Omission in a comparative study of the structure and properties of ClF₂ and Cl₃ radicals by CNDO/2 and INDO methods

Abstract

In a very recent note published in this Journal [I], De and Sannigrahi have performed CNDO/2 and INDO calculations on the ClFz and c 1 3 radicals in order to see whether predictions from such calculations tally with ab initio [2] and experimental findings [3,4]. Furthermore, the authors stated that "the proposed study would also throw some light on the relative performance of the CND0/2 and I N D O methods in the calculation of ground state properties of triatomic interhalogen, in general" [I]. We deem it appropriate to point out that an omission has been made by De and Sannigrahi because seven years ago we had published a paper about the geometry of 21 valence-electron radicals [5]. There, the CNDO/2 method was applied to c 1 3 and ClFl in order to calculate equilibrium geometries. We found that the results of those calculations were in reasonable agreement with existing experimental data, and moreover the predictions of several well known qualitative models for geometry correlation were also discussed.A perusal between both sets of results for the equilibrium geometry of the C13 radical calculated via the CND0/2 method shows the existence of some differences, which we consider appropriate to analyze briefly. In both cases the original scheme of parametrization was used [6]. We made the geometry optimization in two steps [S]: (a) We minimized the energy with respect to the bond length in a linear configuration. The calculated bond lengths were 2.08 8, for CI-CI in C13 and 1 .SO 8, for CI-F in CIF2, which are in agreement with those estimated from covalent radii [7]. (b) With these equilibrium distances, the total energy was then minimized with regard to the bond angle. Clearly, our geometry optimization procedure seems to be rather crude nowadays.De and Sannigrahi do not describe the procedure of minimization as we did, so that perhaps a probable difference in it could be responsible for the different calculated equilibrium geometries. Besides, we lowered the precision to a.u. in our calculation for C13, as explicitly noted in Table 1 of Ref. 5, to achieve convergence in the iterative process. Then, it seems reasonable to suppose that this change too can account for the different calculated equilibrium geometries.A point that deserves special mention is that we carried out our CNDO/2 calculations with and without including d functions in the basis set, while De and Sannigrahi used only s and p functions. The inclusion of d functions could explain the deviations between the calculated equilibrium distances in C13.We think that the contradiction between the calculated angles for C13 when using an identical basis set is due to the previously aforesaid reasons: (i) different techniques of geometry optimization', and (ii) lowering of the precision to 1 0-4 a.u.

Photolysis of Matrix-Isolated HCl−Cl2 Complexes: Electronic Absorption and Emission Spectra Provisionally Assigned to Cl3

“…Following this work, Wright et al suggested that a 370−420 nm emission band, observed when Cl 2 was photodissociated under relatively high-pressure conditions, might also originate from Cl 3 . In parallel with the experimental efforts there have been many attempts to calculate potential energy surfaces for trihalogens using ab-initio ,− and semiempirical methods. ,− …”

“…In parallel with the experimental efforts there have been many attempts to calculate potential energy surfaces for trihalogens using ab-initio 15,[24][25][26] and semiempirical methods. 9,[27][28][29][30][31][32] Recently, additional interest in the trihalogens has been generated by novel time-resolved studies of the Br + I 2 reaction. 1,[33][34][35][36] From the results on this system, it appears that trihalogens, accessed through photodissociation of HX-YZ complexes, may be especially suitable for real-time studies of reaction dynamics.…”

“…The hydrogen atom readily escapes the matrix cage, [37][38][39] while the halogen atoms remain trapped. [40][41][42] To test this method, we have examined Cl 3 as the search for the spectrum could be guided by theoretical calculations 15,[26][27][28][29][30][31][32] and the gas phase results. 14,15…”