See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257707089

Microstructural and Phase Evolution of Compound Layers Growing on α–Iron

During Gaseous Nitrocarburizing

Article in Metallurgical and Materials Transactions A · July 2012

DOI: 10.1007/s11661-012-1100-1

CITATIONS READS

17 286

3 authors, including:

Andreas Leineweber
Technische Universität Bergakademie Freiberg
205 PUBLICATIONS 2,136 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Synthesis and Properties of Zinc oxide (ZnO) View project

Intermetallic Phases View project

All content following this page was uploaded by Andreas Leineweber on 05 June 2014.

The user has requested enhancement of the downloaded file.

Microstructural and Phase Evolution of Compound Layers
Growing on a–Iron During Gaseous Nitrocarburizing
T. WOEHRLE, A. LEINEWEBER, and E.J. MITTEMEIJER

The microstructural evolution of compound layers grown on 1-mm thick a-iron substrates after
nitrocarburizing at 853 K (580 C) in NH3/H2/N2/CO gas mixtures was investigated by light
optical microscopy and X-ray diffraction. The evolution of the microstructure can be divided
into several stages. Starting with the formation of the carbon-rich phase cementite, which
practically does not contain nitrogen, the phase constitution of the compound layer develops
through successive stages of microstructural change into the direction of the nitrogen-richer and
carbon-poorer phases e and c’. These results are the consequences of (1) the kinetics of nitrogen
and carbon uptake at the gas–solid interface and the considerably different solubilities of
nitrogen and carbon in the a-iron substrate and (2) the occurrence of local equilibrium condi-
tions prevailing at the solid–solid interphase boundaries in the compound layer. The change of
the microstructure as a function of depth in the compound layer is shown to be compatible with
so-called ‘‘diffusion paths’’ in the ternary Fe-N-C phase diagram.

DOI: 10.1007/s11661-012-1100-1
 The Minerals, Metals & Materials Society and ASM International 2012

I. INTRODUCTION corresponding phase diagram.[3,4,6,7] At a given temper-

ature and pressure, the diffusion path is defined by the
DIFFUSION-cONTROLLED phase transforma- course of the average composition in a direction
tions in multicomponent solid-state systems are of great perpendicular to the original interface of the diffusion
technological importance in a wide range of applications couple.[4,8] The course of a diffusion path is governed by
including diffusion bonding, solder joints, welded alloys, both the thermodynamics and the kinetics of the
microelectronic devices, and the thermochemical surface interacting components of the system considered.
treatments. All these technologies are based on the In the field of thermochemical surface treatments, an
modification and (meta)stability of microstructures important case of (usually) diffusion-controlled phase
induced by directional long-range diffusion and chem- transformations is met in nitrocarburizing, where nitro-
ical reactions at the surface and/or the interfaces of gen and carbon are supplied simultaneously to the
intimately connected phases; therefore, the properties of surface of ferritic iron-based workpieces at temperatures
multicomponent materials can be varied greatly.[1,2] The usually between 773 K and 853 K (500 C and
driving forces for diffusion in such heterogeneous 580 C).[9–11] If the chemical potentials of nitrogen and
systems are the gradients of the chemical potentials of carbon imposed by the nitrocarburizing agent are
the components in the reaction zone. If the rate of sufficiently high, a compound layer typically composed
diffusion is slow compared with the rate of reaction, i.e., of iron (carbo)nitrides such as e-Fe3(N,C)1 + x and c’-
the rate of reaction is controlled by the rate of diffusion, Fe4(N,C)1-z develops ([possibly] eventually; see the
so-called local thermodynamic equilibria can be results presented subsequently) at the surface of the
assumed to prevail at the solid–solid phase interfaces workpiece.[12–14] Under certain nitrocarburizing condi-
in the reaction zone.[3–5] The establishment of local tions, also h-Fe3C (cementite) can form as an additional
equilibria implies a continuous variation of the chemical phase[11] or even as the only phase[15] in the compound
potentials of the components through the reaction zone, layer. The composition ranges of the relevant Fe-N-C
associated with the possible formation of (one or more) phases are given in Table I. Note that in particular, the e
phases within the reaction zone. The sequence and phase exhibits a wide range in composition.
composition of these phases then is described by a The compound layer (of thickness up to several
so-called diffusion path in (the isothermal section of) the 10 lm) that grew as a result of nitrocarburizing can lead
to a significant improvement of mechanical (resistance
to wear and friction) and chemical (resistance to
T. WOEHRLE, Ph.D. Student, and A. LEINEWEBER, Doctor corrosion) properties of the workpiece.[18,19] Underneath
and Staff Scientist, are with the Max Planck Institute for Intelligent
Systems (formerly the Max Planck Institute for Metals Research), the compound layer, the so-called diffusion zone forms,
Heisenbergstraße 3, D-70569 Stuttgart, Germany. Contact e-mail: where nitrogen and carbon are dissolved interstitially in
[email protected] E.J. MITTEMEIJER, Professor, is with the the ferritic matrix. The diffusion zone is usually up to
Max Planck Institute for Intelligent Systems (formerly the Max Planck several 100 lm thick and induces a considerable
Institute for Metals Research), and with the Institute for Materials
Science, University of Stuttgart, D-70569 Stuttgart, Germany.
improvement in fatigue resistance.[20,21]
Manuscript submitted June 7, 2011. Although several investigations on compound-layer
Article published online April 5, 2012 development on pure a-iron substrates after nitrocarbu-

