Corujeira Gallo 09 PH D
Corujeira Gallo 09 PH D
Corujeira Gallo 09 PH D
A thesis submitted to
DOCTOR OF PHILOSOPHY
January 2009
University of Birmingham Research Archive
e-theses repository
Low temperature plasma surface engineering has proved to be a useful method for increasing the
hardness and wear resistance of austenitic stainless steel without reducing the corrosion resistance of
this alloy. Plasma carburising is of particular interest as it produces a thicker hardened layer than
plasma nitriding, and an equivalent improvement in the tribological and corrosion performance of the
base material.
In this project, the novel active screen (AS) plasma technique was used to carburise austenitic stainless
steel AISI 316. The layers of carbon expanded austenite obtained after active screen treatments were
compared with the one produced by conventional direct current (DC) plasma treatments. The hardening
and wear resistance produced by AS and DC plasma carburising were equivalent. With regard to
corrosion, the AS treated material performed better than its DC counterpart as a consequence of the
The mechanism of AS carburising was comparatively studied with its AS nitriding counterpart. Different
experimental arrangements and two plasma diagnostic techniques were used for this purpose: optical
emission spectroscopy and electrostatic probes. The evidence shows that AS nitriding relies on the
deposition of iron nitrides and the active species in the plasma to produce hardening, whilst AS
Finally, the working principles of AS plasma furnaces are discussed with regard to the generation of
In first place, I would like to express my gratitude to Prof. T. Bell and Dr. H. Dong for giving me the
chance to develop this project in the Surface Engineering Group. I also want to thank Dr. E. Forlerer,
from Comisión Nacional de Energía Atómica (CNEA-CAC), for her continuous assistance and
encouragement.
I would like to extend my appreciation to the following institutions, which made vital financial
contributions to this research project: Roberto Rocca Education Programme, Alßan Programme
(scholarship E05M060379AR), The University of Birmingham, and Universidad Tecnológica Nacional
(UTN-FRBA).
I must acknowledge the help I received from members of the Centre for Electron Microscopy, specially
Ms. M.Q. Chu and Dr. Y.Y. Tse, and also Dr. J. Bowen and Dr. X. Y. Li, for their help with different
microscopy techniques. I am also grateful to Dr. V. R. Machavaram, Dr. L. Wang and Prof. G. F.
Fernando, whose assistance was very important for the development of the optical fibre probe; and I am
enormously thankful to Prof. N. Braithwaite, from the Open University, for his valuable discussions and
advice on electrostatic plasma diagnostic techniques.
It would be unfair not to mention here the constant help and support I received from all the academic
community in Metallurgy and Materials, especially from research fellows, technicians and colleagues.
This project would simply not have been possible without them; and I want to thank Mr. M. T.
Cunningham in particular for keeping an eye on, virtually, every daily issue.
Finally, I would like to express my deepest gratitude to my family and friends, for their joy in the
moments of success and their support in the difficult times.
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION 1
CHAPTER 4: RESULTS 73
1. Processing 73
2. Characterisation of plasma carburised AISI 316 75
3. Study of the mechanisms of active screen plasma surface engineering 78
4. Tribological properties 81
5. Corrosion behaviour 86
6. Plasma diagnostics – Optical emission spectroscopy 90
7. Plasma diagnostics – Ion flux probe 92
TABLES 163
APPENDICES 258
Appendix A 258
Appendix B 263
Appendix C 267
List of tables
Austenitic stainless steels are commonly selected because of their high corrosion resistance, which is
the main advantage of these alloys. On the other hand, low hardness and wear resistance are usually
mentioned among their downsides. Surface engineering processes have been traditionally used to
improve these aspects. Furthermore, thermochemical treatments conducted at low temperature are a
valuable method to improve the tribological performance of austenitic stainless steel, without affecting
the corrosion behaviour of this alloy. This is achieved by the formation of a supersaturated solution of
carbon or nitrogen in austenite, usually called expanded austenite, which confers significant hardening,
increased tribological performance, and corrosion resistance comparable or even superior to the
untreated substrate.
Several processes can be used to produce the so called expanded austenite layers. The evidence
indicates that plasma treatments outperform the conventional gas processes in terms of operative costs
although some intrinsic defects of this technique, such as edge effects, hollow cathode effects and
arcing, have blighted their industrial potential. To this end, a novel active screen plasma technique has
shown its potential to solve the problems of DC plasma treatments. The benefits are principally related
with a considerable improvement in the surface quality and the uniformity of the diffusion layer.
Whilst most of the work on active screen has been focused on nitriding, very little work has been
conducted on active screen plasma carburising. The latter treatments are of interest because they
produce thicker layers of expanded austenite, with a smoother gradient of chemical composition and
mechanical properties. Thus, carbon expanded austenite shows hardening effect, wear resistance and
additional load bearing capacity. Moreover, the better surface quality obtained with active screen
plasma carburising, compared with the DC plasma technique, yields greater improvement in the
Even though the effectiveness of the active screen plasma technology is a fact, the mechanisms
involved in this treatment remain unclear. This makes it difficult to optimise the processing conditions,
and it also generates some scepticism among the potential users. Therefore, the present work has been
directed at developing active screen plasma carburising (ASPC) technology to enhance the corrosion
and wear properties of austenitic stainless steel. The scientific and technological objectives of the
To explore the feasibility of active screen plasma carburising of AISI 316 austenitic stainless
To study the effect of different treatment parameters (i.e. pressure, gas composition and
To assess the wear and corrosion properties of the ASPC treated AISI 316 austenitic
stainless steel.
To develop and use plasma diagnostic techniques to evaluate the active species in ASP and
advance the scientific understanding of carbon mass transfer and carburising mechanism
involved in ASPC.
1. Stainless steel
There is no doubt that the field of materials engineering is now developing faster than at any other time
in the history of mankind, and that the rate of innovation continues to accelerate. The diagram in figure
2.1.1-1 summarises the historical evolution of the different classes of materials [1]. The increasing
number of choices of materials provides a better coverage of the applications’ demands, but this also
represents a great challenge for design engineers to select the optimum one.
The difficulties associated with the selection of engineering materials have been addressed by Ashby in
his book Materials selection in mechanical design. The use of indices based on the materials properties,
relevant for a given application, helps in the identification of the most convenient class of materials.
Furthermore, by plotting these indexes in so called materials selection charts, the range of applicability
for each material class becomes evident. Some of these charts will be used here as a reference frame
for the material of interest to this thesis, i.e. austenitic stainless steel.
With regard to physical and mechanical properties, it is possible to say, from the charts displayed in
figures 2.1.1-2 to 2.1.1-4, that steel covers a wide range of strength values, with a high level of stiffness
and toughness, only penalised by its high density. In addition to this, figure 2.1.1-5 shows that the
relative cost of steel is low compared to other engineering alloys, although this chart is not updated and
extremely valuable material for engineering applications, and stainless steel inherits these general
properties, to some extent. However, it is its high corrosion resistance that makes stainless steel unique.
As it can be seen from the materials selection chart in figure 2.1.1-6, stainless steel shows excellent
corrosion resistance in a wide range of harsh environments, putting it ahead of other engineering alloys.
For over a hundred years, several scientists (including Berthier, Baur, Hadfield, Osmond, Faraday and
Stodart) had collected evidence regarding the corrosion properties of iron-chromium alloys, but it was
not until early in the 20th century that the commercial potential of such material became evident [2, 3].
Harry Brearley is usually referred to as the inventor of a ferritic Fe - 13% Cr alloy firstly called “rustless
steel”, which subsequently became “stainless steel”. Although the discovery of the alloy is disputed by
Strauss, Maurer and Monnartz, Brearley was undoubtedly the first one to identify an application and to
commercialise stainless steel products, in the form of cutlery. In Germany, Maurer and Strauss
developed, in parallel, an austenitic alloy of Fe-Cr-Ni, which was soon applied in the food and chemical
From then onwards, the development of stainless steel was very fast, impelled by the great number of
industrial and domestic applications. Significant work was conducted to research and develop ferritic,
martensitic and the well known austenitic alloys with a number of different alloying elements. In all
cases, high chromium contents in excess of 10% are required to form a stable passive oxide film on the
There is no doubt that the distinct property of stainless steels is their corrosion resistance. As it is usual
in engineering applications, this property is attained at the expense of other factors, such as hardness,
on stainless steel has been focused on improvements of these other mechanical, engineering and
An impressive level of refinement has been achieved in the metallurgical practices to produce stainless
steel with a wide range of properties. In the next section, stainless steels of different compositions and
phases will be introduced, and a summary of their advantages, disadvantages and applications will be
presented.
1.2. Classification
Stainless steels are usually classified according to their phase constitution, i.e. martensitic, ferritic,
austenitic or duplex. The stability of these phases is dictated by the alloying elements, which can be
separated in two main groups: α forming elements and γ forming elements [6]. Two main constituents of
stainless steels, chromium and nickel, head each group, and it is customary to express the contribution
made by all the elements in the alloy as chromium equivalent and nickel equivalent according to the
following equations:
%Cr ⋅ equivalent = (Cr ) + 2 ⋅ (Si ) + 1.5 ⋅ (Mo) + 5 ⋅ (V ) + 5.5 ⋅ ( Al ) + 1.75(Nb) + 1.5(Ti ) + 0.75(W ) (2.1-1)
where the element symbols between brackets represent the alloy composition expressed as mass
fraction (wt%).
gives a rough but generally good map of the phases that can be expected for a given alloy composition
and, consequently, its mechanical properties. In this way, it is possible to see that high chromium
equivalents render martensitic and ferritic structures, while high nickel equivalents stabilise the
austenitic phase. Notwithstanding, the design of the alloy composition is a rather more complicated
matter which involves both technical and economical factors. It is usually intended to produce the best
possible material for a given application at the lowest possible cost and, as it is the case for most
It is important to mention that other phases, intermetallic compounds, carbides and nitrides are
commonly found in the Fe-Cr-Ni system with addition of other alloying elements, e.g. Ti, V, Mo, Co, Mn,
Al, Nb, C, N. Although present in only small fractions, these secondary phases can significantly affect
the metallurgical characteristics, corrosion properties and mechanical properties of stainless steels.
In what follows, each one of the generic categories of stainless steel will be examined in some further
detail.
These alloys typically contain 12% to 17% Cr and 0.1% to 1.0% C with additions of other alloying
elements such as Mo, V, Nb, Al and Cu. The austenitic structure, which is stable over 1050ºC,
undergoes full martensitic transformation on cooling, producing components with high strength and
hardness, but also low toughness. Therefore, it is necessary to temper these alloys at the highest
Alloying elements such as Mo, W, V and Nb are of interest to increase the tempering resistance by
secondary hardening reactions, which increase the strength but also reduce the toughness and the
resistance to stress corrosion. In addition, their tendency to retain δ ferrite has to be balanced by
austenite forming elements. The allowable amount of C and N is limited because these elements
significantly reduce the toughness, so that Ni, Co, Mn and Cu have to be used, thus increasing the cost
of the alloy. In addition, this introduces another detrimental effect since the MS - MF temperature range is
depressed, thus a small fraction of austenite can be retained at grain boundaries and subsequently
transformed to martensite during tempering, causing distortion and embrittlement. The austenite forming
elements also lower the AC1 temperature, and consequently limit the range of tempering conditions.
This rather delicate balance between alloying elements acquires one more degree of complexity when
considering the precipitation hardening effects associated with additions of Ti, Al, Cu and Mo. The
mechanism, independent of the carbon content. The strength and ductility can usually be tailored by
changing the tempering temperature, although some large intermetallic particles (Fe2Mo) and secondary
phases (Cr rich α’-phase, Fe-Cr-Mo χ-phase) can form on over-ageing, leading to severe embrittlement.
In summary, the advantage of martensitic stainless steels is their high strength and high hardness, but
this is achieved at expense of the corrosion resistance. The applications include tools, gears, shafts,
valves, bearings, engine and wear resistant parts in chemical and petrochemical industries,
Martensitic stainless steel components were found to be difficult to produce in the fully hardened
condition, and complex shapes could not be heat treated without severe distortions. Therefore, there
was a requirement for a stainless steel with a preliminary austenitic structure to permit good formability,
which could subsequently undergo martensitic transformation and produce precipitation hardening,
keeping a high tempering and corrosion resistance. It is not difficult to recognise the metallurgical
In order to fulfil the specifications, the alloy requires low C content and 16% to 19% Cr, balanced with
austenite forming elements (typically 4% to 7% Ni and/or Mn) to reduce the tendency to form δ ferrite,
but keeping MS high enough to allow martensitic transformation when necessary. The alloy composition
to produce the right balance between δ ferrite and MS is critical. Furthermore, Mo and V are added for
The martensitic transformation can be induced by refrigeration (sub-zero heat treatment) or by a primary
ageing treatment to precipitate M23C6 and control the amount of C in solution, available for the
martensitic transformation (increase in MS). Cold working is an alternative way, though not usually relied
upon because the transformation is not uniform. Although achievable, the accurate adjustment of the MS
- MF temperature range is difficult and requires stringent control of the alloy composition, with an
obvious impact in the economical viability. Therefore, most of these alloys are manufactured in the fully
The applications are very limited and include the skin of high speed rockets and aircraft, high strength
As their name indicates, these are ferritic alloys up to the melting point, containing 17% to 26% Cr, and
only a small fraction of austenite forms at high temperature depending on the amount of other alloying
elements, mainly C and N. The ferritic structure is stronger than the austenitic one, and the absence of
Ni in their composition represents a significant advantage over the austenitic counterparts in terms of
cost, although they are penalised in corrosion resistance, formability and toughness.
The occurrence of Fe-Cr σ-phase, Fe-Cr-Mo χ-phase and Cr rich α’-phase is of importance in ferritic
stainless steel and can lead to some embrittlement phenomena. Care should also be taken to avoid
alloys with high content of austenite forming elements. Moreover, alloys containing Ni, Co, Mo, Ti, Al
additions precipitate β-phase intermetallic compounds, which together with carbides (M23C6 and M7C3)
and nitrides (M2N) reduce the toughness. This list of embrittlement effects explains why ferritic stainless
steels exhibit ductile to brittle transition temperature, and puts into evidence one of the main weak points
On the other hand, the same mechanisms that reduce ductility also produce strengthening, so ferritic
stainless steels are superior to the austenitic ones in terms of yield and ultimate stresses. The stacking
fault energy is higher, thus ferritic stainless steels have low work hardening rate, which reduces the flow
stresses upon forming but affects the uniform strain and stretch formability. On hot rolling, a problematic
crystallographic texture called roping develops and, while being very ductile, ferritic stainless steels are
although in the most severe conditions they do not perform as well as their austenitic counterparts.
Addition of molybdenum increases their corrosion resistance, and stabilisation with Ti and Nb helps to
avoid grain boundary attack due to precipitation of chromium carbides. Perhaps one of the greatest
advantages offered by ferritic stainless steels is their high resistance to stress corrosion cracking,
Ferritic stainless steels are increasingly displacing austenitic stainless steels in the less demanding
applications, mainly because of economic reasons. Efforts have been made to improve their weldability
and toughness, and thus increase the range of applications, which is for the moment focused on indoor
stainless applications.
Duplex stainless steels were developed in an attempt to fuse in one material the positive aspects of
austenitic and ferritic alloys. Therefore, they are constituted by a ferritic-austenitic or ferritic-martensitic
duplex structure with a typical composition of 18% to 30% Cr, relatively low Ni content and some
additions of Mo, Nb and Ti. The reduction in the amount of Ni provides some cost reduction over the
austenitic stainless steels, while having superior yield and ultimate stresses, and resistance to corrosion
and pitting that compares favourably with the austenitic benchmark alloy.
The presence of ferrite and the induced reduction in the grain size of austenite increases the strength.
Moreover, the austenite undergoes martensitic transformation upon deformation, increasing the strength
to even higher levels than the pure ferritic structures. This is achieved without a significant reduction in
toughness, which is somewhere between the austenitic and ferritic alloys. The physical and mechanical
The relatively high occurrence of secondary phases is perhaps the weakness of duplex stainless steels,
and several embrittlement effects have been reported in the temperature range 300ºC - 1000ºC, related
to Fe-Cr-Mo σ and χ-phases, Fe-Mo-N π-phase, and some degree of partitioning of alloying elements
between the austenitic and ferritic phases. Nevertheless, appropriate heat treatments and cooling rates
The applications of these alloys are expanding and, due to their high corrosion resistance and good
mechanical properties, include: chemical, petrochemical, oil, wood pulp, pharmaceutical, textile and
nuclear industries.
Austenitic stainless steel (ASS) is undoubtedly the most widely used, and currently accounts for 70% to
80% of the stainless steel world production volume, with its emblematic 18/8 composition, i.e. 18% Cr
and 8% Ni. ASS typically contains 18% to 25% Cr and 8% to 20% Ni, making them fully austenitic and
consequently non-magnetic, although martensitic phase transformation may occur during deformation.
Additions of Mo and other alloying elements further increase its outstanding resistance to general and
pitting corrosion, and it is this property that makes it unique, not so its mechanical properties, which are
meagre.
In general, austenitic stainless steel is the first choice in the most severe corrosive conditions. The C
content of these alloys is usually kept at very low levels to avoid precipitation of Cr23C6 carbides at grain
produce the so called stabilised grades. The high Ni and Mo contents make ASS an expensive material,
although its relative cost compares favourably with other non-ferrous alloys, e.g. Ni or Ti based alloys.
Some attempts have been made to use Mn, Cu and N to replace Ni in the alloy composition with some
degree of success, although the corrosion resistance is inferior to the traditional Ni alloyed ASS.
To complete the selection grid, ASS is tough and shows no ductile-brittle transition temperature, it is
easily weldable and, having a low yield stress, it exhibits high ductility and is especially suitable for
stretch forming. Nevertheless, its low resistance to stress corrosion cracking, low ultimate stress and
Being the subject of this research project, the metallurgical aspects, corrosion properties and wear
resistance of ASS, and of AISI 316 in particular, will be discussed in further detail in the following
sections.
The Fe-Ni-Cr ternary phase diagram is the basis for the study of ASS. Additions of Ni and other
austenite stabilising elements expand the γ loop and, with sufficiently large amounts, make the
austenitic structure stable (or metastable) up to room temperature [12]. In general, the effect of alloying
elements depends on the composition of the material to which they are added, and that is the case of
chromium in ASS. Being an α stabilising element, chromium tends to shrink the γ loop, although in the
presence of Ni it retards the γ → α transformation. From the ternary phase diagram in figure 2.1.3-1, it
is possible to conclude that the 18% Cr - 8% Ni composition minimises the amount of Ni for a fully
of interest.
The typical chemical compositions of commercial ASS alloys are shown in table 2.3-1. Differences in
composition and processing conditions have a profound effect in the type and volume fraction of the
secondary phases present in the alloy and, consequently, in their physical and mechanical properties
[14]. A detailed and comprehensive description of the secondary phases observed in ASS is out of the
scope of this review and only the ones most significant to the material under study (i.e. AISI 316) will be
discussed, namely: martensite, carbides and intermetallic phases (σ, χ and Laves phases) [15].
1.3.1. Martensite
By calculating the Cr and Ni equivalents with equations (2.1-1) and (2.1-2), and making use of the
Schaeffler diagram in figure 2.1.2-1, it is evident that the basic 18/8 composition of ASS is on the border
line of the pure austenitic structure domain [5]. The Ni content of these ASS is only enough to reduce
MS to just below room temperature, and substantial martensitic transformation takes place if the material
is cooled to sufficiently low temperatures [4]. The MS temperature is determined by the chemical
composition of the alloy and the following equation can be used as a reference [6]:
where the element symbols between brackets represent the alloy composition expressed as mass
fraction (wt%).
the austenite against martensitic transformation [16]; the effect of Ni being twofold: it stabilises the
austenitic structure and lowers the MS temperature [5]. In this way, AISI 316 exhibits higher stability to
martensitic transformation, compared to the basic 18/8 AISI 304 [17]. However, the γ (fcc) structure is
said to be metastable at room temperature and martensitic phase transformation can occur on plastic
To explain this phenomenon it is useful to recall the thermodynamic basics of the phase transformation:
in order to happen, the transformation has to reduce the free energy of the system and, at the same
time, sufficient energy must be given for the system to initiate the transformation (activation threshold).
The schematics in figure 2.1.3-2 illustrates these principles, so that if TO is the equilibrium temperature
for the γ and α phases, the transformation will occur spontaneously below MS since the thermal energy
at that point is enough to mount the activation threshold ΔGth. Between TO and MS, the γ phase is not
thermodynamically stable, but the thermal energy is insufficient to initiate the transformation. It is in this
region that the transformation can be activated by plastic deformation, taking the name of strain induced
martensitic transformation [16, 18]. Hence, a reference temperature Md30 can be estimated, from the
where Md30 is the temperature at which 50% martensite is obtained under a true strain of 0.3, and the
element symbols between brackets represent the alloy composition expressed as mass fraction (wt%).
as a function of the temperature and the strain [18]. This behaviour was found to be the result of the
competition of two deformation modes. At low temperatures, no slip takes place and all the deformation
is due to martensitic transformation, whilst at higher temperatures, slip becomes dominant and the
The inset of martensite has a marked effect in the corrosion and mechanical properties of AISI 316, and
1.3.2. Carbides
Carbides and, to lesser extent, nitrides have attracted particular attention in high alloy steels, and ASS
is no exception. The mechanical properties and corrosion performance of ASS are significantly affected
by precipitation of carbides and nitrides, being beneficial at times, but catastrophic in some other cases
[6]. The effect of these phases is very much dependent on their composition, stoichiometry, morphology
and precipitation sites [19]; whilst the kinetics of precipitation is mainly governed by the availability of
carbon, nitrogen and carbide/nitride forming elements in the alloy, together with some energetic
As it is evident from figure 2.1.3-4, the solubility of carbon in austenite is reduced by the chromium
content, to about 0.05% in the case of the 18/8 ASS [4]. Therefore, carbon contents in the order of
0.05% or higher will remain in solution in the austenitic structure only if the material is solution treated
and quenched. However, during heat treatments or slow cooling in the temperature range 550ºC -
800ºC, precipitation of carbides takes place. Below 500ºC, precipitation is hindered by the low mobility
according to the reference sequence presented in figure 2.1.3-5. The letter M accounts for transition
elements (e.g. Fe, Cr, Ni, W, Mo) and carbides of mixed composition are usually found, depending on
the temperature of the ageing treatment and the composition of the alloy [5]. However, the chromium
carbide has received special attention because of its implications for the corrosion properties of ASS,
The M23C6 carbide is fcc in structure, with a lattice parameter about three times that of austenite, and
{100}γ || {100}M23C6
<010>γ || <010>M23C6
Regarding location, carbides are prone to follow the precipitation sequence: random grain boundaries,
non-coherent twin boundaries, coherent twin boundaries and, finally, intragranular dislocations [19, 20,
23]. This phenomenon puts into evidence the existence of a carbon content threshold and an energy
threshold inversely proportional to the interfacial energy of the grain boundary [24].
Upon long ageing treatments, cellular precipitates can form at grain boundaries, together with MC1-X and
M6C carbides [5, 23, 25]. Although they are important for the mechanical properties of ASS, and
particularly to its creep resistance, these precipitates are of no significant relevance to this particular
A number of intermetallic compounds may precipitate in ASS, depending on the alloy composition,
during relatively long ageing heat treatments [15]. These usually involve some degree of partition of
alloying elements, and thus affect the mechanical and corrosion properties. Table 2.3-2 lists some of the
most common phases, and figure 2.1.3-6 shows some experimental observations on the time-to-
precipitation in ASS at 550ºC - 600ºC. The formation and evolution of these secondary phases have
been extensively discussed and remain disputed, although evidence indicates that the presence of δ
ferrite has a strong effect in their occurrence and that they nucleate and grow faster through a
The Fe-Cr σ-phase is, perhaps, the most studied intermetallic phase in ASS, and has negative effects in
toughness and corrosion resistance [15]. Increasing contents of Cr, Mo and Si are known to ease
precipitation of σ-phase and the presence of δ ferrite, being richer than austenite in these alloying
elements, provides nucleation sites. Furthermore, the diffusion rate of Cr and Mo is significantly higher
in ferrite than in austenite, thus requiring shorter times and lower temperatures for precipitation of σ-
phase to occur. At the same time, the solubility of carbon in σ-phase is very low, so that this phase
usually precipitates in carbon impoverished regions, i.e. after precipitation of M23C6 carbides [27].
According to calculations done by Sourmail and Bhadeshia [28], the precipitation of σ-phase is more
difficult than M23C6 carbides, and is consequently restricted to high energy sites in the microstructure.
The (η) Fe2M Laves phase (where M usually represents Mo, Nb or Ti) is a hexagonal intermetallic
compound found in minor volume fractions in ASS, and exhibits similar precipitation kinetics to σ-phase,
although there is no general agreement on the temperatures and times required for this precipitation to
take place [25]. Laves phase has been reported as a transient phase, preceding the formation of M6C
χ-phase is a Fe-Cr-Mo intermetallic compound that may precipitate in ASS in minor quantities and after
longer soaking times at relatively high temperatures, typically over 750ºC [23]. Its formation is related to
the presence of M23C6 carbides and δ ferrite in the primary microstructure [29], and it is mainly found in
Other precipitates, although of importance to the properties and applications of ASS, have not been
mentioned in this section because of their low incidence in this particular study, and they will be
introduced at a later stage only when necessary. Some examples are: G-phase silicide, Z-phase
The following section deals with the corrosion properties of ASS, and relates some of the metallurgical
The corrosion resistance is the most important property for ASS, and these alloys were designed and
are usually selected for service under mild to severe operating conditions. A suitable grade of stainless
steel will generally serve well for many years, but the selection of the wrong grade or heat treatment can
lead to catastrophic failures [30]. This is because ASS owe their corrosion resistance to the formation of
a passive film at the surface, which isolates and protects the bulk material from the corrosive
environment. The nature of the passive film is dynamic, and dissolution and repair take place
The existence and protective role of this oxide film was first proposed by Faraday, but confirmation had
to wait for many years until a suitable method was developed to isolate it and to study it [32-34]. From
then onwards, much research work has been conducted on the nature of this passive layer, and the
properties that different alloying elements impart to it, given that the electronic properties of the film are
believed to play an important role on its stability through ion and electron transport reactions [35, 36]. To
this end, the fundamental explanation is still a matter of dispute, but it is widely accepted that the
availability of oxygen is crucial for the formation of the passive film; chromium is the main and most
important constituent, although minor additions of other alloying elements also have an important effect.
The content of chromium is critical to provide the steel with corrosion resistance, and the minimum
amount required is a function of the operating conditions, namely the type of acid, its concentration and
temperature. As a general rule, the onset of a passive film is found at chromium contents in the order of
12% [37], although 17% to 20% is usually required to impart passivity to ASS in many corrosive
environments [2, 31]. In any case, higher chromium contents do not always improve corrosion
resistance, and the opposite trend may be observed in some acids, for example sulphuric and
hydrochloric.
The effect of nickel in the corrosion properties of ASS depends on the presence and quantity of other
alloying elements, as some kind of synergistic effect with chromium has been suggested [2]. For the
particular amounts and proportions found in ASS, nickel is reported to promote repassivation and
Additions of molybdenum produce a marked improvement in the resistance of ASS against reducing
acids, pitting and crevice corrosion [2, 31]. Even though molybdenum is not an integral part of the
passive Cr-rich film, it was found to stabilise it, particularly in the presence of chlorides [36]. On the
other hand, ASS with large amounts of this alloying element are prone to segregation and precipitation
of secondary phases and, consequently its use is limited to moderate contents [39].
Nitrogen is beneficial in reducing the segregation of Mo-containing phases and providing further
resistance to pitting corrosion [31]. Silicon, vanadium, tungsten and copper are respectively found to
provide some beneficial effects to ASS under different service conditions [38]. Figures 2.1.4-1 and 2.1.4-
2 show the schematics of the polarisation curve of ASS in sulphuric acid and the effects of different
The corrosion resistance of ASS is defined not only by the content and proportion of alloying elements,
but also by the way in which these elements are dispersed in the microstructure. In this way,
segregation and precipitation of secondary phases play a very important role in determining the form of
corrosion, i.e. general corrosion or some localised attack of the kind of pitting, crevice, intergranular
Perhaps the most important phenomenon for the corrosion resistance of ASS, and one that has taken
much attention from researchers, is the precipitation of chromium carbides of the form M23C6. The
negative effect is not a product of the carbide itself, but of the depletion of chromium in the area
is usually referred to as sensitisation and may lead to intergranular corrosion and, ultimately, to the
In order to avoid sensitisation, the carbon content of ASS is kept at the lowest possible level, at expense
of the mechanical properties, and other alloying elements (Ti, Nb) might also be added as stabilisers,
i.e. to combine with the remaining carbon and form MC carbides instead of the harmful Cr23C6 [2, 38]. At
the same time, heat treatments or slow cooling in the temperature range 550ºC - 800ºC should also be
avoided.
Precipitation of carbides has also been found to depend on the crystallography of grain boundaries, with
random grain boundaries being more prone to precipitate carbides and also some other secondary
phases, as opposed to the coherent twin boundaries [19, 20, 42]. This is a result of the high number of
atoms common to both grains of the coherent twin boundary, which reduces the energy of the boundary
[43]. More recent studies have reported some correlation between the crystallographic texture [44] and
also the type, distribution and connectivity of grain boundaries [45, 46], with the corrosion behaviour of
the alloy.
These models make use of the theory of Coincident Site Lattice (CSL), which classifies the grain
boundaries according to the misorientation of the neighbouring grains and, consequently, according to
the fraction of coincident atoms between neighbouring lattices [47]. This is usually expressed as the
reciprocal density of coincident sites Σ, and boundaries with high fraction of coincident sites, i.e. low Σ
number, are believed to exhibit special properties [48]. In the case of alloys of low stacking fault energy,
A grain boundary can be defined by five parameters or, in other words it can be said to have five
degrees of freedom: 3 account for the crystallographic relationship between neighbouring lattices and 2
define the plane orientation of the grain boundary [49]. Therefore, the CSL model is not restrictive
enough inasmuch as it fails to account for the plane orientation of the grain boundary [48, 50].
Consequently, a coincident site relationship between neighbouring lattices does not always imply a
coincident site boundary or special boundary [43]. Nevertheless, the CSL theory is useful to categorise
grain boundaries in view of the relative difficulty to define all five degrees of freedom [50].
Finally, the presence of deformation slip bands and also MnS inclusions in the microstructure can
reduce the corrosion performance of ASS [38]. The latter inclusions are reported to be favourable sites
for pitting. Generally speaking, every discontinuity or weak point in the passive film will be susceptible to
H. Brearly, known as the father of stainless steel, used to be called the inventor of the knife that would
not cut. This rather anecdotal comment on the difficulties found in keeping stainless steel knives sharp
is timely when introducing one not so positive aspect of these alloys: their low hardness and wear
resistance.
As opposed to corrosion, the wear properties of ASS are moderate, and it is fair to say that ASS are not
usually selected because of their tribological performance (figure 2.1.5-1). However, it is sometimes
Furthermore, ASS could be selected because of their corrosion resistance or their non-magnetic nature,
at the expense of the wear performance. In any case, the wear properties of ASS are of interest in order
The phenomenon of wear involves several material properties (e.g. strength, hardness, toughness,
thermal conductivity) and also the operating conditions (e.g. contact geometry, load, sliding speed,
temperature). Therefore, it can be said that wear poses a problem with a relatively high level of
complexity. The models and theories proposed in the literature are usually limited to a relatively small
range of materials and conditions, and the diversity of theories and terminology can be misleading.
For the sake of clarity, and at the expense of accuracy, the different wear mechanisms will be
introduced by using the wear map shown in figure 2.1.5-2 [53-55]. The normalised pressure and
~ F
F= (2.1-5)
An ⋅ H
v ⋅ ro
v~ = (2.1-6)
a
where F is the applied load, An is the contact area, H is the material hardness, v is the sliding velocity, ro
is the radius of the contact area and a is the thermal diffusivity. The range of conditions relevant to this
project is that of medium-low velocity and medium-high load (i.e. top left quarter of the map in figure
2.1.5.2).
61]. There is a good degree of coincidence in the experimental results and observations, although not
so with regard to the fundamental explanation of the wear mechanisms. In general, the poor tribological
properties of ASS are associated with damage by galling or seizing [56], especially at high loads and
low sliding speeds, so that it can be said that wear of severe nature takes place by adhesion [62, 63].
In this wear mode, junctions are formed between the sliding elements, at the interacting asperities, and
subsequently deformed until rupture occurs, producing very rough wear tracks [64]. In this way, material
is transferred from one of the contacting surfaces to its counterpart throughout sliding (figure 2.1.5-3).
Furthermore, plastic flow is not confined to the interacting asperities but it extends towards the
subsurface. This leads to the nucleation of cracks, which propagate parallel to the surface and are
ultimately responsible for the detachment of plate like wear debris particles of large size, through a
At low temperatures, below Md, the strain produced at the surface and subsurface of ASS sliders is
sufficient to induce martensitic phase transformation, and it is well established that the formation of
martensite has a significant effect in the wear phenomenon [52, 56, 57, 66]. The thickness of the work
hardened and martensite layer depend on the material and the sliding conditions, namely load and
sliding speed.
Some opposing effects have been reported for this hardened layer, based on two different mechanisms.
On the one hand, martensite is said to increase the hardness of the substrate and, consequently, its
wear resistance [57, 62, 67]. On the contrary, other authors have reported that the austenitic substrate,
even after work hardening, is not strong enough to support the hard but brittle martensite surface layer,
resilience and toughness of the material are of significant importance, and the ratio of hardness to
elastic modulus has been reported to be a more appropriate index than the mere hardness, to rank the
wear resistance [68, 69]. Best tribological performances have been reported for combinations of high
surface hardness and relatively low elastic modulus, to reduce the tendency to plastic deformation and
reduce the mismatch of properties, while keeping deformation within the elastic range.
Mild wear conditions are found at some higher sliding speeds and lower loads, as can be seen in the
wear map in figure 2.1.5-2. The reduced wear rates are associated with the formation of a flat and
smooth protective oxide layer on the surface [57-60], and the onset of a mechanism called oxidational
wear [63]. A competition is established between the removal of the oxide layer through abrasion or
flaking; and the reaction and passivation of the fresh metal surfaces exposed to the environment [70].
