Lignin-Based Polymer Electrolyte Membranes for Sustainable Aqueous Dye-Sensitized Solar Cells

In the quest for sustainable materials for quasi-solid-state (QS) electrolytes in aqueous dye-sensitized solar cells (DSSCs), novel bioderived polymeric membranes were prepared in this work by reaction of preoxidized kraft lignin with poly(ethylene glycol)diglycidylether (PEGDGE). The effect of the PEGDGE/lignin relative proportions on the characteristics of the obtained membranes was thoroughly investigated, and clear structure–property correlations were highlighted. In particular, the glass transition temperature of the materials was found to decrease by increasing the amount of PEGDGE in the formulation, indicating that polyethylene glycol chains act as flexible segments that increase the molecular mobility of the three-dimensional polymeric network. Concurrently, their swelling ability in liquid electrolyte was found to increase with the concentration of PEGDGE, which was also shown to influence the ionic transport efficiency within the membrane. The incorporation of these lignin-based cross-linked systems as QS electrolyte frameworks in aqueous DSSCs allowed the preparation of devices with excellent long-term stability under UV–vis light, which were found to be superior to benchmark QS-DSSCs incorporating state-of-the-art carboxymethylcellulose membranes. This study provides the first demonstration of lignin-based QS electrolytes for stable aqueous DSSCs, establishing a straightforward strategy to exploit the potential of lignin as a functional polymer precursor for the field of sustainable photovoltaic devices.


Fourier-transform infrared spectroscopy
Fourier-transform infrared (FTIR) spectroscopy was performed on a Nicolet 760 FTIR spectrophotometer. FTIR spectra were recorded in transmission mode at room temperature in air by recording 32 accumulated scans at a resolution of 2 cm −1 in the 4000−400 cm −1 wavenumber range.
The obtained signals were normalized to the absorbance observed at 1510 cm -1 , which corresponds to the pure-aromatic skeletal vibrations in lignin.

Differential scanning calorimetry
Differential scanning calorimetry (DSC) was used to investigate the thermal transitions in lignins and lignin-based membranes. The measurements were performed on solid state samples (~10 mg) by using a Mettler Toledo DSC/823e instrument performing three runs (heating/cooling/heating) from -50 °C to 200 °C at a scan rate of 20 °C/min under N2 atmosphere. The determination of the glass transition temperature (Tg) of the tested materials was based on the evaluation of the inflection point of the DSC trace in the last heating ramp.

Thermogravimetric analysis
Thermogravimetric analyses (TGA) were performed on the parent lignins as well as on the ligninbased membranes by means of a Q500 TGA system (TA instruments) from room temperature up to 700 °C at a rate of 10 °C/min. The response under both N2 and air atmospheres was tested.

P-NMR spectroscopy
NMR experiments were carried out on a Bruker Avance 400 spectrometer. Acquisition and data treatment were performed with Bruker TopSpin 3.2 software. In order to quantify the -OH groups in the pristine lignins, 31 P NMR spectra were recorded by inverse gated proton decoupling S3 sequences using 90° pulse flip angle, with a 380-ppm spectral width, 256 scans and a relaxation delay of 5 s. Before analyses, the lignins were dried under vacuum overnight at 40 °C and derivatized according to a procedure previously reported in literature. S1 As phosphitylation reagent, internal standard (IS) and relaxation reagent 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholane, N-hydroxy-5-norbornene-2,3-dicarboxylic acid imide and chromium (III) acetylacetonate were selected, respectively. The integration regions considered in this work were 152.8−152.5 ppm for the IS, 150.0−145.0 ppm for aliphatic hydroxyls, 145.0−137.0 ppm for phenolic hydroxyls and 137.0−134.5 ppm for carboxylic hydroxyls.

Gel permeation chromatography
Gel permeation chromatography (GPC) was used to determine the molecular weight of parent lignins. A Waters 510 HPLC chromatograph was used equipped with a Waters 2410 refractive index detector using THF as eluent. The sample to analyze (concentration in THF 2 mg/mL, dissolution volume 200 µL) was injected into a system of columns (Ultrastyragel models HR4, HR 3 and HR 2, provided by Waters) connected in series. The chromatographic analysis was performed at 30 °C and at a flow rate of 0.5 mL/min. Polystyrene standards in the 10 2 -10 4 g/mol molecular weight range were used to calibrate the GPC system.

