Triplet–singlet energy gaps in iodo-carbenes (I–C–X) : Remarkable discrepancy between theory and experiment Hue Minh Thi Nguyen,ac and Minh Tho Nguyen*a Bala  zs Hajgato  ,ab Tama  s Veszpre  mib a Department of Chemistry, University of L euven, Celestijnenlaan 200F, B-3001 L euven, Belgium. E-mail : minh.nguyen=chem.kuleuven.ac.be b Department of Inorganic Chemistry, Budapest University of T echnology and Economics, 4, H-1521 Budapest, Hungary. E-mail : Gelle  rt te  r veszpremi.inc=chem.bme.hu c Faculty of Chemistry, National University of Education, Hanoi, V ietnam Received 29th June 2000, Accepted 25th September 2000 First published as an Advance Article on the web 30th October 2000 The tripletÈsinglet energy gaps in iodo-carbenes, IÈCÈX with X \ H, F, Cl, Br and I, were computed using the MP2, CCSD(T), CASSCF/CASPT2, MR-SDCI, MR-ACPF and B3LYP methods, with basis sets up to 6-311&&G(3df,2p) and aug-cc-pvTZ and e†ective core potentials. Corrections for relativistic e†ects were also incorporated. Our results indicate that diiodo-carbene is likely to possess a singlet ground state lying (CI 2 ) around 30 kJ mol~1 below the lowest triplet state, thus at variance with that of a recent negative ion photoelectron spectroscopic study (R. L. Schwartz, G. E. Davico, T. M. Ramont and W. C. Lineberger, J. Phys. Chem. A, 1999, 103, 8213). In addition, the present study conÐrms the singlet character of other iodo-carbenes with substantial tripletÈsinglet gaps and also points out the remarkably large Ñuctuations and discrepancies in the absolute values of the energy splittings, not only between experiment and theory, but also between ab initio quantum chemical methods. Carbenes are neutral compounds featuring a diva- (R 1 ÈCÈR 2 ) lent carbon atom and two non-bonded electrons. Owing to the presence of two near-degenerate 2p orbitals at the carbe- noid centre (the in-plane 2p(r) and out-of-plane 2p(p) orbitals), distribution of the two unpaired electrons generates three di†erent electronic states, namely a triplet, an open-shell singlet and a closed-shell singlet state. This constitutes a clas- sical example for the 2-electron-2-orbital problem. The ground electronic state of a carbene usually has either a triplet multiplicity or a closed-shell singlet. The carbene reactivity pattern is inherently dependent on their electronic structure : while closed-shell singlet carbenes usually undergo stereospe- ciÐc insertion or addition reactions, their triplet counterparts often shown non-speciÐc abstraction reactions.1,2 Therefore, knowledge on the identity of ground state and tripletÈsinglet splittings is of fundamental importance in understanding the chemistry of these reactive intermediates. In this context, theo- retical studies have proved to be more than complementary to experimental studies, as witnessed by the long history of the parent methylene.3 It is well known that the substituents R 1 and play a crucial role in favouring one or other electronic R 2 state. While methylene and methylcarbene (CH 2 )3 (CH 3 CH)4 are triplet species, dimethylcarbene and prob- (CH 3 CCH 3 )5,6 ably higher dialkylcarbenes exhibit a singlet ground state. On the contrary, most diarylcarbenes have a persistent triplet multiplicity.7 The singlet carbenes could usually be stabilised by strong p-donor groups such as F, Cl, OR, SR, NR 2 , PR 2 , etc. that reduce the p-electron deÐciency.8h12 In playing with the e†ect of these groups either with a cyclic or open frame- work, it has been possible to make stable, or even isolable, singlet carbenes under ordinary conditions.13 h15 Regarding the halogenated carbenes, they are mainly tran- sient species and could only be detected in situ by trapping agents or spectrometric methods. They have been the subject of several theoretical studies [cf. ref. 16È24 and references therein] that agree with each other, pointing toward a singlet ground state for either mono- or dihalogenated derivatives, even for chlorophenylcarbene in which a competition between two opposite e†ects ends up favouring the closed-shell singlet multiplicity.25 Nevertheless these studies disagreed with each other as to the absolute values of the corresponding tripletÈ singlet energy gaps Two recent experimental (*E ThS ). studies26,27 dealing with thermochemical parameters of halo- carbenes have in particular attracted our attention. On the one hand, Lineberger and coworkers26 evaluated the of *E ThS four dihalocarbenes, namely with X \ F, Cl, Br and I, CX 2 from negative ion photoelectron spectroscopy. On the other hand, Nibbering and coworkers27 studied the standard heats of formation and obtained the values of 226 ^ 25 kJ (* f H) mol~1, for and around 267 and 324 kJ mol~1 for CCl 2 , CClBr and respectively. The latter can be compared CBr 2 , with the corresponding calculated G2 values of 227, 282 and 336 kJ mol derived by Schwartz and Marshall.23 It appears that although the di†erences between theoretical and experi- mental remain substantial, amounting to 15 kJ mol~1, * f HÏs they lie rather within the expected error bars from both sides. In contrast, some remarkable discrepancies between theory and experiment on the values for dihalocarbenes have *E ThS emerged. Table 1 which lists the new experimental results along with the most recent theoretical estimates for the entire series of halocarbenes clearly points out these discrepancies. For CF 2 , the newest experimental of 226 ^ 12 kJ mol~1 can be *E ThS compared with that of either 234 or 247 kJ mol~1 computed at the QCISD(T)/6-311]G(3df,2p) and G2 levels, respec- tively.23 To avoid repeating the reported data, we would refer in particular to ref. 23 and 26 for a more detailed comparison of numerous earlier experimental and theoretical results on the tripletÈsinglet separations of various halocarbenes. While an agreement on the energy gap seems to be reached for the disagreement between available theoretical and CF 2 , experimental values on the splittings of other dihalocarbenes, and is intriguing. In addition, Lineberger CCl 2 , CBr 2 CI 2 , and coworkers26 suggested a qualitative change for the di- DOI : 10.1039/b005208f Phys. Chem. Chem. Phys., 2000, 2, 5041È5045 5041 This journal is The Owner Societies 2000 (