Microwave Plasma Enhanced Chemical Vapour Deposition of Diamond

Microwave plasma enhanced chemical vapour deposition (MWPECVD) is used at Bristol to deposit CVD diamond under varying growth conditions. Films can then be analysed ex situ by Scanning Electron Microscopy (SEM), Laser Raman Spectroscopy (LRS), Field Emission studies, AES, SIMS etc. Previously we did a lot of work studying the gas-phase chemistry of these plasmas using various laser spectroscopic methods and an in situ molecular-beam mass spectrometer. Nowadays our two MWPECVD systems are used for growth experiments and to make devices such as betavoltaic or gammavoltaic devices.

Plasma-assisted deposition methods

Plasma-assisted deposition techniques are very popular methods for growing CVD diamond. Plasma systems offer uniform films over larger substrate areas than hot filament CVD growth, with the possibilty of industrial scale up by using more powerful reactors. The plasma in a microwave system is detached from any reactor surface, hence no impurities from reactor construction materials enter the film bulk during deposition. In hot-filament CVD, incorporation of filament materials into the film occurs during deposition and this (combined with limited deposition area) make hot-filament methods useless for high-purity commercial applications of diamond. However hot-filament reactors are cheaper and simpler than microwave reactors and a lot of useful research is done in them.

Our Microwave systems

Both of our MW systems are top-hat (ASTeX-style) reactors consisting of a microwave generator coupled to the top of a cylindrical, water cooled, stainless steel chamber. A resonant mode, created by the microwaves, is supported in the chamber and the reactant gases are heated and excited to form a plasma ball. The substrate sits ~1 mm below the visible edge of the plasma ball, on top of molybdenum substrate holder. The reactors are also equipped with diagnostic ports.

The 2 kW MW reactor The 3 kW MW reactor
The 2 kW MW reactor used for undoped diamond. The 3 kW MW reactor used for B-doped diamond.

We have 2 reactors, one for boron-doping and one for non-boron work. The boron-reactor has a 3 kW MW power supply and has the usual CVD gases attached to it (CH4, H2, O2, Ar, etc.), plus also 1%B2H6 in H2 for doping. The non-boron reactor has a 2 kW power supply and has the same gases (except B2H6), plus many others for use in plasma experiments, such as N2, NH3, etc.

Both reactors are capable of growing high-quality polycrystalline diamond films at rates of around 10 μm h-1 uniformly over a substrate area of 2×2 cm. Films of many 10's of μm can easily be grown in a few hours, or even 100's of μm by combining together several day's worth of growth runs. The reactors can also grow single-crystal epitaxial layers (undoped or B-doped) onto SCD substrates. This is useful for adding a conducting epitaxial layer to a commercial SCD substrate which is necessary to prevent charging during XPS analysis.

The Plasma Balls

Over the years we have studied plasmas made from many different types of gas mixture. Photos and diagrams of these various plasma balls can be seen below.

standard low-power 1%CH4/H2 plasma
standard low-power 1%CH4/H2 plasma		 100% CO2 plasma
100% CO2 plasma		 50%CO2/50%CH4 plasma
CO2-rich plasmas		 30%CO2/70%CH4 plasma
70% CH4/30% CO2 plasma
The standard low-power gas composition is 1%CH4/H2, which produce a light blue/purple plasma ball due to emission from hydrogen. A 100% CO2 plasma is deep blue with a whitish centre. For CO2-rich plasmas (upto 50% CO2 in CH4) the plasma becomes more light blue in colour, but there is still no diamond growth. Diamond growth is confined to a very narrow window around the 50:50 mixture. With increasing CH4 concentration in CO, the plasma develops a yellow/orange halo around its extremeties due to blackbody emission from macroscopic soot particles as well as excited C2. Close to the wafer surface there is a violet/purple haze. The plasma also starts to deposit soot around the cooler parts of the chamber walls.
100% CH4 plasma
CH4 plasma		 1% CH4 in Ar plasma
Ar CH4 plasma		 High pressure CH4/H2(/Ar) plasma ball - greenish
For 100% CH4 plasmas, the ball is very bright white in the centre with the orange halo now very pronounced and extending a long way into the chamber. The soot production rate is very fast, making it difficult to sustain these plasmas for more than a few minutes. A 1% CH4 in 99% Ar plasma is much colder than other plasmas, but is also quite sooty. These plasmas are used to grow UNCD films. A 5% CH4/10%Ar/85%H2 Ar plasma running at pressures over 150 Torr looks much greener, due to emission from the C2 radical. These plasmas are used to grow polycrystalline diamond films at higher rates.

Low Temperature Diamond CVD Using CO2/CH4 Gas Mixtures

The now well established conditions for diamond growth include the use of substrate temperatures >>700°C and a carbon-containing precursor gas diluted in excess hydrogen (typically <5% CH4 in H2). A major goal in the field of diamond CVD is the lowering of substrate temperatures required for growth, as this could permit the use of a much wider range of substrate materials of industrial importance, such as aluminum, GaAs, nickel and steel.

Many gas mixtures containing varying ratios of O, C and H have been investigated in the search for a viable low temperature diamond deposition process. Work within our group has centred on microwave Plasma CVD diamond deposition using CO2/CH4 gas mixtures. Investigations of both low-temperature growth and gas-phase plasma chemistry have been carried out, results of which are discussed below.

Low Temperature Diamond Growth

It has been found that the highest quality diamond is obtained using the mixture 50%CH4/50%CO2. This gas mixture has also been found to allow diamond growth at substrate temperatures <500°C although a clear reduction in the crystallinity of the films with reduced temperature can be seen.

Lowering substrate temperature also has the effect of lowering film growth rate . Using these growth rate data, an Arrhenius plot can be produced which yields an overall activation energy for diamond deposition of 28 kJ mol-1. This is lower than previously obtained for CH4/H2 systems and hints at a different surface chemistry in the two systems.

growth rate arrhenius plot
Growth rate vs substrate temperature Arrhenius plot of growth rate data

CO2/CH4 plasma chemistry

Studies have shown that CO2/CH4 mixtures containing >50% CO2 result in no film deposition, 50%CO2/50%CH4 gives optimum diamond growth whilst increasing %CH4 above 50% produces increasingly graphitic material. In order to explain these observations molecular-beam mass spectrometry was used to make measurements of plasma species concentrations over a wide range of mixing ratios (0-80% CH4). Computer simulation of these results was then carried out using the CHEMKIN II computer package thus allowing the important plasma reactions to be identified.

  1. CO2 + CH4 → 2CO + 2H2
  2. CO2 + H2 → CO + H2O
  3. H2O + C2H2 → CH4 + CO

Reaction 3 is important because it leads to a peak in CH4 (and therefore CH3) concentration at 50% input CH4. CH3 is believed to be the growth species in diamond CVD using CH4/H2 gas mixtures. The peak in CH3 concentrations at the same composition as the optimum diamond growth suggests that CH3 is also the diamond growth species for CO2/CH4 gas mixtures.

The different activation energy obtained for 50%CO2/50%CH4 compared with CH4/H2 gas mixtures suggests a difference in the growth mechanism. This could be due to a difference in the surface termination of the growing diamond surface which is thought to be performed by H in CH4/H2 systems. Given the relatively high concentration of CO present at 50%CO2/50%CH4 it seems reasonable that at least some of the diamond surface could be terminated with CO (or HCO) in this system.

Related Papers