Sliding wear corrosion of ceramics

Sliding wear corrosion of ceramics

Wear 267 (2009) 599–607 Contents lists available at ScienceDirect Wear journal homepage: www.elsevier.com/locate/wear Sliding wear corrosion of cer...

4MB Sizes 8 Downloads 100 Views

Wear 267 (2009) 599–607

Contents lists available at ScienceDirect

Wear journal homepage: www.elsevier.com/locate/wear

Sliding wear corrosion of ceramics A.J. Gant, M.G. Gee ∗ National Physical Laboratory, Hampton Road, Teddington, Middlesex, TW11 0LW, United Kingdom

a r t i c l e

i n f o

Article history: Received 31 October 2008 Received in revised form 19 February 2009 Accepted 19 February 2009 Keywords: Wear Corrosion Synergy Ceramics

a b s t r a c t This paper describes experiments to examine how sliding wear corrosion experiments can be carried out with simple laboratory equipment that is in common usage in tribological testing laboratories. Experiments were conducted on three ceramics; two aluminas with different alumina volume fractions and one tetragonally stabilised zirconia. Three different test fluids were examined; these were de-ionised water, 0.1 M HCl and 0.1 M NaOH. It was found that the sliding wear corrosion test system performed well, with the two aluminas both exhibiting increasing wear with decreasing pH, and the zirconia showing decreasing wear with decreasing pH. Crown Copyright © 2009 Published by Elsevier B.V. All rights reserved.

1. Introduction

2. Experimental details

In recent years there has been work published detailing a number of aspects of the tribological behaviour of modern advanced technical ceramics. Work has been reported by the research community on sliding, abrasive and erosive wear of technical ceramics [1–28]. It is now well known that the sliding wear of ceramics is affected by tribochemical reaction at the contacting surface [4–6,12,13,19]. Here reaction of the ceramic with the environment (atmosphere or fluid) causes the formation of reaction products that often form protective films on the surface reducing friction and wear. Apart from the work from the group of Valin and Novak [14,16,20,23,25,26], there has been little other work in the important technological area of sliding wear corrosion of ceramics. It was therefore decided to carry out some experiments looking at sliding wear corrosion of ceramics. The three ceramics that were used were chosen for several reasons. Firstly, to give a generic flavour to the work, three oxide ceramics were chosen. The three ceramics differ in applications and composition; Vitox is primarily intended for surgical implants, whereas the two other grades are commonly used in industrial application, Sintox FC being particularly intended to offer resistance to both chemical attack and wear; this latter grade would therefore seem particularly appropriate to subject to the conditions outlined in this paper.

2.1. Materials

∗ Corresponding author. Tel.: +44 2089436374. E-mail address: [email protected] (M.G. Gee).

Three different ceramics were used as the flat samples (Table 1). These were prepared by: (a) Annealing at 1100 ◦ C for 1 h in air. (b) Mounting in a low modulus hardenable two-part resin supplied by Metprep Ltd, Coventry. (c) Grinding on a fixed diamond abrasive pad (10 ␮m; metal bonded). (d) Grinding on a fixed diamond abrasive pad (30 ␮m; resin bonded). (e) Polishing using a nylon cloth in conjunction with a 6 ␮m diamond suspension. The balls were standard zirconia ball bearings supplied by Atlas. Three test fluids were used. These were de-ionised water (W), 0.1 M HCL (H), and 0.1 M NaOH (N). 2.2. Test method The test system that was employed was a Phoenix Tribology TE77 reciprocating test system (Fig. 1). This was modified so that the wear corrosion experiments could be carried out by coating the flat sample holder with non-conducting lacquer, and replacing the sample clamping screws with plastic bolts. A tufnol (insulating polymeric composite) plate was placed beneath the test samples to give a firm, but compliant, non-conductive base to the samples. The sample clamps themselves were also coated with lacquer to

0043-1648/$ – see front matter. Crown Copyright © 2009 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2009.02.015

600

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

Table 1 Materials examined. Material

Manufacturer

Composition

Grain size (␮m)

Density (g cm−3 )

