The effect of substrate temperature on the properties of ... - Me.umn.edu
Journal of Nanoparticle Research 1: 31–42, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands.
The effect of substrate temperature on the properties of nanostructured silicon carbide films deposited by hypersonic plasma particle deposition J. Blum1 , N. Tymiak2 , A. Neuman1 , Z. Wong1 , N.P. Rao1 , S.L. Girshick1 , W.W. Gerberich2 , P.H. McMurry1 and J.V.R. Heberlein1 1 Department of Mechanical Engineering, 2 Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, MN 55455, USA Received 14 July 1998; accepted in revised form 12 January 1999
Key words: nanoparticles, thermal plasma, nanostructural film, particle deposition, silicon carbide, film hardness
Abstract Nanostructured silicon carbide films have been deposited on molybdenum substrates by hypersonic plasma particle deposition. In this process a thermal plasma with injected reactants (SiCl4 and CH4 ) is expanded through a nozzle leading to the nucleation of ultrafine particles. Particles entrained in the supersonic flow are then inertially deposited in vacuum onto a temperature-controlled substrate, leading to the formation of a consolidated film. In the experiments reported, the deposition substrate temperature Ts has ranged from 250◦ C to 700◦ C, and the effect of Ts on film morphology, composition, and mechanical properties has been studied. Examination of the films by scanning electron microscopy has shown that the grain sizes in the films did not vary significantly with Ts . Micro-X-ray diffraction analysis of the deposits has shown that amorphous films are deposited at low Ts , while crystalline films are formed at high Ts . Rutherford backscattering spectrometry has indicated that the films are largely stoichiometric silicon carbide with small amounts of chlorine. The chlorine content decreases from 8% to 1.5% when the deposition temperature is raised from 450◦ C to 700◦ C. Nanoindentation and microindentation tests have been performed on as-deposited films to measure hardness, Young’s modulus and to evaluate adhesion strength. The tests show that film adhesion, hardness and Young’s modulus increase with increasing Ts . These results taken together demonstrate that in HPPD, as in vapor deposition processes, the substrate temperature may be used to control film properties, and that better quality films are obtained at higher substrate temperatures, i.e. Ts ≈ 700◦ C.
Introduction Nanostructured materials have been found to possess mechanical properties, e.g. hardness and yield strength, superior to that of conventional coarse grained materials (Siegel, 1993). There is currently great interest in the synthesis and processing of such materials, particularly in creating functional forms such as films, coatings, composites, etc. We have recently developed a new process for the production of nanostructured
films from a thermal plasma route (Rao et al., 1997a; Rao et al., 1998). Thermal plasma synthesis provides a scalable, continuous flow processing route which is characterized by high temperatures, high energy density, fast reaction kinetics, rapid quenching, high processing rates and is especially well suited to the synthesis of ultrafine particles of non-oxide, hightemperature ceramics such as carbides, nitrides and borides (Kong & Pfender, 1987). The present work concerns the synthesis of nanostructured silicon carbide
32 films, which are formed by depositing freshly synthesized SiC nanoparticles onto a metal (molybdenum) substrate. Such films are potential candidates for wear resistant coatings. In the process being described, hypersonic plasma particle deposition (HPPD), vapor phase precursors are injected into a flowing thermal plasma generated by a dc arc torch. The plasma is then rapidly quenched by supersonic expansion in a ceramic lined nozzle, resulting in the nucleation and growth of nanoparticles. The nanoparticles are further accelerated in the hypersonic free jet downstream of the nozzle, and are deposited by inertial impaction onto a temperature-controlled substrate. The high particle deposition velocity results in the formation of a lightly consolidated nanostructured film, while the small particle residence times (∼ 50 µs) minimize in-flight agglomeration and contamination. A comparison of HPPD with related processes such as thermal plasma spray and thermal plasma chemical vapor deposition has been provided in Rao et al. (1997b). Earlier work on the HPPD process has demonstrated that the grain size in the film correlates well with the size of particles formed in the nozzle (Rao et al., 1997b). For this reason considerable effort was initially focused on controlling particle size in the nozzle, and thus the grain size in the deposit, through control of process parameters such as reactant flow rate, torch power, nozzle length, etc. In order to prevent significant grain growth from occurring, the substrate temperature was generally kept at about 200◦ C by active cooling. Recent experimental work, however, has indicated that the deposition substrate temperature is a key parameter in determining the properties of the product film. This paper describes results from the first experimental exploration of the effects of deposition substrate temperature on the properties of nanostructured films deposited by HPPD. The effect of substrate temperature on deposited films is well known in many other related processes. For example, amorphous silicon carbide films deposited by low pressure plasma enhanced chemical vapor deposition (PECVD) show significantly higher densities at higher deposition substrate temperatures (Delplancke et al., 1991). El Khakani et al. (1994) mention that CVD SiC films deposited at temperatures lower than 800◦ C are generally amorphous, while those formed at higher deposition temperatures are generally crystalline. Substrate temperature has also been found to be an important parameter affecting properties such
as the residual quenching stress in coatings deposited by thermal plasma spray (Bianchi et al., 1995). In this process, it is well known that stresses due to differential expansion are induced when a hot coated workpiece cools down to room temperature. Stresses are also induced when large micron-sized molten particles impacting on the substrate flatten out forming a splat which intertwines with other splats. Substrate temperature affects the wetting angle between the workpiece and the splat as well as the shrinkage of the splat when quenched by contact with the colder workpiece. Since the HPPD process has elements in common with PECVD and conventional thermal spray processes, it is likely that similar deposition substrate temperature effects may be observed in HPPD. In this paper we describe experiments which tested the hypothesis that the deposition substrate temperature may provide a useful means of tailoring the properties of the deposited film. In the sections following, we first describe the experimental apparatus used in the study, discuss the results obtained, and draw conclusions.
Experimental details Figure 1 shows the HPPD reactor which has been discussed in detail elsewhere (Rao et al., 1998). It consists of a thermal plasma reactor exhausting into a low pressure deposition chamber, along with associated systems for pumping, gas handling, power supply, data acquisition, diagnostics, scrubbing and safety. The plasma reactor assembly consists of a dc arc torch coupled to a ceramic reactant injection tube and converging nozzle having an entrance diameter of 15 mm and an exit diameter of 5 mm. Within the deposition chamber, a molybdenum substrate is clamped to a water-cooled holder at a distance 20 mm downstream of the nozzle exit and oriented so that the substrate surface faces the flow issuing from the nozzle. The temperature of the substrate is controlled by injecting a helium–argon gas mixture of controllable composition between the holder and the substrate (Bieberich & Girshick, 1996). Three types of substrates have been used in the experiments reported here: (i) a ‘normal’ substrate, a polished molybdenum disk 20 mm in diameter and 3 mm thick; (ii) a ‘break-apart’ molybdenum substrate similar in size to the ‘normal’ substrate, but provided with a removable central section held together by screws, and (iii) a ‘small’ unpolished molybdenum substrate
33 Table 1. Representative process conditions for HPPD experiments Process parameter
Value ◦
Substrate temperature, C SiCl4 flow rate, slm CH4 flow rate, slm Argon flow rate, slm Hydrogen flow rate, slm Arc current, A Nozzle–plate distance, mm Chamber pressure, torr
Figure 1. Schematic diagram of the HPPD reactor (Rao et al., 1998).
