Augmentation of crevice corrosion resistance of medical grade 316LVM stainless steel by plasma carburising

Corrosion Science 59 (2012) 169–178

Contents lists available at SciVerse ScienceDirect

Corrosion Science journal homepage: www.elsevier.com/locate/corsci

Augmentation of crevice corrosion resistance of medical grade 316LVM stainless steel by plasma carburising Joseph Buhagiar a,⇑, André Spiteri a, Malcolm Sacco b, Emmanuel Sinagra b, Hanshan Dong c a

Department of Metallurgy and Materials Engineering, University of Malta, MSD 2080, Malta Department of Chemistry, University of Malta, MSD 2080, Malta c School of Metallurgy and Materials, The University of Birmingham, Birmingham B15 2TT, UK b

a r t i c l e

i n f o

Article history: Received 20 October 2011 Accepted 28 February 2012 Available online 9 March 2012 Keywords: A. Stainless steel B. Potentiostatic C. Pitting corrosion C. Hardening C. Crevice corrosion C. Carburisation

a b s t r a c t Precipitate free carbon S-phase was produced on the surface of AISI 316LVM medical grade austenitic stainless steel with the use of low temperature direct current and active screen plasma carburising. The treated and untreated alloy was characterised and tested for pitting and crevice corrosion resistance. From this work it can be concluded that when compared to the untreated material, both treatments augment the pitting and crevice corrosion resistance. Using an active screen set-up results in a better surface composition and a higher crevice corrosion resistance than that produced using the direct current plasma carburising treatment. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Austenitic stainless steels, such as AISI 316LVM (ASTM F138), are widely used in medical applications where biocompatibility and corrosion resistance are of utmost importance [1]. Currently the 316LVM alloy is mostly used for temporary implants which are designated to last 6–12 months in an intrahuman environment [2]. These temporary implants are generally bone plates and are locked in place using screws. Some austenitic stainless steel implants, such as hip joint replacement stems and intramedullary rods, are also implanted for up to 10 years. In such environments stainless steel suffers from crevice and pitting corrosion which can cause the implant to fail prematurely [2,3]. In the case of bone plates and screws, crevice corrosion can occur at the interface between the tightly contacted bone plate and the screw on the countersink. Corrosion results in ion release which, when enhanced by mechanical processes such as wear, cause the formation of debris allowing the prosthesis to grow loose and inevitably cause problems. Williams stated that ‘‘The success of any implant is dependent on its bulk and surface properties, the site of implantation, tissue trauma during the surgery and motion at the implant/tissue interface’’ [2]. Hence modifying the surface properties of austenitic stainless steel by enhancing their corrosion and wear resistance properties would in turn lead to an increased range of ⇑ Corresponding author. Tel.: +356 2340 2439; fax: +356 2134 3577. E-mail address: [email protected] (J. Buhagiar). 0010-938X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.corsci.2012.02.023

applications where they would be able to substitute the other more expensive alloys. Surface engineering of austenitic stainless steel by the creation of a precipitate free carbon supersaturated layer called S-phase has improved both the pitting corrosion resistance [4–9] and tribological properties of these alloys [10,11]. A number of techniques can be used to form this layer three of them being: gas [5,7,9], direct current (DC) [1,10] and active screen low temperature plasma surface carburising [10,12–14]. In direct current plasma carburising a carbon bearing gas such as methane is introduced in a chamber and a voltage is applied between the chamber wall (anode) and the work table (cathode). Active screen plasma carburising uses of a cathodic cage also know as a screen, which surrounds the work to be treated. The cage is constructed of a mesh which is subject to a highly cathodic potential and on which plasma is formed. In this case the work table is given a bias, which is a fraction of the potential used to generate the plasma [10,12,13]. Direct current plasma treatments present a number of surface defects and process instabilities such as surface arcing, hollow cathode and edge effects. Active screen low temperature plasma surface alloying is advantageous in so much as many of the defects generated by direct current treatments are minimised or completely eliminated by its application. Its use also results in improved treatment temperature control, material properties and surface quality [12]. Although the effectiveness of the active screen plasma surface treatment is evident, limited work has been

170

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

conducted to compare the corrosion properties of direct current and active-screen treated austenitic stainless steels. Extensive research on S-phase surface engineering of austenitic stainless steel have shown many promising results in terms of effectively improved biocompatibility [1,10], fatigue [10], wear-corrosion [10], fretting-wear [10] and localised corrosion properties [4,10]. These very promising results could pave the way towards long-life, high-performance medical devices [10]. The research work which is related to the localised corrosion resistance of low temperature carburised industrial grade stainless steel in chloride containing environments was carried out by the following authors: Martin et al. [5,9], Heuer et al. [7,8], Sun [6] and García Molleja et al. [15]. In his work Sun [6] concludes that in order to augment the localised corrosion resistance carbon S-phase must have a critical carbon concentration of about 0.25 wt.%. García Molleja et al. [15] carburise an AISI 316L alloy and perform immersion tests in a 5.85 wt.% NaCl solution. Contrary to other findings they report that the carbon S-phase layer had a reduced corrosion resistance when compared to the untreated material. In the works by Martin et al. [5,9] and Heuer et al. [7,8] the enhanced localised corrosion resistance in both 0.6 M NaCl and sea water of carbon S-phase has been reported. Heuer et al. [7,8] challenges the widely accepted theory given by Jargelius-Pettersson [16] by proposing a chemomechanical model of passive film breakdown in order to credit carbon or nitrogen interstitials for this enhanced resistance to localised corrosion. While the theory given by Jargelius-Pettersson [16] can only explain the benefits of nitrogen; on the other hand, the model by Heuer et al. [7,8] includes the beneficial effects of both nitrogen and carbon. Crevice corrosion work on untreated austenitic stainless steel has been researched extensively [17] but the same testing on S-phase treated stainless steel has not been performed. In fact the only reference to crevice corrosion testing of low temperature carburised alloys is related to a Ni-base super alloy rather than stainless steel [18]. Some work has been conducted in the investigation of the hardness and localised corrosion resistance of low temperature plasma surface alloyed biomedical austenitic stainless steels [4,10]. However, most of this work is applied only to the nitrogen S-phase and the crevice corrosion resistance of the Sphase layer created by active screen technology on biomedical stainless steels has never been explored [10]. This study attempts to fill this gap by investigating the pitting and crevice corrosion resistance of untreated and carburised 316LVM stainless steel. The carburising treatments used in this study use the well documented direct current carburising treatment [1,10] and the new active screen carburising treatment [10,12–14].

