Conjugated poly ( azomethine ) s via simple one-step polycondensation chemistry : synthesis , thermal and optoelectronic properties

Three conjugated triphenylamine-based poly(azomethine)s were prepared via well-known polycondensation chemistry using cheap and readily available starting materials and the results were contrasted with rrP3HT. Three functionalized diaminetriphenylamines (TPA(X), X 1⁄4 –H, –OMe, –CN) were polymerized in a simple one-step process with 2,3-dihydrothieno[3,4-b][1,4]dioxine-5,7-dicarbaldehyde (ThOx), with water being the only side product. The resulting polymers (TPA(X)ThOx, X 1⁄4 –H, –OMe, –CN) were characterized by GPC, IR and NMR, and show a good thermal stability. The opto-electronic properties could be tuned by changing the functionalization (X 1⁄4 –H, –OMe, –CN) on the triphenylamine moiety. Photovoltaic devices based on TPA(X)ThOx/PCBM (1 : 2) showed power conversion efficiencies in the range of 0.02–0.04%. TRMC measurements showed that the presence of PCBM as an electron acceptor facilitates the formation of free mobile charges after excitation of the polymer. The low device efficiencies are attributed to a low hole-mobility of the polymer in combination with poor active layer morphology.


Introduction
In this paper we will present our efforts in exploring azomethine-based condensation polymers for photovoltaic applications.The development of conjugated polymers for use in organic photovoltaics (OPVs) has progressed rapidly in the last few decades.OPVs provide an attractive alternative to inorganic silicon-based solar cells because of their ease of processing and potential low cost.In the last few years, a wide range of (record breaking) small bandgap hole-transporting polymers have been developed, mostly synthesized via Suzuki and Wittig type reactions.These chemistries, however, make use of expensive metal catalysts, require inert reaction conditions, and need substantial product purication to obtain the polymers in the required purity, making large-scale application limited if not altogether impossible.An interesting alternative to these traditional conjugated polymers could come from polycondensation polymers.2][3] The azomethine bond is isoelectronic with its vinyl analogue, however, its synthesis is more advantageous as water is the only by-product and (expensive) catalysts are not required.
Previous work from our group showed the rst successful application of a series of conjugated poly(azomethine)s (PAMs) as photovoltaic materials in organic bulk heterojunction solar cells. 1 These polymers were based on triphenylamine and (heterocyclic) aromatic dialdehydes.By varying the aromatic dialdehyde, the bandgap of the polymer could be tuned.The absence of alkyl side-chains and the all-aromatic structure of the polymer backbone ensured excellent thermal properties (T g > 250 C and T 5% d > 400 C).In addition, conjugated PAMs are known to be stable under ambient humid conditions and are also robust under common reducing reagents such as NaBH 3 CN, NaBH 4 , high pressure H 2 /Pd and even diisobutylaluminium hydride. 4To further explore the photovoltaic potential of PAMs and improve the photovoltaic performance compared to the previously reported series of PAMs, a new series of functionalized triphenylamine based poly(azomethine)s was designed.The TPA unit was functionalized to investigate the inuence of substituent effects on the optoelectronic properties of the polymer (Scheme 1).An electron donating methoxy group was introduced in order to push electron density towards the polymer backbone thus reducing its bandgap with respect to the unsubstituted polymer.The introduction of a cyano group is expected to result in an increase in the bandgap compared to the unsubstituted polymer by reducing the electron density in the backbone.A cyclic ether was introduced on the thiophene moiety to further improve the solubility and absorption of the polymers compared to those reported previously. 1 To reduce possible geminate recombination at unreacted functional endgroups all polymers were end-capped with mono-functionalized 4-aminotriphenylamine.The three different TPA(X)ThOx oligomers were synthesized with a number average molecular weight (M n ) of 5000 g mol À1 .In order to investigate whether polymer molecular weight plays a role with respect to processing and device performance, we prepared TPA(OMe)ThOx with a M n of 1000 and 9000 g mol À1 .