Hardness

Balls Zirconia

Atlas Ball and Bearing Co Ltd

97% tetragonal zirconia with 3% Yttria



6.05

1200 HV

Flats Sintox FC (FC) Vitox (V) Technox 3000 (T)

Morgan Advanced Ceramics Ltd Morgan Advanced Ceramics Ltd Dynamic Ceramic Ltd

95% alumina, remainder MgO, CaO, SiO2 99.9% alumina, trace sliica 97% tetragonal zirconia with 3% Yttria

4.0 1.2 1.0

3.72 3.98 6.05

1350 HV1 2000 HV1 1350 HV0.3

through the formula V = m/ where V is the volume loss, m is the mass loss, and  is the density. The mass of the zirconia balls was not measured as it was incorrectly felt that the balls would remain inert during the testing. The width of the wear track on the flat was measured using optical microscopy, taking the average of four measurements as the width. After it was realised that the zirconia balls did not remain inert in the test, an attempt was made to take micrographs of the scars on the balls and to measure their size optically. This proved to be extremely difficult, and scars were found on only four of the ball samples. The scar size on these balls was measured as the average of the width of the scratch parallel and perpendicular to the direction of sliding. 3. Results

Fig. 1. Schematic diagram of reciprocating test system. The reciprocating motion is generated by a scotch yoke arrangement.

prevent metal contact with the test fluid. The ball holder elements were coated with lacquer for the same reason. A new ball was used for each test. The tests that were carried out are listed in Table 2. The stroke that was used in all tests was 4.7 mm, and the test frequency was 2 Hz. The test duration was 10,000 cycles. Tests 12, 13 and 14 were null tests where the ball sample was not loaded against the flat, but was nevertheless reciprocated so that the fluid motion was similar to the loaded tests. The purpose of these tests was to evaluate the corrosion that occurred to the materials from immersion in the fluids alone. The flat samples were cleaned in acetone, dried, and were weighed in an electronic balance before and after testing. Worn surfaces were examined using a Nikon optical measurement microscope, and selected samples were also examined in a Zeiss Supra field emission SEM. The mass loss was calculated to a volume loss Table 2 Tests performed (FC is Sintox FC, V is Vitox, and T is Technox 3000). Test

Flat material

Load (N)

Test fluid

1 2 3 4 5 6 7 8 9 10 11 12 13 14

FC V FC T FC V T T FC V T FC, V and T FC, V and D FC, V and D

20 60 60 60 60 60 60 60 60 60 60 0 0 0

De-ionised water De-ionised water De-ionised water De-ionised water 0.1 M HCl 0.1 M HCl 0.1 M HCl 0.1 M HCl 0.1 M NaOH 0.1 M NaOH 0.1 M NaOH 0.1 M NaOH 0.1 M HCl De-ionised water

Comments

The results of the flat volume loss measurements are given in Table 3. It can be seen that the wear of the Sintox FC and the Vitox both increased as the pH of the test fluid decreased. By comparison, the Technox 3000 wear decreased with decreasing pH. None of the samples tested in the null experiments showed any significant change in mass. The optical micrographs of the wear scars are shown in Figs. 2–4. For both the Sintox FC and Vitox in water the wear scar surface was smoother than the surface of the surrounding area where some grain pluck-out had occurred during the preparation procedure (Fig. 2). The overall impression was that some of the holes created by grain pluck-out had been filled during the wear process. At somewhat higher magnification short grooves were visible that traversed only a few grains of the material. By contrast the Technox 3000 sample tested in water showed a very wide wear track with considerable disruption and damage to the structure of the material. Areas of pull-out of material were evident, together with areas with multiple tears to the surface layers of the material. Fig. 3 shows the wear tracks for the samples tested in 0.1 M HCl. Here the wear track on the Sintox FC sample was visibly deep, with a more disrupted smeared surface than for the sample tested in water. The Vitox sample also showed a wear track that appeared by eye to be deeper the wear track for the sample tested in water. The Technox 3000 sample showed a small wear track relative to the sample tested in water, with layers of material at the base of the wear track that showed light interference patterns when examined at an intermediate magnification. Fig. 4 shows the wear tracks for the samples tested at 0.1 M NaOH. The wear track for the Sintox FC was smaller than for the sample tested in water, with more damage still left from the preparation than for the sample tested in water. The Vitox only showed

Sample slipped Table 3 Flat volume (mm3 ) loss for tests at 60 N.

