Characterization of Electrically Insulating Coatings For Soft Magnetic Composite Materials by Means of Surface Sensitive Analytical Techniques

Characterization of electrically insulating coatings for soft
magnetic composite materials by means of surface

sensitive analytical techniques
by
DIMITRIS NIKAS
Diploma work No. 114/2013

Department of Materials and Manufacturing Technology
CHALMERS UNIVERSITY OF TECHNOLOGY
Gothenburg, Sweden
Diploma work in the Master programme Materials Engineering
Performed at: Department of Materials and Manufacturing Technology

Chalmers University of Technology
SE-412 96 Gothenburg
Supervisor(s): MSc. Christos Oikonomou

Examiner: Professor Lars Nyborg

Chalmers University of Technology,
SE - 412 96 Gothenburg
Characterization of electrically insulating coatings for soft magnetic composite materials by means of surface
sensitive analytical techniques
DIMITRIS NIKAS
© DIMITRIS NIKAS
Diploma work no 114/2013

Sweden
Telephone + 46 (0)31-772 1000
Reproservice
Gothenburg, Sweden 2013
Characterization of electrically insulating coatings for soft magnetic composite materials by means of
surface sensitive analytical techniques
DIMITRIS NIKAS
Abstract
Powder metallurgy (PM) is a cost effective and efficient process for producing net-
shaped parts of high tolerances and homogeneous properties in large production
volumes. These are few of the reasons why this technique is preferred in
manufacturing soft magnetic composite (SMC) materials for electromagnetic
applications. The concept of SMC consists of encapsulating pure iron particles with an
electrically insulating coating which yields compacts of high bulk resistivity and
isotropic magnetic behaviour. Such materials provide great freedom in design and
exhibit improved magnetic performance compared to more traditional laminated
steels and ferrites for a wide range of frequency applications.
In the present thesis, surface characterization of such type of composite powder in as-
received state was performed using surface sensitive analytical techniques.
Information regarding the nature of the surface layer is important in designing the
process and tailoring the properties of a product towards a desired application. The
material under examination was commercially available phosphated iron grade which
was analyzed using X-ray photoelectron spectroscopy (XPS) and high resolution
scanning electron microscopy (HR SEM) coupled with energy dispersive X-ray analysis
(EDX). Based on these techniques, thickness estimation, morphology, coverage,
chemical state and compositional depth profile analysis of the insulating coating was
conducted.
The aim of this investigation was to evaluate the effect of the surface characteristics
and geometry of powder on such type of analyses. For this purpose, gas and water
atomized high purity iron grades, having spherical and irregular shape respectively,
were used. Both kinds of powder were covered with uniform oxides layers of different
thicknesses and they were both divided into different particle size fractions.
Furthermore, standards and flat samples of similar composition to the metal powders
were prepared and evaluated accordingly using additional thermoanalytical and X-ray
diffraction (XRD) techniques in order to acquire reference values.
The results showed that the as-received state appears to have a greater contribution
of oxide thickness than when it is fractioned. This is supposed to be a result of the fact
that the unsieved powder has a significant difference in the particle size and this can
cause more shading effects during the XPS analysis.Moreover, the SMC powder
exhibited a complex surface layer structure consisting of a mixture of iron oxide and
iron phosphate compounds of the order of few nanometres.
Keywords: Powder Metallurgy, soft magnetic composites (SMC), gas and water
atomized iron powder, surface oxides, phosphate coating, XPS, HR SEM, EDX
Acknowledgements
This master degree project was done from January 2013 to June 2013 at Department
of Materials and Manufacturing Technology at Chalmers University of Technology.
At the end of this work I would like to thanks Höganäs AB for providing the powder
materials studied at this project.
I would like to thank my supervisor Christos Oikonomou for the support and guidance
during the entire six month period.
I would also like to thanks my examiner Prof. Lars Nyborg for the useful feedback and
Assoc. Prof. Eduard Hryha for the interesting discussions we had during this project
Finally I would like to thank all the people in the department that helped me during
this time, especially Urban Jelvestam for the help with XPS, Dr. Yiming Yao for the help
with use of SEM, Dr. Kumar Babu Surreddi and MSc. Julio Maistro for the XRD tests and
Dr. Raquel De Oro Calderon and Eduard Hryha for the help with the TGA analysis.
Also many thanks to my classmates for the help and long discussions we had almost
every day at the department during this period that helped me with the present work.
Gothenburg, December 2013
Dimitris Nikas
Table of Contents
Chapter 1................................................................................... 1
Introduction .............................................................................. 1
Chapter 2................................................................................... 3
Theoretical background ............................................................. 3
2.1 Powder metallurgy................................................................…..........2
2.2 Soft magnetic composites powders…………….....................................4
2.2.1 Iron core losses in electromagnetic applications........................5
2.2.2 Advantages and Properties of soft magnetic composites...........6
2.3 Phosphating process..........................................................................7
2.3.1 Kinetics of the phosphating process.............................................9
2.4 X-Ray photoelectron spectroscopy (XPS)..........................................9
2.5 Scanning electron microscopy (SEM) and Energy dispersive X-ray
spectroscopy (EDX)................................................................................11
2.6 X-ray diffraction (XRD).....................................................................13
2.7 Thermogravimetry (TG)...................................................................15
Chapter 3..................................................................................17
Materials and Experimental procedures ..................................17
Chapter 4..................................................................................19
Results ......................................................................................19
4.1 Analysis of the iron oxide standards................................................18
4.2 Analysis of the ABC100.30 powder..................................................25
4.3 Analysis of the Gas atomized powder..............................................40
4.4 Analysis of the phosphate standards...............................................57
4.5 Analysis of the Somaloy 500 powder........................................... ...62
Chapter 5..................................................................................75
Discussion .................................................................................75
5.1 Pure iron powders and standards....................................................76
5.2 Somaloy powder and phosphate standards.....................................78
Chapter 6..................................................................................80
Conclusions and future work ....................................................83
References ................................................................................82
Chapter 1
Introduction
Soft Magnetic Composite (SMC) materials designed for electromagnetic
applications are being produced based on traditional Powder Metallurgical (PM)
techniques. The latter provide profitable and sustainable production routes that yield
net-shaped parts, minimizing thus the required post treatment steps. The concept of
SMC technology is based on encapsulating each individual iron particle with a thin, in
the order of nanometres, electrically insulating surface layer and subsequently
pressing them together in a three dimensional array form of a finished compact. In this
manner, uniform and isotropic 3D magnetic properties are acquired that offer design
freedom in creating unique and innovative application concepts
The SMC products exhibit improved magnetic performance for a wide range of
frequency applications as opposed to more traditional laminated steels and ferrites.
This is due to the fact that they offer higher bulk electrical resistivity by effectively
confining the deleterious effects of eddy currents, especially at high frequency
applications. As opposed to more conventional PM techniques, production of SMC
parts does not require sintering since that would compromise the viability of the
insulating coating. Conversely, a heat treatment at lower temperatures is performed in
order to relax any stresses introduced during compaction, improving thus the magnetic
induction of the part and promoting curing of the surface layer for mechanical
strengthening. Additionally, it is possible by changing the thickness of the layer and the
particle size distribution of the SMC powder to essentially tailor the magnetic
properties of a compact towards a desired application.
From the above, it is obvious that the insulating coating is the paramount
feature of the SMC technology and thorough investigations of its nature and behaviour
are required to improve and tailor the process in producing parts of desired properties.
The objective of this thesis work was to implement techniques and methods that
would provide valid information on the surface chemistry and characteristics of the
surface layer. Surface sensitive analytical techniques, namely X-ray photoelectron
spectroscopy (XPS) and high resolution scanning electron microscopy (HR SEM)
combined with energy dispersive X-ray analysis (EDX) are traditionally used in material
science to investigate, characterize and model surface reactions.
The aim of the current experimental investigations, based on the
aforementioned techniques, was to determine the morphology, homogeneity,
chemical composition, thickness and cohesion of the insulating layer to the iron
particles. For this reason, standards were acquired and analysed in order to accurately
determine the chemical composition and state of the material of interest. Additionally,
thermodynamic considerations regarding the composition and stability of the
insulating coating were addressed and compared with the experimental findings.
Furthermore, the contribution from factors related to particles geometric
characteristics, as size, shape, surface area and morphology, on the analysis of the
surface layer were also evaluated based on a theoretical model. The SMC grade was
divided into different size fractions and compared to metal powders of similar sizes
and various surface characteristics in order to quantify those effects.
1
Chapter 2
Theoretical background
2.1 Powder metallurgy
Powder metallurgy (PM) refers to the processing science of manufacturing complex

shaped parts of high tolerances in large quantities and having low production cost. In
conventional press and sinter PM process this is done by using metal powder as base
material which is then densified to parts of the desired shape through custom made
dies at high pressures. These so called “green bodies” are subsequently sintered at
high temperatures, below the melting point of the material, in order to promote
bonding between the metal particles and thus improve the mechanical properties.
Prealloyed powder allows processing below melting temperature and sometimes even
at solid state, thus avoiding problems like segregation or shrinkage during solidification
in castings.
The PM technology has many advantages over other manufacturing techniques such
as: material saving (around 95% of the material is used), energy per time saving (i.e.
lowest energy consumption per kilogram of finished part), complex shapes, high
reproducibility, wide range of properties and uniqueness in certain fields (e.g. it is the
only process that allows manufacture of cemented carbides, CMCs/MMCs, refractory
metals, structural ceramics, etc). Some of these advantages emerge from the fact that
PM yields the possibility to manufacture more complex microstructures where
porosity or oxide dispersion and homogeneity required are superior as compared to
the results of other processes. Many methods exist for producing PM parts such as hot
or cold isostatic pressing, metal injection molding, powder forging, solid state
sintering, liquid phase sintering, etc. The whole tree of the powder metallurgy
processing appears from the illustration below (Fig 2.1).
Figure 2.1: PM process route (author illustration redrawn from [2]).
3
Fields of application of PM include 1) applications with high cost savings where
complex shapes, small parts or large series are generated 2) applications where
improved properties for better performance are needed and 3) applications where PM
is the unique way as mentioned above.
Powder characterization
All of the above mentioned processes start with a powder therefore it is important to
understand the nature of the material to be able to evaluate the process and the final
product properties. A powder consists of particles, which are defined as the smallest
units of powder that cannot be subdivided further. Typical sizes in metal powder are
between 1µm-0.1mm (smaller than sand). The main attributes of a powder that need
to be considered and are studied in this report are:
1) particle size/distribution
2) particle shape
3) surface area
4) surface characteristics
5) composition, homogeneity and contamination
Also to fully characterize a powder it is necessary to describe how it is manufactured

and understand the various parameters in every step of the process [1].
Manufacturing of iron powders
The main categories of manufacturing techniques are:

 Mechanical Fabrication
 Electrolytic fabrication
 Chemical fabrication
 Atomization fabrication
Mechanical alloying involves fracturing and shearing powdering material in order to

achieve very small particles and then mixing them mechanically by milling using hard
balls in a large jar. In electrolytic fabrication, powder is precipitated at the cathode of
an electrolytic cell whereas in chemical techniques thermal decomposition or
precipitation from a liquid is mostly used.
Since the metal powders studied in the present thesis are manufactured by either gas
or water atomization techniques, a more detailed description of their production route
is given below. The main feature of atomization techniques is that a jet stream of gas
or water of high pressure is directed against the molten metal that is fed as shown in
Fig. 2.2,which gives rapid cooling.
4
Figure 2.2: Simplified schematic of a vertical atomizer.
The main difference in the two types of powder is the shape of the particles. Gas
atomized powder has spherical particles while water atomized has irregularly shaped
particles. This affects the surface roughness of the powder and thus the compaction
capabilities later in the process to produce the final parts.
Figure 2.3: Drawing of gas atomized (left) and water atomized (right) powder particles.
Moreover, water atomization oxidizes the powder significantly more than gas
atomization, affecting in this way its sintering capabilities and mechanical properties.
For this reason, a deoxidation process step is highly necessary after powder production
by heating in hydrogen [3].
2.2 Soft magnetic composites powders
Soft magnetic composite (SMC) powder consists of ferromagnetic particles that are
encapsulated with an electrically insulating thin surface layer. This type of composite is
used primarily for electromagnetic applications. They are typically manufactured from
normal PM techniques under die compaction of the powder mix to the desired
compact shape followed by heat treatment at low temperatures. No sintering step is
5
required since the formation of necks between the particles would damage the
function of the insulating coating.
Thus, preparation of an SMC finished compact includes the following parts:

a) High purity iron powder, produced by water atomization, is preferred as base
material due to its excellent magnetic properties (e.g. magnetic permeability and
induction) and suitable mechanical properties (e.g. higher green strength due to the
mechanical interlocking between the irregular shaped particles). b) Coating of the iron
particles with an electrically insulating layer making sure that good adhesion and
coverage between the particles surface and the layer is achieved.
c) Mixing insulated particles with a lubricant in order to reduce interparticle and
particles to die wall friction.
d) Compacting the powder by conventional die compaction to form the green body
e) Heat treating the green body to a temperature regime between 400-700oC in order
to relax stresses introduced during compaction (i.e. to reduce hysteresis core losses)
and to promote a curing operation that yields mechanical strength to the part.
The schematic below shows how the coated iron particles appear in a soft magnetic
composite (Fig 2.4) [5].
Figure 2.5: Schematic of SMC particles.
2.2.1 Iron core losses in electromagnetic applications
Magnetic properties of the material comprise the most important part when it comes
to electromagnetic applications. There are two main characteristics of an iron core
component. These are its core loss properties and magnetic permeability.
There are three main contributions to the dissipation in magnetic materials:

