Improved performance of dye-sensitized solar cells based on modified kaolin/PVDF-HFP composite gel electrolytes

Abstract

Dye-sensitized solar cells (DSSCs) fabricated with poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP) composite gel electrolytes containing variable amounts of modified kaolin were studied in this work. The kaolin was modified with silane coupling agent γ-aminopropyltriethoxysilane (KH550), and the modified kaolin (M-KL) was characterized by scanning electron microscopy (SEM) and Fourier-transform infrared spectroscopy (FTIR). The results of X-ray diffraction (XRD) and differential scanning calorimetry (DSC) indicated that the crystallinity of polymer membranes decreased with the addition of M-KL nanoparticles. The ionic conductivity and diffusion coefficient (I₃⁻) of polymer gel electrolytes (PGEs) reached optimum values of 9.452 × 10⁻³ S cm⁻¹ and 10.37 × 10⁻⁶ cm² s⁻¹ for 3 wt% M-KL, respectively, which contributed to higher short-circuit current density (J_sc) and photoelectric conversion efficiency (η) of the corresponding DSSCs. The optimum level of η reached 7.48% under the illumination of 100 mW cm⁻², an increase of 16.3% compared with the DSSC without M-KL.

---

1. Introduction

Since the first DSSC was designed by Grätzel in 1991,¹ they have attracted great attention because of their low production cost, easy fabrication and relatively high conversion efficiency. However in DSSCs, conventional liquid electrolytes have disadvantages including solvent leakage, low durability and electrode corrosion, which restrict their commercial viability. Therefore, in order to solve these problems, gel electrolytes have been used as substitutes.^2–6 Polymer gel electrolytes have been used widely for their high ionic conductivity, low vapor pressure, and excellent mechanical and chemical stability.⁷ The common polymers include polyethylene oxide (PEO),⁸ polyacrylonitrile (PAN),⁹ polymethyl methacrylate (PMMA),¹⁰ poly(vinylidene fluoride) (PVDF),¹¹ poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP),¹² and poly(vinylpyridine-co-acrylonitrile) (PVP-AN).¹³ Among them, PVDF-HFP is most often selected as the host polymer electrolyte, because it is photochemically stable and has high ionic conductivity,¹⁴ which can also improve the film-forming property of the gel electrolyte and combinability with the TiO₂ photoelectrodes.¹⁵ In addition, some novel methods have been used to prepare polymer gel electrolyte based devices. Soni et al. studied the effect of self-assembly on triiodide diffusion in water based polymer gel electrolytes, and a remarkable solar to electricity conversion efficiency and good stability of assembled DSSC was obtained using F77 based gel.¹⁶ Bella et al. reported a UV-induced polymerization method to prepare a novel DSSC based polymer gel electrolyte, which exhibited excellent efficiency and durability.¹⁷ Polymer gel electrolytes are usually prepared by incorporating liquid electrolyte into a matrix polymer, which however suffers from difficult assembly, poor mechanical strength and increasing the incidence of short circuit between electrodes. To overcome these shortcomings, polymer membranes were fabricated by immersion precipitation phase inversion,¹⁸ and then soaked into liquid electrolyte to form polymer gel electrolytes. Nanoparticles are usually added into the polymer matrix to reduce the crystallinity of the polymer, and generate a porous, mechanically stable three-dimensional network structure. Thus the anions (iodide/triiodide) can easily migrate through this porous structure, which contributes to the increase of ionic conductivity, and then enhance the photocurrent and the efficiency of DSSCs. Up to now, TiO₂,¹⁹ SiO₂,²⁰ Al₂O₃,²¹ carbon nanoparticles,²² graphite nanoparticles,²³ etc., have been effectively employed as additives to generate polymer nanocomposite electrolytes. Kaolin, Al₄[Si₄O₁₀](OH)₈, is a naturally generating inorganic polymer with a layer structure consisting of siloxane and gibbsite-like layers. The siloxane layer is composed of SiO₄ tetrahedra linked in a hexagonal array. The bases of the tetrahedra are approximately coplanar and the apical oxygen atoms are linked to a second layer containing aluminum ions and OH groups.²⁴ There are already some reports showing that kaolin could be used as nanoparticles in polymers by intercalation and modification process to improve some performances.^25–27 So, we were inspired by these reports and conjectured that kaolin could increase the ionic conductivity of polymer electrolytes similarly to other nanoparticles mentioned above. However, due to the difference of interfacial property between inorganic pristine kaolin and organic polymer, some incompatibilities would arise, and therefore organic surface modification on the kaolin is necessary. The layer thickness of kaolin is only 0.72 nm, and so it first needs to be intercalated by an inorganic salt such as potassium acetate to expand the layer space. The silane coupling agent KH550 was the used to modify the pristine expanded kaolin. KH550 has two types of functional groups, one is used to react with OH group on the surface of kaolin, and the other is combined with polymer segments.²⁸ Therefore, the silane coupling agent plays an important role in bridging kaolin and polymer. The reaction mechanism is depicted in Scheme 1. In this work, we introduced variable amounts of modified kaolin (M-KL) into PVDF-HFP gel electrolyte of DSSCs and studied the performance of the cell for different contents of M-KL. This is the first report for M-KL as an additive to the electrolyte of DSSC and an improved performance was achieved.

Scheme 1 Schematic illustration of the preparation process for the M-KL/PVDF-HFP.

2. Experimental

2.1 Materials

PVDF-HFP (M_n = 130 000) was purchased from Sigma-Aldrich. γ-Aminopropyltriethoxysilane (KH550), potassium acetate, kaolin (99%) and N,N-dimethylformamide (DMF) were purchased from Aladdin. Anhydrous lithium iodide (LiI), iodine (I₂), 1,2-dimethyl-3-propylimidazolium iodide (DMPII), 4-tert-butylpyridine (TBP), and guanidine isothiocyanate (GuSCN) were purchased from HeptaChroma, China. Acetonitrile, valeronitrile and ethyl cellulose (EC) were purchased from J&K. Fluorine-doped SnO₂ conducting glass (FTO, sheet resistivity = 15 Ω square⁻¹), cis-diisothiocyanato-bis(2,2′-bipyridyl-4,4′-dicarboxylato)ruthenium(II) bis(tetrabutylammonium) (N719), and surlyn (thickness = 60 μm) were purchased from Solaronix, Switzerland. TiO₂ nanopowder (P25) was purchased from Degussa AG, Germany.

