- Email: [email protected]
AFM observation of diamond indenters after oxidation at elevated temperatures
Diamond & Related Materials 19 (2010) 1348–1353
Contents lists available at ScienceDirect
Diamond & Related Materials j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / d i a m o n d
AFM observation of diamond indenters after oxidation at elevated temperatures J.M. Wheeler, R.A. Oliver, T.W. Clyne ⁎ Department of Materials Science & Metallurgy, Cambridge University, Pembroke Street, Cambridge CB2 3QZ, UK
a r t i c l e
i n f o
Article history: Received 17 January 2010 Received in revised form 12 June 2010 Accepted 8 July 2010 Available online 24 July 2010 Keywords: Indentation Oxidation Diamond Berkovich and AFM
a b s t r a c t Use of diamond indenter tips at elevated temperatures can cause oxidation and thermomechanical damage, leading to changes in their topography. A Berkovich diamond indenter has been exposed to 450 °C in air, followed by 750 °C and 900 °C in 1 bar of static, commercial purity argon (30–45 ppm O2). The effects of oxidation on the geometry of the indenter were investigated using atomic force microscopy. A Berkovich and a 10 μm tip radius conospheroidal indenter were also examined, after being subjected to 5 years of intermittent use at elevated temperatures (≤400 °C). Significant changes in tip topography were observed, suggesting that commercial purity argon may be an unsuitable atmosphere for high temperature indentation testing. Finally, a mechanism of oxidative etching, which may have potential as a method of sharpening indenters, is also reported. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Nanoindentation testing of materials at elevated temperatures is an increasingly active field of research [1–8]. Berkovich diamond indenters are often used for testing non-ferrous materials at temperatures up to ~ 400 °C and sapphire indenters for ferrous alloys and other materials at temperatures up to ~ 750 °C. Oxidation (erosion) of diamond (indenters) represents a significant problem. Over the temperature range of interest, the oxygen partial pressure required to make oxidation thermodynamically unfavourable is well below the attainable range, even in UHV systems. The key issue is therefore the kinetics of oxidation, which naturally accelerate as the temperature is raised. Diamond indenters are thus not normally used above 400 °C in air, due to relatively rapid oxidation [9,10]. This is a significant limitation, since many materials used in high temperature applications, such as cermets for tool bits, are harder than sapphire. Direct oxidation of diamond produces CO or CO2 gas, which, in flowing oxygen or air, gives a constant etch rate. Oxidation can also lead to formation of a carbon layer on the surface of the diamond, presumably due to reaction with chemisorbed oxygen, and this layer is then converted to CO and/or CO2 gas. These three processes compete at different pressures and temperatures, such that some conditions favour the formation of a thick amorphous carbon layer and others the direct oxidation of a clean diamond surface [9]. This carbon layer formation differs from graphitisation of diamond, which ⁎ Corresponding author. E-mail address: [email protected] (T.W. Clyne). URL: http://www.msm.cam.ac.uk/department/profiles/clyne.php (T.W. Clyne). 0925-9635/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2010.07.004
occurs at much higher temperatures (e.g. N1200 °C) via a phase transformation [11]. All of these processes act to reduce the surface area to volume ratio, blunting conical and pyramidal indenter tips. Elevated temperature micro-hardness testing is often carried out with inert gas or vacuum systems being used to minimise diamond tip oxidation [12]. However, the scale of interest during nanoindentation testing is often appreciably finer, such that the exact topography of specimen and indenter need to be controlled and monitored much more precisely than for micro-indentation. Some dependence has been reported [9] of the oxidation rate on the crystallographic orientation of the exposed face at different oxygen partial pressures — see Fig. 1. In general, however, there is relatively little information available about the oxidative erosion of (natural) diamond in the form of relatively large single crystals. This is in contrast to the situation for various artificial layers and coatings of diamond and diamond-like carbon, which have been quite extensively studied [13,14]. However, the behaviour of such systems, which often have very fine grain sizes and a range of compositions and bond types, is not expected to be close to that of pure, single crystal diamond, of the type used as indenters. The surface topology, specifically the projected area function, of indenter tips is a critical concern for reliable nanoindentation testing [15]. Accurate determination of the area function of an indenter from load-displacement data can be a complicated procedure [16], and these methods for determining area functions yield little morphological data about the tip. Direct observation of tips is necessary for reliable observation of changes in tip morphology and dimensions, which can significantly affect indentation measurements. In the present investigation, atomic force microscopy is used to explore the effects of high temperature oxidation on Berkovich diamond indenters.
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
1349
initial scan of the Berkovich indenter that had been exposed to high temperature oxidation includes a limited amount of sample mount drift. However, improvements to the sample mount corrected this for all subsequent scans. Indenters were cleaned prior to AFM characterisation via manual indentations into polystyrene. This cleaning treatment removes most of the bulk contamination of the samples, but maintains the tips in a similar condition to normal service. Analysis of the data was accomplished using WsXM software from Nanotec [17]. 3. Observed changes in the topography of diamond tips 3.1. Berkovich diamond indenter exposed to successive treatments
Fig. 1. Reported [9] oxidative etch rates of diamond low energy faces, as a function of oxygen partial pressure and temperature, with power law extrapolations.
