Enhancing the photovoltaic performance of dye-sensitized solar cells by modifying TiO₂ photoanodes with exfoliated graphene sheets

Abstract

Exfoliated graphene sheets (EGS) are obtained using simple liquid phase sonication and then mixed with TiO₂ nanoparticles in a dye sensitized solar cell photoanode to achieve higher electrical conductivity and faster electron transfer due to much fewer defects as compared to conventional reduced graphene oxide (RGO): I_D/I_G of 0.256 in EGS as compared with that of 1.128 in RGO. The EGS–TiO₂ photoanode yields a conversion efficiency of 8.24%, over a 19% increase compared with that of the RGO–TiO₂ photoanode, and a 43% increase over that of TiO₂ alone under the same conditions.

---

1. Introduction

Dye-sensitized solar cells (DSSCs), since their pioneering work was reported in 1991,¹ have attracted extensive interests all over the world because of their relatively low cost in fabrication.^2–4 Although the conversion efficiency of DSSCs has achieved 13%,⁵ this is still lower compared to crystalline Si (25.0%)⁶ or thin-film CIGS (Copper Indium Gallium Selenide, 20.5%)⁷ cells.

A typical DSSC comprises a dye sensitized TiO₂ film photoanode, a platinized counter electrode, and a redox electrolyte between the electrodes. As the main component, the photoanode is usually prepared by coating a layer of semiconductor (TiO₂) on conductive glass.⁸ The semiconductors act not only as carriers to adsorb dyes but also as the electron transport materials.^9,10 Photo-generated electrons are injected from the dye into the TiO₂ film where they diffuse to the electrode for collection.⁸ The major bottleneck in achieving a high power conversion efficiency (PCE) is the transport process of photo electrons across the TiO₂ nanoparticles network,^11,12 which competes with the charge recombination.^13,14 To suppress the recombination and enhance the transport, materials of high conductivity are introduced into the photoanode to channel the photoelectrons out to improve the charge-collection efficiency.^15–18

Graphene, a novel two-dimensional (2D) material with sp²-hybridized honeycomb lattice structure,^19–21 is a planar monolayer of carbon atoms arranged with a carbon–carbon bond length of 0.142 nm.²² In graphene, 3 of carbon atom's 4 outer orbital electrons from 3 sigma bonds with its three neighbors in the same plane, leaving the 4th electron out of the plane and free to move.^23,24 Therefore, electrons in the plane behave like massless relativistic particles without crystal lattices' restrictions leading to an extremely high electron mobility in two dimensions at room temperature (250 000 cm² V⁻¹ s⁻¹),^21,25 which makes graphene an ideal carrier material in DSSCs photoanodes. To date, there are a few reports on the enhancement of DSSCs based on graphene,^13,26–31 In those studies, graphene materials hold an advantage of high conductivity to transport photoelectrons from the conduct band of TiO₂ to the current collector, giving rise to improvement of PCE in the device. For examples, Yang et al. introduced graphene with TiO₂ in the photoanode through reduction of graphene oxide by hydrazine.¹³ Due to the unique advantages of graphene, including its high conductivity, suitable work function (4.4 eV) and good interfacial contact with TiO₂ nanoparticles, graphene flakes can act as bridge to shuttle electrons to the current collector and lower recombination. The short-circuit current density (J_SC) and PCE of DSSCs with graphene flakes in photoanode could significantly increase by 45% and 39%, respectively, compared to pure TiO₂ photoanode without graphene. Shu et al. prepared RGO–TiO₂ composites via a solvothermal method and reached a conversion efficiency of 5.5% through reducing the intrinsic resistance and prolong the life-time of the electron transfer in DSSCs.²⁶ Xiang et al. reported an improving of PCE of DSSCs by modifying TiO₂ photoanodes with nitrogen reduced graphene oxide.²⁷ Liu et al. developed a few layered graphene/TiO₂ composite using 10,10′-dibromo-9,9′-bianthryl (C₂₈H₁₆Br₂) as precursor to improve interfacial contact between graphene and TiO₂, which in turn increased the life-time of the photo-induced electron.²⁸ Liu et al. reported a series of TiO₂–graphene stacked photoanodes fabricated by electrospray deposition and found an over 23.6% improvement in PCE.²⁹ In our previous work, RGO reduced with vitamin C in TiO₂ resulted in an enhancement on photovoltaic performance of DSSCs by lower recombination and improves electron life-time in photoanode.³¹ Typically, graphene–TiO₂ composite electrode is prepared by mixing graphene oxide (GO) and titanium dioxide, and then reducing GO to obtain graphene (RGO) via chemistry reducing agent^13,31,32 or hydrothermal reduction.^26,33,34 Those studies report the primary cause for improvement is that graphene with a high conductivity can increase the conductivity of TiO₂. In other words, improving the conductivity of graphene that used in photoanode could further enhance the conversion efficiency of DSSCs. However, the reduction of GO to produce graphene cannot remove all structural defects introduced by the oxidation process and results in lots of basal-plane defects which seriously reduce its conductivity, thus limit the improvement in conversion efficiency. High quality graphene sheets can be obtained by numerous method, such as chemical vapor growth,³⁵ reduction of silicon carbide,³⁶ micromechanical cleavage³⁷ and so on. However, those methods are unsuitable for scalable production of graphene because of the requirements of high throughput and low cost, and what's more, those methods are difficult to prepare uniform graphene–TiO₂ composite which is required in the photoanode of DSSCs. Therefore, it is of prime importance to explore an effective and commercially available technique to produce high quality graphene and uniform graphene–TiO₂ composite that can be used as the photoanode of DSSCs.