4 mm thick with a deposition area 10 mm in diameter. The ‘break-apart’ substrate has been designed to enable easier observation and analysis of the cross-section of deposited films, while the ‘small’ substrate has been designed to allow access to specialized materials analysis tools capable of handling only small substrates. The substrate temperature in all cases has been measured using a sheathed thermocouple probe embedded deep into a thermocouple well drilled into the side of the substrate and accessed through a hole drilled into the substrate clamping ring. With this arrangement, the thermocouple junction lay less than 1 mm below the top surface of the substrate disk. A portion of the thermocouple probe outside the substrate assembly was exposed to the plasma and tended to heat up during a deposition run. We did not attempt to correct for associated ‘finning’ effects in measuring the substrate temperature, and thus the temperatures reported here should be considered to be upper bound values. The finning effects are expected to be greatest for the case of the small substrates of type (iii). The experimental protocol for depositing nanostructured films by HPPD has been reported in previous
250–750 0.2 1.0 35 7.5 225 20 2.5
publications (Rao et al., 1998). In order to synthesize SiC films, SiCl4 vapor and CH4 reactants are injected into an Ar–H2 plasma. Representative operating conditions are listed in Table 1. A movable shutter plate within the deposition chamber protects the substrate assembly and thermocouple from exposure to the plasma before and after the deposition. Deposition times vary from 60–90 s for the experiments reported, with the substrate and deposited film cooling down rapidly to room temperature after deposition. The deposited films have been analyzed by scanning electron microscopy (SEM) to determine grain size and microstructure, by micro-X-ray diffraction (µ-XRD) to obtain phase information, by Rutherford backscattering spectrometry (RBS) to quantify elemental composition, and by micro- and nanoindentation to measure the mechanical properties of the films. The following tools have been used in these analyses: a Hitachi model S800 scanning electron microscope, a Bruker AXS microdiffractometer for µ-XRD, a National Electrostatics Corporation Tandem Ion Accelerator with Phi Evans Analytical End Station for RBS, a Hysitron Triboscope, and Vickers Microhardness tester for mechanical property evaluation. Results obtained from these experiments are presented in the following section.
Results and discussion Figure 2 shows micrographs of films deposited at Ts = 250◦ C and Ts = 700◦ C on ‘normal’ type (i) substrates. We see that the grain sizes and microstructures are similar in both films, indicating that grain growth is negligible even at temperatures as high as 700◦ C. This is not a total surprise considering that films deposited at elevated temperatures are exposed
34
Figure 2. Scanning electron micrographs of two SiC films deposited at low and high substrate temperatures, respectively Ts = 250◦ C, and Ts = 750◦ C.
to those temperatures for periods only on the order of a minute, following which they rapidly cool down to room temperature. What is unexpected, however, is that even this brief exposure to elevated temperatures, while clearly not sufficient to induce grain growth, has the ability to effect other interesting, even desirable, changes in film properties, such as phase, composition, and mechanical properties. For example, it has been
observed that films deposited on cold polished substrates (Ts ≈ 250◦ C) tend to delaminate easily (one film literally disintegrated into bits when probed by a profilometer stylus), while those deposited on hot polished substrates (Ts ≈ 700◦ C) are more adherent and tend to powder off rather than delaminate. With films deposited at lower temperatures it is possible to delaminate pieces of the film to observe the film cross-section
35
Figure 3. SEM images of the layered SiC film deposited at low substrate temperature (Ts = 250◦ C). The layer closest to the substrate appears to be well consolidated and has a ‘glassy’ appearance, while material deposited later appears to be more fractured.
under a microscope. Figure 3 shows SEM images of the cross-section of a low-temperature film (Ts ≈ 250◦ C). The film appears well consolidated and has a glassy appearance close to the substrate, which changes to a more fractured appearance with additional deposition. The substrate temperature also has an effect on the phase of the material synthesized. Figure 4 shows results from micro-X-ray diffraction (µ-XRD) analysis of two films deposited on ‘break-apart’ type (ii) substrates at two different substrate temperatures. The deposit formed at low temperature (Ts = 250◦ C) is largely amorphous, though a few small, barely detectable peaks are seen. These minor peaks, which are also observed in another low-temperature deposit, do not correspond to any of the possible product materials (e.g. β-SiC, α-SiC, Mo2 C) or contaminants (BN from the nozzle, W from the cathode). The only material with peaks at the same locations is a high-pressure metastable phase of silicon observed in one previous study (Leonidova, 1980). The unidentified peaks (labelled provisionally as Si) are found to disappear when the sample is delaminated from the surface. While at present we have no explanation for this observation, the possibility that the low-temperature film contains a metastable silicon phase which has been frozen in during deposition is being investigated further. Efforts are also being made to measure film stress,
since the spontaneous disintegration observed in some films deposited at low Ts suggests that film stresses are high. Very high film stresses (on the order of several GPa) are possible (Moody et al., 1998), and could aid in maintaining a high-pressure metastable phase. Figure 4 clearly shows that a high deposition temperature favors the formation of crystalline phases. The spectrum from a film deposited onto a ‘break-apart’ substrate at the higher substrate temperature (Ts = 700◦ C) reveals the presence of two crystalline phases (β-SiC and Mo2 C). This is representative of all films deposited at these temperatures. During the deposition of this particular high temperature film, the removable part of the ‘break-apart’ substrate glowed red, indicating that it had overheated due to poor thermal contact with the water-cooled holder. The XRD spectrum from this portion of the film has also been plotted in Figure 4 (labelled as ‘Ts > 700◦ C’), and is similar to that obtained from the remainder of the film, suggesting that increases in substrate temperature beyond 700◦ C do not greatly affect the crystallinity of the deposit. The change from amorphous to crystalline SiC with increasing deposition temperatures is consistent with the trend found in other processes (El Khakani et al., 1994). In HPPD, as ‘hot’ particles deposit on the substrate, they rapidly equilibrate to the temperature of the substrate (with picosecond timescales). This substrate
36
Figure 4. X-ray diffraction spectra from nanostructured SiC films deposited on cold (Ts = 250◦ C) and hot (Ts = 700◦ C) substrates. The spectrum labelled ‘Ts > 700◦ C’ was obtained from the film on the hot substrate, formed on a portion of the substrate which overheated due to poor thermal contact with the water-cooled holder.
quenching is particularly intense at low substrate temperatures, and favors the formation of amorphous phases. At higher substrate temperatures, the overall temperature change is smaller, the activation energy barrier is lower and there may be sufficient time for crystallization to occur. In addition to crystalline SiC, a second crystalline phase, Mo2 C, has been found in the high temperature deposits but not in the low temperature deposits. This could be related to either the stability of the oxide film on the molybdenum substrate, or more likely, marginal diffusion kinetics at the lower substrate temperature. As discussed later, RBS analysis of films deposited at similarly high temperatures (Ts ∼ 700◦ C) has shown no molybdenum close to the top surface of the film, within a distance corresponding to the penetration depth of the
probe beam (∼6 µm). This suggests that the Mo2 C is formed in an interface layer close to the molybdenum substrate and may be formed by solid state diffusion of Mo and C. The formation of such an interface layer is likely to improve the bonding between the SiC film and the substrate and could explain the relatively better film adhesion observed at high Ts , based on results from microindentation tests described below. The effect of deposition substrate temperature on chemical composition is seen from results from RBS analysis of three films deposited on the ‘small’ substrate of type (iii). The nominal deposition substrate temperature for these three films ranged from 450 to 700◦ C. Non-Rutherford RBS with a probe beam of 3.8 MeV He++ ions has been used to enhance the carbon signal. The probe beam diameter is about 2 mm.
