Variable temperature spectroelectrochemistry study of silver-doped TiO₂ and its influence on the performance of dye sensitized solar cells

Abstract

Ag-doped TiO₂ nanoparticles are prepared and used as semiconductor materials of photoanodes to improve the performance of dye sensitized solar cells (DSSCs). The results of variable temperature spectroelectrochemistry study show that the conduction band edge of Ag-doped TiO₂ shifts positively, which enhances the driving force of electrons and improves the electron injection efficiency from the LUMO of the dye to the conduction band of TiO₂. The deposition of Ag not only benefits efficient charge transfer but also could minimize the charge recombination process, resulting in a significant photocurrent enhancement. At the optimum Ag concentration of 1.2 at%, the DSSC exhibited a J_sc of 19.75 mA cm⁻², a V_oc of 0.73 V, and a FF of 0.57 with an energy conversion efficiency (η) of 7.34%, indicating a 57% and 30% increase in J_sc and η respectively than that of DSSC based on an undoped TiO₂ photoanode, which gives a J_sc of 12.61 mA cm⁻², a V_oc of 0.75 V, and a FF of 0.59 with a η of 5.64%.

---

1. Introduction

Dye-sensitized solar cells (DSSCs) have gained much attention from both industry and academic research during the past decades due to their great advantage of light weight, low cost and easy processing.^1–4 A DSSC device is a sandwich structure which is comprised of a photoanode, an electrolyte and a counter electrode. Among these components, the photoanode has been receiving greater attention for the reason that the photoanode not only transports photo-induced electrons but also acts as a matrix to adsorb organic dyes, which directly determines the photo-current density.⁵ Accordingly, the semiconductors material used in photoanode play an important role in the working process of DSSCs. Among various semiconductors, nanocrystalline anatase TiO₂ has been identified as a promising photoanode material in DSSCs and has been widely investigated.^6–8 However, the large band gap of TiO₂ (3.2 eV) limits its utilization of visible light and many dyes failed to inject electrons into the conduction band of TiO₂ owing to insufficient electron injection driving force.⁹ Besides, the generated electrons encounter huge amounts of grain boundaries among nanoparticles as they travel through the nanostructured TiO₂ networks inside the photoanode before reaching the collecting electrode. Such a random walk mode can lead to retardation in electron transport and increases their probability to recombine with oxidized dye molecules and electron acceptor in electrolyte. These limit the development of DSSCs. Therefore, proper modifications of TiO₂ semiconductors are desired for efficient solar cells.

Focussing on the nature of the TiO₂ semiconductor, one way of achieving improvements in the properties of DSSCs is by doping the TiO₂ semiconductor. After doping with metal ions some of the photovoltaic properties of the cells can be improved due to the positive band edge movement, faster electron transport times and lower electron-recombination rate after correctly doping TiO₂ with metal ions. For example, Ko et al. reported an improvement of the efficiency of DSSCs after simultaneous doping TiO₂ with aluminium and tungsten;¹⁰ Duan et al. enhanced the performance of DSSCs by doping with metal and F;¹¹ Saad et al. studied the Zn²⁺ concentration in Zn-doped TiO₂ on the dye sensitized solar cells;¹² Lee et al. and Lü et al. managed to improve cell performance with niobium;^13,14 Zhao et al. studied the influence of yttrium dopant on the performance of DSSCs;¹⁵ Bakhshayesh et al. boosted the short circuit current density of DSSCs by one-pot preparation of Sr, Cr co-doped xerogel film;¹⁶ Tian et al. showed retarded charge recombination in nitrogen-doped TiO₂ DSSCs;¹⁷ Malik et al. improved the performance of DSSCs based on Mo and Ni co-doped TiO₂;¹⁸ Wang and Teng enhanced electron transport in cells doped with zinc under low-intensity illumination;¹⁹ Xu et al. studied the influence of ytterbium doping;²⁰ Park et al. improved electron transfer in DSSCs with Nb doped TiO₂/Ag;²¹ Liu et al. improved the efficiency using TiO₂ photoanode structure with gradations in vanadium concentration²² and Tanyi et al. enhanced the efficiency of DSSCs based on Mg and La co-doped TiO₂ photoanodes.²³ These examples indicate that doping the TiO₂ semiconductor is a proper way to enhance the performance of DSSCs. Doping with metal ions could enhance the structural, optical, electronic, and electrochemical properties of TiO₂, which results in enhanced performance of solar cells. This is also confirmed by perovskite solar cells, which also use the mesoporous TiO₂.^24,25 As a noble metal ion, silver doped TiO₂ has been actively reported and applied as photoanode materials with the aim of improving the efficiency of a DSSC.^26–31 However, most of them focus on co-doping Ag with other element such as N, S, P or studying its surface plasmonic resonance effect.^28–31 The spectroelectrochemistry study of single silver element doped TiO₂ and its influence on the performance of DSSC was scarcely reported.

In this study, we report on the fabrication of Ag-doped TiO₂ semiconductors by a hydrothermal process for application in DSSCs. To determine the conduction band edge in transparent mesoporous Ag-doped TiO₂ semiconducting electrodes, a steady-state variable temperature spectroelectrochemical measurement was introduced. To characterize the influence of silver dopant with different concentrations on TiO₂ semiconductor, X-ray diffraction, X-ray photoelectron spectroscopy, scanning electron microscopy, transmission electron microscopy and UV-vis spectroscopy study are also performed. At the optimum Ag concentration of 1.2 at%, the DSSC exhibited a J_sc of 19.75 mA cm⁻², a V_oc of 0.73 V, and a FF of 0.57 with the energy conversion efficiency (η) of 7.34%, indicating a 57% and 30% increase in J_sc and η respectively than that of DSSC based on undoped TiO₂ photoanode. The improved performance of DSSCs with Ag-doped TiO₂ semiconductor is characterized by dark current measurement, electrochemical impedance spectroscopy and open-circuit voltage decay.