The former depends mainly on the load, whilst the latter is a function of the temperature and,
consequently, the sliding speed. Under the right sliding conditions, a thin and tough oxide layer covers
The ductility and resilience of the oxide layer are important to accommodate the deformation of the
substrate [63]. In addition, the hardness and the elastic properties of the underlying material are critical
to provide the necessary support to the layer of oxide. Under stable oxidational wear conditions, the
oxide layer grows up to a critical thickness, at which it becomes unstable, and it finally flakes off by the
The stoichiometry of the oxide depends on the environment and the temperature of formation, i.e. the
temperature at the sliding asperities, and its effect on the wear phenomenon is of significance [58]. At
becomes stable and over 600ºC FeO is found [71]; in the presence of moisture, these oxides may also
form hydroxides [60, 72]. Once again, there are some controversial reports on the effect that each
specific type of oxide has over wear rates and friction, but oxides are generally associated with mild
With regard to friction force, coefficients in the range 0.4 to 0.7 have usually been reported for a wide
range of sliding conditions. An average coefficient of friction of 0.5 can be considered with fairly ample
agreement [56, 60], although some variations and fluctuations in the friction force have also been
mentioned [52]. The fluctuating nature of the friction force is well known [73], and it has been associated
with the interaction of surface asperities [74], wear mode transitions, and formation, trapping and
detachment of large wear debris particles [75]. It is interesting to note that discontinuities in the plastic
deformation at the subsurface, such as twinning and phase transformations, would also yield
This succinct summary of the tribological properties of ASS was intended to provide a reference frame
and introduce the basic principles of wear and friction under dry sliding in air. It is clear that the poor
tribological performance of ASS represents a serious limitation for many applications, which demand
excellent corrosion and wear resistance. Other stainless steels or non-ferrous alloys will continue to
develop and might fill part of this gap, although this is usually achieved at the expense of toughness,
corrosion resistance, ferromagnetism or cost. Surface engineering offers an alternative way to improve
the tribological properties of ASS at the surface level, while keeping the traditional benefits of this
material.
The wide use of surface treatments and coatings to combat wear was foreseen many years ago, based
on the fact that the mere use of wear resistant bulk materials can be inappropriate and expensive [76,
77]. Furthermore, the increasing level of understanding of the importance of the surface properties and
surface condition to the most common failure modes, i.e. wear, fatigue and corrosion, led to the
Several definitions have been proposed, usually implying the application of processes or treatments to
deliberately modify the surface of engineering components, in order to attain adequate material
properties for a given application, thus providing a cost effective enhancement in performance [81, 82].
This definition involves the concept of fitness-for-purpose [83] and it is generally accepted that SE can
Provide flexibility to match substrate and surface properties for a particular application
It is of paramount importance to consider both the surface and the substrate when designing or
selecting a SE treatment, as it has been pointed out by many authors [76-78, 82, 83]. The concentrated
loads developed during the contact and slide of rough surfaces are of significance in tribological
applications [85]. The strength of the substrate and the thickness of the coating or hardened layer define
the load bearing capacity and fatigue response of the system, as shown in figure 2.2.1-1 [86-88]. A
complete understanding of the behaviour of the substrate/surface system under the operating conditions
tempering and case hardening of metal weapons and tools have been quoted in the literature, dating
back to 1400 BC [81]. The innovative aspect of SE is the design approach, and the increasing
awareness of the importance of the surface properties and the surface condition to the performance of
the interacting surfaces [78, 84]. In view of the increasingly demanding operating conditions and the
requests to reduce production costs, many industrial sectors have turned their eyes to SE, making this a
The variety of SE treatments can be appreciated in figure 2.2.1-2. The ever wider range of options, and
the absence of a comprehensive system for selection constitutes a great challenge for design engineers
and the choice of a particular treatment is usually based on previous empirical experience rather than
on a detailed or analytical study [76, 78, 89]. For the purpose of this review, the SE treatments will be
Those which coat the substrate: PVD, CVD, thermal spraying, etc.
Those which modify only the structure of the substrate: induction hardening, flame hardening
Those which modify the chemical composition and the structure of the substrate:
The ones of significance to this project are the thermochemical or diffusion hardening processes, and
Finally, the environmental impact of SE must be taken into account, in view of the increasingly stringent
regulations, if not for our own sake. If the energy content of one particular material is taken as a
applied to it would increase its environmental footprint. However, if the increase in performance is
significant enough, the trend is reversed. SE treatments can increase the life span of the components,
reduce their size and weight, decrease energy losses associated with friction, minimise or even
eliminate the use of polluting lubricating oils, among other environmental benefits [91].
Thermochemical treatments, sometimes referred to as case hardening or cementation, are based on the
modification of the chemical composition of the substrate material. These treatments can be succeeded
by a change in the structure through heat treatment. The formal definition available in BS EN
Thermochemical treatment: Heat treatment carried out in a medium suitably chosen to produce
a change in the chemical composition of the base metal by exchange with the medium.
In the case of diffusion treatment, the definition in that same standard is:
Diffusion treatment: Heat treatment or operation intended to cause the diffusion towards the
interior of the ferrous product of elements previously introduced into the surface (for example,
From these definitions it becomes clear that two main factors will govern the process, namely: the
exchange or absorption reaction with the medium, and the diffusion in the metal [93]. As it is illustrated
in figure 2.2.2-1, the medium will determine the way in which the diffusing elements are delivered to the
account of the media used for carburising will be given in a following section.
Once in the metal, the transport of the absorbed substance takes place by diffusion, and follows Fick’s
laws:
∂C
J = −D (2.2-1)
∂x
∂C ∂ 2C
=D 2 (2.2-2)
∂t ∂x
where J is the flux of diffusing substance, D is the diffusion coefficient, and ∂C/∂x is the concentration
gradient [93, 94]. Therefore, the transport of the substance in solution is driven by its concentration
gradient and the diffusion coefficient which, at the same time, depends on the temperature, the chemical
composition and phase structure of the substrate. For a given alloy, kept at constant temperature in a
medium with a consistent concentration of the substance of interest, the case depth will only depend on
x = a (Dt )
1/ 2
= Kt 1 / 2 (2.2-3)
where x is the case depth, a is a constant, D is the element diffusivity, t is the treatment time and K is a
factor determined by a and D [93]. Higher treatment temperatures yield the same case depth in shorter
time, although there are technical limitations related with life of the furnaces, and metallurgical
considerations regarding the side effects of keeping the substrate material at high temperatures [94].
Consequently, diffusion treatments are slower when compared to other surface deposition techniques
[77], and treatments as long as 72 hours are common practice in industry. On the other hand,
wear and fatigue performance, but also the load bearing capacity [95].
The next section focuses on the application of thermochemical treatments on austenitic stainless steel
substrates.
As it has been mentioned in a previous section, the chemical composition of ASS makes them fully
austenitic up to room temperature, and thus no phase transformation hardening takes place upon
quenching. Consequently, SE treatments are an interesting alternative way to increase the surface
hardness and improve the wear resistance. However, SE of ASS has traditionally been considered bad
practice [30], as it poses two main problems: the passive oxide film and the precipitation of chromium
carbides [96].
The passive chromium oxide film on ASS is stable under a wide range of conditions (figure 2.2.3-1) and
isolates the substrate from the environment. This effect has been of interest for ASS components
exposed to carburising gas mixtures, either in service [97, 98] or for SE purposes [99]. In the latter case,
the oxide layer impairs diffusion of the hardening elements and, consequently, needs to be removed by
applying some sort of surface activation process prior to the SE treatment [100, 101].
Furthermore, traditional SE treatments are conducted at high temperatures, around 500ºC - 600ºC in
the case of nitriding, and 900ºC - 1000ºC for carburising [102, 103]. At these temperatures, and with
increasing availability of nitrogen and carbon from the hardening medium, profuse precipitation of
ASS. However, low temperature SE diffusion treatments with nitrogen and/or carbon have been
reported to increase the surface hardness without affecting or even improving the corrosion resistance
[103, 104].
It is nowadays widely accepted that hard, wear and corrosion resistant surface layers can be produced
on ASS by means low temperature nitriding and/or carburising in a number of different media (salt bath,
gas or plasma), each medium having its own strengths and weaknesses [105]. In order to retain the
corrosion resistance of ASS, these processes are typically conducted at temperatures below 450ºC and
500ºC, for nitriding and carburising respectively (figure 2.2.3-2). The result is a layer of precipitation free
austenite, supersaturated with nitrogen and/or carbon, which is usually referred to as S-phase or
There is general agreement in the literature regarding the thickness and the composition of EA layers.
Maximum thicknesses in the order of 20 μm and 50 μm have been reported for nitriding and carburising
respectively [107], with 20 at.% to 30 at.% of N and 5 at.% to 12 at.% of C correspondingly [109]. In the
case of nitriding, the layer shows a steep gradient in composition (figure 2.2.3-3), with the associated
abrupt change in properties, whilst the composition depth profile for the carburised layer is smoother
[96, 105, 109]. The deviation of the carbon-depth profile in figure 2.2.3-3 compared to the typical profile
shown in figure 2.2.2-1 is attributed to a variation of the diffusion coefficient with the carbon content
[110].
these carbon and nitrogen supersaturated layers [111-115]. The contradiction arises from a genuine
difficulty in the identification of the X-ray diffraction (XRD) patterns, ascribed to the lattice strains and
high density of crystallographic defects intrinsic to this layer of material. To this end, it is generally
accepted that the so called S-phase is an integral part of the substrate with crystalline structure [111,
114]. The evidence indicates that this layer is single-phased, with the same fcc crystal structure of the
parent austenite, although distorted as a result of a high content of interstitial elements in solution [112,
116]. Nevertheless other crystal structures have not been completely disregarded.
With regard to alloying elements, chromium has been considered crucial for the formation of EA, due to
the very high chemical affinity with carbon and nitrogen [101, 106, 109, 117]. Therefore, interstitial C
and N show a high probability to stay in the octahedral interstices of the fcc lattice, surrounding Cr
substitutional atoms, which form trap sites of potential energy [101]. Molybdenum also seems to be
beneficial to the kinetics of carburising, by delaying the precipitation of chromium carbides and thus
allowing higher treatment temperatures [118], and by expanding the octahedral interstices in the fcc
lattice [119].
From the thermodynamic point of view, EA is metastable and it will decompose into the stable austenite
and carbide or nitride phases upon heat treatment [106, 108, 116, 120]. The incubation time for
decomposition is temperature dependent, and nitrogen EA decays faster than carbon EA (figure 2.2.3-
2), so that for prolonged service time, maximum operating temperatures in the order of 200ºC and
the fact that at the low temperatures at which the treatment is conducted, carbon and nitrogen exhibit
diffusion rates significantly higher than the almost immobile substitutional elements, Cr and Ni. Under
Detailed studies and observations with transmission electron microscopes (TEM) have revealed the
precipitation of fine nitrides and needle-like carbides, especially in the outmost layer or close to
structural defects in the microstructure [116, 121, 122]. Calculations of the volume misfit indicate that
the strain energy associated with the precipitation of carbides contributes to keep the carbon in solution,
although at high levels of supersaturation and expansion of the austenite cell, this energy barrier is
The reported nitrides were of the MN type, while the carbides were of the M5C2 χ-kind (where M stands
for metal and is most commonly Cr). In both cases, precipitation seems to follow a very specific and
{001}γ || {001}CrN
<001>γ || <001>CrN
{111}γ || {001}M5C2
<211>γ || <010>M5C2
partitioning between the carbide and the austenitic phase of not only carbon, but nickel as well [123].
supersaturation and expansion of the austenite has been reached, thus reducing the volume misfit
strain energy barrier [123]. At the same time, the observed orientation relationship enables precipitation
of needle-like carbides with highly coherent interfaces, minimising the long range transport of misfit
Tensile tests have shown only minor differences between the untreated and treated ASS, some of the
differences arising from the thermal cycle rather than the SE treatment itself [110, 124]. These
differences can be summarised as a small reduction in the yield strength and in ductility, with a minor
increase in the ultimate strength and elastic modulus; and they were attributed to the existence of high
compressive residual stresses on the treated layer, and to a strain ageing effect produced at the
treatment temperature.
The high compressive residual stresses present on the EA layers predicted obvious increases in the
fatigue life and, in general, improvements of about 25% have been reported [125]. This is the result of a
change in the failure mode, as the hard EA layer suppresses the alternating slip deformation at the
surface [110, 124]. Consequently, the crack initiation site moves inwards, to the interface between the
substrate and the EA layer, which can be explained by the strength-applied stress diagram in figure
2.2.1-1 [126]. It must also be noted that a minor increase in the notch sensitivity has also been reported
The improvement in wear resistance is perhaps the most outstanding feature of EA. The degree of
improvement depends on the sliding conditions, but volume losses between one and two orders of
improvement is attributed to the increased surface hardness, with a typical ratio 4:1 compared to the
untreated ASS [128]. The EA layer prevents the surface from undergoing plastic deformation, and
changes the wear mechanism from adhesion and abrasion, to a mild oxidational wear regime [128,
129]. However, under heavier loads, deformation of the subsurface occurs and leads to catastrophic
failure, through propagation of subsurface cracks and spallation of the EA layer [130]. In this way, the
carbon EA layers, being thicker and tougher than their nitrogen counterparts, show some advantage.
With regard to corrosion, the results vary significantly depending on the testing conditions. Surprisingly,
most researchers found that low temperature nitriding and/or carburising do not harm the corrosion
resistance of ASS, or even improve it. No conclusive explanation has been found for this improved
corrosion behaviour, although it is clear that the benefit stands as long as nitrogen and carbon remain in
solution and EA is free of precipitates [120]. In NaCl solutions, it is generally reported that EA remains
passive under similar or wider range of potential compared to the untreated ASS, carbon EA showing a
marginal advantage over nitrogen EA [96, 131, 132]. Similar or slightly higher initial current densities
have usually been measured on EA, together with the absence of pitting potential, in contrast to what is
usual for ASS [96, 125, 133]. Regarding repassivation, the evidence indicates that the passive film heals
This attractive combination of properties has caught the attention of both research and industrial fields,
which is reflected in an increasing number of technical papers, reports, patents and processes. The
applications include hydraulic pump components, valves, ball bearings, ferrules, etc. [135-143]. In the
As it was previously mentioned, there are a number of different media capable of delivering carbon or
nitrogen to enrich the surface of steel components, although not all of them are suitable for treating ASS
or producing EA. Methods such as pack carburising and salt bath carburising or nitriding strongly
depend on temperature to control the process driving reactions, and their effectiveness is low at the
temperatures required to produce EA (T < 500ºC), especially in the case of carburising [88, 93].
Moreover, they have been widely superseded by gas media because of technical and environmental
reasons.
Gas carburising with endothermic gas is, nowadays, the dominating carburising technology [144] and it
has been successfully employed to produce EA [145, 146]. The main chemical reactions are as follows
[93, 94]:
CH 4 ↔ (C ) + 2H 2 (2.2-4)
CO + H 2 ↔ (C ) + H 2O (2.2-6)
CO 2 + H 2 ↔ H 2O + CO (2.2-7)
CO 2 + CH 4 ↔ 2CO + 2H 2 (2.2-8)
H 2O + CH 4 ↔ CO + 3H 2 (2.2-9)
The carburising potential of the atmosphere (KC) is usually controlled by reaction (2.2-6) according to
where pCO, pH2 and pH2O are the partial pressures of each of the gases involved in the reaction.
There is some degree of secrecy about the available commercial processes to produce EA, but the vast
majority of them are based on gaseous atmospheres containing CO [135, 145, 147]. The results are
satisfactory, and the only drawback of these processes is the need for a prior surface activation to
remove the passive oxide film, which is usually accomplished by using halogenated gases [100, 147,
148]. The low working pressures associated with vacuum carburising are known to remove the oxide
layer from the treated components, although they require high temperatures to do so (900ºC to 1100ºC)
and to activate the process reactions, i.e. thermal decomposition of the hydrocarbon gas under the
catalytic effect of the hot steel surface [93, 149-152]. Another alternative is plasma SE, which removes
the passive oxide film from the surface of ASS components through the sputtering action of highly active
species present in the plasma [88]. Being the subject of this project, they will be treated in further detail
Plasma surface engineering (PSE) is based on the abnormal glow discharge phenomenon that takes
place on parts subjected to cathodic potential (several hundred volts), while being immersed in a low-
pressure gas atmosphere (rough-medium vacuum) [153-157] (figure 2.2.5-1). Active species (i.e.
excited or ionised atoms and molecules) are generated within the plasma and accelerated towards the
cathodic worktable, where the parts to be treated are placed. While most of the energy involved in this
process is released in the form of heat, a small part of it allows nitrogen and/or carbon species to be
PSE the reactions are under electric control, and the conventional equations to calculate the potential of
the atmosphere do not entirely hold [152, 158]. The number of process variables increases:
temperature, time, gas composition, pressure and flow rate, voltage, current, boost-diffusion cycle time
and ratio, just to mention some. This larger number of relatively independent variables makes it more
difficult to control and optimise the treatment, but also permits a higher degree of freedom when
As previously stated, the process variables are not completely independent of each other, and their
interactions are of significant importance. The gas pressure affects the mean free path of the active
species in the plasma [159, 160], and their kinetic energy is proportional to the acceleration produced in
the cathode fall, i.e. the applied voltage [154, 155]. These two variables define, to some extent, the
interaction between plasma and substrate, and the relative importance of some of the phenomena that
take place during the process: sputtering, edge effect and ion implantation [161].
Plasma nitriding (PN) is usually performed in N2-H2 gas mixtures or NH3. On the other hand, CH4-H2
mixtures are mainly used for plasma carburising (PC) purposes, although other carbon containing
gases have proved to be useful in some cases [162]. The dilution of the gas is one of the variables that
determine the flux of C or N transfer into the solid state, i.e. the rate of saturation of the surface [163].
Hydrogen additions are convenient for surface depassivation, especially when treating ASS. Moreover,
the hydrogen content of the gas mixture also increases the tolerance of the process to possible air
leaks in the chamber [164]. Small amounts of argon might be added to increase the sputtering effect
current density, the gas mixture and gas pressure, and even the shape of the components [166]. For
treatments conducted at temperatures below 570ºC, sputtering can be the only source of heat, whilst
additional heating is often required or convenient for higher temperatures [155, 167]. This is because in
order to reach high temperatures by means of sputtering only, very high electric potentials are required,
making the process unstable (arcing transition). However, it was found that much lower power densities
are required for the activation of the treatment reactions, while the temperature can be reached using
Decoupling temperature from electric plasma activation gives more freedom in the selection of the
treatment conditions and better process stability and control [170-172]. Even though the activation
provided by plasma during PSE is obviously responsible for the high rates of carbon and nitrogen mass
transfer to the surface of the sample, the mechanisms involved are still not clear, and the models
proposed in the literature have been severely disputed. Once the surface content reaches the
saturation limit, the diffusion of N and C into the bulk material is a function of temperature, time and
chemical driving force [163]. Depending on these three parameters, the final thickness of the alloyed
layer and concentration gradient of the diffusing elements can vary from a few micrometers up to
Direct current (DC) glow discharge thermochemical processes experienced a rapid uptake in industry in
the 1970s and 1980s [161]. The advantages over conventional thermochemical treatments include:
shorter cycle times, higher energy efficiency, lower gas consumption and environmental friendliness,
just to mention some [156, 163, 170, 173, 174]. However, the higher equipment costs, the lack of
thorough understanding of the process and such intrinsic problems of DC glow discharge plasma as
Moreover, the homogeneity of the temperature within the chamber was not satisfactory for some
applications. Different methods were attempted in order to improve this aspect, making use of external
heat sources like electric resistors or cathodic heaters [168]. These are metal sheets or meshes,
subjected to cathodic potential, which are heated by the ionic bombardment and radiate heat to the rest
of the chamber, thus heating the colder parts of the workload. A further improvement of this concept,
together with the idea of post discharge plasma, became a novel plasma technology in 1999, named
active screen. Active screen plasma (ASP) shows great potential to overcome the problems associated
Among the plasma processes alternative to the DC technique, ASP is of particular interest due to the
simplicity and effectiveness of its concept. In a conventional DC furnace, the worktable is negatively
biased with respect to the wall of the chamber, which is earthed and plays the role of the anode in the
electric circuit. If no supplementary heating elements are used, the workload needs to be subjected to
relatively high electric potentials, in order to reach the treatment temperature by means of sputtering.
The high electric potentials applied imply that the working condition moves closer to the arcing transition
point in figure 2.2.5-1, and renders the treatment unstable, especially when the components are not
Moreover, components with complex geometries and of different sizes may not be heated uniformly,
because of the distortion of the electric field in proximity of sharp edges [176]. The distortion of the
so that parts with small radius of curvature treated in DC furnaces are prone to the so called edge
effect. The local temperatures developed at the edges can be much higher than in the rest of the
At the same time, hollow cathodes are the third type of defects frequently encountered in DC plasma
treatments, and blind holes or dense work loads are vulnerable to this effect [169]. Hollow cathodes
take place when two surfaces exposed to the plasma are separated by a certain distance, depending
on the gas mixture, pressure and applied potential. These process variables define the thickness of the
plasma sheath, as it will be explained in the next section. The overlapping of the plasma sheaths
produces a peculiar electric field that traps electrons, which oscillate between the cathodic surfaces,
increasing the ionisation events and producing a local increase in current density, with the consequent
increase in temperature.
In the case of the active screen technology, the workload is surrounded by a mesh, which plays the role
of the cathode (figure 2.2.6-1). The purpose of the mesh or active screen is twofold [177]:
It radiates heat to the work load: the cage, being the cathode of the circuit, it is subjected to
severe sputtering that increases its temperature; heat is thus transmitted to the workload by
It generates active species, principally excited neutral atoms and molecules, which are believed
The worktable is electrically insulated, or subjected to a low negative bias to ease the flow of active
species to the components [175]. In this way, sputtering does not take place on the surface of the
temperature is more uniform and stable. Thereby, defects associated with direct exposure of the
workload to highly negative potentials are reduced or eliminated. In practice, the AS technique has
proved to be less sensitive to grease and rust on the treated parts [178]. From the aesthetic point of
view, better surface qualities can be obtained, with complete absence of edge effects [179]. Finally, the
local increases in temperature are reduced, producing more even hardening effects and properties
[180].
Apart from its technological importance, AS has raised the unresolved issue of the nitriding and
carburising mechanisms. Much work has been done on AS plasma nitriding (ASPN) in laboratory scale
units, and it was concluded that the principal mass transfer mechanism involved sputtering of iron from
the mesh, reaction of the iron atoms or particles with the active nitrogen in the plasma, and deposition
of the iron nitrides on the worktable with subsequent decomposition and diffusion of free nitrogen [181-
183]. However, the experience on industrial scale units has shown that other mechanisms and effects
must be involved in the process [184, 185]. Furthermore, little work has been conducted up to date on
It is evident that the mechanisms of plasma treatments pose a difficult challenge, which involves not
only metallurgical aspects but also plasma-surface interactions and plasma physics. Therefore, the next
section will introduce some basic but useful concepts of plasma physics and the plasma diagnostic
techniques.
3.1. Plasma
Plasma is usually known as the fourth state of matter and its inception requires a significant amount of
energy, enough to break some atoms and molecules down into positive and negative particles, i.e. ions
and electrons [186]. The amount of charged particles is balanced, so that the system remains
electrically neutral. Contrary to the other phase changes, the transition of matter to the plasma state
takes place gradually over a wide range of conditions, leading to different degrees of ionisation [187].
The energy required to ionise a gas can be applied to the system in the form of heat, although this is
impractical due to the extremely high temperatures involved [188]. On the other hand, cold plasma can
The study of electric discharges through gases dates back to the 18th century and involves famous
researchers such as Faraday, Crookes, Thomson, Langmuir and Stark, just to mention some [186].
Although gases are good insulators under standard atmospheric conditions, they undergo ionisation
when a strong electric field is applied, thus becoming electrically conductive [189]. A small fraction of
charged particles from cosmic or natural radiation is normally present in the system. These particles are
accelerated by the electric field and acquire sufficient energy to collide inelastically with gas molecules,
ripping electrons or attaching to them to form positive and negative ions, respectively. Direct electron
impacts can also excite the molecules to energy levels higher than the ground state but below the
ionisation threshold.
Relatively few collisions occur in a cold plasma formed at low pressure, thus the system is not in local
thermal equilibrium (LTE) [187]. This means that some highly energetic (fast or hot) particles coexist
with others significantly less energetic (slow or cold) ones. The non-LTE is one of the most attractive
conditions, far from equilibrium, depending on the gas composition, pressure, electromagnetic field,
Electrons are significantly lighter than neutral molecules or ions and, consequently, they are highly
mobile. They respond faster than ions to fluctuations in electric and magnetic fields, and acquire higher
energies. Even though most applications rely on the kinetics of ions, electrons are the primary agents in
a plasma and their energy distribution is of significant importance [187, 190]. The energy or velocity of
3
⎛ me ⎞ 2 2 ⎛ m v2 ⎞
f (v ) = 4πne ⎜⎜ ⎟⎟ v exp⎜⎜ − e ⎟⎟dv (2.3-1)
⎝ 2πk BTe ⎠ ⎝ 2k BTe ⎠
where f(v) is the velocity distribution function according to Maxwell, ne is the electron density, me is the
electron mass, kB is the Boltzmann constant, Te is the characteristic electron temperature and v is the
velocity [190]. Equation (2.3-1) represents a distribution where most electrons (bulk) have relatively low
energies, whilst there is a small fraction (tail) of highly energetic electrons (figure 2.3.1-1).
The density of electrons, or other charged particles, is commonly described by the Boltzmann equation:
⎛ e ⋅φ ⎞
ne = n0 exp⎜⎜ ⎟⎟ (2.3-2)
⎝ k BTe ⎠
where ne is the electron density in any position of potential Φ, n0 is the electron density at a point of zero
potential, e is the elementary charge, Φ is the potential at the position of interest, kB is the Boltzmann
The electric brakedown of a gas at low pressure involves several steps, as it can be seen from figure
2.2.5-1. In the first regime, the gas is a good insulator and only the charged particles from background
radiation carry a very small current [192]. When increasing the voltage, the current increases mainly
through acceleration of these charged particles, until a saturation value is reached, depending on the
gas composition and the pressure. If the voltage is raised further, the electrons gain sufficient energy to
collide with the gas molecules inelastically, producing more electrons in an avalanche effect. This
If the voltage is increased further, breakdown occurs and the current increases whilst the voltage can be
reduced markedly. The electric breakdown of a gas is an irreversible process. The ion bombardment on
the work table releases secondary electrons [190, 192], which subsequently multiply through the
forementioned avalanche effect, thus increasing the availability of charge carriers considerably. This
stage receives the name of normal glow discharge, due to the luminous phenomenon observed on the
cathode (emission of light from the decay of excited and ionised atoms or molecules). The current
density remains constant and any increase in current is a consequence of an increase in the area
Once the cathode is completely covered by the glow, a new discharge regime begins, named abnormal
glow discharge. This is characterised by an increase in voltage with increasing current, the latter being a
consequence of a higher current density as the whole surface of the cathode is already involved in the
discharge [192, 193]. It is in this regime that PSE treatments are conducted, thus ensuring that all the
surface of the cathode (i.e. all the components) are exposed to the glow and to the plasma-surface
interactions. The selection of the current density, within the abnormal glow discharge range, depends on
arcs. Arcing involves thermal emission of electrons from hot areas of the cathode, most frequently sharp
asperities, although it also occurs on non-conductive pollution or dirt by the accumulation of electric
According to Paschen’s law, for a given gas and electrode material, the breakdown voltage is a function
of the pressure p and the distance between the electrodes L [192]. This correlation is illustrated in figure
2.3.1-2. On the left side of the minimum point (pL < pLmin), low pressure insulation occurs as a result of
the low density of gas molecules and low frequency of ionising collision; whilst on the right side of the
minimum point (pLmin < pL) high pressure insulation is a consequence of the higher energy losses
experienced by electrons and ions in frequent collisions, mainly with neutral gas molecules [189]. Thus
pLmin can be understood as the combination of pressure and distance between electrodes that yields the
When a direct current (DC) glow discharge forms between two electrodes, the potential difference is
concentrated close to the cathode, and several different regions are observed (figure 2.3.3-1) [187]. The
first region in front of the cathode is the Aston dark space, a thin layer with a high density of electrons
and very strong electric field, where electrons are still being accelerated and have not yet reached
energy levels high enough to produce inelastic collisions with atoms and molecules, thus making this
region appear dark [157]. Immediately next to the Aston dark space is the cathode glow, with a relatively
high density of ions and, in some cases, excited atoms sputtered from the cathode, which account for
the observed glow. The following is a relatively dark region called cathode dark space, although it is also
These regions are not always distinguishable and, in some cases, one of them may mask the others
[157, 192]. In general, the space between the cathode surface and the end of the cathode dark space is
called cathode region or sheath [187]. The difference in potential is mostly concentrated in this part of
the discharge, and the thickness of the sheath will adjust itself to pLmin, following Pashen’s law [157].
The electrons are accelerated across the sheath and acquire energy levels high enough as to excite
and ionise gas molecules in the neighbouring area, called the negative glow, or even beyond.
The negative glow is relatively thick compared to the sheath, and is the brightest region of the
discharge. Because of the high density of ions, the negative glow has a positive space charge, and the
electric field in this region is relatively weak. The glow is more intense on the cathode side, where
electrons are energetic enough to produce excitation and ionisation, and fades away towards the anode
as the electrons slow down due to frequent collisions [157, 187]. The drop in the energy of electrons
reduces the ionisation and excitation events and gives rise to another region, called Faraday dark
space. The space charge in this region is small and the electric field is slightly negative.
The remaining space between the electrodes corresponds to a region of uniform glow called the positive
column [157]. If the distance between electrodes is increased, the positive column lengthens while all
the rest of the regions remain unchanged [192]. The electric field in the positive column is slightly
negative and uniform to lead the electrons towards the anode, and the space charge can be considered
quasi-neutral [187, 194]. Approaching the anode, a luminous region may be seen, called the anode
glow, and immediately next to it a dark zone named anode dark space or anode sheath. The latter is the
This brief summary on DC glow discharges serves as an introduction to the plasma diagnostic
techniques of interest to this project, namely optical emission spectroscopy (OES) and electric probes,
Plasma diagnostic techniques are of vital importance for the understanding, control and optimisation of
plasma processes [195]. The working principles are varied (optic, electric or magnetic) and the
information they obtain is often complementary, so that it is interesting to use more than one technique
at the same time. The local disruption produced by the probes is critical and, although the devices and
working principles are sometimes quite simple, the analysis of the signal usually raises difficulties [196].
The plasma parameters of significance to PSE processes can be summarised as follows [197]:
Type, density and energy of active species, i.e. ions, excited atoms and/or molecules.
Electromagnetic fields, which define the movement, spatial distribution, and time response of
Plasma diagnostic techniques have long been used to further understand the physics of plasma [198,
199]. In addition, once the underlying principles are known, they can also be used to provide real time
information about the process, and to feed control systems [171, 200, 201]. Optical emission
spectroscopy (OES) together with electric probes or collectors, are possibly the most extended plasma
As it was stated in a previous section, inelastic collisions take place within the plasma, mainly through
direct electron impact, producing ionisation and excitation of the gas atoms and molecules [194]:
−
efast + M → M + + eslow
− −
+ eslow (2.3-3)
− −
efast + M → M * +eslow (2.3-4)
where M is an atom or molecule and the superscripts + and * indicate the ionised and excited conditions
respectively, e stands for electron and the subscripts slow and fast indicate the energy of the particle. It
is evident that both M+ and M* are at higher energy levels than the atom or molecule at its ground state,
i.e. its normal equilibrium condition. The energetic species M+ and M* have relatively short lifetimes and
they experience decay processes, which frequently involve the emission of a photon to release the
M * → M + hυ (2.3-5)
where hν is the energy of the emitted photon, being h the Planck constant and v the frequency of the
photon. It is known from the atomic theory that the electron configuration and electronic energy levels
are quantised, i.e. they do not assume any value, but certain allowed energy states [203].
Consequently, if the atomic and molecular energy states are quantised, the energy involved in their
transitions must also be quantised; and this is the case of spontaneous emission, so that the energy of
where Ek is the energy level associated with the energetic states M* or M+, and Ei is the energy level at
In this way, each decay transition emits a photon with certain energy, i.e. at one particular frequency,
giving rise to an emission pattern or spectrum, which can be used to identify the emitting species.
Moreover, the intensity of every emission line is proportional to the number of molecules at each energy
level, and the emission intensity has been used to correlate changes in the dominant ground state with
a treatment parameter [197]. However, quantitative information and absolute concentrations are difficult
The emission lines observed in atomic spectroscopy are due to the electronic transitions, i.e. changes in
the electronic configuration of the excited atoms. In addition to this, polyatomic molecules show a more
complex and finer structure, with different vibrational and rotational energy levels as a result of the
relative movement of the atomic nuclei in the molecule [202]. Although interactions between electronic,
vibrational and rotational energy levels exist, the energy difference is so large that, according to the
Born-Oppenheimer approximation, each level can be considered separately and superimposed to the
others, as shown in figure 2.3.4-1 and represented by the following equation [204]:
ΔE = ΔE E + ΔEV + ΔE R (2.3-7)
where EE, EV and ER are the electronic, vibrational and rotational energy levels respectively.
proposed by Heisenberg indicates that each transition involves a slightly different energy, thus emitting
photons at slightly different frequencies, which account for a natural line broadening [202]. The natural
line width is insignificant compared to the effect of Lorentz and Doppler broadening. The former is a
consequence of the collision of emitting species with other particles and the latter is associated with the
motion of the species while emitting light [131]. Moreover, the experimental broadening of the
The kit for spectroscopic observations typically includes an input optic device (telescope, collimator, or
optical fibres), which gathers light (figure 2.3.4-2). Optical fibres are frequently used to feed the light into
a monochromator, a polychromator or any similar equipment [205], where entrance slits or apertures
define the throughput and bandpass of the system [206]. The light is thus focused onto a wavelength
dispersing element, such as a grating or a prism, and a narrow range of wavelengths is reflected onto a
detector, such as a silicon photodiode or a PMT [207]. The collected spectrum shows the intensity of the
OES is a popular diagnostic technique due to its simplicity, and also because it causes no disruption to
the plasma. It provides useful information on the type and, to some extent, the population of active
species in the plasma. High temperature optical fibres yield good spatial resolution, and have been used
in optical pyrometers to collect light from specific areas of interest [208, 209]. Although valuable, the
information obtained with OES is rather incomplete, and electric probes are necessary to obtain
evidence on the energy and density of electrons [197]. This will be discussed in the next section.