Free swelling capacity
The gravimetric free swelling capacity (FSC) of the materials was determined by placing approximately 0.3 g of dry lignin-based membrane in de-ionized water. After a certain amount of time, the swollen membrane was allowed to drip off for few minutes and the amount of absorbed water was determined gravimetrically following Equation 1, where mdry and mswollen refer to the mass of dry and swollen membrane, respectively. S4 (S1)

Electrochemical and photoelectrochemical characterization
Solar cells performance was evaluated recording tree consecutive photocurrent density photovoltage curves on a Keithley 2420 Source Measure Unit, keeping a scan rate equal to 20 mV s −1 . Cells were irradiated under simulated 1 sun light intensity (100 mW cm −2 , AM 1.5G) after calibration by a silicon photodiode. A VeraSol-2 LED Class AAA solar simulator was used in this work.
Transient photocurrent measurements were carried out employing a modulation (on/off) of the incident light on DSSC devices, while varying the irradiation intensity from 0.2 to 1.0 sun step-bystep.
Open circuit voltage decay (OCVD) measurements were performed using a CH Instruments Inc. Electrochemical impedance spectroscopy (EIS) data were obtained with the same potentiostat, exploring the frequency range between 100 kHz and 0.1 Hz. The amplitude of the AC signal was 10 mV. Spectra were recorded under 1 sun illumination at applied DC potentials equal to the inverse of previously measured Voc values. S2 Z-view software was used to interpolate experimental spectra by using the transmission line (TL) model as equivalent circuit.
The ionic conductivity of lignin-based QS polymer electrolytes, sandwiched between stainless steel blocking electrodes, was measured by means of EIS under scanning frequencies ranging from S5 and the ionic conductivity was calculated by means of the equation σ = l / (Rb⋅A), where l (cm) is the film thickness and A (cm 2 ) the effective contact area between electrolyte and electrode.
Aging tests were carried out following a protocol described in the Results and discussion section, where visible and ultraviolet wavelengths were used to study the photostability of lab-scale cells.
For aging test under visible light the solar simulator was used, while UV irradiation was produced by a Lightingcure spotlight source LC8 (model L9566, from Hamamatsu, wavelength range 240- S6

S2. Pre-oxidation of lignin via Fenton pathway
In a first exploratory step, the use of pristine kraft lignin (PL) was investigated in the attempt to produce hydrophilic membranes via epoxy ring-opening reaction with PEGDGE (Figure 1 in the main text), the latter being selected as highly hydrophilic flexible comonomer, expected to provide excellent water swelling capacity to the resulting lignin-based material. However, obtaining selfstanding thin films with a suitable thickness for use as membrane in DSSCs (≈15-20 µm) was not possible in this case, given the limited accessibility and reactivity of hydroxyl groups in PL, S3,S4 which resulted in poor mechanical integrity of the obtained membranes. In order to overcome this issue and to increase the reactivity of lignin towards PEGDGE to obtain a thin membrane, a preoxidation process for lignin was thus undertaken. To this aim, the Fenton pathway was selected because of its process versatility (it can be carried out at room-to-moderate temperature and atmospheric pressure) and given that the required reagents are readily available, easy to store and handle, safe and environmentally friendly. S5,S6 In addition, this strategy has already been successfully demonstrated on different types of lignins to tune their oxidation degree S7 or to enhance their reactivity, leading to lignin-based crosslinked systems of increased intra-and intermolecular interactions and improved mechanical and thermal response. S8 In this context, a wide range of oxidation conditions have been proposed in the literature to obtain highly hydrophilic membranes able to retain up to 45 g H2O/g polymer for application as soil remediation scaffolds. S9 However, membranes with a more moderate swelling capacity are recommended in the production of DSSCs since changes on their volume during the swelling process might affect the final performance of the device, and mechanical issues in cell sealing may arise if the membrane is too thick. Hence, soft oxidation conditions (see Experimental section for more details) were selected and their effect on the chemical and physical characteristics of the soobtained lignin were thoroughly investigated.

Free swelling capacity
The ability of the lignin-based crosslinked membranes to retain water upon soaking was evaluated in terms of their free swelling capacity (FSC). In particular, membranes obtained at different OL/PEGDGE weight ratios were immersed in water at 25 °C for 120 min and their increase in mass was assessed. As observed in Figure S6, lignin-based membranes incorporating a higher PEGDGE content (i.e., LM_0.5 and LM_0.7) were shown to exhibit the fastest swelling rate, in addition to reaching the highest equilibrium FSC value of around 2.5 gH2O/gmembrane after 2 min. Conversely, a significantly slower swelling rate was observed in LM_1 and LM_2, together with a markedly lower maximum FSC value, found in both cases to be slightly below 1.5 gH2O/gmembrane after 120 min of immersion (it is worth mentioning that 1.5 g of electrolyte is sufficient to assure a complete wetting of both the photoanode and the membrane in DSSC, as will be discussed in the following sections). S10 These different behaviors may be attributed to the twofold effect of the ethoxy-based structure of PEGDGE on the swelling response of these lignin-based membranes, which provides an increasingly more marked hydrophilic character to the material and enables a higher macromolecular network mobility, as previously discussed based on DSC analyses. However, such combined effect on FSC was no longer observed at OL/PEGDGE mass ratios below 0.7, which appears to be a threshold value for water absorption in these systems. It is worth noting that the ultimate FSC values determined for all the lignin-based membranes lie within the typical ranges of common polymeric membranes used for traditional DSSC electrolytes. S11,S12 S11 Figure S6. Free swelling capacity curves in water at 25 °C. Figure S7. Incident photon-to-current efficiency (IPCE) measurements for LM_1-based DSSC devices, in comparison with benchmark liquid-state DSSC devices.

