Copolymers Containing 1-Methyl-2-phenyl-imidazole Moieties as Permanent Dipole Generating Units: Synthesis, Spectroscopic, Electrochemical, and Photovoltaic Properties

New donor–acceptor conjugated alternating or random copolymers containing 1-methyl-2-phenylbenzimidazole and benzothiadiazole (P1), diketopyrrolopyrrole (P4), or both acceptors (P2) are reported. The specific feature of these copolymers is the presence of a permanent dipole-bearing moiety (1-methyl-2-phenyl imidazole (MPI)) fused with the 1,4-phenylene ring of the polymer main chain. For comparative reasons, polymers of the same main chain but deprived of the MPI group were prepared, namely, P5 with diketopyrrolopyrrole and P3 with both acceptors. The presence of the permanent dipole results in an increase of the optical band gap from 1.51 eV in P3 to 1.57 eV in P2 and from 1.49 eV in P5 to 1.55 eV in P4. It also has a measurable effect on the ionization potential (IP) and electrochemical band gap (EgCV), leading to their decrease from 5.00 and 1.83 eV in P3 to 4.92 and 1.79 eV in P2 as well as from 5.09 and 1.87 eV in P5 to 4.94 and 1.81 eV in P4. Moreover, the presence of permanent dipole lowers the exciton binding energy (Eb) from 0.32 eV in P3 to 0.22 eV in P2 and from 0.38 eV in P5 to 0.26 eV in P4. These dipole-induced changes in the polymer properties should be beneficial for photovoltaic applications. Bulk heterojunction solar cells fabricated from these polymers (with PC71BM acceptor) show low series resistance (rs), indicating good electrical transport properties. The measured power conversion efficiency (PCE) of 0.54% is limited by the unfavorable morphology of the active layer.


Introduction
In the past two decades, significant research efforts have been directed towards the elaboration of conjugated donor-acceptor copolymers of -(D-A)-or -(D-A-D)-types [1][2][3][4][5]. These organic semiconductors frequently exhibit interesting physical and chemical properties that are unmatched or, at least, difficult to match by other groups of organic electroactive compounds. In the majority of cases, they are low band gap semiconductors exhibiting appropriate ionization potential (IP) and electron affinity (EA) values to assure their ambipolarity. As a result, they can serve as components of active layers in various types of ambipolar organic field-effect transistors, characterized by well-balanced n-type and p-type electrical transport [6][7][8][9][10][11][12][13][14][15]. Low ionization potential (IP) and high electron affinity (EA) of D-A and D-A-D polymers, together with the presence of several chromophore groups, 1. P1, P4, and P5 are alternating copolymers, whereas P2 is a random copolymer of 4,7dithienobenzimidazole, 4,7-dithienobeznothiadiazole, and 3,6-dithienodiketopyrroloprrole. P3 is a random copolymer of 1,4-dithienobenzene, 4,7-dithienobeznothiadiazole, and 3,6-dithienodiketopyrroloprrole. Note that the only difference between P2 and P3 is the presence of the dipole-bearing moiety, i.e., the 1-methyl-2-phenyl imidazole (MPI) group fused with a phenylene ring in the former. The same applies to P4 and P5 containing the diketopyrrolopyrrole (DPP) acceptor. P1 is different because it contains benzothiadiazole (BT) as an acceptor, whereas in the repeat units of P2 and P3, two acceptor groups can be distinguished, namely, benzothiadiazole (BT) and diketopyrrolopyrrole (DPP).

Synthesis
The synthetic pathway leading to P1-P5 is presented in Scheme 2. Detailed procedures of the synthesis of the prepared monomers (M1-M4) and polymers (P1-P5) can be found in the supplementary data. M1 was synthesized in a two-step reaction of 1-methyl-2-phenyl-4,7-dibromobenzimidazole,1 with the thiophene-2-boronic acid pinacol ester (86% yield), followed by lithiation with n-BuLi and subsequent treatment with trimethyltin chloride (98% yield). M2 was obtained in a similar manner with 90% yield. M3 and M4 were synthesized following standard procedures [36]. Polymers were prepared in a microwave reactor via Stille coupling catalyzed with Pd2(dba)3/P(o-tol)3. All macromolecular products were soluble in chloroform and chlorobenzene. Molecular masses of fractions Scheme 1. Chemical structures of the studied polymers.