1) Hysteresis loss;
2) Eddy current loss;
3) Residual loss.
Residual losses are only important in low induction and high frequency levels and can
be ignored for power applications.
Eddy current losses within an iron core occur due to electrical resistance losses that
the alternating electric field creates. Two main effects are observed in materials when
eddy currents are induced: incomplete magnetization of the material and increase in
core losses. One way to minimize core losses is by using thin laminated materials and
confining in this way the circulation of eddy currents within each laminate.
6
The SMC concept uses composite iron powder which due to its insulating surface layer
allows eddy currents to travel only inside each particle, thus giving relatively high bulk
resistivity. The small non-magnetic distances between every particle also act as air
gaps which decrease the permeability of the bulk material.
As opposed to eddy current losses which are more prominent with increasing
application frequency, hysteresis losses are the main factor in low frequencies and can
be reduced by:
a) Large particle size;
b) High purity iron;
c) Stress relieve annealing step that follows the compaction step.
2.2.2 Advantages and Properties of soft magnetic composites
Soft magnetic composites(SMC) have many advantages over laminated steels that are
generally used today. The most important one is that they exhibit magnetic and
thermal isotropy that yields three dimensional magnetic flux path capabilities.
Furthermore, net shape parts can be produced even in complex configurations using
conventional powder metallurgy compaction processes at a relatively lower cost. The
SMC have similar magnetic saturation induction values to those of wrought steel
laminates and also lower eddy current losses at higher frequencies for a wider range
due to the presence of the insulating coating. Because of these advantages lately they
are preferred as alternative to steel laminates for new electric designs such as electric
motors, fast switching actuators and pulse transformers [6].
As mentioned above, SMC material exhibits significant advantages that make it

appealing for electromagnetic applications due to their properties. Some of these are:
 magnetic and thermal isotropy;
 very low eddy current loss and relatively low total core loss from low to high
frequencies;
 high magnetic permeability at high fields;
 high remanent magnetization;
 high resistivity;
 reduction in size and weight;
 large anisotropy constant;
 low coercivity;
 high Curie temperature.
Additionally, one of the advantages of the SMC technique is the fact that it is possible
to tailor the properties of a component to a desired application. These can be achieved
by changing the particles size distribution or the coating thickness and composition in
order to affect the permeability or core-loss characteristics of the part.
7
2.3 The phosphating process [22]
Phosphating is a well-known metal conversion treatment especially in automotive

industry. It is used to provide the material with a thin layer of a phosphate coating that
gives anticorrosion and wear resistance properties to the surface. It is also used as a
pretreatment procedure to parts that are going to be painted afterwards.
Furthermore, it has been shown that such a coating provides insulation if used on iron
powder thus greatly improving its magnetic properties [11].
Stages of general phosphating process
The stages of the phosphating process are as follows:

 Cleaning. The surface that is to be coated is mechanically cleaned and degreased
usually in an alkaline solution (e.g. this step can be avoided as acetone is used as
solvent).
 Rinsing the material in hot water.
 Cleaning of the surface with an acid (pickling). The acid dissolves possible oxide
films and rust stains that might be present.
 Another step of water rinsing.
 Phosphating either by immersing the material in a phosphate bath or by spraying.
 Water rinsing.
 Drying.
Phosphating chemistry
The main components of a conventional phosphating solution are:
 Water or acetone as the main solvent

 Phosphoric acid (H3PO4)
 Ions (typically cations) of bivalent metals such as Zn2+, Fe2+, Mn2+.
 Accelerator- This substance is needed to increase the coating process rate and
reduce the grain size of the deposited coating. It is usually an oxidizing reagent
such as nitrate, nitrite or peroxide.
When a metallic material, for instance iron, is immersed into a phosphate solution the
iron ions are being provided by the dissolving substrate and a top chemical reaction
takes place on the surface. There the iron dissolution is initiated at the micro anodes
present on the substrate by the free phosphoric acid present in the bath. Hydrogen
evolution occurs at the microcathodic sites [12] according to the following equations:
3Fe + 6H++ 2PO43-= 3Fe2++ 2PO43-+ 3H2 (1)

Fe--> Fe2+ + 2e- (2)
2H+ + 2e---> H2 (3)
8
As these reactions proceed, the pH is locally lowered at the surface. Then, the metal
surface becomes first covered by a thin layer of primary ferrous phosphate which is
water soluble and then a thicker insoluble secondary and tertiary ferrous phosphate
given by the reactions below:
Primary ferrous phosphate (soluble): Fe + 2H3PO4 = Fe(H2PO4)2 + H2 (4)
Secondary ferrous phosphate (insoluble): Fe(H2PO4)2 + Fe = 2FeHPO4+ H2 (5)
Tertiary ferrous phosphate (insoluble): 3Fe2+ + 2PO43- = Fe3(PO4)2 (6)
A simplified schematic of the phosphating process is shown in Figure 5.
Figure 2.5: Schematic of the phosphating process.
The existing ions are hydrogen anions shown in Figure 5 and the resultant coating is
iron-phosphate as indicated by the chemical reactions. The treatment is carried out by
stirring the iron powder continuously in the solution bath in order to expose each
particle [13].
During the process a very small amount of metal is removed from the surface and then
re-deposited in an irregular fashion. However, this etching also has the side effect of
permanently removing some of the metal, since not all of it is re-deposited. The key
parameters that must be controlled to achieve a quality phosphate coating are
concentration, temperature, pH, and dwell time. These parameters also affect the
coating thickness which in turn affects the degree of crystallinity of the coating.
Thinner coatings show an amorphous structure whereas thicker ones tend to give
more crystal structure [14].
During the formation of ferrous phosphate the solution suffers a depletion of the free
acid concentration and as a result the pH in metal/solution interface rises. As indicated
already above, such rise of the pH affects the hydrolytic equilibrium of soluble primary
phosphates and insoluble tertiary phosphates in the solution. Phosphoric acid should
always be added in the solution in order to keep the bath stable and make it effective
for deposition of the phosphate at the microcathodic sites
9
2.3.1 Kinetics of the phosphating process
Several studies of the phosphating kinetics have shown that the formation of the
coating follows four stages, namely:
a) Induction stage
b) Initiation of film growth
c) Exponential main growth stage
d) Linear growth increase
In the induction stage, the surface oxide layer is removed. The initiation happens when
the first nuclei start to form and thus rate of nucleation increases fast. This also
depends on the phosphate bath conditions, the oxidizing agents and the pretreatment
process that the surface has undertaken [22]. Most important part of growth is the
exponential stage.
By using accelerators in the phosphating bath reduction of the induction period and
extension of the linear stage is achieved. It has been reported that the rate of
phosphating is defined by the diffusion of Fe 2+ from the substrate to the
coating/solution interface [15]. The rate of phosphating reaction is a function of
microanodes on the surface:
(7)
where dt is the change in time, FA the surface area of anodes in micro cells and K the
reaction rate constant. The latter reflects the influence of other factors on the process
such as temperature, surface conditions, etc. which can additionally vary [15]. This was
shown elsewere for carbonyl iron powder expressing the variation of the rate constant
according to:
(8)
in which E = 4745 cal/mol is the activation energy of the phosphatizing reaction and k =
143.4 the rate constant coefficient [22].
2.4 X-Ray photoelectron spectroscopy (XPS)
X-Ray Photoelectron Spectroscopy (XPS) is a versatile, well established and widely

applicable spectroscopic technique for studying surfaces [7]. It can be used to acquire
chemical composition and state of the elements present in a material surface, as well
as the surface layer thickness when combined with sputtering operations through ion
gun technique. Its basic principle involves the ejection of a photoelectron from a core
level inside an atom by an X-ray photon of specific energy (Fig 2.6)[8].
10
Figure 2.6: Schematic of the XPS photoemission of a 1s electron.
The instrument is measuring the kinetic energy (EK) of these photoelectrons which
depends on the energy of the X-ray photon (hv). The property that identifies the
parent element and the atomic energy level is the binding energy (EB) of the
photoelectron which is calculated from the above energies using the formula below:
(9)
where is the work function of the spectrometer and is characteristic of its

experimental configuration [9].The main components of an XPS system are the X-ray
source, the stage, the lens system, the analyzer and the detector. These components
are enclosed in an ultra-high vacuum chamber. A schematic of the instrument is shown
in Figure 6.
Figure 2.7: Schematic of the XPS instrument.
The requirement for the vacuum arises for two reasons: firstly electrons that travel
towards the analyzer should encounter as few gas molecules as possible otherwise
they will be scattered and will not reach the detector, while secondly the technique
itself is sensitive to surface contamination. The electrons are focused through a
hemispherical analyzer of a constant voltage by an electrostatic lens. The energy
discrimination of the electrons is done by controlling the pass energy and sweeping the
potential in the lens [10].
11
In the above configuration an argon ion gun can also be used if depth profiling is
needed. This involves the bombardment of the sample with heavy ions. The most
common gas that is used is argon. After ion etching the sample with the Ar+ ions in
specific intervals a small eroded area is created. By analyzing the surface after every
interval and plotting the intensity versus the etching depth the thickness of oxides or
coatings can be determined.
A typical XPS spectrum consists of a set of sharp peaks and a background, both
emerging from the elements present on the surface of the sample. The axes depict
binding energies versus number of electron counts which represents the intensity of a
given element in the sample. The peaks originate from core or Auger electrons that
escape the material without energy loss whereas the background comes from
electrons that have suffered inelastic scattering and thus energy loss. A typical
spectrum appears in Figure 2.8. The notation that is used in XPS uses the electron
orbitals nlj, i.e. 1s1/2, 2s1/2, 2p1/2, 2p3/2, etc. Different electron spin is depicted with a
different peak in the spectrum.
Figure 2.8: XPS survey spectrum of Fe-powder in the as received state.
2.5 Scanning electron microscopy (SEM) and energy dispersive X-ray

