Unless otherwise noted, all materials were used without further purification and were purchased from commercial sources. Deionized water was obtained from a Millipore Direct Q3 system (Millipore, Milford, MA, United States). ¹H and ¹³C nuclear magnetic resonance (NMR) spectra were recorded on a Bruker 400 MHz Avance III HD Nanobay spectrometer. Proton chemical shifts are expressed in parts per million (ppm, δ scale) and are referenced to residual protium in deuterated chloroform (CDCl₃, δ = 7.26 ppm). Carbon chemical shifts are expressed in parts per million (ppm, δ scale) and are referenced to the carbon resonance of CDCl₃. Fourier transform infrared (FTIR) spectra were recorded on a Perkin-Elmer Spectrum-100 FTIR spectrophotometer. Thermal gravimetric analysis (TGA) measurements were obtained on a Perkin Elmer STA 6000 thermal analyser at a rate of 10°C/min under a N₂ flow rate of 20 ml/min. Room temperature ultraviolet-visible (UV-Vis) spectra were recorded on a Thermo Electron Corporation Nicolet Evolution 100. Transmission Electron Microscopy (TEM) images were obtained for the synthesized polymers using a FEI Tecnai T20 TEM while Scanning Electron Microscopy (SEM) images were acquired on a Tescan MIRA3 RISE SEM. Elemental composition data were collected on a FEI NovaNano SEM and photoluminescence spectral traces were obtained from a Horiba Jobin Yvon NanoLog. CV and EIS studies were conducted on a CHI700E-potentiostat where measurements were performed using indium tin oxide (ITO) glass substrate (14–16 Ω/square resistance) as working electrode, platinum wire as counter electrode and silver/silver chloride (Ag/AgCl) as reference electrode in 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF₆)/acetonitrile as electrolyte. The ITO glass substrate was washed with 1% (by volume) Hellmanex III solution, acetone and 2-propylalcohol prior to use. Device characteristics [current density–voltage (J-V)] were measured with an X200 Source Meter Unit from Sciencetech Inc. under an illumination of AM 1.5 G, with light intensity of 100 mW cm⁻².

OPVs fabrication and measurements were done under normal conditions. ITO glass substrates, after washing, were spin-coated with zinc oxide (ZnO) and annealed for 30 min at 100°C. ZnO was prepared using sol-gel method (Ramoroka et al., 2021). In details, ZnO solution was prepared using zinc acetate dihydrate as a precursor, 2-methoxyethanol as a solvent and monoethanolamine as a stabilizing agent. The precursor solution with a concentration of 0.5 M in methoxyethanol was prepared and was used to prepare 0.5 M monoethanolamine solution. The resulting solution was stirred at 60°C for 2 h and after aged for 24 h. The deposition was carried out using spin coating method. ZnO solution was spin coated at 3000 rpm for 60 s and dried at 100°C. The coating processes was repeated two times to increase the thickness of ZnO. After coating, ZnO films were annealed at 300°C for 1 h. The active layer was prepared by dissolving HTB-co-P3HT and PC₇₁BM in chlorobenzene for a total concentration of 25 mg/ml and the solution was stirred overnight at 60°C. The active layer was then spin-coated onto the ZnO/ITO glass substrate and annealed at 100°C for 5 min. Subsequently, a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) layer (0.5–1.0 wt.% in water) was spin-coated and annealed at 150°C for 5 min. Finally, silver conductive paste was used to coat the cathode with drying at 100°C for 5 min to complete the OPVs. The active areas of OPVs are 0.0256 cm².

The structure of HTB was confirmed by NMR spectroscopy. Figures 1A,B show the ¹H-NMR and ¹³C-NMR spectra of HTB, respectively. The integrals on the ¹H-NMR spectrum indicate that a total of 18 protons were detected which corresponds to the number of protons on the HTB structure. The signals at 6.94 ppm and 7.14 ppm correspond to protons at the 4- and 3-positions of the thiophene moiety while the α-position is represented at 7.10 ppm. The ¹³C-NMR spectrum of HTB showed the expected five signals at different chemical shifts. Both the ¹H and ¹³C chemical shifts are in agreement with that reported in the literature (Leitner et al., 2020; Yoshida et al., 2008a; Yoshida et al., 2008b). The ¹H-NMR spectrum of HTB-co-P3HT (Figure 1C) is very different in appearance from that of HTB. The signal at 0.88 ppm is assigned to the CH₃- of the hexyl group, the multiplets at 1.28–1.56 ppm and 2.57–2.80 ppm are attributed to the methylene protons from the hexyl group while the thiophene groups are signaled at 7.01 ppm (Mota et al., 2021; Yadav et al., 2021).

