Monodisperse TiO₂ mesoporous spheres with core–shell structure: candidate photoanode materials for enhanced efficiency dye sensitized solar cells

Abstract

We report a facile synthesis of monodisperse TiO₂ spheres with controllable internal structure and surface areas of 17.9–348.8 m² g⁻¹, which were assembled from porous nanoparticles with average diameters of about 7–16 nm. The internal structures of the as-prepared spheres could be controlled by adjusting the hydrothermal reaction time, while keeping other parameters constant. Finally, novel mesoporous spheres with core–shell structure were assembled. This unique TiO₂ core–shell structure with mesopores has filled a gap in the literature on TiO₂ structures. In order to demonstrate their potential application, dye-sensitized solar cells (DSSCs) with a double layered photoanode structure were fabricated using either the as-obtained TiO₂ spheres with different internal structures or nanometer-sized TiO₂ crystals. The results showed that the DSSCs using the mesoporous spheres with core–shell structure as a top-layer exhibited much higher energy conversion efficiency (9.11%) and short-circuit photocurrent density (18.55 mA cm⁻²), indicating a 38% increase in the energy conversion efficiency compared to those using as-prepared TiO₂ nanocrystals (6.62%). This significant improvement in photoelectric properties was mainly due to the excellent light scattering and dye absorption abilities of the novel structures. The results offer useful guidance in designing photoanode materials.

---

1. Introduction

In recent years, with the depletion of non-renewable energy resource, solar cells have attracted more and more attention. In particular, DSSCs have been the focus of worldwide research due to their obvious superiority.^1,2 Photoanodes, which influence dye loading, light scattering and electron transport,^3–5 have been found to play an important role in the photovoltaic performance of DSSCs. However, few materials with a given structure simultaneously possess excellent dye loading, light scattering and electron transport properties. For example, although one-dimensional (1D) TiO₂ and ZnO, such as nanotubes,^6,7 nanorods^8,9 and nanowires,^10,11 have been regarded as ideal materials for direct electron transport, their inadequate dye adsorption and light scattering still restricts the performance of the corresponding DSSCs.⁴ Moreover, porous electrodes made of small nanoparticles (10–20 nm) will be beneficial to dye adsorption, owing to their large surface areas. However, such a film is usually transparent, which results in high transmittance of visible light.¹² In order to make up for such shortcomings, a top layer is often used to scatter the transmitted light to enhance the light harvest, which has proved effective in boosting photovoltaic efficiencies.^13,14

Many methods have therefore been developed for preparing rough surface microspheres,¹⁵ hollow spherical TiO₂,^16–23 mesoporous TiO₂ aggregates^24–26 or other structures^27–30 that can act as a scattering layer or directly as the light absorption layer. Owing to their large surface area and light scattering effect, these TiO₂ structures have succeeded in improving the energy conversion efficiency of DSSCs. Fang et al. have presented a TiO₂ yolk–shell structure, where the shells were composed of single crystal nanosheets (nanosheet size is ∼150 nm and the thickness is 4–7 nm), that was directly used as the light absorption layer of a photoanode.³¹ Despite these advances, developing new strategies for photoanode materials and structures to increase dye adsorption and light scattering still represents a major scientific challenge. However, monodisperse TiO₂ spheres with both mesoporous and core–shell structures together have not been reported.

In this work, hierarchical TiO₂ mesoporous spheres with core–shell structure have been successfully prepared via a one-step hydrothermal method. The interior structure of these hierarchical spheres could be readily tuned by varying the hydrothermal reaction time. When the reaction time was increased to 22 h, the surface of the monodisperse core–shell TiO₂ spheres assembled from stacked nanoparticles. XRD and SEM images showed that these nanoparticles had a small average size of about 7.2 nm. Moreover, there were many mesopores with diameters of ∼8 nm between the TiO₂ nanoparticles, which resulted in good permeability and a large surface area of 348.8 m² g⁻¹. To reveal how this structure affected their photoelectric properties, the as-synthesized hierarchical spheres with different internal structure and specific surface area were used as double layered photoanodes in DSSCs. These double layered photoanodes were fabricated using the as-prepared TiO₂ products with different internal structures or TiO₂ nanocrystals, which had also been prepared. It was found that hierarchical TiO₂ mesoporous core–shell spheres (HTMCS) displayed excellent photoelectric properties. This enhanced performance can be attributed to their efficient light scattering and dye absorption. The structure-dependent photoelectric properties of these TiO₂ samples might shed light on the future development of advanced electrode materials for DSSCs.