Swelling tests in liquid electrolyte
Swelling tests were performed on the different lignin-based membranes by dipping the preweighted dry materials in the aqueous electrolyte (see Experimental section for details on composition) at 25 ºC for 7 days (time required to achieve swelling equilibrium). After this time period, the soaked membrane was rinsed off gently and weighted.
The volume fraction of polymer in the swollen membrane (ṽ) was determined gravimetrically following Equation S2: where ρs is the density of the swelling liquid (electrolyte), ρd the density of the dry membrane, mw the mass of the swollen wet membrane and md its dry mass.
The value of ρd was estimated at each lignin/PEGDGE wt./wt. ratio considering the following equation: (S3) where wLIG and wPEGDGE are the mass fraction of lignin and PEGDGE in the membrane, respectively, while ρLIG and ρPEGDGE are the density of lignin (0.95 g/cm 3 ) and the density of PEGDGE (1.14 g/cm 3 ).
The volumetric free-swelling capacity (FSCvol) was obtained after the swelling tests according to Table S1 summarizes the main characteristic values obtained from swelling tests for the different lignin/PEGDGE membrane formulations. Table S1. Volume fraction of the solid membrane in the swollen material (ṽ), volume fraction of PEGDGE in the membrane (φPEGDGE) and volumetric free-swelling capacity (FSCvol) of the membrane, all at varying lignin/PEGDGE wt./wt. ratio.  Figure S8. Reaction between PEGDGE and secondary hydroxyl groups resulting from the epoxy-ring opening reaction.

Additional electrochemical characterization
To further investigate the impact of electrolyte mass transport kinetics and macromolecular arrangement on device performance, transient photocurrent measurements under different light intensities were performed on DSSCs incorporating membranes with extreme compositions (LM_0.5 and LM_2, with highest and lowest amount of PEGDGE, respectively). As shown in Figure S9a, the last two photocurrent signals recorded at 0.8 and 1.0 sun highlight a clear mass transport limitation for both devices. This indicates that these electrolyte systems would perform better under medium and low irradiation intensity levels (i.e., below 0.6 sun), in line with the most preferable application scenarios currently proposed for commercial DSSCs that include integration in windows, indoor-environment powering and portable electronics. S13,S14 Interestingly, the transient photocurrent measurements under different light intensities revealed two different device responses for LM_0.5 and LM_2 systems. In particular, the signal associated to the device S15 incorporating the high-PEGDGE-content membrane (LM_0.5) appears to anticipate its photocurrent decay with respect to LM_2 under all illumination conditions. This behavior may be correlated to the more entangled structure likely present in LM_0.5 as a result of the increased concentration of PEGDGE, which is expected to hinder more effectively the diffusion of the redox couple between the electrodes. The same devices were also tested using open-circuit voltage decay (OCVD) measurements to investigate the effect of membrane formulation on recombination rate at the photoanode/electrolyte interface. To this end, solar cells were kept under constant irradiation and open-circuit conditions, then the illumination was switched off and the voltage decay was recorded as a function of time. Photogenerated electrons can undergo recombination under these conditions, thus lowering their population to a dark equilibrium state. In the resulting curves ( Figure S9b), no evident differences were observed between samples incorporating membranes with high (LM_0.5) and low (LM_2) PEGDGE content, indicating a negligible influence of the chemical composition of the QS electrolyte on the recombination phenomena occurring at the electrolyte/photoanode interface. Indeed, these results are in close agreement with the Voc values plotted in Figure 6b in the main text, which were found to remain rather stable irrespective of the OL/PEGDGE relative proportions. Figure S9. a) Transient photocurrent measurements at different light intensities (from left to right, irradiation intensity increases from 0.2 to 1.0 sun) for two DSSCs assembled with different lignin-based QS electrolytes. b) OCVD curves for the same devices.