Panagiotis Giannopoulosa,
Aikaterini K. Andreopoulou*ab,
Charalampos Anastasopoulosa,
Dimitrios Raptisbc,
Georgia Sfyribd,
Joannis K. Kallitsisab and
Panagiotis Lianosbc
aDepartment of Chemistry, University of Patras, University Campus, Rio-Patras, GR26504, Greece. E-mail: andreopo@upatras.gr
bFoundation for Research and Technology Hellas/Institute of Chemical Engineering Sciences (FORTH/ICE-HT), Platani Str., Patras, GR26504, Greece
cDepartment of Chemical Engineering, University of Patras, 26504 Patras, Greece
dDepartment of Physics, University of Patras, 26504 Patras, Greece
First published on 6th January 2016
The development of copolymers based on ionic liquid vinyl monomers of the imidazole family combined with polymerizable chromophoric units is presented herein. For this end, ruthenium complexes bearing polymerizable vinyl groups or ω-end vinyl rr-poly(3-alkyl thiophene) were prepared and copolymerized with the ionic liquid monomers under free radical polymerization conditions, affording chromophore/polyelectrolyte combinations. Homopolymer ionic liquids were also synthesized to select the optimum conditions for the copolymers thereafter. All the monomers and polymers were characterized for their optical properties, and were also structurally characterized using various complementary techniques. Selected copolymers and homopolymers were tested in quasi-solid-state sensitized solar cells based on titania and regioregular poly(3-hexyl thiophene) acting as a hole-transporting semiconducting polymer. The ionic liquid, which is miscible with the hole-conductor and can be deposited with the latter, provides a functionality that, in some cases, supports an increase in open-circuit voltage, thus increasing cell efficiency.
In a typical DSSC, an inorganic semiconductor, usually TiO23–6 or ZnO,7,8 acts as an electron acceptor and transporter. The semiconductor also supports the sensitizer, i.e. a material that absorbs visible light and injects the photogenerated electron into the conduction band of the supporting (acceptor) semiconductor. In DSSCs, a dye plays the role of the sensitizer. In this role, the dye must have anchoring groups to form stable bonds with the semiconductor's surface. Such anchoring groups are mainly carboxylic, although other anchors have been suggested, such as the perfluorophenyl group, which was recently reported by our group.9,10 The remaining functional components of a DSSC are a counter electrode – usually carrying Pt nanoparticles as the catalyst, although various attempts have been reported aimed at the replacement of Pt, mainly with carbon hybrids11 – and an electrolyte.12–14
The electrolyte in DSSCs is very important because it provides the required ionic conductivity and helps dye regeneration through charge transfer reactions with the dye molecules. Liquid electrolytes present a major drawback in the use of DSSCs due to solvent evaporation and a subsequent reduction of device lifetime and performance, especially when the device is not perfectly sealed.15 Thus, the replacement of typical solvents (or solvent mixtures) with solid or quasi-solid electrolytes is a very interesting and active field of research.16 Commonly, ionic liquids, p-type hole-transport organic semiconductors, plastic crystals and polymer gel electrolytes have been used instead of the typical liquid ones. This kind of electrolyte shows long-term stability, although in many cases the efficiency is lower than typical DSSCs.17–25
ILs are widely studied organic salts and have gained much attention for a wide range of applications in various chemical and industrial fields. This is attributed to their unique properties, such as high ionic conductivity, good thermal, chemical and electrochemical stability and low vapour pressure.26,27 Due to these properties, ILs are commonly used as solvents for organic synthesis or catalysis (green solvents) and for polymerization processes, such as free radical polymerization28,29 and atom transfer radical polymerization (ATRP).30,31 Their use is also extended into other applications, such as lithium batteries,32–34 light emitting devices35 and electrochemical capacitors.36 Various IL chemical structures have been reported employing, for example, the imidazolium or pyrrolidinium moieties as cations and hexafluorophosphate or tetrafluoroborate as anions. In the field of DSSCs, imidazolium-based ionic liquids,37–39 or more recently electrolytes based on these ionic liquids,40,41 have been studied.
