Journal of the Tunisian Chemical Society, 2016, 18, 43-51 43

The effect of marl and clay compositions on the Portland cement quality

Sonia Boughanmia*, Islem labidia, Houcine Tissb, Adel Megrichea

a
Université de Tunis El Manar, Faculté des Sciences de Tunis,
UR11ES18 Unité de Recherche de Chimie Minérale Appliquée, 2092 Tunis, Tunisia
b
Laboratory of Production, 7000 Bizerte Cement

(Received: 11 December 2015, accepted: 26 February 2016)

Résumé: La qualité de clinker de ciment Portland dépend de la qualité des matières premières (calcaire, marne,
argile ...) en particulier les marnes et les argiles. Ces matériaux sont les promoteurs de la qualité du ciment et les
principaux contrôleurs du coût de production. L’usine des ciments de Bizerte a réalisé une étude complète des
matières premières dans laquelle une variété d'argiles et de marnes ont été sélectionnés à partir de ses carrières et
préparées pour différentes analyses (analyse chimique, absorption atomique, diffraction des rayons X, fluorescence X)
afin d'élargir la gamme de matériaux existants et d'améliorer la gestion des carrières. Les résultats sont utilisées pour
étalonner un spectromètre ARL 9900 installé dans le laboratoire.

Mots-clés: Clinker, Ciment Portland, matiéres premiéres, Argile, Marne, diffraction des rayons X, Fluorescence X,
Etalonnage, ARL 9900.

Abstract: The quality of Portland cement clinker depends upon the quality of the raw materials (limestone, marl,
clay… ) especially marls and clays. These materials are the proponents of the cement quality and the main controllers
of the production cost. Bizerte Cement factory has carried out a complete study of the raw materials in which a variety
of clays and marls have been selected from its quarries and prepared for different analyses (chemical analysis, atomic
absorption, diffraction X, fluorescence X) in order to expand the range of existing materials and improve the quarries
management.The results are used to calibrate an ARL 9900 spectrometer installed in the laboratory.

Key words: Portland cement clinker, raw materials, Clay, Marls, calibration program, X-ray diffraction, X-ray
fluorescence, calibrate, ARL 9900.

INTRODUCTION The aim of the present paper is to discuss the

Several researches have been made to improve chemistry of the marls and clays and their
the quality of the concrete, most of them are influence on the properties of the raw powder. This
focused on the properties of the anhydrous phases subject was proposed by Bizerte cement factory,
present in the cement and their behavior during the which seeks to extend its quarries. For this reason
hydration. It is known that the reactivity of the we have carried out a systematic study on the
cement and clinker is mainly based on the existing raw materials especially marls and clays.
reactivity of raw meal and therefore the reactivity In fact Portland cement is an equilibrium
of the raw materials [1]. Their use impacts the mixture of raw materials such as limestone, clay
burnability of the raw meal, the chemistry of and other additives, according to their importance
clinkering (additional minor elements), the burning in the manufacture of cement, clays and marls
profile and burner design, the refractory lining, the come in second place after the limestone as raw
chemistry of emission, the properties of the doped materials. The prepared mixture or the raw meal is
clinker and its grindability, as well as the reactivity milled and calcined at 1450°C then the blend
of the final product and its performances [2]. allows to obtain Clinker in which gypsum was

*
Corresponding author, e-mail address : ingenieursonia@gmail.com
44 Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51

