CHEMPHYSCHEM ARTICLES DOI: 10.1002/cphc.201300872
Structurally Stabilized Organosilane-Templated Thermostable Mesoporous Titania Vipin Amoli ,[a, b] Rashmi Tiwari ,[a, b, c] Arghya Dutta,[d] Asim Bhaumik,[d] and Anil Kumar Sinha*[a, b] Structurally thermostable mesoporous anatase TiO2 (m-TiO2) nanoparticles, uniquely decorated with atomically dispersed SiO2, is reported for the first time. The inorganic Si portion of the novel organosilane template, used as a mesopores-directing agent, is found to be incorporated in the pore walls of the titania aggregates, mainly as isolated sites. This is evident by transmission electron microscopy and high-angle annular dark field scanning transmission electron microscopy, combined with electron dispersive X-ray spectroscopy. This type of unique structure provides exceptional stability to this new material against thermal collapse of the mesoporous structure, which is reflected in its high surface area (the highest known for anatase titania), even after high-temperature (550 8C) calcination. Control of crystallite size, pore diameter, and surface area is achieved by varying the molar ratios of the titanium
precursor and the template during synthesis. These mesoporous materials retain their porosity and high surface area after template removal and further NaOH/HCl treatment to remove silica. We investigate their performance for dye-sensitized solar cells (DSSCs) with bilayer TiO2 electrodes, which are prepared by applying a coating of m-TiO2 onto a commercial titania (P25) film. The high surface area of the upper mesoporous layer in the P25–m-TiO2 DSSC significantly increases the dye loading ability of the photoanode. The photocurrent and fill factor for the DSSC with the bilayer TiO2 electrode are greatly improved. The large increase in photocurrent current (ca. 56 %) in the P25–m-TiO2 DSSC is believed to play a significant role in achieving a remarkable increase in the photovoltaic efficiency (60 %) of the device, compared to DSSCs with a monolayer of P25 as the electrode.
1. Introduction Alternative energy sources have received increasing attention over the past decades for the replacement of environmentally damaging and diminishing fossil fuels. Solar energy is one of the most attractive renewable-energy sources. DSSCs based on interpenetrating networks composed of mesoporous, nanocrystalline, wide bandgap semiconductor oxides, such as TiO2, and liquid electrolytes represent a possible alternative to silicon-based solar cells, owing to their ease of fabrication and low cost.[1] Photoexcitation of a dye adsorbed on the surface of TiO2 results in injection of an electron into the conduction
[a] V. Amoli , R. Tiwari , Dr. A. K. Sinha Catalytic Conversion Process Division, Council of Scientific and Industrial Research (CSIR) Indian Institute of Petroleum Mohkampur, Dehradun 248005 (India) Fax: (+ 91) 135-266-203 E-mail: [email protected] [b] V. Amoli , R. Tiwari , Dr. A. K. Sinha CSIR–Network Institute of Solar Energy (CSIR–NISE) Anusandhan Bhawan, New Delhi 110001 (India) [c] R. Tiwari Department of Chemistry, University of Calgary 2500 University Drive N.W. Calgary, Alberta, T2N 1N4 (Canada) [d] A. Dutta, Prof. A. Bhaumik Department of Material Science, Indian Association for the Cultivation of Science Jadavpur, Kolkata, 700032 (India) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201300872.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
band of TiO2. The original state of the dye is restored by the subsequent donation of an electron from the liquid electrolyte in the pores of mesoporous oxide back to the dye. Surface area and porosity are crucial factors in determining the amount of dye adsorbed, which, in turn, affect the performance of DSSCs.[2] To increase the efficiency of a DSSC, a highsurface-area photoanode composed of mesoporous anatase TiO2 is highly desirable, as it maximizes the dye-loading and gives rise to a high photovoltaic efficiency.[2a, 3] Several different techniques have been reported to produce TiO2 nanostructures such as sol gel,[1a, 4] hydrothermal,[5] and template-modified sol gel.[6] There have been several reports regarding the surfactant-assisted synthesis of mesoporous TiO2.[7] Recently, there have been few reports on the synthesis of mesoporous TiO2 employing a mesoporous silica template.[8] These TiO2 structures generally possess a high surface area, better structural stability, an excellent crystalline frame work, and improved photovoltaic performances in DSSCs.[9] Evaporation-induced self-assembly (EISA), explored by Brinker and co-workers, provides an ingenious route to synthesize mesoporous materials.[10] Herein, we report the synthesis of mesoporous anatase TiO2 nanocrystals of high surface areas (as high as 361 m2g1) and with a mean pore diameter of 2.2 nm. We use EISA with a novel organosilane template as a structuredirecting agent. The inorganic Si component of the organosilane template improves the stability of the organic–inorganic
ChemPhysChem 2014, 15, 187 – 194
187
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
(JCPDS card tase TiO2 No. 21:1272).[11] No impurity peaks are observed. The broadening and slight shifting of the (101) peak towards larger 2q values indicates that the average crystallite size decreases with increasing amounts of surfactant, as shown in the Supporting Information (Figure S1). This may be attributed to increased steric repulsion between the growing titania particles caused by the template molecule, which has a suppressing effect on the crystal growth, resulting in better dispersion and size reduction.[12] The broadening of the diffraction peaks is attributed to the size confinement. The average crystallite sizes of T1, T2, T3, and T4 samples, estimated from the full width at half maximum of Scheme 1. The Si component of the organosilane template improves the stability of mesoporous TiO2 structures the (101) peak by using the because of organic–inorganic interactions during synthesis. Scherrer[13] equation, were 5.9, 5.1, 4.9, and 4.3 nm, respectively. The morphology of the different mesoporous titania samcomposite-mesostructured gel when combined with the inorples, prepared through the EISA method, was investigated by ganic Ti precursor (Scheme 1). using field-emission scanning electron microscopy (FESEM), as This is the first report of mesoporous titania uniquely decoshown in Figure 2. The synthesized TiO2 nanostructure aggrerated with isolated SiO2 sites, which gives exceptional thermal stability and very high surface area to these materials. These gates are seen as irregular particles with a size of 10–30 nm. mesoporous crystalline materials are been used, after template Energy dispersive X-ray spectroscopy (EDS) performed near the removal, for the fabrication of a photoanode for DSSCs to impore-walls (given in Figure S2 in the Supporting Information) prove the power-conversion efficiency compared to commerrevealed that the products are composed of Ti, Si, and O. The cial P25. elemental compositions of the obtained materials (near the pore walls) are listed in Table 1. The materials in the bulk region were composed of titania only, whereas the Si atoms
2. Results and Discussion The X-ray diffraction (XRD) patterns of the product, prepared with different precursor and template ratios, are shown in Figure 1. The patterns can be assigned to pure tetragonal ana-
Figure 1. XRD patterns of the mesoporous TiO2 samples T1 (Ti/template = 17), T2 (Ti/template = 8), T3 (Ti/template = 6), and T4 (Ti/template = 4) calcined at 550 8C.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 2. FESEM of mesoporous TiO2 samples calcined at 550 8C. a) T1, b) T2, c) T3, and d) T4.
ChemPhysChem 2014, 15, 187 – 194
188
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 1. Surface and porous characteristics measured from N2 adsorption/desorption isotherms for the mesoporous TiO2 (samples T1, T2, T3, and T4). Sample
Ti/Template
Ti/Si[a]
Surface area [m2 g1]
Mean pore diameter [nm]
Pore volume [cm3 g1]
Porosity [%]
Roughness factor
Particle size [nm]
T1 T2 T3 T4
17 8 6 4
2.16 1.55 1.39 1.11
213.5 295.3 352.6 361.0
4.654 3.428 2.882 2.203
0.2437 0.2531 0.2538 0.1988
48.6 49.4 49.6 43.0
426 578 690 800
7.2 5.2 4.3 4.3
[a] Composition obtained near the pore walls of the mesoporous samples from EDS analysis.
Figure 3. Elemental maps for a) Ti, b) Si, and c) Si and Ti together, obtained by using EDS (Ti + Si: white grey dots = Si, black grey patches = Ti, black area = pores) for sample T3. Inset is a representation of atomically dispersed Si inside the material.
(Figure 4 a, b). As shown in Scheme 1, fine nanoparticles of TiO2 are aggregated around the organosilane micelle mesoporous template, which leads to an interconnected porous network (Figure 4 a, b) after removal of the organic part of the template during calcination. The insets show narrow size distributions of the nanoparticles, with an abundant of size of 4–6 nm and 3– 5 nm for sample T2 and T3, respectively, as also evidenced by XRD analysis. The obtained nanoparticles are well-dispersed, homogeneous, and dense, which can be seen in the Supporting Information (Figure S3). The HRTEM images (Figure 4 c –f) show the presence of multiple individual crystallites with clear lattice fringes. It can be observed that the pores are formed because of the intercrystallite voids of the aggregated titania crystallites. This structure enhances the surface area of the sample. The observed spacing of 0.35 nm between neighboring lattice fringes corresponds to the (101) plane of anatase TiO2.[14] The inset shows corresponding selected-area electron diffraction (SAED) pattern with diffraction spots of (101) and (200) of randomly oriented anatase TiO2 crystallites.