spectroscopy (EDX)
Scanning electron microscopy (SEM) is one of the most commonly used microscopic
techniques when it comes to materials characterization. The main advantage is its high
resolution and depth of focus that provide very detailed topographic contrast. The
basic principle of this method is bombarding the sample with a beam of accelerated
electrons and detecting the various signals that are emitted from the sample area that
is irradiated. The source of electrons is usually a bent tungsten filament which has to
be heated in order to emit electrons Using high voltage of up to 30 keV electrons are
accelerated and directed towards the sample. Magnetic lenses and apertures are also
used for controling the diameter of the beam and to focus it over the sample. The
12
beam scans the surface under examination in a raster scan manner. The instrument
operates under high vacuum conditions so interactions between electrons and gas
molecules are avoided. A schematic of a SEM instrument is shown in Figure 2.9.
Figure 2.9: Schematic of SEM instrument.
There are many different ways that the sample can interact with the electron beam
depending on the penetrating depth (fig 2.10). Different kinds of detectors help collect
the various sample responses and through computer software the respective image on
the screen is provided. Advanced microscopes are also equipped with a Field Emission
Gun (FEG) that yields high quality imaging at low acceleration voltages combined with
higher resolution and brightness.
For normal imaging purposes secondary (SE) or backscattered electrons (BSE) are most
commonly used. Secondary electrons offer a) higher resolution b) higher signal-to-
noise ratio which gives better quality images with good topographic contrast, while the
backscatter electrons provide enhanced elemental contrast. An SEM can also be
equipped with special inlens detector for improved topographical contrast at close
working distances by collecting secondary electron from small angles close to the
beam column [16].
13
Figure 2.10: Interaction volume of sample for various types of signals.
Another possibility within the SEM is to use X-ray spectroscopy. High energy electrons
from the beam have the ability to knock out a core electron from the sample atoms
thus leaving a vacancy to be filled by higher energy electrons. The transition from a
higher energy state to a core level happens with the simultaneous emission of an X-ray
which is characteristic of the atom and the atomic transition. Using the X-ray
spectroscope that is attached to the SEM one can detect such X-rays and the result is
an X-ray spectrum of the area investigated (Fig 11) [16].
Figure 2.11: EDX spectrum of Somaloy 500.
2.6 X-ray diffraction (XRD)
The X-ray diffraction is an analytical technique that helps us to understand how the
atoms or molecules are structured within a crystal. In this technique an X-ray beam is
targeted towards the sample and is diffracted by the atoms into various specific
directions. A detector is measuring the angles and intensities of these beams and the
output is a 3D representation of the crystal electron density. In this manner,
information such as unit cell dimensions, crystalline order, phase identification,
14
residual stresses in the material, among other types of analyses can be performed. In
an X-ray diffraction measurement, the material investigated is mounted on the stage
of the instrument and gradually rotated while the X-ray source is also moving at
specified angle intervals. X-rays are scattered elastically inside the material and while
most of them cancel one another as they add destructively some add constructively
following Braggs law:
2dsinθ=nλ (10)
where d is the spacing between diffracting planes, θ is the incident angle, n is any
integer and λ is the wavelength of the beam [23].
The diffracted x-rays produce a pattern and since most materials have unique
structure, compounds can be identified using already known patterns from available
databases. A typical XRD patterns is shown below:
Figure 2.12: XRD pattern with the database fitting lines present.
A powder X-ray diffractometer consists of an X-ray source (usually an X-ray tube), a

sample stage, a detector and a way to vary the incident angle. The X-ray is focused on
the sample at some angle θ, while the detector opposite the source reads the intensity
of the X-ray it receives at 2θ away from the source path. The incident angle is then
increased over time at a desired step interval while the detector angle always remains
2θ above the source path (Figure 2.13).
15
Figure 2.13: Schematic of a difractometer.
2.7 Thermogravimetry (TG)
Thermogravimetric analysis is the technique that measures the change in mass of a

sample as a function of temperature in a predefined temperature program.
Temperature is then first slowly increased then followed by an isothermal plateau for a
specified time and then a faster cooling down of the sample is performed. The weight
of the sample is measured before the procedure and compared to the weight change
that is happened during the temperature increase. From this data a weight loss curve
is generated. The instrument consists of a sample holder, a high resolution balance and
the furnace all inside a vacuum chamber with possibility to also have a gas
atmosphere. [26]
16
Chapter 3
Materials and Experimental procedures
Materials under examination
In this thesis three types of powder were studied and they were all provided by
Höganas AB, Sweden:
 Commercially available ABC100.30 powder, produced by water atomization
 High purity Fe powder, produced by a pilot gas atomizer
 Commercially available Somaloy 500® soft magnetic composite powder
Additionally, an iron foil, 1mm in thickness, of 99,5% purity and chemical composition
shown in Table 1 produced by Goodfellow Cambridge Ltd. was used for standard
calibration measurements after fine polishing to 1 μm. The samples prepared from the
foil were 0.85mm x 0.85mm in dimensions and they were subsequently annealed at a
maximum temperature of 500 0C for 15 min with a rate of 100C/min in argon 50% /
nitrogen 50% atmosphere. Furthermore, additional samples were also treated under
the same coating process as the soft magnetic composite powder grade.
Table 1: Fe foil composition
Element Fe Mn Si C P S
Content (wt%) 99.5 0.3 0.1 <0.08 <0.4 <0.05
To accurately perform curve fitting operations on the samples under investigation and
assess the effect of the sputtering operations on the compounds present on the
surfaces, measurements were conducted for chemical standards. These included iron
oxides Fe2O3 and Fe3O4 (Sigma-Aldrich) along with iron (III) phosphate dihydrate
(FePO42H2O) and in dehydrated (FePO4) state after heat treatment, all in powder form.
Sieving
To investigate the effect of particle morphology and geometry, all three types of
powder were sieved using the same sieve sizes which were chosen according to their
particle size distributions specifications as received from the manufacturer. The sample
size for every sieving process was 25g and was conducted with the octagon 2000
Digital Sieve Shaker from Endecotts using the 106µm and 53µm aperture size sieve
grids also from Endecotts. Because of the powder fine nature, each sieve operation
lasted 15min under the alternate mode where the apparatus first shakes and then
stops the sieve to leave the particles some time to settle through the meshes. After
17
sieving every powder grade, the samples were categorized in three particle size
fractions:
 106µm and higher
 Between 106 µm and 53 µm
 53 µm and below
Equipment and powder preparation
Measurements were done for each of the three kinds of powder in their “as-received”
state as well for each of the powder size fractions after sieving.
Surface analysis of powder was done at Chalmers University of Technology with X-ray
photoelectron spectroscopy (XPS) using the PHI 5500 instrument.
The XPS samples were prepared by using double sided adhesive conductive carbon
tape as this method allows the particles to stick on the tape without deformation and
alterations of the microstructure.
Along with the XPS measurements, ion etching was also performed on every sample to
get the composition depth profile of each powder. Based on this method it was
possible to estimate the oxide and phosphate thicknesses according to a theoretical
model described earlier by Nyborg et al. [18].. The PHI MULTIPACK software (PERKIN
ELMER, Eden Prairie, Minnesota, USA) was used for the XPS curve fitting analysis.
To characterize the morphology of the metal particles a FEG SEM LEO Gemini 1550
(CARL ZEISS - LEO electron microscopy, GmbH, Germany) was used while chemical
micro analysis on the powder surfaces was conducted using EDX spectroscopy (INCA
Energy) (Oxford Instruments, High Wycombe, England).
For the SEM analyses the samples were prepared by placing small amount of powder
between two soft aluminum plates and pressing at low pressures of 1-2 MPa to
“mechanically” interlock the particles on the plate surfaces. The samples were then
mechanically mounted on a specific holder providing good conductivity for the sample
and then analyzed.
For the XRD analyses the instrument from “BRUKER ANALYTICAL X-RAY SYSTEMS
(AXS)” was used. The source was Cr Ka at 2.29Å wavelength. Scans were performed
from 20o to 120o with 0.05o step and 5 sec/step
The TG instrument used in the present work is a simultaneous TG/DTA/DSC thermal

analyzer STA 449 F1 Jupiter equipped with a high resolution system both in weight and
temperature measurement.
Iron oxides – iron phosphate

The oxides (Fe2O3, Fe3O4) that were used as standards were also mounted on carbon
tape for the XPS analysis whereas the iron phosphate standard (FePO4) was compacted
and placed in the XPS as a compact using a specific holder. For two of these materials
(Fe2O3 and FePO4) electron neutralizer had to be used due to their insulating nature
charging effects.
Samples were also prepared from all the chemical standards for XRD measurements in
order to define their crystalline state and purity.
18
Chapter 4
Results
4.1 Analysis of the iron oxide standards
 Iron(III) oxide (Fe2O3)

The XPS spectrum of the iron (III) oxide gives a clear oxidic peak and the Fe3+ satellite
peak is also visible, see Figure 4.1.
Figure 4.1: XPS high resolution spectrum of the iron Fe2p3 peaks for Fe2O3 at 0nm.
By using curve fitting the iron signal is analyzed to its three main contributions. As
shown in Figure 4.2 the peak at 712.8 eV corresponds to the hydroxide compounds on
the surface, the peak at 710.6 eV is assigned to iron in Fe3+ state while the peak at
709.5 eV is for iron in Fe2+ state. The background removal, method used in this thesis
work was “iterated Shirley”. Charging effects were observed in this sample and
electron flood gun for charge neutralization was used. The adventitious hydrocarbon
with binding energy at 285 eV was also used as reference value for additional software
peak position correction. The oxygen signal shows a clear oxide peak at 529.7 eV and
OH- also present (Figure 4.2) [27].
19
Figure 4.2: XPS analysis of chemical state of iron and oxygen peakswith the help of curve fitting in as-received
state of the Fe2O3.
The binding energy values for the oxidized states of iron are used for the curve fitting
operations on all the samples analyzed in the present study. After excessive sputtering
the metallic state of iron appears (Fig 4.3).
Figure 4.3: Appearance of the metallic state of iron in XPS analysis after excessive ion sputtering at Fe2O3
standard.
To accurately determine the chemical standard purity and composition, XRD technique
was implemented. The findings are shown in Figure 4.4. It is clear that the oxide is
crystalline and the semi-quantitative analysis with DIFFRAC.EVA v.3.0 from BRUKER showed
20
that for the case of Fe2O3 there is approximately 92.5% of rhombohedral hematite and 7.5%
of orthorhombic magnetite.
Figure 4.4: XRD spectrum of Fe2O3 with database references of the known phases.
 Iron(II,III) oxide (Fe3O4)