37 Spectra from a pure silicon sample (silicon wafer) and a pure carbon sample (graphite) have been used to determine the energy calibration of the detection system and also to offset the uncertainty in charge and detector solid angle measurements as well as the collision crosssection deviations from the pure RBS regime. The single crystal silicon sample was tilted and rotated in
order to prevent ion beam channeling. Quantitative data obtained from these measurements are tabulated in Table 2. The corresponding spectra are plotted in Figure 5. As the penetration depth for this high energy beam is on the order of 6 µm, a signal from the molybdenum substrate is not observed for these relatively thick
Table 2. Results from Rutherford backscattering spectrometry for three SiC films, and for Si and C standard reference samples Sample
Atomic ratio
Atomic concentration (1022 atoms/cm3 ) Si
SiC (Ts = 450◦ C) SiC (Ts = 540◦ C) SiC (Ts = 700◦ C) Carbon Silicon Ideal SiC
C
Density (atoms/cm3 )
Density (g/cm3 )
Density (% ideal)
Cl
Si1 C1.04 Cl0.17
3.10
3.22
0.52
6.84 × 1022
2.389
74%
Si1 C1.03 Cl0.16
3.29
3.40
0.54
7.23 × 1022
2.523
78%
Si1 C0.97 Cl0.03
3.82
3.71
0.11
7.64 × 1022
2.581
80%
C Si Si1 C1
— 4.98 4.85
11.30 — 4.85
— — —
1.13 × 1023 4.98 × 1022 9.70 × 1022
— — —
Figure 5. Spectra from Rutherford backscattering analysis of three nanostructured silicon carbide films are stacked together for comparison. Spectra from reference silicon and carbon material are also plotted.
38 films. All three samples have close to ideal Si–C stoichiometry, though noticeable amounts of chlorine are seen in all cases, perhaps adsorbed as HCl. None of the samples have measurable oxygen contamination. The amount of residual chlorine decreases from about 8.5% for the sample deposited at 450◦ C to 1.5% for the sample deposited at 700◦ C, suggesting higher deposition temperatures may be used to drive off the chlorine. As hydrogen is lighter than helium it cannot be detected in backscattering experiments. Forward scattering experiments are planned in the future to determine hydrogen content. The RBS analysis also has allowed us to obtain the first measurements of density for our as-deposited films, based on measured atomic concentrations of component elements in the film. The data presented in Table 2 show that the film density increases gradually with deposition substrate temperature, reaching a value of about 80% of the theoretical density of bulk silicon carbide for the case Ts = 700◦ C. The morphology observed in Figure 2 and the observed trend in density suggest that some sintering may be occurring at higher values of Ts . However it appears that post-processing is required to further densify the film to values on the order of 97% of theoretical density that have been obtained for nanostructured ceramics (Mayo, 1996). The effect of substrate temperature on mechanical properties of the as-deposited films has been tested through microindentation and nanoindentation tests, performed on films deposited on the ‘small’ type (iii) substrates. Indentation with a conventional microhardness tester has been carried out for a zero-order
qualitative assessment of the temperature effects on the adhesion strength of as-deposited films. A Vickers indenter tip with an indentation load of 1 kg has been used to test two films of similar thickness, but deposited at different substrate temperatures. Images of indentation marks obtained with these films are shown in the first two panels of Figure 6, along with the corresponding value of Ts . An image of the indentation mark obtained on the molybdenum substrate is shown for comparison in the third panel of Figure 6. For the sample deposited at Ts = 700◦ C, indentation produced only cracking and partial delamination of the film. In contrast, extensive spalling of the film has been observed for the sample deposited at Ts = 540◦ C. These results indicate that better film properties (specifically adhesion) may be obtained at higher deposition temperatures. For a more quantitative assessment of mechanical properties nanoindentation tests have been performed on three as-deposited films using a Hysitron Triboscope used in conjunction with an atomic force microscope (AFM). This has allowed us to combine nanoindentation with the ability to image the indented area. The nanoindentation tests have been performed with a conical 90-degree indenter having a tip radius of 400 nm. The peak load for all three films is 6000 µN. The substrate temperature ranges from 450◦ C to 700◦ C for these films. Figure 7 shows images of an indented area scanned by the AFM before and after the test. Examples of load vs. displacement curves are seen in Figure 8 for the case of the film deposited at 700◦ C. It is possible to calculate hardness and Young’s modulus from the
Figure 6. SEM images of indentation marks made during microindentation tests on two silicon carbide films deposited at different temperatures. Also shown for comparison is an indentation mark obtained from the molybdenum substrate. The SiC film deposited at Ts = 540◦ C spalls off, while the film deposited at the higher temperature (Ts = 700◦ C) cracks but remains adherent to the substrate.
39
Figure 7. Hysitron AFM images of nanostructured SiC film before and after a nanoindentation test.The indentation mark is clearly seen. The probe tip has a radius of 400 nm, and features in the film finer than this size are not resolved.
Figure 8. Representative load vs. displacement curves obtained at three different locations during nanoindentation tests on a nanostructured SiC film deposited at a substrate temperature of 700◦ C.
40 Table 3. Results from nanoindentation tests performed on three as-deposited SiC films. The hardness, H , and Young’s modulus, E, of bulk SiC and Mo are also listed for comparison Sample
Load, µN
Depth, nm
No. of tests
H [GPa] (Average)
SiC Ts = 450◦ C SiC Ts = 540◦ C SiC Ts = 700◦ C ” ”
6000
210–280
9
13.5
6000
350–410
6
6000
180–210
4000 2000
138–150 110
E [GPa] (Average)
E [GPa] (Range)
5.2–17.9
103.5
89–130
10.7
9.6–16.3
70.1
66.3–78.4
5
20.4
17.0–22.6
157.2
147.5–165.3
5 5
16.9 12.1
15.8–18.1 9.8–15.0
152.2 150.3
144.0–166.5 128–154.1
Standard Values from Askeland (1989); SiC Barsoum (1997) Molybdenum From Metals Handbook (ASME Intntl)
unloading segments of the curve. The method reported by Oliver & Pharr (1992) has been used in the present study. Results from several indentation experiments on the three films are tabulated in Table 3. The films deposited at Ts = 450◦ C and Ts = 700◦ C are of comparable thickness (estimated to be between 6 and 10 µm), while the film deposited at 540◦ C is significantly thicker, about 15 µm. Note that for the film deposited at Ts = 700◦ C, additional indentation tests have been performed using three different peak loads, resulting in different penetration depths. The observed decrease of the measured property values with the decreasing depth of penetration is partly attributable to surface roughness effects. The results could also indicate that the hardness is not constant with depth, and therefore comparisons can only be made for films of comparable thickness, e.g. those deposited at Ts = 450◦ C and Ts = 700◦ C. For these films, it is seen that the mechanical properties of the deposited films improve with increased substrate temperature. It is not entirely clear why the thickest film, i.e. the one deposited at Ts = 540◦ C shows a drop in mechanical property values. Previous work (Rao et al., 1998) has shown that the silicon carbide films generally have a layered structure, believed to be caused by thermal stress induced fracture during deposition. Given that silicon carbide is a poor thermal conductor, and that temperature gradients and the resulting thermal stresses increase with film thickness, it should perhaps not be surprising that poorer mechanical properties are observed for thicker films. The layered structure of
