Strontium titanate nanoparticles as the photoanode for CdS quantum dot sensitized solar cells

Abstract

SrTiO₃ has the potential to be used as the photoanode for quantum dot sensitized solar cells (QDSSCs) since it has very similar band structures to TiO₂ that exhibits the best performance in QDSSCs. In this work, SrTiO₃ nanoparticles (NPs) with a cubic crystal structure were prepared by a hydrothermal method. 120, 70 and 30 nm SrTiO₃ NPs were obtained by using different starting materials. CdS quantum dots (QDs) were coated on three different sized SrTiO₃ NPs using the successive ion layer adsorption and reaction (SILAR) method. QDSSCs based on three different sized SrTiO₃ NPs were fabricated and investigated. Due to the higher surface area compared to larger NPs, the device based on 30 nm SrTiO₃ NPs shows the best open-circuit voltage (V_oc) of 0.76 V and fill factor (FF) of 67% which are relatively high for V_oc and FF reported for CdS-QDSSCs. Comparison studies of light-harvesting efficiency, electron transfer processes and band diagrams were performed to analyze the device performance in the SrTiO₃-QDSSCs and TiO₂-QDSSCs. The 30 nm SrTiO₃ NPs can be combined with TiO₂ NPs (P25) to improve the performance of the CdS-QDSSCs with an increasing percentage of 12.5%, which represents an alternative way to realize high efficiencies in QDSSCs.

---

Introduction

QDSSCs have attracted substantial research interest due to the unique optical and electrical properties of the quantum dots (QDs) as light absorbers for solar cell applications.^1–3 It’s believed that QDSSCs have the possibility of circumventing the Shockley–Queisser efficiency limit for solar cells based on p–n junctions.⁴ In addition, QDs with different sizes also have the capability to match the solar spectrum and improve light absorption. Most recently QDSSCs with efficiencies as high as 8.55% have been achieved.⁵ Various components in QDSSCs can be optimized in the actual process of assembly, the photoanode material as an important area of research has received extensive attention recently. The photoanodes in QDSSCs are usually binary metal oxides, such as TiO₂ or ZnO NPs. In comparison with binary metal oxides, ternary oxides, for example, SrTiO₃, exhibit better corrosion resistance and offer more freedom in tuning of optical and electrical properties.^6,7

SrTiO₃ can be roughly considered as a highly doped anatase TiO₂ because they have similar structures.⁸ SrTiO₃ with a perovskite structure contains titanium atoms in 6-fold octahedral coordination, which is similar to the titanium arrangement in anatase TiO₂. The two materials have exactly the same band gap, but SrTiO₃ has a slightly higher flat band potential, which could lead to higher V_oc for the QDSSCs based on SrTiO₃ compared to that based on TiO₂. With higher flat band potential, SrTiO₃ has the capability of photoassisted electrolysis of water in comparison with anatase TiO₂.⁹ Higher electron mobility in SrTiO₃ (6.35 cm² V⁻¹ s⁻¹)¹⁰ compared to that in TiO₂ (0.1 cm² V⁻¹ s⁻¹) ¹¹ would be favorable for reducing recombination loss and for electron transport. Thus SrTiO₃ is a very promising photoanode in QDSSCs. Burnside et al. evaluated the feasibility of SrTiO₃ NPs with size range of 10–60 nm in the application of dye sensitized solar cells.⁸ However, thus far, there have been few reports on SrTiO₃ applications in QDSSCs.

In this work, ternary oxide SrTiO₃ NPs with different sizes were prepared by a hydrothermal method. Structural characterizations were carried out to examine the size, and crystalline structure of the as-synthesized SrTiO₃ NPs. The QDSSCs based on SrTiO₃ NPs and CdS quantum dots were fabricated and characterized. 30 nm SrTiO₃ NPs show promising device performance. In addition, SrTiO₃ NPs can also be used to improve the device performance of TiO₂-QDSSCs due to the high V_oc and FF of the devices based on SrTiO₃. Thus the first role of SrTiO₃ in this work is that it can be used as the photoanode in QDSSCs. The second role is that SrTiO₃ NPs can also be used to improve the device performance of TiO₂-QDSSCs because TiO₂-QDSSCs have the best device performance till now.

Experimental

Synthesis of SrTiO₃ NPs

SrTiO₃ NPs with different sizes were synthesized by a hydrothermal method following previously published procedures.^8,12,13 120, 70 and 30 nm SrTiO₃ NPs were prepared using titanium dioxide (P25), tetrabutyl titanate, and titanium bis(ammonium lactate) dihydroxide (TALH) as starting materials, respectively. The experimental details can be found in the ESI.†

