Chapter 4 - Results for C/H Systems

4.1 Introduction

As demonstrated in the previous chapter, the technique of molecular beam mass spectrometry is able to provide quantitative measurements of both stable and free radical gas-phase species under conditions typical in the diamond CVD process. Due care has to be taken, however, in the data reduction procedures because the overall system sensitivity is critically dependent on the local temperature, pressure and composition of the gas being sampled. Such variations in the sampling efficiency can be offset by performing simple empirically-based correction procedures. Equipped with a sensitive gas-phase analysis technique and the necessary data reduction procedures for the characterization of a CVD environment, gas phase composition measurements were performed for a variety of source gas mixtures. Extensive work has already been done on the effects that different hydrocarbon precursors (See Table 3.2) have on the CVD diamond growth mechanism by comparing the variation in the composition of the stable gas phase species and CH₃ radicals as a function of filament temperature.^3.1,3.2,3.14 These results are not discussed in detail in this chapter, though the overall conclusions drawn for these hydrocarbon/H₂ systems have been included, in order to enable the reader to identify and understand the changes that occur when other gas phase systems are used, such as C/H/Cl, C/H/N and C/H/P systems (which will be discussed fully in their respective chapters).

4.2 Cracking patterns of CH₄

Figure 4.1 shows how the signal intensities of methane and its dissociation products vary as a function of electron energy in the ionization chamber of the mass spectrometer. By the same linear interpolation method described in Section 3.5 (a), the measured ionization potential (I.P.) for CH₄ and the appearance potentials (A.P.’s) for the different species observed (CH₃, CH₂ and CH) were obtained and compared with literature values (Table 4.1).

Species	Measured I.P. (eV)	Literature Value (eV)^3.21	Measured A.P. (eV)	Literature Value (eV)^3.21
CH4 (m/e=16)	12.4 ± 0.2	12.80	-	-
CH3 (m/e=15)	-	9.96	13.9 ± 0.3	14.3
CH2 (m/e=14)	-	-	17.6 ± 2.4	16.5
CH (m/e=13)	-	-	23.8 ± 1.5	23.4

Table 4.1. I.P.’s and A.P.’s of methane and its observed dissociation fragments.

The A.P.’s of the species observed after electron bombardment of methane relate to the energy required for the following processes:^3.21

A.P. (CH₃): CH₄ ® CH₃⁺ + H + e^-

A.P. (CH₂): CH₄ ® CH₂⁺ + H₂ + e^-

A.P. (CH): CH₄ ® CH⁺ + H + H + H + e^-

Inspection of Figure 4.1 shows that dissociative ionisation of CH₄ produces mainly methyl ions, with only small amounts of CH₂⁺ and CH⁺ ions are formed. This may be related to the number of C-H bonds that need to be broken.

4.3 Cracking patterns of C₂H₂

Figure 4.2 shows how the signal intensity of acetylene and its dissociation products vary as a function of electron energy in the ionization chamber of the mass spectrometer. The I.P. of acetylene and A.P.’s of the dissociation products are summarized in Table 4.2, and compared with the literature values.

Species	Measured I.P. (eV)	Literature Value (eV)^3.21	Measured A.P. (eV)	Literature Value (eV)^3.21
C2H2 (m/e=26)	11.1 ± 0.5	11.41	-	-
C2H (m/e=25)	-	-	17.9 ± 0.5	17.8 ± 0.5
CH (m/e=13)	-	-	22.6 ± 0.4	22.3 ± 0.5

Table 4.2. I.P.’s and A.P.’s of acetylene and its observed dissociation fragments.

Inspection of Figure 4.2 reveals that acetylene contributes by far the greatest signal. Only trace amounts of the C₂H and CH species were observed at high electron energies (above 22 eV).

The A.P.’s of C₂H and CH arise from the following processes:^3.21

A.P. (C₂H): C₂H₂ ® C₂H⁺ + H + e^-

A.P. (CH): C₂H₂ ® CH⁺ + C + H + e^- (or CH⁺ + CH + e^-, neither was given)

4.4 Gas phase composition as a function of filament temperature for 1% CH₄ in H₂

Figure 4.3 shows how the concentration of the major hydrocarbon species [CH₄ (m/e =16), C₂H₄ (m/e =26) and C₂H₂ (m/e =28)] as well as CH₃ radicals change as a function of filament temperature measured 4 mm from the filament for an initial CH₄/H₂ feedstock ratio of 1%. Note that the total carbon balance, defined as (total C fraction measured)/(C fraction in the feed gas), and shown as an upside down triangle in Figure 4.3, decreases as the filament temperature increases, because of the thermal diffusion effects described in Section 3.5 (e).

