Anda di halaman 1dari 5

Surface & Coatings Technology 186 (2004) 277 281

www.elsevier.com/locate/surfcoat

Comparison of expanded austenite and expanded martensite


formed after nitrogen PIII
S. Mandl *, B. Rauschenbach
Leibniz-Institut fur Oberflachenmodifizierung, Permoserstr. 15, 04303 Leipzig, Germany
Available online 25 May 2004

Abstract
The formation of expanded austenite is a well-established phenomenon after energetic nitrogen ion implantation into austenitic steel or
nickel alloys, correlated with a very high surface hardness and wear resistance. Using high-voltage PIII implantation into martensitic stainless
steels, it is also possible to form expanded martensite. In this work, results are compared for martensitic and austenitic stainless steel. Both
groups of materials show a lattice expansion of the base material within the nitrogen diffusion zone of several micrometer at 400 jC.
Nevertheless, the expanded martensite exhibits a significantly higher hardness, reaching up to 2000 HV, compared to 1000 1200 HV for
expanded austenite. In both groups, a wear reduction of several orders of magnitude is observed. An attempt is made to explain the results
within the framework of ion-induced effects.
D 2004 Elsevier B.V. All rights reserved.
Keywords: Expanded austenite; PIII; Expanded martensite; Wear

1. Introduction
Nitrogen implantation into austenitic stainless steel at
elevated temperatures around 350 400 jC and incident
doses of more than 1018 at/cm2 is a well-investigated
process for increasing the hardness and reducing the wear
of these steels while retaining the corrosion resistance [1
3]. Still unresolved is the concentration-dependent diffusion
coefficient, strongly increasing at higher nitrogen concentrations [4] and the anisotropic lattice expansion [5] as well
as a marked influence of the ion energy on the obtainable
nitrogen concentration.
Recently, first reports on a lattice expansion in martensitic
stainless steels were published, showing a similar effect of
lattice expansion and wear reduction [6 8]. However, no
systematic investigation detailing effect of steel grade or an
explanation on the formation mechanism of expanded martensite was published. The main metallurgical difference
between austenites and martensites is the reduced Ni content
in the latter group, thus destabilizing the austenite phase
during cooling [9]. However, addition of nitrogen is supposed to stabilize this phase again [10].
* Corresponding author. Tel.: +49-341-235-2944; fax: +49-341-2352313.
E-mail address: stephan.maendl@iom-leipzig.de (S. Mandl).
0257-8972/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.surfcoat.2004.02.049

The purpose of this work is a rigorous comparison of


PIII nitriding of austenitic and martensitic stainless steels,
looking for similarities and differences in the phase composition, depth profiles, hardness and wear behavior. The
results are used to gain an insight on possible process
opportunities and restrictions. Additionally, further insight
on the actual process from the new results may allow to
exclude certain formation mechanism of these expanded
phases.

2. Experiment
Several austenitic and martensitic stainless steel grades
were used in this experiment: AISI 316Ti (X6CrNiMOTi17.12.2), 321 (X6CrNiTi18.10), 420 (X20Cr13),
440B (X90CrMoV18) and 630 (X5CrNiCuNb17.4). The
first two steels are austenites, the others are martensites.
PIII treatments for all of them were performed at 400 jC
for different nitrogen doses between 2.5 and 9  1018
cm 2. The respective treatment times were between 0.5
and 3 h at voltages ranging from 25 to 40 kV. Additionally,
martensitic grades were treated at 600 jC using similar
parameters, for information on the temperature effect.
Temperature measurements were performed with a pyrometer calibrated against a thermocouple [11]. The pulse

278

S. Mandl, B. Rauschenbach / Surface & Coatings Technology 186 (2004) 277281

width was fixed at 15 As with the repetition rate adjusted to


keep the temperature constant. The voltage-dependent
doses per pulse were obtained from low-dose implantations
into Si, with a medium value of 3.7  1011 nitrogen atoms
per cm2 per pulse at 30 kV [12].
The phase composition was analysed using X-ray diffraction in Bragg Brentano geometry. Selected glancing
angle measurements were performed to obtain information
on possible sub-layers within the structure [13]. Elemental
depth profiles were obtained from glow discharge optical
spectroscopy (GDOS) with an accuracy of better than 0.1
at.% and a reproducibility of 0.1 0.2 at.% [14].
Microhardness measurements were performed with a
Vickers indenter and a load of up to 50 g to obtain depthdepending hardness curves. Dry wear tests were carried out
at room temperature in laboratory air at a humidity of 60
65% using an oscillating ball-on-disk type tribometer [15].
The 3-mm ball, consisting of tungsten carbide, was pressed
to the samples with an applied load of 3 N and moved with
an average velocity of 1.5 cm/s.

