Stainless Steel 2000

SEQ 0003 JOB IOM8420-000-020 PAGE-0003 PRELIMS
REVISED 18OCT01 AT 17:45 BY TF DEPTH: 62.04 PICAS WIDTH 45.03 PICAS
STAINLESS STEEL 2000

Proceedings of an International
Current Status Seminar on
Thermochemical Surface Engineering
of Stainless Steel
Held in Osaka, Japan
November 2000
Edited by Tom Bell
School of Metallurgy & Materials
University of Birmingham, UK
and
Katsuya Akamatsu
Kansai University, Japan
FOR THE INSTITUTE OF MATERIALS

IN ASSOCIATION WITH
THE INTERNATIONAL FEDERATION FOR
HEAT TREATMENT AND SURFACE ENGINEERING
AND THE JAPAN SOCIETY FOR HEAT TREATMENT
SEQ 0004 JOB IOM8420-000-020 PAGE-0004 PRELIMS

B0752
First published in 2001 by
Maney Publishing for
The Institute of Materials
1 Carlton House Terrace
London SW1Y 5DB
The Institute of Materials 2001
All rights reserved
Maney Publishing is the trading name of
W. S. Maney & Son Ltd
Hudson Road
Leeds LS9 7DL
ISBN 1902653491
Typeset in the UK by
Dorwyn Ltd, Rowlands Castle, Hants
Printed and bound in the UK at
The University Press, Cambridge
SEQ 0177 JOB IOM8420-017-001 PAGE-0177 SS2000-17 177-00

Reprinted from SS2000: Thermochemical Processing of Stainless Steel, T.Bell and K.Aakamatsu eds, Maney, 2001
Towards Quantifying the Composition of Expanded

Austenite
M.P. FEWELLa,*, P. GARLICKb, J.M. PRIESTa, P.T. BURKEc,
N. DYTLEWSKIc, K.E. PRINCEc, K.T. SHORTc, R.G. ELLIMANd,
H. TIMMERSd,e, T.D.M. WEIJERSd,e and B. GONGf
aPhysics
and Electronics, University of New England, Armidale NSW 2351, Australia

Microscope Unit, University of New England, Armidale, NSW 2351, Australia
cAustralian Nuclear Science and Technology Organisation, Menai, NSW 2234, Australia
dDepartment of Electronic Materials Engineering, Australian National University, Canberra, ACT 0200,
Australia
eDepartment of Nuclear Physics, Australian National University, Canberra, ACT 0200, Australia
fSurface Science and Technology, School of Chemistry, University of New South Wales, Sydney, NSW 2052,
Australia
bElectron
This paper is dedicated to the memory of Trevor Richard Ophel, Professor of Nuclear
Physics in the Australian National University, who died suddenly on 17th June 2000.
ABSTRACT
Secondary-ion mass spectrometry (SIMS) is an attractive method for elemental depth profiling for
nitrided stainless steel. However, if taken at face value, the results suggest migration of Mo and Cr from
the nitrogen-rich layer. More plausibly, the results indicate the presence of significant matrix effects, and
this conclusion casts doubt on the interpretation of all the SIMS depth profiles, including that of
nitrogen. We explored this question by examining standard samples prepared by nitrogen-ion implantation and comparing assays and depth profiles from SIMS, energy-dispersed X-ray microanalysis (EDX),
heavy-ion elastic-recoil detection analysis (HI-ERDA) and X-ray photoelectron spectroscopy (XPS). We
conclude that there are negligible matrix effects in the SIMS sputter yields of nitrogen and carbon, but
that the yields of MoCs+ and CrCs+ ions are significantly affected by the concentration of nitrogen. Thus,
the shape of the nitrogen and carbon depth profiles from SIMS can be taken at face value. HI-ERDA
gives depth profiles of the same quality as SIMS, but is limited to a relatively thin nitrogen-containing
layer. Results from EDX, HI-ERDA and XPS show that the relative concentrations averaged over the
treated layer of the main metals in the alloy are the same as in the bulk material. Detailed examination of
the peak positions and widths in XPS shows a clear chemical association between the nitrogen in the
treated layer and Cr and Mo, and probably also an NFe association, but none between nitrogen and Ni.
These results support a model of the kinetics of nitriding in austenitic stainless steel.
1.
INTRODUCTION
Stainless steel hardened by nitriding is potentially such a useful material, and plasma nitriding gives such promise of a commercially practicable method of producing it, that much
*Corresponding author (e-mail address: mfewell@metz.une.edu.au).
Present address: School of Chemistry, Physics and Earth Sciences, Flinders University, GPO Box 2100, Adelaide SA 5001,
Australia.
177

Stainless Steel 2000
effort world wide is now being devoted to realising this potential.15 As in many other
fields, a significant issue for this research is the characterisation of treated samples, for the
purpose of comparing the efficacy of different treatment processes. Many diagnostic
methods are used; among them are techniques for determining the composition of the
treated layer. These range from methods familiar in metallurgy, such as glow-discharge
optical spectroscopy, to the most sophisticated techniques developed in modern surface
science. A feature of all methods is the need for calibration: whether one is measuring the
yield of photons, electrons or ions, there is always the question of relating the observed
yield to the concentration of the various species in the material under study. The work
reported in this paper was prompted by just such a question of calibration, in this case in
the results of secondary-ion mass spectrometry (SIMS) on samples of nitrided AISI 316
stainless steel.
Figure 1 shows a typical SIMS elemental depth profile from a nitrided sample. (Experimental details are given in 2.1 and 2.2.) There are many interesting features in
these data, not least the shape of the profiles for carbon and nitrogen. However, the
profiles for the four main alloy elements indicate that the interpretation may not be
straightforward. All four metals show a lower ion-current ratio IMCs+ in the nitrogencontaining layer compared with the bulk. This is to be expected from sputtering theory,
which indicates that the ion-current ratio of a species M should be proportional to its
atomic concentration cM (atom %), and cM values for the metal elements are reduced by
the presence of the nitrogen. But closer inspection shows that the details do not match
this explanation. Table 1 gives the ratio of ion-current ratios in the treated layer to the
bulk for the four main alloy elements. A straightforward interpretation would explain the
value for iron as reflecting a 25 at. % concentration of nitrogen in the treated layer.
However, one would then expect the values to be the same for all the metals, and clearly
they are not. Either there has been migration of Cr and Mo from the treated layer, or the
ion-current ratios of these elements are influenced by the presence of nitrogen differently
from those of Fe and Ni. Given that Fe is the main constituent of the alloy (65 at. %),
there may even be evidence that INiCs+ is behaving in an opposite manner from ICrCs+
and IMoCs+.
Influences of this sort are referred to as matrix effects; their presence clearly complicates the interpretation of depth profiles. One can no longer assume that, for example, a
halving of IMCs+ implies a halving of cM; rather there may be a change in the calibration
factor relating these two. Of the elements profiled in Fig. 1, most interest perhaps lies in
nitrogen and carbon, but there is no reason why the profiles of these should be less
affected by matrix effects than the profiles of the metal elements.
We have explored this problem by preparing samples of AISI 316 stainless steel
containing known concentrations of nitrogen. We also examine the same or similar
nitrided samples with a range of composition-determining techniques. Our results,
though not absolutely conclusive, strongly suggest that the nitrogen profile in Fig. 1 can
be taken at face value, whereas the Mo and Cr profiles cannot. That is, matrix effects are
significant for Mo and Cr, but at most minor for N. We argue that matrix effects are also
likely to be negligible for carbon.
178