2. Experimental procedures 2.1. Tip oxidation After initial AFM characterisation, a commercially obtained natural type IIa diamond Berkovich indenter was exposed to three high temperature treatments in succession. The first consisted of holding the tip at 450 °C in air for 30 min, and the following treatments exposed the tip at 750 °C and 900 °C to 1 bar of static, commercial purity argon (30–45 ppm O2) for 90 min each. For this tip, all observed changes in morphology are attributed to oxidative attack. Additionally, a Berkovich and a 10 μm radius conospheroidal tip, which had been subjected to about 5 years of intermittent use at various temperatures (≤400 °C) in air, were characterised using the AFM, in order to examine the effects of long term exposure to oxidation and mechanical wear from indentation. 2.2. AFM characterisation The Berkovich indenters were characterised using a Veeco Dimension 3100 AFM system with an XY closed-loop scanner, using TappingMode™ with RTESPs (Rotated Tapping Etched Silicon Probes). After locating the surface and apex of the indenter tip, 512 × 512 pixel scans of 10 μm by 10 μm and 2.5 μm by 2.5 μm were performed. The
Results for the initial state of the Berkovich indenter (Fig. 2(a)) show fairly sharp facet edges, as illustrated by the triangular profiles of the contour lines. However, several irregular bumps are also observable. This is partially due to contamination and partially due to surface roughness. Terraces and steps can be seen on the two facets towards the bottom of Fig. 2(b). The steps frame or continue around rough surface irregularities, but are covered by contamination, allowing the actual roughness of the diamond surface to be distinguished from contamination particles. The sample drift, which occurred during scanning, laterally translated the data towards the right of the image and caused the Berkovich to appear skewed. This also changed the apparent slopes of the facets and directions of the steps on the facets. The lowest facet edge may appear flatter due to it drifting with the direction of scan as well. For imaging the tip after it had been held at 450 °C in air for 30 min, a new sample holder was obtained, greatly reducing sample drift. However, some noise is observable in a few scan-lines. This temperature was chosen because it is near the maximum allowable for diamond usage in air. The absence of sample drift-induced skew in the data can be inferred from the roughly equiangular triangles of the contour lines in Fig. 3(a). Facet edge blunting is also observable in the rounded corners of the contour lines. The step features (Fig. 3(b)) now appear to be approximately normal to the facet slopes, but remain absent in the upper facet. The surface roughness is reduced in this condition, but the surface features observable are larger, and extend over several steps. After a 90 min exposure at 750 °C in commercially pure argon (Fig. 4), extensive edge ridging is observable, as well as an approximately flat area at the apex. (The edge ridging was reproducibly observed with different AFM probes used for imaging). The effect of the ridging on the contour lines (Fig. 4(a)) is to reduce significantly the projected area in the region of the corners affected. The ridge of bumps alongside the upper right and lower facet edges is apparently contamination. Interestingly, the lowest
Fig. 2. AFM data, showing the initial state of the Berkovich indenter in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
1350
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
Fig. 3. AFM data showing the Berkovich indenter after 30 min at 450 °C in air in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
facet edge, which occurs at the junction of the two facets with step features, appears to be sharper than in the previous condition. The surface roughness on all facets also appears to be lower. The final treatment to which the Berkovich diamond was subjected (90 min at 900 °C in argon) had a significantly different effect on the diamond, compared with the previous cases. Oxidative attack appears to have occurred uniformly on all facet edges, which can be observed in the uniform rounding of the corners on the contour lines in Fig. 5(a). Rather than smooth blunting, or attack at the apex and facet edges, etch pits were created over the entire surface of the diamond — see Fig. 5(b). This apparent change in oxidative etching/ erosion behaviour could represent a different type of oxidative attack, due to the elevated temperature, or could simply reflect the outcome the same oxidation mechanism occurring more rapidly. For quantitative analysis, cross sectional or projected area functions were extracted from these data, using height histograms. This method takes all the datapoints from within an increment of the z axis, determines the enclosed projected area, and repeats this over the vertical range of the data. From the apex to a depth of 150 nm, area functions were extracted from the 2.5 μm by 2.5 μm scans and areas, from depths of 150 nm to 1500 nm, were extracted from the 10 μm by 10 μm scans. This was done to ensure that the entire projected area for each depth was contained within the bounds of the scan. (A small discontinuity can be observed in the plots at the point where the data from the two scans meet.) Conventional mechanical determinations of the area function were not obtained, since this would've required hundreds of indentations and thus introduced the possibility of mechanical wear in addition to oxidation.