3. Results
Fig. 1 shows XRD data for selected samples. 316Ti
exhibits (Fig. 1a) the clear signature of an expanded

austenite surface layer with the peaks resulting from the


expanded lattice dominating and only very small hints of the
substrate material. No peak corresponding to an expanded
(220) plane is present. Similarly, 630 and 420 show an
expanded martensite (Fig. 1b), most pronounced for the
(110) reflection, but also in clear traces for the high angle
(211) and (220) reflections. No corresponding peak was
found for the (200) substrate peak at 65.0j. Both steel
grades show different intensities and peak shifts, indicating
a lattice expansion and layer thickness depending on the
alloy composition. For both grades, thinner layers than for
the 316Ti in Fig. 1a are obtained.
In contrast, treatment at 600 jC induces a phase segregation into CrN and ferrite, as can be seen in Fig. 1c. The
CrN is present as small precipitates while reflections from
the martensitic base material and the ferrite are indistinguishable at the employed resolution. Please note the
completely different relative intensities in Fig. 1b and c
for the peak at 44.7j and 43.7j.
The respective nitrogen depths profiles are shown in
Figs. 2 and 3. For 316Ti at an incident dose of 4  1018
cm 2, a deep layer of 10 12 Am is obtained from GDOS
measurements (Fig. 2a) with the nitrogen concentration
decreasing in two steps, due to the concentration-dependent
diffusion coefficient in austenitic stainless steel, which
increases by about a factor of 10 up to 10 10 cm2/s beyond

Fig. 1. XRD spectra for (a) 316Ti implanted at 400 jC, (b) 630 and 420 at 400 jC, (c) 630 at 600 jC. The respective incident doses and peak positions
according to powder diffraction files are indicated.

S. Mandl, B. Rauschenbach / Surface & Coatings Technology 186 (2004) 277281

279

Fig. 2. Fe, Cr and N elemental depth profiles for 316Ti (a) and 630 (b) after PIII at 400 jC.

a nitrogen concentration of 5 10 at.% [4]. The relative iron


content increases in conjunction with the nitrogen decrease,
thus keeping the Cr content almost constant. A very similar
shape is observed for the martensitic 630 steel at 400 jC
(Fig. 2b) with a total layer thickness of about 50% from the
austenite 316Ti, albeit for a different treatment time and
incident dose. However, no absolute determination of diffusion coefficients is intended in this presentation. Only a
slight variation in the Cr content is observed, thus again the
Fe is complementary increasing with the diminishing N
content. In contrast, a 1:1 identity between the nitrogen and
chromium content is observed in 630 steel at a treatment
temperature of 600 jC, leading to a layer with both elements
having a constant concentration down to about 12 Am before
a diffusion tail with decreasing nitrogen and constant
chromium down to 20 Am is observed.

The hardness change after PIII treatment at 400 jC is


documented in Fig. 4. The untreated 630 steel is near a
hardness of 500 HV while the untreated 316Ti is around
300 HV. For both groups, an approximate increase by a
factor of 4 is observed, leading to 1700 2000 HV for
martensitic grades and 1000 1200 HV for austenitic
grades. Analyzing the depth-dependent hardness curves
[15], a plateau is visible in the hardness, thus the
presented values are representative for the surface layer
of 5 20 Am. At 50 g load, a final indentation depth of
about 0.5 1 Am is obtained. Increasing the load beyond
that value, lower hardness values are obtained, indicating
increasing influence of the substrate. The specific wear in
Fig. 5 shows an identical tendency with a slight advantage
for the harder martensitic steels. Albeit, the relative
decrease of 2.5 3 orders of magnitude after PIII treatment

Fig. 3. Depth distribution of Fe, Cr and N in 630 after PIII at 600 jC.

280

S. Mandl, B. Rauschenbach / Surface & Coatings Technology 186 (2004) 277281

Fig. 4. Vickers Hardness for selected samples at 50 g load.

is the same within the measurement accuracy, independent


of the initial structure.