2. Experimental

2.1. Materials and reagents

All of the solvents and chemicals used in this work were of reagent grade without further purification. AgNO₃ (>99%) was purchased from Sinopharm. Chemical Reagent Co., Ltd., China. Isopropyl titanate (98%) purchased from J&K CHEMICA Co., Shanghai, China. Propylene carbonate, nitric acid, absolute ethanol, ethyl cellulose, and terpineol were purchased from Jinan Henghua Chemical Co., Shandong, China. Acetonitrile, I₂, LiI, tert-butylpyridine, and lithium perchlorate were purchased from Tianli Chemical Co., Tianjin, China. Ferrocenium, ferrocene (>99%) were purchased from Xilong Chemical Co., Guangdong, China, cis-bis-(isothiocyanato)bis(2,2-bipyridyl-4,4-dicarboxylato)ruthenium(II)bis-tetrabutylammonium (N719) was purchased from Solaronix Company, Switzerland. F-doped SnO₂-coated glass plate (FTO, 90% transmittance in the visible, 15 Ω cm⁻²) was purchased from Acros Organics, Belgium, and cleaned by a standard procedure.

2.2. Synthesis of Ag-doped TiO₂

Undoped TiO₂ and Ag-doped TiO₂ powders were synthesized via a hydrothermal method. Typically, 0.4 mL HNO₃ was first added into an Erlenmeyer flask containing a small amount of AgNO₃ in 60 mL ultrapure water to adjust the pH value. Then, 10 mL of titanium isopropylate (98%) was added dropwise into above solution under vigorous stirring within 20 minutes. After that, the flask was placed in a water bath of 368 K until the mixture was evaporated to 20 mL transparent sol. The sol was then moved into a 25 mL Teflon-lined stainless steel autoclave and kept in an oven at 473 K for 24 h. After cooled at room temperature, the product was centrifuged and washed with distilled water and ethanol. Then the precipitate was vacuum dried at 373 K, calcined at 773 K for 2 h and cooled to ambient temperature, Ag-doped TiO₂ powder was obtained. In order to obtained 0.7 at% Ag–TiO₂, 0.9 at% Ag–TiO₂, 1.0 at% Ag–TiO₂, 1.2 at% Ag–TiO₂ and 1.5 at% Ag–TiO₂, the added amount of AgNO₃ was controlled to Ag⁺/Ti⁴⁺ molar ratio = 0.7 at%, 0.9 at%, 1.0 at%, 1.2 at% and 1.5 at%, respectively. For comparison, undoped TiO₂ powder was prepared under the same condition except that AgNO₃ was not introduced into the reacting system.

2.3. Preparation of film and cell assembly

Pure TiO₂ and Ag-doped TiO₂ paste was prepared according to the literature.³² It was printed onto FTO conductive glass and then dried at 373 K for 5 min. The above process was repeated for six times to make the film thickness was ca. 10 μm. The obtained film was then sintered at 773 K for 15 min. It was dyed after immersed in a solution of 0.3 mM N719 in absolute ethanol for 24 h and then washed with ethanol. The sandwich-type solar cell device was assembled by placing a platinum-coated conductive glass as counter electrode on the dyed photoanode, a drop of liquid electrolyte containing 0.5 M LiI, 0.05 M I₂, 0.1 M 4-tert-butylpyridine (TBP) was added to fill the void between two electrodes and clipped together as open cells for measurement.

The transparent electrodes of Ag-doped TiO₂ for spectroelectrochemistry measurement were prepared by doctor-blade route with the paste that was prepared under the same condition with Ag-doped TiO₂ nanoparticles except that after keeping in an oven at 473 K for 24 h, the obtained product was cooled at room temperature and 0.5–0.6 g carbowax 20 000 was added and stirred for more than 3 days. In all cases, the thin films were annealed at 693 K for 30 min. Similarly, the transparent electrodes of undoped TiO₂ for spectroelectrochemistry were also prepared.

2.4. Characterization

The X-ray powder diffraction (XRD) measurement was obtained on Shimadzu XRD-6000 X-ray diffraction instrument with Cu-Kα radiation. The X-ray photoelectron spectroscopy (XPS) was measured with an ESCALAB-250 spectrometer (Thermo, America) of the Al K_α source in an ultrahigh vacuum (UHV) at 3.5 × 10⁻⁷ Pa. UV-vis spectra was performed with UV-2250 spectrophotometer (Shimadzu, Japan). Scanning electron microscopy (SEM) was taken using a Rili SU 8000HSD Series Hitachi New Generation Cold Field Emission SEM and transmission electron microscope (TEM) was taken using a HitachiH-700 SEM. The amounts of absorbed dye were measured by desorbing the dye from the dye-sensitized films in a 0.1 M NaOH solution in water and ethanol (1 : 1, v/v), and then the concentration of desorbed dye in the NaOH solution was measured using an UV-visible spectrophotometer. Dark current, electrochemical impedance spectroscopy (EIS) and open-circuit voltage decay (OCVD) were recorded by CHI660D Electrochemical Analyzer. The spectroelectrochemical experiments were performed by assembling a potentiostat (Shanghai Chenhua Device Company, China) and a Cary 60 UV-vis spectrophotometer (Agilent Technologies, America), using a Pt wire counter electrode and a Ag reference electrode. The current density–voltage (J–V) curves of DSSCs were recorded by Keithley model 2400 digital source meter under AM1.5G irradiation. The incident light intensity was 100 mW cm⁻² calibrated by a standard silicon solar cell. The working areas of the cells were masked to 0.16 cm². Based on J–V curve, the fill factor (FF) is defined as: FF = (J_max × V_max)/(J_sc × V_oc) where J_max and V_max are the photocurrent density and photovoltage for maximum power output; J_sc and V_oc are the short-circuit photocurrent density and open-circuit photovoltage, respectively, the overall energy conversion efficiency η is defined as: η = (FF × J_sc × V_oc)/P_in where P_in is the power of the incident light. The measurement of the incident photon-to-current conversion efficiency (IPCE) was performed by an EQE/IPCE spectral response system (Newport).