In order to understand the information obtained from electric probe devices, it is useful to make some
brief comments on the electric potential and mobility of charged particles in the plasma. It must be said
that negative ions and double ionisation events are very scarce in the type of plasma of interest to this
project, so that only electrons will be considered as negative charge carriers, and singly ionised atoms
and/or molecules as positive charge carriers [191]. Moreover, the probability of electron-ion
recombination in the volume (equation 2.3-8), i.e. the reverse of electron impact ionisation (equation
2.3-3), is rendered low by the necessary energy and momentum exchange among three particles [194]:
−
eslow −
+ eslow + M + → M + efast
−
(2.3-8)
where M stands for atom or molecule and e for electron, while the fast and slow subscript indicates the
energy or temperature of the particle. Volume recombination events are particularly unusual in gases at
low pressure, thus most of the recombination processes occur at the wall of the vessel that confines the
As it was mentioned in a previous section, the potential between the electrodes is not evenly distributed,
but it is related with the density of charged particles, according to Poisson’s equation [210]:
d 2φ 4πe
= (ne − ni ) (2.3-9)
dx 2 ε0
where Φ is the potential function, x is the distance from the cathode, e is the elementary charge, ne and
ni are the density of electrons and ions respectively and ε0 is the permittivity of the free space. In the
accumulation of ions is rapidly compensated and levelled by the highly mobile thermal electrons [191].
The accumulation of charges in the area surrounding the electrodes, the walls or any other foreign body
immersed in the plasma, produces a shield that confines the perturbation to a distance λD, according to
ε 0 k BTe
λD ≈ (2.3-10)
ni e 2
where λD is the distance of the perturbation or Debye length, ε0 is the permittivity of the free space, kB is
the Boltzmann constant, Te is the electron temperature, ni is the ion density and e is the elementary
charge.
The flow of ions and electrons is governed by diffusion and by the electric field drift, according to the
∂n
Γ = −D + nμE (2.3-11)
∂x
Where Γ is the flux of charges, D is the diffusion coefficient, n is the density of charges, x is the distance
from the cathode, μ is the mobility of the charged particles and E is the electric field.
The flux of positive and negative charges out of the system must remain balanced in order to maintain
its electric neutrality. Moreover, if an insulated electrode is immersed in the plasma, the same amount of
positive and negative charges should arrive at its surface, so that no net flow of charges takes place
[211]. However, as the electrons are highly mobile compared to the ions, the flux of electrons to any
neutral surface exposed to the plasma would be larger than the flux of ions. Consequently, any foreign
electrons, and thus balance the flow of negative and positive charges. This potential is called floating
potential, and depends on the type of gas, the pressure, and the electric discharge parameters.
Electric or electrostatic probes, frequently called Langmuir probes or collectors, are electrodes of known
area and defined shape (figure 2.3.4-3), usually made of tungsten, molybdenum or graphite, and
supported by an insulating ceramic or glass holder [196, 198]. The probes are immersed in the plasma
and electrically biased, sweeping a range of electric potential with respect to the plasma [212]. The I - U
curve is recorded and, providing the experimental settings and conditions are carefully selected and the
correct model is used to evaluate the signal, it gives valuable information on the characteristic energy
Figure 2.3.4-4 shows the typical curve of a Langmuir probe. When the electrode is negatively biased,
the electrons are repelled and only ions are allowed to arrive at the surface of the probe [211]. If the
negative charge is not excessive, the ion flux is independent of the applied voltage, and it can be
measured from this part of the curve, called ion saturation region. On the contrary, when the probe is
positively biased, only electrons are allowed to reach the surface of the probe, thus the electron density
can be calculated. This part of the curve is called electron saturation region [196]. Between these two
extreme cases, the net flow of charge combines both electrons and ions, and the transition points give
To this end, an improved version of the planar Langmuir probe or collector was designed to reduce the
effect of deposition or pollution on the effective area of the electrode [214, 215]. The working principle is
illustrated in figure 2.3.4-5. A capacitor is connected in series with the probe, and it is negatively biased,
interrupted and the capacitor is left to discharge by the arrival of positive charges from the plasma, i.e.
ions. The drain curve of the capacitor is recorded and gives very similar information to the ion saturation
dV
C⋅ = [Γi − Γe ] ⋅ eA (2.3-12)
dt
where C is the capacity connected to the probe, V is the potential across the capacitor, t is the time, Γi
and Γe are the flux of ions and electrons respectively, e is the elementary charge and A the effective
This method is simple, and yields information with minimum signal post-processing. However, care
should be exercised to avoid edge effects on the probe, and to apply the right level of bias to produce
1. Sample preparation
Coupon samples were cut from AISI 316 hot rolled bars (composition shown in table 3.1-1), using a high
precision Struers Accutom-5 cutting machine, fitted with SiC cutting discs. The final dimensions of the
samples were 25.4 mm in diameter and 6 mm in thickness. Both faces of the samples were wet ground
with SiC grinding paper of grit #120 until flat, and one face was ground further up to grit #1200. Samples
were then washed with soapy water and cotton, degreased with acetone in an ultrasonic bath for 10
Some samples were mirror polished with diamond pastes (6 μm and 1 μm) and colloidal silica
suspension, on a Struers DAP-7 automatic polisher. The cleaning procedure after polishing was
Cross sections were cut from the samples with SiC cutting discs on a Struers Accutom-5 cutting
machine, nickel plated when necessary, and mounted on conductive bakelite using MetPrep PA30 or
Opal 400 mounting presses. The grinding, polishing and cleaning procedures were equivalent to the
Whenever it was required, the microstructure was developed by means of an etching solution of
composition 25 ml H2O, 25 ml HNO3 and 50 ml HCl. The sample of interest was immersed in the
solution at room temperature for a few tens of seconds, and immediately rinsed with water and dried
studies. The sample preparation procedure was equivalent to the one described before, with a careful
final polishing with colloidal silica to obtain a smooth and stress-free surface.
These samples were cut from an AISI 316 bar, 4½ inches (114.3 mm) in diameter, using an automatic
cutting saw. Both faces were finely machined until flat and one face was wet ground with SiC grinding
paper up to grit #1200. The final dimensions of the samples were: 114 mm in diameter and 12 mm in
thickness, with a 16 mm through hole at its centre for holding purposes. The cleaning procedure after
Samples were produced from an AISI 316 bar, 10 mm in diameter, for immersion corrosion
experiments. Rods 50 mm long were cut and an effective area 40 mm long and 8 mm in diameter was
finely machined and dry ground with SiC emery paper up to grit #1200. A flange at one end of the rod
(10 mm long and 10 mm in diameter) remained as rolled and was used for holding purposes (figure
3.1.3-1). These samples were subjected to the standard cleaning and degreasing procedures already
explained.
Coupon samples were cut and back thinned with SiC grinding paper until less than 200 μm thick,
leaving the treated surface untouched. Discs 3 mm in diameter were punched and further back thinned
with emery paper, until they were between 50 μm and 80 μm in thickness. Dimpling was conducted from
the back side of the sample in a South Bay Technology 515CE precision dimpling machine, using a
Finally, the samples were mounted on copper discs, using a hot melt adhesive, and ion milled in a
Gatan 691 Precision Ion Polishing System (PIPS) until perforation occurred. The two Ar guns were set
at angles between 8º and 10º to ion mill the sample from a single side, i.e. the side opposite to the
surface treated layer. The ion current was between 8 μA and 14 μA at 4 keV, and the sample was
constantly rotated. Under these conditions, perforation happened within 2 hours. Samples were then
removed from the holder and the remaining hot melt adhesive was cleaned using acetone.
Some powder samples and wear debris were also prepared for TEM observation. The sample
preparation was, in these cases, very simple and consisted only in adding methanol to the glass
container where the samples were kept. A pipette was then used to select particles of suitable size,
Thin discs, 25.4 mm in diameter and 0.5 mm in thickness, were cut from AISI 316 bars under carefully
controlled conditions, in order to obtain flat and parallel surfaces. Both sides of the discs were wet
ground with SiC paper, on a flat glass surface, up to grit #1200, and vibration polished with 6 μm
diamond paste for several hours (approximately 12 hours), until no traces of the grinding marks could be
observed. The previously described cleaning and degreasing procedures were applied and care was
taken to avoid stresses that could bend the discs. Finally, the discs were identified and their thickness,
roughness and flatness (in different directions) were measured in order to select the most suitable ones
for the assessment of the residual stresses developed during the surface engineering treatment.
The sample preparation methods here explained cover most of the types of samples of significance to
this project. Some other kinds of samples of different materials were eventually used: pure iron stripes,
AISI 304 stripes, glass slides, copper coupons, just to mention some. The sample preparation
procedures used in all these cases were very similar to the ones previously explained, and the
Three different plasma furnaces were used for surface engineering treatments (figure 3.2.1-1):
DC Klöckner Ionon 40 kVA unit for plasma carburising in DC and AS experimental arrangements
DC Klöckner Ionon 60 kVA unit for plasma nitriding in DC and AS experimental arrangements
AS Plasma Metal 75 kVA + 15 kVA industrial scale unit for AS plasma nitriding and carburising
The instrumentation was different for each furnace although, in general, it included mass flow meters to
control the gas mixture and gas flow, two pressure gauges of different kinds (Penning, Pirani or
Bourdon) to measure gas pressure, two K-type thermocouples protected by ceramic sheaths, and a
With regard to the procedure for plasma treatments, it could be generally described as follows:
The samples to be treated were cleaned and placed on the work table inside the furnace. In some
cases, a special holder or rig was required to keep the samples in position, for mechanical masking or to
produce the desired arrangement inside the chamber. Once the samples were in place, a vacuum
cleaner was used to remove dust particles that could have been deposited on the surface.
and the vacuum pump was turned on to evacuate the furnace, up to a vacuum level in the order of 1 Pa.
Pure hydrogen gas was then fed into the chamber, and the main power supply was turned on to start
the plasma.
During the initial stage, the electric power (current and voltage) and the gas pressure in the chamber
were increased in steps, as a function of the temperature and the stability of the treatment (onset of
arcing), and the limit of the heating stage was usually set at 300ºC. From this point onwards, the gas fed
into the furnace was not pure H2 any more, but the gas mixture selected for the treatment, i.e. N2-H2 for
nitriding and CH4-H2 for carburising, or CH4-H2-N2 for nitrogen enhanced carburising. All the remaining
treatment conditions were automatically adjusted until the set point values could be reached
(temperature, pressure, gas composition and bias†) and these set point values were kept stable
Once the treatment had finished, the feeding gas valve was closed and the chamber was evacuated.
The samples were left to cool down, under the residual treatment gas pressure, until below 80ºC. The
furnace was then ventilated and opened to remove the samples from the chamber. It was a normal
practice to clean the samples with soapy water and cotton immediately after the treatment, and this will
The principal variables studied here were pressure and bias, although some other treatment conditions
were also assessed, with regard to gas composition (N2 enhanced carburising), treatment time and also
active screen treatments with different arrangements (figures 3.2.1-2 and 3.2.1-3). The process
†
Bias, as a set point, only applies to the active screen furnace and not to DC furnaces.
studied, and it has been found to be limited by metallurgical aspects of the AISI 316 alloy, namely the
Therefore it was the objective of this study to optimise the plasma treatment conditions for a given
treatment temperature in order to produce a thick and uniform expanded austenite layer on AISI 316
samples free from surface defects, such as edge effects and arcing damage; and to increase the
understanding of the mechanisms of active screen treatments, and the main variables involved in the
3. Characterisation techniques
3.1. Hardness
Low force hardness and micro hardness tests were conducted in a Mitutoyo MVK-H1 durometer fitted
with Vickers or Knoop indenters. The load range was 10 g – 1000 g and calibration block #EP9716400
(757.3 HMV0.3) was used to check the condition of the equipment before each measurement. The
reported values are the average of at least six significative measurements and standards BS EN ISO
6507-1, 6507-2, 4516 and 4970 were taken as reference for these tests.
A Micro Materials nanotest system was used to conduct instrumented hardness tests and assess the
mechanical properties of the outmost layer of treated samples (BS EN ISO 14577-1 and 14577-4). Eight
indentations were conducted in each case under the testing conditions shown in table 3.3-1.
Standards BS EN ISO 4287 and 1134-1 were taken as reference for these measurements. Different
pick up settings were selected depending on the objective of the measurement (table 3.3-2).
Contact mode atomic force microscopy was conducted in a Digital Instruments Dimension 3100
nanoscope, using Veeco NanoScope V.5.12r3 software and Veeco NP gold coated silicon nitride
A Philips X’Pert diffractometer was used to obtain the X-ray diffraction patterns (table 3.3-4). The
identification of the diffraction peaks was achieved by comparison with an XRD pattern database
A Leco GDS-750 unit was used to obtain the composition-depth profiles of plasma surface treated
samples. The method and measurement programme were calibrated using standard blocks of known
Optical metallographic observations were conducted in a Leica Leitz DMRX microscope equipped with a
Different scanning electron microscopes were used throughout the project, and for different purposes,
Oxford Instruments Inca EDX detectors were used in all cases, controlled by means of Inca Energy
software. The internal calibration standards were considered valid for the purpose of this project and
only Ni or Co standards were used to correct the position of those reference peaks. The microscope
settings were optimised for the analysis (acceleration voltage, spot size, working distance, etc.)
Crystallographic orientation maps (COMs) were obtained on a Jeol 7000 FEG SEM fitted with an Oxford
Instruments Inca Crystal Mk1 EBSD camera, at the general conditions listed in table 3.3-5. All the data
were collected and analysed using Oxford Instruments Inca Crystal software for EBSD.
image film (3 ¼ in x 4 in), developed and digitalised with 300 dpi resolution.
A custom-made pin-on-disc wear testing apparatus was used to assess the wear resistance of treated
and untreated samples (figure 3.4-1), and standard ASTM G99-05 was taken as a reference. Coupon
samples, 25.4 mm in diameter and 6 mm in thickness with a surface finishing of SiC paper grit #1200,
were rotated on a horizontal plane and worn against a steady WC-Co binder ball, 8 mm in diameter
(table 3.4-1). The tests were conducted under dry conditions in laboratory air at room temperature; the
relative humidity and room temperature were recorded for each test. The tests were 4.5 hours long, the
applied load was 10 N (Hertzian contact stress 1.6 GPa) and the sliding speed 0.031 m/s, giving a total
number of cycles of 17550 (for the coupon) or a sliding distance of 500 m (for the pin).
The matching surfaces were degreased with acetone, mounted on their respective holders and wiped
again with a clean cloth with acetone, to remove any contaminants from their surfaces. The tribological
couple was gently put into contact, and the load was applied 1 minute before sliding began. After the
test, the samples were carefully removed from their respective holders and observed. The wear debris
was collected for inspection and tested with a magnet. The profile of the wear track left on the tested
samples was measured at not less than 6 different positions, by means of a KozakaLab Surfcorder
SE1700 stylus profilometer (table 3.3-2). The obtained profile was integrated for the whole wear track
The friction tests were conducted on a pin-on-disc custom-made machine, equipped with a 100 lb
(45.36 kg) load cell and digital data acquisition card. Standard ASTM G115-04 was considered as a
reference in this case. The AISI 316 discs, 114 mm in diameter and a surface finishing of SiC paper grit
#1200, were mounted and rotated on a horizontal plane against a stationary 8 mm ball made of WC-Co,
AISI 316 untreated or AISI 316 plasma carburised. Tests were conducted under dry conditions in
laboratory air at room temperature, applying a load of 10 N and a sliding speed of 0.05 m/s. Several
tests were conducted on each disc, so that the rotation speed was adjusted according to the diameter of
The matching surfaces were degreased with acetone and gently put into contact. The data collection
started when the load was applied, 15 seconds before sliding began, and continued throughout the 10
minutes duration of the test, at a frequency of 1 data point per second. After the test, the samples were
separated, removed from their holders and inspected. The wear debris was also collected and tested
with a magnet.
Immersion corrosion tests in different boiling acid solutions were conducted to assess the corrosion
resistance of treated and untreated AISI 316 samples. The equipment consisted of a hot plate as heat
source, a 1 litre volume cylindrical flask where the boiling acid and the samples were kept, a flask lid
solutions were used following British Standards BS EN ISO 3651-1 and BS EN ISO 3651-2:
16% mass fraction H2SO4 with 100 g/l CuSO4 (BS 3651-2 Method A)
40% mass fraction H2SO4 with 25 g/l Fe2(SO4)3 5H2O (BS 3651-2 Method C)
The exposure times to the boiling acid varied from one hour up to 240 hours, the latter conducted in five
The samples to be tested were ultrasonically cleaned in acetone and weighed. The coupon samples
were placed on glass Petri dishes to hold them and keep them separated from one another, and they
were immersed in the acid solution at room temperature. The specimens were placed on top of copper
coupons when copper contact was required (BS 3651-2 Method A). In the case of the rod cylindrical
samples, an especial holder was designed and manufactured to hang the samples leaving only their
Once the samples were placed in the flask, the solution was stirred with a magnetic follower, and heated
up to its boiling point (record was kept of the heating time), and kept boiling for the entire testing time.
Once the test had finished, the heat source was disconnected and the samples were immediately
removed from the boiling solution, rinsed with water and ultrasonically cleaned with acetone. Mass loss
measurements were thus conducted in an Avery Berkel analytical balance, with 0.1 mg accuracy.
These tests were conducted only with a 16% mass fraction H2SO4 aqueous solution at room
temperature and only on coupon samples. The polished samples were masked with fast drying acrylic
lacquer before being immersed in the acid solution for 120 hours at room temperature. After the test, the
samples were removed from the acid, rinsed with water and cleaned with soapy water and acetone to
OES spectra were first collected from outside the furnace, through a sight port window, using an
Optronic Lab OL730Q telescope (table 3.6-1). The telescope was placed 80 cm away from the window
and focused on the mesh of the active screen furnace, which resulted in a minimum source diameter of
20 mm approximately (figure 3.6-1). A cylindrical cone was made with white photographic cardboard
panel, its internal surface being covered with an opaque black liner, and placed between the furnace
and the telescope to reduce the stray light from reflections or other light sources. The use of such a
cone also helped to align the telescope and to keep the measuring distance constant.
A 1000 µm core fibre optic cable fed the light into a Jobin Yvon Triax 180 spectrometer, with a 190 mm
focal length and f/3.9 entrance aperture ratio (table 3.6-2). The optical arrangement consisted of
entrance and exit slits, collimating and focusing mirrors, and a grating turret in a cross Czerny-Turner
configuration. The turret was provided with two diffraction gratings, 150 and 1200 groves per millimetre,
for collection of low and high resolution spectra respectively. The low resolution grating allowed a
wavelength range of 1000 nm, while the high resolution grating admitted a total range of 0 nm to 1400
nm in steps of 100 nm. It must be mentioned that no signal could be collected below 250 nm, due to
A Jobin Yvon Spectrum ONE CCD thermoelectrically cooled detector head was used to measure the
light intensity at different wavelengths. In all cases, sufficient time was given for the detector to reach
the operating temperature before data collection (1 hour approximately). The CCD chip had a sensor
array 1024 pixels wide and 256 pixels high, no binning was applied and a flush cycle was run before
data acquisition, to reduce the noise. Data collection was controlled by a Jobin Yvon CCD-3000 unit, 16
bit ADC limit and 166 kHz conversion speed (table 3.6-3). The controller unit was operated from a PC
The automated entrance and exit slits were kept at the same width for optimum throughput. The
selected slit width was generally 0.04 mm for the measurements conducted with the telescope, and 0.08
mm for the spectra acquired with the optical fibre probe (method described below). The weaker signal
obtained with the optical fibres was counterbalanced by the increased slit width, i.e. increased bandpass
and throughput, although this also had an impact on the spectral resolution.
The system was calibrated, before each set of measurements, using the mercury green emission peak
at 546.074 nm (NIST online database, Pearse and Gaydon reference book) from a commercial
fluorescent lamp. In all cases, the Optronic Lab telescope was used to collect light from the lamp, and a
calibration accuracy of ±0.005 nm was usually attained. The manufacturer’s recommendations were
followed to calibrate the spectrometer, and the conditions used are listed in table 3.6-4.
instrument calibration was not severely affected for one set of measurements, on condition that the
grating turret was not rotated to change the dispersing element. Nonetheless, all the spectra were
collected in the direction of increasing wavelength to minimise errors associated with clearances in the
The collection of light from inside the furnace, or intrusive OES (IOES), was achieved by using high
temperature resistant SiO2 optical fibres, 100 / 110 µm core diameter, numerical aperture 0.22 and UV-
Visible wavelength range (figure 3.6-1). A feedthrough was designed and manufactured to allow one
single fibre to have one end inside the furnace and exposed to the plasma, and the other end plugged
into the same Triax 180 spectrometer described before, by means of standard SMA connector (see
These types of measurements were only conducted in the active screen unit and consisted of the
standard heating stage, until the treatment conditions were reached. Next, spectra were collected
sequentially under different treatment conditions, allowing sufficient time for the work table bias to reach
stable values before each spectrum was acquired. Finally, an emission intensity map was produced with
each set of measurements and each map was produced twice for verification purposes.
A planar electric probe, similar to a planar Langmuir collector, was built to measure the characteristic
parameters of the plasma in both DC and AS furnaces (refer to appendix B for construction details).
Before each measurement, the active surfaces of the probe were polished with colloidal silica, cleaned
with soapy water, ultrasonically cleaned with acetone and dried with hot air. The probe was then
means of a standard electrical feedthrough. All ceramic insulators were wet ground and washed before
each experiment to eliminate the layer of conductive material deposited on them. Finally, the insulation
A low noise Instek GPS-6010 DC power supply with variable output voltage (0 V to 60 V) was used to
bias the probe. A Pico 2202 digital oscilloscope and 250 MHz oscilloscope probe, with an impedance of
50 MΩ and a X10 attenuation power, were used to acquire the drain curve of the tantalum capacitors
connected in series with the probe (UMAX = 63 V, C = 4.7 μF). The oscilloscope was connected to the
USB port of a PC and the signal was monitored and stored in the hard drive of a PC using PicoScope
The experiments in the DC furnaces consisted of a standard heating stage until the selected treatment
conditions were reached, followed by the data collection (approximately 1 hour) and the final cooling
down cycle. Two measurements for each condition were conducted, including DC and AS
arrangements, with standard carburising (CH4-H2) and nitriding (N2-H2) atmospheres, keeping
The measurements conducted in the AS furnace consisted of the usual heating stage, followed by data
collection in each of the treatment conditions of interest. A set of measurements in all the range of
selected conditions demanded up to twelve hours, as it was necessary to wait for the furnace conditions
to stabilise before collecting data. The whole range of conditions was divided into two parts, so that the
insulating paper could be changed and the ceramic spacers cleaned. Several curves were collected for
each condition, and the whole range of conditions was assessed up to three times.
subsequently connecting the capacitor to the probe with the active surface exposed to the plasma. The
electric charge stored in the capacitor was such that the probe was biased to a potential some tens of
volts more negative than its floating potential. This negative charge repelled electrons and, in some
Under this condition, all the electrons were repelled and only the positive ions were allowed to reach the
probe surface. Therefore, the drain curve of the capacitor provided information on the flow of positive
ions to the probe. Once the negative potential dropped to low enough values, the high energy electrons
managed to surmount the retarding potential and reached the surface of the probe. The arrival of both
positive and negative charges to the probe produced a change in the discharge rate of the capacitor.
1. Processing
AISI 316 samples were plasma treated in conventional DC furnaces, and these samples were adopted
as bench markers for the treatments carried out in the active screen industrial scale unit. Different
treatment conditions were explored in this AS furnace, in both nitriding and carburising atmospheres. As
a first approach, it must be said that AS processes had to be conducted under different conditions from
their DC counterparts (mainly, lower gas pressures), in order to obtain equivalent results. Furthermore,
the working conditions used for ASPC and ASPN also differed, particularly with regard to the bias level
In these graphs, the bias is expressed in power units (kVA) for simplicity although, as it will be shown
later, only by reporting both the electric potential and the current density can the conditions on the
worktable be unambiguously defined. For one given power set point, the gas mixture and gas pressure
determine, to a great extent, the combination of current density and electric potential on the worktable.
The following equations were found to fit the data points with reasonable accuracy:
p = C ⋅ P −k3 (4.1-1)
C = k 1e − k 2 i (4.1-2)
1 ⎛ p ⋅ P k3 ⎞
i=− ln⎜⎜ ⎟⎟ (4.1-3)
k2 ⎝ k1 ⎠
where i is the current in Amperes, p is the pressure in Pascals, P is the power applied on the work table
in kVA and k1, k2 and k3 are constants, for which values depend on the gas mixture (table 4.1-1).
increase in the bias set point will not alter the actual parameters of the electric discharge. Under these
conditions, the current density is self limited, and the electric potential is frequently very high (ca 900 V).
If the gas pressure is increased, larger current densities can be drawn, and these are associated with
lower electric potentials on the worktable. Furthermore, edge effects occur on samples treated above
some current density threshold, so that the applied bias should be kept below a certain limit in order to
AS treatments conducted within the range of effective conditions, shown in figures 4.1-1 and 4.1-2, were
less susceptible to process instabilities, and produced samples with homogeneous surface appearance.
morphology and properties. Figure 4.1-3 illustrates the arcing, hollow cathode and edge effects
frequently observed on a DC worktable. At the same time, the damage these effects produce on treated
components is apparent from figure 4.1-4, which also shows the comparatively improved surface quality
Further studies within the range of effective treatment conditions revealed differences between the
samples, particularly with regard to layer thickness. Figure 4.1-5 shows the layer thickness, measured
from GDOES nitrogen or carbon composition profiles, as the size of the bubble superimposed on the
map of treatment conditions presented before. It is evident from this figure that the optimum treatment
conditions lay in the low pressure and moderate bias level for ASPC, and even lower pressure and bias
level for ASPN. The layer thickness measured in each case is presented in tables 4.1-2 and 4.1-3, as a
characterisation, wear and corrosion properties was conducted. The ASPC condition for maximum layer
thickness was found at 125 Pa and 15% bias, while the parameters for minimum incidence of edge
effect were 30 Pa and 5% bias. Samples treated under these two conditions were studied in particular
However, it is timely to mention here that some differences in the mean residual stresses (RS) were
evident from the bending of the ASPC treated thin discs. Figure 4.1-6 shows the typical profile of the RS
discs, before and after treatment, collected with a stylus profilometer. The classic Stoney’s formula and
some modified formulas proposed by other authors were used to calculate the mean stress in the
treated layer. The results show that the compressive residual stresses increased with the bias level and
layer thickness (table 4.1-4). Although these results are by no means accurate, they are of use to reveal
differences between the treatment conditions and to understand the behaviour of the samples in other
tests.
AISI 316 samples plasma carburised in a DC furnace were adopted as bench markers for the
treatments carried out under active screen arrangements. All the plasma carburising processes were
conducted within the range of optimised conditions mentioned in the previous section. DC and AS
treated samples showed the typical features and properties reported for carbon expanded austenite, as
Observation under the optical microscope revealed some changes in the microstructure after plasma
carburising. The treated samples showed clear grain contrast due to slip bands, which appeared at
clearly seen in the crystal orientation maps (COMs) obtained with EBSD on the same area of the
sample before and after plasma carburising (figure 4.2-2). The EBSD pattern quality deteriorates after
the plasma diffusion treatment, and this seems to be related with the high density of defects introduced
into the crystal structure. The density of defects increases progressively as the layer grows with
The typical appearance of the treated layer can be seen from the cross section micrograph in figure 4.2-
4. It consists of a white featureless layer which, in contrast to the substrate, is not affected by the
etchant. This higher resistance to the etching agent can be considered as a first rough indication of
With regard to morphology, a significant difference between DCPC and ASPC could be observed under
SEM and AFM. Figure 4.2-5 shows the SEM high magnification micrographs, where some irregular
morphology is seen on the surface of a DCPC sample, compared to the uniform appearance of the
ASPC sample. This aspect was confirmed by AFM, as illustrated in figure 4.2-6, where the DCPC
The TEM studies conducted on the outmost layer of treated specimens revealed a slightly thicker oxide
layer on the ASPC sample compared to the DCPC one (figure 4.2-7). A layer of particles was found to
cover all the surface, and the typical morphology is shown in figure 4.2-8, together with the ring electron
diffraction pattern (EDP) obtained on that same area. The particle size was between 25 nm and 30 nm
for DCPC and between 30 nm and 40 nm for ASPC, whilst the indexing matched some iron and
diffraction pattern corresponding to austenite, which was obtained from the underlying material.
The XRD peaks shown in figure 4.2-10 were obtained from DCPC and ASPC samples and correlate
with the austenite stick pattern in both cases. Only one additional diffraction peak of low intensity was
found at 39.55º, which could be attributed to Hägg Carbide of composition Fe5C2. The peaks are sharp
with a slightly asymmetric pseudo Voigt profile, and they exhibit order dependent widening compared to
the untreated material. However, the reflections are shifted to lower angles, and the shifting varies with
the 2θ angle. The analysis of this shifting and the calculated lattice parameters indicate that DCPC and
ASPC produce the same degree of carbon supersaturation and lattice expansion (figure 4.2-11).
Figure 4.2-12 shows the typical GDOES composition-depth profiles for carbon, corresponding to DCPC
and ASPC treated samples. As it is evident from this graph, there is a high degree of correlation
between both curves in what concerns to the degree of carbon supersaturation, the composition profile,
and also the layer thickness. A drop in carbon content is seen within the first 1 μm of the ASPC sample,
which was found to match a small nitrogen peak. The active screen industrial scale unit was used for
both plasma nitriding and plasma carburising, and it was not possible to remove all traces of nitrogen,
With regard to hardness, figure 4.2-13 shows the Vickers microhardness values obtained on DCPC and
ASPC samples using different loads, which serve as an indication of the load bearing capacity (LBC) of
the treated layer. Once again there is a significant degree of correlation between both techniques. The
ASPC samples showed some higher values at low indentation loads (25 g) which can be explained by
hardness of ASPC at 100 g indentation load was marginally lower than their DCPC counterparts.
The hardness values obtained with higher loads were virtually the same. However, some cracks were
found inside and around the indentations conducted on DCPC specimens (figure 4.2-14). These cracks
appeared at indentation loads higher than 500 g and are not considered as a failure of the treated layer,
The instrumented hardness tests conducted on AISI 316 untreated and plasma carburised samples
confirmed a considerable increase in hardness and a small rise in the elastic modulus of the material
after the plasma diffusion treatment. The curves in figure 4.2-15 illustrate the applied load against the
displacement of the indentor, and the main results are summarised in table 4.2-1.
The set of experiments described in table 3.2-4 was conducted in order to understand the mechanisms
involved in ASPC and the differences between the ASPC and the ASPN treatments. For the purpose of
this assessment, ASPN was considered the yardstick for ASPC, as the former has been studied and
In first place, it must be said that the samples subjected to floating potential (i.e. arrangement C in figure
3.2.1-2) showed no hardening effect, and no signs of diffusion of carbon or nitrogen could be identified.
Even though this evidence is of significance for the understanding of the mechanisms involved in active
screen treatments, no further details will be given, since all the characterisation was identical to the
were dissimilar (arrangements A and B in figure 3.2.1-2). As the first approach, all DC treated samples,
i.e. DCPN, DCPC and DCPCN, showed the characteristic edge effect; the ASPN and ASPCN samples
had a very uniform surface appearance, whilst the ASPC sample exhibited a patchy morphology, which
resembled the shape of the mesh used as active screen in this experiment (figure 4.3-1).
The XRD peaks shown in figure 4.3-2, for AISI 316 samples treated under different conditions, indicate
that austenite is the only phase in all cases. However, the characteristic shifting and widening of the
diffraction peaks of expanded austenite vary for each case. The shifting and widening is clearly seen
after nitriding treatments (both DCPN and ASPN), whereas these effects are weaker on DCPC and
DCPCN specimens, and almost no shifting or widening can be seen on ASPC and ASPCN samples.
The GDOES composition-depth profiles, obtained from this same set of samples, indicate the typical
results after DC treatments: nitriding, carburising and N2 enhanced carburising (figure 4.3-3). On the
other hand, a significantly different response is once again evident after AS treatments. ASPN shows a
similar nitrogen profile to DCPN, although the penetration depth was shallower; a similar result is
observed for ASPCN, but the treated layer was significantly thinner in this case; finally the composition
profile obtained from ASPC samples evidences very shallow carbon diffusion.
The LBC, assessed by means of Vickers microhardness at different indentation loads, confirmed the
GDOES data (figure 4.3-4). The hardness values obtained for the nitrided test pieces were very high at
low loads, and decreased rapidly at higher loads, with the onset of cracks around the indents (figure 4.3-
5). On the other hand, the DCPC samples gave lower hardness values, with a gradual reduction, typical
of the carbon diffusion profile; whilst its ASPC counterpart indicated little hardening effect. A similar
DCPCN show a plateau region, which also fits the GDOES composition profile.
As part of this set of experiments, AISI 304 stripes were treated in AS arrangements using a mesh with
large holes, 25.4 mm in diameter. The hardness measurements conducted on these test pieces are
shown in figure 4.3-6, and indicate an apparent difference between the areas uncovered (holes) and the
areas covered by the cage in the case of nitriding. On the contrary, no differences are seen for
The roughness measurements performed on treated coupons showed some differences, as it can be
seen from table 4.3-1, with DC treated samples showing consistently higher roughness values than their
AS equals. A closer observation of AS treated specimens put into evidence the existence of a layer of
deposited material on them. This was obvious on the copper coupons, some of which exhibited a silver
SEM and AFM detailed studies showed a morphology analogous to the one described in the previous
section, with particles of different sizes depending on the treatment conditions. In general, the ASPN
samples appeared to be covered by a layer of fine particles, 50 nm to 100 nm in diameter, while the
ASPC and ASPCN treated specimens showed a much coarser and irregular structure (figure 4.3-8).
The use of glass slides and copper substrates produced enough contrast with the deposition layer as to
allow for some degree of identification of the present phases with XRD. Figure 4.3-9 shows the data
collected from copper coupons and, in all cases, they show a perfect match with the peaks of the Cu
standard. However, two small additional peaks at 41.69º and 44.89º can be seen after AS treatments,
not resist the cleaning procedure, whilst the deposition produced by ASPN was surprisingly coherent
and adherent (figure 4.3-10). The latter was analysed by means of XRD, and the diffraction peaks
matched several FexN compounds, and showed good agreement with the FeN pattern (figure 4.3-11).
This finding is of relevance for the understanding of the mechanisms of active screen plasma surface
Finally, the mesh used for the experiments on the mechanisms of active screen plasma surface
engineering was observed under the SEM and analysed by EDX to confirm its composition. The results
proved that the mesh used for these experiments was AISI 304, and the typical micrograph and EDX
spectrum are shown in figure 4.3-12. A similar analysis was conducted on the mesh of the AS industrial
unit, and this was found to be made of low carbon steel (figure 4.3-13). In both cases, the deposited
material found on the samples exhibits some interesting correlation with the material of the mesh.