H [GPa] (Range)
26–36
414
1.8–2.5
325
the silicon carbide could also explain the variation of hardness with depth observed for the film deposited at Ts = 700◦ C. In addition to the measured properties of the as-deposited SiC films, Table 3 also lists the range of properties for conventional bulk silicon carbide, and molybdenum. Note that the hardness and Young’s modulus of the as-deposited films are below that of conventional hot-pressed SiC. This too is not particularly surprising, considering that the deposited films are only partially consolidated, having a density of up to 80% of the ideal value. Porosity is known to reduce hardness in nanostructured films. The effect of porosity on Young’s modulus can be evaluated using the empirical relation obtained by Knudsen (1962), Spriggs (1961), Spriggs & Brissette (1962), E = Eo · e−bP
(1)
where Eo is the Young’s modulus of the standard material, E is the modulus of the porous material, P is the porosity, and b is a constant on the order of 4 for ceramic compacts. Assuming Eo = 414 GPA for standard SiC (Askeland, 1989), and b = 4.79 (reported for MgO by Spriggs et al., 1962), Equation (1) has been used to calculate values of the Young’s modulus for the standard material at porosity levels equivalent to the evaluated films. These calculated values of E are shown in Table 4, along with modulus values measured for the corresponding film, taken from Table 3. It is seen that these measured values are in good agreement with the
41 Table 4. Calculated values of Young’s modulus for standard SiC with porosity compared with the corresponding value for the HPPD films Porosity
E [GPa] Standard SiC
E [GPa] HPPD film
20% 26%
160 121
157 104
Acknowledgements This work was partially supported by NSF (CTS9520147), the Engineering Research Center for Plasma-Aided Manufacturing (NSF ECD-87-21545) and the Minnesota Supercomputer Institute. Z. Wong acknowledges support from the NSF-sponsored Undergraduate Research Opportunity Program (UROP) at the University of Minnesota. References
values for the standard material with the same porosity levels. The above results are the first mechanical property measurements for nanostructured films deposited by HPPD. While the results are promising, it is clear that some post-processing (i.e. densification) of our films is necessary for film properties to reach or exceed that of conventional bulk materials. Experiments are planned to accomplish densification by heat treatment and hot isostatic pressing since such treatments have been known to improve the properties of films deposited by related processes such as vacuum plasma spray deposition (Huber et al., 1987), and by plasma spray pyrolysis (Karthikeyan et al., 1997).
Conclusion Examination of several nanostructured SiC films deposited by HPPD at different substrate temperatures supports the view that substrate temperature may be used as a key parameter to tailor the properties of the deposited film. While the effect of substrate temperature on grain size has been found to be insignificant for the range of substrate temperatures studied (250◦ C < Ts < 700◦ C), it has been observed that substrate temperature affects film composition and mechanical properties. Higher substrate temperatures favor the formation of crystalline films, whereas predominantly amorphous films have been formed at low substrate temperatures. Stoichiometric silicon carbide has been formed over a broad range of substrate temperatures (450–700◦ C), but deposition at higher temperatures greatly reduces the level of chlorine contamination. For films of comparable thickness, the substrate temperature has also affected the mechanical properties of the deposited films, with superior hardness and Young’s modulus being obtained at higher substrate temperatures.
Askeland, D.R., 1989. The Science and Engineering of Materials, PWS-KENT Publishing Co., Boston, p. 500. Barsoum, M., 1997. Fundamentals of Ceramics, McGraw Hill. Bianchi, L., P. Lucchese, A. Denoirjean, P. Fauchais & S. Kuroda, 1995. Evolution of quenching stress during ceramic thermal spraying with respect to plasma parameters, Proc. 8th National Thermal Plasma Spray Conference, 11–15 September 1995, Houston, Texas. Bieberich, M.T. & S.L. Girshick, 1996. Control of substrate temperature during diamond deposition, Plasma Chem. Plasma Proc. 16, 157S–168S. Delplancke, M.P., J.M. Powers, G.J. Vandentop, M. Salmeron & G.A. Somorjai, 1991. Preperation and characterization of amorphous SiC : H thin films. J. Vac. Sci. Technol. A9, 450–455. El Khakani, M.A., M. Chaker, A. Jean, S. Boily, & J.C. Keiffer, 1994. Hardness and Young’s modulus of amorphous a-SiC thin films determined by nanoindentation and bulge tests. J. Mater. Res. 9(1), 96–103. Huber, P., H. Kress & H. Kaiser, Vacuum plasma spraying in combination with hot isostatic pressing for the repair of fabrication of superalloy components, In Proc. 12th Nat. Thermal Spray Conf. (Orlando, FL), Sept. 13–17, 1987, pp.155–167. Karthikeyan, J., C.C. Berndt, J. Tikkanen, S. Reddy & H. Herman, 1997. Plasma spray synthesis of nanomaterial powders and deposits. Mat. Sci. Eng. A238, 275–286. Knudsen, F.P., 1962. Effect of Porosity on Young’s Modulus of Alumina. J. Am. Ceram. Soc. 45, 94. Kong, P.C. & E. Pfender, 1987. Formation of Ultrafine β-Silicon Carbide Powders in an Argon Thermal Plasma Jet. Langmuir 3, 259–265. Leonidova, G.G., 1980. New metastable high-pressure phase of doped n-type silicon. Sov. Phys. Solid State 22(3), 559–560. Mayo, M.J., 1996. Processing of nanocrystalline ceramics from ultrafine particles. Int. Mater. Rev. 41(3), 85–115. Metals Handbook, 10th Ed., Vol. 2., ASME International 1990. Moody, N.R., R.Q. Hwang, Venkataraman, J.E. Angelo, D.P. Norwood, & G.W. Gerberich, 1998. Adhesion and fracture of tantalum nitride films. Acta Mater. 46(2), 585–597. Oliver, W.C. & G.M. Pharr, 1992. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation. J. Mater. Res. 7, 1564. Rao, N.P., N. Tymiak, J. Blum, A. Neuman, H.J. Lee, S.L. Girshick, P.H. McMurry & J. Heberlein, 1998. Hypersonic Plasma Particle Deposition of Nanostructured Silicon and Silicon Carbide. J. Aerosol Sci. 29, 707–720.
42 Rao, N.P., H.J. Lee, M. Kelkar, D.J. Hansen, J.V.R. Heberlein, P.H. McMurry & S.L. Girshick, 1997a. Nanostructured materials production by hypersonic plasma particle deposition. Nanostructured Materials 9, 129–132. Rao, N.P., A. Neuman, J. Blum, N. Tymiak, J. Heberlein, P.H. McMurry & S.L. Girshick, 1997b. Thermal plasma spraying of nanostructured coatings by hypersonic plasma particle deposition. Proc. 10th Natl. Thermal spray conf., Indianapolis, IN, 15–19 Sep. 1997. In: C.C. Berndt, ed., Thermal Spray: A united forum for scientific and technological advances, ASM Intl., Materials Park, OH, pp. 329–333.
Siegel, R.W., 1993. Synthesis and properties of nanophase materials. Mater. Sci. Eng. A168. 189–197. Spriggs, R.M., 1961. Expression for effect of porosity on elastic modulus of polycrystalline refractory materials. J. Am. Ceram. Soc. 44, 94. Spriggs, R.M. & L.A. Brissette, 1962. Expressions for shear modulus and Poisson’s Ratio of porous refractory oxides. J. Am. Ceram. Soc. 45, 198. Spriggs, R.M., L.A. Brissette & T. Vasilos, 1962. Effect of porosity on elastic and shear moduli of polycrystalline magnesium oxide. J. Am. Ceram. Soc. 45, 400.