Polymerizable ionic liquid monomers and their respective polymers (PILs) are now considered as a specific field in polymer science. These polymeric ionic liquids show increased thermal stability, combined with interesting electrochemical properties. One of the most studied imidazolium-based PILs is poly[1-(4-vinylbenzyl)-3-butyl imidazolium tetrafluoroborate] (PVBT), which remains stable up to 300 °C,42 is highly polar, is soluble in solvents such as N,N-dimethylformamide (DMF), dimethylsulfoxide (DMSO) or acetonitrile because of its ionic nature, and at the same time is hydrophobic. PILs are identified as exotic polyelectrolytes for various applications, like in DSSCs, in which they could effectively replace the liquid electrolytes used up to now.43–46 Although a lot of progress has been reported in this field, with recent works aiming at the formation of IL-based hybrids,47,48 the use of IL-based polymer gels has so far only resulted in a few successful studies,28,49 with even less employing copolymers.50 Despite this, ongoing research is focusing on this topic, since ILs can overcome solvent electrolyte leakage, whereas the combination of an IL-based electrolyte with a dye could improve cell stability without sacrificing the overall efficiency of the cell.51
An optimum concept for such organic solvent-free quasi-solid-state DSSCs could be the simultaneous incorporation of chromophores and ILs, which could broaden absorption and ion transfer in the final device. These facts urged us to investigate the use of PILs as electrolytes in DSSC applications and to compare them with copolymers based on ILs and suitable dyes. Both organic dyes and metallocomplexes were employed, in order to compare the applicability of organic versus organometallic dyes, even though the drawbacks of ruthenium (or other metal) containing dyes, such as toxicity and availability, needs also to be taken into consideration. Therefore, in the present study, we employed two polymerizable vinyl-ILs, namely 1-(4-vinylbenzyl)-3-methyl imidazolium tetrafluoroborate (VBMImT)52 and 1-vinyl-3-hexyl-imidazolium bromide (VHImBr)53 and polymerizable dye (macro)molecules, a low molecular weight ω-end vinyl terminated regioregular poly(3-alkylthiophene)54 and a vinylphenyl-bipyridine ruthenium complex with two bipyridines as co-ligands,55 in order to prepare homopolymer and copolymer PILs. The polymerizations were carried out under various conditions in order to achieve the best yields for the homopolymers, but also as a first test regarding the forthcoming copolymerizations. Through this route, the homopolymers PVBMImT and PVHImBr were synthesized, followed by the synthesis of copolymers based on the ionic liquids combined with the ω-end vinyl-poly(3-octyl-thiophene) or the (vinylphenyl-bipyridine)-Ru(II)-di(bipyridine) complex. Both the homopolymer and copolymer PILs were structurally characterized and their optical properties evaluated. Such materials were then employed in the fabrication of solar cells and their performance was then tested.
All the solvents and reagents were purchased from Sigma Aldrich, Alfa Aesar or Acros Organics and were used without further purification unless otherwise stated. All the reactions were carried out under an argon atmosphere.
Tetrahydrofuran (THF) was dried and distilled just before use over sodium wire in the presence of benzophenone. Acetonitrile (MeCN) was dried and distilled over calcium hydride (CaH2). 2,2′-Azobis(2-methylpropionitrile) (AIBN) was recrystallized using methanol (MeOH).
Dye sensitizer N719 was obtained from Solaronix, conductive polymer PEDOT:PSS (Baytron P) from H. C. Starck, regioregular poly(3-hexyl thiophene) from Rieke Metals and silver ink (5007E) from DuPont.
Thermogravimetric analysis (TGA) was carried out on ca. 8 mg samples contained in alumina crucibles in a Labsys TM TG apparatus of Setaram under nitrogen and at a heating rate of 10 °C min−1.
