Chapter 5 - Results for C/H/Cl Systems

5.1 Introduction

For hot filament and microwave, rf and dc plasma CVD systems,^3.4,5.1 deposition of high quality diamond films using hydrogen/carbon gas mixtures at any appreciable rates can be achieved only if the substrate temperature is maintained at around 900C. As a result, the choice of substrate is limited to very few materials. Owing to its remarkable chemical, physical and optical properties,^5.2 diamond has great potential for many industrial applications so, by altering the conditions which produce high quality films at lower substrate temperatures, its applications may extend to the less thermally stable materials such as aluminium, stainless steel, glass, copper and plastics.

This has prompted the study of CVD diamond growth using other precursor gas mixtures. For instance, it is possible to grow CVD diamond films using either oxygen containing source gases, e.g. CH₃OH^5.3 and CO,^5.4 in excess H₂ or by the addition of small amounts of O₂^5.5 to the standard hydrocarbon/H₂ source gas mixture. The presence of gas phase oxygen can enhance diamond growth rates and also enable diamond synthesis at lower substrate temperatures. A survey of the different C/H/O source gas mixtures used to produce diamond films is summarised in the well known phase diagram of Bachmann et al.^5.6 The apparent catalytic effect of oxygen in the CVD process enabling diamond growth at substrate temperatures lower than the standard (~900°C) may therefore have possible important implications for coating such lower melting point materials as aluminium, or materials that are unstable at high temperatures in the reducing atmosphere typical of the CVD process. Such advantages cannot be fully exploited in hot filament CVD (HFCVD) reactors, though, because the presence of oxygen rapidly degrades the heated filament. In recent reports however, lower-temperature diamond deposition has also been achieved using either chlorine containing source gases, e.g. CH_4-nCl_n (n = 1-4) diluted in hydrogen^5.7-5.11 or by the addition of small amounts of Cl₂^5.12 or HCl^5.13 to the standard methane/hydrogen gas mixture.

One study,^5.7 using monosubstituted halocarbons CH₃F, CH₃Cl, CH₃Br or CH₃I in H₂ as the input gas mixture in a hot filament reactor, showed that with CH₃Cl the diamond growth rates increased relative to those found using CH₄ as the precursor. It was concluded that the increased reactivity of CH₃Cl compared to the other methyl halides stems from the difficulty, in the one case, of forming atomic fluorine from CH₃F and, at the other extreme, the inability of Br and I atoms to abstract terminating hydrogen atoms from the growing diamond surface. Other studies using chloride source gases (CH₃Cl, CH₂Cl₂, CHCl₃ and CCl₄) both in hot filament reactors^5.8,5.9 and in microwave plasmas^5.10,5.11 have also shown that chloromethanes produce a slightly higher diamond growth rate than with methane at normal substrate temperatures (ca. 900C), and all indicate that this difference is more pronounced at lower substrate temperatures. Recently,^5.12 a study using chlorine with typical H₂/Cl₂/CH₄ ratios of 100/5/1 demonstrated that similar growth rates can be achieved at substrate temperatures 150C lower than that found for typical H₂/hydrocarbon gas mixtures, and that the addition of a few percent of HCl to the standard CH₄/H₂ mixture could cause a tenfold enhancement in growth rates at 670°C.^5.13

A number of possible gas phase and gas-solid heterogeneous reaction mechanisms have been suggested for the apparent catalytic activity of chlorine in the CVD process. The C-Cl bond strengths in the various chloromethanes are weaker than the C-H bond in methane (see Table 5.1) so rupture of the C-Cl bond, by reaction with H atoms, can be achieved more readily to produce active methyl or chloromethyl (CH_3-nCl_n, n=1-3) radicals. The latter species have been suggested as being more effective growth precursors than methyl^5.7 because of the possibility of facile surface dehydrochlorination reactions.^5.8 Alternatively, the catalytic properties of chlorine in the CVD process may be due to reactions with terminal C-H at the diamond surface, or the possibility of forming surface bonded C-Cl. The terminating H and Cl atoms essentially serve the same purpose in suppressing formation of sp² bonding at the growing surface at high substrate temperatures. Abstraction of surface Cl atoms by atomic H, or removal of surface adsorbed H by atomic chlorine, has a significantly lower activation energy than H abstraction of surface H thereby enabling active surface sites to be created at lower substrate temperatures.^5.13

Bond Bond strength (kJ/mol)

H-H 432

Cl-Cl 239

H-Cl 428

H-F 564

H-CH₃ 435

Cl-CCl₃ 305

Cl-CHCl₂ 331

Cl-CH2Cl 346^*

Cl-CH₃ 352

F-CF₃ 540

Table 5.1. Bond dissociation energies ( taken from Reference 5.14 except for ^* which is taken from Reference 5.15).

The aim of the present work is to characterise the gas phase environment, and its variation with filament temperature, when chlorine containing precursors CH_4-nCl_n (n=1-4) in H₂ and CH₄/Cl₂/H₂ mixtures are inputted as the source gases. Such information provides insight into the gas-phase chemical kinetics prevalent in the CVD process. We also wish to understand this behaviour in terms of changes in the gas-phase chemistry and/or the different gas-solid heterogeneous reactions when chlorine is present in the gas mixture.

5.2 Experimental Details

(a) Deposition experiments

Table 5.2 below shows the growth conditions used for the C/H/Cl system. Full experimental details on deposition techniques are given in Chapter 3 (Sections 3.1 and 3.2).

