a. Ultraviolet–visible–near-infrared (UV–vis–NIR) spectra of pristine, F4TCNQ-doped, and anion-exchange-doped (with Li-TFSI) donor–acceptor copolymer thin films based on PDPP-2T-TT-OD, which has a deep HOMO level (−5.5 eV). b, Molecular structure of PDPP-2T-TT-OD. c, Energy-level alignment diagram for PDPP-2T-TT-OD, F6TCNNQ and F4TCNQ, along with the molecular structure of F6TCNNQ. Because of the deep HOMO level of PDPP-2T-TT-OD, charge transfer is not expected following conventional molecular doping with F4TCNQ. Anion-exchange doping with Li-TFSI results in bleaching of the neutral peak and the appearance of PDPP-2T-TT-OD polaron peaks, using F4TCNQ as the initiator dopant. The doping level obtained from anion-exchange doping with Li-TFSI is high compared with that reported for F6TCNNQ doping25, as determined from the intensity ratio of the neutral (815 nm) and polaron peaks (1,400 nm) (the polaron/neutral ratio is about 0.1 for F6TCNNQ molecular doping and about 0.5 for anion-exchange doping with Li-TFSI) d, UV–vis–NIR spectra of TCNQ-doped and anion-exchange-doped (with BMIM-TFSI or Li-TFSI) PBTTT thin films. A bleaching of neutral absorbance was observed only with Li-TFSI. e, Energy-level alignment diagram for PBTTT and TCNQ, along with the molecular structure of TCNQ. The LUMO level of TCNQ is too shallow (−4.5 eV) to produce ground-state charge transfer. Even so, introducing Li-TFSI (that is, anion-exchange doping) promotes efficient doping. This presumably occurs because a slight overlap of tail states between HOMO and LUMO levels could initiate charge transfer between PBTTT and TCNQ, and therefore TCNQ•− is exchanged to TFSI−. Overall, control experiments show that the initiator acceptor is not necessarily a powerful acceptor, and that efficient molecular doping is driven by anion exchange.