The second type of iron oxide standard(Fe3O4) was also examined with XPS. This
oxide can also be formulated as FeO ∙ Fe2O3 and it contains both Fe2+ and Fe3+ ions in
proportions 1:2. The XPS high resolution spectrum show a lower contribution of the
Fe3+ satellite peak since Fe2+ is also present for this oxide (Fig 4.5).
Figure 4.5: XPS high resolution spectrum of the iron Fe2p3 peaks for as-received Fe3O4.
21
By curve fitting the binding energies of these cations are confirmed as shown in Figure
4.6 below.
Figure 4.6: Analysis of chemical state of iron and oxygen with the help of curve fitting in as-received state of the
Fe3O4.
The binding energy of Fe3+ is 710.9 eV and that of Fe2+ is 709.8 eV with the expected
ratio 1:2 between peak areas for these two contributions. The oxygen peak is at 530
eV. No charging was observed in this oxide and the values are very close in accordance
to the previous examined oxide, verifying that the charge correction on Fe 2O3 was
correct [28].
XRD was also performed on this sample for phase determination (fig. 4.7). This oxide
has also crystalline structure as shown from the distinct nature of the Fe3O4 peaks on
the spectrum with only this phase present.
Figure 4.7: XRD spectrum of Fe2O3 with database references of the known phases.
22
 Iron foil standard
XPS analysis was performed on the iron foil in order to obtain a clear iron metal peak
and use its characteristics for curve fitting of signals from the iron powder. A low
resolution survey spectrum of this sample showing an overview of the elements
present on its surface is depicted in Figure 4.8.
As-rec
20 nm
Figure 4.8: XPS survey spectra for the iron foil in as received and 20nm etched state.
As the foil was first polished and then annealed the oxide formed on top was expected
to be very thin in thickness. Some impurity elements already known from the initial
composition were also present on the surface. The etching steps and correspondent
apparent surface composition are in Table 2.
Table 2: Element content (at%) in the surface and ion-etched state of the as-polished iron foil
Content[at%]/Element Fe O C Ca Si Mn
As received 23.3 53.8 18.4 0.6 1.6 1.1
Etched 1nm 44.2 54.1 - 0.43 0.6 1
2 nm 63.2 35.7 - 0.2 0.2 0.8
3 nm 85 14.4 - - 0.3 0.4
5 nm 91.9 7.6 - - 0.3 0.2
10 nm 93.4 6.3 - - - 0.3
20 nm 93.5 6.3 - - - 0.2
The carbon immediately disappears after the first etching step clearly indicating that
its signal originates from contamination on the top layer, while silicon disappears after
1nm of sputtering. The manganese is present in oxide form within the material and
exhibits a slower decrease with etch depth. The apparent concentration of iron almost
reaches 94% while oxygen drops to 6.3% at 20nm etch depth. Figure 4.9 shows the
progress of the elements relevant concentration.
23
Figure 4.9: The element content over etching depth for the iron foil in the as-polished state.
The metallic peak starts to appear after the removal of the surface contamination layer
which eliminates the carbon content and is very clear at 20nm etch depth. Figure 4.10
shows the high resolution XPS spectra for the iron signal at selected etch depths.
20nm
2nm
As-polished
Figure 4.10: XPS spectra in the as-polished state and etching depths of 2nm and 20nm.
24
By using this metallic peak at 20nm all the standards needed for the curve fitting are
acquired. The iron peak has an asymmetric shape and is positioned at 706.7 eV with
full width at half maximum (FWHM) of 1.0eV, showing a very well defined peak (Figure
4.11).
Figure 4.11: Analysis of chemical state of iron with the help of curve fitting in 20nm etching depth.
The oxide thickness for a flat specimen can be calculated using the formula below:
[ ]
where is the attenuation lenght of the iron oxide (1.5nm , [17] ) , θ is the angle
between the normal to the surface and the electron analyser (45o for a flat surface for
the instrument used) , the metal and oxide intensities given from
measurements on the chemical standards and the metal and oxide contribution
of the iron foil at 0nm. For this foil the oxide thickness calculated with this method was
2.5nm, cf. depth profile in Fig. 4.12, which indicates a normalized Fe-metal intensity at
the corresponding etch depth of ~75%
Figure 4.12: Relative intensities of Fe and O signals recorded from as-received Fe foil vs etch depth.
25
4.2 Analysis of the ABC100.30 powder
XPS Analysis of the as-received powder
In the XPS survey spectrum there are three main characteristic peaks. These peaks
come from iron, oxygen and carbon (Figure 4.13).
As-rec
20 nm
Figure 4.13: XPS survey spectrums for the iron foil in as received and 20nm etched state.
Presence of carbon is mostly due to contamination on the surface and after a couple of
etching steps (2nm) the peak that remains is a result of the carbon tape onto which the
powder is mounted (not of interest in the current study). Impurity content (silicon,
manganese) appears to be very low. The apparent composition of this powder in the
as received state is shown in Table 3.
Table 3: Element content (at%) in the surface of the as received and ion etched states of the
ABC100.30 powder
Content[at%]/Element Fe O C N Si Mn
As received 14.9 47.3 33 1.3 1.4 2.2
Etched 1nm 34.3 59.8 3.2 - - 2.6
2 nm 36.8 57.4 3.6 - - 2.2
3 nm 40.8 56 0.6 - - 2.5
4 nm 43.7 49.7 1.9 - - 2.5
5 nm 50.4 43.3 4.1 - - 2.2
6 nm 60.7 37.4 - - - 1.8
7 nm 67.8 29.7 - - - 2.5
8 nm 74.4 23.8 - - - 1.8
10 nm 81.4 16.8 - - - 1.9
20 nm 86.9 9.4 - - - 1.8
30 nm 87.8 6 - - - 2.4
26
Figure 4.14 shows the progress of the elements content as a function of the etch
depth.
Figure 4.14: The element content over etching depth for the ABC100.30 powder in the as-received state.
As mentioned above, after 1nm of ion etching most of the surface contamination is
removed as shown by the decrease in carbon content from 33% to 4%. The iron
content is gradually increasing from 14.9% at the beginning to 87.8% at 30nm etch
depth. Meanwhile, oxygen is decreasing from 47.3% to just 6%. Silicon and manganese
impurity content is very low and thus negligible.
By performing high resolution scans in the XPS on the un-sieved sample, the following
spectra are obtained at selected etch depths as shown in Figure 4.15.
30nm
5nm
As-received
Figure 4.15 XPS spectra in as-received state and etching depths of 5nm and 30nm.
27
The metallic peak is not present on the as received state as it is covered by
contamination and oxide layers. However, it starts to appear after a few etch steps and
becomes clearly obvious after etching 30nm. The binding energy of metallic iron then
is 706.5eV and the oxide which is the Fe2O3 is at 710.5 eV. These energies are obtained
after curve fitting which allows us to distinguish the various chemical states from one
another. An example of curve fitting for the iron peak is shown in Figure 4.16.
Figure 4.16 Analysis of chemical state of iron with the help of curve fitting in the as received state (0nm).
By using curve fitting the contribution of these chemical states, oxidic or metallic, can
be obtained by measuring the area under each curve which depends on the amount of
atoms of the specific chemical state. Due to the destructive nature of the ion etching
process though, careful analysis must take place in order not to misinterpret the
results. For this reason information from the measurements on the standards in
respect to the effect of the ion etching on the compounds of interest is implemented.
Figure 4.17: Estimation of the surface iron oxide layer using the normalized Fe-metalic(Fe2p peak) and
the oxygen (O1s peak) over etch depth.
28
The method used in this thesis to obtain the surface layer thickness of a non-flat
sample such as metal powder uses the iron normalized intensity curve. It has been
shown [17] that the oxide-metal interface can be estimated by taking into account the
relative intensity of the metallic signal versus the etch depth. For surface layers for
which their ratio of thickness (tox) over the attenuation length (λox) of the
photoelectrons through them is 20 ≥ tox/λox≥ 3, the interface can be located in the
range of 60 to 72% of the relative intensity with 65% being a good estimation value for
an unknown layer. When the tox/λox is in the region 4 ≥ tox/λox≥ 0 the estimation is done
at 85% of the iron metallic intensity. This intensity fraction is used for the case of
annealed gas atomized powder.
The relative intensity of iron and oxygen is plotted in Figure 4.17 and from that the
oxide thickness can be estimated with the method described above which for this case
is 6.6nm. Overall, it is shown that there is very pure iron powder, which provides a
good base for evaluation and comparison with the Somaloy 500 powder later
examined.
XPS Analysis of the powder fractions
After sieving, the as-received powder was divided to three size fractions and each one
them was examined in the XPS. The element concentration is shown in the Tables 4 to
6 below for each fraction.
Table 4: Element content (at%) in the surface of the as received and ion etched state of the
+106µm fraction of the ABC100.30 powder
As received 13.2 46.5 36.7 0.3 2.2 1
Etched 1nm 23.1 51.7 23.6 - 0.7 1
2 nm 27.3 51.1 19.6 - 1.1 0.9
3 nm 28.6 49 20 - 1 1.4
4 nm 31.3 46.6 19.1 - 2 1
5 nm 34.9 42 19.3 - 2.8 1
6 nm 40.6 36.6 20 - 1.9 0.9
7 nm 44.9 31 21.2 - 2 1
8 nm 50.2 27.5 21.2 - - 1.1
10 nm 55.4 22 21.5 - - 1.1
20 nm 61.5 15.8 21.9 - - 0.9
30 nm 64.7 12.7 22.3 - - 0.2
Table 5: Element content (at%) in the surface of the as received and ion etched state of the -
106/+53µm fraction of the ABC100.30 powder
As received 12.2 43.3 40.5 0.7 2.4 0.9
Etched 1nm 23.6 51.1 23.1 - 1.1 1.1
2 nm 27.2 51 19.7 - 1.3 0.9
3 nm 29.5 50.1 19.1 - 0.6 0.7
29
4 nm 32 48 18.5 - 0.8 0.7
5 nm 36 44 18 - 1.2 0.9
6 nm 40.9 38.1 18.9 - 1.2 0.8
7 nm 46.3 32.6 19.2 - 1 0.9
8 nm 49.3 27.7 21.3 - 1.2 0.5
10 nm 56.1 21.7 21.8 - - 0.4
20 nm 65.4 14.3 20 - - 0.3
30 nm 64.3 13 22.5 - - 0.1
53µm fraction of the ABC100.30 powder
As received 14.2 48.6 33 1.2 1.8 1.2
Etched 1nm 22 50.7 23.8 - 2.5 1.1
2 nm 27.2 50.3 19.6 - 2.1 0.8
3 nm 29.3 49.6 18.5 - 1.6 0.9
4 nm 31.9 48.1 17 - 2.1 0.9
5 nm 35.7 43.5 17.4 - 2.6 0.9
6 nm 40.9 38.2 18.6 - 1.7 0.7
7 nm 45.7 32.6 19.2 - 1.9 0.7
8 nm 50.2 29.5 19.6 - - 0.8
10 nm 55.9 22.4 21 - - 0.7
20 nm 66.1 15.2 18.1 - - 0.7
30 nm 66.7 13.4 19.7 - - 0.3
(a) (b)
30
(c)
Figure 4.18: The elements content over etch depth for the ABC100.30 powder in the various fractions (a) +106µm
(b)-106/+53µm(c) -53µm.
From the tables and the above graphs it is obvious that the fractions show a similar
behavior in the elemental concentration profiles with the as-received unsieved
powder. The oxide thickness estimations are shown in Figure 4.19.
(a) (b)
(c)
Figure 4.19: Estimation of the surface iron oxide layer using the normalized Fe-metallic (Fe2p peak) and the oxygen (O1s peak)
over etch depth for the various fractions of the ABC100.30 powder (a) +106µm (b)106/+53µm(c) -53µm.
31
The oxide thickness is 6.3nm for all fractions. The results here show that the fractions
give a lower contribution in oxide thickness than the unsieved powder which was
6.6nm.
Figure 4.20 shows a high resolution scan at 5nm of etch depth, which is near the
oxide/metal interfacial region of the unsieved powder and all three fractions. The
intensity is normalized so the scans can be compared without the effect of the
analyzing area. The three fractions have higher metallic peaks than the unsieved
powder which means lower oxide contribution. Hence, analyzing a wider size
distribution means some dispersion effect regarding the thickness estimate.
Figure 4.20: XPS high resolution surveys of the ABC100.30 fractions at 5nm etch depth.
SEM and EDX of the as-received powder
In figure 4.21 the Fe particles are displayed. The irregular shape is typical for water
atomized powders. The various sizes are also visible in (a) and they range between -
53µm to over 200µm (b).
(a) (b)
32
(c) (d)
Figure 4.21: Overview of the ABC100.30 as received water atomized powder.
The surface of this powder is also visible on (c) and (d) and as expected is very clear
with only some very small particulates which have been characterized in previous
studies as particles with different composition [19].
Figure 4.22 shows the appearance of these particulates at higher magnification. The
surface appearing on this figure is typical for this type of powder and is observed in a
similar manner throughout the powder samples.
Figure 4.22: Appearance of the particulate features on the powder surface.
EDX analysis was performed to investigate the chemical composition of these