2. Materials and methods 2.1. Material and surface treatments The material used in this study was an ASTM F138 (Sandvik Bioline 316LVM) austenitic stainless steel which was supplied in the form of an annealed bar of 25 mm in diameter. Its composition can be found in Table 1.

Coupon samples of 6 mm in thickness were cut from the bar and one of the flat surfaces was wet ground using silicon carbide paper from 120 down to 1200 grit. Prior to plasma surface alloying treatments, samples were ultrasonically cleaned in acetone and dried with hot air. Low temperature plasma surface alloying with carbon (carburising) was carried out using a DC 40 kW Klöckner Ionon and a 75 kW Plasma Metal Active Screen plasma furnace. The treatments will be referred to as direct current plasma carburising (DCPC) for the former and active screen plasma carburising (ASPC) for the latter. The coupons were treated in a specially designed sample holder made from a 13 mm thick disc that has a diameter of 200 mm. A thermocouple sheath was radially inserted into the side of the holder. The samples, 25 in total, were placed in recesses machined in the sample holder with their upper surface flush to its surface. Thus the specimens and holder presented a uniform surface with just one edge (the holder circumference) on which the edge effect could occur, and a continuous bulk which allowed a uniform temperature to be attained. Surface treatment conditions and the codes for the samples are given in Table 2. The chosen treatment parameters were based on previous work for producing precipitate free S phase layers on the AISI 316 alloy by DCPC [1] and ASPC [12]. Following the surface treatments, all treated and untreated samples were polished due to the presence of a back-deposited superficial layer as explained in Ref. [19]. Transmission electron microscope observation in our previous work [19] revealed that this back deposited layer consisted of extremely fine equiaxed grains with a diameter of 5–10 nm and with a thickness of 50 nm. Its structure can be assigned to an fcc structured M(N,C) where M = Fe, Cr, Ni, Mo and Mn. Polishing was conducted on a Streurs LaboPol-5 automatic polisher using 6 lm diamond paste with a medium force (mark 3) for 3 min. This was followed by a final polishing at a low force (mark 1) using 1 lm diamond paste for another 3 min. In order to gauge the thickness of material removed a 5 lm GDOES hole was sputtered, measured with a profilometer and then the sample was polished until the mark was no longer visible. Using this polishing technique for all the samples, made sure that less than 5 lm of the layer was removed and the surface finish (Ra) of all the polished samples was between 0.06 and 0.10 lm. 2.2. Characterisation Standard procedures were followed to prepare metallographic specimens to be examined under a Nikon OPTIPHOT-100 optical microscope. This included cross-sectioning normal to the surface, mounting in phenolic resin, wet grinding with silicon carbide paper, polishing and etching in a solution containing 50 ml of HCl (39% conc.), 25 ml of HNO3 (69% conc.) and 25 ml of distilled water [20]. Composition depth-profile analysis was carried out using a LECO GDS-750 QDP Glow Discharge Optical Emission Spectroscopy (GDOES). This equipment was calibrated for all the alloying elements found in stainless steel with special attention to carbon. The surface hardness of the samples was measured using a Mitutoyo MVKH2 micro-hardness tester with a Vickers indenter. A constant load of 0.05 kgf was used, with 10 repeats for each measurement.

Table 1 Composition of the 316LVM material, wt.%. Material

ASTM F138

Composition C

Si

Mn

P

S

Cr

Ni

Mo

Cu

N

Nb

Fe

0.019

0.5

1.87

0.018

0.001

17.43

13.75

2.72

0.06

0.08

–

Bal.

171

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178 Table 2 Process Parameters for the plasma carburising treatments. Treatment

Temperature (°C)

Time (h)

Bias (%)

Pressure (P)

Direct current Active screen

450 450

15 15

N/A 15

400 125

The phase constituents in the as-received, plasma treated surfaces were analysed with an X’Pert Philips X-Radiation diffractometer using Cu-Ka radiation (k = 0.154 nm). The scanning step was of 0.02° at a dwelling time of 3 s. The diffraction patterns obtained were analysed and indexed using X’Pert High Score analytical software for automated powder diffraction. 2.3. Crevice and pitting corrosion testing 2.3.1. ASTM G48 Crevice corrosion tests were performed following the ASTM G48 standard. The solution used contained 170 ml of acidified ferric chloride (14.61 g/l of HCl and 66.51 g/l of FeCl3). A jig, shown in Fig. 1, which exposed only one surface side of the sample to a serrated crevice washer and solution environment for 72 h, was used. After mounting the sample and pouring the solution, the serrated crevice washer was loaded by a bolt with 1.6 N m of torque. The tests were conducted at room temperature (19 °C), the total area subjected to the solution was of 452 mm2 and the creviced area was of 64 mm2. Five repeats per condition were performed and the weight of each sample before and after testing was determined after ultrasonic cleaning using a high precision Precisa balance (resolution of ±0.1 mg).