Recent researches show that the highly conductive graphene can be obtained by liquid-phase exfoliation of graphite flakes.^38–43 The liquid-phase exfoliation was demonstrated by the strategy of tailoring Hansen solubility parameters in various solvent systems.³⁹ Ultrasonication was used to exfoliate the graphite flakes into exfoliated graphene sheets (EGS) in solvent. High quality graphene can be obtained with appropriate ultrasonic power and operation time.⁴⁰ Without the oxidation and reduction process, there are no added basal-plane defects therefore higher conductivity is expected than conventional RGO.^40,41 Furthermore, because of the solvent system, the liquid-phase exfoliation can easily get a uniform graphene suspension which is suitable for producing uniform graphene–TiO₂ composite for the photoanode of DSSCs. Herein, we explore this facile approach of making high quality graphene and directly mix it with TiO₂ in DSSC photoanode for better energy conversion. Graphite flakes are sonicated to EGS in water/acetone with appropriate proportion through simple stirring and centrifugation to form high quality graphene based TiO₂ photoanode in DSSCs. Without using special chemical reagents, we rely on the application of physical pressure to make the graphene connecting with TiO₂. With EGS–TiO₂ photoanode, over 19% and 43% increase in efficiency is achieved respectively over that with RGO–TiO₂ and TiO₂ photoanode. To the best of our knowledge, this is the first attempt in using liquid-phase exfoliated graphene sheets into TiO₂ photoanode to enhance the photovoltaic performance of DSSC and achieves exciting results.

2. Experimental procedures

Fig. 1a shows a schematic representation of the exfoliation process using acetone–water sonication to exfoliate graphene sheets directly from graphite. The exfoliation process of graphene sheets and fabrication of EGS–TiO₂ photoanode was described in Fig. 1b. Graphene Oxide (GO) powders with diameter of 500 nm to 3 μm and 99% of 1 monolayers in thickness was purchased from Nanjing XFNANO Materials Tech Co.,Ltd. Dye N-719 was purchased from Solaronix Co., Ltd., Switzerland.

Fig. 1 (a) Schematic representation of acetone–water ultrasonic exfoliated graphene sheets. (b) Fabrication process for EGS–TiO₂ photoanode.

2.1 Preparation of graphene suspensions

Graphite flakes (cat # 332461, ∼150 μm) was purchased from Sigma-Aldrich. The graphene suspensions were prepared by ultrasonicating graphite flakes (600 mg) in 200 mL DI water/acetone mixtures in a steel vessel using common ultrasonic cleaner (DELTA, D150H) operating at 150 W and 20 kHz for 4 hours at room temperature. Then, centrifugation was followed with a speed of 1000 rpm to remove the maintain graphite and keep the exfoliated graphene in the solution. Finally, the initial concentration of the graphene suspension was 0.2 mg mL⁻¹.

2.2 Preparation of the EGS–TiO₂ composite

Exfoliated graphene–TiO₂ composites were prepared using 3.0 g of commercial TiO₂ nanoparticles (P25) in each sample. During preparation, 30 mL exfoliated graphene suspension was added and stirred for 3 hours in an ultrasonic horn for a homogeneous suspension. Then, centrifugation was carried out several times with water and ethanol. Finally, the remains were dried in a vacuum dryer at 65 °C. The concentration of the EGS suspension was 0.2 mg mL⁻¹, thus there was 0.20 wt% of EGS in the sample. As such, EGS–TiO₂ composite with 0.20 wt% EGS was prepared and is referred to as EGS02. EGS–TiO₂ composites with 0.40, 0.60, 0.80, 1.00 wt% of EGS were also prepared in the same manner and are hereby designated as EGS02, EGS06, EGS08, and EGS10, respectively.

2.3 Fabrication of DSSCs

The EGS–TiO₂ pastes were synthesized following the Grätzel method.⁴⁴ The prepared pastes were spread on Fluorinated Tin Oxide (FTO) glass using the doctor blade method, and dried at 120 °C for 5 min under a pressure of 250 kPa. The film was then annealed for 30 min at 450 °C in air at a heating rate of 5°C min⁻¹, followed by a treatment in 40 mM TiCl₄ aqueous solution at 70 °C for 30 min and then annealing again under the same condition. The film was then assembled into DSSC device, as described previously.³¹ In our previously work,³¹ RGO–TiO₂ incorporated with 0.60 wt% RGO (or 0.75 wt% GO before reduction) showed the best photovoltaic performance. Therefore, a RGO–TiO₂ photoanode with 0.60 wt% RGO and a TiO₂ photoanode were also prepared for comparison.

