Letter pubs.acs.org/NanoLett
Extreme Light Absorption by Multiple Plasmonic Layers on Upgraded Metallurgical Grade Silicon Solar Cells Duck Hyun Lee,† Jae Young Kwon,† Stephen Maldonado,‡ Anish Tuteja,*,† and Akram Boukai*,† †
Department of Materials Science and Engineering and ‡Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109, United States S Supporting Information *
ABSTRACT: We fabricate high-efficiency, ultrathin (∼12 μm), flexible, upgraded metallurgical-grade polycrystalline silicon solar cells with multiple plasmonic layers precisely positioned on top of the cell to dramatically increase light absorption. This scalable approach increases the optical absorptivity of our solar cells over a broad range of wavelengths, and they achieve efficiencies η ≈ 11%. Detailed studies on the electrical and optical properties of the developed solar cells elucidate the light absorption contribution of each individual plasmonic layer. Finite-difference time-domain simulations were also performed to yield further insights into the obtained results. We anticipate that the findings from this work will provide useful design considerations for fabricating a range of different solar cell systems. KEYWORDS: Upgraded metallurgical-grade silicon, flexible solar cell, surface plasmons, block copolymer nanolithography
S
surface plasmon resonance is a particularly exciting methodology for achieving dramatically higher light absorption in ultrathin solar cells. A plasmon is the collective oscillation of conduction electrons stimulated by incident light at the interface between a metal (Ag, Au, Pt) and a dielectric. Both the shape and the size of the metal are key factors in determining the coupling between the metal particle and the dielectric, and thereby the overall enhancement in the optical absorption efficiency caused by surface plasmonic resonance.23 Further, precise three-dimensional placement of plasmonic metal particles can also be used to effectively concentrate light on top of the solar cell, and thereby increase its conversion efficiency.25 Indeed, it has been recently shown that metallic plasmonic nanoparticles can dramatically increase the short circuit current in silicon solar cells.9,23 One methodology for enabling, high-throughput, large-scale assembly of an array of plasmonic metallic nanoparticles on top of a solar cell is block copolymer nanolithography.26−31 In this work, we fabricate multiple plasmonic layers (either double or quadruple plasmonic nanoparticle layers) on top of ultrathin UMG-silicon solar cells using block copolymer nanolithography (see Supporting Information Figure S1). These plasmonic nanoparticle layers act as highly efficient light absorbing coatings that dramatically improve the optical absorption of the developed solar cells. The fabrication details for the solar cells developed here are provided as Supporting Information. Briefly, the 12 μm thin UMG-silicon solar cells were prepared by polishing an UMG-
olar cells must generate electricity at a cost 100 μm). Therefore, improving the efficiency of an ultrathin, indirect bandgap solar cell requires engineering of the cell surface so that the light absorptance can be significantly increased. Several methods have been investigated thus far to increase the light absorptance of thin solar cells, including dielectric antireflection coatings,14,15 surface texturing,16−21 and utilizing surface plasmon resonance.7,22−24 Among these approaches, © 2014 American Chemical Society
Received: December 27, 2013 Revised: February 24, 2014 Published: March 10, 2014 1961
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 1. (a) Optical image of a 12 μm thick UMG-silicon. Because of their thickness, UMG-silicon solar cells have a bending radius as low as ∼10 mm. (b,c) Top (b) and tilted (c) scanning electron microscopy (SEM) images of a nanoporous UMG-silicon substrate patterned using block copolymer lithography. (d,e) Top SEM images of nanoporous UMG-silicon after depositing a 5 nm silver film before (d) and after thermal annealing at 100 °C (e). (f) Representative current density−voltage (J−V) curves of the 12 μm thick UMG-silicon solar cells without (blue) and with (red) silver-based quadruple plasmonic layer under light (100 mW/cm2).
enhanced antireflection and to serve as a substrate for the deposition of two additional plasmonic layers. The two top plasmonic layers were fabricated on top of the spin on glass film using the same procedure as was used earlier to fabricate the bottom two plasmonic layers. The final structure that incorporates the bottom two plasmonic layers, the spin on glass, and the top two plasmonic layers, constitutes our “quadruple plasmonic layer” structure (see Supporting Information Figure S2a). A bare ultrathin silicon substrate has extremely poor light absorption, especially at longer wavelengths (λ > 600 nm). This severely limits the ultimate solar cell performance and efficiency (see Supporting Information Figure S3). However, the fabricated quadruple plasmonic layer dramatically increases the effective optical thickness of the ultrathin silicon substrate. Compared to the thin UMG-silicon substrate, the presence of silver-based quadruple plasmonic layer greatly decreases the reflection of light from the sample surface. As a consequence, the color of the cell becomes significantly darker after the incorporation of the quadruple plasmonic layer (see Supporting Information Figure S4). Figure 1f shows the J−V measurements for a 12 μm thick bare UMG-silicon cell, as well as a similar cell coated with the quadruple plasmonic layer under 100 mW/cm2 (AM 1.5) of light at room temperature. The short circuit current densities (JSC) for the UMG-silicon cell without and with the quadruple plasmonic layer were measured to be 9.53 and 31.16 mA/cm2, respectively. The corresponding solar cell conversion efficiencies (η) were measured to be 2.59 and 10.73%, respectively (see Supporting Information Table S2). Thus the quadruple plasmonic layer increases the efficiency of the thin film UMG-silicon solar cells by over 300%. To better understand the contribution of each constituent within the fabricated solar cells, we also conducted detailed device performance evaluation for six different solar cells: (i) A