Materials and methods
All chemicals were purchased from commercial sources and used as received unless stated otherwise.Lithium chloride and 2,3-dihydrothieno [3,4-b][1,4]dioxine-5,7-dicarbaldehyde were dried in an oven at 150 C before use.rrP3HT reference material was purchased from Rieke metals (M n ¼ 50-70k).The structures were conrmed by 1 H-NMR (Bruker WM-400, 400 MHz) and 13 C-NMR (Bruker WM-400, 100 MHz).All samples were dissolved in deuterated solvents and the recorded spectra were referenced to the solvent (CDCl 3 : 1 H 7.26 and 13 C 77.0 ppm) relative to TMS.Mass spectra were recorded with a Shimadzu QP2010S and infrared spectra were obtained with a Perkin Elmer Spectrum 100.UV-vis spectra were collected using a Perkin Elmer Lambda 35 UV-vis spectrometer and photoluminescence spectra were collected by using a PTI Quantamaster spectrouorimeter.

Electrochemical measurements
Cyclic voltammetry (CV) was measured using a Metrohm Potentiostat (PGSTAT302N) with platinum working and counter electrodes and a Ag/Ag + reference electrode.Experiments were performed in oxygen-free chloroform solution of the polymer with 0.1 M tetrabutylammonium hexauorophosphate as an electrolyte and measured at a scan rate of 10 mV s À1 .

Molecular weight measurements
Gel permeation chromatography (GPC) measurements were performed on a Shimadzu Prominence GPC system equipped with a Shodex LF-804 column.N-Methylpyrrolidone (NMP) with 5 mM of LiBr was used as eluent at a ow rate of 0.5 mL min À1 at 60 C. Data analyses were performed with LabSolutions soware from the refractive index detector data.Quantication was made based on polystyrene standard calibration.

Thermal and calorimetric analysis
Thermogravimetric analysis (TGA) was performed on a Perkin Elmer Pyris diamond TG/DTA under a nitrogen atmosphere with a scan rate of 10 C min À1 .Differential scanning calorimetry was performed on a Perkin Elmer Sapphire DSC at a heating rate of 10 C min À1 .

Device preparation
Solar cells were prepared by covering pre-cleaned ITO patterned glass substrates with PEDOT:PSS (Clevios P, VP Al4083).The active layer, consisting of TPA(X)ThOx and PCBM (1 : 2 by weight), was spin-cast from a chloroform solution (15 mg mL À1 ).A 1 nm thick layer of lithium uoride (LiF) followed by a 100 nm thick aluminum layer were deposited by thermal evaporation under vacuum.Current-voltage characteristics were recorded in a glovebox using a Keithley 2400 source meter under simulated solar light (1000 W m À2 ) from a tungstenhalogen lamp ltered by a Schott GG385 UV lter and a Hoya LB120 daylight lter was used to illuminate the devices.The layer thicknesses were measured on a Veeco Dektak 150 prolometer.

Transmission electron microscopy
Transmission electron microscopy (TEM) was performed using a FEI Tecnai TF20 electron microscope operated at 200 kV.The lms were spin-coated on PEDOT:PSS and transferred to a copper grid.

Time-resolved microwave conductivity
The time-resolved microwave conductivity (TRMC) technique is employed to investigate the formation of mobile charge carriers and their decay with time.By using this technique, the change in conductance of the lm on photoexcitation is recorded on a nanosecond time scale without applying external electrodes.The photoactive layers are placed in a microwave sample holder (cavity) with a response time of 18 ns.Samples were photoexcited with a 3 ns laser pulse from an optical parametric oscillator pumped at 355 nm with the third harmonic of a Q-switched Nd:YAG laser (Vibrant II, Opotek).Photogeneration of mobile charge carriers in the sample leads to an increase of the conductance, DG(t), and consequently to an enhanced absorption of microwave power by the sample.The timedependent change of the conductance is obtained from the normalized change in microwave power (DP(t)/P) reected from the cavity according to, The geometrical dimensions of the cavity and dielectric properties of the media in the microwave cavity determine the sensitivity factor, K. From the maximum change in the conductance (DG max ) and the incident light intensity (I 0 ), the parameter 4Sm denoting the product of the charge carrier generation yield (4) per absorbed photon and the sum of the electron and hole mobilities (Sm) can be calculated using, where b is the ratio between the broad and narrow inner dimensions of the waveguide and e is the elementary charge.F A is the optical attenuation value calculated using eqn (3).For more experimental details regarding TRMC, see ref. 5 and 6.
The optical attenuation value (F A ) of the photoactive layer corresponds to the fraction of incident photons that is absorbed by the sample.This was obtained according to, where I T and I R are transmission and reection, respectively, recorded by a Perkin-Elmer Lambda 900 Vis/NIR spectrophotometer equipped with an integrating sphere.
Synthesis 4,4 0 -Diamine-4 00 -methoxytriphenylamine (2a).A ame-dried 100 mL round-bottom ask equipped with a magnetic stir bar, a reux condenser and a nitrogen inlet was charged with 20 mL dry DMSO, K 2 CO 3 (17.89g, 0.13 mol) and p-anisidine (2.93 g, 0.24 mol).This mixture was stirred under nitrogen for 10 min aer which 1-uoro-4-nitrobenzene (6.8 g, 0.48 mol) was added.The yellow reaction mixture was heated to 150 C for 48 h under nitrogen.The reaction mixture was cooled to room temperature and precipitated in 200 mL cold water.The solids were collected by ltration, washed with water and dried under vacuum at 45 C for 5 h.The product was puried by ltering over a short silica column using dichloromethane and recrystallization using glacial acetic acid.Yield: 3.8 g, 45%; TLC (3 : 1 hexaneethyl acetate), t R 0.4 (one spot).

