GROUP IIIa HYDRIDES XH 2 AND XH 2 – (X = B, Al, Ga): ELECTRON AFFINITIES AND SINGLET-TRIPLET SPLITTINGS REVISITED Ivan ČERNUŠÁK a1, *, Alena ZAVAŽANOVÁ a,b , Juraj RAAB a2 and Pavel NEOGRÁDY a3 a Department of Physical Chemistry, Faculty of Science, Comenius University, Mlynská dolina, SK-842 15 Bratislava, Slovakia; e-mail: 1 cernusak@fns.uniba.sk, 2 raab@fns.uniba.sk, 3 neogrady@fns.uniba.sk b Nuclear Regulatory Authority of the Slovak Republic, Okružná 5, SK-918 64 Trnava, Slovakia; e-mail: alena.zavazanova@ujd.gov.sk Received September 23, 2002 Accepted November 28, 2002 This paper is dedicated to our teachers, colleagues and friends Petr Čársky, Ivan Hubač and Miroslav Urban on the occasion of their 60th birthdays. We are happy to celebrate their contribution to the formation of Czechoslovak school of quantum chemistry – the centre of scientific excellence in the heart of Europe. Geometries, electron affinities (EA) and singlet-triplet (S-T) splittings of XH 2 /XH 2 – molecules (X = B, Al, Ga) are calculated by coupled-cluster methods, using the sequence of basis sets. The EA values and S-T splittings for aluminium and gallium dihydrides are an order of mag- nitude larger (in absolute values) than those for boron. Fo r boron and aluminium dihydrides, two types of extrapolations towards complete basis set limit are applied, leading to EA = 0.24 eV, ST = –0.01 eV (BH 2 ), and EA = 1.10 eV, ST = –0.62 eV. The best calculated values for gallium dihydrides are EA = 1.13 eV and ST = –0.74 eV. All three S-T splittings fa- vour singlet as the ground state, although the S-T splittings of BH 2 – is exceptionally small. In addition, vertical electron affinities and vertical electron detachments are reported for these molecules. Keywords : Hydrides; Electron affinity; Singlet-triplet splitting; Coupled cluster method; Ab initio calculations; Boron; Aluminium; Gallium. Despite rapid development of highly sophisticated computational tech- niques for calculations of correlation energy, the accurate predictions of ex- cited states, singlet-triplet splitting and electron affinities are still a challenge for current theoretical chemistry. The principal requirements for high-quality calculations of these properties are: inclusion of electron corre- lation at computational limits, balanced treatment of both ion and neutral molecule and a large flexible basis set. In practice, one can only approach these criteria due to limitations of computational resources. Collect. Czech. Chem. Commun. (Vol. 68) (2003) Group IIIa Hydrides 75 doi:10.1135/cccc20030075