silicon (99.999% purity) and subsequently doping phosphorus and boron on the top and bottom of the substrate respectively to obtain an n−i−p structure. The as-fabricated ultrathin solar cells were extremely flexible (bending radius as low as 10 mm) due to their thinness (Figure 1a). To fabricate the multiple plasmonic metallic nanoparticle layers on top of the UMGsilicon substrate we utilized a polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA, molecular weight 46k-b-21k) block copolymer. The block copolymer was spin-coated on top of the UMG-silicon substrate and thermally annealed to yield perpendicular self-assembled domains of cylinder forming PMMA block within a PS matrix (see Supporting Information Figure S1). The PMMA domains were then selectively removed, leaving behind a nanoporous polystyrene matrix. Reactive ion etching (SF6/C4F8) was then used to etch into the silicon substrate using the nanoporous polystyrene template as an etching mask. This procedure duplicates the hexagonal arrangement of the block-copolymer template and produces a highly dense and uniform nanoporous array on the surface of the UMG-silicon solar cells (Figure 1b). The diameter and depth of the etched nanopores were maintained at 20 nm, and the pores had a hemiellipsoidal shape (Figure 1c). For the preparation of plasmonic nanoparticles, a metallic (either silver or gold) thin film (t = 5 nm) was deposited over the prepared nanoporous UMG-silicon cells (Figure 1d). The cells were subsequently annealed on a hot plate (T = 100 °C). This caused the deposited metal thin film to dewet and form nanoparticles both within the nanopores (1st plasmonic layer), as well as, outside the nanopores (2nd plasmonic layer). The metal nanoparticles had mean diameters of 16 and 18 nm inside and outside the nanopores, respectively (Figure 1e). These two plasmonic layers together constituted the bottom “double plasmonic layer” on our UMG-silicon solar cell. Next, spin on glass was coated on top of the double plasmonic layer for 1962
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 2. (a) J−V curves for bare UMG-silicon (black), nanoporous UMG-silicon (blue), silver-based double plasmonic layer on UMG-silicon (green), silver-based double plasmonic layer (DPL) along with a top coating of spin on glass (green), UMG-silicon with a silver-based quadruple plasmonic layer (red), and double nanoporous layer (brown). Note the thickness of UMG-silicon is 180 μm. (b) Absorptance spectra for bare UMGsilicon (black), silver-based double plasmonic layer on UMG-silicon (green), and silver-based quadruple plasmonic layer on UMG-silicon (red). (c) Quantum efficiency curves for bare UMG-silicon (black), double nanoporous layer (blue), silver-based double plasmonic layer on UMG-silicon (green), and silver-based quadruple plasmonic layer on UMG-silicon (red). (d) VOC and FF for each sample. (e−g), Power absorption profiles calculated by finite difference time domain (FDTD) model under an incident light of 600 nm along the x−z plane of bare silicon (e), silver-based double plasmonic layer on silicon (f), and silver-based quadruple plasmonic layer on silicon (g). Note the incident light is propagating along the zaxis and is polarized along the x-axis. White dotted lines show surfaces of silicon (near 0 nm) and SiO2 (near 100 nm).
UMG-silicon with the double plasmonic layer, UMG-silicon with the double plasmonic layer and spin on glass, UMG-silicon with the quadruple plasmonic layer, as well as, double nanoporous layer were measured to be 13.67, 15.61, 21.53, 28.27, 34.23, and 20.71 mA/cm2, respectively (see Supporting Information Table S3). Compared with bulk UMG-silicon cell (JSC ∼ 13.67 mA/cm2), the presence of bulk spin on glass layer increased the JSC to 19.17 mA/cm2 due to the antireflection effect (also see Supporting Information Figure S5a).14,15 The nanoporous structures on UMG-silicon cell and the double nanoporous layer slightly increased the JSC to 15.61 and 20.71 mA/cm2, respectively, due to the reduced reflection caused by increased light scattering.32,33 Compared with the nanoporous UMG-silicon (JSC ∼ 15.61 mA/cm2) and double nanoporous layer (JSC ∼ 20.71 mA/cm2), the double plasmonic layer and the quadruple plasmonic layer increased the JSC to 20.71 and
bare UMG-silicon solar cell without any treatment, (ii) a nanoporous UMG-silicon solar cell (nanopores etched using block copolymer nanolithography), (iii) a UMG-silicon solar cell with a silver-based double plasmonic layer, (iv) a UMGsilicon solar cell with a silver-based double plasmonic layer along with a top coating of spin on glass, (v) a UMG-silicon solar cell with a silver-based quadruple plasmonic layer, and (vi) a double nanoporous layer. Please note that in each case the thickness of the UMG-silicon film was 180 μm, and they were doped under the same conditions as the 12 μm cells. This large thickness ensured that there were no transmission losses in all of our measurements (see Supporting Information Figure S3a). The thicknesses of the spin on glass film and the deposited silver thin film were 100 and 5 nm, respectively. Figure 2a shows the J−V curves for the six different solar cells. The JSC for bare UMG-silicon, nanoporous UMG-silicon, 1963
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 3. (a) J−V curves for the silver-based quadruple plasmonic layer on a 180 μm thick UMG-silicon solar cell as a function of spin on glass thickness. Zero nanometers of spin on glass data indicates silver-based double plasmonic layer on UMG-silicon cell. (b) J−V curves for the silverbased quadruple plasmonic layer on a 180 μm thick UMG-silicon solar cell as a function of the deposited silver layer thickness. Zero nanometers of silver data indicates porous UMG-silicon without any plasmonic layer.