4. Discussion
Summarizing the result from Section 3, a nearly identical behavior is observed for austenitic and martensitic
stainless steels after PIII treatment at 400 jC. Both groups
of steel show a lattice expansion concurrent with a
concentration-dependent diffusion coefficient, increasing
with the nitrogen content. The relative increase of the
hardness and decrease of the wear is not a useful measure
to distinguish between them. Furthermore, stress measurements for austenites [16] and martensites show a very high
compressive stress between 1.5 and 1.7 GPa, close to the
plastic flow limit.
Increasing the Cr and Ni content (respective the equivalent values after inclusion of the additional alloying elements [9]) leads to an enlargement of the area for austenite
stabilization, hence the Cr/Ni values for the investigated
austenites are near 20/10. Steel 440B is a high chromium
steel without Ni, whereas 630 is a steel with a reduced Ni
content near 4 at.%. Finally, steel 420 is reduced in
chromium and free of nickel. It is known that nitrogen
incorporation strongly enhances the Ni equivalent ratio
[10]with an initial weighting factor of 30, but decreasing
for higher nitrogen contentso that a transition towards
austenites should be observable starting with 630, and then
followed by 440B and 420. However, as no such transition
is observed despite local nitrogen contents of more than 15
at.%, a chemical, respective metallurgical explanation for
the observed effects must be ruled out.
The observed difference between treatment at 400 and
600 jC may provide a first clue for the real nature of the
transition process towards the expanded surface layers at
lower temperatures, especially as the same high-temperature
process is active in austenitic stainless steels [17]. At
elevated temperature, chromium gets mobile enough to
diffuse through the material, form a chemical bond with
nitrogen and precipitate in CrN crystallites [18]. Thus, the

nitrogen diffusion at higher temperatures can be described


as diffusion in a two-phase system with a very high
diffusivity in the CrN/ferrite and a much lower one in the
austenite or martensite. Hence, the 1:1 ratio of Cr and N is
formed, effectively removing the chromium from a passivating oxide surface network [19] and strongly increasing the
corrosion current.
It can be assumed that the lower chromium mobility near
400 jC with an effective diffusion length of less than a few
atomic distances during the implantation time is the key for
the absence of CrN clusters and the concurrent segregation.
Sparse literature indicates a tendency towards local Cr N
bonds within the austenite framework and nitrogen still on
interstitial sites [18]. Local atomic displacements, respective
a starting Cr segregation near the surface is seen in the Fe/Cr
ratio for the treatments at 400 jC (see Fig. 2), which can
indicate a low diffusivity reaching only a few lattice
distances during the implantation. The formation of an oxide
layer with 200 400-nm thickness as a cause for the segregation can be ruled out. Nevertheless, detailed studies in the
chemical surrounding of Cr or N in the expanded layer
could identify the true microscopic corrosion or passivation
behavior of the expanded austenite and martensite. The
oxide surface layer looses its passivating properties below
a critical concentration of 13 at.% Cr3 + within the oxide
network, thus a partial removal of chromium caused by
isolated Cr N bonds would massively decrease the corrosion protection.
Thus, excluding the dominance of chemical effects, it
may be proposed that the physical mechanism of ion
implantation and subsequent processes can be responsible
for the observed nitrogen agglomeration and distribution in
the near surface region. It is well known that high-dose

Fig. 5. Specific wear for austenites (a) and martensites (b) before and after
PIII treatment.

S. Mandl, B. Rauschenbach / Surface & Coatings Technology 186 (2004) 277281

implantation of nitrogen, carbon or oxygen into metals leads


to an initial lattice expansion to accommodate the surplus
interstitial atoms. The corresponding nitride, carbide and
oxide phases are energetically favorable enough for transition metals from lower groupse.g. Tito start a phase
transition around local concentrations of 10 20 at.%. However, iron nitrides are less stable so that no strong driving
force is present and nitrogen can stay in solid solution even
at the observed high concentrations.
No change in the crystal size is visible from the XRD
data so that hardening via decrease of the grain size can be
ruled out as an explanation for the increased hardness.
Furthermore, grain sizes of less than 50 nm, which are
necessary for an effect of the observed magnitude, would
lead to a much faster nitrogen diffusion along the grain
boundaries [20]. Another explanation, much more likely, for
the outstanding mechanical properties can be the change of
the electronic structure at elevated nitrogen contents. At the
same time, the diffusivity, which depends on the interatomic
potentials and hence the electronic structure, may also
change drastically at higher nitrogen concentrations. Especially for the transition metals, a very strong influence of the
number of d (or f-electrons) on the phonon dispersions is
known [21,22].