Synthesis
The synthetic pathway leading to P1-P5 is presented in Scheme 2. Detailed procedures of the synthesis of the prepared monomers (M1-M4) (Scheme S1-Scheme S3) and polymers (P1-P5) can be found in the supplementary data. M1 was synthesized in a two-step reaction of 1-methyl-2-phenyl-4,7-dibromobenzimidazole,1 with the thiophene-2-boronic acid pinacol ester (86% yield), followed by lithiation with n-BuLi and subsequent treatment with trimethyltin chloride (98% yield). M2 was obtained in a similar manner with 90% yield. M3 and M4 were synthesized following standard procedures [36]. Polymers were prepared in a microwave reactor via Stille coupling catalyzed with Pd 2 (dba) 3 /P(o-tol) 3 . All macromolecular products were soluble in chloroform and chlorobenzene. Molecular masses of fractions soluble in dichloromethane were determined by SEC, indicating M n values in the range from ca. 5.5 × 10 3 to ca. 1.05 × 10 4 g/mol and PDI values ranging from 1.5 to 2.0.

Spectroscopic Properties
Donor-acceptor (D-A) polymers containing phenyl imidazole units have been reported in the literature, including those containing methoxystyryl donors [37]. Although interesting from the point of view of their spectroscopic and electrochemical properties, they showed too large an optical band gap (ca. 2.3 eV) and too low electron affinity |EA| (ca. 2.8 eV) to be considered as the components of organic photovoltaic cells. In this paper, we demonstrate that these parameters can be significantly improved in polymers containing phenylimidazole groups by the appropriate selection of the donor and acceptor units in the polymer main chain.
Absorption spectra of the synthesized polymers strongly depend on the nature of the acceptor units present in the macromolecule. The UV-vis spectrum of P1 is shown in Figure 1. The position of its least energetic band at λmax = 530 nm in the case of the solution (chloroform) spectrum and at λmax = 560 nm for the spectrum of a thin solid film cast from chloroform were typical of copolymers consisting of thiophene and benzothiadiazole units [38]. Replacement of benzothiadiazole (BT) in P1 with a significantly stronger acceptor such as diketopyrrolopyrrole (DPP) yields P4. The least energetic bands in the solution and solid-state spectra of this polymer are bathochromically shifted by 120 and 98 nm, respectively, compared to the corresponding bands in the spectra of P1 (see Figure 1 and Table 1). This is not unexpected since stronger donor-acceptor interactions usually result in the significant narrowing of the optical band gap (Eg opt ). Incorporation of an additional acceptor to the repeat unit is usually beneficial for photovoltaic applications because it may result in a significant broadening of the absorption bands, leading to their better overlapping with the solar spectrum [39]. This broadening, together with a significant bathochromic shift of the absorption band, is very pronounced in the case of P2, which contains two acceptor groups in its repeat unit (BT and DPP). It is also clearly observed for P3. P4, and P5, which do not contain the benzothadiazole chromophore in their repeat units. Their absorption bands are narrower and bathochromically shifted with respect to Scheme 2. Synthesis of polymers P1-P5.

Spectroscopic Properties
Donor-acceptor (D-A) polymers containing phenyl imidazole units have been reported in the literature, including those containing methoxystyryl donors [37]. Although interesting from the point of view of their spectroscopic and electrochemical properties, they showed too large an optical band gap (ca. 2.3 eV) and too low electron affinity |EA| (ca. 2.8 eV) to be considered as the components of organic photovoltaic cells. In this paper, we demonstrate that these parameters can be significantly improved in polymers containing phenylimidazole groups by the appropriate selection of the donor and acceptor units in the polymer main chain.
Absorption spectra of the synthesized polymers strongly depend on the nature of the acceptor units present in the macromolecule. The UV-vis spectrum of P1 is shown in Figure 1. The position of its least energetic band at λ max = 530 nm in the case of the solution (chloroform) spectrum and at λ max = 560 nm for the spectrum of a thin solid film cast from chloroform were typical of copolymers consisting of thiophene and benzothiadiazole units [38]. Replacement of benzothiadiazole (BT) in P1 with a significantly stronger acceptor such as diketopyrrolopyrrole (DPP) yields P4. The least energetic bands in the solution and solid-state spectra of this polymer are bathochromically shifted by 120 and 98 nm, respectively, compared to the corresponding bands in the spectra of P1 (see Figure 1 and Table 1). This is not unexpected since stronger donor-acceptor interactions usually result in the significant narrowing of the optical band gap (E g opt ). Incorporation of an additional acceptor to the repeat unit is usually beneficial for photovoltaic applications because it may result in a significant broadening of the absorption bands, leading to their better overlapping with the solar spectrum [39]. This broadening, together with a significant bathochromic shift of the absorption band, is very pronounced in the case of P2, which contains two acceptor groups in its repeat unit (BT and DPP). It is also clearly observed for P3. P4, and P5, which do not contain the benzothadiazole chromophore in their repeat units. Their absorption bands are narrower and bathochromically shifted with respect to the corresponding bands of P2 and P3. The same trend is observed in the solid-state spectra recorded for thin films cast from chloroform (see Figure 1 and Table 1). Moreover, the absorption bands in the solid-state spectra of the reference polymers (P3 and P5) are structured, showing clear maxima at 653 and 720 nm for P3 and a shoulder at 850 nm, accompanied by two clear maxima at 672 and 738 nm in the case of P5. The presence of these bands can be tentatively attributed to the formation of semi-crystalline phases of these polymers, as was suggested by Li et al. [40] for DPP polymers with alkyl side chains.
bands can be tentatively attributed to the formation of semi-crystalline phases of these polymers, as was suggested by Li et al. [40] for DPP polymers with alkyl side chains.