Solar cell preparation

In the preparation of the SrTiO₃ photoanodes, SrTiO₃ powders were ground in a mortar with a few drops of glacial acetic acid and sufficient amount of ethanol for 20 minutes. Then the ground SrTiO₃ powders were dispersed in 30 mL of ethanol and ultrasonically agitated for 30 minutes to make SrTiO₃ paste. The detailed steps of SrTiO₃ composite paste preparation procedure are described in the ESI (Fig. S1†). After that, the SrTiO₃ paste was spread onto 0.5 × 0.5 cm² (active area 0.25 cm²) conducting FTO glass substrate by the blade method and heated to 150 °C then sintered at 500 °C in a furnace for 30 minutes. The thickness of the photoanode is about 4 μm, which can be found from the cross section scanning electron microscope (SEM) image (Fig. S2†). CdS QDs were coated on the SrTiO₃ photoanode, using the successive ionic layer adsorption and reaction (SILAR) method following the procedure reported previously.¹⁴ Specifically, the SrTiO₃/FTO substrate was successively immersed into 0.05 M Cd(NO₃)₂ in methanol and then into 0.05 M Na₂S in methanol for 2 minutes each. Following each immersion, the substrate was rinsed in pure methanol for 2 minutes to remove excessive precursors and then was dried before the next dipping to finish one CdS coating cycle. The above process was repeated to coat the QDs with different SILAR cycles to get four to eight layers of CdS. QDSSCs were fabricated by clamping a Cu₂S sputtered FTO glass plate onto a CdS QDs coated photoanode and filling the capillary space with the electrolytes (1 M Na₂S + 1 M S, 18 MΩ deionized water). For comparison, the devices based on TiO₂ and SrTiO₃/TiO₂ photoanodes were prepared by the same method.

Measurements

X-ray diffraction (XRD) patterns of all the SrTiO₃ powders were recorded on a RigakuD/max 2550 X-ray diffractometer, using a monochromatized Cu target radiation source at a scanning rate of 4° min⁻¹. SEM images were obtained from a SIRION field-emission scanning electron microscope. The transmission electron microscope (TEM) data were measured on a JEM-2010 with a working voltage of 200 kV. J–V characteristics of the cells were recorded using a Keithley 2400 source meter and a 1.5 AM, 100 mW cm⁻² solar simulator lamp. Incident photon to current conversion efficiency (IPCE) was recorded using a computerized setup consisting of Solar Cell Quantum Efficiency—SolarCellScan100. Absorption spectra were recorded with the Shimadzu UV-1600 Spectrophotometer. Electrochemical impedance spectroscopy (EIS) measurements were performed using an impedance measurement unit (ZAHNER-elektrik IM6) in the frequency range of 0.1–10⁵ Hz, and the applied bias voltage and AC amplitude were set as the open-circuit voltage of the cells and 10 mV between the counter electrode and the working electrode, respectively. The nitrogen adsorption–desorption isotherms were measured at −196 °C with a Gemini VII surface area and porosity system. The specific surface area was estimated by the Brunauer–Emmett–Teller (BET) method.

Results and discussion

Structure and morphology

The crystal structure of the SrTiO₃ NPs was characterized by XRD patterns, as shown in Fig. 1. All the XRD patterns of the SrTiO₃ NPs can be indexed to the cubic structure (JCPDS card 35-734) with well-defined (100), (110), (111), (200), (210), (211), (220) and (310) diffraction peaks. No other crystalline by-products can be observed in these XRD patterns.

Fig. 1 XRD patterns of SrTiO₃ NPs with different sizes.

The XRD data of SrTiO₃ nanoparticles with different sizes look identical, which is consistent with the literature about 20–60 nm SrTiO₃ nanoparticles.^8,12 Only sub-10 nm SrTiO₃ nanoparticles showed the broad XRD peaks.¹³ Thus we believe that the XRD peaks may exhibit the features of broad peaks when the size of SrTiO₃ nanoparticles is below 10 nm.

Fig. 2 shows SEM and TEM images of the as-synthesized SrTiO₃ NPs prepared with different starting materials and the inset shows the size distribution of the corresponding SrTiO₃ NPs. The sizes of the NPs distribute in a wide range for all the samples. From Fig. 2a and b it can be seen that as P25 and Sr(OH)₂ were chosen as starting materials, the products are yielded as cube shaped NPs. The SrTiO₃ NPs have a diameter of 120 ± 32 nm (see inset of Fig. 2a). Fig. 2c shows the SEM image of the SrTiO₃ NPs prepared with tetrabutyl titanate, Sr(NO₃)₂ and PVA. It can be seen that the shapes of the NPs are irregular, and the size of the NPs is 70 ± 8 nm. Fig. 2d shows the corresponding TEM image of Fig. 2c. From the boundary of the 70 nm NPs, it can be observed that they actually consist of a large number of smaller NPs with average size of 10 nm. The size and irregular shape of the as-prepared NPs can be correlated to the additive PVA. As a water-soluble polymer, PVA has been widely used to control the morphology and size in nanomaterial growth.^15,16 Fig. 2e and f present, respectively, the SEM and TEM images of the SrTiO₃ NPs when TALH and Sr(OH)₂·8H₂O are used in the hydrothermal reaction. It is obvious that all the as-prepared NPs are composed of well-defined and relatively regular-shaped particles. The size of SrTiO₃ NPs is 30 ± 8.5 nm, which is comparable to that of TiO₂ (P25).