3. Result and discussion

3.1. Characterization of Ag–TiO₂

Composition and structure of the undoped and doped TiO₂ nanoparticles were investigated by XRD. As shown in Fig. 1, the XRD peaks at 2θ = 25.3°, 37.8°, 48.0°, 62.8°, 70.1° and 75.4° in the spectra of undoped TiO₂ and x at% Ag–TiO₂ samples are easily identified as a relatively high crystallinity anatase form (JCPDS 65-5714). However, a clearly small diffraction peak at 2θ = 27.4° in the spectra of undoped TiO₂, which is easily taken as (110) plane of rutile form was observed, and it disappears with the increased doping amount of Ag, which implies that Ag doping helped to inhibit the phase transformation of TiO₂ from anatase to rutile. Rutile phase is unfavorable for DSSC compared to anatase phase for it can cause many negative effects such as lesser amount of adsorbed dye, owing to a smaller surface area per unit volume of the rutile film compared with that of the anatase film and slower electron transport in the rutile layer than in the anatase layer due to differences in the extent of interparticle connectivity associated with the particle packing density.³³ From this aspect, doping with Ag is benefited for improving properties of TiO₂ and further enhancing the performance of DSSCs. Furthermore, no impurity or secondary peaks in the XRD patterns were detected. However, elemental analysis using X-ray fluorescence (XRF) confirms the presence of Ag in the synthesized x at% Ag–TiO₂ (Table 1). On the basis of the fact that there is no XRD peak for Ag impurities were detected and the existence of Ag for Ag-doped samples were confirmed by XRF, it is implied that the Ag dopant had been doped into TiO₂ through lattice replacement. The diffraction peak of the sample x at% Ag–TiO₂ is obviously broader and the peak intensities of x at% Ag–TiO₂ are weaker with the increase of Ag doping amount, which further confirms the presence of foreign ions of Ag in the crystal lattice.³⁴ The crystalline sizes for undoped TiO₂ and x at% Ag–TiO₂ were in the region of 10–40 nm as determined from the diffraction peak of (101) and Scherer's equation. According to the XRD profiles, the crystal structure and the particle size were not noticeably influenced by the Ag-doping with the content ranging from 0.7 at% to 1.5 at%.

Fig. 1 XRD pattern of pure TiO₂ and x at% Ag–TiO₂.

Table 1 Ag contents in x at% Ag–TiO₂ measured by XRF

Samples (at% of Ag)	Analyzed Ag content in at%
0.7	0.45
0.9	0.63
1.0	0.84
1.2	1.03
1.5	1.21

---

The SEM morphology of different samples in Fig. 2 shows that the crystalline sizes for undoped TiO₂ and at% Ag–TiO₂ were in the region of 10–40 nm, which were in agreement with the size of nanoparticles determined from the diffraction peak of (101) and Scherer's equation and further confirms that the particle size were not noticeably influenced by the Ag-doping with the content ranging from 0.7 at% to 1.5 at%. Fig. 2 also shows that the obtained TiO₂ and x at% Ag–TiO₂ semiconductor materials consist of aggregated nanoparticles without any particular morphology.

Fig. 2 SEM image of (a) undoped TiO₂, (b) 0.7 at% Ag–TiO₂, (c) 0.9 at% Ag–TiO₂, (d) 1.0 at% Ag–TiO₂, (e) 1.2 at% Ag–TiO₂, (f) 1.5 at% Ag–TiO₂.

Further, a typical TEM image of undoped TiO₂ and 1.2 at% Ag–TiO₂ were also recorded and are shown in Fig. 3a and b, respectively. The TEM images further confirm that the crystalline sizes for undoped TiO₂ and 1.2 at% Ag–TiO₂ were in the region of 10–40 nm. Fig. 3c depicts the selected area electron diffraction (SAED) pattern of the 1.2 at% Ag–TiO₂. The pattern clearly reveals bright concentric rings, which were due to the diffraction from the (101), (004), (200), and (105) planes of anatase TiO₂. In the HRTEM image of 1.2 at% Ag–TiO₂ (Fig. 3d), lattice fringes with d-spacing values of 3.52 Å, 1.89 Å, and 1.35 Å were observed, which corresponds to the (101), (200), and (220) planes of anatase TiO₂.

Fig. 3 TEM image of (a) undoped TiO₂, (b) 1.2 at% Ag–TiO₂, (c) SAED pattern and (d) HRTEM image of 1.2 at% Ag–TiO₂.