It is unfortunate that the data for the initial condition were skewed. The lateral drift makes the tip appear elongated and increases its apparent area, while the areas at greater depths from the apex are less affected. This effect is probably apparent in Fig. 6, since the initial condition remains near to the ideal Berkovich areas at higher depths and then rapidly blunts near the tip. In all conditions, at shallow depths, the area functions approximate more closely a conospheroidal tip with a 500 nm radius tip on a 70.3° cone, which is a cone that has the same area-depth relationship as an ideal Berkovich pyramid. A plot comparing the percentage deviation of the measured areas to the initial projected areas for each condition (Fig. 7) is a direct method to compare quantitatively the change in area functions between successive exposures. This was generated by fitting a 6th order polynomial to the projected area function of the initial scan and determining the deviation for all the successive treatments. Fitting was required, since the data intervals on all of the scans were not perfectly coincident. The deviation of the initial scans area function from this fit is also shown, and this demonstrates the fit is suitable for comparing values for depths greater than 30 nm. This type of plot also allows easy extrapolation to measurement errors caused by topography changes, since Hardness and Modulus values are directly proportional to the value and square root of the contact area respectively, i.e. a 25% deviation in area will produce a 25% change in Hardness and a 5% change in Modulus values. It is observed that, for depths greater than 30 nm, the initial condition is significantly closer to an ideal Berkovich pyramid, i.e. a ‘sharper’ indenter. The next condition, 30 min at 450 °C in air, is, surprisingly, the most ‘blunt’ of all the conditions examined, with the greatest deviation
Fig. 4. AFM data showing the Berkovich indenter after 90 min at 750 °C in argon in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
1351
Fig. 5. AFM data showing the Berkovich indenter after 90 min at 900 °C in argon in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
from ideal over all depths. The indenter apparently ‘sharpens’ after the next exposure for 90 min at 750 °C in argon. However, study of the data (Fig. 4) shows that this is due to attack of the facet edges reducing the cross sectional area. This does effectively ‘sharpen’ the indenter. However, it causes a deviation from the ideal three-sided pyramidal shape, which is undesirable for analysis of indentation data. These changes in topography are not expected to change the mechanical properties of the tip, but the blunting will change the relevant contact mechanics towards spherical rather than pyramidal/conical indentation. The final condition of the indenter after 900 °C in argon for 90 min shows a fairly uniform ‘blunting’ of the indenter, from the previous exposure at 750 °C in argon. However, the effect of the etch pitting observed in Fig. 5 on the behaviour during indentation may cause a significant difference between the performance and that of a smooth-faced indenter of similar ‘bluntness’ by increasing the frictional forces. Finally, it should be noted that neither of the high temperature exposure conditions show the extreme blunting which would be expected from the etch rates extrapolated from Fig. 1 for these oxygen partial pressures and temperature. The effect of the tip being in contact with a sample during indentation at elevated temperature will be determined strongly by the sample's mechanical and chemical properties. Strong carbon-acceptor/carbide-former sample materials will have a significant deleterious effect on a diamond tip's topology, as they would absorb carbon from the tip. Soft materials might be expected to act as a shield to the contacting portion of the tip, while sufficiently hard
The two diamond tips examined had been in extended service over several years, on a variety of materials and over a range of temperatures (b400 °C). The outcome is shown in Figs. 8 and 9. The Berkovich indenter (Fig. 8) displays a somewhat stable, uniformly ‘blunted’ appearance, without the ‘blunted’ facet edges observed in Fig. 3. Interestingly, no step features are apparent on any of the facets. The sphericity of the conospheroidal indenter (Fig. 9) remains apparently unaffected after the prolonged service, without any faceting or severe blunting. A crack-like ridge is observed around a depth of 1.5 μm from the apex, and a faint cross-shape centred on the apex is barely distinguishable. This cross shape is possibly a remnant of the original shape of the indenter, prior to mechanical grinding into the conospheroidal geometry. Without data on the initial states of these indenters, it is difficult to speculate on the exact degree of oxidation or blunting which might have taken place. However, it appears that they have remained reasonably stable.
Fig. 6. The projected area, as a function of depth from the apex, for the heat treated tips, together with those for an ideal Berkovich tip and an equivalent area conospheroid with a 500 nm radius tip.
Fig. 7. Deviation of the measured areas, as a function of depth, for the heat treated tips from a polynomial fit of the initial scan. Ratio of the measured area, as a function of depth, for the heat treated tips.
materials might act to abrasively remove the amorphous carbon layer formed by surface oxidation and possibly abrade the surface of the diamond as well. Chipping or fracture of the diamond might occur if the sample is of comparable hardness to the diamond.
3.2. Indenters subjected to prolonged usage
1352
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
Fig. 8. Orthogonal 3D projections of AFM data, taken from a Berkovich indenter used at temperatures b400 °C and shown at 45° inclination aligned to the reference edge in (a) a 10 μm by 10 μm scan and (b) a 2.5 μm by 2.5 μm scan.
