Design of Translucent,
High-Efficiency Photoelectrochemical Solar Cells
Based on Multiple Dye Sensitization of
TiO2 Films
| Abstract Based on the fundamental mechanisms of photosynthesis, a high-efficiency, translucent solar cell has been created. This solar cell uses several photosensitive dyes to initiate a series of regenerative photoelectrochemical reactions. This solar cell was constructed by applying a nanocrystalline TiO2 film onto a conducting glass support, which was coupled with an iodide electrolyte and a platinum counterelectrode. The active surface area of the TiO2 film was enhanced by two applications of ultrafine TiO2 precipitated from aqueous TiCl4, resulting in a 643-fold increase in power production. Experiments were performed to determine the feasibility of a translucent solar cell that would be more practical for applications such as tinted windows and skylights. Thinner, translucent layers of TiO2 (~7 µm) and platinum (30-50 nm) were deposited onto the anode and cathode, respectively. There was no measurable change in cell performance as compared to the opaque cell. On this basis, the feasibility of a translucent photoelectrochemical solar cell is supported. To determine if the absorption spectrum of the solar cell could be increased, experiments were conducted involving dye sensitization of the TiO2. A copper phthalocyanine (CuPh) dye and a zinc porphyrin dye were adsorbed onto separate enhanced TiO2 semiconductor anodes. These dye molecules, acting as photosensitive molecular antennae, absorb light in the red and blue regions of the spectrum, respectively, whereas the TiO2 film absorbs only ultraviolet light. The absorbed CuPh dye increased the overall power output by 252% and the power output at the peak sensitized wavelength (750 nm) by 1500%. The Zn porphyrin dye increased the power production of the cell at the sensitized peak (475 nm) by 265%. Additionally, the resulting incident photon to current conversion efficiency (IPCE) of the dye-sensitized cells was 80% and 77%, respectively. To test the feasibility of multiple dye sensitization, both the CuPh and Zn porphyrin dyes were adsorbed onto the same TiO2 electrode. These dyes sensitized the cell to both the red and blue regions of the visible spectrum, increasing the overall power production by 460% compared to the non-sensitized cell. The IPCE approached 50% at each of the two sensitized spectral peaks. A device based on a simple molecular light absorber has attained a conversion efficiency commensurate with that of conventional silicon-based photovoltaic cells. This research decisively demonstrates that a new, highly-efficient photoelectrochemical solar cell can be developed using the photosynthetic process as a model. Schematic - of solar cell Introduction |
| Experimental
Section Preparation of Nanocrystalline TiO2 Electrodes. Nanocrystalline TiO2 films were prepared by spreading a viscous suspension of colloidal TiO2 particles on a conducting glass support (fluorine-doped SnO2 overlayer). To separate the aggregated particles of the commercial TiO2, the powder (3 g) was ground in a porcelain mortar with water (1 mL) containing acetylacetone (0.1 mL) to prevent reaggregation. After the powder was dispersed, it was diluted by the slow addition of water (4 mL) under continued grinding. A detergent (.05 mL Triton X-100) was then added to facilitate the spreading of the colloid on the substrate.2 The conductive side of the glass was covered on two parallel edges with 10µm-thick adhesive tape, creating a 1in2 active surface area. The tape controlled the thickness of the TiO2 film and provided noncoated areas for electrical contact. The colloid was applied to the conductive side of the glass and distributed with a glass rod rolled over the tape covered edges. After air drying, the electrode was heated for 30 min at 500°C in air. Surface Enhancement. A .2 M TiCl4 solution was prepared at 0°C to prevent immediate precipitation of TiO2 due to the highly exothermic hydrolysis reaction.1 The electrode was soaked overnight in this solution to precipitate ultrafine TiO2 onto the electrode. It was then rinsed with distilled water and heated again for 30 min at 500°C in air. The TiCl4 deposition was repeated to increase the active surface area of the electrode. The surface morphology of the TiO2 film was probed by a scanning electron microscope (SEM) at 10,000x magnification. The performance of the cells constructed from the enhanced electrodes was compared to control cells constructed from the non-enhanced electrodes. Dye Sensitization. Concentration Optimization. Dye solutions of 50 nmol, 5 µmol, and 50 µmol were prepared to determine the optimum concentration for adsorption onto the TiO2 film. Immediately following heating, the TiO2 electrode was immersed for 5 h at room temp. in a solution of copper(II) 1,4,8,11,15,18,22,25-octabutoxy-29H,31H-phthalocyanine in THF and a coadsorbate of deoxycholic acid (80 mM).3 The electrode was then heated for 15 min at 100°C to deadsorb moisture. Multiple Dye Sensitization. Adsorption of the Cu phthalocyanine dye onto the enhanced TiO2 film resulted in a sensitized peak in the far red region of the visible spectrum (725-775 nm). To increase the range of sensitization, a zinc porphyrin dye (8,13-divinyl-3,7,12,17-tetramethyl-21H,23H-porphine-2,18-dipropionic acid) that absorbs light in the blue region of the spectrum (450-500 nm), was chosen. A