Enhanced light to electric power conversion efficiency of photovoltaic cells with a low absorbance was achieved using waveguide integration. We present a proof of concept using a very thin dye-sensitized solar cell which absorbed only a small fraction of the light at normal incidence. The glass substrate in conjunction with the solar cells reflecting back contact formed a planar waveguide, which lead to more than four times higher conversion efficiency compared to conventional illumination at normal incidence. This illumination concept leads to a new type of multi-junction PV systems based on enforced spectral splitting along the waveguide.
© 2008 Optical Society of America
Photovoltaic (PV) cells provide a way for the direct conversion of sunlight into electric power. The absorber thickness in PV cells should be sufficiently large to convert a major part of incident photons into electron-hole pairs and, at the same time, it should be thin enough to minimize recombination and enable efficient carrier collection at the electron and hole conducting media adjacent to the absorber. Thus the absorber thickness is a compromise between the optical absorption coefficient α and the minority carrier diffusion length L.
Nanostructures have been employed to create a manifold enhanced microscopic junction area between the absorber and the electron and hole conductor, respectively, to generate carriers close to a charge separating interface and to prevent recombination within the absorber. [1, 2] The low absorption of an individual absorber layer is compensated by a light path passing through many layers. This concept is particularly well realized in dye-sensitized solar cells (DSSC), where a monolayer of dye on a mesoporous nano-crystalline TiO2 thin film (electron conductor) generates charge carriers upon illumination. [3–5] The film is deposited onto a conducting transparent substrate and the electrical circuit is closed by a redox electrolyte immersed into the pores which transports the positive charge to a Pt back electrode. For many materials an enormous improvement of the light harvesting and charge separation can be achieved using nano-composites. [6–8] However the longer transport path associated with these structures involves significant recombination from the electron conductor into the hole conductor with no benefit to the conversion efficiency.
Optical methods have attracted considerable attention to enhance the photon density in the absorber. Surface texturing has been used to achieve light trapping in Si solar cells  while in DSSCs light scattering on micro-particles or nano-particle clusters has been used to enhance the optical pathway through the active area of the cell.  In polymer solar cells enhanced photon densities due to interference have been observed [11, 12] and thin dielectric coatings have been proposed to enhance the optical field intensity distribution around the charge separating interface. [13, 14] Strong absorption enhancement has been observed in thin silicon on insulator (SOI) waveguides, covered with a layer of metal islands with diameters of tens of nanometers. [15, 16] The absorption enhancement was attributed to the coupling between metal island resonances with waveguide modes supported by the SOI structure. [17, 18] Surface plasmon enhanced absorption was reported for metal island covered thin film and wafer based silicon solar cells  and for a hydrogenated amorphous silicon waveguide slab sandwiched between a periodic array of Au nanowires and a Au substrate.  Furthermore for thin film solar cells novel cell designs have been proposed to enable effective light trapping. 
Here we present a new method to enhance the photovoltaic performance of a thin film solar cell with a thickness far below the optical absorption length α-1 when operated in conjunction with an optical wave guide. Incomplete light harvesting is typically associated with inefficient charge collection beyond a certain cell thickness and is often observed in solid state dye sensitized, polymer and organic solar cells. To proof the concept we used a DSSC based on a ~40 nm thin mesoporous non-scattering TiO2 film which showed low conversion efficiency, η, at normal incidence due to incomplete light absorption. The dye-sensitized TiO2 film was deposited onto an Indium Tin Oxide (ITO) covered glass substrate, which served as a waveguide. Light propagation occurred along the substrate parallel to the surface due to total internal reflection at the glass/air interface and reflection at the Pt back electrode of the DSSC. Operating the DSSC in the waveguide solar cell (WGSC) configuration we succeeded to enhanced the optical path length through the absorber while keeping the electronic path for charge carrier collection at the front (ITO) and back electrode (Pt) unchanged. A clear correlation between the absorption coefficient of the dye and the photocurrent decay as a function of propagation length was observed. While improving a specific class of single junction solar cells, the concept reported below opens the way for a new type of multi-junction systems based on enforced spectral splitting along the waveguide.
