An experimental setup for the deposition of diamond by an acetylene-oxygen combustion flame was developed in which it is possible to control the deposition temperature within 10 °C during long-term experiments. This degree of accuracy is obtained by cooling the back of the substrate holder with an electronically controlled water injection system and using soldered substrates, ensuring a good thermal contact. With this setup a large number of experiments were performed with deposition times of I h. Special attention was given to the growth of diamond layers with uniform thickness and morphology because these properties are considered essential for future applications. It was found that the deposition temperature, the total gas flow, the composition of the gas phase and the position of the substrate in the flame all have major influences on the deposit. Not only the growth rate and the quality of the diamond are strongly influenced by the deposition conditions, but also the preferred crystal habit and the homogeneity.

No abstract

AbrtracL Epitaxial, boron-doped diamond films were grown by hot~filament~asisted chemical vapour deposition (CvD) on {IW) and {IIO). natural diamond substrates.Resistivity measuremenis for 10K < T < SOOK showed a clear transition from band to hopping conduction upon lowering of temperature. in the band conduction regime, the {lo01 films had higher conductivity than the {IIO) samples. The reverse was found in the hopping regime. This is explained by the difference in crystal gmwth mechanisms, leading to higher boron concentrations and lower carrier mobilities for {IlO) samples than for (100) oriented films. Hall effect measurements were performed for the mm1 lightly doped (100) film at a boron level of 2.7 x IOt8 in the band conduction region up to 7SOK. A mobility maximum of @H = S90cm2V-ts-' at 29SK was found, and the wmpensation ratio was determined to be smaller than 0.02. Some preliminaly values for the Hall effective mass of valence band holes are given.

The surface morphologies of numerous homoepitaxial, chemical-vapour-deposition-grown diamond films have been examined by phase-sensitive optical microscopy. The layers were produced by the hot filament technique as well as by the acetylene oxygen combustion method. The {I11} and !1001 faces manifest themselves as F faces below the roughening temperature and grow via steps nucleated at three-dimensional diamond particles or at dislocations. The rate-determining step in '1111 diamond growth is a hindered surface diffusion of the growth species towards the steps. The layer-by-layer growth on the ~ 1001 faces is discussed in terms of (2 x 1) surface reconstruction in combination with the presence of a 4~ screw axis. The II131 face on flame-grown diamonds is made up of strong <~110) chains of bonds which are interconnected by weak forces due to surface reconstruction. The slight curvature of this face points to an F face close to the roughening point. The {1 I01 face is rough, i.e. K,'S type, and no layer growth occurs. The different modes of crystal growth as well as local differences in step spacing are replicated as variations in the intensity of band A and 575 nm cathodoluminescence.

Relatively thick films of diamond were deposited homoepitaxially on cylindrical substrates from an acetylene-oxygen combustion flame. On these single crystals a number of stable F faces are present. For the first time stable facets other than {l 11) and {100} are observed, namely (113) facets. These facets occur under conditions of low surface coverage, e.g. at high deposition temperatures. This confirms the prediction of a previously developed crystal growth model. A preliminary investigation of the relative stability of the (113) faces as function of the deposition temperature is described.