Tribocorrosion behaviour of DLC-coated 316L stainless steel

Abstract

Export Date: 16 June 2009, Source: Scopus

Full text

Page 1

Wear 267 (2009) 860–866
Contents lists available at ScienceDirect
Wear
journal homepage: www.elsevier.com/locate/wear
Tribocorrosion behaviour of DLC-coated 316L stainless steel
M. Azzia,b, M. Paquetteb, J.A. Szpunara, J.E. Klemberg-Sapiehab, L. Martinub,∗
aDepartment of Mining and Materials Engineering, McGill University, Montreal, QC, Canada, H3A 2B2
bDepartment of Engineering Physics, Ecole Polytechnique, Montreal, QC, Canada, H3C 3A7
a r t i c l e i n f o
Article history:
Received 18 September 2008
Received in revised form 4 February 2009
Accepted 4 February 2009
Keywords:
Tribocorrosion
316L
Wear
DLC
a-SiNx:H
EIS
Corrosion
a b s t r a c t
In the present work, we systematically investigate the tribocorrosion behaviour of diamond-like carbon
(DLC)filmson316Lstainlesssteelsubstratesinthecontextoftheirbiomedicalapplications.Twodifferent
bond layers at the interface were particularly studied, namely the plasma nitrided layer and the plasma
depositedamorphoushydrogenatedsiliconnitride(a-SiNx:H).Tribocorrosiontestswereperformedusing
a ball-on-flat tribometer where the sliding contact is fully immersed in NaCl 1wt.% solution. The sample
was connected to a potentiostat: it served as a working electrode and its open circuit potential (OCP) was
monitoredbefore,during,andaftersliding.Electrochemicalimpedancespectroscopy(EIS)wasappliedto
characterize the electrochemical behaviour of the surfaces before and after rubbing. The OCP measured
during sliding was shown to depend on the properties of the protective layer; a decrease in the OCP
indicatesdelaminationoftheprotectivelayerandsubsequentexposureofthesubstratetotheelectrolyte.
We found that the DLC coating with the nitrided bond layer delaminated from the wear track within 50
cycles of sliding, while it resisted the entire tribocorrosion test (1800 cycles) without failure when the a-
SiNx:H bond layer was applied. The EIS results are interpreted in terms of appropriate equivalent circuits.
It is shown that the a-SiNx:H bond layer significantly increases the corrosion resistance by acting as a
corrosion barrier, while the DLC coating assures high wear resistance and low friction. The polarization
resistance of DLC-coated 316L with the a-SiNx:H bond layer was found to be 3.76G?cm2compared to
27.5M?cm2for the same DLC coating without a-SiNx:H.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Tribocorrosion is a term which describes the degradation of
materials that results from the combined action of wear and corro-
sion [1,2]. Examples include degradation of articulation prosthesis
and dental fillers, accelerated corrosion of steel conveyors exposed
to ambient air of high relative humidity, erosion-corrosion of slurry
pipes,andnumerousothers.Whencorrosionandweararesimulta-
neously involved, their synergistic action significantly deteriorates
the performance of the materials in contact.
Over the past years, a number of authors investigated the tri-
bocorrosion behaviour of engineering substrates. Mischler and
co-workers [1,2] studied the effect of applied voltage on the wear
mechanisms of Fe–17Cr stainless steel and carbon steel. Ponthiaux
et al. [3] investigated the tribocorrosion behaviour of 316L stainless
steel. Barril et al. [4] used the potentiostatic polarization tech-
nique to investigate the fretting corrosion behavior of Ti–6Al–4V
in 0.9wt% sodium chloride solutions.
In orthopaedic applications, artificial joints (e.g., hip and knee
prostheses) include bearing surfaces where the material is sub-
∗Corresponding author. Tel.: +1 514 340 5747; fax: +1 514 340 3218.
E-mail address: lmartinu@polymtl.ca (L. Martinu).
jected to sliding wear. The surfaces in contact are immersed in the
body fluid, and therefore, corrosion may also be a concern. For such
application, metals and alloys must possess high tribocorrosion
resistance. Particles generated from wear of prosthetic implants
induce inflammatory reactions that provoke the release of inflam-
matory mediators from macrophages [5]. It is well established that
the cellular response to wear debris depends, among other factors,
on the number, shape, size, surface area, and materials chemistry
of the particles [6,7].
In recent years, diamond-like carbon (DLC) films have been the
subject of extensive investigations due to their potential of attain-
ing a combination of highly desirable properties in the context of
biomedical applications [8–14]. Their high hardness, low friction
and wear, electrical insulation, chemical inertness and good bio-
compatibility make them ideal candidate as protective coatings in
joints replacement [8].
DLC films are usually characterized by an sp2–sp3hybridization
ratio of approximately 0.4, and a hydrogen concentration in the
range of 20–40at.% [9]. Raveh et al. [10,11] studied the relationship
between the structure of the DLC films and their mechanical and
tribological properties. The authors reported that DLC films con-
taining predominantly unbonded hydrogen can resist more severe
conditions of sliding friction than those containing predominantly
bonded hydrogen. Martinu et al. [12] showed that the addition of
0043-1648/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2009.02.006