34.23 mA/cm2, respectively. These measurements clearly show that the presence of the bottom double plasmonic layer, spin on glass, and the top double plasmonic layer significantly increased the JSC. This increase of the short circuit current is attributed to the absorptance enhancement caused by both the light scattering caused by each plasmonic layer, and the antireflection properties of spin on glass.34 To better elucidate the effects of the individual plasmonic layers on the absorptance of our solar cells, we measured the reflectance (R) as a function of wavelength in the range 300 nm ≤ λ ≤ 950 nm for the bare UMG-silicon cell, UMG-silicon cell with the double plasmonic layer, and the UMG-silicon cell with the quadruple plasmonic layer. Figure 2b shows the calculated absorptance curves for the three cells. The absorptance was calculated by subtracting the reflectance values from unity (absorptance = 1 − R). The double plasmonic layer dramatically increased the absorptance of UMG-silicon cell for all wavelengths >400 nm, a result that is consistent with previous work on plasmonic absorption by silver nanoparticles.9,28 In contrast to the silver-based double plasmonic layer, the quadruple plasmonic layer has an additional spin on glass antireflective layer, and a top silver-based double plasmonic layer. The absorptance for the quadruple plasmonic layer increases to ∼98% across the entire wavelength spectrum, making the cell a nearly perfect light absorber. We also performed quantum efficiency measurements in the visible and near IR wavelengths on the UMG-silicon cells. Figure 2c shows the quantum efficiency curves for bare UMG-silicon (black), double nanoporous layer (blue, see Supporting Information Figure S5), silver-based double plasmonic layer on UMGsilicon (green) and silver-based quadruple plasmonic layer on UMG-silicon (red). At a wavelength λ ∼ 800 nm, the bare UMG-silicon cell exhibits ∼41% quantum efficiency, whereas the quantum efficiency for the quadruple plasmonic cell reaches ∼90%, which agrees with the significant increase in JSC caused by silver nanoparticle based quadruple plasmonic layer. To further elucidate the absorptance enhancement effects of the quadruple plasmonic layer on silicon, the electromagnetic field intensities and the power absorption profiles for bare silicon, the double plasmonic layer on silicon, and the quadruple plasmonic layer on silicon were calculated using a finite different time domain (FDTD) method (Lumerical Solutions Inc.). The electromagnetic field calculations also clearly show strong light absorption at the interface between the silver nanoparticles and silicon (see Supporting Information Figure S6 and S7). The total power absorption profile in the
silicon is also greatly enhanced with the presence of both the silver-based double plasmonic layer (Figure 2f) and the quadruple plasmonic layer (Figure 2g). Note it is well-known that plasmon resonance increases light absorption by the nanoparticles, resulting in heating.35 However, the volume integrated total light absorption curves show that the nanoparticle light absorption is negligible, and that most of the light is absorbed by silicon (see Supporting Information Figure S7g). We also compared the effects of the first and second silver-based double plasmonic layers to better understand their individual contributions (see Supporting Information Figure S8). The silver nanoparticles from the first plasmonic layer within the double plasmonic layer were buried within UMG-silicon, which allows for a large interfacial area between the particles and silicon. Therefore, the first plasmonic layer showed more light absorption near the surface of the silver nanoparticles when compared with the second plasmonic layer, which had a much lower interfacial contact area with the substrate. The use of the plasmonic light absorption layers and antireflection layer affects the open circuit voltage (VOC) and fill factor (FF) as well. Surface recombination of carriers results in a decrease of VOC and FF, especially in thin film materials.36−38 The front surface of the cell corresponds to the highest carrier generation region in the solar cell since most of the incident light is absorbed there. Decreasing surface recombination is typically accomplished by reducing the number of dangling silicon bonds at the front surface using a surface passivation layer.36 Figure 2d shows the VOC and FF for the same six samples as Figure 2a. The data shows that the surface passivation caused by the spin-on-glass and the silver nanoparticles may provide a slight increase in both the VOC and FF. Overall, due to the absorption enhancement and surface passivation, the 180 μm thick UMG-silicon cell with the quadruple plasmonic layer showed a very high efficiency of η ∼ 11.5%. To maximize the light concentration and absorption due to the quadruple plasmonic layer, the three-dimensional placement of plasmonic nanoparticles and the surface passivation layer are critical. Figure 3a shows the J−V curves for the silverbased quadruple plasmonic solar cells as a function of spin on glass thickness. The thickness of the deposited silver thin film was fixed at 5 nm. The data shows that the spin on glass film with a thickness of 100 nm leads to the highest enhancement in JSC. When the thickness of spin on glass is higher than 100 nm, the JSC decreased (also see Supporting Information Table S4). 1964
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 4. (a) Schematic illustrations of bare UMG-silicon, spin on glass on UMG-silicon, top silver double plasmonic layer on spin on glass film, and bottom silver double plasmonic layer under spin on glass film. (b,c) J−V curves (b) and efficiency plots (c) for each sample. Note the thickness of UMG-silicon is 180 μm. (d−f) Power absorption profiles calculated by FDTD model under an incident light of 700 nm along the x−z plane of SiO2 (100 nm) on silicon (d), top silver double plasmonic layer (e), and bottom silver double plasmonic layer (f) on silicon. White dotted lines show surfaces of silicon (near 0 nm) and SiO2 (near 100 nm).