2. Experimental details

2.1. Materials synthesis

The hierarchical mesoporous TiO₂ core–shell spheres (designated as HTMCS) were prepared by a one-step hydrothermal process. In a typical synthesis, 27.8 μL of diethylenetriamine (DETA, ≥95%, Sigma-Aldrich) was added to 30 mL isopropyl alcohol (IPA), and the solution was stirred gently for 3 min. 1.2 mL tetra-n-butyl titanate (TBT) was then added dropwise to the solution under stirring. The mixture was then transferred into a 40 mL Teflon-lined stainless steel autoclave, which was heated in an oven at 200 °C for 3–22 h. The autoclave was then removed from the oven and left to cool naturally to room temperature. The solid products were collected by filtration, washed with ethanol five times and dried at 60 °C. The resulting powder was calcined at 450 °C for 2 h with a heating rate of 1 °C min⁻¹ in air to remove organic components.

2.2. Preparation of 15–20 nm TiO₂ nanocrystals

Preparation of 15–20 nm TiO₂ nanocrystals was based on a previous report.³² Typically, 7.5 mL of titanium isopropoxide (Sigma) was added into 45 mL of 0.1 M HNO₃ aqueous solution. The suspension was placed in an 80 °C water bath and stirred for 8 h. The concentrated suspension was then transferred into a 50 mL autoclave and heated at 220 °C for 12 h. The as-prepared TiO₂ colloid particles were collected by centrifugation and kept for later use without drying.

2.3. Fabrication and measurements of DSSCs

To prepare the DSSC working electrodes, hydroxypropyl cellulose (Aldrich) was added to diethylene glycol with a concentration of ∼10 wt%. This paste was added to the wet TiO₂ colloid particles (containing ∼15% net TiO₂) in a proportion of ∼40% TiO₂ by weight. The mixture was vigorously stirred for about 2 h to obtain the slurry for the semi-transparent adsorption layer. The paste was also added to the obtained hierarchical TiO₂ mesoporous core–shell spheres in a ratio of ∼40 wt% TiO₂. After stirring, the slurry for the scattering layer was prepared. The bilayer film was constructed by the doctor-blade method through two-step calcination. Nanocrystals were spread onto fluorine-doped tin oxide (FTO) glass substrate (resistivity 14 Ω per square, Nippon Sheet Glass, Japan) with adhesive tape to control the film thickness. After drying at 125 °C for 6 min, another layer of core–shell spheres was deposited on the semi-transparent layer and the electrodes coated with the TiO₂ paste were gradually heated under an airflow at 325 °C for 5 min, at 375 °C for 5 min, and at 450 °C for 15 min, and, finally, at 500 °C for 15 min.³³ After cooling to 80 °C, the thick film with double-layer structure was loaded with dye by immersing it in a 0.5 mM Ru-dye (cis-dithiocyanate-N,N′-bis(4-carboxylate-4-tetrabutyl ammoniumcarboxylate-2,2′-bipyridine) ruthenium(II)) (N719, Solaronix SA, Switzerland) for 24 h at room temperature. The resulting films with double layer structures were then sandwiched together with platinized FTO counter electrodes and the electrolyte was then injected into the cell from the edges by capillarity. The content of the electrolyte is 0.05 M LiI, 0.05 M I2, 0.5 M 4-tert-butylpyridine (Aldrich) and 0.6 M 1-propyl-3-methylimidazolium iodide (PMII) in 3-methoxypropionitrile. The representative working electrode film area for solar cell performance test was 0.25 cm².

2.4. Measurement of DSSC performance

Current–voltage (I–V) characteristics were measured using a Keithley 2400 Source Meter under one sun AM 1.5G (100 mW cm⁻²) illumination with a solar light simulator (Newport, Model: 94023A). A 450 W xenon arc lamp (Newport, Model: 6280NS) served as a light source and its incident light intensity was calibrated with a NREL (National Renewable Energy Laboratory)-calibrated Si solar cell equipped with an optical filter to approximate AM 1.5G one sun light intensity before each measurement. The incident photon to current efficiency (IPCE) spectra was measured as a function of wavelength from 350 to 800 nm with a spectral resolution of 5 nm using a Spectral Product Zolix DSC300PA. Diffuse-reflectance spectra were measured on the same film samples on a Perkin-Elmer UV-Vis spectrophotometer (SHIMADZU 2550). The electrochemical impedance spectroscopy (EIS) experiments were performed using an electrochemical workstation (Solartron SI1287) at a bias potential of −0.8 V. Intensity-modulated photovoltage spectroscopy (IMVS) and intensity-modulated photocurrent spectroscopy (IMPS) measurements were carried out on an electron lifetime and dispersion test system (PSL-100) with a diode laser light source with variable intensity at 620 nm.