added (a setting retarder) to get the Portland B=(AM+1)*(SR+1)*0.01

Cement [3]. The Portland cement clinker is mainly C=[2.8*(AM+1)*SR+1.65*AM+0.35]*4.071
composed of calcium and aluminum silicate phases. D= -[2.8*(AM+1)*SR+1.65*AM+0.35]*0.01
Clays and Marls participate by the contribution of
silica, aluminum and iron oxide necessary to This formula shows that the lime saturation
perform the main mineralogical cement phases: factor is not a linear function of the alite amount.
Alite (C3S), Belite(C2S), Aluminate (C3A) and The application of this formula using the LSF
Ferrite (C4AF) [4]. These aluminum phyllosilicate versus mineralogy (alite) instead of chemistry
materials contain also other elements such as Na2O, appears to be more efficient in considering the
K2O, Cl and SO3, as impurities which could affect presence of minor elements and their effect on the
the cement quality and even the process of its phase content [8].
fabrication. Such elements can enter in the clinker
composition and then condensate in the mineral 2. The silica ratio (SR)
phases and as a result affect the physico-chemical The SR determines the relationship between the
properties of the material [5]. At high solide and liquid phases in the clinker, it controls
concentrations these impurities could decrease the the burnability [9]. Its value should not exceed
durability of concrete and even cause damage of these limits [2.1 ; 3.4] in the raw meal. The SR
the furnace [6]. ratio controls mainly the composition of clays and
In order to get a stable quality of the clinker, the marls,i.e the higher the SR the higher the amount
quality of the raw material resources and the raw of silicate phases (alite and belite). The usual
meal must be evaluated before undertaking a formula of the SR modulus is shown below
particular strategy for the burning and cooling SR=SiO2/(Al2O3+Fe2O3)
processes. The chemical composition of these raw
materials should be controlled through the ciment 3. The alumina modulus (AM)
production line. This can be assessed by the The AM as described in the following formula:
following modulus: Lime saturation factor (LSF), AM=Al2O3/Fe2O3.
the silica ratio (SR), the alumina modulus (AM) It can determine the viscosity of the liquid phase
and the equivalent alkalis (EA). during the clinkerisation. It ranges between 0.64
and 2.5. A high value of AM makes the alite
1. Lime saturation factor (LSF) training difficult which destroys the early strength
Different formulas of LSF exist in the literature, of concrete.This modulus can be refined by adding
all of them are based on the quaternary diagram a good iron ore [10].
CaO-SiO2-Al2O3-Fe2O3. The most used one is : As mentionned above all these modulus, the lime
saturation factor (LSF) associated with the silica
LSF=100*CaO/[(2.8*SiO2)+ ratio (SR) and the alumina modulus (AM), are well
(1.2*Al2O3+0.65*Fe2O3)]. known as a useful tool to design the raw meal of
Portland cement clinker. In our work we are more
This module gives information about the quality focused on the silica ratio and the alumina
of the produced clinker and the degree of the modulus. Another modulus which can be useful in
burnability, it can also give an information on the this study is the alkali equivalent.
alite and the presence of minor elements introduced
by the use of alternative fuel or raw material, which 4. Alkali equivalent (EA)
highlights the necessity of a relation between LSF An alkali equivalent (EA) calculation, EA=
and the mineralogy of the clinker to fill the gap Na2O + 0.659 K2O, seems necessary in order to
between the process and the product. The following prevent the penetration of alkalis (Na 2O, K2O) into
relations have been established [7]. the body of the refractory bricks which reduces
their durability. Furthermore, alkalis can also cause
LSF=100*[(A+B*C3S)/(C+D*C3S)] a massive ring formation, which can seriously
restrict the movement of material and gas and lead
where the coefficients A, B, C and D are to structure and composition changes in the
represented by the following: refractories [11]. The EA defines the total alkalis
A=7.602*SR*(AM+1)+ 6.718*AM+1.429 converted into Na2O. In some countries, Standards
 Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51 45