2.4 Characterization

The crystal structures of the EGS, the RGO and EGS–TiO₂ were characterized by X-ray diffraction (XRD, PANalytical-X′Pert PRO MPD) with Cu Kα radiation (λ = 1.5405 Å), 30 mA and 40 kV. The morphology and structure of the EGS were investigated by a transmission electron microscopy (TEM, JEOL JEM-2100) and an atomic force microscopy (AFM, Bruker Dimension Edge). Raman spectra of samples were measured using a confocal Raman microscope with a 514 nm laser. The electrical conductivity of EGS and RGO was carried out by a sheet resistance measurement (Napson RG-7B) and the thickness of the films was measured by an Alpha step (KOSAKA-ET3000). The ultraviolet-visible (UV-vis) absorption spectra were obtained by a spectrophotometer (Jasco v670). The current density–voltage (J–V) curves of DSSCs were measured by the Keithley 4200 semiconductor characterization system under AM 1.5 simulated illumination at room temperature. Incident photon-to-electron conversion efficiency (IPCE) spectra were obtained using a 300 W Xe light and a monochromator provided by a QE-mini Quantum Efficiency Measurement System (Enlitech). Electrochemical impedance spectroscopy (EIS) measurements of DSSCs were measured by using a Metrohm Autolab potentiostats (PGSTAT 302N), with frequency range from 0.1 Hz to 100 kHz in the condition of 10 mV AC amplitude under 100 mW cm⁻² irradiation. Ultraviolet photoemission spectroscopy (UPS) was performed by on a ULVAC-PHI PHI 5000 VersaProbe II with a He I (21.22 eV) source.

3. Results and discussion

Fig. 2a is the TEM image of the exfoliated graphene sheets (EGS), showing the flakes being less than 3 layers and several micrometers in lateral size. The number of the layers can also be determined by analysis of the electron diffraction pattern.^40,45 Fig. 2b is the selected area electron diffraction pattern of the exfoliated graphene showing a six-fold symmetry as expected,⁴⁶ and the diffraction peaks are indexed to the Miller-Bravais (hkil) indices. Computational studies have shown that for monolayer graphene, the intensity ratio I_(01̄10)/I_(2̄110) > 1, otherwise, for multilayer graphene with the Bernal AB stacking, I_(01̄10)/I_(2̄110) < 1.^40,45 The diffraction spots on the inner circle are obviously brighter than the outer ring in Fig. 2b, implying I_(01̄10)/I_(2̄110) > 1 which indicates the EGS mostly consist of monolayer. Fig. 3a display a typical atomic force microscopy (AFM) image of EGS on a mica substrate, showing that the EGS has a lateral size of 1.5 μm. The corresponding height cross-sectional profile from the AFM image reveal that the EGS had heights about ∼0.82 nm. Furthermore, based on the analysis of the AFM image for 150 sheets and Gaussian fit, it is very clear from the histogram (Fig. 3b and c) that the maximum population of EGS has a lateral size about 753 nm and a height about 0.77 nm.

Fig. 2 (a) The TEM image of graphene sheets, (b) normal-incidence selected area electron diffraction pattern of the EGS (by using the JEOL 2100 TEM).

Fig. 3 (a) Typical tapping mode AFM image (up) and the corresponding height cross-sectional profile (bottom) of EGS on the mica substrate. Histograms of the lateral size (b) and thickness (c) distribution of EGS, obtained from 150 sheets. The green curve is the Gaussian fit and the maximum distribution of lateral size is 753 nm, correspondingly. The heights of most EGS are in the range of 0.50–1.00 nm.

Fig. 4a plots XRD spectra of the original graphite flakes, the RGO and the EGS. The peak position in EGS at around 25° 2θ corresponding to the (002) plane is almost identical to those in the original graphite flakes, indicating that the graphite lattice parameters remain and the crystal structure is not destroyed. In EGS, no (004) peak at around 55° 2θ is observed, thus the sublattices are almost devoid of long-range order (greater than four layers).^39,47 Fig. 4b plots XRD patterns of EGS–TiO₂ mix with varying amount of EGS, RGO–TiO₂ and pure TiO₂ films. The peaks are originated from the anatase and rutile TiO₂ phases, matching well with that of original TiO₂ nanoparticle (P25). Meanwhile, after decoration with RGO or EGS, there is no obvious peak change at various contents of graphene–TiO₂ composites, which is related to that the EGS/RGO content is too low to be detected by XRD observation.

Fig. 4 XRD patterns of (a) graphite flakes, EGS, and RGO, (b) TiO₂ nanoparticles mix with different amount of EGS/RGO. (c) Raman spectra of RGO (bottom) and EGS (top).

Raman spectrum could be used to characterize the ‘quality’ of graphene. As shown in Fig. 4c, the Raman spectra of EGS includes a G peak at ∼1583 cm⁻¹ corresponding to the first-order scattering of the E_2g vibration mode, a D peak at 1350 cm⁻¹ due to a breathing mode of the k-point phonons of A_1g symmetry first-order zone boundary phonons, and a 2D peak at ∼2700 cm⁻¹ caused by second-order zone boundary phonons, respectively.^48–50 Typically, the defect content can be characterized by the intensity of the D band relative to the G band, I_D/I_G.^39,50 From Fig. 4c, we note that the I_D/I_G in the EGS is ∼0.256 which is significantly lower than that of the RGO (∼1.128), indicating that the EGS contain much fewer defects than RGO. These defects in graphene can be divided into two main types: basal plane defects and edge defects. The edge defects usually depend on the size of the graphene flakes,^48,51,52 because of the smaller flakes have more edges per unit mass. Basal plane defects generally result in broadening of G bands,⁴⁹ which is obvious in the RGO in Fig. 4c (bottom).