Fig. 2 SEM and TEM images of SrTiO₃ NPs with different sizes. (a) and (b) 120 nm SrTiO₃ NPs, P25 and Sr(OH)₂ were chosen as starting materials; (c) and (d) 70 nm SrTiO₃ NPs, synthesized using tetrabutyl titanate and Sr(NO₃)₂ and PVA; (e) and (f) 30 nm SrTiO₃ NPs, prepared using TALH and Sr(OH)₂·8H₂O.

Device performance based on CdS QDs (four to eight cycles)

Current density vs. voltage (J–V) characteristics of the devices sensitized by CdS QDs (four to eight cycles) based on different SrTiO₃ NPs (120, 70 and 30 nm) were obtained under AM 1.5 G illumination with a light intensity of 100 mW cm⁻² and in the dark, as shown in Fig. 3a–c, respectively. For the QDSSCs based on 120 nm SrTiO₃ NPs, as shown in Fig. 3a, the short-circuit photocurrent (J_SC) increases from 0.28 to 0.56 mA cm⁻² and V_oc increases from 0.54 to 0.72 V as the number of coating cycles increases from four to seven. However, J_SC and V_oc decrease as the number of coating cycles increases to eight. The same trend is observed for the QDSSCs based on 70 and 30 nm SrTiO₃ NPs (Fig. 3b and c). J_SC increases from 0.52 to 0.94 mA cm⁻² and V_oc increases from 0.42 to 0.74 V as the number of coating cycles increases from four to seven for the QDSSC based on 70 nm SrTiO₃ NPs. J_SC increases from 0.76 to 1.53 mA cm⁻² and V_oc increases from 0.50 to 0.76 V as the number of coating cycles increases from four to seven for the QDSSC based on 30 nm SrTiO₃ NPs. All devices exhibit the optimum results in the seventh cycle and appear decreased permanently in the eighth cycles. This phenomenon could be explained by the varied IPCE of the QDSSCs and the varied bandgap with CdS particle sizes, which will be discussed in the following text.

Fig. 3 (a) J–V characteristics of QDSSCs based on CdS QDs prepared with different coating cycles and 120 nm SrTiO₃ NPs. (b) J–V characteristics of QDSSCs based on CdS QDs prepared with different coating cycles and 70 nm SrTiO₃ NPs. (c) J–V characteristics of QDSSCs based on CdS QDs prepared with different coating cycles and 30 nm SrTiO₃ NPs.

Fig. 4a shows the IPCE spectra of QDSSCs based on CdS QDs with different coating cycles on 70 nm SrTiO₃ NPs, which could be representative for that based on 120 and 30 nm SrTiO₃ NPs.

Fig. 4 (a) IPCE spectra of QDSSCs based on 70 nm SrTiO₃ NPs and CdS QDs coated using different SILAR cycles. (b) Bandgaps of CdS QDs prepared with different coating cycles and their band alignment with SrTiO₃.

A monotonic increase can be observed as the incident light wavelength is scanned from 600 to 400 nm without exhibiting distinctive excitonic features. The maximum intensity of IPCE increases when the number of the coating cycles increases from four to seven, however, the maximum intensity of IPCE decreases as the number of coating cycles increases to eight, which is well in accordance with the J–V characteristics. The absence of the CdS excitonic features in the IPCE spectra could be explained by the broad size distribution of the SILAR-based QDs.¹⁴ The IPCE spectra also show onset wavelengths shift from 448 nm (four-cycle cell) to 540 nm (eight-cycle cell), which can be attributed to the increased QD size and decreased QD bandgap. According to the literature, the conduction band edge of TiO₂ is ∼4.2 eV below the vacuum level, the conduction band edge of bulk CdS is about 3.98 eV below the vacuum level.^17,18 According to a previous report, SrTiO₃ has the same bandgap of 3.2 eV as TiO₂ but a conduction band edge about 0.2 eV lower than that of TiO₂,⁸ so the conduction band edge of SrTiO₃ is −4.0 eV vs. vacuum. Combining the literature results and IPCE test data, the band diagrams of the SrTiO₃ nanoparticles and CdS QDs are shown in Fig. 4b. Size-dependent band gaps of CdS QDs are estimated by the onset wavelengths of the IPCE spectra (which are due to light absorption of CdS QDs¹⁹). It’s believed that the decrease in the CdS bandgap is mainly caused by a downward shift of the conduction band edge due to the large effective mass of the hole compared to that of the electron.^20,21 From Fig. 4b, it can be observed that when the number of coating cycles increases from four to eight, the conduction band edge difference between CdS QD and SrTiO₃ gradually becomes smaller. The driving force for electron transfer at the QD/SrTiO₃ interface is dependent on the band edge difference. A reduced conduction band difference leads to a decreased driving force and charge injection efficiency.²¹

Actually, as the number of coating cycles increases, on one hand, the light harvesting efficiency of the cell and the quantity of photo-generated electrons gradually increase, leading J_SC to increase; on the other hand, the driving force and charge injection efficiency gradually decrease, leading J_SC to decrease gradually. These two competing mechanisms induce the appearance of an optimum coating cycle (seven cycles) for the J–V response. Note that as the number of coating cycles increases to seven, the light harvesting efficiency of the cell does not increase with the increase of cycle number further, because the size confinement hardly takes efforts any more. In this case, the decreased driving force contributes dominantly to the solar cell performance.