restrict the alkali equivalent up to 0.6 to prevent (HCl)), it must be combined or calcined with an
expansive reactions between the alkali and some other body, forming a decomposable combination
aggregates in the concrete. in contact with acid [13].
In this paper the aim is to perform an analytical A description of the dry and the wet dissolution
program for clays and marls by X ray. This processes is assured in the following paragraph.
program allows to obtain a better management of 1.1. Dry process
the existing resources using the above mentioned 0.5 g of each sample is introduced into a
modulus (LSF, AM, SR), allowing an easy platinum crucible with 1-2g of NaKCO3 added to
assessement of the produced clinker quality. facilitate the melting of sample. The whole is kept
in the furnace at 1000°C for one hour, then the
EXPERIMENTAL SECTION sample is immediately removed and kept into a
1. Samples preparation desiccator to prevent the adsorption of moisture on
The samples of marls and clays are firstly dryed the surface of the powder. After the crucible with
in an oven at a suitable temperature of 110°C for the material is introduced into HCl/ H2O solution.
24 hours to eliminate the water and facilitate the Heated to dry at 80°C the remaining material is
grinding, then the particule size was reduced in a mixed with10 mL of HCl (37%) and boiled for 15
laboratory disc mill, by means of predominantly to 20 minutes with the addition of the hot water,
horizontal vibration, the material is ground by finally the solution is cooled under the hood and
impact and friction and at the same time filtered with precision to minimize the losses.
homogenized. The milled powder usually gives a This solution allows the determination of the
zero refusal on a sieve of 60 microns after amounts of aluminum (Al2O3), magnesium (MgO),
grinding. Finally the samples are kept in numbered iron (Fe2O3), calcium (CaO) by complexometric
holders and prepared for X-ray diffraction analysis, titration with EDTA while the residue will be used
chemical analysis and X-ray fluorescence analysis. for the determination of SiO2 amount by
The classification of the existing oxides of gravimetry [14].
marls and clays is done according to their amonts, As mentioned above the dry process is
we have major, minor and trace elements (see performed using NaKCO3, so the use of this
Table I). method gives incorrect values of the % Na 2O and
The determination of the loss on ignition (LOI) %K2O, therefore another type of dissolution is
is a preliminary step before initiating the chemical needed: the wet process.
analysis, it helps to classify the samples collected 1.2. Wet process
in two main groups clays and marls using the 0.25 g from each sample is introduced in a
below classification. Teflon beaker with 5 mL of HNO3, 5 mL of HClO4
Any chemical analysis of silicates starts with an and 10 mL of HF and maintained on a heater plate
appropriate disintegration of the matter and the at 150°C until dry. After we added 5 mL of HClO 4
dissolution of the attack result. Depending on the and 20 drops of HCl the solution is filtrated, a part
examined silicoaluminous materials, there's a
significant difference in the choice of the means of
Table I: Classification of oxides according to their content
attack which depends largely the success of the
actual analysis [12]. For this reason two
Classification Elements Amount
dissolution process are followed: the dry process
and the wet one. Majors Si(SiO2), Al(Al2O3), Ca (CaO) >10
The dry process is used for the determination of
the SiO2, Al2O3, Fe2O3, CaO, and MgO amount. It Minors MgO, Fe2O3, SO3 0.5-10
is used when a silicate is not easily or hardly
Traces Na2O, K2O <0.5
attacked by acids (especially hydrochloric acid
46 Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51

of the filtrate is used for the determination of the flame photometer installed at Bizerte cement
amount of Al2O3, Fe2O3, CaO, MgO while the laboratory.
other part for the determination of %Na 2O et %
K2O by atomic absorption spectrometry. The other % Na2O = -0, 021 + 0.394 * ANa2O
elements such as chloride and sulfates are % K2O = -0.098 + 0.173 * AK2O
determined respectively by Volhard's method and
by gravimetry. The following formulas are the equations of
straight calibration curves obtained by linear
2. Techniques extrapolation (Beer-Lambert law) of couples of
The milled samples are pressed with a points (%, absorbance) measured on standard
laboratory hydraulic press of 10 tons into solutions.
briquettes for the XRD and XRF analysis, using a
ARL 9900 diffractometer operating with CoKα RESULTS AND DISCUSSION
radiation (λ=1.789). The samples were scanned 1. X-ray diffraction analysis
from 2θ 8° to 80°. Step size is 0.03, the counting The mineralogical phases identification was
time is 1s per step. The tension was set at 40 kV carried out using the X’Pert HighScore Plus. A
and the current at 40 mA. The divergence slit was good interpretation of the diffractograms of clays
fixed at 1º, and the receiving slit was set at 0.0 and marls minerals requires a good knowledge of
mm. This instrument is equipped with a series of their structure to ensure the validity of the license
XRF fixed channels (monochromators), an XRF model and the used parameters. The main
goniometer, and an XRD goniometer integrated in mineralogical present phases are quartz (SiO2),
such a manner that both XRF and XRD calcite, smectite, kolinite and dolomite. All
measurements can be carried out on the same samples have a similar mineralogical composition
sample with the same tube conditions and under except the peaks intensity differs, we note the
vacuum. appearance of new reflection in the last sample it is
The determination of Na (Na2O) and K (K2O) contributed to NaAlSi3O8. The XRD analysis
present in the marls and clays is performed using a provides information about the mineralogical