4. Tribological properties
The wear resistance of AISI 316 untreated (UT), DC plasma carburised (DCPC) and active screen
plasma carburised under two conditions (ASPC 125 Pa – 15% bias and ASPC 30 Pa – 5% bias) were
compared. The treatment conditions selected for this assessment followed the same rationale
mentioned in the section on processing, i.e. conditions of maximum layer thickness and minimum
incidence of edge effect, respectively. Table 4.4-1 summarises the obtained results.
The appearance of the coupon samples after the wear tests evidences the difference in the wear
response of the two materials (figure 4.4-1). The same is observed in the pictures of the counterpart
WC-Co balls, in figure 4.4-2. It is also worth mentioning that a significant amount of wear debris was
after the test in the case of all the treated samples. These three aspects of the wear tests will be
The measured wear volume indicated a marked improvement of all treated samples over the untreated
AISI 316 substrate (figure 4.4-3). The worn volume of the expanded austenite was more than two orders
of magnitude lower than that of the untreated AISI 316. On the other hand, no significant difference
could be appreciated between the DCPC and the ASPC treated samples. All the treated samples
A closer observation of the wear tracks under SEM revealed differences in morphology. The untreated
AISI 316 suffered from abrasive and adhesive wear, while the treated samples developed a protective
oxide layer. The difference in morphology can be appreciated in figure 4.4-4, and the composition of the
oxide layer is evident from the EDX maps in figures 4.4-5 and 4.4-6. The layer is mainly formed by iron
and chromium oxide, but some tungsten is also present (table 4.4-2). The latter evidences the transfer
of material from the counterpart ball, and the formation of the so called third body during sliding. It is
also interesting to note that nickel does not seem to follow the same pattern as chromium and iron.
Regarding the counterpart material, the size and shape of the wear scars on the WC-Co balls matched
the width of the wear tracks left on coupon samples (figures 4.4-2 and 4.4-7). Detailed studies of the
morphology and composition, under the SEM, revealed accumulation of material at the trailing edge
(figure 4.4-8). This is more evident in the ball worn against the plasma carburised material. However,
the SEM pictures and the EDX composition maps (figures 4.4-9 and 4.4-10) indicate that the amount of
untreated AISI 316. This is a clear evidence of severe adhesive wear of the untreated material.
The WC-Co balls worn against untreated AISI 316 showed a significant amount of iron and chromium
on the wear scar. The oxygen signal was also strong, as the tests were conducted in laboratory air, and
the high pressures and temperatures developed in the contact spots favoured oxidation (figure 4.4-9).
On the other hand, the WC-Co balls worn against expanded austenite only displayed minor traces of
iron oxide on the wear scar, and a small amount of material accumulated at the trailing edge (figure 4.4-
10). It is evident that less material is transferred from the coupon sample onto the counterpart ball in this
case.
With regard to the wear debris, obvious differences were found once again between the treated and the
untreated specimens. The latter produced a significant amount of black debris during the sliding wear,
and a simple test with a magnet revealed it responded to the magnetic field. Under the OM, the particles
were 20 μm to 30 μm in size and metallic shiny (figure 4.4-11). The wear tests conducted on plasma
carburised material produced small amounts of reddish-brown debris. This time, the particles were
mostly non-magnetic, and only a few responded weakly to the magnetic field. OM observations showed
the particles had a mean size of 10 μm with a plate-liked shape (figure 4.4-11).
The wear debris was observed under SEM and TEM in order to study their morphology and identify their
nature. In both cases, the debris was found to be a conglomerate of smaller particles, but the particles
were finer in the case of the plasma carburised material than in the untreated AISI 316 (figure 4.4-12).
The ring electron diffraction patterns indicated possible martensite and magnetite matches for the
untreated material (figure 4.4-13). This is in agreement with the appearance and magnetic properties
once again correlates with the previous observations. However, an oxide of mixed nature must be
present in this case, as the EDX spectrum exhibits the unequivocal presence of other alloying elements
XRD was used to identify the phases present in the wear track of AISI 316 specimens. The intensity of
the reflection shown in figure 4.4-15 is not considerable, but it fits the martensite pattern. In addition to
this, microhardness measurements conducted on the wear track gave an average hardness of 560
HMV0.01, which is significantly harder than the austenitic substrate. The hardness profile conducted on
the cross section of the wear track clearly shows the hardening effect and, possibly, martensitic phase
transformation (figure 4.4-16). Strain induced phase transformation is likely taking into account the
severe deformation observed in the subsurface of the untreated material (figure 4.4-17).
EBSD phase and crystal orientation mapping gave some further evidence on the extent of the
deformation and martensitic transformation. Figure 4.4-18 shows a similar SEM cross section image
(without etching), and the EBSD map for austenite and martensite. These images show a 15 μm thick
layer in which strain induced martensitic phase transformation took place, to some extent. Plastic
deformation is evident up to a depth of 60 μm, and some degree of work hardening is also assumed to
The study here presented indicates that work hardening took place during the wear tests of untreated
AISI 316 against WC-Co. Moreover, evidence of strain induced martensitic transformation was found by
results. The electron diffraction pattern obtained with TEM did not match the martensite pattern; the
XRD studies only showed reflections corresponding to austenite; and the microhardness profile
conducted on the cross section was identical to a normal plasma carburised specimen. Only the EBSD
orientation maps showed some evidence of deformation close to the surface (figure 4.4-19). However,
this deformation was attributed not only to the shear stress of the wear test, but also to the distortion
The graphs in figure 4.4-20 show the friction signals obtained during sliding tests under similar
conditions to the wear tests reported before. The results are summarised in table 4.4-4.
The AISI 316 UT couple (UT vs UT) produced an average coefficient of friction of 0.6 with fluctuations,
in agreement with the literature. A large amount of wear debris was released during the tests, which was
metallic in appearance, and magnetic. The WC-Co ball (UT vs WC) yielded lower friction coefficients,
and left some stick-slip markings on the wear track and no significant debris; whilst the PC ball sliding
on AISI 316 (UT vs PC) showed an intermediate behaviour in all these aspects. In general, the
coefficient of friction clearly peaked when sliding started (μstatic), and subsequently decreased
The friction signal for the AISI 316 untreated ball sliding against a PC disc (PC vs UT) exhibits
significant instabilities with very high friction peaks. The peaks in the friction force correlated with
intense noise during the test, and a considerable amount of black magnetic debris was produced. Both
the WC and the PC balls sliding on PC discs (WC vs PC and PC vs PC) yielded lower coefficients of
friction and produced small amounts of reddish brown wear debris, mainly non-magnetic (only some
these tests, followed by a period of low friction and a final increase up to the final steady friction
coefficient, which was higher than the initial relative maximum (μstatic < μdynamic).
5. Corrosion behaviour
The initial immersion corrosion tests conducted in a sulphuric acid solution (16% mass fraction) at room
temperature produced very little corrosion in all the test pieces. The AISI 316 samples were found to
change the appearance to a tarnished grey colour after only 10 minutes of exposure to the acid, while
the PC samples retained their metallic shine throughout the experiment. OM and SEM observations of
the corroded specimens revealed intense corrosion on MnS inclusions in the AISI 316 coupons, and
failure of the outmost deposited layer of carbides in their ASPC equals (figure 4.5-1). The test was
repeated with PC samples polished after treatment, on which no deposition layer remained, and no
The tests conducted at boiling temperature in the same sulphuric acid solution (16% mass fraction)
revealed clear differences between the specimens, as can be seen from figure 4.5-2. Corrosion was
catastrophic on AISI 316, while the layer of EA was left virtually intact. The edges of DCPC 400 and
ASPC 125 test pieces showed some contrast with the rest of the surface area, associated with some
degree of edge effect. On the other hand, the ASPC 030 sample showed a very uniform surface
appearance. The cross sections in figure 4.5-3 illustrate the different behaviour of the AISI 316 UT and
Detailed observation under SEM revealed some localised corrosion on the EA layer. The attack was
particularly intense on MnS inclusions and, once the substrate was exposed, cavities of considerable
bands, even in the early stages of the tests, although corrosion progressed slowly and the layer was still
unbreached after 20 hours of exposure to this corrosive solution (figure 4.5-5). Finally, the grain
boundaries were found to be vulnerable to localised corrosion, and this type of attack was clearly
identified on samples exposed for 15 hours or longer periods of time (figure 4.5-6).
With regard to the preferential corrosion on slip bands and grain boundaries, it must be mentioned that
these were observed more frequently on DCPC 400 and ASPC 125 samples than in the ASPC 030
ones, which were actually free from intergranular corrosion. In all cases, corrosion was more intense
close to the edges of the samples, where the density of slip bands was higher and the grain boundaries
appeared to be weaker. Once the layer of EA was breached, preferential dissolution of the substrate
took place rapidly creating large subsurface pits, with some grain boundaries failing first (figure 4.5-7).
The tests performed on cylindrical specimens yielded similar results, with catastrophic corrosion of the
AISI 316 substrate and mild corrosion of the treated samples, which took place, mainly, through defects
in the EA layer (figure 4.5-8). The AISI 316 samples turned black after only a few minutes immersed in
the corrosive solution, and a considerable amount of gas bubbles formed on their surface, which was
attributed to the cathodic evolution of hydrogen. On the other hand, the PC samples retained their
metallic appearance, although they became dull after the test. The evolution of hydrogen was not so
The localised attack at the centre of the rods is noteworthy, and was attributed to the presence of MnS
stringers; while the extended damage towards the edges of specimen DCPC 400 was believed to be a
consequence of stronger edge effects produced by these treatment conditions. Finally, this tests also
shown in figure 4.5-9. The mass losses of all PC specimens were between 4 and 6 times lower than
Corrosion tests conducted in H2SO4 16% solution with addition of CuSO4 while in contact with metallic
copper, produced very little corrosion of the AISI 316 substrate, though some degree of attack was
found on defects of the EA layer. In this solution, the evolution of hydrogen was replaced by the
cathodic reduction of copper ions, and no gas bubbles were seen to form on the surface of the test
pieces. As it can be seen in figure 4.5-10, the edges of the specimens were more susceptible to
localised corrosion, especially in the case of specimens DCPC 400 and ASPC 125, whereas their ASPC
Closer observation with OM revealed preferential attack on slip bands in all the treated samples (figure
4.5-11). However, grains with higher density of slip bands were more severely attacked, so that
corrosion was more extensive on DCPC 400 specimens, followed by ASPC 125 and ASPC 030. With
regard to grain boundaries, even though no grains were dislodged and no severe intergranular corrosion
was observed, some ditch type grain boundaries were seen on coupons DCPC 400 and ASPC 125; the
SEM and AFM studies of the corroded grain boundaries revealed they were the result of the
coalescence of pits at the intersection of the slip bands with those grain boundaries, to form a
continuous trench (figure 4.5-12). Moreover, EDX composition line scans provided some evidence of
carbon enrichment in the area surrounding the trench, although this is not conclusive as no correlation
could be found with carbide forming elements, particularly chromium (figure 4.5-13).
very low mass losses, two orders of magnitude lower than the values reported for H2SO4 16% solution.
Even though the figures are not conclusive, the corrosion factors in table 4.5-1 show a trend that is
helpful to understand the results and rank the treatments according to their corrosion resistance.
Two other corrosive solutions were adopted in order to assess the susceptibility of the grain boundaries
of the PC samples to localised attack: sulphuric acid 40% solution with iron sulphate, and nitric acid
20% solution. The former test was conducted for 24 hours and produced a very similar attack to the one
observed with copper sulphate. Corrosion was more intense on defects in the EA layer, mainly slip
bands, and some pits coalesced to produce more severe localised attack and form some ditch grain
boundaries. The edge effect was revealed in the case of DCPC 400 and ASPC 125, while the ASPC
In the case of nitric acid, the test was conducted for a total of 240 hours, in 5 periods of 48 hours in
which the solution was renewed. The appearance of these samples differed slightly from previous tests,
principally with regard to the attack on grain boundaries. As it can be seen from figure 4.5-14, the
untreated AISI 316 suffered some mild attack and only the grain boundaries were developed after the
test. On the other hand, all PC samples exhibited localised corrosion on slip bands and pits, which were
larger and more numerous on the DCPC 400 specimens, followed by the ASPC 125 and the ASPC 030.
No grains were dislodged during the test, but ditch type grain boundaries were clearly observed under
AFM, having a depth of, at least, 0.8 μm (figure 4.5-15). In these cases, the grain boundaries had the
appearance of a continuos trench surrounding full grains, and no traces of primitive pits could be
identified. Some severe localised corrosion was found in PC samples, which was attributed to the
it was not generalised, the depth of the pits was enough to be taken into consideration as a failure mode
of the EA layer. In contrast, the corrosion rates calculated from gravimetric measurements indicate that
the corrosion rate of the PC samples becomes comparable to the one of AISI 316 untreated substrate,
The OES spectra collected with the telescope through the sight port of the furnaces were useful to
identify the most common active species involved in electric discharges through different gas mixtures,
i.e. CH4–H2, CH4–N2–H2 and N2–H2. The spectra in figures 4.6-1 to 4.6-3 show some of the differences.
The signals collected through the sight port for DC and AS experimental arrangements were virtually
identical, although the emission intensity was slightly higher in the AS setting, due to the hollow
In the case of plasma carburising (CH4–H2), the spectrum was very scarce in peaks, and only the
atomic hydrogen Balmer series could be seen. The Hα at 656.125 nm dominated this wavelength range,
and only some weak H2 and CH+ peaks completed the spectrum. For the nitrogen enhanced plasma
carburising (CH4–N2–H2), the addition of only 1% N2 changed the shape of the spectra in a considerable
way, and the CN (violet system) and N2+ (first negative system) bands were clearly present, although
the Hα line remained to be the most intense. Finally, the spectra for plasma nitriding were dominated by
the N2+ (first negative system) bands which, under some conditions, were even more intense than the
Hα emission line. Some NH, hydrogen and nitrogen peaks completed the spectra.
conduct with this experimental arrangement, i.e. from outside the furnace. When measurements were
done through the sight port, the telescope was focused on the active screen itself (the mesh), which is
designed to form hollow cathodes. These hollow cathodes changed their position, quantity and intensity
depending on the treatment conditions and, as a result, they also made the collected OES signal erratic.
In conclusion, no clear correlation could be observed between these emission spectra and the AS
treatment conditions.
The optical fibres used to probe the AS industrial unit yielded optical signal directly from the worktable,
which was found to vary consistently with the treatment conditions, although this was achieved at
expense of signal intensity and signal-to-noise ratio. Therefore, only the strongest emission lines and
bands could be used for this purpose, and the Hα and Hβ peaks of the Balmer series at 656.125 nm and
486.003 nm respectively, were selected in the case of plasma carburising‡, whilst the N2+ emission band
at 391.44 nm was chosen for the assessment of plasma nitriding. The emission intensity was recorded,
for each gas mixture, at different working pressures and bias levels, and the information was
The OES maps thus obtained exhibit increasing intensity with increasing level of bias, and this intensity
is higher in the low pressure region. As it was mentioned in a previous section, low working pressures
imposed a limitation on the level of bias and, consequently, on the emission intensity. Increasing the gas
pressure allowed a further increase in the applied bias, and the emission intensity responded
accordingly. However, this was true only until some pressure threshold was reached, and the relative
maximum in emission intensity was found to agree with the conditions identified for maximum layer
‡ The use of the H emission lines was justified because atomic hydrogen is a by-product of the decomposition of the
methane molecule and trials showed some sensitivity of peaks in the Balmer series to the amount of CH4 in the gas mixture.
Nevertheless, the identification of clear and strong carbon containing species would of course be of interest.
levels of bias; this point was found at higher pressures for ASPC than for ASPN (figure 4.6-5).
The optical signal was plotted against different electric parameters and some trends were revealed,
depending on the gas pressure and the gas mixture (figure 4.6-6). The emission intensity plotted as a
function of the bias voltage showed some quadratic trend at high pressure, while the data points
obtained at low pressure deviated from this line. Both ASPC and ASPN exhibited very similar responses
to the electric potential of the worktable. The emission intensity was found to vary linearly with the
electric current, although the slope was affected by the gas pressure and, furthermore, a clear breaking
point was found in ASPN, whereas this was not visible for ASPC. Finally, the emission intensity showed
considerable correlation with the bias power applied on the work table, with slight changes in slope for
The use of the IFP under AS arrangements inside a DC furnace indicated a strong electrostatic coupling
with the active screen electric input. This was concluded from the oscillations in the floating potential
(FP) of the electrostatic probe, which perfectly matched the electric input characteristics of each furnace
(figure 4.7-1); while the same experiments conducted in conventional DC arrangements showed only
background noise on the FP electric signal. Figure 4.7-2 shows the differences in FP between the DC
and AS arrangements, which was seen to depend on the gas mixture and gas pressure used for the
experiment. In addition, the flux of ions was measured and the obtained results indicate appreciably
higher values in the case of the AS setting, although it must be said that these figures bare considerable
error.
electrostatic probe was once again in agreement with the power input of the furnace (figure 4.7-3). The
average value of the FP was also found to vary with the treatment conditions, shifting to more negative
values with increasing bias and pressure (figure 4.7-4). However, a limitation imposed on the applied
bias was once again found at low gas pressure, and the FP values were observed to remain constant or
even decrease slightly once this bias saturation level was reached. Unfortunately, only a reduced range
of treatment conditions could be assessed because of limitations in the probe and the measuring circuit.
An attempt was made to calculate the ion flux from the drain curve of the capacitor of the IFP measuring
circuit. Figure 4.7-5 shows the drain curves obtained with the worktable under floating potential for
ASPC and ASPN at different gas pressures. The different responses were clear, and the capacitor was
drained several times faster in the low pressure range, indicating a higher rate of arrival of positive
charges onto the effective surface area of the probe. Some measurements were also conducted under
normal treatment conditions, although the range was seriously restricted by limitations of the probe and
the measuring circuit, and the errors make these figures unreliable. In general, it can be said that the
calculated ion fluxes increased with bias and decreased with gas pressure (figure 4.7-6).
The current-voltage characteristic curves obtained in the AS furnace indicate a non-linear relationship
typical of abnormal glow discharges, with the current density increasing more rapidly at higher
pressures (figure 5.1-1) [190, 192]. The two constraints found in the operating conditions, namely the
onset of edge effects and the current saturation, are relatively well known phenomena.
The edge effect is a result of the local increase in the electric field, and the consequent higher ionisation
rate and higher current density [218]. This obviously declines when lower electric potentials are applied,
although it has also been reported that low pressure discharges are less sensitive to concentration of
The current saturation can be attributed to a low density of charge carriers and to technical limitations of
the power supply. In the plasma, the density of electrons is approximately the same as the density of
ions (ne ≈ n+), and the current density can be expressed as [220]:
j = en e μ e E = σE (5.1-1)
where j is the current density, e is the elementary charge, ne is the electron density, μe is the electron
mobility, σ is the conductivity of the weakly ionised gas, and E is the electric field. The value of E is
defined by the potential delivered by the power supply (UMAX = 926 V for the AS furnace) and the
distance between the electrodes in the furnace, which impose the first limitation.
consequently, of the frequency of ionising collisions. A coefficient α is widely used to express the
number of ionising collisions undergone by one electron in a 1 cm trajectory along the electric field E.
⎛ B ⎞
α ⎜⎜ − ⎟⎟
= Ae ⎝ E/p⎠ (5.1-2)
p
where α is Townsend first ionisation coefficient, p is the gas pressure, A and B are gas constants and E
is the electric field between the electrodes. A reduction in the pressure increases the E / p ratio and the
mean energy acquired by electrons between collisions, but it also reduces the probability of collision
with gas molecules, which defines the second constraint for equation 5.1-1.
Equation 5.1-2 resembles the experimental relation 4.1-2. Even though no clear physical meaning could
be attributed to each one of the constants, term C seems to account for the avalanche effect, while the
With regard to the bias, it is interesting to note that the real applied power differed from the bias set
point (figure 5.1-2). The deviations were considerable and erratic, although it can be said that at low
pressures there was a trend towards higher power levels than the set point, while at high pressures the
tendency was inverted. The data points for bias current, plotted against bias set point, lay in a straight
line, so that the control system is believed to calculate the bias from the current measured on the work
table (figure 5.1-3). This deviation is not considered to be of fundamental importance as long as it is
optimum treatment conditions at 125 Pa and 15% bias for ASPC, and 75 Pa and 10% bias for ASPN.
However, the applied power exhibits a very similar trend and the optimised treatment conditions
correspond to maxima in the applied power (figure 5.1-4). This is not surprising taking into account that
the electron density increases with the applied power [221]. At low pressure, the bias current saturates
and the power is thus limited, but higher power can be applied at moderate pressures. For current
densities low enough to avoid edge effects, the maximum applied power corresponds to the optimum
condition.
The optimised conditions for ASPC and ASPN also show some qualitative agreement with the Paschen
curves reported for pure H2§ and 25% N2 – 75% H2 [220, 222]. Pure H2 breaks down at higher
potentials and values of p·L (pressure x distance between electrodes) compared to the 25% N2 – 75%
H2 mixture. Furthermore, the ratio pLH2 / pLN2-H2 from the cited references gives an average value of
1.46, while the ratio of optimum pressures for this project pASPC / pASPN yields 1.66, which is in
reasonable agreement. This is a rather speculative interpretation, although it seems sensible to think of
the optimum working conditions found in this project as those closest to the point of highest ionisation
Unfortunately, the full Paschen curves could not be obtained in the present project, thus only very
general conclusions can be drawn. In view of the experimental evidence, the curves are expected to
show a minimum breakdown potential in the region of the optimum treatment conditions reported here,
while they would be shifted to lower potentials and lower p·L values compared to the conventional DC
discharges, due to the additional influx of electrons from the metallic mesh of the AS furnace [223, 224].
§ The 3% CH4 content of the gas mixture used for ASPC does not significantly affect the plasma physicochemistry, so that it
similarity relations [157], indicated the presence of additional charge carriers [190], i.e. the AS furnace
draws larger current and power than the DC furnace for similar discharge conditions (table 5.1-1).
Finally, figure 5.1-5 shows the same I-V characteristic curves of figure 5.1-1, but this time it includes the
value of E / p, represented in the size of the bubble. It is clear from this graph that electrons acquire
significantly larger energies in the low pressure side of the range of assessed conditions, which
accounts for the increased efficiency of the plasma treatment observed in this same pressure region.
The results obtained by means of DCPC and ASPC showed substantial agreement in terms of layer
thickness, although differences were found depending on the ASPC treatment conditions (table 4.1-2).
An estimation of the carbon diffusivity was obtained making use of the following equation:
x = a(Dt )
1/ 2
= Kt 1 / 2 (5.2-1)
where x is the layer thickness, a is a constant, D is the carbon diffusivity, t is the treatment time, and K is
the product of a and D. The obtained values are listed in table 5.2-1 and they are in substantial
agreement with the values reported in the literature [118, 119, 225].
The measurements of layer thickness conducted after different treatment times gave some information
regarding the growth kinetics, as it is illustrated in figure 5.2-1. The data points obtained in this project
are plotted together with those reported by Y. Sun [118] and M. Tsujikawa et al. [119, 225]. The former
On the other hand, the data reported by Tsujikawa et al. [119, 225] show a quadratic dependence of
layer thickness with the square root of time, which agrees with the data points obtained in this project.
This trend has also been identified by other authors [226, 227] who attributed the deviation to the
One additional outcome of these assumptions is the existence of some incubation time for the layer
growth. The linear dependence reported by Sun has a non-zero ordinate at origin, which was
considered to be the consequence of adding methane to the gas mixture during the heating stage. In
contrast, the quadratic dependence found in this project, and also reported by Tsujikawa, indicates
some time will elapse before the layer starts growing, and this feature is of significance in order to
To this end, it is generally accepted that expanded austenite forms under a metastable equilibrium of
carbon in austenite, sometimes referred to as paraequilibrium (PE) [110, 122, 147]. Leaving aside some
equilibrium, in which the substitutional atoms are virtually immobile in front of the interstitial atoms [231,
232]. Other authors have interpreted PE as a complete trapping of interstitials in the substitutional lattice
[233] and, in the particular case of expanded austenite, the chromium content of the alloy seems to play
crystal defects [234]. Maalekian et al. [235] have recently studied the influence of a heavily dislocated
crystal structure on the distribution of carbon in martensitic steel, and Heure et al. [236] conducted
internal friction studies on ASS gas carburised at low temperature and concluded that the obtained
results were consistent with multicarbon interstitial-vacancy complexes. In the present study, COMs
were obtained on the same areas of a sample after different treatment times. Figure 5.2-2 shows the
average EBSD pattern quality and the carbon content from GDOES profiles, obtained after each
treatment. The pattern quality is sensitive to defects in the crystal structure [237] and a qualitative
correlation can be seen, although this evidence is not solid enough to draw conclusions.
With regard to the microstructure, comparative studies conducted on AISI 316 after DCPC and ASPC
showed equivalent results. Figure 4.2-1 illustrates the change observed on the microstructure of AISI
316 after PC. The presence of slip bands has been previously reported [135, 238-240] and it is a clear
indication of plastic deformation of the surface layer under the high compressive stresses developed
upon the expansion of the austenite lattice. Contrary to what Xu et al. have reported [241], very short
incubation times were necessary in the present case, and slip lines were found in AISI 316 samples
after only 1 hour of plasma carburising treatment at 450ºC. The density of slip bands and the height of
The strain is obviously unevenly distributed, and depends on the crystallographic orientation of the
grains, relative to the stress plane [242]. In this way, some grains show no signs of deformation, while
others show one or two sets of parallel slip lines, i.e. one or two slip systems have been activated.
Moreover, a higher density of slip bands is seen close to the grain boundaries, whereas the central
The rotation of the crystallographic structure upon plastic deformation is known to affect the texture of
ASS [246] and the anisotropic strain produced during plasma or ion beam surface treatments has been
found to produce rotation of the crystallographic planes of single crystal and polycrystalline samples
[247-249]. Several authors conducted XRD texture studies on plasma nitrided samples, and reported
differences in the pole figures obtained before and after the surface treatment [113, 250, 251]. Figure
5.2-4 shows the results obtained on PC samples in this project, which show good agreement with the
The crystallographic structure of carbon and nitrogen EA has been extensively studied by means of
TEM and XRD, and there is a significant body of literature and discussion in this regard [106, 111, 114-
116, 252]. TEM studies provide extremely useful but also very localised information, and in the case of
XRD it is difficult to establish the individual contribution of different effects (residual stresses,
dislocations or stacking faults) [112, 114]. Therefore, an attempt was made to use EBSD and AFM to
The slip bands in figure 5.2-5 exhibit some 40 nm steps, which are considerably larger than the 10 nm
limit reported for approximately 45% deformation in a similar material under tension [253]. The number
of emerged crystallographic planes can be estimated taking into account the relationship:
r r
h = nz ⋅ b (5.2-3)
the Burgers vector for the dislocations. Computations for the most common steps, about 30 nm in
height, yielded some 100 planes. Slipping was assumed to take place mainly along the {111} <110>
path, as this is known to be the main slip system in AISI 316 [242, 245, 246, 253].
The deformation mechanism proposed in [253] implies that the balance between the internal friction
necessary to move dislocations and the external applied stress (or the local stress field at a given stress
dislocations. Consequently, if a direct comparison is considered valid, larger stress levels must have
been reached in the present project or the additional amount of dislocations generated by the plasma
The slip system activated in each grain, by the compressive plane stress developed during PC, depends
on the crystallographic orientation, as it can be seen in figure 5.2-6. Grains with one of the {111} planes
parallel to the surface are more prone to cross slip, although the angle of cross slip in some grains
indicate that sliding could have taken place on {100} or {110} planes, when their orientation was
suitable. The domains formed on cross slip, i.e. the chessboard like sub-grain arrangement [245], would
play a role of significance in the distortion and broadening of the XRD peaks, and could eventually
develop a deformation texture [246]. Although the inverse pole figures obtained with EBSD showed
The sub-grain structures are clearly visible after plasma carburising, and some sub-grain boundaries
can be seen from the line scans shown in figure 5.2-7. Moreover, the misorientation map also exhibits
signs of distortion (figure 5.2-8), and the grains with high density of slip bands also show large levels of
upon deformation and they are thus responsible for the misorientation observed within the grains [246].
The AFM scans also indicate that these domains do not remain plane, but they are tilted during sliding
(figure 5.2-5). The changes in the colour orientation code of the EBSD COMs (sample normal direction)
indicate rotation of the crystallographic planes. Although attempts were made to quantify this rotation,
The orientation relationship of grain boundaries was also affected by the plastic deformation, and the
change in Σ3 CSL grain boundaries is illustrated in figure 5.2-9. The degradation of CSL grain
boundaries with increasing strain has also been reported in [245], and it was found to be a consequence
of the rotation of the crystallographic planes away from the original CSL orientation relationship. Some
CSL were seen to degrade faster, presumably because those grains were suitably oriented for
deformation and rotation. However, this trend extended rapidly towards all the microstructure and grain
The assessment of residual stresses developed on AISI 316 PC samples has long been attempted, as it
has a considerable influence on the performance of this material. Several methods have been used to
measure residual stresses, and XRD is among the most accepted ones. Unfortunately, it was not
possible to use the XRD d-sin2ψ technique in this project, so alternative methods were used. However,
some comments will be made with regard to the XRD results obtained in the θ-2θ geometry.
After plasma carburising, the XRD austenite peaks were found to be consistently shifted to lower 2θ
angles and broadened, the latter effect increased with the reflection order. There exist a number of
contributions usually involves some assumptions [249, 254]. Stacking faults and microstrains are
thought to be the main reasons for the peak broadening [107, 112], together with a reduction in the size
of the diffracting domains as a result of the considerable plastic deformation undergone during PC [255,
256]. The peaks are broadened very nearly symmetrically, so that twin faults seem to play only a minor
role [257-259].
The peak shifting can be associated with the high density of dislocations, twins and stacking faults, and
also with the level of residual elastic strains, i.e. residual stresses [107, 114, 257-259]. It is once again
difficult to separate the contribution made by each of these effects, but the XRD diffraction peaks (figure
4.2-10) and the plot of ahkl vs cosθ·cotθ (figure 4.2-11) are consistent with an uneven lattice expansion
(table 5.2-2) and the presence of high compressive residual stresses [107, 112, 114]. An attempt was
made to estimate the residual stress by using the method proposed by Wagner et al. [259]. This method
uses second order reflections to separate the contribution made by stacking faults and residual stresses
to the peak shifting. Thereby, the maximum residual stress estimated in this way was -1.98 GPa (refer
The XRD measurements of residual stresses conducted by Sun et al. [254] on carbon EA yielded
average values in the order of -0.3 GPa to -1.5 GPa within the layer, with maximum compressive
residual stresses at the surface between -1.5 GPa and -3.5 GPa. These values were compared with the
maximum compressive residual stress from the expansion produced by the carbon content of EA,
calculated to be as high as -10 GPa, and plastic deformation was proposed to account for the partial
during PC, from the curvature of thin discs. Calculations yielded values of compressive residual stresses
in the order -0.4 GPa to -0.8 GPa, which are in considerable agreement with the average residual
stresses reported in [254]. The theory of thin plates and shells [260-262] was used to analyse the
bending discs further (refer to appendix C for details on these calculations). The average residual stress
obtained was -0.770 GPa, which is very close to the value obtained with the Stoney’s formula.
However, this calculation assumes a constant residual stress profile within the EA layer, which is far
from the real case. If the stress distribution is approximated by a triangular distribution (figure 5.2-10),
and reducing the moments to the layer-core interface, it yields a value of residual stresses at the surface
of -1.155 GPa. Furthermore, this triangular distribution of residual stresses is still fictitious. If the
hardening effect is considered to be a function of the carbon content of the EA, it is thus sensible to
assume that the residual stresses profile will be similar to the carbon-depth profile obtained from
GDOES. Figure 5.2-10 shows the distribution which, for the same input data used previously, gives a
These levels of stress are obviously higher than the yield stress of AISI 316 (240 MPa) so that a
significant degree of strain hardening must have taken place. It would be thus reasonable to presume
that the residual stresses are distributed according to the strain hardened profile [263]. Microhardness
tests are sensitive to residual strain [264, 265] and indentations conducted at different loads provide
information from different layers of material, depending on the penetration depth. The ratio of
indentation diagonal (D) to penetration depth (h) can be estimated to be approximately 7:1 [266]; and
empirical evaluations have shown that the contribution made by the first layers of underlying material up
to a depth equal to 3·h accounts for the 44% of the measured hardness [267].
considering the hardness values represent a layer of material equal to 1, 2 and 3 times the indentation
depth h. The hardness profile vs 3h is in reasonable agreement with the Knoop hardness profiles,
except for the outmost layer, where errors are known to increase for hardness measurements
conducted on cross sections. Considering a distribution of residual stresses equivalent to the curve
hardness vs 3h, the maximum compressive stress at the surface is found to be -1.61 GPa, in order to
produce the same bending moment as in the previous cases (figure 5.2-10).
The values of residual stresses obtained with this method are in surprisingly good agreement with the
ones measured by XRD. However, it must be said that the former method is expected to yield lower
values than the latter. The reason for this is that the bending discs give information about the residual
stresses of the first kind or macrostresses, whereas XRD measures microstresses [268].
Finally, the curves obtained by instrumented indentations will be briefly discussed from the point of view
of the residual stresses present on the samples. The indentations are affected by the presence of
residual elastic stresses and residual plastic strains, especially in the case of a material that work
hardens [269-271], and the shifting of the curves obtained from EA are consistent with the presence of
Unfortunately, the methodology used to estimate the residual stresses usually requires imaging the
indentation marks, and also the determination of some experimental coefficients, which could not be
done during this project. However, an attempt was made to calculate the residual stresses following the
methodology proposed in [272]. This method assumes the elastic response of the material is not
affected by the residual stresses, and so the shadowed area in figure 5.2-12 would represent the elastic
P1 − P2
σ residual = (5.2-4)
2π tan 2 αh 2
where P1, P2 and h are obtained from figure 5.2-12 and α is an experimental coefficient. Substituting the
values in table 5.2-3, one obtains a residual stress of -9.57 GPa. This compares favourably with the
maximum theoretical stress of -10 GPa calculated by Sun et al. from the carbon content of EA, although
such a high level of stress is highly unlikely. The coefficient α could be a source of error, as this is
experimentally determined and the same value of α as in [272] was used in this calculation, without
further consideration. The method could thus yield some interesting results, although further
3. Tribological properties
The evidence obtained on AISI 316 samples agrees with the results reported in the literature [52, 56, 57,
60, 61]. The rough appearance of the wear track (figure 4.4-4), the large size, shape and metallic nature
of the debris (figure 4.4-11) [273], the occurrence of strain induced martensitic transformation (figure
4.4-15) and the severe plastic deformation of the subsurface (figure 4.4-17) are all signs of severe wear
[274]. The plastic deformation of the subsurface and the formation of a fine microstructure in the
severely distorted layer of material are characteristic features of this wear regime. Moreover, the EBSD
COM revealed that this layer of material develops a strong deformation texture showing {111} or {100}
between asperities at the sliding surfaces [62]. At the same time, some loose particles may become
embedded in one of the surfaces, and produce deformation and abrasion on the opposite surface [277].
