Integration of High-Performance Nanocrystalline TiO2 Photoelectrodes for N719-Sensitized Solar Cells

We report on enhanced performance of N719-sensitized TiO 2 solar cells (DSCs) incorporating size and photoelectron diffusioncontrolled TiO 2 as sensitizer-matched light-scatter layers on conventional nanocrystalline TiO 2 electrodes. The double-layered N719/TiO 2 composite electrode with a high dye-loading capacity exhibits the diffused reflectance of more than 50% in the range of λ= 650–800 nm, even when the films are coupled with the titania nanocrystalline underlayer in the device. As a result, the increased near-infrared light-harvesting produces a high light-to-electricity conversion efficiency of over 9% mainly due to the significant increase of Jsc. Such an optical effect of the NIR-light scattering TiO2 electrodes will be beneficial when the sensitizers with low molar extinction coefficients, such as N719, are introduced in the device.


Introduction
Dye-sensitized solar cells (DSCs) have been explored from both academic and industry world as a promising alternative to the conventional silicon-based photovoltaic devices [1][2][3], and efficiencies of over 12% have been achieved [4].In this cell, a wide band-gap nanostructured semiconductor electrode (typically, TiO 2 ) is employed to support a large surface area, on which dye molecules are assembled to improve the light harvesting of the dyed electrode.A great deal of effort has been focused on the development of effective panchromatic sensitizers or cosensitizers for efficient solar light harvesting in DSC devices.On the other hand, the lightharvesting efficiency of DSC can be promoted by introducing a sophisticated structure in the photoelectrode, in which the photon path length is increased by the confinement of incident solar energy light.Usami made the theoretical finding that lightscattering in DSCs will be optimized if the diameter of TiO 2 nanocrystals is twice the size of the wavelength of incident light [5].The effect has been verified in the DSCs, in which a TiO 2 layer built from submicrometer-sized TiO 2 particles was employed in combination with a transparent nanostructured TiO 2 layer [6][7][8].In addition, TiO 2 (or ZnO) photoelectrodes with submicrometer-sized beads, hollow spheres, and tubes have been widely investigated [9][10][11][12][13][14][15].
As an alternative to such a large particle light-scattering material, TiO 2 layers with periodic pore structures exhibiting photonic band gaps have been applied for DSCs to enhance light harvesting efficiency [16][17][18][19][20][21].Such light scattering TiO 2 layers, however, required an elaborate multistep synthesis.On the other hand, Hore et al. reported scattering of light on disordered spherical voids in TiO 2 electrodes for enhanced DSC performance.The DSC with 400 nm void layers gave a higher efficiency (6.7%) as compared to the cavity-free layer (5.4%) [22].While the overall efficiency was improved, a slight increase of  sc was observed and the appreciable increase of fill factors resulted in the 18% increment of conversion efficiency.Despite a number of nanostructured TiO 2 photoelectrodes reported as mentioned previously, the cell performance originated from spherical void-centered TiO 2 has still been poorly understood.Most importantly, International Journal of Photoenergy little has been reported on an enhancement of near-infrared-(NIR-) light response in IPCE [22,23].As organic dye molecules with large extinction coefficients even in the red region were used, the results represent a compromise between slight light scattering and enhanced mobility of viscous electrolytes in the large hollow cavities [23].In this paper we demonstrate optimized effects of scattering on submicron TiO 2 particles with spherical cavities on quantum efficiencies of light-to-electricity conversion of N719-sensitized solar cells at a NIR region, where the N719 dye has a low extinction coefficient.Accordingly, we prepared a TiO 2 layer containing 600 nm large spherical cavities and combined successively with the conventional nanocrystalline TiO 2 layer.We then applied the hollow-cavity-based TiO 2 /nanocrystalline TiO 2 double composite electrodes to N719-sensitized DSC.The DSC exhibited a maximum conversion efficiency of 9.25% versus 6.38% for the reference cell with single conventional TiO 2 electrodes.We demonstrated that the hollow-cavitybased TiO 2 contributes not only to the enhanced absorption of near-infrared photons but also to effective dye sensitization as evidenced by the remarkable enhancement of  sc in the performance.It should be noted that the effective electron diffusion coefficient of the dye-coated light-scattering TiO 2 layer also plays a crucial role in the improved DSC performance.