The main chemical conversion occurring in the chamber is that of methane to acetylene as the filament temperature is increased, a reaction that is initiated by the reaction of H atoms with methane to produce methyl radicals:

CH₄ + H CH₃ + H₂ (4.1)

Methyl recombination followed by successive H abstraction 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 4.3. An increase in the filament temperature results in higher [CH₃] which is mirrored by increased [C₂H₂]. Therefore the CH₃ radical is an essential intermediate in the formation of acetylene.

4.5 Discussion of errors

Shown in Figure 4.4 are the errors that are inherent in the calculated species concentrations for the various stable hydrocarbon species (and CH₃ radicals) monitored. These values were obtained using standard combination of errors techniques, in which the error in the derived parameter (in this case the species concentrations) is determined from the measured parameters (i.e. the MBMS signal intensities of each of the measured species). Calculations for both stable and free radical species were semi-automated using a spreadsheet program (ASEASYAS). Similar calculations were performed for all the other gas phase systems, including C/H/Cl, C/H/N and C/H/P systems. However, since the gas phase species lie very close to each other on the graphs, their concentrations are presented with no error bars for clarity. The errors shown in Figures 4(a)-(d) are representative of the magnitude observed for all subsequent MBMS experimental runs.

It is worth considering the sources of other uncertainties when interpreting the MBMS data. The factors affecting the error that are inherent in the calculated species mole fractions not only include (1) the sensitivity of the MS and (2) the signal/noise ratio, but also (3) changes in the kinetic energy of the electron bombardment (which will relate to the age of the cathode filaments) and (4) filament/sampling cone distances. These factors will inevitably influence the uncertainties in the run to run reproducibility of the MS system, and hence the accuracy of the measured species mole fractions.

There are also errors that may be inherent in the measured values of the filament temperature. Such errors will arise from uncertainty in the positioning of the pyrometer and the emissivity of the filament, which reflects on the age of and history of the filament. In the latter case, for example, the emissivity of the tantalum filament may be different to that of tantalum carbide which is formed when the filament is carburised by CH₄. Throughout the MBMS experiments conducted in this work the tantalum filament had been previously ‘seasoned’ or carburised with the source gas mixture later being examined.

4.6 Gas composition as a function of filament temperature for a variety of hydrocarbon precursor gases

Similar measurements of the stable gas species were made using different hydrocarbon precursor gases.^3.1,3.2 It has been established that regardless of the hydrocarbon precursor gas used, at filament temperatures between 2000 and 2100°C the subsequent chemistry is such that methane becomes the dominant hydrocarbon species. Furthermore, at filament temperatures near to, and above, the optimum for diamond growth (ca. 2400°C) the relative concentrations of the various stable hydrocarbon species present in the gas mixture, and the way the concentrations vary with temperature, are both remarkably insensitive to the choice of hydrocarbon feedstock gas.

4.7 Analysis of the films grown using 0.5% and 1% CH₄ in H₂

Figure 4.5 shows scanning electron micrographs (top view and cross-section) of a polycrystalline diamond film grown on silicon (100) using 0.5% CH₄ in H₂ under standard deposition conditions (See Section 3.1). Figure 4.6 shows a similar film grown using 1% CH₄ in H₂ gas mixture. The growth rates (Table 4.3) were calculated from the film thickness, determined from cross-sectional SEM images, divided by the time of growth (6 hours). These deposition experiments were carried to allow the comparison between films grown using standard methane/H₂ gas mixtures with those grown using other gas phase systems, such as C/H/Cl, C/H/N and C/H/P systems, following claims that addition of small quantities of these elements has noticeable effects on the growth rate, the morphology and the resulting quality of the diamond films. Full discussion of each of these gas phase systems are given in their respective chapters (5-7).

Figure 4.5(a). Scanning electron micrograph (SEM) of a polycrystalline diamond film grown on silicon using input gas mixtures of 0.5% CH4 in hydrogen.	Figure 4.5(b). Scanning electron micrograph showing the cross sectional view of the film.
Figure 4.6(a). Scanning electron micrograph (SEM) of a polycrystalline diamond film grown on silicon using input gas mixtures of 1% CH4 in hydrogen.	Figure 4.6(b). Scanning electron micrograph showing the cross sectional view of the film.

Gas Mixture	Film Thickness (mm) after 6 hours growth	Deposition Rate (mm/hr)
0.5% CH4 in H2	1.9	0.32
1% CH4 in H2	2.4	0.40

Table 4.3. Diamond film thickness and deposition rate for 0.5% and 1% CH₄ in H₂ gas mixtures.