The theory of adhesive wear is useful to explain such phenomena as subsurface deformation, transfer
of material to the counterpart slider and friction. However, this theory has been extensively challenged
The theory of delamination has been developed to describe wear under conditions of subsurface plastic
flow [65, 278], and accounts for the formation of plate like wear debris particles, similar to the ones
found in this project [75]. The theory predicts the formation of subsurface voids, which would coalesce to
form cracks and, ultimately, to detach metallic debris particles. Although this theory has been
extensively used to analyse wear of metals [75, 274, 275, 279], other authors have questioned its
fundamentals, i.e. the nucleation of cracks in a region of the substrate which is essentially subjected to
high compressive stresses [280, 281]. These same authors proposed an alternative mechanism of wear
known as ratchetting.
According to the ratchetting wear mechanism, the fractures nucleated in the substrate are of ductile
nature, as a result of the severe plastic strain imposed by sliding. If the loads are low and the yield point
of the material is not exceeded, i.e. within the elastic range, the failure mechanism is most probably one
of high cycle fatigue [281]. On the contrary, if the yield threshold is exceeded, plastic deformation takes
place and accumulates in the subsurface region after each loading cycle. Consequently, the total strain
accumulated in the subsurface, after a number of loading cycles, exceeds the strain produced by each
individual loading cycle [280, 282] (figure 5.3.1-2). The latter scenario would lead to failure through a low
fundamental role in the wear process. The understanding of the wear mechanisms of AISI 316 helps to
identify the cause of the significant increase in wear resistance observed after low temperature plasma
The plasma carburised specimens showed a significant improvement in tribological performance when
compared with the untreated AISI 316 (figures 4.4-1 and 4.4-3). Moreover, clear evidence of oxidational
wear mechanism was obtained from the appearance of the wear track (figures 4.4-4 to 4.4-6), the
appearance of the wear debris particles (figures 4.4-11 and 4.4-12) and their nature (figures 4.4-13 and
4.4-14).
Close observation of the wear tracks produced on PC specimens showed brittle failure of the outmost
layer of carbides (figure 5.3.2-1). This was very clear on samples with a thin EA layer, in which the brittle
carbide layer had very little or no support from the substrate [283]. On the other hand, the failure of the
outmost layer of carbides is not so clearly seen on specimens with thick EA layers. The carbides are
obviously removed upon wear, as the EA layer exhibits a very smooth surface with no signs of the
original particulate morphology (figure 5.3.2-2), but no evidence of spalling could be found on the sides
of the wear track. This can be attributed to the support provided by the underlying EA. In any case, the
outmost layer of deposited carbides does not seem to play a significant role in the wear resistance of
EA.
It is important to note that very little plastic strain is seen in the subsurface of worn PC specimens (figure
4.4-19). The deformation is mainly elastic, and plastic strain is only confined to the surface. The ratio of
untreated substrate to 20.48 for the plasma carburised layer. Therefore, according to the theory of
ratchetting, the failure mode would change from low cycle fatigue (ratchetting) to high cycle fatigue
[281]. The predominance of this wear mode depends on contact mechanics considerations [85], namely
the normal load, the friction force and the roughness of the sliding surfaces. The maximum load a
specimen can bear without undergoing plastic deformation is a function of the EA layer thickness, and
failure of plasma carburised components has been reported at high loads, induced by the plastic
In the case of a thick enough EA layer, or a low enough normal load, the stresses developed in the
subsurface are within the elastic range, and the wear process is limited to the surface of the specimen,
where a layer of oxide forms. The formation of this layer of oxides on the tribosurface is an indication of
a mild wear mode, which is commonly called oxidational wear [284]. The flow of heat is of significant
importance to this wear mechanism, as it defines the formation of the oxide layer. Additionally, different
types of oxide will be stable depending on the flash temperature, i.e. the temperature developed at the
contacting asperities [63]. In this way, haematite (α-Fe2O3) is reported to be the stable oxide at
temperatures below 300ºC, whereas magnetite (Fe3O4) is stable between 450ºC and 600ºC [58, 71,
284]. Regardless of the type of oxide, oxidational wear is a mechanism of mild wear.
The formation of the layer of oxide is thus of interest, in view of its relevance to the wear process, and
figure 5.3.2-3 shows the genesis of this layer. In the first place, small fragments are deformed under the
stress developed during sliding. It is important to note that these fragments may include WC particles
pulled out from the counterpart ball, as a result of the abrasion of the Co binder (figure 5.3.2-4) [285].
The fragments are subsequently compacted, to form a coherent layer, although the morphology of the
along the original deformation lines, as it is shown in figure 5.3.2-6. The development of fractures
perpendicular to the sliding direction is the first step in the detachment mechanism. The schematic in
figure 5.3.2-7 shows the different steps of nucleation and propagation of cracks until the final
These observations are in agreement with the literature on wear of EA [128] and also with the
oxidational wear mechanism proposed in [59, 63, 284]. Furthermore, the layer of oxide, i.e. the
existence of a layer of material which is different from the original materials involved in the wear
process, is usually mentioned in the specialised literature as the third body. A further development of the
third body concept in tribological systems has been presented in [286] and involves a state of
equilibrium between competing mechanisms of detachment of particles from the surfaces and ejection
of particles from the contact region. In this way, the protective oxide layer is thought to grow up to a
Figure 5.3.2-8 illustrates the effect of the oxide layer thickness. Even though previous illustrations
clearly showed oxide particles underwent deformation during sliding, the thick oxide glaze was found to
fracture, in a way that resembled a brittle solid [274]. Consequently, the tough and thin oxide layer
seems to grow throughout the sliding process until a relatively thick oxide glaze forms, which finally fails
under the sliding asperities in a brittle manner. The role of the supporting layer must not be overseen,
and the mechanical properties of the underlying EA are of fundamental importance to support the
reduces the plastic deformation of the subsurface and confines it to the surface, and it provides a
convenient support to the protective layer of oxides. Moreover, if a wear mechanism of abrasion or
microabrasion is taken into consideration, a third function can be added to the protective role of the EA
layer. In this case, the hardness ratio of the abrasive particles to the sliding surface is of importance
[129]. The principal abrasive particles in the system are Fe3O4 or α-Fe2O3 oxides (400-750 HV) [72,
274]. Thus, the cause of improvement becomes apparent as the hardness of the sliding surface is
increased from 200 HV to 800-900 HV, which is comparable or even higher than the abrasive particles.
With regard to friction, the interpretation of the friction force and friction coefficient is not unequivocal
and, additionally, there is a large scatter in the friction data reported in the literature for EA [141, 287-
290]. The results obtained in this project are in reasonable agreement with those reported in [128], i.e.
the EA layer exhibits some higher coefficient of friction compared to the untreated AISI 316 substrate.
This is consistent with the presence of a layer of rhombohedral α-Fe2O3 oxide, which is hard and
abrasive when compared to the more ductile cubic oxides, Fe3O4 and FeO [72, 291]. In addition, the
coefficient of friction increases with sliding distance until it reaches its steady state value, which could be
interpreted as an increase of the contact area upon the growth of the oxide layer [292]. Once the steady
condition is reached, the noise in the friction signal is a result of deformation processes, debris
4. Corrosion behaviour
4.1. Corrosion behaviour of AISI 316 untreated substrate in different acid solutions
The resistance of AISI 316 to sulphuric acid solutions is limited, and it mainly depends on the
temperature and concentration of acid in solution [30, 31, 37, 293]. In the particular case of the 16%
attack is only concentrated on weak points in the microstructure, especially MnS stringers (figure 4.5-1)
[38]. However, on increasing the temperature of the solution up to its boiling point, the rapid evolution of
hydrogen makes AISI 316 active and severe corrosion occurs (figures 4.5-3 and 4.5-8) [2].
Addition of oxidising agents, such as Fe3+ and Cu2+ cations, substitute hydrogen in the cathodic
reactions, and consequently stabilise the passive film. This extends the range of corrosion resistance of
ASS to higher temperatures or higher acid concentrations [31]. From the electrochemical point of view,
the potential of stainless steel in these sulphuric acid solutions, with respect to a saturated calomel
electrode, has been reported to change accordingly (figure 5.4.1-1) [2, 37].
With regard to nitric acid, AISI 316 is generally passive under a wide range of conditions [37]. However,
highly oxidising solutions can produce transpassive corrosion, which is particularly intense at grain
boundaries (figure. 4.5-14) [294]. This type of attack develops sensitised grain boundaries, although it
The PC samples exposed to sulphuric acid 16% solution at room temperature exhibited very little
corrosion, and only the outmost layer of material was found to fail (figure 4.5-1). This early stage of
corrosion was confirmed during the tests conducted at boiling temperature, in which some material was
found to peal off the surface of the samples after only a few minutes of exposure. SEM and TEM studies
showed that the recovered fragments were of a similar nature to the outmost layer of material previously
studied (figure 5.4.2-1). The XPS analysis of this outmost layer yielded C, Cr, O and Fe peaks (figure
Further experiments conducted with polished samples, in sulphuric acid solution at room temperature,
showed good corrosion resistance in all cases. This clearly indicates that the corrosion resistance is not
a result of the deposition layer of carbides, nor of a thicker oxide layer formed upon plasma treatment
(figure 4.2-7). In a similar way as it was reported in [297], the primitive oxide layer was removed after
only a few minutes of exposure to the boiling solution of sulphuric acid. Thus, the carbon expanded
austenite was found to be corrosion resistant itself, and the passive film must have been formed, in-situ,
The corrosion resistance of carbon, nitrogen and carbon-nitrogen expanded austenite has been
reported at length [96, 125, 132-134, 298-304]. Most of the evidence is in the form of polarisation curves
This description is in agreement with the results obtained in this project for corrosion in boiling sulphuric
acid solutions, namely: reduced evolution of hydrogen, extended range of passivity and lower corrosion
corrosion resistance, but the experimental details are sparse [305]. The reported mechanism of
dissolution of nitrogen and alkalisation can not explain the good response of the carburising layers [306]
cited in [307]. More recent studies have proposed different mechanisms of passivation, including salt
passive films [308, 309], hydrated oxides [307], modification of the semiconductor characteristics of the
passive film [303] or even the same corrosion mechanism as in the untreated substrate [304].
In reducing solutions, under cathodic potentials (figure 5.4.1-1), the improvement in corrosion resistance
is significant. The evolution of hydrogen is not so readily observed after immersion, and the expanded
austenite clearly shows a wider window of passivity, when compared to the untreated AISI 316
substrate (figure 4.5-8 and 4.5-9). The improvement is limited by the susceptibility of EA layers to
localised corrosion on weak points of their structure (figures 4.5-4 to 4.5-6). These are clearly developed
in more oxidising environments, such as sulphuric acid solutions with Fe3+ and Cu2+ additions, and also
in nitric acid solutions (figures 4.5-11 and 4.5-14). The latter solutions shift the potential of ASS in the
anodic direction and render it passive, so that very mild or negligible general corrosion of the grain
surface occurs, whilst the dissolution of susceptible areas is accelerated, thus increasing contrast [295].
Elongated MnS stringers reaching the surface are known to diminish the corrosion resistance of surface
alloyed ASS [142], and the attack progresses very rapidly after immersion in the boiling sulphuric acid
solution (figure 4.5-4). The presence of slip bands evidences plastic deformation taking place during the
treatment as a result of very high compressive stresses developed by the expansion of the austenite
[238]. These defects, together with some susceptible grain boundaries, are preferential sites for
the substrate, where subsurface pits are formed (figures 4.5-5 and 4.5-6) [311].
Consequently, it can be said that corrosion of PC specimens is controlled by the density of defects in
the expanded austenite layer. Figure 5.4.3-1 shows a schematic of the failure modes observed on
treated samples. The most intense attack takes place on MnS inclusions, but localised corrosion also
occurs on slip bands and grain boundaries. Corrosion progresses at a continuous rate, depending on
the type of defect, until the substrate material is exposed. At this stage, a rapid anodic dissolution of the
substrate takes place and subsurface pits of significant size are formed. In areas where the density of
defects is high, the subsurface pits coalesce and the expanded austenite layer flakes off. In applications
involving corrosion-wear, the formation of cavities in the subsurface layer would significantly affect the
The density of slip bands and of susceptible grain boundaries is higher at the edges than in the central
area of the specimens (figure 5.4.3-2). Furthermore, the density of these defects increases with the
applied power (i.e. ASPC 030, ASPC 125 and DCPC 400 in increasing order). This is attributable to the
strong plasma-surface interactions developed on sharp edges, and the associated local increase in
current density and temperature. The bias levels, thus the power and current densities, selected for
each ASPC treatment account for the observed differences in corrosion performance. In this way, low
bias levels produce uniform surface morphologies on the treated samples and, consequently, a more
even and predictable corrosion behaviour is observed. On the other hand, high bias levels and,
ultimately, DCPC produce edge effects, and the consequent uneven or localised corrosion.
The slip bands observed on the PC samples are non-equivocal evidence of plastic deformation taking
place during the treatment, presumably due to the high residual stresses developed upon the expansion
of the austenite unit cell. The dislocation walls so formed can become subgrain boundaries (figure 5.2-7)
[245]. Moreover, due to the constraints imposed by the neighbouring grains, plastic deformation usually
implies some degree of rotation of the crystallographic planes, as it has been reported in deformation
studies conducted in ASS [242], and also for EA [247-249]. This rotation has a significant effect on the
type of grain boundaries and can, eventually, lead to the development of a deformation texture [245,
246]. Finally, precipitation of carbides is also affected by deformation structures and grain boundary
characteristics [20, 24, 314]. All these features —namely crystallographic texture, grain boundary
characteristics and precipitation of carbides— have a marked effect on the corrosion properties of ASS
The effect of crystallographic texture on the corrosion resistance of ASS was reported in [44], where
closely packed crystallographic planes were found to favour the formation of the passive film, and
reduce the amount of sites for localised corrosion, with the consequent increase in corrosion resistance.
In this project, even though deformation took place, no significant deformation texture seems to have
developed (figure 5.4.4-1). The changes in the inverse pole figures obtained on untreated AISI 316 and
on PC showed no consistent trend, so that texture can not be considered to contribute significantly to
the corrosion resistance of EA. Furthermore, the quantity of favourable sites for localised corrosion is
With regard to the grain boundaries, several studies have recently reported improvements in corrosion
resistance of ASS through an increase in the number fraction of special grain boundaries and a
this project is towards a degradation of CSL orientation grain boundaries (figure 5.4.4-2), which are
converted into stepped boundaries as reported in [245]. This supports the results reported in [316],
where an increase in corrosion resistance was obtained by increasing the frequency of random grain
boundaries. The postulated mechanism is related to the flux of chromium and the precipitation of
Precipitation of chromium carbides is of significant importance to the corrosion resistance of ASS since
it produces a depletion of chromium in the surrounding areas, and the consequent weakening of the
passive film. This process is mainly dependent on the diffusion of chromium and, consequently, on
temperature [314], thus sharp corners are more susceptible to exhibit precipitates due to the edge effect
developed on them (figure 5.4.4-3). However, the deformation structures have been found to play an
important role in carbide precipitation and the high energy grain boundary ledges produced on
intersection of grain boundaries with slip bands, or even the intersection of two slip bands seem to offer
favourable sites for precipitates [19, 314, 317, 318]. Moreover, recent studies have reported early
stages of precipitation at temperatures as low as 250ºC [319]. With regard to carbon expanded
austenite, precipitation of small needles of M5C2 carbides have been reported to be favoured at
crystallographic defects, where diffusion of substitutional elements (mainly Cr and Ni) requires minimum
In this project, evidence of carbides has been obtained by means of XRD, TEM (figure 4.2-8) and EDX
(figure 4.5-13). Furthermore, the selective attack observed in corrosion experiments is consistent with
the type of sensitisation reported in [320] associated with precipitation at deformation structures. Given
the high carbon content of the EA, the presence of favourable sites renders precipitation of carbides
short range diffusion of substitutional elements, so that the chromium depletion is of a local nature, thus
The spectra in figures 4.6-1 to 4.6-3 show the typical optical emission peaks obtained for gas mixtures
corresponding to nitriding, carburising and nitrogen enhanced carburising. Although the existence of N
and C containing species is essential for these surface treatments, the presence of other species —
either as pollutants, as by-products of a primary reaction, or as intentional additions to the gas mixture—
can have a non-trivial effect in the physicochemistry of plasma. Some of these reactions are considered
of importance for the interpretation of the OES results obtained in this project and will be discussed in
Plasma nitriding in N2-H2 and in N2-H2-Ar gas mixtures have been extensively studied and a large
amount of evidence has been gathered [165, 321-327]. The effect of Ar and H in depassivation and
surface activation is usually highlighted. In the case of Ar, this is achieved through increased sputtering
power of the gas mixture, whereas H increases the chemical etching of O species. Moreover, the
dissociation of N2 is increased with small additions of H2 (less than 1%), which was attributed to the
strong interaction (population and depopulation) between vibrational levels of both molecular gases
[328-331]. The electric parameters of the discharge also vary with increasing H2 content of the gas
mixture, and a minimum in discharge impedance is found around 75% N2 − 25% H2.
CH4-H2-Ar plasmas have been extensively studied for deposition of diamond-like carbon [333-338]. In
these studies, atomic H has been clearly identified as a by-product of the decomposition of the CH4
molecule. In addition, a role of significance has been assigned to atomic H in the dissociation of CH4
[336, 337]. Depending on the gas pressure and gas temperature, dehydrogenation of methane by H
The hydrogen abstraction efficiencies of CO and OH have been reported to be even higher than atomic
H [338, 339]. Bearing in mind these same principles, and due to technical limitations to do experiments
with other gases (CO, CO2 or O2), small additions of N2 were used. Contrary to what could be expected
from the dissociation enthalpies listed in table 5.5-1 [340], a strong interaction between N and C is
evident from the relatively high activity of CN groups in figure 4.6-2. The interaction between C and N to
form CN species has also been reported in [341], and this mechanism would involve a chemical driving
force, similar to the one reported in [339] for dissociation of CH4 with small additions of CO. This result
highlights the importance of chemical interaction between the species present in the plasma phase.
The emission intensity maps obtained by means of the optical fibre probe yielded good information on
the plasma under different treatment conditions (figure 4.6-5). The selection of the emission bands used
for the purpose of the emission intensity maps was very much influenced by the signal-to-noise ratio.
However, there are good reasons to correlate these bands with the N and C activity in the plasma. In
the case of nitriding, transition N2+ (B2Σu+ − X2Σg+) is usually the most intense, with a band head at
391.44 nm [342, 343]. Moreover, this so called first negative system of N2 has been related to
dissociation of molecular nitrogen, in view of the high energy required for direct dissociation [345, 346].
The emission intensity is known to increase with the electron density (ne) and the electron temperature
(Te) and, as a consequence, the emission intensity increases with the power (figure 4.6-6) [347]. The
electron temperature is particularly important, assuming that ionisation takes place mainly through
single collisions with energetic electrons. However, for a molecular gas, and in particular for N2, step
ionisation from excited levels must also be considered [348-351]. An example of stepwise excitation
proceeds as follows:
e − + N 2 ( X ) → e − + e − + N 2+ ( X ) (5.5-1)
e − + N 2+ ( X ) → e − + N 2+ (B ) (5.5-2)
where X indicates the ground state and B the excited state, and this or similar mechanisms would be of
importance at high electron densities and low electron temperatures. This is believed to be the case at
the high pressure range assessed in this project (P > 250 Pa), where the electron temperature has been
reported to drop due to the increasing number of collisions, and the electron energy distribution function
Figure 5.5.2-1 illustrates the relative increase in the emission intensity of the nitrogen first negative
system, with respect to the hydrogen Balmer series, at increasing gas pressure in ASPN conditions.
Considering the energy required for direct electron impact excitation of Hα and N2+ are 12.74 eV [346]
and 18.5 eV [357] respectively, it seems plausible that, in the high pressure range, the N2+ is being
populated through stepwise ionisation, which lowers the energy threshold to 10.5 eV [357]. Moreover,
these changes in the electron energy distribution function (EEDF) and the ionisation processes with gas
In the case of ASPC, the absence of intense peaks corresponding to C containing species limits the
analysis to the hydrogen Balmer series. As it was mentioned in section 5.1 of this chapter, atomic
hydrogen is not only a by-product of the decomposition of the methane molecule, but it also plays an
active role in the dehydrogenation of CHX groups. Therefore, its study could be considered to represent,
to some extent, the C activity in the plasma. Nevertheless, the study of CHX emission peaks in the
The intensity maps obtained under carburising atmospheres (figure 4.6-5) exhibit one interesting
feature: in the low pressure region, the emission intensity was found to increase with the applied bias up
to a maximum value, near the current saturation region, followed by a marginal decrease on further
increasing the bias. This was contrary to the expectations, as higher applied power was supposed to
yield higher emission intensity. The trend was initially considered an experimental error, but the same
phenomenon was observed on repetition of the experiment in CH4-H2 gas mixtures, although it was not
so evident or not observable at all, in N2-H2. A possible explanation for this phenomenon became
apparent from the results of the ion flux probe, discussed in the next section.
The flux of ions measured in the AS industrial furnace under both nitriding and carburising atmospheres
shows the same trend, i.e. the ion flow increases with the bias level, and decreases with gas pressure
(figures 4.7-5 and 4.7-6). The effect of bias seems self-evident, as an increase in the negative potential
applied to the worktable increases the electric field, thus the ion drift. In the case of pressure, two
changes in the EEDF which affect the production of ions, as discussed in the previous section [220].
The trend followed by the floating potential of the probe (i.e. the potential at which equal number of
electrons and ions arrive at the probe, thus yielding zero net current) is noteworthy. The floating
potential initially dropped to more negative values when bias was applied, followed by a reverse in the
trend on further increasing the bias, and this behaviour was clearer in CH4-H2 than in N2-H2 gas
mixtures (figure 4.7-4). Although the phenomenon was initially considered an artefact, its resemblance
On further analysing the OES and IFP signals and the electric parameters of the AS furnaces during the
experiments, it was found that the reduction in emission intensity and the shifting of the floating potential
to more positive values, correlated with considerable reductions in the power applied to the mesh of the
AS furnace. This reduction was imposed by the automatic control system of the furnace in order to keep
the temperature stable, to compensate for the increasing power applied on the worktable. In view of
these observations, the cathodic mesh in the AS furnace could be considered an additional source of
With regard to the measurements conducted with the electric probe in DC and AS experimental
arrangements inside a conventional DC furnace, similar trends were observed: the AS arrangement
produced larger ion fluxes, and changed the floating potential (figure 4.7-2). However, the latter varied
depending on the gas mixture. For the carburising atmosphere, the floating potential dropped to more
negative values, whereas under nitriding atmospheres the potential shifted in the positive direction.
parameters were applied in both the DC and the AS experiments, the electron temperature is thought to
remain constant. Therefore, the differences may be attributed to an increase in the electron density,
which is, once again, consistent with the results reported for triode discharges [359-361]. In the
particular case of nitrogen containing atmospheres, higher electron densities would render step
ionisation processes more profuse, thus increasing the density of ions and shifting the floating potential
to more positive values (i.e. less negative floating potentials are required to balance the flow of electrons
and ions).
However, there is an additional factor that should be considered in the case of an AS furnace, namely
the geometry of the cathodic mesh. The mesh in an AS arrangement is negatively biased and it
surrounds the workload, so that there is an electric field which tends to confine electrons within the
operative volume of the furnace. This concept is by no means new, and there is a large body of
literature on the electrostatic confinement of plasma in a variety of ways, and using different electrode
A considerable increase in the plasma density is reported by confining charged particles, as the path of
electrons from the cathode to the anode is extended due to reflections on the cathodic surfaces [223,
362]. The confinement effect is restricted to the low energy electrons (slow or cold electrons), whilst the
high energy electrons (fast or hot electrons) would escape the trapping electric field [366]. The
dimension and the geometry of the chamber are of significant importance in this regard. Furthermore,
the confinement is consistently reported to be more effective in the low pressure range [221, 363, 365],
in which the negative glow extends to fill the whole volume within the confining field.
because it is in this region of the gas discharge where the active species are produced. Therefore, the
high electron density developed inside the mesh of an active screen furnace facilitates the creation of
active species. Furthermore, the electron density has already been proposed as a parameter to predict
and control the carbon potential of the plasma [171]. Consequently, the parameters of the electric
discharge affect the chemistry of the gas (increase activity through ionisation), and this, at the same
time, affects the parameters of the electric discharge (changes in EEDF observed in molecular gases).
It is thus proposed that the AS arrangement contributes to the creation of additional primary electrons
and confinement of charged particles within the operative volume of the chamber, yielding plasma
conditions more suitable for nitriding and carburising. The electric parameters applied on the workload
can be thus controlled independently to produce the required degree of ion bombardment on the
It is convenient to begin the discussion on the mechanisms of active screen plasma surface engineering
by talking about nitriding, as this has been more extensively studied than carburising. The evidence
found in the former will be comparatively discussed with respect to the latter.
One of the most interesting findings of the experiments on the mechanisms of active screen plasma
surface engineering is the layer of material deposited on the ASPN treated samples (figure 4.3-7). This
deposition layer covered the AS plasma nitrided specimens with a morphology that depended on the
slides were especially useful for XRD studies (figures 4.3-10), and the pattern shows a very clear
coincidence with FeXN nitrides and a remarkable correlation with FeN (figure 4.3-11). This experimental
finding is of significance considering that the deposition layer has been deemed in the literature as one
The sputtering and redeposition mechanism was first proposed by Kölbel and has been supported and
reviewed by several authors ever since [154, 159, 160, 371, 372]. The model attributes the nitriding
mechanism to iron atoms which are sputtered from the specimen, and subsequently react with the
nitrogen in the plasma. Unstable FeN nitrides would thus be redeposited on the specimen surface, and
would decay into lower Fe2-3N and Fe4N nitrides, releasing nitrogen atoms free to diffuse (figure 5.7.1-
3). In the case of active screen plasma nitriding, the mesh was proposed to be the source of sputtered
iron atoms [181-183], in view of the absence of ion bombardment or sputtering on specimens treated at
Further evidence on the importance of the transferred material is apparent from the results obtained
under an AS arrangement using a mesh with big holes. In the case of ASPN, the area of the sample
covered by the mesh exhibited significant hardening effect, while the area under the hole showed little
signs of nitriding (figure 4.3-6). The appearance of both areas of the sample is very similar, and only a
minor change in the metallic shine is observed, the area covered by deposited material being dull grey,
whereas the area under the hole retained its metallic shine. In terms of composition, the EDX analysis of
the deposited material revealed alloying elements such as Cr and Ni in addition to Fe (figure 5.7.1-4), in
agreement with the composition of the AISI 304 mesh (figure 4.3-12).
extensively discussed and reported in the literature on active screen plasma nitriding [180, 373-376]. To
this end, and given the evidence produced during this project, sputtering and deposition is an irrefutable
fact, although its relevance for the treatment is still an open question, and it has been seriously
confronted [185, 377]. The results obtained in this project indicate that the treatment conditions for the
thickest deposition layer of iron nitride differ from those ones for the thickest layer of nitrogen expanded
austenite (figure 5.7.1-5). Indeed, the very concept of active screen plasma nitriding was not developed
on the theory of sputtering and redeposition, but on the post-discharge nitriding, i.e. the generation of
active species on the mesh and their transfer to the worktable through a carefully designed gas flow
[177].
From the experience gathered in this project, the contribution made by the gas flow to the transfer of
active species from the mesh towards the worktable is arguable. This is evident considering that
experiments conducted inside DC furnaces have been widely successful without introducing any special
modifications in the gas flow [180-183, 374, 378]. Furthermore, experimental arrangements with plate
lids, which severely restrict gas flow, have produced similar or even better results than those with mesh
lids [182, 378]. In addition, the distance active species can travel as such, is limited due to their
considerably short lifetimes. This was observed in the experiments conducted without a lid in [182] and
also in the experiments conducted in this project using a mesh with big holes, in which the hardening
effect was limited to regions of the specimens lying in proximity to the active screen mesh (figure 4.3-6).
Consequently, the post-discharge mechanism seems to be limited to relatively small scales and, in fact,
it has been found difficult to produce satisfactory hardening in large scale active screen units without
applying some level of bias on the treated components [184, 185, 377, 379]. Nevertheless, the
efficiency of the plasma treatment showed a substantial degree of correlation with the optical emission
intensity of active species. At the same time, the emission intensity was affected by the electric
discharge conditions on the mesh. Consequently, the influence of the mesh on the treatment becomes
apparent, and it is here proposed that the role played by the mesh in an AS furnace is threefold: it
radiates heat, it is an additional source of electrons in a similar way to a triode discharge; and it confines
These last two roles of the mesh in an AS furnace are not mutually exclusive, but synergetic. The
beneficial effect of an additional source of electrons has been reported in experiments conducted in
triode discharges [359-361] and was discussed in a previous section. However, it has also been noted
that the plasma density is not uniform in these arrangements [380]. On the contrary, the geometrical
arrangement of active screen furnaces produces an even distribution of electrons around the workload.
Furthermore, under suitable conditions, the confining negative electric field generated by the mesh
would also contribute to produce a higher electron density and more uniform distribution of electrons
The theory is consistent with the results obtained in this project, and those ones reported in the literature
on active screen [185], triode discharges [359, 360] and electrostatic confinement of plasmas [223, 362,
363]. The additional electrons provided by the mesh allow operating the furnace at lower pressures, at
which the electrostatic confinement is also more effective. The higher electron density implies an
increased number of collisions with gas molecules, which will eventually lead to a larger number of
active species and a higher plasma density. This can finally be translated into an increased plasma
379]. This issue recalls the ancient discussion on the effect of ions versus the role of neutrals in plasma
nitriding [381-383]. The equipment used in this project did not allow for a complete decoupling of the
voltage and current discharge characteristics of the worktable, which are related with the ion energy and
ion flux. Therefore, the results are not conclusive, although there is a clear increase in the hardening
effect and the EA layer thickness with both increased voltage and current, i.e. with increased applied
power.
Even though it was possible to nitride samples at floating potential, this condition did not produce the
thickest layer of EA. Furthermore, the adhesion of the deposition layer was not satisfactory and the layer
of iron nitrides was found to peel off in some regions (figure 5.7.1-6). It can be thus concluded that a
small level of bias is not only desirable, but also necessary to optimise the treatment conditions. The
deposition layer seems to be beneficial, but it is not the only mechanism operating in ASPN treatments.
The evidence indicates that the active species in the plasma are of fundamental importance to control
the nitriding potential. Therefore, the optimisation of the treatment would require uncoupling the plasma
generation and heat source from the treated components [361, 380], and carefully selecting the
discharge conditions applied on the worktable, in order to favour the plasma-surface interactions, which
The results obtained for AS plasma carburising indicate similar trends to AS plasma nitriding, although
the operating mechanisms are clearly different (figure 4.3-10). The first significant distinction observed is
the complete absence of a coherent and dense deposition layer. The glass slides treated in ASPC
conditions only exhibited some fine powder on their surfaces which, according to the EDP obtained in
possible from the deposition film grown on the walls of the furnace —i.e. on the anodic walls lying some
100 mm in front of the cathodic mesh— after several consecutive carburising treatments. The
morphology is obviously coarser (figure 5.7.2-2), but it is clearly formed by conglomeration of smaller
particles, and the XRD result indicates a very good correlation with the Fe3C pattern.
A thin layer of carbides was found on ASPC treated AISI 316 samples (figure 4.2-8), and it has been
reported in the literature [384]. However, the role of these carbides in the mechanism of plasma
carburising treatments seems to be negligible. Therefore, plasma carburising seems to arise as a result
of the step decomposition of the methane molecule according to the following reactions [163, 385]:
e − + CH 4 → CH 3+ + H + 2e − (5.7-1)
e − + CH 3 → CH 2+ + H + 2e − (5.7-2)
e − + CH 2 → CH + + H + 2e − (5.7-3)
e − + CH → C + + H + 2e − (5.7-4)
The ions would thus be adsorbed at the cathodic surface and would release C atoms free to diffuse into
the bulk material. The fact that CH and C ions or excited species were not abundant in the OES spectra,
collected during the ASPC treatments, leads to the assumption that CHX species with x > 1 would be
responsible for the hardening effect. This assumption implies that a significant amount of hydrogen is
adsorbed with the carbon, which was suggested in [386] and confirmed in [387].
The quantity of reports on plasma carburising is sparse, although there is a high degree of agreement in
the evidence. In general, the carburising is attributed to reactions under electric control, which
correlation between ions and excited species, and the carbon mass flow has been established;
moreover, the electron density has been proposed as a convenient variable to control the carbon
potential of the plasma [171]. All this evidence is in favour of the model proposed for AS plasma surface
engineering, in which higher electron densities would increase the number of collisions and the plasma
reactivity.
However, this explanation is not completely satisfactory in view of the difficulty to produce hardening on
samples left at floating potential. This was initially thought to be a consequence of the presence of
passivating oxide at the surface of AISI 316 specimens. Therefore, some experiments were conducted
applying a surface activation stage at the beginning of the treatment, consisting of bombardment of the
surface with argon ions, but the results were equally negative. Furthermore, the high hydrogen content
of the gas mixture is supposed to chemically etch the oxides formed on stainless steel samples, and to
increase the tolerance of the treatment to residual oxygen gas [388]. Nevertheless, the results were
negative in all the attempts, thus it can be concluded that the application of bias is essential for ASPC.
The equipment used in this project did not allow for an independent selection of potential and current,
but the hardening effect was found to increase with both, i.e. with the applied power.