It should be noted that the least energetic bands in the solution and solid-state spectra of the reference polymers (P3 and P5) are bathochromically shifted with respect to the corresponding bands of the permanent dipole-bearing polymers (P2 and P4), indicating better conjugation in the former (see Figure 1 and Table 1). Molar absorption coefficients of P2 and P3 are higher than those determined for P1, P4, and P5. It should, however, be stressed that they contain more chromophore groups in their repeat units. For the same reason, the molar absorption coefficient of P2 is higher than that determined for P3 since in the repeat unit of the former, an extra chromophore (1-methyl-2-phenyl benzimidazole) is present.  Optical band gaps, Eg opt , determined from the absorption onset in the spectra of thin films, are listed in Table 1. As expected, the largest band gap, 1.84 eV, is observed for P1, i.e., the only polymer containing a weaker acceptor (BT) instead of a stronger DPP one. Optical band gaps of P2 and P4 are similar (1.57 and 1.55 eV) and consistently larger by  It should be noted that the least energetic bands in the solution and solid-state spectra of the reference polymers (P3 and P5) are bathochromically shifted with respect to the corresponding bands of the permanent dipole-bearing polymers (P2 and P4), indicating better conjugation in the former (see Figure 1 and Table 1). Molar absorption coefficients of P2 and P3 are higher than those determined for P1, P4, and P5. It should, however, be stressed that they contain more chromophore groups in their repeat units. For the same reason, the molar absorption coefficient of P2 is higher than that determined for P3 since in the repeat unit of the former, an extra chromophore (1-methyl-2-phenyl benzimidazole) is present.
Optical band gaps, E g opt , determined from the absorption onset in the spectra of thin films, are listed in Table 1. As expected, the largest band gap, 1.84 eV, is observed for P1, i.e., the only polymer containing a weaker acceptor (BT) instead of a stronger DPP one. Optical band gaps of P2 and P4 are similar (1.57 and 1.55 eV) and consistently larger by 60 meV compared to the gaps of the reference polymers, i.e., P3 and P5. The E g opt of P5 is slightly lower than the band gap of a very similar polymer reported by Li et al. [41], which differs from P5 by shorter alkyl substituents attached to the DPP group.
Thin films of P1 show weak photoluminescence with a maximum at ca. 720 nm (1.72 eV). P2 and P3, as well as P4 and P5, emit at the same λ max. of 942 nm (1.32 eV) and Molecules 2022, 27, 915 6 of 15 998 nm (1.24 eV), respectively. In all cases, the Stokes shifts are rather large, exceeding 150 nm. Note that the emission bands of the polymers bearing the permanent dipole (P2 and P4) are broadened on their higher energy side (see Figure 2). 60 meV compared to the gaps of the reference polymers, i.e., P3 and P5. The Eg opt o slightly lower than the band gap of a very similar polymer reported by Li et al. [41], differs from P5 by shorter alkyl substituents attached to the DPP group.
Thin films of P1 show weak photoluminescence with a maximum at ca. 720 nm eV). P2 and P3, as well as P4 and P5, emit at the same λmax. of 942 nm (1.32 eV) and 9 (1.24 eV), respectively. In all cases, the Stokes shifts are rather large, exceeding 15 Note that the emission bands of the polymers bearing the permanent dipole (P2 an are broadened on their higher energy side (see Figure 2).

Electrochemical Properties
Redox properties of conjugated polymers are of crucial importance for their e chemical and electronic applications. For these reasons, cyclic voltammetry (CV) an ferential pulse voltammetry (DPV) are routinely used for the characterization of new jugated polymers and for the determination of their ionization potential (IP) and el affinity (EA). In Figure 3, cyclic voltammograms of P1-P5 are presented, whereas t termined redox potentials are collected in Table 2.