Corrosion test Corrosion test Corrosion test

NaOH Water HCl

Sintox FC

Vitox

Technox 3000

−0.003 0.696 3.167

0.010 0.025 0.131

0.240 0.198 0.015

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

601

Fig. 2. Optical micrographs of surfaces of samples exposed to sliding wear-corrosion in water, (a) and (b) Sintox FC, (c) and (d) Vitox, (e) and (f) Technox 3000. A marks areas of pluck-out, B marks grooving, and C “tearing” in the zirconia.

a small amount of damage, but examination of the end of the wear scar showed a build up of debris. The Technox 3000 sample showed disruption to the structure of the wear surface similar to the sample tested in water. Examination of the worn samples in the SEM showed more detail of the mechanisms that had occurred. In Fig. 5a–c it can be seen that the apparent reduction in porosity in the surface of the wear track is because the surface layers of the sample have been polished revealing the underlying two phase structure of the material. Here the second phase gives a lighter contrast than the main alumina phase because of the presence of elements with higher atomic number such as calcium. Some short grooving is also visible in these micrographs, and it is clear that this cracking often stops at specific grain boundaries. Some cracking is also evident between the alumina grains, and there is also damage to some of the second phase regions. The Vitox sample tested in water showed little damage apart from some short grooving. The Technox 3000 sample tested in water showed a highly disrupted surface (Fig. 5e and f). The tear like features observed optically were shown to be areas where material had been removed form the surface of the sample, with particulate material at the base of the groove that was created in this process. The surface of the Sintox FC surface tested in acid was covered by layers of material that had a plate-like structure with the plates separated by considerable cracking (Fig. 6a–c). Considerable small scale cracking structure was also visible within these plates in bands parallel to the wear direction (Fig. 6c). The worn surface of the Vitox sample was covered with a large number of very small cracks (Fig. 6d). SEM examination of the worn surface of the Technox showed very little damage at high magnification (Fig. 6e). The worn surface of the Sintox FC sample tested in alkali showed little change compared to the original prepared surface (Fig. 7).

Table 4 gives the values of track width and pin scar width. The size of the pin wear scars agreed with the measurements of wear track widths within the likely error of measurement. However, for the tests on the two alumina materials, the simple picture of increasing wear with decreasing pH as shown by mass changes was not found. For both of these flat materials the largest width for both aluminas was found for the tests in water, with a reduction in width for the HCl tests. The track width results for the Technox flat were consistent with the mass loss results. For the measured wear scars on the pins, the ratios of the wear scar diameters parallel and perpendicular to the sliding direction were also calculated. If this parameter is close to 1 (the scar is circular) it shows that little depth of wear has occurred to the flat sample relative to the depth of wear that has taken place to the ball. If this parameter deviates from 1 (the scar is not a circle), this shows that a significant depth of wear has taken place to the flat as well as the ball. Fig. 8 shows the micrographs of the wear scars that could be found. Both of the wear scars on the zirconia balls tested against the alumina samples in water were flat and showed similar features, with the appearance of layers of material over the edges of the sample in the sliding direction (Fig. 8a and b). The zirconia pin tested against the Vitox flat in acid was relatively smooth, but had some small deposits of material on the surface (Fig. 8c). The zirconia pin tested against the Technox flat in alkali showed a relatively rough surface with some obvious surface damage (Fig. 8d). 4. Discussion The test protocol that has been developed is clearly successful for the examination of the sliding wear corrosion. The choice of test conditions for successful experiments needs to be carefully

602

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

Fig. 3. Optical micrographs of surfaces of samples exposed to sliding wear-corrosion in 0.1 M HCl, (a) and (b) Sintox FC, (c) and (d) Vitox, (e) and (f) Technox 3000.