X-ray photoelectron spectroscopy (XPS) was performed on the undoped and doped powders to compare the chemical state of the elements and explore how the doped silver cations interact with TiO₂. The full scan XPS spectra of binding energy in Fig. 4a indicated that the material contained Ti, O, Ag elements together with trace contamination of C. In Fig. 4b, the peak position of the O 1s core levels of 1.0 at% Ag–TiO₂ and 1.5 at% Ag–TiO₂ shift to lower energy compared with those of undoped TiO₂, indicating that the chemical environment of the elements has been changed.^35,36 In x at% Ag–TiO₂, some O–Ti was replaced by lower binding energy bind of O–Ag, therefore, the peak position of the O 1s core levels of x at% Ag–TiO₂ shifts to lower energy, it is more obviously when the doping amount up to 1.5%. Similarly, two peaks of undoped-TiO₂ at 457.9 and 463.6 eV, which were ascribed to Ti 2p_3/2 and Ti 2p_1/2, respectively, also shifted toward low binding energy (Fig. 4c). This phenomenon was attributed to the change of the local chemical environment of Ti ions influenced by Ag incorporation and the formation of Ag–O–Ti bonds on the surface of TiO₂.³⁷ For Ag 3d spectrum (Fig. 4d), two distinct peaks are observed at 367.9 and 373.9 eV. This doublet was assigned to Ag 3d_5/2 and Ag 3d_3/2 levels, respectively, coming from Ag⁺ ions. Although silver species was not observed in the XRD diagrams of x at% Ag–TiO₂ even the concentration of Ag is up to 1.5% (as shown in Fig. 1), it is certain that some Ag ions have successfully substituted Ti ions in the TiO₂ lattice during the solvothermal process, and some Ti–O bonds in TiO₂ have been replaced by Ag–O bonds.

Fig. 4 X-ray photoelectron spectroscopy of undoped-TiO₂, 1.0 at% Ag–TiO₂ and 1.5 at% Ag–TiO₂. (a) Full scan XPS spectrum, (b) O 1s spectrum, (c) Ti 2p spectrum and (d) Ag 3d spectrum.

3.2. Band edge movement analysis of Ag–TiO₂

As important electronic parameters in DSSCs, the message of band edge positions of photoanode material in DSSCs is extremely important because it could be used to estimate the driving force for electron injection. Fig. 5 shows the UV-vis absorption spectroscopy of different samples. In Fig. 5, a broad absorption peak attributed to the intrinsic band gap of TiO₂ (≈3.2 eV) is observed in the region of 300–400 nm, which was associated with the electronic excitation of the valence band O 2p electron to the conduction band Ti 3d level. After doped with Ag, the band edge of this absorption shifts toward long wavelength, which indicates that the band gap is narrowed by doping with Ag. And it is more obvious with the increase doping amount of Ag. Also, the red shift of light adsorption indicated that the band gap energy (E_g) of Ag-doped TiO₂ was reduced. This is because that E_g of semiconductor materials could be calculated using the following formula:where λ is the adsorption wavelength from UV-vis spectroscopy in nanometers. The narrowed band gap and reduced band gap energy are all good for absorbing sunlight and enhance its photovoltaic performance.³⁸

Fig. 5 UV-vis absorption spectra of undoped-TiO₂ and x at% Ag–TiO₂.

As analysis by optical spectra, the narrowed band gap and reduced band gap energy means band edge movement occurs in Ag-doped TiO₂. However, the details and mechanism are still unknown by optical spectra analysis. Therefore, a steady-state spectroelectrochemical measurement was introduced to determine the conduction band edge position and the band edge movement in the semiconductor materials of x at% Ag–TiO₂. Spectroelectrochemistry is a method used to ascertain the conduction band position in semiconductor materials. As supposed by Michael Grätzel, the absorbance at 780 nm is proportional to the density of electrons in the conduction band.³⁹ Therefore, the method of spectroelectrochemistry could firstly determine the concentration of electrons in the conduction band by measuring the absorption of photons with band gap energy in the semiconducting film and then ascertain the conduction band position. The absorbance of undoped TiO₂ and x at% Ag–TiO₂ thin films deposited on an FTO substrate measured at 780 nm as a function of applied potential are shown in Fig. 6. As shown in Fig. 6, the absorbance of undoped TiO₂ and x at% Ag–TiO₂ thin films all follows a rule of exponential functions of applied potential and the absorption improved with the increase of the applied negative potential. More electrons are obtained in Ag-doped TiO₂ films than that of undoped TiO₂ film at the same applied potential (absolute value > 300 mV).

Fig. 6 Optical absorbance at 780 nm of undoped TiO₂ and x at% Ag–TiO₂ photoelectrode films.

The onset of x at% Ag–TiO₂ films are also smaller (absolute value) than that of undoped TiO₂ film. The smaller onset means a lower Fermi level of the x at% Ag–TiO₂, which indicates a positive shift of the electron quasi-Fermi level. Usually, the conduction band edge will shift simultaneously by an equal displacement of quasi-Fermi level relative to the I₃⁻/I⁻ Fermi level, therefore, the conduction band edge of x at% Ag–TiO₂ also shifts toward the positive direction.⁴⁰ This positive shift of conduction band edge is the main contribution to the increase of absorbance for x at% Ag–TiO₂. Meanwhile, combined with optical spectra analysis, it is easy to found that the narrowed band gap of x at% Ag–TiO₂ is also caused by its positive shift of conduction band edge. Herein, band edge movement of Ag-doped TiO₂ is confirmed by spectroelectrochemical measurement, and this spectroelectrochemical measurement also confirms that the decreased bottom of conduction band of TiO₂ after doping with Ag has narrowed the band gap of TiO₂ and improved driving force of injected electron, resulting in increased TiO₂ (e⁻) concentration in conduction band.

