Introduction

Solar energy can solve the problem of energy crisis. Although silicon-based and thin-film solar cells are having good efficiencies, they are costly, which make them away from common people’s reach. Moreover, toxic materials such as cadmium used in thin-film solar cells are harmful. DSSCs are low cost, environment friendly and hence, consumer adaptable solution.

TiO2 has been widely used as a photoanode material in DSSCs, since its discovery by O’Regan and Gratzel (1991) as compared to other wide band gap semiconducting metal oxides such as SnO2 (Bedja et al. 1994; Ferrere et al. 1997), ZnO (Redmond et al. 1994; Rensmo et al. 1997; Rao and Bahadur 1997; Keis et al. 1999), Nb2O5 (Sayama et al. 1998; Guo and Aegerter 1999), CeO2 (Turkovic and Crnjak 1997) and SrTiO3 (El Zayat et al. 1998). The DSSCs are based on the sensitization of wide band gap material to visible and near-infrared light by adsorbed dye molecules to the photoanode. Efforts of researchers are focused on the solar conversion efficiency improvement. Some undesirable reactions like recombination of injected electrons either with the oxidized sensitizer or with the oxidized redox couple at TiO2 surface (Nazeeruddin et al. 2011) result in loss in the cell efficiency.

In order to reduce recombination rate, different research groups are using different nano-composites or bilayered photoanode. Tripathi and Chawla (2014) used a CeO2–TiO2 nano-composite photoanode sensitized with betacyanin natural dye in order to reduce recombination rate by providing inherent energy barrier. Upadhayay et al. (2014) have used natural dye-sensitized CeO2–TiO2 admixed photoanode to reduce the charge recombination rate by providing energy barrier at the interface between the photoanode and electrolyte. According to Mangesh et al. (2009), mixing oxides of rare earth elements, specially CeO2 with TiO2, reduces the recombination rate at the TiO2 interface. For this they have used photocatalytic behavior of CeO2–TiO2 system for the degradation of methylene blue. Greene et al. (2007) tried ZnO–TiO2 core shell nano-rod combination for solar cells. Maheshwari and Venkatchalam (2014) studied enhanced efficiency and improved photocatalytic activity of 1:1 composite mixture of TiO2 nanoparticles and nanotubes in DSSCs. Kang et al. (2002) studied the improvement of the photocurrent and conversion efficiency of DSSCs by homogeneously incorporating large titanium silicate particles (~500 nm long) in the TiO2 film electrodes. Yu et al. (2012) developed a new type of bilayered photoanode with TiO2 and cubic CeO2 nanoparticles as mirror-like scattering thin layers via screen printing technique for DSSCs. Rai et al. (2014) have used CeO2 quantum-dot-functionalized ZnO nano-rod photoanode for DSSC application.

In the present study, our aim is to improve the efficiency of CeO2-based DSSCs. For this, we have used low-cost RB sensitized bilayered TiO2–CeO2 photoanode. The LUMO level of RB dye is at −3.7 eV (Oku et al. 2011), CB of CeO2 is at −0.53 eV (Elaziouti et al. 2014) and CB of TiO2 is at −0.29 eV (Gratzel 2001). Thus, the electron transfer from LUMO level of RB dye to CB of TiO2 is facilitated through CB of CeO2. Also, the electron transfer directly from LUMO level of RB dye to the CB of TiO2 is possible. Moreover, CeO2 acts as a blocking layer to reduce the electron recombination. Although the efficiency is less for low-cost dyes, cost per watt will be reduced, if we achieve moderate efficiency.

Experimental procedure

Bilayered TiO2–CeO2 photoanode fabrication

Nano-titanium dioxide (TiO2) powder was purchased from Nvis Technology. Cerium nitrate (Ce(NO3)3·6H2O) and RB dye were purchased from HPLC, and ammonium hydroxide solution (20 %, NH4OH) was purchased from Thomas baker.

To make TiO2 paste, 0.5 g of TiO2 powder was mixed with 0.4 g ethyl cellulose, 2.5 g anhydrous terpineol and 5 ml ethyl alcohol. The mixture was properly mixed with mortar and pestle to form a uniform paste. Then, this paste was deposited on fluorine-doped tin oxide (FTO) glass using doctor blade method. After 20–30 min of drying, samples were annealed for 1 h at 250 °C.

The nano-crystalline CeO2 powder was prepared using Ce(NO3)3·6H2O and NH4OH, where Ce(NO3)3·6H2O was used as a source of Ce4+ and NH4OH was the precipitant. The precipitated powder obtained after evaporation of water was annealed at 450 °C to obtain nano-crystalline CeO2 powder.

By using above-synthesized CeO2 powder, CeO2 paste was prepared by the similar procedure as that of TiO2 paste was prepared. And the CeO2 film was deposited on the annealed TiO2 layer, using doctor blade method. After 20–30 min of drying, samples were kept for annealing for 1 h at 450 °C. These samples were characterized by SEM (Model No. JEOL–JSM6360-A). TiO2 and CeO2 films were separately characterized by using XRD pattern (Model No. D-8 Advance Bruker AXS, Germany) equipped with a monochromator Cu-Kα radiation source (λ = 1.54 Å) and DRS by using UV–Vis absorption spectrophotometer (Jasco Model: V-670).