The effect of substrate temperature on the properties of nanostructured silicon carbide films deposited by hypersonic plasma particle deposition J. Blum1 , N. Tymiak2 , A. Neuman1 , Z. Wong1 , N.P. Rao1 , S.L. Girshick1 , W.W. Gerberich2 , P.H. McMurry1 and J.V.R. Heberlein1 1 Department of Mechanical Engineering, 2 Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, MN 55455, USA Received 14 July 1998; accepted in revised form 12 January 1999
Key words: nanoparticles, thermal plasma, nanostructural film, particle deposition, silicon carbide, film hardness
Abstract Nanostructured silicon carbide films have been deposited on molybdenum substrates by hypersonic plasma particle deposition. In this process a thermal plasma with injected reactants (SiCl4 and CH4 ) is expanded through a nozzle leading to the nucleation of ultrafine particles. Particles entrained in the supersonic flow are then inertially deposited in vacuum onto a temperature-controlled substrate, leading to the formation of a consolidated film. In the experiments reported, the deposition substrate temperature Ts has ranged from 250◦ C to 700◦ C, and the effect of Ts on film morphology, composition, and mechanical properties has been studied. Examination of the films by scanning electron microscopy has shown that the grain sizes in the films did not vary significantly with Ts . Micro-X-ray diffraction analysis of the deposits has shown that amorphous films are deposited at low Ts , while crystalline films are formed at high Ts . Rutherford backscattering spectrometry has indicated that the films are largely stoichiometric silicon carbide with small amounts of chlorine. The chlorine content decreases from 8% to 1.5% when the deposition temperature is raised from 450◦ C to 700◦ C. Nanoindentation and microindentation tests have been performed on as-deposited films to measure hardness, Young’s modulus and to evaluate adhesion strength. The tests show that film adhesion, hardness and Young’s modulus increase with increasing Ts . These results taken together demonstrate that in HPPD, as in vapor deposition processes, the substrate temperature may be used to control film properties, and that better quality films are obtained at higher substrate temperatures, i.e. Ts ≈ 700◦ C.
Introduction Nanostructured materials have been found to possess mechanical properties, e.g. hardness and yield strength, superior to that of conventional coarse grained materials (Siegel, 1993). There is currently great interest in the synthesis and processing of such materials, particularly in creating functional forms such as films, coatings, composites, etc. We have recently developed a new process for the production of nanostructured
films from a thermal plasma route (Rao et al., 1997a; Rao et al., 1998). Thermal plasma synthesis provides a scalable, continuous flow processing route which is characterized by high temperatures, high energy density, fast reaction kinetics, rapid quenching, high processing rates and is especially well suited to the synthesis of ultrafine particles of non-oxide, hightemperature ceramics such as carbides, nitrides and borides (Kong & Pfender, 1987). The present work concerns the synthesis of nanostructured silicon carbide
32 films, which are formed by depositing freshly synthesized SiC nanoparticles onto a metal (molybdenum) substrate. Such films are potential candidates for wear resistant coatings. In the process being described, hypersonic plasma particle deposition (HPPD), vapor phase precursors are injected into a flowing thermal plasma generated by a dc arc torch. The plasma is then rapidly quenched by supersonic expansion in a ceramic lined nozzle, resulting in the nucleation and growth of nanoparticles. The nanoparticles are further accelerated in the hypersonic free jet downstream of the nozzle, and are deposited by inertial impaction onto a temperature-controlled substrate. The high particle deposition velocity results in the formation of a lightly consolidated nanostructured film, while the small particle residence times (∼ 50 µs) minimize in-flight agglomeration and contamination. A comparison of HPPD with related processes such as thermal plasma spray and thermal plasma chemical vapor deposition has been provided in Rao et al. (1997b). Earlier work on the HPPD process has demonstrated that the grain size in the film correlates well with the size of particles formed in the nozzle (Rao et al., 1997b). For this reason considerable effort was initially focused on controlling particle size in the nozzle, and thus the grain size in the deposit, through control of process parameters such as reactant flow rate, torch power, nozzle length, etc. In order to prevent significant grain growth from occurring, the substrate temperature was generally kept at about 200◦ C by active cooling. Recent experimental work, however, has indicated that the deposition substrate temperature is a key parameter in determining the properties of the product film. This paper describes results from the first experimental exploration of the effects of deposition substrate temperature on the properties of nanostructured films deposited by HPPD. The effect of substrate temperature on deposited films is well known in many other related processes. For example, amorphous silicon carbide films deposited by low pressure plasma enhanced chemical vapor deposition (PECVD) show significantly higher densities at higher deposition substrate temperatures (Delplancke et al., 1991). El Khakani et al. (1994) mention that CVD SiC films deposited at temperatures lower than 800◦ C are generally amorphous, while those formed at higher deposition temperatures are generally crystalline. Substrate temperature has also been found to be an important parameter affecting properties such
as the residual quenching stress in coatings deposited by thermal plasma spray (Bianchi et al., 1995). In this process, it is well known that stresses due to differential expansion are induced when a hot coated workpiece cools down to room temperature. Stresses are also induced when large micron-sized molten particles impacting on the substrate flatten out forming a splat which intertwines with other splats. Substrate temperature affects the wetting angle between the workpiece and the splat as well as the shrinkage of the splat when quenched by contact with the colder workpiece. Since the HPPD process has elements in common with PECVD and conventional thermal spray processes, it is likely that similar deposition substrate temperature effects may be observed in HPPD. In this paper we describe experiments which tested the hypothesis that the deposition substrate temperature may provide a useful means of tailoring the properties of the deposited film. In the sections following, we first describe the experimental apparatus used in the study, discuss the results obtained, and draw conclusions.
Experimental details Figure 1 shows the HPPD reactor which has been discussed in detail elsewhere (Rao et al., 1998). It consists of a thermal plasma reactor exhausting into a low pressure deposition chamber, along with associated systems for pumping, gas handling, power supply, data acquisition, diagnostics, scrubbing and safety. The plasma reactor assembly consists of a dc arc torch coupled to a ceramic reactant injection tube and converging nozzle having an entrance diameter of 15 mm and an exit diameter of 5 mm. Within the deposition chamber, a molybdenum substrate is clamped to a water-cooled holder at a distance 20 mm downstream of the nozzle exit and oriented so that the substrate surface faces the flow issuing from the nozzle. The temperature of the substrate is controlled by injecting a helium–argon gas mixture of controllable composition between the holder and the substrate (Bieberich & Girshick, 1996). Three types of substrates have been used in the experiments reported here: (i) a ‘normal’ substrate, a polished molybdenum disk 20 mm in diameter and 3 mm thick; (ii) a ‘break-apart’ molybdenum substrate similar in size to the ‘normal’ substrate, but provided with a removable central section held together by screws, and (iii) a ‘small’ unpolished molybdenum substrate
33 Table 1. Representative process conditions for HPPD experiments Process parameter
Value ◦
Substrate temperature, C SiCl4 flow rate, slm CH4 flow rate, slm Argon flow rate, slm Hydrogen flow rate, slm Arc current, A Nozzle–plate distance, mm Chamber pressure, torr
Figure 1. Schematic diagram of the HPPD reactor (Rao et al., 1998).