Fig. 1. Schematic cross-sectional diagram of crevice corrosion jig including: (A) fluid container; (B) base outer enclosure; (C) sample enclosure; (D) crevice corrosion plunger; (E) lid; (F) lock nut; and (G) torque applicator screw.

Gas composition (%) CH4

H2

3 3

97 97

2.3.1.1. Electrochemical monitoring. The open-circuit potential (OCP) of the corroding samples was monitored using a Gamry Ref600 potentiostat configured with a 3-electrode setup that consisted of a platinum counter electrode, a reference electrode (SCE) and a working electrode (sample). 2.3.1.2. Standard preparation and solution analysis. Pure chromium metal (1.00 g) was placed into a 500 ml round bottomed flask. To this were added 100 ml concentrated hydrochloric acid and heated to boiling. Acidified ferric chloride was then added slowly to the mixture. The contents were then refluxed for 4 h until the chromium was completely dissolved. The mixture was then allowed to cool overnight and transferred to a 1-l volumetric flask and then diluted by acidified ferric chloride solution to give a standard stock solution of 1000 lg/ml. A set of standards were produced from this stock solution by dilution using acidified ferric chloride solution. The chromium concentrations ranged from 5–50 lg/ml. This range was determined from weight loss analysis of the 316LVM corrosion test sample. The same procedure was adopted for the preparation of a set of nickel standards from nickel powder (1.00 g). The extracts from the corrosion tests were analysed for chromium and nickel using a Varian AA280Z Zeeman Flame Atomic Absorption Spectroscopy (FAAS). The analysis for chromium was performed using an acetylene/nitrous oxide flame, whilst that for nickel was conducted in an acetylene/air flame. This was necessary in order to eliminate interference from the presence of iron in the analysed solution. The extracts were also tested for chromium and nickel using X-ray Fluorescence spectrometry using a Bruker Ranger 2 instrument which is capable of analysing solutions. The standards were introduced to the instrument in the appropriate liquid-sample holders. The atmosphere in the sample chamber was regulated by helium at 1024 mbar. The composition of the standards and the samples were input into the onboard computer of the instrument. Florescence spectra were measured for chromium at 30 kV using a Ag (5.0 lm), 40 kV using an Al (500 lm) filter and 20 kV with no filter. The onboard computer software then adjusted for matrix interferences and used the calibration curve to give values of chromium concentrations in the samples. In the case of nickel, an energy of 20 kV was used without any filter. 2.3.2. ASTM F746 The critical pitting potential (CCP) was measured in compliance with the ASTM F746 standard. Although most of the standard procedure was kept unchanged, the apparatus, test solution and crevice former used were not as specified by the standard. These changes were put in place due to the shape and size of the treated surface to be tested. The apparatus used was the same crevice corrosion cell that was described in the previous section. A Gamry Ref600 potentiostat was connected to a 3-electrode setup consisting of a SCE reference electrode, a platinum counter electrode and a working electrode (sample). The specimen was pressed against a serrated crevice washer (contact area 64 mm2) that was loaded by a bolt with 1.6 N m of torque in order to create a crevice situation that had an exposed surface area of 388 mm2.

172

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

The test was conducted in 170 ml Ringer’s solution (0.1540 M NaCl; 0.0056 M KCl; 0.0043 M CaCl2; 0.0024 M NaHCO3) at a temperature of 37 °C ± 1 °C in order to mimic the body conditions. In order to prepare one litre Ringer’s solution eight, one quarter strength LAB100Z Ringer’s solution tablets supplied by Lab M (UK) were used. The total area subjected to the solution was of 452 mm2 and the creviced area was of 64 mm2. Three repeats for every test surface were carried out. Before each test, the solution was deaerated with nitrogen at a flow rate of 150 cm3/min for 30 min. The open circuit potential (OCP) was left to stabilise for one hour and the final value, E1, recorded. The potential was then potentiostatically shifted to +800 mV (SCE) to stimulate pitting and crevice corrosion. Current was then monitored to establish if localised corrosion was stimulated or not. When current is observed to remain very small or decrease rapidly with time during the initial 20 s it is then observed for an additional 15 min. The low current value or decrease in current indicates that localised corrosion is not stimulated. If localised corrosion cannot be stimulated after 15 min the test is terminated, and the material under test is considered to have a very high localised corrosion resistance in the test environment conditions. Otherwise in the case of localised corrosion stimulation, which is characterised by increasing polarising currents or by current densities exceeding 500 lA/cm2 the potential is shifted to E1 to determine if the specimen will repassivate or if localised corrosion will continue to propagate. If the current increases with time it means that repassivation does not take place at E1 and the test is terminated and E1 is considered to be the critical voltage. Repassivation takes place if at E1 the current is observed to drop rapidly to zero or to low values. This current is then observed for 15 min. If repassivation at E1 is monitored during the 15 min, the stimulation step at +800 mV (SCE) is repeated and then the potential is changed rapidly to a potential that is 50 mV (SCE) more noble than E1. The test is terminated at the incremental potential which causes the localised corrosion to continue to propagate instead of repassivating.