Table 4 Track width and pin scar size measurements. Material

Fluid

Track width (mm)

Sintox Sintox Sintox Vitox Vitox Vitox Technox Technox Technox

NaOH Water HCl NaOH Water HCl NaOH Water HCl

0.62 1.10 0.51 0.28 1.05 0.54 2.46 2.13 0.62

Pin Scar Diameter (mm)

Ratio of wear scar diameter parallel and perpendicular to sliding direction

1.09

1.00

1.11 0.67 2.73

0.99 0.97 0.95

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

603

Fig. 4. Optical micrographs of surfaces of samples exposed to sliding wear-corrosion in 0.1 M NaOH, (a) and (b) Sintox FC, (c), (d) and (e) Vitox, (f) and (g) Technox 3000. Arrows mark “tears” in material.

604

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

Fig. 5. Worn surfaces of samples tested by wear-corrosion in water examined by SEM, (a), (b) and (c) Sintox FC, (d) Vitox and (e) and (f) Technox 3000. A marks pluck out, B marks grooving, C marks gross material removal.

managed so that the balance between wear and corrosion is a realistic simulation of the industrial application. If the load and speed are too high, then wear can become very large so that corrosion does not contribute markedly to the overall material removal. If the load and speed are too low then the test times can become very large. This does introduce the consideration of the design of wear corrosion experiments. The overt introduction of the extra variables due to the fluid environment make an already complex test even more complex. One issue that needs particular care is the need to accelerate corrosion in many cases. This is a reasonable need, and can be successfully carried out as long as care is taken in the control of variables. An important check to take is to ensure that the mechanisms that are obtained in the laboratory sliding wear corrosion test are the same as those seen in the industrial application by examining both with techniques such as microscopy and chemical analysis, and to ensure that the magnitude of the damage is similar. The relative response of the three different materials relative to the different test fluids is informative. The two aluminas both show

a sensitivity to pH with increasing wear with decreasing (more acidic) pH. However, the purer alumina wears considerable less in the water and acid tests than the 95% alumina. Both aluminas wear almost insignificantly in the tests presented in this paper in the alkali fluid. The Technox 3000 material may be de-stabilised by the presence of the water and the high temperature generated at the wear interface. This may explain the rapid wear of this material in water and in alkali conditions. When wear and corrosion act together to cause a higher wear loss than either of the two phenomena together, the term synergy describes this amplification. This relationship is described by the formulaV = W + C + Swhere V is the volume of material loss, W is the contribution to material loss from wear alone, C is the contribution to volume loss from corrosion alone, and S is the contribution to material wear from the combined action of wear and corrosion. Since the contribution of corrosion alone, C, was found to be negligible, this formula can be rewritten as V =W +S

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

Fig. 6. Worn surfaces of samples tested by wear-corrosion in 0.1 M HCl examined by SEM, (a), (b) and (c) Sintox FC, (d) Vitox , (e) Technox 3000.

Fig. 7. Worn surfaces of Sintox FC sample tested by wear-corrosion in alkali examined by SEM.

605

606

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

Fig. 8. Wear scars on zirconia balls tested against, (a) Vitox tested in water, (b) Sintox tested in water, (c) Vitox tested in acid, (d) Technox tested in alkali.

