Ab Initio Calculations of the Trigonal and Zero-Field Splittings in Trischelated Diketonato Complexes of Trivalent Chromium Carl Ribbing, Kristine Pierloot, and Arnout Ceulemans* Division of Quantum Chemistry, University of Leuven, Celestijnenlaan 200F, B-3001 Leuven, Belgium ReceiVed February 12, 1998 The ground and excited states of the neutral tris(1,3-propanedionato)chromium(III) d 3 complex are investigated with quantum chemical methods. Trigonal splittings are calculated and compared with experiment for the first two excited quartet states 4 T 2g and 4 T 1g . The effect of spin-orbit coupling is also introduced, and a reassignment is proposed for three absorption bands in the doublet region. Trigonal symmetry-induced mixing between the e g and t 2g shells is found to be responsible for the unusually large zero-field splitting of the 4 T 2g ground state. Orbitals pictures are presented which illustrate the role of the phase-coupling effect from the unsaturated ligands. 1. Introduction The tris(acetylacetonato)chromium(III) complex, Cr(acac) 3 , has been the subject of numerous spectroscopic, magnetooptical, and theoretical studies. 1 Especially the π -bonding role of the conjugated bidentate chain and the possible occurrence of a phase-coupling effect have attracted attention. 2 On the basis of a detailed analysis in the framework of the angular overlap model (AOM) Atanasov et al. concluded that the trigonal splitting of the first spin-allowed band, 4 T 2g r 4 A 2g , by some 800 cm -1 was indeed due to the phase-coupling effect. 3 In a further study 4 it was argued that the unusually large zero- field splitting (ZFS) of the ground state 5 by 1.2 cm -1 could only be explained by assuming an anisotropic spin-orbit coupling mechanism. The low-temperature emission spectrum was assigned to phosphorescense from the 2A h ( 3 / 2 ) spin-orbit component of a trigonal 2 E state with predominant 2 E g octahedral parentage. 1,4 The extremely large ZFS 6 and unusual g tensor 7,8 of the doublet state remained poorly understood. Progress in correlated methods of computational chemistry currently allows us to perform electronic structure calculations that can throw a new light on these issues. In this paper we report the first investigation of a trischelated Cr(III) complex with the unsubstituted 1,3-propanedionato ligand, PDO - , using the complete active space methods CASSCF and CASPT2. The conclusions corroborate the results by Atanasov et al. concerning the phase-coupling origin of the trigonal splitting. 3,4 In addition a new assignment for the doublet bands is suggested. 2. Computational Details The calculations were performed with the MOLCAS program system, 9 and the effective symmetry was limited to the C 2 subgroup of the actual D 3 point group. Optimized orbitals were obtained with the CASSCF method for each state. The active space was restricted to the five d orbitals. It was noticed during optimization that the orbitals acquired some ligand character, but only to a minor extent. Orbitals of e, a 1 , and a 2 type were prevented from mixing by using the supersymmetry option in the CASSCF program. To introduce further correlation into the calculations a CASPT2 9,10 calculation was performed for each state, using the CASSCF wave function as reference. In the CASPT2 calcula- tions all electrons, except 1s, 2s, and 2p on chromium and 1s on carbon and oxygen, were correlated (95 electrons). ANO type basis sets 11 were chosen as follows: Cr(17s12p9d4f/ 6s4p3d1f), O(10s6p/3s2p), C(10s6p/3s2p), and H(7s/2s). The geometry of the Cr(PDO) 3 complex was determined in the following way. First we optimized the free PDO - ligand with the constraint that the O-O distance is 2.77 Å (see below). This produced a flat C 2V geometry for the PDO - ligand. If the numbering of the carbons along PDO is O-C 1 -C 2 -C 3 -O, the geometry is the following: distances (Å), O-O ) 2.77, O-C 1 ) 1.33, C 1 -C 2 ) 1.39, C 1 -H 1 ) 1.10, C 2 -H 2 ) 1.08; angles (deg), O-C 1 -C 2 ) 124.6, O-C 1 -H 1 ) 119.1, C 1 -C 2 -H 2 ) 117.8. If the PDO - ligands are placed around the Cr 3+ ion so that the oxygens are in exact octahedral positions, an O-O distance of 2.77 Å corresponds to a Cr-O distance of 2.77/2 ) 1.96 Å. For the closely related Ga(acac) 3 and Cr(acac) 3 complexes the geometry is available from X-ray diffraction data 12-14 and the coordination of the oxygens is very close to (1) Scho ¨nherr, T. Top. Curr. Chem. 1997, 191, 87. (2) Ceulemans, A.; Vanquickenborne, L. G. Pure Appl. Chem. 1990, 62, 1081. (3) Atanasov, M. A.; Scho ¨nherr, T.; Schmidtke, H. H. Theor. Chim. Acta 1987, 71, 59. (4) Atanasov, M.; Scho ¨nherr, T. Inorg. Chem. 1990, 29, 4545. (5) Elbers, G.; Remme, S.; Lehmann, G. Inorg. Chem. 1986, 25, 896. (6) Scho ¨nherr, T.; Eyring, G.; Linder, R. Z. Naturforsch. 1983, 38, 736. (7) Fields, R. A.; Haindl, E.; Winscom, C. J.; Khan, Z. H.; Plato, M.; Mo ¨bius, K. J. Chem. Phys. 1984, 80, 3082. (8) Fields, R. A.; Winscom, C. J.; Haindl, E.; Plato, M.; Mo ¨bius, K. Chem. Phys. Lett. 1986, 124, 121. (9) Andersson, K.; Fu ¨ lscher, M. P.; Karlstro ¨ m, G.; Lindh, R.; Malmqvist, P.-Å.; Olsen, J.; Roos, B. O.; Sadlej, A. J.; Blomberg, M. R. A.; Siegbahn, P. E. M.; Kello ¨, V.; Noga, J.; Urban, M.; Widmark, P.-O. MOLCAS, Version 3; Department of Theoretical Chemistry: Chemistry Center, University of Lund, P.O.B. 124, S-221 00 Lund, Sweden, 1994. (10) Andersson, K.; Malmqvist, P.-Å.; Roos, B. O.; Sadlej, A. J.; Wolinski, K. J. Phys. Chem. 1990, 94, 5483. (11) Pierloot, K.; Dumez, B.; Widmark, P.-O.; Roos, B. O. Theor. Chim. Acta 1995, 90, 87. (12) Dymock, K.; Palenik, G. J. Acta Crystallogr. 1974, 30, 1364. (13) Morosin, B. Acta Crystallogr. 1965, 19, 131. 5227 Inorg. Chem. 1998, 37, 5227-5232 S0020-1669(98)00161-X CCC: $15.00 © 1998 American Chemical Society Published on Web 09/01/1998