December 1, 2013 / Vol. 38, No. 23 / OPTICS LETTERS
5071
Ultra-broadband performance enhancement of thin-film amorphous silicon solar cells with conformal zig–zag configuration Zhenhai Yang, Aixue Shang, Yaohui Zhan, Cheng Zhang, and Xiaofeng Li* Institute of Modern Optical Technologies & Collaborative Innovation Center of Suzhou Nano Science and Technology, Jiangsu Key Lab of Advanced Optical Manufacturing Technologies & MOE Key Lab of Modern Optical Technologies, Soochow University, Suzhou 215006, China *Corresponding author: [email protected] Received September 13, 2013; revised October 8, 2013; accepted October 28, 2013; posted October 29, 2013 (Doc. ID 197541); published November 25, 2013 An ultrathin amorphous silicon solar cell with conformal zig–zag nanoconfiguration is studied from both lighttrapping and light-conversion perspectives. The design improves the front antireflection property, optimizes the rear metallic reflector, and elongates the optical path inside the photoactive layer. Compared to conventional nanoconfigurations, this system shows significant absorption enhancement in the whole amorphous silicon band and exhibits extremely low sensitivity to light polarization. The nano-optimization indicates that the short-circuit current density (light-conversion efficiency) of the 200-nm-thick solar cell can be 16.88 mA∕cm2 (13.38%), showing an enhancement factor of 32.90% (33.53%) from the planar system. © 2013 Optical Society of America OCIS codes: (040.5350) Photovoltaic; (350.4238) Nanophotonics and photonic crystals. http://dx.doi.org/10.1364/OL.38.005071
Solar cells that convert electricity from sunlight have attracted huge interest on the road to clean and renewable energy. The cost of currently available crystalline silicon solar cells is still high because of their thick (>200 μm) active layer. Nevertheless, thin-film solar cells (TFSCs) are considered to be very promising as they can absorb over 90% of sunlight in the material band with much thinner active layers (typically 520 nm. This is because the rear nanostructured reflectors can significantly increase
Fig. 2. (a) Absorption spectra P abs of a-Si:H layer and (b) reflection spectra of the entire device, where Λ 400 nm and H Λ∕6.
the optical path of the long-wavelength light (close to the bandgap). A simple comparison between the rose and zig–zag systems shows the superiority of the latter in light trapping (see P abs when λ > 650 nm). This provides direct optical evidence of the finding in [15], showing the potential of zig–zag nanostructures in outperforming other typical designs. In fact, the better performance of a zig–zag design arises from a slight modification to the rose type by breaking the geometric symmetry (Fig. 1). However, the above zig–zag enhancement is just confined to the LWB (with short-wavelength band, SWB, almost unaffected), where the solar incidence penetrates through the active layer to strongly interact with the back reflector. Although a front “planar” antireflection coating can efficiently decrease the light-reflection loss, space still exists for further improvement in the SWB, in order to achieve a full-spectrum absorption enhancement. Figure 2(a) (blue solid curve) exhibits the absorption spectrum of the considered cell with a conformal zig–zag design. It is obvious that the newly designed solar cell obtains an ultrabroadband performance enhancement (in the whole material absorption band); especially, P abs ∼ 100% within the region of λ ∼ 500 nm. This is the result of the joint effects of better front antireflection, broadband light reflection in the rear of the cell, and a significantly elongated optical path inside the photoactive layer [13]. For further insight into the enhancement mechanism, the corresponding reflection spectra R are given in Fig. 2(b). The energy loss in the SWB comes directly from the reflection of the device surface, while that in the LWB is resulted from the back metallic reflector, material absorption, and cavity resonance, which yield the localized peaks. With a planar Ag reflector, as shown in Fig. 2(b), the device reflection at