Pressure 20 Torr

Gases (1) CH_4-nCl_n (n = 1-4), (2) 1% CH₄ in Cl₂/H₂, the amount of chlorine varying from 1-4%.

Total gas flow rate 100 sccm

Substrate type Si (100) substrates (manually abraded)

Substrate temperature ~ 900°C

Filament temperature 2300-2400°C (filament current 6-6 A)

Filament/substrate distance 4 mm

Deposition time 6 hours

Table 5.2. Deposition conditions used for C/H/Cl system.

(b) Film analysis

The as-grown diamond films were investigated by scanning electron microscopy (SEM) and Auger electron spectroscopy (AES). The AES analysis was carried out using a Perkin-Elmer PHI 595 scanning Auger microprobe.

(c) Gas phase composition measurements

Gas-phase product distributions were monitored as a function of the filament temperature using the differentially pumped molecular beam mass spectrometer (MBMS). Filament temperatures were measured using a two-color optical pyrometer (Land Infrared) and the filament-to-sampling orifice distance was held at 4 mm for all readings. The absolute concentrations of the species monitored are determined by direct room temperature calibration with mixtures of known composition. Full description of the molecular beam mass spectrometer and the calibration procedures are given in Chapter 3 (Sections 3.3 to 3.5).

A series of quantitative measurements of the stable gas-phase species and CH₃ radicals was made using, respectively, chloromethane (CH₃Cl), dichloromethane (CH₂Cl₂), trichloromethane (CHCl₃) or tetrachloromethane (CCl₄) in H₂ (See Figures 5.4 (a) to 5.4 (d) ). The signal intensities of the species monitored have been converted into mole fractions using the step-by-step procedure described in Chapter 3 (See Section 3.6). The following results are therefore presented as species concentrations as a function of filament temperature with no correction being made as a result of thermal diffusion (experimental data sheets for each precursor gas can be found in the appendix that follows this chapter).

Detection of all stable carbon and chlorine containing species, other than C₂H₄, was made with an ionizing electron energy of 15.6 eV. An electron energy of 13.6 eV was used for C₂H₄ in order to minimize interference from unwanted ions with the same mass-to-charge ratio or those arising from fragmentation of other ionic species (See Section 3.5 (b) ). For the same reasons, all CH₃⁺ signal measurements were made using an ioniser voltage of 13.6 eV. However corrections have been made to the CH₃⁺ signal arising from the dissociative ionization of CH₃Cl (See Section 3.6 (b) ).

As previously mentioned (See Section 3.5 (c) ), introduction of a heavy species, such as Cl₂, in the gas mixture results in mass discrimination effects in the molecular beam,^3.11 particularly when the concentration is greater than ~1%. Consequently, the lighter species, such as CH₄, are preferentially excluded from the centre of the beam by the heavier Cl₂ molecules (See Figure 3.20). When performing gas-phase composition measurements using CH₄/Cl₂/H₂ containing gas mixtures, a minor modification had to be made to the direct room temperature calibration procedure in order to determine accurate concentrations for each species monitored. This was achieved by measuring the signal loss for all the species monitored during the calibration process when Cl₂ is added to the gas mixture. For example, if 1% Cl₂ was introduced to the standard 1% CH₄/H₂ during sampling, then the same concentration of Cl₂ was added to that gas mixture when calibrating for methane. The attenuated CH₄⁺ signal therefore constitutes 1% in the gas-phase when 1% Cl₂ is present in the process gas mixture, allowing for the mass discrimination effects. Similar corrections were made when 2% Cl₂ and 4% Cl₂ were added to the CH₄/H₂ gas mixture during MBMS measurements.

5.3 Analysis of the diamond films

Scanning electron micrographs of diamond films grown on silicon when using 1% each of CH_4-nCl_n (n=1-4) in H₂ as the process gas mixture are shown in Figure 5.1. The SEM images in Figure 5.2 show that polycrystalline diamond films were also grown successfully using up to 3 % Cl₂ in a 1% CH₄/H₂ gas mixture. Further addition of chlorine (~4%) leads to a significant change in the film morphology (See Figure 5.2(d) ) with a concomitant reduction in the film quality and growth rates.

Figure 5.3 (a) shows a typical AES spectrum of the diamond film surface grown using a 2% Cl₂ was introduced in a 1% CH₄ in H₂ gas mixture. The vertical scale on the AES spectrum is a differential scale used in order to display the peaks more clearly. The fine structure of the KLL carbon peak (Figure 5.3 (b) ) is characteristic of that found with natural diamond.^5.16, 5.17 The peak is shifted slightly to lower energies (by ca. 4 eV) compared with the peak obtained from the very poor quality diamond when 4% Cl₂ was used for growth. This energy shift arises mainly from charging effects on the diamond relative to the more conducting amorphous or graphitic carbon.

Also evident on the spectrum is a small peak at an electron energy of 180 eV indicative of the presence of chlorine on the diamond surface. After argon ion sputtering for 30 seconds, resulting in etching of the diamond film to a depth of ~120 Å, the chlorine was no longer detectable indicating that this chlorine is present only on the film surface and not in the bulk diamond. In AES, the concentration of any particular element can be calculated from its peak intensity multiplied by a sensitivity factor. The sensitivity factors are determined empirically and are unique to each element. The relative intensity of the surface Cl peak suggests a chlorine concentration of ~0.5%; however, this is an average value obtained over the top four diamond monolayers. Given that the chlorine is present only on the diamond surface then about 2% of the surface carbon atoms are terminated by chlorine rather than hydrogen.