Device performance based on SrTiO₃ NPs with different sizes

Because the QDSSCs based on CdS QDs prepared with seven cycles demonstrate the optimum device permanence, in the following experiments the CdS QDs are all fixed at seven cycles. Fig. 5a shows the influence of SrTiO₃ particle size on the J–V characteristics of seven-cycle cells. It can be seen that when the size decreases from 120 to 30 nm, the short circuit photocurrent density J_SC increases almost three times, from 0.56 to 1.53 mA cm⁻². Concomitantly, the V_oc increases from 0.72 to 0.76 V, and the FF increases from 0.48 to 0.67. The power conversion efficiency increases from 0.19% to 0.78%. The detailed device performance parameters are listed in Table 1. From Fig. 5a, the device with 30 nm SrTiO₃ NPs as the photoanode shows the best device performance of 0.78% with the highest V_oc and FF reported in typical CdS-QDSSCs.^22,23

Fig. 5 (a) J–V characteristics of QDSSCs based on CdS QDs prepared with seven coating cycles and different sized SrTiO₃ NPs. (b) UV-vis absorption spectra of CdS QDs prepared with seven coating cycles on SrTiO₃ NPs with different sizes. The inset shows the photographs of CdS QDs prepared with seven coating cycles on 120, 70 and 30 nm SrTiO₃ NPs. (c) IPCE spectra of QDSSCs based on SrTiO₃ NPs with different sizes and CdS QDs coated using seven SILAR cycles.

Table 1 Parameters obtained from J–V measurements of the QDSSCs prepared with seven coating cycles

Architecture	Voc (V)	JSC (mA cm^−2)	FF	η (%)
SrTiO3 (30 nm)	0.76	1.53	0.67	0.78%
SrTiO3 (70 nm)	0.74	0.94	0.51	0.35%
SrTiO3 (120 nm)	0.72	0.56	0.48	0.19%

---

In order to reveal the origin of device performance changing with SrTiO₃ particle size, the absorption spectra and IPCE measurements based on the three kinds of best performance devices were designed.

The absorption spectra of CdS QDs, which are obtained by subtracting the blank (no QDs) device absorption from a working device absorption, are presented in Fig. 5b. It can be seen that the absorption intensities of CdS QDs increase when the size of SrTiO₃ NPs decreases, indicating that smaller SrTiO₃ NPs have higher QD coverage. A similar phenomenon was also observed for the QDSSCs based on TiO₂.²⁴ The inset of Fig. 5b shows the photographs of different sized SrTiO₃ NPs coated with seven-cycle CdS QDs. The coating of CdS QDs over SrTiO₃ is accompanied by color changes visible to the naked eye. The color becomes darker as smaller SrTiO₃ NPs are utilized for CdS QD coating, which confirms higher QD coverage on smaller SrTiO₃ NPs. Higher QD coverage leads to better device performance due to improved light harvesting efficiency of the cell. Fig. 5c shows the IPCE spectra of seven-cycle cells based on 120, 70 and 30 nm SrTiO₃ NPs. The maximum IPCE values increase from 5.7% to 15.6% as the size of NPs decreases following the same trend of J–V of QDSSCs (Fig. 5a). The IPCE spectra of CdS QDs are consistent with the corresponding absorption spectra, as shown in Fig. 5b. The onset wavelengths are very similar for devices based on the SrTiO₃ NPs of different sizes, which can be observed in both IPCE and absorption spectra (Fig. 5b and c), indicating that the sizes of the QDs are the same for seven-cycle QDs. Although the average size of SrTiO₃ NPs changes, however, the size of QDs is only related to the number of SILAR coating cycles. This explains why different sized SrTiO₃-based QDSSC devices have the same ideal performance appearing in the seventh cycles.

Different intensities in IPCE and absorption spectra can be attributed to different QD adsorptions on SrTiO₃ NPs, and it is the dominant factor to influence the overall performance of the devices.

TiO₂/SrTiO₃ composite-based QDSSC device performance

It was reported that incorporating different materials and using their respective advantages had a great influence on the device performance.^25–27 In order to take advantage of the high V_oc and FF characteristics of the device based on 30 nm SrTiO₃ NPs and the high current characteristics of TiO₂ based QDSSCs. QDSSCs based on the TiO₂/SrTiO₃ composite system at different ratios of TiO₂ to SrTiO₃ were fabricated and compared. Fig. 6a shows J–V curves of QDSSCs based on the TiO₂/SrTiO₃ composites under AM 1.5 G illumination with a light intensity of 100 mW cm⁻². The corresponding solar cell parameters are summarized in Table 2. Compared with pure TiO₂, the QDSSC exhibits higher V_oc and FF but slightly small J_SC when the ratio of TiO₂ to SrTiO₃ is 9 : 1 leading to the energy conversion efficiency being enhanced from 1.6% to 1.8% with an improvement of 12.5%. As the ratio of TiO₂ to SrTiO₃ decreases further, V_oc and FF increase but J_SC decreases, resulting in decreased conversion efficiency.