the LWB is very high since the light undergoes the shortest optical path under normal injection without being sufficiently absorbed. Such a high reflection can be greatly decreased with nonplanar nanodesigns (e.g., rose, and zig–zag) due to the excitation of modes with tilted propagation angles, many of which may satisfy the total internal reflection (TIR) condition at the top surface to greatly improve the lighttrapping performance. Nevertheless, rear rose and zig– zag designs do not modify the reflection when λ < 520 nm, for the reason mentioned above. Under a conformal zig– zag scheme, however, substantial reduction of reflection at the SWB can be seen, which can be ascribed to the better antireflection of the surface [according to Fig. 2(b), R can be very close to 0 when 480 < λ < 600 nm]. A lower reflection in the LWB is also observed to correspond to the better P abs in Fig. 2(a). This is because the titled interface between TCO and a-Si:H lets TIR occur more easily (i.e., under larger incident angles of the light reflected back from the rear zig–zag metallic reflector). The strengthened light confinement thus leads to a much longer optical path for efficient interaction between the light and the photoactive material. It is also worth noting that, although the conformal zig–zag structure is not symmetrical as is the rose, the modified system does not exhibit noticeable polarization dependence. In this study, both the transverse electric (TE) and transverse magnetic (TM) polarizations have been simulated. We find that the results for both TE
December 1, 2013 / Vol. 38, No. 23 / OPTICS LETTERS
and TM are almost the same (a slight difference occurs in the region with very long wavelength beyond the bandgap; not shown here). This reveals that the zig–zag structure is polarization insensitive with great potential in broadband absorption enhancement; therefore, it is expected to provide opportunity for an excellent lighttrapping design. Although a promising light-trapping is achieved, further improvement is still possible by conducting a thorough nano-optimization on the zig–zag structure, i.e., through controlling the period Λ and the height H. However, a conclusion is hard to draw based on the complex spectral behavior of the absorption curve; we thus have to find a simple and direct way to rate the designs. In fact, the spectral behavior can be integrated into a unique parameter J sc (short circuit current density in mA∕cm2 ) with the solar spectrum taken into account [23]. In many cases, J sc values obtained in this way are based on the condition of a perfect internal quantum process (i.e., internal quantum efficiency IQE 100%). A realistic prediction of solar cell performance must consider the detailed internal carrier (electron and hole) process with generation, diffusion, drift, and recombination mechanisms taken into consideration [24]. Based on an electrical modeling technique, we simulate the internal quantum process and obtain the IQE spectrum, which can be used to convert the light absorption to a realistic EQE and, thus, J sc . To find a globally maximal J sc , we sweep H from 50 to 140 nm for typical Λ values (from 200 to 800 nm with steps of 100 nm) to find the best H design for each Λ. A number of three-dimensional electromagnetic simulations yield the results shown in Fig. 3(a), where the maximal J sc for each Λ has been inserted into the bar. It is shown that a relatively large period is needed for the conformally configured TFSCs in order to obtain a high J sc , which can be as high as 16.88 mA∕cm2 (when Λ 600 nm and H 130 nm), with an enhancement of ∼32.9% from the planar. The EQE spectra under planar (dashed) and optimal conformal zig–zag (solid) designs are plotted in Fig. 3(b). Compared to the P abs spectra shown in Fig. 2(a), it is obvious that the assumption of IQE 100% is indeed inaccurate [18]. In fact, the surface
Fig. 3. Electrical response of the a-Si:H TFSCs with the conformal zig–zag structure, with (a) J sc as a function of period, (b) EQE spectra, and (c) photocurrent and power densities as functions of forward electrical bias V.