3.3. Step (terrace) features An intriguing feature discovered during the present study is the steps (or terraces) seen on the surfaces of two of the indenters. The absence of the steps on the extensively-used Berkovich indenter, and on the upper face of the successively oxidised Berkovich indenter, is also curious. A similar effect has been observed previously on some faces of plasma etched CVD diamond coatings [18]. These steps seem likely to have been formed by the preferential exposure of low energy planes. This was explored via examination of the spherical indenter (Figs .9 and 10). By taking high resolution scans at successively greater distances from the apex, the angle of the tangent surface relative to the internal structure was effectively varied. It can be seen (Fig. 10) that the spacing of the step features is dependent upon the slope of the surface. At the bottom right corner of Fig. 10(d), the steps are spaced ~ 30 nm apart, while at the top left corner of Fig. 10(a) the step spacing is ~110 nm. This suggests that surfaces which do not exhibit step features may be closely aligned to low energy planes of the diamond. It's clear that further systematic work is needed on this, where the facet angle relative to the low energy planes can be varied in a controlled manner. It is worth noting that, during oxidation, these steps probably translate along the surface, since the step edge is obviously more vulnerable to erosion than other regions. If a step encountered a surface asperity during oxidation-induced translation, the irregular shape of the asperity would probably have a high surface energy, causing it to be preferentially consumed and reduced to a planar surface. This might effectively ‘polish’ planar surfaces, as was observed in Figs. 2(b)–4(b). If neighbouring facets are appropriately aligned to each other, then this same action might sharpen the edges of the pyramid, as was observed in the lower edge in Fig. 4(a). It is possible to imagine an indenter geometry, with an appropriate relationship to the crystallographic orientation of the diamond cubic structure, becoming ‘sharpened’ during oxidation. This might produce
diamond indenters etched to near theoretical sharpness, in a similar manner to the method by which AFM probes are prepared, using facet-selective wet-etching [19]. Indenter geometries suitable for this type of sharpening may be limited to a cube-corner geometry oriented about a particular crystallographic axis, but might also favor 4-sided geometries. This might yield a Vickers tip without a line of conjunction. Further work is clearly needed in this area. 4. Conclusions The following conclusions can be drawn from this work. (a) Oxidation of diamond is thermodynamically favoured over the complete attainable range of temperature and oxygen partial pressure. The kinetics of oxidation accelerate with increasing temperature and oxygen pressure. Relatively rapid oxidation (erosion) occurs at temperatures ≥450 °C in ambient air. In the temperature range of current interest for nanoindentation, i.e. up to about 800 °C, very low levels of oxygen, lower than those in commercial purity argon, are necessary in order for diamond erosion rates to be negligible. These levels are, however, attainable in vacuum systems, or in atmospheres of high purity inert gas. (b) Various observations have been made concerning the types of contamination and erosion that can occur on indenter tips used for nanoindentation under a range of conditions. Attempts have been made to rationalise and explain these observations. It is noted that some of the observed changes in tip topology would be likely to require a secondary, non-mechanical evaluation of area functions by a technique such as AFM. (c) Among the effects that have been observed is the formation of steps, or terraces, on the surfaces of some indenters, even in some cases when the exposure times and temperatures have not been excessive. It seems likely that this is due to the
Fig. 9. (a) Orthogonal 3D projection of AFM data taken from a 15 μm by 15 μm scan of a 10 μm radius conospheroidal indenter, used at temperatures b 400 °C, shown at 45° inclination, and (b) the derivative image, with high resolution scan regions shown.
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
1353
Fig. 10. 1 μm by 1 μm derivative images of AFM data (a)–(d), taken from regions of increasing distance from the apex as highlighted in Fig. 9(b).
preferential exposure of low energy crystallographic planes. There is certainly a crystallographic element to the way that these features form. While the step heights are in most cases quite small (~few nm), the features might be of significance in some cases. (d) It has been observed that, while erosion is obviously expected to be deleterious in many cases, there may be scope for exploiting these effects under some circumstances. For example, it has been shown that erosion can in some cases take place so as to “sharpen” the edges of an indenter tip, and possibly to produce a better-defined tip shape than is possible just by conventional machining. It seems likely that, in order for this approach to be useful, a relationship would need to be preselected between the crystallographic orientation of the diamond single crystal and the geometry of the indenter to expose faces with favorable oxidation characteristics. Acknowledgements This material is based upon work partially supported under a National Science Foundation Graduate Research Fellowship. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author and do not necessarily reflect the views of the National Science Foundation. JMW is grateful for financial support from the NSF and from AWE. RAO is grateful for support from the Royal Society. Contributions to
the work reported here have been made by RJ Stearn and AW Rayment, of the Materials Science Department in Cambridge University. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
J.F. Smith, S. Zheng, Surface Engineering 16 (2000) 143. B. Beake, J. Smith, Philosophical Magazine A 82 (2002) 2179. B. Beake, S. Goodes, J. Smith, Zeitschrift fur Metallkunde 94 (2003) 1. A.C. Lund, A.M. Hodge, C.A. Schuh, Applied Physics Letters 85 (2004) 1362. C.A. Schuh, J.K. Mason, A.C. Lund, Nature Materials 4 (2005) 617. A.J. Muir Wood, T.W. Clyne, Acta Materialia 54 (2006) 5607. R. Goodall, T.W. Clyne, Acta Materialia 54 (2006) 5489. S. Tin, A. Sawant, Scripta Materialia 58 (2008) 275. T. Evans, C. Phaal, The Kinetics of the Diamond-Oxygen Reaction, Proceedings of the 5th Conference on Carbon, 1961. T. Evans, Changes Produced by High Temperature Treatment of Diamon, in: J.E. Field (Ed.), The Properties of Diamond, Academic Press, New York, 1979, p. 403. J.E. Field, The Properties of Natural and Synthetic Diamond. Academic Press, 1992. M.G.S. Naylor, T.F. Page, Journal of Microscopy 130 (1983) 345. J. Li, Q. Zhang, S.F. Yoon, J. Ahn, Q. Zhou, S. Wang, D. Yang, Carbon 41 (2003) 1847. A. Joshi, R. Nimmagadda, Journal of Materials Research 6 (1991) 1484. G. Aldrich-Smith, N.M. Jennett, U. Hangen, Zeitschrift Fur Metallkunde 96 (2005) 1267. K. Herrmann, N.M. Jennett, W. Wegener, J. Meneve, K. Hasche, R. Seeman, Thin Solid Films 377 (2000) 394. I. Horcas, R. Fernandez, J.M. Gomez-Rodriguez, J. Colchero, J. Gomez-Herrero, A.M. Baro, Review of Scientific Instruments 78 (2007) 013705. M. Wolfer, J. Biener, B.S. El-dasher, M.M. Biener, A.V. Hamza, A. Kriele, C. Wild, Diamond and Related Materials 18 (2009) 713. O. Wolter, T. Bayer, J. Greschner, Journal of Vacuum Science and Technology B 9 (1991) 1353.