4 mm thick with a deposition area 10 mm in diameter. The ‘break-apart’ substrate has been designed to enable easier observation and analysis of the cross-section of deposited films, while the ‘small’ substrate has been designed to allow access to specialized materials analysis tools capable of handling only small substrates. The substrate temperature in all cases has been measured using a sheathed thermocouple probe embedded deep into a thermocouple well drilled into the side of the substrate and accessed through a hole drilled into the substrate clamping ring. With this arrangement, the thermocouple junction lay less than 1 mm below the top surface of the substrate disk. A portion of the thermocouple probe outside the substrate assembly was exposed to the plasma and tended to heat up during a deposition run. We did not attempt to correct for associated ‘finning’ effects in measuring the substrate temperature, and thus the temperatures reported here should be considered to be upper bound values. The finning effects are expected to be greatest for the case of the small substrates of type (iii). The experimental protocol for depositing nanostructured films by HPPD has been reported in previous
250–750 0.2 1.0 35 7.5 225 20 2.5
publications (Rao et al., 1998). In order to synthesize SiC films, SiCl4 vapor and CH4 reactants are injected into an Ar–H2 plasma. Representative operating conditions are listed in Table 1. A movable shutter plate within the deposition chamber protects the substrate assembly and thermocouple from exposure to the plasma before and after the deposition. Deposition times vary from 60–90 s for the experiments reported, with the substrate and deposited film cooling down rapidly to room temperature after deposition. The deposited films have been analyzed by scanning electron microscopy (SEM) to determine grain size and microstructure, by micro-X-ray diffraction (µ-XRD) to obtain phase information, by Rutherford backscattering spectrometry (RBS) to quantify elemental composition, and by micro- and nanoindentation to measure the mechanical properties of the films. The following tools have been used in these analyses: a Hitachi model S800 scanning electron microscope, a Bruker AXS microdiffractometer for µ-XRD, a National Electrostatics Corporation Tandem Ion Accelerator with Phi Evans Analytical End Station for RBS, a Hysitron Triboscope, and Vickers Microhardness tester for mechanical property evaluation. Results obtained from these experiments are presented in the following section.
Results and discussion Figure 2 shows micrographs of films deposited at Ts = 250◦ C and Ts = 700◦ C on ‘normal’ type (i) substrates. We see that the grain sizes and microstructures are similar in both films, indicating that grain growth is negligible even at temperatures as high as 700◦ C. This is not a total surprise considering that films deposited at elevated temperatures are exposed
34
Figure 2. Scanning electron micrographs of two SiC films deposited at low and high substrate temperatures, respectively Ts = 250◦ C, and Ts = 750◦ C.
to those temperatures for periods only on the order of a minute, following which they rapidly cool down to room temperature. What is unexpected, however, is that even this brief exposure to elevated temperatures, while clearly not sufficient to induce grain growth, has the ability to effect other interesting, even desirable, changes in film properties, such as phase, composition, and mechanical properties. For example, it has been
observed that films deposited on cold polished substrates (Ts ≈ 250◦ C) tend to delaminate easily (one film literally disintegrated into bits when probed by a profilometer stylus), while those deposited on hot polished substrates (Ts ≈ 700◦ C) are more adherent and tend to powder off rather than delaminate. With films deposited at lower temperatures it is possible to delaminate pieces of the film to observe the film cross-section
35
Figure 3. SEM images of the layered SiC film deposited at low substrate temperature (Ts = 250◦ C). The layer closest to the substrate appears to be well consolidated and has a ‘glassy’ appearance, while material deposited later appears to be more fractured.
under a microscope. Figure 3 shows SEM images of the cross-section of a low-temperature film (Ts ≈ 250◦ C). The film appears well consolidated and has a glassy appearance close to the substrate, which changes to a more fractured appearance with additional deposition. The substrate temperature also has an effect on the phase of the material synthesized. Figure 4 shows results from micro-X-ray diffraction (µ-XRD) analysis of two films deposited on ‘break-apart’ type (ii) substrates at two different substrate temperatures. The deposit formed at low temperature (Ts = 250◦ C) is largely amorphous, though a few small, barely detectable peaks are seen. These minor peaks, which are also observed in another low-temperature deposit, do not correspond to any of the possible product materials (e.g. β-SiC, α-SiC, Mo2 C) or contaminants (BN from the nozzle, W from the cathode). The only material with peaks at the same locations is a high-pressure metastable phase of silicon observed in one previous study (Leonidova, 1980). The unidentified peaks (labelled provisionally as Si) are found to disappear when the sample is delaminated from the surface. While at present we have no explanation for this observation, the possibility that the low-temperature film contains a metastable silicon phase which has been frozen in during deposition is being investigated further. Efforts are also being made to measure film stress,
since the spontaneous disintegration observed in some films deposited at low Ts suggests that film stresses are high. Very high film stresses (on the order of several GPa) are possible (Moody et al., 1998), and could aid in maintaining a high-pressure metastable phase. Figure 4 clearly shows that a high deposition temperature favors the formation of crystalline phases. The spectrum from a film deposited onto a ‘break-apart’ substrate at the higher substrate temperature (Ts = 700◦ C) reveals the presence of two crystalline phases (β-SiC and Mo2 C). This is representative of all films deposited at these temperatures. During the deposition of this particular high temperature film, the removable part of the ‘break-apart’ substrate glowed red, indicating that it had overheated due to poor thermal contact with the water-cooled holder. The XRD spectrum from this portion of the film has also been plotted in Figure 4 (labelled as ‘Ts > 700◦ C’), and is similar to that obtained from the remainder of the film, suggesting that increases in substrate temperature beyond 700◦ C do not greatly affect the crystallinity of the deposit. The change from amorphous to crystalline SiC with increasing deposition temperatures is consistent with the trend found in other processes (El Khakani et al., 1994). In HPPD, as ‘hot’ particles deposit on the substrate, they rapidly equilibrate to the temperature of the substrate (with picosecond timescales). This substrate
36
Figure 4. X-ray diffraction spectra from nanostructured SiC films deposited on cold (Ts = 250◦ C) and hot (Ts = 700◦ C) substrates. The spectrum labelled ‘Ts > 700◦ C’ was obtained from the film on the hot substrate, formed on a portion of the substrate which overheated due to poor thermal contact with the water-cooled holder.
quenching is particularly intense at low substrate temperatures, and favors the formation of amorphous phases. At higher substrate temperatures, the overall temperature change is smaller, the activation energy barrier is lower and there may be sufficient time for crystallization to occur. In addition to crystalline SiC, a second crystalline phase, Mo2 C, has been found in the high temperature deposits but not in the low temperature deposits. This could be related to either the stability of the oxide film on the molybdenum substrate, or more likely, marginal diffusion kinetics at the lower substrate temperature. As discussed later, RBS analysis of films deposited at similarly high temperatures (Ts ∼ 700◦ C) has shown no molybdenum close to the top surface of the film, within a distance corresponding to the penetration depth of the
probe beam (∼6 µm). This suggests that the Mo2 C is formed in an interface layer close to the molybdenum substrate and may be formed by solid state diffusion of Mo and C. The formation of such an interface layer is likely to improve the bonding between the SiC film and the substrate and could explain the relatively better film adhesion observed at high Ts , based on results from microindentation tests described below. The effect of deposition substrate temperature on chemical composition is seen from results from RBS analysis of three films deposited on the ‘small’ substrate of type (iii). The nominal deposition substrate temperature for these three films ranged from 450 to 700◦ C. Non-Rutherford RBS with a probe beam of 3.8 MeV He++ ions has been used to enhance the carbon signal. The probe beam diameter is about 2 mm.