particulates. This type of chemical microanalysis is regarded only quantitatively due to
the large interaction volume and the geometry of the surface under examination.
33
Spectrum[at%] C O Al Si Cr Mn Fe
Spectrum 1 14.4 19 0.3 1.2 1.4 3.5 60.3
Spectrum 2 27.1 7.1 - - - - 65.8
Spectrum 3 15.4 2.5 - - - - 82.1
Figure 4.23: SEM image of the powder surface and results of the EDX measurements.
Figure 4.23 shows the EDX analysis performed in various points on the powder surface.
The carbon signal appears to be large as a result of the surface contamination.
34
Spectrum 1 includes some of the small particulates in the interaction volume and it is
supposed that these are mostly impurity oxides containing silicon and/or manganese.
Spectrum 2 aims at a larger particulate and from the analysis it appears to be an iron
oxide whereas spectrum three aim at an area where there are no particulates and the
expected iron oxide layer is then detected.
SEM and EDX of the powder fractions
Figure 4.24 shows SEM images at low magnifications of the three sieved fractions at
specific particles size ranges as mentioned earlier, in order to show an overview of the
samples and to evaluate the quality of the sieving operation.
(a) (b)
20μm WD=8.0mm EHT=10kV Signal=InLens Mag=500X 100μm WD=8.0mm EHT=10kV Signal=InLens Mag=400KX
(c)
20μm WD=8.0mm EHT=10kV Signal=InLens Mag=1.00KX
Figure 4.24: SEM overview of the of the ABC100.30 powder (a) +106µm (b)106/+53µm(c) -53µm.
35
 +106μm
(a) (b)
(c) (d)
Figure 4.25: Overview of the ABC100.30 +106μm fraction.
 -106/+53μm
(a) (b)
36
(c) (d)
Figure 4.26: Overview of the ABC100.30 -106/+53μm fraction.
 -53μm
(a) (b)
(c) (d)
Figure 4.27 Overview of the ABC100.30 -53μm fraction.
37
Figures 4.25, 4.26 4.27 show an overview of the ABC100.30 fractions. A smooth surface
is shown with only the same small particulates appearing as before. The EDX analysis
performed reveals that the particulates are silicon and manganese based oxides but
most of the powder surface is ok and smooth. On those smooth surface regions the
EDX chemical microanalysis shows concentration of iron, carbon as contaminant and
some oxygen from surface oxides (Fig 4.28, 4.29 4.30).
Spectrum[at%] C O S Mn Fe
Spectrum 1 13.8 5 - 0.5 80.7
Spectrum 2 25.6 2.7 - - 71.7
Spectrum 3 20.3 4.5 19 18.9 37.4
38
Spectrum[at%] C O Mn Fe
Spectrum 1 21 6.4 0.4 72.2
Spectrum 2 22.4 3 74.7
39
Spectrum[at%] C O Mn Fe
Spectrum 1 26 10.1 0.4 63.5
Spectrum 2 30.2 2.8 - 67
40
4.3 Analysis of the gas atomized iron powder
 As received
XPS analysis of the powder fractions
The condition of the as received powder was not suitable for examination in terms of
particle size distribution, thus a high estimation error would be expected in all our
analyses that would give no significant information. Hence, only the sieved fractions
were examined. The composition of the grade is similar to the ABC100.30 with iron,
oxygen as the main elements and some impurity elements also present. Carbon exists
as surface contaminant (eg. hydrocarbon) at first and then the signal remains constant
due to the carbon tape that the powder is mounted on. The progress of the elemental
apparent atomic concentration and their plots appear below.
+106µm fraction of the gas atomized powder
Content[at%]/Element Fe O C N Si Mn Na
As received 11 40.7 43.6 1.3 2.4 0.3 0.6
Etched 1nm 29.8 57.2 9.9 - 2.4 0.4 0.4
2 nm 30.1 54.3 15.6 - - - -
5 nm 34.3 53 12.7 - - - -
10 nm 42.7 44.8 12.6 - - - -
15 nm 48.9 38.5 12.6 - - - -
20 nm 52.8 34.8 12.4 - - - -
25 nm 55.1 32.6 12.2 - - - -
30 nm 55.7 31 13.3 - - - -
40 nm 60.5 28.8 10.7 - - - -
50 nm 62.2 26.7 11.1 - - - -
70 nm 65.8 22.2 12 - - - -
100 nm 69.9 18.4 11.7 - - - -
106/+53µm fraction of the gas atomized powder
As received 12.8 44.1 40.3 0.6 1.4 0.2 0.8
Etched 1nm 25 55.5 17.1 0.2 2 - 0.2
5 nm 32.5 54.7 10.8 - 1.8 - 0.2
10 nm 43 46.4 8.6 - 1.3 - 0.1
15 nm 50.2 40.1 8.4 - 1.8 - -
20 nm 54.9 35.7 7.6 - 0.5 - -
25 nm 58.5 34 7 - 1 - -
30 nm 60.9 31.5 6.6 - - - -
40 nm 64.6 29.6 5.9 - - - -
50 nm 67 26.1 6.9 - - - -
70 nm 71.4 22.4 6.2 - - - -
100 nm 75.6 18.1 6.4 - - - -
41
53µm fraction of the gas atomized powder
As received 13.4 42.5 38 0.3 1.9 0 0.9
Etched 1nm 26.2 54.7 17.4 - 1.2 0.3 0.3
2 nm 31.4 54.5 11.9 - 1.4 0.3 0.5
3 nm 32.3 53.6 10.9 - 2.4 0.3 0.5
4 nm 32.7 53.3 11.2 - 2.1 0.3 0.4
5 nm 33.8 52 11 - 2.8 0.3 0.2
6 nm 35.8 51.3 10.5 - 2.2 0.2 -
10 nm 41.5 46.1 10 - 2.2 0.2 -
15 nm 51 38.8 8.4 - 1.7 0.1 -
20 nm 57.1 32.5 8.8 - 1.6 - -
25 nm 62.5 28.6 7.5 - 1.5 - -
30 nm 62.8 25.5 10.3 - 1.3 - -
40 nm 71.6 21.6 6.8 - - - -
50 nm 73.8 18.4 7.8 - - - -
60 nm 74.9 17.1 8 - - - -
(a) (b)
(c)
Figure 4.31: The element content over etching depth for the gas atomized powder of the various fractions (a) +106µm
(b)106/+53µm(c) -53µm.
42
For the +106µm fraction iron goes up to almost 70% after extended ion etching
whereas in the other two cases it reaches to 75%. From the etching depth it appears
that this powder has thicker oxide than the ABC100.30 and more etch steps were
needed to get a clear iron signal.
Typical XPS spectra at different etch depths are shown in Figure 4.32. All three
fractions appear to have similar behavior.
100nm
5nm
As-received
Figure 4.32: XPS spectra in as-received state and etch depths of 5nm and 100nm for the middle fraction (+106/-
53µm) of the gas atomized powder. Other fractions exhibit similar behavior when etched.
Using curve fitting the iron signal contributions are determined. The iron peak with a
binding energy of 706.6 eV in all fractions is in very good agreement with the iron foil
standard along with the signals for the Fe3+ contribution and the Fe2+ contribution (Fig
4.33).
Figure 4.33: Analysis of chemical state of iron with the help of curve fitting for the 5nm etch depth. The metal
state is the small peak on the right side of the figures and the rest are peaks for the iron oxides.
43
Figure 4.34: XPS high resolution spectra of gas atomized of different as received fractions after 10nm of ion
etching.
Figure 4.34 shows a comparison between the three fractions. It is clear from the
figures above that the +106µm fraction appears to have a slight higher metallic peak
and a lower oxidic one so less oxide thickness is expected. Using the metallic peak
intensity the oxide thickness was calculated and is 12.3nm for the +106µm fraction,
14.7nm for the middle fraction and 14nm for the -53µm fraction (figure 4.35).
(a) (b)
44
(c)
Figure 4.35: Estimation of the surface iron oxide layer using the normalized Fe-metalic (Fe2p peak) and the oxygen (O1s
peak) over etching depth for the various fractions of the gas atomized powder (a) +106µm (b)106/+53µm(c) -53µm.
SEM and EDX

Below are shown overview SEM images of the three fractions with specific particles
size measured to help determine the quality of the sieving and size distribution of the
fractions. The spherical particles are typical for gas atomized powder.
(a) (b)
100μm WD=8.5mm EHT=15kV Signal=InLens Mag=150X 100μm WD=8.5mm EHT=10kV Signal=InLens Mag=300X
(c)
10μm WD=10mm EHT=10kV Signal=InLens Mag=800X
Figure 4.36: SEM overview of the of the as-received gas atomized powder (a) +106µm (b)106/+53µm(c) -53µm.
45
A first look at this powder revealed a few heavily oxidized particles like the one
showing in figure 4.37 at high magnification. The thick oxide layer appears to be
fractured and scattered with a clear surface appearing below.
Figure 4.37: Gas atomized powder particle in the as received state.
A more representative appearance of this powder is shown in the SEM images of the
various fractions below.
 +106μm
(a) (b)
46
(c) (d)
Figure 4.38: Overview of the gas atomized powder in as received state: +106μm fraction.
 -106/+53μm
(a) (b)
(c) (d)
Figure 4.39: Overview of the gas atomized powder in as received state:-106/+53μm fraction.
47
 -53μm
(a) (b)
(c) (d)
Figure 4.40: Overview of the gas atomized powder in as received state:-53μm fraction.
Figures 4.38, 4.39 and 4.40 show overview (a) and higher magnification image (d) of
the three size fractions of the gas atomized powder in as-received state. A smooth
surface is obvious for all fractions with some very small particulates also present on
the surface. Some of these were examined with EDX, but due to the large interaction
volume this was not possible for the really small particulate features.
48
Spectrum[at%] C O Na Si Fe
Spectrum 1 24.9 4.8 - - 70.3
EDX analysis performed on the surface reveals that is pure iron with some line
formations visible, possibly an oxide layer that is cracked (figure 4.41). Carbon is also
present due to surface contamination.
49
 Annealed in argon/nitrogen atmosphere
XPS analysis of the powder size fractions
After the gas atomized powder was annealed it was analyzed in the XPS. Below is the
progress of the element composition as a function of ion etching for the different
samples. It is clear that the oxide layer is reduced in thickness compared to before
annealing (cf. Tables 7-9). The impurity elements are also present.
+106µm fraction of the gas atomized powder
As received 18.2 50.5 28.1 0.1 2.3 0.4 0.4
Etched 1nm 31.3 50.3 15.3 - 2.7 0.3 -
2 nm 42 36.7 19.1 - 2 0.3 -
3 nm 49.7 28.5 19.8 - 1.8 0.2 -
5 nm 54.7 22.9 21.4 - 0.8 0.3 -
10 nm 60.6 20 19.3 - 0.1 0.1 -
20 nm 64.3 17.2 18.5 - - - -
106/+53µm fraction of the gas atomized powder
As received 15.3 51.7 28.4 0.2 3.9 0.2 0.3
Etched 1nm 25.9 50.5 19.5 - 3.9 0.2 -
2 nm 34 45.3 16.2 - 4.1 0.4 -
3 nm 44.1 37.3 15.5 - 2.9 0.2 -
5 nm 53.9 26.8 17.1 - 2 0.2 -
7 nm 57.8 22.8 17.3 - 2.1 0.1 -
10 nm 61.6 20.5 14.2 - 3.5 0.1 -
20 nm 69.9 17.6 12.6 - - - -
53µm fraction of the gas atomized powder
As received 15.5 44.9 34.9 0.4 3.2 0.6 0.5
Etched 1nm 29.9 49.7 16.5 - 3.3 0.4 0.2
2 nm 40.7 42.7 12.5 - 3.6 0.3 0.1
3 nm 51.1 31.4 14.5 - 2.6 0.4 -
4 nm 57.5 25.1 15.3 - 2.1 0.1 -
5 nm 58.7 22.5 16.01 - 2.6 0.1 -
10 nm 63.1 18.4 16.3 - 2 0.2 -
20 nm 73.7 14.6 11.8 - - - -
50
(a) (b)
(c)
Figure 4.42: The element content over etching depth for the annealed gas atomized powder in the various
fractions (a) +106µm (b)106/+53µm(c) -53µm.
Typical XPS spectra of etch depths are shown in Figure 4.43. All three fractions appear
to show similar behavior.
51
20nm
2nm
As-received
Figure 4.43: XPS spectra in as-received state and etch depths of 5nm and 100nm from the middle fraction (-
106/+53µm) of the gas atomized powder. Other fractions exhibit similar behavior.
Using curve fitting the iron signal contributions are determined. The iron peak appears
on the right side of the images in Figure 4.44 with a binding energy of 706.6 eV for all
fractions, same as the as-received powder fractions and in very good agreement with
the iron foil standard. The iron metal peak also appears clearly before ion etching (Fig
4.43-4.44).
Figure 4.44: Analysis of chemical state of iron with the help of curve fitting before ion etching. The metal state is
the peak on the right side of the figures and the rest are peaks for iron oxides.
52
Figure 4.45: XPS high resolution spectra of the Fe signal peak for gas atomized as annealed powder fractions at
1nm etching depth.
The oxide layer thickness was calculated and as indicated by the XPS spectra it is very
thin. For the +106µm fraction it is 2.7nm, for the -106/+57µm is 4.2nm and for the -
53µm is 2.9nm. The diagrams show the relative intensity of the metallic iron peak and
the oxide layer thickness calculation (Fig 4.46). It should however be noted that this
represents the thickness estimate related to the Fe-oxide layer. After ion etching to
>20nm there is still substantial presence of oxygen (see Tables 10-12). It may therefore
be assumed that part of the surface is also covered by thicker oxide products.
(a) (b)
53
(c)
Figur 4.46: Estimation of the surface iron oxide layer using the normalized Fe-metallic (Fe2p peak) and the oxygen (O1s peak)
over etching depth for the various fractions of the annealed gas atomized powder (a) +106µm (b)106/+53µm(c) -53µm.
SEM and EDX
Below are then images (Figs 4.47, 4.48, 4.49) of the various fractions of the gas
atomized powder in the as-annealed state. After the annealing, smooth, supposedly
clean surfaces are observed but there are also particulate features on the surface.
 +106μm
(a) (b)
54
(c) (d)
Figure 4.47: Overview of the annealed gas atomized powder +106μm fraction.
 -106/+53μm
(a) (b)
(c) (d)
Figure 4.48: Overview of the annealed gas atomized powder -106/+53μm fraction.
55
 -53μm
(a) (b)
(c) (d)
Figure 4.49: Overview of the annealed gas atomized powder -53μm fraction.
EDX analysis (Fig 4.50) showed that these particulates contain manganese, chromium
and magnesium together with oxygen (Spectrum 2). The rest of the surface appears to
show only iron. As indicated by XPS the oxide layer is thin and thus it cannot be
detected with EDX (Spectrum 1). In both cases, carbon is present due to surface
contamination.
56
Spectrum[at%] C O Mg Cr Mn Fe
Spectrum 1 55.7 - - - - 44.3
Spectrum 2 22.1 29.3 0.6 5.2 3 39.9
57
4.4 Analysis of the phosphate standards
 Iron(III) phosphate dihydrate (FePO42H2O)