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, JULY 2012—2401

rizing have been performed,[11,22–25] a systematic and nitrogen solubility in ferrite compared with that of
general description of the compound-layer evolution is carbon in ferrite is shown to have an important impact
still lacking, undoubtedly because of the complexity of on the microstructure of the developing compound layer.
the occurring microstructures that are far more compli-
cated than in the case of pure nitriding.
Previous studies on the microstructural evolution of II. EXPERIMENTAL
the compound layer growing on pure a-iron substrates
after salt-bath nitrocarburizing at 853 K (580 C)[22] and Rectangular specimens of dimensions 20 9 25 9
after gaseous nitrocarburizing at 848 K (575 C)[23] and 1 mm3 were cut from cold-rolled cast iron plates (Alfa
853 K (580 C)[24] reported the initial formation of Fe3C Aesar, Ward Hill, MA; 99.98 wt pct), ground, polished
(cementite) at the surface of the substrate, which was (final stage, 1-lm diamond suspension), cleaned ultra-
followed by the subsequent formation of the nitrogen- sonically in ethanol, and recrystallized in hydrogen for
rich e-phase. For prolonged treatment times, the com- 2 hours at 973 K (700 C), followed by slow cooling.
pound layer was composed only of (1) both e and c’[22,23] Immediately before nitrocarburizing the specimens were
or (2) c’.[24] The initial nucleation of cementite at the polished (final stage, 1-lm diamond suspension) and
substrate surface was explained as a consequence of a cleaned with ethanol.
much faster absorption rate of carbon than that of Nitrocarburizing experiments were performed at
nitrogen at the initial stage of nitrocarburizing, implying 853 K (580 C) in a vertical quartz-tube furnace mea-
that the solubility limit of carbon in ferrite is reached suring 28 mm in diameter. The process temperature was
faster than that of nitrogen.[22,23] controlled within ±1 K in the middle of the furnace,
The current work aims to investigate systematically, by where the specimen was placed. The nitrocarburizing
performing dedicated experiments with specifically pre- atmosphere was composed of NH3 (99.999 vol pct) as
pared specimens (Section II), the observed initial forma- nitriding species, CO (99.997 vol pct) as carburizing
tion of the carbon-rich phase cementite at the surface of species, and H2 (99.999 vol pct) and N2 (99.999 vol pct)
the a-iron substrate after nitrocarburizing as well as the as inert gases (all gases from Westfalen AG, Münster,
subsequent evolution of the compound layer. It will be Germany). The flow rate of each gas component was
shown that all different microstructures occurring suc- adjusted separately by mass-flow controllers. To avoid
cessively within the developing compound layer (Section significant (thermal) dissociation of ammonia and pos-
III) can be understood on the basis of local thermody- sible side reactions in the gas atmosphere (but see
namic equilibria at the interfaces between the solid phases footnote to Table II), which would affect the composi-
formed (cf. Section IV). In particular, the relatively high tion of the gas mixture, a constant overall gas-flow rate

Table I. Composition Ranges and Crystal Structure of Some Relevant Phases in the Ternary System Fe-N-C at a Temperature
of 853 K (580 °C)[16,17]

Phase Crystal Structure N Content (at. pct) C Content (at. pct)

a-Fe Fe: Body-centered-cubic type, N and C in octahedral interstices 0 to 0.37 0 to 0.02
c’-Fe4N1–z* Fe: Face-centered cubic-type, N ordered in octahedral interstices 19.4 to 20 <0.7
e-Fe3(N,C)1 + x* Fe: Hexagonal close-packed-type, N ordered in octahedral interstices 15 to 33 0 to 8
h-Fe3C* Fe: Complicated orthorhombic, C in bicapped trigonal prisms ~0 25
*These phases are metastable with respect to the decomposition into iron and nitrogen and/or carbon.

.
3=2
Table II. Composition of the Nitrocarburizing Atmosphere, the Corresponding Nitriding Potential* rN ¼ PNH3 PH2
and Treatment Time Used for Experiments of Series A, B, and C at 853 K (580 °C)

Series Treatment Time (h) NH3 (vol pct) H2 (vol pct) CO (vol pct) N2 (vol pct) rN (atm1/2)
A 4 13.23 57.95 20 8.82 0.30
4 15.44 57.95 20 6.61 0.35
4 17.64 57.95 20 4.41 0.40
4 22.05 57.95 20 – 0.50
B 0.25 15.44 57.95 20 6.61 0.35
0.5 15.44 57.95 20 6.61 0.35
2 15.44 57.95 20 6.61 0.35
4 15.44 57.95 20 6.61 0.35
8 15.44 57.95 20 6.61 0.35
24 15.44 57.95 20 6.61 0.35
C 4 13.23 57.95 20 8.82 0.30
*The nitriding potential is related to the chemical potential of nitrogen in a NH3/H2-containing gas mixture,[12] and it determines at a given
temperature the iron-nitride phase in equilibrium with a purely nitriding atmosphere. In contrast to the nitriding potential, no defined carburizing
potential can be given for the NH3/H2/CO/N2 gas mixtures as applied in the current work because the carburizing potential is hypothetically
infinite.[12,15] However, an ‘‘effective’’ carburizing potential prevails on nitrocarburizing because of the side reactions in the gas atmosphere.[26]