IR spectra were recorded on an ALPHA-P diamond ATR spectrometer of Bruker Optics GmbH.
Mass spectra were recorded on a Waters Acquity ESI-MS system with a Waters 2966 Photodiode Array detector coupled to a Waters Micromass ZQ mass spectrometer. A single-quadrupole Quattro micro mass spectrometer (ACQUITY SQ Detector) equipped with an electrospray ionization (ESI) interface was used for analytical detection. ESI-MS was operated in positive mode under the following operating parameters: capillary voltage, 3.5 kV; cone voltage, 30 V; source temperature, 80 °C; desolvation temperature, 300 °C; desolvation gas (nitrogen), 5500 L h−1; cone gas (nitrogen), 50 L h−1. All the data were acquired and processed using Masslynx 4.1 software (Waters Corp., MA, Milford, USA).
Size exclusion chromatography (SEC) measurements were carried out at 25 °C at a flow rate of 1 mL min−1, using a polymer lab chromatographer equipped with two Ultra Styragel linear columns (104 Å, 500 Å) and a UV detector. Polystyrene standards were used for the calibration and the eluent was CHCl3.
UV-Vis spectra were recorded using a Hitachi U-1800 spectrophotometer. Photoluminescence (PL) spectra were recorded using a Perkin-Elmer LS45 luminescence spectrometer, after excitation at the UV-Vis absorption maxima.
For solar cell characterization, illumination of the samples was made with a PECCELL PEC-L01 Solar Simulator set at 100 mW cm−2. J–V characteristic curves were recorded under ambient conditions with a Keithley 2601 source meter controlled by Keithley computer software (LabTracer).
Transmission electron microscopy (TEM) measurements were performed on a JEOL JEM2100 operating at 200 kV. Sample preparation for TEM examination involved the preparation of dilute solutions of the samples in CHCl3. A drop of the solution was placed on 3 mm carbon-coated copper grids (Electron Microscopy Sciences) and the samples were dried in air for 2 days.
a/a | Monomer | Solvent | Monomer/AIBN ratio |
---|---|---|---|
Homopolymers | |||
1 | VBMImT | DMF | 20/1 |
2 | VBMImT | DMSO | 15/1 |
3 | VBMImT | MeCN (dry) | 15/1 |
4 | VHImBr | DMF | 20/1 |
5 | VHImBr | CHCl3 | 15/1 |
6 | VHImBr | MeCN (dry) | 15/1 |
![]() |
|||
Copolymers | |||
1 | VBMImT/v-P3OT | MeCN (dry)/CHCl3 | 15/1 |
2 | VHImBr/v-P3OT | CHCl3 | 15/1 |
3 | VBMImT/Ru(bpy)2(styrylbpy) | MeCN (dry) | 11/1 |
The first part of Schemes 1 and 2 show the synthetic procedures for the preparation of the ionic liquid monomers VBMImT and VHImBr, respectively.
The successful formation of the desired monomers was confirmed by 1H NMR (Fig. S1a and S2a†) and ESI-MS (Fig. S3†). The assignment of the NMR peaks to the monomers' protons show the absence of any by-products. This is also proven by the mass spectra, which show the M+ ion mass, as the ESI-MS instrument was tuned in positive ion mode. In the case of VBMImT, an intense m/z of 117.38 [M+] and 118.39 [M + H+] could be observed, besides the ones at 199.32 [M+] and 200.33 [M + H+] attributed to the product. This peak is due to the fragment that occurs when the methylimidazolium unit is cleaved from VBMImT. IR spectroscopy of the monomers was also carried out, as shown in Fig. S4a and S4b,† for the two monomers, respectively, and in comparison to the starting materials in each case. In both ionic liquid monomers, characteristic bands from the presence of the imidazolium unit can be observed. More specifically, the symmetric CC stretching vibration around 1565–1570 cm−1, the aromatic C–H stretch (3050–3100 cm−1), the C
N stretch (1630–1650 cm−1), the C–N stretch (1160–1170 cm−1) and the in-plane C–C vibrations of the imidazolium ring (around 1580 cm−1) exist in both products. On the other hand, the VBMImT monomer presents two peaks at 760 cm−1 and 1150 cm−1, respectively, due to the BF4− anion. The VHImBr monomer shows an aliphatic C–H peak around 2850 cm−1, because of the hexyl chain, and a broad peak at 3300–3500 cm−1, due to the hydration of the ionic liquid.58–61
After the successful formation of the homopolymers, respective copolymers based on the ionic liquid vinyl monomers and vinyl-functionalized dye molecules, either a ω-end vinyl-polythiophene macromonomer (Scheme 3) or a ruthenium complex (Scheme 4) were formed.