2.2 Preparation of modified kaolin (M-KL)

The preparation of M-KL included two steps: exfoliation and modification. In the process of exfoliation, 3 g kaolin and 4.5 g potassium acetate were mixed with 3 mL deionized water and ground vigorously until the mixture became mushy. Then the mushy materials were exposed to the air for 24 h. After rinsing with anhydrous ethanol for 2 h, the mixture was dried at 80 °C for 24 h, and then the kaolin–potassium acetate intercalation complex was obtained. The obtained complex was adequately washed with deionized water and anhydrous ethanol in turn at room temperature for 30 min. After drying at 60 °C for 24 h, the exfoliated kaolin was obtained. For modification, the silane coupling agent KH550 hydrolysate was prepared first using a volume ratio of KH550, ethanol and water of 20 : 78 : 8 wt%). Then the exfoliated kaolin was sprayed with the prepared hydrolysate and refluxed at 80 °C for 12 h. Finally, the whole kaolin suspension was filtered off and washed with deionized water to obtain the modified kaolin.

2.3 Preparation of polymer gel electrolyte (PGE)

The M-KL/PVDF-HFP composite membranes were prepared by immersion precipitation phase inversion. 0.5 g PVDF-HFP was dissolved in 4.5 g DMF solvent under continuous stirring at 80 °C. The pre-dried M-KL nanoparticles were slowly introduced into the polymer solution with different weight percent (0–5 wt% relative to PVDF-HFP). The mixture solution was stirred continuously for about 7 h until the nanoparticles were dispersed thoroughly. Then, the obtained viscous polymer solutions were slowly poured into the self-assembly mold, which could be used to control the thickness of polymer membranes. A few minutes later, the mold was immersed into the deionized water to conduct the phase inversion process. Finally, the wet membranes were dried under vacuum at room temperature for 6 h to obtain the M-KL/PVDF-HFP composite membranes. The obtained polymer membranes were soaked in the liquid electrolyte containing 0.5 M LiI, 0.05 M I₂, 0.1 M DMPII, 0.5 M 4-TBP and 0.1 M GuSCN in acetonitrile–valeronitrile (85 : 15 volume ratio) for 24 h.

2.4 Photoanode preparation and DSSCs assembly

1 g TiO₂ nanopowder and 0.5 g EC were added to a mixture of 30 mL ethanol and 4.06 g terpineol. The mixture was milled for 4 h in a ball mill to obtain a suspension slurry, and then TiO₂ pastes were obtained after evaporating the solvent. TiO₂ paste was coated onto FTO conductive glass using a doctor blade and was sintered at 500 °C for 15 min. After cooling, the films were immersed in TiCl₄ solution at 70 °C for 45 min and sintered at 500 °C for 30 min. Then, it was sensitized with 0.3 mM N719 dye in ethanol solution for 22 h at room temperature to obtain the photoanode. The DSSC was fabricated by sandwiching the polymer gel electrolyte composite membrane between a photoanode and a platinum counter electrode, and heat-sealing the perimeter of the cell using surlyn while pressurizing it.

2.5 Characterization and measurements

The surface morphology of U-KL and M-KL were examined using a field emission scanning electron microscope (SEM, S-4800, Hitachi Ltd, Japan). The Fourier-transform infrared (FTIR) spectra of U-KL and M-KL were obtained using a Nicolet 380 (Thermo Fisher, USA). X-Ray diffraction (XRD) analysis of polymer membranes was conducted using an X-ray diffractometer (D/MAX-2500, Japan) with Cu-Kα radiation at a scanning rate of 5° min⁻¹ over a 2θ interval from 3 to 50°. Differential scanning calorimetry (DSC, Q20, M/s, TA Instruments, USA) measurements of polymer membranes were carried out at a heating rate of 5 °C min⁻¹ in the temperature range of 25 to 200 °C. The morphology of polymer membranes was examined using SEM (S-4800, Hitachi Ltd, Japan). The porosity of polymer membranes was measured by the n-butanol uptake method.²⁹ The membrane was soaked in 1-butanol for 2 h and the weight of the membrane was measured before and after absorption immersion. The porosity was calculated using eqn (1):

Porosity = (M_t − M₀)/ρV × 100% (1)

where M_t and M₀ are the weight of the wet and dry membrane, respectively. V is the apparent volume of the membrane and ρ is the density of 1-butanol. In order to measure the electrolyte uptake, polymer membrane was immersed in liquid electrolyte for 2 h. The uptake was determined by using eqn (2):

Electrolyte uptake = (M_wet − M_dry)/M_dry × 100% (2)

where M_dry and M_wet are the weight of the membrane before and after soaking in the liquid electrolyte, respectively.³⁰ AC impedance (Nyquist plots) and linear sweep voltammetry (LSV) curves of polymer gel electrolytes were measured by a CHI660 electrochemical workstation after PGEs were sandwiched between two Pt electrodes. The former was measured over the frequency range of 1 Hz to 100 kHz with amplitude of 10 mV, the latter was measured with a scan rate of 10 mV s⁻¹ from −1 to 1 V. The area (A) of the Pt electrode was 0.16 cm² and the thickness (L) of the membrane was controlled to 0.01 cm by the self-assembly mold mentioned above. Photocurrent–photovoltage (I–V) curves of DSSC were measured under simulated AM 1.5 irradiation (100 mW cm⁻²) and the photocurrent–voltage (J–V) characteristics were recorded on a Keithley 2400 Source meter (solar AAA simulator, oriel China, calibrated using a standard crystalline silicon solar cell). EIS measurements of DSSCs were performed using a CHI660 electrochemical workstation over a frequency range of 1 mHz to 100 kHz at open circuit under 100 mW cm⁻².