Fabrication of solar cell

All annealed films were kept immersed in 0.3 mM RB dye in ethanol for 24 h to adsorb dye on the surface of photoanode. To prepare the counter electrode, the FTO was washed with acetone, water and ethanol. After removing contaminants, carbon-coated counter electrode was prepared on the conductive side of the FTO substrate by using mild flame of candle.

To fabricate the solar cell, few drops of electrolyte solution (iodide/tri-iodide redox mediator) were added to dye-loaded TiO2–CeO2 photoanode before covering it with counter electrode (carbon-coated FTO). Then, both the photoanode and the counter electrodes were clamped together using binder clips. Performance of the cells was studied by using the photocurrent density–voltage (JV) characteristics. The cells were tested for voltage against time for different loads. EIS analysis (Bode plot) was performed for RB-sensitized single-layered TiO2, CeO2 photoanode and bilayered TiO2–CeO2 photoanode. To test the stability of such DSSCs, the V OC and J SC values were measured for 10 days.

Results and discussion

XRD analysis of TiO2 and CeO2

The presence nano-crystalline anatase TiO2 was confirmed by XRD pattern (Fig. 1a). As per Joint Committee on Powder Diffraction Standards (JCPDS 21-1272), there are ten peaks at 2θ = 25.3°, 36.9°, 37.8°, 38.5°, 48.0°, 53.9°, 55.0°, 62.0°, 62.7° and 78.2° with corresponding miller planes at (101), (110), (004), (200), (105), (211), (204), (116), (220) and (215). The presence of nano-crystalline CeO2 was confirmed by XRD pattern (Fig. 1b). As per JCPDS No. 81-0792, there are eight peaks observed at 2θ = 28.6°, 33.1°, 47.5°, 56.4°, 59.1°, 68.1°, 76.8° and 79.1° which are assigned to diffraction from the (111), (200), (220), (311), (222), (400), (331) and (420) planes, respectively, for cubic structure of CeO2.

Fig. 1
figure 1

XRD pattern of a TiO2 b CeO2 samples

Full size image

The average crystalline size (D) of the TiO2 and CeO2 was estimated by using Scherrer formula (Azaroff 1968).

$$D = \frac{0.89\lambda }{\beta \cos \theta }$$
(1)

where λ is wavelength of X-rays (λ = 1.54 Å), β is full width in radians at half maximum of diffraction peaks, and θ is the Bragg’s angle of the X-ray pattern at maximum intensity. The crystalline size of TiO2 and CeO2 was found as ~24 and ~10 nm, respectively.

Optical properties of TiO2 and CeO2 films

The DRS absorption data obtained from UV–Vis spectrophotometer is used to calculate the band gap. As shown in Fig. 2, the band gap of TiO2 (annealed at 250 °C) and CeO2 (annealed at 450 °C) films was found to be 3.20 and 3.11 eV, respectively. Figure 2 shows the absorption peaks nearly at 320 and 350 nm for TiO2 and CeO2, respectively.

Fig. 2
figure 2

Optical absorption spectra and band gap of TiO2 and CeO2 films

Full size image

SEM and EDS of TiO2–CeO2 photoanode

According to Gratzel (2003), the heart of the DSSC system is a mesoporous oxide layer composed of nanometer-sized particles which have been sintered together to allow for electronic conduction to take place. The surface morphology plays the most important role in dye adsorption in DSSCs. The SEM micrographs, as shown in Fig. 3, are obtained to study the surface morphology of TiO2–CeO2 photoanode. The sample possesses network of aggregated sphere like morphology with pores of size roughly in the range 60–80 nm. The morphology of photoanode is porous and rough, useful for maximum dye adsorption. EDS confirms the presence of Ti and Ce (Fig. 3), which supports the XRD results.

Fig. 3
figure 3

SEM images and EDS of bilayered TiO2–CeO2 photoanode

Full size image

JV characteristics of DSSC

A charge transfer process in DSSC based on RB-sensitized TiO2–CeO2 photoanode can be explained in similar way as explained by Rai et al. (2014). The charge transfer process is shown in Fig. 4. Initially, by absorbing the photons the RB dye gets sensitized. And electrons from highest occupied molecular orbital (HOMO) get transferred to excited state, i.e., LUMO. The CB position of CeO2 (−0.53 eV) (Elaziouti et al. 2014) lies below LUMO level of RB dye (−3.7 eV) (Oku et al. 2011). Hence, the electrons in LUMO level are injected quickly into the CB of CeO2, get transferred to the CB of TiO2 (−0.29 eV) (Gratzel 2001) and finally transferred to FTO substrate where they are utilized for the conduction. Since CB of CeO2 is higher than that of TiO2, the layer of CeO2 acts as blocking layer for electrons to recombine either with dye or with electrolyte. The corresponding mechanism of transportation of electrons is represented according to Nazeeruddin et al. (2011) by Eqs. (2)–(5).