The synthesized HTB and HTB-co-P3HT were characterized by FTIR (Figure 2). The bands at 3,104 and 2,919 cm⁻¹ in the spectrum of HTB are attributed to C–H stretching while the band at 816 cm⁻¹ is ascribed to the in-plane bending vibration of the aromatic ring. The band at 1,094 cm⁻¹ results from C–S vibrations. Additionally, the band at 694 cm⁻¹ is due to the in-plane bend of the C–S–C of thiophene. The remaining bands at 1,435 and 1,301 cm⁻¹ are assigned to the C = C and C–C of the thiophene ring, respectively (Agrawal et al., 2013). After formation of HTB-co-P3HT, strong bands at 2,926 and 2,855 cm⁻¹ appear owing to the vibration of the C–H bond of the hexyl group (Tamanai et al., 2013). The C = C stretching vibration shifts to 1,461 cm⁻¹ (from 1,435 cm⁻¹) subsequent to the attachment of P3HT at the α-position of thiophene (Naureen et al., 2021) and indicates the successful attachment of P3HT on HTB.

Figure 3 shows a TGA thermogram of HTB-co-P3HT obtained from the co-polymerization of HTB and 3HT. The decomposition of HTB-co-P3HT takes place in three stages. No weight loss was observed at a temperature range between 50°C and 130°C indicating the absence of adsorbed water on the surface of the sample (Joseyphus and Nair, 2010). In the second stage, weight loss of approximately 8% occurred in the 138°C–312°C temperature range and could be attributed to the decomposition of small molecular weight HTB-co-P3HT and newly formed HTB. Weight loss of 25% in the third stage (temperature range of 379°C–455°C) was observed and represents the decomposition of HTB-co-P3HT. Onset and maximum degradation temperature were found to be 150°C and 460°C, respectively. The remaining weight can be rationalized as solid residues (Khalaji and Das, 2015; Warad et al., 2020).

The UV-Vis spectra of HTB and HTB-co-P3HT were recorded in different solvents at room temperature as depicted in Figure 4. One band was observed in the absorption spectra of HTB-co-P3HT while two bands appeared for HTB. The first band in the spectrum of HTB was assigned to the π→π* transitions and the second band results from the n→π* transitions (Leitner et al., 2020; Ramoroka et al., 2021). The conjugated P3HT chains on the surface of HTB are responsible for the single band in the spectrum of HTB-co-P3HT (Fei et al., 2015; Ramoroka et al., 2021). The absorption responses of the two compounds are collected in Table 1. The parameter dissimilarities are a consequence of the polarity differences of the solvents. The maximum wavelength and lowest band gap values were obtained in toluene. Because toluene stabilizes excited state by interacting with it resulting to a shift to longer wavelength (Sancho et al., 2011). Optical band gaps for HTB were determined to be 3.55, 3.54 and 3.56 eV, while for HTB-co-P3HT they were quantified as 2.38, 2.34 and 2.36 eV in chlorobenzene, toluene and chloroform, respectively. Optical band gaps of 2.15, 2.12 and 2.13 eV for thin films of HTB-co-P3HT prepared in chlorobenzene, toluene and chloroform, respectively were also recorded. These results indicate a bathochromic shift when chlorine atoms are absent on the solvent molecules. Since J _SC is inversely proportional to optical band gap, the highest J _SC can be achieved in toluene (Jeong et al., 2021) and heightened generation of electron-hole pairs in HTB-co-P3HT will preferentially also occur in toluene (Fan et al., 2021).

The morphologies of HTB-co-P3HT in different solvents were characterized using TEM (Figure 5). The HTP-co-P3HT particles are indicated by the high-contrast dark regions. It was expected that chloroform, with its low boiling point and lowest surface tension (Birnie, 2013; Zhao et al., 2015), would show the most agglomeration as the HTB-co-P3HT particles would not have enough time to settle (fast evaporation) and would therefore lead to high reabsorption rates and decreasing V _OC (Liu et al., 2020). However, the same effect was observed for the co-polymer in toluene and to a lesser extent in chlorobenzene. Consequently, it is difficult to observe the effect of solvent on the morphology of HTB-co-P3HT owing to the similar agglomeration of the polymer particles.