3. Results and discussion

3.1 Surface morphology of U-KL and M-KL

The scanning electron microscopic (SEM) images of U-KL and M-KL are shown in Fig. 1. Fig. 1 shows the surface morphology of unmodified kaolin (U-KL) and modified kaolin (M-KL) at magnifications of ×2000 and ×30 000, respectively. From Fig. 1(a), it is seen that the U-KL particles exhibit irregular granules with dimensions from 1 to 10 μm in diameter, as well as obvious agglomeration. Moreover, there is apparent kaolin lamella structure as shown in Fig. 1(a′). After surface treatment with silane coupling agent KH550, agglomeration between the kaolin particles is weakened, and the modified kaolin (M-KL) is distributed more equably and exhibits smaller dimensions of under 1 μm in diameter, as shown in Fig. 1(b) and (b′). These observations suggest that M-KL may show better dispersibility in PVDF-HFP than U-KL.

Fig. 1 SEM images of (a) U-KL (×2k), (a′) U-KL (×30k), (b) M-KL (×2k) and (b′) M-KL (×30k).

3.2 FTIR analysis of U-KL and M-KL

Fig. 2 shows the FTIR spectra of pristine kaolin (U-KL) and silane modified kaolin platelets (M-KL). In the FTIR spectra of U-KL, the peaks in the regions of 3600–3200 and 1640 cm⁻¹ correspond to –OH stretching and bending vibrational absorptions, which suggests the existence of entrapped water molecules in the interlayer galleries of kaolin, and may be due to that water has not been removed completely (used to wash out the kaolin–potassium acetate intercalation complex in the processing of kaolinic exfoliation). U-KL should have presented the inner surface –OH stretching of Si–OH and the outer surface –OH stretching of Al–OH of the kaolin structure at around 3690 and 3620 cm⁻¹,³¹ respectively, but the –OH stretching vibration absorption of the residual water as mentioned above, may have obscured these characteristic peaks. The Si–O bending vibration peaks are assigned between 1095 and 472 cm⁻¹, and the peaks at 810 and 744 cm⁻¹ correspond to the Si–O–Si symmetric stretching vibration peak and Al–O stretching vibration peak, respectively, additionally, the Si–O–Al bonds stretching vibration is observed at 571 cm⁻¹.³²

Fig. 2 FTIR spectra of U-KL and M-KL.

After treating with KH550, the FTIR spectra exhibit new peaks at 2923 and 2850 cm⁻¹, which are due to the –CH₂ and –CH structure in the silane coupling agent. Moreover, after KH550 modification, the intensity of the –OH peaks at 3600–3200 and 1640 cm⁻¹ are reduced, which indicates that silane moieties have been attached on the surface of kaolin platelets through condensation reaction. The peak observed at 1564 cm⁻¹ corresponds to –NH₂ bending vibration of the functional amino group present in the silane coupling agent.³³ These indicate that kaolin has been modified by the silane coupling agent successfully.

3.3 XRD studies of M-KL/PVDF-HFP composite membranes

The variation in crystallinity of PVDF-HFP membranes as a result of the addition of different amounts of M-KL was studied by X-ray diffraction (XRD). Fig. 3 shows the XRD patterns of PVDF-HFP membranes, containing 0, 1, 2, 3, 4 and 5 wt% of M-KL, and also shows the reference XRD patterns for pure M-KL. The diffraction peaks at 18.2, 20 and 26.5° are observed as the characteristic peaks of crystalline phases of PVDF-HFP in Fig. 3(a).³⁴ According to Fig. 3(a)–(f), the diffraction intensities of M-KL/PVDF-HFP first decrease and then increase with the increase of amount of the M-KL, and reaches its minimum value when PVDF-HFP contains 3 wt% M-KL. At low concentration levels of M-KL (<3 wt%), M-KL nanoparticles disturb the crystalline regions and decrease the crystalline phase of PVDF-HFP. This behavior is due to the large surface area of M-KL nanoparticles, which inhibits the recrystallization of the PVDF-HFP host. The addition of M-KL nanoparticles could lead to a decrease in the intermolecular interaction between polymer chains, and hence increases the amorphous region.³⁵ At higher filler contents (>3 wt%), however, the agglomeration of M-KL particles leads to bigger particles and lower surface area. This mechanism assumes that the surface area of filler is the key factor in determining the crystallinity.³⁶ High concentration of M-KL also leads to well-defined crystallite regions. This reasoning is in line with the finding of Rajendran et al.³⁷ Further, beyond this optimum concentration, the added filler particles may catalyze aggregation of polymer chains and thus increase the rate of recrystallization processes.³⁸

Fig. 3 XRD patterns of (a) PVDF-HFP polymer and of M-KL/PVDF-HFP composite with (b) 1 wt%, (c) 2 wt%, (d) 3 wt%, (e) 4 wt% and (f) 5 wt% of M-KL, and (g) M-KL nanopowder.

The major peaks at 16.4, 21.3 and 26° for crystalline phases of pure M-KL are observed in Fig. 3(g).³⁹ The diffraction peaks of M-KL have lower intensity, which indicate that they mainly exist in amorphous form. This may be due to that the crystalline regions of M-KL decrease after calcination.⁴⁰ At the same time, the concentration of M-KL relative to PVDF-HFP is low. Thus, the characteristic peaks of M-KL can not be observed in the XRD patterns of M-KL/PVDF-HFP, as shown in Fig. 3(b)–(f).

3.4 DSC studies of M-KL/PVDF-HFP composite membranes

Differential scanning calorimetry (DSC) thermograms of M-KL/PVDF-HFP composite membranes in a wide range of temperatures from 25 to 200 °C are shown in Fig. 4. From the DSC thermograms, the melting temperatures of M-KL/PVDF-HFP (T_m) and melting enthalpy of M-KL/PVDF-HFP (ΔH_m) are determined. Here, the value of T_m and ΔH_m were obtained with the help of TA Universal Analysis software (USA).