(from the template) decorate the pore walls. It is also observed that the concentration of silica in the prepared samples [coming from the octadecyldimethyl(3-trimethoxysilylpropyl)ammonium chloride (ODAC) template] increases with increasing template concentration. Selected elemental maps (Figure 3) were recorded to identify the dispersion of SiO2 on the surface of the TiO2 matrix. It can be seen that Si is highly dispersed in the material. The grey/ black-colored regions represent titania patches and the grey/ white dots represent silica. The Si sites are seen to be atomically dispersed and are decorating the surface of the titania patches (Scheme 1). This is a unique nanostrucutre for porous anatase titania, in which Si is atomically incorporated in the structure through a chemical method. Transmission electron microscopy (TEM) images of the T2 and T3 samples are shown in Figures 4 a and b, and high-resolution TEM (HRTEM) images are shown in Figures 4 c–f. Very fine well-interconnected particles, with a diameter of only few nanometres Figure 4. TEM images of a) T2 and b) T3 with the inset showing particles-size distribution. HRTEM images showing (ca. 5 nm), are predominantly intercrystallite voids of c) T2 and d) T3. e, f) HRTEM images showing (101) anatase crystallites (insets show correseen forming a porous network sponding FFT patterns). 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 187 – 194
189
CHEMPHYSCHEM ARTICLES
Figure 5. a) Representative HAADF-STEM image of the T3 sample. The corresponding compositional maps for b) Ti, c) Si, and d) Si and Ti together, obtained by using EDS.
www.chemphyschem.org mesopores, showing that the sample porosity is somewhat homogenous. The pore-size distribution analysis by using the Barrett–Joyner–Halenda (BJH) method showed a narrow pore-size distribution, with mean pore diameter of 2–5 nm. The template concentration variation helps to control the pore size and surface area of these mesoporous titania materials. As the amount of template is increased during the synthesis, the surface area of the mesoporous TiO2 increases, as shown in Table 1. The surface area increased (213–360 m2g1) and the pore size decreased (4.6–2.2 nm) with decreasing Ti/template molar ratio (from 17 to 4) during the synthesis (Table 1). This increase in surface area and decrease in pore size can be correlated with the decrease in the particle size with increasing template concentration, as observed from XRD analysis. This may be caused by increased steric hindrance from the template adsorbed on the surface of TiO2 during synthesis, which retards particle agglomeration and results in a decrease in the particle size[15] that, in turn, increases the surface area. A decrease in the mean pore diameter and pore volume is observed with an increase in the amount of ODAC, which may be attributed to two reasons. 1) As the particle size decreases, the intercrystallite void size also decreases, and 2) as there is an increase in the amount of ODAC, more SiO2 is embedded into the mesoporous TiO2 framework pore walls, which results in a decrease in the pore size and the pore volume.[16] EDS spectra show an increased amount of SiO2 in the TiO2 matrix with increasing ODAC content (from 2.9 to 7.6 at %). The loss of hysteresis in sample T4 may be caused by an ink-bottle effect from the increased amount of SiO2 that is present in the TiO2 framework, which reduces the pore diameter and pore volume significantly. The large surface area of T4 has limitations of reduced pore volume and pore diameter. There is no detectable effect on the crystalline phase (observed in the XRD pattern)
The composition of the sample was determined by using scanning transmission electron microscopy (STEM)–EDS. Figure 5 a shows representative high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) image of the sample. A 3D porous integrity of interconnected particles was observed. The compositional EDS mappings showing elemental distributions for Ti, Si, and Ti + Si, over the same area highlighted in HAADF-STEM image, are shown in Figures 5 b–d. The overlap of Si and Ti elemental maps (Figure 5 d) shows that Si is atomically dispersed along the pore walls, over the titania patches given in the Supporting Information (Figure S4). The presence of atomically dispersed Si along the pore walls of the titania framework improves the thermal stability of the mesoporous titania structures, resulting in anatase crystallization without the collapse of mesoporous structures. The specific surface area and pore-size distribution of the mesoporous titania were characterized by using nitrogen gas sorption. Figure 6 shows the isotherms of the calcined mesoporous TiO2 samples. Mesoporous TiO2 exhibits type IV isotherms with a hysteresis loop, which is a typical characteristic of mesoporous materials. For the mesoporous TiO2 samples, a sharp increase in the adsorption volume of N2 was observed between a relative partial pressure (P/P0) of 0.4–0.7, which may be attrib- Figure 6. N2 adsorption/desorption isotherm (left) and BJH pore-size distribution (right) of the mesoporous TiO2 uted to the sharp uptake in the samples. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 187 – 194
190
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
analysis (see Figure S7 in the Supporting Information), without changing its porous structure (Figure 7). The specific surface area decreased after treatment with NaOH aqueous solution (Table 2), which may be attribSample Surface area Pore volume Mean pore Porosity Particle uted to the adsorption of Na + on the TiO2 surface, lead[m2g1] [cm3g1] diameter [nm] [%] size [nm] ing to the formation of Na +OTi bonds, as observed T3 352.6 0.2538 2.88 49.6 4.3 during the NaOH hydrothermal treatment of other TiO2 after NaOH 266.1 0.2569 3.86 49.9 5.8 nanostructures by Dmitry et al.[18] In our case, we did treatment after HCl/ 313.3 0.2901 3.76 53.0 4.9 not employ hydrothermal treatment, but the very high distilled-water surface area (360 m2g1) of TiO2 nanostructures may washing offer a large number of active sites for Na + adsorption. There is a large increase in the mean pore diameter and pore volume during NaOH treatment, which may be attributed to leaching of the silica present in the TiO2 or the morphology of the mesoporous TiO2 after NaOH/HCl matrix, because it is shown by HRTEM that the pores are intertreatment (see Figures S5 and S6 in the Supporting Informacrystallite voids, which may be formed during the sintering tion). There is a slight increase in particle size from 4.3 to process. The surface area increased significantly upon treat4.9 nm after treatment (Table 2). ment with aqueous HCl solution. The treatment with dilute In our case, silica is atomically incorporated in the product, HCl is an ion-exchange phenomenon, in which TiONa and mainly decorating as isolated sites, particularly at the pore TiOH bonds react with acid and the formation of new TiO walls of the titania aggregates, which produce a protective Ti bonds can occur. The amount of dye on the surface of three effect and limit the grain growth (to ca. 4.3 nm) after calcinaphotoelectrodes was measured by the method described in tion at 550 8C. After NaOH/HCl treatment, owing to silica etchref. [19] The amount of N719 dye on the different TiO2 photoing, this protective effect is diminished, resulting in the strucelectrodes is summarized in Table 3. tural rearrangement of TiO2 units and slight grain growth Figure 8 shows the photocurrent density behavior of the during further drying, as explained in the literature.[17] The shape and size of the resulting material, after treating with DSSCs, made from different working electrodes, as a function NaOH (2 m) followed by 0.1 m HCl, is shown by the FESEM of voltage. Device A, device B, and device C were tested under images in the Supporting Information (Figure S6). The NaOH similar simulated irradiation (global AM 1.5, 100 mW cm2), and HCl treatment process results in the removal of any silica with an active area of 0.16 cm2. Table 3 summarizes the perforpresent in the mesoporous sample (T3), as confirmed by EDS mance parameters of the different DSSCs. The TiO2 electrode made from P25 nanoparticles showed a short-circuit current density of 8.13 mA cm2 and an energy conversion efficiency of 3.49 %, whereas the P25–m-TiO2 electrode achieved a current density of 12.73 mA cm2 and a conversion efficiency of 5.58 %, indicating a 60 % increase in the conversion efficiency compared to the P25 electrode, despite of similar open circuit voltage. A large increase in the photocurrent current (ca. 56 %) in the P25–m-TiO2 system, as compared to the bare P25 system, is believed to play a significant role in determining the overall efficiency of the device. The higher photocurrent in case of the P25m-TiO2 photoanode is attributed to two factors. First, the higher dye-loading capabilities of the high-surface-area upper mesoporous layer in the P25–m-TiO2 photoanode increases the Figure 7. N2 physiosorption (left) and BJH pore-size distribution (right) of TiO2 (sample T3) with NaOH treatment amount of light harvested, (top) followed by 0.1 m HCl treatment (bottom). Table 2. Physiological characteristics of mesoporous TiO2 (sample T3) before and after NaOH and HCl/distilled-water treatment, measured from N2 adsorption-desorption isotherms.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 187 – 194
191
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 3. Photovoltaic performance parameters of DSSC made from P25 nanoparticles (device A), P25–m-TiO2 (device B), and P25–m-TiO2 after NaOH and HCl treatment (device C). Solar cell
Jsc [mA cm2]
Voc [V]
FF [%]
h [%]
Adsorbed dye [mol cm2]
device A device B device C
8.13 12.36 12.73
0.694 0.698 0.698
61.7 60.3 62.5
3.49 5.22 5.58
0.87 107 1.65 106 1.71 106
Figure 8. Current–voltage curves of DSSCs made from A) P25 nanoparticles, B) P25–m-TiO2 (sample T3), C) P25–m-TiO2 (sample T3) after NaOH and HCl treatment.