This need for ion bombardment to produce hardening in H2-CH4 gas mixtures does not weaken the
potential benefits offered by the AS technique. The uncoupling of the plasma and heat sources from the
treated components, provides an additional degree of freedom to optimise the treatment. Furthermore,
the possibility to increase the electron density, the plasma density and the plasma uniformity within the
operative volume of the furnace, is clearly an interesting feature of AS. Finally, the transfer of material
The evidence produced throughout this project and the discussion presented in the last chapter can be
summarised as follows:
1. It is possible to generate carbon expanded austenite in austenitic stainless steel using active
2. Based on systematic experimental work, effective AS processing windows have been identified to
3. Optimal treatment conditions for maximum layer thickness and no incidence of edge effect were
identified at 125 Pa and 15% bias for ASPC, and 75 Pa and 10% bias for ASPN.
4. The surface quality of active screen treated samples could be significantly improved compared to
the DC counterparts. Defects such as edge effects, hollow cathodes and arcing were minimised
or eliminated.
5. The layer of carbon expanded austenite produced in the active screen furnace was equivalent to
the one obtained in DC furnaces in terms of layer thickness, composition, phase constitution and
hardness. Improved surface quality was consistently obtained under the right AS treatment
conditions.
6. The microstructure of the treated material was severely deformed and distorted. Degradation of
the CSL boundaries took place, with a trend towards a randomisation of the grain boundaries.
maximum compressive residual stresses were calculated to be in the order of -1.6 GPa. Residual
stresses were found to increase with the ion bombardment and the layer thickness.
8. Under the conditions assessed in this project, plasma carburised AISI 316 exhibited a resistance
9. The dominant wear mechanisms were found to be changed from severe adhesive and
delamination wear for the untreated AISI 316, to mild oxidational wear for the plasma carburised
material.
10. The carbon expanded austenite can avoid subsurface plastic deformation, increase resistance to
abrasion by hard particles or asperities, and provide support to the protective oxide layer.
11. The coefficient of friction exhibited by carbon expanded austenite was higher than the untreated
12. Plasma carburised AISI 316 showed higher corrosion resistance than the untreated AISI 316 in
boiling 16% sulphuric acid solutions. Furthermore, active screen plasma carburising produced
13. The outer layer of carbides played no protective role against corrosion, but it failed shortly after
being immersed in acid solutions. The improved corrosion resistance was attributed to the carbon
The weak points of the microstructures include: MnS inclusions, deformation slip bands and grain
boundaries. Sharp edges were particularly susceptible areas to develop the latter two defects.
15. Optical emission spectroscopy, in the UV-visible range, was a valuable technique to identify
active species in N2-H2 gas mixtures, although it was severely limited in the case of diluted CH4-
H2 atmospheres.
16. The emission intensity of nitrogen ionic species and hydrogen atomic species showed a
considerable correlation with the treatment effectiveness. Therefore, the intensity maps obtained
with the optical fibre probe were useful to identify the optimum range of treatment conditions.
17. The interpretation of the ion flux probe signal indicates that the plasma inside an active screen
furnace was considerably affected by the electric parameters applied on the cathodic mesh. The
flow of ions increased with increasing bias and decreased with increasing gas pressure.
18. Sputtering and redeposition was found to be a mechanism of significance for active screen
plasma nitriding, although the active plasma species and some moderated ion bombardment
19. Ion bombardment was a necessary condition for active screen plasma carburising. The activation
provided by the plasma was of importance for the decomposition of methane, whereas sputtering
20. The role of the cathodic mesh in an active screen furnace seems threefold: it radiates heat, it is
an additional source of electrons in a similar way to a triode discharge; and it confines low energy
The present project contributed to the understanding of active screen plasma processing, the
characterisation of carbon expanded austenite, its tribological properties and corrosion behaviour.
Evidence was produced on the mechanisms involved in the active screen plasma treatments, and the
diagnostic techniques allowed some basic characterisation of the plasma. From the results obtained in
this project, the following topics are suggested for future studies:
The metallic mesh of AS furnaces may be replaced by other electrode arrangements, which
could be equally or even more effective in producing additional electrons and plasma
confinement, and show further flexibility and simplicity, e.g. segmented hollow cathodes.
Plasma diagnostic with optical fibre assisted OES and electric probes provide useful and
complementary information about the plasma. Further improvements of these techniques would
be valuable, for example to produce spatial distribution maps of active species, electron density
1. Ashby, M. F.: Materials selection in mechanical design, 2nd Ed. (1999) Oxford, Butterworth
Heinemann
2. Streicher, M. A.: Stainless steels: past, present and future in 'The metallurgical evolution of
stainless steel' 1st Ed. (1979) London, The Metals Society, 442-475
3. Keown, S. R.: The history of alloy steels with special reference to Sheffield contributions,
Perspectives in Metallurgical Development (1984) Sheffield, The Metals Society, 36-46
4. Bhadeshia, H. and Honeycombe, R.: Steels: microstructures and properties, 3rd Ed. (2006)
London, Butterworth-Heinemann
5. Pickering, F. B.: Introduction: The metallurgical evolution of stainless steel in 'The metallurgical
evolution of stainless steel' 1st Ed. (1979) London, The Metals Society, 1-42
6. Pickering, F. B.: Physical metallurgical development of stainless steel, Stainless Steel '84
(1984) Göteborg, London Institute of Metals, 2-28
7. Oshima, T., Habara, Y. and Kuroda, K.: Efforts to save nickel in austenitic stainless steels, ISIJ
International (2007) 47, 3, 359-364
8. British Standard DPC 08/30166437 DC, 2008, "Stainless steels: chemical composition", BSI,
London, www.bsi-global.com
9. Voronenko, B. I.: Austenitic-ferritic stainless steel: a state-of-the-art review, Metal Science and
Heat Treatment (1997) 39, 9-10, 428-437
10. New 200-series steels: An opportunity or a threat to the image of stainless steel?, International
Stainless Steel Forum (2005) November, 1-14
11. Shanina, B. D., Gavriljuk, V. G., Berns, H. and Schmalt, F.: Concept of a new high-strength
austenitic stainless steel, Steel Research (2002) 73, 3, 105-113
12. Leslie, W. C. and Hornbogen, E.: Physical metallurgy of steels in 'Physical metallurgy' 4th Ed.
(1996) Oxford, Elsevier North Holland, 1556-1620
13. Handbook Committee, A. S. M.: Metals Handbook Vol 8: Metallography, structures and phase
diagrams, 8th Ed. (1973) Novelty, Ohio, ASM American Society for Metals
14. Marshall, P.: Austenitic stainless steel: microstructure and mechanical properties, 1st Ed.
(1984) London, Elsevier Applied Science Publishers
15. Villanueva, D. M. E., Junior, F. C. P., Plaut, R. L. and Padilha, A. F.: Comparative study on
sigma phase precipitation of three types of stainless steels: austenitic, superferritic and duplex,
Materials Science and Technology (2006) 22, 9, 1098-1104
16. Angel, T.: Formation of martensite in austenitic stainless steels in 'The metallurgical evolution of
stainless steel' 1st Ed. (1953) London, The Metals Society, 402-412
18. Fang, X. F. and Dahl, W.: Strain hardening and transformation mechanism of deformation-
induced martensite transformation in metastable austenitic stainless steels, Materials Science
and Engineering A: Structural Materials: Properties, Microstructure and Processing (1991)
A141, 189-198
19. Hong, H. U., Rho, B. S. and Nam, S. W.: Correlation of the M23C6 precipitation morphology
with grain boundary characteristics in austenitic stainless steel, Materials Science and
Engineering A: Structural Materials: Properties, Microstructure and Processing (2001) A318,
285-292
20. Trillo, E. A. and Murr, L. E.: A TEM investigation of M23C6 carbide precipitation behaviour on
varying grain boundary misorientations in 304 stainless steels, Journal of Materials Science
(1998) 33, 5, 1263-1271
21. Bruemmer, S. M. and Charlot, L. A.: Development of grain boundary chromium depletion in type
304 and 316 stainless steel, Scripta Metallurgica (1986) 20, 1019-1024
22. Jack, D. H. and Jack, K. H.: Carbides and nitrides in steel, Materials Science and Engineering
(1973) 11, 1-27
23. Sourmail, T.: Precipitation in creep resistant austenitic stainless steels, Materials Science and
Technology (2001) 17, 1, 1-14
24. Trillo, E. A. and Murr, L. E.: Effects of carbon content, deformation and interfacial energetics on
carbide precipitation and corrosion sensitization in 304 stainless steel, Acta Materialia (1999)
47, 1, 235-245
25. Williams, T. M.: Precipitation in neutron-irradiated type 316, Stainless Steel '84 (1984)
Göteborg, London Institute of Metals, 403-412
26. Padilha, A. F., Escriba, D. M., Materna-Morris, E., Rieth, M. and Klimenkov, M.: Precipitation in
AISI 316L(N) during creep tests at 550 and 600 C up to 10 years, Journal of Nuclear Materials
(2007) 362, 132-138
27. Farrar, R. A.: Influence of microsegregation on phase transformations and properties of type
316 weld metals at elevated temperatures, Stainless Steel '84 (1984) Göteborg, London
Institute of Metals, 336-342
28. Sourmail, T. and Bhadeshia, H.: Modelling simultaneous precipitation reactions in austenitic
stainless steels, Computer Coupling of Phase Diagrams and Thermochemistry (2003) 27, 2,
169-175
29. Slattery, G. F. and O'riordan, P.: Microstructural transformations in stress relieved type 316
stainless steel weld metal, Metallography (1980) 13, 59-70
31. Davis, J. R.: Atmospheric and aqueous corrosion in 'ASM Speciality Handbook: Stainless
steels' 1st Ed. (1996) Materials Park, Ohio, ASM International, 133-153
32. Evans, U. R.: The passivity of metals: Part I - The isolation of the protective film, Journal of
Chemical Society (1927) May, 1020-1040
33. Vernon, W. H. J., Wormwell, F. and Nurse, T. J.: The surface film on chromium-nickel 18-8
stainless steel, Metallurgia (1944) 31, 181, 19-22
34. Dye, T. G., Fursey, A., Lloyd, G. O. and Robson, M.: Stripping thin oxide films from metals,
Journal of Physics E: Scientific Instruments (1968) 1, 2, 463-464
35. Sudesh, T. L., Wijesinghe, L. and Blackwood, D. J.: Characterisation of passive films on 300
series stainless steels, Applied Surface Science (2006) 253, 1006-1009
36. Montemor, M. F., Simoes, A. M. P., Ferreira, M. G. S. and Da Cuhna Belo, M.: The role of Mo in
the chemical composition and semiconductive behaviour of oxide films formed on stainless
steels, Corrosion Science (1999) 41, 17-34
37. Sedriks, A. J.: Corrosion of stainless steels, 1st Ed. (1979) New York, John Wiley and sons
38. Sedriks, A. J.: Metallurgical aspects of passivation of stainless steels, Stainless Steel '84 (1984)
Göteborg, London Institute of Metals, 125-133
39. Brandis, H. and Kiesheyer, H.: Influence of chromium and molybdenum on chloride corrosion of
stainless steels, Stainless Steel '84 (1984) Göteborg, London Institute of Metals, 217-221
40. Dayal, R. K., Parvathavarthini, N. and Raj, B.: Influence of metallurgical variables on
sensitisation kinetics in austenitic stainless steels, International Materials Review (2005) 50, 3,
129-155
41. Matula, M., Hyspecka, L., Svoboda, M., Vodarek, V., Dagbert, C., Galland, J., Stonawska, Z.
and Tuma, L.: Intergranular corrosion of AISI316L steel, Materials Characterization (2001) 46,
2-3, 203-210
42. Barbi, N. C. and Judd, G.: Microstructure and microsegregation effects in the intergranular
corrosion of austenitic stainless steel, Metallurgical Transactions A: Physical Metallurgy and
Materials Science (1972) 3, 11, 2959-2964
43. Brandon, D. G., Ralph, B., Ranganathan, S. and Wald, M. S.: A field ion microscope study of
atomic configuration at grain boundaries, Acta Metallurgica (1964) 12, 7, 813-821
44. Ravi Kumar, B., Singh, R., Mahato, B., De, P. K., Bandyopadhyay, N. R. and Bhattacharya, D.
K.: Effect of texture corrosion behaviour of AISI 304L stainless steel, Materials Characterization
(2005) 54, 2, 141-147
46. Michiuchi, M., Kokawa, H., Wang, Z. J., Sato, Y. S. and Sakai, K.: Twin-induced grain boundary
engineering for 316 austenitic stainless steel, Acta Materialia (2006) 54, 19, 5179-5184
47. Randle, V.: Twinning-related grain boundary engineering, Acta Materialia (2004) 52, 14, 4067-
4081
48. Rohrer, G. S., El Dasher, B. S., Miller, H. M., Rollett, A. D. and Saylor, D. M.: Distribution of
grain boundary planes at coincident site lattice misorientations, Interfacial Engineering for
Optimized Properties III (2004) San Francisco, CA, USA, Materials Research Society, 265-275
49. Brandon, D. G.: The structure of high angle grain boundaries, Acta Metallurgica (1966) 14, 11,
1479-1484
50. Reed, B. W., Minich, R. W., Rudd, R. E. and Kumar, M.: The structure of the cubic coincident
site lattice rotation group, Acta Crystallographica A: Foundations of Crystallography (2004) A60,
263-277
51. Chivers, T. C.: Nuclear tribology: a personal perspective, Tribology International (1986) 19, 5,
225-233
52. Smith, A. F.: The friction and sliding wear of unlubricated 316 stainless steel at room
temperature in air, Wear (1984) 96, 3, 301-318
53. Lim, S. C. and Ashby, M. F.: Wear mechanism maps, Acta Metallurgica (1987) 35, 1, 1-24
54. Ashby, M. F. and Lim, S. C.: Wear mechanism maps, Scripta Metallurgica et Materialia (1990)
24, 805-810
55. Lim, S. C., Ashby, M. F. and Brunton, J. H.: Wear rate transitions and their relationship to wear
mechanisms, Acta Metallurgica (1987) 35, 6, 1343-1348
56. Hsu, K.-L., Ahn, T. M. and Rigney, D. A.: Friction, wear and microstructure of unlubricated
austenitic stainless steels, Wear (1980) 60, 1, 13-27
57. Dumbleton, J. H. and Douthett, J. A.: The unlubricated adhesive wear resistance of metastable
austenitic stainless steel containing silicon, Wear (1977) 42, 2, 305-332
58. Allen, C. B., Quinn, T. F. J. and Sullivan, J. L.: The oxidational wear of high chromium ferritic
steel on austenitic stainless steel, Transactions of the ASME Journal of Tribology (1985) 107, 2,
172-179
59. Smith, A. F.: The influence of surface oxidation and sliding speed on the unlubricated wear of
316 stainless steel at low load, Wear (1985) 105, 2, 91-107
60. Smith, A. F.: The friction and sliding wear of unlubricated 316 stainless steel in air at room
temperature in the load range 0.5 - 90N, Wear (1986) 110, 2, 151-168
62. Archard, J. F. and Hirst, W.: The wear of metals under unlubricated conditions, Proceedings of
the Royal Society of London A: Mathematical and physical sciences (1956) 236, 1206, 397-410
63. Quinn, T. F. J.: Review of oxidational wear. Part I: The origins of oxidational wear, Tribology
International (1983) 16, 5, 257-271
64. Sarkar, A. D.: Friction and wear, 1st Ed. (1980) London, Academic press
65. Suh, N. P.: An overview of the delamination theory of wear, Wear (1977) 44, 1, 1-16
66. Zandrahimi, M., Reza Bateni, M., Poladi, A. and Szpunar, J. A.: The formation of martensite
during wear of AISI304 stainless steel, Wear (2007) 263, 1-6, 674-678
67. Wear resistant surfaces in engineering, 1st Ed. (1986) London, Great Britain Department of
Trade and Industry, H.M.S.O
68. Leyland, A. and Matthews, A.: On the significance of the H/E ratio in wear control: a
nanocomposite coating approach to optimised tribological behaviour, Wear (2000) 246, 1-2, 1-
11
69. Leyland, A. and Matthews, A.: Design criteria for wear-resistant nanostructured and glassy-
metal coatings, Surface and Coatings Technology (2004) 177-178, 317-324
70. Lancaster, J. K.: The formation of surface films at the transition between mild and severe
metallic wear, Proceedings of the Royal Society of London A: Mathematical and physical
sciences (1963) 273, 1355, 466-483
71. Sakrani, S. B. and Sullivan, J. L.: Iron oxide films in tribological surfaces of alloy steel, Third
International Conference on Thin Film Physics and Applications (1998) Shanghai, China, SPIE -
The International Society for Optical Engineering, 175-179
72. Godfrey, D.: Iron oxides and rust (hydrated iron oxides) in tribology, Lubrication Engineering
(1999) 55, 2, 33-38
73. Rabinowicz, E.: Friction fluctuations in 'Fundamentals of friction: macroscopic and microscopic
processes' 1st Ed. (1992) London, Kluwer Academic Publishers, 25-34
74. Stolarski, T. A.: Basic principles of tribology: friction in 'Tribology in machine design' 1st Ed.
(1999) Oxford, Butterworth Heinemann, 13-19
75. Barwell, F. T.: The contribution of particle analysis to the study of wear of metals, Wear (1982)
90, 1, 167-181
76. Smart, R. F.: Selection of surfacing treatments, Tribology International (1978) 11, 2, 97-104
77. Hurricks, P. L.: Some aspects of the metallurgy and wear resistance of surface coatings, Wear
(1972) 22, 3, 291-319
79. Roberts, W. H.: Some current trends in tribology in the UK and Europe, Tribology International
(1986) 19, 6, 295-311
80. Whelan, R. C., Howlett, S., Keown, S. R. and Vaughan, D. A. J.: The CEST surface engineering
initiative in 'Surface engineering and heat treatment: past, present and future' 1st Ed. (1991)
London, The Institute of Metals,
81. Bell, T.: Surface engineering: past, present and future, Surface Engineering (1990) 6, 1, 31-40
82. Jacobs, M. H.: Surface engineering of materials, Materials and Design (1993) 14, 1, 33-37
83. Strafford, K. N. and Subramanian, C.: Surface engineering: an enabling technology for
manufacturing industry, Journal of Materials Processing Technology (1995) 53, 393-403
84. Roberts, W. H.: Surface engineering and tribology in general engineering in 'Surface
engineering and heat treatment: past, present and future' 1st Ed. (1991) London, The Institute
of Metals,
85. Bhushan, B.: Contact mechanics of rough surfaces in tribology: multiple asperity contact,
Tribology Letters (1998) 4, 1, 1-35
86. Priestner, R. and Priestner, D. M.: Importance of substrate in surface engineering, Surface
Engineering (1991) 7, 1, 53-59
87. Mao, K., Sun, Y. and Bell, T.: Contact mechanics of engineering surfaces: state of the art,
Surface Engineering (1994) 10, 4, 297-306
88. Thelning, K. E.: Steel and its heat treatment, 2nd Ed. (1984) London, Butterworths
89. Syan, C. S., Matthews, A. and Swift, K. G.: Knowledge-based expert systems in surface coating
and treatment selection for wear reduction, Surface and Coatings Technology (1987) 33, 105-
115
90. Elliot, T. L.: Surface hardening, Tribology International (1978) 11, 2, 121-125
91. James, A. S., Thomas, K., Mann, P. and Wall, R.: The role and impacts of surface engineering
in environmental design, Materials and Design (2005) 26, 7, 594-601
92. British Standard EN 10052:1994, "Vocabulary of heat treatment terms for ferrous products",
BSI, London, www.bsi-global.com
93. Committee on gas carburizing, A. S. M.: Carburizing and carbonitriding, 1st Ed. (1977) Metals
Park, Ohio, American Society for Metals
94. Parrish, G. and Harper, G. S.: Production gas carburising, 1st Ed. (1985) Oxford, Pergamon
96. Sun, Y., Li, X. and Bell, T.: Low temperature plasma carburising of austenitic stainless steels for
improved wear and corrosion resistance, Surface Engineering (1999) 15, 1, 49-54
97. Yin, R.: Carburization of 310 stainless steel exposed at 800-1100C in 2% CH4/H2 gas mixture,
Corrosion Science (2005) 47, 1896-1910
98. Christ, H. J.: Experimental characterization and computer based description of the carburization
behaviour of the austenitic stainless steel AISI 304L, Materials and Corrosion (1998) 49, 4, 258-
265
99. Ueda, Y., Kanayama, N., Ichii, K., Oishi, T. and Miyake, H.: Effect of nitrogen on the plasma
(ion)-carburized layer of high nitrogen austenitic stainless steel, Surface and Coatings
Technology (2005) 200, 1-4, 521-524
100. Somers, M., Christiansen, T. and Ller Per, M.: Case-hardening of stainless steel European
Patent 1521861 (2004) EU
101. Parascandola, S., Moller, W. and Williamson, D. L.: Successful nitriding of austenitic stainless
steel: the diffusion mechanism of nitrogen and the role of the surface oxide layer, Stainless
Steel 2000 (2000) Osaka, Maney - IofM - IFHTSE - JSHT, 201-214
102. Ueda, Y., Kanayama, N., Ichii, K., Oishi, T. and Miyake, H.: Metallurgical characteristics of the
plasma (ion)-carburized layer of austenitic stainless steel SUS 316L, Surface and Coatings
Technology (2005) 193, 1-4, 50-54
103. Zhang, Z. L. and Bell, T.: Structure and corrosion resistance of plasma nitrided stainless steel,
Surface Engineering (1985) 1, 2, 131-136
104. Bell, T. and Sun, Y.: Process for the treatment of austenitic stainless steel articles UK Patent
9715180.7 (1998) UK
105. Bell, T.: Surface engineering of austenitic stainless steel, Surface Engineering (2002) 18, 6,
415-422
106. Li, X. Y.: Low temperature plasma nitriding of 316 stainless steel - Nature of S-phase and its
thermal stability, Surface Engineering (2001) 17, 2, 147-152
107. Sun, Y., Li, X. and Bell, T.: Structural characteristics of low temperature plasma carburised
austenitic stainless steel, Materials Science and Technology (1999) 15, 10, 1171-1178
108. Li, X. Y., Thaiwatthana, S., Dong, H. and Bell, T.: Thermal stability of carbon S phase in 316
stainless steel, Surface Engineering (2002) 18, 6, 448-452
109. Christiansen, T. and Somers, M.: Characterisation of low temperature surface hardened
stainless steel, Struers Journal of Materialography (2006) 9, 1-17
111. Fewell, M. P., Mitchell, D. R. G., Priest, J. M., Short, K. T. and Collins, G. A.: The nature of
expanded austenite, Surface and Coatings Technology (2000) 131, 1-3, 300-306
112. Christiansen, T. and Somers, M.: On the crystallographic structure of S-phase, Scripta
Materialia (2004) 50, 35-37
113. Marchev, K., Hidalgo, R., Landis, M., Vallerio, R., Cooper, C. V. and Giessen, B. C.: The
metastable m phase layer on ion-nitrided austenitic stainless steels Part 2: crystal structure and
observation of its two-directional orientation anisotropy, Surface and Coatings Technology
(1999) 112, 1-3, 67-70
114. Sun, Y., Li, X. Y. and Bell, T.: X-ray diffraction characterisation of low temperature plasma
nitrided austenitic stainless steels, Journal of Materials Science (1999) 34, 19, 4793-4802
115. Mingolo, N., Tschiptschin, A. P. and Pinedo, C. E.: On the formation of expanded austenite
during plasma nitriding of an AISI 316L austenitic stainless steel, Surface and Coatings
Technology (2006) 201, 7, 4215-4218
116. Li, X. Y.: Characterisation of low temperature plasma nitrided austenitic stainless steel, PhD
thesis (1999) Metallurgy and Materials - The University of Birmingham
117. Sun, Y. and Bell, T.: Low temperature plasma nitriding characteristics of precipitation hardening
stainless steel, Surface Engineering (2003) 19, 5, 331-336
118. Sun, Y.: Kinetics of low temperature plasma carburizing of austenitic stainless steels, Journal of
Materials Processing Technology (2005) 168, 189-194
119. Tsujikawa, M., Noguchi, S., Yamauchi, N., Ueda, N. and Sone, T.: Effect of molybdenum on
hardness of low-temperature plasma carburized austenitic stainless steel, Surface and Coatings
Technology (2007) 201, 9-11, 5102-5107
120. Li, X. Y. and Dong, H.: Effect of annealing on corrosion behaviour of nitrogen S phase in
austenitic stainless steel, Materials Science and Technology (2003) 19, 10, 1427-1434
121. Ernst, F., Cao, Y. and Michal, G. M.: Carbides in low-temperatue-carburized stainless steel,
Acta Materialia (2004) 52, 6, 1469-1477
122. Cao, Y., Ernst, F. and Michal, G. M.: Colossal carbon supersaturation in austenitic stainless
steels carburized at low temperature, Acta Materialia (2003) 51, 14, 4171-4181
123. Ernst, F., Cao, Y., Michal, G. M. and Heuer, A. H.: Carbide precipitation in austenitic stainless
steel carburized at low temperature, Acta Materialia (2007) 55, 6, 1895-1906
124. Tokaji, K., Kohyama, K. and Akita, M.: Fatigue behaviour and fracture mechanism of a 316
stainless steel hardened by carburizing, International journal of fatigue (2004) 26, 543-551
126. Ceschini, L. and Minak, G.: Fatigue behaviour of low temperature carburised AISI 316L
austenitic stainless steel, Surface and Coatings Technology (2007) 202, 9, 1778-1784
127. Akita, M. and Tokaji, K.: Effect of carburizing on notch fatigue behaviour in AISI 316 austenitic
stainless steel, Surface and Coatings Technology (2006) 200, 20-21, 6073-6078
128. Qu, J., Blau, P. J. and Jolly, B. C.: Tribological properties of stainless steel treated by colossal
carbon supersaturation, Wear (2007) 263, 1-6, 719-726
129. Sun, Y. and Bell, T.: Dry sliding wear resistance of low temperature plasma carburised
austenitic stainless steel, Wear (2002) 253, 5-6, 689-693
130. Sun, Y. and Bell, T.: Effect of layer thickness on the rolling-sliding wear behavior of low-
temperature plasma-carburized austenitic stainless steel, Tribology Letters (2002) 13, 1, 29-34