The WGSC was deposited onto an Indium Tin Oxide (ITO) covered glass slide (25×75 mm2, thickness of 0.7 mm) with a sheet resistance of 8–12 Ω/square (Delta Technologies), which served as a waveguide. 12 conducting patches (terminals) with a size of 25×5 mm2, separated by 0.5–1 mm wide insulating gaps were created by local removal of ITO using concentrated HCl for etching. The TiO2 paste consisting of nanocrystals with an average diameter of ~12 nm was prepared in an autoclave at 150 °C for 5 hours and afterwards diluted with ethanol. A ~40 nm thin and nearly non-scattering TiO2 film was deposited onto the ITO terminals by the doctor blade method, using adhesive tape as a spacer. After drying in air the films were sintered at 400 °C for 30 min before they cooled down to 80 °C and were directly immersed into an ethanol solution containing 5 mM cis-dithiocyanato bis(4,4′-dicarboxylic acid-2,2′-bipyridine) ruthenium(II) dye, commonly termed N3, which was purchased from Dyesol. The electrode was kept in dye-solution for at least 24 hours before it was rinsed with ethanol and dried in air. Pt sputtered onto conducting Fluorine doped Tin Oxide (FTO) covered glass was used as a counter electrode and back reflector. The reflecting Pt electrode was pressed mechanically onto the TiO2 electrode, separated by thin Teflon tape, which was placed at the edges. An I-/I- 3 redox electrolyte (0.6 M 1-butyl-3-methylimidazolium iodide, 0.5 M 1-methyl benzimidazole, 0.1 M I2), dissolved in propylene carbonate was filled in between the electrodes. The low vapor pressure of propylene carbonate allowed measurements without further sealing of the cell.
All solar cell measurements were performed with a solar simulator (Newport, class A) at AM1.5G illumination (100 mW/cm2). For waveguide measurements the light was coupled into the glass substrate using a homemade prism from Perspex with an acute angle of 5°, which was attached with index matching oil to the glass substrate (Fig. 1(a)). An aperture with an area of 2×10 mm2 in front of the prism was used to define the area of incident light while I-V curves at each terminal were recorded with a potentiostat (Autolab). Reference measurements of each terminal were performed at normal incidence, where the entire cell area was illuminated (Fig. 1(b)). To exclude experimental artifacts such as evaporation of electrolyte during the measurements, several terminals were measured at the beginning and end of the of the measurement series, which did not show significant deviations with time. For measurements with monochromatic illumination band pass filters were placed in front of the aperture.
Current-voltage curves of the WGSC for different waveguide lengths are presented in Fig. 2(a), where the active length of the WGSC was defined by the number of terminals connected in parallel. Terminals T1, T2 and T3, did not show significant photocurrents (and were disconnected for further measurements), indicating that light coupling into the waveguide occurred at terminal T4 (schematically shown in Fig. 1(a)). A continuous increase of the short circuit current density with increasing WGSC length is observed, while the open circuit voltage (V oc) decreases. This can be explained when the I-V characteristics of the device is approximated by a diode equation. In this case the V oc is proportional to ln(J sc/J 0 + 1), where J sc is the short circuit current density and J 0 is the diodes reverse bias saturation current density. Lower light intensities further away from the place where light is coupled into the waveguide generate lower photocurrents while J 0 remains constant, which subsequently leads to smaller open circuit voltages at individual terminals. When connected in parallel, terminals with low V oc decrease the overall V oc. Nevertheless the photocurrent and the conversion efficiency η steadily increase with increasing WGSC length, as shown in Fig. 2(b). The maximum conversion efficiency of 0.52% is reached when terminals four to twelve (T4–T12) are connected in parallel, which is 4.3 times larger compared to the conversion efficiency of 0.12% at normal incidence, shown by the dashed line. It should be noticed that terminal T4 already shows a larger conversion efficiency in the WGSC configuration compared to normal incidence, which we attribute to the enhanced optical path of photons through the absorber due to the shallow angle under which light is coupled in (Fig. 1(a)). We further note that the aperture size in the WGSC was used to calculate current- and power-densities while the illuminated terminal area was used for normalization at normal incidence.