Fig. 4 DSC thermograms of M-KL/PVDF-HFP composite membranes containing (a) 0 wt%, (b) 1 wt%, (c) 2 wt%, (d) 3 wt%, (e) 4 wt% and (f) 5 wt% M-KL.

At the melting temperature (T_m), polymers begin to melt macroscopically and polymer chains start to move at the microscopic level, so intermolecular interaction between polymer chains is an important factor to determine the T_m. The incorporated nanoparticles act as solid plasticizers and significantly reduce the cohesive forces between the long range polymer chains of macromolecular polymers.⁴¹ M-KL nanoparticles interact with PVDF-HFP and reduce the intermolecular interaction between polymer chains, thus T_m decreases with the filler added at a lower concentration (<3 wt%). When the content of M-KL in PVDF-HFP is more than 3 wt%, however, the value of T_m is increased with the elevated amount of the M-KL. The reason is that excess M-KL may act as a crystallization nucleating agent and catalyze reaggregation of polymer chains,⁴² so the polymers begin to melt at higher T_m. The value of T_m reaches its minimum (135.89 °C) when the content of M-KL is 3 wt%.

It is necessary to study the crystallinity (χ%) of M-KL/PVDF-HFP, because this is the key factor to influence the ionic conductivity of polymer electrolyte membranes. The relative percentage of crystallinity χ% was calculated using eqn (3):

χ% = ΔH_m/ΔH⁰_m × 100% (3)

where, ΔH⁰_m is the heat of fusion of pure PVDF, which is equal to 105 J g⁻¹ when the material is assumed to be 100% crystalline,⁴³ ΔH_m is heat of fusion of M-KL/PVDF-HFP membranes. The value of melting enthalpy (ΔH_m) and percentage of crystallinity (χ%) and their changes with composition are shown in Table 1. The values in thermal parameters for M-KL/PVDF-HFP membranes are lower than for bare PVDF-HFP, which may be due to the presence of organic functional groups such as CH₂, NH₂ and Si–O–Si, which may interrupt the polymer chains conformation and coordinate to the salt systems. Such new interactions formed by KH550 molecules will destroy the structural conformation of polymer. The addition of M-KL (<3 wt%) nanoparticles results in broadening of the melting peak, which suggests that there is an disturbance in the crystallization of PVDF-HFP and an enhancement in the free volume regions. When the amount of M-KL is over 3 wt%, the percentage of crystallinity is increased again, because of the recrystallization processes in PVDF-HFP. The minimum value of χ% obtained at 3 wt% is decreased by 21.9% compared with pure PVDF-HFP. Its more amorphous nature may help to increase electrolyte uptake and ionic conductivity. The change of crystallinity according to DSC is consistent with the results of XRD studies.

Table 1 Thermal parameters derived from DSC for PVDF-HFP membranes (PMs) containing different amounts of M-KL

PM	ΔHm/J g^−1	Tm/°C	χ (%)
PVDF-HFP	23.78	140.68	22.65
PVDF-HFP/1 wt% M-KL	22.03	139.07	20.98
PVDF-HFP/2 wt% M-KL	20.52	138.30	19.54
PVDF-HFP/3 wt% M-KL	18.57	135.89	17.69
PVDF-HFP/4 wt% M-KL	19.85	136.71	18.90
PVDF-HFP/5 wt% M-KL	20.69	138.69	19.71

---

In addition, there are few small peaks observed around 50 and 90 °C which is probably due to the crystal phase transition of the ferroelectric to the paraelectric phase in the PVDF copolymer corresponding to the Curie temperature.⁴⁴

3.5 Morphology of M-KL/PVDF-HFP composite membranes

Fig. 5 shows the surface and cross-section morphology of bare PVDF-HFP membrane and M-KL/PVDF-HFP composite membranes. As shown in Fig. 5(a) and (b), there are only small diameter sparsely distributed pores on the surface of bare PVDF-HFP membrane, and there is an obvious three-dimensional network structure without finger-like structures in cross-section, which suggests that the formed pores are only superficial and are not interconnected. After M-KL was embedded into PVDF-HFP, there are obvious changes on the surface and cross-section of polymer membranes. Fig. 5(c)–(h) show the increased pore size and quantities on the surface and finger-like interconnected structure in the cross-section with the incorporation of M-KL. With increased amounts of fillers, changes of the characteristics of the pores are not apparent, but cross-sectional morphology changed obviously, with finger-like structures gradually becoming wider, bigger and more interconnected. The polymer membranes were produced by immersion precipitation phase inversion, and in the process of phase inversion M-KL had an effect on delaying the time of phase separation of polymer, and promoting a sufficient exchange process between the coagulation bath and organic solvent DMF, resulting in a large number of interconnected pores in the polymer membranes.⁴⁵ These pores can provide new channels for electrolyte ions transport in PVDF-HFP. However, with a further increase of the amount of M-KL (>3 wt%), there are agglomerated M-KL nanoparticles on the surface of polymer membranes, and the phenomenon becomes more obvious when the amount of M-KL reaches 5 wt%, as shown in Fig. 5(i) and (k). Due to the high concentration of M-KL, they are not dispersed in the polymer thoroughly, and some remain on the surface of membranes and even cover existing pores, resulting in the decrease of ionic diffusion channels. Furthermore, the cross-sectional morphology of polymer membranes incorporated with excess M-KL shown in Fig. 5(j) and (l) further illustrate that the residual M-KL has blocked the diffusion and exchange between organic solvent DMF and coagulation bath water in the process of phase inversion, so that the obtained pores are not interconnected thoroughly.