Towards Quantifying the Composition of Expanded Austenite
Fig. 1 Relative ion-current ratios IMCs+ = IMCs+/ICs+ as a function of sputter time and hence
depth for a sample of AISI-316 stainless steel nitrided in the UHV chamber for 3 h at 400C in a
N2 + 25% H2 gas mixture of total pressure 0.53 Pa without pretreatment. The vertical line shows
the thickness of the treated layer determined from a cross-sectional electron micrograph of the same
sample (Fig. 4a). The broken line shows the carbon ion-current ratio in the bulk, to emphasise the
manner in which carbon is piled up at the interface between the treated layer and the bulk.
Table 1 Ratio of ion-current ratios IMCs+ in the treated layer to those in the bulk for the main alloy
elements (data from Fig. 1).
IMCs+ (treated layer)
Element
IMCs+ (bulk)
Cr
Fe
Ni
Mo
0.625 0.005
0.748 0.003
0.816 0.005
0.338 0.005
In addition, we note the strengths and weaknesses of each technique used in its application to nitrided stainless steel. This includes questions of the depth that can be profiled, the
depth and concentration resolution possible, mass discrimination, and the extent to which
the composition of the sample may be altered by the technique. Not least, we also highlight
the information that can be obtained on aspects other than an assay of the sample.
179

2.
EXPERIMENTAL METHODS AND RESULTS
We examined a total of twelve samples of one grade of austenitic stainless steel, nitrided by
exposure to a low-pressure rf plasma at elevated temperature, and several untreated and
nitrogen-implanted samples as controls. Our experience in conducting nitriding in lowpressure rf plasmas610 gives us confidence in the reproducibility of the process. As far as
practicable, samples were examined by more than one technique. The following subsections detail the nitriding process, the diagnostic techniques employed and the results
obtained.
2.1.
SAMPLES AND THE NITRIDING PROCESS
The material examined is AISI 316 austenitic stainless steel. Because its density is required
for the analysis of the results from several of the techniques reported herein, we determined
this by direct measurements on an accurately machined cylinder of our stock material and
also by a method based on Archimedes principle. After correction for the buoyancy of air,
the two methods agreed within uncertainty. The mean value of the density is 7950 8 kg
m3.
The nitriding chambers, sample preparation and process conditions have been described
previously.610 In brief, discs of stainless steel are parted from a rod, surface ground and
polished, ultrasonically cleaned in ethanol, and loaded onto the stage of the nitriding
chamber. All of our chambers are fabricated from stainless steel and pumped by turbomolecular pumps. In each, the chamber volume contains a single-turn inductively
coupled antenna also fabricated of stainless steel. With an appropriate filling pressure, a
diffuse rf glow discharge fills the chamber when rf power is applied to the antenna.
Samples may be treated either with or without a glow-discharge pretreatment step. If
without, the sample is heated to the working temperature in vacuum prior to filling the
chamber with nitrogen or a nitrogen/hydrogen mixture and commencing the nitriding.
Where a pretreatment is included, the chamber is filled with hydrogen or a hydrogen/argon
mixture and an rf glow discharge is maintained as the sample is being heated. At the end of
the pretreatment, the chamber is pumped out and filled with nitrogen for the nitriding
step. In all cases, the sample cools in vacuum following the treatment. The ranges of
treatment parameters used for samples studied herein are given in Table 2.
The samples were treated either in the hot-wall chamber at the Australian Nuclear
Science and Technology Organisation (ANSTO) or the UHV chamber at the University
of New England (UNE).8,10 In the first, the whole volume is at the working temperature,
which is monitored by both a thermocouple and an optical pyrometer. In the UHV
chamber, the sample stage is heated and the temperature is measured by a thermocouple
embedded in the stage. This is calibrated prior the treatment to allow for the temperature
drops across the vacuum gap between the sample and the stage.
All processes use flowing gas. Where a gas mixture is used, the percentages quoted in
Table 2 refer to partial pressures. Both the UNE and ANSTO systems show differential
pumping, with hydrogen being pumped the most efficiently. The effect is not small: to
obtain equal partial pressures of nitrogen and hydrogen in our chambers, the hydrogen
180


Table 2 Nitriding conditions used for samples included in this study.
rf power and frequency

sample bias
pretreatment (where applied):
gas
total pressure
temperature
pretreatment time
nitriding step:
gas
total pressure
temperature
treatment time
UHV chamber
Hot-wall chamber
80 W at 13.56 MHz
0V
250 W at 13.56 MHz

250 V
H2
133 mPa
1.04.0 h
Ar + 50% H2
400 mPa
rising from ambient to nitriding temperature
1.02.0 h
N2 or N2 + 25% H2
400530 mPa
300400C
1.04.0 h
N2
4001000 mPa
350450C
3.05.0 h
flow rate must be about three times that of the nitrogen. The actual values in each chamber
were determined prior to the treatments by noting the relationship between flow rate and
pressure for each gas separately.
2.2.
SECONDARY-ION MASS SPECTROMETRY
The present work extends a previous study of nitrided stainless steel by secondary-ion mass
spectrometry (SIMS),11 the principal finding of which is that the sputter rate (crater depth
vs sputter time) is constant in nitrided AISI 316 stainless steel. As in the earlier work, we
used the Cameca IMS 5F magnetic-sector mass spectrometer located at ANSTO, employing a 10 keV Cs+ beam with the sample biased at +4.5 kV and counting caesium binary
ions MCs+, where M is the species being monitored. Matrix effects in the yields of MCs+
ions are much less pronounced than those in the atomic ions M+,12 but are not necessarily
absent.13 The caesium beam was rastered over a square 250 m on a side and sputtered
ions were collected from circular area 60 m in diameter centred on the raster pattern.
During the determination of a depth profile, several elements were monitored by counting
for 1.0 s at each chosen value of mass, in repetitive sequence from the lowest mass to the
highest.
Figure 1 shows an example of such a depth profile. The quantity plotted is the ioncurrent ratio IMCs+: the measured current IMCs+ of the various MCs+ ions divided by the
measured Cs+-ion current. (Dividing by the Cs+-ion current eliminates effects of drifts in
the primary beam current and reduces matrix effects.)13,14 The sputter yields for the seven
elements shown encompass all the elements expected in nitrided stainless steel, except for
the 2.0 at. % of Si, which was not detectable. The depth scale shown at the top of the
figure was calibrated after the profiling by measuring the depth of the sputter crater with a
Tencor Alpha-Step stylus profilometer and relying on the previous observation of constant
sputter rate in this material.11 The vertical line near the nitrogen profile shows the
thickness of the treated layer as determined by electron microscopy (see 2.4 and Fig. 4a).
181

As mentioned in 1, sputtering theory1214 can be summarised as predicting the increase in the ion-current ratios of Cr, Fe, Ni and Mo at the boundary between the treated
layer and the bulk to be compensated for entirely by the reduction in the ion-current ratio
of the nitrogen. This applies only in the absence of matrix effects, and is not the case in the
data shown in Fig. 1. Hence, to extract elemental concentrations from IMCs+ values, it is
necessary either to examine standard samples containing known amounts of nitrogen, or to
find another technique to determine the composition of each region of the nitrided
samples. Both approaches are explored in the following subsections.
2.3.
STANDARDS FOR SECONDARY ION MASS SPECTROMETRY
To prepare standards for calibrating the SIMS measurements of nitrogen concentration, we

implanted samples of AISI 316 stainless steel with known fluences of nitrogen using a beam
of 2.00 MeV N+ atomic ions from the 3 MV single-ended Van der Graaff accelerator at
ANSTO. The relatively high implantation energy was chosen to ensure that the region of
high nitrogen concentration is well away from the surface. During the implantation, the
beam current was kept below 0.5 A to avoid elevated sample temperature and so minimise
diffusion of the implanted nitrogen. The implanted area was 3.0 mm 3.2 mm. This was
defined by apertures upstream of the target chamber, with the beam defocused to more than
fill them. The area was determined by irradiating a polymer film in the target position, and
measuring the extent of the damaged area using an optical microscope. The total implanted
fluence was determined by beam-current integration, using the whole scattering chamber as
an isolated Faraday cup. The fluence is expected to be accurate to 5%. Four samples were
prepared, with fluences ranging from 3.0 1015 to 3.0 1017 ions cm2.
The resulting samples were analysed by SIMS in the usual manner; Fig. 2a shows an
example. As in Fig. 1, the depth scale was obtained by direct measurement of the depth of
the sputter crater. The concentration cM of element M can be related to the ion-current
ratio IMCs+ by
IMCs+ = cMSMCs+,
(1)
where SMCs+ is the calibration factor. Values of cN and hence SNCs+ were extracted from the
data in Fig. 2a in the usual way;15 Fig. 3 shows the result. At nitrogen concentrations above 2
at. %, the calibration factor is constant. Rather higher values of SNCs+ occur at very low
concentrations, but this is of little consequence for the interpretation of the nitrogen profile
of a nitrided sample. It must be noted that the highest concentration studied, 6.9 atom %, is
somewhat below the 2030 at. % expected in a nitrided sample, so that the possibility of
variations in the calibration factor at very high values of cN is not ruled out. Nevertheless, the
evidence of this work is that, for the nitrogen concentrations studied, the SIMS calibration
factor for NCs+ from nitrided austenitic stainless steel is at most weakly dependent on
nitrogen concentration for concentration values of interest.
The nitrogen-implanted samples give information only on the SIMS calibration factor
for nitrogen. To explore those for the other elements, we used other methods of elemental
quantification, as detailed in the next three subsections.
182

