................................283
3. IMPLEMENTATION OF DOCUMENTED QUALITY SYSTEMS IN THE CEMENT
INDUSTRY..................................................................................................................... 284
3.1 General..................................................................................................................... 284
3.2 Specific European Situation......................................................................................284
4. PRINCIPLES OF THE ISO 9000 QUALITY SYSTEM....................................................285
5. ASPECTS OF QUALITY SYSTEM DOCUMENTATION ACCORDING TO ISO 9002....286
5.1 General..................................................................................................................... 286
6. EXAMPLES OF ISO 9000 QUALITY SYSTEMS IN HOLCIM GROUP COMPANIES. . .287
6.1 Queensland Cement Ltd. (QCL) Australia.................................................................287
6.2 Cementos Boyaca (Columbia)..................................................................................287
6.3 HISALBA.................................................................................................................. 287
6.4 Groupe Origny (France}............................................................................................287
6.5 Conclusions.............................................................................................................. 288

P. Bürki, HMC/TC/Materials Division
Formalized and documented quality systems are increasingly applied in the cement industry,
often applying the models given in the ISO 9000 standard series. The article addresses the
internal and external reasons for this trend and outlines the basic features of ISO 9000 and
the criteria for an appropriate documentation. Examples of application and documentation’s
from Holcim. Group plants are discussed.

INTRODUCTION
The cement industry has been traditionally subjected to an extensive product
standardization. The compliance with these standards formed the only generally valid quality
liability towards the consumer. The way in which the products are produced and in which the
quality is controlled has been entirely up to the producer.
In many other industries, another method of quality commitment in the supplier/customer
relationship has been adopted: in place of product standardisation - which is practically
impossible for products undergoing rapid developments - the manufacturer has to provide
evidence that he disposes of o quality system that enables him to plan and realize a state-of-
the-art quality level throughout the production cycle of the product. Efforts to find an
internationally agreed upon concept of quality systems and a generally accepted terminology
eventually resulted in the ISO 9000 standard series which were approved in 1987, and which
were adopted in many national standards as well as in the European standards (EN 29 000).
It is a fad that today many of the supplier/ customer relationship in Europe are based on
these standards.
REASONS FOR IMPLEMENTATION OF DOCUMENTED QUALITY SYSTEMS

Various reasons taken individually or in combination, can lead a company to implement a
quality system along with a corresponding documentation.
2.1 Internal reason
2.1.1 Confidence, transparency efficiency.

A well documented quality system provides transparency and confidence with regard to all
quality related activities, both to those directly involved in production, marketing, and quality
assurance, and to the company management
2.2 External reasons
2.2.1 Market.
Customers can require the existence of a (certified) quality system as a prerequisite for a
delivery contract. This particularly applies for large customers or for customers in the public
sector.
2.2.2 Product liability
For the protection of the individual private consumer, fault independent regulations for
damages caused by defective products are increasingly being introduced. While such
consumer oriented product liability legislation has been established in the USA for quite
some time, similar laws have been, or are being introduced in Europe following the
guidelines issued by the EC - Commission - as well as in other parts of the world.
A documented quality system and the availability of quality data records can be considered
as effective preventive measures for avoiding the release of defective products and are
moreover valuable evidence for the defence of a producer in case he is involved in a product
liability litigation.

2.2.3 Certification of products conforming to standards.
There is a trend largely promoted by consumer-oriented organisations - that the
conformance to standards of a given product should be certified by official authorities
(certification body, approved body). The scheme for certification may not only involve
external product control, but also the requirement of a documented quality system. This
aspect may become the <<driving force>> for implementation of standard quality
systems based on national or international standards (e.g. ISO 9000) in the European
cement industry, as further outlined in the subsequent chapter.
IMPLEMENTATION OF DOCUMENTED QUALITY SYSTEMS IN THE CEMENT

INDUSTRY
3.1 General
Until recently, there was little external pressure for establishing documented quality systems
in the cement industry. Nevertheless, a number of plants have established a company- wide
quality system and documentation on a volunteer basis, with a view to improve internal
transparency and efficiency and/or to use it for promotional purposes to gain a competitive
advantage.
Increasingly, however, external factors are pushing the cement industry, at least in certain
regions, to adopt documented quality systems, practically exclusively according to ISO 9002:
The British Cement Industry has subjected itself to a certification scheme by the British
Standard Institute (BST). Labels for <<quality assured cement>> are granted to the products
of cement manufacturers certified on the basis of a ISO 9002 quality system.
In Australia, the cement and construction industries are subject to the generally valid state
regulation that <<all firms supplying manufactured goods to (governmental) agencies, must
have adequate quality assurance systems...>>. In practice, again, the ISO 9002 system is
adopted.
Very recently (1991), API—the American Petroleum Institute—has introduced the demand
for a documented quality system (Q1) similar—though not identical—to ISO 9002 for
granting their monogram on well cement.
3.2 Specific European Situation

In the context of the European cement standardization, (EN 196; ENV 197) a scheme for
the certification of conformity of cement complying with the requirements of ENV 197 has
been proposed (document entitled <<European Cement Certification>>). The proposed
scheme takes account of, and refers to two EC regulations, namely the Construction
Products Directive'' and the <<Global Approach to Certification and Testing>>, which
attempt to harmonize within Europe the regulations, the standards and the means for
verifying compliance with standards.

The European Cement Certification scheme defines the tasks of the manufacturer and of
approved bodies, in order to obtain product certification. Among the tasks of the
manufacturer belongs the implementation of a quality system based on EN 29002 (ISO
9002), to be documented in form of a quality manual. It is foreseen in the scheme, that
approved inspection bodies regularly inspect the production facilities, quality control
procedures and testing equipment with a view to approve and verify the procedures defined
in the quality manual. This implies a considerable intensification of external control for
assuring conformity to standards, which has so far been largely based on auto-control and
external control of finished products. It has to be pointed out that this certification scheme is
still much debated and an approval in the majority of the member countries not yet in sight.
It is noteworthy that Spain has adopted in form of national standards, not only the draft EN
cement specifications (ENV 197), but also a certification scheme largely identical to the
proposed <<European Certification Scheme>>.
PRINCIPLES OF THE ISO 9000 QUALITY SYSTEM

The ISO standard series 9000 on quality systems has become effective in 1987. It describes
models and elements for quality systems in different contractual situations and for the
internal quality management.
ISO 9000: Quality management and quality assurance standards—

Guidelines for selection and use.
ISO 9001: Quality systems—Model for quality assurance in design,
development, production, installation and servicing
ISO 9002: Quality systems—Model for quality assurance in production and
installation
ISO 9003: Quality systems—Model for quality assurance in final inspection
and test.
ISO 9004 Quality management and quality system elements—guidelines
For the purposes of the cement industry, the ISO 9002 is usually applicable. The main points
of emphasis of these standards can be summarized as follows:
Company commitment to quality; quality responsibility on top management level.
Clear definition of quality relevant procedures, including organisation and responsibilities,
throughout the production process and the various departments involved
Clear definitions of measures in case of non-conformity to quality targets
Training of personnel
Requirement for quality audits
Implementation of a continuously updated quality documentation.
The ISO 9001-9003 Standards for quality systems explicitly describe models for
quality assurance. Their application has to be made company (or industry) specific.

ASPECTS OF QUALITY SYSTEM DOCUMENTATION
ACCORDING TO ISO 9002
5.1 General
A quality system has to be documented in order to make it visible, applicable, and auditable.
The ISO 9000 explicitly demand the establishment of a quality system documentation.
It is recommended to apply the <<pyramid structure>> for the organisation of the quality
system documentation (Fig. 1 ). Usually, three document levels are distinguished:
Quality Manual
Quality Procedures (for Quality Guidelines)
Work instruction and test descriptions
Fig 1 Structure of Quality System Documentation

The <<pyramid-structure>> offers the following advantages:
The quality manual may be kept relatively short. It provides a transparent guideline to the
concepts of the quality system, its organisation and procedures. It provides the
connections to the other document levels by referencing.
The quality manual largely covers the organisational aspects of the quality system and
may thus be made accessible to externals (customers) without disclosing technical
know-how.
The quality manual is not subject to frequent modifications.
On the workshop level, only parts of the quality procedures and works instructions are
required, which can be selectively distributed.
When applying the ISO 9000 standards, the structure of the quality manual should, but need
not, follow that of the standards, i.e. the list of contents should correspond to that of the
corresponding standard. If a different structure is applied, which may be better adapted to
the company's requirements, a correlation has to be made with the clauses of the standards.
For the administration of the quality manual and all other quality system documents, it is
advisable to
use loose sheets to facilitate exchange of pages or chapters
indicate date and stage of revision on each page or chapter number the pages only
within a chapter
to have a control on the internal and external document distribution.
EXAMPLES OF ISO 9000 QUALITY SYSTEMS IN HOLCIM GROUP COMPANIES
6.1 Queensland Cement Ltd. (QCL) Australia

(see article on page 38, this issue)
6.2 Cementos Boyaca (Columbia)

The implementation of a company-wide quality system at Cementos Boyacà started some
four years ago, on an entirely voluntary basis. The reasons were, on one hand, a recognised
lack of efficiency and of personnel motivation in most sections of the plant, and on the other
hand, the objectives of the Company in the frame of PMI (Project Marketing Innovation) and
TOC (Total Offer Concept), necessitating a market and quality orientation of the company.
As a definite target and motivating factor to improve the company-wide quality assurance,
the attainment of the national quality award <<Premio Nacional de Calidad>> was aimed at.
This goal was reached in 1991, when the quality award was granted to Cementos Boyacà.
(see ‘Holderbank’ NEWS 4/91, p.5).
6.3 HISALBA
As already mentioned in chapter 3.2, the Spanish national standard is very close to the
proposed <<European Certification Scheme>>. The cement plants of the HISALBA group,
therefore, have certified quality systems according to ISO 9002. A comprehensive
information on his topic was recently published by HISALBA (Calidad y Automatization
HISALBA).
6.4 Groupe Origny (France}

In the context of the French scheme for approving cement trade marks (AFNOR), all plants
of the Origny group have had to dispose of quality manuals (<<plans qualité>>) for many

years to give evidence of existing quality systems. The requirements on the manuals have
not been very high; the document could be kept very short and had admittedly little
significance for the practical day-to-day quality assurance in the plant
Recently, the Groupe Origny took the decision to reelaborate these <<plans qualité>> in
such a way that they would become practice relevant, binding working tools. Among others,
the following objectives were put forward:
Quality improvements of final and intermediate products should be achieved by increased
efficiency, at minimum investment costs. Quality deficiencies should, without
exceptions, be detected (and eliminated) internally and not by customers.
Preparation for the anticipated higher requirements on quality systems and
documentation within the future European Certification scheme (ISO 9002).
Maintenance of company image and market position in view of similar efforts to
implement documented quality systems being made by competitors
With these objectives in mind, Origny made the interesting approach to describe in detail all
quality relevant procedures from quarrying to cement dispatch, by applying just two types of
formalized sheets. These sheets contain instructions on routine procedures, frequencies of
sampling and testing, targets, acceptance and alarm limits, instruction in case of exceeding
the limits, etc. and are distributed to all concerned and continually updated in order to strictly
conform with actual practice at all times. About 90 sheets were required to cover the entire
process of the Origny plant.
Origny intends to incorporate the entity of these sheets in the future quality documentation
as required by ISO 9002.
6.5 Conclusions
From the yet limited experience within the Holcim Group - and the cement industry in
general - it is an undisputed fact that the implementation and documentation of a quality
system according to ISO 9002 is an immense task. It has also become evident that an
implementation can only be successful (and durable) with the uncompromising support of
the company management.
The opinions regarding the benefit of such a system are still divided. The main criticism
expressed by the sceptics is that the written procedures are, in the best case, declared
intentions, over -emphasising in their very nature the quality aspects of cement
manufacturing. In practice, it is claimed, pragmatic considerations of other production or cost
related aspects at a given situation and point in time, may result in decisions not congruent
to the instructions contained in the quality documentation. Furthermore, it is pointed out that
it is questionable whether the considerable effort will always be made to update the system
and the documentation in the course of changes arising in the process, in the materials
used, and in the organisation of the company.

On the other hand, the astonishingly good acceptance in companies where such systems
have been introduced, is quite remarkable. The benefits often put forward better
organisation, clarified responsibilities, well defined decision criteria etc. resulting in better
plant operation, improvements of intermediate and final product quality and hence in
strengthened market position—indicate that significant weak-points in these respects exist in
many companies. It thus appears that the detailed scrutinizing of all activities and
procedures which is mandatory for establishing a documented quality system, is by itself
already very valuable, as many deficiencies can already be detected and rectified at this
stage. Attention is also drawn to todays more rapid fluctuations in personnel in the context of
which well-documented procedures and instructions serve as most valuable training tools.
The experience is not yet sufficient to permit definite conclusions on the above argument
The present trend, however, is in the direction of increased application of ISO 9002
standards as a consequence of market or standard requirements. It is therefore advisable
that the companies closely follow the developments and start their preparation in time. In a
situation where the implementation of a complete ISO 9002 quality systems is not yet very
urgent, a stepwise elaboration starting on the work instruction level is a recommendable
practical approach, as it was carried out by Origny. The ultimate application of a complete
ISO 9002 quality system will, in many cases, not be a question of choice but of survival.

Chapter 9
Quality Control: Concept, Control Plan, Procedures

Quality Control: Concept, Control Plan, Procedures
1. INTRODUCTION............................................................................................................293
2. QUALITY CONTROL CONCEPT...................................................................................293
3. CONTROL PLAN........................................................................................................... 294
4. GENERAL PROCEDURES AND CONTROL PARAMETERS.......................................297
4.1 Supplied Materials....................................................................................................299
4.2 Quarry....................................................................................................................... 299
4.3 Pre-blending............................................................................................................. 300
4.4 Raw Meal.................................................................................................................. 300
4.5 Kiln Feed.................................................................................................................. 303
4.6 Hot meal................................................................................................................... 305
4.7 Clinker...................................................................................................................... 306
4.8 Ground Cement........................................................................................................316
4.9 Dispatched Cement..................................................................................................317
4.10 Ground Coal.....................................................................................................317
5. UNIFORMITY................................................................................................................. 318
6. EXTENDED QUALITY CONTROL.................................................................................321

INTRODUCTION
The principle aim of this paper is to highlight aspects concerning quality concept, plans and
procedures that meet with today's process and market requirements. It is certainly not
possible to make a generally valid statement as to the most suitable way of carrying out
quality control and to define the most adequate degree of sophistication and automation for
every plant. This must be part of the quality planning in each individual company and plant.
QUALITY CONTROL CONCEPT

Quality control normally complements the process control, with the aim being to monitor the
product quality at the different production stages, so that the required quality level is finally
achieved.
Fig. 1 Quality Control.
Tested properties may include chemical composition, moisture, granulometry as well as

other properties relevant to the specific material.
Special procedures are being applied for the delivery control of combustible or non
combustible waste materials.

CONTROL PLAN
The control plan shows how the general control concept for each raw material, intermediate
and final products is to be realised.
A detailed control plan for quality control is set up by a thorough analysis of the following
questionnaire:
What should be examined?
What information is required for the control?
How often must the test be performed?
How accurate must the result of the testing be?
In every particular plant, the concept of quality control, with its objectives and targets, its
detailed programme of testing , its technical features of the control system and its decision
criteria should be carefully formulated in written and graphical form.
Thereby, the quality control procedure is clear, and adaptation to changed situations,
whenever necessary, is facilitated. The enclosed figures 2, 3 and 4 are examples of such
documentation.

Fig. 2 Inspection and Testing Plan.

Fig. 3 Quality Control Plan (extract).
Fig. 4 Plan Qualite.

GENERAL PROCEDURES AND CONTROL PARAMETERS
The following sections give a general survey of the most important control positions for raw
materials, intermediate and final products in a cement plant (emphasis on dry process).
Reference is being made to the schematic plant layout in figure 5.
Fig. 5 Cement Plant - Material Flow.

Fig.6 Sampling Tower

Fig.7 Automatic on-line Control of Raw Mix Composition
In a concise form, the purpose of control, typical control parameter and typical techniques for
sampling and testing are given, as well as additional comments where appropriate. In view
of the large variety of materials involved, of plant layouts, processes, sampling and testing
methods, the survey does not apply to all cases and does not attempt to be complete.
4.1 Supplied Materials

Purpose
The testing of supplied materials such as raw mix correctives, gypsum, mineral additives
(slag, fly ash, etc.), combustibles, serves to verify their compliance with delivery contracts
and hence their suitability for the anticipated utilisation.
Procedure
The extent of sampling and testing varies strongly, depending on type of material, its source
and delivery contracts. Whenever possible, sampling should be done from conveyed
material rather than from trucks or stockpiles.
4.2 Quarry
Purpose
Control of the materials composition prior to excavation enables selective quarrying to
achieve
correct composition of (integrated) stockpiles
medium to long-term uniformity of stockpile composition
optimum utilisation of materials and equipment.

Control parameter
Chemical composition, often with emphasis on specific elements such as SiO2, MgO,
SO3, alkalis, C1.
Sampling
Usually blast drill hole dust.
Process interaction
Based on manual or computerised data evaluation:
selection of area to be blasted
selection of (blasted) material to be fed to the crusher.
Alternatives
Information on the chemical structure of the quarry can also be obtained from analysis of
crushed stone by online bulk material analysis (neutron activation technology) or mechanical
sampling followed by conventional analysis.
4.3 Pre-blending
Purpose
Blending of raw components on (integrated) pre-blending stock piles to a given target
composition.
Control parameter
Chemical composition, moisture.
Process interaction
Selection and dosage of material fed to the crusher.
Procedure
Today, two different concepts are applied to monitor and integrate the composition of the
material fed to a pre-blending bed.
Sampling tower
Mechanical sampling of 1-2% of the materials feed and subsequent steps of sample
comminution, drying where required, and splitting in a sample tower (Fig. 6), to receive
typically 1 kg of fine, sample per hour for analysis.
Bulk material analysis
Continuous chemical analysis of the entire or partial materials stream by means of a bulk
material analyser applying the technology of prompt gamma neutron activation analysis
PGNAA (see reports MA 89/3702/E and VA 92/5975).
4.4 Raw Meal

Purpose
The raw meal control serves the following objectives:
Blending of components to obtain the target composition of raw meal and clinker.
Achievement of target meal fineness to obtain an appropriate clinker burnability.
Achievement of a sufficient uniformity which, together with the homogenisation in the
subsequent silo, results in a high kiln feed uniformity.

Control parameters
Chemical analysis, and parameters derived thereof such as lime saturation, silica ratio,
alumina ratio, potential clinker mineral phases C3S, C3A, etc.
Fineness, usually in terms of sieve residues (e.g. on 200 µm and 90 µm sieves)
Residual moisture content.
Process interaction
Adjustment of feed rate to the mill
off-line: manual, based on mix calculation
on-line: automatic, based on blending programme (fig. 7).
To be observed: Dynamics of grinding process (retention time, internal
circulation, dead time of analysis).
Sampling
Usually, automatic compositing sampling from a chute or an airslide is applied.
Standard samplers, typically screw samplers or airslide samplers are used, with or
without integrated mixing facilities.
Frequency
Typically every ½ to 2 hours.
Testing
Chemical composition: X-ray fluorescence
complexometric titration
carbonate titration
Fineness: Sieving, laser granulometer
Alternative techniques
Bulk material analysis (PGNA technology) can be applied to continuously monitor the
composition of the mill feed or the raw meal.
Comments
The set points of the raw meal composition have to be such that the target composition of
the clinker is being obtained. Allowance has therefore, to be made for:
the coal ash incorporation in clinker
the kiln dust absorption in the raw mill (kiln dust composition can significantly deviate
from the kiln feed composition)
the discarding of kiln dust
systematic errors in sampling and analysis.
The primary target is the clinker composition. If any of the above factors change, the raw
meal set-point has to be adjusted accordingly.

A practical example for a case where such changes have evidently occurred after a long kiln
stop, but without any reaction on the raw meal set point, is given below.
Plant with oil firing Monthly average values.
Raw meal Clinker Clinker
LSF LSF CaOF
December 99.1 98.1 1.4
January 98.0 96.8 0.8
February 97.8 97.1 0.9
Kiln stop
June 96.0 99.7 2.1
July 97.0 100.0 1.9
August 97.2 100.4 2.1
As an illustration for the compositional range of industrial clinkers, the lime saturation, silica
ratio and alumina ratio of the clinkers produced in the Holcim group are graphically displayed
in figures 8 and 9
Fig. 8 Clinker Composition (Alumina).

Fig. 9 Clinker Composition (Lime Saturation).
4.5 Kiln Feed

Purpose
Kiln feed control has only informative character. It provides guideline information (LS, SR,
AR) for kiln operation on currently processed material, as well as indications on the raw
material blending and homogenising efficiency. Figure 10a and 10b are an illustration of a
poor kiln feed uniformity, caused by poor raw mill control and low homogenisation silo
efficiency.

Fig. 10a Kiln Feed.
Fig. 10b Raw Meal.

Procedure
Usual sampling positions are in front of a kiln feed airlift (e.g. with screw sampler) or from a
belt weigh feeder (manually). Sampling frequency varies according to the plant specific
requirements form e.g. 2 samples per shift to no control at all. Normally, the same chemical
and fineness parameters are determined as for the raw meal.
4.6 Hot meal

Purpose
The control of the hot meal composition can have various objectives:
Determination of concentration of volatile elements (SO3, alkalis, C1) in the case of kiln
systems affected by build-up formation in the pre-heater/kiln inlet area. The level of
volatile elements can provide indications or, in severe cases, can serve as control
parameters for the operation of the kiln, with the objective to reduce the tendency for
build-up formation. Moreover, they indicate changes in the absolute and relative input of
circulation elements with raw meal and fuel, and corrective actions can - if possible - be
initiated.
Determination of non-burnt combustibles introduced with the kiln feed or the secondary
firing. This can be of interest for assessing the combustion efficiency of pre-calciners, or
in the context of problems associated with circulating elements (locally reducing burning
conditions).
Determination of the degree of pre-calcination in pre-calciner and pre-heater kilos.
The hot meal control can be carried out on a regular basis (e.g. 1 per shift), or in the context
of specific investigations, depending on the prevailing plant situation.
Sampling
Hot meal is usually sampled manually from the meal duct downstream the lowest cyclone,
with a cup welded to a steel rod. For determination of remaining carbon, quenching of the
sample within the sampling cup is advisable. Fig. 11 shows a sketch of a sampler with a lid
to facilitate quenching without water spilling into the sample.
Fig. 11 Hot Meal Sampler.

Analysis
Accurate and sufficiently rapid analyses of the elements SO3, alkalis and C1 require
analytical instruments which do not necessarily belong to the standard equipment of a plant
laboratory. Suitable methods are:
SO3 : LECO sulphur analyser
K2O, Na2O : Flame photometry; atomic absorption spectroscopy
C1 : Potentiometric titration.
The determination of these elements in pressed tablets by Xray fluorescence spectroscopy a
very fast method available in most plant laboratories is less accurate. Reasons for the
inaccuracy are strong matrix effects and the lack of suitable calibration standards. Hot meals
applied as calibration standards are chemically unstable (hydration, carbonisation) and
hence not suitable. Better suited are artificial calibration standards prepared on the basis of
raw meal, possibly also of clinker.
However, for practical purposes, the accuracy achievable with the XRFmethod can still be
acceptable for a correct assessment of the prevailing situation regarding the levels of
circulating elements.
4.7 Clinker
Purpose
The objectives of clinker control are:
Assurance of an adequate clinker quality in view of the final product cement
Support for an efficient kiln operation at low heat consumption.

The primary criterion of the clinker quality is the free lime content; the main consequences of
inadequate free lime as a result of too weak or too hard burning are:
in the case of too high free lime:
Loss in strength potential
Increase of cement expansion (Le Châtelier, autoclave), to the extent of nonconformity
with specifications
Disturbances in cement grinding
in case of too low free lime:
Loss in cement reactivity
Excessive heat consumption
Poor grindability.
An adequate target free lime content has to be established individually for each production
line (and clinker type). The following statistical figures for the average free lime content of
the clinkers from the Holcim Group plants illustrate the range encountered in industrial
clinkers (Fig. 12):
CaOf: x 0.96%
s 0.45%
min. 0.20%
max. 2.16% (n = 71)

Fig. 12 Free Lime in Industrial Clinkers (Holcim Group Plants 1988).
The target free lime content should be such that it permits a stable and economic kiln
operation and that an appropriate cement quality can be achieved. Upper limits have to be
defined, above which the cement quality will be seriously affected; clinker not meeting this
internal specification should be stored separately and returned to the process in a controlled
way.
The extent of clinker control too has to be established individually for each situation,
depending on the uniformity of materials flow and general stability of kiln operation.
Frequent control (hourly or even more often), possibly with automated systems, can though
need not necessarily in all cases lead to reduced fuel consumption and hence appreciable
cost savings. The signals of the automatically determined control parameters (free lime, litre-
weight) can be transmitted to and be utilised in kiln automation systems.
The free lime determination can be substituted or complemented for process control
purposes by litre-weight or alkali content determination. In a number of cases good
correlation’s are also found with NOX measurement.
Clinker sampling
Clinker is still mostly sampled manually from a conveyor belt. Automated clinker sampling
can either be done along the clinker transport or inside of grate coolers (Fig. 13).
Fig. 13 Clinker Sampling.

Free lime determination
In addition to the established laboratory methods for manual free lime determination by
titration (ethylene glycol method), there are different systems for automated free lime
determination:
Conductometric free lime determination
Automatic free lime analysers, particularly designed for installation outside the laboratory,
e.g. close to the clinker cooler, are offered by two suppliers in France, namely ISYS and
ITECA. The layout is shown in Fig. 13.
In a usual arrangement, clinker samples are taken underneath the cooler grate. Free lime is
determined by extraction of the ground clinker with ethylene glycol and measurement of the
resulting conductivity of the solution. A measuring cycle takes about 6 to 10 minutes. The
results are transmitted to the central control room or laboratory.
Such systems are, with good experience, fairly widely applied in French and Belgium plants.
It should be noted that the conductivity measurement is influenced by soluble alkali
sulphates. The calibration has to account for this, though a more or less uniform alkali
sulphate content in clinker is a prerequisite for reliable measurements.
Free lime determination by X-ray diffraction
This method is applied in a number of plant laboratories, usually in connection with the X-ray
fluorescence spectroscopy and quite often integrated in automatic systems. The chemical
composition and the free lime content can be determined on one and the same pressed
tablet by XRF and XRD respectively. For the free lime determination, the peak intensity or
integrated intensity of the CaO 200-reflection (or 220-reflection) is measured.
Calibration curves have to be established individually for each plant; due to the chemical
instability of CaOfree containing calibration standards, secondary standards containing
NiO are usually applied.
The accuracy of the method as compared to the wet chemical methods is not excellent but
mostly sufficient for practical purposes (standard deviation of linear regression in the order of
0.1 0.15% CaOf).
It has to be noted that only CaOfree is determined, in contrast to the wet chemical
or conductometric method, where the sum of CaOf and Ca(OH) 2 is measured.
Litre-weight determination
The determination of the clinker litre-weight on a sieved fraction typically 4 8 mm is a fairly
fast method frequently used for clinker control on an hourly or two hourly basis. It is usually
complemented by free lime determination on shift or daily composite samples.
The litre -weight determination can only be applied, if there is a sufficiently good correlation
with the free lime content. From Fig. 14 and 15 it can be derived that the free lime content is
generally a reliable indicator of the degree of burning, whereas the litre-weight is not in all
cases.

Fig. 14 Free Lime Content Soft and well burnt Clinkers.

Fig. 15 Litre-Weight (Soft and well burnt Clinker).
In those cases where the litre-weight is a good indicator of the degree of burning, the
correlation with free lime can vary in dependence with the clinker composition (Fig. 16).
Furthermore, the linear increase of the litre-weight with harder burning can inverse at very
hard burning (Fig. 17).

