1

Figure 1.1. The Diamond Lattice Figure 1.2. The Graphite Lattice.

At ambient temperatures and pressures, graphite, the sp²-bonded allotrope of carbon (Figure 1.2), is the stable allotrope of carbon. The standard enthalpies of formation of diamond and graphite differ only by 2.9 kJ mol^-1,^1.4 but a large activation barrier rules out simple thermal activation as a means of driving the graphite®diamond interconversion. Since diamond is the densest allotrope of carbon, at high pressure, diamond must be the stable form of solid carbon. This is the scientific basis for the high pressure high temperature (HPHT) techniques by which the so called ‘industrial diamond’ has already been synthesised commercially for over 30 years.^1.1,1.2,1.5 The advent of Chemical Vapour Deposition (CVD) of diamond has dramatically increased the possibility of exploitation of its other extraordinary properties and potential wide-ranging industrial applications. This technique uses, as process gases, a standard hydrocarbon gas typically methane) in an excess of hydrogen.^1.3,1.6,1.7 The resultant diamond produced by this technique can display properties comparable to those of natural diamond.

1.2 Historical overview of the CVD diamond process

The synthesis of diamond has been pursued ever since the French chemist Antoine Laurent Lavoisier discovered that diamond was a form of crystalline carbon.^1.8 The first report describing successful diamond synthesis using HPHT methods was published in 1955,^1.9 in which diamond is crystallised from metal solvated carbon at pressures and temperatures in the region of 55,000 atmospheres and ~1800°C.

Experiments with low pressure synthesis had been ongoing in both the United States and USSR during the same period. The first patents emerged as early as 1962 by Eversole of the Union Carbide Corporation^1.10 who reported diamond synthesis under reduced pressure by the CVD technique. His process involved the growth of diamond on diamond substrates with the use of carbon-containing gases at pressures of less than 1 atmosphere and temperatures of around 800-1000°C.^1.11 Eversole had actually succeeded in late 1952, even before results of HPHT techniques were published. This technique, however, was not very effective since large amounts of graphite were co-deposited with diamond, and growth rates were extremely low. Eversole’s work was followed up by Angus and co-workers who provided evidence of the use of atomic hydrogen as an etchant for graphite.^1.12 In the mid-1970’s, the Soviet group led by Spitzyn, Bouilov and Derjaguin showed that growth on non-diamond substrates, such as copper, silicon and tungsten, was possible.^1.13 This is a significant achievement as growth had only previously been possible on diamond substrates. The Japanese built on both these developments to devise a ‘hot filament’ CVD technique^1.14 in which atomic hydrogen was incorporated in to the growth process to give good quality films on a number of substrates with significantly better deposition rates, making diamond film growth commercially viable for the first time. Widespread interest developed from here, and at present a variety of CVD deposition techniques have been established.

1.3 The diamond CVD technique

Chemical vapour deposition involves a gas-phase chemical reaction taking place above a solid surface, which results in deposition onto that surface. All the CVD techniques employed to deposit diamond films require a means of activating gas-phase carbon-containing precursors. This involves mainly thermal (e.g. hot filament) or plasma (microwave or R.F.) activation, or use of a combustion flame (oxyacetylene torch).^1.7 Figure 1.3 shows two of the most common deposition techniques and gives some indication of typical operating conditions.

Figure 1.3. Examples of two most common types of low pressure CVD techniques to deposit diamond. (a) Hot filament reactor, and (b) Microwave plasma enhanced reactor.

Typically any hydrogen/hydrocarbon source gas mixture may be used, subject to the C/H ratio being less than ~0.03. Common to each deposition technique and source gas mixture employed in the synthesis of CVD diamond is the requirement of high gas temperatures which generates atomic hydrogen and produces reactive carbon species. Furthermore the substrate temperature must be maintained in the range 1000-1400K. The resulting films are polycrystalline, with a morphology that is sensitive to the precise growth conditions (See Section 1.4). The deposition rate differs from one CVD deposition technique to another, and it is generally observed that higher deposition rates can be achieved at a cost, and that is the ‘quality’ of the films. The quality of a CVD diamond film is normally judged by a number of measures such as the ratio of sp³ (diamond) to sp²-bonded (graphite) carbon in the sample, the composition (e.g. C-C versus C-H bond content) and the crystallinity. In general, combustion methods deposit diamond at high rates (typically 100 mm/hr), but are restricted to small substrate areas, and with poor process control leading to poor quality films. On the other hand, hot filament and plasma methods produce high quality films but at much slower rates (0.1-10 mm/hr). One of the main aims in current CVD research is to increase the growth rates to economically viable rates (hundreds of mm/hr) without compromising film quality.