2.5. Characterizations of HTMCS

The crystal structures of the as-prepared products were investigated by X-ray diffraction (XRD) (Rigaku TTRIII, with Cu Kα1 radiation). Morphologies and microstructures were examined by field-emission scanning electron microscopy (FESEM, JEOL JSM-7500F, operated at an acceleration voltage of 15 kV). Transmission electron microscopy (TEM) measurements were obtained on a JEOL JEM-2100 microscope operated at 200 kV. Specific surface areas were measured using the Brunauer–Emmett–Teller (BET) equation based on nitrogen adsorption isotherms obtained with a Micromeritics Gemini VII apparatus (Surface Area and Porosity System).

3. Results and discussion

3.1 Structure and morphology

Detailed morphology and structural features of the TiO₂ samples were observed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). It can be seen from Fig. 1a that the TiO₂ mesoporous core–shell spheres are monodisperse and have diameters of about 850 ± 40 nm. The higher magnification SEM images clearly revealed that the surfaces of the spheres were assembled from tiny porous nanoparticles (Fig. 1c and S1d, ESI†). In addition, it could be seen that these spherical products were all core–shell structures (Fig. 1(b) and (d), S1a and b, ESI†), and that the thickness of the shells was about 100 nm (Fig. S1c, ESI†). Furthermore, from the higher resolution TEM image (Fig. 1e and S1e, ESI†), we could identify that the surface (the edge of the core–shell sphere) was composed of nanoparticles and there were many mesopores with diameters of ∼8 nm between these TiO₂ nanoparticles (highlighted by arrow in Fig. 1e and S1e, ESI†), which may greatly enhance the surface-to-volume ratio of the core–shell spheres. A high-resolution TEM (HRTEM) image (Fig. 1f) was taken of the indicated area of Fig. 1e. The inter-plane spacings of 0.48 nm and 0.35 nm imply that the particles are growing along the 〈001〉 direction and enclosed within {101} facets.

Fig. 1 (a and c) SEM images of the hierarchical TiO₂ mesoporous core–shell spheres with different magnifications, (b, d and e) TEM and (f) HRTEM images of as-prepared core–shell spheres.

3.2 Growth control and proposed mechanism for the formation of the spheres

Time-dependent experiments were carried out to shed light on the growth mechanism of this unique core–shell structure. The precipitates collected at different time intervals were characterized by SEM and TEM, as shown in Fig. 2. After the complex precursor was introduced into the hydrothermal system and maintained at 200 °C for 3 h, it can be observed that monodisperse TiO₂ spheres were formed (designated as T1, Fig. 2A₁), Fig. 2A₂ and A₃ show higher magnification SEM images, which clearly revealed that the spheres have a smooth surface and a diameter of about 700 nm. The T1 spheres were solid spheres (Fig. 2A₄). When the hydrothermal process was prolonged to 12 h, the surface of the spheres become rough, as shown in Fig. 2B_1–3 (designated as T2). The surface of the spheres was composed of tiny nanoparticles and had an excellent porosity, as shown by a high-resolution TEM (HRTEM) image obtained of the marked fringe of the TiO₂ nanoparticles (shown in Fig. S1f, ESI†), from which the lattice fringes could be observed clearly and the lattice spacing was 0.351 nm, corresponding to the (101) plane of anatase TiO₂. This result confirms the single-crystal structure of the nanoparticles. The T2 spheres were still solid (Fig. 2B₄). When the reaction was carried out for 22 h, the surfaces of the spheres become shaggier and the convex shapes were more apparent (designated as T3, Fig. 2C_1–3). Furthermore, the solid spheres were transformed to mesoporous core–shell structure spheres (Fig. 2C₄, S1d and e, ESI†).

Fig. 2 SEM and TEM images of the three step crystalline TiO₂ sphere products at different reaction time. A_1–4: 3 h (T1); B_1–4: 12 h (T2), and C_1–4: 22 h (T3).

The X-ray diffraction (XRD) pattern of these TiO₂ spheres (Fig. 3a) showed well-resolved diffraction peaks corresponding to the reflections of anatase TiO₂ material (JCPDS card no. 73-1764). The crystal size estimated from the full width at half maximum of the (101) peak using the Scherrer equation: D = Kλ/β cos θ, where λ is the wavelength of the X-ray radiation (λ = 0.15418 nm), K is the Scherrer constant (K = 0.89), θ is the position of X-ray diffraction peak, and β is the full width at half-maximum (fwhm), which indicated that these TiO₂ nanocrystals were 15.8, 8.2 and 7.2 nm (T1, T2 and T3) in diameter, in good agreement with the diameters measured from SEM images (Fig. 2). The well-defined and sharp Bragg peaks with high intensity indicate the good crystallinity and phase purity of the samples. The peaks at 25.35, 37.95, 48.19, 54.27, 55.21, 62.84, 69.01, 70.45 and 75.27° can be assigned to (101), (004), (200), (105), (211), (204), (116), (220) and (215) crystal planes of anatase TiO₂, respectively.