Fig. 16 Correlation between Free Lime content; Litre-weight and Silica Modulus.
Fig. 17 Correlation between Free Lime content and Litre-weight for over burnt
Clinker

At a Holcim Group plant in Switzerland, the manual litre-weight determination has
successfully been replaced by an automatic online system, measuring the apparent clinker
density with a nuclear densitometer. The system is schematically shown in Fig. 18a, along
with the correlation of automatic and manual litre-weight measurement (Fig. 18b). It is worth
noting that the density measurement is used as the primary set point parameter for the
automated kiln control system.
Fig. 18a On-line Litre-Weight Determination at Eclepens
Circuit diagram for the continuous measurement of the apparent density of clinker.

Fig 18b Manual and Automatic measurements.
Other control parameters

In a few plants, the levels of K2O or SO3 in clinker have been found to be the most sensitive
clinker parameter with respect to the burning degree. The clinker is analysed in these cases
on a hourly or two hourly basis by XRF spectroscopy.
Moreover, a correlation between the NOx level in the kiln gas after the pre-heated and the
free lime content of the clinker was established in several investigations (Fig. 19). However,
this correlation can be disturbed by various factors influencing the NOx measurement, and
can, therefore, not make a clinker control altogether redundant.

Fig. 19 NO and Free Lime.
4.8 Ground Cement

Purpose
The control of the ground cement serves to ensure the designed cement quality in terms of
proportioning of constituents and achievement of the required mechanical and physical
properties. Since the determination of the latter is time consuming, the process control has
to be based on chemical, fineness and possibly temperature parameters, for which a
correlation with the cement performance has been established.
Control parameters
For process control:
• Fineness:Blaine
Residues (e.g. R45 ~m)
Parameter from particle size
distribution (slope, size
fraction)
• SO3: for gypsum dosage
• Selected elements: for dosage of mineral additives
• Temperature: to control gypsum dehydration
For quality control
Standard chemical, mechanical and physical testing; usually on composite sample
(e.g. daily).
Process interaction
Adjustment of feed rate and of separator speed, usually in off-line, but increasingly also in
online mode.
Sampling
Automatic or manual sampling from chute, airslide or conveyor belt. Various standard types
of samplers are used, typically screw samplers or air-slide sampler.
Frequency
Typically one sample per 1 to 4 hours.
Testing methods
SO3: XRF, LECO
Chemical composition: XRF
Blaine: - manual or semiautomatic Blaine meter
- on-line Blaine meter
- on-line Gamma granulometer
Particle size distribution: Laser granulometer.
4.9 Dispatched Cement
Purpos
e
Sampling and testing of the dispatched cement serves as a final verification of the cement
quality. Testing of dispatched cement can be a requirement of product conformity
certification schemes (so-called auto-control), or it can be requested in the context of
delivery contracts to specific customers. Samples can be retained for a certain period
without testing, as a reference in case of complaints.
Control parameter
Chemical composition, standard mechanical and physical properties.
Sampling
Manual or semiautomatic from silo outlet, loading device or bags. Sampling procedures
described in pertaining cement standards.
Frequency
Daily or according to applicable sampling schemes. The European standard ENV 197
requests a minimum of 2 samples per week and type of cement for assessing standard
conformity.
4.10 Ground Coal
Purpose
The control of fine coal is required to ensure adequate flame properties (fineness), to
verify the calorific value, as well as to verify ash and sulphur content which influence
clinker composition.

Control parameters
For process control:
Fineness: sieve residues R90 µm, R200 µm
Moisture
for verification and trend analysis:
calorific value
proximate analysis (moisture, volatiles, fixed C, ash)
S, Cl, if required H, N (ultimate analysis)
ash composition.
Sampling
Manual slitsampler or automatic screw sampler.
Frequency
Every 1 to 4 hours for fineness. Other tests on composite sample (e.g. daily).
Testing
♦ Fineness : Sieves
♦ Calorific value : Calorimeter
♦ Proximate analysis : According to pertaining standards
: Automatic proximate analyser
♦ Ultimate analysis : According to pertaining standards
♦ Sulphur : Gravimetric (fusion)
♦ Chloride : Titration.
UNIFORMITY
Significance
Uniformity is one of the most important prerequisites for smooth and efficient kiln operation.
Excessive fluctuations can easily upset the often less than robust balance of appropriate
burning behaviour, thus giving rise to operational disturbances. Furthermore, since the
clinker should always be adequately burnt, unnecessarily hard burning is practised for much
of the production time.
Consequences are:
high fuel consumption
increased tendency to ring and build-up formation
reduced brick lifetime
fluctuations in product quality

Sources of fluctuations
♦ Raw components: chemical and mineralogical composition
♦ Raw meal: chemical and mineralogical composition, fineness;
disturbances in feed rate to the mill
♦ Combustible: ash content, sulphur content, calorific value, fineness
♦ Dust return: different modes of return during direct or indirect
operation
♦ Feed rates: equipment related fluctuations at constant settings
(Kiln feed,
combustibles)
Abrupt manual adjustment of kiln operation parameters
Abrupt and excessive adjustment initiated by automatic control loops.
Acceptable variability of kiln feed composition
Since the reduction of material variability is associated with costs for homogenisation and
control, an acceptable, optimised target has to be established for each specific situation.
Raw meals or clinkers of an extreme composition (e.g. LS close to 100, high SR, low AR) or
with unfavourable fineness characteristics require more stringent variation limits than easy to
burn clinkers. What eventually can be achieved depends on the complexity of the raw
material situation, the pre- blending and homogenising facilities as well as on the efficiency
of the quality control system.
Some values which are considered to provide a good basis for kiln operation, even when
using difficult materials, are given below.
Guide values for feed variations
Accuracy of the feed system: 1 to 1.5%
Regular calibration of the system mandatory
Guide values for good chemical uniformity of the kiln feed
(for random samples over 24 h; on raw meal rather than clinker basis)
Without analytical error Including analytical error

Titration method s = 0.2 % CaCO3 s = 0.25 % CaCO3
max. s = 0.3 % CaCO3 s = 0.35 % CaCO3
XRF method: s = 0.2 % CaO s = 0.15 % CaO
max. s = 0.17 % CaO s = 0.20 % CaO

SiO2 s = 0.07 % SiO2 s = 0.2 % SiO2
Al2O3 s = 0.07 % Al2O3 s = 0.2 % Al2O3

Fe2O3 s = 0.04 % Fe2O3 s = 0.2 % Fe2O3
LS s = 0.7 s = 1.2
SR s = 0.04 s = 0.06
AR s = 0.06 s = 0.08
Means of improving uniformity
The kiln feed uniformity is a result of various factors along the preparation process, starting
from the raw material deposits and going through several stages of homogenisation and
blending.
Possibilities for improvements can be found in the following areas:
Exploitation planning
Medium to long-term exploitation planning, based on accurate raw materials inventory.
Application of computer based methods.
Quarry scheduling
Short-term quarry scheduling, based on blast hole dust analysis. Application of computer
based methods.
Raw material and coal pre-blending
Blending control
Blending control at integrated pre-blending stockpile and raw mill. Appropriate selection
of sampling regime and analytical methods. Various degrees of automation from manual
to online, or even to ‘real-time’ with bulk material analyser.
Raw meal homogenisation

EXTENDED QUALITY CONTROL
In addition to the basic test methods for routine control, a large variety of sophisticated
methods and procedures is applied for material technological investigations in connection
with special manufacturing and application problems and in research and development.
Characterisation of raw materials, clinker, additives, etc. is not only limited to chemical
analysis, but also the mineralogical and textural features are important. Thermo-analytical
(DTA, TGA), microscopic (light and electron microscope) and X-ray diffraction (XRD)
methods are used as routine procedures in this context. Microprobe is a powerful tool to
determine the elemental composition of individual clinker minerals. Burnability and
volatility tests are used to assess the raw mix behaviour in the industrial kilos.
Although most of these methods are only applied in central or research laboratories, some of
them are also used in a number of plant laboratories for routine quality control, such as X-ray
diffraction to monitor the quartz and clay minerals of raw materials and to examine free lime
in clinker; or differential thermal analysis to control the dehydration of gypsum in cement
during the grinding process. However, if these specialised testing procedures are promoted
too much, there is the danger that they might finally be included in the standard
specifications. This has to be avoided because it would make the quality control procedures
too sophisticated and expensive for the majority of plants.
In addition to the traditional tasks and methods of quality control, in many plants the scope of
control is significantly extended, not only due to more stringent process conditions and to
higher quality requirements and special requests by the customers. The following tasks
require extended application of already established quality control methods or even
introduction of new techniques:
Refractories testing, to ensure the material quality of the installed kiln brick lining which
requires apart from visual inspection extended chemical analysis, simple physical and
thermal tests.
Quality control of additives for production of blended cements; depending on the kind of
additive, its source and supply conditions, more or less extensive chemical,
mineralogical and performance tests have to be carried out; with the increasing use of
industrial by-products, such as fly ashes, slags, etc., efforts are also undertaken to
establish reliable and rational methods for their control and for their determination in
blended cements.
The Blaine value for control and characterisation of cement fineness with respect to grinding
process and cement quality is only of limited significance. therefore, grain size analysis
by laser diffraction or automatic sedigraphs is used increasingly for quality and process
control; in the future, the laser diffraction technique might even be used in online
installations.
Use of waste materials as raw mix components and particularly as combustibles (e.g.
tyres, colliery wastes, battery casings, chemicals, etc.) require adaptation of existing or
new testing methods.
Last but not least environmental concerns and the resulting pressure upon industries force
plants to extend their quality control into the area of emission control (and, in some
instances, even emission control); this implies development and refinement of
techniques for gas analysis especially for SO2 , NOX, but also for hydrocarbons as well
as for dust determination and collection; analysis of traces in the ppm or even the ppb-
range for toxic elements such as Ti, Cd, Pb, Hg, etc., nowadays is not an unusual task
for the cement industry.

Chapter 10
Basic Principles of Sampling and Testing

1. Basic Principles of Sampling and Testing...................................................................324
2. Sampling Methods and Equipment..............................................................................340

Basic Principles of Sampling and Testing
1. INTRODUCTION............................................................................................................325
2. SAMPLING REGIME.....................................................................................................325
2.1 Basic considerations.................................................................................................325
2.2 Guidelines for sample size and sampling frequency................................................. 327
3. RELIABILITY OF TEST RESULTS................................................................................331
4. ANNEX 1........................................................................................................................ 334

INTRODUCTION
When dealing with bulk materials, the properties of a certain quantity of material, such as the
contents of a preblending bed, a shift's cement production or a coal delivery, cannot be
determined on the entire quantity, but have to be estimated on the basis of measurements
carried out on a limited number of individual samples. The reliability of such an estimate
depends on the heterogeneity of the material and on the procedure of sampling, sample
preparation and analysis. A high reliability - as may be required for production or market
purposes - is usually associated with comprehensive sampling, with precise preparation
procedures and analytical instruments, and may thus incur considerable costs. The objective
of a well designed testing procedure is, therefore, to extract the necessary information on
quality parameters of a material at the lowest possible costs.
This paper reviews various aspects relevant to the design of an appropriate sampling
regime, and summarizes the main factors influencing the reliability of test results. Emphasis
is given to practical aspects rather than to the theory and statistics behind them; for the
latter, reference should be made to the paper "Statistics" contained in the seminar
documentation.
SAMPLING REGIME
2.1 Basic considerations

In order to decide on an adequate sampling regime and corresponding equipment, some
considerations are necessary concerning the objectives of sampling, the representativity of
samples from bulk materials and the material characteristics to be determined.
2.1.1 Objectives
Bulk materials can be sampled either for a mere determination of the characteristics of a
certain basic material quantity or for the purpose of process control.
Characterization of a given basic material quantity
Examples of such materials are:
♦ Area of quarry to be blasted
♦ Materials delivered to the plant (coal, slag, gypsum) ♦
Kiln feed
♦ Dispatched cement
The sampling regime has to be adapted to the given variability of the material and to the
required precision of the measurement. The complexity of sampling can vary from
occasional manual grab samples to sampling regimes prescribed in norms (bulk cement) or
agreed upon between seller and purchaser.
Control of process
Examples:
♦ Preparation of preblending bed ♦
Raw and cement mill control ♦
Clinker control

The sampling regime, in particular the sampling frequency, has to be adapted to the target
variability of the output material, to the variability of the input materials and to the dynamics
of the process, i.e. how fast and how precise corrections can be made and how fast they can
be observed (dead time) . This type of sampling can readily be automated, which permits a
comfortable precision to be attained.
2.1.2 Representativity of samples
In order to obtain meaningful information from a sample on the characteristics of the basic
material quantity, the sample must exhibit these characteristics in a representative way.
Representative means that the value for a certain characteristic determined on the sample
differs only randomly - and not systematically - from the corresponding mean value of the
basic material quantity. For representative sampling, it is normally a prerequisite that all
volume sectors of the material have an equal chance of being sampled, and that all particle
sizes are correctly represented.
Representative sampling of bulk material is often rendered difficult, especially when
stationary material is concerned; some of the reasons are:
Large volumes can be inaccessible when sampling materials on stockpiles, in silos, on
ships or on trucks
Stationary materials can be stratified due to sequential deposition of different materials
Materials with a broad particle size range tend to segregate, i.e. fines concentrate on the
bottom of a material heap. If coarse and fine particles differ in characteristics of concern
(e.g. composition), considerable inhomogeneities will occur, which can be an important
source of sampling error. Excessive moisture behaves in a similar way.
Separation effects also occur when materials with a broad particle size range are thrown off
a quarry face or drop from one conveyor belt onto a second one running in
perpendicular direction. In the latter case, strongly inhomogenous material streams
result.
A similar effect occurs when two material streams combine on a conveyor belt, e.g.
returned filter dust added to the finish cement conveyor belt.
Sampling from stationary materials should be avoided whenever possible, and replaced by
sampling during loading or discharge over conveyor systems. If this cannot be done,
preference should be given to traverse sampling, i.e. a primary sample should embrace the
whole cross section.
Examples for traverse sampling are:
Blast hole sampling in the quarry
Auger sampling of a railway car
Sampling from large stockpiles can be done according to the recommendations given in
norms (e.g. ASTM D 75-82 for sampling aggregates) or according to special agreements.
For the reasons mentioned above, due attention must be paid to representative sampling of
conveyed materials, independent whether manual or automatic equipment is used.

2.1.3 Selection of characteristics to be measured
To keep costs to a minimum, it may be adequate to limit the number of characteristics to be
determined. Preference should be given to characteristics easy to determine, which strongly
correlate with other material characteristics or with process conditions.
Examples are:
Ash content of coal correlating with calorific value

CaO- and Fe2O3-content correlating with remaining elements
in raw-meal
Literweight of clinker correlating with sinterzone
conditions
The validity of correlation has to be periodically verified (e.g. on daily composite samples).
It is evident that the chosen characteristics must not be altered during sample preparation.
2.2 Guidelines for sample size and sampling frequency

The larger the number of samples taken and the larger the extracted quantity per primary
sample, the more accurate the sampling of a basic material quantity.
Thus, the minimum number and quantity of samples, which are required for a specific
precision of the mean value of a characteristic, have to be determined. They depend upon:
the grain size
the heterogeneity of the material
In a first approximation, the following applies:
The minimum size of a primary sample can be related to the maximum grain size of the
material. The German DIN specification states that the minimum weight of the primary
sample must be at least 10 times the weight of the largest grain. This third power
relationship between sample mass and maximum grain size is usually applied in linearised
graphical form, as depicted in Fig. 1.

Fig. 1 Maximum grain size versus sample quantity
This graph mainly applies for coarse grained raw materials. For homogeneous fine ground
materials such as raw meals or cements, the recommended sample quantity is a few
hundred grams to 1 kg. For homogenous coarse grained materials, such as clinker, 1 to 3 kg
are sufficient.
It may be noted that various national norm specifications relevant to the mining industry
exist. The sample size / grain size relationship can differ considerably from that in Fig. 1, as
illustrated by the example of manual sampling with norm scoops shown in Fig. 2.

Fig. 2: Japanese Industrial Standard M 8100
Sampling scoops and their capacities in kg for sampling bulk solids

Top Scoop Width Approx. Weight (kg) of scooped
Size sample at various bulk
densities
(mm) designa (mm) (ml) 0.75 1.5 2.0
tion
3 3 3
t/m t/m t/m
5 5 50 75 0.06 0.11 0.15
10 10 75 125 0.09 0.19 0.25
15 15 100 200 0.15 0.3 0.4
20 20 100 300 0.22 0.45 0.6
30 30 150 400 0.3 0.6 0.8
40 40 150 800 0.6 1.2 1.6
50 50 260 1700 1.3 2.5 3.4
The number of primary samples - or the frequency of sampling - directly bears on the
sampling precision, since the sampling variance of the mean of n-samples is n-times smaller
than that of a single sample.
Hence, if the variance is known for a given basic material quantity - e.g. determined by means
of a sampling experiment - the number of samples can be calculated for any desired sampling
precision. This rule, however, only applies if all samples come from the same population. If the
mean of the observed characteristic changes with time (trend!), samples taken at different times
can no longer be regarded as originating from the same population.
Fig. 4 illustrates a reasonable sampling frequency with respect to a given materials
variability, as well as too frequent and insufficient sampling.
Automatic sampling normally ensures a more than sufficient sampling frequency.

Fig. 4 Sampling Frequency
A different way to determine optimum size and number of primary samples is based on a
sampling experiment in which a sufficient number of simultaneous pairs of samples of
strongly different quantities are taken.
A rough evaluation of the test results - visualized in Fig. 5 - may already indicate whether
much more precision is gained with a larger sample size or not. An exact evaluation can be
made applying the sampling theory of Yisman [1]; an example is given in Annex 1.

Fig. 5
RELIABILITY OF TEST RESULTS

As illustrated in Fig. 6, sampling and testing procedures consist of several independent
steps. The reliability of the final test result depends on the reliability of each individual step.
Three types of errors can occur at each stage, namely systematic errors, random errors and
gross errors. They are characterised as follows:
Systematic errors:
Systematic errors cause a bias of the estimated value for a certain characteristic with
respect to the (unknown) true value. They are caused by incorrect (non- representative)
sampling and sample preparation, by the malfunctioning of instruments, by inadequate
calibration, or by person specific peculiarities in performing the tests.

Fig. 6
Systematic errors can be detected by cross-checking with alternative, independent testing

procedures.
Examples:
Campaign of comprehensive manual sampling to test representativity of automatic
sampler
Norm-conform CaOf-determination to cross-check conductometric determination.
Systematic errors can either be of relevance - which requires their detection and elimination
-, or they can be irrelevant, usually under the condition that the error is approximately
constant.
Example:
A systematic error in determining the C3A-content may be irrelevant when producing
cement of ASTM Type I, but rather problematic when producing Type II or Type V.

Random errors:
Random errors arise due to the inherent heterogeneity of the sampled material and due to
the imprecision of sampling, sample preparation and analytical techniques. The amount of
random errors may be expressed by statistical measures of variability, such as variance or
standard deviation (refer to paper "Statistics").
Random errors cannot be eliminated, but they can be reduced by :
Repetition, i.e. extraction of more primary samples, performance of measurements
on several test samples, repeating measurements on individual test samples.
2
The variance (s ) of the mean value of n measurements is n-times smaller than the
variance of a single measurement
2
s2 (x)=s (x)
n
Applying more precise procedures and instruments for sample preparation and analysis
Random errors of a sequence of independent testing activities add up to the total random
error of the final result according to the law of error propagation (see Fig. 6).
s total = s 2 + s 2 + .....s 2
1 2 n
From this law of error propagation, the following important conclusions can be derived:
an error which has been committed in the testing procedure cannot be compensated by
increased precision in the subsequent testing steps
the individual errors of the sequence should be as small as possible and of similar size:
There is no point in increasing the precision of one step as long as others are decisive
for the overall error. Therefore, it is necessary to determine the precision of the individual
steps in each control.
Gross errors
Gross errors arise from the wrong performance of procedures, the wrong reading of scales,
errors in copying data, miscalculations etc. Such errors can considerably influence statistical
evaluations. They may show up as outliers in a series of data.

LITERATURE:
[1] Merks J.W.: Sampling and weighing of bulk solids.
Trans Tech. Publications, 1st Ed., Clausthal-Zellerfeld, Germany (printed in USA)
Additional literature:
Hasler R.:
Some aspects of process technology of raw material preparation.
HMC-report PA 74/6167/E
Gy P.:
L'échantillonnage de minérais en vrac.
Mémoires du Bureau de Recherches Géologiques et Minières No.
67, Paris 1971
ANNEX 1
OPTIMUM SIZE AND NUMBER OF INCREMENTS ACCORDING TO THE THEORY OF
VISMAN [1]
The total variance of a sampling and testing procedure is split up into the terms:
2
(d ) +
s2 (t )= s (c) + +s2 (p) + s2 (a)
s2 am ⋅ n n
=s2(s)+s2 (p)+s2 (a)
in which 2 = total variance of mean value

s (t)
2 = sampling variance
s (s)
2 = composition variance
s (c)
2 = distribution variance
s (d)
2 = preparation variance
s (p)
2 = analysis variance
s (a)
am = size of primary sample (kg)
n = number of primary samples
The composition variance and the distribution variance can be determined by means of a
sampling experiment, in which a number ( ~ 30) of pairs of primary samples differing strongly
in size are taken and analysed.
=am(1) ⋅ am(2) ⋅ [s2 (1) − s2 (2)]
s2 ()
c am(2) − am(1)
)=s2 (2) − am(1) ⋅ [s (1) − s (2)]
2 2
s2
( am(2) − am(1)
d

2 = composition variance
in which s (c)
2 = distribution variance
s (d)
2 = variance for set of small samples
s (1)
2 = variance for set of large samples
s (2)
am(1) = average mass of small samples
am(2) = average mass of large samples
The optimum size of primary samples is then given by
s2 (c)
m(opt )
s2 (d )
and the optimum number of primary samples to obtain a required total variance:
n= s2 (d ) + s2 (c) / m(opt ) =s2 (d ) + s2 (c) / m(opt )
s2 (t ) − [s2 (p) + s2 (a)] s2 (s)
Example (compiled from [1]):
A coal delivery is sampled by means of 24 pairs primary samples of about 10Og and 50 kg
size. The primary samples of small size are entirely dried and ground to analytical fineness.
The primary samples of large size are dried, crushed and split and only a partial samples
ground to analytical fineness. The ash content is determined in duplicate. The results are
given in tables A1 and A2.

Table A1: SAMPLING EXPERIMENT WITH SMALL SAMPLES
Sample Mass Ash on Dry Basis Mean
Number in g First Second ∆
1 89 14.20 14.23 0.03 14.22
2 126 13.27 13.66 0.06 13.69
3 152 18.44 18.42 0.02 18.43
4 109 15.87 15.75 0.12 15.81
5 149 13.71 13.81 0.10 13.76
6 87 17.10 17.13 0.03 17.12
7 110 13.64 13.70 0.06 13.67
8 142 19.71 19.64 0.07 19.68
9 111 16.34 16.31 0.03 16.32
10 140 12.49 12.38 0.11 12.44
11 121 12.19 12.21 0.02 12.20
12 150 10.73 10.75 0.02 10.74
13 160 23.24 23.20 0.04 23.22
14 96 15.71 15.80 0.09 15.76
15 105 21.83 21.89 0.06 21.86
16 132 11.78 11.73 0.05 11.76
17 141 12.14 12.05 0.09 12.10
18 123 19.84 19.67 0.17 19.76
19 135 8.34 8.37 0.03 8.36
20 150 12.80 12.85 0.05 12.82
21 108 14.02 13.91 0.11 13.96
22 99 9.35 9.31 0.04 9.33
23 102 12.46 12.38 0.08 12.42
24 129 13.15 13.18 0.03 13.16
Mean 123.9 0.063 14.69
Variance 0.0031 14.1448
Standard Deviation 0.0558 3.7610

Coefficient of Variation 25.6
Variance of the mean 0.5894
Standard deviation of the mean 0.7677

Coefficient of variation of the mean 5.2%
Sampling variance : total variance minus variance of analysis
2
s (s) = 14.1448 -0.0031 = 14.1417

Table A2: SAMPLING EXPERIMENT WITH LARGE SAMPLES
Sample Mass Ash on Dry Basis Mean
Number in g First Second ∆
1 53.0 13.61 13.34 0.27 13.48
2 53.2 13.15 13.46 0.31 13.30
3 55.9 14.34 14.20 0.14 14.27
4 41.1 15.33 15.26 0.07 15.30
5 46.1 14.17 14.51 0.34 14.34
6 40.6 13.60 13.41 0.19 13.50
7 48.8 14.76 14.38 0.38 14.57
8 50.2 14.25 14.72 0.47 14.48
9 55.7 12.54 12.29 0.25 12.40
10 48.1 13.02 13.21 0.19 13.12
11 43.4 14.33 14.74 0.41 14.54
12 48.4 13.70 13.63 0.07 13.66
13 45.3 12.80 12.68 0.12 12.74
14 40.1 13.26 13.64 0.38 13.45
15 41.8 14.09 14.41 0.32 14.25
16 48.0 13.52 13.28 0.24 13.40
17 56.2 15.44 15.11 0.33 15.28
18 50.3 15.07 15.28 0.21 15.18
19 48.0 14.40 14.26 0.14 14.33
20 44.3 13.06 13.22 0.16 13.14
21 48.5 13.22 13.47 0.25 13.34
22 41.6 14.26 13.90 0.36 14.08
23 39.9 13.18 13.12 0.06 13.15
24 52.3 13.44 13.56 0.12 13.50
Mean 47.53 0.241 13.87
Variance 0.0455 0.7903
Standard Deviation 0.2134 0.8890

Coefficient of Variation 6.6
Variance of the mean 0.0329
Standard deviation of the mean 0.1815

Coefficient of variation of the mean 1.3%
2
s (s) = 0.7903 -0.0455 = 0.7448
calculation of composition and distribution variance:
2 = 123.9 . 47530 . (14.1417 - 0.7448) = 1664.2
s (c)
47530 - 124
s2(d) = 0.7448 - 123-9 . (14.1417 - 0.7448) = 0.71
47530 - 124
Hence:

Optimal mass m(opt) = 1664.2 = 2343g 2.3 kg
0.71
(entire primary sample can be prepared to analytical fineness, thus only analytical
variance to be considered)
Optimal number of primary sample:
n= 0.71 + 1664.2/2343 = 0.71 + 0.71 = 1.42
2 2 2 2
s (t) - s (a) s (s) s (s)
Error of mean Coeff. of s s2(t) s2(s) = n
(95% conf intev.) Variation s2(t - 0.0031

± 10% 5% 0.7 0.49 0.4869 3
± 5% 2.5% 0.35 0.12 0.1194 12
± 2% 1% 0.14 0.02 0.0165 86
± 1% 0.5% 0.07 0.0049 0.0018 (800
)*
Analytical variation outweighs sampling variation, i.e. error of mean is determined by
analytical precision and not by sampling regime.