131. Martin, W. C. and Wiese, W. L.: Atomic, molecular, and optical physics handbook in 2.1st Ed.
(2002) Gaithersburg, NIST,
132. Thaiwatthana, S., Li, X. Y., Dong, H. and Bell, T.: Corrosion wear behaviour of low temperature
plasma alloyed 316 austenitic stainless steel, Surface Engineering (2003) 19, 3, 211-216
133. Aoki, K. and Kitano, K.: Surface hardening for austenitic stainless steels based on carbon solid
solution, Surface Engineering (2002) 18, 6, 462-464
134. Dong, H., Qi, P.-Y., Li, X. Y. and Llewellyn, R. J.: Improving the erosion-corrosion resistance of
AISI 316 austenitic stainless steel by low-temperature plasma surface alloying with N and C,
Materials Science and Engineering A: Structural Materials: Properties, Microstructure and
Processing (2006) 431, 137-145
135. Farrell, K., Specht, E. D., Pang, J., Walker, L. R., Rar, A. and Mayotte, J. R.: Characterization of
a carburized surface layer on an austenitic stainless steel, Journal of Nuclear Materials (2005)
343, 123-133
136. Tanaka, S., Ueda, K., Mitamura, N. and Oohori, M.: The development of an austenitic stainless
steel bearing with high corrosion resistance and high non-magnetic property, Journal of ASTM
International (2006) 3, 2, 1-7
137. Bos, M. J.: Case hardening of austenitic stainless steel pump components, World Pumps
(1998) March, 30-34
138. Ceschini, L., Lanzoni, E., Sambogna, G., Bordiga, V. and Schild, T.: Tribological behaviour and
corrosion resistance of Kolsterized AISI 316L austenitic stainless steel: existing applications in
the automotive industry, Journal of ASTM International (2006) 3, 2, 1-9
139. Williams, P. and Marx, S.: Low temperature case hardening processes US Patent 6,093,303
(2000) United States of America
141. Ueda, N., Yamauchi, N., Sone, T., Okamoto, A. and Tsujikawa, M.: DLC film coating on plasma-
carburized austenitic stainless steel, Surface and Coatings Technology (2007) 201, 9-11, 5487-
5492
142. van der Jagt, R. H., Kolster, B. H. and Gillham, M. W. H.: Anti-wear/corrosion treatment of
finished austenitic stainless steel components: the Hardcor process, Materials and Design
(1991) 12, 1, 41-46
143. Lebrun, J. P., Poirier, L., Hertz, D. and Lincot, C.: Environmentally friendly low temperature
plasma processing of stainless steel components for nuclear industry, Stainless Steel 2000
(2000) Osaka, Maney - IofM - IFHTSE - JSHT, 361-375
144. Grafen, W. and Edenhofer, B.: New developments in thermo-chemical diffusion processes,
Surface and Coatings Technology (2005) 200, 5-6, 1830-1836
145. Christiansen, T. and Somers, M.: Low temperature gaseous nitriding and carburising of
stainless steel, Surface Engineering (2005) 21, 5-6, 445-455
146. Somers, M. and Christiansen, T.: Kinetics of microstructure evolution during gaseous
thermochemical surface treatment, Phase Equilibria and Diffusion (2005) 26, 5, 520-528
147. Michal, G. M., Ernst, F. and Heuer, A. H.: Carbon paraequilibrium in austenitic stainless steel,
Metallurgical Transactions A: Physical Metallurgy and Materials Science (2006) 37A, 6, 1819-
1824
148. Aoki, K., Shirahata, T., Tahara, M. and Kitano, K.: Low temperature gas carburising for
austenitic stainless steels: the NV-Pionite process, Stainless Steel 2000 (2000) Osaka, Maney -
IofM - IFHTSE - JSHT, 389-405
149. Kula, P., Pietrasik, R. and Dybowski, K.: Vacuum carburizing - Process optimization, Journal of
Materials Processing Technology (2005) 164-165, 876-881
150. Ryzhov, N. M., Smirnov, A. E., Fakhurtdinov, R. S., Mulyakaev, L. M. and Gromov, V. I.: Special
features of vacuum carburizing of heat-resistant steel in acetylene, Metal Science and Heat
Treatment (2004) 46, 5-6, 230-235
151. Birch, B. J. and Ellis, B.: Twenty five years of vacuum heat treatment in 'Surface engineering
and heat treatment: past, present and future' 1st Ed. (1991) London, The Institute of Metals,
152. Hoffmann, F. and Clausen, B.: Low pressure carburizing with and without plasma assistance,
ASTRA (2004) Hyderabad, India, International Advanced Research Centre for Powder
Metallurgy and New Materials (ARCI),
153. Berghaus, B.: Method of ionitriding objects made of iron and steel GB Patent 1202810 (1968)
UK
155. Booth, M., Farrell, T. and Johnson, R. H.: Theory and practice of plasma carburising, Materials
and Design (1984) 5, 3, 139-148
156. Grube, W.: Direct-current glow discharge furnace for high-rate carburizing, Journal of Vacuum
Science and Technology (1979) 16, 2, 335-338
157. Reece Roth, J.: Industrial plasma engineering II, 1st Ed. (1997) London, IOP Publishing
158. Jacobs, M. H. and Law, T. J.: Plasma carburising: operating theory and practice with a large
industrial unit, Heat Treatment of Metals (1984) 3, 56-66
159. Wells, A. and Strydom, I. l. R.: Sputtering and redeposition of cathode material during plasma
nitriding, Surface Engineering (1986) 2, 4, 263-267
160. Ruset, C., Ciuca, S. and Grigore, E.: The influence of the sputtering process on the constitution
of the compound layers obtained by plasma nitriding, Surface and Coatings Technology (2003)
174-175, 1201-1205
161. Edenhofer, B.: Physical and metallurgical aspects of ionitriding - Part 2, Heat Treatment of
Metals (1974) 2, 59-67
162. Haruman, E., Bell, T. and Sun, Y.: Compound layer characteristics resulting from plasma
nitrocarburising in atmospheres containing carbon dioxide gas additions, Surface Engineering
(1992) 8, 4, 275-282
163. Edenhofer, B., Conybear, J. G. and Legge, G. T.: Opportunities and limitations of plasma
carburising, Heat Treatment of Metals (1991) 18, 1, 6-12
164. Maliska, A. M., Egert, P., de Souza, A. R., Speller, C. V. and Klein, A. N.: Influence of oxygen in
a plasma nitriding process, Journal of Materials Science (1997) 32, 23, 6375-6382
165. Avni, R. and Spalvins, T.: Nitriding mechanisms in Ar-N2, Ar-N2-H2 and Ar-NH3 mixtures in DC
glow discharges at low pressures (less than 10 torr), Materials Science and Engineering (1987)
95, 237-246
166. Ruset, C.: The influence of pressure on temperature uniformity in the plasma nitriding process,
Heat Treatment of Metals (1991) 18, 3, 81-84
167. Hick, A. J.: The industrial application of ionic carburising, Heat Treatment of Metals (1984) 3,
56-66
168. Rembges, W. and Oppel, W.: Process control of plasma nitriding and plasma nitrocarburizing in
industry, Surface and Coatings Technology (1993) 59, 1-3, 129-134
169. Janosi, S., Kolozsvary, Z. and Kis, A.: Controlled hollow cathode effect: New possibilities for
heating low-pressure furnaces, Metal Science and Heat Treatment (2004) 46, 7-8, 310-316
171. Okumiya, M., Tsunekawa, Y., Matsumoto, T., Tanaka, K. and Nagai, T.: Carbon content control
in plasma carburizing using process parameter and probe methods, Surface and Coatings
Technology (2003) 174-175, 1171-1174
172. Staines, A. M. and Bell, T.: Technological importance of plasma-induced nitrided and carburized
layers on steel, Thin Solid Films (1981) 86, 2-3, 201-211
173. Bell, T., Sun, Y. and Suhadi, A.: Environmental and technical aspects of plasma
nitrocarburising, Vacuum (2000) 59, 1, 14-23
174. Bell, T. and Dearnley, P. A.: Environmental issues in surface engineering and related industrial
sectors, Surface Engineering (1994) 10, 2, 123-128
175. Georges, J.: Nitriding process and nitriding furnace therefor US Patent 5,989,363 (1999) United
States of America
176. Alves Jr., C., de Araujo, F. O., Ribeiro, K. J. B., da Costa, J. A. P., Sousa, R. R. M. and de
Sousa, R. S.: Use of cathodic cage in plasma nitriding, Surface and Coatings Technology
(2006) 201, 6, 2450-2454
177. Georges, J.: Industrial thermochemical treatments and plasma nitriding (2002)
http://www.plasma-metal.lu, 13 July 2008
178. Li, C. X., Dong, H. and Bell, T.: A feasibility study of plasma nitriding of steel with an oxide layer
on the surface, Journal of Materials Science (2006) 41, 18, 6116-6118
179. Li, C. X., Georges, J. and Li, X. Y.: Active screen plasma nitriding of austenitic stainless steel,
Surface Engineering (2002) 18, 6, 453-458
180. de Sousa, R. R. M., de Araujo, F. O., Ribeiro, K. J. B., Mendes, M. W. D., da Costa, J. A. P. and
Alves Jr., C.: Cathodic cage nitriding of samples with different dimensions, Materials Science
and Engineering A: Structural Materials: Properties, Microstructure and Processing (2007) 465,
223-227
181. Zhao, C., Li, C. X., Dong, H. and Bell, T.: Study on the active screen plasma nitriding and its
nitriding mechanism, Surface and Coatings Technology (2006) 201, 6, 2320-2325
182. Li, C. X., Bell, T. and Dong, H.: A study of active screen plasma nitriding, Surface Engineering
(2002) 18, 3, 174-181
183. Zhao, C., Li, C. X., Dong, H. and Bell, T.: Study of the active screen plasma nitriding,
Transactions of Materials and Heat Treatment (2004) 25, 1, 330-333
184. Hubbard, P., Dowey, S. J., Doyle, E. and McCulloch, D.: Influence of bias and in situ cleaning
on through cage (TC) or active screen plasma nitrided (ASPN) steels, Surface Engineering
(2006) 22, 4, 243-247
186. D'Agostino, R., Favia, P., Oehr, C. and Wertheimer, M. R.: Low temperature plasma processing
of materials: past, present and future, Plasma Processes and Polymers (2005) 2005, 2, 7-15
187. Bogaerts, A., Neyts, E., Gijbels, R. and van der Mullen, J.: Gas discharge plasmas and their
applications, Spectrochimica Acta B: Atomic Spectroscopy (2002) 57, 4, 609-658
188. Sree Harsha, K. S.: Principles of vapor deposition of thin films, 1st Ed. (2006) London, Elsevier
189. Kuffel, E., Zaengl, W. S. and Kuffel, J.: Chapter 5: Electrical breakdown in gases in 'High
voltage engineering fundamentals' 2nd Ed. (2000) Oxford, Elsevier Newnes, 281-366
190. Druyvesteyn, M. J. and Penning, F. M.: The mechanism of electrical discharges in gases of low
pressure, Reviews of Modern Physics (1940) 12, April 1940, 87-174
191. Langmuir, I.: Electrical discharges in gases at low pressures, Journal of the Franklin Institute
(1932) 214, 275-298
192. Sree Harsha, K. S.: Chapter 4: Cold plasma discharges in 'Principles of vapour deposition of
thin films' 1st Ed. (2006) London, Elsevier, 259-365
193. Morse, P. M.: A theory of the electric discharge through gases, Physical review (1928) 31,
1003-1017
194. Braithwaite, N. S.: Introduction to gas discharges, Plasma Sources Science and Technology
(2000) 9, 4, 517-527
195. ITER: Measurement of plasma parameters, Nuclear Fusion (1999) 39, 12, 2541-2575
196. Demidov, V. I., Ratynskaia, S. V. and Rypdal, K.: Electric probes for plasmas: the link between
theory and instrument, Review of Scientific Instruments (2002) 73, 10, 3409-3439
197. Hancock, G.: Diagnostic of active species in plasmas, Surface and Coatings Technology (1995)
74-75, 10-14
198. Mott-Smith, H. M. and Langmuir, I.: The theory of collectors in gaseous discharges, Physical
review (1926) 28, 727-763
199. Found, C. G. and Langmuir, I.: Study of a neon discharge by the use of collectors, Physical
review (1932) 39, 237-253
200. Cooke, K. E., Goodyear, A., Hampshire, J. and Teer, D. G.: An investigation of an industrial
coating environment with planar probe technology, Surface and Coatings Technology (2004)
188-189, 750-755
202. Modest, M. F.: Chapter 10: Radiative properties of molecular gases in 'Radiative heat transfer'
2nd Ed. (2003) London, Academic Press, 288-360
203. Halliday, D. and Resnick, R.: Fisica: parte II (Physics: part II), 2nd Ed. (1977) Buenos Aires,
John Wiley and Sons
204. Svanberg, S.: Atomic and molecular spectroscopy: basic aspects and practical applications, 3rd
Ed. (2001) New York, Springer
205. Olesik, J. W.: ICP-OES in 'Encyclopedia of materials characterization' 1st Ed. (1992) London,
Butterworth-Heinemann, 633-644
206. Harrison, G. R., Lord, R. C. and Loofbourow, J. R.: Practical spectroscopy, 2nd Ed. (1948) New
York, Prentice Hall
208. Polk, J. E., Marrese-Reading, C. M., Thornber, B., Dang, L., Johnson, K. L. and Katz, I.:
Scanning optical pyrometer for measuring temperatures in hollow cathodes, Review of Scientific
Instruments (2007) 78, 9, 0931011-0931018
209. Korotchenko, V. and Matthews, A.: Substrate temperature monitoring in plasma assisted
processes, Vacuum (1986) 36, 1-3, 61-65
210. Langmuir, I.: The interaction of electron and positive ion space charges in cathode sheaths,
Physical review (1929) 33, 954-967
211. Langmuir, I.: The pressure effect and other phenomena in gaseous discharges, Journal of the
Franklin Institute (1923) 196, 751-762
212. Swift, J. D. and Schwar, M. J. R.: Electrical probes for plasma diagnostics, 1st Ed. (1970)
London, Iliffe Books Ltd
213. Loeb, L. B.: Fundamental processes of electrical discharge in gases, 1st Ed. (1939) London,
Chapman and Hall
214. Booth, J. P. and Braithwaite, N. S.: Method and device for measuring an ion flow in a plasma
US Patent 5,936,413 (1999) United States of America
215. Braithwaite, N. S., Booth, J. P. and Cunge, G.: A novel electrostatic probe method for ion flux
measurements, Plasma Sources Science and Technology (1996) 5, 4, 677-684
218. Boeuf, J. P.: A two dimensional model of DC glow discharges, Journal of Applied Physics
(1988) 63, 5, 1342-1349
219. Spyrou, N., Peyrous, R., Soulem, N. and Held, B.: Why Paschen's law does not apply in low
pressure gas discharges with inhomogeneous fields, Journal of Physics D: Applied Physics
(1995) 28, 4, 701-710
220. Raizer, Y. P.: Gas discharge physics, 1st Ed. (1997) London, Springer
221. Toader, E. I., Graham, W. G., Mahony, C. M. O. and Steen, P. G.: Characterization of a high
density direct current reflex discharge plasma source operating in Ar and N2, Review of
Scientific Instruments (2002) 73, 8, 2974-2980
222. Meek, J. M. and Craggs, J. D.: Electrical breakdown of gases, 1st Ed. (1953) Oxford University
Press
223. Metel, A.: Plasma immersion ion implantation based on glow discharge with electrostatic
confinement of electrons, Surface and Coatings Technology (2002) 156, 1-3, 38-43
224. Lisovskiy, V. A., Yakovin, S. D. and Yegorenkov, V. D.: Low pressure gas breakdown in uniform
DC electric field, Journal of Applied Physics (2000) 33, 21, 2722-2730
225. Tsujikawa, M., Egawa, M., Ueda, N., Okamoto, A., Sone, T. and Nakata, K.: Effect of
molybdenum and copper on S-phase layer thickness of low-temperature carburized austenitic
stainless steel, Surface and Coatings Technology (2008) 202, 22-23, 5488-5492
226. Williamson, D. L., Wilbur, P. J., Fickett, F. R. and Parascandola, S.: Role of ion-beam
processing time in the formation and growth of the high-nitrogen phase in austenitic stainless
steel, Stainless Steel 2000 (2000) Osaka, Maney - IofM - IFHTSE - JSHT, 333-352
227. Moller, W., Parascandola, S., Telbizova, T., Gunzel, R. and Richter, E.: Surface processes and
diffusion mechanisms of ion nitriding of stainless steel and aluminium, Surface and Coatings
Technology (2001) 136, 1-3, 73-79
228. Hillert, M. and Agren, J.: Reply to comments on "On the definition of paraequilibrium and
orthoequilibrium", Scripta Materialia (2005) 52, 87-88
229. Speer, J. G., Matlock, D. K., DeCooman, B. C. and Schroth, J. G.: Comments on "On the
definitions of paraequilibrium and orthoequilibrium" by M. Hillert and J. Agren, Scripta
Materialia, 50, 697-9 (2004), Scripta Materialia (2005) 52, 83-85
230. Hillert, M. and Agren, J.: On the definitions of paraequilibrium and orthoequilibrium, Scripta
Materialia (2004) 50, 697-699
231. Enomoto, M. and Aaronson, H. I.: Derivation of general conditions for paraequilibrium in multi-
component systems, Scripta Metallurgica (1985) 19, 1-3
233. Aziz, M. J.: Model for solute redistribution during rapid solidification, Journal of Applied Physics
(1982) 53, 2, 1158-1168
234. Wert, C. A. and Frank, R. C.: Trapping of interstitials in metals, Annual Reviews of Materials
Science (1983) 13, 139-172
235. Maalekian, M. and Kozeschnik, E.: A thermodynamic model for carbon trapping in lattice
defects, Computer Coupling of Phase Diagrams and Thermochemistry (2008) 32, 4, 650-654
236. Heuer, A. H., Ernst, F., Kahn, H., Avishai, A., Michal, G. M., Pitchure, D. J. and Ricker, R. E.:
Interstitial defects in 316L austenitic stainless steel containing "colossal" carbon concentrations:
An internal friction study, Scripta Materialia (2007) 56, 1067-1070
237. Zaefferer, S.: The electron backscatter diffraction technique - a powerful tool to study
microstructures by SEM, JEOL News (2004) 39, 1, 10-15
238. Yu, Z. W., Wang, L., Xu, X. L. and Gao, Y. Z.: In situ SEM and AFM observations of surface
morphology of plasma nitrided layer of austenitic stainless steel, Materials Science and
Technology (2005) 21, 8, 921-924
239. Borgioli, F., Fossati, A., Galvanetto, E., Bacci, T. and Pradelli, G.: Glow discharge nitriding of
AISI 316L austenitic stainless steel: influence of treatment pressure, Surface and Coatings
Technology (2006) 200, 18-19, 5505-5513
240. Sun, Y. and Haruman, E.: Influence of processing conditions on structural characteristics of
hybrid plasma surface alloyed austenitic stainless steel, Surface and Coatings Technology
(2008) 202, 17, 4069-4075
241. Xu, X., Yu, Z., Wang, L., Qiang, J. and Hei, Z.: Phase depth distribution characteristics of the
plasma nitrided layer on AISI 304 stainless steel, Surface and Coatings Technology (2003) 162,
2-3, 242-247
242. Monteiro, S. N.: Influence of grain orientation on the dislocation substructure in austenitic
stainless steel, Metallurgical Transactions A: Physical Metallurgy and Materials Science (1975)
6, 4, 938-940
243. Byun, T. S.: On the stress dependence of partial dislocation separation and deformation
microstructure in austenitic stainless steel, Acta Materialia (2003) 51, 11, 3063-3071
244. Lee, E. H., Byun, T. S., Hunn, J. D., Yoo, M. H., Farrell, K. and Mansur, L. K.: On the origin of
deformation microstructures in austenitic stainless steel - Part 1: Microstructures, Acta
Materialia (2001) 49, 16, 3269-3276
245. Cizek, P., Whiteman, J. A., Rainforth, W. M. and Beynon, J. H.: EBSD and TEM investigation of
the hot deformation substructure characteristics of a type of 316L austenitic stainless steel,
Journal of Microscopy (2004) 213, 285-295
247. Abrasonis, G., Riviere, J. P., Templier, C., Declemy, A., Pranevicius, L. and Milhet, X.: Ion
beam nitriding of single and polycrystalline austenitic stainless steel, Journal of Applied Physics
(2005) 97, 8, 8353101-8353110
248. Abrasonis, G., Riviere, J. P., Templier, C., Declemy, A., Muzard, S. and Pranevicius, L.: A
comparative study of ion beam nitriding of single-crystalline and polycrystalline 316L austenitic
stainless steel, Surface and Coatings Technology (2005) 196, 1-3, 262-266
249. Riviere, J. P., Templier, C., Declemy, A., Redjdal, O., Chumlyakov, Y. and Abrasonis, G.:
Microstructure of expanded austenite in ion-nitrided AISI 316 single crystals, Surface and
Coatings Technology (2007) 201, 19-20, 8210-8214
250. Mandl, S. and Rauschenbach, B.: Anisotropic strain in nitrided austenitic stainless steel, Journal
of Applied Physics (2000) 88, 6, 3323-3329
251. Marchev, K., Landis, M., Vallerio, R., Cooper, C. V. and Giessen, B. C.: The m-phase layer on
ion nitrided austenitic stainless steel (III): an epitaxial relationship between the m-phase and the
parent phase and a review of structural identifications of this phase, Surface and Coatings
Technology (1999) 116-119, 184-188
252. Ichii, K. and Fujimura, K.: Structure of the ion-nitrided layer of 18-8 stainless steel, Stainless
Steel 2000 (2000) Osaka, Maney - IofM - IFHTSE - JSHT, 13-22
253. Frechard, S., Martin, F., Clement, C. and Cousty, J.: AFM and EBSD combined studies of
plastic deformation in a duplex stainless steel, Materials Science and Engineering A: Structural
Materials: Properties, Microstructure and Processing (2006) A 418, 312-319
254. Sun, Y. and Chin, L. Y.: Residual stress evolution and relaxation in carbon S-phase layers on
AISI 316 austenitic stainless steel, Surface Engineering (2002) 18, 6, 443-447
255. Wagner, C. N. J.: X-ray study of low-temperature cold work in silver and aluminum, Acta
Metallurgica (1957) 5, 9, 477-482
256. Paterson, M. S.: X-ray diffraction by face-centered cubic crystals with deformation faults,
Journal of Applied Physics (1952) 23, 805-811
257. Mikkola, D. E. and Cohen, J. B.: Effects of thermal-mechanical treatments on faulting in some
fcc alloys, Journal of Applied Physics (1962) 33, 3, 892-898
258. Adler, R. P. I. and Wagner, C. N. J.: X-ray diffraction study of the effects of solutes on the
occurrence of stacking faults in silver-base alloys, Journal of Applied Physics (1962) 33, 12,
3451-3458
259. Wagner, C. N. J., Baisseau, J. P. and Aqua, E. N.: X-Ray diffraction study of plastically
deformed copper, Transactions of the Metallurgical Society of AIME (1965) 233, 7, 1280-1286
261. Sglavo, V. M., Bonafini, M. and Prezzi, A.: Procedure for residual stress profile determination by
curvature measurements, Mechanics of Materials (2005) 37, 8, 887-898
262. Roll, K.: Analysis of stress and strain distribution in thin films and substrates, Journal of Applied
Physics (1976) 47, 7, 3224-3229
263. Loh, N. L. and Siew, L. W.: Residual stress profiles of plasma nitrided steels, Surface
Engineering (1999) 15, 2, 137-142
264. Johnson, K. L.: The correlation of indentation experiments, Journal of Mechanics and Physics of
Solids (1969) 18, 2, 115-126
265. Richer, G. C.: Indentation hardness, Metallurgia (1945) 31, 186, 296-299
266. British Standard EN ISO 4516:2002, "Metallic and other inorganic coatings - Vickers and Knoop
microhardness tests", BSI, London, www.bsi-global.com
267. Buckle, H.: Use of the hardness test to determine other material properties in 'The science of
hardness testing and its research applications' Ed. (1971) Metals Park, Ohio, American Society
for Metals, 453-494
268. van Houtte, P. and de Buyser, L.: The influence of crystallographic texture on diffraction
measurements of residual stress, Acta Metallurgica et Materialia (1993) 41, 2, 323-336
269. Doerner, M. F. and Nix, W. D.: A method for interpreting the data from depth-sensing
indentation instruments, JOurnal of Materials Research (1986) 1, 4, 601-609
270. Suresh, S. and Giannakopoulos, A. E.: A new method for estimating residual stresses by
instrumented sharp indentation, Acta Materialia (1998) 46, 16, 5755-5767
271. Carlsson, S. and Larsson, P. L.: On the determination of residual stress and strain fields by
sharp indentation testing - Part I: theoretical and numerical analysis, Acta Materialia (2001) 49,
12, 2179-2191
272. Wang, Q., Ozaki, K., Ishikawa, H., Nakano, S. and Ogiso, H.: Indentation method to measure
the residual stress induced by ion implantation, Nuclear Instruments and Methods in Physics
Research B: Beam Interactions with Materials and Atoms (2006) 242, 1-2, 88-92
273. Stachowiak, G. P. and Podsiadlo, P.: Automated classification of wear particles based on their
surface texture and shape features, Tribology International (2008) 41, 1, 34-43
274. Hutchings, I. M.: Tribology: Friction and wear of engineering materials, 1st Ed. (1992) London,
Arnold
275. Rigney, D. A. and Glaeser, W. A.: The significance of near surface microstructure in the wear
process, Wear (1978) 46, 1, 241-250
277. Glaeser, W. A.: Friction and wear, IEEE Transactions on parts, hybrids and packaging (1971)
PHP 7, 2, 99-105
278. Suh, N. P.: The delamination theory of wear, Wear (1973) 25, 1, 111-124
279. Straffelini, G., Trabucco, D. and Molinari, A.: Sliding wear of austenitic and austenitic-ferritic
stainless steels, Metallurgical Transactions A: Physical Metallurgy and Materials Science (2002)
33A, 3, 613-624
280. Kapoor, A., Johnson, K. L. and Williams, J. A.: A model for the mild ratchetting wear of metals,
Wear (1996) 200, 1-2, 38-44
281. Johnson, K. L.: Contact mechanics and the wear of metals, Wear (1995) 190, 2, 162-170
282. Kang, G.: Ratchetting: recent progresses in phenomenon observation, constitutive modeling
and application, International journal of fatigue (2008) 30, 1448-1472
283. Podgornik, B. and Vizintin, J.: Influence of substrate treatment on the tribological properties of
DLC coatings, Diamond and Related Materials (2001) 10, 12, 2232-2237
284. Quinn, T. F. J.: Physical analysis for tribology, 1st Ed. (1991) Cambridge University Press
285. Shipway, P. H. and Hogg, J. J.: Dependence of microscale abrasion mechanisms of WC-Co
hardmetals on abrasive type, Wear (2005) 259, 1-6, 44-51
286. Fillot, N., Iordanoff, I. and Berthier, Y.: Wear modeling and the third body concept, Wear (2007)
262, 7-8, 949-957
287. Li, C. X. and Bell, T.: Sliding wear properties of active screen plasma nitrided 316 austenitic
stainless steel, Wear (2004) 256, 11-12, 1144-1152
288. Foerster, C. E., Serbena, F. C., da Silva, S. L. R., Lepienski, C. M., de M. Siqueira, C. J. and
Ueda, M.: Mechanical and tribological properties of AISI 304 stainless steel nitrided by glow
discharge compared to ion implantation and plasma immersion ion implantation, Nuclear
Instruments and Methods in Physics Research B: Beam Interactions with Materials and Atoms
(2007) B257, 1-2, 732-736
289. Manova, D., Hirsch, D., Richter, E., Mandl, S., Naumann, H. and Rauschenbach, B.:
Microstructure of nitrogen implanted stainless steel after wear experiment, Surface and
Coatings Technology (2007) 201, 19-20, 8329-8333
290. Sun, Y. and Kappaganthu, S. R.: Effect of nitrogen doping on sliding wear behaviour of
stainless steel coatings, Tribology Letters (2004) 17, 4, 845-850
291. Vergne, C., Boher, C., Gras, R. and Levaillant, C.: Influence of oxides on friction in hot rolling:
Experimental investigations and tribological modelling, Wear (2006) 260, 9-10, 957-975
293. Monypenny, J. H. G.: Stainless iron and steel, 3rd Ed. (1951) London, Chapman and Hall
294. Fauvet, P., Balbaud, F., Robin, R., Tran, Q. T., Mugnier, A. and Espinoux, D.: Corrosion
mechanisms of austenitic stainless steels in nitric media used in reprocessing plants, Journal of
Nuclear Materials (2008) 375, 52-64
295. Streicher, M. A.: General and intergranular corrosion of austenitic stainless steels in acids,
Journal of The Electrochemical Society (1959) 106, 3, 161-180
296. Castle, J. E. and Clayton, C. R.: The use of x-ray photo-electron spectroscopy in the analysis of
passive layers on stainless steel, Corrosion Science (1977) 17, 7-26
297. Shieu, F. S., Deng, M. J. and Lin, S. H.: Microstructure and corrosion resistance of a type 316L
stainless steel, Corrosion Science (1998) 40, 1267-1279
298. Thaiwatthana, S., Li, X. Y., Dong, H. and Bell, T.: Mechanical and chemical properties of low
temperature plasma surface alloyed 316 austenitic stainless steel, Surface Engineering (2002)
18, 2, 140-144
299. Li, C. X. and Bell, T.: Corrosion properties of active screen plasma nitrided 316 austenitic
stainless steel, Corrosion Science (2004) 46, 1527-1547
300. Spies, H.-J., Eckstein, C. and Zimdars, H.: Structure and corrosion behaviour of stainless steels
after plasma and gas nitriding, Surface Engineering (2002) 18, 6, 459-461
301. Nosei, L., Farina, S., Avalos, M., Nachez, L., Gomez, B. J. and Feugeas, J.: Corrosion
behaviour of ion nitrided AISI 316L stainless steel, Thin Solid Films (2008) 516, 6, 1044-1050
302. Sun, Y.: Enhancement in corrosion resistance of austenitic stainless steel by surface alloying
with nitrogen and carbon, Materials Letters (2005) 59, 3410-3413
303. Martin, F. J., Lemieux, E. J., Newbauer, T. M., Bayles, R. A., Natishan, P. M., Kahn, H., Michal,
G. M., Ernst, F. and Heuer, A. H.: Carburization-induced passivity of 316L austenitic stainless
steel, Electrochemical and Solid-State Letters (2007) 10, 12, C76-C78
304. Nikam, V. V., Reddy, R. G., Collins, S. r., Williams, P. C., Schiroky, G. H. and Henrich, G. W.:
Corrosion resistant low temperature carburized SS 316 as bipolar plate material for PEMFC
application, Electrochimica Acta (2008) 53, 2743-2750
305. Leitao, E., Silva, R. A. and Barbosa, M. A.: Electrochemical impedance spectroscopy of
nitrogen and carbon sputter coated 316L stainless steel, Corrosion Science (1997) 39, 333-338
306. Baba, H., Kodama, T. and Katada, Y.: Role of nitrogen on the corrosion behavior of austenitic
stainless steels Corrosion Science (2002) 44, 2393-2407
308. Flis, J. and Kuczynska, M.: Effect of low temperature plasma nitriding on corrosion of 304L
stainless steel in sulfate and chloride solutions, Journal of The Electrochemical Society (2004)
151, 11, B573-B580
309. Flis, J., Kuczynska-Wydorska, M. and Flis-Kabulska, I.: The effect of molybdenum on corrosion
of low temperature nitrided stainless steel in sulphate-chloride solution, Journal of Solid State
Electrochemistry (2006) 10, 689-695
310. Corujeira Gallo, S. and Dong, H.: Corrosion behaviour of direct current and active screen
plasma carburised AISI 316 stainless steel in boiling sulphuric acid solutions, Corrosion
Engineering Science and Technology (2009) Article in Press
311. Picard, S., Memet, J. B., Sabot, R., Grosseau-Poussard, J. L., Riviere, J. P. and Meilland, R.:
Corrosion behaviour, microhardness and surface characterisation of low energy, high current
ion implanted austenitic stainless steel, Materials Science and Engineering A: Structural
Materials: Properties, Microstructure and Processing (2001) A303, 163-172
312. Dearnley, P. A.: Corrosion wear response of S phase coated 316L, Surface Engineering (2002)
18, 6, 429-432
313. Dong, H., Sun, Y. and Bell, T.: Enhanced corrosion resistance of duplex coatings, Surface and
Coatings Technology (1997) 90, 1-2, 91-101
314. Murr, L. E., Advani, A., Shankar, S. and Atteridge, D. G.: Effects of deformation (strain) and
heat treatment on grain boundary sensitization and precipitation in austenitic stainless steels,
Materials Characterization (1990) 39, 2-5, 575-598
315. Bi, H. Y., Kokawa, H., Wang, Z. J., Shimada, M. and Sato, Y. S.: Suppression of chromium
depletion by grain boundary structural change during twin-induced grain boundary engineering
of 304 stainless steel, Scripta Materialia (2003) 49, 219-223
316. Wasnik, D. N., Kain, V., Samajdar, I., Verlinden, B. and De, P. K.: Resistance to sensitization
and intergranular corrosion through extreme randomization of grain boundaries, Acta Materialia
(2002) 50, 18, 4587-4601
317. van Aswegen, J. S. T., Honeycombe, R. W. K. and Warrington, D. H.: Precipitation on stacking
faults in Cr-Ni austenitic steels, Acta Metallurgica (1964) 12, 1, 1-13
318. Parvathavarthini, N., Dayal, R. K. and Gnanamoorthy, J. B.: Influence of prior deformation on
the sensitization of AISI type 316LN stainless steel, Journal of Nuclear Materials (1994) 208,
251-258
319. Wasnik, D. N., Dey, G. K., Kain, V. and Samajdar, I.: Precipitation stages in a 316L austenitic
stainless steel, Scripta Materialia (2003) 49, 135-141
321. Figueroa, C. A., Weber, S., Czerwiec, T. and Alvarez, F.: Oxygen, hydrogen, and deuterium
effects on plasma nitriding of metal alloys, Scripta Materialia (2006) 54, 1335-1338
322. Baldwin, M. J., Haydon, S. C. and Fewell, M. P.: Metastable states and nitriding plasmas,
Surface and Coatings Technology (1997) 97, 1-3, 97-101
323. Walkowicz, J.: On the mechanisms of diode plasma nitriding in N2-H2 mixtures under DC-
pulsed substrate biasing, Surface and Coatings Technology (2003) 174-175, 1211-1219
324. Rusnak, K. and Vicek, J.: Emission spectroscopy of the plasma in the cathode region of N2-H2
abnormal glow discharges for steel surface nitriding, Journal of Physics D: Applied Physics
(1993) 26, 4, 585-589
325. Ricard, A.: Production of active plasma species for surface treatments, Journal of Physics D:
Applied Physics (1997) 30, 16, 2261-2269
326. Ricard, A., Czerwiec, T., Belmonte, T., Bockel, S. and Michel, H.: Detection by emission
spectroscopy of active species in plasma-surface processes, Thin Solid Films (1999) 341, 1-2,
1-8
327. Walkowicz, J., Smolik, J. and Miernik, K.: Research on physico-chemical bases of the ion
nitriding process control with the use of plasma spectroscopic diagnostics, Vacuum (2000) 56,
1, 63-69
328. Loureiro, J. and Ricard, A.: Electron and vibrational kinetics in an N2-H2 glow discharge with
application to surface processes, Journal of Physics D: Applied Physics (1993) 26, 2, 163-176
329. Ferreira, C. M., Gordiets, B. F. and Tatarova, E.: Kinetic theory of low-temperature plasmas in
molecular gases, Plasma Physics and Controlled Fusion (2000) 42, 12B, B165-B188
330. Bastien, F., Wu, J. H., Goguillon, P. and Marode, E.: Mechanism of a nitrogen abnormal glow
discharge. Computation and measurements of the spatial light distribution, Journal of Physics
D: Applied Physics (1990) 23, 7, 813-822
331. Gordiets, B. F., Ferreira, C. M., Pinheiro, M. J. and Ricard, A.: Self-consistent kinetic model of
low-pressure N2-H2 flowing discharges: I. Volume processes, Plasma Sources Science and
Technology (1998) 7, 3, 363-378
332. Rie, K.-T., Menthe, E. and Wohle, J.: Optimization and control of a plasma carburizing process
by means of optical emission spectroscopy, Surface and Coatings Technology (1998) 98, 1-3,
1192-1198
333. Pastol, A. and Catherine, Y.: Optical emission spectroscopy for diagnostic and monitoring of
CH4 plasmas used for a-C.H deposition, Journal of Physics D: Applied Physics (1990) 23, 7,
799-805
335. Schulz-von der Gathen, V., Ropcke, J., Gans, T., Kaning, M., Lukas, C. and Dobele, H. F.:
Diagnostic studies of species concentrations in a capacitively coupled RF plasma containing
CH4-H2-Ar, Plasma Sources Science and Technology (2001) 10, 3, 530-539
336. Heintze, M., Magureanu, M. and Kettlitz, M.: Mechanism of C2 hydrocarbon formation from
methane in a pulsed microwave plasma, Journal of Applied Physics (2002) 92, 12, 7022-7031
337. Menningen, K. L., Childs, M. A., Toyoda, H., Ueda, Y., Anderson, L. W. and Lawler, J. E.: CH3
and CH densities in a diamond growth DC discharge, Contributions to Plasma Physics (1995)
35, 4-5, 359-373
338. Wolden, C. A. and Gleason, K. K.: On the pressure limits of diamond chemical vapor
deposition, Diamond and Related Materials (1996) 5, 12, 1503-1508
339. Teii, K., Ito, H., Hori, M., Takeo, T. and Goto, T.: Kinetics and role of C, O, and OH in low-
pressure nanocrystalline diamond growth, Journal of Applied Physics (2000) 87, 9, 4572-4579
340. Atkins, P. W.: Phisical chemistry, 3rd Ed. (1986) Oxford University Press
341. Escobar-Alarcon, L., Arrieta, A., Camps, E., Romero, S., Muhl, S. and Camacho-Lopez, M. A.:
a-CN thin film properties as a function of laser ablation plasma parameters, Diamond and
Related Materials (2007) 16, 4-7, 1291-1294
342. Walkowicz, J., Supiot, P., Smolik, J. and Grushin, M.: The influence of the N2-H2 mixture
composition on the spectroscopic and temporal behaviour of glow discharge characteristics in
pulse-supplied nitriding processes, Surface and Coatings Technology (2004) 180-181, 407-412
343. Cleugh, D.: Plasma species analysis for in situ assessment of surface treatments, Surface
Engineering (2002) 18, 2, 133-139
344. Kim, Y. M., U., K. J. and Han, J. G.: Investigation on the pulsed DC plasma nitriding with optical
emission spectroscopy, Surface and Coatings Technology (2002) 151-152, 227-232
345. Kim, Y.-K., Baek, J.-M. and Lee, K.-H.: Plasma diagnostics of pulsed d.c. glow discharge
combined with ICP for deep nitriding process, Surface and Coatings Technology (2001) 142-
144, 321-327
346. Kulakowska-Pawlak, B. and Zyrnicki, W.: Spectroscopic investigations into plasma used for
nitriding processes of steel and titanium, Thin Solid Films (1993) 230, 2, 115-120
347. Hugon, R., Henrion, G. and Fabry, M.: Diagnostics of a DC pulsed-plasma assisted nitriding
process, Surface and Coatings Technology (1993) 59, 1-3, 82-85
348. Behringer, K. and Fantz, U.: Spectroscopic diagnostics of glow discharge plasmas with non-
Maxwellian electron energy distributions, Journal of Physics D: Applied Physics (1994) 27, 10,
2128-2135
350. Herion, G., Hugon, R., Fabry, M. and Scherentz, V.: Reactivity of a DC-pulsed plasma: plasma
diagnostics and nitrided sample analysis, Surface and Coatings Technology (1997) 97, 1-3,
729-733
351. Petitjean, L. and Ricard, A.: Emission spectroscopy study of N2-H2 glow discharge for metal
surface nitriding, Journal of Applied Physics (1984) 17, 5, 919-929
352. Lombardi, G., Hassouni, K., Benedic, F. and Mohassed, F.: Spectroscopic diagnostics and
modeling of Ar/H2/CH4 microwave discharges used for nanocrystalline diamond deposition,
Journal of Applied Physics (2004) 96, 11, 6739-6751
354. Boffard, J. B., Lin, C. C. and DeJoseph Jr., C. A.: Application of excitation cross sections to
optical plasma diagnostics, Journal of Physics D: Applied Physics (2004) 37, 12, 143-161
355. Ljepojevic, N. N. and Labat, J. M.: Selfconsistent calculation of the hydrogen plasma
parameters. I. Stationary plasmas, Journal of Physics D: Applied Physics (1982) 15, 4, 621-629
356. Langmuir, I. and Jones, H. A.: Collisions between electrons and gas molecules, Physical review
(1928) 31, 357-404
357. Hugon, R., Fabry, M. and Henrion, G.: The influence of the respective durations of the
discharge and the afterglow on the reactivity of a DC pulsed plasma used for iron nitriding,
Journal of Applied Physics (1996) 29, 3, 761-768
358. Herzberg, G.: Molecular spectra and molecular structure: III. Electronic spectra and electronic
structure of polyatomic molecules, 1st Ed. (1966) London, van Nostrand Company
359. Leyland, A., Fancey, K. S., James, A. S. and Matthews, A.: Enhanced plasma nitriding at low
pressures: a comparative study of DC and RF techniques, Surface and Coatings Technology
(1990) 41, 3, 295-304
360. Leyland, A., Fancey, K. S. and Matthews, A.: Plasma nitriding in a low pressure triode
discharge to provide improvements in adhesion and load support for wear resistant coatings,
Surface Engineering (1991) 7, 3, 207-215
361. Michel, H., Czerwiec, T., Gantois, M., Ablitzer, D. and Ricard, A.: Progress in the analysis of the
mechanisms of ion nitriding, Surface and Coatings Technology (1995) 72, 1-2, 103-111
362. Andreu, J., Sardin, G., Esteve, J. and Morenza, J. L.: Electrostatic confinement effects on a hot
cathode DC glow discharge in silane, Journal of Physics D: Applied Physics (1987) 20, 11,
1479-1483
364. Rozsa, K., Janossy, M., Bergou, J. and Csillag, L.: Noble gas mixture CW hollow cathode laser
with internal anode system, Optics communications (1977) 23, 1, 15-18
365. Arslanbekov, R. R., Kudryavtsev, A. A. and Movtchan, I. A.: Cage discharge: theory and
experiment, IEEE Transactions on Plasma Science (1996) 24, 3, 1079-1094
366. Arslanbekov, R. R., Kudryavtsev, A. A. and Tobin, R. C.: On the hollow-cathode effect:
conventional and modified geometry, Plasma Sources Science and Technology (1998) 7, 3,
310-322
367. Major, F. G., Gheorghe, V. N. and Werth, G.: Charged particle traps: physics and techniques of
charged particle field confinement, 1st Ed. (2004) Berlin, Springer
368. Ingram, S. G. and Braithwaite, N. S. J.: Ion and electron energy analysis at a surface in an RF
discharge, Journal of Physics D: Applied Physics (1988) 21, 10, 1496-1503
369. Muratore, C., Walton, S. G., Leonhardt, D., Fernsler, R. F., Blackwell, D. D. and Meger, R. A.:
Effect of plasma flux composition on the nitriding rate of stainless steel, Journal of Vacuum
Science and Technology (2004) A22, 4, 1530-1535
370. Kuppusami, P., Sundararaman, D. and Raghunathan, V. S.: Comparative study of plasma
nitriding behaviour of a type 316 stainless steel and a microalloyed steel, Surface Engineering
(1993) 9, 2, 137-141
371. Wells, A. and Strydom, I. l. R.: Influence of sputtering and redeposition on compound layer
growth during plasma nitriding, Surface Engineering (1988) 4, 1, 55-59
372. Korhonen, A. S., Molarius, J. M. and Sulonen, M. S.: Plasma processing in nitrogen containing
atmospheres: 10 years' development, Surface Engineering (1988) 4, 1, 44-50
373. Hubbard, P., McCulloch, D. G., Doyle, E. D., Dowey, S. J. and Georges, J. N.: A fundamental
contribution to a study of the active screen plasma nitriding process, IFHTSE (2006) Vienna,
Austria,
374. de Sousa, R. R. M., de Araujo, F. O., Ribeiro, K. J. B., Dumelow, T., P., d. C. J. A. and Alves, C.