5. Summary and conclusions


Expanded austenitic and martensitic stainless steels are
formed after nitrogen PIII near 400 jC. Both groups of
materials show a lattice expansion depending on the steel
composition, a nitrogen diffusivity increasing with nitrogen
content and an identical increase, respective decrease, in
relative hardness and relative specific wear. The Cr and Ni
equivalent in the steel grades and hence the stability of the
austenitic phase plays no role in the evolution. It is proposed
that the insertion of energetic nitrogen ions leads to lattice
expansion, coupled with an increased diffusion, thus making
the process a purely physical one dominated by the ion
solid interaction. Secondary effects of thermal activation of
diffusing chromium atoms limit the available temperature
range and lead to a segregation into CrN precipitates and
ferrite at higher temperatures. The influence of a chemical

281

Cr N attraction on the local structure and the corrosion


properties needs further investigations.

Acknowledgements
E. Richter, Forschungszentrum Rossendorf, is acknowledged for providing some of the wear and hardness data.

References
[1] D.L. Williamson, J.A. Davis, P.J. Wilbur, Surf. Coat. Technol. 103/
104 (1998) 178.
[2] E. Menthe, K.-T. Rie, Surf. Coat. Technol. 116/119 (1999) 193.
[3] S. Mandl, B. Rauschenbach, Defect Diffus. Forum 188/190 (2001)
125.
[4] S. Mandl, B. Rauschenbach, J. Appl. Phys. 91 (2002) 9737.
[5] S. Mandl, B. Rauschenbach, J. Appl. Phys. 88 (2000) 3323.
[6] S.K. Kim, J.S. Yoo, J.M. Priest, M.P. Fewell, Surf. Coat. Technol.
163/164 (2003) 380.
[7] A. Leyland, D.B. Lewis, P.R. Stevenson, A. Matthews, Surf. Coat.
Technol. 62 (1993) 608.
[8] C. Blawert, B.L. Mordike, U. Rensch, G. Schreiber, H. Oettel, Surf.
Eng. 18 (2002) 249.
[9] A.L. Schaffler, Metal Prog. 56 (1949) 680.
[10] V.G. Gavriljuk, H. Berns, High Nitrogen Steels, Springer, Berlin,
1999.
[11] D. Manova, F. Scholze, S. Mandl, H. Neumann, B. Rauschenbach,
Surf. Coat. Technol. doi:10.1016/j.surfcoat.2004.04.003.
[12] D. Manova, S. Mandl, B. Rauschenbach, Plasma Sources Sci. Technol. 10 (2001) 423.
[13] S. Mandl, R. Gunzel, C. Hammerl, E. Richter, B. Rauschenbach, W.
Moller, Surf. Coat. Technol. 136 (2001) 176.
[14] T. Asam, Surf. Coat. Technol. 116/119 (1999) 310.
[15] S. Mandl, R. Gunzel, E. Richter, W. Moller, Surf. Coat. Technol.
100/101 (1998) 372.
[16] S. Sienz, S. Mandl, B. Rauschenbach, Surf. Coat. Technol. 156 (2002)
185.
[17] X. Li, M. Samandi, D. Dunne, G. Collins, J. Tendys, K. Short, R.
Hutchings, Surf. Coat. Technol. 85 (1996) 28.
[18] D.L. Williamson, O. Ozturk, R. Wei, J.P. Wilbur, Surf. Coat. Technol.
65 (1994) 15.
[19] E. McCafferty, Corr. Sci. 44 (2002) 1393.
[20] W.P. Tong, N.R. Tao, Z.B. Wang, J. Lu, K. Lu, Science 299 (2003)
686.
[21] F. Willaime, Adv. Eng. Mater. 3 (2001) 283.
[22] J. Wong, M. Krisch, D.L. Farber, F. Occelli, A.J. Schwartz, T.-C.
Chiang, M. Wall, C. Boro, R. Xu, Science 301 (2001) 1078.

Anda mungkin juga menyukai