To test the electrical conductivity, we measured the sheet resistance of the EGS and RGO film. The films were obtained by dying the 10 mL suspension on glass slide at 60 °C for 30 min. Three different sites of the sample were chosen to measure. The films thickness was measured by Alpha step as 225 ± 3 nm for EGS sample and 262 ± 4 nm for RGO sample. The average sheet resistance of the EGS film was measured to be 3.85 ± 0.05 kΩ and the RGO was 6.30 ± 0.03 kΩ, or average conductivity of 1150 ± 30 S m⁻¹ for EGS, almost twice as high as that of the RGO (606 ± 12 S m⁻¹).

To investigate the effect of the EGS on the performance of DSSCs, the photoanodes of EGS–TiO₂ and RGO–TiO₂ and TiO₂ alone were fabricated under strict identical conditions. The energy conversion efficiency was determined using photocurrent density–voltage (J–V) curves (Fig. 5a). EGS02 through EGS10 represents respectively 30, 60, 90, 120, and 150 mL of exfoliated graphene suspension mixed with the same amount (3.0 grams) of TiO₂ particles, thus indicating increasing amount of EGS in the photoanode. The relative photovoltaic parameters of the DSSCs are summarized in Table 1. Fig. 5b shows the incident photon-to-electron conversion efficiency (IPCE) curves of DSSCs with different amount of EGS/RGO. The J_SC can be theoretically calculated by integrating the product of IPCE (λ) and incident light intensity P_in(λ) of the device over the wavelength (λ) of the incident light,

J_SC = q/hc∫dλ(λ × IPCE(λ) × P_in(λ)) (1)

where q is the electron charge. Therefore, we could compare the J_SC experimentally obtained from the J–V curves vs. J_SC calculated from the spectral IPCE spectra. As shown in Table 1, the mismatch factor between J_SC (IPCE) and J_SC from J–V curves was around 0.94, which is within measurement error possibly occurring in ambient conditions. As shown in Table 1, the energy conversion efficiency of RGO–TiO₂ was 6.91%, that of TiO₂ is 5.77% and the best is 8.24% (EGS06), an increase of 19.2% over the RGO case and 42.8% over the pure TiO₂ case. Various cell parameters also peak at EGS06 as shown in Fig. 5 with increasing EGS content.

Fig. 5 (a) J–V curves of photoanodes containing varying amount of EGS/RGO. (b) Incident photon-to-electron conversion efficiency (IPCE) curves with different amount of EGS/RGO. (c) Plot of photovoltaic parameters as a function of different amounts of EGS. (d) Spectra of dye adsorption upon the addition of different amount of EGS/RGO. (e) EIS measurements of DSSCs with various EGS/RGO–TiO₂ electrodes under 100 mW cm⁻² radiation. The inset is the equivalent circuit model for fitting Nyquist plot of EIS measurement. EGS02 through EGS10 represents 0.20, 0.40, 0.60, 0.80, and 1.00 wt% EGS mixed with TiO₂ particles, respectively, thus indicating increasing amount of EGS in the photoanode.

Table 1 Photoelectric parameters of photoanodes containing varying amounts of EGS/RGO^a

Name of cells	Anode with wt% EGS	VOC (V)	JSC (mA cm^−2)	FF (%)	η (%)	JSC (IPCE) (mA cm^−2)	JSC (IPCE)/JSC	RCR (Ω)	RT (Ω)	Cμ (μF)
a Values are listed for open-circuit voltage (VOC), short-circuit current (JSC), fill factor (FF), energy conversion efficiency (η), calculated JSC from spectral IPCE spectra (JSC (IPCE)) and EIS fitting parameters (RCR, RT, Cμ). EGS02 through EGS10 represents 0.20, 0.40, 0.60, 0.80, and 1.00 wt% EGS mixed with TiO2 particles, respectively, thus indicating increasing amount of EGS in the photoanode.
TiO2	0	0.704	12.59	65.07	5.77	11.91	0.945	31.21	25.34	107.33
EGS02	0.20	0.694	13.70	66.81	6.35	12.94	0.944	30.46	17.02	115.24
EGS04	0.40	0.692	15.46	67.84	7.26	14.73	0.952	29.20	11.11	122.92
EGS06	0.60	0.690	17.31	69.04	8.24	16.15	0.932	28.65	7.63	168.73
EGS08	0.80	0.683	14.90	62.25	6.34	14.02	0.941	25.60	6.28	108.24
EGS10	1.00	0.665	11.37	56.55	4.28	10.82	0.952	14.12	5.56	101.10
RGO–TiO2	0.60 wt% RGO	0.697	14.52	68.22	6.91	13.74	0.945	30.57	11.34	165.87

---

The overall efficiency η was evaluated by the short-current density (J_SC), the open-circuit photovoltage (V_OC), the fill factor of the cell (FF), and the incident light power density (P_in), namely,^53,54

It should be noted that the overall conversion efficiency of DSSCs is affected significantly by the open-circuit voltage (V_OC), the short-circuit current density (J_SC), and the fill factor (FF). The V_OC value of the cells shows a decrease trends with increasing amount of EGS in photoanodes as shown in Table 1 and Fig. 5. For DSSCs, the V_OC values are mainly obtained as the difference between the electron's Fermi level in the TiO₂ (ε_CB) and the Nernst potential of the redox couple (ε_redox) in the electrolyte and can be described as,⁵⁵

where N_e is the effective density of states at the TiO₂ conduction band (CB) edge, γ is the characteristic constant of TiO₂ tailing states, κ_B is the Boltzmann constant, e is the elementary charge, n is the electron density in TiO₂, and T is the temperature. In principle, V_OC is determined by the energy difference between the quasi-Fermi level of TiO₂ under illumination and the chemical potential of redox couples in the electrolyte, which is closely related to charge recombination process at the TiO₂/dye/electrolyte interface.^53,56 After decorating with EGS, the charge could get into the external circuit more easily with less charge recombination, and it would cause less electron density (n) in TiO₂ CB, leading to a decrease in V_OC.