Page 2

M. Azzi et al. / Wear 267 (2009) 860–866
861
Table 1
Experimental conditions of sample pre-treatment, nitriding, DLC and a-SiNx:H deposition, and the corresponding hardness and Young’s modulus.
Operation Pressure (mTorr)Gaseous mixture (sccm)Bias voltage (V)Process duration (min)Hardness (GPa)Young’s modulus (GPa)
Ar sputter-cleaning
Nitriding
a-SiNx:H deposition
DLC deposition
100
100
100
100
Ar (10)
N2(50)
SiH4(6.6), N2(20), Ar (30)
CH4(40), Ar (10)
−800
−500
−400
−500
156 220
220
160
140
180
10
30
15
17
18
argon to the hydrocarbon feed gas and the application of a com-
bination of microwave and radio frequency power during plasma
deposition substantially enhance gas phase processes such as dis-
sociation of CH4and formation of the sp3component in the DLC,
leading to a substantial decrease in the mechanical stress.
Another important factor that influences the protective charac-
ter of DLC films is their adhesion to the substrate. Many authors
pointed out the importance of the bond layers for enhanced adhe-
sion of the DLC films. Snyders et al. [13] suggested a simple nitrided
bond layer. They reported significantly improved adhesion and dry
wear resistance of DLC on 316L stainless steel. Chandra and co-
workers [14] successfully tested amorphous hydrogenated silicon
(a-Si:H) bond layer on Ti–6Al–4V.
One should keep in mind that pores are often present, even
in good quality DLC films [15]. In this context, electrochemical
impedance spectroscopy (EIS) was shown to be a sensitive tech-
niqueindetectingporesanddefectsevenatthenano-scale[16–19].
In addition, different circuits were used to simulate the impedance
spectra, incorporating elements representing microstuctural fea-
turessuchasopenporesthatallowtheinfiltrationoftheelectrolyte
through the film. Lillard et al. [20] used the EIS technique to study
the breakdown mechanism of DLC-coated nickel in chloride solu-
tion.Theauthorsreportedtheinitiationofbreakdownatthebottom
of the pores inside the DLC film.
In this work, we systematically investigate the tribocorrosion
behaviourofDLC-coated316LstainlesssteelexposedtotheRinger’s
solution in the context of biomedical applications. Two different
bond layers are studied, namely amorphous hydrogenated silicon
nitride(a-SiNx:H)andplasmanitridedlayers.Electrochemicaltech-
niques such as open circuit potential (OCP) measurements and EIS
are applied for the assessment of the corrosion resistance.
2. Experimental methodology
2.1. Coating deposition
DLC coatings were deposited on 316L stainless steel substrates
(25mm×25mm×1.2mm) using a turbo-molecularly pumped
radio frequency (13.56MHz) PECVD system, equipped with a 15cm
diameter electrode where a self-induced DC bias voltage, VB, devel-
ops. Two different methods for interface engineering between the
316L substrate and the DLC film were used in this study: (1) nitrid-
ing of the stainless steel surface, and (2) deposition of amorphous
hydrogenated silicon nitride bond layer.
The316Lsubstratesweremechanicallypolishedusing1?malu-
minasuspension.Afterpolishingthespecimenswereultrasonically
cleaned in acetone (15min) and isopropanol (15min), and then
introduced into the deposition chamber. Prior to nitriding or depo-
sition, the substrates were cleaned with Ar plasma sputtering for
15min to remove the native oxide eventually formed after polish-
ing. The experimental conditions of the substrate pre-treatment,
the interface engineering and the DLC film deposition are summa-
rized in Table 1, as well as the mechanical properties obtained from
the nano-indentation measurements.
2.2. Tribocorrosion test
Tribocorrosion experiments were performed using a linear
reciprocating ball-on-flat tribometer described in Ref. [21]. During
thetests,a3/16??(4.75mm)diameteraluminaballrubsonthespeci-
mensurfaceimmersedinthetestsolution.Thesampleservedasthe
working electrode and its potential was controlled using Autolab
PGSTAT302 potentiostat equipped with a frequency response anal-
yser. The counter electrode was made of coiled platinum, and the
standardcalomelelectrodeSCE(+241mVversusstandardhydrogen
electrode) was used as a reference for the potential measurements.
Sliding tests were carried out with 9N normal force applied
using a compression spring. This corresponds to a maximum
Hertzian contact pressure of 1.18GPa, calculated based on the con-
tact between the alumina ball and the 316L flat surface. The stroke
length was 10mm, and the sliding frequency was 1Hz. Ringer’s
solution of pH ≈6.6 was used as an electrolyte: its composition was
9g/l NaCl, 0.4g/l KCl, 0.17g/l CaCl2and 2.1g/l NaHCO3in distilled
water.
The sequence of operations during the tribocorrosion test is
schematically illustrated in Fig. 1 that shows an evolution of the
OCP. First, the sample was immersed in the Ringer’s solution for
Fig. 1. Sequence of operations during the tribocorrosion test illustrated by the OCP evolution.