There are some common features in the electrochemical properties of all studie ymers. They all undergo one-step, partially reversible oxidation at relatively low tials. The reduction of P1-P4 is also partially reversible, P5 being the only exception reduction is purely irreversible. P1 undergoes oxidation at the same potential as P its reduction potential is lower by 120 mV compared to that of P4. Since both pol differ by the acceptor unit nature only, the more difficult reduction of P1 reflec weaker electroaccepting properties of benzothiadiazole compared to d pyrrolopyrrole. P2 and P3 contain both (BT and DPP) acceptor units and, for this r undergo a two-step reduction. By comparison with the data obtained for P1, P4, a it can be concluded that the cathodic peak at higher potentials corresponds to the tion of DPP, whereas that at lower potentials corresponds to the reduction of BT presence of the permanent dipole in P2 and P4 meaningfully affects their redox prop decreasing their oxidation potential compared to P3 and P5 and increasing the red potential in the case of P4 (see Table 2).

Electrochemical Properties
Redox properties of conjugated polymers are of crucial importance for their electrochemical and electronic applications. For these reasons, cyclic voltammetry (CV) and differential pulse voltammetry (DPV) are routinely used for the characterization of new conjugated polymers and for the determination of their ionization potential (IP) and electron affinity (EA). In Figure 3, cyclic voltammograms of P1-P5 are presented, whereas the determined redox potentials are collected in Table 2.   There are some common features in the electrochemical properties of all studied polymers. They all undergo one-step, partially reversible oxidation at relatively low potentials. The reduction of P1-P4 is also partially reversible, P5 being the only exception whose reduction is purely irreversible. P1 undergoes oxidation at the same potential as P4, but its reduction potential is lower by 120 mV compared to that of P4. Since both polymers differ by the acceptor unit nature only, the more difficult reduction of P1 reflects the weaker electroaccepting properties of benzothiadiazole compared to diketopyrrolopyrrole. P2 and P3 contain both (BT and DPP) acceptor units and, for this reason, undergo a two-step reduction. By comparison with the data obtained for P1, P4, and P5, it can be concluded that the cathodic peak at higher potentials corresponds to the reduction of DPP, whereas that at lower potentials corresponds to the reduction of BT. The presence of the permanent dipole in P2 and P4 meaningfully affects their redox properties, decreasing their oxidation potential compared to P3 and P5 and increasing the reduction potential in the case of P4 (see Table 2).
From the oxidation and reduction potentials of the studied polymers, it is possible to calculate their ionization potentials (IPs) and electron affinities (EAs), provided that the first oxidation process corresponds to the oxidation of the repeat unit to a radical cation, whereas the first reduction process leads to its transformation into a radical anion. It should be, however, noted that the electrochemically determined IPs and EAs comprise several contributions that may alter their values. These are, for example, interactions of the ionized macromolecules with their environment, which depend on their chemical nature, polarization, and solvation energy. Thus, an important question arises of whether IP and EA values derived from electrochemical data properly are correlated with those obtained by direct methods, i.e., ultraviolet photoelectron spectroscopy (UPS) and inversed photoelectron spectroscopy (IPES). Sworakowski et al. [42][43][44], based on numerous experimental electrochemical and spectroscopic data published in the literature, found a clear correlation between electrochemical and photoelectron spectroscopy data. The data collected in Table 2 show that the presence of the permanent dipole lowers IP since its values determined for P2 and P4 are lower than the corresponding IP values found for P3 and P5. EA values are less affected by the presence of the permanent dipole but the lowering of their values in P2 and P4 is measurable.
Assuming the validity of Koopmans theorem (E HOMO = −IP, E LUMO = EA), it is possible to calculate the so-called "electrochemical band gap" (E g CV ). These gaps are very similar for P3 and P5, being, however, slightly narrower for P2 and P4. E g CV values are consistently larger than optical E g opt. for all studied polymers (P1-P5). The difference between the electrochemical and the optical band gaps can be treated as an effective singlet exciton-binding energy E b [35]: E b = E g CV − E g opt . Willems et al. estimated E b for different polymers containing DPP units and found that E b is equal to ca. 0.44 eV, independent of the specific polymer [45]. This value was clearly higher than that determined for P5 (0.38 eV, see Table 2), i.e., the polymer in which DPP was the sole acceptor unit. By comparing P2 with P3 and P4 with P5, it becomes clear that the presence of permanent dipole-bearing units in the polymer chain lowers the exciton binding energy by 100-120 mV. Thus, the obtained results clearly indicate that the local dipoles should play a beneficial role in the charge generation associated with photovoltaic phenomena.