The short-circuit photocurrent is another parameter that is fundamental to the efficiency of DSSCs. A good J_SC is based on three parameters: good charge injection efficiency from the excited dye to TiO₂, a large amount of dye loading onto the TiO₂ surface, and the excellent charge collection/extraction efficiency in the photoanode. First of all, for dye N-719 used here, it has been proved in many papers^57–59 that the injection from the excited dye molecule to the semiconductor's CB is fast enough to compete with other processes. The UV-vis spectroscopy was used to assess the impact of EGS–TiO₂ film on dye absorption (i.e., loading).⁶⁰ After preparing the photoelectrodes, N-719 dye was adsorbed for 24 hours and desorbed with 1 mM KOH solution for UV-vis analysis. As shown in Fig. 5, the amount of dye adsorbed onto the surfaces of photoanodes decreased in the order: EGS06 > EGS04 > EGS08 > EGS02 > TiO₂ > EGS10. The EGS06 adsorbed the highest amount of dye among the tested photoanodes, because the EGS formed sheets between the TiO₂ particles to increase the specific surface area resulted in an increase in dye loading.^13,30 However, the addition of excess EGS resulted in the formation of graphene sheets shielding TiO₂ particles, which decreased the dye loading amount. Therefore, the amount of dye adsorbed is related to the current density and show a good agreement with the trends of J_SC, as shown in Fig. 5. With appropriated amount, the addition of EGS increased the dye absorption and enhanced the energy conversion efficiency of DSSCs. However, for RGO–TiO₂ and EGS06, it does not show any clear difference of the dye loading amount on the Fig. 5. Therefore, the charge collection and transportation efficiency of the injected electrons is critical to the photocurrent of the cell based on EGS and RGO. A good conductor with high electrical transport property is widely accepted that it could act as a high-way to capture and transport electrons from the conductor band of TiO₂ to electron collector.^13,30,31 After decorating with EGS, the photo electrons in the TiO₂ CB could get into the external circuit more easily, and this will be proved by the electrochemical impedance spectrum (EIS) data.

EIS was carried out to give evidence of the graphene effect on boosting the charge transport and reducing the recombination. As shown in Fig. 5, the Nyquist plot contains two main semicircles in the high and medium-frequency regions, where the semicircles in low-frequency region corresponding to redox couple diffusion in the electrolyte are absent. The large semicircle in the medium frequency region relates to the electron transport in the TiO₂ film and the back transfer at the TiO₂/electrolyte interface.^17,53,61 According to the previous study,^17,62 the radius of arc in the medium-frequency region reflects resistance to charge transfer at the aforementioned interface. In Fig. 5, it is clearly showed that the semicircle size for the EGS–TiO₂ photoanode in the medium-frequency region exhibits significant reduction with increasing amount of EGS than that for the pure TiO₂ photoanode, indicating faster electron transfer process through the graphene bridge. This gives rise to reduced V_OC values as increasing EGS content, as shown in Table 1.

A transmission line model, as shown in Fig. 5, is used to characterize the measured EIS of TiO₂ and graphene–TiO₂ based DSSCs, aimed at simultaneously estimating the charge recombination and transport resistance in the photoanode.^63–65 This model can be expressed as following formula without considering the effects of the Pt counter electrode or the diffusion in the electrolyte,¹⁷

where R_CR and R_T are the charge recombination and transport resistance in the TiO₂ network, ω_d = 1/R_TC_μ and ω_k = 1/R_CRC_μ represent characteristics transport and recombination frequency, C_μ is the chemical capacitance. ω is the angular frequency, respectively. A lower transport resistance leads to a faster interfacial charge transfer process, a higher recombination resistance indicating low charge recombination at TiO₂/electrolyte interface, and a higher chemical capacitance indicates increased dye adsorption, with those three parameters give a higher η.^66,67

Taking into account of the electron transfer at Pt/electrolyte interface and the diffusion process in electrolyte, a full equivalent circuit model of DSSCs is presented in Fig. 5 (inset in Fig. 5) to explain the transfer and recombination process of the photo-excited electrons in the cells. Corresponding to the transmission line model, the starting point in the circuit (R_S) is the series resistance, including the sheet resistance of the FTO glass, the contact resistance of the cell, respectively. R_Pt and C_Pt are the charge transfer resistance and the double-layer capacitance at the counter electrode I⁻₃ at the electrolyte/Pt/FTO glass interface, DX1 is complicated element in the equivalent circuit model (DX type TL-Open), and W1(Ws) represents Warburg impedance (short), respectively.⁶⁸ Therefore, the parameters including R_CR, R_T and C_μ could be obtained by fitting Nyquist plot and are summarized in Table 1. Obviously, for photoanode decorating with EGS, the charge transport resistance (R_T) is significantly reduced to a low level, compared with that of the pure TiO₂ photoanode (25.34 Ω). However, with too much EGS, the charge recombination resistance (R_CR) is also reduced, indicating higher charge recombination and lower V_OC of DSSCs, which, in turn, eliminate the benefit of low R_T of EGS. Meanwhile, EGS06 achieve the highest C_μ (168.73 μF) among those cells with or without EGS, indicating the largest amount of dye adsorption, which is in agreement with the UV-vis spectroscopy as discussed above. Moreover, the charge transport resistance of EGS06 is 7.63 Ω, just 67.2% of the RGO–TiO₂ (11.34 Ω), without significantly change the R_CR and C_μ. This result can be used to explain the difference of the short-circuit current between the EGS06 and RGO–TiO₂. The high conductive EGS can facilitate charge transport in photoanodes more easily than RGO to improve the photo current even with the same dye loading amount.