Sampling Methods and Equipment
1. SAMPLING OF STATIONARY MATERIALS..................................................................341

1.1 Field sampling...........................................................................................................341
1.2 Sampling of bulk cement...........................................................................................342
2. SAMPLING OF CONVEYED MATERIALS....................................................................342
2.1 Evaluation of the 3 categories...................................................................................343
2.2 Various possibilities of collecting a transverse sample..............................................344
2.3 Deviations from the ideal case..................................................................................346
2.4 Rules for the sampler................................................................................................347
2.5 Recapitulation of sampling rules...............................................................................352
3. CRITICAL ANALYSIS OF THE BEST KNOWN SAMPLERS........................................352
3.1 Apparatus for transverse sampling with linear movement.........................................352
3.2 Equipment for Transverse Sampling with a Rotary Motion........................................ 358
3.3 Equipment for Longitudinal Samples........................................................................360
3.4 Equipment for Quasi-Point Samples.........................................................................363
3.5 Experimental Assessment of Samplers.....................................................................365
4. HOW TO IMPLEMENT AN ADEQUATE SAMPLING SYSTEM.....................................366

SAMPLING OF STATIONARY MATERIALS
Sampling of stationary material should, whenever possible, be replaced by sampling on
conveyed material. In a number of cases, however, sampling of stationary material cannot
be avoided. Such cases are reviewed in the following:
1.1 Field sampling

The mineralogical and chemical composition of sedimentary rocks generally vary far more
rapidly in the vertical direction (stratigraphic sequence) than they do in the horizontal
direction of the layer; therefore a vertically oriented sampling is required.
1.1.1 Surface sampling

In the sequence of increasing reliability, the following methods of surface sampling on an
exposed face can be applied:
unweighted spot sample
for each rock layer of a rock complex only one sample is taken at a definite spot.
Information: only a rough judgement of the investigated material possible.
aa) weighted spot samples
more reliable information is achieved as the spot samples of each layer are weighted up
according to their stratigraphic thickness.

chip sample
small material samples are broken out of the wall face at regularly spaced intervals.
Information: reliable information on the chemical and mineralogical composition.
trench sample
the sample is obtained by cutting a trench of uniform width and depth across the face of the
exposed rock and collecting the excavated material.
Information: reliable information on chemical and mineralogical composition as well as on
moisture content.
1.1.2 Drill hole sampling
Dust drilling sample. Mainly hard, compact dry material is recovered from hammer, tri-
cone or rotary drilling.
Proper sampling requires a dust collector attached to the drilling equipment. The often
encountered procedure of dust collection around the borehole is often a source of
systematic errors due to the loss of the fine particles.
Core drilling sample

the core sample is obtained by rotary drilling with core barrel tipped with a steel or diamond
bit. This method can also be applied below the water table. A representative sample for
chemical-mineralogical investigations can only be guaranteed on a core recovery of over
90%. The cores can also be used to determine the technological properties of the rock.
1.2 Sampling of bulk cement

In addition to sampling at the material surface, it is also required to collect material from the
interior using a tube sampler. For cement sampling tube samplers are often specified in the
standard specifications e.g. ASTM C183.
SAMPLING OF CONVEYED MATERIALS

This term applies to all types of material which are transported continuously; for instance
gravel on a belt conveyor, meal in an airslide or slurry in a slurry pipe.
There are three principal ways in which a material can be taken from a material flow:
Transverse sample consists of the whole of the material flow during short, predetermined
time interval.

Longitudinal sample consists of a part of the material flow which is continuously
deflected from the main stream.
Quasi-point sample is created by the branching off of a part of the material flow during a
certain period of time.
A visual impression of these three categories can be obtained by imagining the samples
taken from a stationary belt conveyor.
or drawn 3 - dimensionally:
2.1 Evaluation of the 3 categories
2.1.1 Transverse samples

This sampling method is, in principle, correct as the average value of the individual collected
samples is almost the true average value of the "lot". The term "lot" should be understood as
that quantity to which the measurements refer, as for example the heap of a mixing bed, a
shipload, a rail wagon, a silo etc.

2.1.2 Longitudinal samples
If all particle sizes were represented in their correct relationship in that part of the stream
sampled, this type of sampling would yield a true picture of the "lot". In reality, however, the
force of gravity causes a segregation of the particle sizes to a lesser or greater extent, and
therefore excludes the fulfillment of the conditions mentioned above.
2.1.3 Quasi-point samples
This type of sample is only a combination of the preceding two types. It incorporates no new
advantages, but embodies the disadvantages of both the other categories.
The transversal sample therefore appears to be the only one promising success. It is the
only one that can fulfill the condition of having an equal chance of sampling all the particle
size fractions.
2.2 Various possibilities of collecting a transverse sample

The various possibilities are illustrated below in the form of a three-dimensional illustration of
the material flow.
A) Rectangular cut

B) Horizontal inclined cut
C) Vertical inclined cut
D) Alternating horizontal inclined cut

E) Alternating vertical inclined cut
From a theoretical viewpoint, these five cuts are of comparable value. The differences lie in
the ease of their practical application.
Type A): Can only be realised on a stationary conveyor belt or by means of a travelling
sampler.
Type B): Can be realised by means of a rotating sampler. The sampling aperture always
moves in the same direction through the material.
Type C): A bucket wheel sampler belongs to this type.
Type D): This is the most commonly found type of equipment. The reciprocating slit vessel
and the reciprocating chute types are typical examples.
Type E): This can only be realised by a radially mounted slit vessel; this type of equipment is
generally too large.
2.3 Deviations from the ideal case

The particles do not drop one behind the other, but in a more or less compact stream
and can therefore influence each other.
The particles make contact, not with a stationary separating wall, but with a moving one
and can therefore be deflected at the moment of impact.
The opening of the sampler is not of an infinite size. Processes on the one separating
wall can therefore be influenced by those on the opposite one.
The separating walls are not infinitely thin and can be bent in the course of time.
The particles themselves can break on impact. This would be of importance only when a
screen analysis is to be prepared from the sample.
The design of samplers must be aimed at overcoming all these difficulties. What form these
solutions take and what important rules must be respected, is described in the next chapter.

2.4 Rules for the sampler
(1) Slit width and speed are inter-dependent.
Width of slit B ≥ 3 d + 10 mm
d = diameter of the largest particle.
This ensures that even for fine meal, the width of the slit does not fall under 10 mm.
Speed w ≥ 300 mm/s
In special cases this can be increased to 400 mm/s, provided the slit width is increased to 4
d + 10 mm.
The conditions for the slit width must be respected for all slit dimensions.
The edges of the sample apertures should be fitted with interchangeable sacrificial
elements made of a wear resistant steel.
The slit aperture must transverse the entire cross sectional area of the material flow. The
return and stopping positions must be placed outside of this area.

The speed of the sampler during its journey through the material must be constant,
during any single traverse as well as from one traverse to another.
Quick starting electric motor with excess capacity. This drive is more suitable than a
pneumatic or hydraulic cylinder, as the latters speed cannot be easily regulated.
An acceleration zone must be provided outside the material stream.
Fig. Motor starts too slowly
Fig. Motor too week

(5) The shape of the sampler aperture must be suited to the movement.
linear movement → the side walls must be parallel
rotating movement → the side walls must be positioned radially.
(6) Measures must be taken to avoid blocking and overflowing of the sampler.
Slit vessel cutter:
The capacity of the vessel should be twice as large as the nominal sample volume. In the
filled vessel there should be a remaining free height equivalent to d or at least 100 mm.
Chute type cutter:

The vertical walls must be of sufficient height to prevent the "splashing" of particles. To
ensure a flawless material extraction, the inclination of the walls must be chosen to suit the
properties of the wettest materials encountered. Care should be taken that there are no
obstructions. Corners, joints and protruding screws are often the cause of build-ups.
Stainless steel should be used for sticky materials and if necessary provision for heating
should be made.

The integrity of a sample must assured. Once a sample has been taken, it must be
protected against all alterations, especially from the loss or addition of particles.
Loss of dust:
To reduce the drop height, the sampler should be mounted as close to the discharge point
as possible. To avoid the formation of dust, the material can be slightly moistened before it
reaches the sampler.

Addition of dust:
The whole sampling apparatus should be protected against dust. The aperture in the
sampler can be closed by a lid in its static position.
(8) Calculation of the single sample volume m.

straight movement:
B
m=M w
M = material flow [kg/s]
B = slit aperture [m]
w = speed [m/s]
rotary movement :
θ
m=M ϕ
= aperture angle [red] ϕ
= angular speed [rad/s]
Note
A sampler designed in accordance with the previous rules has a lower rated capacity. It
would be wrong to try to decrease the sample quantity by an increase in speed or y a
reduction of the slit width.

2.5 Recapitulation of sampling rules
Fundamental Principle No. 1 Principle No. 2 Principle No. 3

principles which All sectors of the All particle sizes There must be no
must be adhered material stream must have the alteration to the
to. must be sampled same chance of sample once it has
for an uniform time. being sampled. been taken.
Measures to 1) Correct shape 1) Minimum slit 1) Prevention of
maintain the of the sampler size 3d + 10mm blockages and
adherence to the aperture. overfilling.
fundamental 2) The whole of the 2) Speed generally 2) No material loss.
principles.
material flow must less than 300 mm/s 3) No material
be covered.
Stopping points additions.
must lie outside the
material flow.
3) The speed must 3) Sampling to take 4) Protection
stay constant place at a location against any

during the traverse where the fine alteration of the
and from traverse particles are not composition.
to traverse. thrown up as dust.
CRITICAL ANALYSIS OF THE BEST KNOWN SAMPLERS
1 Apparatus for transverse sampling with linear movement
3.1.1 Sampling from a stopped conveyor belt

The sample is separated from the remaining material by a frame. This system is the closest
approach to the theoretical model of a right- angle cut. As the material is stationary, L only
needs to conform to the condition of L = 1.5 d + 100 mm. The minimum sample therefore is
smaller.
M
m1= v (1 .5d + 100mm)

m1 = minimum sample from the stationary belt
M
m2= w (3d + 10mm)
m2 = smallest sample at material discharge point

m1=w 1.5d + 0.1
m2 v 3d + 0.01
v = belt speed
This function for w = 0.4 m/s appears as follows:
It can be observed that with larger particle sizes and at higher belt speeds the minimum
sample quantity tends towards the sample from the stationary belt.
In practice it is, however, impossible to stop large conveyor belts for the purpose of
sampling.
The sketch below shows a set-up which makes it possible to extract samples from a
conveyor, fully automatically, without having to stop the main conveyor.

Fig. Sampling belt

For a certain pre-set time, the material is deflected by the flap onto the sample belt. The
sample belt is stopped after the return of the flap. The sample can now be taken off the
sample conveyor and the process can be easily automated. The speed of the sample
conveyor can be freely selected and kept high. This makes it possible to reduce a sample
down to 1/lOth of the size of those from a sampler at the discharge end.
3.1.2 Hammer samplers
This equipment takes samples from a running conveyor. The hammer is designed as a
scoop to ensure that the material is not Pushed aside. Brushes fixed to the end of the
hammer clean the fine material from the belt. This type of sample is well-suited for non-sticky
material.

3.1.3 Sampler with flaps
The sectional illustrations show that the principle of uniform sampling of all material layers is
not maintained. One might think that this systematic error could be avoided by a faster
operation of the flap and by a longer residence time in the sampling position. In the first
case, an unwelcome violent movement of the material would result whilst in the second
case, the sample quantity will be very large or, with longer intervals, the integration error
becomes unacceptable.

3.1.4 Slit vessel - sampler
Properly designed equipment of this type can fulfill all conditions.
3.1.5 Chute sampler

The chute sampler, like the slit vessel sampler fulfills all require meets. It is suitable for the
sampling of slurry as well as of dry material. The chute is of a simpler design than the slit
vessel, but requires more headroom.

3.2 Equipment for Transverse Sampling with a Rotary Motion
3.2.1 Slit vessel on a radial swing
This type, when suitably adjusted for the materials properties, is precise and produces no
systematic errors. The equipment is of excellent design by virtue of the small drop height of
the material and allows the possibility to remove the sample immediately.
3.2.2 Vézin - Sampler
The design concept of this equipment is correct. It is unsuitable for installation at a discharge
point, but rather for small material flows of relatively fine grained dry solids or for slurries.
One advantage is the good separation of the sample from the remaining material.

3.2.3 Sampler with rotating distributor
This type has been developed from the Yezin sampler by means of exchanging the rotating
and stationary parts, and is therefore correct in its concept and principle. Due to the
relatively high position of the distributor tube, the material reaches the sampler aperture at a
high velocity. This is unfortunate when taking the sample and creates some dust.
Furthermore, a demixing process does take place in the drop tube due to the considerable
centrifugal forces present.
Further variations of this type are shown in the sketches below.
All these types suffer from the disadvantage that the whole mass of the undivided sample
has to be accelerated.

3.2.4 Slides and flaps in ducts
In the case of the transport of bulk materials having a low bed depth, this sampling
equipment may be sufficient.
The material lying on the flap is extracted whilst the oncoming material is stopped during the
sampling. Provided the material does not stick, and provided that the dimensions have been
well-chosen, this sampler is just right and delivers reproducible samples.
3.3 Equipment for Longitudinal Samples
3.3.1 Multi-stage sample divider

This type of equipment has been extensively used in the past and is still in frequent use to
reduce the quantity of samples. It consists of several vertically orientated samplers. That
alone has two disadvantages:
The equipment requires considerable height for its installation and
the dust which is created by the partial free fall of the material is ducted upwards due to the
chimney effect of the closed vertical casing, so that the sample on the bottom is lacking
in the finest particle sizes.

A known representative of this type is the "Koulakoff-Divider". It consists of a series of
rotating and stationary double cones.
The ratio of partition can be widely varied through an increase or decrease of the number of
steps. It can also obtain very large values. By means of precise tests, Pierre Gy was able to
establish a systematic error of up to 20% using this equipment, caused by selective
sampling of the various particle sizes.
Because of this considerable systematic error, the Koulakoff divider cannot be
recommended for use in sample preparation.

3.3.2 Sampling from screw conveyor
This type of sampling screw continuously takes part of the whole of the flow, therefore
providing a longitudinal sample. As the material layer in a screw conveyor is
heterogeneously distributed, this sampler will create a systematic error and therefore can not
be recommended.

3.3.3 Sampling from a meal transfer chute
3.4 Equipment for Quasi-Point Samples

The same comments which were made for the category of section 3.3, are also valid for this
class. The samples are only representative if the material is homogeneous which is mostly
not the case shortly after a transfer station or a change in direction. Whether equipment of
this type should be declined as a matter of principle depends mainly on the use to which it is
put. In the sampling of fine meal and small grained bulk material, in which the units to be
measured are not a function of the particle size, the resulting systematic errors can be
ignored. In each case, however, it would be more elegant to install a sampler more suited to
the material conditions. There now follows a few examples of meal samplers, which obtain
quasipoint samples.

3.5 Experimental Assessment of Samplers
This chapter does not go into the process of defining the precision of the results for samplers
or sampler systems, but limits itself to referring to the already mentioned book by Pierre Gy.
We only wish to touch briefly on two elements of precision namely representativeness and
distribution of accidental errors.
The term representativeness has already been mentioned and signifies the difference of the
true mean value µ of the material and the mean value of the single measurements. A
representative sample is only obtained if the total sample error has a random variation with
the mean value of 0 or in other words, if no systematic error appears. The distribution of the
random error describes the reliability of the sampling. It is independent of the
representativeness and can, for an example, be very small, even if and µ are not the
same. The illustration below attempts to explain these two expressions.
It is necessary to keep both terms strictly separated, and the appropriate methods must be
applied to determine them. It must be further stated that here an experimental assessment of
a sampling plant will cost a great deal of work and care, when done correctly.

HOW TO IMPLEMENT AN ADEQUATE SAMPLING SYSTEM
For the design or selection of an appropriate material sampler which satisfies the plant-
specific requirements, the following aspects must be considered:
flow sheet, layout and detailed drawings of the area in question
definition of the proposed sampling point and its position within the circuit
space available
definition of potential criteria
process-oriented criteria with regard to the potential sampling point
flow rate of material (t/h)
velocity of material (m/s)
pressure (in case of a closed system)
rate of sampling (kg/h)
frequency of sampling
physical description of material samples
type of material
screen analysis
density
moisture or water content
stickiness, dustiness, viscosity (slurry)
etc.

Chapter 11
Chemical and Physical Methods of the Laboratories

1. Chemical and Physical Methods of the Laboratories.................................................369
2. Quality Assurance.......................................................................................................387

Chemical and Physical Methods of the Laboratories.
1. SAMPLE PREPARATION..............................................................................................373
1.1 METHOD.................................................................................................................. 373
1.2 APPLICATION:......................................................................................................... 373
1.3 TIME OF PREPARATION.........................................................................................373
1.4 MAINTENANCE AND PERSONNEL........................................................................373
1.5 SUPPLIERS:............................................................................................................373
1.6 COST OF EQUIPMENT............................................................................................373
2. GRAVIMETRY................................................................................................................ 374
2.1 METHOD.................................................................................................................. 374
2.2 APPLICATION..........................................................................................................374
2.3 TIME OF ANALYSIS.................................................................................................374
2.5 EQUIPMENT:............................................................................................................374
3. TITRATION OF CARBONATES.....................................................................................374
3.1 METHOD.................................................................................................................. 374
3.2 APPLICATION..........................................................................................................375
3.3 TIME OF ANALYSIS.................................................................................................375
3.5 SUPPLIERS............................................................................................................. 375
4. COMPLEXOMETRIC TITRATION.................................................................................375
4.1 METHOD.................................................................................................................. 375
4.2 APPLICATION..........................................................................................................375
4.3 TIME OF ANALYSIS.................................................................................................375
4.5 SUPPLIERS............................................................................................................. 375

5. SPECIFIC SURFACE ACCORDING TO BLAINE..........................................................375
5.1 METHOD.................................................................................................................. 375
5.2 APPLICATION..........................................................................................................376
5.3 TIME OF ANALYSIS.................................................................................................376
5.5 SUPPLIERS............................................................................................................. 376
6. DETERMINATION OF SULPHUR AND CARBON DIOXIDE.........................................377
6.1 METHOD.................................................................................................................. 377
6.2 APPLICATION..........................................................................................................377
6.3 TIME OF ANALYSIS.................................................................................................377
6.5 SUPPLIERS............................................................................................................. 377
6.7 ATOMIC ABSORPTION SPECTROMETRY..............................................................378
6.8 METHOD.................................................................................................................. 378
6.9 APPLICATION..........................................................................................................378
6.10 TIME OF ANALYSIS.........................................................................................378
6.11 MAINTENANCE AND PERSONNEL................................................................378
6.12 SUPPLIERS......................................................................................................378
6.13 COST OF EQUIPMENT....................................................................................378
7. ATOMIC EMISSION SPECTROMETRY (BY INDUCTIVELY COUPLED PLASMA)......378
7.1 METHOD.................................................................................................................. 378
7.2 APPLICATION..........................................................................................................378
7.3 TIME OF ANALYSIS.................................................................................................378
7.5 SUPPLIERS............................................................................................................. 379
8. X-RAY FLUORESCENCE (XRF)...................................................................................379
8.1 ANALYSIS:................................................................................................................379
8.2 METHOD.................................................................................................................. 379
8.3 PRINCIPLE............................................................................................................... 379
8.4 APPLICATION..........................................................................................................379
8.5 TIME OF ANALYSIS.................................................................................................379
8.6 LABOUR................................................................................................................... 379
8.7 SUPPLIERS............................................................................................................. 380

9. X-RAY DIFFRACTION (XRD)........................................................................................380
9.1 ANALYSIS................................................................................................................. 380
9.2 METHOD.................................................................................................................. 380
9.3 APPLICATION..........................................................................................................380
9.4 TIME OF ANALYSIS.................................................................................................380
9.5 STAFF...................................................................................................................... 380
9.6 SUPPLIERS:............................................................................................................380
10. LIGHT MICROSCOPY.............................................................................................380
10.1 ANALYSIS........................................................................................................380
10.2 METHOD..........................................................................................................380
10.3 APPLICATION..................................................................................................381
10.4 TIME OF PREPARATION.................................................................................381
10.5 STAFF.............................................................................................................. 381
10.6 SUPPLIERS......................................................................................................381
11. THERMAL SHOCK TESTING OF BASIC REFRACTORIES..................................382
11.1 METHOD..........................................................................................................382
11.2 APPLICATION..................................................................................................382
11.3 TIME OF ANALYSIS.........................................................................................382
11.4 MAINTENANCE AND PERSONNEL................................................................382
12. THERMAL ANALYSIS - DIFFERENTIAL THERMAL ANALYSIS (DTA).................382
12.1 METHOD..........................................................................................................382
12.2 PRINCIPLE.......................................................................................................382
12.3 APPLICATION..................................................................................................383
12.4 TIME OF ANALYSIS.........................................................................................383
12.5 PERSONNEL....................................................................................................383
12.6 SUPPLIERS......................................................................................................383
13. THERMAL ANALYSIS - THERMO-GRAVIMETRIC ANALYSIS (TGA)...................383
13.1 METHOD..........................................................................................................383
13.2 PRINCIPLE.......................................................................................................383
13.3 APPLICATION..................................................................................................384
13.4 TIME OF ANALYSIS.........................................................................................384
13.5 PERSONNEL....................................................................................................384
13.6 SUPPLIERS......................................................................................................384

14. THERMAL ANALYSIS - DIFFERENTIAL SCANNING CALORIMETRY (DSC).......384
14.1 METHOD..........................................................................................................384
14.2 PRINCIPLE.......................................................................................................384
14.3 APPLICATION..................................................................................................384
14.4 TIME OF ANALYSIS.........................................................................................384
14.5 PERSONNEL....................................................................................................384
14.6 SUPPLIERS......................................................................................................384

SAMPLE PREPARATION
1.1 METHOD
Much work has to be done before a sample of any material is ready for analysis. The usual
sequence is as follows:
weighing and registration of the received sample
drying in a laboratory oven
splitting of the sample to obtain smaller, representative samples
coarse crushing in a laboratory jaw crusher
re-splitting into smaller portions
fine grinding with a swing grinding mill
sieving with different sets of sieves (60 um for analytical purposes)
fine re-grinding
homogenizing of the final samples with a Turbula-mixer
filling into bottles and labeling the samples
1.2 APPLICATION:
All types of material from raw material to cement and concrete
1.3 TIME OF PREPARATION
2- 3 samples per hour can be prepared to 60 un fineness depending on hardness and size
of the original sample.
1.4 MAINTENANCE AND PERSONNEL

Installation and service of mills, crushers and mixers can be done by the in-house workshop
and electricians. The work has to be done by trained personnel .
1.5 SUPPLIERS:
A. Fritsch & Co., Retsch Laborgeräte, Maschinen fabrik Herzog, Bachofen AG
Maschinenfabrik for Turbula-mixer
1.6 COST OF EQUIPMENT

Approx. SFr. 140'000 to 150'000 for complete preparation laboratory

GRAVIMETRY
2.1 METHOD
Gravimetric analysis or quantitative analysis by weight is the process of isolating and
weighing an element or a definite compound of the element in as pure a form as possible.
The element or compound is separated from a weighed portion of the substance being
examined. A large proportion of the determination in gravimetric analysis is concerned with
the transformation of the element or ion to be determined into a pure stable com pound
which can be readily converted into a form suitable for weighing. The weight of the element
or ion may be readily calculated from a knowledge of the formula of the compound and the
atomic weights of the constituent elements.
The separation of the element or of the compound containing it may be effected in a number
of ways, such as: precipitation methods, volatilisation or evolution methods, electro-
analytical methods and miscellaneous physical methods.
One of the most important methods in gravimetry is precipitation, i.e. the constituent being
determined is precipitated from solution in a form which is almost insoluble in water so that
no appreciable loss occurs when the precipitate is separated by filtration and weighed.
2.2 APPLICATION
Determination of all elements in all kinds of material. The alkalis sodium and potassium are
very difficult to determine by gravimetry.
2.3 TIME OF ANALYSIS

Very time consuming! At least five hours for the determination of the four main elements - or
4 samples per day

Minimum of maintenance required for balances and furnaces. Routine work should be
executed by skilled and qualified personnel.
2.5 EQUIPMENT:
No special instruments and apparatus are needed apart from the normal equipment of an
analytical laboratory such as glassware, furnaces, balances, etc.
TITRATION OF CARBONATES
3.1 METHOD
A sample of a raw material (e.g. limestone) or raw meal is dissolved in a known excess of
hydro chloric acid:
CaCO3 + 2HCl → Cal2 + H2O + CO2
The excess of acid is then determined by a so called back titration with sodium hydroxide
solution
HCl + NaOH → NaCl + H2O
The end point of the back titration can be found visually by means of an indicator (dye stuff)
which changes its colour or by measuring an electrical parameter, i.e. the potential
difference between two electrodes dipping into the solution. The result is calculated with a
programmable pocket or desk calculator.
3.2 APPLICATION
All types of raw material and raw meal; including all carbonates which are expressed as
CaCO3 by calculation.

10 minutes per sample

Instruments have to be installed and introduced to the operator by the supplier. Routine work
can be performed by trained personnel.
3.5 SUPPLIERS
e.g. Metrohm, Mettler, Radiometer, etc.

approx. SFr. 22'000 for semiautomatic titration apparatus.
COMPLEXOMETRIC TITRATION
4.1 METHOD
The complexometric determination of many elements is based on a titration with a solution of
EDTA (Ethylene Diamine Tetra Acetic Acid) using many special indicators. The end point of
the titra tion is found either visually or with the aid of a spectrophotometer which is more
sensitive to colour changes of many of the indicators. The results are calculated from the
consumption of the EDTA solution.
4.2 APPLICATION
Ca, Fe, Al, and Mg in cement, clinker, raw meal and raw material such as limestone, clay
and sand. Titanium and manganese may be determined also by complexometric titration
whereas silica has to be analysed by gravimetry or colorimetry.

3 - 4 hours per sample including silica, or 6 - 7 samples per day

Installation and implementation by the supplier or a chemist. Routine work requires
laboratory assistant.
4.5 SUPPLIERS
of spectrophotometer and semiautomatic titrators potentiographs: Metrohm and Mettler

ca: SFr. 20'000.
SPECIFIC SURFACE ACCORDING TO BLAINE
5.1 METHOD
The Blaine air permeability apparatus is used to draw a definite quantity of air through a
prepared bed of cement of definite porosity. Number and size of the pores in the prepared
bed (cement layer of 2.7 9 and 1.5 cm thickness) is a function of the size of the particles and
determines the rate of flow through the bed. The cement is filled into a cylindrical cell and
pressed by a plunger to form the bed. The cell is attached to the U-tube of a manometer,
one arm of which is now slowly evacuated until the liquid reaches the top mark on the glass
tube. After opening a valve, the liquid flows back again and passing two other marks, a
definite volume of air is pressed through the bed of cement.
The time which is required for the meniscus of the liquid to pass from one mark to the other
is measured. It is necessary to use a complicated formula for the calculation of the specific
surface, there fore it is advantageous to split this formula into three factors, one of them
depending only on the density of the cement, the second one depending on the temperature
and the third one being the square root of the stopped time. The first two factors may be
taken from a diagram which has to be established once for every Blaine apparatus.
5.2 APPLICATION
Cement only! Although the method has been used for the determination of the fineness of
various other materials, it should be clear that in general relative rather than absolute values
are obtained.

Double determination for one material takes approx. 20 minutes if the density of the cement
is known. Otherwise, additional 45 min. are needed for the determination of the density.

Trained personnel.
5.5 SUPPLIERS
e.g. Schiltknecht, Zürich; Tonindustrie for semi automatic instrument.

SFr. 800 to 6'000 for semiautomatic instrument

DETERMINATION OF SULPHUR AND CARBON DIOXIDE
6.1 METHOD
The solid sample is mixed with metal powder (copper, iron or tungsten) as an accelerator
and burnt in a high frequency or resistance induction furnace at cat 1500°C in a stream of
pure oxygen. The gases produced in the chemical reaction during combustion are sulphur
dioxide, carbon dioxide and water steam. With different analysers it is possible to determine
either SO2 or CO2 or even both compounds. The simplest and cheapest method uses a
titration with potassium iodate executed in an automatic burette which is connected with the
induction furnace. Another possibility is the use of an infrared analyser instead of a burette.
The IR analyser determines the sulphur dioxide by its absorption of infrared light and
displays the result as percent SO3. - Carbon dioxide can be determined with another model
which gives the result as total carbon. Instruments for the determination of both SO3 and
carbon by infrared light are also on the market.
The Leco Corporation has another type of instrument for carbon analysis which does not
apply infrared absorption but a heat conductivity cell.
6.2 APPLICATION
Total sulphur and carbon in all types of solid material, with special instruments or by means
of accessories and chemical reagents also in fuels such as coal and oil.

6 - 12 minutes per sample with sulphur titrator depending on material. 3 - 4 minutes per
sample with infrared or heat conductivity analyser

Instruments have to be installed by supplier or specialist. Routine work can be done by
trained personnel. Service by the supplier is very important for infrared and heat conductivity
analysers.
6.5 SUPPLIERS
LECO Corporation (USA) for: Titrator DB 64 and IR 232 for SO3, IR 212 and WR 12 for
carbon, CS 244 for sulphur and carbon
Leybold-Heraeus (Germany): Series 302 for sulphur and carbon

Titrator ca. SFr. 26'000
Infrared analyser ca. SFr. 50'000 to 70'000
Infrared analyser for C + S ca. SFr. 90'000

6.7 ATOMIC ABSORPTION SPECTROMETRY
6.8 METHOD
The sample has to be dissolved in acid, generally hydrochloric acid. This solution is then
sprayed into a very hot flame (ca. 2000.C, for some elements 3000.C) where the solvent
evaporates and the chemical compounds dissociate into their atomic components. Most
atoms remain in the ground state, a fact which is important for atomic absorption.
The light of a so-called hollow cathode lamp (HCL), the cathode of which consists of the
element to be determined is led through the flame with the cloud of mostly neutral
atoms. The atoms are able to absorb this HCL-light so that the intensity of the light beam
is reduced. The absorption is a measure of the number of atoms in the flame which is
proportional to the concentration of the element in the solution or at least the sample.
6.9 APPLICATION
All types of material and all elements of interest in cement chemistry except sulphur, fluorine,
chlorine and phosphorus.

approx. 4 hours for one analysis including Si, Al, Fe, Ca, Mg, K, Na or 6 analyses per day.