Fig. 2 (a) Relative ion-current ratios INCs+ = INCs+/ICs+ of NCs+ ions as a function of sputter
time and hence depth for a sample of AISI-316 stainless steel implanted with 1.00 1017 nitrogen
atoms cm2 and (b) calculated depth distribution for 2.00 MeV nitrogen ions in AISI 316 stainless
steel using the computer code TRIM. (Panel (b) is discussed in 3.1.)
Fig. 3 SIMS calibration factors SNCs+ determined as described in the text as a function of
nitrogen concentration cN.
2.4.
ENERGY-DISPERSED X-RAY MICROANALYSIS USING A SCANNING

ELECTRON MICROSCOPE
We have previously reported the use of scanning electron microscopy (SEM) of etched
sectioned samples as a means of measuring the thickness of the treated layer.11 Examples
are shown in Fig. 4. These were taken with the JEOL JSM-5800LV instrument at UNE.
The samples were prepared by transverse sectioning, casting in epoxy resin, grinding and
polishing, etching for 10 s in Marbles solution, and coating with carbon to minimise
charging under the electron beam.
The SEM is also equipped with an X-ray detector to monitor fluorescence from the
beam spot. We used this to perform chemical analysis in selected regions on several
samples, a technique referred to as energy-dispersed X-ray microanalysis (EDX). The UNE
system comprises an Oxford Instruments Si(Li) detector optimised for x rays from light
elements and viewing at 35 to the sample surface. We used a 20-keV electron beam. The
fluorescence yield was calibrated against a cobalt standard just prior to commencement of
183

Fig. 4 Scanning electron micrographs of cross sections of (a) the same sample as in Fig. 1 and (b) a
sample nitrided in pure N2 in the hot-wall chamber for 3.0 h at 350C following a 1.75 h
pretreatment in H2/Ar. The light circular marks show locations at which X-ray microanalysis was
performed.
the EDX analysis and every two hours thereafter. Initial tests showed the need to collect
high-statistics X-ray spectra if the presence of nitrogen was to be detectable at all. This
precluded area or even line scans. Instead, X-ray spectra were recorded at a variety of
locations in the treated layer and the bulk.
Figure 5 shows observed spectra. The carbon peak comes mainly from the coating, and
in addition from deposition under the electron beam during the collection of the spectrum.
The nitrogen K line can be discerned by comparison of the treated-layer spectrum with
the spectrum from regions in the bulk, as shown in the inset; the line is too weak to provide
a useful measure of nitrogen concentration. The energy of the electron beam was too low
for efficient formation of K-shell holes in Mo, but the L line from this element is clear, as
is the K line from Si which, according to the specification of AISI 316, may be present at
up to 2 atom %. Manganese may also be present at the same level; the line marked Mn in
Fig. 4 is a blend of the Mn K and Cr K lines. Line intensities were converted to element
concentrations using standard X-ray fluorescence yields.
The images in Fig. 4 were taken after the EDX analysis. They show circular beam burns
indicating the locations where X-ray spectra were collected. A total of nine locations in the
bulk region of three samples were examined; all show the same composition within
uncertainties. In the treated layer, sixteen locations were examined, including the three in
the interface region shown in Fig. 4b, and again all showed the same composition. Further,
there is negligible difference between the two regions. (Details are presented in 3.2.)
Because nitrogen and oxygen could not be included in the assay of the treated layer, the
assays for the heavier elements in the two regions are directly comparable. This means that
the differences in the ratio of sputter yields shown in Table 1 reflects variations in SIMS
calibration factors. That for the MoCs+ ion is particularly sensitive to the concentration of
nitrogen in the matrix.
184

Fig. 5 X-ray fluorescence spectrum from locations in the treated layer. The spectrum shown is the
sum of all spectra recorded from three samples, one of which is the same sample as in Figs 1 and 4a.
The X-ray peaks are identified. The inset shows an expanded view of the low-energy part of the
spectrum, compared with a summed spectrum from locations in the bulk shown as full points. The
only difference between the two spectra is the occurrence of the nitrogen K line, as shown.
This conclusion assumes that the EDX calibration factors are not influenced by nitrogen
concentration. Support for this assumption was sought from other methods of quantitative
analysis: heavy-ion elastic-recoil detection and x-ray photoelectron spectroscopy.
2.5
ELASTIC-RECOIL DETECTION WITH HEAVY IONS
Heavy-ion elastic-recoil detection analysis (HI-ERDA) is an ion-beam analysis technique

for the compositional depth-profiling of materials. In this technique, incident ions penetrate into a sample, interacting with atoms and much less frequently with nuclei in the
sample. Some recoil ions from nuclear scattering are ejected through the sample surface,
permitting their detection. Typically, a wide range of species with a wide range of energies
is ejected, so that particle identification is required. This is achieved either by time-of-flight
measurements or by measurements of energy loss in a gas-ionisation detector. The numerous interactions of both the incident and the recoil ions with electrons in the sample
cause loss of kinetic energy. Consequently, the energy with which the recoil ions emerge
from the surface is a function of the depth of the nuclear collision, with lower-energy
recoils originating from deeper inside the sample. By measuring the energy spectra of the
recoil ions, stoichiometric information can be correlated with depth.
The HI-ERDA in this work was carried out at the Australian National University using
beams of 200 MeV and 241 MeV gold ions from the 14UD Pelletron accelerator. The
beam was incident at 22.5 to the sample surface. Recoil ions were detected with a
position-sensitive gas-ionisation detector that had been developed for the specific demands
of HI-ERDA with heavy-ion beams.1618. The detector was centred at a recoil angle of 45
185

Fig. 6 Two dimensional plot of ion yield against energy loss E and total energy E for elastic
recoils following 200-MeV 197Au bombardment of a sample treated in pure N2 in the UHV
chamber for 3.0 h at a pressure of 400 mPa and a temperature of 300C. The magnitude of the ion
yield is indicated in gray-scale. Identified species are noted.
with a net detection solid angle of 3.50 0.05 msr. Propane was passed through the
detector at a constant pressure of 7.0 kPa. The energy E of the recoil ions was measured
directly using a grid electrode18 with an energy resolution of 1.6% for 28.6 MeV 16O ions.
The anode of the detector is divided into three segments, allowing simultaneous energyloss measurements for particles with different ranges in the gas. The first segment, adjacent
to the entrance window, provides a measurement of the initial energy loss E of ions
entering the detector. The middle segment is further subdivided in a saw-tooth pattern to
provide information on the lateral position of the ion, so allowing compensation for the
kinematic variation of energy with recoil angle. Finally, the third segment of the anode
gives a measurement of the residual energy Eres of the ions. It is also used to detect protons.
Combining the signals from the short E section with those from the grid electrode in a
two-dimensional spectrum allows the identification of the atomic number of a recoil ion.
This is shown in Fig. 6 for data from a nitrided stainless steel sample; all the elements
indicated can be readily identified and separated. The signal from Mn, which is nominally
present at a level of 2 at. %, is obscured by those from the much more abundant elements
Cr and Fe. Hydrogen is detected simultaneously but identified separately using the Eres
signal, and is therefore not shown in Fig. 6.
A total of seven nitrided samples and one untreated sample were examined. For each,
energy spectra for each element were obtained by gating on appropriate areas in the EE
plots. Figure 7 shows these for N and C extracted from the data in Fig. 6 and following
correction for the kinematic energy spread over the detector acceptance angle. The nitrogen energy spectrum clearly terminates below E 10 MeV, reflecting the interface between
186