Fig. 3 (a) XRD patterns and (b) nitrogen sorption isotherms of as-prepared crystalline TiO₂ spheres.

The specific surface areas of the calcined TiO₂ spheres were characterized using the nitrogen gas sorption technique, and the typical isotherms are shown in Fig. 3b. For the sample T1, a type II isotherm was observed, indicating the nonporous character of its dense structure,³⁴ and its specific surface area was only 17.9 m² g⁻¹. With further prolonging of hydrothermal reaction time, type IV isotherms, with a sharp capillary condensation step at high relative pressures (P/P_o = 0.75) and H1 type hysteresis loops, were observed for the calcined samples T2 and T3, which indicate their relatively large pore sizes.³⁵ The specific surface areas of T2 and T3 were calculated to be 187.3 and 348.8 m² g⁻¹ by the BET method. Using the Barrett–Joyner–Halenda (BJH) method and the desorption branch of the nitrogen isotherm, the calculated pore size distribution indicated that T3 had an average pore size of 8 nm (Fig. S2, ESI†), suggesting that it is a mesoporous material.

On the basis of the above results, a growth mechanism of the precursor spheres was tentatively proposed (see Fig. 4) to illustrate the role of DETA as a structure-directing agent affecting the control of morphology and monodispersity in the hydrothermal synthesis. Monodisperse precursor spheres were proposed to form through a cooperative assembly process involving long-chain diethylenetriamine and Ti species/oligomers.³⁶ On hydrolysis of TBT, the resultant titanium species and their oligomers most likely participate in hydrogen-bonding interactions with amino groups of the DETA to form inorganic–organic composites at a molecular level. Such hybrid composites contain hydrophobic long-chain alkenyl groups, and thus tend to self-organize into silk-like hybrid micelles to reduce the interfacial free energy.^37,38 Meanwhile, further hydrolysis and condensation of the titanium species/oligomers associated with the hybrid micelles results in inter-connection and short-range packing of the hybrid micelles to form a new liquid condensed phase, rich in DETA and titanium oligomers. As the titanium oligomers further polymerize, the condensed phase becomes denser over time, accompanied by the formation of a mesostructured inorganic framework as a result of the gelation transformation. To minimize the surface free energy, the condensing phase forms spherical shapes as in conventional colloid formation processes.^39,40 The spherical shapes then aggregate coarsely to form the precursor spheres (Fig. 4a). As the reaction proceeds, hydrolysis reactions would occur preferentially on the spheres to release Ti species for nucleation on the surface of the spheres (Fig. 4b) due to its higher surface free energy under hydrothermal conditions. As the large amount of ammonium ions, which was provided by DETA in the reaction system, could hinder the growth of lateral facets of the anatase structure,⁴¹ the anatase TiO₂ particles selectively grow along the 〈001〉 direction to form convex shapes composed of tiny nanoparticles (see Fig. 1f). Finally, the inner spheres began to decompose and shrink to serve as a reservoir of Ti species and the surface of the T3 spheres observably form convex shapes (Fig. 4c).

Fig. 4 Schematic representation of the formation mechanism of the crystalline TiO₂ spheres, through a cooperative self-assembly process in the presence of DETA as the structure-directing agent. Images (a), (b) and (c) are the sphere products of the three steps.

It is well known that hierarchical TiO₂ submicrometer-sized spheres can be used to enhance light scattering, electrolyte permeation, dye adsorption and electron transport, and thus photovoltaic performance. We compared four cells with different films. A semi-transparent film composed of a single layer of TiO₂ nanocrystals with a thickness of ∼15 μm was labeled as Film-1. The as-prepared spheres from each of the three steps were deposited as the scattering layer (∼5 μm) on TiO₂ nanocrystalline film (∼10 μm) to fabricate the bilayer films, which were then labeled as Film-2, Film-3 and Film-4, respectively. The corresponding DSSCs assembled from these photoanodes were denoted Cell-1, Cell-2, Cell-3 and Cell-4, respectively.

The resulting N719 sensitized TiO₂ solar cells were characterized by measuring the current–voltage behavior under standard AM 1.5G simulated sunlight (power density of 1000 W m⁻²). The typical photocurrent density–photovoltage (J–V) characteristics of the DSSCs based on the four photoanodes are shown in Fig. 5 and the corresponding photovoltaic characteristics are summarized in Table 1. Since the filling factor (FF) and open-circuit photo-voltage (V_oc) varied little among the four cells, the variation of the energy conversion efficiency (η) mainly came from the change in the short-circuit photocurrent density (J_sc). Compared with Cell-1, the J_sc of Cell-2 was further increased to 16.36 mA cm⁻² and the η of Cell-2 was improved to 7.87%, which was attributed to increased light scattering capacity of the large spheres on the top-layer. Moreover, the short-circuit photocurrent density (J_sc) of Cell-3 and Cell-4 increased to 17.50 and 18.55 mA cm⁻², leading to the energy conversion efficiency (η) improving to 8.61 and 9.11%, respectively. This is mainly due to the light scattering capacity and increasing specific surface area of the top-layer. Furthermore, both the J_sc and η of the DSSCs based on Film-4 were improved by 32% and 38%, respectively, compared to those of Film-1 (single layer of TiO₂ nanocrystals), which will be discussed later in detail.