5073
recombination and other carrier loss mechanisms result in a lower EQE (EQE < P abs ) due to the imperfect internal quantum process [24]. From Fig. 3(b), we can see that the EQE of the conformal zig–zag structure has been significantly improved over the planar in the whole spectral range. Taking λ 300 nm, for example, EQE 40.8% (28.4%) for conformal zig–zag (planar), with an efficiency increase of 12.4%. At wavelengths close to the bandgap, e.g., when λ 700 nm, EQE can be close to 50%, much higher than the planar, whose EQE 13.8%. The peak EQE can be 91.4% at λ 520 nm. We then properly modify the carrier transport model in order to obtain the dark current response [19]. With previously calculated J sc and the dark current densities under various forward electrical bias (V), the current– voltage characteristics of the designed a-Si:H TFSCs can be obtained. The simulation results are plotted in Fig. 3(c), where the output current and power densities for planar and optimized conformal zig–zag designs are shown. Based on these data, we can further obtain the parameters of maximum output power density (P max ), open-circuit voltage (V oc ), fill factor [FF P max ∕ J sc V oc ], and light-conversion efficiency (η P max ∕ P sun , where P sun is the overall incident light power density from the solar). It is found that, compared to a planar system, our design improves J sc from 12.70 to 16.88 mA∕cm2 (enhanced 32.90%) and η from 10.02% to 13.38% (enhanced 33.53%), while simultaneously keeping V oc ∼ 0.92 V and FF > 85%. Finally, we briefly compare the optical absorption properties of the TFSCs under the conformal zig–zag design with the conventional planar configuration. Shown in Fig. 4 are the absorption patterns of these two systems with typical wavelengths, under which the cell shows the better antireflection [e.g., Fig. 4(b1) at λ 400 nm], elongated light path [e.g., Figs. 4(b2) at λ 520 nm and 4(b3) at λ 690 nm], and promising rear metallic nanostructured reflector [Figs. 4(b2) and 4(b3)], respectively. Actually, the carrier generation pattern [25] is straightforwardly similar to the absorption distribution. As can be seen from the figures, the conformal nanostructure
Fig. 4. Optical absorption distributions inside a-Si:H TFSCs with (a1)–(a3) planar and (b1)–(b3) conformal zig–zag configurations under various wavelengths. The carrier generation patterns are very similar to these figures after a typical treatment by considering that one photon generates one electron–hole pair [19,24].
5074
OPTICS LETTERS / Vol. 38, No. 23 / December 1, 2013
substantially modifies the absorption characteristics inside the solar cells, leading to many spatially localized absorption peaks and a higher P abs after spatial integration. In particular, as displayed in Figs. 4(a3) and 4(b3), the absorption capability of a-Si:H under such a wavelength close to the bandgap has been extremely improved, which is the result of the excitation of many high-order modes from the zig–zag rear structure together with the waveguide confinement inside the a-Si:H cavity. This way, the optical path can be greatly increased for more efficient light absorption by the photoactive layer in the bandgap-proximity region. In summary, we investigated both the light-trapping and electrical responses of a-Si:H TFSCs under a conformal zig–zag configuration. The optical absorption and light-conversion efficiency were studied and compared extensively with planar, rose, and normal zig– zag designs, based on the spectral response, polarization sensitivity, reflection property, J sc , EQE, and other typical parameters. It was found that the conformal zig–zag system shows whole-spectrum (in the material band) enhancement with independence of incident light polarization. We explained that this promising feature comes from the joint effect of improved front antireflection, the special rear nanoreflector, and the greatly elongated optical path inside the active layer. EQE response and I– V curves were also studied through electrical simulation. Based on the broadbandly enhanced EQE, J sc and η can be improved by 32.90% and 33.53%, respectively. This work is supported by NSFC (61204066, 91233119), the “Thousand Young Talents Program” of China, and Priority Academic Program Development (PAPD) of Jiangsu Higher Education Institutions. We thank Professor Yu Qiu for advice on revising the paper. References 1. A. V. Shah, H. Schade, M. Vanecek, J. Meier, E. Vallat-Sauvain, N. Wyrsch, U. Kroll, C. Droz, and J. Bailat, Prog. Photovoltaics 12, 113 (2004). 2. M. Zeman, R. A. C. M. M. van Swaaij, J. W. Metselaar, and R. E. I. Schropp, J. Appl. Phys. 88, 6436 (2000).