Fig. 2. Carbon chemical profiles and species absorbed (inset) for the direct current and active screen carburised material.

2.3.3. Potentiodynamic testing Potentiodynamic scans were performed on all the sample types using a Gamry Ref600 potentiostat and a 3-electrode configuration consisting of a SCE reference electrode, a platinum counter electrode and a working electrode (sample). The sample to be tested was placed against a teflon knife-edge seal leaving a theoretical working area of 452 mm2 on the sample surface in contact with the testing solution. All tests were conducted in Ringer’s solution, deaerated for 30 min with nitrogen at a flow rate of 150 cm3/ min, at a constant temperature of 37 °C ± 1 °C which mimics the body fluid. The OCP was left to stabilise for 1 h and this was followed by a potentiodynamic sweep at a scan rate of 1 mV/s at a voltage range of 100 mV (SCE) before the OCP to 1000 mV (SCE) vs. reference. Three repeats of each test were performed to prove repeatability of the results and no crevice corrosion was visible after testing. 3. Results

Fig. 3. XRD profile for as-treated (unpolished) and polished surfaces of direct current and active screen carburising treatments compared to the untreated 316LVM alloy. X-Ray source: Cu Ka radiation.

3.1. Material characterisation The first 5 lm of the layer, indicated by the line A–A in Fig. 2 was removed by polishing in order to make sure that only the carbon S-phase layer was tested. Fig. 3 shows a superimposition of XRD scans for treated samples in the polished and unpolished condition. The areas indicated by the arrows in Fig. 3, mark the location of a number of carbides. Due to the large amount of peaks

close to each other these could not be identified, however it is know that in that range of angles possible carbides include cementite (Fe3C) and Hägg-carbide (Fe5C2). For the polished samples of both treatment types a well defined peak without a trailing tail at a higher relative intensity when compared to the unpolished treated alloy was noted. The area previously marked with an arrow is now absent indicating that any possible carbides formed in the

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

surface had been removed and that only a carbon S-phase layer is present on the surface. The removal of 5 lm is essential in order to make sure that all the testing was being done on a precipitate free and contaminant free S-phase. Being a medical grade stainless steel, with possible biomedical applications, precipitates and contaminants are highly undesirable as they can be the cause of failure due to corrosion. All the results presented in this work are on treated and untreated samples that had 5 lm of their surface removed by polishing. This material removal can be easily performed by implant manufacturers using a controlled lapping technique after the carburising process. Fig. 2 shows the carbon depth profiles and the species absorbed for the treated 316LVM stainless steel. From these plots it is clear that more carbon was absorbed during the direct current treatment than during the active screen treatment. Utilising the untreated, polished austenitic structure as a benchmark, Fig. 3 shows a clear asymmetric shift to lower angles. This is the characteristic signature of S-phase and gives evidence that S-phase has been created in the surface of the 316LVM alloy. When comparing the active screen treated samples with those treated by the direct current treatment it is clear that the former has both a higher peak intensity and also a slightly greater peak shift to lower angles. The cross-section microstructure of the S-phase layer formed during direct current and active screen plasma carburising can be seen in Fig. 4(a) and (b), respectively. The layers appear to be bright and featureless and this is another indication that a precipitate free S-phase layer has been formed. The layer thickness was of 22 and

Fig. 4. Cross-sectional microstructure of the S-phase layer formed using (a) direct current and (b) active screen plasma carburising.

173

19 lm for the direct current and active screen treated alloy, respectively. The value quoted was a result of a five point average. The surface microhardness results obtained from the polished 316LVM stainless steel material in the untreated, direct current treated and active screen treated conditions are plotted in Fig. 5. Both active screen and direct current treated surfaces had similar surface hardness with an improvement of over than 260% when compared to the untreated alloy.

3.2. Pitting and crevice corrosion resistance 3.2.1. ASTM G48 Crevice corrosion resistance was evaluated by immersing a sample which was pressed against a crevice former with a fixed torque value, in an acidified ferric chloride solution for 72 h. The average weight loss for the untreated and treated alloy is plotted in Fig. 6. It can be noted that for both carburising treatments, in particular for the active screen treatment, the crevice corrosion resistance was increased drastically. In fact the weight loss for the direct current and active screen treated alloys was of 39% and 98%, respectively, less than for the untreated material. Clearly, active-screen plasma is more effective than DC plasma in improving the crevice corrosion resistance of 316LVM. The crevice corrosion resistance of the untreated material was found to be low with localised corrosion attack occurring under all the 12 ‘‘feet’’ of the crevice former. Fig. 7(a) shows the extent of the attack. The direct current treated samples suffered from a much lower degree of attack when compared to the untreated material. The attack on the sample can be seen in Fig. 7(b) where the crevices formed under the crevice former were very small, shallow and never on the entire surface. The active screen-treated sample had the best crevice corrosion resistance of all the three types of samples. In fact as it can be seen in Fig. 7(c) after the 72 h immersion test no crevices were visible on the surface. The surface remained polished and for most samples no material loss was experienced. The plot shown in Fig. 8 is the OCP measured throughout the test for the three sample conditions. The potential for the untreated material stabilised after 14 h from the starting point, keeping a steady potential value till the end of the 72 h. The curve for the direct current treated sample was of a decaying nature. An

Fig. 5. Surface micro-hardness for untreated, direct current and active screen plasma carburised AISI 316LVM. Error bar: largest deviation from the mean of 10 values.