Table II: Chemical analyses results of the sample A

Test number SiO2 Al2O3 Fe2O3 CaO MgO Na2O K2O CL SO3 LOI

1 44.55 14.68 7.34 5.11 5.92 0.026 0.66 0.042 0.264 16.82

2 15.14 7.18 6.08 5.23

3 14.93 5.82 5.34 6.44

4 15.59 7.2 5.53 6.58 17.82

5 44.7 13.56 7.02 5.42 6.5

6 16.61 6.95 5.9 5.95

7 16.62 7.98 5.91 5.57 16.32

8 16.62 8.46 4.7 6.49

9 44.46 15.6 6.54 6.4 6.92

Average 44.57 15.48 7.14 5.8 6.10 0.026 0.66 0.042 0.26 16.98
standard
0.121 0.98 0.76 0.65 0.54 - - - - 0.62
deviation
Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51 47

Figure : X-ray diffraction patterns of A,B and D samples

Figure : X-ray diffraction patterns of F,G H and I samples

Figure : X-ray diffraction patterns of C and E samples

48 Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51

Figure : X-ray diffraction patterns of sample J

composition, so it informs about the chemical 3. X-ray fluorescence analysis

composition of samples which makes easy the To measure the X-ray intensity on the various
choice of reactive and the samples dissolution. elements a pellet was prepared for each sample and
analyzed by X-ray fluorescence three times in
2. Chemical analysis results order to have the average and standard deviation of
A serie of chemical analyses was performed on three determinations (table IV).
each sample in order to minimize the error and Static given on the calculation of the average
increase the fidelity of analysis. and standard deviation have shown an acceptable
The average of the analyses of each sample are relative standard deviation except for Cl which
summarized in table III, a calculation of the recorded a value of 2.78. The average x-ray
maximum and minimum is assured to determine intensities of the analysed elements for marls and
the extent of the prepared range. clays samples are summarized in the table V.

Table III: Chemical composition of samples

Samples SiO2 Al2O3 Fe2O3 CaO MgO K 2O Na2O Cl SO3 LOI

A 44.57 15.48 7.14 5.80 6.10 0.66 0.026 0.042 0.264 16.98
B 43.81 14.09 6.75 7.71 6.14 0.73 0.034 0.053 0.569 19.27
C 56.74 9.77 4.97 8.16 2.26 0.43 0.035 0.047 1.522 7.46
D 42.47 18.14 5.12 8.00 5.76 0.34 0.022 0.046 0.171 16.88
E 46.30 17.77 9.43 6.66 2.21 0.65 0.025 0.045 0.044 15.48
F 36.25 13.24 6.12 16.45 2.82 0.27 0.034 0.044 3.196 22.79
G 30.57 7.89 3.67 28.95 4.02 0.18 0.03 0.039 1.807 31.58
H 32.48 8.67 5.05 26.32 3.2 0.2 0.024 0.046 0.967 27.94
I 23.74 7.05 2.96 36.79 3.2 0.15 0.039 0.04 0.699 27.27
J 61.34 20.63 1.52 1.782 2.49 0.07 0.717 0.025 0.171 1.655
MAX 61.34 20.63 9.43 36.79 6.14 0.66 0.717 0.053 3.196 31.58
MIN 23.74 7.05 1.52 1.782 2.21 0 .07 0.02 0.025 0.044 1.655
 Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51 49