![]() | ||
Scheme 3 Synthesis of polythiophene–ionic liquid random copolymers poly(P3OT-co-VBMImT) and poly(P3OT-co-VHImBr). |
For the polythiophene containing copolymers of Scheme 3, their 1H NMR characterization (Fig. 1a and b), using CDCl3, revealed the disappearance of the vinyl groups (located at 5.2, 5.6 and 6.8 ppm), which demonstrates the full conversion of the monomers to the polymers. The conversion of the ionic liquid monomers was further confirmed by ATR spectroscopy (Fig. 2). The existence of a small peak at 2900 cm−1 in PVBMImT but not in VBMImT is due to the conversion of vinyl groups (CH2CH bond) to the polymeric chain (CH2–CH bond). This alternation could not be observed among VHImBr and PVHImBr, because VHImBr bears an alkyl hexyl chain, which absorbs in the same region. After the copolymerization of the IL vinyl monomers with vinyl-P3OT, the specific peak is more intense because of the thiophene's octyl chain. Vibrations from the imidazolium group can be observed at 1040 and 1140 cm−1 due to the C–N bond and near 1550 cm−1 due to C
N.59,62 Furthermore, bands appearing in the area 700–900 cm−1 can also be ascribed to contributions from the ring bending modes of the imidazolium cation.63
![]() | ||
Fig. 2 ATR spectra of poly(P3OT-co-VBMImT) (a) and poly(P3OT-co-VHImBr) (b) in comparison to the starting materials and the respective homopolymers. |
Efforts were also made for the characterization of the copolymers via GPC. However, due to the ionic nature and the nitrogen-containing heterocycles of the polymeric ionic liquid blocks, the molecular weights were substantially underestimated, leading to even lower molecular weights than the initial vinyl polythiophene macromonomer. In particular, for the poly(P3OT-co-VBMImT) copolymer, an Mn = 6.500 was found, while for the copolymer poly(P3OT-co-VHImBr)], its Mn was in the range of 4.800, both of which are significantly lower than the initial vinyl-P3OT macromonomer's Mn of 21.000.
The polythiophene–ionic liquid copolymers were also characterized by thermogravimetric analysis (TGA). Specifically, as seen in Fig. S5,† both copolymers show increased thermal stability when compared to the ionic liquid monomers and homopolymers, and are almost completely degraded around 450 °C (VBMImT and PVBMImT) and 350 °C (VHImBr and PVHImBr). On the other hand, vinyl-P3OT is thermally stable up to 500 °C, leading to copolymers with higher degradation temperatures.
Regarding the optical properties of the synthesized compounds, UV-Vis absorption and PL emission spectra were carried out in solution. Vinyl-P3OT and the corresponding copolymers poly(P3OT-co-VBMImT) and poly(P3OT-co-VHImBr) were dissolved in CHCl3. VHImBr is also soluble in chloroform, in contrast with the other monomeric ionic liquid, VBMImT, for which acetonitrile was used instead. Finally, both homopolymers were dissolved in MeOH. The two copolymers demonstrated similar behaviours in their absorption spectra, presenting a broad peak around 450 nm. As can be seen in Fig. 3, this peak is slightly shifted to a lower wavelength in the copolymers' cases in comparison to the pure vinyl-P3OT. The intense peak at 280 nm in poly(P3OT-co-VBMImT) is attributed to VBMImT and is shifted around 25 nm in comparison to the monomeric VBMImT.