Fig. 4
figure 4

Schematic of processes involved in RB-sensitized bilayered TiO2–CeO2 -based DSSC

Full size image
$$S_{\text{adsorbed}} + h\nu\rightarrow S_{\text{adsorbed}}^{*}$$
(2)
$$S_{\text{adsorbed}}^{*} \to S_{\text{adsorbed}}^{ + } + {\text{e}}_{\text{injected}}^{ - }$$
(3)
$$I_{3}^{ - } + 2 {\text{e}}_{\text{cathode}}^{ - } \to 3 I_{\text{cathode}}^{ - }$$
(4)
$$S_{\text{adsorbed}}^{ + } + \frac{3}{2}I^{ - } \to S_{\text{adsorbed}} + \frac{1}{2}I_{3}^{ - }$$
(5)
$$S_{\text{adsorbed}}^{ + } + {\text{e}}_{{{\text{TiO}}_{2} }}^{ - } \to S_{\text{adsorbed}}$$
(6)
$$I_{3}^{ - } + 2 {\text{e}}_{{{\text{TiO}}_{2} }}^{ - } \to 3 I_{\text{anode}}^{ - }$$
(7)

where \(S_{\text{adsorbed}}\), \(S_{\text{adsorbed}}^{*}\) and \(S_{\text{adsorbed}}^{ + }\) correspond to ground state, excited and oxidized molecules of RB dye, respectively.

The injected electrons diffuse through CeO2–TiO2 porous network and transfer through external load toward counter electrode. The oxidized dye is quickly reduced back to its original state by reduced redox species \(I^{ - }\) in the electrolyte which in turn become the oxidized redox species \(I_{3}^{ - }\).

Some undesirable reactions may take place in case of single-layered wide band gap material resulting in losses in the cell efficiency. They are the recombination of the injected electrons (Fig. 4 dotted lines) either with oxidized sensitizer (Eq. 6) or with the oxidized redox couple at the monolayer wide band gap material surface (Eq. 7). The recombination reactions (Eqs. 6 and 7) are reduced in bilayered TiO2–CeO2 photoanode.

Figure 5 shows the J–V characteristics of DSSCs based on TiO2–CeO2 photoanode. The performance of cell is improved as the recombination reactions are reduced. The cell performance was observed for the optimized dye adsorption time (DAT) of 24 h. The best-performing cell shows V OC ~500 mV and J SC ~0.29 mA/cm2 with fill factor of 62.12 %. Table 1 summarizes the performance of cell under light (30 mW/cm2) and diffused light (0.3 mW/cm2) conditions. As given in Table 2, there is 66.67 and 38.10 % increase in V OC and J SC values, respectively, compared to RB-sensitized CeO2 photoanode. We are getting better results as compared to Turkovic and Crnjak (1997), in which they have discussed ruthenium dye-sensitized CeO2 showing V OC ~60 mV and J SC ~25 nA/cm2. Comparing with other groups (Tripathi and Chawla 2014; Upadhyay et al. 2014; Yu et al. 2012) work on bilayered TiO2–CeO2 or mixed nano-composite photoanode, we are getting less values of V OC and J SC. As we are using low-cost, easily available and environment-friendly RB dye, the cost per unit watt will reduce.

Fig. 5
figure 5

Photocurrent density–voltage (J–V) characteristics of DSSCs based on RB sensitized bilayered TiO2–CeO2 photoanode

Full size image
Table 1 Solar cell parameters for RB-sensitized bilayered TiO2–CeO2 photoanode
Full size table
Table 2 Comparison of solar cell parameter of RB-sensitized TiO2–CeO2 photoanode with RB-sensitized CeO2 photoanode under light (30 mW/cm2)
Full size table

Figure 6 shows the voltage against time graphs for different loads, which indicate that the cells are having good performance. Figure 7 shows the EIS analysis (Bode plot) for RB-sensitized single-layered TiO2, CeO2 and bilayered TiO2–CeO2 photoanode. For bilayered photoanode, there is negative shift in frequency as compared to single-layered photoanode. Thus, there is increase in lifetime of electrons in bilayered photoanode, confirming reduction in recombination reactions. As a result, there is improvement in the performance of the cell. The cells were tested for stability for 10 days, which showed stable performance.

Fig. 6
figure 6

Voltage–time graph at different loads of DSSCs based on RB-sensitized bilayered TiO2-CeO2 photoanode

Full size image
Fig. 7
figure 7

Bode plot for RB-sensitized single-layered TiO2, CeO2 and bilayered TiO2–CeO2 photoanode

Full size image

Conclusion

In conclusion, the RB sensitized bilayered nano-crystalline TiO2–CeO2 photoanode was successfully fabricated and tested. The best performance of cell for 24 h DAT shows V OC ~500 mV and J SC ~0.29 mA/cm2 with FF ~62.12 %. There is 66.67 and 38.10 % increase in V OC and J SC values, respectively, compared to CeO2 photoanode. The performance of cell can be further improved by controlling different parameters. The cells show good performance for different loads and are tested for 10 days having the stable performance.