This observation of optimal spin on glass condition of ∼100 nm agrees well with previously reported results.14,39 Figure 3b presents the J−V curves for the silver-based quadruple plasmonic solar cells as a function of silver thin film thickness. The thickness of spin on glass was fixed at 100 nm. The JSC for the silver-based quadruple plasmonic solar cells varied significantly with the thickness of silver thin film. The best JSC of 34.23 mA/cm2 was attained at a silver thickness of 5 nm. The JSC decreased with further increase of silver thickness, for example, a 15 nm thin silver film decreased the JSC to 7.09 mA/cm2 (see Supporting Information Table S5), which is much lower than the bare UMG-silicon solar cell. This variation is likely because with increasing thickness, it becomes difficult to cause the dewetting of the deposited silver film.40 To further differentiate between the light absorption effects of the silver nanoparticles and the antireflection effects of the spin on glass film, we fabricated and measured the properties of four different solar cells: (i) a bare UMG-silicon cell, (ii) a UMG-silicon cell with a spin on glass layer, (iii) a UMG-silicon cell with a spin on glass layer, covered on top with a silverbased double plasmonic layer and (iv) a UMG-silicon cell with a silver-based double plasmonic layer covered on top with a spin on glass film (Figure 4a). Figure 4b,c shows the J−V curves and efficiency measurements for the four cells. Comparing across the obtained data, it is clear that the surface
passivation from the spin on glass film enhanced both the VOC (∼0.53) and FF (∼0.63) for the UMG-silicon cell (also see Supporting Information Table S6). Further, the absorptance enhancement caused by the spin-on-glass film increased JSC to 19.17 mA/cm2, resulting in an overall efficiency increase of 2.4%. The addition of the top silver-based double plasmonic layer (cell iii) and the bottom silver-based double plasmonic layer (cell iv) increased the JSC to 25.07 and 28.28, mA/cm2 respectively. The change of VOC (∼0.53) and FF (∼0.63) among the three cells (ii, iii, and iv) was negligible. Thus the top and bottom silver-based double plasmonic layers increased the absolute efficiency of the cells by ∼2.1 and ∼3.2%, respectively. Note that the total efficiency improvements caused by the bottom silver-based double plasmonic layer (3.20%), spin on glass (2.40%), and top silver-based double plasmonic layer (2.10%) add up to essentially the overall efficiency enhancement caused by the quadruple plasmonic layer on UMG-silicon (7.50%). This trend also agrees with the calculated power absorption profiles of SiO2 on silicon (Figure 4d), top silver-based double plasmonic layer on silicon (Figure 4e), and bottom silver-based double plasmonic layer on silicon (Figure 4f). The larger increase in the magnitude of JSC caused by the bottom double plasmonic layer when compared with the top double plasmonic layer can be explained by the difference in their respective environments. The bottom silver double 1965
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 5. (a,b) J−V curves (a), as well as FF and efficiency plots (b) for bare UMG-silicon, gold/silver quadruple plasmonic layers, and silver/silver quadruple plasmonic layer. Note the thickness of UMG-silicon is 180 μm. (c,d) Field-intensity profiles calculated by FDTD model under an incident light of 400 nm along the x−z plane of gold/silver quadruple plasmonic layer (c) and silver/silver quadruple plasmonic layer (d) on silicon. White dotted lines show surfaces of silicon (near 0 nm) and SiO2 (near 100 nm).
electromagnetic field intensity maps for the Au/Ag cell (Figure 5c) and the Ag/Ag cell (Figure 5d) clearly show the difference in light absorption intensities between the silver and gold nanoparticles. The power absorption profiles (see Supporting Information Figure S9) also support this hypothesis. The results from this work illustrate that multiple plasmonic layers composed of double or quadruple metallic plasmonic nanoparticle layers can be used to fabricate ultrathin solar cells with nearly perfect light absorption. Our quadruple plasmonic layers were fabricated on ultrathin (12 μm), low-cost UMGsilicon substrates and exhibit significantly enhanced JSC, VOC, and FF over bulk cells, resulting in overall solar cell efficiencies of almost 11%. The developed cells fabricated using scalable manufacturing techniques are flexible and use relatively impure, ultrathin silicon, thereby potentially addressing both the material and installation costs for silicon based solar cells. Systematic studies on the optical and electrical properties of the quadruple plasmonic layers elucidate the light absorption effects of each individual component within the developed module. FDTD calculations help explain the mechanisms behind the enhanced absorption caused by each double plasmonic layer and the spin on glass. The results reported here can provide useful design considerations for future work on upgraded metallurgical grade silicon cells, as well as, a range of other classes of solar cells.