The iron phosphate dihydrate was examined in the XPS and evaluated in a similar way
as the oxides. This time the interest was on the phosphorous peak and curve fitting
provided with the binding energy of the phosphate. The binding energy of
phosphorous was 133.8 eV that is very close to the values reported in literature [20]
and is shown in Figure 4.48. The iron compounds are also visible and as reported in
literature they are slightly shifted to higher binding energies compared to the iron
oxide standards [26].The Fe2+ is at 709.7 eV and the Fe3+ is at 712 eV. The specimen
was also charging so again adventitious carbon charge correction had to be used to
position the Fe peaks.
Figure 4.51: Analysis of chemical state of iron phosphate dihydrate with the help of curve fitting in as-received
state. Phosphorous left and iron right.
This phosphate powder was found to also contain sodium (NaPO 4) hence the small
peak on the right side of Figure 4.51(left). The iron phosphate dihydrate sample was
analyzed with XRD and this indicated that it was amorphous with no characteristic
peaks appearing in the spectrum (Fig. 4.52).
Figure 4.52: XRD spectrum of FePO42H2O.
58
 Iron(III) phosphate dehydrated (FePO4)
The iron phosphate dihydrate was put in the DSC for the dehydration process. The
mass percentage /temperature curve reveals that the sample lost 17.77% of its mass
(Fig 4.53) and the onset was observed at 116.2 oC. Assuming that the weight loss is due
to structural water in the compound the percentage that should have been observed is
19.28%. Hence, the dehydration process was almost complete with a very small
amount of bound water still remaining.
Figure 4.53: Dehydration DSC curve.
After the dehydration the compacted phosphate powder was put to the XPS. The curve
fitting applied to the phosphorous signal (Fig 4.54) gave the binding energy of the
phosphate at 133.7 eV which is even more close to the literature values as expected.
The oxygen peak is also shown with the binding energy for oxygen in the phosphate
being 531.6 eV and for O2- in oxide at 529.3 eV. Again the small sodium phosphate
contribution on the phosphorous signal is also present.
Figure 4.54: Analysis of chemical state of the dehydrated iron phosphate peaks with the help of curve fitting in as-
received state. Phosphorous left and oxygen right.
59
 Phosphated iron foil
XPS measurements was done on the phosphated iron foil with the purpose of
obtaining a clear phosphorous peak and use its characteristics for the curve fitting
necessary for the Somaloy 500 powder. A survey of this measurement is shown in
Figure 4.55.
20nm
As-received
Figure 4.55 XPS survey spectra for the phosphated iron foil in as-received and 20nm etched state.
Main elements that are investigated are iron, oxygen, phosphorous and carbon. The
etch steps and the compositions appear in Table 13 below.
Table 13: Element content (at%) in the surface and ion-etched state of the as-polished iron
foil
Content[at%]/Element Fe O C P
As received 8.5 47.5 43.6 0.5
Etched 1nm 31.2 63 3.4 2.4
2 nm 34.7 63 - 2.3
3 nm 38.7 59.4 - 1.9
5 nm 60.6 39.1 - 0.3
7 nm 82.4 13.7 - -
10 nm 90.2 6.4 - -
20 nm 88 6.3 - -
The carbon is immediately removed after the first etch step and disappears after the
second. The concentration of iron reaches 88% while oxygen drops to 6.3% after 20nm
of etching. Phosphorous is revealed after the removal of the surface contamination
and disappears after 5nm. Figure 4.56 shows the progress of the elements apparent
concentration with etch depth.
60
Figure 4.56: The element content over etch depth for the phosphated iron foil in the as-polished state.
The metallic peak of Fe starts to appear after the removal of the surface contamination
layer and is very clear at 20nm etch depth. Figure 4.57 shows the XPS spectra for
selected etch depths.
20nm
2nm
As-received
Figure 4.57: XPS spectra of phosphated iron foil before and after ion etching (2nm and 20nm).
After the removal of the surface contamination, curve fitting was used to define the
phosphorus peak position. It is at 133.8 eV as shown in Figure 4.58.
61
Figure 4.58: Analysis of chemical state for phosphated iron foil phosphorous peak with the help of curve fitting of
P2p XPS peak at 1nm etch depth.
The oxide or coating layer thickness for a flat specimen is approximately given by the
etch depth at which the coating layer intensity reaches 50% of its maximum value.
Considering this for the P intensity the thickness is estimated at 3.7 (Figure 4.59).
Figure 4.59: Estimation of the surface iron oxide layer using the normalized P peak over etching depth.
62
4.5 Analysis of the Somaloy 500 powder
XPS analysis of the as-received powder
In the XPS survey spectrum of the Somaloy powder there are four main characteristic
peaks. These are from iron, oxygen, carbon and phosphorous.
30nm
As-received
Figure 4.60: XPS survey spectrum for the phosphated iron foil in as-received and 30nm etched state.
The apparent surface composition of this powder in the as received state is shown in
Table 14 below.
Somaloy 500 powder
Content[at%]/Element Fe O C P N Si Mn
As received 4.6 56.6 24.2 11.7 1.6 0.9 0.5
Etched 1nm 10.9 58.4 15.7 13.3 - 1.2 0.5
2nm 13.2 57.7 15.6 12.2 - 0.8 0.6
5nm 18.5 56.5 15.1 9.3 - - 0.6
7nm 24.5 53.1 13.7 8.1 - - 0.7
10nm 32.6 45.6 14.7 6.5 - - 0.5
12nm 39.4 39.4 15.5 5.2 - - 0.6
15nm 44.8 33.1 17.2 4.2 - - 0.7
20nm 54 26.1 15.9 3.5 - - 0.5
30nm 59 19.9 19 2.2 - - -
50nm 64.9 14.4 17.7 - - - -
Carbon is present at all steps due to the carbon tape used for mounting the powder.
Iron increases with etch depth up to around 65% while oxygen and phosphorous
decrease with increasing etch depth. Phosphorous disappear after 30nm of ion etching
63
as the coating is etched away. The progress of the element concentration with etch
depth is shown in Figure 4.61.
Figure 4.61: The element content over etch depth for the Somaloy 500 powder in the as-received state.
The high resolution XPS spectra at Fe3p region in figure 4.62 show the intensity for
different etching depths.
50nm
5nm
As-received
Figure 4.62: XPS spectra in the as-received state and etch depths of 5nm and 50nm.
64
No metallic peak appears before ion etching. Curve fitting of Fe and P signals shows
the contributions before ion etching in Figure 4.63. The left peak component is the
hydroxide species on the surface, the middle one is Fe3+ and the right is the Fe2+ with
the binding energies being 712.8, 711.6 and 709.9 eV respectively. The phosphorous
peak in the P2p area is at 133.9eV in accordance with previous studies [19].
Figure 4.63: Analysis of chemical state of iron (left) and phosphorous (right) with the help of curve fitting in the as
received state(0nm).
Curve fitting was also applied for the oxygen peak and the progress of the two
contributions with etch depth is shown in Figure 4.64. The phosphate component
remains steady while the O2- increases with etch depth and then decreases again. The
binding energy of oxygen in phosphate is 531.6 eV and that of O2- in oxide is 530.2 eV
Figure 4.64: Analysis of chemical state of oxygen with the help of curve fitting at three etch depths.
65
The coating thickness was calculated using two approaches that are shown in figure
4.65.
Figure 4.65: Estimation of the phosphate coating using the normalized Fe-metallic (Fe2p peak), the oxygen (O1s
peak) and the phosphorous (P2p peak) over etch depth.
The first uses the relative metallic peak intensity as before; here 65% owing to greater
thickness, and gives 11.1nm (with the oxide thickness also included) and the second
uses the phosphorous peak intensity at 55% relative intensity (due to the fact that P
doesn’t go to zero) and it gives 5.9nm.
XPS analysis of the powder fractions
The surface coomposition of elements for the sieved fractions of the Somaloy 500
powder are shown in Tables 15-17 below.
+106µm fraction
As received 5 54.4 26.2 11.3 2.3 0.5 0.4
Etched 1nm 12.1 61.4 12.5 12.9 - 1 0.2
2nm 14.4 61.1 10.9 12.3 - 1 0.3
5nm 19.6 57.4 13.6 9.3 - - 0.1
7nm 25 55 11.3 8.5 - - 0.1
10nm 31.3 48.1 14.1 6.4 - - 0.2
12nm 37.4 42.3 14.3 5.8 - - 0.3
15nm 43.9 36 15.8 4.1 - - 0.2
20nm 53.7 27.6 14.8 3.8 - - 0.1
30nm 62.5 21.5 14.2 1.8 - - 0.1
50nm 66.9 17 16.1 0 - - 0.4
66
Table 16 Element content (at%) in the surface of the as received and ion etched state of the -
106/+53µm fraction
As received 4.8 56.2 24 12.1 1.7 0.8 0.5
Etched 1nm 10.4 59 14.8 14 - 1.2 0.6
2nm 12.8 58.8 12.8 13.1 - 1.9 0.6
5nm 17.8 57.7 12.9 10.9 - - 0.8
7nm 22.9 53.9 13.9 8.6 - - 0.7
10nm 29.3 47.4 15.3 7.3 - - 0.6
12nm 36.6 42.3 14.4 6.2 - - 0.6
15nm 41.4 35.6 17.4 4.9 - - 0.8
20nm 53.5 29 14.8 2.4 - - 0.4
30nm 59.6 21.4 16 2.7 - - 0.3
50nm 63.5 17.5 17.6 1.4 - - 0.5
53µm fraction
As received 4.3 52.6 27.6 11.7 2 1.2 0.5
Etched 1nm 11.2 61.2 11.5 14.8 - 0.8 0.6
2nm 13.7 60.1 10.8 13.5 - 1.3 0.8
5nm 19.1 58.1 11.9 10.3 - - 0.6
7nm 24.4 54.8 11.5 8.7 - - 0.7
10nm 32.1 46.8 13.4 7 - - 0.7
12nm 37.9 41 15.1 5.5 - - 0.6
15nm 44.5 33.9 16.7 4.5 - - 0.5
20nm 55.1 27 14.7 2.8 - - 0.4
30nm 59.3 19.9 18 2 - - 0.8
50nm 66.4 15.6 16.9 1.1 - - 0.5
(a) (b)
67
(c)
Figure 4.66: The element content over etch depth for the Somaloy 500 powder of the various fractions (a)
+106µm (b)106/+53µm(c) -53µm.
The above graphs indicate that the fractions show a similar behavior in the element
concentration with the as-received unsieved powder. Here 50% of the P intensity is
used. A small difference seems to appear in the phosphate coating thickness with the
as-received powder having lower value than the fractions.
(a) (b)
(c)
Figure 4.67: Estimation of the phosphate coating using the normalized Fe-metallic(Fe2p peak) ,the oxygen (O1s peak)
and the phosphorous (P2p peak)over etch depth for the various fractions (a) +106µm (b)106/+53µm(c) -53µm.
68
When using the iron metallic peak intensity the thickness starts from 11.5 nm, 11.3 nm
for the middle fraction and 11.0 nm for the small one. With the phosphorous intensity
these values are 7.8, 7.1 and 6.9 nm respectively following the same trend. Overall the
fractions give almost the same contribution in coating thickness as the unsieved
powder which was 11.1 nm using the iron peak intensity.
SEM and EDX of the as-received powder
The Somaloy 500 SEM images show a clean surface with some few particulates of
around 0,5μm size (fig 4.66). Irregular shape of the particles is visible as a water
atomized powder is used as base powder for the Somaloy 500.
(a) (b)
(c) (d)
Figure 4.68: Overview of the Somaloy powder.
The EDX examination showed that the particulates on the surface contain aluminum
and silicon. The rest of the analysis sites revealed the expected iron, oxygen
phosphorous (Fig 4.69). Carbon again is present on the surface as contamination
(hydrocarbon species).
69
Spectrum[at%] C O Al Si P Fe
Spectrum 1 6.5 7.7 - - - 85.8
Spectrum 2 4.2 53.3 1.6 10.6 1.5 29.8
Spectrum 3 6.8 14.5 - - 1.2 77.5
Spectrum 4 6.8 10 - -- 1.3 81.9
70
SEM and EDX of the powder fractions
Below appear the SEM images of the Somaloy 500 fractions. Similar surface
characteristics are apparent for all fractions. Smooth surface is obvious with only the
evaporation lines from the solidification process visible and some few particulates
present.
 +106μm
(a) (b)
(c) (d)
Figure 4.70: Overview of the Somaloy powder, +106μm fraction.
71
 -106/+53μm
(a) (b)
(c) (d)
Figure 4.71: Overview of the Somaloy powder, -106/+53μm fraction.
 -53μm
(a) (b)
72
(c) (d)
Figure 4.71 Overview of the Somaloy powder, -53μm fraction.
Another feature that appeared in specific sites of this powder is shown in figure 4.72. A
high concentration of flaky type particulates is visible.
Figure 4.72: SEM image of Somaloy 500 powder on a flaky particulates site.
EDX analysis on this showed particulates enriched in silicon, aluminum and manganese
as well as clear sites with iron, oxygen and phosphorous concentration (fig 4.74).the
signal in the spectra appear low due to the low accelerating voltage used but voltage is
enough to excite the X-rays from the elements expected.
73
Examination of the flaky particulates is shown on fig 4.73. Phosphorous signal is much
higher than areas around this spot so it is assumed that these particulates are part of
the phosphate coating.
Spectrum[at%] C O Mg Si P Ca Mn Fe
Spectrum 1 2.1 60.3 3.4 10.5 0.6 10 13.1 -
Spectrum 2 5 21 - - 1.7 - - 72.3
Spectrum 3 4.6 18.1 - - 3.7 - - 73.4
Spectrum 4 21.5 25.3 - - 5.1 - - 48.1
Spectrum 5 18.2 8.8 - - 1.8 - - 71.20
Figure 4.73 SEM image of the powder surface and results of the EDX measurements.
74
Spectrum[at%] C O Al Si P Mn Fe
Spectrum 1 7.8 71 1.6 - 1.5 - 18.1
Spectrum 2 33.4 5.8 - - - - 60.9
Spectrum 3 8 49.2 - 6.5 3.3 22.4 10.6
Spectrum 4 7 23 - - 4.5 - 65.7
75
Chapter 5
Discussion
5.1 Pure iron powders and standards
For a valid characterization of a material using XPS, good reference values for peak
positions are necessary. Such can be obtained from analyses of chemical standards
fabricated with well characterized and known techniques. For this purpose, the oxides
and phosphates were examined during this thesis work in order to establish the peak
positions and peak characteristics such as peak width and peak shape asymmetry of
the XPS signals by compounds of interest. Furthermore, the ion etching technique is
used in order to acquire compositional depth profiles, and by means of which the
surface layer thickness is also estimated. Here, the analysis of the chemical standards
also provides information on their behavior under etching. Additionally, from the flat
foil samples that were prepared accordingly to match the elemental composition for
given surfaces of interest it is also possible to acquire the precise etch rate for specific
compounds provided their thickness is controlled and pre-known.
The results from the analyses showed that the two types of iron oxides that were
examined are in agreement with reference values in literature (Figures 4.2 and
4.6)[24,27]. Although there was prominent charging for the case of Fe2O3 standard, the
charge correction was correct as the oxide peak value was similar to the one obtained
from the Fe3O4 which was not charging and according to the XRD analyses both of
these standards had expected compounds as major constituents (Figures 4.4 and 4.7).
From the analyses on the chemical iron standards it was also shown that after
extensive ion etching a progression of the initial oxidized state of iron to lower state is
taking place and signal on the metallic iron position is also emerging and stabilizing at a
fixed ratio with respect to the oxide state (Figures 5.1 and 4.3). This result is an artifact
of the ion etching and should not be considered as actual presence of pure metallic
iron in the oxide sample.
Figure 5.1: Oxidized vs metallic state of iron with etch depth.
77
Analysis on the iron foil showed that the iron oxide surface layer is 2.54 nm thick,
taking into account the etch rate calibration with Ta2O5, (Figure 4.12) based on the iron
metal normalized signal. This layer thickness is in the range expected for Fe-oxide
formed at room temperature. The chemical compositional profile indicated the
presence of manganese thick oxide particles (Figure 4.9) which were not removed
entirely by the etching process, assuming that they are thicker and distributed within a
uniform iron oxide layer.
The analysis of the ABC100.30 water atomized powder showed that the surface and its
compositional depth profiles resemble the ones acquired from the heat treated iron
foil. This is implies that all information extracted from the analysis of the latter can be
used for the characterization of the water atomized powder as well. The estimation of
the surface oxide layer thickness was given by the iron relative signal. Under the
annealing treatment of the powder to reduce the oxide layer from its production, the
impurity elements such as manganese diffuse on the surface and form oxide
particulates as shown in the SEM and EDX analysis. These particulates contribute to
the total oxide on the surface.
The XPS analysis of the ABC100.30 showed that in the as-received state there is a
slightly greater oxide thickness shown from the depth profiling compared to that for
fractioned powder samples. This could be due to the fact that the unsieved powder
has a great difference in the particle size distribution which ranges from a few microns
to 200 μm. This can cause shading effects during the XPS analysis because of the
geometry of the experimental arrangement, since the ion beam cannot etch the entire
oxide layer and thus more oxide contribution is obtained after ion etching (Fig 5.2).
Figure 5.2: XPS geometry and possible shadowing effect.
Results from the analyses on the sieved fractions for this powder are very close to one
another, indicating that when a narrower particle size distribution is examined the
deviation is minimal. Furthermore, it is shown that the surface layer thickness is
independent of the particle size, which is in accordance to the theoretical model [17],
while the degree of shape irregularity and surface roughness of the ABC100.30 is also
consistent along its size fractions.
78
XPS analysis of the gas atomized powder in the as-received state showed thicker oxide
layer. The large size particles (+106μm) appear to have thinner oxide (12.32nm) than
the other two with the middle fraction having the thickest (14.65nm).
The SEM and EDX analysis of the pure iron gas atomized powder in the as received
state showed the presence of inhomogeneous surface characteristics. The presence of
certain amount of more oxidized powder particles (Figure 4.37) is expected to give an
overestimation of the oxide layer thickness for all samples. For this reason an
annealing treatment was performed and as shown later from subsequent imaging and
chemical microanalysis on the annealed gas atomized powder, this gave a much more
homogeneous surface which mainly consists of only an iron oxide layer and some
dispersed particulates of silicon and manganese oxides (Figure 4.50).
The XPS analysis of the annealed gas atomized powder, as seen from the normalized
iron intensities and from the thickness determination method, has shown that the
large size particles (+106μm) appear to have thinner oxide layer as compared to the
other two fractions (Figure 4.44 and 4.45). This could be explained using the previous
argument from the un-sieved ABC100.30 powder as compared to its sieved fractions. A
wider particle size distribution on the sample can lead to overestimation of the true
layer thickness. This can also be seen in Figure 4.36 that for the large size fraction the
sample has a more homogeneous surface while for the other two fractions a rougher
surface is evident with a wider particle size distribution available for analysis. This
concept is also portrayed in Figure 5.3 where a theoretical narrow distribution of
spherical particles is compared against a wider one. The line above the particles
indicates the surface roughness profile and the dashed line the theoretical flat surface
Figure 5.3: Particle size distribution in the large size fraction of the annealed gas atomized (a) and in the other
two (b).
The middle fraction appears to have the thickest oxide layer compared to the other
two. This could be explained with the above argument as the middle size fraction has a
wider particle size distribution than the other two and thus shading effects are more
prone to occur for this fraction
5.2 Somaloy powder and phosphate standards