Fig. 5 Surface (left) and cross-section SEM images (right) of M-KL/PVDF-HFP composite membranes containing (a, b) 0 wt%, (c, d) 1 wt%, (e, f) 2 wt%, (g, h) 3 wt%, (i, j) 4 wt% and (k, l) 5 wt% M-KL

The porosity and electrolyte uptake measurements of M-KL/PVDF-HFP composite membranes are shown in Fig. 6. It is observed that the porosity of polymer membranes increases first and then decreases with the increase of M-KL content, and reaches a maximum of about 76% for 3 wt% M-KL. From observation of Fig. 5, it can be easily seen that pores in the polymer membranes become more numerous and larger with the increase of M-KL content, and then some pores are blocked by self-agglomerative M-KL nanoparticles with further increase in M-KL content, resulting in the reduction of the porosity. Similarly, the electrolyte uptake of polymer membranes increases from about 80 to 276% as the M-KL content is increased from 0 to 3 wt% and thereafter it decreases. This is mainly related to the change of porosity, and the large interconnected pores could provide space for storing liquid electrolyte. The electrolyte uptake is not only related with the porosity of the membranes, but also to the crystallinity of the M-KL/PVDF-HFP, because the amorphous regions could also offer storage for liquid electrolyte. From XRD and DSC studies, the changes of the crystallinity are also in accordance with the electrolyte uptake of polymer membranes. In addition, the enhancement in electrolyte uptake with the addition of fillers may have some correlation with M-KL nanoparticle content, given their higher affinity toward liquid electrolytes and ionic liquid solvents.⁴⁶

Fig. 6 Effects of M-KL content in the M-KL/PVDF-HFP composite membranes on porosity and electrolyte uptake.

3.6 Electrochemical properties of the polymer gel electrolytes (PGEs)

The typical Nyquist plots of polymer gel electrolytes (PGEs) containing variable amounts of M-KL at room temperature are illustrated in Fig. 7. The Nyquist plots show typical impedance spectra of a polymer electrolyte layer sandwiched between two blocking electrodes.⁴⁷ The bulk resistance of the electrolytes R_b was obtained from the intercept of an inclined straight line with the real axis in the Nyquist plot. After the bulk resistance R_b was determined, the ionic conductivity σ of the PGEs can be calculated following eqn (4):

σ = L/R_bA (4)

where L is the thickness of the polymer electrolyte membrane and A is the area of the electrode in contact with electrolyte membrane.

Fig. 7 Nyquist plots of PGEs containing variational amounts of M-KL.

Table 2 shows the values and variation of the ionic conductivity of PGEs containing different amounts of M-KL. The ionic conductivity of PGEs increases with the addition of M-KL and reaches a maximum value of 9.452 × 10⁻³ S cm⁻¹ for 3 wt%, an increase of 105.3% compared with 4.603 × 10⁻³ S cm⁻¹ for pure PVDF-HFP gel electrolyte. XRD and DSC studies have confirmed that with the addition of M-KL, the nanoparticles interacting with polymer chain, increased the localized amorphous regions, and promoted the migration ability of electrolyte ions along the polymer segments. Moreover, the addition of M-KL into the polymer matrix generated a three-dimensional, mechanically stable and porous network structure, which facilitated the anions (iodide/triiodide) to move into the porous structures within the electrolyte more easily. In addition, the interconnecting pores and large free volume in the polymer help to uptake more electrolytes, which contributes to the increase of ionic conductivity. On the contrary, the ionic conductivity of PGEs decreases if the addition of M-KL exceeds the optimum level (>3 wt%). As studied in XRD and DSC, high concentration of M-KL could catalyze reaggregation of polymer chains, and increased the crystalline phase of polymer electrolytes, which blocked ion transfer in the electrolyte. The porosity and uptake of PGEs also decreased when adding excess M-KL, resulting in the decrease of mobile charge carriers in the electrolyte. In addition, high concentrations of nanoparticles agglomerate more easily, and the miscibility between PVDF-HFP and M-KL become worse, therefore the process of phase separation is aggravated,⁴⁸ which is detrimental to the ionic conductivity.

Table 2 Electrochemical parameters of PGEs containing variational amounts of M-KL

PGE	Rb/Ω	σ/10^−3 S cm^−1	Ilim/mA	Dss/10^−6 cm^2 s^−1
0 wt% M-KL	13.66	4.603	0.908	4.931
1 wt% M-KL	11.52	5.458	1.212	6.582
2 wt% M-KL	8.966	7.013	1.635	8.879
3 wt% M-KL	6.652	9.452	1.909	10.37
4 wt% M-KL	7.831	8.029	1.741	9.455
5 wt% M-KL	9.964	6.310	1.570	8.526

---

In this study, the concentration of iodide (I⁻) is higher than triiodide (I₃⁻), thus the transfer of I₃⁻ plays a decisive role on ionic migration. Fig. 8 shows the linear sweep voltammetry (LSV) curves for PGEs containing different amounts of M-KL at room temperature. From the cathodic steady-state limiting currents (I_lim), which could be derived from LSV curves, the diffusion coefficient (D_ss) of I₃⁻ could be obtained using eqn (5):⁴⁹

D_ss = I_limL/2nAFc (5)

where L is the thickness of the polymer electrolyte membrane, n is the number of electrons involved in the electrochemical reaction, A is the area of the electrode, F is the Faraday constant, c is the concentration of I₃⁻.

Fig. 8 LSV curves of PGEs containing variational amounts of M-KL.