Table IV: X-ray intensity of the sample A

SiO2 Al2O3 Fe2O3 CaO MgO SO3 K 2O Na2O Cl

Essay 1 272.448 109.403 466.864 85.852 40.820 1.953 74.315 1.486 0.383
Essay 2 271.541 109.015 465.403 85.764 40.698 1.971 74.144 1.527 0.401
Essay 3 271.380 109.072 465.676 85.811 40.590 1.980 74.078 1.519 0.403
Average 271.79 109.16 465.98 85.81 40.70 1.97 74.18 1.52 0.40
standard deviation 0.58 0.21 0.78 0.04 0.12 0.01 0.12 0.01 0.01
Relative standard
0.21 0.19 0.17 0.05 0.28 0.70 0.16 0.37 2.78
deviation

These values will be used for the calibration curves concentration already done by chemical analysis
determination of each element by linear regression based on their relative intensities obtained by X-
based on the determination of the gradient a the ray fluorescence intensities. It should cover
order originally b and R correlation coefficient for different levels ranging from 0 to 100% for each
the regression line. All curves presented an R> assayed element but in front of the possible causes
0.97 will be accepted in cases of major elements of error experienced by an X-ray spectrometry
for minors and trace elements R will be accepted analysis system leaving very complicated the
from 0.94. uncertainties determination requires users
4. Tracing calibration curve spectrometer to intersect the curve calibration of
The treated samples are selected according to the instrument in several segments representing the
the amount of each chemical element quantified in tangent to the curve. The Figures 5, 6, 7 and 8
order to have a range covering the variations of show the calibration curves of SiO2 CaO Fe2O3 and
clay and marl suspected to be analyzed in Bizerte K2O.
cement laboratory. These analyses will be used to
prepare calibration curves to calibrate an ARL 5.The results of the calculated modulus
9900 spectrometer. The calibration curves The established analytical program used to
represent the variations of different oxides control the modulus: AM SR and EA in the raw

Table V: The average of X-rays intensities of the various elements analyzed for each sample

Samples SiO2 Al2O3 Fe2O3 CaO MgO SO3 K 2O Na2O Cl

A 271.541 109.163 465.981 85.809 40.702 1.968 74.179 1.51 0.395

B 268.449 102.641 439.441 96.437 47.871 1.123 82.681 1.829 0.506

C 344.846 71.557 241.729 81.453 17.753 1.349 48.900 1.907 0.456

D 256.311 138.118 257.934 94.701 38.410 2.025 38.910 1.260 0.451

E 280.636 125.759 578.218 70.070 18.232 1.122 73.588 1.430 0.440

F 219.128 84.121 339.973 207.729 23.731 16.206 31.241 1.687 0.434

G 182.845 45.497 147.913 418.900 20.720 20.347 20.886 1.390 0.392

H 194.594 60.880 245.905 331.488 18.653 9.481 24.284 1.968 0.467

I 143.524 34.842 110.866 526.220 15.587 9.071 18.291 1.181 0.404

J 360.428 108.490 11.419 17.276 6.131 0.854 9.120 35.368 0.255

50 Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51

Table VI: The results of calculated modulus for the • AM> 1.6
differents samples The melting phase which based on alumina
becomes highly viscous that is cause a casting
Samples AM SR EA
during the Clinkerisation which usually accom-
A 2.17 1.97 0.46 panied with a decrease in the density of the melting
B 2.1 2.10 0.51 phase and increase of the crustin thinness. As
result the heat transfer from the flame to the ferrule
C 1.96 3.84 0.31 increase and the refractory bricks become more
D 3.54 1.82 0.24 stressed. Consequently the risk of cement kiln stop
increase and the clinker will be like a slight
E 1.88 1.7 0.45
drosses.
F 2.16 1.87 0.22 • 0.64 <AM <1.38
G 2.15 2.64 0.15
The liquid phase becomes very tacky and the
H 1.71 2. 36 0.15 crusting becomes very thick and dense. As result
I 2.83 2.73 0.14 the crusting falls down accompanied by the tearing
off the bricks. Furthermore the clinker will be on
J 13.57 2.77 0.76 the form of large blocks that prevent the well
functioning of the cooler and makes difficult the
getting of the desired clinker.
powder. The modulus are calculated for the SR>2.8
different studied marls and clays (see table VI).
The clay or marl fraction is usually controlled In this case the crystalline phase predominates
by the silica modulus which is generally desired the reactivity of the raw powder decrease which
between 2.1 and 2.8 in the following an SR requires an increase in the heat capacity during the
examination. clinkerisation. The clinker becomes dusty forming
SR<2.1 a screen material which generally accompanied by