plasmonic layer is surrounded by higher refractive index materials, silicon (refractive index, n ∼ 3.5) and SiO2 (n ∼ 1.5), which prevent significant dispersion of light.41 In comparison, the top silver-based double plasmonic layer is surrounded by air (n ∼ 1) and SiO2 (also see Supporting Information Figure S8). The larger mismatch in refractive indices leads to increased scattering at the SiO2/Si interface, which results in a larger increase of JSC. Besides silver, other metallic nanoparticles can also be used to fabricate effective plasmonic light absorpting layers. Gold nanoparticles have been previously considered as good plasmonic materials.42−44 We compared the effects of utilizing silver and gold nanoparticle layers within the quadruple plasmonic layer on the overall efficiency of the UMG-silicon solar cells. Figure 5a,b shows the J−V curves, FF, and overall efficiency plots for two different UMG-silicon cells with quadruple plasmonic layers. The first cell (denoted as the Au/Ag cell) has a bottom double plasmonic layer composed of gold nanoparticles, a spin on glass layer, and a top double plasmonic layer composed of silver nanoparticles. As silver nanoparticles are known to have higher light absorption efficiency than gold nanoparticles in the short wavelength regime,28 we used silver nanoparticles for the top double plasmonic layer. This is because the top plasmonic layer typically absorbs the short wavelength light. For the second cell (denoted as the Ag/Ag cell), both the bottom and top double plasmonic layers are composed of silver. The JSC for Au/Ag cell and the Ag/Ag cell were measured to be 24.82 and 34.23 mA/ cm2, respectively (see Supporting Information Table S7). The fill factors for both the cells were 0.66 and 0.63, respectively. Although Au/Ag cell showed a slightly higher value for the fill factor (probably due to the antioxidation properties of gold), the Ag/Ag cell had a higher value for the short circuit current due to significantly higher light absorption. The calculated
■
ASSOCIATED CONTENT
S Supporting Information *
Experimental section, supporting figures, and tables. This material is available free of charge via the Internet at http:// pubs.acs.org. 1966
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
■
Letter
(25) Li, X.; Choy, W. C.; Huo, L.; Xie, F.; Sha, W. E.; Ding, B.; Guo, X.; Li, Y.; Hou, J.; You, J.; Yang, Y. Adv. Mater. 2012, 24, 3046−3052. (26) Zschech, D.; Kim, D. H.; Milenin, A. P.; Scholz, R.; Hillebrand, R.; Hawker, C. J.; Russell, T. P.; Steinhart, M.; Gosele, U. Nano Lett. 2007, 7, 1516−1520. (27) Lee, D. H.; Shin, D. O.; Lee, W. J.; Kim, S. O. Adv. Mater. 2008, 20, 2480−2485. (28) Shin, D. O.; Jeong, J. R.; Han, T. H.; Koo, C. M.; Park, H. J.; Lim, Y. T.; Kim, S. O. J. Mater. Chem. 2010, 20, 7241−7247. (29) Cheng, J. Y.; Mayes, A. M.; Ross, C. A. Nat. Mater. 2004, 3, 823−828. (30) Park, M.; Harrison, C.; Chaikin, P. M.; Register, R. A.; Adamson, D. H. Science 1997, 276, 1401−1404. (31) Park, S.; Lee, D. H.; Xu, J.; Kim, B.; Hong, S. W.; Jeong, U.; Xu, T.; Russell, T. P. Science 2009, 323, 1030−1033. (32) Spinelli, P.; Verschuuren, M. A.; Polman, A. Nat. Commun. 2012, 3, 692(1)−692(5). (33) Parker, A. R.; Townley, H. E. Nat. Nanotechnol. 2007, 2, 347− 353. (34) Zhou, W.; Tao, M.; Chen, L.; Yang, H. J. Appl. Phys. 2007, 102, 103105(1)−103105(9). (35) Baffou, G.; Quidant, R.; Girard, C. Appl. Phys. Lett. 2009, 94, 153109(1)−153109(3). (36) Nelson, J. The physics of solar cells; Imperial College Press: London, 2003. (37) Das, U. K.; Burrows, M. Z.; Lu, M.; Bowden, S.; Birkmire, R. W. Appl. Phys. Lett. 2008, 92, 063504(1)−063504(3). (38) Green, M. A. Prog. Photovoltaics 1999, 7, 327−330. (39) Cid, M.; Stem, N.; Brunetti, C.; Beloto, A. F.; Ramos, C. A. S. Surf. Coat. Technol. 1998, 106, 117−120. (40) Maier, S. A.; Atwater, H. A. J. Appl. Phys. 2005, 98, 011101(1)− 011101(10). (41) Primak, W. Appl. Opt. 1971, 10, 759−763. (42) Maier, S. A.; Kik, P. G.; Atwater, H. A. Appl. Phys. Lett. 2002, 81, 1714−1716. (43) Wei, Q. H.; Su, K. H.; Durant, S.; Zhang, X. Nano Lett. 2004, 4, 1067−1071. (44) Derkacs, D.; Lim, S. H.; Matheu, P.; Mar, W.; Yu, E. T. Appl. Phys. Lett. 2006, 89, 093103(1)−093103(3).
AUTHOR INFORMATION
Corresponding Authors
*E-mail: (A.T.) [email protected]. *E-mail: (A.B.) [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank A. Blosse for the silicon raw materials. We acknowledge financial support from the National Science Foundation (Grant CBET-1066447). A.T. acknowledges financial support from the Air Force Office of Scientific Research (AFOSR) for financial support under grant FA955011-1-0017. This work used the Lurie Nanofabrication Facility at University of Michigan, a member of the National Nanotechnology Infrastructure Network funded by the National Science Foundation.
■
REFERENCES
(1) Bagnall, D. M.; Boreland, M. Energy Policy 2008, 36, 4390−4396. (2) Ginley, D.; Green, M. A.; Collins, R. MRS Bull. 2008, 33, 355− 364. (3) Surek, T. J. Cryst. Growth 2005, 275, 292−304. (4) Lewis, N. S. Science 2007, 315, 798−801. (5) Buonassisi, T.; Istratov, A. A.; Marcus, M. A.; Lai, B.; Cai, Z. H.; Heald, S. M.; Weber, E. R. Nat. Mater. 2005, 4, 676−679. (6) Boukai, A.; Haney, P.; Katzenmeyer, A.; Gallatin, G. M.; Talin, A. A.; Yang, P. D. Chem. Phys. Lett. 2011, 501, 153−158. (7) Shir, D.; Yoon, J.; Chanda, D.; Ryu, J. H.; Rogers, J. A. Nano Lett. 2010, 10, 3041−3046. (8) Yoon, J.; Baca, A. J.; Park, S. I.; Elvikis, P.; Geddes, J. B.; Li, L. F.; Kim, R. H.; Xiao, J. L.; Wang, S. D.; Kim, T. H.; Motala, M. J.; Ahn, B. Y.; Duoss, E. B.; Lewis, J. A.; Nuzzo, R. G.; Ferreira, P. M.; Huang, Y. G.; Rockett, A.; Rogers, J. A. Nat. Mater. 2008, 7, 907−915. (9) Kwon, J. Y.; Lee, D. H.; Chitambar, M.; Maldonado, S.; Tuteja, A.; Boukai, A. Nano Lett. 2012, 12, 5143−5147. (10) De Wolf, S.; Szlufcik, J.; Delannoy, Y.; Perichaud, I.; Hassler, C.; Einhaus, R. Sol. Energy Mater. Sol. Cells 2002, 72, 49−58. (11) Kwapil, W.; Wagner, M.; Schubert, M. C.; Warta, W. J. Appl. Phys. 2010, 108, 023708(1)−023708(6). (12) Yu, K. J.; Gao, L.; Park, J. S.; Lee, Y. R.; Corcoran, C. J.; Nuzzo, R. G.; Chanda, D.; Roger, J. A. Adv. Energy Mater. 