Similar AES results indicating high quality diamond with only a trace of chlorine on the film surface were obtained for all the as-grown films where the Cl atom input fraction was 0.06 or below (i.e. £3% Cl₂ in the process gas mixture). However, for growth conditions where the Cl fraction was greater than 0.06 the fine structure of the KLL carbon peak showed that the resulting films were of poor quality with significant deposition of non-diamond carbon phases. Also incorporated into the film were appreciable amounts of tantalum and chlorine, indicating degradation of the filament at higher chlorine concentrations.

a)	b)
c)	d)

Figure 5.1. Scanning electron micrographs (SEM’s) of diamond films grown on silicon using input gas mixtures of 1% each of: (a) CH₃Cl; (b) CH₂Cl₂; (c) CHCl₃; and (d) CCl₄ in hydrogen.

a)	b)
c)	d)

Figure 5.2. Scanning electron micrographs (SEM’s) of diamond films grown on silicon using input gas mixtures of 1% CH₄ in H₂ with additional Cl₂: (a) 1%; (b) 2%; (c) 3%; and (d) 4%.

5.4 Gas composition versus filament temperature for a variety of chlorine containing precursor gases in H₂

(a) Chloromethane (1% CH₃Cl in H₂)

The effects of the variation of filament temperature between room temperature (25°C) and 2690°C on the mole fractions of the major dissociation products of chloromethane gases [CH₄ (m/e =16), C₂H₂ (m/e =26), C₂H₄ (m/e =28) and HCl (m/e =36 for H³⁵Cl)] were examined for an initial feedstock of 1% CH₃Cl in H₂ measured 4 mm from the filament at 20 Torr (Figure 5.4 (a) ). Inspection of Figure 5.4 (a) shows that there is a rapid decrease in the concentration of CH₃Cl for filament temperatures above ~1700°C; at the same time there is significant rise in the CH₄ and HCl mole fractions such that these become the dominant species at ~2100°C. However, at filament temperatures above ~2100°C the CH₄ concentration decreases steadily, and a subsequent increase in the C₂H₂ and CH₃ concentrations is observed.

(b) Dichloromethane (1% CH₂Cl₂ in H₂)

Figure 5.4 (b) shows how the distribution of the same major product gases vary as a function of filament temperature for an initial CH₂Cl₂/H₂ feedstock ratio of 1% measured at 4 mm from the filament. The gas phase composition as a function of filament temperature for dichloromethane are remarkably similar to that for chloromethane. The concentration of CH₂Cl₂ decreases rapidly for filament temperatures above ~1700°C, and a corresponding rise in the HCl concentration is observed reaching a maximum value at ~2100°C. Similarly, at this filament temperature the dominant hydrocarbon species is methane, while at higher temperatures acetylene (and CH₃ radicals) are formed. Note that the absolute concentrations of the carbon- or chlorine-containing species in the vicinity of the filament diminish at higher temperatures due to thermal diffusion effects (See Section 3.5 (e) ).

(c) Trichloromethane (1% CHCl₃ in H₂)

Similar trends were observed for the general product distribution as a function of filament temperature for 1% CHCl₃ in H₂ (Figure 5.4 (c) ). At ~2100°C filament temperature the concentration of CHCl₃ is reduced to produce primarily HCl and CH₄ species. Above this filament temperature, a steady increase in the C₂H₂ and CH₃ concentrations is observed. A fall in the absolute concentrations of the carbon- or chlorine-containing species around the filament at higher temperature is again a result of thermal diffusion effects.

(d) Tetrachloromethane (1% CCl₄ in H₂)

Inspection of Figure 5.4 (d) reveals that the variation of dissociation products with filament temperature for 1% CCl₄ in H₂ is also very similar to that of the other chlorine containing precursor gases. The C³⁵Cl₄⁺ signal (m/e=152) could not be monitored as its mass-to-charge ratio is beyond the measurable range of the mass spectrometer.

Mechanism

Qualitatively, the CH₄/C₂H₄/C₂H₂ product distribution as a function of temperature for all the chlorine containing precursor gases strongly resembles that obtained when methane is used as the precursor.^3.1 When the filament temperature is ~2100C, methane is the dominant carbon containing species in the sampled gas, whilst at higher temperatures acetylene production becomes increasingly important.

Comparison of the observed trends in the relative gas concentrations of the hydrocarbon species using chlorine containing precursors with those found for CH₄/H₂^3.1 suggests that a similar reaction mechanism is taking place, namely the chemical conversion of methane to acetylene as the filament temperature is increased, a reaction which is initiated by the dissociation of methane yielding methyl radicals:

CH₄ + H CH₃ + H₂ (4.1)

Methyl recombination followed by successive H abstraction reactions yields acetylene. As the filament temperature increases, the increasing H atom concentration drives the equilibrium from CH₄, through C₂H₆ and C₂H₄ to C₂H₂. Due to its transient nature (and thus very low steady-state concentration) under conditions rich in H atoms, no C₂H₆ was detected in the gas phase. Quantitative measurements of the absolute concentrations of methyl radicals were made simultaneously and are displayed in Figures 5.4 (a) to 5.4 (d). An increase in the filament temperature results in higher [CH₃] which is mirrored by increased [C₂H₂]. As with methane and other hydrocarbon precursor gases^3.1, it is clear that the CH₃ radical is also an essential intermediate in the formation of acetylene when chlorine containing input precursors are used.