It would normally be expected that the contribution to material loss, W, should be constant even if the test fluid and thus the pH is altered. Since for any of the materials that were tested the volume of wear varied considerably with the test solution it is therefore clear that there is a major variation in the contribution of synergy to the total material loss as the test fluid is changed. Thus the material loss increased from 0 to 3 mm3 for the Sintox, increased from 0.01 to 0.13 mm3 for the Vitox, and decreased from 0.24 to 0.015 mm3 as the pH decreased. However, without an independent measure of the action of wear alone, the contribution of synergy cannot be fully quantified. The reason for the increase in material loss in corrosive fluids is not completely clear from the experiments presented in this paper, but there are a number of possible mechanisms. These include the attack of second phases or grain boundaries by the corrosive media, undermining the mechanical strength of the material so that wear occurs more readily. It is known that alumina can be attacked by acids, particularly at the elevated temperatures generated in the wear interface [29]. Another likely mechanism is chemical reaction of the materials with the fluid to give softer surface layers that are more easily removed by the mechanical action that results in wear. Stabilised zirconia can undergo phase transformation in the presence of aqueous fluids that is enhanced by elevated temperatures [30,31]. Kalin and Novak [20,25] found that under acidic conditions the friction coefficient for sliding wear experiments on zirconia was half or less the value found for neutral or alkaline conditions. This is possibly due because hydroxide films formed on zirconia in acidic conditions are maybe more stable than those formed in alkaline conditions [32]. This would lead to the higher friction coefficients observed under neutral and alkali conditions, in turn giving a higher propensity to hydrothermal phase transformation through the higher temperatures generated at the wear

interface under these conditions. The observations given in this paper a fully consistent with this interpretation and the observations of Kalin and Novak [20,25]. It is significant that the wear track width results appear to be at odds with the flat mass change results. The reason for this apparent discrepancy must be that the zirconia pins are not inert in these experiments as was originally believed when the experiments were started. In the tests of the zirconia balls against the alumina flats in water it is clear that considerable wear took place to the zirconia balls, leading to the wide wear tracks, but still with little wear to the alumina materials. This observation reinforces the point, initially missed in this investigation, that it is always important to consider all of the different materials that are involved in a wear contact. Wear is always dependent on the many aspects that are involved in wear and friction as a mechanical system, as originally described by Czichos [33]. 5. Conclusions A test protocol has been developed that makes use of common tribological testing equipment to enable sliding wear corrosion experiments to be made. Three ceramics were tested to validate the test system. It was found that the two aluminas tested both gave increased wear with decreasing pH, with the pure alumina wearing much less than the 95% alumina. The zirconia ceramic showed the opposite behaviour with decreasing wear with decreasing pH. An inconsistency between mass change measurements for the wear flats and wear track width measurements showed that the zirconia balls also wore in the experiments, showing the importance of adopting a systems approach to wear and considering all the contacting surfaces in the wear system.

A.J. Gant, M.G. Gee / Wear 267 (2009) 599–607

Acknowledgement The authors would like to acknowledge support from the DTI Materials Performance Programme for this work. References [1] A.G. Evans, D.B. Marshall, Wear Mechanisms in Ceramics, in: D.A. Rigney (Ed.), Fundamentals of Friction and Wear of Materials, ASM Metals, Pittsburgh; Pa, 1981, pp. 439–452. [2] D.H. Buckley, K. Miyoshi, Friction and wear of ceramics, Wear 100 (December) (1984) 333–353. [3] R.W. Rice, Micromechanics of microstructural aspects of ceramic wear, Ceramic Engineering and Science Proceedings 6 (7–8) (1985) 940–958. [4] M.G. Gee, The formation of aluminum hydroxide in the sliding wear of alumina, Wear 153 (1) (1992) 201–227. [5] K.H. Lee, K.W. Kim, Effects of humidity and sliding speed on the wear properties of Si3 N4 ceramics, Materials Science and Engineering A: Structural Materials Properties Microstructure and Processing 186 (1–2) (1994) 185–191. [6] M.G. Gee, N.M. Jennett, High resolution characterisation of tribochemical films on alumina, Wear 193 (2) (1996) 133–145. [7] S.M. Hsu, M.C. Shen, Ceramic wear maps, Wear 200 (1–2) (1996) 154–175. [8] A. Krell, D. Klaffke, Effects of grain size and humidity on fretting wear in finegrained alumina, Al2 O3 /TiC, and zirconia, Journal of the American Ceramic Society 79 (5) (1996) 1139–1146. [9] H.Y. Liu, S.M. Hsu, Modeling of microfracture-induced wear and wear transition in sliding of polycrystalline alumina ceramics, Wear 195 (1–2) (1996) 169–177. [10] Y.S. Wang, S.M. Hsu, Wear and wear transition modeling of ceramics, Wear 195 (1–2) (1996) 35–46. [11] Y.S. Wang, S.M. Hsu, Wear and wear transition mechanisms of ceramics, Wear 195 (1–2) (1996) 112–122. [12] R. Wasche, D. Klaffke, In situ formation of tribologically effective oxide interfaces in SiC-based ceramics during dry oscillating sliding, Tribology Letters 5 (2–3) (1998) 173–190. [13] D. Klaffke, M. Hartelt, F. Koesling, Tribological characterisation of ceramic materials—Friction and wear under reciprocating sliding motion, Materialprufung 41 (7–8) (1999) 286–293. [14] S. Novak, M. Kalin, T. Kosmac, Chemical aspects of wear of alumina ceramics, Wear 250 (1–12) (2001) 318–321.