The generated photocurrent as a function of the terminal number is shown in Fig. 3(a) on a logarithmic scale. The J sc at the individual terminals follows a single exponential decay with increasing propagation length at white light illumination, indicating that light absorption and charge carrier generation follows Lambert-Beer law:
The photon flux Φ at a distance x from the point where light is coupled into the waveguide is defined by the incident photon flux Φ0 and an effective absorption coefficient α which takes the waveguide structure into consideration. Figure 3(a) also shows the J sc as a function of terminal number at monochromatic illumination at 400 nm, 550 nm and 600 nm. The transmission bands of the filters are shown in Fig. 3(b) together with the absorption spectrum of the dye-sensitized TiO2 electrode measured at normal incidence. A decreasing slope of the current decay with increasing terminal number is observed from 400nm, with the steepest slope, to 600 nm with a shallow slope (Fig. 3(a)), which corresponds well to the higher absorption coefficient at 400 nm and decreasing α at 550 nm and 600 nm. We note that the larger α at 400 nm is also partly due to light scattering caused by a larger scattering cross section of nanoparticles for photons with a short wavelength. The results in Fig. 3 show that light propagation is strongly influenced by the absorption coefficient of the WGSC, which is relevant for the design of more sophisticated solar cell assemblies with waveguide integration.
While we succeeded to improve the conversion efficiency of a DSSC with a very low absorbance by a factor of 4.3, we also achieved conversion enhancement with a medium efficient DSSC, though the enhancement factor was lower. For these measurements the conducting substrate was not divided into terminals, the electrode size was approximately 20×70 mm2 and the TiO2 film thickness was 6 µm. Similar to the measurements of the WGSC with a very thin film, a lower V oc and a larger J sc were observed in the waveguide configuration. However the conversion efficiency increased only by a factor of 1.25, from 3.1% to 3.9%.
4. Discussion and outlook
In high performance DSSCs with conversion efficiencies above 10%, alternative methods are usually used to enhance the optical path way of photons through the absorber material. Large particles of TiO2 are dispersed throughout the mesoporous film for light scattering within the absorber layer. Consequently, the optical light path is enhanced similar to the WGSC, extending the effective film thickness while the electronic path length for carrier collection remains unchanged.
Even though an increase beyond the current efficiency limit of DSSCs was not achieved using waveguide enhancement it has a great potential for thin film polymer cells or solid state DSSCs, where light scattering does not sufficiently increase the photon path length.
A possible design for a large area panel, based on the waveguide concept, is shown in Fig. 4(a), where the panel thickness (waveguide length) depends on the absorber thickness and its absorption coefficient. Furthermore the waveguide can be used to design more sophisticated third generation photovoltaic devices where the solar spectrum is divided into spectral bands which are converted into electrical power by solar cells, optimized for the specific band. Figure 4(b) shows schematically a possible design, based on optical concentration in conjunction with two WGSCs mounted behind each other on the same waveguide. A lens focuses the light onto the waveguide entrance and energy rich photons of the solar spectrum (shown as dashed blue rays) are absorbed in the first WGSC which has a large bandgap and thus a large α for photons above bandgap, while α is very small for photons with energies below the bandgap. These photons (shown as dashed-dotted red rays), propagate along the waveguide to meet the lower bandgap WGSC and are absorbed there. In such a configuration both cells can be operated independently at maximum power and current matching of both cells is not required. While Fig. 4(b) shows an example where the solar spectrum is split into two bands, it is in general possible to increase the number of WGSCs such that a cascade of cells with decreasing bandgaps can be placed behind each other. We note that optical concentration is not a prerequisite of the presented design.
In summary we have shown that macroscopic optical components such as waveguides can be used to increase the optical path length through a solar cell absorber without changing the electronic path of generated charge carriers. As a result the absorption probability is increased while the travel distance of the generated electrons and holes to the front and back electrode remains below the minority carrier diffusion length. Using the concept of waveguide enhanced PV we were able to increase the light to electric power conversion efficiency of a DSSC consisting of a very thin mesoporous TiO2 film by a factor of 4.3.
The authors thank David Cahen (Weizmann Institute of Science, Rehovot, Israel) for fruitful discussions.