3. Conclusions Mesoporous, nanocrystalline, thermally stable anatase titania with a high surface area (up to 360 m2g1 after calcination at 550 8C) and narrow pore-size distribution (4.4–2.2 nm) was produced through a simple and reproducible modified sol-gel and evaporation-induced self-assembly method by using a novel organosilane template. Crystallographic phases, particle size, and morphology of the TiO2 nanostructures were characterized by using XRD, FESEM, TEM, and Brunauer–Emmett–Teller (BET) techniques. Detailed characterization established that the Si atoms from the organosilane are atomically dispersed along the pore walls of the titania matrix. This unique structure and composition provides the required structural and thermal stability of this new material, which is reflected in a surface area that is the highest known for anatase titania, even after hightemperature (550 8C) calcination. After removal of silica from the pore walls, the structural integrity, porosity, and surface areas are maintained. This new material has the potential for practical applications because of its large surface area and porous structure. These mesoporous TiO2 particles improve the dye loading through their high surface area. A high light-tocurrent conversion efficiency of 5.58 % could be achieved when applied as a photoanode in a DSSC, as compared to a photoanode consisting of P25 nanoparticles.
Experimental Section which is highly desirable for achieving a better performance of the device, as described in many reports.[2a, 3, 21] In particular, the cell short-circuit current (Jsc) is directly proportional to the light-harvesting ability of the cell, which, in turn, is strictly dependent on the dye concentration on the TiO2 adsorptive surface.[20] Second, the enhancement in photocurrent in the P25– m-TiO2 system is thought to be a result of improved interparticle connectivity, owing to the presence of smaller (ca. 5 nm) high-surface-area particles among the larger P25 (ca. 25 nm) particles. During the photoanode sintering process, described above, these smaller particles may serve as a binding agent/filling agent between the larger P25 nanoparticles, and, at the same time, these mesoporous particles can offer a large surface area for dye adsorption. There are several reports that have obtained results similar to ours, in which the interparticle connectivity among TiO2 particles is improved by several techniques, such as the post treatment of sintered electrode by using TiCl4,[21] incorporating P25 into the mesoporous TiO2 matrix, and using 5 nm pure anatase nanoparticles as a nanoglue by mixing it with P25 nanoparticles.[22] There is an increase in the short-circuit current density, fill factor, and photoconversion efficiency for the photoelectrode made from mesoporous TiO2 obtained after NaOH/HCl treatment, as compared to the untreated sample, which may be attributed to the increase in pore diameter and porosity, which favors the diffusion of the redox electrolyte (I/I3) through the nanosized pores in the mesoporous TiO2 film, resulting in an increased fill factor that improved the efficiency of the DSSC.[23] 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Transparent conducting oxide (TCO) glass (Solaronix), Ruthenium 535 dye [Ru(bpy)2(NCS)2H4, N719; Solaronix], iodolyte TG-50 (Solaronix), and platisol (Solaronix) were used as received. Titanium (IV) isopropoxide [Ti {OCH (CH3)2}4, 95 % Alfa Aesar] was used as a source of titanium. ODAC (60 % in methanol; Gelest, Inc), glacial acetic acid (Merck), ethanol (Merck, anhydrous), Triton X-100 [C14H22O(C2H4O)n, Loba Chemie], P25 (Sigma–Aldrich, particle size of 27 nm, specific surface area of 57 m2g1, and the phase composition of 80 % anatase and 20 % rutile), 67–72 % HNO3 (Merck) were used without further purification.
Synthesis of Mesoporous TiO2 Particles ODAC was added as a mesopore template for the synthesis of mesoporous titania. In a typical synthesis, ODAC (2 mL) was dissolved in anhydrous ethanol (40 mL) and stirred for 30 min at room temperature (solution 1). Titanium isopropoxide (TIP; 40 mm) was dissolved in a mixed solvent composed of 69–72 wt % HNO3 acid (6.4 mL) and anhydrous ethanol (20 mL) (solution ). Solution 1 was added drop-wise to solution 2, with continuous stirring. Next, additional ethanol (20 mL) was added. The final gel composition was 1:3.8:0.06:25.5 TIP/HNO3/ODAC/EtOH. The solution mixture was stirred for 24 h at room temperature. After that, the solution was transferred to a petri dish and solvent evaporation was carried out at 70 8C for 48 h, followed by calcination in air in a furnace at 550 8C for 6 h. A series of mesoporous TiO2 samples (T1, T2, T3, and T4) were prepared by changing the Ti/template (molar) ratios to 17, 8, 6, and 4, respectively. The organic part of the template was removed by calcination. The silica from the template present in the as-synthesized mesoporous titania samples was removed by treating the samples in a 2 m NaOH solution at 80 8C for 24 h under continuous stirring. The samples were then washed with ChemPhysChem 2014, 15, 187 – 194
192
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
0.1 m HCl and then with distilled water before using as the materials as the anode in a DSSC.
Preparation of the TiO2 Paste for the DSSC
in which Jsc is a short-circuit current density (mA cm2), Voc is an open-circuit voltage (V), Pin is an incident light power, FF is the fill factor, h is an overall energy-conversion efficiency, and Jmax (mA cm2) and Vmax (V) are the current density and voltage in the J–V curve, respectively, at the point of maximum power output.
TiO2 paste was prepared by mixing the synthesized TiO2 powder (0.25 g) with acetic acid (0.5 mL), which was then ground mechanically for 15 min. Next, a mixture of water and ethanol (1:3) (15 mL) was added dropwise, and this mixture was ground mechanically. Finally, Triton X-100 (0.5 mL) was added to facilitate the spreading of the paste on the substrate. The resultant slurry was ground for another 15 min to obtain a homogeneous paste. A similar method was adopted for the preparation of a paste with P25 nanoparticles.
For dye-loading measurements, photoanodes with a dye-loaded area of 4 cm2 and a thickness of approximately 7 mm were dipped into 10 mL of a 0.1 m solution of NaOH in ethanol/H2O (1:1 v/v), until complete desorption of the N719 dye occurred. The alkaline solution containing the fully desorbed dye was carefully measured by using UV/Vis spectra (Hitachi U-2900, UV/Vis Double Beam spectrophotometer); the Beer–Lambert law [Eq. (2)] was used to calculate the number of adsorbed N719 dye molecules by using the absorption value at 515 nm.
Device Fabrication
A ¼ elc
Three working electrodes [device A with P25 nanoparticles, device B with one layer of P25 and second layer of m-TiO2 (sample T3, without NaOH/HCl treatment), and device C with one layer of P25 and second layer of m-TiO2 (sample T3, after NaOH/HCl treatment)] were fabricated. The doctor-blade technique was used for coating the TiO2 paste on the TCO glass. The TiO2-film-coated TCO samples were sintered at 450 8C for 30 min. For sensitization, the films were impregnated with 0.3 mm N719 dye in ethanol for 24 h at room temperature. The counter electrode was obtained by depositing a thin layer of platisol T/SP, which can be squeegee-printed by using a polyester mesh, and heat-treated in air for 30 min at 450 8C. One drop of iodolyte TG-50 electrolyte was put on the surface of the dye-loaded TiO2 photoanode, which penetrated inside the porous TiO2 via capillary action. A Pt-coated fluorine–tin oxide electrode was then clipped onto the top of the TiO2 working electrode to form the complete DSSC. The active area of each DSSC was about 0.16 cm2. A similar fabrication procedure is adopted for all DSSCs.