Thin Films Morphology
Intuitively, it seems clear that the presence of permanent dipoles in the macromolecules of conjugated polymers should significantly modify their capability of aggregation, and, by consequence, it should lead to different supramolecular organizations and morphologies of the resulting solution cast films. This is an important aspect of conjugated polymers solution processing, especially in view of their photovoltaic applications, since morphology is one of the dominant factors influencing electrical transport properties in organic electronic devices, including photovoltaic cells. Thus, one of the main problems to be discussed in this research is the effect of fusion of the permanent dipole-bearing groups (methyl-2-phenyl imidazole, MPI) with the polymer main chain on the morphology of the resulting polymers. Hence, detailed comparative AFM observations of monolayers of P4 and monolayers of its dipole-free analog, i.e., P5 were performed. Figure 4 shows the AFM images (a-c) and the corresponding line-profile (d) of the reference polymer (P5) layers deposited directly from its solution in chlorobenzene (optimal solvent for the tested group of polymers). A representative image of the larger area (5 µm × 5 µm) is presented in the image in Figure 4a. As a consequence of the polymer concentration gradient, a transition is clearly observed: from a monolayer (in the upper-left corner of the image) through to an incomplete bilayer (dominant in the middle part of the image) to three-layer coverage (bottom-right corner). Subsequent higher resolution images provide more details. The bilayer surface is presented in Figure 4b. As expected, three levels can be easily distinguished in this case: small areas of the uncovered substrate (darker parts), macromolecules in the monolayer (grey areas dominating the image), and locally observed individual molecules or their small aggregates in the second layer (bright areas of elongated shape). Additionally, the corresponding profile of the selected line is presented in Figure 4d, which confirms the same thickness of the first and the second layer (ca. 2 nm). The AFM image of the area where the monolayer strongly dominates is shown in Figure 4c. The above observations demonstrate the tendency of P5 to form films of layered structures, characterized by a small distance between the chains of adjacent molecules in each layer. This is evidenced by clearly visible areas of densely packed polymers in the monolayer as well as in the bilayer. These morphological features clearly indicate that multilayer films of P5, showing a compact structure, can be deposited in these conditions. groups (methyl-2-phenyl imidazole, MPI) with the polymer main chain on the morphology of the resulting polymers. Hence, detailed comparative AFM observations of monolayers of P4 and monolayers of its dipole-free analog, i.e., P5 were performed. Figure 4 shows the AFM images (a-c) and the corresponding line-profile (d) of the reference polymer (P5) layers deposited directly from its solution in chlorobenzene (optimal solvent for the tested group of polymers). A representative image of the larger area (5 µm × 5 µm) is presented in the image in Figure 4a. As a consequence of the polymer concentration gradient, a transition is clearly observed: from a monolayer (in the upperleft corner of the image) through to an incomplete bilayer (dominant in the middle part of the image) to three-layer coverage (bottom-right corner). Subsequent higher resolution images provide more details. The bilayer surface is presented in Figure 4b. As expected, three levels can be easily distinguished in this case: small areas of the uncovered substrate (darker parts), macromolecules in the monolayer (grey areas dominating the image), and locally observed individual molecules or their small aggregates in the second layer (bright areas of elongated shape). Additionally, the corresponding profile of the selected line is presented in Figure 4d, which confirms the same thickness of the first and the second layer (ca. 2 nm). The AFM image of the area where the monolayer strongly dominates is shown in Figure 4c. The above observations demonstrate the tendency of P5 to form films of layered structures, characterized by a small distance between the chains of adjacent molecules in each layer. This is evidenced by clearly visible areas of densely packed polymers in the monolayer as well as in the bilayer. These morphological features clearly indicate that multilayer films of P5, showing a compact structure, can be deposited in these conditions.  As already stated, the main chain of P4 is the same as that of P5. The only difference between these polymers is the presence of the permanent dipole-bearing group (1-methyl-2-phenyl imidazole) fused with the phenylene ring of the main chain in the case of P4. The presence of this dipole results in the distinctly different supramolecular aggregation patterns of this polymer. The first dissimilarity concerns the selection of the solvent for thin layer deposition. In the case of P4, chloroform has to be used for casting. It is, therefore, tempting to compare the topologies of the monolayers of P4 and P5, deposited from chloroform, in the same conditions. Chloroform is more polar than chlorobenzene, and the effect of this higher polarity additionally increases the effect of intermolecular interactions in the case of P5. As a consequence, the monolayer of P5, deposited from this solvent, is characterized by randomly distributed larger areas of very tightly packed macromolecules (Figure 5a). More polar P4, on the other hand, forms, in the same conditions, a homogeneous network of loosely packed macromolecules (Figure 5b,c). The resulting mesh is characterized by much larger distances between adjacent molecules (approx. 110-160 nm).
