“…We used this semiempirical method instead of a more accurate ab initio method to compute the geometries for several reasons: first, the sizes of some of the systems studied prevented the use of ab initio molecular quantum-mechanical methods to perform optimizations, second, we have found by comparison with our previous results for C 6 H 6 , C 10 H 8 , and C 14 H 8 [73] that differences in HOMA, NICS, and PDI indices computed at the HF/6-31G* (HOMA and PDI) and HF/6-31 G* (NICS) levels by use either of the B3LYP/6-31G* or of the AM1 geometries are generally small, and third, the AM1 method yields reliable results for the geometries of C 60 , [87,88] C 70 , [88] and also for the series of PAHs studied. We used this semiempirical method instead of a more accurate ab initio method to compute the geometries for several reasons: first, the sizes of some of the systems studied prevented the use of ab initio molecular quantum-mechanical methods to perform optimizations, second, we have found by comparison with our previous results for C 6 H 6 , C 10 H 8 , and C 14 H 8 [73] that differences in HOMA, NICS, and PDI indices computed at the HF/6-31G* (HOMA and PDI) and HF/6-31 G* (NICS) levels by use either of the B3LYP/6-31G* or of the AM1 geometries are generally small, and third, the AM1 method yields reliable results for the geometries of C 60 , [87,88] C 70 , [88] and also for the series of PAHs studied.…”