Characterization Crystalline-structure and phase identification of the samples was performed by using XRD measurements on a Bruker D8 Advance diffractometer, operating in the reflection mode with CuKa radiation (40 KV, 40 mA). Surface morphology of the samples was obtained by using FESEM (Quanta 200F, Netherlands). EDS spectroscopy combined with SEM was used to analyze the composition of the samples. The BET specific surface area, BJH-calculated average pore size, and pore volume of the TiO2 nanocrystals were obtained from the N2 adsorption/desorption isotherms recorded on a BELSORP max instrument (Japan) at 77 K. The samples were degassed and dried under a vacuum system at 150 8C for 4–5 h prior to the measurement. TEM studies were performed by using a JEOL 2010F TEM/STEM microscope, operating at 200 kV. This instrument was equipped with a JEOL high-angle annular dark field detector and an Oxford X-Max Silicon Drift X-Ray EDS (XEDS) detector, enabling the acquisition of STEM images and compositional analysis, either in spot mode or through elemental mapping using a 0.5 nm electron probe. Photovoltaic performance of the assembled DSSCs was measured by completely integrated current–voltage (I–V) Test Station (PVIV-211 V) with a Kiethley 2420 source meter and a 94023 A Oriel Sol3 A, class AAA solar simulator (power output = 100 mW cm2, lamp power = 450 W) equipped with an AM 1.5 filter. The solar conversion efficiency was calculated by using the formula: h ð%Þ ¼ ðV max Jmax Þ=Pin 100 ¼ ðV oc Jsc FFÞ=Pin 100 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ð1Þ
ð2Þ
in which A is the absorption of the UV/Vis spectra at 515 nm, e is the molar absorptivity (14100 m1 cm1), l is the path length of light beam, and c is the concentration of the dye. The porosity of the TiO2 film was calculated according to Equation (3):[19a] ðPÞ ¼ V p =ð11 þ V p Þ
ð3Þ
in which (P) is the porosity, Vp is the specific pore volume (cm3g1), and 11 is the reciprocal of the density of anatase TiO2 (11 = 0.257 cm3g1). The roughness factor (R) per unit of film thickness and the particle size (D) were calculated using equation Equations (4) and (5), respectively:[19a] ðRÞ ¼ 1ð1-PÞS
ð4Þ
ðDÞ ¼ 6000=ðS 1Þ
ð5Þ
in which 1 is the density of anatase TiO2 (3.89 cm3g1), P is the porosity of the sample, and S is the specific surface area (m2g1).
Acknowledgements The director of the Indian Institute of Petroleum is acknowledged for approving the work. V.A. thanks CSIR India for a Junior Research Fellowship. The authors thank the Analytical Science Division at the Indian Institute of Petroleum for analytical services. The Ministry of New and Renewable Energy, New Delhi is acknowledged for research funding. Keywords: mesoporous materials · micelles · surface area · thermostable · titania [1] a) B. O. Regan, M. Grtzel, Nature 1991, 353, 737; b) M. A. Khan, M. S. Akhtar, O. Yang, Solar Energy 2010, 84, 2195. [2] a) S. Kambe, K. Murakoshi, T. Kitamura, Y. Wada, S. Yanagida, H. Kominami, Y. Kera, Sol. Energy Mater. Sol. Cells 2000, 61, 427; b) S. Agarwala, M. Kevin, A. S. W. Wong, C. K. N. Peh, V. Thavasi, G. W. Ho, Appl. Mater. interfaces 2010, 2, 1844. [3] K. Hou, B. Tian, F. Li, Z. Bian, D. Zhao, C. Huang, J. Mater. Chem. 2005, 15, 2414.
ChemPhysChem 2014, 15, 187 – 194
193
CHEMPHYSCHEM ARTICLES [4] a) S. Lee, I. Cho, J. H. Lee, D. H. Kim, D. W. Kim, J. Y. Kim, H. Shin, J. Lee, H. S. Jung, N. Park, K. Kim, M. J. Ko, K. S. Hong, Chem. Mater. 2010, 22, 1958; b) D. Chen, F. Huang, Y. Cheng, R. A. Caruso, Adv. Mater. 2009, 21, 2206. [5] a) B. Liu, E. S. Aydil, J. Am. Chem. Soc. 2009, 131, 3985; b) F. Sauvage, D. Chen, P. Comte, F. Huang, L. Heiniger, Y. Cheng, R. A. Caruso, M. Graetzel, ACS Nano 2010, 4, 4420. [6] a) J. Jiu, F. Wang, M. Sakamoto, J. Takao, M. Adachi, Sol. Energy Mater. Sol. Cells 2005, 87, 77; b) H. Choi, Y. J. Kim, R. S. Varma, D. Dionysiou, Chem. Mater. 2006, 18, 5377. [7] a) T. Ren, Z. Yuan, B. Su, Chem. Phys. Lett. 2003, 374, 170; b) R. Linacero, J. Aguado-Serrano, M. Rojas-Cervantes, Journal of Materials Science 2006, 41, 2457. [8] a) J. Chen, Z. Hua, Y. Yan, A. A. Zakhidov, R. H. Baughmand, L. Xu, Chem. Commun. 2010, 46, 1872; b) S. S. Kim, H. I. Lee, J. K. Shon, J. Y. Hur, M. S. Kang, S. S. Park, S. S. Kong, J. A. Yu, M. Seo, D. Li, S. S. Thakur, J. M. Kim, Chem. Lett. 2008, 37, 140; c) T. Leshuk, S. Linley, G. Baxter, F. Gu, Appl. Mater. Interfaces 2012, 4, 6062. [9] a) K. Hwang, D. W. Cho, J. Lee, C. Im, New J. Chem. 2012, 36, 2094; b) B. Yu, X. Jiang, J. Yin, Nanoscale 2013, 5, 5489. [10] J. Brinker, Y. Lu, A. Sellinger, H. Fan, Adv. Mater. 1999, 11, 579. [11] S. Ding, J. S. Chen, Z. Wang, Y. L. Cheah, S. Madhavi, X. Hub, X. W. Lou, J. Mater. Chem. 2011, 21, 1677. [12] S. Wong, P. Huang, C. Wu, B. Xu, C. Chiang, H. Yen, Appl. Phys. Lett. 2006, 89, 153111. [13] B. D. Cullity, Elements of X-ray Diffraction, Addison Wesley Pub, Notre Dame, 1978. [14] J. Ye, W. Liu, J. Cai, S. Chen, X. Zhao, H. Zhou, L. Qi, J. Am. Chem. Soc. 2011, 133, 933.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org [15] U. Diebold, Surface Science Reports 2003, 48, 53. [16] F. Sayilkan, M. Asilturk, S. Sener, S. Erdemoglu, M. Erdemoglu, H. Sayilkan, Turk. J. Chem. 2007, 31, 211. [17] a) J. B. Joo, Q. Zhang, M. Dahl, I. Lee, J. Goebl, F. Zaera, Y. Yin, Energy Environ. Sci. 2012, 5, 6321; b) J. B. Joo, M. Dahl, N. Li, F. Zaera, Y. Yin, Energy Environ. Sci. 2013, 6, 2082; c) J. B. Joo, Q. Zhang, I. Lee, M. Dahl, F. Zaera, Y. Yin, Adv. Funct. Mater. 2012, 22, 166. [18] D. V. Bavykin, V. N. Parmon, A. A. Lapkin, F. C. Walsh, J. Mater. Chem. 2004, 14, 3370. [19] a) J. T. Park, D. K. Roh, R. Patel, E. Kim, D. Y. Ryu, J. H. Kim, J. Mater. Chem. 2010, 20, 8521; b) E. Dell’Orto, L. Raimondo, A. Sassella, A. Abbotto, J. Mater. Chem. 2012, 22, 11364; c) H. Xu, X. Tao, D. Wanga, Y. Zheng, J. Chen, Electrochimica Acta 2010, 55, 2280. [20] J. Jiu, F. Wang, M. Sakamoto, J. Takao, M. Adachi, Sol. Energy Mater. Sol. Cells 2005, 87, 77. [21] a) P. M. Sommeling, B. C. O’Regan, R. R. Haswell, H. J. P. Smit, N. J. Bakker, J. J. T. Smits, J. M. Kroon, J. A. M. van Roosmalen, J. Phys. Chem. B. 2006, 110, 19191; Regan, R. R. Haswell, H. J. P. Smit, N. J. Bakker, J. J. T. Smits, J. M. Kroon, J. A. M. van Roosmalen, J. Phys. Chem. B. 2006, 110, 19191; b) B. C. O’Regan, J. R. Durrant, P. M. Sommeling, N. J. Bakker, J. Phys. Chem. C. 2007, 111, 14001. [22] Y. Li, W. Lee, D. Lee, K. Kim, N. Park, M. J. Ko, Appl. Phys. Lett. 2011, 98, 103301. [23] Q. Zhang, G. Cao, Nano Today 2011, 6, 91.