In order to explain the influence of electronic structure of EGS in DSSCs, the UPS spectra was measured. Fig. 6a shows the UPS spectra around the secondary electrons threshold region of the EGS sample and the inset shows the full UPS spectrum. The work function (Φ) is determined from the secondary electron threshold as Φ = hν − E_th, where hν and E_th are the photon energy of excitation light (He I discharge lamp, 21.22 eV) and the secondary electron threshold energy, respectively. The secondary electron threshold was determined by extrapolating two solid lines from the background and straight onset in the secondary electron threshold region of the UPS spectra.⁶⁹ From Fig. 6a, the onset point of UPS spectra in the EGS is about 16.71 eV, suggesting that the work function of the EGS sample is about 4.51 eV, which is between the CB of TiO₂ and FTO. Fig. 6b shows the energy level diagram of EGS–TiO₂ DSSCs. Under illumination, dye molecular (N719) was excited by light and electrons transit from HOMO to LUMO. The LUMO and HOMO are the lowest unoccupied molecular orbit and highest occupied molecular orbit, respectively. The CB of semiconductor TiO₂ receives the electrons from the LUMO of photo-excited dye. Because the conductivity of EGS is billions of times larger than that in TiO₂ (∼10⁻¹⁰ to 10⁻¹ S m⁻¹, respectively^70,71), and the EGS is homogeneous in the film, the excited electrons from TiO₂ can be collected and transferred to the EGS without any obstruction. The electrons can rapidly transport to the conductive substrate (FTO) through EGS, resulting in reducing the R_T of the cells and increasing the photo-generated current. However, EGS have a continuous band structure like metal, electrons can stay at continuous energy levels near the Fermi level, accelerating recombination to the dye or back reaction to tri-iodide in the electrolyte, leading to decrease the R_CR and exhibit a trend of decreasing V_OC.

Fig. 6 (a) The UPS spectra of the EGS around the secondary electrons threshold region. The onset point of UPS spectra in EGS is about 16.71 eV, suggesting that the work function of EGS is 4.51 eV. The inset shows the full UPS spectrum. (b) Energy diagrams of DSSCs incorporating EGS.

In addition, the good fill factor is based on a low total series resistance of the cells. After the addition of the excellent conductor-EGS, the interior resistance of the cells can be reduced to a lower level to achieve a higher fill factor. However, the excessive graphene can act as a kind of recombination center accelerating recombination of electrons to the dye or back reaction to the electrolyte.^13,72 Therefore, the fill factor exceeded to the highest value 69.04% with the optimal amount of EGS.

4. Conclusions

For the first time, direct sonication of graphite produced exfoliated graphene sheets (EGS) were mixed with nanosized TiO₂ particles in preparation of photoanode in dye-sensitized solar cells. Energy conversion measurements indicate that the mixed EGS–TiO₂ photoanode results in as much as over 19% and 43% better than RGO–TiO₂ and TiO₂-only counterparts under strict identical preparation conditions. The increase is believed to come from the increased conductivity of EGS owing to fewer basal plane defects as there is no chemical reduction process in comparison with the case of RGO, and to the much better connectivity to the case of TiO₂-only photoanode.

Acknowledgements

The authors gratefully acknowledge funding from the National Nature Science Foundation of China (No. 51301041), the Ministry of Science and Technology (MOST) of Taiwan (104-2221-E-131-011), and Ming Chi University of Technology for financial and research support.