The method has to be introduced by a specialist or chemist. Routine work can be executed
by trained personnel.
Service by supplier should be guaranteed.
6.12 SUPPLIERS
Perkin Elmer, Beckman, Varian, Hilger Watts, Jarell Ash, Instrumentation Laboratory, Pye
Uni cam (Philips)

SFr. 50'000 to 120'000 depending on spectrometer single or double beam, etc.)
ATOMIC EMISSION SPECTROMETRY (BY INDUCTIVELY COUPLED PLASMA)
7.1 METHOD
The sample has to be dissolved in acid very often after a fusion procedure. This solution is
then sprayed into a very hot argon plasma, a physical flame of more than 6000.C where the
solvent evaporates and the chemical compounds dissociate into their atomic components.
The atoms are excited by the high temperature of the plasma and are then able to emit light
of certain wave lengths which can be measured and which is proportional to the
concentration of the elements to be determined.
7.2 APPLICATION
All types of material and all elements of interest in cement chemistry with the exception of
the halogens.

approx. 1 hour for determination of 30 to 40 elements and including preparation of the
solution.
The method has to be introduced by a specialist and chemist. Routine work can be executed
by very well trained personnel.
Very important it that good and prompt service is guaranteed by the supplier.
7.5 SUPPLIERS
ARL, Jarell-Ash, Instrumentation Laboratory, Jobin-Yvon, Kontron, Philips.

SFr. 180'000 to more than 300'000 depending on the instrument: sequential or multi-channel
spectrometer.
X-RAY FLUORESCENCE (XRF)
8.1 ANALYSIS:
Qualitative and quantitative determination of chemical elements with atomic number higher
than 9.
8.2 METHOD
Wave lengths dispersive method or non-dispersive technique (energy dispersive)
8.3 PRINCIPLE
The sample (in form of pressed powder tablet or fused bead) is irradiated by primary X-
radiation from an X- ray tube. Atoms and ions of the various elements in the sample
are thereby excited and emit their characteristic radiation.
Wave length dispersive method
The fluorescence radiation from the excited sample is dispersed on special analysing
crystals according to Bragg's law (compare with dispersion of sunlight by a prism into the
bands of colour of the visible spectrum). The resolved characteristic radiation of the
different elements is analysed individually and measured by appropriate detectors. The
intensity of the characteristic radiation is thereby a direct measure of the chemical
concentration of the corresponding element.
Energy dispersive technique
The fluorescence radiation is measured using a semi-conductor together with an
analyser which has an electrical response proportional to the quantum energy of the
radiation to be measured.
8.4 APPLICATION
Determination of the chemical composition of raw material, raw meal, clinker and cement.
Plant laboratories are mostly equipped with so -called wave length dispersive multi
-channel units and make use predominantly of pressed powder tablets; central
laboratories often prefer the installation of sequential spectrometers to increase flexibility
of analysis.

According to the application used: 3 to 20 minutes for 8 elements excluding preparation.
8.6 LABOUR
1 well trained man

8.7 SUPPLIERS
Philips, Siemens, ARL, Rigaku (Ortec, EDAX)

SFr. 250'000 - 400 '000
X-RAY DIFFRACTION (XRD)
9.1 ANALYSIS
Qualitative and quantitative determination of crystalline solids
9.2 METHOD
A monochromatic X-ray beam is directed upon a fine powdered sample. By varying the
angle between sample and incoming beam, the X-rays are reflected from the various
crystallographic planes of the minerals present. The reflected X-rays contain the following
information:
Angle: identification of mineral
Intensity: concentration of individual mineral
9.3 APPLICATION
Raw meal: content of quartz, clays, dolomite, etc.
Clinker: content of free lime, periclase, sulphates, aluminate, ferrite and modifications of
C3S, C2S, C3A, etc.

Quantitative 10 minutes (computerised)
Qualitative 1 hour
9.5 STAFF
Skilled laboratory assistant, mineralogist, chemist
9.6 SUPPLIERS:
Philips (Netherlands), Siemens (W. Germany), Seiffert and Co. (W. Germany), CGR
(France), Rigaku (USA, Japan)

SFr. 80'000 to 200'000
LIGHT MICROSCOPY
10.1 ANALYSIS
Determination of microstructure and mineralogical composition
10.2 METHOD
Samples are examined on:
polished sections using reflected light
thin sections and loose powders using transmitted light
Qualitative information:
phase association
texture
structure Quantitative
information:
phase composition
crystallite / grain / granule size
10.3 APPLICATION
Determination of actual clinker composition, e.g. percentage of elite, belite, ferrite,
aluminate, free lime, periclase and sulphates
Determination of clinker microstructure to ascertain conditions present during production.
Characterisation of raw mix components (e.g. limestone) Investigation of refractory and
concrete damages.
10.4 TIME OF PREPARATION

Preparation: approx. 3 hours:
Interpretation: 1 - 3 hours
10.5 STAFF
Skilled laboratory assistant, mineralogist, chemist
10.6 SUPPLIERS
Wild & Leitz (Switzerland), etc.

SFr. 20'000 to 60'000

THERMAL SHOCK TESTING OF BASIC REFRACTORIES
11.1 METHOD
Test prisms (35 x 35 x 200 mm) are cut from the supplied brick(s) using a diamond
impregnated saw, such as to give a full hot to cold face section. For dolomite products, the
cutting should be performed dry; with magnesite bricks wet cutting is possible but requires
immediate drying afterwards.
For shock testing, the prepared prisms are heated at 1350.C in a laboratory furnace, for a
period of up to 15 minutes, including heating up time. They are then placed vertically (i.e.
end-on) upon a water-cooled copper plate, on which they are allowed to cool for 10 minutes.
After this time, the now cooled prisms are checked for spelling and crack development. They
are then returned to the furnace for more heating up and subsequent cooling cycles until 20
such treatments have been performed, assuming of course that previous destruction has not
taken place.
11.2 APPLICATION
Testing of new refractory types and quality control on already delivered ones, in cases where
doubt exists as to the quality supplied.

For the cutting of the prisms about 1 hour is required with drying taking place overnight. The
actual thermal testing requires about 9 hours, the actual number of samples tested
depending principally on the capacity of the 1350°C furnace.

No special requirements with regard to type of equipment and its maintenance. The work
can be performed by semi-skilled personnel after a short training period.

approx. SFr 10'000 for diamond saw, high temperature furnace and cooling plate.
THERMAL ANALYSIS - DIFFERENTIAL THERMAL ANALYSIS (DTA)
12.1 METHOD
Qualitative and semi quantitative determination of caloric changes due to exo- and
endothermic reactions in materials subjected to heating; max. temperature: 1500°C.
12.2 PRINCIPLE
Approx. 100 mg of sample are weighed into a Pt -crucible and placed into the DTA furnace. A
second crucible with Al2O3 is used as reference. Both, sample and reference are heated in
the same furnace according to a selected temperature gradient. The temperature of sample
and reference is measured continuously and the difference is recorded.
To relate temperature difference with energy it is necessary to perform a calibration in the
temperature range of interest prior to the measurement.

12.3 APPLICATION
Identification of clay minerals in cement raw materials, investigations on de-carbonisation of
carbonates and formation of clinker minerals. Behaviour of circulating elements under
different atmospheric conditions.

Depending on type of application: 30 minutes to 3 hours.
12.5 PERSONNEL
Skilled laboratory assistant
12.6 SUPPLIERS
Perkin Elmer, Stanton Redcroft, Dupont, Netsch

SFr. 70'000 - 130'000
Purchase price depends on whether DTA is combined with DCS and/or TGA.
THERMAL ANALYSIS - THERMO-GRAVIMETRIC ANALYSIS (TGA)
13.1 METHOD
Quantitative determination of weight losses or weight gains in materials subjected to heating;
max. temperature 1000°C.
13.2 PRINCIPLE
Up to 1g of sample (generally approx. 50 mg) are weighed into a Pt-crucible which hangs on
filament of the TGA-balance. The sample is heated in a platinum resistance furnace at a
selected heat rate. The weight of the sample is determined and recorded continuously.
Weight changes are given in absolute values or in percent of initial weight

13.3 APPLICATION
Investigation of decomposition, dehydration and oxidation accompanied by weight changes,
e.g. determination of degree of hydration and carbonisation in hydrated clinker or cement
(open clinker storage, coating in cement mill or silos); determination of crystal water, e.g. of
gypsum, hemihydrate etc.; determination of fillers (components) in blended cements;
oxidation of sulphides and organic compounds.

Depending on type of application: 30 minutes to 3 hours
13.5 PERSONNEL
13.6 SUPPLIERS
Mettler, Perkin Elmer, Dupont, Stanton Redcroft

SFr. 60'000 - 80'000 Purchase price may depend on whether TGA is combined with DSC
and/or DTA.
THERMAL ANALYSIS - DIFFERENTIAL SCANNING CALORIMETRY (DSC)
14.1 METHOD
Qualitative and quantitative determination of caloric changes in materials subjected to iso
thermal or non-isothermal heating; max. temperature approx. 600°C.
14.2 PRINCIPLE
Approx. 10 - 100 mg of sample are weighed into a Al or Pt crucible and placed into the
sample furnace of the DSC. An empty crucible is used as reference and placed into the
reference furnace. Both, sample and reference are subjected to the same temperature
program and equal atmospheric conditions. Differences in temperature between sample and
reference - due to caloric changes - are compensated by supplying the required amount of
electrical energy. The heat of reaction is determined by integration of the energy flow. The
kinetics of the reaction may also be determined.
14.3 APPLICATION
Transformation of phases: melting, crystallisation, evaporation, volatilisation, changes in
modifications. Chemical reactions: decomposition, oxidation, dehydration, hydration, e.g.
determination and optimization of gypsum phases in cement, hydra tion of cement,
determination of sulfides in raw materials.

Depending on type of application: 30 minutes to 1 day.
14.5 PERSONNEL
14.6 SUPPLIERS
Perkin Elmer, Mettler, Dupont, Stanton, Redcroft
SFr. 60'000 - 150'000
Purchase price depends on whether DSC is combined with DTA and/or TGA
Ref. PRO MA.86/0015/2

Quality Assurance
1. SCOPE OF TESTING....................................................................................................388
2. TYPES OF XRF-SPECTROMETERS............................................................................389
3. X-RAY FLUORESCENCE SPECTROMETERS LOW COST EQUIPMENT (200 W).....389
4. ANALYTICAL TRENDS IN XRF.....................................................................................390
5. ADDITIONAL METHODS...............................................................................................390
6. ADDITIONAL ANALYTICAL EQUIPMENT....................................................................390
7. ADDITIONAL ANALYTICAL EQUIPMENT....................................................................390

SCOPE OF TESTING
♦ Production control
♦ Quality Control (due to standard specifications)
Material technological investigations
project planning
application
research & development
METHODS OF CHEMICAL CONTROL

Criteria
Method Investment Information Times of analysis Personnel
Titration Very little ∑ Of carbonates Minutes Few, trained
Gravimetry Little Main elements LS, Many hours Several,
volumetry SR, AR; BOGUE trained
Complexo ~ $10'000 CaO, Al2O3, Fe2O3, Approx. 2 - 3 Hrs. Several,
metry MgO trained
colorimetry
AAS ~ $30'000 All elements except Approx. 2 - 3 Hrs. Several +
S, Cl, F, P specialist
XRF ~$250' 000 All elements + Minutes Few +
complete specialist
automation (off or
online)

Fig. 1 Standard deviations for clinker analyses by different analytical
techniques.
TYPES OF XRF-SPECTROMETERS
Wavelength dispersive (WLD)
Sequential Mode
Simultaneous (`Mull) Channel`) mode
Energy dispersive (ED)
Low cost (200 W) WLD Systems
Bench-top XRF analysers
Continuous (`on-line`) XRF analysers
X-RAY FLUORESCENCE SPECTROMETERS LOW COST EQUIPMENT (200 W)

ADVANTAGES:
Moderate purchase costs
Lower operational cost
Simple operation and maintenance

DISADVANTAGES:
Less sensitive
Increased! measuring time
Frequent re-calibration
No trace element analyses
ANALYTICAL TRENDS IN XRF

♦ Trace element analyses
♦ Chemical analysis of liquids ♦
Analysis of carbon
♦ Sulphide/sulphate analysis
ADDITIONAL METHODS
Sulphur analyser
Carbon analyser
Coal analyser
calorimeter
proximate analysis
Free lime
Conductometric
X-ray diffraction
ADDITIONAL ANALYTICAL EQUIPMENT

♦ delivery control of waste ♦
emission control
XRF (additional channels or sequential mode) or
AAS → trace elements in liquids/solids
Ion chromatography or ion-sensitive electrodes
halogens (chlorine, fluorine, bromine)
flash point determination
viscometer
gas chromatography
organic compounds (PCB's etc.) bomb calorimeter (automatic)
ADDITIONAL ANALYTICAL EQUIPMENT

gas chromatography with mass spectrometer
• emission control on stack gas (organic compounds, volatile toxic elements etc.)

Chapter 12
Automation in Quality Control

1. INTRODUCTION............................................................................................................393
2. CENTRALISED CONTROL SYSTEMS.........................................................................394
2.1 The Automated Control System at Obourg (Lit. 1)....................................................394
2.2 The Control Concept at Gmunden (Lit. 2).................................................................400
3. DECENTRALISED CONTROL SYSTEMS....................................................................402
3.1 Raw Meal Control at Altkirch.....................................................................................402
3.2 Clinker Control.......................................................................................................... 404
3.3 Physical Cement Testing...........................................................................................408
4. BULK MATERIAL CONTROL BY NEUTRON ACTIVATION ANALYSIS....................... 408

First Chemist’s Colloquium of
North American Holcim Group Plants
January 17-19, 1990 in St. Louis, MO
by Dr. P. Bürki,
Holcim Group Support Ltd.
Technical Centre
Summary:
Automated quality control systems find increasing application in the cement industry. The
main objectives are to intensify the control in order to achieve more uniform materials
properties and to reduce risks associated with inadequate quality and to reduce shift
personnel in quality control. Automated systems are widely applied for raw mix blending and
clinker control, to a lesser extent so far on cement grinding, and little for physical cement
testing.
The paper reviews the concept and main features of such systems on the hand of examples
mainly from Holcim group plants, and briefly discusses the potential and limits of neutron
activation analysis.
Report No. MA 89/3702/E
INTRODUCTION
There is definitely a trend to increase the automation in quality control procedures in the
cement industry, whereby the degree of automation can vary from individual automated
testing equipment to complex integrated on-line systems.
The objectives of automation, and possible drawbacks are summarised in the following
figure.
Fig. 1 BENEFITS AND DRAWBACKS OF AUTOMATED CONTROL SYSTEMS
BENEFITS:
Increased product quality and reduced operation costs through improved uniformity:
high sampling frequency
more sampling points
round-the-clock control
optional on-line link to process computer
high standard of sample preparation and analysis
elimination of operator related errors
Liberation of personnel from routine work for more demanding duties.
Reduction of personnel.

DRAWBACKS:
High investment costs
Susceptible to {minor) technical faults
Requirement of back-up methods
Highly qualified personnel required for maintenance
An attempt to classify various concepts of automation is made in Figure 2.
FIG. 2 CONCEPTS OF AUTOMATED CONTROL SYSTEMS
Centralised:
sample transport to central test laboratory
Decentralised:
sample transport to testing equipment installed close to production unit
continuous analysis on full or partial, by-passed materials stream
The present paper does not aim at giving a systematic, well balanced review of current
automation systems, but to highlight some selected aspects of automation, mainly taken
from examples of such systems installed in Holcim Group plants.
CENTRALISED CONTROL SYSTEMS
2.1 The Automated Control System at Obourg (Lit. 1)
2.1.1 Objective
Obourg is one of the few remaining wet plants in western Europe. The tradition of using low-
cost ash rich coal has made it possible to produce at costs competitive to dry plants. The
lay-out is shown in Figure 3.

Fig. 3 Preparation of Clinker at Obourg.

The reasons for investing into a complex automated control system were:
High variability in chemical composition; occasional quality problems due to excessive
C3A content.
Intention to increase the use of ash rich coal, aiming at an ash content in clinker of 20%
(including non-combustible components injected into the sinter zone), which increases
the average C3A content.
Clinker targets: LS :
98 ±1
C3A : 12 ±1 %
C4AF : 11.5 ±1 %
Such an extreme clinker composition does not tolerate significant fluctuations!
Intention to change the slurry preparation from a batch to a continuous mode, to
eliminate frequent bottlenecks in slurry production.
The realisation of these objectives necessitates very close blending control for both slurry
and clinker production. After a detailed evaluation, it was decided:
to automate slurry and clinker control up to the stage of calculation of feed corrections
(excluding feeder adjustments)
to adopt a centralised concept
to realise the investment as a turnkey project.
A corresponding contract was made with Polysius to supply their POLAB-system, whereby
detailed guarantee conditions for system performance and availability were agreed upon.

2.1.2 Aspects of hardware
Fig. 4 Block Diagram of the Polab system.
Figure 4 shows the elements of the POLAB system. Features of particular interest are:
Slurry sampling (Figure 5)
Emphasis was given to obtaining a representative sample of the slurry, which still
contains an appreciable quantity of large grains; hence the complex preparation plant.
Clinker sampling
The sampling from the cooler (second screen plate) reduces the dead time of the control
loop.
Laboratory automate
Sample preparation by grinding and pelletising was selected, one reason being the
scarce experience available at that time with automated fused lead preparation.
XRF-spectrometer
The eight channel simultaneous spectrometer is exclusively used for the POLAB system.
Free lime analyser
The POLAB is complemented with a free lime analyser DIA 2000 (ethylene glycol
extraction method), which was formerly installed at the kilns.

Fig. 5 Slurry sampling and preparation plant.

2.1.3 Aspects of software
The POLAB blending control software is responsible for calculating the components feed
rate both for slurry preparation and for clinker production.
One aspect of the clinker ‘blending’ control is worth commenting on:
A number of the 4 or more components injected at the kiln hood into the sinter zone affect
both the ash and the heat input. The blending control has therefore not only to meet the
clinker composition targets, but also the target of the kiln heat requirement. The latter is
introduced and continuously adapted by the kiln operator through a terminal in the control
room.
2.1.4 Link to kiln automation
Recently, LINKMAN kiln automation systems were implemented for both kilns. The essential
tasks of the system are to determine the kiln heat requirement and to control and
automatically adjust the feed rates of all materials fed to the kiln. Because the POLAB and
the LINKMAN have obviously to fulfill complementary tasks, the two respective suppliers
were obliged to establish a link between these systems, with the aim of having an integrated
kiln automation, simultaneously fulfilling set points for kiln operation and clinker composition.
Figure 6 illustrates, in a simplified way, the concept of data exchange between the two
systems.
Fig. 6 Link of COLAB and LINKMAN
The link has been in successful operation since October 1989.

2.1.5 Experiences with the POLAB
Fig. 7 C3A: Daily Mean Sample.
The experiences can be summarised as follows:

The ash content in clinker could be raised close to 20%.
The clinker uniformity could be significantly improved; C3A can be controlled within +
1%, as compared to previous +3% (see Figure 7).
The slurry preparation and control concept are successful; production bottlenecks are
eliminated. On-line control of the components flow rate is anticipated.
The K2O-content of clinker turned out to be a rapid and significant indicator of the
burning degree.
Technical difficulties, mainly in the sector of slurry sampling and preparation - causing
delay in commissioning - could be solved in a co-operative manner.
The guaranteed systems availability of 95% could eventually be met.
The formation of personnel to operate and maintain the system became far more
demanding and time consuming than envisaged. Preventive maintenance and rapid
trouble-shooting are essential to achieve and maintain the required high availability.
2.2 The Control Concept at Gmunden (Lit. 2)

The automation system at the Hans Hatschek cement plant at Gmunden, Austria, was
supplied by Siemens (with Pfaff and Herzog as subcontractors). It will integrate the raw meal
(already realised), clinker and cement control in three stages.
The concept is shown in Figure 8 (simplified)
Fig. 8 Automatic Concept (Gmunder Zementwerke Hans Hatschek)

In comparison to the system at Obourg, the following aspects are worth commenting.
The sample fine-grinding is done at the place of sampling. Advantages: no bottlenecks in
sample preparation in the central laboratory; less risk of sample contamination. Both
untreated (for fineness determination) and fine ground samples are sent to the
laboratory.
The sequential XRF-spectrometer is also used for general analytical purposes. The
software ensures that samples arriving automatically are given priority over
manually introduced samples.
The clinker free lime will be determined by X-ray diffraction, on samples prepared in the
same way as for XRF-analysis.
The future automated cement control shall include the following control parameters:
Fineness
Presently evaluated alternatives are:
Automated Blaine-metre, AB500 by Pragoinvest, Czechoslovakia
Laser-Granulometer, either with wet-dispensing system (e.g. CILAS) or dry
dispensing system (e.g. Sympatec). The former would necessitate recuperation
of ethanol.
A major criteria will be the correlation between possible fineness parameters (Blaine,
size fraction) and standard mortar strength.
Chemical composition by XRF-analysis. In addition to SO3 determination for gypsum
dosage, the determination of additional guide-elements for fly-ash and slag proportioning
is anticipated.
DECENTRALISED CONTROL SYSTEMS
3.1 Raw Meal Control at Altkirch

Until recently, on-line raw meal control was carried out by means of an XEG (X-ray emission
gauge) system installed at the exit of the raw mill. The raw meal composition was
continuously determined by XRF-analysis directly on a by-passed stream of raw meal.
As the analytical accuracy has always been affected by matrix effects due to the fluctuating
quartz content in the raw meal, one requirement for the necessary replacement of the
system was to include conventional XRF-spectroscopy on fused beads. For reasons related
to the plant layout, the new system should again be installed close to the raw mill, accepting
the risks associated with a very limited systems supervision.

The new on-line control system was supplied as a turnkey project by Phillips France (Figure
9).
Fig. 9 Ciments D’Origny
Rep Désignation Fournisseurs

1 Vanne de dérivation KIVAIR
2 Qualdoris 30 MINEMET
3 Ecluse rotative Devance Industries
4 Vibrant AEG
5 Qualminos MINEMET
6 Ventilateur Delta neu
7 Transport pneumatique PFAFF
8 Transport pneumatique PFAFF
9 Mapp x Phillips
10 Dépoussiérage PFAFF
11 Perl x -2- Philips
12 Refroidisseur SCU 4E F&R Cooling Ltd.
13 Spectrométre Philips
14 Argon Méthane Alphagaz
15 Récupération farine PFAF

The following features deserve comment:
Application of a robot for sample manipulation (MAPP'X):
Although in this application the task of the robot is limited to control the weighing of
sample and fusing agent, the principal advantage of laboratory robots is seen in their
flexibility to be programmable for a wide range of conventional laboratory equipment,
equipment which can also be manually operated.
Application of a low-power (200 W) XRF-spectrometer PW1660.
In contrast to previous instruments, the performance of the latest generation of low-
power XRF-spectrometers (PW 1660, Oxford QX) appears to be quite satisfactory,
unless special emphasis is laid on the speed of analysis and on analysis of heavy or very
light elements.
3.2 Clinker Control
3.2.1 Conductometric free lime determination

Automatic free lime analysers, particularly designed for installation outside the laboratory,
e.g. close to the clinker cooler, are offered by two suppliers in France, namely ISYS and
ITECA. The layout is shown in Figure 10.

Fig. 10 ITECA Free Lime Analyzer.

In a usual arrangement, clinker samples are taken underneath the cooler grate. Free lime is
determined by extraction of the ground clinker with ethylene glycol and measurement of the
resulting conductivity of the solution. The results are transmitted to the central control room
or laboratory.
Such systems are, with good experience, fairly widely applied in French and Belgium plants,
including the plants of Ciments d'Origny and Ciments d'Obourg (Lit. 3).
3.2.2 Automatic litre weight determination
At the Eclepens plant of the ‘Societe de Ciments et Betons’, Switzerland, the manual litre
weight determination has successfully been replaced by an automatic on-line system,
measuring the apparent clinker density with a nuclear densitometer (Lit. 4). The system is
schematically shown in the next section -Figure 11a, along with the correlation of automatic
and manual litre-weight measurement (Fig. 11b).
3.2.3 On-line Litre-weight Determination of Eclepens
Fig. 11a Circuit diagram for continuous measurement of the apparent density of
clinker.

Kiln
clinker from grate
vibrating screen
measuring cell
overflow
gamma-ray source
variable vibrating funnel
cooling system
measuring electronics
cooling fan
process computer
analogue recorder
monitor screen
kiln control
clinker silo
clinker crusher
clinker to silo
Fig. 11b Development of manual and automatic measurements of density over a

period of 4 hours.
It is worth noting that the density measurement is used as the primary set-point parameter
for the automated kiln control system.

3.3 Physical Cement Testing
Automated, computer controlled equipment for physical cement testing has been available
on the market for several years, in particular from the German company RK TONI TECHNIK.
Such equipment comprises individual testing instruments up to integrated testing systems
for mortar preparation, VICAT-testing, specimen storage and strength determination. A
comprehensive review is given in Lit. 5.
While automated testing instruments such as Blaine-meter, laser granulometer, VICAT-
testing apparatus, penetrometer etc. are frequently used in plant laboratories, integrated
testing systems are not often applied we are presently not aware of any such installation in
Holcim Group plants. It would appear that the potential for rationalisation and cost-savings is
too small with the usual testing frequency in plant laboratories.
BULK MATERIAL CONTROL BY NEUTRON ACTIVATION ANALYSIS

The technique of prompt gamma neutron activation analysis (PGNAA) has already
existed for a few years for on-line analysis of bulk materials. ~ Commercial equipment is
presently available from Gamma Metrics, USA.
The principle of the method, and some characteristics relevant to its application are
summarised in Figure 12 to 15.
Fig. 12a How Activation Analysis Works.

Fig. 12b Principles of Operation.
Fig. 13 PROMPT GAMMA NEUTRON ACTIVATION ANALYSIS (PGNAA)
Application:
On-stream analysis of crushed raw materials or coal
Blending control upon filling of integrated pre-blending becs
Blending control on feed to raw mill
Blending control on raw meal

Advantages;
Analysis on entire material stream maximum 1000 t/h
Analysis on crushed material Ømaximum 100 mm
Continuous, real-time analyses
Multi-element analysis, incl. moisture
Permits rapid reactions in control loops
Low maintenance requirements
High availability
Disadvantages:
High investment costs, excluding installations for material transport
High service costs
Problematic material flow with moist and sticky materials
Relatively high threshold concentrations for certain elements (Na2O, K2O, MgO)
Low precision at short measuring times
Subject to regulations regarding handling of radioactive sources
Apart from the Phoenix plant in USA, the longest experience with such equipment has been
made at the cement plant of the ‘Compagnie des Ciments Belge’, in Belgium, where it is
installed ahead of the raw mill for raw mix blending control. The experience is apparently
very satisfactory, both in achieving a uniform raw meal and a very high systems availability
(Lit. 6).
Nevertheless, the still limited industrial experience does not yet permit a definite assessment
of the potential and restrictions of the PGNAA application in the cement industry. A thorough
assessment for each individual project is, therefore, still necessary.
In the context of assessing the PGNA-analyser for the pre-blending bed control at the future
Norte plant of Cementos Apasco, Mexico, HMC has carried out analyser tests with six
different plant specific raw material samples at the Gamma Metrics premises.
The tests revealed that:
a linear calibration is possible over a wide range of concentration (Figure 16a)
the standard deviation of repeated measurements is - as expected rather poor (Figure
16b)
the analytical results are not affected by the water content of the material.
As a first conclusion, the analyser is considered to be suitable for the control of the filling of
pre-blending beds. On the other hand the consequences of the high scatter of short-term
analysis for the application in raw mill control - especially with raw components of fluctuating
composition - need to be further clarified.