Received: September 23, 2013 Published online on December 4, 2013
ChemPhysChem 2014, 15, 187 – 194
194
Structurally Stabilized Organosilane-Templated Thermostable Mesoporous Titania Vipin Amoli ,[a, b] Rashmi Tiwari ,[a, b, c] Arghya Dutta,[d] Asim Bhaumik,[d] and Anil Kumar Sinha*[a, b] Structurally thermostable mesoporous anatase TiO2 (m-TiO2) nanoparticles, uniquely decorated with atomically dispersed SiO2, is reported for the first time. The inorganic Si portion of the novel organosilane template, used as a mesopores-directing agent, is found to be incorporated in the pore walls of the titania aggregates, mainly as isolated sites. This is evident by transmission electron microscopy and high-angle annular dark field scanning transmission electron microscopy, combined with electron dispersive X-ray spectroscopy. This type of unique structure provides exceptional stability to this new material against thermal collapse of the mesoporous structure, which is reflected in its high surface area (the highest known for anatase titania), even after high-temperature (550 8C) calcination. Control of crystallite size, pore diameter, and surface area is achieved by varying the molar ratios of the titanium
precursor and the template during synthesis. These mesoporous materials retain their porosity and high surface area after template removal and further NaOH/HCl treatment to remove silica. We investigate their performance for dye-sensitized solar cells (DSSCs) with bilayer TiO2 electrodes, which are prepared by applying a coating of m-TiO2 onto a commercial titania (P25) film. The high surface area of the upper mesoporous layer in the P25–m-TiO2 DSSC significantly increases the dye loading ability of the photoanode. The photocurrent and fill factor for the DSSC with the bilayer TiO2 electrode are greatly improved. The large increase in photocurrent current (ca. 56 %) in the P25–m-TiO2 DSSC is believed to play a significant role in achieving a remarkable increase in the photovoltaic efficiency (60 %) of the device, compared to DSSCs with a monolayer of P25 as the electrode.
1. Introduction Alternative energy sources have received increasing attention over the past decades for the replacement of environmentally damaging and diminishing fossil fuels. Solar energy is one of the most attractive renewable-energy sources. DSSCs based on interpenetrating networks composed of mesoporous, nanocrystalline, wide bandgap semiconductor oxides, such as TiO2, and liquid electrolytes represent a possible alternative to silicon-based solar cells, owing to their ease of fabrication and low cost.[1] Photoexcitation of a dye adsorbed on the surface of TiO2 results in injection of an electron into the conduction
[a] V. Amoli , R. Tiwari , Dr. A. K. Sinha Catalytic Conversion Process Division, Council of Scientific and Industrial Research (CSIR) Indian Institute of Petroleum Mohkampur, Dehradun 248005 (India) Fax: (+ 91) 135-266-203 E-mail: [email protected] [b] V. Amoli , R. Tiwari , Dr. A. K. Sinha CSIR–Network Institute of Solar Energy (CSIR–NISE) Anusandhan Bhawan, New Delhi 110001 (India) [c] R. Tiwari Department of Chemistry, University of Calgary 2500 University Drive N.W. Calgary, Alberta, T2N 1N4 (Canada) [d] A. Dutta, Prof. A. Bhaumik Department of Material Science, Indian Association for the Cultivation of Science Jadavpur, Kolkata, 700032 (India) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201300872.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
band of TiO2. The original state of the dye is restored by the subsequent donation of an electron from the liquid electrolyte in the pores of mesoporous oxide back to the dye. Surface area and porosity are crucial factors in determining the amount of dye adsorbed, which, in turn, affect the performance of DSSCs.[2] To increase the efficiency of a DSSC, a highsurface-area photoanode composed of mesoporous anatase TiO2 is highly desirable, as it maximizes the dye-loading and gives rise to a high photovoltaic efficiency.[2a, 3] Several different techniques have been reported to produce TiO2 nanostructures such as sol gel,[1a, 4] hydrothermal,[5] and template-modified sol gel.[6] There have been several reports regarding the surfactant-assisted synthesis of mesoporous TiO2.[7] Recently, there have been few reports on the synthesis of mesoporous TiO2 employing a mesoporous silica template.[8] These TiO2 structures generally possess a high surface area, better structural stability, an excellent crystalline frame work, and improved photovoltaic performances in DSSCs.[9] Evaporation-induced self-assembly (EISA), explored by Brinker and co-workers, provides an ingenious route to synthesize mesoporous materials.[10] Herein, we report the synthesis of mesoporous anatase TiO2 nanocrystals of high surface areas (as high as 361 m2g1) and with a mean pore diameter of 2.2 nm. We use EISA with a novel organosilane template as a structuredirecting agent. The inorganic Si component of the organosilane template improves the stability of the organic–inorganic
ChemPhysChem 2014, 15, 187 – 194
187
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
(JCPDS card tase TiO2 No. 21:1272).[11] No impurity peaks are observed. The broadening and slight shifting of the (101) peak towards larger 2q values indicates that the average crystallite size decreases with increasing amounts of surfactant, as shown in the Supporting Information (Figure S1). This may be attributed to increased steric repulsion between the growing titania particles caused by the template molecule, which has a suppressing effect on the crystal growth, resulting in better dispersion and size reduction.[12] The broadening of the diffraction peaks is attributed to the size confinement. The average crystallite sizes of T1, T2, T3, and T4 samples, estimated from the full width at half maximum of Scheme 1. The Si component of the organosilane template improves the stability of mesoporous TiO2 structures the (101) peak by using the because of organic–inorganic interactions during synthesis. Scherrer[13] equation, were 5.9, 5.1, 4.9, and 4.3 nm, respectively. The morphology of the different mesoporous titania samcomposite-mesostructured gel when combined with the inorples, prepared through the EISA method, was investigated by ganic Ti precursor (Scheme 1). using field-emission scanning electron microscopy (FESEM), as This is the first report of mesoporous titania uniquely decoshown in Figure 2. The synthesized TiO2 nanostructure aggrerated with isolated SiO2 sites, which gives exceptional thermal stability and very high surface area to these materials. These gates are seen as irregular particles with a size of 10–30 nm. mesoporous crystalline materials are been used, after template Energy dispersive X-ray spectroscopy (EDS) performed near the removal, for the fabrication of a photoanode for DSSCs to impore-walls (given in Figure S2 in the Supporting Information) prove the power-conversion efficiency compared to commerrevealed that the products are composed of Ti, Si, and O. The cial P25. elemental compositions of the obtained materials (near the pore walls) are listed in Table 1. The materials in the bulk region were composed of titania only, whereas the Si atoms
2. Results and Discussion The X-ray diffraction (XRD) patterns of the product, prepared with different precursor and template ratios, are shown in Figure 1. The patterns can be assigned to pure tetragonal ana-
Figure 1. XRD patterns of the mesoporous TiO2 samples T1 (Ti/template = 17), T2 (Ti/template = 8), T3 (Ti/template = 6), and T4 (Ti/template = 4) calcined at 550 8C.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 2. FESEM of mesoporous TiO2 samples calcined at 550 8C. a) T1, b) T2, c) T3, and d) T4.
ChemPhysChem 2014, 15, 187 – 194
188
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 1. Surface and porous characteristics measured from N2 adsorption/desorption isotherms for the mesoporous TiO2 (samples T1, T2, T3, and T4). Sample
Ti/Template
Ti/Si[a]
Surface area [m2 g1]
Mean pore diameter [nm]
Pore volume [cm3 g1]
Porosity [%]
Roughness factor
Particle size [nm]
T1 T2 T3 T4
17 8 6 4
2.16 1.55 1.39 1.11
213.5 295.3 352.6 361.0
4.654 3.428 2.882 2.203
0.2437 0.2531 0.2538 0.1988
48.6 49.4 49.6 43.0
426 578 690 800
7.2 5.2 4.3 4.3
[a] Composition obtained near the pore walls of the mesoporous samples from EDS analysis.
Figure 3. Elemental maps for a) Ti, b) Si, and c) Si and Ti together, obtained by using EDS (Ti + Si: white grey dots = Si, black grey patches = Ti, black area = pores) for sample T3. Inset is a representation of atomically dispersed Si inside the material.
(Figure 4 a, b). As shown in Scheme 1, fine nanoparticles of TiO2 are aggregated around the organosilane micelle mesoporous template, which leads to an interconnected porous network (Figure 4 a, b) after removal of the organic part of the template during calcination. The insets show narrow size distributions of the nanoparticles, with an abundant of size of 4–6 nm and 3– 5 nm for sample T2 and T3, respectively, as also evidenced by XRD analysis. The obtained nanoparticles are well-dispersed, homogeneous, and dense, which can be seen in the Supporting Information (Figure S3). The HRTEM images (Figure 4 c –f) show the presence of multiple individual crystallites with clear lattice fringes. It can be observed that the pores are formed because of the intercrystallite voids of the aggregated titania crystallites. This structure enhances the surface area of the sample. The observed spacing of 0.35 nm between neighboring lattice fringes corresponds to the (101) plane of anatase TiO2.[14] The inset shows corresponding selected-area electron diffraction (SAED) pattern with diffraction spots of (101) and (200) of randomly oriented anatase TiO2 crystallites.