4.8 Appendix: Experimental Data

The following tables show actual experimental data recorded for 1% CH₄ in H₂. Table 4.4 shows the signal intensities measured for the stable hydrocarbon species as well as CH₃ radicals as a function of filament temperature. Table 4.5 shows signal intensities of the stable species after background subtraction and MS sampling efficiency corrections. The mole fractions of these species were then determined by direct room temperature calibration, and are shown in Table 4.6. Methyl radical concentrations were also determined using a slightly different procedure (See Section 3.6 (b) ). Error calculations using ASEASYAS are displayed in the tables for five selected temperature readings.

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

MS Probe Parameters: -6% (DISCRIM), -20% (DELTAM), -40% (RES’N), 2550V (SEM), 3.0V (CAGE) 140mA (EMISS).

MS Pressure = 8x10^-7 Torr

Table 4.4. Signal Intensity vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	CH3 (13.6eV)	H2 (15.6eV)
23	1470 ± 32	8 ± 2	3 ± 1	0 ± 0	141000 ± 3000
800	1008	5	4	6	119000
1400	832	6	3	6	108000
1600	707 ± 17	6 ± 1	3 ± 1	5 ± 1	100000 ± 4000
1810	602	14	7	6	95000
1910	549 ± 16	18 ± 3	10 ± 2	5 ± 1	93000 ± 4000
2030	520	19	7	6	91000
2120	465	20	7	5	88000
2210	423 ± 15	21 ± 4	10 ± 2	5 ± 1	87000 ± 3000
2310	381	26	11	6	83000
2400	305	34	13	9	83000
2490	230 ± 9	55 ± 5	15 ± 2	8 ± 1	79000 ± 4000
2575	168	67	15	11	77000
2655	140	73	18	8	74000
Background Signal	4 ± 1	3 ± 1	1 ± 1	0 ± 0	3560 ± 500

Table 4.5. Signal Intensity vs. Filament Temperature after background subtraction and MS efficiency corrections

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	CH3 (13.6eV)
23	1466 ± 108	5.0 ± 3	2.0 ± 2	0.0 ± 0
800	1195	2.3	3.6	2.2
1400	1090	3.9	2.6	3.4
1600	1002 ± 98	4.3 ± 4	2.9 ± 3	3.0 ± 2
1810	899	16.5	9.0	5.3
1910	837 ± 90	23.1 ± 8	13.8 ± 5	4.2 ± 2
2030	811	25.1	9.4	6.1
2120	750	27.7	9.8	5.0
2210	690 ± 73	29.6 ± 10	14.8 ± 5	5.4 ± 2
2310	652	39.8	17.3	7.7
2400	521	53.6	20.8	13.4
2490	412 ± 54	94.7 ± 19	25.5 ± 7	12.9 ± 3
2575	307	119.8	26.2	19.3
2655	265	136.6	33.2	14.5

Room temperature calibration of stable species

1% CH₄ in H₂ at 20 Torr = 1466 c/s ± 33 (15.6eV)

1% C₂H₂ in H₂ at 20 Torr = 3260 c/s ± 44 (15.6eV)

1% C₂H₄ in H₂ at 20 Torr = 1131 c/s ± 30 (13.6eV)

1% CH₃ in H₂ at 20 Torr = 606 c/s ± 39 (13.6eV)*

Table 4.6. Species Mole Fractions vs. Filament Temperature

Fil. Temp (°C)	CH4 (15.6eV)	C2H2 (15.6eV)	C2H4 (13.6eV)	CH3 (13.6eV)
23	1.000 ± 0.098	0.002 ± 0.001	0.002 ± 0.002	0.000 ± 0.001
800	0.813	0.001	0.003	0.004
1400	0.741	0.001	0.002	0.006
1600	0.682 ± 0.083	0.001 ± 0.001	0.003 ± 0.003	0.005 ± 0.003
1810	0.611	0.005	0.008	0.009
1910	0.570 ± 0.075	0.007 ± 0.003	0.012 ± 0.005	0.007 ± 0.004
2030	0.552	0.008	0.008	0.010
2120	0.510	0.008	0.009	0.008
2210	0.469 ± 0.061	0.009 ± 0.003	0.013 ± 0.005	0.009 ± 0.004
2310	0.444	0.012	0.015	0.013
2400	0.354	0.016	0.018	0.022
2490	0.280 ± 0.043	0.029 ± 0.006	0.023 ± 0.007	0.021 ± 0.006
2575	0.209	0.037	0.023	0.032
2655	0.181	0.042	0.029	0.024

* The error for 1% CH₃ signal may be larger than the value quoted here since there will be errors arising from the determination of the beam component of the sampled gas (~0.35 of the total signal). See Section 3.5 (i) for details of the measurements.