Fig. 7 Energy spectra of (a) carbon and (b) nitrogen recoils projected from the data shown in Fig.
6. The smooth curves show simulations using RUMP.
the nitrogen-rich layer and the underlying bulk material. The stoichiometry of the samples
was determined by fitting theoretical simulations to the energy spectra using the code
RUMP,19 as shown by the broken lines in Fig. 7. Fits such as those shown in Fig. 7 for
each element give the relative concentrations; results are compiled in 3.2. All samples
studied give the same assay, within uncertainties.
All samples have a thin oxidised surface layer (1017 atoms cm2). This is evident in Fig.
7a, which is typical of all the samples. In this case, the surface layer contains 5 at. %
carbon and 18 at. % oxygen. Underneath this layer, the oxygen content is much reduced,
at 0.2 at. %. For carbon, the situation is more complex. Below the surface layer, the
concentration decreases to a constant level of 0.4 at. %. At the interface between the
treated and the bulk the carbon concentration piles up to a maximum of 0.8 at. % and
then decreases to a constant 0.4 at. % in the substrate material. The region of relatively
high carbon concentration is 1.5 1018 atoms cm2 in thickness. The RUMP simulation
shown as the full curve in Fig. 7b corresponds to a peak nitrogen concentration of 27.9
0.7 at. %. A detailed depth profile for nitrogen is presented in 3.1.
A feature of HI-ERDA is its sensitivity to hydrogen.201 This is of interest in the
examination of nitrided stainless steel because of the use of hydrogen in precleaning and as
a gaseous additive during the nitriding step.9,22 Whether or not there is any retained
hydrogen in the treated layer is relevant to the question of how the hydrogen aids the
nitriding process.
Protons recoiling from the sample were detected simultaneously with heavy ions using
the transmission technique.20 The intensity of the detected hydrogen was related to the
hydrogen concentration in the sample using tabulated stopping powers. The samples
examined included some that had been exposed to hydrogen during the treatment. All
samples, whether exposed to hydrogen or not and including the untreated control sample,
showed the same [H]:[Fe] concentration ratio within uncertainty; the mean value is (1.3
0.7) 103. It can thus be concluded that exposure to hydrogen during the treatment,
187

whether in a pretreatment step or as part of the process gas in the nitriding step, does not
lead to measurable retention of additional hydrogen in the treated layer.
2.6.
X-RAY PHOTOELECTRON SPECTROSCOPY COMPOSITIONS
X-ray photoelectron spectroscopy (XPS) is not primarily a technique for compositional

analysis; its strength is the provision of information on the chemical states of atoms in a
sample, that is, on the measurement of binding energies of core-level electrons. However,
the technique can also be used to give an indication of the surface composition of a sample
by measuring photoelectron peak areas, which are then normalised by standard sensitivity
factors and instrument transmission functions. Here, we look at both aspects of the XPS
data. XPS is very much a surface-analysis technique; facilities for sputter etching are
commonly provided, as with the instrument used in the present study, but it is usually not
practicable to penetrate more than several 100 nm below the as-presented surface. In the
present analysis, both chemical identification and atomic quantification are restricted to a
surface layer up to about 40 nm thick. This is sufficient to examine the treated layer below
the oxygen- and carbon-rich region of surface contamination.
Treated and untreated samples were examined using the ESCALAB 200i-XL X-ray
photoelectron spectrometer at the University of New South Wales. This instrument uses a
200 W double-focused monochromated Al K X-ray source. The X-rays are incident on the
sample surface at an angle of 58 to the sample normal. The area analysed is approximately
rectangular with dimensions 0.8 mm 0.5 mm. Photoelectrons emitted normal to the
sample surface are detected with six electron channel multipliers in parallel. Photoelectron
spectra were acquired at a pass energy of 20 eV. Under these conditions, the resolution of the
electron spectrometer is 0.55 eV for Ag 3d5/2 photoelectrons. Sputter etching of the sample
surface was carried out with a 5.0 keV Ar+ beam. When activated, the ion beam is incident
on the sample at an angle of 50 to the surface. The sputter rate was not measured but
previous experience indicates a sputter rate of order 0.2 nm C1, where the charge refers to
the integrated beam current: the total charge delivered as Ar+ ions to the surface.
Three samples were examined, two nitrided and one untreated. Spectra were taken from
the as-presented surfaces and following many successive episodes of sputtering up to a total
delivered Ar+ charge of 300 C. Figure 8 shows a photoelectron spectrum obtained from a
nitrided sample at a location in the region of surface oxidation, which a depth profile shows
was not penetrated after about 120 C of delivered Ar+.
A comparison of depth profiles from treated and untreated samples shows a thicker
oxidised surface layer in the untreated sample. The treated samples have a region of very
high nitrogen concentration immediately below the oxygen-rich surface layer. The nitrogen concentration peaks at 32.6% after 18 C of delivered Ar+ (Fig. 8) and then settles
down to a steady value of 20% deeper into the sample. The depth profile of Cr is very
similar to that of N, showing the same pattern of surface enrichment. Molybdenum is also
surface-enriched: its surface concentration after 18 C of delivered Ar+ is 4.9%, more than
twice the value deep into the treated sample. On the other hand, iron and nickel are both
surface-deficient but, unlike Fe, Ni shows a peak in its concentration profile below the
188

Fig. 8 X-ray photoelectron spectrum (Al K X-rays) from a sample treated in the hot-wall
chamber with a 2 h pretreatment followed by nitriding at 600 mPa and 400C for 5.0 h. The main
peaks are identified; the others are, at low binding energy, 3s, 3p and Mo 4p photoelectrons, and
elsewhere various Auger-electron lines. The sample was sputter etched for an integrated Ar+ beam
current of 18 C before recording the spectrum. The sputter rate is expected to be of order 0.2 nm
C1.
layer rich in N and Cr. For the untreated sample, Mo, Fe and Ni all show surface
deficiency, due mainly to the high concentration of oxygen, but Cr is slightly surfaceenriched. The maximum Cr concentration appears 18 C of delivered Ar+. These results
are compared with those of the other techniques in 3.2.
2.7.
X-RAY PHOTOELECTRON SPECTROSCOPY CHEMICAL BINDING
The previous subsection indicates the ability of XPS to give concentration profiles with
high depth resolution in the very near-surface region. The second capability of XPS is the
determination of the chemical relationship between nitrogen and the four metals through a
detailed examination of the binding energies of the Cr 2p3/2, Mo 3d5/2, Fe 2p3/2 and Ni
2p3/2 photoelectron lines. Figure 9 shows regions of the photoelectron spectra relevant to
the lines mentioned for near-surface and deeper layers of a treated and an untreated
sample. For chromium, the binding energies of 2p3/2 electrons in Cr(0) and Cr(III) are
574.2 eV and 576.6 eV respectively.23 Hence the most prominent component in the
spectrum from the surface region of the untreated sample (Fig. 9a(i )) is probably due to
Cr2O3. For the treated sample, the Cr 2p3/2 peak starts at 575.3 eV on the very surface
(not shown in Fig. 9) and its width is clearly larger than that of the Cr(0) peak shown in
Fig. 9a(iv). After 18 C of etching (trace (ii )), the peak has shifted to 574.9eV. This
indicates the presence of CrN.236 The only other likely candidate is Cr2N, but the Cr
2p3/2-electron binding energy for this species is even higher.24 Deep in the sample (trace
(iii )), the peak position has shifted again, this time to 574.2 eV, the same value as Cr(0),
but the peak width remains larger than that from Cr(0) deep in the untreated sample (trace
(iv)). This is attributed to ion-beam-induced reduction of the nitride.
189