2402—VOLUME 43A, JULY 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

of 13.5 mm seconds1 through the quartz tube (calcu- 1 g NaOH, and 1 g KOH per 100 mL H2O) at 333 K
lated for the gas volume at room temperature) was (60 C), which allows phase identification by discrimi-
chosen. After nitrocarburizing, the specimens were native staining of the carbon-containing phases
quenched by dropping them out of the furnace into e-Fe3(N,C)1 + x and, in particular, cementite (Fe3C).
nitrogen-flushed water. The c’ phase remains unstained because of the low
To investigate the microstructural evolution of the solubility of carbon, as indicated in Table I. Optical
compound layer after nitrocarburizing, two series of microscopy was performed on cross sections of speci-
experiments were performed in NH3/H2/N2/CO gas mens using a Zeiss Axiophot microscope (Carl Zeiss,
mixtures at 853 K (580 C): (1) using a constant Oberkochen, Germany).
treatment time of 4 hours and a variable gas composi- X-ray diffractograms for identification of the phases
tion (cf. Table II; series A) and (2) using a variable present in the compound layer were recorded from the
treatment time and a constant gas composition (cf. surface region of the compound layers using a PANa-
Table II; series B). lytical X’Pert Multi-Purpose Diffractometer (PANalyt-
Additionally, experiments were performed using ical, Almelo, the Netherlands) equipped with a graphite
wedge-shaped specimens obtained by dedicated grinding monochromator in the diffracted beam (Co-Ka radia-
of original 20 9 25 9 1 mm3 specimens, which led to a tion) and Bragg-Brentano geometry. To obtain better
linear thickness variation from 1 mm to 0.02 mm. Using crystallite statistics, the samples were rotated around
such specimens and the normal rectangular their surface normal during the measurement.
20 9 25 9 1 mm3 specimens that were presaturated Microhardness measurements (micro-Vickers) on the
with nitrogen, the influence of nitrogen saturation of specimen cross section were carried out in the ferrite
the ferrite substrate on the microstructural evolution of substrate directly underneath the compound layer using
the compound layer could be revealed. The wedge- a Leica VMHT MOT micro-indentation system (Leica
shaped specimens as well as the nitrogen-presaturated AG, Wetzlar, Germany) with a load of 50 gf (ffi0.49 N)
rectangular specimens were nitrocarburized in a (rela- and a dwell time of 10 seconds (higher loads occasion-
tively) strongly carburizing atmosphere (cf. Table II; ally caused localized cracking of the compound layer).
series C).* The nitrogen presaturation was achieved by Each hardness value presented is an average value of
four measurements. The nitrogen content in ferrite (not
well measurable with electron probe microanalysis; note
*The nitrocarburizing atmosphere used for series C leads to the the small nitrogen-solubility limit in ferrite of approx-
formation of a single-phase cementite layer on 1-mm-thick, rectangu-
lar, nonsaturated a-iron substrates.[15,26] imately 0.36 at. pct at 853 K [580 C][17]) was deter-
mined using a linear dependence of hardness on nitrogen
content in ferrite as reported in Reference 32. The same
nitriding of the rectangular iron specimens for 66 hours experimental conditions as described in Reference 32
at 853 K (580 C) in a NH3/H2 gas mixture containing were used in the current work to ensure the applicability
10.9 vol pct NH3 and 89.1 vol pct H2. These nitriding of the reported dependence of hardness on nitrogen
parameters had been chosen such that a maximum content in ferrite.
solubility of nitrogen in ferrite was obtained without
forming an iron-nitride compound layer.[27] After nitrid-
ing, X-ray diffraction was applied to verify that iron III. RESULTS
nitrides had not formed. The total nitrogen content in
the presaturated substrate was determined by weighing In the following, a shorthand notation for the
the specimen before and after nitriding[28] using a specification of the microstructure of the compound
Toledo UMX2 high-precision balance (Mettler-Toledo layer is introduced. For this notation, the compound
International, Inc., Columbus, OH). The nitrogen con- layer is divided into separate sublayers perpendicular to
tent in the presaturated samples was approximately the surface normal, where each sublayer comprises a
0.37 at. pct, which is close to the nitrogen-solubility laterally homogeneous phase constitution, at least over a
limit in ferrite at 853 K (580 C); these data are reported lateral distance of several 10 lm. In ternary systems (as
in the literature as approximately 0.36 to 0.38 at. the Fe-N-C system) at a constant temperature and
pct.[17,29] pressure, a sublayer comprising a (depth) range in
The nitrocarburized specimens were cut into two composition can be composed either of a single phase or
pieces, which were used for optical microscopy and of two different phases (single-phase or dual-phase
X–ray powder diffraction, respectively. sublayers) if the phase rule is obeyed. The overall phase
The specimens for optical microscopy analysis were constitution of the compound layer can then be denoted
covered with an electrodeposited nickel layer using a by the sequence of sublayers (separated by ‘‘/’’), starting
Watts bath[13] at 333 K (60 C) to prevent damage of the from the surface of the material and ending at the
compound layer during sample preparation. Embedding compound layer/substrate interface. For example, an
(Polyfast; Struers GmbH, Willich, Germany) was fol- e/e + c’ compound layer is composed of an e single-
lowed by grinding, polishing (final stage 1 lm diamond phase sublayer adjacent to the surface of the compound
suspension), and etching in 1 vol pct Nital containing layer followed by an e + c’ dual-phase sublayer adjacent
0.1 vol pct HCl.[30] Some cross sections were treated to the ferritic substrate. Note that no specific indication
additionally with a Groesbeck reagent[31] (4 g KMnO4, is used for the a-iron substrate.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, JULY 2012—2403

A. Influence of the Nitriding Potential locally confined h/e double-layers. After 2 hours, a
on the Compound-Layer Evolution; Series A laterally inhomogeneous microstructure was obtained:
Nitrocarburizing of pure a-iron substrates at the At some locations, the compound layer is composed of a
lowest applied nitriding potential of rN = 0.3 atm1/2 massive cementite layer (Figure 2(a)) that, proceeding
(series A) showed the formation of a closed cementite laterally, changes to a h/e double layer and to a nitrogen-
layer at the surface of the substrate (cf. Figure 1(a)), richer e/e + c’ double layer (Figure 2(b)). After contin-
which is in agreement with the experiments reported in ued nitrocarburizing (4 hours and 8 hours), the laterally
Reference 15. An increased nitriding potential inhomogeneous microstructure of the compound layer
(rN = 0.35 atm1/2) resulted in the development of an reveals an e/e + c’ double layer (Figure 2(c)) and also an
inhomogeneous microstructure.** Besides a monophase e + c’ dual-phase layer (Figure 2(d)). After long treat-
ment times (24 hours), the microstructure consisted of
an e + c’ dual-phase layer (Figure 2(e)) and also of a
**Inhomogeneous means that the phase constitution of the micro- single-phase c’ layer (Figure 2(f)).
structure and, thus, the composition in the compound layer varies as The preceding results show a complex time depen-
function of the lateral position on the specimen cross section. dence of the microstructure of the compound layer.
According to the current observations, the position/height of the
specimen in the furnace did not influence the distribution of the dif-
With increasing treatment time, the microstructure
ferent microstructures: The linear gas flow rates in the furnace are high evolves from a carbon-rich and virtually nitrogen-free
enough so that the effective chemical potentials of N and C are con- cementite layer to nitrogen-richer e + c’ and c’ layers.
stant along the surface of the specimen during the heat treatment. In both series (A and B), the development of the
microstructure of the compound layer proceeds from a
cementite layer adjacent to the surface, a complex dual- carbon-rich phase to nitrogen-rich phases. Whereas such a
phase microstructure of intrinsically nitrogen-poor development for series A seems to be linked with an
cementite and nitrogen-richer e is present in the com- increasing value of the imposed nitriding potential, it is
pound layer (cf. Figures 1(b) and (c)). A reproduction of unclear at this stage (but see what follows) why a change in
this experiment (rN = 0.35 atm1/2) resulted again in treatment time under constant nitrocarburizing conditions
the formation of an inhomogeneous microstructure invokes a similar microstructural development for series B.
showing, at different lateral positions, an e/e + c’double
layer (Figure 1(d)) and a h/e double layer (Figure 1(e)). C. Role of Nitrogen Saturation of the Substrate
At one location, the h sublayer had transformed into e in
association with a locally increased thickness of the An initially nonsaturated, a-iron substrate can
become saturated with nitrogen during nitrocarburizing,
compound layer (dashed arrow in Figure 1(e)). At
given sufficient treatment time, i.e., a (very) thin
higher nitriding potentials (rN = 0.4 atm1/2 and
rN = 0.5 atm1/2), the microstructure is composed of substrate is much faster saturated with nitrogen than a
a laterally homogeneous e/e + c’double layer. In partic- thick substrate (note that the carbon solubility in the
a-Fe substrate can be neglected; see Table I). Against
ular, the morphology of the compound layer generated
at the highest nitriding potential of rN = 0.5 atm1/2 this background, wedge-shaped substrates with a linear
approaches the well-known morphology of an e/c’dou- variation in thickness from 1 mm to 0.02 mm (cf.
Section II) were used for nitrocarburizing.
ble layer obtained after pure nitriding (Figure 1(g) and
Nitrocarburizing of wedge-shaped a-iron substrates for
see Reference 33).
Evidently, the previous results reveal a strong effect of 4 hours at 853 K (580 C) (cf. Table II, series C) showed
a strong variation in the microstructure of the compound
the nitriding potential on the microstructural evolution
layer as a function of the substrate thickness (Figure 3).
of the compound layer. With increasing nitriding
After nitrocarburizing, the thick end of the specimen is
potential (rN = 0.3 atm1/2 to rN = 0.5 atm1/2), cor-
covered with a massive cementite layer as shown in
responding with an increasing content of nitrogen in the
Figure 3(b) (in agreement with the result for the rectan-
compound layer, the microstructure evolves from a
gular 1-mm-thick, nonsaturated substrate nitrocarbu-
carbon-rich and virtually nitrogen-free cementite layer
rized under the same conditions; cf. series A, Figure 1(a)).
to a nitrogen-rich e/e + c’ double layer.
The formation of nitrogen-richer e at the cementite layer/
substrate interface had occurred close to the thin end of
B. Influence of the Treatment Time
the substrate, i.e., an h /h + e double layer occurs at this
on the Compound-Layer Evolution; Series B
location (Figure 3(c)). At the thin end of the substrate, a
Nitrocarburizing of pure a-iron substrates as a h/e double layer was obtained (Figure 3(d)). Evidently,
function of time was performed at 853 K (580 C) using the amount of e in the compound layer increases from the
a nitriding potential of rN = 0.35 atm1/2 (series B), i.e., thick end to the thin end of the substrate.
an intermediate value of rN with respect to the range of Nitrocarburizing of rectangular, 1-mm-thick, non-
rN values covered by the experiments of series A saturated and nitrogen pre-saturated a-iron substrates
(Section III–A). For short treatment times (15 minutes), for 4 hours at 853 K (580 C) (cf. Table II, series C) led
a closed intrinsically nitrogen-poor cementite layer had to the formation of strikingly dissimilar microstructures
developed at the surface of the substrate. Slightly longer of the compound layer. After nitrocarburizing of the
nitrocarburizing-times (30 minutes) induced the forma- nonsaturated substrate, a single-phase h layer was
tion of the nitrogen-richer e phase in some regions at the obtained (Figure 4(a) and see Figure 1(a)), whereas
cementite/substrate interface leading to the formation of after nitrocarburizing of the nitrogen presaturated