Thermal properties
The thermal properties of the 5K polymers were studied by thermogravimetric analysis (TGA) under an inert nitrogen atmosphere with a heating rate of 10 C min À1 (Fig. 1).TPA(CN) ThOx shows the lowest thermal stability with a T 5% d at 235 C. TPA(OMe)ThOx as well as the unsubstituted TPA(H)ThOx show a higher thermal stability with T 5% d at 291 and 321 C respectively.Although the thermal stabilities of this new series are signicantly lower than those of the previously reported poly-(azomethine)s, the polymers still show a high thermal stability, compared to traditional OPV polymers.Mass spectrometry indicates that the weight loss below 350 C can mainly be attributed to the degradation of the cyclic ether functionalized thiophene moiety.Although the substituents on the triphenylamine moieties have a signicant inuence on the thermal stability of the polymer, the (functionalized) triphenylamine itself is stable up to at least 400 C. 7 Glass transition temperatures could not be observed with differential scanning calorimetry (DSC).The absence of a glass transition (T g ) is most likely due to the relative low degradation temperatures.The T g for unsubstituted thiophene-triphenylamine-based azomethines is typically in the range of 250 to 300 C. 1 Scheme 2 Synthesis of the triphenylamine-based poly(azomethine)s (TPA(X)ThOx) with the different substituents on the triphenylamine unit.The polymers were subsequently end-capped with 4-aminotriphenylamine.

Optical properties
The absorption spectra of the three polymers as well as that of rrP3HT are measured both in solution and in the lm and are shown in Fig. 2. The three TPA(X)ThOx poly(azomethine)s all show two clear absorption peaks in the measured absorption spectrum.As expected, functionalization of the TPA moiety with an electron withdrawing group resulted in a blue shi in the maximum of the absorption spectrum of TPA(CN)ThOx compared to that of the unfunctionalized TPA(H)ThOx.The introduction of an electron donating methoxy group on the other hand did not result in the expected red shi, however, it broadens the absorption spectrum compared to that of TPA(H)-ThOX.The extinction coefficients of the TPA(X)ThOx poly-(azomethine)s were found to be within the same order of magnitude and are comparable to that of rrP3HT.Molecular weight, however, did not signicantly inuence the absorption spectrum of TPA(OMe)ThOx (Fig. 3), as absorption spectra of polymers are known not to change signicantly aer approximately ve repeating units.
The absorption spectra of the polymers in the solid state showed a red shi of only a few nanometers (<15 nm) compared to those measured in solution (Fig. 4).The bandgap of the polymer was determined by using the onset of the absorption spectrum (Table 2).The bandgap of TPA(OMe)ThOx is found to be 1.9 eV, which is similar to rrP3HT.TPA(H)ThOx and TPA(CN)-ThOx show, as expected from the theory, a slightly larger bandgap of 2.0 and 2.1 eV, respectively.