Fig. 9 Expanded sections of X-ray photoelectron spectra in the regions of the (a) Cr 2p3/2, (b) Mo
3d doublet, (c) Ni 2p3/2 and (d) Fe 2p3/2 peaks taken from (i ) an untreated sample without sputter
etching (Mo 3d and Fe 2p3/2) or after 18 C of etching (Cr 2p3/2 and Ni 2p3/2), (ii ) a treated
sample after 18 C of etching, (iii ) a treated sample after 210 C of etching and (iv) an untreated
sample after 210 C of etching. Electron binding energies of some features are indicated.
Molybdenum behaves similarly to chromium. The binding energy of Mo 3d5/2

photoelectrons (the member of the doublet at lower binding energy) is: Mo(0) 227.6 eV,
Mo(II) 228.3 eV, Mo(IV) 229.3 eV.23 Values for nitrides of Mo are not known to us. The
untreated sample shows Mo(0) peaks both at the surface (Fig. 9b(i )) and at depth (Fig.
9b(iv)), with a FWHM of 0.7 eV. The 3d5/2 peak in the spectrum from the surface (trace
(i )) has a tail extending to higher binding energy; this is attributed to Mo(VI) oxide. For
the treated sample after 18 C of etching (trace (ii )), the 3d5/2-photoelectron peak is at
228.6 eV with a FWHM of 1.2 eV. This could be due to either the mixture of Mo(II) and
Mo(IV) or to nitrides of Mo. However, the value of the FWHM is not large enough for the
first explanation, in view of the binding energy difference between these two oxidation
states. This leaves nitrides of Mo as the most likely explanation of the Mo 3d5/2 spectrum
in trace (ii ). The association of a region enriched in Mo with the presence of N supports
the suggestion of Mo nitrides. Also, the layer of surface oxidation on the treated sample is
thinner and the surface concentration of Mo higher compared with the untreated sample,
so Mo(0) would be seen were the atoms not bonded to N. Deep into the treated sample
(trace (iii )), the peak has shifted down to 228.0 eV. This is again due to ionbombardment-induced reduction of the Mo nitrides.
190

The Ni 2p3/2 photoelectron peaks (Fig. 9c) are not affected by the treatment. This
together with the surface depletion of the element indicates that nickel is not bonded to
nitrogen. The situation concerning iron, shown in Fig. 9d, is not so clear. Figure 9d(i ),
taken from the untreated sample without etching, shows binding energies of 707.0 eV and
710.5 eV for 2p3/2 photoelectrons from Fe(0) and Fe oxide respectively, the latter being
most likely Fe(III). A comparison of the near-surface spectrum from the treated sample
(Fig. 9d(ii )) with both spectra from the untreated sample(traces (i ) and (iv)) shows a peak
shift of 0.20 eV. It is not clear whether this is due to nitrides of Fe, although the value of
the shift is of the right magnitude.23,256 Further evidence for nitrides of Fe comes from
the widths of the 2p3/2-photoelectron peaks. The peak in trace (ii ) is clearly broader than
that in trace (iv). Even deep in the treated sample (trace (iii )), the peak is 0.19 eV wider
than at the corresponding position in the untreated sample (trace (iv)), although the 0.2eV peak shift apparent near the surface (trace (ii )) has disappeared. Depth nprofiling
shows that the N concentration in the treated sample is the highest after 18 C of etching,
the same amount of etching as in Fig. 9d(ii ). This suggests some chemical relationship
between Fe and N. Taken together, the data indicate a significant possibility of iron
nitrides in a very thin surface layer of the treated sample, probably no more than 5 nm
thick. Such a layer has been identified by transmission electron microscopy.27 This suggestion does not contradict the results of X-ray diffraction because such a thin layer of iron
nitride would not be detectable by XRD.
3.
3.1
DISCUSSION
MATRIX AND OTHER EFFECTS IN SIMS
The great attraction of SIMS lies in its visually striking results: data such as those in Fig. 1
invite immediate and straightforward interpretation. Also, the data are relatively simple to
collect; depth resolution is good, being limited by the depth of penetration of the Cs+ ions;
and depths of several m can be profiled in a practicable amount of machine time.
However, the question of matrix effects intervenes, and this prompted the other studies
reported here. As noted in 1, the data in Table 1 suggest at face value that some Mo and a
lesser amount of Cr migrated out of the treated layer during the treatment. However, EDX
contradicts this, showing the same ratios [Cr]:[Fe]:[Ni]:[Mo] in the treated layer as in the
bulk. As discussed in the next subsection, HI-ERDA and XPS support the EDX result.
Hence we have evidence for a strong SIMS matrix effect for Mo and a lesser one for Cr.
On the other hand, the analysis of implanted samples (2.3) suggests a negligible matrix
effect for N, at least at high concentration. Another way of looking for matrix effects is to
compare depth profiles from different techniques. Figure 10 shows nitrogen depth profiles
from the same sample determined by SIMS and HI-ERDA. The degree of similarity is very
close, providing good evidence of the absence of SIMS matrix effects in the case of nitrogen.
Internal evidence in Fig. 1 suggests a negligible matrix effect also for carbon. The broken
line on the carbon profile in Fig. 1 shows the concentration in the bulk. The area above
this line in the interface region equals the missing area below the line in the treated layer.
191

Fig. 10 Nitrogen depth profiles from the same sample as in Figs 6 and 7 according to (a) SIMS
and (b) HI-ERDA. The HI-ERDA depth profile was obtained by dividing the measured energy
spectrum (Fig. 7b) by a calculated spectrum based on a constant nitrogen content of 27.9 atom %.
The filled points show the level of agreement with a simulation using RUMP.
This is only circumstantial evidence, but it nevertheless points strongly to a constant SIMS
sensitivity factor; for otherwise it would require a finely balanced variation to produce the
observed equality in areas. All carbon depth profiles examined show similar features, and all
show the equality of areas.
A comparison of Figs 2a and 2b indicates another aspect of the SIMS results. Figure 2b
shows the results of a calculation of the range distribution of 2.00 MeV nitrogen ions in
AISI 316 stainless steel, carried out using the code TRIM.28 The calculated depth of the
peak of the profile agrees well with the measured value in fig. 2a, but the shape does not:
TRIM indicates a asymmetric distribution with a tail toward the surface; the SIMS data
show a tail in the opposite direction. Thermal diffusion of the implanted nitrogen might be
a possible explanation for this, despite the efforts taken to keep the substrate temperature
down during implantation. It is, however, hard to see how diffusion could convert the peak
shape in Fig. 2b to that in Fig. 2a. Alternatively, there may be roughening of the floor of
the sputter crater, so that the value of INCs+ reflects the nitrogen distribution at a range of
depths. This effect can be significant in some materials,15 and usually increases with the
depth of the crater. However, in previous work on nitrided stainless steel, many the sputter
craters were examined with a stylus profilometer;11 all showed reasonably flat floors.
A third explanation for the shape of the depth profile in Fig. 2a may lie in impact-driven
rearrangement of the elemental distribution caused by the Cs+ beam, in particular by
knocking nitrogen atoms further into the sample. A TRIM calculation gives the range of
5.5 keV Cs+ ions in stainless steel as 50 nm. Hence, the material just beneath the floor of
the sputter crater will be homogenised to about that depth. This represents more than just
an effective instrumental resolution profile to be convoluted with the actual depth profile;
it provides a source of nitrogen, perhaps only slowly diluted, for implanting further into
the sample. Whatever the explanation of the difference in shape between Figs 2a and 2b,
the fact of the difference must be borne in mind when interpreting a SIMS depth profile.
192