Contents lists available at ScienceDirect
Diamond & Related Materials j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / d i a m o n d
AFM observation of diamond indenters after oxidation at elevated temperatures J.M. Wheeler, R.A. Oliver, T.W. Clyne ⁎ Department of Materials Science & Metallurgy, Cambridge University, Pembroke Street, Cambridge CB2 3QZ, UK
a r t i c l e
i n f o
Article history: Received 17 January 2010 Received in revised form 12 June 2010 Accepted 8 July 2010 Available online 24 July 2010 Keywords: Indentation Oxidation Diamond Berkovich and AFM
a b s t r a c t Use of diamond indenter tips at elevated temperatures can cause oxidation and thermomechanical damage, leading to changes in their topography. A Berkovich diamond indenter has been exposed to 450 °C in air, followed by 750 °C and 900 °C in 1 bar of static, commercial purity argon (30–45 ppm O2). The effects of oxidation on the geometry of the indenter were investigated using atomic force microscopy. A Berkovich and a 10 μm tip radius conospheroidal indenter were also examined, after being subjected to 5 years of intermittent use at elevated temperatures (≤400 °C). Significant changes in tip topography were observed, suggesting that commercial purity argon may be an unsuitable atmosphere for high temperature indentation testing. Finally, a mechanism of oxidative etching, which may have potential as a method of sharpening indenters, is also reported. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Nanoindentation testing of materials at elevated temperatures is an increasingly active field of research [1–8]. Berkovich diamond indenters are often used for testing non-ferrous materials at temperatures up to ~ 400 °C and sapphire indenters for ferrous alloys and other materials at temperatures up to ~ 750 °C. Oxidation (erosion) of diamond (indenters) represents a significant problem. Over the temperature range of interest, the oxygen partial pressure required to make oxidation thermodynamically unfavourable is well below the attainable range, even in UHV systems. The key issue is therefore the kinetics of oxidation, which naturally accelerate as the temperature is raised. Diamond indenters are thus not normally used above 400 °C in air, due to relatively rapid oxidation [9,10]. This is a significant limitation, since many materials used in high temperature applications, such as cermets for tool bits, are harder than sapphire. Direct oxidation of diamond produces CO or CO2 gas, which, in flowing oxygen or air, gives a constant etch rate. Oxidation can also lead to formation of a carbon layer on the surface of the diamond, presumably due to reaction with chemisorbed oxygen, and this layer is then converted to CO and/or CO2 gas. These three processes compete at different pressures and temperatures, such that some conditions favour the formation of a thick amorphous carbon layer and others the direct oxidation of a clean diamond surface [9]. This carbon layer formation differs from graphitisation of diamond, which ⁎ Corresponding author. E-mail address: [email protected] (T.W. Clyne). URL: http://www.msm.cam.ac.uk/department/profiles/clyne.php (T.W. Clyne). 0925-9635/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2010.07.004
occurs at much higher temperatures (e.g. N1200 °C) via a phase transformation [11]. All of these processes act to reduce the surface area to volume ratio, blunting conical and pyramidal indenter tips. Elevated temperature micro-hardness testing is often carried out with inert gas or vacuum systems being used to minimise diamond tip oxidation [12]. However, the scale of interest during nanoindentation testing is often appreciably finer, such that the exact topography of specimen and indenter need to be controlled and monitored much more precisely than for micro-indentation. Some dependence has been reported [9] of the oxidation rate on the crystallographic orientation of the exposed face at different oxygen partial pressures — see Fig. 1. In general, however, there is relatively little information available about the oxidative erosion of (natural) diamond in the form of relatively large single crystals. This is in contrast to the situation for various artificial layers and coatings of diamond and diamond-like carbon, which have been quite extensively studied [13,14]. However, the behaviour of such systems, which often have very fine grain sizes and a range of compositions and bond types, is not expected to be close to that of pure, single crystal diamond, of the type used as indenters. The surface topology, specifically the projected area function, of indenter tips is a critical concern for reliable nanoindentation testing [15]. Accurate determination of the area function of an indenter from load-displacement data can be a complicated procedure [16], and these methods for determining area functions yield little morphological data about the tip. Direct observation of tips is necessary for reliable observation of changes in tip morphology and dimensions, which can significantly affect indentation measurements. In the present investigation, atomic force microscopy is used to explore the effects of high temperature oxidation on Berkovich diamond indenters.
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
1349
initial scan of the Berkovich indenter that had been exposed to high temperature oxidation includes a limited amount of sample mount drift. However, improvements to the sample mount corrected this for all subsequent scans. Indenters were cleaned prior to AFM characterisation via manual indentations into polystyrene. This cleaning treatment removes most of the bulk contamination of the samples, but maintains the tips in a similar condition to normal service. Analysis of the data was accomplished using WsXM software from Nanotec [17]. 3. Observed changes in the topography of diamond tips 3.1. Berkovich diamond indenter exposed to successive treatments
Fig. 1. Reported [9] oxidative etch rates of diamond low energy faces, as a function of oxygen partial pressure and temperature, with power law extrapolations.