2-phenyl imidazole) fused with the phenylene ring of the main chain in the case of P4. The presence of this dipole results in the distinctly different supramolecular aggregation patterns of this polymer. The first dissimilarity concerns the selection of the solvent for thin layer deposition. In the case of P4, chloroform has to be used for casting. It is, therefore, tempting to compare the topologies of the monolayers of P4 and P5, deposited from chloroform, in the same conditions. Chloroform is more polar than chlorobenzene, and the effect of this higher polarity additionally increases the effect of intermolecular interactions in the case of P5. As a consequence, the monolayer of P5, deposited from this solvent, is characterized by randomly distributed larger areas of very tightly packed macromolecules (Figure 5a). More polar P4, on the other hand, forms, in the same conditions, a homogeneous network of loosely packed macromolecules (Figure 5b,c). The resulting mesh is characterized by much larger distances between adjacent molecules (approx. 110-160 nm). Summing up this part of the research, significant differences in the supramolecular organization of both polymers should be emphasized. The P4 polymer forms layers with much weaker packing compared to reference polymer P5. This is a direct consequence of electrostatic interaction between the polar MPI groups presented in this macromolecule. This may have a negative impact on the electrical properties of the layers it forms.

Photovoltaic Properties
Before discussing in detail the photovoltaic properties of the newly developed polymers containing 1-methyl-phenyl imidazole, it is instructive to justify their application as donors in organic cells, together with PC71BM acceptor. As already mentioned, by selecting appropriate electron-rich and electron-poor units in the polymer main chain, we could adjust the crucial material parameters of the donor component of the cell. These are ionization potential (IP) and electron affinity (EA). IP values of P1-P5 are in the range of 4.92 to 5.09 eV, whereas the |EA| values vary from 2.99 to 3.22 eV. IP and |EA| of PC71BM are equal to 5.87 and 3.91 eV, respectively [46]. It is generally accepted that in properly operating photovoltaic cells, the difference between the values of |EA| of the donor and |EA| of the acceptor should be superior to 0.3 eV if excitons are generated in the donor phase. By symmetry, the difference between IP of the donor and IP of the acceptor should also be superior to 0.3 eV if excitons are formed in the acceptor phase [47][48][49]. Nanocomposites of P1-P5 with PC71BM excessively fulfill these conditions. Moreover, as shown in Section 2.3, the presence of a permanent dipole in P1, P2, and P4 lowers the exciton-binding energy by ca. 30%, facilitating, in this manner, the generation of free charge carriers. Finally, our investigations of the luminescent properties of blends of PC71BM with the synthesized polymers showed the effective quenching of the luminescence in this two-phase system (see Figure S6 in the supplementary information). Thus, by estimating the above-discussed thermodynamic and spectroscopic properties of P1, P2, and P3, it quickly became Summing up this part of the research, significant differences in the supramolecular organization of both polymers should be emphasized. The P4 polymer forms layers with much weaker packing compared to reference polymer P5. This is a direct consequence of electrostatic interaction between the polar MPI groups presented in this macromolecule. This may have a negative impact on the electrical properties of the layers it forms.

Photovoltaic Properties
Before discussing in detail the photovoltaic properties of the newly developed polymers containing 1-methyl-phenyl imidazole, it is instructive to justify their application as donors in organic cells, together with PC 71 BM acceptor. As already mentioned, by selecting appropriate electron-rich and electron-poor units in the polymer main chain, we could adjust the crucial material parameters of the donor component of the cell. These are ionization potential (IP) and electron affinity (EA). IP values of P1-P5 are in the range of 4.92 to 5.09 eV, whereas the |EA| values vary from 2.99 to 3.22 eV. IP and |EA| of PC 71 BM are equal to 5.87 and 3.91 eV, respectively [46]. It is generally accepted that in properly operating photovoltaic cells, the difference between the values of |EA| of the donor and |EA| of the acceptor should be superior to 0.3 eV if excitons are generated in the donor phase. By symmetry, the difference between IP of the donor and IP of the acceptor should also be superior to 0.3 eV if excitons are formed in the acceptor phase [47][48][49]. Nanocomposites of P1-P5 with PC 71 BM excessively fulfill these conditions. Moreover, as shown in Section 2.3, the presence of a permanent dipole in P1, P2, and P4 lowers the exciton-binding energy by ca. 30%, facilitating, in this manner, the generation of free charge carriers. Finally, our investigations of the luminescent properties of blends of PC 71 BM with the synthesized polymers showed the effective quenching of the luminescence in this two-phase system (see Figure S6 in the supplementary information). Thus, by estimating the above-discussed thermodynamic and spectroscopic properties of P1, P2, and P3, it quickly became clear that they should have been considered promising candidates for application in organic photovoltaic cells. P2 and P4 were selected for these investigations because of their better matching of the solar spectrum.