References and links
1. I. Kaiser, K. Ernst, C. H. Fischer, R. Könenkamp, C. Rost, I. Sieber, and M. C. Lux-Steiner, “The eta-solar cell with CuInS2: A photovoltaic cell concept using an extremely thin absorber (eta),” Sol. Energy Mater. Sol. Cells 67, 89–96 (2001). [CrossRef]
2. D. Kieven, T. Dittrich, A. Belaidi, J. Tornow, K. Schwarzburg, N. Allsop, and M. Lux-Steiner, “Effect of internal surface area on the performance of ZnO/In2S3/CuSCN solar cells with extremely thin absorber,” Appl. Phys. Lett. 92, 153107–153103 (2008). [CrossRef]
3. M. Grätzel, “Photovoltaic performance and long-term stability of dye-sensitized meosocopic solar cells,” C. R. Chimie 9, 578–583 (2006). [CrossRef]
4. F.-T. Kong, S.-D. Dai, and K.-J. Wang, “Review of recent progress in dye-sensitized solar cells,” Adv. OptoElectron. 2007, doi:10.1155/2007/75384 (2007). [CrossRef]
5. F. O. Lenzmann and J. M. Kroon, “Recent advances in dye-sensitized solar cells,” Adv. OptoElectron. 2007, doi:10.1155/2007/65073 (2007). [CrossRef]
6. G. Yu, J. Gao, J. C. Hummelen, F. Wudl, and A. J. Heeger, “Polymer photovoltaic cells: Enhanced efficiencies via a network of internal donor-acceptor heterojunctions,” Science 270, 1789 (1995). [CrossRef]
7. H. Hoppe and N. S. Sariciftci, “Morphology of polymer/fullerene bulk heterojunction solar cells,” J. Mater. Chem. 16, 45 (2006). [CrossRef]
8. C. Lévy-Clément, R. Tena-Zaera, M. A. Ryan, A. Katty, and G. Hodes, “CdSe-sensitized p-CuSCN/nanowire n-ZnO heterojunctions,” Adv. Mater. 17, 1512–1515 (2005). [CrossRef]
9. P. Campbell and M. A. Green, “Light trapping properties of pyramidally textured surfaces,” J. Appl. Phys. 62, 243–249 (1987). [CrossRef]
10. Z. Zhang, S. Ito, B. O’Regan, D. Kuang, S. M. Zakeeruddin, P. Liska, R. Charvet, P. Comte, M. K. Nazeeruddin, P. Péchy, R. Humphry-Baker, T. Koyanagi, T. Mizuno, and M. Grätzel, “The electronic role of the TiO2 light-scattering layer in dye-sensitized solar cells,” Z. Phys. Chem. 221, 319 (2007). [CrossRef]
11. T. Stubinger and W. Brutting, “Exciton diffusion and optical interference in organic donor-acceptor photovoltaic cells,”J. Appl. Phys. 90, 3632–3641 (2001). [CrossRef]
12. H. Hoppe, S. Shokhovets, and G. Gobsch, “Inverse relation between photocurrent and absorption layer thickness in polymer solar cells,” Phys. Status Solidi (RRL) 1, R40–R42 (2007). [CrossRef]
13. B. O’Connor, K. H. An, K. P. Pipe, Y. Zhao, and M. Shtein, “Enhanced optical field intensity distribution in organic photovoltaic devices using external coatings,” Appl. Phys. Lett. 89, 233502–233503 (2006). [CrossRef]
15. H. R. Stuart and D. G. Hall, “Absorption enhancement in silicon-on-insulator waveguides using metal island films,” Appl. Phys. Lett. 69, 2327–2329 (1996). [CrossRef]
16. H. R. Stuart and D. G. Hall, “Island size effects in nanoparticle-enhanced photodetectors,” Appl. Phys. Lett. 73, 3815–3817 (1998). [CrossRef]
17. B. J. Soller, H. R. Stuart, and D. G. Hall, “Energy transfer at optical frequencies to silicon-on-insulator structures,” Opt. Lett. 26, 1421–1423 (2001). [CrossRef]
18. H. R. Stuart and D. G. Hall, “Enhanced dipole-dipole interaction between elementary radiators near a surface,” Phys. Rev. Lett. 80, 5663 (1998). [CrossRef]
19. S. Pillai, K. R. Catchpole, T. Trupke, and M. A. Green, “Surface plasmon enhanced silicon solar cells,” J. Appl. Phys. 101, 093105–093108 (2007). [CrossRef]
21. S.-B. Rim, S. Zhao, S. R. Scully, M. D. McGehee, and P. Peumans, “An effective light trapping configuration for thin-film solar cells,” Appl. Phys. Lett. 91, 243501–243503 (2007). [CrossRef]