2404—VOLUME 43A, JULY 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 1—Optical micrographs showing cross sections (after etching with 1 vol pct Nital containing 0.1 vol pct HCl and staining with an alkaline
potassium permanganate solution) of compound layers grown on a-Fe substrates by nitrocarburizing for 4 h at 853 K (580 C) for various
nitriding potentials (series A): (a) cementite layer, rN = 0.30 atm1/2; (b) to (c) massive cementite layer and e/h + e double layer, rN = 0.35
atm1/2; (d) to (e) e/e + c’ double layer and h/e double layer, rN = 0.35 atm1/2. The dashed arrow in (e) indicates transformation of the h sub-
layer into e phase; (f) e/e + c’ double layer, rN = 0.4 atm1/2; (g) e/e + c’ double layer, rN = 0.5 atm1/2.

substrate, a h/e double layer had developed (Figure 4(b)). ture, a nitrogen content of approximately 0.35 at. pct
The corresponding X-ray diffraction patterns are given in occurs, which is practically equal to the nitrogen-
Figure 5. solubility limit in ferrite at 853 K (580 C) of
To determine the degree of nitrogen saturation in the approximately 0.36 at. pct.[17] Underneath the massive
a-iron substrate, in relation to the specific microstruc- cementite layer, a remarkably lower nitrogen content of
ture of the compound layer on top of the substrate, 0.25 at. pct occurs in the ferrite substrate. The lower
hardness measurements were performed in the diffusion nitrogen content underneath the cementite layer can be
zone, at positions laterally apart, directly underneath the ascribed to an inhibited diffusion of nitrogen through
compound layer, with a microstructure that varied from cementite compared with the diffusion of nitrogen
a carbon-rich cementite single layer to a nitrogen-rich e/ through iron (carbo)nitrides.[23,32,34]
e + c’ double-layer (cf. Figures 2(a) and (b); rectangular
specimen nitrocarburized in an 15.44 pct NH3,
57.95 pct H2, 20 pct CO, and 6.61 pct N2 gas mixture; IV. DISCUSSION
series B). The results are shown in Figure 6; the local
A. Kinetics: Origin of the Microstructural Evolution
nitrogen concentration in the ferrite substrate was
of the Compound Layer
calculated from the measured hardness values as
described in Section II. The hardness measurements As will be demonstrated in the following discussion,
show that in the ferrite substrate underneath the the time-dependent (series B) and atmosphere-depen-
compound layer, where e is present in the microstruc- dent (series A) evolutions of the microstructure of the

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, JULY 2012—2405

Fig. 2—Optical micrographs showing cross sections (after etching with 1 vol pct Nital containing 0.1 vol pct HCl) of compound layers grown on
a-Fe substrates by nitrocarburizing for various treatment times at 853 K (580 C) (series B): (a) to (b) cementite layer and e/e + c’ double layer
obtained after 2 h, (c) to (d) e/e + c’ double layer and e + c’ dual-phase layer obtained after 8 h, and (e) to (f) e + c’ layer and c’ layer obtained
after 24 h.

