Search

The use of CVD diamond in electronics has very stringent requirements. For a CVD diamond industry to become viable it is mandatory to obtain very large growth rates (> 5 µm/h), all the while maintaining extremely high purity, a crystalline defect density as low as possible, and large usable surface areas. At the same time, one must keep the stress level within the growing crystal below acceptable limits to avoid crack formation and preserve the crystal structural integrity. These imperatives imply to work to improve both the plasma deposition process and the CVD diamond crystal growth. In this paper, we propose a three-pronged approach: (i) We use detailed plasma models to establish the influence of process parameters (in particular deposition pressure) on plasma chemistry in order to optimize film growth rate and diamond quality; (ii) We emphasize the need for careful substrate pre-treatment and selection (including choosing a single-sector face) to minimize defects in the growing films; (iii) We employ a 3D geometrical model to predict the crystal shape under given growth conditions, and exploit this knowledge to devise a growth strategy maximizing the usable film surface area while minimizing stresses inside the films.

The quality of single crystal diamond obtained by microwave CVD processes has been drastically improved in the last 5 years thanks to surface pretreatment of the substrates [A. Tallaire, J. Achard, F. Silva, R.S. Sussmann, A. Gicquel, E. Rzepka, Physica Status Solidi (A) 201, 2419-2424 (2004); G. Bogdan, M. Nesladek, J. D'Haen, J. Maes, V.V. Moshchalkov, K. Haenen, M. D'Olieslaeger, Physica Status Solidi (A) 202, 2066-2072 (2005); M. Yamamoto, T. Teraji, T. Ito, Journal of Crystal Growth 285, 130-136 (2005)]. Additionally, recent results have unambiguously shown the occurrence of (110) faces on crystal edges and (113) faces on crystal corners [F. Silva, J. Achard, X. Bonnin, A. Michau, A. Tallaire, O. Brinza, A. Gicquel, Physica Status Solidi (A) 203, 3049-3055 (2006)]. We have developed a 3D geometrical growth model to account for the final crystal morphology. The basic parameters of this growth model are the relative displacement speeds of (111), (110) and (113) faces normalized to that of the (100) faces, respectively alpha, beta, and gamma. This model predicts both the final equilibrium shape of the crystal (i.e. after infinite growth time) and the crystal morphology as a function of alpha, beta, gamma, and deposition time. An optimized operating point, deduced from the model, has been validated experimentally by measuring the growth rate in (100), (111), (110), and (113) orientations. Furthermore, the evolution of alpha, beta, gamma as a function of methane concentration in the gas discharge has been established. From these results, crystal growth strategies can be proposed in order, for example, to enlarge the deposition area. In particular, we will show, using the growth model, that the only possibility to significantly increase the deposition area is, for our growth conditions, to use a (113) oriented substrate. A comparison between the grown crystal and the model results will be discussed and characterizations of the grown film (Photoluminescence spectroscopy, EPR, SEM) will be presented. (C) 2008 Elsevier B.V. All rights reserved.