As well as the ionic conductivity, the ionic diffusion coefficient also increases with adding M-KL nanoparticles and reaches an optimum value of 10.37 × 10⁻⁶ cm² s⁻¹ for 3 wt% M-KL in the PGEs, increasing by 110.3% (4.931 × 10⁻⁶ cm² s⁻¹) compared with PGEs without M-KL, as shown in Table 2. There are two pathways for I⁻ and I₃⁻ to migrate in the PGEs. One is migrating through interconnected pores; the other is migrating through amorphous regions in the PGEs. Incorporation of M-KL creates large interconnected pores and free volume space between the polymer chains, which provides new pathways for charge carriers to move. The high porosity and uptake also contribute to the increase of the ionic diffusion coefficient. In addition, the negative charges in the M-KL nanoparticles, such as –O⁻, weaken the attractive force between the cations (Li⁺) and anions (I⁻) present in the electrolyte salt, so that the ionic diffusion coefficient is increased. The value of the diffusion coefficient of I₃⁻ decreases when the M-KL content exceeds the optimum level. Higher concentration of M-KL causes aggregation of high molecule weight polymer chains, resulting in an increase in crystallinity, and hindering ion transport in the PGEs. As Fig. 5 shows, high content of M-KL causes agglomeration on the surface of polymer membranes more easily and covers existing pores, leading to the decrease of ion transport channels, which is unfavorable for the ionic diffusion coefficient. Moreover, excess filler could not disperse in the PGEs effectively, increasing the viscosity of electrolyte, so reducing ionic movement, causing a decrease of the diffusion coefficient.

3.7 Photovoltaic performance of DSSCs

The J–V curves of the DSSCs with the gel electrolyte based on the PVDF-HFP membrane containing varied amounts of M-KL at a light intensity of 100 mW cm⁻² are presented in Fig. 9. The corresponding photovoltaic parameters of all DSSCs are summarized in Table 3, which show that the photovoltaic performance of all DSSCs with M-KL are superior to that without M-KL. The short-circuit current density (J_sc) first increases and then decreases with increasing M-KL content, and reaches an optimum value of 13.51 mA cm⁻² for 3 wt%. The value of J_sc mainly depends on the electron injection rate from the excited state dye (D⁺) to the TiO₂ conduction band and the charge recombination rate of conduction band electrons of dye sensitized TiO₂ and I₃⁻. In this study, the charge recombination rate is related to the concentration of polymer, which is a constant value, so it has no effect on J_sc. Incorporated M-KL nanoparticles decrease the crystallinity of polymer membranes, and the large amorphous regions are in favor of the migration of charge carriers, which contribute to accelerating the process of the reduction of oxidation state of dye and regenerating the normal state of dye, resulting in an increase of J_sc. Beyond the optimum content (>3 wt%) J_sc decreases because of the decrease of diffusion coefficient of I₃⁻, which is due to the increase of crystallinity of polymer membranes and hindrance from the reaggregation of polymer molecule chains. However, the open circuit voltage (V_oc) of DSSCs keeps nearly constant. The value of V_oc is the difference between Fermi level of TiO₂ and redox potential of electrolyte, which is influenced mainly by the molar ratio of I⁻/I₃⁻. In this case, I⁻/I₃⁻ molar ratio is invariant, thus V_oc shows a constant value around 0.8 V. Photoelectric conversion efficiency (η) is a significant index in measuring the photovoltaic performance of DSSCs. From Table 3, it is observed that with the addition of 3 wt% M-KL, η (as defined by eqn (6)⁵⁰) reaches its optimum level, increasing by 16.3% from 6.43 without M-KL to 7.48.

η = J_scV_ocFF/P_in (6)

In this equation, FF is the fill factor and P_in is the incident light intensity, which is constant in this study. The addition of M-KL nanoparticles has little or no effect on V_oc, FF or P_in, respectively, thus, it is obvious that the value of η is only determined by J_sc, that is to say, J_sc is the main factor that promoted the increase of η.

Fig. 9 J–V curves of DSSCs containing gel electrolytes based on the PVDF-HFP membrane containing varying amounts of M-KL.

Table 3 Photovoltaic parameters from J–V curves and simulated data from the EIS of DSSCs with varying amounts of M-KL in the PGEs

PGE	Jsc/mA cm^−2	Voc/V	FF	η (%)	R1/Ω	R2/Ω	Rdiff/Ω
0 wt% M-KL	11.30	0.81	70.06	6.43	3.719	29.15	24.29
1 wt% M-KL	11.84	0.80	70.42	6.71	3.782	27.86	22.73
2 wt% M-KL	12.73	0.81	69.04	7.13	3.598	25.24	20.05
3 wt% M-KL	13.51	0.81	68.54	7.48	3.684	22.35	16.48
4 wt% M-KL	12.87	0.80	70.64	7.29	3.545	24.25	19.10
5 wt% M-KL	12.00	0.81	70.12	7.02	3.580	26.56	21.36

---

3.8 EIS analysis of DSSCs

The charge transfer resistances of the DSSCs were analyzed using EIS measurements. Nyquist plots of DSSCs with PVDF-HFP composite membranes containing different contents of M-KL and the equivalent circuit model of the DSSCs are shown in Fig. 10. As shown, the spectra exhibit three distinct semicircles, which can be denoted in terms of R₁, R₂ and R_diff, which correspond to the charge transfer resistance at the interface of FTO/TiO₂ and counter electrode/electrolyte, the charge transfer resistance R_t in the TiO₂ membrane and charge recombination resistance R_ct at the interface of TiO₂/electrolyte, and the diffusion resistance of I⁻/I₃⁻ redox couple or Warburg resistance within the electrolyte, respectively.^51,52 R_s is denoted as the external circuit resistance whose value is determined from the intercept of the plots (left) with the real axis. The simulated data of all resistances are summarized in Table 3.

Fig. 10 EIS spectra of the DSSCs with the gel electrolyte based on the PVDF-HFP membrane containing variational amounts of M-KL.