The XPS analysis of the phosphated iron foil showed that the surface layer consists of
an iron phosphate compound which shows a sharp substrate interface when
examining the relative intensities of the iron and phosphorus signal. The chemical state
of the surface layer of the phosphated iron foil, based on the peak values of iron and
phosphorus, is in agreement with the iron phosphate dihydrate chemical standard
(Figure 4.51) and close to the dehydrated phosphate (Figure 4.54) and literature data
[11]. Here, it is shown that it consists of a XRD amorphous or very fine crystallized iron
79
phosphate compound and the iron is present in a mixture of ferrous and ferric states.
The thickness of the surface layer was determined at the 50% of the relative P intensity
3.7 nm
From the SMC powder XPS analysis the chemical state and thickness of the surface
insulating layer were determined for both un-sieved and sieved fractions. As for the
case of the phosphated plate the peak position values for phosphorus and iron are in
agreement with the standards tested showing again the presence of XRD amorphous
iron phosphate in ferric and ferrous states. Additionally, the chemical compositional
profile for this grade in combination with the chemical microanalysis of the EDX layer
showed similarities with the ABC100.30 (Table 13) indicating the presence of a thin
iron based oxide layer and dispersed oxide particulates. Furthermore, the oxygen
signals from the SMC powder indicate the presence of more than one compound in the
surface layer (Figure 4.64). From the curve fittings of the oxygen signals for different
etch depths using reference values from the analyzed standards it can be concluded
that the layer is a complex mixture of iron phosphate and iron oxide in ferrous and
ferric states. These results indicate that the relative intensity of the phosphorous signal
should be used in order to determine the thickness of the phosphate coating at the 50-
55% of its value. Moreover, the metallic iron relative intensity at 65%, according to the
theoretical model, and oxygen relative intensity at 50% are in agreement and
determine the whole complex surface layer thickness.
Based on these results, the phosphate coating is in the range of 7 to 8 nm for all sieved
fractions of the SMC powder, much more than the one on the iron plate as expected
due to the difference in specific surface area between those two samples. Based on
the arguments though created earlier for the thickness of the homogeneous surface
layers for the water atomized iron powder fractions, one would expect that surface
layer on the SMC fractions would all follow the same trend and have similar values if
they were coated in a homogeneous manner. Conversely, the smaller size fraction has
the lowest thickness and the coarsest one the highest estimated both by phosphorus
or iron signal (Figure 4.67). Also, the un-sieved SMC sample shows lower thickness
value only to the coarsest size fractions when taking into account the phosphorus
signal while it shows higher values only for the smaller size fraction based on the iron
signal (Figure 4.65).These findings can be interpreted based on the effect of the
phosphating process. Shorter of time of the phosphating process would mean that the
oxide layer cannot be totally reduced as is suggested from the phosphating theory
kinetics and would form a thin amorphous type of compound. With that in mind the
phosphoric acid removes some of the oxide layer but then the coating starts to grow
on top of it.
If it is assumed that the SMC base powder is a similar to ABC100.30 then the values of
the oxide thickness match the above suggestion. For example the oxide thickness
calculated for the ABC100.30 powder was ~ 6 nm. A small part of this oxide layer is
removed during the phosphating process and then the coating grows on top of what is
left which is given from the phosphorous intensity (~ 6-7 nm again). Thus the total
layer thickness given from the iron intensity should be around 10-12 nm (in this case
11nm). The above can be also explained by the phosphating process factors that cause
the powder to oxidize.
The SEM and EDX analysis of the SMC grade showed highest intensity of the phosphate
coating at places of high degree of irregularity on a powder particle at regions with
concave surface (Figure 4.72) while in more convex type of regions the coating
thickness is smaller (Figure 4.74, spectrum 5). Additionally, the larger size particles
exhibit a more irregular shape than the finer ones. Hence, difference in the coating
80
thickness with increasing particle size can be expected as well as the intermediate
value of the un-sieved fraction between the finer size and the rest (Figure 5.4)
Figure 5.4: Schematic of phosphate concentration in different areas of a particle.
Moreover, with the SEM and EDX analyses it was observed that high concentration of
the iron phosphate compounds in the concave regions appear as flaky particulates or
small crystals in the nanometer scale (Figure 4.72). This can be explained by the fact
that during the phosphating process the solution has a better flow in convex areas
than concave ones where it is formed as flaky particulates.
Another explanation could be that these particulates are a result of the handling
process of the powder during sample preparation and thus a more thorough
investigation of this effect should be performed
Overall the present work compliments the theoretical model for calculating this type of
layers with experimental data. It also provides with a good basis for characterizing such
types of powder and estimating the coating thickness.
81
Chapter 6
Conclusions and future work
In the present study a standard and robust analysis method was developed for
characterizing insulating layers on metal powder, SMC powder by means of X-ray
photoelectron spectroscopy(XPS). For a valid and accurate analysis of such surfaces,
chemical standards were acquired and analyzed. Iron oxides were examined and their
main peak contributions were identified. An interesting observation here was the fact
that after excessive ion etching a signal on the metallic iron position is emerging and
stabilizes at a fixed ratio with respect to the oxide state. This result is an artifact of the
ion etching technique and should not be considered as actual presence of pure
metallic iron. Examination of an annealed pure iron foil provided the necessary data
for the metallic iron peak determination.
The analysis of the water atomized iron powder showed that the surface and its depth
compositional profiles resemble the ones acquired for the heat treated iron foil. The
powder has a clean surface with only some small particulates on the surface. The XPS
analysis showed that in the as-received state there is a higher apparent contribution of
oxide thickness than when it is fractioned due to the fact that the unsieved powder has
a great difference in the particle size and this can cause a lot of shading effects during
the XPS analysis because of the geometrical characteristics of the experimental
arrangement.
The XPS analysis of the gas atomized powder in as-received state showed a thick oxide
layer due to the presence of heavily oxidized powder particles and for this reason an
annealing treatment was performed. For the annealed gas atomized powder the
normalized iron intensities versus etch depth gave the thicknesses of the oxide layer.
The results show that the large size particles (+106μm) appear to have somewhat
thinner oxide layer as compared to the other two fractions. This could be explained
using the previous argument from the un-sieved water atomized powder as compared
to its sieved fractions. A wider particle size distribution of the sample can lead to
overestimation of the oxide layer thickness.
Phosphate standards and a phosphated iron foil were examined in order to establish
references to compare with the elements of the coating of the metal powder. The
chemical state of the surface layer of the phosphated iron foil, based on the peak
values of iron and phosphorous, is in agreement with the iron phosphate dihydrate
chemical standard and close to the dehydrated phosphate.
From the SMC powder XPS analysis, the chemical state and thickness of the surface
layer were determined for both un-sieved and sieved powder fractions whereby the
contribution to thickness from the phosphate layer and the Fe-oxide could be shown.
These findings can be interpreted as the effect of the phosphating process or the
sample preparation of the powder samples.
83
Future work
 Auger electron spectroscopy (AES)