liquid cell spectrophotometer was used to evaluate the absorption spectra of the dyes. The Zn porphyrin was also dissolved in THF and deoxycholic acid (80 mM). The Cu and Zn dyes were adsorbed onto separate electrodes in 5 µmol concentrations to compare their relative power production. Additional electrodes were prepared with both dyes adsorbed in 1µmol concentrations onto a single enhanced TiO2 film. It was theorized that sensitization of the TiO2 with several dyes that have complementary absorption spectra would broaden the spectral sensitivity of the cell and therefore increase power production. Design of a Translucent Solar Cell. A thinner layer of TiO2 (~7 µm) was coated onto the conducting glass support. The aqueous solution of TiO2 was diluted by a factor of 10 to create a translucent film. A thinner layer of platinum (30-50 nm) was electroplated onto a conducting glass counterelectrode. The performance of the cells created from these counterelectrodes was compared to those with a thicker layer of platinum (100 nm). Transparent counterelectrodes, with no Pt coating, were also constructed. The performance of these cells was measured to determine if an optically transparent solar cell could be designed with an overall power efficiency comparable to that of a translucent cell. Although not essential, complete transparency would broaden the range of practical applications. Preparation and Testing. The counterelectrode was coupled with the dye-senstized, translucent TiO2 film, supported by the conducting glass sheet. A teflon spacer was placed between the electrodes to prevent a short-circuit. Two wired metal clamps were attached to opposite ends of the two electrodes. A thin layer of electrolyte (0.5 M KI, 40 mM I2 in 80% ethylene carbonate, 20% propylene carbonate) was attracted into the inter-electrode space by capillary forces.4 All 97 of the electrodes created in each stage described above were assembled in this manner to evaluate their performance. The dye-sensitized TiO2 film was illuminated by a Schoeffel 1000-W Xe lamp and monochromator through its conducting glass support. A test protocol consisting of both white light and monochromatic illumination from 450 to 850 nm, with loads from 1 to 10,000 W, in addition to both open and closed circuits, was performed on each cell variation. The photocurrent was recorded by a digital multimeter as a voltage drop across a 1000-W resistor. |
| Results and
Discussion Semiconductor Surface Enhancement. Inspection of the micrographs produced by the SEM showed that the surface of the TiO2 electrode was composed of ultrafine particles in close contact with each other. The particle size was measured to be 100-200 nm, indicating a highly porous surface for dye adsorption. The micrographs also showed that no significant aggregation occurred during the preparation or application of the TiO2. The TiO2 film prepared from colloidal suspension can be characterized as a three-dimensional, porous lattice of nanocrystallites. Compared to the control cell with a non-enhanced TiO2 film, the cell with a single treatment of TiCl4 exhibited a 7-fold increase in overall power production. The changes in surface characteristics produced by the TiCl4 treatment were too fine to be resolved by the SEM at 10,000x magnification. A second treatment of TiCl4 further increased the active surface area of the electrode. This cell exhibited a 643-fold increase in overall power production as compared to the control cell. (See Figure 2) As with the first TiCl4 treatment, there was no detectable change in surface characteristics at 10,000x magnification. Performance of the Translucent Design. In addition to creating a translucent solar cell, it was expected that a 30% thinner layer of TiO2 (~7µm) would improve charge transfer between the sensitizing dye and the TiO2, ultimately reducing internal recombination losses. No significant change in performance, however, resulted from the thinner layer of TiO2. Therefore, other mechanisms must also control electron injection into the TiO2 semiconductor. Resistance may exist primarily at the dye-TiO2 interface and not between the TiO2 particles. A thinner layer of platinum (30-50 nm) on the counterelectrode had no effect on cell performance. The complete absence of Pt, however, significantly reduced the overall power production of the cell, indicating that Pt has a catalytic or enabling effect on electron transfer to the iodide electrolyte. (See Figure 3) Further research is required to develop and evaluate alternatives to electroplated platinum. Thinner, more translucent Pt layers may be applied by vacuum electrodeposition. It may also be possible to incorporate the Pt into a porous glass substrate, thereby creating an optically transparent counterelectrode. Photosensitive Dye Selection. The Cu phthalocyanine dye was chosen due to its chemical structure. It was anticipated that the four additional carbon rings, as compared to the porphyrin molecule chlorophyll, would further distribute the electrons and lower the electronic energy levels of the molecule. The stabilized ring structure also suggested that the molecule would be very stable under heat and have low chemical reactivity, which is characteristic of phthalocyanine compounds.5 Spectral analysis confirmed that the absorbance curve was shifted to the red region of the spectrum (725-775 nm). The selection of the solvent from which the