For example, inspection of Figure 5.4 (b) shows that the concentration of dichloromethane diminishes rapidly for filament temperatures above ~1700C as the relatively weak C-Cl bonds are broken in preference to C-H bonds by Cl abstraction with H atoms, e.g.

CH₂Cl₂ + H == CH₂Cl + HCl

CH₂Cl + H == CH₂ + HCl, etc. (5.1)

This may reflect a role for direct pyrolysis of, for example,

CCl₄ ® Cl + CCl₃ (5.2)

followed by the equilibrium

Cl + H₂ == H + HCl (5.3)

The key point here would be that pyrolytic H-H bond fission (in the gas phase or on the filament surface) would not be necessary to initiate the whole reaction sequence.

A corresponding rise in the HCl concentration is observed reaching a maximum value at the same temperature as methane (~2100C). At this temperature almost all of the dichloromethane has been reduced. Similar trends were observed for all of the other CH_4-nCl_n, (n=1-4) species, resulting in the reduction of the precursors to produce primarily methane and HCl, (see Figs. 5.4 (e) and 5.4 (f) ). Figure 5.4 (g) shows how the acetylene concentrations vary as a function of filament temperature for input gas mixtures containing 1% CH_4-nCl_n, (n=0-4) in H₂. For all the precursors, the C₂H₂ concentration rises sharply above ~2100C as the H atom concentration increases. However, there is a significant increase in the C₂H₂ concentration for CCl₄ and, to a lesser extent, for CHCl₃ at lower temperatures (~1900C-2100C). As an alternative, we might envisage the case where, for CHCl₃ for example, a single C-Cl bond is cleaved to form the CHCl₂ radical which could then combine with a CH₃ radical or react with CH₄ to produce CHCl₂CH₃ (+ H in the latter case). Subsequent elimination of two HCl molecules, reactions which are known to have reasonably low activation barriers^5.18, would yield acetylene. We should point out, however, that radical-radical encounters are likely to be relatively infrequent in the gas phase environment, nor have we been able to detect this intermediate, possibly due to its instability in the presence of atomic hydrogen. However, close inspection of Figure 5.4 (h), which shows how the methyl radical concentrations vary as a function of filament temperature for different chlorocarbon input gas mixtures, indicates that production of appreciable amounts of acetylene at lower filament temperatures for the CCl₄/H₂ gas mixture may be due to the presence of high concentrations of the methyl radicals in the gas phase. In this case there is a greater steady-state [CH₃] and the increased likelihood of methyl recombination, followed by consecutive H abstractions yields larger [C₂H₂]. Since the formation of [C₂H₂] depends on k[CH₃]², the detection of large amounts of gas-phase acetylene in a hot filament CVD reactor is generally taken as an indicator of steady state [CH₃] and thus of fast diamond film growth.

5.5 Gas composition versus filament temperature for various CH₄/Cl₂/H₂ containing gas mixtures

Similar measurements of the stable gas species were made using different CH₄/Cl₂/H₂ gas mixtures subject to the same process conditions as the chlorine containing hydrocarbon experiments.

Figure 5.5 shows how the distribution of the major stable product gases [CH_4, C₂H₂, HCl and Cl₂ (m/e =70 for ³⁵Cl₂)] change as a function of filament temperature for (a) 1%, (b) 2% and (c) 4% Cl₂ in the CH₄/Cl₂ gas mixtures in H₂ measured 4 mm from the filament. Almost all the molecular chlorine is dissociated when the filament temperature is ~1000C. The Cl atoms then undergo rapid reaction with hydrogen to form HCl.

Cl + H₂ == H + HCl (5.3)

The absolute measured concentrations of both HCl and methane decrease with temperature due to the thermal diffusion effects mentioned previously. At temperatures of ~2100C and above there is a striking similarity in the measured C/Cl product distribution with that obtained with 1% CH₂Cl₂ in H₂ as the input gas mixture (c.f. Figure. 5.4 (b) ). In these two cases the relative C/Cl input mole fractions in H₂ are identical.

Figure 5.5 shows the temperature variation of: (d) [CH₄], (e) [C₂H₂], (f) [HCl] and (g) [CH₃] for different input concentrations of Cl₂ (0%, 1%, 2%, 4%) in the CH₄/Cl₂ gas mixtures in H₂. As with the chlorine containing hydrocarbons the observed trends for the hydrocarbon species at and above filament temperatures ~2100C indicate that the reaction mechanism is congruent with that found for methane as the precursor, namely the chemical conversion of methane to acetylene via methyl recombination followed by successive H abstractions.

5.6 Gas composition versus filament temperature using 1% CF₄ in H₂

No significant concentrations of CH₄, C₂H₄, C₂H₂ or HF were detected at any filament temperature when using a 1:100 CF₄/H₂ mixture as the precursor. The difficulty in producing atomic fluorine or methyl radicals, or in an H atom abstracting an F atom from CF₄ suggests that CF₄ is not a suitable precursor for diamond growth by HFCVD and, indeed, growth studies using CH₃F^5.7 in a hot filament reactor show much reduced growth rates. Further evidence that conditions are not suitable for diamond growth comes from the analysis of the tantalum filament after deposition. For all hydrocarbon precursors investigated thus far^3.5 and for the CH_4-nCl_n, (n=1-4) species studied here, the filament carburises and forms a stable gold coloured tantalum carbide coating after a few minutes of deposition, indicating the presence of carbon radical species. In the case of CF₄ a grey film (presumably tantalum fluoride) forms on the filament and the filament temperature tends to be rather unstable. We note, however, that C/H/F gas mixtures have been used to grow diamond successfully by microwave plasma CVD, and MBMS studies^5.19 of such systems indicate that HF is the major sink for fluorine.