Page 3

862
M. Azzi et al. / Wear 267 (2009) 860–866
1h in order to reach a stable potential (Region I). Next, EIS was per-
formedtocharacterizetheelectrochemicalbehaviourofthesurface
before sliding wear (Region II). Subsequently, the alumina ball was
loaded on the sample surface and the sliding test was initiated. The
sliding was stopped after 1800 cycles. During and after the slid-
ing wear test, the OCP was continuously monitored (Regions III and
IV). Finally, EIS was performed to characterize the electrochemical
behaviour of the surface after sliding (Region V). The EIS spectra
were acquired over the frequency range from 105Hz to 10−2Hz, at
OCP, with an AC amplitude of 10mV.
Tribocorrosion behaviour of the bare 316L stainless steel sub-
strates (SS), DLC-coated 316L with nitrided (3h) bond layer
(SS/N3h/DLC), and of the DLC-coated 316L with the a-SiNx:H bond
layer(SS/a-SiNx:H/DLC)wasinvestigated.Thetribocorrosionexper-
iments were repeated three times for the same surface condition
to validate the results. SS, SS/N3h/DLC and SS/a-SiNx:H/DLC nota-
tions will be used in the rest of this paper to represent the above
mentioned samples.
2.3. Surface characterization
Field emission scanning electron microscope (FESEM, Philips
XL30) equipped with an energy dispersive spectrometer (EDS)
was used to perform cross-section analysis of the coated sub-
strates and to characterize the worn surfaces. A Rigaku Rotaflex
X-ray diffractometer (XRD) was used to characterize the crystal
structure of the deposited films. XRD spectra were acquired at a
grazing angle of 1.5◦, using CuK? radiation of 1.54nm wavelength,
under 40kV voltage and 40mA current. Chemical compositions
of the DLC and a-SiNx:H films were measured by the elastic
recoil detection (ERD) technique in the time-of-flight regime
[10].
3. Results and discussion
3.1. Characterization of the as-deposited coatings
Fig. 2 shows the SEM cross-section micrographs of SS/N3h/DLC
and SS/a-SiNx:H/DLC. For SS/a-SiNx:H/DLC, the thicknesses of the
a-SiNx:H and DLC layers were approximately 350nm and 650nm,
respectively. For SS/N3h/DLC, the thickness of the DLC film was
also about 650nm. The white particles shown in Fig. 2(a) are pol-
ishing solution residues. Previous study of the nitrided layer [13]
showed that the nitriding process at these conditions results in
the formation of a relatively thick (1–2?m), hard (15GPa) metal
nitride layer at the stainless steel surface. Chemical compositions
determined by ERD of the DLC and a-SiNx:H films are presented in
Table 2.
Table 2
Chemical composition of the DLC and a-SiNx:H films measured by ERD.
C (at.%)Si (at.%) N (at.%)H (at.%)
DLC
a-SiNx:H
85
–
–– 15
1735 48
Fig. 3. OCP measurements before sliding of SS, SS/N3h/DLC, and SS/a-SiNx:H/DLC in
Ringer’s solution.
XRD analysis has shown that both a-SiNx:H and DLC layers are
amorphoussincenodiffractionpeakswereobservedinthespectra.
3.2. Open circuit potential measurements
3.2.1. OCP before sliding
The OCP evolutions during the first hour of immersion of SS,
SS/N3h/DLC, and SS/a-SiNx:H/DLC in Ringer’s solution are pre-
sented in Fig. 3. For all samples, the OCP increased and stabilized
after 1h, indicating a stable electrochemical condition at the sur-
face.TheOCPstabilizedat−90mVforSS,−125mVforSS/N3h/DLC,
and −70mV for SS/a-SiNx:H/DLC.
3.2.2. OCP during and after sliding
The OCP measurements during and after sliding of SS,
SS/N3h/DLC, SS/a-SiNx:H/DLC, and SS/a-SiNx:H are shown in Fig. 4.
For SS, a sudden decrease of the OCP was observed when rubbing
started. The potential dropped to approx. −400mV, and was fluc-
tuating in phase with the reciprocating motion of the alumina ball.
When rubbing ceased, the OCP first steeply increased and then
Fig. 2. SEM cross-sectional micrographs of (a) SS/N3h/DLC and (b) SS/a-SiNx:H/DLC.

