Angular selective semi-transparent photovoltaics - OSA Publishing

Angular selective semi-transparent photovoltaics Brian Roberts, D. M. Nanditha, M. Dissanayake, and P.-C. Ku* Department of Electrical Engineering and Computer Science, University of Michigan, 1301 Beal Avenue, Ann Arbor, Michigan 48109, USA * [email protected]

Abstract: Conventional semi-transparent photovoltaics suffer from an inherent tradeoff between the amount of visible light transmitted versus absorbed, reducing energy conversion efficiency when higher transparency is desired. As a solution to lift this tradeoff, we propose a wavelength and angular selective reflector and demonstrate a potential implementation utilizing high aspect ratio metal nanoparticles. Using the anisotropy in the localized surface plasmon resonance wavelength, the proposed device can selectively harness sunlight incident at an elevated angle, increasing the power conversion efficiency by a factor of 1.44, while maintaining 70 percent optical transparency at normal incidence. ©2012 Optical Society of America OCIS codes: (350.6050) Solar energy; (250.5403) Plasmonics.

References and links 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

J. Benemann, O. Chehab, and E. Schaar-Gabriel, “Building-integrated PV modules,” Sol. Energy Mater. Sol. Cells 67(1-4), 345–354 (2001). R. Lunt and V. Bulovic, “Transparent, near-infrared organic photovoltaic solar cells for window and energyscavenging applications,” Appl. Phys. Lett. 98(11), 113305 (2011). D. H. W. Li, T. N. T. Lam, W. W. H. Chan, and A. H. L. Mak, “Energy and cost analysis of semi-transparent photovoltaic in office buildings,” Appl. Energy 86(5), 722–729 (2009). R. F. Bailey-Salzman, B. P. Rand, and S. R. Forrest, “Semitransparent organic photovoltaic cells,” Appl. Phys. Lett. 88(23), 233502 (2006). D. M. N. M. Dissanayake, B. Roberts, and P. C. Ku, “Plasmonic backscattering enhanced inverted photovoltaics,” Appl. Phys. Lett. 99(11), 113306 (2011). C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles (Wiley, 1983). J. Byun, J. I. Lee, S. Kwon, G. Jeon, and J. K. Kim, “Highly ordered nanoporous alumina on conducting substrates with adhesion enhanced by surface modification: universal templates for ultrahigh-density arrays of nanorods,” Adv. Mater. (Deerfield Beach Fla.) 22(18), 2028–2032 (2010). M. F. Cansizoglu, R. Engelken, H. W. Seo, and T. Karabacak, “High optical absorption of indium sulfide nanorod arrays formed by glancing angle deposition,” ACS Nano 4(2), 733–740 (2010). S. B. Chaney, S. Shanmukh, R. A. Dluhy, and Y.-P. Zhao, “Aligned silver nanorod arrays produce high sensitivity surface-enhanced Raman spectroscopy substrates,” Appl. Phys. Lett. 87(3), 031908 (2005). H. A. Atwater and A. Polman, “Plasmonics for improved photovoltaic devices,” Nat. Mater. 9(3), 205–213 (2010). A. F. Oskooi, D. Roundy, M. Ibanescu, P. Bermel, J. D. Joannopoulos, and S. G. Johnson, “MEEP: A flexible free-software package for electromagnetic simulations by the FDTD method,” Comput. Phys. Commun. 181(3), 687–702 (2010). B. Roberts and P.-C. Ku, “Broadband characteristics of surface plasmon enhanced solar cells,” in 2010 35th IEEE Photovoltaic Specialists Conference (Institute of Electrical and Electronics Engineers, New York, 2010), pp. 002952–002954. A. D. Rakic, A. B. Djurišic, J. M. Elazar, and M. L. Majewski, “Optical properties of metallic films for verticalcavity optoelectronic devices,” Appl. Opt. 37(22), 5271–5283 (1998). E. Lioudakis, A. Othonos, I. Alexandrou, and Y. Hayashi, “Optical properties of conjugated poly(3hexylthiophene)/[6,6]-phenylC61-butyric acid methyl ester composites,” J. Appl. Phys. 102(8), 083104 (2007). B. Smits, “An RGB to spectrum conversion for reflectances,” J. Graphics Tools 4(4), 11–22 (1999). U.S. Department of Energy Benchmark, “U.S. Department of Energy Commercial Reference Building Models of the National Building Stock” (National Renewable Energy Laboratory, 2011). http://www.nrel.gov/docs/fy11osti/46861.pdf.