References

1. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737–740.
2. L. Sun, S. Zhang, X. Sun and X. He, J. Nanosci. Nanotechnol., 2010, 10, 4551–4561.
3. J. Gong, J. Liang and K. Sumathy, Renewable Sustainable Energy Rev., 2012, 16, 5848–5860.
4. S. Zhang, X. Yang, Y. Numata and L. Han, Energy Environ. Sci., 2013, 6, 1443.
5. S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, B. F. Curchod, N. Ashari-Astani, I. Tavernelli, U. Rothlisberger, M. K. Nazeeruddin and M. Gratzel, Nat. Chem., 2014, 6, 242–247.
6. T. Saga, NPG Asia Mater., 2010, 2, 96–102.
7. M. A. Green, K. Emery, Y. Hishikawa, W. Warta and E. D. Dunlop, Progress in Photovoltaics: Research and Applications, 2015, 23, pp. 1–9.
8. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595–6663.
9. L. Peter, J. Electroanal. Chem., 2007, 599, 233–240.
10. H. Zhang, X. Lv, Y. Li, Y. Wang and J. Li, ACS Nano, 2010, 4, 380–386.
11. K. Zhu, N. R. Neale, A. Miedaner and A. J. Frank, Nano Lett., 2007, 7, 69–74.
12. Y. B. Tang, C. S. Lee, J. Xu, Z. T. Liu, Z. H. Chen, Z. He, Y. L. Cao, G. Yuan, H. Song, L. Chen, L. Luo, H. M. Cheng, W. J. Zhang, I. Bello and S. T. Lee, ACS Nano, 2010, 4, 3482–3488.
13. N. Yang, J. Zhai, D. Wang, Y. Chen and L. Jiang, ACS Nano, 2010, 4, 887–894.
14. Q. Liu, Y. Zhou, Y. Duan, M. Wang, X. Zhao and Y. Lin, J. Alloys Compd., 2013, 548, 161–165.
15. L. J. Brennan, M. T. Byrne, M. Bari and Y. K. Gun'ko, Adv. Energy Mater., 2011, 1, 472–485.
16. D. Chen, H. Zhang, Y. Liu and J. Li, Energy Environ. Sci., 2013, 6, 1362.
17. M. Zhu, X. Li, W. Liu and Y. Cui, J. Power Sources, 2014, 262, 349–355.
18. J. Zhi, H. Cui, A. Chen, Y. Xie and F. Huang, J. Power Sources, 2015, 281, 404–410.
19. V. Singh, D. Joung, L. Zhai, S. Das, S. I. Khondaker and S. Seal, Prog. Mater. Sci., 2011, 56, 1178–1271.
20. F. Bonaccorso, L. Colombo, G. Yu, M. Stoller, V. Tozzini, A. C. Ferrari, R. S. Ruoff and V. Pellegrini, Science, 2015, 347, 1246501.
21. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, M. I. Katsnelson, I. V. Grigorieva, S. V. Dubonos and A. A. Firsov, Nature, 2005, 438, 197–200.
22. C. Lee, X. Wei, J. W. Kysar and J. Hone, Science, 2008, 321, 385–388.
23. S. Stankovich, D. A. Dikin, G. H. Dommett, K. M. Kohlhaas, E. J. Zimney, E. A. Stach, R. D. Piner, S. T. Nguyen and R. S. Ruoff, Nature, 2006, 442, 282–286.
24. D. Li and R. B. Kaner, Science, 2008, 320, 1170–1171.
25. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666–669.
26. W. Shu, Y. Liu, Z. Peng, K. Chen, C. Zhang and W. Chen, J. Alloys Compd., 2013, 563, 229–233.
27. Z. Xiang, X. Zhou, G. Wan, G. Zhang and D. Cao, ACS Sustainable Chem. Eng., 2014, 2, 1234–1240.
28. Y. Liu, Y. Cheng, W. Shu, Z. Peng, K. Chen, J. Zhou, W. Chen and G. S. Zakharova, Nanoscale, 2014, 6, 6755–6762.
29. J. Liu, X. Fu, D.-P. Cao, L. Mao, J. Wang, D.-H. Mu, B.-X. Mi, B.-M. Zhao and Z.-Q. Gao, Org. Electron., 2015, 23, 158–163.
30. S.-B. Kim, J.-Y. Park, C.-S. Kim, K. Okuyama, S.-E. Lee, H.-D. Jang and T.-O. Kim, J. Phys. Chem. C, 2015, 119, 16552–16559.
31. H. Ding, S. Zhang, J.-T. Chen, X.-P. Hu, Z.-F. Du, Y.-X. Qiu and D.-L. Zhao, Thin Solid Films, 2015, 584, 29–36.
32. H. N. Kim, H. Yoo and J. H. Moon, Nanoscale, 2013, 5, 4200–4204.
33. P. Wang, J. Wang, X. Wang, H. Yu, J. Yu, M. Lei and Y. Wang, Appl. Catal., B, 2013, 132–133, 452–459.
34. M. Zhu, X. Li, W. Liu and Y. Cui, J. Power Sources, 2014, 262, 349–355.
35. K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, J. H. Ahn, P. Kim, J. Y. Choi and B. H. Hong, Nature, 2009, 457, 706–710.
36. C. Berger, Z. Song, X. Li, X. Wu, N. Brown, C. Naud, D. Mayou, T. Li, J. Hass, A. N. Marchenkov, E. H. Conrad, P. N. First and W. A. de Heer, Science, 2006, 312, 1191–1196.
37. K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich, S. V. Morozov and A. K. Geim, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 10451–10453.
38. Y. Hernandez, V. Nicolosi, M. Lotya, F. M. Blighe, Z. Sun, S. De, I. T. McGovern, B. Holland, M. Byrne, Y. K. Gun'Ko, J. J. Boland, P. Niraj, G. Duesberg, S. Krishnamurthy, R. Goodhue, J. Hutchison, V. Scardaci, A. C. Ferrari and J. N. Coleman, Nat. Nanotechnol., 2008, 3, 563–568.