Preparation of DSCs.
A 60 nm dense TiO 2 layer was deposited on a patterned ITO-coated glass substrates (10 Ω/◻, Nippon Sheet Glass) by spin coating technique with tetraisopropyl orthotitanate in ethanol, as reported previously [24].The TiO 2 films, served as a blocking layer in DSCs, were annealed in air at 500 ∘ C for 30 minutes.Following this, a TiO 2 paste with 20 nm particle size was deposited by doctor blade technique on the blocking layer, where the TiO 2 paste was prepared according to the previous report [25].A transparent TiO 2 underlayer was achieved after the calcination at 500 ∘ C for 30 minutes.Functional polystyrene spheres (PS) (560 nm in size) were synthesized according to our previous method [26].The resulting latex spheres were used directly without purification.The polystyrene spheres (PS) were fabricated via batch emulsion polymerization, where styrene was mixed with methyl methacrylate (MMA) and acrylic acid (AA).For the preparation of the spherical cavity-based light-scattering TiO 2 layer, as prepared spheres were dispersed in the TiO 2 colloid with the particle size of about 20 nm as a coating paste for the preparation of the overlayer.The PS sphere, the previously mentioned TiO 2 colloid, and PEG were mixed in water with ratio of 1/0.5/0.3 (TiO 2 /PS/PEG, by weight) and homogenized with a stirring bar and ultrasonication.The resulting paste was deposited as a light-scattering layer on the underlayer, followed by the same calcination procedure mentioned previously.Finally, the double-layered TiO 2 film was impregnated with 0.2 M TiCl 4 aqueous solution at room temperature overnight and then calcinated again at 500 ∘ C for 30 minutes.After cooling to about 120 ∘ C, the films were emerged overnight in 1 : 1 v/v acetonitrile and tertbutyl alcohol solvent mixture containing N719 dye at a concentration of 0.3 mM.A Pt-sputtered FTO glass was used as a counter electrode.The electrolyte was composed of 0.6 M 1-propyl-3-methylimidazolium iodide, 0.1 M LiI, 0.04 M I 2 , and 0.5 M tert-butylpyridine in the mixture of acetonitrile and valeronitrile (1 : 1, by volume).

Characterization and Photoelectric Measurements.
The SEM images of titania films were obtained with a fieldemission SEM (JEOL JSM-4800, Japan).The optical reflection measurement was performed using an UV-vis spectrophotometer (U-3010, HITACHI) equipped with an integrating sphere.
The incident photon-to-current conversion efficiency (IPCE) for solar cells was measured by using a commercial setup (PV-25 DYE, JASCO) equipped with a monochromator.A 300 W Xenon lamp was employed as a light source for the generation of a monochromatic beam.Calibrations were performed with a standard silicon photodiode.IPCE is defined by IPCE() = hc  sc /e, where h is Planck's constant, c is the speed of light in a vacuum,  is the electronic charge,  is the wavelength in meters (m),  sc is the short-circuit photocurrent density (A m −2 ), and  is the incident radiation flux (W m −2 ).The photocurrent-voltage (I-V) characteristics were recorded using a computer-controlled Keithley 2400 source meter under air mass (AM) 1.5 simulated illumination (100 mw cm −2 , Oriel, 67005).The active area of the samples was about 0.2 cm 2 , controlled by an aperture mask to prevent extra light from coming through the lateral space.I-V data were well within experimental error.

Results and Discussion
As shown in Figure 1(a), the latex PS sphere holds a hard PS core with an elastomeric PMMA/PAA shell and presents a monodispersity with the size of about 560 nm (Figure 1(b)).The TiO 2 overlayer was prepared through a simple mixing of submicrometer-size polystyrene (PS) spheres and nano sized TiO 2 particles, followed by subsequent coating and calcinations on a transparent TiO 2 underlayer.The PS spheres vanish during the calcination process.
The SEM image in Figure 2(a) shows a top surface image of the TiO 2 overlayer.It is very clear that homogeneous spherical voids spread on the whole surface with the size of around 600 nm.In order to investigate the light-scattering ability of the overlayer, UV-vis reflectance spectra were recorded for a 6 m thick spherical-cavity-based TiO 2 film with and without sensitization by using a standard N719 dye, and an 8 m-thick nanostructured TiO 2 film was employed for comparison, as shown in Figure 3.For the film without the dye loading, the reflectance of the overlayer is more than 70% in the whole UV-vis region, indicating that more than 90% of the light reflecting efficiency can be achieved by the overlayer when considering the absorption and reflection by the FTO glass substrate.By contrast the reflectance of the transparent nanostructured TiO 2 underlayer is less than 30%.After the dye adsorption, the reflectance was significantly decreased in the spectral range by around 535 nm for both the layers, where N719 dye has the maximum absorption, indicating that the reduction is mainly ascribed to the light absorption by the dye.Most importantly, the diffused reflectance of the TiO 2 /dye composite film is found to be more than 50% in the range of  = 650-800 nm after the dye loading for the overlayer.
We tested the double composite films as DSC photoelectrodes (denoted as sample 1) consisting of a 8 m thick nanostructured TiO 2 underlayer and a 6 m thick overlayer, in which a single-layer TiO 2 electrode only with the 8 m   4 and Table 1.The conversion efficiency () was improved from 6.38% for sample 2 to 9.25% for sample 1, corresponding to the 45% increment.The improved efficiency was mainly caused by the increase of the short-circuit current density ( sc ) from 13.04 mA cm −2 to 18.40 mA cm −2 , while the  oc s and fill factors are the same or comparable for the two samples.