3. Y. M. Song, J. S. Yu, and Y. T. Lee, Opt. Lett. 35, 276 (2010). 4. M. Wellenzohn and R. Hainberger, Opt. Express 20, A20 (2012). 5. W. Bai, Q. Gan, F. Bartoli, J. Zhang, L. Cai, Y. Huang, and G. Song, Opt. Lett. 34, 3725 (2009). 6. X. Meng, E. Drouard, G. Gomard, R. Peretti, A. Fave, and C. Seassal, Opt. Express 20, A560 (2012). 7. J. Krc, M. Zeman, S. L. Luxembourg, and M. Topic, Appl. Phys. Lett. 94, 153501 (2009). 8. K. Söderström, G. Bugnon, R.-J. Haug, S. Nicolay, and C. Ballif, Sol. Energy Mater. Sol. Cells 101, 193 (2012). 9. K. X. Wang, Z. Yu, V. Liu, Y. Cui, and S. Fan, Nano Lett. 12, 1616 (2012). 10. S. Pillai and M. A. Green, Sol. Energy Mater. Sol. Cells 94, 1481 (2010). 11. T. L. Temple and D. M. Bagnall, Prog. Photovolt. Res. Appl. 21, 600 (2013). 12. H. Tan, L. Sivec, B. Yan, R. Santbergen, M. Zeman, and A. H. M. Smets, Appl. Phys. Lett. 102, 153902 (2013). 13. V. E. Ferry, M. A. Verschuuren, M. C. van Lare, R. E. I. Schropp, H. A. Atwater, and A. Polman, Nano Lett. 11, 4239 (2011). 14. J. Gjessing, E. S. Marstein, and A. S. Sudbø, 26th European Photovoltaic Solar Energy Conference and Exhibition (2011), pp. 2759–2763. 15. J. Gjessing, A. S. Sudbø, and E. S. Marstein, J. Appl. Phys. 110, 033104 (2011). 16. ASTM, “Reference solar spectral irradiance: AM 1.5 Spectra,” http://rredc.nrel.gov/solar/spectra/am1.5/. 17. Comsol Multiphysics, http://www.comsol.com/. 18. X. Li, N. Hylton, V. Giannini, K.-H. Lee, N. J. Ekins-Daukes, and S. A. Maier, Opt. Express 19, A888 (2011). 19. X. Li, N. Hylton, V. Giannini, K.-H. Lee, N. J. Ekins-Daukes, and S. A. Maier, Prog Photovolt. Res. Appl. 21, 109 (2013). 20. Y. Zhan, X. Li, and Y. Li, IEEE J. Sel. Top. Quantum Electron. 19, 4000208 (2013). 21. X. Li and Y. Zhan, Appl. Phys. Lett. 102, 0211101 (2013). 22. E. D. Palik, Handbook of Optical Constants of Solids (Academic, 1985). 23. J. Grandidier, D. M. Callahan, J. N. Munday, and H. A. Atwater, Adv. Mater. 23, 1272 (2011). 24. J. Nelson, The Physics of Solar Cells (Imperial College, 2003). 25. M. G. Deceglie, V. E. Ferry, A. P. Alivisatos, and H. A. Atwater, Nano Lett. 12, 2894 (2012).
5071
Ultra-broadband performance enhancement of thin-film amorphous silicon solar cells with conformal zig–zag configuration Zhenhai Yang, Aixue Shang, Yaohui Zhan, Cheng Zhang, and Xiaofeng Li* Institute of Modern Optical Technologies & Collaborative Innovation Center of Suzhou Nano Science and Technology, Jiangsu Key Lab of Advanced Optical Manufacturing Technologies & MOE Key Lab of Modern Optical Technologies, Soochow University, Suzhou 215006, China *Corresponding author: [email protected] Received September 13, 2013; revised October 8, 2013; accepted October 28, 2013; posted October 29, 2013 (Doc. ID 197541); published November 25, 2013 An ultrathin amorphous silicon solar cell with conformal zig–zag nanoconfiguration is studied from both lighttrapping and light-conversion perspectives. The design improves the front antireflection property, optimizes the rear metallic reflector, and elongates the optical path inside the photoactive layer. Compared to conventional nanoconfigurations, this system shows significant absorption enhancement in the whole amorphous silicon band and exhibits extremely low sensitivity to light polarization. The nano-optimization indicates that the short-circuit current density (light-conversion efficiency) of the 200-nm-thick solar cell can be 16.88 mA∕cm2 (13.38%), showing an enhancement factor of 32.90% (33.53%) from the planar system. © 2013 Optical Society of America OCIS codes: (040.5350) Photovoltaic; (350.4238) Nanophotonics and photonic crystals. http://dx.doi.org/10.1364/OL.38.005071
Solar cells that convert electricity from sunlight have attracted huge interest on the road to clean and renewable energy. The cost of currently available crystalline silicon solar cells is still high because of their thick (>200 μm) active layer. Nevertheless, thin-film solar cells (TFSCs) are considered to be very promising as they can absorb over 90% of sunlight in the material band with much thinner active layers (typically 520 nm. This is because the rear nanostructured reflectors can significantly increase
Fig. 2. (a) Absorption spectra P abs of a-Si:H layer and (b) reflection spectra of the entire device, where Λ 400 nm and H Λ∕6.