174

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

Fig. 6. Weight loss for untreated, direct current and active screen plasma carburised AISI 316LVM following a 72 h crevice corrosion test (adapted ASTM G48-03) in acidified ferric chloride solution. Error bar: largest deviation from the mean of five values.

initial high potential showed good corrosion resistance, however as time elapsed, the value of the potential started to decrease. A higher negative gradient slope was experienced in the first 42 h and following this point the curve started to stabilise ending with a final OCP value of 196 mV (SCE). The excellent crevice corrosion resistance of the active screen treated alloy is also evidenced by the OCP plot. A high initial potential, very similar to that of the initial value of the direct current treated sample was recorded but as opposed to what happened for the direct current treated sample the potential increased with time, stabilizing in a short span of time. Fig. 9(a) and (b) shows the amount of chromium and nickel, respectively leached out during the 72 h test. The scatter shown in these graphs is asymmetric about the mean because it has been calculated from the largest deviation from the mean of five values. The two techniques, XRF and AAS, were compared to a gravimetric method of measurement and seem to agree. From these plots it is clear that both direct current and active screen treated samples showed a huge improvement in reducing the ion release. The highest improvement was for the active screen treated samples where the values of chromium and nickel leached were below the detection limit of both XRF (2 lg/ml) and AAS (0.2 lg/ml) equipment. Cross-sections of the post-corroded samples were produced on the three sample conditions. The crevices seen on the untreated material, Fig. 10(a), were approximately 200 lm deep and were consistent along the crevice former face. The crevice attack on the direct current treated sample, Fig. 10(b) shows an S-phase layer breach which then allowed further accelerated material attack under the layer. As it was explained before, no attack was seen on the active screen treated alloy and the etched cross-section microstructure shows no evidence of crevice attack (Fig. 10(c)). 3.2.2. Potentiodynamic tests Throughout this test which involved a one hour OCP test followed by a potentiodynamic tests in a more intra-human-like environment no crevice formers were used since only the pitting corrosion resistance was under scrutiny. The OCP results, not shown in this work, for all the three sample conditions were very similar. The values typically ranged between 285 mV (SCE) and 310 mV (SCE) for all the samples tested. As it can be seen in Fig. 11, the untreated material had an initial low passive current density, with evidence of metastable pitting at

Fig. 7. Macrographs of corroded (a) untreated, (b) direct current and (c) active screen plasma carburised AISI 316LVM sample after a 72 h corrosion test (ASTM G48-03) in acidified ferric chloride solution.

150 mV (SCE) and a pitting breakdown potential of 450 mV (SCE). The direct current and active screen treated samples behaved in

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

175

Fig. 8. Open circuit potential for the 72 h crevice corrosion test (ASTM G48-03) in acidified ferric chloride solution.

almost an identical manner. The scans show low passive current densities and no evidence of pitting. This means that both treatments have increased the pitting corrosion resistance of the 316LVM austenitic stainless steel in the intra-human-like environment. 3.2.3. Critical pitting potential – ASTM F746 Throughout this test which involved a 1 h OCP test followed by a series of simulation phases in an intra-human-like environment crevice formers were used since the aim of this test is to evaluate the resistance of an alloy to both pitting and crevice corrosion. Fig. 12 shows the ultimate and penultimate repassivation phase curves of a typical test on the untreated 316LVM material. These curves show that since the last repassivation phase curve at 350 mV (SCE) reached the current density limit, the one ahead of it, in this case 300 mV (SCE), is the critical pitting potential. Three repeated tests were performed and the average critical potential in deaerated Ringer’s solution was of 370 ± 80 mV (SCE). This means that any pit or crevice generated after this voltage would not be able to repassivate. On the other hand the direct current and active screen treated samples behaved differently from the untreated material. In the simulation phase, keeping the voltage at a constant 800 mV (SCE), both treated samples behaved in a similar manner with current density tending to zero in a short span of time (Fig. 12). This means that the treated alloys had very high pitting and crevice corrosion resistance. 4. Discussion 4.1. Untreated vs. treated From the results presented in this work it is clear that the creation of carbon S-phase in the surface of 316LVM augments the pitting and crevice corrosion resistance. However the reason why the S-phase is beneficial to the corrosion resistance has not yet been established with certainty. Pitting reactions on stainless steels are nucleated at manganese sulphide inclusion sites and although the removal of these secondary inclusions by the plasma process is possible this is not the likely cause for the improved crevice corrosion resistance.

Fig. 9. Amount of (a) chromium and (b) nickel leached out of the metal into the solutions after the 72 h crevice corrosion tests (ASTM G48-03) in acidified ferric chloride solution. Error bar: largest deviation from the mean of five values.

The sputtering cleaning effect for both the DC and active screen processes would occur at the very outside surface of the alloy. This effect, if any, would have been completely eliminated by the fact that the surface of the treated alloy had 5 lm polished off from the surface. Heuer et al. [7] have reported that the removal of MnS inclusions by electropolishing prior to the low temperature carburising treatment does not lead to the level of improvements achieved by interstitial hardening. In fact the carburising of 316L specimens both with and without removal of MnS inclusions resulted in similar increases in pitting potential. The biomedical austenitic stainless steel used in this study is governed by two standards ASTM F138 and BS 7252-1 which demand a very low sulphur content, the maximum value being 0.01 wt.%. These standards also dictate that the microcleanliness of the steel should be determined by Method A of Test Method ASTM E 45 and that the inclusion content for sulphides determined at the billet stage, from a billet not exceeding 15 cm thickness shall not exceed the 1.5 max limit for thin inclusion and the 1.0 max limit for heavy inclusions. The material certificate of our steel had a 10 times lower value of sulphur than that suggested by the standard and the sulphide inclusion content was zero for both thin and heavy inclusions. In order to produce a steel with these

176

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

Fig. 10. Cross-sections of (a) untreated, (b) direct current and (c) active screen plasma carburised corroded samples after the 72 h crevice corrosion test (ASTM G48-03) in acidified ferric chloride solution.