5.7 Discussion

As with the hydrocarbon precursor gases,^5.4 chlorine containing hydrocarbons are reduced to methane at very similar temperatures, the other major product being hydrogen chloride which is formed in near stoichiometric amounts. At higher filament temperatures the temperature variation of the relative CH₄/C₂H₄/C₂H₂ product distribution generally mirrors that found for the hydrocarbon precursors at similar C/H₂ input ratios. Only with CHCl₃ or CCl₄ do we observe significant acetylene formation at considerably lower filament temperatures which may indicate the presence of transient chlorine-containing radical species supplementing the CH₃ radicals as precursors to acetylene production. For the CCl₄/H₂ gas mixture however, formation of appreciable amounts of acetylene at lower filament temperatures may be a consequence of high concentrations of the methyl radicals present in the gas phase.

Comparison of Figures 5.6 (a) and 5.6 (b) indicates that at these filament temperatures (2300°C) both the chlorocarbon and Cl₂ input precursors are completely dissociated producing HCl in near stoichiometric proportions. Thermal diffusion effects again account for the reduction in the concentration of HCl measured 4 mm from the filament. At such filament temperatures the H atom concentration is of the order of 1-2%^5.20 and abstraction of Cl from the relatively weak Cl-Cl and C-Cl bonds (see Table 5.1) by H atoms to form HCl is predicted on thermodynamic grounds. Indeed, reduction of the chloromethanes to produce HCl is observed to be largely complete at filament temperatures of ~1800°C and that of Cl₂ at temperatures less than 900°C.^5.21

Figure 5.6 (a). Gas composition versus Cl atom fraction input in the feed gas using 1% each of CH_4-nCl_n (n=0-4) in H₂ at 2300C, measured 4 mm from the filament.

Figure 5.6 (b). Gas composition versus Cl atom fraction input in the feed gas using 1% CH₄ in H₂ with additional Cl₂ varying from 0-4% at 2300C, measured 4 mm from the filament.

The breakdown of the chlorocarbon precursors to form HCl produces a concomitant increase in the concentration of methane such that at 2300°C it is the dominant carbon-containing species. The H atom concentration at these temperatures may also initiate H abstraction from methane yielding significant amounts of methyl radicals. Methyl recombination to form C₂H₆ followed by successive H abstractions produces significant amounts of (C₂H₂), the other main carbon containing species observed (see Figure 5.6). No C₂H₆ and only small concentrations of C₂H₄ were detected, presumably because their thermodynamic instability in the presence of atomic hydrogen leads to low steady-state concentrations. Previous gas-phase composition studies on CH₄/H₂ mixtures^3.1 indicate that C₂H₂ is the most stable two-carbon species as the H atom concentration increases, and at standard growth temperatures a CH₄ to C₂H₂ ratio of ~8:1 is observed.

Previous reports^5.7,5.8,5.13 have speculated upon the relative importance of: (a) modification of the gas-phase chemistry enhancing the production of diamond precursor species, or (b) fast reaction of atomic hydrogen with surface bonded C-Cl and atomic Cl with surface bonded C-H which open up active surface growth sites at reduced substrate temperatures, as possible explanations for the observations that addition of small amounts of chlorine in the CVD process encourages diamond growth at lower substrate temperatures.

The inability to detect the CH₂Cl radical in the present MBMS study suggests that its concentration must be at least an order of magnitude smaller than that of CH₃ at filament temperatures which produce optimum diamond growth (2200-2400C). This is presumably because the residual CH₄ concentration at these elevated temperatures is several times higher than that of CH₃Cl even when CH₃Cl is the input precursor; for dichloromethane and trichloromethane no significant concentration of CH₃Cl is measurable. This is to be expected since thermodynamic and kinetic arguments suggest that the C-Cl bond is likely to be broken in preference to the C-H bond. Therefore, we do not believe that gas-phase chemistry involving chlorohydrocarbon radicals can be primarily responsible for enhanced growth rates at lower substrate temperatures. However, it is possible that the addition of the CH₂Cl radical to the diamond surface and the subsequent elimination of HCl may occur more readily than addition of the CH₃ radical followed by elimination or successive abstraction of hydrogen at lower substrate temperatures.