The photovoltaic properties of P2 and P4 were tested in the bulk heterojunction-type solar cells of the following architecture: ITO/PEDOT:PSS/polymer:PC 71 BM/InAl. The blends were solution-processed from chloroform. Under standard 1.5 AM illumination (100 mW/cm 2 ), the investigated solar cells generated current density in the range of a few mA/cm 2 (see Figure 6). P4 turned out to be a more effective component of the bulk heterojunction since P4-based cells were ca. twice more efficient than the P2-based ones. The absolute value of current density |J| decreased with increasing forward voltage, indicating high leakage. The characteristics were analyzed, assuming the parasitic resistance model (series resistance and shunt resistance). The data of electrical parameters of P2:PC 71 BM and P4:PC 71 BM based solar cells are presented in Table 3. The cells had 8 pixels each; thus, the presented data correspond to average values, and their uncertainties are given.
The photovoltaic properties of P2 and P4 were tested in the bulk heterojunction-type solar cells of the following architecture: ITO/PEDOT:PSS/polymer:PC71BM/InAl. The blends were solution-processed from chloroform. Under standard 1.5 AM illumination (100 mW/cm 2 ), the investigated solar cells generated current density in the range of a few mA/cm 2 (see Figure 6). P4 turned out to be a more effective component of the bulk heterojunction since P4-based cells were ca. twice more efficient than the P2-based ones. The absolute value of current density |J| decreased with increasing forward voltage, indicating high leakage. The characteristics were analyzed, assuming the parasitic resistance model (series resistance and shunt resistance). The data of electrical parameters of P2:PC71BM and P4:PC71BM based solar cells are presented in Table 3. The cells had 8 pixels each; thus, the presented data correspond to average values, and their uncertainties are given.  Voc-open-circuit voltage, Jsc-short circuit current, rsh and rS-shunt and series resistivities, FFfill factor, PCE-power conversion efficiency.
The measured low rS values (3-6 Ω⋅cm 2 ) suggest the good mobility of charge carriers. It should be, however, noted that both devices show rather low values of short circuit current. The photocurrent measured for the blends with P4 is ca. twice higher than that for the blends with P2, but it is ca. three times lower than the value reported by Li et al. [40] for a polymer very similar to P5 (analog of P5 without the permanent dipole) or for PTB7-based (poly [4,8-  V oc -open-circuit voltage, J sc -short circuit current, r sh and r S -shunt and series resistivities, FF-fill factor, PCE-power conversion efficiency. The measured low r S values (3-6 Ω·cm 2 ) suggest the good mobility of charge carriers. It should be, however, noted that both devices show rather low values of short circuit current. The photocurrent measured for the blends with P4 is ca. twice higher than that for the blends with P2, but it is ca. three times lower than the value reported by Li et al. [40] for a polymer very similar to P5 (analog of P5 without the permanent dipole) or for PTB7-based (poly [4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo [1,2- ) of similar architecture [50]. The absorption spectra of P2 and P4, although different, cover a very similar spectral range; thus, the measured difference of the photocurrent cannot be related to the better harvesting of the solar flux and the greater number of excitons formed in the blend with P4 compared to that fabricated with P2. Moreover, the IP and EA values (which correspond to HOMO and LUMO levels) for P2 and P4 are similar. Low values of the measured photocurrent in both cases should, therefore, originate from weak light absorption of the blends and high leakage of the layer (low r sh value). This leakage can be caused by the poor morphology of the layer. Monolayers of neat polymers form a kind of mesh with a large distance (above 100 nm) between adjacent aggregates (see Figure 5), and this tendency is maintained in thicker films and in the blends with fullerene derivatives.
The open-circuit voltage (V oc ) of the fabricated cells was lower than expected, considering the values of the HOMO and LUMO energies, and lower than predicted from the relation between the open-circuit voltage and the oxidation potential, as proposed by Willems et al. [45]. It can be supposed that the presence of the permanent dipoles in the macromolecules of P2 and P4 may provoke energetic disorder. Different local positions of these dipoles may cause a broad distribution of the electronic density of states, which could be more pronounced at the donor-acceptor interfaces. Thus, V oc losses can be related to the distribution of electronic states and the presence of interfacial energetic traps. The distribution of electronic states and energetic traps is dependent on the blend morphology. Thus, the morphological disorder may affect the material properties, leading to low values of the measured electrical parameters.
External quantum efficiency (EQE) spectra were obtained by measuring the photocurrent in the short circuit mode (I SC ). The EQE was of the order of 20%, which means that one out of five photons was successfully transformed into a charge carrier, reaching the cathode. Since the absorbing layers were relatively thin (ca. 100 nm), only half of the photons were absorbed and about half of the photoexcited electrons arrived at the cathode. The cells with P4 were about two times more efficient than those with P2. As indicated in Figure 7, absorption bands of the donor and acceptor components of the bulk heterojunctions (marked by arrows) are located in the visible part of the spectrum (for the shape of the spectra of P2 and P4, see Figure 1; the spectrum of PC 71 BM can be found in [50]). This means that both components can efficiently absorb photons, and the interface charge transfer can proceed in both directions. The absorption efficiencies of P2 and P4 were lower than that determined for PC 71 BM. Different spectral ranges of measurements should, however, be taken into consideration. In the 2 eV range, significantly more solar photons are expected than in the 3 eV range, so the former is more important for the power conversion efficiency (PCE).