Fig. 11. Potentiodynamic curves for untreated, direct current and active screen plasma carburised AISI 316LVM in deaerated Ringer’s solution.

Fig. 12. Critical pitting potential curve (repassivation phase) for untreated, direct current and active screen plasma carburised AISI 316LVM specimens in deaerated Ringer’s solution.

required cleanliness, vacuum melting had to be employed by the steel manufacturer. This would imply that the influence of MnS inclusions on the results would be very minimal. The importance of interstitial content towards localised corrosion resistance is further highlighted by the fact that in High-N stainless steels (0.25–0.5 wt.% N) governed by the standards ASTM

F1586 allow a higher heavy sulphide inclusion content (1.5) than ASTM F138. In fact the High-N alloys do not have to be produce by vacuum melting but by electroslag remelting. This lower restriction in inclusion content is due to the interstitial nitrogen which gives the extra localised corrosion resistance to the alloy. The work by Ningshen et al. [21] highlights the beneficial effect

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

of nitrogen addition on the pitting resistance in the presence of chloride ions. From their polarisation and Mott–Schottky plots a correlation between the semiconducting nature of the passive film and pitting corrosion resistance was observed. The widely accepted theory related to S-phase on the improvement of pitting and crevice corrosion is not related to carbon but to nitrogen. The most common mechanism, given by JargeliusPettersson [16], is the formation of ammonia inside the pit that neutralises the acidic solution which is dissolving the metal inside the pit. The validity of this theory is directly dependent on the presence of nitrogen interstitials. This is not the case for this study as only carbon was added through a carbon-bearing gas in the treatment chamber. Furthermore, to ensure that no nitrogen contamination was present, the top 5 lm of the layer had been removed by polishing. This is because nitrogen S-phase would form near the surface should nitrogen exist in the gas mixture for plasma carburising [10]. One theory proposed by Martin et al. [5], applicable to both carbon and nitrogen interstitials, is based on oxygen vacancies. The interstitials residing in the oxygen vacancy sites of the passive film have the ability to move. Oxygen vacancy is the key factor which dictates charge transport within the passive film. Due to the mobility of the interstitials in the oxygen vacancies being either carbon or nitrogen, the passive film becomes less resistive. Consequently the nitrogen or carbon interstitials keep the oxygen vacancy in motion and thus slow down pit formation and subsequent corrosion [5]. Heuer et al. [7,8] proposed a chemomechanical model that is also applicable to both carbon and nitrogen interstitials. This model explains the loss of corrosion resistance by a mechanical failure of the passive film. They suggest that the stress in the passive oxide film, at a critical passive film thickness (3 nm for 0.6 M NaCl solutions), induces the formation of surface undulations which develop as the passive layer undergoes simultaneous growth due to anodic oxidation and dissolution in chloride containing solutions. Pitting occurs when the amplitude of the undulations approaches the passive layer thickness, and under the applied potential, dielectric breakdown can occur in the thinnest regions of the film. In their work utilising very similar treatment conditions to ours, Heuer et al. [7,8] noticed that the carburised specimens always had a thinner passive film along the applied potential range when compared to untreated 316L. This implied that as the film thickness increased with applied potential the untreated alloy always reached the critical thickness before the treated sample. Therefore the critical potential for pitting corrosion is always lower for the untreated material rather than the carburised. In agreement with Heuer et al. [7] that the levels of localised corrosion improvement could only be attributed to the high amount of interstitial carbon and not to some other phenomenon together with the incompatibility of Jargelius-Pettersson’s [16] theory to explain the carbon interstitial effect, the theories outlined by Martin et al. [5] and Heuer et al. [7,8] were found to be the most plausible. The carbon supersaturation of medical grade stainless steel showed positive corrosion resistance results giving way to the belief that these two theories are plausible. 4.2. Active screen plasma vs. direct current plasma The pitting and crevice corrosion resistance of the biomedical 316LVM alloy has been augmented with both low temperature plasma carburising treatments. In vitro corrosion tests, Figs. 11 and 12, conducted in Ringer’s solution have revealed that both treatments created a surface that was highly corrosion resistant. On the other hand these tests could not be used to identify whether direct current or active screen plasma carburising was the superior treatment. The ranking of the two treatments was possible with the use of a more aggressive test (ASTM G48)