Figure 6 shows the SEM images of HTB-co-P3HT films prepared in chlorobenzene (A), toluene (B) and chloroform (C) by drop coating onto ITO coated glass substrates. The SEM images revealed that the surface of HTB-co-P3HT is smooth, and that the morphology changes trend the changes in solvent. In one instance (chlorobenzene) there was some agglomeration of HTB-co-P3HT particles while in another (chloroform and toluene), no agglomeration was observed. The EDS spectra of HTB-co-P3HT (Figure 7) revealed the presence of carbon (C) and sulfur (S), arising from the synthesized HTB-co-P3HT polymer with little differentiation associated with the change in solvents. What was deemed to be an impurity chlorine (Cl) can be attributed to the oxidant, FeCl₃, employed during the polymerization process.

Photoluminescence studies of HTB and HTB-co-P3HT in different solvents (Figure 8) present excitation spectra that are blue-shifted compared with the analogous emission spectra, suggestive of the loss of energy by the photogenerated excitons via either non-radiative or radiative processes before recombination (Chang et al., 2013). The emission spectra of HTB exhibited peaks at 365, 364 and 366 nm while excitation spectra of HTB exhibited peaks at 329, 331 and 331 nm resulting in Stokes shifts of 36, 33 and 35 nm in chlorobenzene, toluene and chloroform, respectively. Larger Stokes shift for HTB-co-P3HT were determined to be 130, 133 and 131 nm in chlorobenzene, toluene and chloroform, respectively, which can result from efficient energy transfer to lower energy levels and planarity enhancement of the P3HT polymer chain in HTB-co-P3HT (Hu et al., 2015). Meanwhile, in both cases, the relatively non-polar toluene recorded the smallest Stokes shift values as expected (Hu et al., 2015). The intense emission peaks of HTB-co-P3HT suggest high recombination of holes and electrons (Liu et al., 2019). There being little difference between the band positions at 545 nm in chlorobenzene, 546 nm in toluene and 543 nm in chloroform indicates weak solvatochromism and suggests that the ground state energy levels of HTB-co-P3HT are independent of the solvent polarity (Ekbote et al., 2017).

Photoexcited charge transfer in the heterojunction between HTB-co-P3HT and PC₇₁BM was investigated using photoluminescence. Emission spectra of HTB-co-P3HT and PC₇₁BM blends are shown in Figure 8. The photoluminescence intensity of HTB-co-P3HT:PC₇₁BM is lower than that of HTB-co-P3HT, indicating an increase in charge dissociation of photoexcited excitons and reduction of recombination activities in the former. The re-emergence of photoluminescence intensity of HTB-co-P3HT, quenched after blending with PC₇₁BM, are 89, 80 and 93% in chlorobenzene, toluene and chloroform, respectively. These results indicate that the greatest charge separation takes place in chloroform, the most polar of the solvents.

The effect of solvent used in the preparation of the HTB-co-P3HT thin films was studied using CV. Figure 9A shows the cyclic voltammograms of HTB-co-P3HT thin films prepared using chloroform, toluene and chlorobenzene. The HOMO and LUMO energy levels were determined and compared with the energy levels of P3HT and PC₇₁BM and the data are collected in Table 2. The difference in the HOMO and LUMO energy levels (E_g ^CV) of HTB-co-P3HT in toluene are less than those in the chlorinated solvents and results in a narrow electrochemical band gap. These observations can be attributed to the polarity and the boiling point characteristics of the solvents as increased polarity and decreased boiling points tend to result in augmented aggregation and energy band gaps (Niefind et al., 2019; Salaiman et al., 2019). The differences in polarity of the chloroform, toluene and chlorobenzene solvents might influence their interaction with the aromatic backbone and/or side groups of the polymer chain materialising as varying degrees of aggregation (Rahman et al., 2009; El-Bashir, 2019). This agglomeration factor may have contributed to the higher LUMO energy levels and wider electrochemical band gaps in chloroform and chlorobenzene. Further, the LUMO energy levels of HTB-co-P3HT in toluene, chloroform and chlorobenzene are −0.61, −0.73 and −0.83 eV above the LUMO energy levels of PC₇₁BM, respectively. These LUMO offset (ΔE_LUMO) values are significantly smaller in toluene than in chlorobenzene and chloroform which indicates that charge transfer from the LUMO of HTB-co-P3HT in toluene can occur more efficiently than if chlorobenzene and chloroform were used. However, high energy loss will still occur thereby lowering V _OC as ΔE_LUMO values are much higher than the threshold value of 0.3 eV (Gautam et al., 2017).