At filament temperatures above 2000C most of the chlorine is rapidly abstracted by H atoms from the parent chlorohydrocarbon to produce HCl in amounts proportional to the Cl mole fraction in the source gas. Similarly Cl₂ is almost entirely reduced to HCl at filament temperatures above 1000°C. However, in this case it is atomic chlorine produced by the thermal dissociation of Cl₂ which reacts rapidly with molecular hydrogen to form HCl. It has been proposed that the ability of chlorine to play a part in CVD diamond deposition stems from the interchangeability of atomic Cl with atomic H in an HCl/H₂ atmosphere. Calculations by Bai et al.^5.7 indicate that for the reaction,

k_f

H + HCl === Cl + H₂ (5.3)

k_b

the equilibrium constant (K_eq) is close to unity over a large temperature range and the forward and backward rate constants (k_f and k_b) are large (See Table 5.3), thereby allowing rapid equilibration at typical substrate temperatures. Therefore, the relative concentration ratio of Cl/H atoms is proportional to the Cl fraction in the source gas mixture regardless of the form of chlorine in the input mixture.

	log kf	log Keq
Reaction	T= 500 K 800 K 1200 K	T= 500 K 800 K 1200 K
H + HF H2 + F	-0.7 5.2 8.4	-13.9 -8.6 -5.7
H + HCl H2 + Cl H + CH4 ® CH3 + H2 H + CH3Cl ® CH3 + HCl H + CH3Cl ® CH2Cl + HCl Cl + CH4 ® CH3 + HCl Cl + CH3Cl ® CH2Cl + HCl	11.7 12.2 12.5 8.9 11.0 12.1 9.6 11.0 12.2 11.7 12.1 12.8 12.2 12.6 12.8	0.2 0.03 -0.1 1.0 1.2 1.4 9.9 6.8 5.0 3.2 2.6 2.2 0.8 1.2 1.4 3.0 2.6 0.0

Table 5.3. Equilibrium and forward rate constants vs. temperature (taken from Reference 5.7).

The MBMS results combined with the AES analysis^5.22 of the as-grown diamond films are consistent with the premise that Cl atoms play some part in gas-surface reactions involved in the production of active surface growth sites at reduced substrate temperatures. Chemical kinetics calculations^5.13 have shown that the rate of H abstraction by Cl atoms from a (110) diamond surface at 670°C is some 60 times faster than abstraction by gas-phase H atoms. As indicated above, the Cl atom concentration at standard filament temperatures is directly related to the Cl fraction in the input gas, regardless of the form of the chlorine precursor. Our chlorine-assisted diamond growth studies shows that if the Cl input fraction is too high (>0.06) the quality of the diamond grown at normal substrate temperatures (~900°C) is reduced, probably because a large fraction of surface sites are activated, leading to the graphitisation of the diamond surface. These findings broadly agree with the diamond growth domain indicated in the C/H/Cl ternary gas-phase composition diagram of Bachmann et al.^5.23 However, at lower substrate temperatures the presence of such chlorine concentrations in the process gas mixture would result in an increased deposition rate due to the greater efficiency of surface hydrogen abstraction by Cl atoms.

The inability of fluorine containing precursors to participate in the hot filament CVD of diamond results from the strength of the C-F bond. Our gas-phase composition findings when using 1% CF₄ in H₂ indicate that very little CH₄ or HF are produced; further, even if we could produce atomic fluorine, the equilibrium

H + HF == H₂ + F (7)

lies far to the left in a hydrogen atmosphere (See Table 5.3).

5.8 References

5.1 M.N.R. Ashfold, P.W. May, C.A. Rego and N.M. Everitt, Chem. Soc. Rev., 23, 21 (1994).

5.2 J.C. Angus, Annu. Rev. Mater. Sci., 21, 221 (1991).

5.3 Y. Hirose and Y. Terasawa, Jpn. J. Appl. Phys., 25, L519 (1986).

5.4 Y. Saito, K. Sato, K. Gomi and H. Miyadera, J. Mater. Sci., 25, 1246 (1990).

5.5 T. Kawato and K. Kondo, Jpn. J. Appl. Phys., 26, 1429 (1986).

5.6 P.K. Bachmann, D. Leers and H. Lydtin, Diamond and Relat. Mater., 1, 12 (1991).

5.7 B.J. Bai, C.J. Chu, D.E. Patterson, R.H. Hague and J.L. Margrave, J. Mater. Res., 8, 233 (1993).

5.8 F. C.-N. Hong, G.-T. Liang, J.-J. Wu, D. Chang and J.-C. Hsieh, Appl. Phys. Lett., 63, 3149 (1993).

5.9 F. C.-N. Hong, J.-J. Wu, C.-T. Su and S.-H. Yeh, in Advances in New Diamond Science and Technology, MYU, Tokyo, p.85 (1994).

5.10 C.H. Chu and M.H. Hon, Diamond Relat. Mater., 2, 311 (1993).

5.11 T. Nagano and N. Shibata, Jpn. J. Appl. Phys., 32, 5067 (1993).

5.12 C. Pan, C.J. Chu, J.L. Margrave and R.H. Hague, J. Electrochem. Soc., 141, 3246 (1994).

5.13 N.J. Komplin, B.J. Bai, C.J. Chu, J.L. Margrave and R.H. Hague, in Diamond Materials (The Electrochemical Society proceedings Series, Pennington, NJ, 1993) Vol. 93-17, p.385.