Fig. 5 J–V curves of the DSSCs based on different photoanodes measured under one sun illumination (AM 1.5G, 100 mW cm⁻²). -■- Cell-1, -▼- Cell-2, -◆- Cell-3 and -●- Cell-4.

Table 1 Photovoltaic data of four different TiO₂ films, measured under AM 1.5G one sun illumination (100 mW cm⁻²) and simulative value of resistance (R_s, R₁ and R₂) from EIS spectra calculated by equivalent circuit as shown in Fig. 8. Data from I–V measurements, EIS spectra and UV-Vis adsorption.

Sample	Jsc (mA cm^−2)^a	Voc (mV)^a	FF (%)^a	η (%)^a	Rs (Ω)	R1 (Ω)	R2 (Ω)	Dye adsorbed (×10^−7 mol cm^−2)
a Jsc: short-circuit photocurrent density. Voc: open-circuit photovoltage; η: total power conversion efficiency; FF: fill factor
Cell-1	14.05	749	0.63	6.62	15.9	9.4	62.8	0.88
Cell-2	16.36	751	0.64	7.87	15.7	9.2	102.1	0.75
Cell-3	17.50	757	0.65	8.61	16.0	8.9	111.5	1.29
Cell-4	18.55	756	0.65	9.11	15.8	9.0	145.6	1.94

---

It is well known that, as the TiO₂ photoanode is the carrier of the dye and the dye is the harvester of the photon, the photocurrent is strongly related to the light-harvesting capability of TiO₂ photoanode.⁴² As a result, promoting dye loading and light scattering of the electrodes are two equally important tasks for light-harvesting. High specific surface area photoanode materials lead to high dye adsorption amount, therefore HTMCS film, which has a surface area of 348.8 m² g⁻¹, may be an excellent photoanode material. The amount of N719 dye adsorbed on these four films was investigated by thorough desorption in 1 mM NaOH solution and then measuring the UV-Vis absorption spectra of the resulting solutions. The dye loading of Cell-4 (HTMCS top-layer) was 2.2, 2.6 and 1.5 times higher than that of Cell-1, 2, 3 (see Table 1), respectively. The higher dye loading of Cell-4 was ascribed to the larger specific surface area of the film.

Hierarchical spheres with submicron size consisting of nanoparticles have been proven to enhance the light scattering effect, hence leading to higher photocurrent and better photovoltaic performance. To further investigate the light scattering property of the HTMCS material and TiO₂ nanocrystal bases on the four films, UV-vis diffuse reflectance spectra of the four films were measured. Fig. 6a shows the UV-Vis reflectance spectrums of the four films without dye loading. Before the dye adsorption, comparatively, all four films have high diffuse reflection in the visible range of 400–450 nm, but a distinctly rapid decrease is observed for the single layer of TiO₂ nanocrystals film (Film-1) in the wavelength range from 450 to 800 nm. This would cause the unabsorbed light to penetrate through the film without being back scattered to enhance light absorption. In contrast, the reflectances of the double layers (Film-2, 3, 4) are much higher than those of Film-1 in the wavelength range from 450 to 800 nm, pointing out the key role of the TiO₂ spheres in dominating the light scattering behavior.⁴³ However, the light scattering ability of Film-3 was better than Film-2 in the wavelength range from 450 to 800 nm, mainly due to the surface of top-layer spheres being rougher in Film-3 than those in Film-2. Film-4 displayed the best light scattering capacity due to the core–shell structure.

Fig. 6 UV-Vis reflectance spectra of the four films without N719 dye loading (a) and UV-Vis absorption spectra with N719 dye loading (b). The inset of (b) gives the UV-Vis absorption spectrum of N719. -■- Film-1, -▼- Film-2, -◆- Film-3 and -●- Film-4.

