PACS 71.55.Ht, 71.55.Mb The usual p-type dopants of SiC, B, and Al, do not produce really shallow levels. In fact, boron can give rise to a secondary very deep acceptor level as well. The picture is complicated by hydrogen which is readily incorporated during in-growth doping into p-type material and can passivate both dopants but to a different degree. First principle calculations are reported regarding the interaction of hydrogen with B and Al in SiC. The results explain why hydrogen is incorporated in much higher amounts into B-doped than into Al-doped samples, and also reveal the influences of hydrogen on boron to produce the shallower acceptor. It will be shown that hydrogen incorporation during growth does not influence Al. Finally an Al -N -Al complex is proposed as a shallower acceptor in SiC.Introduction Silicon carbide is a wide band gap semiconductor intended for use in high power, high temperature electronics. This fact increases somewhat the tolerance for what can be termed a shallow level, still the traditional p-type dopants, boron and aluminium, seem to behave rather unsatisfactorily in SiC. Boron can produce a deep acceptor state about 0.55 eV above the valence band, as well as a shallower one at 0.30 eV [1], and it is not clear why the former appears. Aluminium gives only one level at about 0.21 eV [2], which is still too high for planned high-frequency devices operating at moderate temperatures. (Note that SiC comes in many different polytypes, the most important ones being the cubic 3C and the hexagonal 4H and 6H. The numbers here denote the number of inequivalent Si-C bilayers stacked on each other within one period along the cubic [111] or the hexagonal [0001] direction. Unless otherwise noted, values quoted refer to the 4H polytype.) Doping of SiC is complicated by two facts. For one, in a compound semiconductor dopants can choose between the two sublattices where they can behave rather differently. Second, the strong bonds and tight structure of SiC precludes diffusion doping. Therefore, doping may be achieved in growth during chemical vapour deposition (CVD) of homoepitaxial layers, or by implantation. The former procedure introduces hydrogen impurities, the latter causes intrinsic point defects -in both cases the dopants may form complexes.As for the site selection, the behaviour of Al is relatively predictable: it prefers the silicon site (Al Si ), where it fits in more conveniently, and produces the "shallow" acceptor level mentioned above. Boron, on the other hand, has been reported to occupy both, the silicon and the carbon site, alas, with preference for the Si site [3]. If SiC is grown under Si-rich conditions in CVD (offer of empty C sites for B), the boron incorporation is substantially lower than in the case of C-rich conditions (offer of empty Si sites). The latter (C-rich growth) produces the "shallow" acceptors mentioned above while the "deep" ones were observed in the former (Si-rich growth). Therefore, although no conclusive proof has yet been