3.2
COMPARISON OF THE TECHNIQUES
Like SIMS, EDX microanalysis produces straightforwardly interpretable results, but unlike
SIMS, it seems to be relatively unaffected by variations in calibration factors. However,
EDX has several features limiting its application. First, it requires the preparation of a cross
section of the sample, which involves substantial machining and polishing. Perhaps this
might alter the composition of the surface layer under observation, although the degree of
consistency with the other methods suggests not for the case of stainless steel. Secondly, the
depth resolution is limited to the size of the beam spot. Thus, it is not applicable to samples
with thin nitrided layers, and finely resolved depth profiles cannot be obtained. Thirdly,
because of interference with L lines from Fe and the decrease of X-ray yield with Z,
assaying for nitrogen is very difficult and oxygen is impossible. The last limitation is, of
course, applicable mainly to iron and steel samples.
HI-ERDA is also little affected by composition-dependent variations in its calibration
factors. The use of energetic very heavy ions, such as 200 MeV gold ions, allows the
simultaneous detection of all elements in the periodic table with comparable sensitivity. Light
elements (Z <14) are fully separated from their neighbours in the periodic table, unlike with
Rutherford backscattering. HI-ERDA also can profile across the whole treated layer with
reasonable depth resolution.29 The total depth probed is determined not only by the energy
and atomic number of the incident species, but also by the mass and atomic number of the
recoils, since these must escape from the sample to be detected. The straggling of species
emerging from depth coupled with the energy resolution of the detector limits the range of
energies over which particle identification is possible. The limitation is readily apparent in
Fig. 6: as the recoil energy E falls, a point is reached where the signatures of the different
species merge. In the present case, the depth accessible for Cr, Fe and Ni is limited to 0.5
m because they are so close in Z, and the intensity of these groups completely obliterates
Mn. For nitrogen, the profiling depth is mainly dependent on the range of gold nuclei in
stainless steel. Gold ions with a bombarding energy of 241 MeV can access layers up to 1
m thick. This is somewhat of a restriction in examining nitrided stainless steel; samples with
thicker layers can be assayed but not profiled over the whole layer. In contrast, the only
limitation on the thickness of the layer that can be profiled by SIMS is the length of time that
one has access to the machine, although the effects of roughening of the crater floor may
impose limitations on the composition resolution at large depths.
Of techniques employed here, HI-ERDA has the highest sensitivity for hydrogen. With
SIMS, there is the difficulty of resolving HCs+ from the very intense Cs+ group, and yields
of H+ and H are low and likely to suffer from significant matrix effects. None of the other
techniques has any sensitivity to hydrogen at all.
XPS can provide high-resolution depth profiles, but practical considerations limit this to
the outermost few 100s of nm. As discussed in the next paragraph, calibration factors seem
to be less reliable for XPS than for EDX. On the other hand, XPS provides unique
information on the chemical state of the atoms in the sample.
Table 3 collates assays of bulk material from the techniques that provide this information and compares this with the specifications of AISI 316. (The official specifications are
193


Table 3 Official composition of AISI 316 as specified in wt %, with a conversion to at. %, compared
with results from locations in the bulk material obtained with techniques used in this work (balance: Fe).
Specification*
Element
H
C
O
Si
Cr
Mn
Ni
Mo
(wt %)
(at. %)
<0.08
<0.37
<1.00
16.018.0
<2.00
10.014.0
2.03.0
<1.98
17.019.3
<2.03
9.413.4
1.11.8
EDX
HI-ERDA
(at. %)
(at. %)
18 C
210 C
2.9
48.6
2.7
3.9
0.50 0.10
17.3 1.0
14.2
16.7
9.3 1.0
1.30 0.13
4.3
1.1
10.0
3.1
0.08 0.02
0.40 0.10
0.40 0.05
17.10 0.14
1.41 0.06
9.56 0.19
2.39 0.14
XPS (at. %)
*In addition, P < 0.045 wt %, S < 0.030 wt %.

The HI-ERDA and XPS values were measured on an untreated sample.
The uncertainty in the values is 0.2 percentage points. The C values refer to the delivered charge from the Ar+ sputtering
beam; the sputter rate is of order 0.2 nm C1.
given as wt %; these were converted to at. % by taking into account the extremes of all the
ranges.) By and large, the EDX, HI-ERDA and deeper (210 C) XPS results agree within
their uncertainties except for Mo. As noted in 2.6, Mn was not resolved from Cr and Fe
in the HI-ERDA, and so its contribution is included in those elements. The near-surface
(18 C) XPS results shows the composition of the oxidised region at the surface. Elevated
concentrations of carbon and oxygen in this region were also seen with HI-ERDA, as
shown for carbon in Fig. 7a, although the values were different (5 at.% for C and 18 at.%
for O). This probably reflects different oxidation responses of the two samples concerned.
The situation with Mo is interesting: only HI-ERDA give a result in agreement with the
specifications of AISI 316. This may possibly indicate calibration problems in EDX and
XPS for this element. In the EDX analysis, the concentration of Mo is derived from the
areas of the unresolved L-line multiplet, whereas well resolved K lines are used for the other
elements. In XPS, the Mo results rely on the analysis of the third-shell photoelectrons,
whereas all the other elements use electrons from the second shell. Perhaps the discrepancies are to be found in these distinctions.
Table 4 presents assays of the treated layer. The EDX results are must be rescaled before
comparison with the others, because of the insensitivity to N. As presented, they are
directly comparable with the assays of the bulk. It is evident that the relative proportions of
the main alloying elements are the same in the treated layer as in the bulk. The differences
in the C, N and O concentrations between HI-ERDA and the 210-C XPS results may
simply reflect real differences in the two samples examined. The differences in Cr, Ni and
Mo are more difficult to rationalise. XPS suggests that there is some fractionation of the
metals in the outermost regions of the treated layer; perhaps this can account for the
observations. However, XPS depth profiling also relies on sputtering, and so experiences
many of the problems of SIMS: roughening of the crater floor, preferential sputtering and
sputter-driven implantation. Whatever the cause, comparison of the results in Tables 3 and
194


Table 4 Compositions measured in the treated layer (at. %, balance: Fe).
XPS
Element
H
C
N
O
Si
Cr
Mn
Ni
Mo
EDX
0.41 0.10
16.82 0.17
1.26 0.09
9.76 0.33
2.40 0.22
HI-ERDA
0.08 0.02
0.80 0.05*
27.9 1.2*
0.16 0.02
0.39 0.05
12.4 0.7
6.7 0.7
0.90 0.10
18 C
210 C
4.5
32.6
15.8
2.7
19.6
2.1
25.2
10.7
2.5
4.9
14.4
1.8
*Peak values.
The uncertainty in the XPS values is 0.2 percentage points. The C values refer to the
delivered charge from the Ar+ sputtering beam; the sputter rate is of order 0.2 nm C1.
4 gives the clear impression of greater scatter in the XPS results, especially when compared
with claimed uncertainties, than in the results from other methods.
3.3
ELEMENTAL DEPTH PROFILE FOR NITRIDED AISI 316 STAINLESS STEEL
We have taken SIMS depth profiles on dozens of samples; they all show the same features:
reduced ion-current ratios for Mo and Cr, carbon piled up at the interface between the
treated layer and the bulk and a characteristic shape to the nitrogen profile. On the
evidence presented here the variations observed for Mo and Cr do not correspond to
varying compositions of these species: the ratios [Cr]:[Fe]:[Mo] are the same in the treated
layer as in the bulk. However, the shapes of the carbon and nitrogen profiles can be taken
seriously. To set these on an absolute scale, they must be calibrated at some point on the
profile. For nitrogen, this can be done by analysis of implanted standard samples, by HIERDA or some other technique. In fact, for the sample shown in Figs 1 and 4a, it proved
most convenient to use some old results from proton backscattering.
Proton backscattering (pBS) is essentially the same as Rutherford backscattering, but we
used bombarding energies well above the Coulomb barrier because of the thickness of the
nitrogen-rich layer. Also for this reason, we choose protons rather than the often-used He+
ions. Details of the method and representative spectra have been published
elsewhere.7,22,30 The spectra were analysed with the computer program RUMP19 and
checked with the computer program SIMNRA.31 The observed spectra can be reproduced
using no more than four layers with constant nitrogen concentration in each layer. This
rather crude representation of the depth profile is the best that can be obtained from the
data. However, the nitrogen concentration at the surface is well determined; the value
obtained is 22.0 0.5 at. % and this provides a means of calibrating the SIMS sputter
profile.
195