Fig. 6 (a) J–V characteristics of QDSSCs based on CdS QDs prepared with seven coating cycles and TiO₂/SrTiO₃ composites with different ratios. (b) Nitrogen adsorption–desorption isotherms of 30 nm SrTiO₃ NPs and TiO₂.

Table 2 Parameters obtained from J–V measurements of the QDSSCs based on seven-cycle CdS QDs and TiO₂/SrTiO₃ composites. The average and standard deviation (s.d.) values are obtained from a batch of 6 identically processed TiO₂/SrTiO₃ composite-based QDSSCs

Architecture	Voc (V)	JSC (mA cm^−2)	FF	η (%)
TiO2 (100%) SrTiO3 (0%)	0.58 ± 0.03	6.6 ± 0.30	0.41 ± 0.01	1.6% ± 0.15
TiO2 (90%) SrTiO3 (10%)	0.60 ± 0.02	6.0 ± 0.24	0.48 ± 0.02	1.8% ± 0.18
TiO2 (80%) SrTiO3 (20%)	0.62 ± 0.02	5.0 ± 0.13	0.48 ± 0.01	1.5% ± 0.16
TiO2 (50%) SrTiO3 (50%)	0.70 ± 0.03	3.0 ± 0.15	0.49 ± 0.05	1.0% ± 0.20
TiO2 (20%) SrTiO3 (80%)	0.73 ± 0.05	2.0 ± 0.22	0.56 ± 0.02	0.82% ± 0.15
TiO2 (0%) SrTiO3 (100%)	0.76 ± 0.06	1.53 ± 0.11	0.67 ± 0.03	0.78% ± 0.10

---

Mao et al. reported the enhanced photovoltaic response of the TiO₂/SrTiO₃ composite compared to pure SrTiO₃ and TiO₂.²⁸ Improved separation efficiency of photogenerated carriers due to the matched energy levels of TiO₂ and SrTiO₃ were considered as one possible reason for the enhanced photovoltaic response. Reduced recombination rate of the composite was also proposed to explain the improved photovoltaic response, which was manifested by photoluminescence spectra.²⁸ In our case, the electrons will be injected into both SrTiO₃ and TiO₂ as SrTiO₃ was incorporated into TiO₂. The flat band potential of SrTiO₃ is higher than that of TiO₂, which leads to the higher V_oc of the device based on TiO₂/SrTiO₃ composite. In addition, according to Mao’s results, the recombination rate in the TiO₂/SrTiO₃ composite can be reduced compared to pure TiO₂, which results in larger FF. The power conversion efficiency is determined by J_SC, FF and V_oc. The J_SC of the QDSSCs based on the TiO₂/SrTiO₃ composite is smaller than that for pure TiO₂. However, FF and V_oc can be increased due to the band alignment and reduced recombination rate. Thus it is very possible that the power conversion efficiency of QDSSCs based on TiO₂ can be improved as an appropriate amount of SrTiO₃ is used in the composite.

Meng et al. investigated CdS/CdSe-sensitized solar cells controlled by the structural properties of compact porous TiO₂ photoelectrodes, and achieved a high FF (61–65%). They explained that was due to extending the light path length and decreasing the electron recombination caused by the scattering layer.²⁹ It was also reported that a high FF (56%) could be obtained by using ordered multimodal porous carbon as the counter electrode in QDSSCs, and that was attributed to the low charge transfer resistance and Nernst diffusion impedance of the counter electrodes.³⁰ Emin et al. obtained a high FF of 60% and attributed that to the low series resistances of the FTO/TiO₂ films.³ In our study (see Table 3), the series resistance of FTO/SrTiO₃ (35.1 Ω) is very similar to that of FTO/TiO₂ (35.6 Ω). However, the charge transfer resistance of the device based on SrTiO₃ is 924 Ω which is smaller than that of TiO₂ (1149 Ω). Thus we think the low charge transfer resistance leads to high FF for the device based on SrTiO₃ in comparison with the TiO₂-based device. Kang et al. reported a high V_oc of 0.77 V by reducing the recombination process via a novel photoanode architecture of “pine tree” ZnO nanorods on a Si nanowires hierarchical branched structure.³¹ Self-assembled monolayers were used as recombination barriers to realize high efficiency of the devices and the high V_oc.³² In our recent work, we used rare earth ions to modulate the band gap of the photoanode in dye sensitized solar cells to increase V_oc.³³ We think the main reason for the high V_oc of the devices based on SrTiO₃ is higher flat band potential.