Fig. 16a Calibration of SiO2 with PGNAA

Fig. 16b PGNAA Reproducibility of SiO2
REFERENCES
Lit. 1 : F. Lamproye; Automatic supervision of clinker production in Obourg Cement Works,
‘Holderbank’ News 5/1989
Lit. 2 : W. Gerger; Automatic quality control of raw meal in an Austrian cement works,
Zement-Kalk-Gips No. 6/1989 (German) No. 8/1989 (English)
Lit. 3 : F. Lamproye; Automatic free lime control: a step towards efficient kiln operation,
Report MA85/3288/E (29th Technical Meeting, 1985)
Lit. 4 : A. Daina, T. Guignard; Automatic determination of the specific weight (litre weight) of
clinker, ‘Holderbank ‘ News 4/1989
Lit. 5 : C. Voos, O. Scheer; Modern ways of cement testing: presented at the General
Technical Committee Meeting of the Portland Cement Association, 1989, Lancaster,
Pennsylvania, USA (available through J. Campbell, Ciment St.Laurent)
Lit. 6 : P. Bürki; Inspection of GAMMA METRICS ‘Bulk Material Analyser’ installed at the
plant of S.A. Compagnie des Ciments Belges, Report MA 89/3582/E

Chapter 13
Statistics

Statistics
1. INTRODUCTION ............................................................................................................. 418

2. SIMPLE DATA DESCRIPTION....................................................................................... 419
2.1 Graphical Representation......................................................................................... 420
2.2 Statistical measures ................................................................................................. 426
2.2.1 The p-quantile ................................................................................................... 426
2.2.2 The Box Plot...................................................................................................... 428
2.2.3 Measures of location ......................................................................................... 429
2.2.4 Measures of Variability...................................................................................... 431
2.2.5 Statistical program packages ............................................................................ 433
2.2.6 Interpretation of the standard deviation............................................................. 433
2.2.7 Outliers.............................................................................................................. 434
3. The normal distribution (ND)........................................................................................ 436

4. Confidence limits........................................................................................................... 440
4.1 Confidence limits for the mean................................................................................. 440
4.2 Confidence limits for the median .............................................................................. 441
4.3 Confidence limits for the standard deviation ............................................................ 441
4.4 Other methods for the construction of confidence limits .......................................... 441
5. Standard Tests............................................................................................................... 442

5.1 General Test Idea..................................................................................................... 442
5.2 Test Procedures ....................................................................................................... 448
5.2.1 z-Test ................................................................................................................ 449
5.2.2 Sign Test ........................................................................................................... 451
5.2.3 Signed-rank test ................................................................................................ 452
5.2.4 Wilcoxon-Test ................................................................................................... 453
5.2.5 t-Test ................................................................................................................. 454
5.2.6 Median-Tests .................................................................................................... 455
5.2.7 X2-Test .............................................................................................................. 455
5.2.8 F-Test................................................................................................................ 455
5.3 Sample Size Determination...................................................................................... 457

6. Data Presentation and Interpretation .......................................................................... 459
6.1 Intelligible Presentation ............................................................................................ 460
6.2 Interpretation ............................................................................................................ 460
6.3 Data Interpretation related to Problems in Cement Application ............................... 461
6.4 Control Charts .......................................................................................................... 470
6.5 Comparative representation ..................................................................................... 473
7. CORRELATION AND REGRESSION............................................................................. 477

7.1 Correlation coefficient............................................................................................... 477
7.2 Linear Regression .................................................................................................... 481
7.2.1 Regression line with slope 0: ............................................................................ 483
7.2.2 Regression line with y-intercept 0 ..................................................................... 483
7.2.3 Intercept = 0. Slope = 0..................................................................................... 484
7.2.4 Comparison of models ...................................................................................... 484
7.2.5 Standard deviation of the estimates.................................................................. 485
7.2.6 Coefficient of determination .............................................................................. 486
7.2.7 Transformations before Regression Analysis ................................................... 487
7.2.8 Multiple and non-linear regression .................................................................... 488
8. STATISTICAL INVESTIGATIONS AND THEIR DESIGN............................................... 493

8.1 The Five Phases of a Statistical Investigation.......................................................... 493
8.2 Sample Surveys and Experiments ........................................................................... 494
8.3 Fundamental Principles in Statistical Investigations................................................. 494
8.3.1 Experiments ...................................................................................................... 494
8.3.2 Sample Surveys ................................................................................................ 496
9. OUTLOOK....................................................................................................................... 498
9.1 Time series and growth curves analysis .................................................................. 498
9.2 Categorical and Qualitative Data Analysis ............................................................... 498
9.3 Experimental Designs and ANOVA.......................................................................... 499
9.4 Multivariate Methods ................................................................................................ 500
9.5 Nonparametric Methods ........................................................................................... 501
9.6 Bootstrap and Jack-knife Methods ........................................................................... 501
9.7 Simulation and Monte Carlo Method ........................................................................ 501
9.8 General Literature .................................................................................................... 502
10. STATISTICAL PROGRAM PACKAGES ................................................................. 502

PREFACE
For appropriate process and quality control in the cement and concrete industry, a large
number of data are derived. Optimum benefit is, however, only achieved if these are
adequately processed and interpreted. Statistics is one of the important means to make best
use of the data be it by application of numerical methods and/or by graphical representation.
The present handbook describes the relevant basic definitions, formulae and applications of
statistical methods which are useful in the cement industry. Emphasis is put on adequate
data description and graphical representation to ensure reasonable processing and
interpretation of statistical data. Most of the described procedures are illustrated with
practical examples.
Chapters 1 and 2 are concerned with the basic ideas of statistics, the rules for the
representation of a given set of data (graphical representation, numerical measures) and the
treatment of outliers.
Chapters 3 to 8 present some statistical methods, useful for decision making,
experimentation and process control.
Of special practical significance is the Application Section (chapter 5.2), which includes a
procedure manual with a general check list and a collection of important test procedures.
Chapter 9 gives an outwork to more sophisticated statistics.
Appendix I contains a selection of practical examples to illustrate applicability and
interpretation of the demonstrated methods. A useful work sheet to construct frequency
tables and to check the data for normality is given in Appendix II. Further Appendices
contain: the required statistical tables for the determination of confidence limits and the
application of test procedures, a list of recommended literature and an index of examples
used in the text.
A subject index (English, German, French) of statistical terms is provided in Appendix VI.
The copyright for this documentation is reserved by Holcim Group Support Ltd. The right to
reproduce it entirely or in part in any form or by any means is subject to the authorisation of
Holcim Group Support Ltd.

INTRODUCTION
Statistics is concerned with methods for collecting, organising, summarising, presenting and
analysing data, obtained by measurement, counting or enumeration.
With descriptive statistics a given set of observations is summarised or presented to get a
quick survey of the corresponding phenomena.
In a more sophisticated analysis of representative samples statistical inference (inductive
statistics) allows conclusions to be drawn about the entire population.
A sample is considered to be representative only if it is drawn from the population at random.
Such a group of n observations is called a sample of size n.
Some main topics of interest in statistical analysis are:
Location of the data:
Where are the observations located on the numerical scale? This question leads to the
use of central values as the mean or the median.
Variability or dispersion:
Problems concerning the degree to which data tend to spread about an average value.
Correlation:
Degree of dependence between paired measurements, e.g.: Is there a real dependence
of mortar strength on the alkali content in clinker?
Regression:
Fitting lines and curves to express the relationship between variables in a mathematical
form especially used for prediction and calibration.
Splitting variability:
Looking for the importance of several causes to the variability of observations. Variability
arises due to different components of random errors. Special experiments must be
designed to split these components.
The use of statistical methods and the interpretation of statistical results requires a certain
comprehension of variability. If ten pieces of coal from a delivery are analysed as to their
water content, the results are not identical with the water content of the full quantity. Rather
we have ten different results in a certain range. This variability of data arises not only from
the fact that the pieces really have different water contents, but also because the results are
influenced by three different types of errors which may occur in every set of observations:
Random errors cannot be avoided. They are due to imprecise measurement, rounding of
data, environmental effects and not identically repeatable preparation. The amount of
random errors may be expressed by statistical measures of variability.
Systematic errors lead to a bias in location, but not necessarily in dispersion. A bias in
dispersion is obtained if several systematic errors are mixed.
Example: Every laboratory assistant has usually his own systematic error. This error may be
relevant or not and it may change with time. But we always expect a greater variability of the
observations if more than one person have performed the measurements. On the other
hand, we recognise in this example that a small variability does not necessarily lead to a
better average value. Environmental effects may be systematic too (e.g. air humidity).

Gross errors are "wrong" values in the set of observations. Experience shows that 5% to
10% of gross errors have to be expected in a data set. Reasons may be: wrong reading of
scales, errors in copying data, data not legible, miscalculations, gross error in measurement.
Gross errors have a considerable influence on statistical results.
Careful measurement and data handling is, therefore, important. Data have to be
inspected for outliers before a statistical analysis is performed.
Conclusion: The deviation of the single results from the true mean of the full quantity
originate from real differences between the samples on one hand, and from the occurrence
of several types of errors on the other.
It is often stated that "everything can be proved with statistics", this reasoning clearly is
wrong; false or often only misunderstood statistics originate from insufficient representations,
application of wrong procedures or assumptions, or misleading interpretation of the results.
Especially graphical representations can easily be manipulated. It is therefore essential that
the applicant of statistical techniques knows what he can and what he cannot do! An
amusing booklet dealing with such statistical "lies" is "How to Tell the Liars from the
Statisticians".
In our days of growing computer use, a lot of powerful (and sometimes less powerful or even
poor) statistical software packages are available, leading to extensive use of statistics
procedures by nonstatisticians or people not having enough statistical background. These
packages manage nearly any instruction without being able to decide whether the statistical
procedure is appropriate. This bears a great danger of use of statistics by statistical
amateurs or ignorants. It is therefore absolutely necessary that the user of statistics has a
solid statistical education.
SIMPLE DATA DESCRIPTION

In this section some descriptive methods are presented to get a quick survey of the data.
The methods are illustrated with an example of concrete strength.
Example 1
The following data represent the compressive strength of 90 cubes (20 x 20 x 20 cm) of
concrete. The data listed in chronological order of measurement is called set of
observations.
35.8 39.2 36.8 32.4 30.7 30.8 23.5 22.8 23.7 31.7 34.6 27.6 29.9 28.4 29.3
33.0 37.6 38.1 33.3 38.9 37.1 33.3 33.4 36.4 44.3 48.9 40.1 43.4 35.4 36.6
32.8 34.1 37.4 27.9 30.2 32.0 45.3 45.8 41.0 26.1 27.9 24.4 35.3 34.5 36.1
30.1 40.2 37.9 25.0 23.0 27.8 33.5 34.2 30.0 29.0 35.2 35.8 23.9 34.9 31.5
35.9 39.7 39.4 32.4 33.6 35.2 32.8 30.2 31.6 28.5 28.5 30.3 31.4 31.8 35.5
27.1 24.5 20.9 24.6 27.2 31.7 32.2 38.6 32.8 37.8 36.8 35.3 41.9 34.4 35.5
Generally a set of n observations is denoted by x1, x2, ..., xn where the index j of
xj corresponds to the number of the observation in the set.

2.1 Graphical Representation
Although the set of observations gives the complete information about the measurements, it
is little informative for the reader. A better survey is obtained by grouping the data in classes
of equal length, presented in a frequency table with tally.
Rules for classification:
The mid points should be impressive values with a few number of digits.
Choose 8 to 20 classes of equal length (approx. √n; classes)
Class boundaries should not coincide with observed values (if possible)
mid-point class tally absolute relative

frequency frequency
20.0 18.75 - 21.25 I 1 0.011
22.5 21.25 - 23.75 llll 4 0.044
25.0 23.75 - 26.25 llll l 6 0.067
27.5 26.25 - 28.75 llll llll 9 0.100
30.0 28.75 - 31.25 llll llll 10 0.111
32.5 31.25 - 33.75 llll llll llll llll 19 0.212
35.0 33.75 - 36.25 llll llll llll ll 17 0.189
37.5 36.25 - 38.75 llll llll ll 11 0.122
40.0 38.75 - 41.25 llll ll 7 0.078
42.5 41.25 - 43.75 llll 2 0.022
45.0 43.75 - 46.25 ll 3 0.033
47.5 46.25 - 48.75 0 0.000
50.0 48.75 - 51.25 l 1 0.011
Total 90 1.000
2
In this representation we recognise directly a minimum of 20.0 N/mm , a maximum of
2
50.0 N/mm , and an average value between 32.5 and 35.0. The measurements are
distributed symmetrically about the average value.
The graphical representation of the tally with rectangles is called a histogram.

Fig. 1 Histogram.
The area of the rectangles must be proportional to the tally not the height of the bar, e.g. if
two classes contain the same number of observations and the width of class 1 is twice the
one of class 2, then the height in class 1 is half the height in class 2. Same number of
observations same area of rectangle!
Two further informative graphs are the frequency curve and the cumulative frequency curve.
They are constructed as follows:
Frequency curve
plot the class frequency (absolute or relative) against the class midpoint (again take into
account the note for the histogram, same number = same area).
Fig. 2 Frequency Curve.

Cumulative frequency curve
cumulate the absolute frequencies less than the upper class boundary for every class
plot the cumulative frequency in percent against the upper class boundary.
Fig. 3 Cumulative Frequency Curve.
From this graph we can determine the portion of observations smaller than any given
strength value x.
2
Example: Percentage of the observations smaller than 30.0 N/mm : 26 %. Inverse
2
problem: Half of the measurements are smaller (resp. greater) than 33.5 N/mm .
These two frequency curves are especially suited to compare two or more different
distributions with one another.
Fig. 4 Compare Distributions.

Stem and Leaf Plot
The stem and leaf plot is very similar to the histogram but allows to reconstruct the individual
data. [The stem and leaf plot is shown in the next picture, the comment does not belong to it,
but shows how the data are reconstructed.]
STEM LEAF COMMENT

20 9 20.9
21 no observation
22 8 22.8
23 0579 23.0, 23.5, 23.7, 23.9
24 456 24.4, 24.5, 24.6
25 0
26 1 etc.
126899
455
039
0122378
456778
0244888
033456
124569
2233455889
14688
14689
169
247
1
029
42
4
3
38

Small Samples
Individual values are marked on the scale
Fig. 5 Small Sample Table.
A further attractive possibility to describe graphically the distribution of a variable, the Box
Plot, will be given in the next section. Often it is appropriate (especially in quality control) to
plot the observations in chronological order to show a possible change of the level during the
experiment.
Fig. 6 Chronological Order.
If only few data are available, the single values may be represented as points on the
measurement scale (cf. example A1, Appendix I).
The scatter diagram or scatter plot is used to represent the relationship between two
variables (paired observations concerning the same individual sample). The following
2
diagrams show the dependence of concrete strength in N/mm from cement/water ratio
after 2 days and 28 days. Different symbols can be used to discriminate groups of
observations. In the present case two groups are considered: Portland cement (PC) and
blended Portland cement (BPC).

Fig. 7 Scatter Diagram.
If we are dealing with more than 2 variables a suggested graphical representation is the
draftmansplot, which plots all pairwise scattergrams in one picture.

2.2 Statistical measures
As statistics deals with the behaviour of random variables, i.e. variables which take certain
values with certain probabilities, we are interested in describing this behaviour of the random
variable by a few characteristic measures. The first measures we will be considering are the
p quantiles (or p fractiles), which are roughly said those values for which 100.p %
( 0 <= p <= 1) of the data are smaller or equal to. The p quantiles allow a full description of
the data, similar e.g. to the cumulative frequency curve from Section 2.1. The use of the p
quantiles gives way to construct a simple and attractive graphical representation of the data,
the box plot.
2.2.1 The p-quantile
xp is called the p-quantile (or p-fractile) if
100 • p%(0 <= p <= 1)
of the values of the random variable are smaller or equal to xp.

Based on a set of observations from a random variable x, a first method to estimate
these quantiles is given by the cumulative frequency curve.
Given a sample x(i), i=1,..,n the empirical cumulative distribution function (cdf) is defined as
F(x): = (number of the x(i)'s smaller or equal to x) / n
and the following relations between the p quantile y(p) and the cdf F(x) hold
F(y(p)) = p
Y(p) = min {x(i) : F(x(i)) >= p}
The p quantile y(p) is roughly said the value for which 100*p% of the observed data are
smaller or equal to.

Estimation of P quantile v(P)
Denote by z(i), i=l, ..,n the ordered sequence of the x(i)'s, e.g.
z(1) <= z(2) <= z(3) <= .... '= z(n 1) '= z(n)
and let p(i) = (i 0.5)/n.
Then the P quantile can be computed as follows:
If P equals a p(i), then z(i) is the P quantile
Y(P) = z(i)
Otherwise compute
j(P) = nP + 0.5
and split this value j(P) into a whole number j and the remaining part B (e.g. 1.347 is split
into j=1 and B=0.347). The P quantile is then
y(P)=(1− B)z( j )+Bz( j + 1)
An example
Consider the following ordered sample
i Z(I) p(i)
1 156 0.0417
2 158 0.1250
3 159 0.2083
4 160 0.2917
5 161 0.3750
6 161 0.4583
7 163 0.5417
8 166 0.6250
9 166 0.7083
10 168 0.7917
11 172 0.8750
12 174 0.9583
12.5 % quantile (P=0.125)

As P=p(2), our 12.5 %-quantile y(.125) equals the second ordered observation, thus y(.125)
= 158
10 % quantile (P=O.1)
Compute j(P)=nP + 0.5 = 12*0.1 + 0.5 = 1.7
As this yields no whole number our 10 % quantile is (j=1 and B=0.7)
y(.1) = (1 B) z(1) + B z(2)
= 0.3*156 + 0.7*158 = 157.4

25 %-quantile (P=0.25)
Compute j(P)=nP + 0.5 = 12*0.25 + 0.5 = 3.5
y(.25) = 0.5*159 + 0.5*160 = 159.5
50 % quantile (P=0.5)
Compute j(P)=nP + 0.5 = 12*0.5 + 0.5 = 6.5
y(.5) = 0.5*161 + 0.5*163 = 162
Some of the quantiles have special names, e.g.
50 % quantile median (M)
25 % quantile lower quantile (LQ)

75 % quantile upper quantile (UP)
10 % quantile lower decile (LD)

90 % quantile upper decile (UD)
OQ UQ Inter quantile range (IQ)
2.2.2 The Box Plot

Today the box-plot is perhaps the most frequently used graphical representation for
univariate data. It offers a condensed picture of the data's distribution and shows location,
variability and extremes of the data.
In order to construct the box plot we need three quartiles:
the median
the lower quantile
the upper quantile
The construction is very simple:
where L = max (LQ 1.5 x IQ, min xi)

U = min (UQ + 1.5 x IQ, max xi)
IQ = UQ IQ
x denotes extreme values.

This construction is based on the fact that for a normal distribution approximately 1 % of the
data lie outside the interval (L,U).
An alternative method also frequently used, would be not to use L and U but instead use the
lower and upper decile. Especially for non normal distributions this representation seems
preferable.
2.2.3 Measures of location
The arithmetic mean is the mean of all observations and corresponds to the centre of gravity
in physics.
x + x + ...x 1 n
= 1 2 n = ∑ xi
nn i =1
= 35.8 + 39.2 + 36.8 + . . . + 34.4 + 35.5 = 33.199N / mm2
90
The median is the central value. Half of the observations are smaller, respectively greater
than the median.
Place the measured values in an ordered array starting with the smallest value
X (1) ≤ X(2) ≤ ... ≤ X(n)
From this array we obtain the median by
~ =x if n is odd
x (n+1)
2
~ 1
x = x +x if n is even
n n
2 (
2) (
2 +1)
For a great number of observations the ordering is very laborious. In this case the median
may be determined graphically from the cumulative frequency curve by looking for the point
on the x axis corresponding to a cumulative frequency of 50 %.
Mean or Median?
If the histogram is symmetrical, mean and median are approximately the same. In this case the
mean is a better estimate of central tendency if the distribution is not too long tailed.
If the histogram is skew, mean and median are different.
Use the mean if you are interested in the centre of gravity or the sum of all observations
Use the median if you are interested in the centre, i.e. with equal probability a future
observation will be smaller, resp. greater than the median.

Example 2
The following histogram represents the income of 479 persons
In this extremely skew distribution the arithmetic mean is quite different from the median.
Whereas the median represents a typical income, the mean can be used in a projection to
estimate the total income of the population if the sample is drawn at random.
The trimmed mean is used to estimate the mean of a symmetrical distribution if gross errors
in the data are suspected.
To calculate the α-trimmed mean the α-percent largest and smallest values are deleted. The
trimmed mean is then the arithmetic mean of the remaining observations. Notation for the
5%-trimmed mean: xtrimmed 0 05
Usual values for α: 5 % to 10 %
The weighted mean is used if certain weighting factors wi are associated with the
observation xi. Reasons may be:
- samples of unequal weights
- observation are measured with unequal precision

w x
x= ∑ i i
w
∑wi

Schematic survey on the use of measures of location:
Type of distribution
Measure to be used: arithmetic mean x
Measure to be used: trimmed mean xtrimmed
~
Measure to be used: median x
Measure to be used: median or arithmetic mean. The choice depends on the interest of the
user.
Remark: The median is equal to the 50%-trimmed mean.
2.2.4 Measures of Variability

The variance is the mean square deviation of single observations from the mean.
n
1
2
s = ∑(xi − x ) 2
n − 1i =1
For practical computations use
s 2= 1∑ 2
− 1 (∑ x ) 2
i i
n x n
−1
The standard deviation is the positive square root of the variance.

= + s2
In our example:
n
1
s= ∑(xi − x )2 = 55.596N / mm2
n − 1 i =1
The coefficient of variation is a relative measure of the variability (relative to the mean).
s
v= x
v is often used for comparing variabilities. It is especially useful if s increases proportionally

with the mean x , i.e. v = constant.
Example of compressive strength:
5.596 ==
0.169 16.9%
33.199
The range is the difference between the largest and the smallest value.
R=X -x =X -X
(n) (1) max min
Useful with small sample sizes.

Note: The range increases in general with increasing sample size n (greater probability to
get extreme values).
If the variables under consideration are assumed to be normally distributed then
the range can be used to provide an estimate of the standard deviation, according to
R
s= k where k ~= n for 3 ≤ n ≤ 10

For sample size n >10 divide the set of observations in random sub samples of size n',
calculate the arithmetic mean R of the range in these sub-samples, and estimate s
according to
R
s~ k
using a k -value corresponding to n'. . For a first rough estimation the following k values
may be used:
k~4 for n = 20
k ~ 4.5 for n = 50
k~5 for n = 100
The interquartile range is the difference between the upper and lower quartiles.
Q=x −x
.75 .25
4
If the distribution is symmetric s ≈ Q. 3
2.2.5 Statistical program packages

There are a lot of statistical programs on PCs, microcomputers and mainframes that offer
the possibility to compute these statistical measures including quartiles and even to plot the
box plots. The perhaps most prominent under many others are SAS, SYSTAT/SYGRAPH,
SPSS, BMDP, STATGRAPHICS.
2.2.6 Interpretation of the standard deviation
In the histogram of example 1 (chapter 2.1) frequency of observations decreases on both
sides of the mean. The distribution law seems to be symmetrical. In the present case we
consider the 90 measurements to be a random sample of a specific distribution model,
the normal distribution (or Gaussian distribution). This type of distribution is represented
by a symmetrical, bell shaped curve, and often observed in practical applications.
The normal distribution is determined by the mean and the standard deviation.
For this distribution about 68 % of the observation are expected in the interval mean ±1
standard deviation.
If the distribution is not normal, the standard deviation allows no direct interpretation. In this
case we can determine the interval that contains a certain portion of the observations
graphically from the cumulative frequency curve.
Another possibility to get an interpretation of the standard deviation in a skew distribution is
“normalising” of the distribution by transformations. In the case of example 2 (chapter 2.2.3),
the logarithms of incomes show approximately a normal distribution. So called variance-
stabilising transformations are especially used to fulfil normality conditions in higher
statistical analysis.
For further information consult e.g. Natrella (1963).
Further properties of the normal distribution are given in section 3.
2.2.7 Outliers
As mentioned in the introduction we have to expect about 5 % to 10 % gross errors in a set
of observations. Most of them may not be recognisable in the region of all other values.
Some may be extremely outlying values with an important influence on statistical results.
In modern statistics robust methods are studied, which are not sensitive to a certain portion
of gross errors, as for example the median or the trimmed mean. More sophisticated
robust procedures are in general rather complicated.
If classical measures as the standard deviation and the arithmetic mean are used, we have
to check the data for outliers and to eliminate them from the set of observations (cf. example
A2, Appendix I).
Note: If outliers are eliminated, they must be recorded separately in the report.
To detect outliers, check the data plot (tally, histogram or chronological order) for suspicious
values. If outliers are suspected and the data show a normal distribution, use the Dixon
criterion for rejecting observations (n<26).

The Dixon Criterion
Procedure
Choose a, the probability or risk we are willing to take of rejecting an observation that
really belongs in the group.
If
3≤n≤7 Compute r10
8 ≤ n ≤ 10 Compute r11
11 ≤ n ≤ 13 Compute r21
14 ≤ n ≤ 25 Compute r22
where rij is computed as follows
r If X(n) is suspect If X(1) is suspect
ij
r (X −X ) (X −X ) ( X − X(1) ) ( X − X )
10 (n) (n−1) (n ) (1) (2) (n) (1)
r (X −X ) (X −X ) (X −X ) (X −X )
11 (n) (n−1) (n) (2) (2) (1) (n−1) (1)
r (X −X ) (X −X ) (X −X ) (X −X )
21 (n) (n−2) (n) (2) (3) (1) (n−1) (1)
r (X −X ) (X −X ) (X −X ) (X −X )
22 (n) (n−2) (n) (3) (3) (1) (n−2) (1)
3) Look up r1−α / 2 for the rij from Step (2), in Table A 2

4) If rij > ri −α / 2 reject the suspect observation; otherwise, retain it.
In the case of a sample size n>25 use the following procedure:

1) Choose α, the probability or risk we are willing to take of rejecting an observation that
really belongs to the group
x(n) − x(1)
2) Calculate ZB =
s
3) Look for ZT ,1−α in Table A 3, Appendix III
(4) If Zb > ZT ,1−α reject the suspect observation, otherwise retain it.
Note: The presented outlier rejecting rules are only valid in normal distributions. In a skew
distribution, the elimination of outliers is very dangerous and should be avoided. In this case
the reason for the extreme observation must be known.
A check for outliers is not necessary if the trimmed mean or the median is used and if we are
only interested in a location measure.
For further tests on outliers cf. "Wissenschaftliche Tabellen Geigy, Statistik".
The standard deviation is extremely sensitive to outliers.
A check is therefore important, because it is not allowed to calculate a standard deviation
from a trimmed set of observations.

THE NORMAL DISTRIBUTION (ND)
As mentioned in section 2, the normal distribution is a theoretical model of a statistical
universe (population), defined by the mean µ and the standard deviation σ. Graphically it is
represented by a smooth, symmetric mean x and the standard deviation s of the samples
are estimated for the true, but unknown values µ and σ.
By the standardisation formula
x−µ x−x
Z= (respectively if n is large)
σ s
the ND is transformed in a normalised form with mean 0 and standard deviation 1.
Formulae for the standard normal distribution:
Density function (bell shaped curve):
1 z2
f (z) = π exp(− 2 )(−∞ < z < ∞)
2
Especially of interest is the area under the curve (distribution function), which corresponds for every given
value z to the probability of an observation to be smaller than z.
z
1
z2
φ(z) = π ∫exp(− 2 )dz,φ(−z) = 1− φ(z)
2 −∞
Numerical values for z and φ(z) are given in Table A 1, Appendix III. The probability to observe
a measured value between two given limits T1 and T2 can be calculated as follows:
transform the limits T1 and T2 in a standardised form
calculate the probability with help of Table A 1 by
F(z1,z2 ) = φ(z2 ) − φ(z1)

Interpretation of the standard deviation in a normal distribution:
We expect about 95 % of the observation between µ-2σ and µ+2σ . More general we have:
The statement that the values (of the population) lie between µ+z1-α/2.. and µ-z1-α/2 .. is
right with the probability S = 1 - α and wrong with probability α. One sided or two sided
regions may be considered. The corresponding z-values are taken from a table of the
standard normal distribution.
Usual percentiles for the statistical confidence S:
a) Two sided, S = 1− α;zα 2,z1−(α / 2)
S(%) 90 95 99 99.9
-z α/2 = z1 -- α/2 1.64 1.96 2.58 3.29
b) One sided, S = 1− α;zα ,z1−α )
S(%) 90 95 99 99.9
-z α = z1 -- α 1.28 1.64 2.33 3.09
How to check normality?