sensitizing dye is adsorbed is critical to photosensitization. The solvent must provide adequate solubility to dissolve the high molecular weight, low polarity dye molecule. Excessive affinity, however, will inhibit the adsorption of the dye onto the TiO2 electrode.3 THF proved to be an acceptable vehicle for dye adsorption onto the TiO2 film. Dye Concentration and Cell Performance Optimization. Cells produced with dye concentrations of 5 µmol exhibited 50% greater power production than those produced from the 50 nmol and 50 µmol concentrations. It is theorized that the 50 nmol concentration did not provide a complete monolayer of coverage and the 50 µmol concentration resulted in multiple dye layers which created an insulating effect. On a semiconductor surface, only the first adsorbed monolayer results in efficient electron injection into the semiconductor; thick dye layers result in electron-dye recombination.3 On a flat surface, however, the light absorption by a single monolayer is at most a few percent, resulting in a weak sensitization effect. In a porous film consisting of nanometer-sized TiO2 particles, the effective surface area can be enhanced 1000-fold, providing more adsorption sites for the photosensitive dye, thereby making light absorption efficient with only a monolayer of dye. The Cu phthalocyanine dye absorbs light in the visible spectrum, primarily the red region (725-775 nm), whereas TiO2 absorbs only ultraviolet light. As a result of the extended spectral sensitivity, the photosensitized cell exhibited a 252% increase in overall power production and a 1500% increase at the sensitized peak as compared to the enhanced, non-sensitized cell. (See Figure 4) The incident monochromatic photon to current conversion efficiency (IPCE), defined as the number of electrons generated by the light in the external circuit divided by the number of incident photons, is plotted as a function of excitation wavelength. (See Figure 5) This value was obtained from the photocurrents by means of the following equation: At the peak sensitized
wavelength (750 nm), the photocurrent efficiency was 80%.
This remarkable quantum efficiency offers great promise
for future solar energy applications. |
| Conclusions Improvement in cell performance resulted from experiments on semiconductor surface area, dye selection and concentration, and multiple dye sensitization. The colloidal semiconductor TiO2 was found to be an effective photoconductor, although its absorbance is limited to the ultraviolet region of the spectrum. Adsorbed moisture was found to be a strong inhibitor of cell performance, and therefore the deadsorption of moisture from the TiO2 surface is vital to cell function. To enhance cell performance, it was necessary to maximize surface area, creating more adsorption sites for the sensitizing dye. The photosensitization of wide-bandgap semiconductors by adsorbed dyes has become more practical for solar cell applications with the recent development of porous, nanocrystalline films, such as those prepared in this research.1 The TiCl4 treatment improved the performance of the control cell by increasing the active surface area of the TiO2 film, thereby increasing the photon conversion efficiency. A second treatment of TiCl4 further increased the active surface area, allowing the monolayer of dye to sensitize the TiO2 much more effectively. Using the Cu phthalocyanine dye as a photosensitizer, a 252% increase in broad spectrum power production was observed. Additionally, a 1500% increase in power production was observed at the peak sensitized wavelength of 750 nm. The photocurrent spectra of the sensitized cells directly correspond to the absorption spectra of the dyes, indicating the specificity of both of the sensitizing dyes. The IPCE values of 80% and 77% for the cells individually sensitized by the Cu phthalocyanine and Zn porphyrin dyes, respectively, demonstrate that the dyes are highly efficient photosensitizers in the visible spectrum. The sensitivity of the dyes in the visible spectrum complements the sensitivity of TiO2 in the ultraviolet region, allowing the cell to utilize more of the available light energy. To further broaden the spectral sensitivity of the cell, a single TiO2 film can be sensitized by both dyes, optimizing the cell performance. Without sacrificing performance, the cell can be designed to be translucent, increasing the range of practical applications. The 7-µm-thick TiO2 coating on the anode and a 30-50 nm thick platinum coating on the cathode created a translucent cell that performed at an efficiency equal to that of an opaque cell. The impressive results described above provide confirmation of the hypothesis that a regenerative photoelectrochemical solar cell is a viable potential means of energy production. This research decisively demonstrates that a new, highly-efficient photoelectrochemical solar cell can be developed using the photosynthetic process as a model. There are numerous areas worthy of further investigation. The concentrations of the Cu phthalocyanine and Zn porphyrin dyes may be adjusted to optimize broad spectrum cell performance. Different counterelectrodes, such as conducting polymers, carbon, and other metals, may be used as conductive surfaces. Dyes may be combined or specifically synthesized to absorb a broad spectrum of light. Acknowledgements References |
[Home] - [Intellectual] - [Abstract] - [WH Paper]