Fig. 6b presents the UV-Vis absorption spectra of the four films loaded with N719 dye. For all photoanodes, the absorption wave peak appears at about 525 nm, which matches the peak of N719 dye (shown in Fig. 6b inset), indicating that the main peak can be ascribed to the light adsorption of the dye molecules attached on the TiO₂ nanomaterials. Drastic absorption decreases were observed for Film-4, composed of HTMCS, indicating that it had better dye adsorption capacity than the other films (Fig. 6b). In the long wavelength region, the dye-absorbed Film-2, 3, 4 still retain a substantially higher absorption than that the Film-1, and the curve of Film-4 is the highest, as HTMCS has the largest specific surface area. This further supports our argument that the presence of the spheres (T1, T2, T3) in the top-layer of the photoanode plays a very important part in increasing the short-circuit current density and the energy conversion efficiency. However, the differences between the four samples in light scattering efficiency may not be proportional to the difference in IPCE, since the IPCE can be affected by a number of parameters.⁴⁴

Fig. 7 shows IPCE spectra for the four kinds of DSSCs. It can be observed that, for all DSSCs, the maximum of the IPCE appears at 525 nm (the peak of the N719 adsorption). Amongst the DSSCs, Cell-1 had the lowest IPCE value, which could be mainly explained by TiO₂ nanocrystals having low specific surface area and poor light scattering property. Cell-2 and Cell-3 had better IPCE values and, similarly to our diffuse reflectance and specific surface area results, the IPCE value of Cell-3 was higher than Cell-2. Notably, Cell-4 reveals the highest IPCE values over the whole spectral region, which can be ascribed to its relatively superior specific surface area, which allows the dye molecules to distribute uniformly on the surface of the nanomaterial. Moreover, the improved light scattering ability of HTMCS also contributes greatly to its excellent incident photon to current efficiency.

Fig. 7 IPCE curves of TiO₂ electrodes prepared from the four TiO₂ samples. -■- Cell-1, -▼- Cell-2, -◆- Cell-3 and -●- Cell-4.

The EIS of the four DSSCs were measured in the dark under a forward bias of −0.8 V.^45,46 As shown in Fig. 8, two semicircles were observed in the Nyquist plots. The smaller and larger semicircles in the Nyquist plots are attributed to the charge transfer at the counter electrode–electrolyte interface and the TiO₂–dye–electrolyte interface, respectively. The sheet resistance (R_s) of the substrate, charge transfer resistance of the counter electrode (R₁) and recombination resistance (R₂) were analyzed by Z-view software using an equivalent circuit containing a constant phase element (CPE) and resistors (R)⁴⁷ (Fig. 8, inset). As shown in Fig. 8 and depicted in Table 1, the cells based on these four photoanode structure devices reveal similar R_s and R₁ of 15.9 ± 0.2 and 9.2 ± 0.3 Ω, respectively, due to the use of the same counter electrode (Pt/FTO glass) and electrolyte. However, the recombination resistance (R₂) of Cell-4 (145.6 Ω) is larger than that of other cells (62.8, 102.1 and 111.5 Ω). This reveals that the core–shell material structure and the double layers of photoanode structure have a slower electron recombination process.

Fig. 8 Nyquist plots from electrochemical impedance spectra of the four films measured in the dark at −0.8 V bias. The inset illustrates the equivalent circuit simulated to fit the impedance spectrum.

After carefully analyzing and fitting the EIS results, we resorted to intensity modulated photocurrent/photovoltage spectroscopy (IMPS/IMVS).^48–50 In these studies, we used a diode laser light source with variable intensities at 620 nm, Fig. 9 represents the main results in the form of the electron diffusion coefficient (D_n) and electron lifetime (τ_r) as a function of open-circuit voltage (V_oc) for these four DSSCs based on different scattering layers. Clearly recognizable are the bias exponentially dependences of the electron diffusion coefficient (D_n) and τ_r on V_oc, which are in accordance with the well-known trapping–detrapping model customarily used to describe electron transport/recombination in DSSCs.⁵¹ On the whole, the small variations in the slopes of the four cells are thought to arise from the different backbones and porosities of the corresponding photoanode films. Notably, whereas τ_r appears to be insensitive to the morphology of our photoanode films, D_n depends very much on the morphology.

Fig. 9a shows that the D_n curves of Cell-3 and Cell-4 still lie above that of Cell-1, indicating faster electron transport in the former cells than in the latter. This can be explained by the high crystallinity and uniformity of the spheres, which has already been shown above by XRD and TEM. Specially, the curve of Cell-4 (T3 contained scattering layer) is the highest, which would indicate both interior conductivity and interfacial conductivity though the (101) planar contacts between nanocrystals. This significant improvement of electron transport can be understood by invoking one basic cause, which is related to the close packing between the microspheres, resulting in very good interior contacts between the constituent nanoparticles. These contacts create a sort of highway to facilitate electron transport.

Fig. 9 Electron diffusion coefficients (a) and electron lifetimes (b) as a function of the open-circuit voltage for the four DSSCs. -■- Cell-1, -▼- Cell-2, -◆- Cell-3 and -●- Cell-4.