the optical path of the long-wavelength light (close to the bandgap). A simple comparison between the rose and zig–zag systems shows the superiority of the latter in light trapping (see P abs when λ > 650 nm). This provides direct optical evidence of the finding in [15], showing the potential of zig–zag nanostructures in outperforming other typical designs. In fact, the better performance of a zig–zag design arises from a slight modification to the rose type by breaking the geometric symmetry (Fig. 1). However, the above zig–zag enhancement is just confined to the LWB (with short-wavelength band, SWB, almost unaffected), where the solar incidence penetrates through the active layer to strongly interact with the back reflector. Although a front “planar” antireflection coating can efficiently decrease the light-reflection loss, space still exists for further improvement in the SWB, in order to achieve a full-spectrum absorption enhancement. Figure 2(a) (blue solid curve) exhibits the absorption spectrum of the considered cell with a conformal zig–zag design. It is obvious that the newly designed solar cell obtains an ultrabroadband performance enhancement (in the whole material absorption band); especially, P abs ∼ 100% within the region of λ ∼ 500 nm. This is the result of the joint effects of better front antireflection, broadband light reflection in the rear of the cell, and a significantly elongated optical path inside the photoactive layer [13]. For further insight into the enhancement mechanism, the corresponding reflection spectra R are given in Fig. 2(b). The energy loss in the SWB comes directly from the reflection of the device surface, while that in the LWB is resulted from the back metallic reflector, material absorption, and cavity resonance, which yield the localized peaks. With a planar Ag reflector, as shown in Fig. 2(b), the device reflection at the LWB is very high since the light undergoes the shortest optical path under normal injection without being sufficiently absorbed. Such a high reflection can be greatly decreased with nonplanar nanodesigns (e.g., rose, and zig–zag) due to the excitation of modes with tilted propagation angles, many of which may satisfy the total internal reflection (TIR) condition at the top surface to greatly improve the lighttrapping performance. Nevertheless, rear rose and zig– zag designs do not modify the reflection when λ < 520 nm, for the reason mentioned above. Under a conformal zig– zag scheme, however, substantial reduction of reflection at the SWB can be seen, which can be ascribed to the better antireflection of the surface [according to Fig. 2(b), R can be very close to 0 when 480 < λ < 600 nm]. A lower reflection in the LWB is also observed to correspond to the better P abs in Fig. 2(a). This is because the titled interface between TCO and a-Si:H lets TIR occur more easily (i.e., under larger incident angles of the light reflected back from the rear zig–zag metallic reflector). The strengthened light confinement thus leads to a much longer optical path for efficient interaction between the light and the photoactive material. It is also worth noting that, although the conformal zig–zag structure is not symmetrical as is the rose, the modified system does not exhibit noticeable polarization dependence. In this study, both the transverse electric (TE) and transverse magnetic (TM) polarizations have been simulated. We find that the results for both TE
December 1, 2013 / Vol. 38, No. 23 / OPTICS LETTERS
and TM are almost the same (a slight difference occurs in the region with very long wavelength beyond the bandgap; not shown here). This reveals that the zig–zag structure is polarization insensitive with great potential in broadband absorption enhancement; therefore, it is expected to provide opportunity for an excellent lighttrapping design. Although a promising light-trapping is achieved, further improvement is still possible by conducting a thorough nano-optimization on the zig–zag structure, i.e., through controlling the period Λ and the height H. However, a conclusion is hard to draw based on the complex spectral behavior of the absorption curve; we thus have to find a simple and direct way to rate the designs. In fact, the spectral behavior can be integrated into a unique parameter J sc (short circuit current density in mA∕cm2 ) with the solar spectrum taken into account [23]. In many cases, J sc values obtained in this way are based on the condition of a perfect internal quantum process (i.e., internal quantum efficiency IQE 100%). A realistic prediction of solar cell performance must consider the detailed internal carrier (electron and hole) process with generation, diffusion, drift, and recombination mechanisms taken into consideration [24]. Based on an electrical modeling technique, we simulate the internal quantum process and obtain the IQE spectrum, which can be used to convert the light absorption to a realistic EQE and, thus, J sc . To find a globally maximal J sc , we sweep H from 50 to 140 nm for typical Λ values (from 200 to 800 nm with steps of 100 nm) to find the best H design for each Λ. A number of three-dimensional electromagnetic simulations yield the results shown in Fig. 3(a), where the maximal J sc for each Λ has been inserted into the bar. It is shown that a relatively large period is needed for the conformally configured TFSCs in order to obtain a high J sc , which can be as high as 16.88 mA∕cm2 (when Λ 600 nm and H 130 nm), with an enhancement of ∼32.9% from the planar. The EQE spectra under planar (dashed) and optimal conformal zig–zag (solid) designs are plotted in Fig. 3(b). Compared to the P abs spectra shown in Fig. 2(a), it is obvious that the assumption of IQE 100% is indeed inaccurate [18]. In fact, the surface
Fig. 3. Electrical response of the a-Si:H TFSCs with the conformal zig–zag structure, with (a) J sc as a function of period, (b) EQE spectra, and (c) photocurrent and power densities as functions of forward electrical bias V.