2. Experimental procedures 2.1. Tip oxidation After initial AFM characterisation, a commercially obtained natural type IIa diamond Berkovich indenter was exposed to three high temperature treatments in succession. The first consisted of holding the tip at 450 °C in air for 30 min, and the following treatments exposed the tip at 750 °C and 900 °C to 1 bar of static, commercial purity argon (30–45 ppm O2) for 90 min each. For this tip, all observed changes in morphology are attributed to oxidative attack. Additionally, a Berkovich and a 10 μm radius conospheroidal tip, which had been subjected to about 5 years of intermittent use at various temperatures (≤400 °C) in air, were characterised using the AFM, in order to examine the effects of long term exposure to oxidation and mechanical wear from indentation. 2.2. AFM characterisation The Berkovich indenters were characterised using a Veeco Dimension 3100 AFM system with an XY closed-loop scanner, using TappingMode™ with RTESPs (Rotated Tapping Etched Silicon Probes). After locating the surface and apex of the indenter tip, 512 × 512 pixel scans of 10 μm by 10 μm and 2.5 μm by 2.5 μm were performed. The
Results for the initial state of the Berkovich indenter (Fig. 2(a)) show fairly sharp facet edges, as illustrated by the triangular profiles of the contour lines. However, several irregular bumps are also observable. This is partially due to contamination and partially due to surface roughness. Terraces and steps can be seen on the two facets towards the bottom of Fig. 2(b). The steps frame or continue around rough surface irregularities, but are covered by contamination, allowing the actual roughness of the diamond surface to be distinguished from contamination particles. The sample drift, which occurred during scanning, laterally translated the data towards the right of the image and caused the Berkovich to appear skewed. This also changed the apparent slopes of the facets and directions of the steps on the facets. The lowest facet edge may appear flatter due to it drifting with the direction of scan as well. For imaging the tip after it had been held at 450 °C in air for 30 min, a new sample holder was obtained, greatly reducing sample drift. However, some noise is observable in a few scan-lines. This temperature was chosen because it is near the maximum allowable for diamond usage in air. The absence of sample drift-induced skew in the data can be inferred from the roughly equiangular triangles of the contour lines in Fig. 3(a). Facet edge blunting is also observable in the rounded corners of the contour lines. The step features (Fig. 3(b)) now appear to be approximately normal to the facet slopes, but remain absent in the upper facet. The surface roughness is reduced in this condition, but the surface features observable are larger, and extend over several steps. After a 90 min exposure at 750 °C in commercially pure argon (Fig. 4), extensive edge ridging is observable, as well as an approximately flat area at the apex. (The edge ridging was reproducibly observed with different AFM probes used for imaging). The effect of the ridging on the contour lines (Fig. 4(a)) is to reduce significantly the projected area in the region of the corners affected. The ridge of bumps alongside the upper right and lower facet edges is apparently contamination. Interestingly, the lowest
Fig. 2. AFM data, showing the initial state of the Berkovich indenter in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
1350
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
Fig. 3. AFM data showing the Berkovich indenter after 30 min at 450 °C in air in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
facet edge, which occurs at the junction of the two facets with step features, appears to be sharper than in the previous condition. The surface roughness on all facets also appears to be lower. The final treatment to which the Berkovich diamond was subjected (90 min at 900 °C in argon) had a significantly different effect on the diamond, compared with the previous cases. Oxidative attack appears to have occurred uniformly on all facet edges, which can be observed in the uniform rounding of the corners on the contour lines in Fig. 5(a). Rather than smooth blunting, or attack at the apex and facet edges, etch pits were created over the entire surface of the diamond — see Fig. 5(b). This apparent change in oxidative etching/ erosion behaviour could represent a different type of oxidative attack, due to the elevated temperature, or could simply reflect the outcome the same oxidation mechanism occurring more rapidly. For quantitative analysis, cross sectional or projected area functions were extracted from these data, using height histograms. This method takes all the datapoints from within an increment of the z axis, determines the enclosed projected area, and repeats this over the vertical range of the data. From the apex to a depth of 150 nm, area functions were extracted from the 2.5 μm by 2.5 μm scans and areas, from depths of 150 nm to 1500 nm, were extracted from the 10 μm by 10 μm scans. This was done to ensure that the entire projected area for each depth was contained within the bounds of the scan. (A small discontinuity can be observed in the plots at the point where the data from the two scans meet.) Conventional mechanical determinations of the area function were not obtained, since this would've required hundreds of indentations and thus introduced the possibility of mechanical wear in addition to oxidation.
It is unfortunate that the data for the initial condition were skewed. The lateral drift makes the tip appear elongated and increases its apparent area, while the areas at greater depths from the apex are less affected. This effect is probably apparent in Fig. 6, since the initial condition remains near to the ideal Berkovich areas at higher depths and then rapidly blunts near the tip. In all conditions, at shallow depths, the area functions approximate more closely a conospheroidal tip with a 500 nm radius tip on a 70.3° cone, which is a cone that has the same area-depth relationship as an ideal Berkovich pyramid. A plot comparing the percentage deviation of the measured areas to the initial projected areas for each condition (Fig. 7) is a direct method to compare quantitatively the change in area functions between successive exposures. This was generated by fitting a 6th order polynomial to the projected area function of the initial scan and determining the deviation for all the successive treatments. Fitting was required, since the data intervals on all of the scans were not perfectly coincident. The deviation of the initial scans area function from this fit is also shown, and this demonstrates the fit is suitable for comparing values for depths greater than 30 nm. This type of plot also allows easy extrapolation to measurement errors caused by topography changes, since Hardness and Modulus values are directly proportional to the value and square root of the contact area respectively, i.e. a 25% deviation in area will produce a 25% change in Hardness and a 5% change in Modulus values. It is observed that, for depths greater than 30 nm, the initial condition is significantly closer to an ideal Berkovich pyramid, i.e. a ‘sharper’ indenter. The next condition, 30 min at 450 °C in air, is, surprisingly, the most ‘blunt’ of all the conditions examined, with the greatest deviation
Fig. 4. AFM data showing the Berkovich indenter after 90 min at 750 °C in argon in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
1351
Fig. 5. AFM data showing the Berkovich indenter after 90 min at 900 °C in argon in (a) topography image, with contour lines at 50 nm intervals, and (b) an orthogonal 3D projection of AFM data shown at 45° inclination aligned to a reference edge.