(from the template) decorate the pore walls. It is also observed that the concentration of silica in the prepared samples [coming from the octadecyldimethyl(3-trimethoxysilylpropyl)ammonium chloride (ODAC) template] increases with increasing template concentration. Selected elemental maps (Figure 3) were recorded to identify the dispersion of SiO2 on the surface of the TiO2 matrix. It can be seen that Si is highly dispersed in the material. The grey/ black-colored regions represent titania patches and the grey/ white dots represent silica. The Si sites are seen to be atomically dispersed and are decorating the surface of the titania patches (Scheme 1). This is a unique nanostrucutre for porous anatase titania, in which Si is atomically incorporated in the structure through a chemical method. Transmission electron microscopy (TEM) images of the T2 and T3 samples are shown in Figures 4 a and b, and high-resolution TEM (HRTEM) images are shown in Figures 4 c–f. Very fine well-interconnected particles, with a diameter of only few nanometres Figure 4. TEM images of a) T2 and b) T3 with the inset showing particles-size distribution. HRTEM images showing (ca. 5 nm), are predominantly intercrystallite voids of c) T2 and d) T3. e, f) HRTEM images showing (101) anatase crystallites (insets show correseen forming a porous network sponding FFT patterns). 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 187 – 194
189
CHEMPHYSCHEM ARTICLES
Figure 5. a) Representative HAADF-STEM image of the T3 sample. The corresponding compositional maps for b) Ti, c) Si, and d) Si and Ti together, obtained by using EDS.
www.chemphyschem.org mesopores, showing that the sample porosity is somewhat homogenous. The pore-size distribution analysis by using the Barrett–Joyner–Halenda (BJH) method showed a narrow pore-size distribution, with mean pore diameter of 2–5 nm. The template concentration variation helps to control the pore size and surface area of these mesoporous titania materials. As the amount of template is increased during the synthesis, the surface area of the mesoporous TiO2 increases, as shown in Table 1. The surface area increased (213–360 m2g1) and the pore size decreased (4.6–2.2 nm) with decreasing Ti/template molar ratio (from 17 to 4) during the synthesis (Table 1). This increase in surface area and decrease in pore size can be correlated with the decrease in the particle size with increasing template concentration, as observed from XRD analysis. This may be caused by increased steric hindrance from the template adsorbed on the surface of TiO2 during synthesis, which retards particle agglomeration and results in a decrease in the particle size[15] that, in turn, increases the surface area. A decrease in the mean pore diameter and pore volume is observed with an increase in the amount of ODAC, which may be attributed to two reasons. 1) As the particle size decreases, the intercrystallite void size also decreases, and 2) as there is an increase in the amount of ODAC, more SiO2 is embedded into the mesoporous TiO2 framework pore walls, which results in a decrease in the pore size and the pore volume.[16] EDS spectra show an increased amount of SiO2 in the TiO2 matrix with increasing ODAC content (from 2.9 to 7.6 at %). The loss of hysteresis in sample T4 may be caused by an ink-bottle effect from the increased amount of SiO2 that is present in the TiO2 framework, which reduces the pore diameter and pore volume significantly. The large surface area of T4 has limitations of reduced pore volume and pore diameter. There is no detectable effect on the crystalline phase (observed in the XRD pattern)
The composition of the sample was determined by using scanning transmission electron microscopy (STEM)–EDS. Figure 5 a shows representative high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) image of the sample. A 3D porous integrity of interconnected particles was observed. The compositional EDS mappings showing elemental distributions for Ti, Si, and Ti + Si, over the same area highlighted in HAADF-STEM image, are shown in Figures 5 b–d. The overlap of Si and Ti elemental maps (Figure 5 d) shows that Si is atomically dispersed along the pore walls, over the titania patches given in the Supporting Information (Figure S4). The presence of atomically dispersed Si along the pore walls of the titania framework improves the thermal stability of the mesoporous titania structures, resulting in anatase crystallization without the collapse of mesoporous structures. The specific surface area and pore-size distribution of the mesoporous titania were characterized by using nitrogen gas sorption. Figure 6 shows the isotherms of the calcined mesoporous TiO2 samples. Mesoporous TiO2 exhibits type IV isotherms with a hysteresis loop, which is a typical characteristic of mesoporous materials. For the mesoporous TiO2 samples, a sharp increase in the adsorption volume of N2 was observed between a relative partial pressure (P/P0) of 0.4–0.7, which may be attrib- Figure 6. N2 adsorption/desorption isotherm (left) and BJH pore-size distribution (right) of the mesoporous TiO2 uted to the sharp uptake in the samples. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 187 – 194
190
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
analysis (see Figure S7 in the Supporting Information), without changing its porous structure (Figure 7). The specific surface area decreased after treatment with NaOH aqueous solution (Table 2), which may be attribSample Surface area Pore volume Mean pore Porosity Particle uted to the adsorption of Na + on the TiO2 surface, lead[m2g1] [cm3g1] diameter [nm] [%] size [nm] ing to the formation of Na +OTi bonds, as observed T3 352.6 0.2538 2.88 49.6 4.3 during the NaOH hydrothermal treatment of other TiO2 after NaOH 266.1 0.2569 3.86 49.9 5.8 nanostructures by Dmitry et al.[18] In our case, we did treatment after HCl/ 313.3 0.2901 3.76 53.0 4.9 not employ hydrothermal treatment, but the very high distilled-water surface area (360 m2g1) of TiO2 nanostructures may washing offer a large number of active sites for Na + adsorption. There is a large increase in the mean pore diameter and pore volume during NaOH treatment, which may be attributed to leaching of the silica present in the TiO2 or the morphology of the mesoporous TiO2 after NaOH/HCl matrix, because it is shown by HRTEM that the pores are intertreatment (see Figures S5 and S6 in the Supporting Informacrystallite voids, which may be formed during the sintering tion). There is a slight increase in particle size from 4.3 to process. The surface area increased significantly upon treat4.9 nm after treatment (Table 2). ment with aqueous HCl solution. The treatment with dilute In our case, silica is atomically incorporated in the product, HCl is an ion-exchange phenomenon, in which TiONa and mainly decorating as isolated sites, particularly at the pore TiOH bonds react with acid and the formation of new TiO walls of the titania aggregates, which produce a protective Ti bonds can occur. The amount of dye on the surface of three effect and limit the grain growth (to ca. 4.3 nm) after calcinaphotoelectrodes was measured by the method described in tion at 550 8C. After NaOH/HCl treatment, owing to silica etchref. [19] The amount of N719 dye on the different TiO2 photoing, this protective effect is diminished, resulting in the strucelectrodes is summarized in Table 3. tural rearrangement of TiO2 units and slight grain growth Figure 8 shows the photocurrent density behavior of the during further drying, as explained in the literature.[17] The shape and size of the resulting material, after treating with DSSCs, made from different working electrodes, as a function NaOH (2 m) followed by 0.1 m HCl, is shown by the FESEM of voltage. Device A, device B, and device C were tested under images in the Supporting Information (Figure S6). The NaOH similar simulated irradiation (global AM 1.5, 100 mW cm2), and HCl treatment process results in the removal of any silica with an active area of 0.16 cm2. Table 3 summarizes the perforpresent in the mesoporous sample (T3), as confirmed by EDS mance parameters of the different DSSCs. The TiO2 electrode made from P25 nanoparticles showed a short-circuit current density of 8.13 mA cm2 and an energy conversion efficiency of 3.49 %, whereas the P25–m-TiO2 electrode achieved a current density of 12.73 mA cm2 and a conversion efficiency of 5.58 %, indicating a 60 % increase in the conversion efficiency compared to the P25 electrode, despite of similar open circuit voltage. A large increase in the photocurrent current (ca. 56 %) in the P25–m-TiO2 system, as compared to the bare P25 system, is believed to play a significant role in determining the overall efficiency of the device. The higher photocurrent in case of the P25m-TiO2 photoanode is attributed to two factors. First, the higher dye-loading capabilities of the high-surface-area upper mesoporous layer in the P25–m-TiO2 photoanode increases the Figure 7. N2 physiosorption (left) and BJH pore-size distribution (right) of TiO2 (sample T3) with NaOH treatment amount of light harvested, (top) followed by 0.1 m HCl treatment (bottom). Table 2. Physiological characteristics of mesoporous TiO2 (sample T3) before and after NaOH and HCl/distilled-water treatment, measured from N2 adsorption-desorption isotherms.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 187 – 194
191
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 3. Photovoltaic performance parameters of DSSC made from P25 nanoparticles (device A), P25–m-TiO2 (device B), and P25–m-TiO2 after NaOH and HCl treatment (device C). Solar cell
Jsc [mA cm2]
Voc [V]
FF [%]
h [%]
Adsorbed dye [mol cm2]
device A device B device C
8.13 12.36 12.73
0.694 0.698 0.698
61.7 60.3 62.5
3.49 5.22 5.58
0.87 107 1.65 106 1.71 106
Figure 8. Current–voltage curves of DSSCs made from A) P25 nanoparticles, B) P25–m-TiO2 (sample T3), C) P25–m-TiO2 (sample T3) after NaOH and HCl treatment.