177

conducted in acidified ferric chloride solution. All the results and evidence obtained from the G48 test leads to the conclusion that the S-phase layer formed by active screen plasma carburising had a much superior crevice corrosion resistance when compared to the direct current treatment. When comparing results the confidence level is a very important factor. The large amounts of repeats, the utilisation of different techniques to confirm a result together with the excellent specimen-to-specimen reproducibility increases our level of confidence in our results. According to Speidel [22] carbon has a very good influence on the pitting and crevice corrosion resistance of stainless steel. The ‘‘Measure of Alloying for Resistance to Corrosion’’ (MARC = % Cr + 3.3Mo + 20N + 20C 0.5Mn 0.25Ni) can be used in order to rank different alloys. Using the values of carbon wt.% at 5 lm it is possible to compare the two treatments. The wt.% of carbon after the direct current treatment was of 2.4 wt.% and, when applied to the MARC equation, gives a value of 73. Similarly the wt.% of carbon after the active screen treatment was of 2.1 wt.% which gave a value of 67 when applied to the MARC equation. The values obtained should rank the direct current treated surface as the best, followed by the active screen treated surface. This implies that the alloying criterion is not the single factor that affects crevice corrosion resistance. The work by Sun [6] has indicated that a minimum of 0.25 wt.% carbon is required for improved pitting corrosion resistance. Therefore one would assume that the higher the carbon wt.% at the surface would be better. Our work is indicating otherwise probably due to some other effect that would influence the localised corrosion resistance of the stainless steel. It is very well known that active screen plasma technology has reduced the problems that were associated with direct current plasma treatments. As explained by Corujeira Gallo and Dong [14] the bias in the active screen set-up has a direct effect on the corrosion resistance properties of active screen treatments. Low bias gives a uniform surface with predictable corrosion properties, whilst increasing bias leads to a direct current approach with uneven corrosion surface properties further leading to edge effect defects. In fact the limitations such as hollow cathode, arcing and edge effect synonymous with direct current plasma treatments have been greatly reduced [12]. During this study special care was taken to diminish the deleterious effects associated with direct current treatments and no direct evidence of these effects could be seen on the specimen. Also as explained previously 5 lm from the top layer of both DC and active screen treated specimens were polished off. The deleterious effects mentioned before would occur at the very close surface therefore the removal of the top layer from both samples should place both treatments on par-levels. In direct current treatments the specimens are heated directly by the plasma formed on them whilst in active screen the specimens are heated by the plasma that is formed on the cage. This makes the active screen treatment a more controllable process in terms of temperature. Also the mechanisms of charge carriers in active screen is different to that of direct current [12]. Our speculation is that the differences between the treatments lies in either a very slight temperature difference due to the nature of the heating or due to a different carbon potential in the near surface of the specimens. Both factors would affect the amount of diffused carbon into the surface of the stainless steel. The XRD plots (Fig. 3) together with the cross-sectional micrographs (Fig. 4) give evidence that a precipitate free layer has been created by both treatments. The only slight difference between the two treatments lies in the fact that the peaks of the active screen treated samples are slightly shifted to lower angles and have a higher intensity when compared to those of the direct current treated specimens. Li et al. [23] have shown that the relaxation of stresses within the S-phase layer result in shifting of peaks to slightly higher angles and this is attributed to the precipitation of

178

J. Buhagiar et al. / Corrosion Science 59 (2012) 169–178

carbides within the layer. This precipitation theory is strengthened further by the fact that the relative peak intensities for the direct current treated alloys are lower than those of active screen. These precipitated carbides are in very small volumes and cannot be detected by the XRD and it is hypothesised that these carbides might be one of the reasons for the discrepancy in crevice corrosion resistance noted. Summarising the above arguments it is therefore hypothesised that the direct current treatment introduced a higher amount of carbon in the material either due to a slightly higher temperature or due to a more efficient carbon species delivery process. This resulted in higher carbon species absorbed and thicker layer thicknesses. This higher amount of carbon eventually resulted in some minor chromium carbide precipitations within the layer. These small precipitates which cannot be detected by XRD would eventually result in poorer localised corrosion resistant properties when compared to the active screen specimens. 5. Conclusions (1) The medical grade 316LVM stainless steel could be effectively low temperature carburised using both active screen and direct current treatments. The creation of a precipitate free carbon S-phase on the surface of 316LVM augments the pitting and crevice corrosion resistance of the untreated alloy. Also the carbon S-phase generated using an active screen process technique, has a superior crevice corrosion resistance to that produced using the direct current treatment. (2) The difference in crevice corrosion resistance between active screen and direct current treated samples cannot be related to carbon content but might be due to the deleterious effects that direct current treatments are synonymous with and active screen treatments are not. XRD has indirectly revealed that carbides in small volumes might have also precipitated in the direct current treated samples. These carbides which might not be present after active screen treatment can reduce the crevice corrosion resistance of the treated alloy. (3) The higher pitting resistance of the treated surfaces opposed to the untreated material in Ringer’s solution was experienced by the difference in the critical potential that was calculated to be of 370 mV (SCE) for untreated to over 800 mV (SCE) for the treated specimens. Similarly potentiodynamic sweeps for untreated AISI 316LVM in Ringer’s solution showed pitting at 450 mV (SCE) whilst no pitting was seen up to a value of 1000 mV (SCE) for both treated samples.

Acknowledgements The authors would like to thank the University of Malta Research Fund Committee for the financial support and ERDF (Malta) for the financing of the testing equipment through the project: ‘‘Developing an Interdisciplinary Material Testing and Rapid Prototyping R&D Facility (Ref. No. 012)’’ and ‘‘ Strengthening of Analytical Chemistry, Biomedical Engineering and Electromagnetics RTDI Facilities (Ref. No. 018)’’. One of the authors, André Spiteri, wants to thank the KSU student opportunity fund and the student affairs