Table 3 Fitting results of the Nyquist plot

Photoanode	Rs (Ω)	R1 (Ω)	CPE1	R2 (Ω)	CPE2
SrTiO3	35.1	924	1.073 × 10^−5	1528	2.318 × 10^−4
TiO2	35.6	1149	1.052 × 10^−5	3874	5.956 × 10^−4

---

J_SC is mainly determined by light-harvesting efficiency and charge injection efficiency. Light-harvesting efficiency is correlated with the adsorption capacity of the photoanode, and adsorption capacity is determined by higher specific surface area. Nitrogen adsorption measurements are designed to measure the specific surface area of samples, the nitrogen adsorption–desorption isotherms based on SrTiO₃ (30 nm) and TiO₂ (P25) samples are plotted in Fig. 6b. The BET surface areas for the SrTiO₃ and TiO₂ are reported to be 40 and 51 m³ g⁻¹ respectively by the nitrogen adsorption–desorption measurements. Thus the adsorption ability of 30 nm SrTiO₃ is lower than that of P25 due to the difference in surface areas. The surface areas of 120 nm and 70 nm SrTiO₃ NPs are 7.1 and 15.7 m³ g⁻¹ respectively. This further explains the lower J–V and IPCE of the devices based on 120 nm and 70 nm SrTiO₃ NPs compared to that based on 30 nm SrTiO₃ NPs shown in Fig. 5a and c.

To study the adsorption ability of SrTiO₃ compared to TiO₂, oxide photoanodes coated with CdS QDs by the SILAR process were put into 50 mL dilute nitric acid for two hours to dissolve the CdS QDs and to produce Cd²⁺. A VA Computrace (TEA 4000) was used to detect the concentration of Cd²⁺, and the results are listed in Table S1.† It can be observed that the concentration of Cd²⁺ increases when the ratio of TiO₂ to SrTiO₃ increases. The concentration of Cd²⁺ taken by SrTiO₃ is lower compared to that of TiO₂, which indicates that SrTiO₃ has lower adsorption ability. Thus the low J_SC of the QDSSC based on SrTiO₃ could be partially attributed to the low adsorption ability of SrTiO₃ compared to TiO₂.

EIS study of SrTiO₃-QDSSCs and TiO₂-QDSSCs

A comparison study of the electron transfer processes in the SrTiO₃-QDSSCs and TiO₂-QDSSCs was carried out by EIS measurements. Typical Nyquist plots of SrTiO₃-QDSSCs and TiO₂-QDSSCs are shown in Fig. 7. The inset shows the equivalent circuit which is used to fit the Nyquist plot according to ref. 34. A smaller semicircle in the high frequency region (around 100 Hz) and a larger semicircle in the low frequency region (around 1 Hz) can be clearly observed. Each semicircle indicating a charge transfer process exhibited by a resistance capacitance (R–C) parallel circuit. For a more precise fitting, the capacitance element is replaced by a constant phase element (CPE). R_s, R₁ and R₂ values are listed in Table 3, which are determined by fitting according to the equivalent circuit. R_s is the device ohmic series resistance, which is contributed by the sheet resistance of the substrates, resistivity of the electrolyte, and electrical contacts of the cell. R₁ and CPE₁ stand for the charge transfer resistance and double layer capacitance at the electrolyte/counter electrode interface.³⁴ R₂ and CPE₂ represents the recombination charge transfer resistance and chemical capacitance at the photoanode/QDs/electrolyte interface.²⁹ The R_s value for the SrTiO₃-QDSSCs is 35.1 Ω, which is similar to that of TiO₂-QDSSCs (35.6 Ω), this indicates that SrTiO₃-QDSSCs and TiO₂-QDSSCs have the same device configuration. The R₁ value for the SrTiO₃-QDSSCs is 924 Ω, which is a little smaller compared to that of TiO₂-QDSSCs (1149 Ω). The R₂ values of the SrTiO₃-QDSSC and TiO₂-QDSSC are 1528 Ω and 3874 Ω respectively. The higher R₂ implies a lower recombination of electrons between the electrolyte and the conduction band of the photoanode than is expected.²⁹ This result reveals that the recombination of electrons in SrTiO₃-QDSSCs is higher than TiO₂-QDSSCs, leading to a lower J_SC in SrTiO₃-QDSSCs.

Fig. 7 Nyquist plots of QDSSCs. The inset is the equivalent circuit. The scattered points are experimental data and the solid lines are the fitting curves.

The driving force analysis between SrTiO₃/TiO₂ and CdS QD

Fig. 8 shows the band diagrams and the estimated conduction band edges of SrTiO₃, TiO₂ and CdS QDs (seven cycles). According to previous discussion and described in Fig. 4b, the conduction band edge of the CdS QDs is estimated as ∼0.1 and ∼0.3 eV higher than that of the SrTiO₃ and TiO₂ respectively, which will make the charge injection from QDs to SrTiO₃ or TiO₂ possible. The conduction band edge of SrTiO₃ is 0.2 eV higher than that of TiO₂, leading to the higher position of the Fermi level in SrTiO₃, which is related to the larger V_oc of the QDSSCs based on SrTiO₃ according to the V_oc analysis in theory as follows.