3. Conclusions Mesoporous, nanocrystalline, thermally stable anatase titania with a high surface area (up to 360 m2g1 after calcination at 550 8C) and narrow pore-size distribution (4.4–2.2 nm) was produced through a simple and reproducible modified sol-gel and evaporation-induced self-assembly method by using a novel organosilane template. Crystallographic phases, particle size, and morphology of the TiO2 nanostructures were characterized by using XRD, FESEM, TEM, and Brunauer–Emmett–Teller (BET) techniques. Detailed characterization established that the Si atoms from the organosilane are atomically dispersed along the pore walls of the titania matrix. This unique structure and composition provides the required structural and thermal stability of this new material, which is reflected in a surface area that is the highest known for anatase titania, even after hightemperature (550 8C) calcination. After removal of silica from the pore walls, the structural integrity, porosity, and surface areas are maintained. This new material has the potential for practical applications because of its large surface area and porous structure. These mesoporous TiO2 particles improve the dye loading through their high surface area. A high light-tocurrent conversion efficiency of 5.58 % could be achieved when applied as a photoanode in a DSSC, as compared to a photoanode consisting of P25 nanoparticles.
Experimental Section which is highly desirable for achieving a better performance of the device, as described in many reports.[2a, 3, 21] In particular, the cell short-circuit current (Jsc) is directly proportional to the light-harvesting ability of the cell, which, in turn, is strictly dependent on the dye concentration on the TiO2 adsorptive surface.[20] Second, the enhancement in photocurrent in the P25– m-TiO2 system is thought to be a result of improved interparticle connectivity, owing to the presence of smaller (ca. 5 nm) high-surface-area particles among the larger P25 (ca. 25 nm) particles. During the photoanode sintering process, described above, these smaller particles may serve as a binding agent/filling agent between the larger P25 nanoparticles, and, at the same time, these mesoporous particles can offer a large surface area for dye adsorption. There are several reports that have obtained results similar to ours, in which the interparticle connectivity among TiO2 particles is improved by several techniques, such as the post treatment of sintered electrode by using TiCl4,[21] incorporating P25 into the mesoporous TiO2 matrix, and using 5 nm pure anatase nanoparticles as a nanoglue by mixing it with P25 nanoparticles.[22] There is an increase in the short-circuit current density, fill factor, and photoconversion efficiency for the photoelectrode made from mesoporous TiO2 obtained after NaOH/HCl treatment, as compared to the untreated sample, which may be attributed to the increase in pore diameter and porosity, which favors the diffusion of the redox electrolyte (I/I3) through the nanosized pores in the mesoporous TiO2 film, resulting in an increased fill factor that improved the efficiency of the DSSC.[23] 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Transparent conducting oxide (TCO) glass (Solaronix), Ruthenium 535 dye [Ru(bpy)2(NCS)2H4, N719; Solaronix], iodolyte TG-50 (Solaronix), and platisol (Solaronix) were used as received. Titanium (IV) isopropoxide [Ti {OCH (CH3)2}4, 95 % Alfa Aesar] was used as a source of titanium. ODAC (60 % in methanol; Gelest, Inc), glacial acetic acid (Merck), ethanol (Merck, anhydrous), Triton X-100 [C14H22O(C2H4O)n, Loba Chemie], P25 (Sigma–Aldrich, particle size of 27 nm, specific surface area of 57 m2g1, and the phase composition of 80 % anatase and 20 % rutile), 67–72 % HNO3 (Merck) were used without further purification.
Synthesis of Mesoporous TiO2 Particles ODAC was added as a mesopore template for the synthesis of mesoporous titania. In a typical synthesis, ODAC (2 mL) was dissolved in anhydrous ethanol (40 mL) and stirred for 30 min at room temperature (solution 1). Titanium isopropoxide (TIP; 40 mm) was dissolved in a mixed solvent composed of 69–72 wt % HNO3 acid (6.4 mL) and anhydrous ethanol (20 mL) (solution ). Solution 1 was added drop-wise to solution 2, with continuous stirring. Next, additional ethanol (20 mL) was added. The final gel composition was 1:3.8:0.06:25.5 TIP/HNO3/ODAC/EtOH. The solution mixture was stirred for 24 h at room temperature. After that, the solution was transferred to a petri dish and solvent evaporation was carried out at 70 8C for 48 h, followed by calcination in air in a furnace at 550 8C for 6 h. A series of mesoporous TiO2 samples (T1, T2, T3, and T4) were prepared by changing the Ti/template (molar) ratios to 17, 8, 6, and 4, respectively. The organic part of the template was removed by calcination. The silica from the template present in the as-synthesized mesoporous titania samples was removed by treating the samples in a 2 m NaOH solution at 80 8C for 24 h under continuous stirring. The samples were then washed with ChemPhysChem 2014, 15, 187 – 194
192
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
0.1 m HCl and then with distilled water before using as the materials as the anode in a DSSC.
Preparation of the TiO2 Paste for the DSSC
in which Jsc is a short-circuit current density (mA cm2), Voc is an open-circuit voltage (V), Pin is an incident light power, FF is the fill factor, h is an overall energy-conversion efficiency, and Jmax (mA cm2) and Vmax (V) are the current density and voltage in the J–V curve, respectively, at the point of maximum power output.
TiO2 paste was prepared by mixing the synthesized TiO2 powder (0.25 g) with acetic acid (0.5 mL), which was then ground mechanically for 15 min. Next, a mixture of water and ethanol (1:3) (15 mL) was added dropwise, and this mixture was ground mechanically. Finally, Triton X-100 (0.5 mL) was added to facilitate the spreading of the paste on the substrate. The resultant slurry was ground for another 15 min to obtain a homogeneous paste. A similar method was adopted for the preparation of a paste with P25 nanoparticles.
For dye-loading measurements, photoanodes with a dye-loaded area of 4 cm2 and a thickness of approximately 7 mm were dipped into 10 mL of a 0.1 m solution of NaOH in ethanol/H2O (1:1 v/v), until complete desorption of the N719 dye occurred. The alkaline solution containing the fully desorbed dye was carefully measured by using UV/Vis spectra (Hitachi U-2900, UV/Vis Double Beam spectrophotometer); the Beer–Lambert law [Eq. (2)] was used to calculate the number of adsorbed N719 dye molecules by using the absorption value at 515 nm.
Device Fabrication
A ¼ elc
Three working electrodes [device A with P25 nanoparticles, device B with one layer of P25 and second layer of m-TiO2 (sample T3, without NaOH/HCl treatment), and device C with one layer of P25 and second layer of m-TiO2 (sample T3, after NaOH/HCl treatment)] were fabricated. The doctor-blade technique was used for coating the TiO2 paste on the TCO glass. The TiO2-film-coated TCO samples were sintered at 450 8C for 30 min. For sensitization, the films were impregnated with 0.3 mm N719 dye in ethanol for 24 h at room temperature. The counter electrode was obtained by depositing a thin layer of platisol T/SP, which can be squeegee-printed by using a polyester mesh, and heat-treated in air for 30 min at 450 8C. One drop of iodolyte TG-50 electrolyte was put on the surface of the dye-loaded TiO2 photoanode, which penetrated inside the porous TiO2 via capillary action. A Pt-coated fluorine–tin oxide electrode was then clipped onto the top of the TiO2 working electrode to form the complete DSSC. The active area of each DSSC was about 0.16 cm2. A similar fabrication procedure is adopted for all DSSCs.