As shown, the value of R₁ is almost same in all cases, because additional M-KL has little effect on the interface of FTO/TiO₂, counter electrode/electrolyte and external circuit. R₂ and R_diff, however, change substantially with the addition of M-KL. In the DSSCs based polymer gel electrolyte, semiconductor TiO₂ nanoparticles are coated with polymer gelatinizer, blocking the electron injection and charge recombination process. In this study, the amount of polymer gelatinizer PVDF-HFP is constant, thus the charge recombination rate is the same in the DSSCs, and charge recombination resistance R_ct is invariable. At low concentration of M-KL, with increasing content of M-KL, the crystallinity of polymer is decreased and the polymer segments move more easily, at the same time, the interconnected pores become larger. All these factors contribute to increasing the charge carriers’ diffusion rate in the polymer electrolyte, resulting in increasing the regeneration rate of oxidation state of dye and the electron injection rate from the excited state of dye (D⁺) to the TiO₂ conduction band, which diminish the value of charge transfer resistance R_t. Thus, R₂ decreases with adding an appropriate amount of M-KL. When adding further M-KL, increased viscosity of polymer and reaggregated polymer chains hinder the migration of electrolyte ions, and M-KL may interact with charge carriers, which block the charge transfer process, resulting in the increase of R₂. There is no doubt that R_diff changes with the variation of M-KL, and its variation behaviour is similar to R₂. The smallest R₂ and R_diff of the DSSCs are obtained in the case of the cell with 3 wt% M-KL in the PGEs. This result indicates that the addition of an appropriate amount of M-KL not only reduces the diffusion resistance of electrolyte but also improves the charge transfer reaction at the TiO₂ electrode, resulting in the highest photoelectric conversion efficiency in the DSSC with M-KL/PVDF-HFP composite membrane containing 3 wt% M-KL.

3.9 OCVD measurements of DSSCs

The open-circuit voltage decay (OCVD) method was applied to measure carrier lifetime in the DSSCs. The V_oc decay curve and the electron lifetime (τ_n) curves in the DSSCs with PVDF-HFP composite membranes containing different contents of M-KL are shown in Fig. 11(a) and (b), respectively. The value of τ_n can be calculated from eqn (7):⁵³

τ_n = −k_BT/e(dV_oc/dt)⁻¹ (7)

where k_B is the Boltzmann constant, T is the temperature, e is the elementary charge, and the value of dV_oc/dt can be obtained from the V_oc decay curve.

Fig. 11 (a) Voltage decay curve and (b) electron lifetime curve of DSSCs with gel electrolytes based on the PVDF-HFP membrane containing varying amounts of M-KL.

It is obvious that the V_oc decay curve among all DSSCs almost coincide and all DSSCs have the same electron lifetime. That is to say, the added M-KL has no obvious effect on carrier lifetime. The electron lifetime is determined by the charge recombination rate which is affected by the concentration of gelatinizer PVDF-HFP. The amount of gelatinizer is invariable in the measurement. So these DSSCs have the same charge recombination rate, electron lifetime, R_ct and V_oc. This conclusion is in accordance with previous analyses.

4. Conclusions

M-KL/PVDF-HFP composite membranes containing various amounts of M-KL nanoparticles were characterized through a series of measurements, which was shown to improve the performance of the corresponding DSSCs. The addition of appropriate M-KL not only reduced the crystallinity but also increased the porosity of polymer membranes, which promoted the diffusion rate of charge carriers, resulting in enhancement in ionic conductivity (σ) and diffusion coefficient (D_ss). The DSSC assembled with the composite membrane containing 3 wt% M-KL nanoparticles exhibited a remarkably high photoelectric conversion efficiency (η) of 7.48% at 100 mW cm⁻², an increase of 16.3% compared with the DSSC without M-KL. These results provide a new avenue for improving the efficiency of DSSC with other appropriate nanoclays or their hybrid nanoparticles, which is in our plans for future work.

Acknowledgements

This work was supported by the Key project of the National Natural Science Foundation of China for international academic exchanges (No. 21476162), and by National Key Technologies R&D Program (No. 51020105010).