Fig. 8 Band alignment of SrTiO₃, TiO₂ and CdS QDs prepared with seven cycles. (Note that band positions are for reference only and not drawn to precise scale.)

According to experimental results and theoretical calculations, although the V_oc is determined by many factors inside the cell, the open circuit voltage of QDSSCs can be estimated by:^35,36

V_oc = E_g − Δ₁ − Δ₂,

where E_g is the bandgap of CdS QDs, Δ₁ represents the energy difference between the conduction band of the CdS QDs and the conduction band of the semiconductor oxide (TiO₂ or SrTiO₃); Δ₂ is the energy difference between the oxidized redox potential of the electrolyte and the valence band of the CdS QDs. Compared with the conventional TiO₂, SrTiO₃ has a reduced Δ₁, this will lead to improvement in the V_oc of the device. Therefore, the QDSSCs based on SrTiO₃ have a higher open circuit voltage than that based on TiO₂. However, Δ₁ can determine the driving force for the charge injection from QDs to SrTiO₃. So it will also result in a small driving force for the charge injection from QDs to SrTiO₃. In addition, the charge recombination process happens synchronously when Δ₁ decreases. These two factors would partially contribute to the low device performance compared to TiO₂-QDSSCs. In a word, the low J_SC of SrTiO₃-QDSSCs can be attributed to the smaller surface area of the SrTiO₃ NPs, the smaller driving force due to the higher flat band potential and the smaller recombination charge transfer resistance of the device. Thus both higher V_oc and lower J_SC originate from the flat band potential of SrTiO₃ which leads to the lower device performance.

Conclusions

In summary, in this work, SrTiO₃ NPs with different sizes were synthesized by a hydrothermal method and utilized as the photoanode in QDSSCs. CdS QDs were deposited on the NPs using the SILAR approach. Device performance of the QDSSCs was studied, and the highest V_oc of 0.76 V and FF of 67% were observed for the cells prepared with seven coating cycles on 30 nm SrTiO₃ NPs. IPCE measurements were carried out to inspect CdS bandgaps and alignment with SrTiO₃ NPs. Through nitrogen adsorption measurements, it was proven that 30 nm SrTiO₃ NPs have a surface area of 40 m³ g⁻¹, which is smaller than that of P25 (51 m³ g⁻¹). Improved device performance of TiO₂-QDSSCs was observed as 30 nm SrTiO₃ NPs were incorporated into TiO₂, which can be attributed to the high V_oc of SrTiO₃-QDSSCs caused by their band structure and larger FF due to a reduced recombination rate. That points to a new method to improve the solar cell efficiency. EIS measurements for SrTiO₃-QDSSC and TiO₂-QDSSC were also carried out to study electron transport and recombination processes in the QDSSCs. Smaller recombination charge transfer resistance was observed for the SrTiO₃-QDSSC compared to the TiO₂-QDSSC, which can result in lower device performance. The first scientific significance of QDSSC fabrication with the SrTiO₃ is SrTiO₃ nanoparticles can be used as the photoanode for QDSSCs, and the devices based on SrTiO₃ show high V_oc and FF compared to other QDSSCs. The second one is that TiO₂/SrTiO₃ composites were developed to improve the TiO₂-QDSSC device performance by using the high V_oc and FF of SrTiO₃-QDSSCs.

Acknowledgements

This work was supported by the Major State Basic Research Development Program of China (973 Program) (no. 2014CB643506), the National Natural Science Foundation of China (Grant no. 11374127, 11304118, 61204015, 81201738, 81301289, 61177042, and 11174111), Program for Chang Jiang Scholars and Innovative Research Team in University (no. IRT13018). The China Postdoctoral Science Foundation Funded Project (2012M511337 and 2013T60327).