607

[15] R. Wasche, D. Klaffke, Wear of multiphase SiC based ceramic composites containing free carbon, Wear 249 (3–4) (2001) 220–228. [16] M. Kalin, S. Novak, J. Vizintin, Wear and friction behavior of alumina ceramics in aqueous solutions with different pH, Wear 254 (11) (2003) 1141–1146. [17] A.J. Gant, M.G. Gee, A.T. May, The evaluation of tribo-corrosion synergy for WC/Co hardmetals in low stress abrasion, Wear 256 (2004) 500–516. [18] A.J. Gant, M.G. Gee, A.T. May, Microabrasion of WC-Co hardmetals in corrosive media, Wear (2004) 954–962. [19] R.S. Gates, S.M. Hsu, Tribochemistry between water and Si3 N4 and SiC: induction time analysis, Tribology Letters 17 (3) (2004) 399–407. [20] S. Novak, M. Kalin, The effect of pH on the wear of water-lubricated alumina and zirconia ceramics, Tribology Letters 17 (4) (2004) 727–732. [21] R. Wasche, D. Klaffke, T. Troczynski, Tribological performance of SiC and TiB2 against SiC and Al2 O3 at low sliding speeds, Wear 256 (7–8) (2004) 695–704. [22] U. Dulias, K.H. Zum Gahr, Investigation of Al2 O3 and SSiC-ceramic under lubricated, reciprocating sliding contact and cavitation erosion, Materialwissenschaft Und Werkstofftechnik 36 (3–4) (2005) 140–147. [23] S. Novak, G. Drazic, M. Kalin, Structural changes in ZrO2 ceramics during sliding under various environments, Wear 259 (1–6) (2005) 562–568. [24] K. Poser, K.H. Zum Gahr, J. Schneider, Development of Al2 O3 based ceramics for dry friction systems, Wear 259 (1–6) (2005) 529–538. [25] M. Kalin, et al., Wear mechanisms associated with the lubrication of zirconia ceramics in various aqueous solutions, Journal of the European Ceramic Society 26 (3) (2006) 223–232. [26] M. Kalin, S. Novak, J. Vizintin, Surface charge as a new concept for boundary lubrication of ceramics with water, Journal of Physics D: Applied Physics 39 (15) (2006) 3138–3149. [27] K.H. Zum Gahr, M. Mathieu, B. Brylka, Friction control by surface engineering of ceramic sliding pairs in water, Wear 263 (2007) 920–929. [28] M. Belmonte, et al., Wear of aligned silicon nitride under dry sliding conditions, Wear 266 (1–2) (2009) 6–12. [29] G. Richards, Alumina ceramics, Transactions of British Ceramic Society 80 (1981) 120–124. [30] N. Ohmichi, K. Kamioka, K. Ueda, Phase transformation of zirconia ceramics by annealing in hot water, Journal of Japanese Ceramic Society 107 (1999) 128–133. [31] K. Yasuda, et al., Influence of Y2 O3 distribution on the rate of tetragonal to monoclinic phase transformation of yttria stabilised zirconia during hydrothermal aging, Journal of Material Science 34 (1999) 3597–3604. [32] R. Morrell, Shriers Corrosion Handbook, in press. [33] H. Czichos, A Systems Approach to Tribology, Elsevier, 1978.