Next, we move on to discussing the lifetime recombination time (τ_r) in the four DSSCs (Fig. 9b).⁵² The electrons with I₃⁻ ions in Cell-3 possess the highest τ_r values among the four cells, while the τ_r values of Cell-4 lie slightly below those of Cell-3 with low J_sc. Extrapolating to the high J_sc region, the τ_r curves of Cell-4 would lie above those of Cell-3. This result is presumably ascribable to there being more defects in Film-4 (T3 contained scattering layer), because of the core–shell structure of T3. However, due to the much larger specific surface area of Film-4, which allow higher dye loading than that of T2 (see Table 1), and the excellent scattering ability of Film-4, the slightly decreased τ_r has not compromised the performance of Cell-4 as a whole.

4. Conclusions

In summary, monodisperse TiO₂ spheres with controllable internal structure were simply synthesized by a one-step hydrothermal method. The spheres were assembled from tiny porous nanoparticles with average diameters of about 7–16 nm. The monodisperse spheres transition from solid spheres with smooth surfaces into core–shell spheres with mesoporous surfaces. This unique TiO₂ core–shell structure with mesopores filled a gap in the literature on TiO₂ structures. Double-layer photoanodes for DSSCs were constructed using either these monodisperse TiO₂ spheres or TiO₂ nanocrystals. Due to their high scattering ability and specific surface areas, the mesoporous HTMCS can greatly enhance light harvesting, which leads to a sharp increase in the photon-to-current conversion efficiency (9.11%). Compared to a single layer photoanode fabricated with as-prepared TiO₂ nanocrystals, a 38% increase in the conversion efficiency was observed due to the introduction of the HTMCS top-layer.

Acknowledgements

The financial support of National Science Fund of China (no. 60906036, 61074172, 61134010, 61377058) and Program for Chang Jiang Scholars and Innovative Research Team in University (no. IRT13018) is gratefully acknowledged.