1. Introduction Semi-transparent photovoltaics (PVs) have received recent attention as a solution for building integrated photovoltaics (BIPVs) [1–4]. Integrating a semi-transparent PV onto a building

#160125 - $15.00 USD (C) 2012 OSA

Received 19 Dec 2011; revised 30 Jan 2012; accepted 1 Feb 2012; published 9 Feb 2012 13 February 2012 / Vol. 20, No. 4 / OPTICS EXPRESS 265

window can harness a large amount of incident sunlight that is otherwise reflected or converted to heat, as well as reducing unwanted glare. However, a conventional semitransparent PV suffers from an inherent tradeoff between the amount of visible light that can be converted into photocurrent versus the amount that must be transmitted through the window for daylight illumination, aesthetics, and other benefits. In this paper, we propose an angle selective back reflector layer for semi-transparent PVs, with the goal of allowing more window transparency for normally incident light while facilitating enhanced harvesting of direct sunlight entering from an elevated angle, as shown in Fig. 1. As light passes through the semi-transparent PV layer, it either goes through the angular selective reflector or is reflected back to the PV absorber again, according to the angle of incidence. The absorption enhancement can be further tuned by controlling the distance between the PV absorber and the angular selective back reflector via the optical cavity effect [5].

Fig. 1. Angle selective photovoltaic window, designed for transparency to normally incident light and harvesting of angled light, including direct sunlight.

2. Proposed implementation To show that the proposed angle selective PV device can be potentially realized for a large area in a scalable fashion, we considered an array of anisotropically shaped subwavelength metal nanoparticles. The scattering resonance of an anisotropically shaped metal nanoparticle depends on the polarization direction of the light with which it is illuminated [6]. The polarization is in turn determined by the angle of incoming light, providing a method of angular selectivity. For elliptical metal particles, a short wavelength localized surface plasmon (LSP) response is obtained for light polarized along a minor ellipse axis and a red-shifted LSP response is seen for light polarized along the major axis [6]. Using this principle, a monolayer of nanoparticles can be designed to scatter different portions of the spectrum as the angle (and thus the polarization direction) of the incident light is changed. Our proposed backscattering layer is shown schematically in Fig. 2. An array of elliptical silver nanorods are placed behind a planar thin film PV, with the long ellipse axes approximately normal to the PV layer. Such a structure can potentially be fabricated using a number of techniques such as electrodeposition with an anodized aluminum oxide template or oblique angle deposition [7–9]. The spectral locations of the long and short axis LSP resonances are determined by the silver nanorod shape and aspect ratio, the surrounding dielectric material, and to a lesser extent, the size of the particle relative to the optical wavelength [6]. We chose 120 nm by 50 nm elliptical particles to demonstrate a clear resonance splitting in simulations. Particles are spaced 200 nm apart, chosen as a compromise between maintaining a transparency of ca. 70% while having adequate metal nanoparticle density for significant scattering effects. For each particle to scatter as much light as possible, they are spaced out far enough that their resonant scattering cross sections do not overlap. #160125 - $15.00 USD (C) 2012 OSA


Adequate spacing also mitigates coupled resonance effects. We note that periodic spacing can also affect scattering via Bloch-type modes [10], though we limit the scope of this paper to focus only on the anisotropic structure of each scatterer. As shown in Fig. 2, metal particles are oriented with a 15-degree rotation with respect to the normally incident light. The additional rotation allows fine tuning of the angles of light that are targeted for maximum backscatter, though at the cost of a small component of the long wavelength characteristics emerging at normal incidence. Advantageously, metallic nanostructures can be placed completely outside the electrical and optical paths of the PV cell itself, eliminating any adverse effects on charge transport and collection. Furthermore, the independent design of the angle-selective reflector and the PV cell allows one to control the spacing between them, potentially realizing additional absorption enhancement via the optical cavity effect [5].

Fig. 2. Proposed geometry for semi-transparent thin film photovoltaic (including transparent electrodes (TE), electron/hole blocking layers (EBL/HBL) and thin absorbing region) with metal particle back scattering layer.

3. Numerical analysis We theoretically analyze the optical properties of the metal nanoparticle scattering layer using full-field finite-difference time-domain (FDTD) optical simulations [11, 12]. A 3-dimensional volume representing the active region of the PV cell is defined, containing the geometry shown in Fig. 2. Five nano-ellipses are simulated across the x direction of the region. In the y direction, periodic boundary conditions are used to reduce computation time. Perfectly matched layers (PML) boundaries are enforced in the x and z directions [11]. Objects are modeled by their complex dielectric constant ε=ε’+iε”. For silver particles, the frequency dependence of ε is approximated by a Lorentz-Drude model fit to tabulated data [13]. The 60 nm thick PV absorber layer is modeled with a refractive index n=1.8 and an absorption coefficient α=5×104/cm, approximating the absorption band of an organic polymer PV system [14]. This absorption has been treated as constant across all wavelengths purely to improve the clarity of these results, and we note that the principles of our device can be extended to any thin film material system of interest. A broadband pulse of spatially coherent planar light is propagated through the structure, starting from the transparent electrode layer. The initial interface between air and the dielectric stack, assumed to be further from the PV layer than the coherence length of sunlight, is neglected for computational simplicity. The Fourier-transformed optical flux entering each object in the simulation, given by the Poynting vector, is used to determine absorption spectra. The response to angled incident light is simulated by angling the geometry within the x-z plane with respect to the planar source. 4. Results Figure 3 shows results of the electromagnetic simulation of the device active region in the spectral domain. Normally incident light (dashed curves of Fig. 3(a)) shows a drop in transmission (blue dashed curve) and significant metal nanoparticle absorption (red dashed #160125 - $15.00 USD (C) 2012 OSA