References

1. B. O’Regen and M. Grätzel, Nature, 1991, 353, 737–740.
2. P. Wang, S. M. Zakeeruddin, I. Exnar and M. Grätzel, Chem. Commun., 2002, 2972–2973.
3. W. Kubo, S. Kambe, S. Nakade, T. Kitamura, K. Hanabusa, Y. Wada and S. Yanagida, J. Phys. Chem. B, 2003, 107, 4374–4381.
4. P. Wang, S. M. Zakeeruddin, R. Humphry-Baker and M. Grätzel, Chem. Mater., 2004, 16, 2694–2696.
5. B. O’Regan, F. Lenzmann, R. Muis and J. Wienke, Chem. Mater., 2002, 14, 5023–5029.
6. W. U. Huynh, J. J. Dittmer and A. P. Alivisatos, Science, 2002, 295, 2425–2427.
7. A. L. Sharma and A. K. Thakur, Ionics, 2010, 16, 339–350.
8. Y. J. Ren, Z. C. Zhang, S. B. Fang, M. Z. Yang and S. M. Cai, Sol. Energy Mater. Sol. Cells, 2002, 71, 253–259.
9. M. A. K. L. Dissanayake, C. A. Thotawatthage, G. K. R. Senadeera, T. M. W. J. Bandara, W. J. M. J. Jayasundara and B. E. Mellander, J. Appl. Electrochem., 2013, 43, 891–901.
10. H. X. Yang, M. L. Huang, J. H. Wu, Z. Lan, S. C. Hao and J. M. Lin, Mater. Chem. Phys., 2008, 110, 38–42.
11. A. K. Arof, M. F. Aziz, M. M. Noor, M. A. Careem, L. R. A. K. Bandara, C. A. Thotawatthage, W. N. S. Rupasinghe and M. A. K. L. Dissanayake, Int. J. Hydrogen Energy, 2014, 39, 2929–2935.
12. T. Asano, T. Kubo and Y. Nishikitani, J. Photochem. Photobiol., A, 2004, 164, 111–115.
13. M. Y. Li, S. J. Feng, S. B. Fang, X. R. Xiao, X. P. Li, X. W. Zhou and Y. Lin, Electrochim. Acta, 2007, 52, 4858–4863.
14. D. Saikia, C. C. Han and Y. W. Chen-Yang, J. Power Sources, 2008, 185, 570–576.
15. H. S. Yang, O. A. Ileperuma, M. Shimomura and K. Murakami, Sol. Energy Mater. Sol. Cells, 2009, 93, 1083–1086.
16. S. S. Soni, K. B. Fadadu, R. L. Vekariya, J. Debgupta, K. D. Patel, A. Gibaud and V. K. Aswal, J. Colloid Interface Sci., 2014, 425, 110–117.
17. F. Bella, D. Pugliese, J. R. Nair, A. Sacco, S. Bianco, C. Gerbaldi, C. Baroloc and R. Bongiovanni, Phys. Chem. Chem. Phys., 2013, 15, 3706–3711.
18. H. Strathmann and K. Koch, Desalination, 1977, 21, 241–255.
19. C. L. Chen, T. W. Chang, S. C. Su, H. S. Teng and Y. L. Lee, J. Power Sources, 2014, 247, 406–411.
20. P. Wang, S. M. Zakeeruddin and M. Grätzel, J. Fluorine Chem., 2004, 125, 1241–1245.
21. N. Jeon and D. W. Kim, J. Nanosci. Nanotechnol., 2013, 13, 7955–7958.
22. J. Zhang, H. W. Han, S. J. Wu, S. Xu, Y. Yang, C. H. Zhou and X. Z. Zhao, Solid State Ionics, 2007, 178, 1595–1601.
23. Y. L. Lee, Y. J. Shen and Y. M. Yang, Nanotechnology, 2008, 19, 455201.
24. R. L. Frost, J. Kristof, E. Mako and K. T. Kloprogge, Langmuir, 2000, 16, 7421–7428.
25. M. Alexandre and P. Dubois, Mater. Sci. Eng., R, 2000, 28, 1–63.
26. M. B. A. Bakar, Y. W. Leong, A. Ariffin and Z. A. M. Ishak, J. Appl. Polym. Sci., 2008, 110, 2770–2779.
27. X. Wang, Q. Y. Wu and Z. N. Qi, Polym. Int., 2003, 52, 1078–1082.
28. L. Y. Xie, X. Y. Huang, K. Yang, S. T. Li and P. K. Jiang, J. Mater. Chem. A, 2014, 2, 5244–5251.
29. S. W. Choi, J. R. Kim, S. M. Jo, W. S. Lee and Y. R. Kim, J. Electrochem. Soc., 2005, 152, A989–A995.
30. Y. H. Ding, P. Zhang, Z. L. Long, Y. Jiang, F. Xu and W. Di, J. Membr. Sci., 2009, 329, 56–59.
31. W. S. Chow and S. S. Neoh, J. Appl. Polym. Sci., 2009, 114, 3967–3975.
32. Y. P. Wang, X. H. Gao, R. M. Wang, H. G. Liu, C. Yang and Y. B. Xiong, React. Funct. Polym., 2008, 68, 1170–1177.
33. S. J. Park, B. J. Kim, D. I. Seo, K. Y. Rhee and Y. Y. Lyu, Mater. Sci. Eng., A, 2009, 526, 74–78.
34. A. M. Stephan, K. S. Nahm, T. P. Kumar, M. A. Kulandainathan, G. Ravi and J. Wilson, J. Power Sources, 2006, 159, 1316–1321.
35. X. D. Ma, X. F. Qian, J. Yin and Z. K. Zhu, J. Mater. Chem., 2002, 12, 663–666.
36. M. Forsyth, D. R. MacFarlane, A. Best, J. Adebahr, P. Jacobsson and A. J. Hill, Solid State Ionics, 2002, 147, 203–211.
37. S. Rajendran and T. Uma, Mater. Lett., 2000, 44, 208–214.
38. F. Capuano, F. Croce and B. Scrosati, J. Electrochem. Soc., 1991, 138, 1918–1922.
39. X. N. Liu, N. Hu, H. T. Zhang, C. Y. Ha and Z. Z. Yang, Sci. China, Ser. B: Chem., 2005, 48, 326–333.
40. M. Gabor, M. Toth, J. Kristof and G. K. Komaromi-Hiller, Clays Clay Miner., 1995, 43, 223–228.
41. J. Zhang, Y. Yang, S. J. Wu, S. Xu, C. H. Zhou, H. Hu, B. L. Chen, X. D. Xiong, B. Sebo, H. W. Han and X. Z. Zhao, Nanotechnology, 2008, 19, 245202.
42. D. N. Bikiaris, N. P. Nianias, E. G. Karagiannidou and A. Docoslis, Polym. Degrad. Stab., 2012, 97, 2077–2089.
43. K. Nakagawa and Y. Ishida, J. Polym. Sci., Polym. Phys. Ed., 1973, 11, 1503–1533.
44. H. Kodama, Y. Takahashi and T. Furukawa, Jpn. J. Appl. Phys., 1999, 38, 3589–3595.
45. A. Bottino, G. Camera-Roda, G. Capannelli and S. Munari, J. Membr. Sci., 1991, 57, 1–20.
46. Y. Ran, Z. L. Yin, Z. Y. Ding, H. J. Guo and J. Yang, Ionics, 2013, 19, 757–762.
47. G. B. Appetecchi and S. Passerini, Electrochim. Acta, 2000, 45, 2139–2145.
48. W. X. Yu, Z. D. Zhao, W. T. Zheng, Y. M. Song, B. Li, B. H. Long and Q. Jiang, Mater. Lett., 2008, 62, 747–750.
49. P. Wang, S. M. Zakeeruddin, P. Comte, I. Exnar and M. Grätzel, J. Am. Chem. Soc., 2003, 125, 1166–1167.
50. M. Grätzel, Progr. Photovolt.: Res. Appl., 2000, 8, 171–185.
51. C. Longo, J. Freitas and M. De Paoli, J. Photochem. Photobiol., A, 2003, 159, 33–39.
52. K.-M. Lee, V. Suryanarayanan and K.-C. Ho, J. Power Sources, 2008, 185, 1605–1612.
53. K. Wijeratne, J. Akilavasan, M. Thelakat and J. Bandara, Electrochim. Acta, 2012, 72, 192–198.

---