Electrochemical impedance spectra obtained for HTB and HTB-co-P3HT in different solvents are depicted in Figure 10. Nyquist plots were used to study charge carrier migration in the electrochemical cells. It is well-known that the charge transfer resistance (R_ct) is related to the diameter of the semicircle at the interface of an electrode and sample and that the smaller the diameter of the semicircle, the faster the diffusion of charges through the sample to the electroactive surface of the electrode. The Nyquist plot (Figure 10A) showed two semicircles, where the one at higher frequency corresponds to the resistance of the electrode whereas the other identifies with the ionic diffusion resistance of the electrolyte, charge transfer resistance or the resistance of the sample/electrolyte interphase layer (Mei et al., 2018). Both semicircles from the Nyquist spectra were fitted and the results for the lower frequency trace are collected in Table 3. This study revealed that HTB and HTB-co-P3HT each display different electrochemical behaviour depending on the solvent. The smallest R_ct of 939.70 and 109.70 Ω were obtained for HTB and HTB-co-P3HT in chlorobenzene, respectively. This result indicates that the most efficient electron-hole pair separation and the fastest electron transfer at the sample/electrode interface occurs in chlorobenzene (Huang et al., 2013). The very noticeable reduction in the R_ct of HTB-co-P3HT when compared to HTB in chlorobenzene reveals the increased electron transfer between the electrode and electrolyte as a result of the presence of P3HT chains on the surface of HTB. It is noteworthy that the frequency of the maximum peak shifts to the left after modification of HTB with P3HT in both toluene and chloroform but remains unshifted in chlorobenzene (Figure 10B). The frequency shifts in chloroform and toluene indicate reduction of the recombination rate (Pei et al., 2016). Estimations of the electron lifetimes for HTB and HTB-co-P3HT revealed the highest values of 1.91 and 2.30 ms in chloroform. Conclusively, the most reduced rate of recombination and longer electron lifetimes were obtained in chloroform. Figure 10C shows the impedance vs frequency plot of HTB and HTB-co-P3HT prepared in different solvents. It was observed that HTB-co-P3HT has subjacent impedance at lower frequencies than HTB denoting faster electron conduction in HTB-co-P3HT, an observation consistent with the lower R_ct obtained for HTB-co-P3HT.

Figure 11 shows the current density-voltage (J-V) curves of HTB-co-P3HT, as a donor material, in different solvents and the key photovoltaic parameters for OPVs fabricated in ambient atmosphere are presented in Table 4. The maximum PCE was obtained in chlorobenzene, followed by toluene, then chloroform. This winning performance in chlorobenzene is due to fast electron transfer as confirmed by the prior detailed EIS studies. However, the small FF and V _OC values were realized for chloroform and toluene-prepared HTB-co-P3HT signifying the highest hole and electron charge recombination. This is supported by the higher R_ct values obtained from EIS. Photoluminescence revealed that sufficient charges are separated at the interface of HTB-co-P3HT and PC₇₁BM film prepared in chloroform. But they turn to recombine because of poor conductivity of chloroform-prepared HTB-co-P3HT:PC₇₁BM film and agglomeration of HTB-co-P3HT observed on TEM images resulting in lower V _OC and PCE. As all OPVs were fabricated using silver paste as the top electrode and in the ambient air, some degree of degradation may have occurred. The architecture of the OPVs fabricated in this study was as follows: glass|ITO|ZnO|HTB-co-P3HT:PC₇₁BM|PEDOT:PSS|Ag paste. It has been revealed that spin-coating PEDOT:PSS results in the presence of pinholes (Ramoroka et al., 2021) and that direct contact between Ag paste and HTB-co-P3HT:PC₇₁BM led to diffusion of Ag paste into HTB-co-P3HT:PC₇₁BM active layer resulting in an increase in charge recombination (Kim et al., 2015). In addition, since our devices were fabricated in air, oxidized sulfur species may have formed in the active layer (Kettle et al., 2016) and consequently may act as charge carrier trapping sites (Rafique et al., 2020). Series resistance (R_S) and shunt resistance (R_Sh) were determined from the inverse of slopes near V _OC and J _SC , respectively (Gholizadeh, et al., 2017; Diantoro et al., 2018). The R_S and R_Sh values are given in Table 4. High values of R_S estimated lead to a decrease in I _SC and maximum voltage, while low values of R_Sh lead to a decrease in V _OC and maximum current. Therefore, R_S and R_Sh are responsible for reduction of the FF (Wang et al., 2012).