Characterization Crystalline-structure and phase identification of the samples was performed by using XRD measurements on a Bruker D8 Advance diffractometer, operating in the reflection mode with CuKa radiation (40 KV, 40 mA). Surface morphology of the samples was obtained by using FESEM (Quanta 200F, Netherlands). EDS spectroscopy combined with SEM was used to analyze the composition of the samples. The BET specific surface area, BJH-calculated average pore size, and pore volume of the TiO2 nanocrystals were obtained from the N2 adsorption/desorption isotherms recorded on a BELSORP max instrument (Japan) at 77 K. The samples were degassed and dried under a vacuum system at 150 8C for 4–5 h prior to the measurement. TEM studies were performed by using a JEOL 2010F TEM/STEM microscope, operating at 200 kV. This instrument was equipped with a JEOL high-angle annular dark field detector and an Oxford X-Max Silicon Drift X-Ray EDS (XEDS) detector, enabling the acquisition of STEM images and compositional analysis, either in spot mode or through elemental mapping using a 0.5 nm electron probe. Photovoltaic performance of the assembled DSSCs was measured by completely integrated current–voltage (I–V) Test Station (PVIV-211 V) with a Kiethley 2420 source meter and a 94023 A Oriel Sol3 A, class AAA solar simulator (power output = 100 mW cm2, lamp power = 450 W) equipped with an AM 1.5 filter. The solar conversion efficiency was calculated by using the formula: h ð%Þ ¼ ðV max Jmax Þ=Pin 100 ¼ ðV oc Jsc FFÞ=Pin 100 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ð1Þ
ð2Þ
in which A is the absorption of the UV/Vis spectra at 515 nm, e is the molar absorptivity (14100 m1 cm1), l is the path length of light beam, and c is the concentration of the dye. The porosity of the TiO2 film was calculated according to Equation (3):[19a] ðPÞ ¼ V p =ð11 þ V p Þ
ð3Þ
in which (P) is the porosity, Vp is the specific pore volume (cm3g1), and 11 is the reciprocal of the density of anatase TiO2 (11 = 0.257 cm3g1). The roughness factor (R) per unit of film thickness and the particle size (D) were calculated using equation Equations (4) and (5), respectively:[19a] ðRÞ ¼ 1ð1-PÞS
ð4Þ
ðDÞ ¼ 6000=ðS 1Þ
ð5Þ
in which 1 is the density of anatase TiO2 (3.89 cm3g1), P is the porosity of the sample, and S is the specific surface area (m2g1).
Acknowledgements The director of the Indian Institute of Petroleum is acknowledged for approving the work. V.A. thanks CSIR India for a Junior Research Fellowship. The authors thank the Analytical Science Division at the Indian Institute of Petroleum for analytical services. The Ministry of New and Renewable Energy, New Delhi is acknowledged for research funding. Keywords: mesoporous materials · micelles · surface area · thermostable · titania [1] a) B. O. Regan, M. Grtzel, Nature 1991, 353, 737; b) M. A. Khan, M. S. Akhtar, O. Yang, Solar Energy 2010, 84, 2195. [2] a) S. Kambe, K. Murakoshi, T. Kitamura, Y. Wada, S. Yanagida, H. Kominami, Y. Kera, Sol. Energy Mater. Sol. Cells 2000, 61, 427; b) S. Agarwala, M. Kevin, A. S. W. Wong, C. K. N. Peh, V. Thavasi, G. W. Ho, Appl. Mater. interfaces 2010, 2, 1844. [3] K. Hou, B. Tian, F. Li, Z. Bian, D. Zhao, C. Huang, J. Mater. Chem. 2005, 15, 2414.
ChemPhysChem 2014, 15, 187 – 194
193
CHEMPHYSCHEM ARTICLES [4] a) S. Lee, I. Cho, J. H. Lee, D. H. Kim, D. W. Kim, J. Y. Kim, H. Shin, J. Lee, H. S. Jung, N. Park, K. Kim, M. J. Ko, K. S. Hong, Chem. Mater. 2010, 22, 1958; b) D. Chen, F. Huang, Y. Cheng, R. A. Caruso, Adv. Mater. 2009, 21, 2206. [5] a) B. Liu, E. S. Aydil, J. Am. Chem. Soc. 2009, 131, 3985; b) F. Sauvage, D. Chen, P. Comte, F. Huang, L. Heiniger, Y. Cheng, R. A. Caruso, M. Graetzel, ACS Nano 2010, 4, 4420. [6] a) J. Jiu, F. Wang, M. Sakamoto, J. Takao, M. Adachi, Sol. Energy Mater. Sol. Cells 2005, 87, 77; b) H. Choi, Y. J. Kim, R. S. Varma, D. Dionysiou, Chem. Mater. 2006, 18, 5377. [7] a) T. Ren, Z. Yuan, B. Su, Chem. Phys. Lett. 2003, 374, 170; b) R. Linacero, J. Aguado-Serrano, M. Rojas-Cervantes, Journal of Materials Science 2006, 41, 2457. [8] a) J. Chen, Z. Hua, Y. Yan, A. A. Zakhidov, R. H. Baughmand, L. Xu, Chem. Commun. 2010, 46, 1872; b) S. S. Kim, H. I. Lee, J. K. Shon, J. Y. Hur, M. S. Kang, S. S. Park, S. S. Kong, J. A. Yu, M. Seo, D. Li, S. S. Thakur, J. M. Kim, Chem. Lett. 2008, 37, 140; c) T. Leshuk, S. Linley, G. Baxter, F. Gu, Appl. Mater. Interfaces 2012, 4, 6062. [9] a) K. Hwang, D. W. Cho, J. Lee, C. Im, New J. Chem. 2012, 36, 2094; b) B. Yu, X. Jiang, J. Yin, Nanoscale 2013, 5, 5489. [10] J. Brinker, Y. Lu, A. Sellinger, H. Fan, Adv. Mater. 1999, 11, 579. [11] S. Ding, J. S. Chen, Z. Wang, Y. L. Cheah, S. Madhavi, X. Hub, X. W. Lou, J. Mater. Chem. 2011, 21, 1677. [12] S. Wong, P. Huang, C. Wu, B. Xu, C. Chiang, H. Yen, Appl. Phys. Lett. 2006, 89, 153111. [13] B. D. Cullity, Elements of X-ray Diffraction, Addison Wesley Pub, Notre Dame, 1978. [14] J. Ye, W. Liu, J. Cai, S. Chen, X. Zhao, H. Zhou, L. Qi, J. Am. Chem. Soc. 2011, 133, 933.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org [15] U. Diebold, Surface Science Reports 2003, 48, 53. [16] F. Sayilkan, M. Asilturk, S. Sener, S. Erdemoglu, M. Erdemoglu, H. Sayilkan, Turk. J. Chem. 2007, 31, 211. [17] a) J. B. Joo, Q. Zhang, M. Dahl, I. Lee, J. Goebl, F. Zaera, Y. Yin, Energy Environ. Sci. 2012, 5, 6321; b) J. B. Joo, M. Dahl, N. Li, F. Zaera, Y. Yin, Energy Environ. Sci. 2013, 6, 2082; c) J. B. Joo, Q. Zhang, I. Lee, M. Dahl, F. Zaera, Y. Yin, Adv. Funct. Mater. 2012, 22, 166. [18] D. V. Bavykin, V. N. Parmon, A. A. Lapkin, F. C. Walsh, J. Mater. Chem. 2004, 14, 3370. [19] a) J. T. Park, D. K. Roh, R. Patel, E. Kim, D. Y. Ryu, J. H. Kim, J. Mater. Chem. 2010, 20, 8521; b) E. Dell’Orto, L. Raimondo, A. Sassella, A. Abbotto, J. Mater. Chem. 2012, 22, 11364; c) H. Xu, X. Tao, D. Wanga, Y. Zheng, J. Chen, Electrochimica Acta 2010, 55, 2280. [20] J. Jiu, F. Wang, M. Sakamoto, J. Takao, M. Adachi, Sol. Energy Mater. Sol. Cells 2005, 87, 77. [21] a) P. M. Sommeling, B. C. O’Regan, R. R. Haswell, H. J. P. Smit, N. J. Bakker, J. J. T. Smits, J. M. Kroon, J. A. M. van Roosmalen, J. Phys. Chem. B. 2006, 110, 19191; Regan, R. R. Haswell, H. J. P. Smit, N. J. Bakker, J. J. T. Smits, J. M. Kroon, J. A. M. van Roosmalen, J. Phys. Chem. B. 2006, 110, 19191; b) B. C. O’Regan, J. R. Durrant, P. M. Sommeling, N. J. Bakker, J. Phys. Chem. C. 2007, 111, 14001. [22] Y. Li, W. Lee, D. Lee, K. Kim, N. Park, M. J. Ko, Appl. Phys. Lett. 2011, 98, 103301. [23] Q. Zhang, G. Cao, Nano Today 2011, 6, 91.
Received: September 23, 2013 Published online on December 4, 2013
ChemPhysChem 2014, 15, 187 – 194
194