1.4 The CVD diamond film

The surface morphology resulting from diamond CVD depends critically on C/H ratio in the source gas mixture, as well as the substrate temperature. At low substrate temperatures and low CH₄ concentration, diamond films are produced with predominantly triangular {111} facets with many twin grain boundaries (See Figure 1.4).

Figure 1.4. Surface morphology of a diamond film grown on Si. The film is polycrystalline, with twinning and many crystal defects present.

{100} facets, appearing as both square and rectangular forms, become more evident as the concentration of CH₄ in the source gas mixture, and/or the substrate temperature, is increased. Cross-sections of these microcrystalline films shows the growth to be mainly columnar in nature (See Figure 1.5)

Figure 1.5. Cross section through a 6.7 mm-thick diamond film on Si, showing the columnar characteristics of the growth up from the surface.

At even higher CH₄ concentrations the crystalline morphology disappears altogether; instead a film composed of an aggregate of nanocrystals and disordered graphite is formed, as shown in Figure 1.6.

Figure 1.6. Nanocrystalline film, exhibiting ‘ballas’-type morphology, typical of diamond grown under high (>2%) CH₄ concentrations. This film is much smoother than the microcrystalline film.

1.5 The choice of substrates suitable for growing CVD diamonds

The most commonly used substrates used to deposit CVD diamonds are Si wafers, but other substrate materials may also be used. What these substrate materials share in common is the need to withstand the temperature window (1000-1400°C) required for diamond growth. This precludes the use of existing CVD techniques to diamond-coat materials such as plastics or low-melting metals like alumimium. It is also useful that the substrate be capable of producing a carbide. Initial stages of CVD diamond growth on non-diamond substrates would involve the formation of a carbide interfacial layer upon which the diamond then grows. It is difficult to grow on materials with which carbon is ‘too reactive’, i.e. those with which carbon exhibit a high mutual solubility (e.g. transition metals such as iron, cobalt, etc.). Substrates like Si, Mo and W are by far the most popular because these materials form carbides, but only as a localised interfacial layer due to their modest mutual solubility with carbon under typical CVD process conditions. The carbide layer can be envisaged as the ‘adhesive’ which promotes growth of the CVD diamond, and aids its adhesion by (partial) relief of stresses at the interface. The popularity of Si as the substrate material is also due to its comparatively low thermal expansion coefficient, which again serves to minimise the build up of stress in the CVD diamond films.

1.6 Mechanism of CVD growth

A diamond CVD process is schematically summarised in Figure 1.7.^1.16 Gaseous precursors, typically 1% CH₄ in H₂ at ~20 Torr, flow onto the reactor and gas-phase reactions are initiated by the hot filament or plasma. The reactants, products, and reactive species are transported throughout the reactor by diffusion and convection. On the surface, adsorption, diffusion, reaction and desorption of various species occur giving rise to the nucleation of diamond particles, suppression of graphitic carbon, and ultimately the growth of a continuous diamond film.

Figure 1.7. A Schematic diagram showing the principle elements in the complex diamond CVD process: flow of reactants into the reactor, activation of the reactants by thermal and plasma processes, reaction and transport of the species to the growing surface, and surface chemical processes depositing diamond and other forms of carbon (Adapted from Reference 1.16).