37 Spectra from a pure silicon sample (silicon wafer) and a pure carbon sample (graphite) have been used to determine the energy calibration of the detection system and also to offset the uncertainty in charge and detector solid angle measurements as well as the collision crosssection deviations from the pure RBS regime. The single crystal silicon sample was tilted and rotated in
order to prevent ion beam channeling. Quantitative data obtained from these measurements are tabulated in Table 2. The corresponding spectra are plotted in Figure 5. As the penetration depth for this high energy beam is on the order of 6 µm, a signal from the molybdenum substrate is not observed for these relatively thick
Table 2. Results from Rutherford backscattering spectrometry for three SiC films, and for Si and C standard reference samples Sample
Atomic ratio
Atomic concentration (1022 atoms/cm3 ) Si
SiC (Ts = 450◦ C) SiC (Ts = 540◦ C) SiC (Ts = 700◦ C) Carbon Silicon Ideal SiC
C
Density (atoms/cm3 )
Density (g/cm3 )
Density (% ideal)
Cl
Si1 C1.04 Cl0.17
3.10
3.22
0.52
6.84 × 1022
2.389
74%
Si1 C1.03 Cl0.16
3.29
3.40
0.54
7.23 × 1022
2.523
78%
Si1 C0.97 Cl0.03
3.82
3.71
0.11
7.64 × 1022
2.581
80%
C Si Si1 C1
— 4.98 4.85
11.30 — 4.85
— — —
1.13 × 1023 4.98 × 1022 9.70 × 1022
— — —
Figure 5. Spectra from Rutherford backscattering analysis of three nanostructured silicon carbide films are stacked together for comparison. Spectra from reference silicon and carbon material are also plotted.
38 films. All three samples have close to ideal Si–C stoichiometry, though noticeable amounts of chlorine are seen in all cases, perhaps adsorbed as HCl. None of the samples have measurable oxygen contamination. The amount of residual chlorine decreases from about 8.5% for the sample deposited at 450◦ C to 1.5% for the sample deposited at 700◦ C, suggesting higher deposition temperatures may be used to drive off the chlorine. As hydrogen is lighter than helium it cannot be detected in backscattering experiments. Forward scattering experiments are planned in the future to determine hydrogen content. The RBS analysis also has allowed us to obtain the first measurements of density for our as-deposited films, based on measured atomic concentrations of component elements in the film. The data presented in Table 2 show that the film density increases gradually with deposition substrate temperature, reaching a value of about 80% of the theoretical density of bulk silicon carbide for the case Ts = 700◦ C. The morphology observed in Figure 2 and the observed trend in density suggest that some sintering may be occurring at higher values of Ts . However it appears that post-processing is required to further densify the film to values on the order of 97% of theoretical density that have been obtained for nanostructured ceramics (Mayo, 1996). The effect of substrate temperature on mechanical properties of the as-deposited films has been tested through microindentation and nanoindentation tests, performed on films deposited on the ‘small’ type (iii) substrates. Indentation with a conventional microhardness tester has been carried out for a zero-order
qualitative assessment of the temperature effects on the adhesion strength of as-deposited films. A Vickers indenter tip with an indentation load of 1 kg has been used to test two films of similar thickness, but deposited at different substrate temperatures. Images of indentation marks obtained with these films are shown in the first two panels of Figure 6, along with the corresponding value of Ts . An image of the indentation mark obtained on the molybdenum substrate is shown for comparison in the third panel of Figure 6. For the sample deposited at Ts = 700◦ C, indentation produced only cracking and partial delamination of the film. In contrast, extensive spalling of the film has been observed for the sample deposited at Ts = 540◦ C. These results indicate that better film properties (specifically adhesion) may be obtained at higher deposition temperatures. For a more quantitative assessment of mechanical properties nanoindentation tests have been performed on three as-deposited films using a Hysitron Triboscope used in conjunction with an atomic force microscope (AFM). This has allowed us to combine nanoindentation with the ability to image the indented area. The nanoindentation tests have been performed with a conical 90-degree indenter having a tip radius of 400 nm. The peak load for all three films is 6000 µN. The substrate temperature ranges from 450◦ C to 700◦ C for these films. Figure 7 shows images of an indented area scanned by the AFM before and after the test. Examples of load vs. displacement curves are seen in Figure 8 for the case of the film deposited at 700◦ C. It is possible to calculate hardness and Young’s modulus from the
Figure 6. SEM images of indentation marks made during microindentation tests on two silicon carbide films deposited at different temperatures. Also shown for comparison is an indentation mark obtained from the molybdenum substrate. The SiC film deposited at Ts = 540◦ C spalls off, while the film deposited at the higher temperature (Ts = 700◦ C) cracks but remains adherent to the substrate.
39
Figure 7. Hysitron AFM images of nanostructured SiC film before and after a nanoindentation test.The indentation mark is clearly seen. The probe tip has a radius of 400 nm, and features in the film finer than this size are not resolved.
Figure 8. Representative load vs. displacement curves obtained at three different locations during nanoindentation tests on a nanostructured SiC film deposited at a substrate temperature of 700◦ C.
40 Table 3. Results from nanoindentation tests performed on three as-deposited SiC films. The hardness, H , and Young’s modulus, E, of bulk SiC and Mo are also listed for comparison Sample
Load, µN
Depth, nm
No. of tests
H [GPa] (Average)
SiC Ts = 450◦ C SiC Ts = 540◦ C SiC Ts = 700◦ C ” ”
6000
210–280
9
13.5
6000
350–410
6
6000
180–210
4000 2000
138–150 110
E [GPa] (Average)
E [GPa] (Range)
5.2–17.9
103.5
89–130
10.7
9.6–16.3
70.1
66.3–78.4
5
20.4
17.0–22.6
157.2
147.5–165.3
5 5
16.9 12.1
15.8–18.1 9.8–15.0
152.2 150.3
144.0–166.5 128–154.1
Standard Values from Askeland (1989); SiC Barsoum (1997) Molybdenum From Metals Handbook (ASME Intntl)
unloading segments of the curve. The method reported by Oliver & Pharr (1992) has been used in the present study. Results from several indentation experiments on the three films are tabulated in Table 3. The films deposited at Ts = 450◦ C and Ts = 700◦ C are of comparable thickness (estimated to be between 6 and 10 µm), while the film deposited at 540◦ C is significantly thicker, about 15 µm. Note that for the film deposited at Ts = 700◦ C, additional indentation tests have been performed using three different peak loads, resulting in different penetration depths. The observed decrease of the measured property values with the decreasing depth of penetration is partly attributable to surface roughness effects. The results could also indicate that the hardness is not constant with depth, and therefore comparisons can only be made for films of comparable thickness, e.g. those deposited at Ts = 450◦ C and Ts = 700◦ C. For these films, it is seen that the mechanical properties of the deposited films improve with increased substrate temperature. It is not entirely clear why the thickest film, i.e. the one deposited at Ts = 540◦ C shows a drop in mechanical property values. Previous work (Rao et al., 1998) has shown that the silicon carbide films generally have a layered structure, believed to be caused by thermal stress induced fracture during deposition. Given that silicon carbide is a poor thermal conductor, and that temperature gradients and the resulting thermal stresses increase with film thickness, it should perhaps not be surprising that poorer mechanical properties are observed for thicker films. The layered structure of
H [GPa] (Range)
26–36
414
1.8–2.5
325
the silicon carbide could also explain the variation of hardness with depth observed for the film deposited at Ts = 700◦ C. In addition to the measured properties of the as-deposited SiC films, Table 3 also lists the range of properties for conventional bulk silicon carbide, and molybdenum. Note that the hardness and Young’s modulus of the as-deposited films are below that of conventional hot-pressed SiC. This too is not particularly surprising, considering that the deposited films are only partially consolidated, having a density of up to 80% of the ideal value. Porosity is known to reduce hardness in nanostructured films. The effect of porosity on Young’s modulus can be evaluated using the empirical relation obtained by Knudsen (1962), Spriggs (1961), Spriggs & Brissette (1962), E = Eo · e−bP
(1)
where Eo is the Young’s modulus of the standard material, E is the modulus of the porous material, P is the porosity, and b is a constant on the order of 4 for ceramic compacts. Assuming Eo = 414 GPA for standard SiC (Askeland, 1989), and b = 4.79 (reported for MgO by Spriggs et al., 1962), Equation (1) has been used to calculate values of the Young’s modulus for the standard material at porosity levels equivalent to the evaluated films. These calculated values of E are shown in Table 4, along with modulus values measured for the corresponding film, taken from Table 3. It is seen that these measured values are in good agreement with the
41 Table 4. Calculated values of Young’s modulus for standard SiC with porosity compared with the corresponding value for the HPPD films Porosity
E [GPa] Standard SiC
E [GPa] HPPD film
20% 26%
160 121
157 104
Acknowledgements This work was partially supported by NSF (CTS9520147), the Engineering Research Center for Plasma-Aided Manufacturing (NSF ECD-87-21545) and the Minnesota Supercomputer Institute. Z. Wong acknowledges support from the NSF-sponsored Undergraduate Research Opportunity Program (UROP) at the University of Minnesota. References
values for the standard material with the same porosity levels. The above results are the first mechanical property measurements for nanostructured films deposited by HPPD. While the results are promising, it is clear that some post-processing (i.e. densification) of our films is necessary for film properties to reach or exceed that of conventional bulk materials. Experiments are planned to accomplish densification by heat treatment and hot isostatic pressing since such treatments have been known to improve the properties of films deposited by related processes such as vacuum plasma spray deposition (Huber et al., 1987), and by plasma spray pyrolysis (Karthikeyan et al., 1997).