Before using the characteristics of the ND the validity of the model has to be checked.
Method: Draw the cumulative frequency curve in the normal probability paper (Appendix II).
Normality can be assumed if the resulting curve is approximately a straight line between
5% and 95%.

Example: Normal probability plot of compressive strength data
Conclusion: The observations of compressive strength follow approximately a normal

distribution.
Short cut rule for rejecting normality: If no negative values are allowed in the observations
and the coefficient of variation is greater than 30 %, the distribution is not normal.

CONFIDENCE LIMITS
4.1 Confidence limits for the mean

The arithmetic mean is an estimate for the true, unknown mean of the population of all
possible observations. We may be interested in the precision of this estimation. Therefore
we calculate a confidence interval that contains the true value with high probability
(confidence level).
a) If the distribution is normal:
Two sided confidence interval at confidence level 1-α
x±t . s
(1−α / 2;f ) n
with
f = n-1 degrees of freedom
t1-α/2;f given in Table A 4, Appendix III (Student t-Distribution)
Interpretation:
With probability 1-α the true mean µ lies between the two confidence limits.
For n>50 t1-α/2;f may be replaced by z1-α/2 given on page 15 (corresponding to normal
deviates of Table A 1, Appendix III). In our example of compressive strength we calculate
the 95% confidence interval by replacing the t- by the z- value:
x±t . s ~ s
0.975;89
n =x±z n
0.975
→33.2 ± 1.96 5.6= 33.2 ± 1.16

90
b) If the distribution is not normal:
Approximate confidence intervals can be obtained by
x±z . s
1−α / 2 n
This approximation is derived from the central limit theorem and can be used if
n > 10s2
2
x
c) Range method:
In the case of a normal distribution, confidence limits may be calculated by use of the range
instead of the standard deviation (often used in quality control for small sample sizes n).
±λ R
1−α / 2
Values for are given in Table A 6, Appendix III.

4.2 Confidence limits for the median
Confidence limits for the median can be determined directly from the ordered array of
observations.
x ≤ x ≤ .... ≤ x ≤x
(1) (1) (n−1) (n )
(1-α)-confidence interval:
x < Median < x
(k) (n-k+1)
n−1 z n − 1 rounded to the next lower integer value

with k = − 1−α / 2
2 2
z1-α/2 is given in Table A 1, Appendix III.
4.3 Confidence limits for the standard deviation

If the population follows a normal distribution, two sided confidence limits are given by
s f ≤σ≤s f
2 2
x x
1−α / 2;f α / 2;f
with f = n 1 degrees of freedom
,x2
/ 2;f
critical values of the chi-square distribution given in Table A 5, Appendix III
x2
1−α / 2;f α
Confidence intervals are reduced with increasing sample size n, i.e. the more observations
available, the better is the estimation.
The degree of improvement for the arithmetic mean can be derived from the fundamental
central limit theorem. If several samples of size n are drawn from the same population, the
arithmetic means of these samples are approximately normal with mean µ and standard
deviation s/√n, the so called standard error of the mean (SEM).
For increasing n to infinity, the standard error of the mean tends to zero, i.e. the estimation
tends to be absolutely precise if no systematic errors are present.
4.4 Other methods for the construction of confidence limits

Sometimes it happens that we have to deal with very complicated functions of random
variables, for which we can't derive or know the underlying distribution function, but we
would like to have confidence limits for the values of this complicated function. A newer
method, the Bootshap method (B. Efron, 1983), allows to obtain such results, but is rather
intensive in calculation.

STANDARD TESTS
5.1 General Test Idea

Tests are used to make decisions in the case of incomplete information. We want to know if
some observed difference is significant or only an effect of random errors.
Significant:
We decide that a difference really exists. The probability of this decision to be false is known
and can be chosen by the decision maker. A usual choice of this error probability is 5%.
Not significant:
The observed difference may be realized by chance alone. The data give no argument to
suppose the existence of a real difference. If such a difference really exists the sample size
n is too small to detect it.
Example 3:
At a cement plant A, 618 titration samples gave the mean value x1 = 82.67 and the
calculated standard deviation s1 = 4.13 .
At a later stage a further series of 525 samples were taken and after processing, yielded the
information of calculated mean value x2 = 85.58 and calculated standard deviation
s2 = 3.79
Your decision problem is:
State whether a significant change has occurred under plant conditions!
Example 4:
The production rate of a cement mill in tons/hr was measured as:
28.3 27.2 29.3 26.7 29.9 24.6 25.0 30.0 26.3 27.8
After modification, the production rate of the same mill in tons/hr was measured as:
28.0 30.0 30.5 26.0 31.0 30.3 24.6 25.4 26.7 29.3
Your decision problem is:
Has the modification made a significant improvement?
There are many problems in which we are interested in whether the mean (or another
parameter value) exceeds a given number, is less than a given number, falls into a certain
interval, etc.
Instead of estimating exactly the value of the mean (or another parameter), we thus want to
decide whether a statement concerning the mean (or other parameter value) is true or false,
i.e. we want to test a hypothesis H0 about the mean (or another value).
In example 3 the hypothesis H0 could be
“No significant change has occurred in the plant conditions"
In example 4 the hypothesis H0 could be
“No significant improvement has been made by the modification”
We will solve the two given decision problems later in section 5.2 (test procedures).
Now we will consider another specific example introducing at the same time the
important parts of all similar decision problems.

Example 5:
In the manufacture of safety razor blades the width is obviously important. Some variation in
dimension must be expected due to a large number of s mall causes affecting the production
process. But even so the average width should meet a certain specification. Suppose that
the production process for a particular brand of razor blades has been geared to produce a
mean width of 0.700 inches. Production has been underway for some time since the cutting
and honing machines were set for the last time, and the production manager wishes to know
whether the mean width is still 0.700 inches, as intended.
We call set of all blades coming from the production line in a certain time interval (t, t+h) the
oo
statistical population to be studied. For example t = January 5th, 0 and t+h = January
00
6th, 0 , if the statistical population we are interested in is the set of all blades produced on
January 5th.
If the production process was initially set up on that day to give a mean width of 0.700 inches,
we can say that the hypothesis H0 "the produced mean width of the regarded population is
0.700" should be tested. In symbols this is: µ0 = 0.700 = hypothesized mean.
Accepting the Hypothesis H0:
Suppose we draw a simple random sample of 100 blades from the production line. We
measure each of these carefully and find the mean width of the sample to be 0.7005 inches.
The standard deviation in the sample turns out to be 0.010 inches. That is,
n = 100
= 0.7005 inches s
= 0.010 inches
For the hypothesis µ0 = 0.700 to be true, the sample mean x = 0.7005 inches would have to
be drawn from the sampling distribution of all possible sample means whose overall mean is
0.700 inches.
Now the important question arises: If the true mean of the population really were 0.700
inches, how likely is it that we would draw a random sample of 100 blades and find their
mean width to be as far away as 0.7005 inches or farther? In other words, what is the
probability that a value could differ by 0.0005 inches or more from the population mean by
chance alone?
If this is a high probability, we can accept the hypothesis that true mean is 0.700 inches,
because it is very easy to get it (high probability).
If the probability is low, however, the truth of the hypothesis becomes questionable because
the sample we got is in reality very seldom.
To get at this question, compute the standard error of the mean from the sample:
s 0. 010
sx = n = 100 = 0.001inches

Since the difference between the hypothetical mean and the observed sample mean is
0.0005 inches and the standard error of the mean is 0.001 inches, the difference equals 0.5
*
standard errors. By consulting Table A-1 (zp = 0.5) we find that the area within this interval
around the mean of a normal curve is 38%, so that 100 - 38 = 62 % of the total area falls
outside this interval (cf. dashed lines below).If 0.700 inches were the true mean, therefore,
we should nevertheless expect to find that about 62% of all such possible means would, by
chance alone, fall as far away as 0.5 sx or farther.
Therefore, the probability is 62% that our particular sample mean could fall this far away.
This is a substantial reason to accept the hypothesis and attribute to mere chance the
appearance of a 0.7005 inches mean in a single random sample of 100 blades.
Naturally we reject in the same time the contrary of H0 namely that:
0 :µ ≠ 0.700
In case H0 can not be rejected, avoid saying “it is proved that H0 is correct”. or "there are
no differences in the means”, say "there is no evidence that H0 is not true" or "there is no
evidence H0 should be rejected”.
Rejecting the Hypothesis H0:
Later, after production has gone on for some time, the query again arises:
Is it reasonable to believe that the true mean width of blades produced remains 0.700
inches? Since the process was adjusted to yield that figure the hypothesis still seems
reasonable. We could then test it by taking another random sample of 100 blades.
This time the standard deviation is still 0.010 inches, so the standard error of the mean is still
0.001 inches, but the mean is now 0.703 inches:
In order to test the hypothesis that the true mean of the population is 0.700 inches, we again
go through the same line of reasoning. If the true population mean really were 0.700 inches,
how likely is it that we should draw a random sample of 100 blades and find their sample
mean to be as far away as 0.703 inches?
Since the difference between the hypothetical mean of 0.700 inches and the actual sample
mean of 0.703 inches is 0.003 inches, and the standard error of the mean is 0.001 inches,
the difference is equal to three standard errors of the mean (i.e. 0.003/0.001 = 3).
*
zp = 0.5 p = 0.69 = 69%. Therefore 31% fall outside to the right. The same portional fall outside to the left of -zp
Now if 0.700 inches really were the population mean, we know from Table A -1, Appendix III
that 99.7% of all possible sample means, for random samples of 100, would fall within three
standard errors around 0.700 inches. Hence, the probability is only 0.3% that we would get a
sample mean falling as far away as ours does.
We have two choices:
We may continue to accept the hypothesis (i.e. leave the production process alone), and
attribute the deviation of the sample mean to chance.
We may reject the hypothesis as being inconsistent with the evidence found in the
sample (hence, correct the production process).
Either of two things is true and we have to make a decision between them:
the hypothesis is correct, and an exceedingly unlikely event has occurred by chance
alone (one which would be expected to happen only 3 out of 1000 times); or
the hypothesis is wrong
Type I and Type II Errors
Understandably, the question can be raised: What critical value should we select for the
probability of getting the observed difference (x - µ 0) by chance, above which we should
accept the hypothesis H0 and below which we should reject it? This value is called the
critical probability or level of significance
The answer to this question is not simple, but to explore it will throw further light on the
nature and logic of statistical decision making. Let's study the following example:
in H0 is true H0 is false
In reality H0 is true H0 is false
We decide
Accept H0 3 right decision 2 error II
accept a true accept a
hypothesis false
hypothesis
Reject H0 1 error I 4 error II
reject a true accept a
hypothesis false
hypothesis
Example:
H0: the parliament building burns
Decision maker is the commander of the fire brigade.
Another expression for error I is: error of first kind
Another expression for error II is: error of second kind
If we ask here, what is worse
error I or error II,
then error I naturally costs a lot of money because the fire will destroy the whole parliament
building. Error II only moves the fire-brigade.
In a long run of cases which the hypothesis is in fact true (although we do not know it is true,
for otherwise there would be no need to test it), we will necessarily either be wrong as in 1 or
right as in 3.

That is to say, if we make an error it will be of Type I.
Suppose we should adopt 5% as the critical probability. accepting the hypothesis when the
probability of getting the observed difference by chance exceeds 5% and rejecting the
hypothesis when this probability proves to be less than 5%. This amounts to the decision to
accept the hypothesis when the discrepancy of the sample mean is less than 1.96 standard
deviations, and to reject the hypothesis when the discrepancy is more than 1.96 standard
deviations.
If we take 1% instead of 5%, as above, we will get as limit 2.58 standard deviations.
In fact, the percentage of cases in which we would expect to make an error of the first kind is
precisely equal to the critical probability adopted.
(The probability of error I will be abbreviated quite often by α).
Just significant probability level:
In many studies the critical probability is used to describe the statistical significance of a
sample result. For example, an economist collects some data on, say, interest rates and the
demand for money. He hypothesizes some relationship and wishes to see if the data support
his thesis. He tests the hypothesis to rule out the alternative that the observed relationship
occurred by pure chance. He reports his sample results as "significant at the 1 percent
level". Such a statement is a report to the reader that has the following meaning:
if we were to set up a statistical hypothesis
if we were to test this hypothesis using a critical probability of 1%
then we would reject the hypothesis and rule out a chance relationship
Significance levels of 10%, 5%, 1%, 0.1% are often used in reporting sample data. The
smallest of these probability values is chosen at which the hypothesis can be rejected.
So we see now what is basic for every statistical test:
we need a clear hypothesis H0
we have to know by which statistic we want to test H0 (in our example it was x )
we need a criterion C for decision making in the following form:
reject H0 if C applies
accept H0 if C does not apply
The criterion C is usually given in form of a critical limit, called significance limit, which
should not be exceeded by the calculated test statistic.
To perform such a statistical test, it is necessary to select the risk to commit a type I error,
i.e. the significance level α (see section 5.2, p. 28).
Often a relevant difference has to be detected with a certain probability, i.e. with a
predetermined type II error β. This is only possible with a certain sample size n, as is further
outlined in section 5.3.

If we reject a hypothesis when it should be accepted, we say that a Type I error has been
made. If, on the other hand, we accept a hypothesis when it should be rejected, we say that
a Type II error has been made. In either case a wrong decision or error in judgement has
occurred.
In order for any tests of hypotheses or rules of decision to be good, they must be designed
so as to minimize errors of decision. This is not a simple matter since, for a given sample
size, an attempt to decrease one type of error increases the other type. In practice one type
of error may be more serious than the other, and so a compromise should be reached in
favour of a limitation of the more serious error. The only way to reduce both types of error is
to increase the sample size, which may or may not be possible.
Note that from the philosophy of testing there is always the possibility to reject H 0 although H0
effectively is true (the probability for this is α, the type I error probability). So if you are testing
say 100 “correct” datasets (i.e. for which H0 holds) to a significance level α = 5% you will expect
about five results that reject H0, although it holds. This has to be considered if a lot of
(statistical) tests are made on the same data material (see e.g. Multiple Testing .......
.”Simultaneous Statistical Inference”, Miller, 1981).
One-sided and two-sided tests
In example 5 we are interested in values of significantly higher or smaller than µ 0. Any such test
which takes account of departures from the null hypothesis H0 in both directions is called
a two-sided test (or two-tailed test H0 :µ = µ0,A : µ ≠ µ0 )
However, other situations exist in which departures form H0 in only one direction is of
interest. In example 4 we are interested only in an improvement of production rate due to the
modification of the cement mill and so a one -sided test is appropriate. It is important that the
decision maker should decide if a one-tailed or two-tailed test is required before the
observations are taken.

5.2 Test Procedures
Performance of a statistical test requires in advance the recognition and formulation of the
present decision problem (test situation). Use the following checklist for applications of
statistical tests:
CHECKLIST
Is the decision problem concerned with the mean. median. standard deviation.
correlation coefficient or other?
Formulate a clear hypothesis H0.
Is it a one-sample or a two-sample problem?
One-sample problem:
Mean, median or standard deviation of a given sample of size n is compared with a
hypothetical value (example 5).
Two-sample problem:
The decision problem is concerned with differences between two given samples.
For the two-sample problem: Are the observations independent or paired? Observations
are paired if every value in the first sample can be attached definitely to a value in the
other sample. Paired comparisons are in general considerably more efficient than a
comparison of independent samples.
Is the sample drawn from a normal distribution?
Decide between one-sided or two-sided test.
One-sided tests are used if differences only in one direction may occur or are of interest.
Two-sided tests are used if no preliminary information is available, in what direction the
sample may differ.
Choose the test procedure in the table “TEST CONCERNED WITH” (sample sizes n<50
are considered to be small).
Choose the significance level a. Usual choices are α = 5% or α = 1% (error of type I).
Calculate the test statistic T of the chosen test (cf. the formulae of the following pages).
Look for the significance limit Tp for the test statistic in the corresponding table with p = 1-
α/2 (two-sided test) or p = 1-α (one-sided test).
Decision: If the calculated test statistic exceeds the significance limit, reject the hypothesis H 0
(the observed difference is significant at the level α). Otherwise there is no reason to reject
the hypothesis H0 and the difference is considered to be not significant.

TEST CONCERNED WITH
test situation median mean general normal mean
distribution standard
deviation
0ne-sample problem x2-test
small n (n<50) sign test signed-rank-test t-test
large n (n≥50) sign test z-test z-test x2-test
Two-sample problem
independent
small n median test Wilcoxon-test t-test F-test
large n median test z-test z-test F-test
paired
small n sign test signed-rank-test
t-test -
large n sign test z-test z-test -
2
More than two x -test Krukskal- Wallis-
Analysis of Bartlett-test
samples test variance (ANOVA)
Friedmann-test
(These tests are not treated i n this paper)
Selected test statistics
5.2.1 z-Test
Used to test means with large sample size n.
One-sample-problem:
Given is a sample of size n with arithmetic mean x and standard deviation s, we test the
hypothesis that the mean (estimated by x ) is equal to a given or target value µ0
Hypothesis H0 µ = µ0
Test statistic z = (x − µ0 ) n
σ
In general σ is not known. It can be replaced by s for large sample sizes.
Two independent samples:
Given are two samples of size n1 and n2 with arithmetic means x1,x2 and standard
deviations s1, s2.
Hypothesis H0 µ1 = µ21 −
Test statistic z = x x2
s2 s 2 1 + 2
n n 2
1
Paired comparison
Given is a sample of n paired observations xi, yj. Calculate the arithmetic mean d of all
the differences di = yi - xi and the standard deviation sd.
Hypothesis H0 µx = µy respectively µd = 0
Test statistic d n
z=
sd

Significance limits
Often used significance limits for z1-α (one-sided test) and z1-α/2 (two-sided test) are
z0 95 = 1.645 z0.975 = 1.960
z0 99 = 2.326 z0 995 = 2.576
For other significance levels α look for z1-α in Table A-1.
Decision
The difference is significant if
z ≥z two sided test
1−α / 2
z ≥z
1−α
z ≤ −z one-sided test
1−α
Example
In the example 1 we test the two-sided hypothesis H0: µ
2
= 35.0 N/mm = µ0 (one-sample problem).
= 33.2
s = 5.6
n = 90
= 0.05
z = (33.2 − 35.0) 90 = −3.05
5.6
Izl = 3.05 ' 1.96 = Z0.975
2
Decision: The sample mean differs significantly from the standard strength 35.0 N/mm .
in example 3 (chapter 5.1) we test the hypothesis whether a change in plant conditions
has occurred or not (two independent samples):
H0: µ1 = µ2
n1 = 618, n2 = 525
x1 = 82.67, x2 = 85.58
s1 = 4.13, s2 = 3.79
= 0.05 (two-sided test)
85.58-82.67
z= 85.58 − 82.67
(4.13)2 + (3.79)2
618 525
= 12.41 > 1.96 = z0.975
Decision: Highly significant difference between the two samples of titration values.

5.2.2 Sign Test
Test for the median in one-sample problems and paired comparisons.
One-sample problem: Given is a sample of size n.
We test the hypothesis that the median of the population is equal to a given value m0
H0: median = m0
Test statistic:
Count the number of observations smaller than m0 and those larger than m0. The
test statistic k is then the smaller of the two numbers.
Paired comparison: Given are n paired observations xi, yj. We
test the hypothesis that xi and yj have the same median.
Test statistic:
For each pair of observations xi, yj record the sign of the difference yi - xi . The
test statistic k is then the number of occurrences of the less frequent sign.
Significance limits
Calculate kα / 2 z n − 1 (two-sided test).
=n−1 − 1−α / 2
2 2
For the one-sided test replace α/2 by α. Zp is the significance limit of the z-test
z0.95 = 1.645 z0.975 = 1.960
Decision
If k is less than kα /2 (resp. kα) conclude that the medians are different, otherwise, there is
no reason to believe that the medians differ. 44
Example
Test the hypothesis that 50% of the population has an income of more than 3000 (example
2).
H0: m0 = 3000
k = 217
n = 479
= 0.05 two-sided test
k = 478− 1.96 478 = 217.6
0.025
2 2
The test is just significant at the 5% level. Because the sample median is 2700, we conclude
that less than 50% of the population has an income of 3000.

5.2.3 Signed-rank test
Test for the mean in symmetrical distributions with small sample sizes n (one-sample or
paired comparison).
One-sample problem: Given is a sample of size n from a symmetrical distribution.
Hypothesis H0 : µ = µ0
Disregarding signs, rank the di according to their numerical value, i.e., assign rank 1 to the
smallest observation, assign the rank of 2 to the di which is next smallest, etc. In case of ties,
assign the average of the ranks which would have been assigned if the di's had differed only
slightly. (If more than 20% of the observations are involved in ties, this procedure should not
be used).
To the assigned ranks 1, 2, 3, etc., prefix a + or a - sign, according to whether the
corresponding di is positive or negative. The test statistic T is then the sum of these signed
ranks.
Paired comparisons: Given are n paired observations xi, yi. The
hypothesis is tested that both have the same mean.
Test statistic: Compute di = yi - xi for each pair of observation and continue as in a).
Significance limits
Look for T0.95 or T0.975 in Table A-7. If the number m of differences di exceeds 20 perform
a z-test with
zT = T
m(m + 1)(2m + 1)
6
Decision
Conclude that the means differ if
T≥T (or zT ≥ z ) two-sided test
1−α / 2 1−α / 2
T ≥ T1−α (or zT ≥ z1−α ) one-sided test

T < −T1−α (or zT < −z1−α ) one-sided test

5.2.4 Wilcoxon-Test
Test for a comparison of means in two independent samples:
Given are two independent samples of size n1 and n2.
We test the hypothesis that both samples have the same mean
H0:µ1 = µ2
Test statistic:
Combine the observations from the two samples, and rank them in order of increasing size
from smallest to largest. Assign the rank of 1 to the lowest, a rank of 2 to the next lowest,
etc. (Use algebraic size, i.e., the lowest rank is assigned to the largest negative number, if
there are negative numbers). In case of ties, assign to each the average of the ranks which
would have been assigned if the tied observations had differed only slightly. (If more than
20% of the observations are involved in ties, this procedure should not be used).
Let n1 = smaller sample
n2 = larger sample
n = n1 + n2
Compute R, the sum of the ranks for the smaller sample. (If the two samples are equal in
size, use the sum of the ranks for either sample).
Compute W = 2R - n1(n + 1)
Significance limits
Look for W0.95 or W0.975 in Table A-8. For sample sizes not mentioned in the table perform
a z-test with
ZW = W
n1n2 (n + 1)
3
Decision
Conclude that the means differ if
W two-sided
≥W (or z ≥ z )
1−α / 2 w 1−α / 2
W≥W (or zw ≥ z1−α ) one-sided

1−α
W < −W1−α (or zw < −z1−α ) one-sided

Example
In example 4 (chapter 5.1) we are interested in an improvement of production rate after
a modification of the cement mill.
= 0.05 (one-sided test)
Combined sample (values in italics correspond to the sample after modification):
tons/hr 24.6 24.6 25.0 25.4 26.0 26.3 26.7 26.7 27.2 27.8
rank 1.5 1.5 3 4 5 6 7.5 7.5 9 10
tons/hr 28.0 28.3 29.3 29.3 29.9 30.0 30.0 30.3 30.5 31.0
rank 11 12 13 14 15 16.5 16.5 18 19 20

The sum R of ranks in the second sample is
R = 1.5 + 4 + 5 + 7.5 + 11 + 13 + 16.5 + 18 + 19 + 20 = 115.5
W = 2 . 115.5 - 10 . 21 = 21
W = 21 < 46 = W0.95
Decision
The improvement is not significant. We have no reason to assume a real improvement due
to the modification.
5.2.5 t-Test
This test is used for comparisons of means in normal distribution with small sample sizes n.
If the distribution is not known to be normal, prefer rank tests (Wilcoxon, signed ranks).
The test situation is the same as in 5.2.1 (z-test) refer for remarks on hypothesis and
decision. Instead of z1-α look for significance limits t1-α in Table A-4, with n-1 degrees of
freedom (df) in the one-sample case or for paired comparison, and with n1 + n2 - 2 degrees
of freedom (df) for the two-sample case respectively.
One-sample problem: Given is a sample of size n drawn from a normal distribution.
t = (x − µ0 n
s
Two independent samples: Given are two independent samples of size n1 and n2 from two
normal distributions with means µ1, µ2 and equal standard deviation σ.
2 −
1
nn
t= x x 1 2
(n − 1)s2 + (n
1 1
2 2
− 1)s n1 + n2
2
(n1 + n2 − 2)
Paired comparison: Given are n pairs of observations xi, yi, when x and y follow a
normal distribution.
d
t= n
sd

5.2.6 Median-Tests
Median tests for two independent samples are not given here in an explicit form. For small
2
sample sizes n Fisher's test for 2 x 2 contingency tables and for large n x -test for
contingency tables may be used (Natrella 1963, Noether 1971).
In a comparison of two independent samples we are often not interested in differences only
of means or medians, but generally in location differences of the two samples.
In this case use the tests for means, which are more efficient than median tests (Wilcoxon-
test or z-test).
5.2.7 X2-Test
2
Several X -tests are available for different test situations. The test we explain here is used to
compare the standard deviation s of a sample with hypothetical value σ0.
2
X -tests are also used for
Goodness of fit test
Independence test
Loglinear models
(see e.g. Haberman, 1978)
Given is a sample of size n drawn from a normal distribution with mean µ and standard
deviation σ. We test the hypothesis that the standard deviation σ, estimated by s from the
sample is equal to a hypothetical standard deviation σ0.
(n − 1)s2
H0: σ = σ0 Test statistic: x2 =
σ 02
Significance limits
2
Look for significance limit X p,m, in Table A-6 with m = n-1 degrees of freedom (df).
Decision
If X2 ≥ X2 1-α/2;m 2
or X ≤ X
2
α/2;m conclude: σ = σ0 (two-sided)
If X2 ≥ X2 1-α;m conclude: σ > σ0 (one-sided)
2 2 (one-sided)
X ≤X α;m conclude: σ < σ0
Otherwise we have no reason to believe that σ differs from σ0.
5.2.8 F-Test
Comparison of the standard deviation in two independent samples.
Given are two independent samples of size n1 and n2 with standard deviations s 1 and
s2. Both samples are drawn from a normal distribution. We test the hypothesis that both
populations have the same standard deviation.
H0: σ1 = σ2
Test statistic:
Let be s1 > s2, then compute
s2
1
F=
s22

Significance limits
Look for F1-α;m1,m2 (one-sided) or F1-α/2;m1,m2 (two-sided) in Table A-9 with m1 = n1-1 and
m2 = n2-1 degrees of freedom.
Decision
Conclude σ1 ≠ σ2 if F ≥ F1-α/2;m1,m2 (two-sided)
or σ1 > σ2 if F ≥ F1-α;m1,m2 (one-sided)

5.3 Sample Size Determination
As stated at the end of chapter 5.1, the probability β of making a type II error in a given
test with significance level α (type I error) depends upon the sample size n.
How can we determine the sample size that is necessary to hold the type II error within
certain boundaries (probability β)? Or in other words: What sample size n is necessary
to detect a relevant difference with great probability (1-β)?
Let us first examine a one-sided one-sample-test for testing a mean:
Suppose we are interested in the mean µ of a random variable X. We want to test the
hypothesis H0: µ = µ0 against the alternative hypothesis A: µ > µ0 with a z-test. The
error-type-one shall be α.
The probability to detect a deviation of ∆ from µ0 shall be at least (1 - β), if this deviation
exceeds a preselected relevant difference δ (β is now an upper bound for the type II error).
The following figure shows the distribution of the sample mean x under H0 and under the
(special) alternative hypothesis A: µ = µ0 + δ:
Interpretation of the graph:

CHo is the criteria (significance limit) of the test:
The hypothesis H0 is rejected with probability α even if it is true.
On the other hand, the probability to accept H0: µ = µ0, even if the true mean is µ = µ0 + δ
(type II error), is greater than β.
The relevant difference δ may only be detected with probability (1-β), if CHo ≤ CA; in this
case the risk to commit a type II error is ≤ β.
The variance σ 2
n of the two distributions becomes smaller with increasing n and hence CA
moves to the right and CHo to the left. Therefore, we can find an n such that CHo ≤ CA:
C
H0 A
≤ C
z ⋅σ + µ≤ σ + µ + δ =− z 6+ µ + δ (1)
z
1−α 0 0 1−β 0
n n
βn
σ
δ (z1−α + z1−β ) ≤ n
or n ≥ σ 2 (z +z )2 (2)
1−α 1−β
δ2
n corresponds with the required sample size to detect a desired difference δ with probability
(1-β) by a z-test with significance level α.
For the two-sided one-sample-test we only have to replace α by α/2, and we get:
n ≥ σ 2 (z +z )2 (3)
1−α / 2 1−β
δ2
Note: β is not replaced by β/2!
In a similar (but somewhat more complicated) way we find lower bounds for the sample
sizes n1 and n2 satisfying our conditions in the one-sided two-sample-test:
2 2
(4)
n≥ σ σ1 + σ 2 (z 2
2 + z1−α )
δ2 1−β
n ≥ σ1 n
1 2
σ2
For equal standard-deviations (σ1 = σ2 = σ) we obtain from (4):
n = n 2 ≥ 2σ 2 (z +z (5)
) 2
1 1−β 1−α
δ2
In the two-sided two-sample-test we simply have to replace α by α/2 in (4) or (5).
The following table shows short rules for determining sample sizes when α = β = 5% and σ1.
σ2 = σ in the two-sample-case. The numbers in parentheses refer to the formulae from
which the rules are derived.
one sample two samples
one-sided 2 n=n 2
n ≥ 11σ ≥ 11σ
2 1 2
δ δ2
(2) (5; for σ1 ≠ σ2 see (4))
two-sided 2 n=n 2
n ≥ 13σ ≥ 26σ
1 2
δ2 δ2
(3) (5)
with δ = relevant difference to be detected with probability (1-β)
= β = 5%
Note:
2 2
The variance σ is usually not known, but often some knowledge about σ is available
2
from former experiments (standard deviation). Otherwise σ may be estimated in a
pilot study.
In the case of small sample sizes (t-test, Wilcoxon-test), add 5% to the calculated n.