committee for funding his research trip to the University of Birmingham. In addition, the authors wish to express their appreciation to Dr. X.Y. Li, Dr. J. Chen, Mr. T. Bell, Dr. W. Li, Dr. J.C. Betts, Mr. D. Vella, Mr. J. Camilleri, Ing. M. Fenech, Mr. A. Agius, Mr. S. Zerafa, Mr. C. D’Amato and Mr. J. Grech. References [1] J. Buhagiar, T. Bell, R. Sammons, H. Dong, Evaluation of the biocompatibility of S-phase layers on medical grade austenitic stainless steels, J. Mater. Sci.-Mater. Med. 22 (2011) 1269–1278. [2] D.F. Williams, Medical and dental materials, in: R.W. Cahn, P. Haasen, E.J. Kramer (Eds.), Materials Science and Technology – A Comprehensive Treatment, VCH, Cambridge, UK, 1992. [3] U. Kamachi Mudali, T.M. Sridhar, N. Eliaz, Badev Raj, Failures of stainless steel orthopaedic devices – causes and remedies, Corros. Rev. 21 (2003) 231–267. [4] M.K. Lei, X.M. Zhu, In vitro corrosion resistance of plasma source ion nitrided austenitic stainless steels, Biomaterials 22 (2001) 641–647. [5] F.J. Martin, E.J. Lemieux, T.M. Newbauer, R.A. Bayles, P.M. Natishan, H.K.G.M. Michal, F. Ernst, A.H. Heuer, Carburisation induced passivity of 316L austenitic stainless steel, Electrochem. Solid State Lett. 10 (2007) C76–C78. [6] Y. Sun, Corrosion behaviour of low temperature plasma carburised 316L stainless steel in chloride containing solutions, Corros. Sci. 52 (2010) 2661– 2670. [7] A.H. Heuer, H. Kahn, F. Ernst, G.M. Michal, D.B. Hovis, R.J. Rayne, F.J. Martin, P.M. Natishan, Enhanced corrosion resistance of interstitially hardened stainless steel: implications of a critical passive layer thickness for breakdown, Acta Mater. 60 (2012) 716–725. [8] A.H. Heuer, H. Kahn, P.M. Natishan, F.J. Martin, L.E. Cross, Electrostrictive stresses and breakdown of thin passive films on stainless steel, Electrochim. Acta 58 (2011) 157–160. [9] F.J. Martin, P.M. Natishan, E.J. Lemieux, T.M. Newbauer, R.J. Rayne, R.A. Bayles, H. Kahn, G.M. Michal, F. Ernst, A.H. Heuer, Enhanced corrosion resistance of stainless steel carburized at low temperature, Metall. Mater. Trans. 40A (2009) 1805–1810. [10] H. Dong, S-phase surface engineering of Fe–Cr, Co–Cr and Ni–Cr alloys, Int. Mater. Rev. 55 (2010) 65–98. [11] Y. Sun, E. Haruman, Tribocorrosion behaviour of low temperature plasma carburised 316L stainless steel in 0.5 M NaCl solution, Corros. Sci. 53 (2011) 4131–4140. [12] S. Corujeira Gallo, H. Dong, Study of active screen plasma processing conditions for carburising and nitriding austenitic stainless steel, Surf. Coat. Technol. 203 (2009) 3669–3675. [13] J. Georges, D. Cleugh, Active screen plasma nitriding, in: T. Bell, K. Akamatsu (Eds.), Stainless Steel 2000: Thermochemical Surface Engineering of Stainless Steel, Maney, Osaka, Japan, 2000, pp. 377–387. [14] S. Corujeira Gallo, H. Dong, Corrosion behaviour of direct current and active screen plasma carburised AISI316 stainless steel in boiling sulphuric acid solutions, Corros. Eng. Sci. Technol. 46 (2011) 8–16. [15] J. García Molleja, L. Nosei, J. Ferrón, E. Bemporad, J. Lesage, D. Chicot, J. Feugeas, Characterization of expanded austenite developed on AISI 316L stainless steel by plasma carburization, Surf. Coat. Technol. 204 (2010) 3750–3759. [16] R.F.A. Jargelius-Pettersson, Electrochemical investigation of the influence of nitrogen alloying on pitting corrosion of austenitic stainless steels, Corros. Sci. 41 (1999) 1639–1664. [17] F.J. Martin, K.E. Lucas, E.A. Hogan, Experimental procedure for crevice corrosion studies of Ni–Cr–Mo alloys in natural seawater, Rev. Sci. Instrum. 73 (2002) 1273–1276. [18] R. Sharghi-Moshtaghin, H. Kahn, Y. Ge, X. Gu, F.J. Martin, P.M. Natishan, R.J. Rayne, G.M. Michal, F. Ernst, A.H. Heuer, Low-temperature carburization of the Ni-base superalloy IN718: improvements in surface hardness and crevice corrosion resistance, Metall. Mater. Trans. 41A (2010) 2022–2032. [19] X.Y. Li, J. Buhagiar, H. Dong, Characterisation of dual S phase layer on plasma carbonitrided biomedical austenitic stainless steels, Surf. Eng. 26 (2010) 67– 73. [20] G.F. Vander Voort, Metallography: Principles and Practice, ASM International, Ohio, 1999. [21] S. Ningshen, U. Kamachi Mudali, V.K. Mittal, H.S. Khatak, Semiconducting and passive film properties of nitrogen-containing type 316LN stainless steels, Corros. Sci. 49 (2007) 481–496. [22] M.O. Speidel, Nitrogen containing austenitic stainless steels, Mat-wiss. u. Werkstoiftech 37 (2006). [23] X.Y. Li, S. Thaiwatthana, H. Dong, T. Bell, Thermal stability of carbon S phase in 316 stainless steel, Surf. Eng. 18 (2002) 448–452.