from ideal over all depths. The indenter apparently ‘sharpens’ after the next exposure for 90 min at 750 °C in argon. However, study of the data (Fig. 4) shows that this is due to attack of the facet edges reducing the cross sectional area. This does effectively ‘sharpen’ the indenter. However, it causes a deviation from the ideal three-sided pyramidal shape, which is undesirable for analysis of indentation data. These changes in topography are not expected to change the mechanical properties of the tip, but the blunting will change the relevant contact mechanics towards spherical rather than pyramidal/conical indentation. The final condition of the indenter after 900 °C in argon for 90 min shows a fairly uniform ‘blunting’ of the indenter, from the previous exposure at 750 °C in argon. However, the effect of the etch pitting observed in Fig. 5 on the behaviour during indentation may cause a significant difference between the performance and that of a smooth-faced indenter of similar ‘bluntness’ by increasing the frictional forces. Finally, it should be noted that neither of the high temperature exposure conditions show the extreme blunting which would be expected from the etch rates extrapolated from Fig. 1 for these oxygen partial pressures and temperature. The effect of the tip being in contact with a sample during indentation at elevated temperature will be determined strongly by the sample's mechanical and chemical properties. Strong carbon-acceptor/carbide-former sample materials will have a significant deleterious effect on a diamond tip's topology, as they would absorb carbon from the tip. Soft materials might be expected to act as a shield to the contacting portion of the tip, while sufficiently hard
The two diamond tips examined had been in extended service over several years, on a variety of materials and over a range of temperatures (b400 °C). The outcome is shown in Figs. 8 and 9. The Berkovich indenter (Fig. 8) displays a somewhat stable, uniformly ‘blunted’ appearance, without the ‘blunted’ facet edges observed in Fig. 3. Interestingly, no step features are apparent on any of the facets. The sphericity of the conospheroidal indenter (Fig. 9) remains apparently unaffected after the prolonged service, without any faceting or severe blunting. A crack-like ridge is observed around a depth of 1.5 μm from the apex, and a faint cross-shape centred on the apex is barely distinguishable. This cross shape is possibly a remnant of the original shape of the indenter, prior to mechanical grinding into the conospheroidal geometry. Without data on the initial states of these indenters, it is difficult to speculate on the exact degree of oxidation or blunting which might have taken place. However, it appears that they have remained reasonably stable.
Fig. 6. The projected area, as a function of depth from the apex, for the heat treated tips, together with those for an ideal Berkovich tip and an equivalent area conospheroid with a 500 nm radius tip.
Fig. 7. Deviation of the measured areas, as a function of depth, for the heat treated tips from a polynomial fit of the initial scan. Ratio of the measured area, as a function of depth, for the heat treated tips.
materials might act to abrasively remove the amorphous carbon layer formed by surface oxidation and possibly abrade the surface of the diamond as well. Chipping or fracture of the diamond might occur if the sample is of comparable hardness to the diamond.
3.2. Indenters subjected to prolonged usage
1352
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
Fig. 8. Orthogonal 3D projections of AFM data, taken from a Berkovich indenter used at temperatures b400 °C and shown at 45° inclination aligned to the reference edge in (a) a 10 μm by 10 μm scan and (b) a 2.5 μm by 2.5 μm scan.
3.3. Step (terrace) features An intriguing feature discovered during the present study is the steps (or terraces) seen on the surfaces of two of the indenters. The absence of the steps on the extensively-used Berkovich indenter, and on the upper face of the successively oxidised Berkovich indenter, is also curious. A similar effect has been observed previously on some faces of plasma etched CVD diamond coatings [18]. These steps seem likely to have been formed by the preferential exposure of low energy planes. This was explored via examination of the spherical indenter (Figs .9 and 10). By taking high resolution scans at successively greater distances from the apex, the angle of the tangent surface relative to the internal structure was effectively varied. It can be seen (Fig. 10) that the spacing of the step features is dependent upon the slope of the surface. At the bottom right corner of Fig. 10(d), the steps are spaced ~ 30 nm apart, while at the top left corner of Fig. 10(a) the step spacing is ~110 nm. This suggests that surfaces which do not exhibit step features may be closely aligned to low energy planes of the diamond. It's clear that further systematic work is needed on this, where the facet angle relative to the low energy planes can be varied in a controlled manner. It is worth noting that, during oxidation, these steps probably translate along the surface, since the step edge is obviously more vulnerable to erosion than other regions. If a step encountered a surface asperity during oxidation-induced translation, the irregular shape of the asperity would probably have a high surface energy, causing it to be preferentially consumed and reduced to a planar surface. This might effectively ‘polish’ planar surfaces, as was observed in Figs. 2(b)–4(b). If neighbouring facets are appropriately aligned to each other, then this same action might sharpen the edges of the pyramid, as was observed in the lower edge in Fig. 4(a). It is possible to imagine an indenter geometry, with an appropriate relationship to the crystallographic orientation of the diamond cubic structure, becoming ‘sharpened’ during oxidation. This might produce