Page 4

M. Azzi et al. / Wear 267 (2009) 860–866
863
Fig. 4. OCP measurements during and after sliding test of SS, SS/N3h/DLC, SS/a-
SiNx:H and SS/a-SiNx:H/DLC. Normal load: 9N, frequency: 1Hz.
progressively returned to a steady state value. During sliding the
frictioncoefficientwasabout0.3.ForSS/N3h/DLC,theOCPdropped
within the first 50 cycles of sliding to reach an average value of
−650mV. During the first few cycles, the friction coefficient was
0.08. Subsequently, it progressively increased to reach 0.3 after
about 50 cycles. When sliding ceased, the evolution of OCP was
similar to the one measured on SS; a steep increase followed by a
progressive stabilization. In the case of SS/a-SiNx:H/DLC, the OCP
remained constant during and after sliding, and the fluctuations of
the OCP during rubbing were small compared to the ones observed
ontheothersamples.Inaddition,thefrictioncoefficientwasalmost
constant (0.08) throughout the entire sliding test. For SS/a-SiNx:H,
the OCP dropped, at the onset of rubbing, to an approx. value of
−800mV, and then increased when rubbing ceased.
The SEM images of the wear tracks of the different investi-
gated samples are shown in Fig. 5. It is clearly seen that the DLC
film delaminated from SS/N3h/DLC whereas on SS/a-SiNx:H/DLC,
the DLC film had no tendency to delaminate. For SS/a-SiNx:H, the
a-SiNx:Hfilmwascompletelyremovedfromtheweartrack.Inaddi-
tion, a stainless steel transfer layer was observed on the alumina
ball after all the tribocorrosion tests except the one performed on
SS/a-SiNx:H/DLC where the ball was found almost intact.
The sharp drop of the OCP observed at the onset of rubbing on
SS is due to the removal of the oxide layer from the wear track
area and the subsequent exposure of fresh active materials to the
aqueous environment. Under these conditions, 316L stainless steel
undergoesanodicdissolutionfollowedbytheformationofapassive
oxide film on the surface according to the following reactions [22]:
M → Mn++ne−
mM + nH2O → MmOn+H++xe−,
where M designates the metallic materials of the stainless steel,
mainly Cr and Fe. The oxidation reactions (whether dissolution or
oxide formation) take place in the wear track and generate elec-
trons which must be consumed by the cathodic reaction in order
for the oxidation reactions to proceed. The main cathodic reaction
supporting the passivation reaction under these conditions is the
reduction of the dissolved oxygen [22]:
(1)
(2)
O2+2H2O + 4e−→ 4OH−
Ponthiaux et al. [3] measured the OCP of 316 stainless steel sub-
strateimmersedinaerated0.5Msulphuricacidandrubbingagainst
an alumina ball counterpart. They reported an OCP of approxi-
mately −800mV vs. SCE during sliding compared to −400mV vs.
SCE in our case. The difference can be explained by the cathodic-
to-anodic (CTA) ratio during sliding. In our tests, the total exposed
area was 283.5mm2and the wear track area was only 3mm2
(300?m×10mm), as shown in Fig. 5(a). This results in a CTA ratio
of 93, which is very high compared to the CTA in the Ponthiaux’s
experiments where a large part of the area outside the wear track
was insulated with cataphoretic paint to reduce the cathodic reac-
tion activities. Berradja et al. [23] studied the fretting behaviour
of 304 stainless steel; they also reported a decrease of the poten-
tial at the onset of fretting experiments, and referred this drop to
the exposure of the fresh active material to the environment. They
measured only −200mV vs. SCE since the depassivated area in the
(3)
Fig. 5. SEM images of the wear tracks after 1800 cycles of sliding on (a) SS, (b) SS/N3h/DLC, (c) SS/a-SiNx:H/DLC, and (d) SS/a-SiNx:H.