Figure 5: A calibration curve of SiO2 in a series of Figure 7: A calibration curve of Fe2O3 in a series of
clays and marls samples clays and marls samples.

Figure 6: A calibration curve of CaO in a series of Figure 8: A calibration curve of K2O in a series of clays
clays and marls samples and marls samples
 Sonia Boughanmi and al., Journal Tun. Chem. Soc., 2016, 18, 43-51 51

a decrease in production further an increase in the [3] H.F.W. Taylor Cement Chemistry Academic Press
specific fuel consumption. New York 1990.
The treated samples present a very variable [4] Michel Venuat, Michel Papadakis « Contrôle et
silica modulus value extending from 1.70 to 3.84 essais des ciments mortiers bétons » 61 boulevard
Saint-Germain- PARIS Ve 1961.
to fix these values the cement factory usually use a
[5] V. K. Klassen, E. P. Ermolenko, D.A. Michin and
correction elements which are mainly based on A. G. Novosyolov Problem of Impurity of Salts of
iron or alumina material. Regarding Bizerte Alkali Metals in Cement Raw Materials Middle-
cement factory the iron ore is added to lower the East Journal of Scientific Research 17 (8) (2013)
silica modulus thanks to its high content of Fe2O3. 1130-1137.
For the alkali equivalent (EA) it represent an [6] J. Nievoll, S. Jorg K.Dosinger and J. Corpus Sulfur
acceptable value exept for the last sample which Spurrite and Rings Always a Headache for the
define as a feldspath such as the main phase is Cement Klin Operateur? RHI Bulletin 2007 pp.35-
NaAlSi3O8 which explain its highest value. 38.
[7] F. Sorrentino Lime saturation factor new insight
Cement Wapno Beton 2 (2008) 82–88.
CONCLUSION
[8] F. Sorrentino Modern Design of Raw Mix
The program prepared can perform a lot of Proportioning of Portland Cement Clinker to be
analyses in a few seconds that ‘s can save the lost Published at the 13th International Congress on the
time usually spent in the sample preparation phase Chemistry of Cement Madrid, 2011.
and the chemical analysis. Furthermore all [9] Mohamed A. Aldieb Hesham G. Ibrahim Variation
analyzes can be performed in better condition of of feed chemical composition and it is effect on
repetability and reproducibility. Thanks to this clinker formation - Simulation process
program Bizerte Cement Factory has integrated an Proceedings of the world congres on engineering
abandoned marls and preserved others to ensure and computer science 2011 Vol II.
the longevity of its quarries. Clays and marls [10] Nicholas Winter Uderstanding cement December
2012.
represent the active part in the cement production
[11] Hegde S. B.Chaturvedi S.1995 Effect of alkalis on
regardless of the used limestone it will be the alumina refractories in cement rotary kilns
proper clay or marl and the necessary correction in Refractories 36 184-188.
order to maintain the good quality of cement. [12] I.A.Voinovitch, J. Debras-Guedon et J. Louvrier
L’analyse des silicates Paris 1962.
REFERENCES [13] Dr Carl Friedheim et L.Gautier Précis d’analyse
[1] A K. Chatterjee Chemistry and engineering of the chimique quantitative, les substances
clinkerization process - Incremental advances and minerals,Paris et Liége.
lack of breakthroughs cement and concrete [14] The French standard CEN 196-2 published by the
research 41(2011) 624-641. French Association of normalization ''AFNOR''en
[2] F. Sorrentino Chemistry and engineering of the 2006.
production process: State of the art Cement and
Concrete Research 41 (2011) 616–623.