2013, 3, 1401− 1406. (13) Jeong, S.; McGehee, M. D.; Cui, Y. Nat. Commun. 2013, 4, 2950(1)−2950(7). (14) Zhao, J.; Green, M. A. IEEE Trans. Electron Devices 1991, 38, 1925−1934. (15) Richards, B. S. Prog. Photovoltaics 2004, 12, 253−281. (16) Tian, B. Z.; Zheng, X. L.; Kempa, T. J.; Fang, Y.; Yu, N. F.; Yu, G. H.; Huang, J. L.; Lieber, C. M. Nature 2007, 449, 885−890. (17) Hu, L.; Chen, G. Nano Lett. 2007, 7, 3249−3252. (18) Zhu, J.; Yu, Z.; Burkhard, G. F.; Hsu, C. M.; Connor, S. T.; Xu, Y.; Wang, Q.; McGehee, M.; Fan, S.; Cui, Y. Nano Lett. 2009, 9, 279− 282. (19) Inomata, Y.; Fukui, K.; Shirasawa, K. Sol. Energy Mater. Sol. Cells 1997, 48, 237−242. (20) Blakers, A. W.; Wang, A.; Milne, A. M.; Zhao, J.; Green, M. A. Appl. Phys. Lett. 1989, 55, 1363−1365. (21) Haase, C.; Stiebig, H. Appl. Phys. Lett. 2007, 91, 061116(1)− 06116(3). (22) Barnes, W. L.; Dereux, A.; Ebbesen, T. W. Nature 2003, 424, 824−830. (23) Atwater, H. A.; Polman, A. Nat. Mater. 2010, 9, 205−213. (24) Ding, I.-K.; Zhu, J.; Cai, W.; Moon, S.; Cai, N.; Wang, P.; Zakeeruddin, S. M.; Gratzel, M.; Brongersma, M. L.; Cui, Y.; McGehee, M. D. Adv. Energy Mater. 2011, 1, 52−57. 1967
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Extreme Light Absorption by Multiple Plasmonic Layers on Upgraded Metallurgical Grade Silicon Solar Cells Duck Hyun Lee,† Jae Young Kwon,† Stephen Maldonado,‡ Anish Tuteja,*,† and Akram Boukai*,† †
Department of Materials Science and Engineering and ‡Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109, United States S Supporting Information *
ABSTRACT: We fabricate high-efficiency, ultrathin (∼12 μm), flexible, upgraded metallurgical-grade polycrystalline silicon solar cells with multiple plasmonic layers precisely positioned on top of the cell to dramatically increase light absorption. This scalable approach increases the optical absorptivity of our solar cells over a broad range of wavelengths, and they achieve efficiencies η ≈ 11%. Detailed studies on the electrical and optical properties of the developed solar cells elucidate the light absorption contribution of each individual plasmonic layer. Finite-difference time-domain simulations were also performed to yield further insights into the obtained results. We anticipate that the findings from this work will provide useful design considerations for fabricating a range of different solar cell systems. KEYWORDS: Upgraded metallurgical-grade silicon, flexible solar cell, surface plasmons, block copolymer nanolithography
S
surface plasmon resonance is a particularly exciting methodology for achieving dramatically higher light absorption in ultrathin solar cells. A plasmon is the collective oscillation of conduction electrons stimulated by incident light at the interface between a metal (Ag, Au, Pt) and a dielectric. Both the shape and the size of the metal are key factors in determining the coupling between the metal particle and the dielectric, and thereby the overall enhancement in the optical absorption efficiency caused by surface plasmonic resonance.23 Further, precise three-dimensional placement of plasmonic metal particles can also be used to effectively concentrate light on top of the solar cell, and thereby increase its conversion efficiency.25 Indeed, it has been recently shown that metallic plasmonic nanoparticles can dramatically increase the short circuit current in silicon solar cells.9,23 One methodology for enabling, high-throughput, large-scale assembly of an array of plasmonic metallic nanoparticles on top of a solar cell is block copolymer nanolithography.26−31 In this work, we fabricate multiple plasmonic layers (either double or quadruple plasmonic nanoparticle layers) on top of ultrathin UMG-silicon solar cells using block copolymer nanolithography (see Supporting Information Figure S1). These plasmonic nanoparticle layers act as highly efficient light absorbing coatings that dramatically improve the optical absorption of the developed solar cells. The fabrication details for the solar cells developed here are provided as Supporting Information. Briefly, the 12 μm thin UMG-silicon solar cells were prepared by polishing an UMG-
olar cells must generate electricity at a cost 100 μm). Therefore, improving the efficiency of an ultrathin, indirect bandgap solar cell requires engineering of the cell surface so that the light absorptance can be significantly increased. Several methods have been investigated thus far to increase the light absorptance of thin solar cells, including dielectric antireflection coatings,14,15 surface texturing,16−21 and utilizing surface plasmon resonance.7,22−24 Among these approaches, © 2014 American Chemical Society
Received: December 27, 2013 Revised: February 24, 2014 Published: March 10, 2014 1961
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 1. (a) Optical image of a 12 μm thick UMG-silicon. Because of their thickness, UMG-silicon solar cells have a bending radius as low as ∼10 mm. (b,c) Top (b) and tilted (c) scanning electron microscopy (SEM) images of a nanoporous UMG-silicon substrate patterned using block copolymer lithography. (d,e) Top SEM images of nanoporous UMG-silicon after depositing a 5 nm silver film before (d) and after thermal annealing at 100 °C (e). (f) Representative current density−voltage (J−V) curves of the 12 μm thick UMG-silicon solar cells without (blue) and with (red) silver-based quadruple plasmonic layer under light (100 mW/cm2).