5.14 Tables of Physical and Chemical Constants, edited by G.W.C. Kaye and T.H. Laby, (Wiley, New York 1989).

5.15 Ref. Data on Atoms, Molecules and Ions, edited by A.A. Radzig and B.M. Smirnov, Vol. 31 (Springer, Berlin 1985).

5.16 P.G. Lurie and J.M. Wilson, Surface Sci., 65, 476 (1977).

5.17 Q.S. Chia, C.M. Younes, P.G. Partridge, G.C. Allen, P.W. May and C.A. Rego, J. Mater. Sci., 29, 6397 (1994).

5.18 W. Ho, R.B. Barat and J.W. Bozzelli, Combustion and Flame, 88, 265 (1992).

5.19 C. Fox and W.L. Hsu, private communication.

5.20 D.S. Dandy and M.E. Coltrin, J. Appl. Phys., 32, 5067 (1993).

5.21 C.A. Rego, R.S. Tsang, P.W. May, M.N.R. Ashfold and K.N. Rosser, J. Appl. Phys., 79, 7264 (1996).

5.22 R.S. Tsang, C.A. Rego, P.W. May, J. Thumim, M.N.R. Ashfold, K.N. Rosser, C.M. Younes and M.J. Holt, Diamond Relat. Mater., 5, 359 (1996).

5.23 P.K. Bachmann, H.J. Hagemann, H. Lade, D. Leers, F. Picht, D.U. Wiechert and H. Wilson, Mater. Res. Soc. Symp. Proc., 339, 267 (1994).

5.9 Appendix:

(I) Ionization potentials (I.P.) of the various precursor gases used (taken from Reference 3.19)

Precursor Gas	Chemical Formula	Ionization Potential (eV)
Methane	CH4	12.98
Chloromethane	CH3Cl	11.28
Dichloromethane	CH2Cl2	11.35
Trichloromethane	CHCl3	11.42
Tetrachloromethane	CCl4	11.47
Hydrogen Chloride	HCl	12.74
Chlorine	Cl2	11.48
Tetrafluoromethane	CF4	<15.00

(II) Experimental Data

The following experimental values have been converted from raw MBMS data into species mole fractions as a function of filament temperature for chlorine containing precursors CH_4-nCl_n (n=1-4) in H₂ and CH₄/Cl₂/H₂ gas mixtures. The species concentration are presented with no correction being made as a result of thermal diffusion. Typical uncertainties in these values can be estimated by comparison with Figures 4.4 (a) to (d).

1% CH₃Cl in H₂ at 20 Torr vs. Filament Temperature

MS Probe Parameters: -6% (DISCRIM), 0% (DELTAM), -40% (RES’N), 2650V

(SEM), 3.0V (CAGE), 120mA (EMISS).

MS Pressure = 8x10^-7 Torr

Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	HCl (15.6eV)	CH3Cl (15.6eV)	CH3 (13.6eV)
25	0.008	0.002	0.003	0.000	1.000	0.000
900	0.007	0.002	0.003	0.010	0.740	0.006
1070	0.007	0.002	0.003	0.020	0.590	0.007
1400	0.010	0.002	0.002	0.022	0.530	0.008
1700	0.056	0.002	0.006	0.043	0.390	0.008
1820	0.122	0.003	0.005	0.083	0.310	0.009
1960	0.254	0.003	0.005	0.145	0.120	0.008
2060	0.299	0.003	0.003	0.204	0.060	0.008
2160	0.332	0.005	0.004	0.212	0.040	0.012
2270	0.295	0.007	0.006	0.211	0.020	0.018
2370	0.257	0.014	0.009	0.217	0.010	0.022
2480	0.185	0.025	0.012	0.230	0.004	0.025
2580	0.133	0.033	0.008	0.243	0.000	0.035
2690	0.095	0.031	0.007	0.243	0.000	0.038

1% CH₂Cl₂ in H₂ at 20 Torr vs. Filament Temperature

MS Probe Parameters: -6% (DISCRIM), -20% (DELTAM), -40% (RES’N), 2600V (SEM), 3.0V (CAGE), 140mA (EMISS).

MS Pressure = 1.4x10^-6 Torr

Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	HCl (15.6eV)	CH2Cl2 (15.6eV)	CH3 (13.6eV)
25	0.004	0.002	0.000	0.000	1.000	0.000
900	0.004	0.002	0.002	0.015	0.599	0.003
1250	0.004	0.002	0.002	0.024	0.495	0.004
1480	0.005	0.002	0.003	0.028	0.432	0.006
1750	0.012	0.002	0.001	0.035	0.330	0.005
1860	0.032	0.003	0.002	0.070	0.273	0.005
1980	0.341	0.005	0.002	0.457	0.028	0.004
2070	0.365	0.008	0.005	0.549	0.016	0.006
2140	0.361	0.009	0.006	0.564	0.009	0.009
2240	0.326	0.014	0.006	0.524	0.002	0.011
2380	0.233	0.029	0.010	0.511	0.002	0.017
2505	0.151	0.046	0.010	0.471	0.000	0.026
2615	0.098	0.049	0.007	0.422	0.000	0.027

1% CHCl₃ in H₂ at 20 Torr vs. Filament Temperature

MS Probe Parameters: -6% (DISCRIM), -20% (DELTAM), -40% (RES’N), 2600V (SEM), 3.0V (CAGE) 140mA (EMISS).