Furthermore, in order to quantitative analysis the effect of conduction band edge movement of TiO₂ after doping with Ag on TiO₂ (e⁻) concentration, another steady-state spectroelectrochemical measurement of 1.2 at% Ag–TiO₂ with variable temperature was also applied. Fig. 7 shows the exponential distribution of the TiO₂ (e⁻)s extracted from the variable temperature spectroelectrochemical data and are plotted as the chemical capacitance versus applied potential. In Fig. 7, the area below the exponential distribution curve stands for charge quantity in the conduction band of 1.2 at% Ag–TiO₂ and undoped TiO₂ at different temperature. It was obvious that the charge quantity in the conduction band of 1.2 at% Ag–TiO₂ was larger than that of undoped TiO₂ at any testing temperature. This is consistent with the results of spectroelectrochemical measurement above that Ag doping could reduce the bottom of conduction band of TiO₂, increase TiO₂ (e⁻) concentration and absorbance in conduction band. From Fig. 7 it could also be found that the charge quantity in the conduction band for both 1.2 at% Ag–TiO₂ and undoped TiO₂ increased with the raise of temperature, which could be easily interpreted by the relationship of TiO₂ (e⁻) concentration and temperature as predicted by eqn (2).

where k_B is Boltzmann's constant, T is test temperature, E_cb is the energy of the conduction band edge, E_F is the Fermi level energy and N_c is the states density of effective conduction band.

Fig. 7 Chemical capacitance for TiO₂ and 1.2 at% Ag–TiO₂ thin-film electrodes immersed in the standard electrolyte with 100 mM LiClO₄, for TiO₂ (e⁻) as a function of applied potential.

To obtain the quantity data of the band edge movement in 1.2 at% Ag–TiO₂, the absorbance (A) for TiO₂ and 1.2 at% Ag–TiO₂ with different applied potential were plotted as a function of reciprocal temperature and shown in Fig. 8. The solid lines in Fig. 8 represent the results of a linear fit to eqn (3) for the data obtained at potentials between −0.45 and −0.70 V vs. Ag/AgCl.

Fig. 8 Plot of ln(A) vs. T⁻¹ each line represents a different potential from −0.45 to −0.70 V vs. Ag/AgCl in 50 mV increments from bottom to top. (a) Undoped TiO₂ and (b) 1.2 at% Ag–TiO₂.

As shown in Fig. 8, the absorbance increased with the increase of both temperature (moving from right to left of the plot) and applied potential (moving from down to up in the figure). From Fig. 8 and eqn (3), it can be conclude that the E_cb value of TiO₂ and 1.2 at% Ag–TiO₂ at different applied potentials could be calculated by the slope of lines in Fig. 8. Therefore, to get the E_cb value and further illustrate the band edge movement of Ag-doped TiO₂, a global fit to all the data was performed to get the E_cb value at different applied potentials and the fit results are collected in Table 2. The fit results shown in Table 2 also indicate that the absolute values of E_cb increased as the increase of applied potential for both TiO₂ and 1.2 at% Ag–TiO₂. What's more, E_cb of 1.2 at% Ag–TiO₂ is lower than that of undoped TiO₂ at any applied potentials. This semi-quantitative analysis of variable temperature spectroelectrochemical measurement further confirms that conduction band edge of TiO₂ shift positively after doping with Ag, which also verified the qualitative analysis results of conduction band edge movement by optical absorption and fixed temperature spectroelectrochemical measurement.

Table 2 Fit results of absorbance vs. temperature to eqn (3)

EF (mV vs. Ag/AgCl)	Ecb/q (mV vs. Ag/AgCl)
	1.2 at% Ag–TiO2	TiO2
−450	−802.37	−790.31
−500	−814.92	−823.73
−550	−864.66	−856.89
−600	−890.25	−882.62
−650	−920.06	−910.77
−700	−949.31	−940.85

---

3.3. Photovoltaic performance of DSSCs based on Ag–TiO₂

In order to figure out the influence of Ag doping on the photoelectric properties of TiO₂, different photoanodes based on TiO₂ and x at% Ag–TiO₂ were prepared and employed into N719 sensitized solar cells. The current–voltage (J–V) characteristic of the DSSCs devices based on different photoanodes under illumination (AM1.5G, 100 mW cm⁻²) are shown in Fig. 9, and the corresponding cells performance are summarized in Table 3. As shown in Fig. 9 and Table 3, the V_oc decreases after doping with Ag at any amount levels. The J_sc increases when the TiO₂ semiconductor doped with Ag, reaching a limit coinciding with an initial Ag quantity of 1.2 at%, after which the effect is reversed and the J_sc begins to fall. The decrease in efficiency at a high Ag loading (1.5 at%) was due to the free standing/excess Ag in the composite, which could oxidize to Ag(I) and erode to the redox electrolyte.⁴¹ The oxidation of the Ag would have acted as a new recombination center, thus reducing the number of charge carriers, which led to a decrease in the J_sc and V_oc. Consequently, the overall conversion efficiency of the DSSC deteriorated. However, DSSC based on 1.2 at% Ag–TiO₂ photoanode exhibited a J_sc of 19.75 mA cm⁻², a V_oc of 0.73 V, and a FF of 0.57 with the energy conversion efficiency (η) of 7.34%, indicating a 57% and 30% increase in J_sc and η respectively than the DSSC based on undoped TiO₂ photoanode, which gives a J_sc of 12.61 mA cm⁻², a V_oc of 0.75 V, and a FF of 0.59 with a η of 5.64%.