J.: Ionic nitriding in cathodic cage of AISI 420 martensitic stainless steel, Surface Engineering
(2008) 24, 1, 52-56
375. Ahangarani, S., Sabour, A. R. and Mahboubi, F.: Surface modification of 30CrNiMo8 low-alloy
steel by active screen setup and conventional plasma nitriding methods, Applied Surface
Science (2007) 254, 5, 1427-1435
376. Cleugh, D.: Effects of Rare Earth additions on plasma nitriding of EN40B steel, PhD thesis
(2003) Metallurgy and Materials - University of Birmingham
377. Doyle, E., Dowey, S., Hubbard, P. and McCulloch, D.: A study of key processing parameters in
the active screen plasma nitriding (ASPN) of steels, IFHTSE (2005) Riva-Del Garda, Italy,
379. Doyle, E., Wong, Y., Dowey, S., Cairney, J. and Munroe, P.: Nitriding response of high speed
steels using active screen plasma nitriding (ASPN), IFHTSE (2005) Riva-Del Garda, Italy,
380. Czerwiec, T., Michel, H. and Bergmann, E.: Low-pressure, high-density plasma nitriding:
mechanisms, technology and results, Surface and Coatings Technology (1998) 108-109, 182-
190
381. Hudis, M.: Study of ion-nitriding, Journal of Applied Physics (1973) 44, 4, 1489-1496
382. Tibbets, G. G.: Role of nitrogen atoms in "ion-nitriding", Journal of Applied Physics (1974) 45,
11, 5072-5073
383. Brokman, A. and Tuler, F. R.: A study of the mechanisms of ion nitriding by the application of a
magnetic field, Journal of Applied Physics (1981) 52, 1, 468-471
384. Li, J. L., O'Keefe, T. J. and James, W. J.: Iron carbide films formed by plasma deposition and
plasma carburizing, Materials Science and Engineering B: Solid-State Materials for Advanced
Technology (1990) B7, 15-23
385. Edenhofer, B., Grafen, W. and Muller-Ziller, J.: Plasma-carburising - A surface heat treatment
process for the new century, Surface and Coatings Technology (2001) 142-144, 225-234
386. Lampe, T., Eisenberg, S. and Laudien, G.: Compound layer formation during plasma nitriding
and plasma nitrocarburising, Surface Engineering (1993) 9, 1, 69-76
387. Li, S. and Manory, R. R.: Comparison of compound layer nucleation mechanisms in plasma
nitriding and nitrocarburizing: the effect of CHn species, Journal of Materials Science (1999)
34, 5, 1045-1049
388. Jauberteau, I., Cinelli, M. J., Cahoreau, M., Jauberteau, J. L. and Aubreton, J.: Expanding
microwave plasma for steel carburizing: role of the plasma impinging species on the steel
surface reactivity, Journal of Vacuum Science and Technology (2000) A18, 1, 108-114
Carburising • • • •
Treatment
N2 enhanced
• • • •
carburising
Nitriding • • • •
Notation: DC - Direct Current; FP - Floating Potential; ASsh - Active Screen mesh with small holes
(Ø= 8mm); ASbh – Active Screen mesh with large holes (Ø= 25.4mm)
Map
Crystal map size 512 pixels x 512 pixels
EBSD Camera
Acquisition mode Live
Sensitivity High
Integration time* 54 ms
Video gain* 8
Background
Average frames 10
Smooth filter size 11 x 11
Background correction Division
Solver set up
Minimum indexed bands 4
Maximum solution error 5
Depth of search for solution Optimum
Hough resolution High accuracy
Filter size 9 (Def)
Detect minimum number of bands 5
Hough radius 48
Min relative intensity 0.10
Reflective planes 37
Texture {001} <110>
* The values here reported are approximate as these parameters are sample
sensitive, and were altered in each case.
Table 3.6-3: Jobin Yvon Spectrum ONE detector and CCD 3000 controller unit specifications
Variable Value
Detector type CCD
Sensor array 1024 pixels x 256 pixels
Cooling method Thermoelectric (Peltier)
Operating temperature -30ºC
ADC limit 16 bit
Controller speed 166 kHz
Table 4.1-2: ASPC layer thickness (GDOES) after treatment under different conditions
ASPC Pressure [Pa]
T = 450ºC, t = 10 h
3%CH4 – 97%H2 30 80 125 250 400
0 0.5 μm 0.5 μm 0.5 μm
5 14.0 μm 16.5 μm 2.0 μm 0.5 μm
Bias level 10 14.0 μm 17.0 μm 9.0 μm
[%] 15 14.0 μm 16.0 μm 20.0 μm 13.0 μm 0.75 μm
20 14.0 μm 16.0 μm
25 14.0 μm 18.0 μm
Table 4.1-3: ASPN layer thickness (SEM) after treatment under different conditions
ASPN Pressure [Pa]
T = 400ºC, t = 10 h
25%N2 – 75%H2 30 75 100 125 150 250
0 6.2 μm
Bias level
5 6.5 μm
[%]
10 6.6 μm 5.8 μm 5.2 μm 5.4 μm 5.0 μm
Table 4.2-1: Instrumented indentation results from AISI 316 UT and AISI 316 PC samples
Maximum depth [nm] Plastic depth [nm] Hardness [GPa] Elastic Modulus [GPa]
Sample E/H
Average Std. dev. Average Std. dev. Average Std. dev. Average Std. dev.
AISI 316 2,156.97 41.79 2,093.27 42.68 1.95 0.08 206.22 6.64 105.75
ASPC 1,030.46 23.96 881.72 23.77 10.54 0.59 215.91 11.97 20.48
Table 4.3-1: Roughness measured on Cu samples after treatment under different conditions
Ra [um] DC AS
Nitriding 0.310 μm 0.139 μm
Carburising 0.232 μm 0.187 μm
N2 enhanced carburising 0.234 μm 0.175 μm
Theoretical Measured
Sample label Pressure Bias Roughness Hardness
worn volume* worn volume
AISI 316 N/A N/A 0.0801 μm 200.20 HMV0.025 3.1924 mm3 2.7130 mm3
DCPC 400 400 Pa N/A 0.0837 μm 837.37 HMV0.025 0.0075 mm3 0.0052 mm3
ASPC 125 125 Pa 15% 0.0772 μm 866.54 HMV0.025 0.0150 mm3 0.0059 mm3
ASPC 030 30 Pa 5% 0.0716 μm 857.05 HMV0.025 0.0120 mm3 0.0054 mm3
*Note: the theoretical worn volume was calculated, assuming that no significant wear of the pin took place,
using the following equation:
⎡ ⎤
( )
1
⎛ d ⎞ ⎛d ⎞
V = 2πR ⎢r 2 sin −1 ⎜ ⎟ − ⎜ ⎟ 4r 2 − d 2 2 ⎥
⎣ ⎝ 2r ⎠ ⎝ 4 ⎠ ⎦
AISI 316 PC
Element AISI 316
(oxide layer)
C [wt%] 7.4 1.5
O [wt%] 39.2 43.0
Cr [wt%] 9.7 10.0
Mn [wt%] 0.9 0.8
Fe [wt%] 35.8 34.0
Co [wt%] 0.4 0.2
Ni [wt%] 5.7 6.2
Mo [wt%] 0.7 0.6
W [wt%] 0.2 3.7
*Note: the corrosion factor was calculated according to the following equation:
ΔM treated
CF =
ΔM AISI 316
Table 5.1-1: Electric parameters in DC and AS furnaces under similar discharge conditions
Furnace Pressure Current Voltage Power
DC 400 Pa 5.0 A 550 V 2.75 kVA
AS 150 Pa 6.0 A 540 V 3.24 kVA
DC 125 Pa 2.0 A 800 V 1.60 kVA
AS 50 Pa 2.2 A 896 V 1.97 kVA
x = a(Dt )
1/ 2
= Kt 1 / 2
Table 5.2-3: Parameters used for estimation of residual stresses from the instrumented indentation test
Variable Symbol Value
Maximum applied load P1 200 mN
Load at intersection point B P2 40 mN
Experimental constant α 65.3º
Depth of residual indentation h 750 nm
Reference: (i) single bond, (ii) double bond, (iii) triple bond, (a) aromatic
Figure 2.1.1-1: Historical evolution of materials and their relative importance [1]
Figure 2.1.3-3: Stress induced martensitic transformation of ASS at different temperatures [16]
Figure 2.1.3-6: Ageing times for phase precipitation in AISI 316 at 550ºC and 600ºC [26]
Figure 2.1.4-3: Metallurgical features detrimental to the corrosion resistance of ASS [38]
Interacting asperities
Sliding direction
Normal load
Transferred material
Sliding direction
Figure 2.1.5-3: Transfer of material between interacting asperities and deformation of the
subsurface
Figure 2.2.1-1: Schematic of the stress and mechanical strength diagram for a cross section of
a surface engineered component between the neutral axis and the surface [88]
Figure 2.2.1-3: Energy content vs strength chart for different materials [1]
Figure 2.2.3-3: GDOES composition depth profile of AISI 316 plasma carburised and nitrided
Townsend regime
Breakdown voltage
VB Corona E
Glow to arc transition
Potential, U [V]
D H
High intensity
I arcs
C
Saturation G
regime F' F
B K
A Low intensity J
Normal glow Abnormal glow arcs
Background ionisation
DCPC ASPC
Vessel Vessel
Active Screen
Worktable Worktable
Bias
Earth power supply Earth
Main power
supply
Main power
supply
Figure 2.2.6-1: DC and AS furnaces for plasma carburising (DCPC and ASPC)
Figure 2.3.4-4: typical current voltage curve obtained with an electric probe [157]
Figure 3.2.1-1: (a) Klöckner Ionon DC furnace and (b) Plasma Metal active screen industrial unit
Main work table (-) Main work table (-) Main work table (-)
Figure 3.2.1-2: (a) DC, (b) AS and (c) FP arrangements inside a conventional DC plasma furnace
used to study the mechanisms of active screen plasma treatments.
DC
FP
DC
Figure 3.2.1-3: DC, ASsh, ASbh, and FP arrangements inside a conventional DC plasma furnace
used to study the mechanisms of active screen plasma treatments.
DC: direct current; AS: active screen; FP: floating potential; ASsh: AS mesh with small holes (Ø = 8 mm),
ASbh: AS mesh with large holes (Ø = 25.4 mm)
Pin holder
Vertical rotation axis
Applied load
Figure 3.4-1: Schematics of the pin-on-disc apparatus used for wear and friction tests
Coil condenser
Cooling water inlet
Clamp
Cylindrical flask
Figure 3.5-1: Schematics of the apparatus used for immersion corrosion in boiling acid solutions
DC arrangement AS arrangement
Probe Probe
DC DC DC
Figure 4.1-6: Typical profile of the thin discs used for the assessment of residual stresses
Figure 4.2-1: Optical micrographs of the same area before and after plasma carburising
Figure 4.2-2: EBSD Crystal Orientation Maps of the same area before and after plasma
carburising
DCPC ASPC
Oxide layer
Figure 4.2-7: TEM micrograph and diffraction pattern (amorphous) of the oxide layer on DCPC
and ASPC samples (top view)
Figure 4.2-8: TEM micrograph and ring electron diffraction pattern on a plasma carburised specimen
Figure 4.2-10: XRD pattern obtained on DCPC and ASPC, and austenite reference pattern
Figure 4.2-11: Lattice parameter ahkl vs cosθ.cotθ of expanded austenite and AISI 316 UT
Figure 4.2-12: GDOES composition-depth profile of DCPC and ASPC treated samples
Figure 4.2-14: Cracks inside a low-force (500g) Vickers indentation on a DCPC sample
Figure 4.3-1: Appearance of the samples treated under DC and AS arrangements in nitriding,
carburising and nitrogen enhanced carburising atmospheres
Figure 4.3-6: Vickers microhardness tests on different parts of the AISI304 stripe specimens
treated in an AS arrangement with big holes in the mesh
AS: Active screen, PN: plasma nitriding, PC: plasma carburising, PCN: nitrogen enhanced carburising, H: hole area
Figure 4.3-7: Optical micrographs of the deposition layer on ASPN, ASPC and ASPCN copper
samples
ASPN ASPN
ASPC ASPC
ASPCN ASPCN
Figure 4.3-8: Morphology of AS nitrided, carburised and N2 enhanced carburised samples observed under
SEM and AFM
Figure 4.3-11: XRD peaks of the deposition layer on ASPN treated glass slides
Figure 4.3-12: Micrograph and EDX spectrum of the mesh material used in the AS experimental
settings inside DC furnaces.
Figure 4.3-13: Micrograph and EDX spectrum of the mesh material used in the AS industrial unit
Figure 4.4-1: Typical appearance of the wear tracks on AISI 316 UT and AISI 316 PC
Figure 4.4-2: Appearance of the WC-Co ball worn against AISI 316 UT and AISI 316 PC
Figure 4.4-4: SEM images of the wear tracks on AISI 316 UT and AISI 316 PC
Cr Fe
Ni O W
Figure 4.4-5: EDX maps on the wear track left on AISI 316 PC worn against WC-Co ball
Figure 4.4-7: SEM images of the wear scars of WC-Co balls worn against AISI 316 UT and
AISI 316 PC
Figure 4.4-8: SEM images of the material accumulated at the trailing edge of the wear scar of
WC-Co balls worn against AISI 316 UT and AISI 316 PC
O W Co
Figure 4.4-9: EDX mapping of the wear scar left on the WC-Co ball worn AISI 316 UT
Fe O
Figure 4.4-10: EDX map of the wear scar left on the WC-Co ball worn against AISI 316 PC
Figure 4.4-11: Wear debris from untreated and carburised AISI 316 respectively
Figure 4.4-12: SEM BSE micrograph of wear debris from AISI 316 UT and AISI 316 PC
Figure 4.4-13: Electron diffraction patterns on wear debris from AISI 316 UT and AISI 316 PC
Figure 4.4-14: TEM micrograph of wear debris from AISI 316 PC and the EDX spectrum
Figure 4.4-16: Hardness profile measured on the cross section of one AISI 316 UT sample
worn against WC-Co
Figure 4.4-17: SEM cross section image of the wear track of AISI 316 UT worn against WC-Co
Figure 4.4-18: EBSD crystal orientation maps of the distorted layer, and martensite phase map on
the cross section of the wear track of one AISI 316 UT specimen
The arrow indicates the sliding direction
Figure 4.4-19: EBSD crystal orientation maps on cross sections of an AISI 316 PC specimen
The arrow indicates the sliding direction
Figure 4.5-1: Immersion corrosion test in H2SO4 16% solution at room temperature for 120 h
AISI 316 – Attack on a MnS inclusion; ASPC – Deposition layer of carbides pealing off
Figure 4.5-2: Surface appearance of AISI 316, DCPC 400 , ASPC 125 and ASPC 030
specimens corroded in boiling H2SO4 16% solution
Figure 4.5-3: Cross section of AISI 316 UT and PC specimens exposed to boiling H2SO4 16%
solution for 1 hour
Figure 4.5-5: Cross section of a PC specimen corroded in boiling H2SO4 16% solution for 1 h
Localised attack on slip bands
Figure 4.5-7: Cross section of a PC specimen corroded in boiling H2SO4 16% solution for 1 h
Corrosion progressing through grain boundaries of the AISI316 substrate once the EA layer
had been penetrated
AISI 316
DCPC400
ASPC125
ASPC030
Figure 4.5-8: Rod samples used for corrosion tests in boiling H2SO4 16% solution
Figure 4.5-9: Mass loss specimens after immersion corrosion test in boiling H2SO4 16%
solution for 20 hours
Figure 4.5-10: Coupon samples exposed to boiling H2SO4 16% solution with CuSO4 for 20 h
Figure 4.5-11: Optical micrographs of specimens corroded in boiling H2SO4 16% solution with
CuSO4 for 20 h
Figure 4.5-12: SEM micrograph of pits coalescing at the intersection of slip bands with a grain
boundary, and AFM image of the attack on slip bands and formation of ditch grain boundary
Figure 4.5-13: SEM micrograph of a ditch type grain boundary and EDX Carbon line scan on a
similar boundary
Figure 4.5-14: SEM micrographs of samples corroded in boiling HNO3 20% solution for 240 hours
Figure 4.5-16: Localised attack on samples exposed to boiling HNO3 20% solution for 240 h
Figure 4.5-17: Corrosion rates in boiling HNO3 20% solution for different periods of time
Figure 4.6-1: Low and high resolution spectra Figure 4.6-2: Low and high resolution spectra for Figure 4.6-3: Low and high resolution spectra for
for plasma carburising nitrogen enhanced plasma carburising plasma nitriding
Figure 4.6-5: Summary of the OES intensity maps obtained under ASPC and ASPN treatment
conditions
Figure 4.7-2: Floating potential and ion flux measured under DC and AS experimental
arrangements inside a DC furnace, in carburising and nitriding gas mixtures
Figure 4.7.3: Oscillations observed in the FP of the IFP inside the AS industrial unit
Figure 4.7-5: IFP drain curves obtained at floating potential in ASPC and ASPN conditions
Figure 4.7-6: Ion flux calculated from the drain curves of the IFP in ASPC and ASPN conditions
Figure 5.1-2: Bias set point and real bias power applied for ASPC and ASPN conditions
Figure 5.1-3: Bias current vs bias set point for ASPC and ASPN
Figure 5.1-5: Characteristic I-V curves for ASPC and ASPN with E/p as bubble size
Figure 5.2-2: EBSD pattern quality vs carbon concentration at plateau (GDOES carbon profile)
Figure 5.2-4: XRD pole figures obtained on AISI 316 UT and AISI 316 PC
Figure 5.2-6: AFM plane view of PC samples and crystal mimic obtained from EBSD COMs
AISI 316 UT
Figure 5.2-7: Misorientation measured by EBSD on the same area of the sample before and after
plasma carburising
Figure 5.2-8: EBSD crystal orientation (normal direction) before and after PC respectively
Figure 5.2-11: Hardness profiles calculated from LBC and Knoop hardness profile
Figure 5.2-12: Estimation of residual stresses from instrumented indentation tests according to
[238]
Figure 5.3.1-2: Signs of accumulated plastic strain (3 steps of deformation) at the edge of the
wear track of AISI 316 UT, and schematics of the deformation in the subsurface region.
Figure 5.3.2-2: Smooth surface of the underlying EA layer, and protective oxide layer formed
during sliding wear
3 4
Figure 5.3.2-6: Crack propagation along the original deformation morphology in the protective
oxide layer
Figure 5.3.2-7: Process of detachment of wear debris from the protective oxide layer. SEM
image and schematics
Element Weight %
C 19.1
O 7.7
S 0.3
Cr 17.0
Mn 1.2
Fe 48.0
Ni 1.4
Cu 3.1
Mo 2.2
Figure 5.4.2-1: SEM micrograph and chemical composition according to EDX of material pealed
off from a PC sample at early stages of a corrosion test in boiling sulphuric acid
Figure 5.4.2-2: XPS spectrum obtained from the surface of an ASPC sample
Figure 5.4.4-1: Inverse pole figures obtained by EBSD on the same area of the sample before
and after PC
Figure 5.4.4-2: Fraction number of special grain boundaries before and after PC
Figure 5.4.4-3: Edge effect revealed by microhardness measurements across the surface of a
DCPC sample
Figure 5.7.1-1: Morphology of the deposition layer under different ASPN treatment conditions
SEM image Fe
Cr Ni
Figure 5.7.1-4: Deposition layer on a glass slide treated in an AS arrangement with an AISI 304
mesh
Figure 5.7.1-6: Detachment of the deposition layer of iron nitrides AS plasma nitrided at floating
potential
Figure 5.7.2-2: SEM micrograph and XRD pattern from carbide film deposited on the walls of
the AS furnace
Aomolin Ltd UV-Vis silica optical fibre of core diameter 100/110 μm (NA 0.22) was selected for the
construction of the probe. The fibre was cut into 5 m long strings, and four strings were threaded
together through the stainless steel shell. A bundle of four optical fibres was necessary in order to
increase the intensity of the signal, which was otherwise too weak, especially when the high resolution
grating was used, i.e. with larger spectral dispersion. This method was good enough for the purpose of
the current work, although alignment was problematic, and probes made with single fibres of larger core
diameter could be useful for future applications.
The shell of the feedthrough was made of a stainless steel tube, with a quick fit type flange on one end,
which allowed an easy attachment to any of the standard instrumentation ports of the furnace (figure
A.1). An insulating and temperature resistant ceramic adhesive, Aremco Ceramcast 586 (ZrSiO4), was
used to fill the gap between the optical fibres and the stainless steel tube, and the feedthrough was
cured in an oven following the thermal cycle suggested by the supplier (table A-1). Finally, the outer side
of the feed through was sealed with an epoxy resin to reduce the leaks and not to affect the quality of
the vacuum in the furnace.
The optical fibre feedthrough was then installed in the active screen furnace and the fibres were placed
on top of ceramic insulators. It is worth noting that if the optical fibres are subjected to the cathodic
potential of the work table during the treatment, severe arcing takes place, with the associated instability
of the process and risk of permanent damage or destruction of the optical fibres.
The polymeric buffer coating of the fibres did not resist the operating conditions, but a layer of metallic
material was deposited on the fibres after only a few hours of exposure to the plasma. This metallic
An insulating ceramic tube was used to aim the bundle of fibres in the direction of interest and to reduce
deposition of sputtered material on their open ends. Nonetheless, deposition still blocked the fibres after
several hours of exposure to the plasma and, as a result, it was necessary to cleave the fibres before
each experiment, and to shorten the data collection time as much as possible in order to get intense and
comparable signals.
The other ends of the optical fibres were stripped from the buffer polymer coating, cleaved and fitted
with an SMA connector, to feed the signal into the Triax 180 spectrometer.
For the assessment of the transmission efficiency before and after exposure to the plasma environment,
their respective transmission efficiencies, in the range 200 nm to 1100 nm, were recorded and
compared. The exposed and unexposed optical fibres were fitted with Bullet bare fibre connectors and
plugged into an Ocean Optics HL-2000-FHSA halogen light source (table A-2) and an Ocean Optics
HR2000 high resolution miniature fibre optics spectrometer (table A-3).
In addition, the deposition layer was found to coat the open end of the optical fibres after only a few
hours of exposure to the plasma, and this coating hindered light from being collected. For this reason,
[mm]
Figure A-1: Dimensions of the stainless steel shell for optical fibre feedthroughs
Figure A-2: Deposition layer observed on an optical fibre exposed to the plasma in an AS
furnace
Table A-3: Ocean Optics HR2000 high resolution fibre optic spectrometer - Technical specifications
Variable Value
Detector Sony ILX511 linear CCD array
Detector Range 200 nm -1100 nm
Pixel size 14 µm x 200 µm
Optical bench f/4, 101 mm focal length
Gratings Choose from 13 different gratings
Pixel elements 2048
Entrance Aperture 5, 10, 25, 50, 100 or 200 µm wide slits or fibre (no slit)
Order-sorting filters Installed longpass and bandpass filters
Stray Light <0.05% at 600 nm; <0.10% at 435 nm
Dynamic Range 2 x 108 (system); 2000:1 for a single scan
Fibre Optic Connector SMA 905 to single-strand optical fibre 0.22 NA)
Data Transfer Rate Full scans into memory every 13 ms
Integration Time 3 ms to 65 s
A planar electric probe was made of copper, with a central disc as the actual probe and two guarding
rings surrounding it (figure B-1). The guarding rings were meant to avoid edging effects on the central
disc, making the sheath in front of it flat and even. The surface area of the central disc and the inner
guarding ring are the same, while the exposed area of the outer guarding ring is twice as large (figure B-
2). They were finely turned and the active surfaces were wet ground with silicon carbide grinding paper,
and polished with activated colloidal silica.
The isolation between the central disc and the guarding rings was achieved using DuPont Nomex
Aramid paper type 410 (table B-1). This type of temperature resistant insulating paper was found
convenient in view of its thermal resistance, dielectric and mechanical properties and low cost. From two
to four layers of 0.25 mm thick insulating paper were used in all cases and new insulating sheets and
washers were used for each experiment. The paper was found to resist the treatment conditions here
studied, but its use is not recommended for longer periods of time or higher temperatures.
The central disc was silver soldered to a copper wire and connected with the measuring circuit, through
a standard electric feedthrough, by means of gold coated pin connectors (figure B-3). The guarding
rings were not connected to the measuring circuit due to a technical inconvenient, and remained at
floating potential during the experiments.
The probe was mounted on ceramic insulating tubes, to support it and insulate it from the high cathodic
potential of the work table (up to 900 V DC). Two ceramic tubes were attached together by means of
steel bolts, which also played the role of spacers for copper wires to be connected to the probe (figure
B-1). To avoid deposition of conductive material on the live parts of the measuring circuit, the copper
wire was insulated with ceramic beads throughout its length, and it was also guided inside an insulating
Regarding the measuring circuit, tantalum capacitors were selected in view of their low electric charge
leak rate in air, allowing retention of charge for longer times. The 63 V and 4.7 μF tantalum capacitors
were connected to a 60 V DC power supply with a 10 Ω resistor in series. Heavy duty toggle switches
were used for the connection and disconnection of the probe and the power supply. Finally, high speed
0.8 A fuses were used to protect the measuring circuit in case of electric fault, and the operating panel
was conveniently earthed.
A low noise Instek GPS-6010 DC power supply with variable output voltage (0 V to 60 V) was
connected to charge the capacitors and, subsequently, to bias the probe (table B-2). A Pico 2202 digital
oscilloscope and 250 MHz oscilloscope probe, with an impedance of 50 MΩ and an attenuation power
of 10, were used to acquire the drain curve of the tantalum capacitors (table B-3). The oscilloscope was
connected to the USB port of a PC and the signal was monitored and stored in the hard drive of a PC
using PicoScope 5.15.6 (32 bit) data collection software.
Probe
Insulating paper
Inner guarding ring
Work table
Detail A
Ø35.8
Ø25.6 Ø40
Ø25
Ø36.4
4.3
Ø5.6
2
4
27
19
2.3
6.1
12.1
Ø20
Ø5 Ø19.4 Ø30
[mm]
Figure B-2: IFP - Dimensions of the central disc, inner and outer guarding rings
A B C D E F G
0.93 1.57 1.57 0.53 0.43 0.032 0.25 [in]
23.6 39.9 39.9 13.5 11.0 0.81 6.3 [mm]
Figure B-3: Lesker EFT0043038A feedthrough – Stainless steel KF25 flange, 4 copper pins
Table B-1: DuPont Nomex Aramid Insulation paper type 410 – Technical specifications
Variable Value
Nominal thickness 0.25 mm
Dielectric strength (AC rapid raise) 32 kV/mm
Dielectric Strength (full wave impulse) 63 kV/mm
Dielectric constant at 60 Hz 2.7
Dissipation factor at 60 Hz 6
1 E S d S2 1
σ0 = − (C.1)
6 1 − ν S δ EA R
where σ0 is the mean value of residual stress through the thickness of the layer, ES is the elastic
modulus of the substrate, νS is Poisson’s ratio of the substrate, dS and δEA are the thickness of the
substrate and the layer of expanded austenite respectively, and R is the radius of curvature of the disc
measured after the treatment (figure C-1).
1 E S d S2 1 ⎡ ⎛ E EA (1 − ν S ) ⎞ δ EA ⎤
σ0 = − ⎢1 + ⎜ 4 − 1⎟ ⎥ (C.2)
6 (1 − ν S ) δ EA R ⎣ ⎜⎝ E S (1 − ν EA ) ⎟⎠ d S ⎦
where the first part of this formula is equivalent to the Stoney’s formula described above, and the term
between brackets is the proposed correction. Subscripts S and EA account for substrate (AISI 316) and
expanded austenite respectively. The elastic modulus of EA was obtained from the instrumented
hardness tests and the Poisson’s ratio was considered equal to the untreated substrate.
1 E S d S2 ⎛ δ EA ⎞
σ0 = − ⎜1 − ⎟⎟ (C.3)
6 R δ EA ⎜⎝ dS ⎠
where the symbols have the same physical meaning as explained for the previous formulas.
where ahkl is the lattice parameter measured from the XRD data, a0 is the lattice parameter of the
annealed specimen, Ghkl is the fault parameter (table C-1), α is the stacking fault probability, S1 is the
elastic constant (table C-1), σ is the residual stress, m is the slope of the line in figure 4.2-11 and the
value of cosθ·cotθ corresponding to ahkl, which is the extrapolation function used in the same graph.
The second order reflections allow distinguishing both effects, namely the stacking fault and the residual
stress, by making use of the following equation:
where the superscript app refers to the apparent lattice parameter from the extrapolation of the trend
line in figure 4.2-11 to cosθ·cotθ value of zero. For reflections (111) and (222), the second and third
terms on the right side of the equation are equal so that this relation can be derived:
Δa111 − Δa 222
= −0.0518 ⋅ α (C.6)
a0
The value of α can be obtained from equation C.6 and inserted in equation C.4, which allows calculating
the residual stress (table C-2)
σ x = σ y = σ (z ) (C.8)
d
⎛ d⎞
m x , m y = ∫ σ ( z ) ⋅ ⎜ z − ⎟dz ≠ 0 (C.10)
0 ⎝ 2⎠
where fx, fy and mx, my are the forces and moments per unit width respectively and d is the thickness of
the disc.
On condition that the displacements associated with the bending of the discs are small compared to
their thickness, and that the cross sections of the discs remain plane and rotate only with respect to the
neutral axes, the theory of pure bending of plates applies, and the strains ε can be defined as:
σx σy
εx = −ν (C.11)
E E
σy σx
εy = −ν (C.12)
E E
where E is the Young’s modulus and ν is Poisson’s ratio. The bending moments per unit width can thus
be defined as:
Ed 3 ⎛1 1⎞
mx = ⋅ ⎜ −ν ⎟ (C.13)
(
12 1 − ν 2 ) ⎜r
⎝ x r y ⎟⎠
Ed 3 ⎛1 1⎞
my = ⋅ ⎜ −ν ⎟ (C.14)
(
12 1 − ν 2 ) ⎜r
⎝ y r x ⎟⎠
And in this particular case, the disc is bent into a sphere sector, so that both radii of curvature are equal:
1 1
= = −ϕ ⇒ m x = m y (C.15)
rx ry
If we now consider the mean compressive residual stresses as the external load responsible for the
bending of the disc:
where σavg is the average residual stress, δ is the EA layer thickness, and replacing the values listed in
table C-3, σavg is found to be -0.770 GPa (sample ASPC P125B15).
If the stress distribution is approximated by a triangular distribution, and reducing the moments to the
layer core interface it is obtained:
⎛ δ ⎞ ⎛2 ⎞ ⎛δ ⎞
⎜ σ surf ⎟ ⋅ ⎜ δ ⎟ = (σ avg δ ) ⋅ ⎜ ⎟ (C.17)
⎝ 2⎠ ⎝3 ⎠ ⎝2⎠
3
σ surf = σ avg (C.18)
2
Using the previous calculated values, equation C.18 yields σsurf = -1.155 GPa. A similar procedure is
followed to calculate the residual stress profiles from GDOES carbon distribution and LBC strain
hardening, yielding maximum residual stresses of -1.22 GPa and -1.61 GPa respectively.
Figure C-1: Schematics of the thin discs used for the assessment of residual stresses (cross
section): δ - expanded austenite layer thickness; d - disc thickness
Table C-2: Data used for the calculation of the residual stress on sample ASPC P125B15
Table C-3: Parameters used for the calculation of residual stresses from the bending of thin discs
Parameter Symbol Value
Young’s modulus E 205 GPa
Poisson’s ratio ν 0.305
Disc thickness d 530 μm
Layer thickness δ 20 μm
Radius of curvature r 0.8962 m
1. Corujeira Gallo, S., Machavaram, V. R., Fernando, G. F. and Dong, H.: Probing active screen
plasma using high-temperature optical fibre assisted OES method, PSE 2008 (2008) Garmisch-
Partenkirchen, Germany,
3. Mezin, A.: Coating internal stress measurement through the curvature method: A geometry-
based criterion delimiting the relevance of Stoney's formula, Surface and Coatings Technology
(2006) 200, 18-19, 5259-5267
4. Wang, Q., Ishikawa, H., Nakano, S., Ogiso, H. and Akedo, J.: Curvature-based residual stress
measurement for ion-implanted stainless-steel sheet, Vacuum (2004) 75, 3, 225-229
5. Wagner, C. N. J., Baisseau, J. P. and Aqua, E. N.: X-Ray diffraction study of plastically
deformed copper, Transactions of the Metallurgical Society of AIME (1965) 233, 7, 1280-1286
6. Timoshenko and Woinowsky-Krieger: Theory of plates and shells, 2nd Ed. (1959) London,
McGraw Hill
7. Sun, Y., Li, X. Y. and Bell, T.: X-ray diffraction characterisation of low temperature plasma
nitrided austenitic stainless steels, Journal of Materials Science (1999) 34, 19, 4793-4802