5073
recombination and other carrier loss mechanisms result in a lower EQE (EQE < P abs ) due to the imperfect internal quantum process [24]. From Fig. 3(b), we can see that the EQE of the conformal zig–zag structure has been significantly improved over the planar in the whole spectral range. Taking λ 300 nm, for example, EQE 40.8% (28.4%) for conformal zig–zag (planar), with an efficiency increase of 12.4%. At wavelengths close to the bandgap, e.g., when λ 700 nm, EQE can be close to 50%, much higher than the planar, whose EQE 13.8%. The peak EQE can be 91.4% at λ 520 nm. We then properly modify the carrier transport model in order to obtain the dark current response [19]. With previously calculated J sc and the dark current densities under various forward electrical bias (V), the current– voltage characteristics of the designed a-Si:H TFSCs can be obtained. The simulation results are plotted in Fig. 3(c), where the output current and power densities for planar and optimized conformal zig–zag designs are shown. Based on these data, we can further obtain the parameters of maximum output power density (P max ), open-circuit voltage (V oc ), fill factor [FF P max ∕ J sc V oc ], and light-conversion efficiency (η P max ∕ P sun , where P sun is the overall incident light power density from the solar). It is found that, compared to a planar system, our design improves J sc from 12.70 to 16.88 mA∕cm2 (enhanced 32.90%) and η from 10.02% to 13.38% (enhanced 33.53%), while simultaneously keeping V oc ∼ 0.92 V and FF > 85%. Finally, we briefly compare the optical absorption properties of the TFSCs under the conformal zig–zag design with the conventional planar configuration. Shown in Fig. 4 are the absorption patterns of these two systems with typical wavelengths, under which the cell shows the better antireflection [e.g., Fig. 4(b1) at λ 400 nm], elongated light path [e.g., Figs. 4(b2) at λ 520 nm and 4(b3) at λ 690 nm], and promising rear metallic nanostructured reflector [Figs. 4(b2) and 4(b3)], respectively. Actually, the carrier generation pattern [25] is straightforwardly similar to the absorption distribution. As can be seen from the figures, the conformal nanostructure
Fig. 4. Optical absorption distributions inside a-Si:H TFSCs with (a1)–(a3) planar and (b1)–(b3) conformal zig–zag configurations under various wavelengths. The carrier generation patterns are very similar to these figures after a typical treatment by considering that one photon generates one electron–hole pair [19,24].