Cyclic voltammetry
The redox behavior of TPA(X)ThOx poly(azomethine)s was studied by cyclic voltammetry (Fig. 5).A small oxidation wave with an onset of around +0.37 V vs. Ag/Ag + can be observed in all three voltammograms.This oxidation is assumed to originate from the thiophene part of the polymer, as this oxidation is seen for all three polymers at the same potential and does not arise Fig. 4 Normalized UV-vis absorption spectrum of the 5000 g mol À1 TPA(X)ThOx polyazomethine series in the solid state; spin-coated from pyridine on quartz.rrP3HT spin-coated from chloroform on quartz is added for reference purposes.
Fig. 2 UV-vis absorption spectrum of the 5000 g mol À1 TPA(X)ThOx polyazomethine series in solution at a concentration of 30 mg L À1 in pyridine.rrP3HT at a concentration of 30 mg L À1 in chloroform is included for reference purposes.from the solvent or experimental conditions.The second larger oxidation wave shows a shi correlating with the electronic effect of the substituent on the TPA moiety; the electron donating methoxy group leads to a lower oxidation potential for TPA(OMe)ThOx, while the electron withdrawing cyano group results in a higher oxidation potential for TPA(CN)ThOx compared to that of the unsubstituted TPA(H)ThOx.
TPA(OMe)ThOx and TPA(CN)ThOx, both bearing a substituent at the para-position on the triphenylamine moiety, showed reversible oxidations, whereas TPA(H)ThOx, lacking a substituent at the para-position, showed a more irreversible behaviour (Fig. 6).0][11] The change in intensity of the observed signal during the measurement is attributed to the deposition of material on the electrodes, resulting in an increased signal.
The positions of the HOMO levels of the polymers were calculated from the onsets (+0.37 V, À4.8 eV) of the oxidation relative to Ag/Ag + and are shown in Table 2.The HOMO level determined for TPA(X)ThOx poly(azomethine)s is very similar to P3HT, which is found at À4.9 eV according to the literature. 8rom combining the determined HOMO energy levels with the optical bandgaps, which were calculated from the UV-vis absorption onsets, the energy of the LUMO levels were estimated (Fig. 7).

Computational study
To gain insight into the electronic structure of the TPA(OMe)-ThOx polymer, density functional theory (DFT) geometry optimizations (B3LYP, 6-31G(d,p)) were performed on oligomeric model structures; a tetramer ($1K oligomer) of TPA(OMe)ThOx (Fig. 8a) and compared with our previously published TPAthiophene polyazomethine (TPA(H)Th), which lacks both the methoxy group on the TPA unit and the cyclic ether on the thiophene moiety (Fig. 8b).The values for the HOMO energy levels of the polymers are extracted from tting the calculated HOMO energy of the different length oligomers versus N, the minimum number of double bonds to go from one extremity of the oligomer to another, with an Kuhn-type expression found in ref. 12 The calculated energy difference between both structures resembles the energy difference found with cyclic voltammetry (DE exp ¼ 0.5 eV versus DE th ¼ 0.32 eV).The addition of the cyclic ether and the methoxy group destabilizes the HOMO of TPA(OMe)ThOx somewhat with respect to that of TPA(H)Th, which appears to be due to the electron donating character of both substituents.Nevertheless, the HOMO appears to be fairly delocalized over a large part of the optimized tetramers, suggesting that both (planar) model compounds are conjugated and the effect of the methoxy group and the cyclic ether is only minimal.

Device fabrication
To evaluate the photovoltaic performance of the TPA(X)ThOx oligomer series, bulk heterojunction solar cells were fabricated by spin-coating a blend of polymer:[60]PCBM on ITO coated glass, on which a layer of electron blocking PEDOT:PSS was spin-coated, followed by a thermally evaporated LiF:Al top electrode.Processing conditions were screened in order to optimize the device parameters.The best performing devices were spin-coated from a chloroform solution with a 1 : 2 TPA(X)-ThOx:PCBM ratio, at a concentration of 15 mg mL À1 .The relatively low concentration of the spin-coating solutions is a result of the limited solubility of the oligomers.Higher concentrations resulted in residual undissolved material causing defects in the lm.A spin speed of 2000 rpm resulted in an active layer thickness of $100 nm.Annealing at different temperatures had no signicant inuence on the device performance.TPA(H)-ThOx, without a substituent on the triphenylamine moiety, showed slightly better performance compared to the other two TPA(X)ThOx oligomers.Although devices obtained from the 1K TPA(OMe)ThOx oligomer can perform better than those from the longer chain 5K oligomer, the overall efficiency is still very poor compared to state-of-the-art devices (I-V curve is shown in Fig. 9).We were not able to obtain reproducible devices from the 9K TPA(OMe)ThOx polymer because of the limited solubility of the polymer.Many of the fabricated devices showed leakage currents, which can be explained by defects in the lm, most likely caused by the limited solubility of this polymer.The results of the device testing are summarized in Table 3.The 1K TPA(OMe)ThOx polymer was further investigated to gain insight into the origin of the unexpected low performance of these polymers.The methoxy-substituted polymer was selected for it has the lowest bandgap and its synthesis is most convenient for scale up.