![]() | ||
Fig. 3 UV-Vis spectra of poly(P3OT-co-VBMImT) (a) and poly(P3OT-co-VHImBr) (b) in comparison to the starting materials and the respective homopolymers. |
Fig. 4 shows the PL spectra of the copolymers versus the vinyl-P3OT macromonomer in chloroform solutions, after excitation at the P3OT absorption maximum. The PL spectra of the respective ionic liquid homopolymers are given in the inset of Fig. 4. For both the copolymers, emissions due to the P3OT blocks are detected at 564 nm, which is slightly blue-shifted compared to the net vinyl-P3OT, which emits at 570 nm.
Scheme 4 presents the copolymerization reaction of the Ru(bpy)2(styrylbpy) monomer55 with the monomeric VBMImT. Since both the ruthenium complex and the monomeric ionic liquid are soluble in acetonitrile, MeCN was used as the polymerization solvent, while the IL/initiator ratio was 11/1 (Table 1).
The structural characterization of the copolymer was carried out using 1H NMR and ATR spectroscopies (Fig. 5 and S6a†). Thermogravimetric analysis of poly(Ru-co-VBMImT) showed increased thermal stability in comparison to the ionic liquid monomer and homopolymer (Fig. S6b†), which is attributed to the increased thermal stability of the ruthenium complex monomer.
UV-Vis and PL spectroscopies of acetonitrile solutions of the Ru(bpy)2(styrylbpy) and of the copolymer poly(Ru-co-VBMImT) were employed for their optical properties evaluation, as shown in Fig. 6, respectively. In the absorption spectra of the poly(Ru-co-VBMImT), no significant alternation was observed in comparison to the monomer complex Ru(bpy)2(styrylbpy) in the region above 260 nm (Fig. 6a). On the other hand, the copolymer showed an intense peak around 210 nm, which does not exist in the monomeric ruthenium complex, and is attributed to VBMImT. At 320 nm, both the monomeric complex and the copolymer show an absorption due to the π–π* LC (ligand-centered) transitions of 2,2′-bipyridine, and also an absorption at around 450 nm, which is characteristic of the d–π* MLCT (metal to ligand charge transfer) of the Ru(bpy)2(styrylbpy) complexes.55 The photoluminescence examination in solution of both the copolymer poly(Ru-co-VBMImT) and of the monomer complex Ru(bpy)2(styrylbpy) showed a broad peak around 600 nm, after excitation at 450 nm, due to the ruthenium complex's d–π* MLCT absorption band (Fig. 6b).
![]() | ||
Fig. 6 UV-Vis spectra (a) and PL spectra (b) after excitation at 450 nm of VBMImT, PVBMImT, Ru(bpy)2(styrylbpy) and poly(Ru-co-VBMImT) in MeCN solutions. |
A few different polymeric ionic liquids have been employed and tested in quasi-solid-state DSSC configuration. Fig. 7 and Table 2 show current–voltage data for the studied cells. It must first be underlined that the cells showed an acceptable performance when the polymeric ionic liquid content was relatively low. Thus, optimized data were obtained when the P3HT/polymeric ionic liquid ratio was 9:
1 w/w. On the other hand, most polymeric ionic liquids did not show any substantial influence on the cell performance compared to pure P3HT. However, one of the prepared samples demonstrated a marked improvement of cell efficiency. Indeed, as seen in Fig. 7 and Table 2, the introduction of PVHImBr resulted in about a 30% improvement of efficiency. This increase was due to the large increase in the open-circuit voltage of the cell and the improvement of the fill factor, despite the fact that a substantial decrease in the current density was caused by the introduction of the ionic liquid. The decrease in the current was most probably due to the fact that the above cells function by hole collection and transfer through the hole-conducting P3HT. The interference of the ionic agent most probably disrupts some of the hole-transporting channels. On the contrary, the presence of the ionic agent establishes a potential gradient that results in an open-circuit voltage increase. Such a voltage increase was also observed in a previous publication treating a solid-state quantum-dot-sensitized solar cell where a tiny quantity of Na2S was added as an ionic agent.66 Even though the current dropped in the present case, the voltage increase had a marked beneficial effect on cell performance. The advantage of the presently introduced materials is that they can be mixed with P3HT to eliminate solubility problems.