MS Pressure = 1.4x10^-6 Torr

Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	HCl (15.6eV)	CHCl3 (15.6eV)	CH3 (13.6eV)
25	0.003	0.003	0.001	0.000	1.000	0.000
900	0.003	0.002	0.002	0.040	0.524	0.005
1470	0.004	0.002	0.002	0.069	0.325	0.008
1730	0.011	0.003	0.002	0.111	0.224	0.008
1850	0.024	0.005	0.003	0.183	0.199	0.014
1920	0.042	0.008	0.003	0.226	0.128	0.011
2010	0.268	0.009	0.006	0.643	0.024	0.008
2070	0.305	0.011	0.006	0.742	0.016	0.007
2150	0.329	0.007	0.003	0.777	0.000	0.009
2270	0.313	0.009	0.007	0.737	0.000	0.011
2360	0.266	0.014	0.005	0.731	0.000	0.018
2440	0.207	0.028	0.009	0.634	0.000	0.023
2535	0.140	0.040	0.008	0.610	0.000	0.035
2610	0.080	0.043	0.008	0.571	0.000	0.046

1% CCl₄ in H₂ at 20 Torr vs. Filament Temperature

MS Probe Parameters: -6% (DISCRIM), -20% (DELTAM), -40% (RES’N), 2600V (SEM), 3.0V (CAGE) 140mA (EMISS).

MS Pressure = 7x10^-7 Torr

Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	HCl (15.6eV)	CH3 (13.6eV)
25	0.004	0.001	0.001	0.063	0.007
1360	0.007	0.001	0.002	0.121	0.012
1698	0.009	0.003	0.002	0.167	0.014
1910	0.025	0.006	0.003	0.298	0.028
2010	0.332	0.015	0.006	1.162	0.043
2130	0.369	0.019	0.006	1.184	0.060
2200	0.341	0.020	0.003	1.093	0.063
2250	0.306	0.019	0.007	1.035	0.062
2328	0.294	0.022	0.005	0.973	0.059
2395	0.231	0.026	0.009	0.980	0.045
2580	0.162	0.041	0.008	0.871	0.036
2670	0.069	0.042	0.008	0.758	0.028

1% CH₄ & 1% Cl₂ in H₂ at 20 Torr vs. Filament Temperature

MS Probe Parameters: -3% (DISCRIM), -20% (DELTAM), -40% (RES’N), 2600V (SEM), 3.0V (CAGE) 140mA (EMISS).

MS Pressure = 8x10^-7 Torr

Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	HCl (15.6eV)	Cl2 (15.6eV)	CH3 (13.6eV)
23	1.000	0.001	0.000	0.000	1.000	0.000
800	0.683	0.000	0.001	0.973	0.131	0.002
950	0.533	0.000	0.001	0.886	0.023	0.003
1330	0.467	0.000	0.003	0.710	0.000	0.002
1630	0.383	0.002	0.004	0.669	0.000	0.003
1830	0.354	0.007	0.002	0.607	0.000	0.005
1980	0.315	0.010	0.005	0.535	0.000	0.005
2120	0.286	0.010	0.004	0.469	0.000	0.007
2250	0.235	0.024	0.002	0.425	0.000	0.008
2350	0.152	0.037	0.002	0.403	0.000	0.015
2450	0.122	0.045	0.004	0.372	0.000	0.017
2560	0.075	0.048	0.004	0.349	0.000	0.013

1% CH₄ & 2% Cl₂ in H₂ at 20 Torr vs. Filament Temperature

MS Probe Parameters: -3% (DISCRIM), -20% (DELTAM), -40% (RES’N), 2600V (SEM), 3.0V (CAGE) 140mA (EMISS).

MS Pressure = 8x10^-7 Torr

Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	HCl (15.6eV)	Cl2 (15.6eV)	CH3 (13.6eV)
23	1.000	0.001	0.000	0.000	2.000	0.000
800	0.818	0.000	0.000	2.191	0.148	0.003
950	0.723	0.000	0.000	2.011	0.059	0.004
1130	0.614	0.000	0.000	1.798	0.036	0.001
1410	0.557	0.000	0.001	1.699	0.012	0.003
1710	0.476	0.005	0.002	1.560	0.013	0.008
1910	0.394	0.008	0.003	1.273	0.000	0.009
2050	0.355	0.012	0.002	1.251	0.000	0.009
2160	0.308	0.016	0.001	1.007	0.000	0.009
2270	0.275	0.021	0.003	0.911	0.000	0.014
2400	0.191	0.033	0.003	0.837	0.000	0.020
2510	0.091	0.060	0.001	0.822	0.000	0.016
2600	0.053	0.050	0.001	0.616	0.000	0.015

1% CH₄ & 4% Cl₂ in H₂ at 20 Torr vs. Filament Temperature

MS Probe Parameters: -3% (DISCRIM), -20% (DELTAM), -40% (RES’N), 2600V (SEM), 3.0V (CAGE) 140mA (EMISS).

MS Pressure = 6.5x10^-7 Torr

Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	HCl (15.6eV)	Cl2 (15.6eV)	CH3 (13.6eV)
23	1.000	0.000	0.000	0.000	4.000	0.000
800	0.733	0.000	0.000	3.863	0.203	0.001
970	0.611	0.000	0.000	3.468	0.052	0.000
1150	0.485	0.000	0.000	2.683	0.027	0.000
1430	0.446	0.000	0.000	2.459	0.009	0.000
1700	0.407	0.003	0.002	2.427	0.000	0.006
1860	0.473	0.006	0.002	2.063	0.000	0.009
2010	0.327	0.011	0.001	1.935	0.000	0.013
2130	0.231	0.013	0.000	1.745	0.000	0.010
2250	0.198	0.015	0.001	1.427	0.000	0.011
2350	0.141	0.036	0.002	1.301	0.000	0.013
2470	0.075	0.053	0.004	1.163	0.000	0.017
2600	0.030	0.041	0.006	0.828	0.000	0.014