1.7 Role of atomic hydrogen in the CVD process

As previously mentioned, graphite is thermodynamically the more stable form of solid carbon at ambient pressures and temperatures. Atomic hydrogen is believed to change the relative free energies of small graphitic and diamond nuclei, thus providing a lower energy pathway for the conversion of graphite to diamond. Therefore diamond is the predominant product of the CVD process.^1.12,1.17 H atoms are produced either thermally (at the filament) or via electron bombardment; once formed they are believed to play a number of crucial roles in the CVD process:

n They undergo H abstraction reactions with stable gas-phase hydrocarbon molecules, leading to the formation of highly reactive carbon-containing radical species. Unlike the stable hydrocarbon molecules, these active radical species can diffuse to the substrate surface and react. As a result, the all important C-C bonds are formed, leading to the propagation of the diamond lattice.

n Atomic hydrogen etches both diamond and graphite but, under typical CVD conditions, the rate of diamond growth exceeds its etch rate whilst the reverse is true for all other forms of carbon (e.g. graphite). This is believed to be the basis for the preferential deposition of diamond over graphite.

n H-atoms terminate the ‘dangling carbon bonds on the growing diamond surface and prevent them form cross-linking, thereby reconstructing to a graphite-like surface.

1.8 Gas-phase chemistry involved in the CVD process

The initiation of the gaseous chemistry is dominated by dissociation of molecular hydrogen into atomic hydrogen, which is achieved either by thermal dissociation on a hot filament or by electron impact dissociation in a plasma. The subsequent gaseous chemistry is driven by reactions of atomic hydrogen with hydrocarbon species and a series of complex reactions among the hydrocarbon species. Figure 1.8 shows a schematic of this complicated ‘chemical soup’ of the major elements, reactive atoms and radicals.

Figure 1.8. A schematic diagram showing the major elements involved in the complex mix of reactions. The reactions involve rapid hydrogen transfer reactions amongst the C₁ and C₂ species, and to a lesser extent the bimolecular hydrocarbon reactions forming C₂ and higher species.^1.16

Since typical process pressures are below 1 atm H atom recombination is slow, and a super-equilibrium concentration of hydrogen atoms is present.^1.18 Furthermore, since the major species in the gas phase are H and H₂, and the total hydrocarbon concentration is about 1%, the hydrogen transfer reaction rates are generally much greater than those describing the bimolecular hydrocarbon reaction rates.^1.19-1.21

1.9 Characterisation of the CVD process

There is considerable interest in understanding the mechanism involved in the CVD of diamond films, which could lead to greater control of a potentially beneficial process. However, the precise mechanism leading ultimately to the growth of a CVD diamond film is still not very well understood. Several groups have attempted to determine the gas-phase species involved in the process, but no definitive answers have been reached. Several techniques, both optical and physical diagnostic methods have been employed, such as optical spectroscopy,^{1.20,1.22-1.24} resonance-enhanced multiphoton ionisation (REMPI),^1.25 gas chromatography,^1.26,1.27 cavity ring down spectroscopy (CRDS)^1.29,1.30 and mass spectrometry.^1.31-1.34 It is mass spectrometry which is of most interest to the present work. Using a mass spectrometer it is possible to detect a number of species simultaneously, and accurately quantify their composition (See Chapter 3). Harris and his co-workers^1.34 utilised a quartz probe to sample gaseous products from a hot filament CVD diamond chamber during deposition, which were then analysed using a mass spectrometer. However, a direct measurement of radical species such as methyl (CH₃) or ethenyl (C₂H) radicals was not possible because of recombination of these species in the probe used to sample the gases. An attempt was made to estimate the concentrations of both radicals by calculating ions from measurements of the concentrations of the likely recombination products, namely C₂H₂ and C₂H₆ respectively. The main drawback to this method is relying on the fact that they assume that there are no alternative recombination reactions that could occur with other radical species. A means of detecting radical species directly from the growth region is therefore of fundamental importance in gaining a true identification of the gas-phase species responsible for the propagation of diamond during the CVD process. The solution is the molecular beam mass spectrometry (MBMS) system developed by Hsu et al.^1.35 which has enabled the detection of stable as well as free radical species. A three stage differential pumping system was utilised to cause a large pressure difference to exist between the CVD chamber (20 Torr) and the mass spectrometer (10^-8 Torr). Gas from the reaction chamber is extracted through a small orifice (typically 100 mm in diameter) to the high vacuum region of the mass spectrometer. The pressure difference induces the formation of a supersonic molecular beam that effectively freezes out reactions, even between reactive species such as free radicals. This beam has an unobstructed path towards the ionisation chamber of the mass spectrometer where, ultimately, the observable species are analysed.

