The as-produced isomer mixture of the organic photovoltaic acceptor material bisPCBM has recently been purified into 19 components by peak-recycling HPLC and their frontier orbitals have been characterised by UV-Vis absorption spectroscopy and cyclic voltammetry. OPV devices based on the purified isomers may now be made, but before meaningful conclusions can be made about why some isomers are better that others in these devices, we mush first know which particular isomer is used in which OPV device. Here we present the structural identification of more than half (11) of the 19 isomers by ¹³C NMR and UV-Vis spectroscopy.

Via UV-Vis spectroscopy the 19 isomers of bisPCBM were classified into 7 different groups, with each group containing 2 or 3 members. Each of the 7 groups corresponds to 1 of the 7 bond types identified by Hirsch (trans-1, trans-2, trans-3, trans-4, e, cis-3, cis-2). By examining the molecular structure of each of the 19 isomers, we conclude that 11 isomers have C₁ symmetry, 3 isomers have Cs symmetry, 3 have C₂ symmetry, 1 has C_2h symmetry and 1 has C_2v symmetry; and we have confirmed this prediction experimentally by ¹³C NMR spectroscopy. The two high-symmetry isomers, C_2h and C_2v, which correspond to the two polar isomers (where the two addends bridge the 1,9 and 52,60 bond of C₆₀with the two addends having opposing relative orientations) correspond to the first and third HPLC fractions by retention time, respectively.

Unfortunately, the other isomers cannot be identified from the NMR spectra alone. However, as the UV-Vis spectra and 7 groups identified above reduce the uncertainty to no more than 3 isomers, when coupled with the symmetry identified from NMR spectra, some of the isomers may be confidently identified. For example, the 13^th HPLC fraction belongs to the cis-2 group which contains one C_s and two C₁ isomers. The NMR spectra shows that this 13^th fraction is the C_s isomer, which involves the 32,33 bond of C₆₀. By this method, the other C_s and C₂isomers were also unambiguously confirmed.

The remaining unidentified isomers all have C₁ symmetry and cannot be further identified by this method. For these isomers, the HOMOs and HOMO-LUMO gaps were calculated by DFT and TD-DFT methods. By comparing the experimental and calculated results, some, but not all, the C₁ isomers were tentatively identified with reasonable confidence. For the remaining 8 C1 isomers the DFT calculations were not able to adequately distinguish, so the identification of these isomers will have to wait for single crystal analysis. A full description of the identification of the isomers will be presented in the meeting.