Fig. 9 Current density versus voltage curves of the DSSCs based on undoped TiO₂ and x at% Ag–TiO₂ photoanodes.

Table 3 Photoelectric properties and dye loading amount of different DSSCs

Photoanodes	Jsc (mA cm^−2)	Voc (V)	FF	η (%)	Dye loading amount (10^−7 mol cm^−2)
Undoped-TiO2	12.61	0.75	0.59	5.64	1.21
0.7 at% Ag–TiO2	16.93	0.73	0.60	6.58	1.20
0.9 at% Ag–TiO2	17.54	0.74	0.60	6.90	1.21
1.0 at% Ag–TiO2	18.73	0.73	0.55	7.11	1.21
1.2 at% Ag–TiO2	19.75	0.73	0.57	7.34	1.20
1.5 at% Ag–TiO2	14.96	0.73	0.62	6.75	1.21

---

The decrease in the V_oc is due to the conduction band edge movement of Ag–TiO₂, as discussed above, the movement of conduction band edge potential (E_cb) of Ag-doped TiO₂ decreases the energy gap between the Fermi level (E_F) and the potential of the I⁻/I₃⁻ redox species (E_red) in the electrolyte, resulting in the phenomenon that the V_oc of Ag-doped TiO₂-based DSSC decreased. The V_oc value can be described as eqn (4):

V_oc = |E_F − E_red| (4)

Assuming that E_red of I⁻/I₃⁻ in the electrolyte does not change with the dopant, it is anticipative that the V_oc relies on the E_F, which is related to the E_cb position. So it is definite that the lower conduction band edge of TiO₂ results in less V_oc.⁴⁰

Although dye loading plays an important role in determine the photo-current density in DSSCs, the increase of J_sc in this study is not due the amount of dye adsorbed in the semiconductor films because the adsorption does not changed in the Ag-doped semiconductors, as shown in Table 3. Therefore, the increase of J_sc is due to several causes. First, as shown in Scheme 1, the conduction band edge moves to lower position after silver doping. This indicates the driving force for electron injection increases, which enhances the electron injection efficiency from the LUMO of the dye to the conduction band of TiO₂. In addition, the inclusion of Ag nanoparticles resulted in a change in the Fermi energy level. The electrons in the conduction band of the TiO₂ could be effectively captured by the Ag until a Fermi level equilibrium was obtained and the charge recombination process was minimized, which improved the DSSC performance. What's more, the increased J_sc is confirmed by IPCE response of cells, since they are related with the eqn (5):

where e is the elementary charge and ϕ_ph.AM1.5G is the photon flux at AM1.5G, 100 mW cm⁻² irradiation.⁴² Fig. 10 collected the IPCE spectra of different devices. As shown in Fig. 10, upon doping TiO₂ with Ag, the IPCE spectrum of cells are enhanced in the region of 300–750 nm, which contributes to the increase of J_sc.

Scheme 1 Schematic on operation principles and photoinduced charge transfer process of DSSCs containing Ag–TiO₂.

Fig. 10 IPCE spectra of the DSSCs based on different photoelectrodes.

In order to illustrate the minimized charge recombination process in DSSCs based on Ag-doped TiO₂, dark current measurement was firstly used to characterize the prepared DSSCs devices. Dark current measurement in DSSC can not be related directly to the back electron transfer process, since the electrolyte concentration in the films and the potential distribution across the nanoporous electrode in dark are different than those under illumination.⁴³ However, a comparison of dark current between the investigated cells can provide useful information regarding the back electron transfer process. Therefore, dark current measurement of DSSCs has been considered as a qualitative technique to describe the extent of the back electron transfer.⁴⁴ Fig. 11 shows the dark current–voltage characteristics of the DSSCs based on different photoelectrodes with the applied bias from 0 to +0.70 V. By comparing the curves in Fig. 11, it is found that when the concentration of Ag in the semiconductor is ≤1.2 at%, the dark current of the DSSCs based on x at% Ag–TiO₂ increased much slower than that of DSSC based on undoped TiO₂ when potential was greater than +0.25 V. In other words, under the same potential bias, when the potential was ≥0.25 V, the dark current for the DSSCs based on x at% Ag–TiO₂ was noticeably smaller than that for the DSSC based on undoped TiO₂. However, it is different when the concentration of Ag in the semiconductor is up to 1.5 at%, which indicates that a proper doping amount of silver in TiO₂ semiconductor can successfully suppress the electron back reaction with oxide dye cation and/or the oxidized form of the redox shuttle, resulting in enhanced short circuit current density.

Fig. 11 Dark current of the DSSCs based on different photoelectrodes.

EIS analysis was also used to study the interfacial charge transfer process in DSSCs based on undoped-TiO₂ and x at% Ag–TiO₂, respectively. The measurements were scanned from 0.1 to 100 kHz at room temperature with an applied bias voltage of 0.75 V. The Nyquist plots for the devices under dark condition are shown in Fig. 12a. The diameters of the medium-frequency semicircle increased when the concentration of Ag in the semiconductor is ≤1.2 at%, implying that the recombination reaction between the conduction band electrons in TiO₂ film and electrolyte is better inhibited by doping with silver.⁴⁵ While, when the concentration of Ag in the semiconductor exceeds 1.2 at%, the diameters of this semicircle decreased is also due to the oxidation of the Ag caused by excessive silver, which accelerate the recombination of injected electrons. This is consistent with the result of dark current measurement. The Nyquist plots for the devices under standard AM1.5G solar irradiation are shown in Fig. 12b. The large medium-frequency semicircles are assigned to the charge transfer processes at TiO₂/dye/electrolyte interface, whose radius are decreased after doping with ≤1.2 at% Ag, suggesting a decrease of the electron transfer impedance at this interface. However, when the doping amount exceeds 1.2 at%, the electron transfer impedance at this interface increased due to the formation of more recombination centers in the semiconductor. The increase of charge recombination impedance and the decrease of electron transfer impedance in DSSCs are all benefit for performance improving.

