A 2,2-dimethyl-2H-benzimidazole (22MBI) pulling unit has been synthesized as a potential high-solubility substitute for benzothiadiazole and incorporated into a push−pull-type copolymer used for decent-efficiency (∼3%) organic photovoltaic devices. We herein replace the two methyl side groups of 22MBI by longer alkyl (ethyl, butyl, and hexyl) side chains to further improve the solubility. However, the copolymers replaced by the new pulling units, 2,2-diethyl/dibutyl/dihexl-2H-benzimidazole (22EBI/22BBI/22HBI), lose favorable optical characteristics and exhibit negligible (<0.5%) power conversion efficiency. Intrigued by this anomalous side-group effect of 2,2-dialkyl-2H-benzimidazole (22BI), we carry out time-dependent density functional theory calculations on a series of 22BI-based copolymers with various lengths of 2,2-dialkyl side chains (methyl, ethyl, butyl, and hexyl), but no discernible difference in equilibrium structure nor in electronic structure is found between them. We hence formulate a hypothesis that 22BI may either isomerize into 1,2-dialkylbenzimidazole (12BI) or lose one of its alkyl (≥ethyl) chains as olefin to become 2-dialkylbenzimidazole (2BI) because these aromatic products, unlike the quinoidtype 22BI, would exhibit unfavorable electronic structure for organic photovoltaic (OPV) applications. Indeed, the absorption spectra measured for the 22BI-based copolymers with long dialkyl side chains are best reproduced by the calculations on 2BI-and 12BI-based copolymers, and the two side products are calculated to be more stable than 22BI, indicating the spontaneity of the proposed reactions. The activation barriers are prohibitively high (>29 kcal/mol) but could be reduced down to 8 kcal/mol in the presence of palladium-based polymerization catalysts. Indeed, the presence of the predicted 12BI-containing side products is confirmed by NMR spectra. A temperature-dependent polymerization experiment shows that 22MBI is in fact subject to the same type of isomerization when the temperature is raised to 150°C above the original polymerization temperature (90−110°C ), further supporting the hypothesis from our calculations and explaining the observed anomalous side-group effect.