compound layer can be explained by assuming that local nitrogen and of 0.018 at. pct for carbon,[17] i.e., both
equilibria prevail at the solid–solid phase interfaces, solubilities differ by a factor of 20.
whereas solid–gas equilibrium does not necessarily The phase changes observed at the surface of the
prevail at the surface of the compound layer. The compound layer as function of time (series B, Figure 2)
time-dependent microstructural evolution of the com- indicate that at the surface, no ‘‘local equilibrium’’
pound layer can be caused by (1) the finite rates of occurs between the gas phase and the solid phase, at
nitrogen and carbon transfer from the gas phase to the least not for short and intermediate treatment times,
a-Fe substrate (before and after compound-layer for- which is discussed subsequently.
mation) and (2) the finite thickness of the plate-like The uptake of nitrogen and carbon from the nitro-
specimens used in the current experiments, implying a carburizing gas atmosphere in the ferrite substrate is
gradual saturation of the initially unsaturated substrate controlled by (1) the rate of NH3 and CO dissociation
after nitrocarburizing. As a consequence, in the case of a (possibly catalyzed) at the substrate surface, leading to
gas atmosphere of constant composition, the micro- adsorbed nitrogen and carbon atoms, (2) the rate of
structure of the compound layer will evolve as function recombination (and possibly desorption) of the
of treatment time. In the current case, the characteristic adsorbed atoms, and (3) the diffusion rates of, subse-
quently absorbed, nitrogen and carbon in the substrate
away from the surface. Finite values for the nitrogen
A time-invariant phase constitution (microstructure) of the com- and carbon uptake imply that a local equilibrium,à as
pound layer is expected to occur only for the case of a constant
nitrogen and carbon concentration at the surface of the compound
layer (implying local equilibrium or stationary state at the gas–solid à
interface) in combination with (1) a nitrogen and carbon saturated For example, if the recombination and desorption of adsorbed
substrate or (2) an infinitely thick substrate with initially constant nitrogen atoms from the specimen surface cannot be neglected com-
nitrogen and carbon content. In such cases (and assuming local equi- pared with the rate of NH3 dissociation, a stationary state instead of a
librium at the solid–solid interfaces in the compound layer), the local equilibrium occurs, i.e., the chemical potential of nitrogen in the
so-called Boltzmann transformation can be applied to Fick’s second solid at the surface of the specimen is lower than that pertaining to the
law, implying a time-invariant diffusion path (lateral gross composi- nitrocarburizing atmosphere.[12,36]
tion as a function of depth) and a parabolic growth of the compound-
layer thickness.[4,7,35]
prescribed by the chemical potentials of nitrogen and
carbon, as pertaining to the composition of the nitro-
evolution of phases in the compound layer can be carburizing atmosphere, cannot be established at once
described as a consequence of the large difference of the (i.e., at t = 0).
solubilities of nitrogen and carbon in the a-iron sub- The rate of carbon transfer from CO gas to ferrite has
strate (see Figure 7(b)) that affects the local solid–solid been found to be much faster than the rate of nitrogen
equilibrium at the compound layer/substrate interface. transfer from NH3 gas.[37] It can be concluded that the
As shown in the Fe-N-C phase diagram at 853 K relatively low solubility of carbon in ferrite (cf. Table I)
(580 C) (Figure 7(b)), the a-iron single-phase field in combination with the relatively fast uptake of carbon
exhibits a maximum solubility of 0.36 at. pct for (recognizing that the diffusivities of carbon and nitrogen

2406—VOLUME 43A, JULY 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 3—Optical micrographs showing cross sections (after etching with 1 vol pct Nital containing 0.1 vol pct HCl and staining with an alkaline
potassium permanganate solution) of compound layers obtained on a wedge-shaped a–iron specimen by nitrocarburizing for 4 h at 853 K
(580 C) using a 13.23 vol pct NH3, 57.95 vol pct H2, 20 vol pct CO, 8.82 vol pct N2, (rN = 0.30 atm1/2) gas mixture: (a) schematic illustration
of the wedge-shaped specimen showing the lateral positions of the optical micrographs shown in (b) to (d); (b) cementite layer at the thick end of
the sample; (c) h/h + e double layer close to the thin end of the sample; and (d) h/e double layer at the thin end of the sample.

in a-iron do not differ much[38]) causes the ferrite observation and interpretation along these lines for
substrate to be much faster saturated with carbon than the initial formation of cementite was presented in
with nitrogen at the early stage of nitrocarburizing.§ References 22 and 23.
After continued nitrocarburizing, the surface-adjacent
§
region of the substrate becomes enriched with nitrogen
Even in the case of finite but about equal transfer rates of nitrogen gradually by the continuous diffusion of nitrogen
and carbon, the solubility limit of carbon will be surpassed first as a
consequence of the large difference of the solubility limits of nitrogen through (the grain boundaries of) the cementite layer
and carbon. as discussed in detail in Reference 32. The nucleation of
e at the cementite layer/substrate interface is then
induced if the nitrogen concentration in ferrite at
Hence, at the surface of the substrate, the initial the layer/substrate interface exceeds a critical value
formation of nitrogen-poor cementite instead of a (Figure 7(b), point B): First, a local equilibrium of h, e,
nitrogen-richer (carbo)nitride can occur (Figure 7(b), and a and then, after further nitrogen enrichment of the
point A), although the presence of the latter phase at the substrate, a local equilibrium of a and e, prevails. Note
surface would comply with thermodynamic equilibrium that e can be formed also by the conversion of cementite,
with the gas atmosphere or with the establishment of a as demonstrated first in Reference 39.
stationary state (see footnoteà), as suggested by the The role of the substrate saturation with nitrogen on
phase constitution at the surface of the compound layer the microstructural evolution of the compound layer
after long treatment times (cf. Section III–B). The first is supported fully by the results obtained from the

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, JULY 2012—2407

Fig. 4—Optical micrographs (after etching with 1 vol pct Nital containing 0.1 vol pct HCl and staining with an alkaline potassium permanga-
nate solution) showing cross-sections of compound layers obtained by nitrocarburizing for 4 h at 853 K (580 C) using a 13.23 vol pct NH3,
57.95 vol pct H2, 20 vol pct CO, 8.82 vol pct N2 (rN = 0.30 atm1/2) gas mixture. (a) Cementite layer grown on a pure a-iron substrate. (b) h/e
double layer grown on a nitrogen presaturated a-iron substrate.

Fig. 5—Exemplary X-ray diffraction patterns (Co-Ka radiation) of compound layers obtained by nitrocarburizing for 4 h at 853 K (580 C)
using a 13.23 vol pct NH3, 57.95 vol pct H2, 20 vol pct CO, 8.82 vol pct N2, (rN = 0.30 atm1/2) gas mixture. (a) Cementite layer grown on a
pure a-iron substrate (cf. Fig. 4(a)), (b) h/e double layer grown on a nitrogen presaturated a-iron substrate (cf. Fig. 4(b)).