1.10 Summary of the work performed on hot filament CVD systems.

Much research relevant to the present work has been performed by Harris, who utilised mass spectroscopic technique. The overall conclusion is that CH₃ and C₂H₂ are the most likely candidates responsible for diamond deposition, because they are the two main observable product species present under optimum growth conditions. This results accord well with chemical kinetics modelling^1.22 which predicts comparatively large concentrations of both species. Subsequent work^1.34 by the same authors involved measuring the mole fractions of the stable species CH₄ and C₂H₂ in the diamond growth environment. As previously mentioned, recombination at the probe used to sample the gas precludes the detection of CH₃, but an indirect method was employed to determine the methyl concentration by measuring the C₂H₆ and C₂H₄ species concentrations. However, these species were detected in low quantities, presumably because of their transient nature at high H atom concentrations, [H]. At optimum growth conditions, C₂H₂ is the dominant C₂ species in the gas phase. Since the formation of [C₂H₂] depends on [CH₃], the detection of gas-phase acetylene is now generally taken as an indicator of high steady state [CH₃] and hence of diamond growth.

Further research^1.31 was performed to examine the importance of methyl radicals and acetylene. Aided with a chemical kinetics model^1.22 it was proposed that CH₃ and/or CH₄ is an effective diamond growth precursor. Furthermore they observed that the diamond films deposited using CH₄/H₂ source gas mixtures were of better quality and higher growth rates than when C₂H₂ was the precursor gas. In addition, it was believed that C₂H₂ was possibly responsible for the deposition of non-diamond carbon phases. This observation, coupled to the fact that the effective reaction probability for diamond growth (defined as the number of carbon atoms deposited as diamond divided by the number of collisions between a growth species and the diamond surface) due to CH₃ was greater than that due to C₂H₂, led the authors to believe that CH₃ is the major growth precursor in hot filament CVD process. In a later report^1.36 it was found that the rate of diamond deposition was directly proportional to the calculated mole fraction of CH₃ present in the reactor, but no such relationship was observed for the calculated C₂H₂ mole fraction. This supports the view that methyl radicals are important intermediates in the CVD diamond growth mechanism.

The development of the MBMS system by Hsu^1.37 has enabled the detection of radical species as well as stable hydrocarbon species in the CVD growth process. The apparatus allowed measurements of H, CH₃, CH₄ and C₂H₂ as a function of CH₄ concentration in the source gas mixture. Results show that hydrocarbon species other than those above were present in trace amounts regardless of input CH₄ concentration, and therefore are unlikely to be involved in the growth mechanism. At low [CH₄] in the feed gas the dominant product is that of acetylene although at higher [CH₄] methane becomes the main hydrocarbon species. A rise in the CH₃ concentration was also observed with [CH₄], though to a lesser extent at high methane input. The measured H atom mole fraction, by contrast, drops rapidly with increasing CH₄ feed, presumably due to filament poisoning effects (See Section 8.7, Chapter 8). As a result, reactions which require the presence of [H] in the gas phase, such as the formation of CH₃ and H abstraction reactions of the C2 hydrocarbon species (ultimately forming C₂H₂) will occur much less readily, leading to a decrease on the absolute concentration of both CH₃ and C₂H₂. This is accompanied by an increase in the graphitic content in the deposited diamond films with methane feed, which confirms the importance of the role of H atoms in etching graphite and other non-diamond carbon phases (See Section 1.7).

1.11 References

1.1 J.E. Field, Properties of Natural and Synthetic Diamond, Academic Press,

London, 1992.

1.2 G. Davies, Diamond, Adam Hilger, Bristol, 1984.

1.3 M.N. Yoder in Diamond Films and Coatings, R.F. Davis ed. (Noyes Publications

1993).