enhanced antireflection and to serve as a substrate for the deposition of two additional plasmonic layers. The two top plasmonic layers were fabricated on top of the spin on glass film using the same procedure as was used earlier to fabricate the bottom two plasmonic layers. The final structure that incorporates the bottom two plasmonic layers, the spin on glass, and the top two plasmonic layers, constitutes our “quadruple plasmonic layer” structure (see Supporting Information Figure S2a). A bare ultrathin silicon substrate has extremely poor light absorption, especially at longer wavelengths (λ > 600 nm). This severely limits the ultimate solar cell performance and efficiency (see Supporting Information Figure S3). However, the fabricated quadruple plasmonic layer dramatically increases the effective optical thickness of the ultrathin silicon substrate. Compared to the thin UMG-silicon substrate, the presence of silver-based quadruple plasmonic layer greatly decreases the reflection of light from the sample surface. As a consequence, the color of the cell becomes significantly darker after the incorporation of the quadruple plasmonic layer (see Supporting Information Figure S4). Figure 1f shows the J−V measurements for a 12 μm thick bare UMG-silicon cell, as well as a similar cell coated with the quadruple plasmonic layer under 100 mW/cm2 (AM 1.5) of light at room temperature. The short circuit current densities (JSC) for the UMG-silicon cell without and with the quadruple plasmonic layer were measured to be 9.53 and 31.16 mA/cm2, respectively. The corresponding solar cell conversion efficiencies (η) were measured to be 2.59 and 10.73%, respectively (see Supporting Information Table S2). Thus the quadruple plasmonic layer increases the efficiency of the thin film UMG-silicon solar cells by over 300%. To better understand the contribution of each constituent within the fabricated solar cells, we also conducted detailed device performance evaluation for six different solar cells: (i) A
silicon (99.999% purity) and subsequently doping phosphorus and boron on the top and bottom of the substrate respectively to obtain an n−i−p structure. The as-fabricated ultrathin solar cells were extremely flexible (bending radius as low as 10 mm) due to their thinness (Figure 1a). To fabricate the multiple plasmonic metallic nanoparticle layers on top of the UMGsilicon substrate we utilized a polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA, molecular weight 46k-b-21k) block copolymer. The block copolymer was spin-coated on top of the UMG-silicon substrate and thermally annealed to yield perpendicular self-assembled domains of cylinder forming PMMA block within a PS matrix (see Supporting Information Figure S1). The PMMA domains were then selectively removed, leaving behind a nanoporous polystyrene matrix. Reactive ion etching (SF6/C4F8) was then used to etch into the silicon substrate using the nanoporous polystyrene template as an etching mask. This procedure duplicates the hexagonal arrangement of the block-copolymer template and produces a highly dense and uniform nanoporous array on the surface of the UMG-silicon solar cells (Figure 1b). The diameter and depth of the etched nanopores were maintained at 20 nm, and the pores had a hemiellipsoidal shape (Figure 1c). For the preparation of plasmonic nanoparticles, a metallic (either silver or gold) thin film (t = 5 nm) was deposited over the prepared nanoporous UMG-silicon cells (Figure 1d). The cells were subsequently annealed on a hot plate (T = 100 °C). This caused the deposited metal thin film to dewet and form nanoparticles both within the nanopores (1st plasmonic layer), as well as, outside the nanopores (2nd plasmonic layer). The metal nanoparticles had mean diameters of 16 and 18 nm inside and outside the nanopores, respectively (Figure 1e). These two plasmonic layers together constituted the bottom “double plasmonic layer” on our UMG-silicon solar cell. Next, spin on glass was coated on top of the double plasmonic layer for 1962
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 2. (a) J−V curves for bare UMG-silicon (black), nanoporous UMG-silicon (blue), silver-based double plasmonic layer on UMG-silicon (green), silver-based double plasmonic layer (DPL) along with a top coating of spin on glass (green), UMG-silicon with a silver-based quadruple plasmonic layer (red), and double nanoporous layer (brown). Note the thickness of UMG-silicon is 180 μm. (b) Absorptance spectra for bare UMGsilicon (black), silver-based double plasmonic layer on UMG-silicon (green), and silver-based quadruple plasmonic layer on UMG-silicon (red). (c) Quantum efficiency curves for bare UMG-silicon (black), double nanoporous layer (blue), silver-based double plasmonic layer on UMG-silicon (green), and silver-based quadruple plasmonic layer on UMG-silicon (red). (d) VOC and FF for each sample. (e−g), Power absorption profiles calculated by finite difference time domain (FDTD) model under an incident light of 600 nm along the x−z plane of bare silicon (e), silver-based double plasmonic layer on silicon (f), and silver-based quadruple plasmonic layer on silicon (g). Note the incident light is propagating along the zaxis and is polarized along the x-axis. White dotted lines show surfaces of silicon (near 0 nm) and SiO2 (near 100 nm).
UMG-silicon with the double plasmonic layer, UMG-silicon with the double plasmonic layer and spin on glass, UMG-silicon with the quadruple plasmonic layer, as well as, double nanoporous layer were measured to be 13.67, 15.61, 21.53, 28.27, 34.23, and 20.71 mA/cm2, respectively (see Supporting Information Table S3). Compared with bulk UMG-silicon cell (JSC ∼ 13.67 mA/cm2), the presence of bulk spin on glass layer increased the JSC to 19.17 mA/cm2 due to the antireflection effect (also see Supporting Information Figure S5a).14,15 The nanoporous structures on UMG-silicon cell and the double nanoporous layer slightly increased the JSC to 15.61 and 20.71 mA/cm2, respectively, due to the reduced reflection caused by increased light scattering.32,33 Compared with the nanoporous UMG-silicon (JSC ∼ 15.61 mA/cm2) and double nanoporous layer (JSC ∼ 20.71 mA/cm2), the double plasmonic layer and the quadruple plasmonic layer increased the JSC to 20.71 and
bare UMG-silicon solar cell without any treatment, (ii) a nanoporous UMG-silicon solar cell (nanopores etched using block copolymer nanolithography), (iii) a UMG-silicon solar cell with a silver-based double plasmonic layer, (iv) a UMGsilicon solar cell with a silver-based double plasmonic layer along with a top coating of spin on glass, (v) a UMG-silicon solar cell with a silver-based quadruple plasmonic layer, and (vi) a double nanoporous layer. Please note that in each case the thickness of the UMG-silicon film was 180 μm, and they were doped under the same conditions as the 12 μm cells. This large thickness ensured that there were no transmission losses in all of our measurements (see Supporting Information Figure S3a). The thicknesses of the spin on glass film and the deposited silver thin film were 100 and 5 nm, respectively. Figure 2a shows the J−V curves for the six different solar cells. The JSC for bare UMG-silicon, nanoporous UMG-silicon, 1963
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 3. (a) J−V curves for the silver-based quadruple plasmonic layer on a 180 μm thick UMG-silicon solar cell as a function of spin on glass thickness. Zero nanometers of spin on glass data indicates silver-based double plasmonic layer on UMG-silicon cell. (b) J−V curves for the silverbased quadruple plasmonic layer on a 180 μm thick UMG-silicon solar cell as a function of the deposited silver layer thickness. Zero nanometers of silver data indicates porous UMG-silicon without any plasmonic layer.