wedge-shaped specimens (series C; Figure 3) nitrocar- (practically) saturated with nitrogen (for more details,
burized under conditions where (initially) pure cementite see Section IV–C).
layers were obtained on the usual, 1-mm-thick, rectan- Based on the preceding discussion, the occurrence of
gular ferrite substrates (see series A). Indeed, at the thick similar stages in the compound-layer evolution of series
end of the wedge-shaped specimen, the substrate was A (variable gas composition with constant treatment
not saturated with nitrogen after nitrocarburizing, time; Section III–A) compared with series B (variable
which led to the formation of a single-phase cementite treatment time with constant gas composition; Section
layer. At some distance away, toward the thin end (rim) III–B) can now be understood as well: The substrate
of the specimen, nitrogen saturation of the substrate was becomes faster saturated with nitrogen after applying a
reached resulting in the development of a h + e dual- higher nitriding potential (cf. series A and cf. Figures 1
phase layer adjacent to the substrate. At the thin end and 2).
(rim) of the specimen, the substrate was saturated fast
with nitrogen resulting in the formation of a h/e double
B. Thermodynamics: Microstructure and Diffusion Paths
layer (Section IV–C).
Within the Compound Layer
The effect of nitrogen saturation on the development
of the microstructure of the compound layer is also As described subsequently, the observed phase con-
confirmed by nitrocarburizing of a nonsaturated, rect- stitution and interphase boundaries within the com-
angular specimen and a nitrogen presaturated, rectan- pound layer are in agreement with the metastable§§
gular specimen using a strongly carburizing atmosphere
(series C; Figure 4). The nonsaturated substrate showed §§
the formation of a massive cementite layer, whereas a h/ ‘‘Metastable’’ because h, e, and c’ are considered not to decompose
into graphite, nitrogen gas, and a-Fe as required for ‘‘stable’’ phase
e double layer had developed on the nitrogen presatu- equilibria.
rated substrate.
The hardness measurements in the ferrite underneath
a laterally inhomogeneous (h/e and e/h + e) compound ternary Fe-N-C phase diagram at 853 K
layer (cf. series B, Section III–C; Figure 6) show that the (580 C),[16,17,40,41] i.e., ‘‘local equilibria’’ prevail at the
formation of e–(carbo)nitride occurs only at locations of interphase boundaries in the compound layer. As a
the layer/substrate interface where the substrate is preparation for this discussion, the observations on the

2408—VOLUME 43A, JULY 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 6—Vickers hardness and the corresponding nitrogen concentration in the ferrite substrate directly underneath the laterally inhomogeneous
compound layer obtained after nitrocarburizing using the following process parameters: 15.44 pct NH3, 57.95 pct H2, 20 pct CO, 6.61 pct N2
(rN = 0.35 atm1/2), 2 h, and 853 K (580 C). For the inset in the bottom part of this figure, see the discussion of Fig. 9.

time-dependent development of the microstructure have carbon in ferrite–– (see Table I). It is concluded that the
been summarized schematically in Figure 8.
After the initial formation of a single-phase h layer at ––
the surface (stage 1), the subsequent formation of e at Moreover, as shown in Figure 7, the formation of a single-phase c’
layer from a cementite and/or an e containing compound layer is
the cementite layer/substrate interface occurs and results possible only if the carbon content in the a-iron substrate at the layer/
in either a h/e (stage 2a) double layer or a h/h + e double substrate interface is decreased.
layer (stage 2b). Then, the amount of e in the compound
layer increases strongly by the growth of e into the
substrate and the simultaneous conversion of cementite carbon has escaped through the surface into the gas
into e– at the surface-adjacent region of the compound atmosphere.
Now, starting from the phase at the surface of the
compound layer and its composition and ending at the
–
The possible conversion of cementite into e was first reported in a-iron substrate, different diffusion paths, representing
Reference 39 after nitriding of pearlitic substrates at 843 K (570 C). the change of the lateral gross composition and phase
constitution of the compound layer as function of depth
layer (stages 3 and 4), for which the development ends at constant time, can be indicated schematically in the
with the presence of a single-phase e layer (stage 5). isothermal section of the ternary Fe-N-C phase diagram
Next, the c’ phase is formed at the interface between the at 853 K (580 C) for different stages of compound-
e layer and the substrate leading to an e/e + c’ double layer growth. With reference to Figure 8, proposals of
layer (stages 6a and 6b) and then to an e/c’ double layer these diffusion paths have been made in Figure 7(a);
(stage 7). The final stage of the compound layer observe the full lines with arrows (if possible
development involves increasing the amount of c’ until the diffusion paths were simplified by straight lines).
a single-phase c’ layer results (stage 8). The occasional The time-dependent change in the microstructure of the
formation of an e + c’ dual-phase layer (Figures 2(d) compound layer implies that the diffusion path changes
and (e)) is not considered in Figure 8 as a separate stage. correspondingly during nitrocarburizing. It follows that
It should be recognized that the single stages can overlap the microstructural development as observed within the
in time, as indicated by the laterally inhomogeneous compound layer can be understood as a straightforward
nature of the microstructure of the compound layer (cf. consequence of (1) the depth dependence of the gross
Figures 1 and 2, and Section IV–C). composition at the stage of nitrocarburizing considered
In the final stage of compound-layer evolution, a and (2) the thermodynamics of the Fe-N-C system,
single-phase c’ layer is obtained (Figure 2(f) and stage 8 determining the solid–solid local equilibria. Conse-
in Figure 8). The c’-phase can contain only little carbon quently, kinetic constraints have not to be invoked to
because of its narrow composition range (Table I). So explain the microstructure within the compound layer.
the question emerges, where the appreciable amount A coexistence of cementite and c’ (i.e., cementite being
carbon taken up initially in the compound layer has directly in contact with c’ as reported for a temperature
gone. The uptake of carbon by the a-iron substrate can of 823 K [550 C] in Reference 15) was never observed
be excluded because of the small solubility limit of in the current study pertaining to a temperature of

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, JULY 2012—2409

Fig. 7—Isothermal section of the metastable Fe-N-C phase diagram at 853 K (580 C). The time-dependent microstructural evolution of the
compound layer from carbon-rich to nitrogen-rich phases after nitrocarburizing at 853 K (580 C) is illustrated schematically by diffusion paths
featuring, at a given time, the phase constitution and the lateral gross composition from the top to the bottom of the compound layer (solid
lines with arrows for continuously changing gross composition, dashed lines for jumps in gross composition). (a) Isothermal section with phase
boundaries according to Reference 24 (The invariant transition reaction c’ + Fe3C fi a + e was found to occur at a (single) temperature in the
range between 833 K an 843 K (560 C to 570 C), as determined by a series of experiments in Reference 24. However, according to calculations
based on the CALPHAD database (2008), the invariant transition reaction a + c’ + e fi c would occur between 848 K and 853 K (575 C and
580 C), which is not in agreement with Reference 24 and with the observed evolution of the compound-layer microstructure.) showing the entire
relevant range of composition. (b) The time-dependent change of the gross composition at the surface is indicated by a dotted line with arrows.
The iron-rich corner of the isothermal section shows the phase boundaries of the a-iron single-phase field.[17] The time-dependent change of the
nitrogen and carbon content in the ferrite substrate at the surface and subsequently at the layer (cementite)/substrate (ferrite) interface after
nitrocarburizing has been indicated schematically by the dotted lines with arrows. For the points A and B, see the text. Within the a phase, the
diffusion path is not shown because the starting point (substrate surface) and the end point (substrate center) of the diffusion path change with
time.