DATA PRESENTATION AND INTERPRETATION
(Chambers, Cleveland, Kleiner and Mikey, 1983).
The problem of data presentation and interpretation is common to cement manufacturers
and users. In every stage of cement, aggregate, concrete production and application, data
concerning materials, equipment, energy consumption, market situation, costs, etc. are
generated. All this information must increase the knowledge about what happens in the
process and market, thus providing the basis of decision. Moreover, communication between
supplier and consumer has to rely on this information. We consider it, therefore, vital not only
for optimum manufacture of cement and cement based products, but also with respect to the
mutual relation between manufacturer and user that appropriate attention is given to data
handling, i.e. presentation and evaluation.
We may differentiate between three levels of information:
data needed for direct process and quality control (routine decision, off- or on-line)
data required for day-to-day management decisions based on quality reports
additional data allowing long-term improvements and developments.
With respect to presentation and interpretation of test results, the main problems for plant
management are:
how to organize the required information flow so that decision can be made at all levels of
competence
to reproduce and evaluate the corresponding data in a specific situation.
The relevant data of al1 domains cannot be made available without a functional information
system. But availability alone does not necessarily result in a rational decision based on the
data. To achieve this, two further aspects must be considered: first, the data should be
presented in an intelligible form, i.e. a high transparency of the results enables the
management to recognize certain relationships or critical results in time. Secondly, the
decision maker must be aware of accuracy, significance and relevance of the considered
data in order to provide a realistic interpretation.
Availability of data
A quick availability of data does not necessarily require a fully integrated data based system
with electronic data processing, but is rather a matter of organization. Important data (e.g.
for process and quality control) should circulate with little loss of time. Consequently, an
appropriate reporting system (organization) must be established. Additional data should not
disappear in some drawer where a later retrieval is impossible or at least very inconvenient.
To avoid such a disorder, all data are recorded in a similar way, including a note on the
circumstances of measurement and provenance of samples and test results. Remarks about
circumstances and provenance are used to judge the comparability of different sets of
observations. A standardised recording procedure makes data surveying easy and simplifies
a later change to electronic data processing.

6.1 Intelligible Presentation
The main aim of data presentation is transparency rather than secrecy, i.e. graphs should be
employed instead of tables. A graphical presentation gives a quick survey on relevant
information such as changes in time, extreme values, relationship between variables. A short
survey on frequently used graphs can be found in section 6.4. It is recommended to produce
the graphs directly at the source of the data in the form of tallies and/or control charts. A
control chart may already be included in the laboratory journal and reports, directly behind
the columns for sample identification and test results. The possibility of obtaining a quick
survey by consulting a well presented graph will not only help the management in its
decision making, but transparency will also improve interest and motivation of the personnel
at any level of competence.
6.2 Interpretation
First of all. we must know how much we can rely on the data.
Test results are never "true" values, but are rather subject to errors of three types (see also
section 1):
random errors caused by sampling, imprecise measurement, environmental effects, etc.
systematic errors caused by bias in sampling or process measurement, by the use of
inadequate experimental design
gross errors caused by recording wrong or not comparable values.
Usually, we expect 5 to 10% gross errors in a set of observations. Random errors may be
denoted as reproducibility and expressed as the corresponding standard deviation. In
general, these errors are underestimated. Reproducibility should be known for every
important analyzing method. Systematic and gross errors are difficult to characterize. They
should be minimized by careful experimentation and data handling. A periodic check for
systematic errors should be done (calibration, comparison with a standard, inter-laboratory
test).
How to compare several data sets or groups of data?
Before any comparison is made, we have to seriously check the comparability of data. Often,
data sets differ in provenance of samples or circumstance of measurement, so that a
comparison ma y be impossible. Even the fact that sample 1 was measured by laboratory
assistant Miller and sample 2 by Brown will lead to a biased comparison if there is any
relevant systematic error between the two persons. If the data sets are comparable and we
observe a certain difference, we have to ask the following questions before taking any
action:
Is the difference significant?
Is the difference relevant?
The problem of significance is answered by a statistical test. If it is not significant, we have
no reason to take any action because the observed difference may occur by chance alone. If
the test indicates a significant difference, it is not necessarily relevant for the problem we are
concerned with. Of course, the decision whether or not it is relevant is not a statistical
problem. Perhaps a decision maker is alarmed when an observed difference, considered to
be relevant in the present problem, does not lead to a significant test result. In this case, the
sample size used may be too small or the testing procedures may not be sufficiently
sensitive to solve the given problem.

Consequently, experience in product manufacture and application is necessary to assess
the relevance of difference between target quality and experimental values. However, to
judge whether it is significant requires a training in statistical technique, particularly test
procedures. Decisions based simply on either practical experience of many years
(relevance) or on statistical technique and procedures (significance) will lead to too frequent
and unnecessary actions in both the manufacturing process and product application.
6.3 Data Interpretation related to Problems in Cement Application

If confronted with the problem of finding the reasons for poor quality of cement related
products (concrete, asbestos cement, etc.), the cement is often suspected to be the cause.
This can be explained by the fact that cement is the binding agent, and in the majority of
cases constitutes the most expensive component, although it is known that the effect of
other components, proportioning and curing conditions etc. are of great importance. A
further reason to inspect the cement first is the difficulty to specify the other effects while
cement is well defined by its chemical and mineralogical composition and its fineness.
To find out the causes for inferiority or changes in quality, a collaboration of consumer,
producer and statistician is indispensable: the statistician may be omitted in uncomplicated
problems if delegates of consumer or producer are well trained in statistics. In a first
retrospective analysis of available data, parallel changes of parameters and quality are
studied. This is done by drawing scatter diagrams and performing a regression analysis. The
disadvantage of retrospective studies is the difficulty to find out causal relationships. The
effects of several variables are mixed and cannot be separated due to their causal origin. On
the other hand, an observed correlation indicates a possible causal effect. The decision
about what cause is really responsible must be made by an experienced specialist and not
by a statistical test. Often a decision is not possible because dependencies are too complex.
In this case, a special experiment has to be planned and performed with a controlled
variation of suspected variables and careful elimination of interfering effects (prospective
analysis ). In contrast to the retrospective analysis, an adequately designed experiment
renders it possible to evaluate and judge causal effects with statistical methods. A short
survey on the use and interpretation in regression/correlation analysis and designed
experiments is given in chapter 7.

Survey on Graphical Data Representation
Representation of one sample - documentation
Frequency table / Tally
Given is a sample of large size n. Observations are grouped into classes of equal length and
marked in the tally.
Example: 90 values of concrete strength

Histogram
Graphical representation of the tally. Visualization of minimum, maximum, center and shape
of the distribution.
Area of rectangles corresponds to absolute or relative frequencies in the classes.

Frequency curve
Relative (or absolute) frequencies plotted against class mid point. Every point on the curve
corresponds to the relative (or absolute) frequency of observations falling in this class.
Cumulative frequency curve

Cumulated relative frequencies plotted against upper class boundaries. Every point on the
curve corresponds to the portion of values which are smaller than any given strength x.
2
Example: Half of the measurements are smaller (respectively greater) than 33.5 N/mm

Stem-and-leaf Diagram
Similar to the histogram, but allows to see the individual data. Each observation is a leaf of a
stem.
Representation of small samples

For small sample sizes n the individual values are plotted directly on the measurement
scale. Suitable to detect outliers and skewed distribution.
Example: K2O-content of 10 cements

Time-plot
Often it is appropriate to draw the observations in chronological order to show a possible
change of the level in time.
Useful in annual reports.
Cumulative time-plot
Cumulated values are plotted against time. Used to show deviations from a cumulative
target.
Example: Actual clinker production is cumulated every month and compared with a target.

Box-Plot (Box-and Whisker-Plot)
Special and frequently used plot to numerize data. Shows median, lower and upper quartile
(forming the box) and two lines (the whiskers) extending to the extremes of the data. If
unusual extreme values occur, the whiskers extend only to those points that are within 1.5
times the inter-quartile range.
Scatter-plot or x-y plot

Plots two variables against each other.

Draftmensplot
Plots two or more variables against each other!

Flury-Riedevyl faces and Chemoff faces
Faces are a method to show multivariate data graphically. Each variable is mapped to one or
more of the face parameters. Other methods to represent multivariate data are the star
symbol plot, the sun ray plot, castles and trees.

6.4 Control Charts
Control charts are special time-plots to show a possible change of the characteristics in a
production process. The essence of a control chart is its clearness and intelligibility. In a
quick survey it is possible to recognize natural subgroups, within which variation is likely to
be random but among which assignable causes of one sort or another may cause non-
random variation.
Summarized statistical errors (such as means or standard deviations) and confidence limits
determined from overall sample data can be very misleading, if the data are not free from the
effects of assignable causes. In cement industry annual means and standard deviations are
mostly misleading due to a nonstationarity of the process, i.e. the occurrence of systematic
changes at the production level. (Example: mean and standard deviation of the lime
saturation of raw meal and/or clinker in annual reports. Changes in the target value required
by the process or the product lead to a high non-interpretable standard deviation.)
If possible, the data should originally be collected with this subgrouping in mind; in such
cases, the analyses can be simplified by arranging for an equal number of observations in
each subgroup. If the data must be analyzed as they come, subgrouping may still be
possible with knowledge of the data's source, i.e. obvious changes in the process conditions
must be registered together with the data.
Control limits
To judge a significant change at the production level, the charts are completed with control
limits. As a general rule it can be assumed that such a change is present, if a sample point
falls outside of 3 σ-limits, where σ is the standard deviation of a homogeneous production
phase.
This engineering rule has been found to work well. No exact probability is given for a chance
variation beyond 3 σ-limits, but in general it is very s mall, in the order of perhaps 0.3%.
It may be of advantage to use two limits, such as a 2 σ-warning-limit and a 3 σ-action-
limit. In this case, an action should already take place when two subsequent values lie
between the two limits.

The use of 3 σ-limits bases on pure statistical considerations. It may occur that these limits
are in conflict with the required tolerances of an external standard (e.g. process, standard or
market requirements). If the required tolerances are smaller than the statistical ones, a
modification of the process is necessary to improve its precision. In the opposite case, the
statistical limits may be used in order to detect changes as early as possible.
Control Charts
Several types of charts may be used to detect different types of changes in the process:
Control chart
Cause of Change Mean Range Standard Cumulative
X R deviations
Gross error (blunder) 1 2 - 3
Shift in average 2- - 1
Shift in variability - 1 2 -
Slow fluctuation (trend) 2 - - 1
Rapid fluctuation (cycle) -1 2 -
= not appropriate / 3 = least useful
2 = useful / 1 = most useful
x − chart
Samples of fixed size n are taken from time to time. The arithmetic mean of the samples are
plotted versus time.
Control limits (for warning and/or action) can be computed with the help of special
procedures.

R-chart of s-chart
Given are small samples of fixed size n. The range or the standard deviation of the sample is
plotted.
Values exceeding a certain control limit indicate an increase of variability.
Chart of individual values

Instead of means, individual values ma y be plotted in the order of measurement.
CUSUM-chart
Cumulative -sum-chart. The deviations of individual values from a target value
are cumulated.
This chart is especially sensitive to slow fluctuations in the process.

6.5 Comparative representation
Frequency polygon
Comparison of the distribution of two samples. Relative frequencies corresponding to the
histogram are plotted on a line graph against the mid-points of the classes.
Example: Mortar strength in the first and the second half-year.

Cumulative Frequency plot
Comparison of the distribution of two samples. Relative cumulative frequencies are plotted
against upper class limit.

Grouped Box-Plots
Shows two or more box- plots in the same graph. May be used to show the distribution of
different variables or the distribution of the same variable in different groups (samples). The
notched box plot may be used for the latter comparison.

Time-plot
Comparison of simultaneous changes of several variables in an observed period.
Example: Silica- and alumina ratio of a cement type in a specific year.
Scatter-plot
Illustration of the relationship between two variables.
Example: Relation between strength and density of concrete. The relationship is not linear.
Don't use the correlation coefficient in this case, because it is only a measure for linear
dependence.
Comparison of small samples

Individual values are marked on the scale separately for each group of observations.
Example: Inter-laboratory-test concerned with elite-content.

Asymmetrical Flury-Riedwyl Faces
This is a special version of the faces. E.g. the variables before and after treatment,
respectively, are mapped to the same face parameter on the left and right face side,
respectively. 0r the left side shows the specification values of the variables and the right side
the actual measurements.
CORRELATION AND REGRESSION
7.1 Correlation coefficient

The degree of linear dependence between two random variables X and Y can be expressed
by the correlation coefficient rxy.
sxy
rxy =
sxsy
n
1
with sxy = ∑(xi − x )(yi − y ) the covariance of X and Y
n − 1 i =1
and sx, sy standard deviations of X and Y.
For practical computations use
s = 1 x y 1 ( x )( y )
xy n − 1∑ i i n ∑ i ∑ i

The following figure shows the scatter diagrams for several degrees of dependence.
Properties of rxy:
rxy = ranges from -1 to +1
rxy = +1 : All measured values lie on an increasing line -1 :
All measured values lie on a decreasing line
0 : No linear relationship between the measured values
rxy is a measure of linear dependence. If the scatter diagram indicates a non-linear
relationship, then the correlation coefficient will be misleading and should not be
calculated.
rxy is very sensitive (not robust) against outliers.
The necessity of drawing a scatter diagram is illustrated in the following figures. Completely
different graphs may result with equal correlation coefficients (r = 0.82).
rxy can be close to zero even though the variables are clearly non-linear dependent and rxy
is not defined if sx or sy is zero.
Interpretation of rxy
A high correlation coefficient between two variables does not necessarily indicate a causal
dependence. There may be a third variable not under control which is causing the
simultaneous change in the first two variables, and which produces a spuriously high
correlation coefficient. In order to establish a causal relationship it is necessary to run a
carefully controlled experiment (see chapter 8). Unfortunately it is often impossible to control
all the variables which could possibly be relevant to a particular experiment, so that the
experimenter should always be on the lookout for spurious correlation.
The following is an example for confusion of correlation with causation:
The following figure shows the population of Oldenburg at the end of each of 7 years plotted
against the number of storks observed in the corresponding year. Although in this example
few would be led to hypothesize that the increased number of storks caused the observed
increase of population, investigators are sometimes guilty of this kind of mistake in other
contexts. Correlation between two variables Y and X often occurs because they are both
associated with a third factor W. In the stork example, since the human population Y and the
number of storks X both increased with time W over this 7-year period, it is readily
understandable that a correlation appears when they are plotted together as Y versus X.
Fig.: A plot of the population of Oldenburg at the end of each year against the
number of storks observed in that year.

By using sound principles of experimental design and, in particular, randomization, data can
be generated that provide a more sound basis for deducing causality.

7.2 Linear Regression
If the scatter diagram indicates that the two variables are linearly related, then we may want
to predict the value of one of the variables from a given value of the other variable. For this
purpose a regression line is fitted to the data (method of least squares).
Model assumptions
yj =α+βj+Ej (1)
where the residuals Ej's are independent normal random variables with mean 0 and constant
variance σ2.
= y -intercept β
= slope
x = independent variable - Y dependent variable
This assumption is essential if parameter reductions (see later), i.e. reduced models are
tested and the assumptions have to be checked when a model is fitted.
The estimates of the parameters α and β in model (1) are obtained by the method of Least
Squares (MLS), i.e. the estimates are obtained by minimizing the sum of squared residuals
n 2 n
ˆ2
=
∑(y j − α − βx j ) ∑E j
j =1 j =1
with respect to α and β !

This method is very wide spread in statistics and the estimates have good statistical
properties under the above model assumptions! Today there are also procedures which do
not minimize the sum of squared residuals but use other criteria (Robust Methods!).
In the following let
1
sxx = ∑(x j − x )2 = ∑ x j 2 − n (∑ x j )2 = (n − 1)sx 2
1
syy = ∑(y j − y )2 = ∑y j 2 − n (∑y j )2 = (n − 1)sy 2
1
sxy = ∑(x j − x )(y j − y ) = ∑ x j y j − n ∑xjyj = (n − 1)sxy
The resulting estimated regression line is

ˆ ˆ s
xy
yˆ = αˆ + β , where β =s
xx
ˆ
αˆ = y + βx
with a resulting minimal sum of squares (MSSQ) smin = syy − s2 xy with n-2 degrees of
s
xx
freedom (2 parameters α and β are estimated). yˆ is the predicted value for the dependent
variable y.

gives us the possibility to validate the model assumptions (Independence,
Note:
In the mentioned problem there are two regression lines, one to predict y from x and one to
predict x from y. The two lines are not identical and therefore it is not allowed to invert
the regression equation.
The prediction equation is valid only in the observed range of observations. Extrapolations
may give misleading results. We have no information that linearity holds outside of the
present observations.
For each pair (xj, yj) we can therefore compute the predicted value yˆ of y by
ˆ
means of the regression function yˆ = αˆ + β x i.e.
ˆ
yˆ j = αˆ + β x j
and the estimated residual
eˆj = y − yˆ j
αˆ 2 = S min /(n − 2) is an estimate for the variance of the residuals. The analysis of the residuals eˆj
Normality and constant Variability). The first check is done by plotting the residuals against the
predicted yj. The residuals eˆj should be randomly scattered around the line e = 0 and
show no pattern.
For example the following structures of the residuals would indicate, that the model is not
correctly specified (e.g. that the variables x and y should be transformed before calculation
of the regression line).
The second check would be a normal probability plot of the residuals.

If the residuals suggest that the assumptions do not hold, then further investigations are
necessary (suitable transformations, other influencing variables, time-dependence => growth
curves, time-series-analysis).
After confirmation of the assumption and calculation of the regression line (sometimes also
before calculation) it is possible that we want to investigate whether perhaps the y-intercept
equals zero, or the slope equals zero or whether there is no relation between y and x (α = β
ˆ
0) or whether the α or β equal some specified values α0 or β0 (e.g. if a value of β = 1.1 may
be replaced by β0 = 1.0 ). This means that we want to test if a reduced (simplified)
model is sufficient to describe the relationship between y and x. For this it is essential that
the model assumptions hold.
7.2.1 Regression line with slope 0:
Model: yj = α + Ej
Estimate αˆ for α : αˆ = y (= mean of the yj's)
Minimal Sum of Squares (MSQ): Smin = Syy
Degrees of freedom (df): df = n-1
7.2.2 Regression line with y-intercept 0

Model: yj = βxj + Ej
x y
ˆ
β=∑
ˆ
j j1
2
Estimate β for β: x
∑ j
MSQ: Smin = y 2
− ( x y )2 .
∑ j ∑ j j
2
x
∑ j
df = n-1
7.2.3 Intercept = 0. Slope = 0

Model: yj = Ej
no parameters to estimate
MSQ = ∑y j 2
df = n
Other models:
= α0 + βx y
= α + β0 x y
= α 0 + β0 x
The estimates and MSQ can be obtained by differentiation of the corresponding Sum of
Squared Residuals!
These special models are only useful if the data really speaks for them!
7.2.4 Comparison of models

This is done by performing an Analysis of Variance (ANOVA). ANOVA is a widely spread
statistical technique. (Comparison of more than two means, Experimental Designs etc.) .
ANOVA compares the Minimal Sum of Squares of the models to each other. The basic
results for the comparison are filled in a table, the ANOVA-Table.
ANOVA-TABLE
Model MSQ df
0 min 0
H0 S df
A Smin df
0
Reduction S min - Smin df0 - df
Ho denotes the null hypothesis, A the alternative model. By H0 the test-statistic
0
F = (S min − Smin ) (df 0 − df )
S df
min
0
is distributed according to a F-distribution with m1 = (df - df) and m2 = df degrees
of freedom.
0
Significance limits: Look up F 0
in Table A-9 with m1 = (df -df) and m2 = df degrees
1−α;(df −df ),df
of freedom.
Decision : If F > F then conclude that the simplification to the null-model H0 is not
0
1−α;(df −df ),df
permitted, therefore the alternative model has to be used; otherwise (F≤F1-α) there is
no evidence that the null-model should be rejected.

7.2.5 Standard deviation of the estimates
The ANOVA of the null hypotheses H’0 : y = α and H'’0 : y = βx, resp. against the alternative
A : y = α + βx allows us to compute the standard deviations (sd) of the estimated parameters
ˆ
αˆ and β in the alternative model A : y = α + βx: let F(β = 0) be the calculated F-statistic for the
test H’0 against A (e.g. ANOVA for β = 0) and correspondingly F(α = 0) for H’’ 0 against A. Then
sd(αˆ ) = αˆ
F(α = 0)
ˆ
ˆ β
sd(β ) = F(β = 0)
so that the regression line of the alternative (full) model is
ˆ
y= αˆ + β x
(sd (αˆ )) (sd ( β ˆ ))
In order to obtain the standard deviation in a model A : y = βx, this model has to be
compared with the null-model H0 : y = 0.
For A : y = α, compare A to H0 : y = 0
For A : y = α0+βx, compare A to H0 : y = α0
For A : y = α+β0x, compare A to H0 : y = β0x
1-α confidence intervals for the estimated parameters can be obtained by calculating
αˆ ± t1−α / 2,msd(αˆ )
and
ˆ ˆ
β ± t1−α / 2,msd(β )
where m equals the degrees of freedom in the alternative model and t1_a/2,m can be looked
up in Table A-4.

7.2.6 Coefficient of determination
The prediction of y with a regression line is more or less accurate dependent on the degree
of linear dependence. How well does the regression line fit the data? A measure to express
the relative accuracy of prediction compared with the total variation of y is the coefficient of
2
determination r . In the case of a regression line the coefficient of determination is the
square of the correlation coefficient
s2 S2
xy xy
r 2 = rxy2 = =S S
sx2 sy2
xx yy
The total variation of y can be partitioned into two components. Total variation = explained
variation + unexplained variation
where
total variation ∑(yi − y )2
unexplained variation
= ∑(yi − yˆ ) 2 (observed - predicted)
explained variation
∑(yˆ − y )2
The coefficient of determination is the ratio
r 2 = explainedvariation
totalvariation
2
In this form r is defined also for non-linear curves and multiple regression.

7.2.7 Transformations before Regression Analysis
Kind of transformation Type of relation Regression after
transformation
y' = ln y, x(y > 0) y = αeβx y' = ln+ βx
(exponential)
y, x' = ln x,(x > 0) y = α + β ln x y = α + βx'
(logarithmic)
y' = ln y, x' = ln x,(x,y > 0) y = αx β y' = lnα + βx'
(exponentiation)
y' = 1 y , x(y ≠ 0) y = (α + βx)−1 y' = α + βx
y, x' = 1 x (x ≠ 0) y=α+β1 y = α + βx'
x
y' = 1 y , x' = 1 x (x,y ≠ 0) y= x y' = α + βx'
αx + β
(hyperbolic)
y' = ln y, x' = 1 x (y > 0, x ≠ 0) y = e β x y' = lnα + βx'
α
− x y = 1 y' = α + βx'
y' = 1 y , x' = e (y ≠ 0)
α + βe−x
Model Overview for Linear Regression

7.2.8 Multiple and non-linear regression
Often the variable to be predicted is not only dependent on one but on several independent
variables.
Model:
Yj = β0 + β1x1j + β2 x2 j + L+ βp xpj + E j
2
where Ej are independent normal random variables with mean 0 and constant variance σ .
Again these assumptions have to be checked after a model has been fitted.
In this case prediction can be improved in a multiple (or multivariate) regression expressed
as
ˆ ˆ ˆ ˆ ˆ
yˆ = β 0 + β 1x1 + β 2 x2 + β 3 x3 + L+ β p xp
Or the scatter diagram shows a non-linear relationship which demands to fit a non -linear
curve to the data. Often a logarithmic or some other transformation of one or both variables
may lead to a linear relationship, and a regression line can be fitted to the transformed data.
In more complex situations further methods are available.
For multiple, non-linear and mixed (multiple/non-linear) procedures consult literature (ref.
Chatfield 1975)
Flury + Riedwyl (1988) give a very good praxis oriented introduction to multiple linear
regression and multivariate analysis, including Discriminant Analysis, Principal Components,
Identification and Specification analysis.
Example 6
The following data represent
water requirement (%) and
grain fraction 10-32 µ (%)
of n = 34 different cements.
x y x y x y x y
32.9 25.0 38.6 27.2 29.8 23.3 37.2 26.8
35.4 26.3 32.5 25.4 30.3 25.6 42.2 25.6
33.1 24.9 31.3 25.1 36.0 25.2 35.8 24.6
38.7 27.9 35.1 26.9 30.2 24.5 39.1 25.2
33.8 24.8 32.2 24.4 35.6 24.0
32.9 25.0 31.5 24.8 40.6 26.4 40.1 26.6
29.6 24.0 31.1 25.4 40.3 25.2 43.2 27.4
29.7 25.0 38.8 25.6 38.1 25.5 36.7 24.5
36.1 26.2 42.1 27.0 34.7 25.8
The scatter diagram shows a positive linear relationship between water requirement and
grain size. Regression lines to predict x or y are clearly distinct. The lines are identical if rxy is
1 or -1.