Fig. 12 Nyquist plots of DSSCs based on different photoelectrodes measured (a) in dark, (b) under standard AM1.5G solar irradiation.

Furthermore, the OCVD technique has been employed as a powerful tool to study interfacial recombination processes in the TiO₂ DSSCs between photo-injected electrons and the electrolyte in the dark.^46,47 It can provide some quantitative information on the electron recombination rate. Fig. 13 shows the OCVD decay curves of the DSSCs based on different photoelectrodes. It was observed that the OCVD response of the DSSCs with x at% Ag–TiO₂ (x ≤ 1.2) photoelectrode were much slower than that with undoped-TiO₂ photoelectrode, especially in the shorter time domain (within 30 s). Since the decay of the V_oc reflects the decrease in the electron concentration, which is mainly caused by the charge recombination,⁴⁸ the cell with x at% Ag–TiO₂ (x ≤ 1.2) photoelectrode has a lower electron recombination rate than that of the cell with undoped-TiO₂ photoelectrode. This further confirmed the results of EIS analysis in dark conditions.

Fig. 13 Open-circuit voltage decay curves of the DSSCs based on different photoelectrodes.

4. Conclusions

Ag-doped TiO₂ was prepared by an easy hydrothermal route and employed into ruthenium dye N719 sensitized solar cells. It was found that the Ag is incorporated into the structure of the TiO₂. After doping with Ag, the conduction band edge moves to lower position. This positive shift of conduction band edge implies a decrease in the open-circuit voltage of DSSCs. However, lower movement of conduction band edge provides stronger driving force for electrons injection from the LUMO of the dye to the conduction band of TiO₂. The deposition of Ag not only benefits for efficient charge transfer but also could effectively suppress the charge recombination between injected electrons and oxide dye cation and/or the oxidized form of the redox shuttle, resulting in a significant photocurrent enhancement. The DSSC based on 1.2 at% Ag–TiO₂ photoanode exhibited a J_sc of 19.75 mA cm⁻², a V_oc of 0.73 V, and a FF of 0.57 with the energy conversion efficiency (η) of 7.34%, indicating a 57% and 30% increase in J_sc and η respectively than that of DSSC based on undoped TiO₂ photoanode.

Acknowledgements

This work was supported by National Natural Science Foundation of China (Grant 21571042, 21171044 and 21371040) and the National key Basic Research Program of China (973 Program, 2013CB632900).