2410—VOLUME 43A, JULY 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 8—Schematic illustration of the progressive microstructural stages of the compound-layer evolution on pure a-iron as function of time upon
nitrocarburizing at 853 K (580 C). The development starts with the occurrence of a carbon-rich phase and proceeds toward nitrogen-rich phases
along diffusion paths compatible with the Fe-N-C phase diagram; see Fig. 7(a). (1) single-phase h layer, (2a) h/e double layer, (2b) h/h + e layer,
(3) h + e/h/h + e double layer, (4) e/h + e/e layer, (5) single-phase e layer, (6a) e with some c’ developing at regions close to the compound layer/
substrate interface, (6b) e/e + c’ double layer, (7) e/c’ double layer, and (8) single-phase c’ layer.

853 K (580 C). This finding is in agreement with the discussion and the diffusion paths in Figure 7(a)), a
ternary Fe-N-C phase diagram[17,24] where a two-phase lateral inhomogeneity of the microstructure is induced
field of cementite and c’ does not exist at 853 K (580 C) by the dramatic increase in the diffusive flux of
(cf. Figure 7(a)). The two-phase equilibrium a + e
obtained in the microstructure of the compound layer
after nitrocarburizing of nitrogen presaturated sub- *The diffusional flux of carbon into the a-iron substrate can be
neglected because of the small solubility limit of carbon in ferrite; see
strates (series C; cf. Figure 4(b)) indicates clearly the Table I.
temperature of 853 K (580 C) to be above the single
temperature for the invariant reaction h + c’ fi e + a in
the Fe-N-C system, which is consistent with experimen- nitrogen* through the compound layer at those regions,
tal results reported in Reference 24, where this temper- where the initially formed cementite layer has already
ature has been claimed to lie in the range of 833 K to been locally transformed into the e phase. Nitrogen
843 K (560 C to 570 C). diffusion through cementite (formed in the first stage of
nitrocarburizing) might vary locally depending on the
local microstructure (e.g., the local presence of an
C. Kinetics: Origin of Lateral Inhomogeneity initially high grain-boundary density might enhance the
of the Microstructure of the Compound Layer local flux of nitrogen through the cementite layer; cf.
Apart from the thermodynamically prescribed multi- Reference 26). Consequently, along the h/a interface, a
phase nature of the compound layer (see the preceding variation occurs for the time needed to attain the critical

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, JULY 2012—2411

Fig. 9—Schematic illustration of a mechanism enhancing the lateral inhomogeneity of the microstructure of the compound layer. (a) Cross-sec-
tional view of the compound layer showing fast diffusion of nitrogen through e at locations where the initially formed cementite layer has been
converted locally into e (the diffusion of nitrogen through cementite is comparably slow[32]). A diffusional carbon flux in the ferrite substrate is
neglected because of the small carbon solubility in ferrite. (b) Following the short-circuit diffusion illustrated in (a) and (b), the lateral diffusion
of nitrogen underneath the cementite layer leads to a lateral saturation of the substrate at the h/e transformation front and, thus, to the sub-
sequent formation of e, i.e., the h/e transformation front proceeds laterally at the specimen surface. (c) Top view of the specimen shown in (b).

concentration of nitrogen in the substrate for the V. CONCLUSIONS

formation of e at the h/a interface (cf. point B in
1. The microstructural evolution of the compound
Figure 7(b)). After prolonged nitrocarburizing, this e
layer developing after nitrocarburizing of a-iron
might grow from the h/a interface through cementite to
starts with the formation of carbon-rich cementite
the surface of the compound layer. The growth of the e
and develops into the direction of the nitrogen-
phase might occur additionally by the conversion of h
richer and carbon-poorer phases e and c’.
into e at the surface of the compound layer because of
2. Local equilibrium (or stationary state) conditions
the gradually increasing chemical potential of nitrogen
do not necessarily prevail at the gas–solid interface,
at the surface (cf. the discussion in Section IV–A).
at least not in the beginning of nitrocarburizing.
Because the diffusivity of nitrogen through e is much
3. The initial formation of cementite and the subse-
higher than through cementite, ‘‘short-circuit diffusion’’
quent formation of e are the consequence of (1) a
of nitrogen through an e ‘‘channel’’ becomes possible
solubility of nitrogen in a-iron much larger than
(cf. Figure 9(a), dashed arrow in Figure 1(e)). Such a
that of carbon and (2) a rate of nitrogen transfer
nitrogen short-circuit diffusion, followed by a lateral
from the gas phase into the solid much lower than
diffusion of nitrogen through e and a underneath the
that of carbon, i.e., kinetics governs the developing
cementite layer, will result in a laterally proceeding
composition and phase constitution at the gas–solid
saturation of the substrate with nitrogen at the trans-
and compound layer–substrate interfaces.
formation front between h and e (cf. Figure 6). This
4. Cementite acts as barrier for the inward diffusion of
mechanism contributes to a laterally proceeding trans-
nitrogen. This leads to a retarded growth of the
formation of the cementite layer into an e-rich com-
compound layer at locations where cementite is
pound layer (cf. Figure 9(b) and (c), and see
present, as well as initiates the lateral inhomogene-
Figures 1(b) through (e), 2(a) through (d)). The (local)
ity of the microstructure as a consequence of
development of the e phase at the surface of the
‘‘short-circuit’’ diffusion of nitrogen through e
compound layer can be observed macroscopically
‘‘channels’’ traversing the originally closed cement-
because the reflectivity of the cementite phase regions
ite layers.
at the surface is higher than that of the e phase regions.
5. Local equilibrium conditions do prevail within the
It is noted that the thickness of the compound layer
compound layer at solid–solid phase boundaries.
depends on the amount of cementite in the microstruc-
The change of the microstructure (phase constitu-
ture: cementite acts as a barrier for the inward diffusion
tion and lateral gross composition) as a function of
of nitrogen and carbon, in particular compared with the
depth in the compound layer is compatible with
e phase.[9,15,23,32,33,42] Indeed, a retarded growth of the
so-called ‘‘diffusion paths’’ in the ternary Fe-N-C
overall compound layer and, thus, a smaller compound-
phase diagram, i.e., thermodynamics governs the
layer thickness was observed (also locally) for micro-
microstructure within the compound layer. The use
structures exhibiting (also locally) a high amount of
of diffusion paths is a powerful method to visualize
cementite.

