Journal of Physical Chemistry B, Vol.106, No.42, 10816-10827, 2002
A confocal Raman imaging study of an optically transparent boron-doped diamond electrode
Confocal micro-Raman imaging was used to investigate the structural and, to some extent, the chemical inhomogeneities that exist in a high-quality, free-standing, polished, optically transparent boron-doped chemical-vapor-deposition (CVD) diamond disk. A number of boron-related Raman lines centered at about 610, 925, 1045, 1375, and 1470 cm(-1) were evidenced and found to vary from one region of the disk to another. These lines have previously been reported in the literature for other boron-doped crystals or films obtained using completely different growth conditions. Even if their origin is still somewhat uncertain, their presence enabled the construction of spatial maps of the boron signal intensity over the surface. It was confirmed that boron is incorporated into CVD diamond with significant inhomogeneities. In most cases, the boron signal intensity tracked the structure of the grains. The high quality of the disk also allowed a detailed analysis of the grain boundaries to be made. However, a unique description of these grain boundaries in terms of their microstructure and chemical composition could not be developed. There is some evidence that nondiamond phases are incorporated into the grain boundaries, but their Raman signature is quite distinct from that usually reported. Another remarkable feature of this disk is that it is nearly photoluminescence-free. Photoluminescence, as excited by the 514.5 or 457.9 nm lines of an argon ion laser, was only detected within some of the nondiamond amorphous carbon-rich grain boundaries. The low photoluminescence background was also a common feature of a heavily boron-doped, polycrystalline film deposited on silicon, another sample that was also examined in this work for comparison. Despite its high quality, the optically transparent disk is far from being stress-free. Localized stress variations were observed. The strained regions are not systematically associated with the incorporation of nondiamond carbon. Grain boundaries and, most probably, planar defects have been identified as the stress sources. However, it remains difficult to explain the large compressive stress observed for some of the crystals.