If r = 0 the lines are perpendicular, i.e. the best prediction of y is the arithmetical mean y . In
the same way x is the best prediction of x if no dependence exists.
Computations
In a first step five auxiliary sums are provided:
A = Σxi = 1'205.3 n = 34
2 = 43'253.19
B = Σx i
C = Σyj = 867.1
2 = 22'150.93
D = Σy i
E = Σ xjyj = 30'828.44
AC1
s
xy = E− = 2.719545
n−1 n
s2 = 1 B − A2 = 15.918333
n
x
n−1
s2 = 1 D − C2 = 1.131203
y
n−1 n
x = A n = 35.45
= C n = 25.05
Sxy = 89.7550
Sxx = 525.3050
Syy = 37.3297
Smin = 21.9973
df = 32

σˆ 2 = 0.6874
σˆ = 0.8291
Correlation coefficient: rxy = S = 0.64
xy
S S
xx yy
S
ˆ xy
Regression line: β = S = 0.171

xx
αˆ = y − = 19.45
bx
Test: α = 0
Model MSQ df
β 178 l527 33
H0 : y = x
A : y = α + βx 21.9973 32
Reduction 156.1553 1
F= 156.1553 = 227.16 = F(α = 0)
21.9973 32
F.95;1,32 = 4.17 (Table A-9)
Simplification not permitted
Test β = 0
Model MSQ df
Ho : y = α 37.3297 33
A : y = α + βx 21.9973 32
Reduction 15.3324 1
F= 15.3324 = 22.304 = F(β = 0)

21.9973 32
F95,1,32 = 4.17 (Table A-9)
→ Simplified model not allowed !
We obtain the following standard deviations for the parameter estimates αˆ ˆ
and β
ˆ
sd(αˆ ) = a = 19.45 = 1.2905
F(α = 0) 227.16
ˆ ˆ 0.171
β
sd(β ) = F(β = 0) = 22.30 = 0.0362
and the regression line is:

= 19.45+ 0.171x σˆ = 0.8291
(1.29) (0.036)
ˆ
95% confidence intervals for αˆ and β
From table A-4: t.975,32 = 2.042
αˆ : 19.45 ± 2.042 * 1.29 = (16.82, 22.08)
ˆ
β : 0.171± 2.042 * 0.036 = (0.097, 0.245)
2
Coefficient of determination: r = 0.41
Comment:
Only 41% of the total variation of Y can be explained by the dependence of the grain
fraction. The prediction can be improved when further variables are considered in a multiple
regression. The coefficient of determination increases to 69% if C3A- and alkali-content is
included in the equation
y = 17.44 + 0.15x1 + 0.26 x2 + 0.52x3
x1 = grain fraction 10-32µ%
x2 = C3A- content %
x3 = K20 + Na2O%

The improvement is visible in a scatter diagram of observed against predicted values.

STATISTICAL INVESTIGATIONS AND THEIR DESIGN
In practice numerous problems and questions cannot be answered unless special
quantitative investigations are performed. This applies to every practical field, starting with
the raw material exploitation and ending with the sale of the finished product. In many cases,
it is not possible to attain complete information as an effective basis for decision making. For
this reason, it is becoming more and more common to resort to statistical investigations in
form of sample surveys and experiments.
The proper performance and interpretation of statistical investigations is essential. The so-
called “lies of statistics” generally have to be attributed to the wrong selection of data, a
misinterpretation, or procedures which do not relate to the objective. Experience has shown
that very rarely data are manipulated intentionally. However, people who employ “wrong
statistics”, are convinced, due to a lack of in-depth knowledge, that their argumentation is
objective and correct.
8.1 The Five Phases of a Statistical Investigation

Despite the variety of potential applications, a statistical investigation can roughly be divided
into five phases in the following order:
Formulation of problem:
exact definition of purpose for which information is acquired
Planning:
purpose-oriented planning of investigation according to statistical principles with the
objective to acquire optimum information at given expenses. In case of doubt and
problems contact an experienced statistician.
Performance:
procurement of data strictly in accordance with the established planning
Evaluation:
summary and presentation of results; inference from sample to population
Conclusions:
realistic interpretation of results
Essential basic principle:
Statistical investigations will only provide truly efficient results if knowledge in the
investigated field is optimally combined with thorough statistical knowledge. Usually this will
lead to teamwork, since the investigator who lacks extensive knowledge in statistics is
usually just as unsuccessful in his attempt to carry out the investigation on his own as the
statistician who is entrusted with the entire problem complex.

8.2 Sample Surveys and Experiments
Statistical investigations are subdivided into two groups sample surveys and experiments -
which differ significantly with respect to their objective and interpretation.
Experiment
In an experiment response variables are investigated, which depend on various factors. The
results are produced by a controlled variation of the factors of primary interest, whereas all
other effects on the response have to be eliminated.
Factors that are being studied may be quantitative (e.g. temperature, concentration) or
qualitative (e.g. type of ample preparation, origin of samples).
In an experiment, causal effects on the response variable can be determined quantitatively if
the necessary provisions have been made in the planning and performance.
Sample survey
Samples are taken from one or several populations (the set of sample units is called sample
of size n). Certain interesting characteristics are measured in order to evaluate
(estimate)unknown characteristics of the populations or to compare populations. Contrary to
the experiment, the results of a sample survey do not permit any conclusions regarding
cause and effect. This is often ignored when correlations between two characteristics are
assessed.
Reliable results of both experiments and sample surveys can only be obtained if the
respective statistical principles have been adhered to in the planning and performance of the
investigation.
8.3 Fundamental Principles in Statistical Investigations
8.3.1 Experiments
The five criteria of a good experiment are:
The experiment should serve a well defined purpose
In a first step the problem has to be clearly identified for each experiment, and the
hypotheses to be investigated have to be determined.
Factors. which are not of primary interest. should not influence the results
Influencing factors, which are not included in the investigation, must be under control at
fixed levels. Otherwise several effects become mixed up and cannot be separated by
statistical methods. Systematic effects of factors, which cannot be controlled, are
eliminated by random allocation of samples to treatments or factor levels and by random
measuring sequence (randomization).
Practical performance of randomization: In example A 1 we assign a number to each
sample, No. 1 - 8 to tablets and No. 9 - 16 to beads. Each number is written on a leaflet
and mixed in a box. The order of measurement is given by blind drawing of the leaflets
from the box.
The experiment should be free of systematic errors
This requirement is partially connected with 2). Systematic errors, due to any change of
effects in time, are eliminated by randomization. An often underestimated source of
systematic errors is the prejudice of the experimenter. In order to avoid this, mainly
“blind" experiments should be performed. The samples are coded with random numbers,
the decoding key only known to a confidential person, who herself is not in a position to
perform analyses.
The experiment should provide a measure of its precision
An estimate of the precision is obtained by a replication of the experiment. For each step
of the experiment several measurements should be taken. In example A 1 it would be
impossible to state whether a systematic difference exists, if only one tablet and one
bead are measured. (Appendix I)
Precision and efficiency of an experiment should be high enough to reach the set goals There
are various measures to improve the precision and efficiency of an experiment:
Reduction of variability by using homogeneous materials and by carefully controlling all
the factors, as well as by strictly observing the analytical regulations.
Increasing the number of replications; this will lead to the problem of determining the
sample volume required to achieve the desired precision.
Blocking: Measurements are performed in homogeneous groups (blocks). Such blocks
may be 'measurements made by the same operator' or 'measurements made on the
same day'.
A special blocking procedure is the paired comparison. In example A 3 both
laboratories measure the compressive strength on samples drawn from the same
cement bag.
Block experiments should be balanced and randomized, i.e. the number of
measurements and treatments is equal in each block and assignment of samples or
treatments to blocks is random.
Special cases of blocking: paired comparison.
Analysis of covariance: factors, which are not subject to the experimenter's control,
are recorded in order to eliminate their effects in a later analysis.
The principles above are valid for any experimentation. Special designs for various problems
are given in the literature. Problems may be: Effects of one or several factors on a response,
additivity of factor effects, splitting of components of variability in a measurement procedure
in order to achieve a prescribed precision at minimum cost, inter-laboratory tests, calibration,
evaluation of systematic errors etc.
Note:
In contrast to a regression analysis with a given set of observations, a carefully designed
experiment allows the evaluation of causal relationships. In regression the effects are
usually mixed and can only be separated due to some mathematical model and not due to
their real origin (ref. Interpretation of rxy section 7.1).

STRUCTURE OF AN EXPERIMENT
Recognize problem
Describe problem in detail and determine hypotheses
Define experimental area; contact statistician; determine factors, levels and number of
replications; elaborate exact analytical regulations; determine a reference basis if
required. It is often necessary to analyse what happens if no treatment is applied (e.g.
Placebo in clinical experiments).
Establish experimental design, taking into account methods to improve precision
(blocking).
Randomization:
random selection of sample units
random allocation of treatments
coding of samples
Specify variables, which cannot be maintained constant (covariables)
Determine number of replications
Determine procedural organization
Determine procedures of statistical data analysis; exact regulation concerning the
manner in which data are to be supplied (form)
Perform experiment
Evaluate results (statistical data analysis)
Draw conclusions
Take measures
Step 2) to 7) are designated as experimental planning . The results of this planning
should absolutely be recorded in an experimental plan. Since experiments are usually
quite expensive, careful planning will definitely be worthwhile.
The request to prepare a written experimental plan entails significant advantages:
the necessity to formulate statements in precise terms
clear conditions and thus less uncertainty in the performance of the experiment
the circumstances of the experiment can be reconstructed, if later on the results are
used for comparison with new results.
8.3.2 Sample Surveys
Sample surveys and experiments are performed in an analogous way. Differences are due
to the different situation. While the purpose of an experiment is to directly produce results,
sampling is carried out to analyse existing characteristics of a population.
Special attention should be paid to the following points:
Target population and sampled population
The target population is the aggregate about which the investigator is trying to make
inferences from his sample. It is usually helpful to focus the attention on differences
between the population actually sampled and the population that is attempted to be
studied.

Sample unit
The sample units have to be described in clear terms. They should not overlap and the
sum of all the sample units must be equal to the investigated population.
In the cement production for instance, if an individual sample is analysed, the results and
interpretations are different from those of an analysis performed on a daily composite
sample.
Representativity and sampling method
Samples only supply unbiased results if they represent the entire population. It is often
mistakenly believed that the large sample size n ensures its representativity, or it is
simply maintained that the sample is representative in order to prevent rejection of the
results.
Representativity can only be ensured by an appropriate, random sampling method,
giving each sample unit an equal chance to be included in the sample (exceptions in
special cases: sampling with unequal probabilities, systematic sampling).
In literature a number of selection procedures are described. Depending upon the problem
situation, various methods ma y be considered. They differ in their practicability, financial
consequences (expenses) and efficiency.
In most cases an optimum selection procedure for a specific problem situation can be found.
STRUCTURE OF A SAMPLE SURVEY
Recognize problem. What information is required about what populations?
Detailed definition of problem
definition of target populations
what information with what accuracy?
Determine the sample population. Where, when and how are the sample units extracted
from the population? Determine the variables to be recorded
Determine sampling method (equal probability sampling, stratified sampling, cluster
sampling, multi-phase sampling, multi-stage sampling, systematic sampling, unequal
probability sampling)\
Determine sample scheme (in accordance with the required accuracy)
Outline procedural organization
Determine procedures of statistical data analysis; exact regulation concerning the
manner in which data are to be supplied
Perform sampling
If possible perform measurements in random sequence
Evaluation and presentation
Conclusions
Take measures

OUTLOOK
This section will give a short outlook to further and perhaps more sophisticated statistical
applications. It also contains references for further readings. All these applications require
more than elementary knowledge of statistical methods.
9.1 Time series and growth curves analysis

This is the analysis of chronological observations, (e.g. daily temperatures at a certain
location or observations over a certain time period of a patient obtaining a certain
medicament). The characteristic of these data are that the observations are not independent
from each other and the method of linear regression cannot be applied.
Literature:
Box, G.E.P. and Jenkins, G.M. (1976). Time Series Analysis, Forecasting and Control,
second edition. San Francisco: Holden-Day, Inc.
Cox, D.R. and Lewis, P.A.W. (1966). The Statistical Analysis of Series of Events. London:
Methuen.
Nelson, C.R. (1973). Applied Time Series Analysis for Managerial Forecasting. San
Francisco: Holden-Day, Inc.
9.2 Categorical and Qualitative Data Analysis

This is the analysis of counts. Many experiments contain qualitative and nonmetric variables,
e.g. when analyzing the number of smokers and non-smokers beyond male and female
persons. Smoking behaviour and sex have no natural numeric values associated with them.
Used statistical methods for such data are
Crosstabulation
Contingency tables
Goodness of fit tests
Log linear models
Logistic regression
Correspondence analysis
Literature:
Agresti, A. (1984). Analysis of ordinal categorical data. New York: Wiley-Interscience.
Bishop, Y.M.M., Fienberg, S.E. and Holland, P.W. (1975). Discrete Multivariate Analysis:
Theory and Practice. Cambridge, MA: MIT Press.
Haberman, S.J. (1978). Analysis of Qualitative Data, Vol. 1: Introductory Topics. New York:
Academic Press.
Haberman, S.J. (1979). Analysis of Qualitative Data, Vol. 2: New Developments. New York:
Academic Press.

9.3 Experimental Designs and ANOVA
This is a wide field containing among others:
One-Way ANOVA
Multifactor ANOVA
Analysis of Nested Designs
Full and Fractional Designs
Response Surface Analysis
Covariance Analysis
and a lot more.
Literature:
Box, G.E.P., Hunter, W.G. and Hunter, J.S. (1978). Statistics for Experimenters. New York:
Wiley.
Neter, J. and Wassermann, W. (1974). Applied Linear Statistical Models. Homewood,
Illinois: Richard E. Irvin, Inc.

9.4 Multivariate Methods
This title involves the analysis of multivariate data. It is not appropriate to analyze
multivariate data univariate, because correlations among the different variables are not
taken into account.
Statistical Methods:
Correlation Analysis
Covariance Analysis
Multiple Regression
Principal Components
Factor Analysis
Discriminant Analysis
Cluster Analysis
Canonical Correlations
Multidimensional Scaling
and a lot more.
Literature:
Sieber, G.A.F. (1984). Multivariate Observations. New York: Wiley.
Flury, B. and Riedwyl, H. (1988). Multivariate Statistics. A practical approach. London, New
York: Chapman and Hall.
Johnson, R.A. and Wichern, D.W. (1982). Applied Multivariate Statistical Analysis. London:
Prentice-Hall.
Morrison, D.F. (1976, 2nd edition). Multivariate Statistical Methods. New York: McGraw-Hill.
Everitt, B.S. (1980). Cluster Analysis, 2nd edition, London: Heinemann Education Books,
Ltd.

9.5 Nonparametric Methods
Parametric methods base on certain assumptions on the data (e.g. normality of residuals
in linear regression, normality of the observations in testing etc.). If these assumptions do
not hold, it is often more efficient to use methods that do not use a particular underlying
distribution function, so called nonparametric methods (e.g. the Wilcoxon Test known for
Section 5.2.4 is a nonparametric test). Some nonparametric procedures exist for:
Tests of binary sequences
Tests for randomness
Tests for location
Comparison of 2 samples
Rank correlation analysis
Goodness of fit tests
Literature:
Conover, W.J. (1980). Practical nonparametric statistics, 2nd edition. New York: John Wiley
and Sons, Inc.
Lehmann, E.L. (1975). Nonparametrics. San Francisco: Holden-Day, Inc.
Gibbons, J.D. (1976). Nonparametric Methods for Quantitative Analysis. New York: Holt,
Rinehart and Winston.
Hollander, M. and Wolfe, D.A. (1973). Nonparametric Statistical Methods. New York: Wiley.
9.6 Bootstrap and Jack-knife Methods

In dealing with complicated functions of data, it is mostly not possible to derive the
underlying distribution function. Bootstrap and Jackknife sometimes provide the possibility to
derive the distribution functions, statistical measures and a lot more at such complicated
data.
Literature:
Efron, B. (1982). The Jackknife, the Bootstrap and Other Resampling Plans. Philadelphia:
Soc. for Industrial and Applied Math.
9.7 Simulation and Monte Carlo Method

By use of simulation and Monte Carlo methods it is often possible to investigate and analyse
complex functions of random variables or systems of events. Simulation incorporates also
the generation of random variables.
Literature:
Fishman, G.S. (1978). Principles of Discrete Event Simulation. New York: John Wiley &
Sons, Inc.
Rubinstein, R.Y. (1981). Simulation and the Monte Carlo Method. New York: John Wiley &
Sons, Inc.

9.8 General Literature
Seber, G.A.F. (1984). Multivariate Observations. New York: Wiley.
Belsley, D.A. Kuh, E. and Welsh, R.E. (1980). Regression Diagnostic: Identifying Influential
Data and Sources of Collinearity. New York: Wiley.
Chambers, J.M., Cleveland, W.S., Kleiner, B. and Tukey, P.A. (1983). Graphical Methods for
Data Analysis. Boston: Duxbury Press.
Flury, B. (1980). Construction of an asymmetrical face to represent multivariate data
graphically. Tech. Rep. No. 3, University of Berne, Dept of Statistics.
Schupbach, M. (1984). ASYMFACE - Asymmetrical Faces on IBM-PC. Tech.Rep. No.16.
University of Berne, Dept of Statistics.
Draper, N.R. and Smith, H. (1981), 2nd ed.). Applied Regression Analysis. New York: Wiley.
Grant, E.L. and Leavenworth, R.S. (1980). Statistical Quality Control, fifth ed. New York:
McGraw Hill.
Montgomery, D.C. (1985). Introduction to Statistical Quality Control. New York: Wiley.
Recommended literature pp. 90 ff
Huff, D. (1974). How to lie with statistics.
Miller, R.G. (1981). Simultaneous Statistical Inference, 2nd ed. New York: Wiley.
Morrison, D.F. (1983). Applied Linear Models. Englewood Cliffs, NJ: PrenticeHall Inc.
STATISTICAL PROGRAM PACKAGES

Very complete and sophisticated packages are (all programs are available for PC's under
DOS):
BMDP Statistical Software Ltd., Cork Technology Park. Cork,
Ireland (phone: 021-542722)
SAS: SAS Inst. GmbH, Cary NC, USA
(phone: (919)467-8000)
SPSS SPSS Inc., Chicago Il, USA (phone: (312)329-3300)
SYSTAT SYSTAT Inc., Evanston Il, USA (phone: (312)864-5670)
STATGRAPHICS STSC Inc., Rockville MD, USA (phone: (301)984-5000)
A review of 49 statistical packages is given in the PC magazine, Vol. 8, No. 5, March 1989.

Appendix I
Example A1:
In the application of XRF- analysis of raw meal it was required to know to what extent the
results are dependent on sample preparation, by pressed powder tablets and by fusion with
Litetraborate, respectively.
With 8 preparations of tablets and beads of the same raw meal the following results for CaO
and Si02 were obtained.
CaO SiO2
Tablets Beads tablets beads
42.63 42.42 13.87 14.15
42.59 42.44 14.01 13.86
42.63 42.45 14.04 13.92
42.80 42.42 14.09 13.83
43.02 42.59 14.26 14.01
42.61 42.49 13.70 13.78
42.40 42.62 13.89 14.24
42.79 42.26 13.77 13.88
We are interested in differences between the preparation methods. To get a quick survey on
the data we mark every observation on the measurement scale for tablets and beads
respectively.

Looking at the graph we suppose a significant difference of CaO-results between
beads and tablets. For Si 02 results no difference is visible.
The final decision whether the differences are significant is made with an appropriate test.
Test situation:
We test the hypothesis that the mean results of both preparation methods are the same
H0: µtablets = µbeads
Two independent samples with small sample sizes N = 8
two-sided test: we have no idea in what direction a possible difference may occur before the
observations are taken.
Test procedure: Wilcoxon-test at α = 5% level
As an alternative representation we can mark the data on both side of one line. In this case
ranks of the observations can be read directly from the graph.
The ranksum of beads is R = 45.5

W = 2 . 45.5. - 8 . 17 = -45
|W| = 45 => 38 = W0.975 (from Table A-8, Appendix III).
Decision:
The difference of CaO results between tablets and beads is significant.
|W| = 1 < 38 = W0.975

There is no significant difference for SiO2 results.

Example A2:
Control of homogenization efficiency.
In intervals of 30 minutes, 20 samples are taken from the raw ma serial stream both before
and after homogenization.
Measured values: CaCO3 content
Total capacity of homogenization silo: 1000 tons
Sample No. CaCO3 before CaCO3 after
homogenization homogenization
1 76.4 76.9
2 76.7 77.0
3 77.5 77.0
4 77.5 76.8
5 76.5 76.9
6 76.0 76.8
7 76.8 76.9
8 77.0 77.0
9 77.1 77.0
10 77.2 76.8
11 77.2 76.9
12 77.1 76.9
13 76.9 76.9
14 76.7 77.0
15 76.8 77.0
16 77.0 76.0
17 77.1 76.8
18 77.0 76.9
19 76.7 77.0
20 76.8 76.9
The graph is chronological order (time-plot) shows a systematical variation of CaCO3
content before homogenization. Except for one outlying value the variation is small after
homogenization.

Before Homogenization
After Homogenization

Using the Dixon criterion to check for outliers we compute
r 22
= x −x =76.8 − 76.0 = 0.8
(3) (1)
x −x 77.0 − 76.0
(n−2) (1)
with x(1) the smallest, X(3) the third smallest and X(n-2) the largest value.
Since r22 = 0.8 > r0.995 = 0.562 (Table A-2, Appendix III) the extreme value is considered
to be a real outliner (with significance level α = 1%). It is deleted for further analysis.
Statistical data description
CaCO3 before CaCO3 after homogenization

homogenization
n 20 (20) 19
x 76.9 (76.87) 76.92
x 76.95 (76.9) 76.9
s 0.36 ( 0.22) 0.076 one outlier deleted: 76.0
v 0.0046 ( 0.0028) 0.0010
xmax 77.5 (77.0) 77.0
xmin 76.0 (76.0) 76.8
R 1.5 ( 1.0) 0.2
Comment:
Mean and median are almost equal. The distribution of observations seems to be therefore
symmetrical about the mean either before and after homogenization. Due to homogenization the
standard deviation is reduced from 0.36 to 0.08 corresponding to a factor of four to five.
Note:
Elimination of the outliner reduces the standard deviation from 0.22. to 0.08 !

Example A3:
In order to control mortar strength a national control laboratory takes every month a sample
of cement in a plant. It was supposed that these control results tend to be lower than results
of internal quality control.
To find out a suspected syste ma tic error between results of control office and plant laboratory
during a year the samples were measured in both laboratories:
Plant x Control y Difference
d=x-y
January 548 526 22
February 540 524 16
March 574 591 -17
Apri1 469 477 -8
May 540 531 9
June 515 431 84
July 520 485 35
August 531 476 55
September 530 490 42
October 464 452 12
November 524 498 26
December 519 516 3
The strength values are paired. Each pair of observation is concerned with the same
sample. To test for a significant difference we use the singed-rank test at a significance level
of α = 5% (two-sided)
H0: µ = 0
Plot of differences di = xi - yi
Sum of signed ranks:

T = (-6) + (-2) + 1+ 3+ 4+ 5+ 7+ 8+ 9+ 10+ 11+ 12 = 62
Since T = 62 > 52 = T0 975 we decide that the difference between results is systematic.

Example A4:
The following data represent 115 measurements of Schmidt-haamer-strength (abutment of a
bridge).
A detailed interpretation of these data requires their representation in a frequency table and
histogram. With the probability paper we check the data for normality.
The cumulative frequency curve plotted on the probability paper is not linear and the
distribution is therefore not normal.
Histogram and tally show that even values of Schmidt-hammer strength are more
frequent than odd values. This seems to be caused by a reading error of scale. Probably
only even values are marked on the measurement scale.

EXTENDED PROBABILITY PAPER
Table A-1 Standard Normal Distribution - Values of P
Values of P corresponding to zp for the normal curve.

z is the standard normal variable. The value of P -zp equals one minus the value of P for +zp,
e.g. the P for -1.62 equals 1-0.9474 = .0526.

Table A-2 Critical values for the Dixon Criterion

Table A-3 Critical values for Outlier-tests, (large sample size n)
Table A-4 Critical values of the t-distribution

2
Table A-5 Critical values of the x -distribution

Table A-6 Values of λ1-α for confidence limits (Range method)

Table A-7 Critical values of the signed-rank test

Table A-8 Critical values of the Wilcoxon-test

Table A-9 (continued) Critical values of the F-distribution : F0 975

Recommended Literature
The following list gives a selection of books on applied statistics for further reading. The list
is restricted to books which are easily comprehensible for users without profound knowledge
in mathematics.
C. Chatfield "Statistics for Technology" Chapman and Hall, London
(1975), 350 pages
General survey on statistical methods with good
comments on the interpretation of statistical results.
Special chapters on regression, design of experiments and
quality control
Noether "Introduction to Statistics: A Fresh Approach" Houghton
Mifflin Company, Boston (1971) 230 pages
Modern statistical methods, especially for the analysis of
experiments. Good explanation of basic statistical ideas
and problems based on intuition.
M.G. Natrella "Experimental Statistics" US Department of Commerce,
NBS Handbook 91 (1963)
"Cookbook" with many test procedures and designs of
experiments. Often it is somewhat difficult to know what
procedure has to be chosen.
M.R. Spiegel "Theory and Problems of Statistics" Schaum's Outline
Series,(1961) New York 350 pages
Definitions and procedures of classical statistics
accompanied by many examples (875 solved problems)

Chapter 14
Case Study

Case Studies
Case studies are not included here


V2 Materials Technology 2

Uploaded by

Copyright:

Available Formats

V2 Materials Technology 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

V2 Materials Technology 2

Uploaded by

Copyright:

Available Formats

What are the main reaction pathways encountered during clinker formation?

What are the main reaction pathways encountered during clinker formation?

What are the main reaction mechanisms involved in clinker burning?

What are the main reaction mechanisms involved in clinker burning?

Volume 2

1. Clinker Burning: Material Technological Aspects....................................................... 2

Cement Manufacturing Course Volume 2 - Page 1

Clinker Burning: Material Technological Aspects

Cement Manufacturing Course Volume 2 - Page 2

Cement Manufacturing Course Volume 2 - Page 3

1.1 Important Features

1.3 Basic Aspects of the Material Changes Involved in Clinker Formation

Table 1 Important aspects of the understanding of chemical reactions

23 REACTION PATHWAYS ENCOUNTERED DURING CLINKER FORMATION

2.1 General Considerations

Cement Manufacturing Course Volume 2 - Page 5

2.2 Basic Sequence of Reactions

Reactants + products Alite + Belite

Table 2 Sequence of reactions occurring in a rotary kiln

Cement Manufacturing Course Volume 2 - Page 6

2.3 Mineralogical and Chemical Characteristics of Raw Mixes

Table 3 Minerals frequently found in cement raw mixes

Cement Manufacturing Course Volume 2 - Page 7

Cement Manufacturing Course Volume 2 - Page 8

Table 5 Intermediate minerals encountered during clinker production

Table 6 Reasons for the formation of the intermediate products

2.5 Liquid Phase

Cement Manufacturing Course Volume 2 - Page 9

Temperature °C Quantity of liquid phase

Cement Manufacturing Course Volume 2 - Page 10

Cement Manufacturing Course Volume 2 - Page 11

2.5.3 Viscosity of Liquid Phase

Cement Manufacturing Course Volume 2 - Page 12

Figure 3 Minerals identified at different locations (long wet kiln)

Cement Manufacturing Course Volume 2 - Page 13

Figure 4 Sequence of compound formation according to chemical composition

Cement Manufacturing Course Volume 2 - Page 14

Figure 5 Quasi-quantitative variation of minerals with temperature

Cement Manufacturing Course Volume 2 - Page 15

Type Example from clinker burning

reaction type Example from clinker burning

Type Example from clinker burning

Cement Manufacturing Course Volume 2 - Page 16

3.2.1 Structural Changes

Figure 6 Arrangement of the atoms in low and high quartz

Cement Manufacturing Course Volume 2 - Page 17

c) alite and belite transformations

Cement Manufacturing Course Volume 2 - Page 18

(solid) (solid) (gas)

Figure 8 Equilibrium dissociation pressure of calcite with temperature

Cement Manufacturing Course Volume 2 - Page 19

Figure 9 Equilibrium dissociation pressure of calcite and spurrite with

melts This aspect will be further considered in paper 13.

Cement Manufacturing Course Volume 2 - Page 21

Figure 11 Sequence of appearance of phases in the solid state reaction of CaO

0 KINETICS OF CLINKER BURNING

4.1 General Aspects

This phenomenon of the rate of reaction can be schematically represented as follows:

Cement Manufacturing Course Volume 2 - Page 24