International Journal of Photoenergy
It should be noted that, in the overlayer, the submicro-meter sized voids are surrounded by the nanostructured TiO 2 film with the TiO 2 particles of 20 nm.The dye molecules can be absorbed on the overlayer to contribute to the increment of  sc in the double layer.In order to evaluate the effect of dye absorption in the overlayer on  sc , the one-layer nanostructured TiO 2 electrode with a thickness of 10 m was prepared (sample 3), by considering the amount of the dye loading for fair comparison (the porosity of TiO 2 overlayer : ∼80%).The amounts of the absorbed dyes were determined for the three different electrodes by measuring the eluted dye concentration from the TiO 2 electrodes as shown in Table 1.
The dye adsorption capacity was comparable for samples 1 and 3, suggesting that the scattering layer maintains a high dye-adsorption capacity.As shown in Table 1, sample 3 with the 10 m thick TiO 2 electrode gave 14.86 mA cm −2 of  sc , which was much lower than that of sample 1.A further comparison between sample 2 and sample 3 showed only a slight increase (1.82 mA cm −2 ) in  sc .On the other hand, the reflectance for sample 1 was significantly higher across the visible part of the spectrum than that of a single-layer TiO 2 cell as mentioned previously.Therefore, these results indicated that the marked increment in  sc for sample 1 as compared to 3 is ascribed to the stronger light scattering, particularly at a red to near-infrared light region.It is thus postulated that both the light-scattering property and high dye-loading capacity from the integrated nanocrystalline composite electrodes resulted in the significant increase of IPCE as shown in Figure 5.In addition, it should be noted that the unaffected and effective electron diffusion coefficient in the double-layered TiO 2 is essential for the improved IPCE in sample 1.

Conclusions
We applied a facile colloidal template method for the preparation of spherical-cavity-based TiO 2 /nanocrystalline TiO 2 double composite layers as a strong near-infrared lightscattering layer for N719-sensitized solar cells, which is particularly useful for the significant gain of photocurrents.The DSC fabricated by using double-layered electrodes composed of the scattering layer and the nanocrystalline TiO 2 layer produced a higher conversion efficiency of 9.25% versus 6.38% for the reference cell with a single TiO 2 electrode with the comparable thickness.The results obtained from our facile photon capture approach could open up new perspectives of exploiting low-cost and high-efficiency DSCs with a wide range of newly developed dye molecules and electrolytes.

Figure 2 (
b) shows a top surface image of the TiO 2 transparent layer, which presents a porous structure comprised of 20 nm sized TiO 2 particles.The cross-sectional image of the corresponding double-layer electrode is shown in Figure2(c), where the 8 m thick transparent nanoporous TiO 2 underlayer and the 6 m thick TiO 2 overlayer are formed without cracks.

Figure 1 :Figure 2 :
Figure 1: (a) A schematic structure of the functionalized polystyrene spheres.(b) A SEM image of the PS spheres with a diameter of 560 nm.

− 2 )Figure 4 :Figure 5 :
Figure 4: Photocurrent and voltage curves for samples 1 and 2. Both measurements were performed under AM 1.5 G with the active area of 0.2 cm 2 .

Table 1 :
Photovoltaic characteristics of samples 1, 2, and 3 measured under AM1.5 illumination (100 mW cm −2 ).Each data represents the average of three cells with active area of 0.2 cm 2 .