References

1. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737–740.
2. M. Grätzel, Acc. Chem. Res., 2009, 42, 1788–1798.
3. S. K. Balasingam, M. Lee, M. G. Kang and Y. Jun, Chem. Commun., 2013, 49, 1417–1487.
4. P. P. Das, S. Mukhopadhyay, S. A. Agarkar, A. Jana and P. S. Devi, Solid State Sci., 2015, 48, 237–243.
5. C. K. Dong, W. C. Xiang, F. Z. Huang, D. C. Fu, W. C. Huang, U. Bach, Y. B. Cheng, X. Li and L. Spiccia, Nanoscale, 2014, 6, 3704–3711.
6. E. L. Unger, S. J. Fretz, B. Lim, G. Y. Margulis, M. D. McGehee and T. D. P. Stack, Phys. Chem. Chem. Phys., 2015, 17, 6565–6571.
7. L. Tu, H. Pan, H. Xie, A. Yu, M. Xu, Q. Chai, Y. Cui and X. Zhou, Solid State Sci., 2012, 14, 616–621.
8. T. Chen, S. Wang, Z. Yang, Q. Feng, X. Sun, L. Li, Z. S. Wang and H. Peng, Angew. Chem., Int. Ed., 2011, 50, 1815–1819.
9. Z. S. Wang, Y. Cui, Y. Dan-oh, C. Kasada, A. Shinpo and K. Hara, J. Phys. Chem. C, 2008, 112, 17011–17017.
10. K. H. Ko, Y. C. Lee and Y. J. Jung, J. Colloid Interface Sci., 2005, 283, 482–487.
11. Y. D. Duan, J. Zheng, M. Xu, X. Song, N. Fu, Y. Fang, X. Zhou, Y. Lin and F. Pan, J. Mater. Chem. A, 2015, 3, 5692–5700.
12. S. K. M. Saad, A. A. Umar, M. Y. A. Rahman and M. M. Salleh, Appl. Surf. Sci., 2015, 353, 835–842.
13. S. Lee, J. H. Noh, H. S. Han, D. K. Yim, D. H. Kim, J. K. Lee, J. Y. Kim, H. S. Jung and K. S. Hong, J. Phys. Chem. C, 2009, 113, 6878–6882.
14. X. Lü, X. Mou, J. Wu, D. Zhang, L. Zhang, F. Huang, F. Xu and S. Huang, Adv. Funct. Mater., 2010, 20, 509–515.
15. B. Zhao, J. Wang, H. Li, H. Wang, X. Jia and P. Su, Phys. Chem. Chem. Phys., 2015, 17, 14836–14842.
16. A. M. Bakhshayesh and N. Bakhshayesh, J. Colloid Interface Sci., 2015, 460, 18–28.
17. H. Tian, L. Hu, C. Zhang, W. Liu, Y. Huang, L. Mo, L. Guo, J. Sheng and S. Dai, J. Phys. Chem. C, 2010, 114, 1627–1632.
18. A. Malik, S. Hameed, M. J. Siddiqui, M. M. Haque, K. Umar, A. Khan and M. Muneer, J. Mater. Eng. Perform., 2014, 23, 3184–3192.
19. K. P. Wang and H. Teng, Phys. Chem. Chem. Phys., 2009, 11, 9489–9496.
20. W. W. Xu, S. Y. Dai, L. H. Hu, L. Y. Liang and K. J. Wang, Chin. Phys. Lett., 2006, 23, 2288–2291.
21. J. T. Park, W. S. Chi, H. Jeon and J. H. Kim, Nanoscale, 2014, 6, 2718–2729.
22. Z. Liu, Y. Li, C. Liu, J. Ya, L. E. W. Zhao, D. Zhao and L. An, ACS Appl. Mater. Interfaces, 2011, 3, 1721–1725.
23. A. R. Tanyi, A. I. Rafieh, P. Ekaneyaka, A. L. Tan, D. J. Young, Z. Zheng, C. Vijila, G. S. Subramanian and R. L. N. Chandrakanthi, Electrochim. Acta, 2015, 178, 240–248.
24. D. H. Kim, G. S. Han, W. M. Seong, J. W. Lee, B. J. Kim, N. G. Park, K. S. Hong, S. Lee and H. S. Jung, ChemSusChem, 2015, 8, 2392–2398.
25. M. Yang, R. Guo, K. Kadel, Y. Liu, K. O'Shea, R. Bone, X. Wang, J. He and W. Li, J. Mater. Chem. A, 2014, 2, 19616–19622.
26. J. J. Mock, M. Barbic, D. R. Smith, D. A. Schullz and S. J. Schullz, J. Chem. Phys., 2002, 116, 6755–6759.
27. C. Wen, K. Ishikawa, M. Kishima and K. Yamada, Sol. Energy Mater. Sol. Cells, 2000, 61, 339–351.
28. Y. Gao, P. Fang, F. Chen, Y. Liu, Z. Liu, D. Wang and Y. Dai, Appl. Surf. Sci., 2013, 265, 796–801.
29. K. C. Lee, S. J. Lin, C. H. Lin, C. S. Tsai and Y. J. Lu, Surf. Coat. Technol., 2008, 202, 5339–5342.
30. S. P. Lim, A. Pandikumar, H. N. Lim, R. Ramaraj and N. M. Huang, Sci. Rep., 2015, 5, 11922.
31. S. H. Hwang, H. Song, J. Lee and J. Jang, Chem.–Eur. J., 2014, 20, 12974–12981.
32. P. Y. Reddy, L. Giribabu, C. Lyness, H. J. Snaith, C. Vijaykumar, M. Chandrasekharam, M. Lakshmikantam, J. H. Yum, K. Kalyanasundaram, M. Grätzel and M. K. Nazeeruddin, Angew. Chem., Int. Ed., 2007, 46, 373–376.
33. N. G. Park, J. van de Lagemaat and A. J. Frank, J. Phys. Chem. B, 2000, 104, 8989–8994.
34. J. T. Tang, X. M. Chen, Y. Liu, W. Gong, Z. S. Peng, T. J. Cai, L. J. Luo and Q. Deng, Solid State Sci., 2013, 15, 129–136.
35. B. Xin, L. Jing, Z. Ren, B. Wang and H. Fu, J. Phys. Chem. B, 2005, 109, 2805–2809.
36. X. Chen, L. Song, L. You and L. Zhao, Appl. Surf. Sci., 2013, 271, 248–252.
37. Y. C. Cao, Z. Y. Zhao, J. Yi, C. S. Ma, D. C. Zhou, R. F. Wang, C. Li and J. B. Qiu, J. Alloys Compd., 2013, 554, 12–20.
38. M. Saif and M. S. A. Abdel-Mottaleb, Inorg. Chim. Acta, 2007, 360, 2863–2874.
39. G. Rothenberger, D. Fitzmaurice and M. Grätzel, J. Phys. Chem., 1992, 96, 5983–5986.
40. J. Yu, Y. L. Yang, R. Q. Fan, L. Li and X. Y. Li, J. Phys. Chem. C, 2014, 118, 8795–8802.
41. G. Zhao, H. Kozuka and T. Yoko, Sol. Energy Mater. Sol. Cells, 1997, 46, 219–231.
42. A. Hagfeldt, G. Boschloo, L. C. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595–6663.
43. A. Zaban, A. Meier and B. A. Gregg, J. Phys. Chem. B, 1997, 101, 7985–7990.
44. S. Ito, P. Liska, P. Comte, R. Charvet, P. Pechy, U. Bach, L. Schmidt-Mende, S. M. Zakeeruddin, A. Kay, M. K. Nazeeruddin and M. Graetzel, Chem. Commun., 2005, 25, 4351–4353.
45. W. Zhang, R. Zhu, X. Liu, B. Liu and S. Ramakrishna, Appl. Phys. Lett., 2009, 95, 043304.
46. J. Bisquert, A. Zaban, M. Greenshtein and I. Mora-Sero, J. Am. Chem. Soc., 2004, 126, 13550–13559.
47. K. Fan, W. Zhang, T. Y. Peng, J. N. Chen and F. Yang, J. Phys. Chem. C, 2011, 115, 17213–17219.
48. H. Yua, S. Q. Zhang, H. J. Zhao, G. Willb and P. Liu, Electrochim. Acta, 2009, 54, 1319–1324.

---