39. M. Yi, Z. Shen, X. Zhang and S. Ma, J. Phys. D: Appl. Phys., 2013, 46, 025301.
40. A. Ciesielski and P. Samori, Chem. Soc. Rev., 2014, 43, 381–398.
41. I. W. P. Chen, Y.-S. Chen, N.-J. Kao, C.-W. Wu, Y.-W. Zhang and H.-T. Li, Carbon, 2015, 90, 16–24.
42. E. Quesnel, F. Roux, F. Emieux, P. Faucherand, E. Kymakis, G. Volonakis, F. Giustino, B. Martín-García, I. Moreels, S. A. Gürsel, A. B. Yurtcan, V. D. Noto, A. Talyzin, I. Baburin, D. Tranca, G. Seifert, L. Crema, G. Speranza, V. Tozzini, P. Bondavalli, G. Pognon, C. Botas, D. Carriazo, G. Singh, T. Rojo, G. Kim, W. Yu, C. P. Grey and V. Pellegrini, 2D Mater., 2015, 2, 030204.
43. F. Bonaccorso, N. Balis, M. M. Stylianakis, M. Savarese, C. Adamo, M. Gemmi, V. Pellegrini, E. Stratakis and E. Kymakis, Adv. Funct. Mater., 2015, 25, 3870–3880.
44. S. Ito, P. Chen, P. Comte, M. K. Nazeeruddin, P. Liska, P. Péchy and M. Grätzel, Progress in Photovoltaics: Research and Applications, 2007, 15, pp. 603–612.
45. Z. Li, P. Wu, C. Wang, X. Fan, W. Zhang, X. Zhai, C. Zeng, Z. Li, J. Yang and J. Hou, ACS Nano, 2011, 5, 3385–3390.
46. C. L. Jia, M. Lentzen and K. Urban, Science, 2003, 299, 870–873.
47. C. J. Shih, A. Vijayaraghavan, R. Krishnan, R. Sharma, J. H. Han, M. H. Ham, Z. Jin, S. Lin, G. L. Paulus, N. F. Reuel, Q. H. Wang, D. Blankschtein and M. S. Strano, Nat. Nanotechnol., 2011, 6, 439–445.
48. A. Ferrari and J. Robertson, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 14095–14107.
49. A. C. Ferrari, J. C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. S. Novoselov, S. Roth and A. K. Geim, Phys. Rev. Lett., 2006, 97, 187401.
50. L. M. Malard, M. A. Pimenta, G. Dresselhaus and M. S. Dresselhaus, Phys. Rep., 2009, 473, 51–87.
51. G. Cravotto and P. Cintas, Chemistry, 2010, 16, 5246–5259.
52. G. Xie and Y. Shen, Phys. Chem. Chem. Phys., 2015, 17, 8822–8827.
53. Q. Wang, S. Ito, M. Gratzel, F. Fabregat-Santiago, I. Mora-Sero, J. Bisquert, T. Bessho and H. Imai, J. Phys. Chem. B, 2006, 110, 25210–25221.
54. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595–6663.
55. B. J. Song, H. M. Song, I. T. Choi, S. K. Kim, K. D. Seo, M. S. Kang, M. J. Lee, D. W. Cho, M. J. Ju and H. K. Kim, Chemistry, 2011, 17, 11115–11121.
56. H. Yang, G. H. Guai, C. Guo, Q. Song, S. P. Jiang, Y. Wang, W. Zhang and C. M. Li, J. Phys. Chem. C, 2011, 115, 12209–12215.
57. J. B. Asbury, E. Hao, Y. Wang, H. N. Ghosh and T. Lian, J. Phys. Chem. B, 2001, 105, 4545–4557.
58. W. R. Duncan and O. V. Prezhdo, J. Phys. Chem. B, 2005, 109, 365–373.
59. S. Y. Brauchli, F. J. Malzner, E. C. Constable and C. E. Housecroft, RSC Adv., 2014, 4, 62728–62736.
60. X. Pan, Y. Zhao, S. Liu, C. L. Korzeniewski, S. Wang and Z. Fan, ACS Appl. Mater. Interfaces, 2012, 4, 3944–3950.
61. K. Pan, Y. Dong, W. Zhou, G. Wang, Q. Pan, Y. Yuan, X. Miao and G. Tian, Electrochim. Acta, 2013, 88, 263–269.
62. N. J. Bell, Y. H. Ng, A. Du, H. Coster, S. C. Smith and R. Amal, J. Phys. Chem. C, 2011, 115, 6004–6009.
63. J. van de Lagemaat, N. G. Park and A. J. Frank, J. Phys. Chem. B, 2000, 104, 2044–2052.
64. X. Yin, Z. Xue and B. Liu, J. Power Sources, 2011, 196, 2422–2426.
65. F. Fabregat-Santiago, J. Bisquert, L. Cevey, P. Chen, M. Wang, S. M. Zakeeruddin and M. Gratzel, J. Am. Chem. Soc., 2009, 131, 558–562.
66. Y.-F. Chan, C.-C. Wang and C.-Y. Chen, J. Phys. Chem. A, 2013, 1, 5479.
67. Y.-F. Chan, C.-C. Wang, B.-H. Chen and C.-Y. Chen, Progress in Photovoltaics: Research and Applications, 2013, 21, pp. 47–57.
68. J. Qiu, F. Zhuge, K. Lou, X. Li, X. Gao, X. Gan, W. Yu, H.-K. Kim and Y.-H. Hwang, J. Mater. Chem., 2011, 21, 5062.
69. S. Y. Kim, K. Hong, K. Kim, H. K. Yu, W.-K. Kim and J.-L. Lee, J. Appl. Phys., 2008, 103, 076101.
70. K. Madhusudan Reddy, S. V. Manorama and A. Ramachandra Reddy, Mater. Chem. Phys., 2003, 78, 239–245.
71. R. Asmatulu, A. Karthikeyan, D. C. Bell, S. Ramanathan and M. J. Aziz, J. Mater. Sci., 2009, 44, 4613–4616.
72. X. Dang, H. Yi, M. H. Ham, J. Qi, D. S. Yun, R. Ladewski, M. S. Strano, P. T. Hammond and A. M. Belcher, Nat. Nanotechnol., 2011, 6, 377–384.

---