diamond indenters etched to near theoretical sharpness, in a similar manner to the method by which AFM probes are prepared, using facet-selective wet-etching [19]. Indenter geometries suitable for this type of sharpening may be limited to a cube-corner geometry oriented about a particular crystallographic axis, but might also favor 4-sided geometries. This might yield a Vickers tip without a line of conjunction. Further work is clearly needed in this area. 4. Conclusions The following conclusions can be drawn from this work. (a) Oxidation of diamond is thermodynamically favoured over the complete attainable range of temperature and oxygen partial pressure. The kinetics of oxidation accelerate with increasing temperature and oxygen pressure. Relatively rapid oxidation (erosion) occurs at temperatures ≥450 °C in ambient air. In the temperature range of current interest for nanoindentation, i.e. up to about 800 °C, very low levels of oxygen, lower than those in commercial purity argon, are necessary in order for diamond erosion rates to be negligible. These levels are, however, attainable in vacuum systems, or in atmospheres of high purity inert gas. (b) Various observations have been made concerning the types of contamination and erosion that can occur on indenter tips used for nanoindentation under a range of conditions. Attempts have been made to rationalise and explain these observations. It is noted that some of the observed changes in tip topology would be likely to require a secondary, non-mechanical evaluation of area functions by a technique such as AFM. (c) Among the effects that have been observed is the formation of steps, or terraces, on the surfaces of some indenters, even in some cases when the exposure times and temperatures have not been excessive. It seems likely that this is due to the
Fig. 9. (a) Orthogonal 3D projection of AFM data taken from a 15 μm by 15 μm scan of a 10 μm radius conospheroidal indenter, used at temperatures b 400 °C, shown at 45° inclination, and (b) the derivative image, with high resolution scan regions shown.
J.M. Wheeler et al. / Diamond & Related Materials 19 (2010) 1348–1353
1353
Fig. 10. 1 μm by 1 μm derivative images of AFM data (a)–(d), taken from regions of increasing distance from the apex as highlighted in Fig. 9(b).
preferential exposure of low energy crystallographic planes. There is certainly a crystallographic element to the way that these features form. While the step heights are in most cases quite small (~few nm), the features might be of significance in some cases. (d) It has been observed that, while erosion is obviously expected to be deleterious in many cases, there may be scope for exploiting these effects under some circumstances. For example, it has been shown that erosion can in some cases take place so as to “sharpen” the edges of an indenter tip, and possibly to produce a better-defined tip shape than is possible just by conventional machining. It seems likely that, in order for this approach to be useful, a relationship would need to be preselected between the crystallographic orientation of the diamond single crystal and the geometry of the indenter to expose faces with favorable oxidation characteristics. Acknowledgements This material is based upon work partially supported under a National Science Foundation Graduate Research Fellowship. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author and do not necessarily reflect the views of the National Science Foundation. JMW is grateful for financial support from the NSF and from AWE. RAO is grateful for support from the Royal Society. Contributions to
the work reported here have been made by RJ Stearn and AW Rayment, of the Materials Science Department in Cambridge University. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
J.F. Smith, S. Zheng, Surface Engineering 16 (2000) 143. B. Beake, J. Smith, Philosophical Magazine A 82 (2002) 2179. B. Beake, S. Goodes, J. Smith, Zeitschrift fur Metallkunde 94 (2003) 1. A.C. Lund, A.M. Hodge, C.A. Schuh, Applied Physics Letters 85 (2004) 1362. C.A. Schuh, J.K. Mason, A.C. Lund, Nature Materials 4 (2005) 617. A.J. Muir Wood, T.W. Clyne, Acta Materialia 54 (2006) 5607. R. Goodall, T.W. Clyne, Acta Materialia 54 (2006) 5489. S. Tin, A. Sawant, Scripta Materialia 58 (2008) 275. T. Evans, C. Phaal, The Kinetics of the Diamond-Oxygen Reaction, Proceedings of the 5th Conference on Carbon, 1961. T. Evans, Changes Produced by High Temperature Treatment of Diamon, in: J.E. Field (Ed.), The Properties of Diamond, Academic Press, New York, 1979, p. 403. J.E. Field, The Properties of Natural and Synthetic Diamond. Academic Press, 1992. M.G.S. Naylor, T.F. Page, Journal of Microscopy 130 (1983) 345. J. Li, Q. Zhang, S.F. Yoon, J. Ahn, Q. Zhou, S. Wang, D. Yang, Carbon 41 (2003) 1847. A. Joshi, R. Nimmagadda, Journal of Materials Research 6 (1991) 1484. G. Aldrich-Smith, N.M. Jennett, U. Hangen, Zeitschrift Fur Metallkunde 96 (2005) 1267. K. Herrmann, N.M. Jennett, W. Wegener, J. Meneve, K. Hasche, R. Seeman, Thin Solid Films 377 (2000) 394. I. Horcas, R. Fernandez, J.M. Gomez-Rodriguez, J. Colchero, J. Gomez-Herrero, A.M. Baro, Review of Scientific Instruments 78 (2007) 013705. M. Wolfer, J. Biener, B.S. El-dasher, M.M. Biener, A.V. Hamza, A. Kriele, C. Wild, Diamond and Related Materials 18 (2009) 713. O. Wolter, T. Bayer, J. Greschner, Journal of Vacuum Science and Technology B 9 (1991) 1353.