5074
OPTICS LETTERS / Vol. 38, No. 23 / December 1, 2013
substantially modifies the absorption characteristics inside the solar cells, leading to many spatially localized absorption peaks and a higher P abs after spatial integration. In particular, as displayed in Figs. 4(a3) and 4(b3), the absorption capability of a-Si:H under such a wavelength close to the bandgap has been extremely improved, which is the result of the excitation of many high-order modes from the zig–zag rear structure together with the waveguide confinement inside the a-Si:H cavity. This way, the optical path can be greatly increased for more efficient light absorption by the photoactive layer in the bandgap-proximity region. In summary, we investigated both the light-trapping and electrical responses of a-Si:H TFSCs under a conformal zig–zag configuration. The optical absorption and light-conversion efficiency were studied and compared extensively with planar, rose, and normal zig– zag designs, based on the spectral response, polarization sensitivity, reflection property, J sc , EQE, and other typical parameters. It was found that the conformal zig–zag system shows whole-spectrum (in the material band) enhancement with independence of incident light polarization. We explained that this promising feature comes from the joint effect of improved front antireflection, the special rear nanoreflector, and the greatly elongated optical path inside the active layer. EQE response and I– V curves were also studied through electrical simulation. Based on the broadbandly enhanced EQE, J sc and η can be improved by 32.90% and 33.53%, respectively. This work is supported by NSFC (61204066, 91233119), the “Thousand Young Talents Program” of China, and Priority Academic Program Development (PAPD) of Jiangsu Higher Education Institutions. We thank Professor Yu Qiu for advice on revising the paper. References 1. A. V. Shah, H. Schade, M. Vanecek, J. Meier, E. Vallat-Sauvain, N. Wyrsch, U. Kroll, C. Droz, and J. Bailat, Prog. Photovoltaics 12, 113 (2004). 2. M. Zeman, R. A. C. M. M. van Swaaij, J. W. Metselaar, and R. E. I. Schropp, J. Appl. Phys. 88, 6436 (2000).
3. Y. M. Song, J. S. Yu, and Y. T. Lee, Opt. Lett. 35, 276 (2010). 4. M. Wellenzohn and R. Hainberger, Opt. Express 20, A20 (2012). 5. W. Bai, Q. Gan, F. Bartoli, J. Zhang, L. Cai, Y. Huang, and G. Song, Opt. Lett. 34, 3725 (2009). 6. X. Meng, E. Drouard, G. Gomard, R. Peretti, A. Fave, and C. Seassal, Opt. Express 20, A560 (2012). 7. J. Krc, M. Zeman, S. L. Luxembourg, and M. Topic, Appl. Phys. Lett. 94, 153501 (2009). 8. K. Söderström, G. Bugnon, R.-J. Haug, S. Nicolay, and C. Ballif, Sol. Energy Mater. Sol. Cells 101, 193 (2012). 9. K. X. Wang, Z. Yu, V. Liu, Y. Cui, and S. Fan, Nano Lett. 12, 1616 (2012). 10. S. Pillai and M. A. Green, Sol. Energy Mater. Sol. Cells 94, 1481 (2010). 11. T. L. Temple and D. M. Bagnall, Prog. Photovolt. Res. Appl. 21, 600 (2013). 12. H. Tan, L. Sivec, B. Yan, R. Santbergen, M. Zeman, and A. H. M. Smets, Appl. Phys. Lett. 102, 153902 (2013). 13. V. E. Ferry, M. A. Verschuuren, M. C. van Lare, R. E. I. Schropp, H. A. Atwater, and A. Polman, Nano Lett. 11, 4239 (2011). 14. J. Gjessing, E. S. Marstein, and A. S. Sudbø, 26th European Photovoltaic Solar Energy Conference and Exhibition (2011), pp. 2759–2763. 15. J. Gjessing, A. S. Sudbø, and E. S. Marstein, J. Appl. Phys. 110, 033104 (2011). 16. ASTM, “Reference solar spectral irradiance: AM 1.5 Spectra,” http://rredc.nrel.gov/solar/spectra/am1.5/. 17. Comsol Multiphysics, http://www.comsol.com/. 18. X. Li, N. Hylton, V. Giannini, K.-H. Lee, N. J. Ekins-Daukes, and S. A. Maier, Opt. Express 19, A888 (2011). 19. X. Li, N. Hylton, V. Giannini, K.-H. Lee, N. J. Ekins-Daukes, and S. A. Maier, Prog Photovolt. Res. Appl. 21, 109 (2013). 20. Y. Zhan, X. Li, and Y. Li, IEEE J. Sel. Top. Quantum Electron. 19, 4000208 (2013). 21. X. Li and Y. Zhan, Appl. Phys. Lett. 102, 0211101 (2013). 22. E. D. Palik, Handbook of Optical Constants of Solids (Academic, 1985). 23. J. Grandidier, D. M. Callahan, J. N. Munday, and H. A. Atwater, Adv. Mater. 23, 1272 (2011). 24. J. Nelson, The Physics of Solar Cells (Imperial College, 2003). 25. M. G. Deceglie, V. E. Ferry, A. P. Alivisatos, and H. A. Atwater, Nano Lett. 12, 2894 (2012).