The interest here is to determine the homogeneity of the coating and its thickness
using point analysis. Also etch rate of the iron oxide for fixed as well as varying angle of
ion incidence could be achieved and thus provide more accurate thickness calculation
in XPS.
 Narrower particle size distribution in fractions
By choosing more sieve sizes a more narrow size distribution in each fraction could be
achieved and thus more accurate determination of the effect of surface roughness
could be obtained.
 Ion etching angle dependence in XPS
The theoretical model of calculating oxide and layer coatings is also affected by the
angle between the normal to the surface under consideration and the ion incident
beam. By measuring a flat sample in XPS and progressively changing the angle this
dependence could be estimated experimentally as well.
84
References
[1] German, Randall M. Powder Metallurgy Science. New Jersey : Metal Powder
Industries Federation, 1984.
[2] Association, European Powder Mettalurgy. EPMA. [Online]
http://www.epma.com/New_non_members/intro-home.htm.
[3] AB, Hoganas. hoganas. [Online] http://hoganas.com/en/Segments/PM-
Components/Downloads/Handbooks/.
[4] Werner Schatt,Klaus-Peter Wieters. Powder Metallurgy-Processing and
Materials. Shrewsberry : European Powder Mettalurgy Association, 1997.
[5] Aerospace, GKN. gkn. [Online]
http://www.gkn.com/hoeganaes/media/Pages/product_data_sheets.aspx.
[6] Soft magnetic composite materials (SMCs). H. Shokrollahi, K. Janghorban.
Shiraz,Iran : Journal of Materials Processing Technology, 2007, Vol. 189.
[7] Wagner, Johanna M. X-ray Photoelectron Spectroscopy. New York : Nova
science publishers, 2011.
[8] D Briggs, M.P. Seah. Practical surface analysis,Vol 1,Auger and X-ray
photoelectron spectroscopy. West Sussex,England : John willey & sons Ltd, 1983.
[9] John F Watts, John Wolstenholme. An introduction to surface analysis by XPS
and AES. West Sussex,England : John Wiley & Sons Ltd, 2003.
[10] Olefjord, I. Surface characterization by XPS. Goteborg : Department of
Engineering Metals,Chalmers Univeristy of Technology.
[11] Phosphating of bulk a-iron and its oxidation resistance at 400 C. S. Rebeyrat,
J.L. Grosseau-Poussara, J.F. Silvain, B. Panicaud, J.F. Dinhut. Bordeaux :
Applied Surface Science, 2002, Vol. 199.
[12] Dr. Dmitri Kopeliovich . Substances and technologies. SubsTech. [Online]
http://www.substech.com/dokuwiki/doku.php?id=phosphate_coating.
[13] Properties of iron-based soft magnetic composite with iron phosphate–silane
insulation coating. A.H. Taghvaei, H. Shokrollahi, K. Janghorbana.
Department of Materials Science and Engineering, Engineering School, Shiraz
University, Shiraz, Iran : Journal of Alloys and Compounds, 2009, Vol. 481.
[14] Surface pretreatment. [Online]
http://www.surfacepretreatment.com/default.asp?id=2.
[15] Effect of Zinc Phosphating on Corrosion Control for Carbon. Dr. Mohammed
Hliyil Hafiz,Junan Sttar Kashan,Ali Shaker Kani. 5, s.l. : Engineering and
technology, 2008, Vol. 26.
[16] Joseph I. Goldstein, Charles E. Lyman, Dale E. Newbury, Eric Lifshin,
Patrick Echlin, Linda sawyer, David C. Joy, Joseph R. Michael. Scanning
electron microscopy and X-ray microanalysis. New York : LKluwer Academic/
Plenum publishers, 2003.
[17] Thickness determination of oxide layers on Spherically-shaped metal powders by
ESCA. L. Nyborg, A. Nulund, I. Olefjord. 110-114, Göteborg : Surface and
interface analysis, 1988, Vol. 12.
[18] The influence of phosphate on repassivation of 304 stainless steel in neutral
chloride solution. M.Lakatos-Varsanyi, F. Falkenberg, I. Olefjord. 1-2, s.l. :
Elsevier Science Ltd., 1997, Vol. 43.
[19] Characterization of surface oxides on water-atomized steel powder By XPS/AES
depth profiling and nano-scale lateral surface analysis. D.Chasoglou,E. Hryha,
M. Norell, L. Nyborg. Gothenburg : Elsevier, 2013, Vol. 218.
[20] Barbaux Y., Dekiouk M., Le Maguer D., Gengembre L., Huchette D., and
Grimblot J. s.l. : Appl. Catalysis A 90, 1992, Vol. 51.
85
[21] Phosphate Conversion Coatings, Corrosion Fundamentals, Testing, and
Protection. K. Ogle, M. Wolpers. p 712–719, s.l. : ASM Handbook, ASM
International, 2003, Vol. 13A.
[22] Kinetics of the phosphatizing reaction on carbonyl powders. Yu. G. Kir'yanov,
T. A. Zavina,V.G Syrkin.
[23] Ewald, P. P. Fifty Years of X-Ray Diffraction. Utrecht : s.n., 1962
[24] Seyama H., Soma M. 97 , s.l. : J. Electron Spectrosc. Relat. Phemon, 1987, Vol.
42.
[25] An XPS study of phosphate glasses. P.K Brow, C.M Arens, X. Yu, E.Day. 3,
s.l. : Physics and chemistry of glasses, 1993, Vol. 35.
[26] Heal, G. R. Principles of Thermal Analysis and Calorimetry. Salford : Peter
Haines, 2002.
[27] Studies of the oxidation of iron by air after being exposed to water vapour using
angle-resolved x-ray photoelectron spectroscopy and QUASES. A. P.
Grosvenor,B. A. Kobe, N. S. McIntyre. London : SURFACE AND
INTERFACE ANALYSIS, 2004, Vol. 36.
[28] Investigation of multiplet splitting of Fe 2p XPS spectra and bonding in iron
compounds. A. P. Grosvenor,B. A. Kobe, M. C. Biesinger and N. S. McIntyre.
London : SURFACE AND INTERFACE ANALYSIS, 2004, Vol. 36.
86

Characterization of Electrically Insulating Coatings For Soft Magnetic Composite Materials by Means of Surface Sensitive Analytical Techniques

Uploaded by

Copyright:

Available Formats

Characterization of Electrically Insulating Coatings For Soft Magnetic Composite Materials by Means of Surface Sensitive Analytical Techniques

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

Characterization of Electrically Insulating Coatings For Soft Magnetic Composite Materials by Means of Surface Sensitive Analytical Techniques

Uploaded by

Copyright:

Available Formats

Characterization of electrically insulating coatings for soft

magnetic composite materials by means of surface

Diploma work No. 114/2013

Diploma work in the Master programme Materials Engineering

Performed at: Department of Materials and Manufacturing Technology

Supervisor(s): MSc. Christos Oikonomou

Examiner: Professor Lars Nyborg

Diploma work no 114/2013

Gothenburg, December 2013

Powder metallurgy (PM) refers to the processing science of manufacturing complex

Figure 2.1: PM process route (author illustration redrawn from [2]).

Also to fully characterize a powder it is necessary to describe how it is manufactured

Manufacturing of iron powders

The main categories of manufacturing techniques are:

Mechanical alloying involves fracturing and shearing powdering material in order to

2.2 Soft magnetic composites powders

Thus, preparation of an SMC finished compact includes the following parts:

Figure 2.5: Schematic of SMC particles.

2.2.1 Iron core losses in electromagnetic applications

There are three main contributions to the dissipation in magnetic materials:

2.2.2 Advantages and Properties of soft magnetic composites

As mentioned above, SMC material exhibits significant advantages that make it

Phosphating is a well-known metal conversion treatment especially in automotive

Stages of general phosphating process

The stages of the phosphating process are as follows:

The main components of a conventional phosphating solution are:

 Water or acetone as the main solvent

3Fe + 6H++ 2PO43-= 3Fe2++ 2PO43-+ 3H2 (1)

Primary ferrous phosphate (soluble): Fe + 2H3PO4 = Fe(H2PO4)2 + H2 (4)

Secondary ferrous phosphate (insoluble): Fe(H2PO4)2 + Fe = 2FeHPO4+ H2 (5)

Tertiary ferrous phosphate (insoluble): 3Fe2+ + 2PO43- = Fe3(PO4)2 (6)

A simplified schematic of the phosphating process is shown in Figure 5.

Figure 2.5: Schematic of the phosphating process.

2.4 X-Ray photoelectron spectroscopy (XPS)

X-Ray Photoelectron Spectroscopy (XPS) is a versatile, well established and widely

where is the work function of the spectrometer and is characteristic of its

Figure 2.7: Schematic of the XPS instrument.

Figure 2.8: XPS survey spectrum of Fe-powder in the as received state.

2.5 Scanning electron microscopy (SEM) and energy dispersive X-ray

Figure 2.9: Schematic of SEM instrument.

Figure 2.11: EDX spectrum of Somaloy 500.

2.6 X-ray diffraction (XRD)

A powder X-ray diffractometer consists of an X-ray source (usually an X-ray tube), a

2.7 Thermogravimetry (TG)

Thermogravimetric analysis is the technique that measures the change in mass of a

Materials under examination

Table 1: Fe foil composition

Equipment and powder preparation

The TG instrument used in the present work is a simultaneous TG/DTA/DSC thermal

Iron oxides – iron phosphate

4.1 Analysis of the iron oxide standards

 Iron(III) oxide (Fe2O3)

 Iron(II,III) oxide (Fe3O4)

XPS Analysis of the powder fractions

SEM and EDX of the as-received powder

Figure 4.21: Overview of the ABC100.30 as received water atomized powder.

Figure 4.22: Appearance of the particulate features on the powder surface.

EDX analysis was performed to investigate the chemical composition of these

SEM and EDX of the powder fractions