Conclusion Examination of several nanostructured SiC films deposited by HPPD at different substrate temperatures supports the view that substrate temperature may be used as a key parameter to tailor the properties of the deposited film. While the effect of substrate temperature on grain size has been found to be insignificant for the range of substrate temperatures studied (250◦ C < Ts < 700◦ C), it has been observed that substrate temperature affects film composition and mechanical properties. Higher substrate temperatures favor the formation of crystalline films, whereas predominantly amorphous films have been formed at low substrate temperatures. Stoichiometric silicon carbide has been formed over a broad range of substrate temperatures (450–700◦ C), but deposition at higher temperatures greatly reduces the level of chlorine contamination. For films of comparable thickness, the substrate temperature has also affected the mechanical properties of the deposited films, with superior hardness and Young’s modulus being obtained at higher substrate temperatures.
Askeland, D.R., 1989. The Science and Engineering of Materials, PWS-KENT Publishing Co., Boston, p. 500. Barsoum, M., 1997. Fundamentals of Ceramics, McGraw Hill. Bianchi, L., P. Lucchese, A. Denoirjean, P. Fauchais & S. Kuroda, 1995. Evolution of quenching stress during ceramic thermal spraying with respect to plasma parameters, Proc. 8th National Thermal Plasma Spray Conference, 11–15 September 1995, Houston, Texas. Bieberich, M.T. & S.L. Girshick, 1996. Control of substrate temperature during diamond deposition, Plasma Chem. Plasma Proc. 16, 157S–168S. Delplancke, M.P., J.M. Powers, G.J. Vandentop, M. Salmeron & G.A. Somorjai, 1991. Preperation and characterization of amorphous SiC : H thin films. J. Vac. Sci. Technol. A9, 450–455. El Khakani, M.A., M. Chaker, A. Jean, S. Boily, & J.C. Keiffer, 1994. Hardness and Young’s modulus of amorphous a-SiC thin films determined by nanoindentation and bulge tests. J. Mater. Res. 9(1), 96–103. Huber, P., H. Kress & H. Kaiser, Vacuum plasma spraying in combination with hot isostatic pressing for the repair of fabrication of superalloy components, In Proc. 12th Nat. Thermal Spray Conf. (Orlando, FL), Sept. 13–17, 1987, pp.155–167. Karthikeyan, J., C.C. Berndt, J. Tikkanen, S. Reddy & H. Herman, 1997. Plasma spray synthesis of nanomaterial powders and deposits. Mat. Sci. Eng. A238, 275–286. Knudsen, F.P., 1962. Effect of Porosity on Young’s Modulus of Alumina. J. Am. Ceram. Soc. 45, 94. Kong, P.C. & E. Pfender, 1987. Formation of Ultrafine β-Silicon Carbide Powders in an Argon Thermal Plasma Jet. Langmuir 3, 259–265. Leonidova, G.G., 1980. New metastable high-pressure phase of doped n-type silicon. Sov. Phys. Solid State 22(3), 559–560. Mayo, M.J., 1996. Processing of nanocrystalline ceramics from ultrafine particles. Int. Mater. Rev. 41(3), 85–115. Metals Handbook, 10th Ed., Vol. 2., ASME International 1990. Moody, N.R., R.Q. Hwang, Venkataraman, J.E. Angelo, D.P. Norwood, & G.W. Gerberich, 1998. Adhesion and fracture of tantalum nitride films. Acta Mater. 46(2), 585–597. Oliver, W.C. & G.M. Pharr, 1992. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation. J. Mater. Res. 7, 1564. Rao, N.P., N. Tymiak, J. Blum, A. Neuman, H.J. Lee, S.L. Girshick, P.H. McMurry & J. Heberlein, 1998. Hypersonic Plasma Particle Deposition of Nanostructured Silicon and Silicon Carbide. J. Aerosol Sci. 29, 707–720.
42 Rao, N.P., H.J. Lee, M. Kelkar, D.J. Hansen, J.V.R. Heberlein, P.H. McMurry & S.L. Girshick, 1997a. Nanostructured materials production by hypersonic plasma particle deposition. Nanostructured Materials 9, 129–132. Rao, N.P., A. Neuman, J. Blum, N. Tymiak, J. Heberlein, P.H. McMurry & S.L. Girshick, 1997b. Thermal plasma spraying of nanostructured coatings by hypersonic plasma particle deposition. Proc. 10th Natl. Thermal spray conf., Indianapolis, IN, 15–19 Sep. 1997. In: C.C. Berndt, ed., Thermal Spray: A united forum for scientific and technological advances, ASM Intl., Materials Park, OH, pp. 329–333.
Siegel, R.W., 1993. Synthesis and properties of nanophase materials. Mater. Sci. Eng. A168. 189–197. Spriggs, R.M., 1961. Expression for effect of porosity on elastic modulus of polycrystalline refractory materials. J. Am. Ceram. Soc. 44, 94. Spriggs, R.M. & L.A. Brissette, 1962. Expressions for shear modulus and Poisson’s Ratio of porous refractory oxides. J. Am. Ceram. Soc. 45, 198. Spriggs, R.M., L.A. Brissette & T. Vasilos, 1962. Effect of porosity on elastic and shear moduli of polycrystalline magnesium oxide. J. Am. Ceram. Soc. 45, 400.