1.4 F.P. Bundy, J. Geophys. Res., 85, 6930 (1980).

1.5 J.E. Field, The Properties of Diamond, Academic, New York, 1979.

1.6 S. Matsumoto, Y. Sato, M. Tsutsumi and N. Setaka, J. Mater. Sci., 17, 3106

(1982).

1.7 F.G. Celii and J.E. Butler, Ann. Rev. Phys. Chem., 42, 643 (1991), and references

therein.

1.8 A.L. Lavoisier, Memoire Acedemie des Science, 564 (1772).

1.9 F.P. Bundy, H.T. Hall, H.M. Strong and R.H. Wentorf, Nature, 176, 51 (1955).

1.10 W.G. Eversole United States Patent No. 303187 and 303188 (1962).

1.11 R.C. DeVries, Annu. Rev. Mater. Sci., 17, 161 (1987).

1.12 J.C. Angus, H.A. Will and W.S. Stanko, J. Appl. Phys., 39, 2915 (1968).

1.13 B.V. Spitzyn, L.L. Bouilov and B.V. Derjaguin, J. Cryst. Growth, 52, 219 (1981).

1.14 S. Matsumoto, Y. Sato, M. Kamo and N. Setaka, Jpn. J. Appl. Phys., 21, L183

(1982).

1.15 M.N.R. Ashfold, P.W. May, C.A. Rego, and N.M. Everitt, Chem. Soc. Rev., 23,

21 (1994).

1.16 J.E. Butler and R.L. Woodin, Phil. Trans. R. Soc. Lond. A, 342, 209 (1993).

1.17 T.R. Anthony, Vacuum, 41, 1356 (1990).

1.18 D.G. Goodwin, Appl. Phys. Lett., 59, 277 (1991).

1.19 S.J. Harris, J. Appl. Phys., 65, 3044 (1989).

1.20 D.G. Goodwin and G.G. Gavillet, J. Appl. Phys., 68, 6393 (1990).

1.21 M. Frenklach and H. Wang, Phys. Rev. B, 43, 1520 (1991).

1.22 S.J. Harris, A.M. Weiner and T.A. Perry, Appl. Phys. Lett., 53, 1605 (1988).

1.23 L.R. Martin and M.W. Hill, Appl. Phys. Lett., 55, 2248 (1989).

1.24 S.M. Myers and A.F. Zoltowski, Appl. Phys. Lett., 43, 3425 (1989).

1.25 F.G. Celii and J.E. Butler, J. Appl. Phys., 71, 2877 (1992).

1.26 F.G. Celii and J.E. Butler, Appl. Phys. Lett., 54, 1031 (1989).

1.27 M.E. Coltrin, R.J. Kee and J.A. Miller, J. Electrochem. Soc., 133, 1206 (1986).

1.28 D.N. Belton and S.J. Harris, J. Chem. Phys.,76, 2371 (1992).

1.29 E.H. Wahl, T.G. Owano, C.H. Kruger, P. Zalicki, Y. Ma and R.N. Zare, Diamond.

Relat. Mater., 5, 373 (1996).

1.30 P. Zalicki, Y. Ma, R.N. Zare E.H. Wahl, J.R. Dadamio, T.G. Owano and C.H.

Kruger, Chem. Phys. Lett., 234, 269 (1995).

1.31 S. J. Harris and L.R. Martin, J. Mater. Res., 5, 2313 (1990).

1.32 S. J. Harris, Appl. Phys. Lett., 56, 2298 (1990).

1.33 M. Frenklach and K.E. Spear, J. Mater. Res., 3, 133 (1988).

1.34 S.J. Harris and A.M. Weiner, J. Appl. Phys., 67, 6520 (1990).

1.35 W.L. Hsu and D.M. Tung, Rev. Sci. Instrum., 63, 4138 (1992).

1.36 S.J. Harris and A.M. Weiner, Thin Solid Films, 212, 201 (1992).

1.37 W.L. Hsu, Appl. Phys. Lett., 59, 1427 (1991).

Chapter 1 - Introduction