Page 5

864
M. Azzi et al. / Wear 267 (2009) 860–866
Fig. 6. Cathodic polarization curves of SS, SS/N3h/DLC and SS/a-SiNx:H/DLC in
Ringer’s solution.
fretting experiment was very small compared to the sliding wear
test.
For SS/N3h/DLC, the drop in the OCP indicates that the coat-
ing was completely removed after 50 cycles of sliding to leave
behind a fresh active material in the wear track exposed to the
electrolyte. However, the average OCP during sliding was rel-
atively low (−650mV) compared to the one measured on SS
(−400mV). This is due to the fact that the cathodic reaction rate
was lowered by the presence of the DLC film on the unworn
area. Indeed, the cathodic polarization curves of SS, SS/N3h/DLC
and SS/a-SiNx:H/DLC, presented in Fig. 6, show that the cathodic
current density, i (A/cm2), of the DLC-coated 316L is lower than
the one of SS. For example, at −400mV, the current densities
were 4×10−6A/cm2, 5×10−7A/cm2, and 8×10−10A/cm2for SS,
SS/N3h/DLC, and SS/a-SiNx:H/DLC, respectively. In addition, these
curves show that the a-SiNx:H film significantly reduces the rate of
the cathodic reaction, and therefore explains the low value of the
OCP measured on SS/a-SiNx:H during the sliding test.
Snyders et al. [13] conducted dry sliding tests on the same
SS/N3h/DLC samples, deposited in the same reactor and under the
same conditions. They reported a high wear resistance without
delamination up to 22N normal load, using the same counterface
material. This completely different behaviours under dry and wet
conditions clearly revealed the effect of corrosion processes on the
wear resistance of DLC coating. Lillard et al. [20] reported that the
corrosionreactionsatthebottomoftheporesarethefirststepinthe
breakdown of DLC coatings in chloride solution. This is believed to
be the reason for the delamination of the DLC film during the slid-
ing test in Ringer’s solution, the solution infiltration through the
Fig.7. EISspectraandmodelledcurvesofSS,SS/N3h/DLCandSS/a-SiNx:H/DLCafter
1h immersion in Ringer’s solution.
pores in the DLC may weaken the DLC/metal interface and reduce
the adhesion of the DLC film.
For SS/a-SiNx:H/DLC, the coating did not delaminate from the
wear track. This explains why the OCP was relatively high and did
not drop during the sliding test. The a-SiNx:H layer significantly
reduced the charge transfer between the metallic substrate and the
electrolyte, as it will be shown later in this work, by acting as a
barrier layer between the substrate and the electrolyte. Therefore,
it reduced the possibility of weakening the interface between the
substrate and the film. The combination of a-SiNx:H and DLC films
was appropriate for the wet wear conditions: the first layer sig-
nificantly reduces the corrosion rate, while the second layer, due
to its high hardness, assures wear resistance and low friction, and
protects the a-SiNx:H from mechanical wear.
TheincreaseoftheOCP,whenrubbingceases,isduetotherepas-
sivation of the stainless steel material. It was reported [1–3] that in
the absence of sliding wear the stainless steel surface repassivetes
and the OCP increases.
3.3. Electrochemical impedance spectroscopy (EIS)
3.3.1. EIS before the sliding test
The EIS spectra of SS, SS/N3h/DLC, and SS/a-SiNx:H/DLC after
1h of immersion in Ringer’s solution are shown in Fig. 7. The
value of the impedance |Z| is plotted on the left y-axis and
the phase shift Arg(Z) is plotted on the right y-axis. The Arg(Z)
plot of SS/a-SiNx:H/DLC is shifted towards high angles at high
frequencies indicating a more capacitive response of the film.
In addition, the impedance of SS/a-SiNx:H/DLC is almost three
orders of magnitude larger than the impedance of SS, while it
Fig. 8. Equivalent circuits used for EIS spectra simulation: (a) basic (Randle) circuit; (b) double capacitance circuit.