Notes and references

1. D.-M. Li, L.-Y. Cheng, Y.-D. Zhang, Q.-X. Zhang, X.-M. Huang, Y.-H. Luo and Q.-B. Meng, Sol. Energy Mater. Sol. Cells, 2014, 120, 454–461.
2. G. Seo, J. Seo, S. Ryu, W. Yin, T. K. Ahn and S. I. Seok, J. Phys. Chem. Lett., 2014, 5, 2015–2020.
3. S. Emin, M. Yanagida, W. Peng and L. Han, Sol. Energy Mater. Sol. Cells, 2012, 101, 5–10.
4. O. E. Semonin, J. M. Luther, S. Choi, H.-Y. Chen, J. Gao, A. J. Nozik and M. C. Beard, Science, 2011, 334, 1530–1533.
5. C.-H. M. Chuang, P. R. Brown, V. Bulović and M. G. Bawendi, Nat. Mater., 2014, 13, 796–801.
6. B. Tan, E. Toman, Y. Li and Y. Wu, J. Am. Chem. Soc., 2007, 129, 4162–4163.
7. Y. Zhang, H. Zhang, Y. Wang and W. Zhang, J. Phys. Chem. C, 2008, 112, 8553–8557.
8. S. Burnside, J.-E. Moser, K. Brooks, M. Grätzel and D. Cahen, J. Phys. Chem. B, 1999, 103, 9328–9332.
9. H. O. Finklea, Semiconductor electrodes, Elsevier Science Ltd, 1988.
10. R. Moos, W. Menesklou and K. H. Härdtl, Appl. Phys. A, 1995, 61, 389–395.
11. P. Tiwana, P. Docampo, M. B. Johnston, H. J. Snaith and L. M. Herz, ACS Nano, 2011, 5, 5158–5166.
12. X. Wei, G. Xu, Z. Ren, C. Xu, G. Shen and G. Han, J. Am. Ceram. Soc., 2008, 91, 3795–3799.
13. K. Fujinami, K. Katagiri, J. Kamiya, T. Hamanaka and K. Koumoto, Nanoscale, 2010, 2, 2080–2083.
14. H. Lee, M. Wang, P. Chen, D. R. Gamelin, S. M. Zakeeruddin, M. Graetzel and M. K. Nazeeruddin, Nano Lett., 2009, 9, 4221–4227.
15. H. Yu, X. Xu, X. Chen, T. Lu, P. Zhang and X. Jing, J. Appl. Polym. Sci., 2007, 103, 125–133.
16. D. Guin, S. V. Manorama, J. N. L. Latha and S. Singh, J. Phys. Chem. C, 2007, 111, 13393–13397.
17. H. Lee, H. C. Leventis, S.-J. Moon, P. Chen, S. Ito, S. A. Haque, T. Torres, F. Nueesch, T. Geiger, S. M. Zakeeruddin, M. Graetzel and M. K. Nazeeruddin, Adv. Funct. Mater., 2009, 19, 2735–2742.
18. T. Lv, L. Pan, X. Liu, T. Lu, G. Zhu, Z. Sun and C. Q. Sun, Catal. Sci. Technol., 2012, 2, 754–758.
19. Q. Dai, J. Chen, L. Lu, J. Tang and W. Wang, Nano Lett., 2012, 12, 4187–4193.
20. K. Tvrdy, P. A. Frantsuzov and P. V. Kamat, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 29–34.
21. G. Rothenberger, D. K. Negus and R. M. Hochstrasser, J. Chem. Phys., 1983, 79, 5360–5367.
22. T. Bora, H. H. Kyaw and J. Dutta, Electrochim. Acta, 2012, 68, 141–145.
23. Q. Zhang, Y. Zhang, S. Huang, X. Huang, Y. Luo, Q. Meng and D. Li, Electrochem. Commun., 2010, 12, 327–330.
24. A. Kongkanand, K. Tvrdy, K. Takechi, M. Kuno and P. V. Kamat, J. Am. Chem. Soc., 2008, 130, 4007–4015.
25. B. C. O’Regan, S. Scully, A. C. Mayer, E. Palomares and J. Durrant, J. Phys. Chem. B, 2005, 109, 4616–4623.
26. A. Kay and M. Gratzel, Chem. Mater., 2002, 14, 2930–2935.
27. K. Tennakone, G. K. R. Senadeera, V. P. S. Perera, I. R. M. Kottegoda and L. A. A. De Silva, Chem. Mater., 1999, 11, 2474–2477.
28. Y. T. Zheng, Z. L. Zhang and Y. L. Mao, J. Alloys Compd., 2013, 554, 204–207.
29. Q. Zhang, X. Guo, X. Huang, S. Huang, D. Li, Y. Luo, Q. Shen, T. Toyoda and Q. Meng, Phys. Chem. Chem. Phys., 2011, 13, 4659–4667.
30. S.-Q. Fan, B. Fang, J. H. Kim, B. Jeong, C. Kim, J.-S. Yu and J. Ko, Langmuir, 2010, 26, 13644–13649.
31. P. Sudhagar, T. Song, D. H. Lee, I. Mora-Seró, J. Bisquert, M. Laudenslager, W. M. Sigmund, W. I. Park, U. Paik and Y. S. Kang, J. Phys. Chem. Lett., 2011, 2, 1984–1990.
32. P. Ardalan, T. P. Brennan, J. R. Bakke and S. F. Bent, in Photovoltaic Specialists Conference (PVSC), 2010 35th IEEE, 2010, pp. 000951–000954.
33. C. Miao, C. Chen, Q. Dai, L. Xu and H. Song, J. Colloid Interface Sci., 2015, 440, 162–167.
34. Z. Huang, G. Natu, Z. Ji, P. Hasin and Y. Wu, J. Phys. Chem. C, 2011, 115, 25109–25114.
35. H. J. Snaith, Adv. Funct. Mater., 2010, 20, 13–19.
36. J. Zhang, H.-B. Li, S.-L. Sun, Y. Geng, Y. Wu and Z.-M. Su, J. Mater. Chem., 2012, 22, 568–576.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra11960f

---