The open-circuit voltage (Voc) of the fabricated cells was lower than expected, considering the values of the HOMO and LUMO energies, and lower than predicted from the relation between the open-circuit voltage and the oxidation potential, as proposed by Willems et al. [45]. It can be supposed that the presence of the permanent dipoles in the macromolecules of P2 and P4 may provoke energetic disorder. Different local positions of these dipoles may cause a broad distribution of the electronic density of states, which could be more pronounced at the donor-acceptor interfaces. Thus, Voc losses can be related to the distribution of electronic states and the presence of interfacial energetic traps. The distribution of electronic states and energetic traps is dependent on the blend morphology. Thus, the morphological disorder may affect the material properties, leading to low values of the measured electrical parameters.
External quantum efficiency (EQE) spectra were obtained by measuring the photocurrent in the short circuit mode (ISC). The EQE was of the order of 20%, which means that one out of five photons was successfully transformed into a charge carrier, reaching the cathode. Since the absorbing layers were relatively thin (ca. 100 nm), only half of the photons were absorbed and about half of the photoexcited electrons arrived at the cathode. The cells with P4 were about two times more efficient than those with P2. As indicated in Figure 7, absorption bands of the donor and acceptor components of the bulk heterojunctions (marked by arrows) are located in the visible part of the spectrum (for the shape of the spectra of P2 and P4, see Figure 1; the spectrum of PC71BM can be found in [50]). This means that both components can efficiently absorb photons, and the interface charge transfer can proceed in both directions. The absorption efficiencies of P2 and P4 were lower than that determined for PC71BM. Different spectral ranges of measurements should, however, be taken into consideration. In the 2 eV range, significantly more solar photons are expected than in the 3 eV range, so the former is more important for the power conversion efficiency (PCE). Taking into account the IP and |EA| values of P2 and P4 and their absorption spectra matching the solar spectrum, reduced exciton dissociation energy due to the presence of permanent dipoles, and efficient photoluminescence quenching in the presence of PC 71 BM, much better cell parameters should be expected. Morphological studies of the active layers clearly showed that unfavorable morphology was the main reason for the cell parameters worsening. Modifications of the processing conditions were, therefore, undertaken. In particular, BHJ layers of P2 and P4 with PC 71 BM were processed from different solvents (chloroform or chlorobenzene) using different additives such as diiodobenzene (DIO) and o-dichlorobenzene (DCB). Although for individual cells, higher PCE values were obtained, the average values calculated for the set of eight cells did not improve (see Table S1 in the supplementary information). Again, unfavorable film morphology was the main reason for this lack of cell parameter improvement (see Figures S8, S9, S11, and S12). It is clear that further improvement of the cell parameters requires the elaboration of processing methods, leading to better BHJ morphologies. Such investigations are in progress.

Materials and Methods
Detailed procedures of the synthesis of the prepared monomers (M1-M4) and polymers (P1-P5) can be found in the supplementary materials. This involves: (i) description of all reagents and procedures used in the preparation of the monomers and the studied polymers together with spectroscopic (NMR, IR) data of intermediate and final products as well as their elemental analyses; (ii) absorption spectra of the studied polymers in chlorobenzene solutions and in solidstate films prepared from chlorobenzene solutions; (iii) description of cyclic voltammetry (CV) and differential pulse voltammetry (DPV) experiments; (iv) description of AFM measurements; (v) device preparation and characterization; (vi) morphology of the active layer investigated with an optical microscope using Nomarsky contrast.

Conclusions
To summarize, we have synthesized new donor-acceptor-type conjugated copolymers modified by the fusion of a permanent dipole-bearing moiety (1-methyl-2-phenyl imidazole, MPI) to the 1,4-phenylene group of the main chain. For comparative reasons, we prepared periodic copolymers of the same main chain but deprived of the MPI groups. The presence of the permanent dipole increases the optical band gap of the studied polymers but lowers their ionization potential (IP) and their electrochemical band gap. Moreover, their exciton binding energy is also reduced, as derived from a combination of spectroscopic and electrochemical investigations. These beneficial, dipole-induced changes of the polymers' properties, combined with appropriate IP and |EA| and good spectral matching with the solar spectrum, are not reflected in the parameters of the tested bulk heterojunction-type solar cells. The measured PCE values are strongly limited by the unfavorable morphology of the active layers.