2412—VOLUME 43A, JULY 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

 the relation between the microstructure of the com- 13. P.F. Colijn, E.J. Mittemeijer, and H.C.F. Rozendaal: Z. Metal-
pound layer and (the isothermal section of) the cor- lkd., 1983, vol. 74, pp. 620–27.
14. J. Slycke and L. Sproge: Surf. Eng., 1989, vol. 5, pp. 125–40.
responding phase diagram. 15. T. Gressmann, M. Nikolussi, A. Leineweber, and E.J. Mittemeijer:
6. The courses of microstructural change of the com- Scripta Mater., 2006, vol. 55, pp. 723–26.
pound layer as observed after increasing the treat- 16. H. Du: J. Phase Equil., 1993, vol. 14, pp. 682–93.
ment time at a constant gas composition and as 17. J. Kunze: Nitrogen and Carbon in Iron and Steel, Akademie-Verlag
Berlin, Germany, 1990.
observed after increasing the nitriding potential at a 18. T. Bell: Heat Treat. Met., 1975, vol. 2, pp. 39–49.
constant treatment time are identical. 19. C. Dawes and D.F. Tranter: Heat Treat. Met., 1985, vol. 3,
7. A two-phase equilibrium e + a occurs in the pp. 70–76.
Fe-N-C system at 853 K (580 C) as validated by 20. T. Bell and D.H. Thomas: Metall. Trans. A, 1979, vol. 10A,
the formation of h/e double layers on top of nitro- pp. 79–84.
21. E.J. Mittemeijer: J. Heat Treat., 1983, vol. 3, pp. 114–19.
gen presaturated a-iron substrates. 22. M.A.J. Somers, P.F. Colijn, W.G. Sloof, and E.J. Mittemeijer: Z.
Metallkd., 1990, vol. 81, pp. 33–43.
23. H. Du, M.A.J. Somers, and J. Ågren: Metall. Mater. Trans. A
2000, vol. 31A, pp. 195–211, 801.
24. M. Nikolussi, A. Leineweber, E. Bischoff, and E.J. Mittemeijer:
Int. J. Mater. Res., 2007, vol. 98, pp. 1086–92.
ACKNOWLEDGMENT 25. T. Woehrle, A. Leineweber, and E.J. Mittemeijer: HTM J. Heat
Treat. Mater., 2010, vol. 65, pp. 243–48.
The authors are grateful to Dr. Thomas Gressmann 26. M. Nikolussi, A. Leineweber, and E.J. Mittemeijer: J. Mater. Sci.,
for his help with the nitrocarburizing experiments and 2009, vol. 44, pp. 770–77.
the scientific discussions. 27. E. Lehrer: Z. Elektrochem., 1930, vol. 36, pp. 383–92.
28. S.S. Hosmani, R.E. Schacherl, and E.J. Mittemeijer: Acta Mater.,
2006, vol. 54, pp. 2783–92.
29. H.A. Wriedt, N.A. Gokcen, and R.H. Nafziger: Bull. Alloy. Phase.
REFERENCES Diagr., 1987, vol. 8, pp. 355–77.
30. A. Wells: J. Mater. Sci., 1985, vol. 20, pp. 2439–45.
1. F.J.J. Van Loo, M.R. Rijnders, K.J. Ronka, J.H. Gulpen, and 31. G. Petzow: Metallographic Etching, ASM International, Materials
A.A. Kodentsov: Solid State Ionics, 1997, vol. 95 (1–2), pp. 95– Park, OH, 1999.
106. 32. M. Nikolussi, A. Leineweber, and E.J. Mittemeijer: Phil. Mag.,
2. P.M. Unterweiser: Source Book on Nitriding, ASM, Materials 2010, vol. 90, pp. 1105–22.
Park, OH, 1977. 33. M.A.J. Somers and E.J. Mittemeijer: Metall. Mater. Trans. A,
3. A.A. Kodentsov, G.F. Bastin, and F.J.J. Van Loo: J. Alloys 1995, vol. 26A, pp. 57–74.
Compd., 2001, vol. 320, pp. 207–17. 34. T. Sone, E. Tsunasawa, and K. Yamanaka: Trans. Japan Inst.
4. F.J.J. Van Loo: Progr. Solid State Chem., 1990, vol. 20, pp. 47–99. Met., 1981, vol. 22, pp. 237–43.
5. J.M.G. Vilar and J.M. Rubı́: Proc. Natl. Acad. Sci. U S A, 2001, 35. J. Crank: The Mathematics of Diffusion, Oxford Science Publica-
vol. 98, pp. 11081–84. tions, Oxford, U.K., 1975.
6. Y.H. Sohn and M.A. Dayananda: Acta Mater., 2000, vol. 48, 36. E.J. Mittemeijer and J.T. Slycke: Heat Treat. Met., 1996, vol. 3,
pp. 1427–33. pp. 67–71.
7. J.S. Kirkaldy and D.J. Young: Diffusion in the Condensated State, 37. H.J. Grabke: Archiv Eisenhütten., 1975, vol. 46, pp. 75–81.
The Institute of Metals, London, U.K., 1987. 38. M. Weller: Mater. Sci. Forum, 2001, vols. 366–368, p. 95.
8. G. Ghosh: J. Electron. Mater., 1998, vol. 27, pp. 1154–60. 39. E.J. Mittemeijer, W.T.M. Straver, P.F. Colijn, P.J. van der Schaaf,
9. M.A.J. Somers: Heat Treat. Met., 2000, vol. 4, pp. 92–102. and J.A. van der Hoeven: Scripta Metall., 1980, vol. 14, pp. 1189–
10. J. Slycke: Nitrieren und Nitrocarburieren, AWT, Darmstadt, 92.
Germany, 1996, pp. 19–28. 40. J. Kunze: Härterei-Tech. Mitt., 1996, vol. 51, pp. 348–55.
11. M.A.J. Somers and E.J. Mittemeijer: Surf. Eng., 1987, vol. 3, 41. J. Slycke, L. Sproge, and J. Ågren: Scand. J. Metall., 1988, vol. 17,
pp. 123–37. pp. 122–26.
12. E.J. Mittemeijer and M.A.J. Somers: Surf. Eng., 1997, vol. 13, 42. H. Du and J. Ågren: Z. Metallkd., 1995, vol. 86, pp. 522–29.
pp. 483–37.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, JULY 2012—2413

View publication stats