The thin SiC layer often observed between a Si substrate and a diamond
him has led to the belief that the formation of this SiC layer preced
es diamond nucleation. However, the nucleation of diamond on a clean,
undamaged SiC surface is difficult. Some of the defects introduced by
abrading the surface of many substrate materials with diamond powder b
efore deposition appear to serve as nucleation sites. Atomic-scale def
ects (vacancies, interstitials) and impurities can be introduced by io
n implantation. Single-crystal 6H SIC with the external face composed
of Si or C were implanted with silicon (150 keV) and carbon (55 keV) i
ons at fluences between 2 x 10(14) ions cm(-2) and 2 x 10(15) ions cm(
-2). Nucleation did not occur on as-received or implanted samples. Abr
asion with diamond powder using an ultrasonic cleaner bath caused prof
use nucleation on all samples. It was found that nucleation density on
abraded samples decreased when implanted with increasing fluence of i
ons. Fluences of 2 x 10(14) Si+ ions cm(-2) (150 keV) and 4 x 10(14) C
+ ions cm(-2) (55 keV) did not effect the nucleation of diamond on sil
icon carbide. However, a fluence of 2 x 10(15) Si+ ions cm(-2) complet
ely suppresses the nucleation of diamond. Rutherford backscattering sp
ectrometry/channeling studies on the high-fluence-implanted silicon ca
rbide sample showed the presence of an amorphous surface layer suggest
ing that CVD diamond does not nucleate on amorphous SiC.