curve) around the 400 nm LSP resonance associated with the minor axis of the metal particles. Much of the visible spectrum after 450 nm is transmitted with approximately 70% intensity, with the exception of a small perturbation at 700 nm, associated with the particles not being perfectly vertically aligned. Transverse Magnetic (TM) polarized light at 45 degrees (solid curves) shows significant interaction with the major ellipse axis and the corresponding 700 nm LSP resonance. There is a broad corresponding peak in improved absorption by the PV material (black solid curve), and a drop in device transmission (blue solid curve). A factor of 1.7 in improved PV layer absorption is realized for TM light near the redshifted LSP resonance. Finally, as shown by Fig. 3(b), we observe that the influence of the redshifted LSP resonance increases with increasing angle of incident light.

Fig. 3. (a) Performance spectra for the device of Fig. 2 under normal illumination (dashed lines) and TM polarized illumination at an angle of 45 degrees (solid lines). Short wavelength LSP resonance at 400 nm dominates normal illumination. Angled illumination interacts with the 700 nm resonance, increasing PV absorption by backscattering. (b) PV absorption component, increasing with incident angle of TM polarized light.

A number of tradeoffs and limitations must be considered in determining device performance. The inclusion of any metal nanostructure is accompanied by intrinsic optical losses (red curves of Fig. 3(a)). For such structures to be effective, the advantages of spectral control, used here to enable angular selectivity, must outweigh the optical losses. Furthermore, the metal particles only target the TM components of angled incident light. It is possible to design plasmonic structures to scatter both polarizations of angled light (nanodisks rather than nanorods, for example), though such structures would not be as highly transparent to normally incident light. While the scattering mechanism is optimized for TM light only, both polarizations realize some improved absorption of angled light due to the longer angled path length through the PV layer. Additionally, the Fresnel equations dictate that angled light near Brewster's angle will have a more significant TM component after transmission from the air into the dielectric layers on the front surface of the PV cell, making targeting of TM light more advantageous. As a demonstration of the merit of proposed structure, we use the FDTD transmission spectra to simulate the appearance of an RGB image as seen through the window at different incident angles [15]. The image maintains good optical quality and brightness at normal incidence, with transmission degraded only at high incident angles, as shown in Fig. 4(a). Lastly, we weight the simulated performance spectra by the AM1.5 solar spectrum and the angular availability of direct sunlight [16], including the contributions from both TM and TE polarized light. In Fig. 4(b), we plot the tradeoff between total absorption of solar power (x axis) and transmission of visible light at normal incidence (y axis). The blue line marks the upper transparency limit for a given amount of PV absorption in a conventional transparent PV structure. Most importantly, the angular selective harvesting of light with this structure improved solar absorption while maintaining transparency, with a simulated improvement factor of 1.44.

#160125 - $15.00 USD (C) 2012 OSA


Fig. 4. (a) Simulated images, as seen through the transmission spectra for normal and 60 degree incident unpolarized light. (b) Tradeoff between transmission of normally incident visible light and PV absorption of solar radiation for a conventional semi-transparent BIPV window, with contours showing factors of improvement and a point indicating the performance with a selective scattering layer. The integration ranges for transmission and PV absorption calculations are from 480 – 650 nm and 480 – 780 nm, respectively.

Having established that metal nanorods can be employed as a selective backscattering layer for semi-transparent PV, we note that further study is possible for improvement of these results. In this work, we have used a highly idealized model of a thin-film absorbing layer, and, for clarity, chose silver nano-rods that demonstrate well-spaced LSP resonances. Optimizing a nano-rod backscattering layer for a particular thin-film material system requires tuning the particle size and aspect ratio to maximize overlap between the resonant scattering and the material absorption band [14]. The spacing between the PV stack and the backscattering array can also be tuned to ensure that incident and backscattered light interfere constructively in the PV film to maximize absorption, which requires a detailed model of the relevant layers [5]. We seek to extend the results of this study to real thin-film PV systems in future work, with the goal of realizing designs with comparable levels of transparency and absorption enhancement. 5. Conclusion In conclusion, we have discussed the motivation for angular selectivity of light harvesting in semi-transparent BIPVs, and have described and numerically simulated a metal nanostructured back reflector structure to realize this behavior. By integrating the backscattering layer with an idealized thin-film PV device, we showed that the PV power conversion efficiency can be improved by a factor of 1.44 while maintaining up to 70% optical transparency for normally incident light. Acknowledgments This work is supported as part of the Center for Solar and Thermal Energy Conversion, an Energy Frontier Research Center funded by the U.S. Department of Energy Office of Science, Office of Basic Energy Sciences under Award Number DE-SC0000957. We thank Prof. Max Shtein at the University of Michigan for fruitful discussions.

#160125 - $15.00 USD (C) 2012 OSA