Photoluminescence
The photoluminescence (PL) quenching, to determine charge transfer characteristics, of the 1K TPA(OMe)ThOx polymer by [60]PCBM was investigated both in solution and in lm.When TPA(OMe)ThOx was excited at its absorption maximum at 480 nm, a photoluminescence peak around 600 nm was observed (Fig. 10).However the photoluminescence efficiency was found to be around two orders of magnitude lower than P3HT, which is considered as a weakly uorescent material  itself.4][15][16] Skene and co-workers showed that suppressed uorescence of azomethines can be the result of the deactivation of the excited state by internal conversion processes involving bond rotation. 15The fast deactivation of the excited state could be in competition with the charge transfer decreasing the amount of free charges that would be formed.
Fluorescence from the lm showed a small red-shi of approximately 20 nm compared to that measured in solution, which can be explained by a better ordering of the polymer chains in the solid-state.Nevertheless also in the solid-state, the photoluminescence of TPA(OMe)ThOx is very weak compared to that of P3HT.Due to the low uorescence efficiency, quenching by [60]PCBM could not be observed, either in solution or in the lm.

Time resolved microwave conductivity
In order to study dissociation of excited states into free mobile charges, we record microwave photoconductance transients for the TPA(OMe)ThOx polymer and its blend with the electron acceptor PCBM (Fig. 11).Upon photoexcitation of these thin lms with a 3.5 ns laser pulse the transient shows a clear rise followed by a slow decay on a time scale of hundreds of nanoseconds.Due to the 18 ns response time of the microwave cavity the maximum of the photoconductance signal is observed at about 20 ns.It can be clearly observed that the amplitude of the photoconductance for the blend is increased by a factor of 25 as compared to that for the polymer alone.This demonstrates that the excited state dissociates into free mobile charge carriers, due to electron transfer from the polymer to PCBM.From the maximum change in conductance (DG max ) the product of the quantum yield per absorbed photon (4) and the sum of the charge carrier mobilities (Sm) was calculated using eqn (2).The 4Sm value at lower uences remains constant, yielding a value of 3 Â 10 À4 cm 2 V À1 s À1 .However, at higher uences a reduction of the 4Sm values is observed that has been previously attributed to higher order recombination processes. 17The 4Sm value observed for this blend is approximately two orders of magnitude smaller compared to that of P3HT:PCBM blends (Fig. 12). 18This difference can be both due to a smaller yield of charge carriers and/or lower mobility.

Mobility study
Hole-only devices were prepared for the 1K TPA(OMe)ThOx polymer by evaporating gold, instead of a LiF/Al, top electrode on top of the active layer.The current-voltage characteristics (Fig. 13) can be t assuming space-charge-limited currents with mobility exhibiting a Poole-Frenkel-type eld dependence.The modeled mobilities are very low, for example, with a mobility of approximately 10 À8 cm 2 V À1 s À1 at a eld of 3 Â 10 5 V cm À1 .Under working conditions this low hole mobility will almost certainly lead to unbalanced charge transport, explaining for a large part of the poor efficiency obtained with photovoltaic devices comprising this polymer.

Transmission electron microscopy
The morphology of the active layer was studied by transmission electron microscopy (TEM).A P3HT sample was used for reference purposes and the morphology observed was very similar to what has been published for rrP3HT/PCBM (Fig. 14B).
The lms from a blend of 1K TPA(OMe)ThOx with PCBM (1 : 1 and 1 : 2) are amorphous and very homogeneous, as can be seen in Fig. 14A.No morphological features were visible on the image, indicating a completely homogeneous lm.The polymer might dissolve the PCBM, resulting in a homogeneous mixed active layer.It is known from the literature that polymers with aromatic rings and atoms signicantly larger than carbon, like sulphur, are good solvents for C 60 . 19Good mixing of the polymer with PCBM prevents phase separation, which leads to a poor interpenetrating network, which is an explanation for the low mobility as well as the poor device efficiency.