Fig. 11 Elemental depth profile for the same sample as in Figs 1 and 4a constructed as described in
the text. Uncertainties are shown at selected locations on the depth profiles; in some cases these are
smaller than the symbols.
The result is shown in Fig. 11, together with the results for the other elements. The
profile for carbon is from SIMS, normalised to the concentration in the bulk determined
by HI-ERDA (Table 3). The constant hydrogen profile is also from HI-ERDA. The
concentrations of Si, Cr and Ni in the bulk material are taken as the means of the values
from HI-ERDA in Table 3 and EDX in Tables 3 and 4. The value for Mn comes from
EDX alone and, in view of the doubts about the EDX results for Mo, the Mo value is from
HI-ERDA alone. The concentrations of all these elements in the treated layer are determined by simply scaling to allow for the carbon and nitrogen profiles.
As can be seen from Fig. 11, the nitrogen concentration is not uniform across the treated
layer. It is highest at the surface, falling gradually to about half of the surface value at the
interface with the bulk. After this point, the nitrogen concentration decays exponentially
into the bulk. This non-error-function behaviour has been widely observed with a variety
of profiling techniques.1,25,3248 It has been described with a trapping model, in which
nitrogen is trapped in the treated layer, presumably by chemical binding with Cr.36 The
treated layer is then viewed as a region of occupied trap sites through which further
nitrogen can diffuse quickly to the interface. That is, the nitrogen diffusion coefficient
varies with nitrogen concentration.33,36 Our observation of a chemical association between
nitrogen and Cr, Mo and probably Fe supports this view. Recent work using 14N and 15N
as successive nitriding agents46 has required revision of the model to include the possibility
of nitrogen escape from shallow traps.49 This could find justification in the range of
binding energies observed in our XPS work, with nitrogen being at most only weakly
bound to Fe.
The displacement of carbon by the incoming nitrogen is a ubiquitous feature of our
depth profiles. This behaviour has been observed in nitriding of stainless steel by plasmaimmersion ion implantation,34 in a pulsed dc glow35 and by ion implantation with a
nitrogen beam.41
196

4.
CONCLUSIONS
This work seeks to establish a method of placing an absolute concentration scale on SIMS
elemental depth profiles of nitrided stainless steel. The task was separated into two parts:
establishing the reliability of relative measurements of concentration (the shape of the
depth profile) and then obtaining a reliable absolute concentration at one point in the
profile. From a combination of the use of standard samples prepared by nitrogen-ion
implantation and comparisons of assays and depth profiles from SIMS, EDX, HI-ERDA
and XPS, we conclude that there are negligible matrix effects in the SIMS sputter yields of
nitrogen and carbon, but that the yields of MoCs+ and CrCs+ ions are significantly affected
by the concentration of nitrogen, with the matrix effect being most marked for Mo. Thus,
the shape of the nitrogen and carbon depth profiles from SIMS can be taken at face value.
HI-ERDA gives depth profiles of the same quality as SIMS, but is limited to a relatively
thin nitrogen-containing layer.
For the metals in the AISI 316 alloy, results from EDX, HI-ERDA and XPS show that
their relative concentrations averaged over the treated layer are the same as in the bulk
material. Hence, the depth profile of these elements, and also Si, can be obtained by simply
scaling the values in the bulk to allow for the dilution caused by the presence of nitrogen
and, to a much lesser extent, carbon. The SIMS carbon profile is placed on an absolute
scale by using the concentration in the bulk measured by HI-ERDA. For the nitrogen
profile, its absolute value could be determined with ion-implanted standards, by HI-ERDA
if the treated layer is thin enough or, as in the present work, by using the surface nitrogen
concentration determined from proton backscattering.
Detailed examination of the peak positions and widths in XPS leads to the conclusion
that there is a clear chemical association between the nitrogen in the treated layer and Cr
and Mo, and none between nitrogen and Ni. The case of iron is less certain, but the weight
of evidence supports an NFe association with quite low binding energy. These results
support the model of the kinetics of nitriding in austenitic stainless steel, which views the
nitrogen as residing in traps shallow enough for atom exchange with incoming nitrogen.
ACKNOWLEDGEMENTS
The authors gladly acknowledge the contributions of the late Professor T.R. Ophel to this
work. Had he lived to see the drafting of this paper, he would have been a co-author. This
research was supported by the Australian Institute of Nuclear Science and Engineering, the
Australian Research Council and by an Australian Postgraduate Research Award.
REFERENCES
1. R. Wei, J.J. Vajo, J.N. Matossian, P.J. Wilbur, J.A. Davis, D.L. Williamson and G.A. Collins,
A Comparative Study of Beam Ion Implantation, Plasma Ion Implantation and Nitriding of
AISI 304 Stainless Steel, Surf. Coat. Technol., 1996, 83, 235242.
2. A. Ricard, The Production of Active Plasma Species for Surface Treatments, J. Phys., 1997,
D30, 22612269.
197


3. D.L. Williamson, J.A. Davis and P.J. Wilbur, Effect of Austenitic Stainless Steel Composition
on Low-Energy, High-Flux, Nitrogen Ion Beam Processing, Surf. Coat. Technol., 1998, 103
104, 178184.
4. G.A. Collins, R. Hutchings, K.T. Short and J. Tendys, Ion-Assisted Surface Modification by
Plasma Immersion Ion Implantation, Surf. Coat. Technol., 1998, 103104, 212217.
5. T. Czerwiec, N. Renevier, H. Michel, Low Temperature Plasma-Assisted Nitriding, Surf.
Coat. Technol., 2000, 131, 267277.
6. G.A. Collins, R. Hutchings, K.T. Short, J. Tendys and C.H. Van der Valk, Development of a
Plasma Immersion Ion Implanter for the Surface Treatment of Metal Components, Surf. Coat.
Technol., 1996, 84, 537543.
7. M.J. Baldwin, G.A. Collins, M.P. Fewell, S.C. Haydon, S. Kumar, K.T. Short and J. Tendys,
The Low-Pressure Rf Plasma as a Medium for Nitriding Iron and Steel, Jpn. J. Appl. Phys.,
1997, 36, 49414948.
8. J.M. Priest, M.J. Baldwin, M.P. Fewell, S.C. Haydon, G.A. Collins, K.T. Short and J. Tendys,
Low Pressure Rf Nitriding of Austenitic Stainless Steel in an Industrial-Style Heat-Treatment
Furnace, Thin Solid Films, 1999, 345, 113118.
9. S. Kumar, M.J. Baldwin, M.P. Fewell, S.C. Haydon, K.T. Short, G.A. Collins and J. Tendys,
The Effect of Hydrogen on the Growth of the Nitrided Layer in Rf-Plasma-Nitrided Austenitic Stainless Steel AISI 316, Surf. Coat. Technol., 2000, 123, 2935.
10. M.P. Fewell, J.M. Priest, M.J. Baldwin, G.A. Collins and K.T. Short, Nitriding at Low
Temperature, Surf. Coat. Technol., 2000, 131, 284290.
11. M.J. Baldwin, S. Kumar, J.M. Priest, M.P. Fewell, K.E. Prince and K.T. Short, PlasmaNitrided AISI-316 Stainless Steel Examined by Scanning Electron Microscopy and Secondary
Ion Mass Spectrometry, Thin Solid Films, 1999, 345, 108112.
12. H. Gnaser, Improved Quantification in Secondary-Ion Mass Spectrometry Detecting MCs+
Molecular Ions, J. Vac. Sci. Technol., 1994, A12, 452456.
13. W. Bock, H. Gnaser, H. Oechsner, Secondary-Neutral and Secondary-Ion Mass Spectrometry
Analysis of TiN-Based Hard Coatings: An Assessment of Quantification Procedures, Anal.
Chim. Acta, 1994, 297, 277283.
14. H. Gnaser: Correlations in Secondary-Ion Yields from Cs-Implanted Semiconductors, Surf.
Sci., 1995, 342, 319326.
15. R.G. Wilson, F.A. Stevie and C.W. Magee, Secondary Ion Mass Spectrometry: A Practical
Handbook for Depth Profiling and Bulk Impurity Analysis, Wiley Interscience, 1989.
16. H. Timmers, R.G. Elliman, G.R. Palmer, T.R. Ophel and D.J. OConnor, The Development
of a Facility for Heavy-Ion Eleastic-Recoil Detection Analysis at the Australian National
University, Nucl. Instr. Meth., 1998, B136138, 611615.
17. T.R. Ophel, H. Timmers and R.G. Elliman, Non-Linearities and Entrance Window Effects in
Large Solid-Angle Gas Ionisation Detectors, Nucl. Instr. Meth., 1999, A423, 381393.
18. H. Timmers, T.R. Ophel and R.G. Elliman, Simplifying position-sensitive gas-ionisation
detectors for heavy ion elastic recoil detection, Nucl. Instr. Meth., 2000, B161163, 1928.
19. L.R. Doolittle, Algorithms for the Rapid Simulation of Rutherford Backscattering Spectra,
Nucl. Instr. Meth., 1985, B9, 344351; http://www.genplot.com/doc/rump.htm.
20. R.G. Elliman, H. Timmers, T.R. Ophel, T.D.M. Weijers, L.S. Wielunski and G.L. Harding,
Simultaneous Hydrogen Detection with an ERD Gas Ionisation Detector, Nucl. Instr. Meth.,
2000, B1613, 2314.
21. T.D.M. Weijers, K. Gaff, H. Timmers, T.R. Ophel and R.G. Elliman, Heavy-Ion ElasticRecoil Detection Analysis of Doped-Silica Films for Integrated Photonics, Nucl. Instr. Meth.,
2000, B161163, 624628.
22. J.M. Priest, M.J. Baldwin and M.P. Fewell, The Action of Hydrogen in Low-Pressure RfPlasma Nitriding, accepted for publication in Surf. Coat. Technol.
23. C.D. Wagner, A.V. Naumkin, A. Kraut-Vass, J.W. Allison, C.J. Powell and J.R. Rumble Jr,
NIST X-ray Photoelectron Spectroscopy Database, NIST Standard Reference Database 20, Ver.
3.0, http://srdata.nist.gov/xps/.
198