Notes and references

1. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737.
2. G. E. Tulloch, J. Photochem. Photobiol., A, 2004, 164, 209.
3. Y. C. Qiu, W. Chen and S. H. Yang, Angew. Chem., Int. Ed., 2010, 49, 3675.
4. D. H. Chen, F. Z. Huang, Y. B. Cheng and R. A. Caruso, Adv. Mater., 2009, 21, 2206.
5. J. Van de Lagemaat and A. J. Frank, J. Phys. Chem. B, 2001, 105, 11194.
6. G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese and C. A. Grimes, Nano Lett., 2006, 6, 215.
7. K. Zhu, N. R. Neale, A. Miedaner and J. A. Frank, Nano Lett., 2007, 7, 69.
8. B. Liu and E. S. Aydil, J. Am. Chem. Soc., 2009, 131, 3985.
9. J. J. Wu and S. C. Liu, J. Phys. Chem. B, 2002, 106, 9546.
10. X. J. Feng, K. Shankar, M. Paulose and C. A. Grimes, Angew. Chem., Int. Ed., 2009, 48, 8095.
11. E. Hosono, S. Fujihara, K. Kakiuchi and H. Imai, J. Am. Chem. Soc., 2004, 126, 7790.
12. S. Hore, C. Vetter, R. Kern, H. Smit and A. Hinsch, Sol. Energy Mater. Sol. Cells, 2006, 90, 1176.
13. P. Wang, S. M. Zakeeruddin, P. Comte, R. Charvet, R. Humphry-Baker and M. Grätzel, J. Phys. Chem. B, 2003, 107, 14336.
14. F. Gao, Y. Wang, D. Shi, J. Zhang, M. K. Wang, X. Y. Jing, R. Humphry-Baker, P. Wang, S. M. Zakeeruddin and M. Grätzel, J. Am. Chem. Soc., 2008, 130, 10720.
15. L. Yang, Y. Lin, J. Jia, X. Li, X. Xiao and X. Zhou, Microporous Mesoporous Mater., 2008, 112, 45.
16. H. J. Koo, Y. J. Kim, Y. H. Lee, W. I. Lee, K. Kim and N. G. Park, Adv. Mater., 2008, 20, 195.
17. D. Wu, F. Zhu, J. Li, H. Dong, Q. Li, K. Jiang and D. Xu, J. Mater. Chem., 2012, 22, 11665.
18. X. Wu, G. Q. Lu and L. Wang, Energy Environ. Sci., 2011, 4, 3565.
19. M. Samadpour, S. Gimenez, A. I. Zad, N. Taghavinia and I. Mora-Sero, Phys. Chem. Chem. Phys., 2012, 14, 522.
20. Y. Liu, S. Wang, Z. Shan, X. Li, J. Tian, Y. Mei, H. Ma and K. Zhu, Electrochim. Acta, 2012, 60, 422.
21. J. Jiang, F. Gu, W. Shao and C. Li, Ind. Eng. Chem. Res., 2012, 51, 2838.
22. H. Wang, M. Miyauchi, Y. Ishikawa, Y. Li, L. Li, X. Li, Y. Bando and D. Golberg, J. Am. Chem. Soc., 2011, 133, 19102.
23. H. Pang, H. Yang, C. X. Guo, J. Lu and C. M. Li, Chem. Commun., 2012, 48, 8832.
24. S. R. Gajjela, C. Yap and P. Balaya, J. Mater. Chem., 2012, 22, 10873.
25. K. Y. Yan, Y. C. Qiu, W. Chen, M. Zhang and S. H. Yang, Energy Environ. Sci., 2011, 4, 2168.
26. J. Xi, Q. Zhang, K. Park, Y. Sun and G. Cao, Electrochim. Acta, 2011, 56, 1960.
27. Y. C. Park, Y. J. Chang, B. G. Kum, E. H. Kong, J. Y. Son, Y. S. Kwon, T. Park and H. M. Jang, J. Mater. Chem., 2011, 21, 9582.
28. W. G. Yang, F. R. Wan, Q. W. Chen, J. J. Li and D. S. Xu, J. Mater. Chem., 2010, 20, 2870.
29. H. Dai, Y. Zhou, L. Chen, B. Guo, A. D. Li, T. Yu and Z. G. Zou, Nanoscale, 2013, 5, 5102.
30. M. D. Wei, Y. Konishi, H. S. Zhou, M. Yanagida, H. Sugihara and H. Arakawa, J. Mater. Chem., 2006, 16, 1287.
31. W. Q. Fang, X. H. Yang, H. J. Zhu, Z. Li, H. J. Zhao, X. D. Yao and H. G. Yang, J. Mater. Chem., 2012, 22, 22082.
32. C. J. Barbe, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover and M. Grätzel, J. Am. Ceram. Soc., 1997, 80, 3157.
33. S. Ito, T. N. Murakami, P. Comte, P. Liska, C. Grätzel, M. K. Nazeeruddin and M. Grätzel, Thin Solid Films, 2008, 516, 4613.
34. S. Lowell, J. E. Shields, M. A. Thomas and M. Thommes, Characterization of Porous Solids and Powders: Surface Area, Pore Size and Density, Kluwer Academic Publishers, 2004, p. 347.
35. M. Kruk and M. Jaroniec, Chem. Mater., 2001, 13, 3169.
36. A. Monnier, F. Schüth, Q. Huo, D. Kumar, D. Margolese, R. S. Maxwell, G. D. Stucky, M. Krishnamurty, P. Petroff, A. Firouzi, M. Janicke and B. F. Chmelka, Science, 1993, 261, 1299.
37. P. T. Tanev, M. Chibwe and T. J. Pinnavaia, Nature, 1994, 368, 321.
38. P. T. Tanev and T. J. Pinnavaia, Science, 1995, 267, 865.
39. C. Z. Yu, J. Fan, B. Z. Tian and D. Y. Zhao, Chem. Mater., 2004, 16, 889.
40. W. Stöber, A. Fink and E. J. Bohn, J. Colloid Interface Sci., 1968, 26, 62.
41. T. Sugimoto, X. P. Zhou and A. Muramatsu, J. Colloid Interface Sci., 2003, 259, 53.
42. Y. C. Qiu, W. Chen and S. H. Yang, J. Mater. Chem., 2010, 20, 1001.
43. L. Yang, Y. Lin, J. G. Jia, X. R. Xiao, X. P. Li and X. W. Zhou, J. Power Sources, 2008, 182, 370.
44. A. C. Fisher, L. M. Peter, E. A. Ponomarev, A. B. Walker and K. G. U. Wijayantha, J. Phys. Chem. B, 2000, 104, 949.
45. J. Bisquert, J. Phys. Chem. B, 2003, 107, 13541.
46. T. Hoshikawa, M. Yamada, R. Kikuchi and K. Eguchi, J. Electroanal. Chem., 2005, 577, 339.
47. J. Bisquert, F. Fabregat-Santiago, I. Mora-Seró, G. Garcia-Belmonte and S. Giménez, J. Phys. Chem. C, 2009, 113, 17278.
48. G. Schlichthorl, S. Y. Huang, J. Sprague and A. J. Frank, J. Phys. Chem. B, 1997, 101, 8141.
49. X. Zhang, V. Thavasi, S. G. Mhaisalkar and S. Ramakrishna, Nanoscale, 2012, 4, 1707.
50. Q. Wang, Z. Zhang, S. M. Zakeeruddin and M. Grätzel, J. Phys. Chem. C, 2008, 112, 7084.
51. A. J. Frank, N. Kopidakis and J. Van de Lagemaat, Coord. Chem. Rev., 2004, 248, 1165.
52. W. Yuan, H. Zhao, H. Y. Hu, S. Wang and G. L. Baker, ACS Appl. Mater. Interfaces, 2013, 5, 4155.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra01956c

---