Conclusion
In conclusion, three new poly(azomethine)s with different substituents on the triphenylamine unit have been synthesized via a simple one-step polycondensation reaction.Although the functionalizations with respect to the previously reported poly(azomethine)s were expected to increase the solubility of these new polymers, we found that their solubility was actually very low.Because of the aromatic units and the absence of functional groups, the polymers showed good thermal stability   but no glass transition (T g ).By changing the substituent on the triphenylamine moiety the bandgap of the polymer was inuenced.The electronic properties of the polymers were studied by cyclic voltammetry and are found to be similar to those of rrP3HT.Photovoltaic devices were fabricated comprising the three TPA(X) ThOx polymers in combination with [60]PCBM.Power conversion efficiencies of around $0.02% were found, with the highest efficiency of 0.04% for the further optimized devices comprising the 1K TPA(OMe)ThOx polymer.The low efficiencies were explained by the low hole mobility of the polymer in combination with a poor morphology of the active layer.In addition, the weak uorescence of the pristine polymer indicates the fast internal decay of the excited state, which will be in competition with dissociation into free charges.TRMC measurements showed that free mobile charges are produced by photoexcitation, albeit less efficiently than for rrP3HT-based devices.

Fig. 3
Fig.3Normalized UV-vis absorption spectrum of the TPA(OMe)ThOx polyazomethine with a M n of 1K, 5K and 9K in pyridine.Fig.1TGA thermograms of the 5000 g mol À1 TPA(X)ThOx polyazomethine series recorded under a nitrogen atmosphere with a heating rate of 10 C min À1 .

a
Measured in a quartz cuvette in chloroform.b Spin-cast on a quartz substrate from pyridine.c Estimated from the onset of the absorption spectrum of the lm, 1240/l onset .d Determined by cyclic voltammetry, assuming F c is 4.8 eV below vacuum.e Estimated by subtracting the band-gaps from the HOMO energy levels.f Electrochemical data from ref. 8.

Fig. 5 Fig. 6 Fig. 7
Fig. 5 Normalized cyclic voltammogram of the TPA(X)ThOx in chloroform containing 0.1 M TBAPF 6 measured with a scan rate of 10 mV s À1 .Fig. 6 Evolution of the cyclic voltammograms over 15 scans of TPA(H)ThOx in chloroform containing 0.1 M TBAPF 6 measured with a scan rate of 10 mV s À1 .

Fig. 9 I
Fig. 9 I-V curve of 1K TPA(OMe)ThOx in the dark and under illumination.

Fig. 10
Fig. 10 Room temperature PL spectra of 1K TPA(OMe)ThOx and P3HT at a concentration of 13 mg L À1 in chloroform.TPA(OMe)ThOx was excited at 480 nm and P3HT at 450 nm.

Fig. 11
Fig.11Photoconductance transients obtained for the neat TPA(OMe)ThOx polymer and a TPA(OMe)ThOx:PCBM blend.Both thin films were prepared by spin-coating from chloroform.Traces were recorded at a 480 nm excitation wavelength using a fluence of 8.9 Â 10 14 photons per cm 2 .

Fig. 13
Fig.13Linear-log and log-log plot to determine the field independent mobility of TPA(OMe)ThOx by fitting the equation for space-charge limit current to the I-V characteristic of a hole only device in the high field region.

Fig. 12
Fig. 12 Product of the quantum yield per absorbed photon (4) and the sum of the charge carrier mobilities (Sm) versus incident intensity normalized to the optical absorption of a TPA(OMe)ThOx polymer and a TPA(OMe)ThOx:PCBM blend.

Table 1
Molecular weight of TPA(X)ThOx poly(azomethine)s determined by GPC a Could not be determined because of limited solubility.

Table 2
Summary of the optoelectronic properties of TPA(X)ThOx

Table 3
Device characteristics of polymer:[60]PCBM devices prepared from a 1 : 2 chloroform solution under simulated AM1 conditions a No reproducible devices could be obtained because of the limited solubility of the polymer.