24. I.L. Singer and J.S. Murday, Chemical State of Ion-Implanted Nitrogen in Fe18Cr8Ni Steel,
J. Vac. Sci. Technol., 1980, 17, 327329.
25. E. Menthe, K.-T. Rie, J.W. Schultze and S. Simson, Structure and Properties of PlasmaNitrided Stainless Steel, Surf. Coat. Technol., 1995, 7475, 412416.
26. G.S. Chang, J.H. Son, S.H. Kim, K.H. Chae, C.N. Wang, E. Menthe, K.-T. Rie and Y.P. Lee,
Electronic Structures and Nitride Formation on Ion-Implanted AISI 304L Austenitic Stainless
Steel, Surf. Coat. Technol., 1999, 112, 291294.
27. M.P. Fewell, D.R.G. Mitchell, J.M. Priest, K.T. Short and G.A. Collins, The Nature of
Expanded Austenite, Surf. Coat. Technol., 2000, 131, 300306.
28. J.F. Zeigler, J.P. Biersack and U. Littmark, The Stopping and Range of Ions in Matter, 2nd edn,
Pergamon, 1996; http://www.research.ibm.com/ionbeams/.
29. R.G. Elliman, H. Timmers, G.R. Palmer and T.R. Ophel, Limitations to Depth Resolution in
High-Energy Heavy-Ion Elastic-Recoil Detection Analysis, Nucl. Instr. Meth.,1998, B136
138, 649653.
30. M.J. Baldwin, M.P. Fewell, S.C. Haydon, S. Kumar, G.A. Collins, K.T. Short and J. Tendys,
Rf-Plasma Nitriding of Stainless Steel, Surf. Coat. Technol., 1998, 98, 11871191.
31. http://ibaserver.physics.isu.edu/sigmabase//programs/simnra44.html.
32. K. Ozybaysal and O.T. Inal, Structure and Properties of Ion-Nitrided Stainless Steels, J.
Mater. Sci., 1986, 21, 43184326.
33. A. Saker, Ch. Leroy, H. Michel and C. Frantz, Properties of Sputtered Stainless Steel
Nitrogen Coatings and Structural Analogy with Low Temperature Plasma Nitrided Layers of
Austenitic Steels, Mater. Sci. Eng., 1991, A140, 702708.
34. M. Samandi, B.A. Sheddon, D.I. Smith, G.A. Collins, R. Hutchings and J. Tendys, Microstructure, Corrosion and Tribological Behaviour of Plasma Immersion Ion-Implanted Austenitic Stainless Steel, Surf. Coat. Technol., 1993, 59, 261266.
35. A. Leyland, D.B. Lewis, P.R. Stevenson and A. Matthews, Low Temperature Plasma Diffusion
Treatment of Stainless Steels for Improved Wear Resistance, Surf. Coat. Technol., 1993, 62,
608617.
urk, R. Wei and P.J. Wilbur, Metastable Phase Formation and
36. D.L. Williamson, O. Ozt
Enhanced Diffusion in F.c.c. Alloys under High Dose, High Flux Nitrogen Implantation at
High and Low Ion Energies, Surf. Coat. Technol., 1994, 65, 1523.
37. F. Bodart, Th. Briglia, C. Quaeyhaegens, J. DHaen and L.M. Stals, Investigation of Nitride
Phases in Stainless Steels by Nuclear Reaction Analysis and X-Ray Diffraction, Surf. Coat.
Technol., 1994, 65, 137141.
urk and D.L. Williamson, Phase and Composition Depth Distribution Analyses of Low
38. O. Ozt
Energy, High Flux N Implanted Stainless Steel, J. Appl. Phys., 1995, 77, 38393850.
39. R. Wei: Low Energy, High Current Density Ion Implantation of Materials at Elevated Temperatures for Tribological Applications, Surf. Coat. Technol., 1996, 83, 218227.
40. S.J. Bull, A.M. Jones and A.R. McCabe, Improving the Mechanical Properties of Steels using
Low Energy, High Temperature Nitrogen Ion Implantation, Surf. Coat. Technol., 1996, 83,
257262.
41. C. Blawert, A. Weisheit, B.L. Mordike and F.M. Knoop, Plasma Immersion Ion Implantation
of Stainless Steel: Austenitic Stainless Steel in Comparison to AusteniticFerritic Stainless
Steel, Surf. Coat. Technol., 1996, 85, 1527.
42. S. Leigh, M. Samandi, G.A. Collins, K.T. Short, P. Martin and L. Wielunski, The Influence of
Ion Energy on the Nitriding Behaviour of Austenitic Stainless Steel, Surf. Coat. Technol., 1996,
85, 3743.
43. K. Marchev, C.V. Cooper, J.T. Blucher and B.C. Giessen, Conditions for the Formation of a
Martensitic Single-Phase Compound Layer in Ion-Nitrided 316L Austenitic Stainless Steel,
Surf. Coat. Technol., 1998, 99, 225228.
44. S. Mandl, R. Gunzel, E. Richter and W. Moller, Nitriding of Austenitic Stainless Steels using
Plasma Immersion Ion Implantation, Surf. Coat. Technol., 1998, 100101, 372376.
199


45. E. Camps, S. Muhl, S. Romero and J.L. Garcia, Microwave Plasma Nitrided Austenitic
AISI-304 Stainless Steel, Surf. Coat. Technol., 1998, 106, 121128.
46. S. Parascandola, R. Gunzel, R. Grotzschel, E. Richter and W. Moller, Analysis of Deuterium
Induced Nuclear Reactions Giving Criteria for the Formation Process of Expanded Austenite,
Nucl. Instr. Meth., 1998, B136138, 12811285.
47. E. Menthe and K.-T. Rie, Further Investigation of the Structure and Properties of Austenitic
Stainless Steel after Plasma Nitriding, Surf. Coat. Technol., 1999, 116119, 199204.
48. B. Larisch, U. Brusky and H.-J. Spies, Plasma Nitriding of Stainless Steels at Low Temperatures, Surf. Coat. Technol., 1999, 116119, 205211.
49. D.L. Williamson, private communication.
200

Stainless Steel 2000

Uploaded by

Copyright:

Available Formats

Stainless Steel 2000

Uploaded by

Document Information

Original Description:

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

Stainless Steel 2000

Uploaded by

Copyright:

Available Formats

SEQ 0003 JOB IOM8420-000-020 PAGE-0003 PRELIMS

REVISED 18OCT01 AT 17:45 BY TF DEPTH: 62.04 PICAS WIDTH 45.03 PICAS