Composition of the hole-transporting/ionic liquid layer | Voc (V) | Jsc (mA cm−2) | FF (%) | η (%) | |
---|---|---|---|---|---|
1 | P3HT alone | 0.42 | 17.0 | 35 | 2.5 |
2 | P3HT/poly(P3OT-co-VHImBr) | 0.42 | 7.0 | 30 | 0.9 |
3 | P3HT/PVBImT | 0.29 | 11.2 | 33 | 1.1 |
4 | P3HT/PVHImBr | 0.55 | 14.1 | 42 | 3.2 |
As can be seen in Fig. 8a, the net P3HT formed uniform films that presented some degree of organization into fibrillary structures due to the close packing and crystallization of the high molecular weight P3HT.1 The casting procedure for the preparation of the TEM samples, of course, does not allow ordering to a great extent, but the method for preparing the P3HT film on the DSSC device, which is accompanied by thermal treatment, should have a more prominent effect on the crystallization of the P3HT macromolecular chains.
On the contrary, the 9:
1 blend of P3HT/poly(P3OT-co-VHImBr) shown in Fig. 8b did not present any fibrillar morphological features. The insertion of even a small percentage of the poly(P3OT-co-VHImBr) copolymer to the P3HT matrix obviously led to the compatibilization of the P3HT and the PIL fractions and, moreover, suppressed the crystallization ability of the net P3HT.
Surprisingly, the P3HT/PVHImBr 9:
1 blend (see Fig. 8c) presented the same crystallization tendency of the net P3HT (see Fig. 8a), while some darker regions were observed. When looking deeper into those darker regions, we noticed that they consisted of dark spots of about 2 nm and up to 6 nm diameter (inset of Fig. 8c). These higher contrast spots were attributed to the charged imidazolium cation and Br anion of the PVHImBr ionic liquid homopolymer and, more specifically, to the high contrast of the Br atoms under the electron beam.67 Therefore, it can be concluded that the presence of the homopolymer PVHImBr did not disturb the crystallization and ordering of the P3HT and that it was quite uniformly dispersed in selective regions along the polythiophene matrix.
The abovementioned findings constitute an explanation of the better DSSC device performance of the P3HT/PVHImBr blend over the P3HT/poly(P3OT-co-VHImBr) blend, since for P3HT it is known that charge transport is promoted along its crystalline domains, which are interconnected with amorphous regions.1,68 Thus, the net P3HT and the P3HT/PVHImBr blend worked nicely in the DSSC devices, whereas the P3HT/poly(P3OT-co-VHImBr) blend with no ordering into crystalline species did not show an analogous DSSC performance.
The homopolymer PILs and copolymer PILs with the thiophene- or ruthenium-based dyes were prepared via radical polymerizations. Their adequate solubility in common organic solvents allows their processability and introduction as an additive of the net organic hole-transporting layer of the cell. Although further improvements are, of course, necessary to increase the efficiency of the solid-state DSSCs, the methodology reported herein shows a promising route for the enhancement of the cells performance. Perhaps the most important finding of the present study is that the presence of the PIL fraction into the hole-transporting P3HT layer improves the device performance as long as it does not disturb the development of crystalline domains that promote the necessary charge carrier transport.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra27374a |
This journal is © The Royal Society of Chemistry 2016 |