34.23 mA/cm2, respectively. These measurements clearly show that the presence of the bottom double plasmonic layer, spin on glass, and the top double plasmonic layer significantly increased the JSC. This increase of the short circuit current is attributed to the absorptance enhancement caused by both the light scattering caused by each plasmonic layer, and the antireflection properties of spin on glass.34 To better elucidate the effects of the individual plasmonic layers on the absorptance of our solar cells, we measured the reflectance (R) as a function of wavelength in the range 300 nm ≤ λ ≤ 950 nm for the bare UMG-silicon cell, UMG-silicon cell with the double plasmonic layer, and the UMG-silicon cell with the quadruple plasmonic layer. Figure 2b shows the calculated absorptance curves for the three cells. The absorptance was calculated by subtracting the reflectance values from unity (absorptance = 1 − R). The double plasmonic layer dramatically increased the absorptance of UMG-silicon cell for all wavelengths >400 nm, a result that is consistent with previous work on plasmonic absorption by silver nanoparticles.9,28 In contrast to the silver-based double plasmonic layer, the quadruple plasmonic layer has an additional spin on glass antireflective layer, and a top silver-based double plasmonic layer. The absorptance for the quadruple plasmonic layer increases to ∼98% across the entire wavelength spectrum, making the cell a nearly perfect light absorber. We also performed quantum efficiency measurements in the visible and near IR wavelengths on the UMG-silicon cells. Figure 2c shows the quantum efficiency curves for bare UMG-silicon (black), double nanoporous layer (blue, see Supporting Information Figure S5), silver-based double plasmonic layer on UMGsilicon (green) and silver-based quadruple plasmonic layer on UMG-silicon (red). At a wavelength λ ∼ 800 nm, the bare UMG-silicon cell exhibits ∼41% quantum efficiency, whereas the quantum efficiency for the quadruple plasmonic cell reaches ∼90%, which agrees with the significant increase in JSC caused by silver nanoparticle based quadruple plasmonic layer. To further elucidate the absorptance enhancement effects of the quadruple plasmonic layer on silicon, the electromagnetic field intensities and the power absorption profiles for bare silicon, the double plasmonic layer on silicon, and the quadruple plasmonic layer on silicon were calculated using a finite different time domain (FDTD) method (Lumerical Solutions Inc.). The electromagnetic field calculations also clearly show strong light absorption at the interface between the silver nanoparticles and silicon (see Supporting Information Figure S6 and S7). The total power absorption profile in the
silicon is also greatly enhanced with the presence of both the silver-based double plasmonic layer (Figure 2f) and the quadruple plasmonic layer (Figure 2g). Note it is well-known that plasmon resonance increases light absorption by the nanoparticles, resulting in heating.35 However, the volume integrated total light absorption curves show that the nanoparticle light absorption is negligible, and that most of the light is absorbed by silicon (see Supporting Information Figure S7g). We also compared the effects of the first and second silver-based double plasmonic layers to better understand their individual contributions (see Supporting Information Figure S8). The silver nanoparticles from the first plasmonic layer within the double plasmonic layer were buried within UMG-silicon, which allows for a large interfacial area between the particles and silicon. Therefore, the first plasmonic layer showed more light absorption near the surface of the silver nanoparticles when compared with the second plasmonic layer, which had a much lower interfacial contact area with the substrate. The use of the plasmonic light absorption layers and antireflection layer affects the open circuit voltage (VOC) and fill factor (FF) as well. Surface recombination of carriers results in a decrease of VOC and FF, especially in thin film materials.36−38 The front surface of the cell corresponds to the highest carrier generation region in the solar cell since most of the incident light is absorbed there. Decreasing surface recombination is typically accomplished by reducing the number of dangling silicon bonds at the front surface using a surface passivation layer.36 Figure 2d shows the VOC and FF for the same six samples as Figure 2a. The data shows that the surface passivation caused by the spin-on-glass and the silver nanoparticles may provide a slight increase in both the VOC and FF. Overall, due to the absorption enhancement and surface passivation, the 180 μm thick UMG-silicon cell with the quadruple plasmonic layer showed a very high efficiency of η ∼ 11.5%. To maximize the light concentration and absorption due to the quadruple plasmonic layer, the three-dimensional placement of plasmonic nanoparticles and the surface passivation layer are critical. Figure 3a shows the J−V curves for the silverbased quadruple plasmonic solar cells as a function of spin on glass thickness. The thickness of the deposited silver thin film was fixed at 5 nm. The data shows that the spin on glass film with a thickness of 100 nm leads to the highest enhancement in JSC. When the thickness of spin on glass is higher than 100 nm, the JSC decreased (also see Supporting Information Table S4). 1964
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 4. (a) Schematic illustrations of bare UMG-silicon, spin on glass on UMG-silicon, top silver double plasmonic layer on spin on glass film, and bottom silver double plasmonic layer under spin on glass film. (b,c) J−V curves (b) and efficiency plots (c) for each sample. Note the thickness of UMG-silicon is 180 μm. (d−f) Power absorption profiles calculated by FDTD model under an incident light of 700 nm along the x−z plane of SiO2 (100 nm) on silicon (d), top silver double plasmonic layer (e), and bottom silver double plasmonic layer (f) on silicon. White dotted lines show surfaces of silicon (near 0 nm) and SiO2 (near 100 nm).
This observation of optimal spin on glass condition of ∼100 nm agrees well with previously reported results.14,39 Figure 3b presents the J−V curves for the silver-based quadruple plasmonic solar cells as a function of silver thin film thickness. The thickness of spin on glass was fixed at 100 nm. The JSC for the silver-based quadruple plasmonic solar cells varied significantly with the thickness of silver thin film. The best JSC of 34.23 mA/cm2 was attained at a silver thickness of 5 nm. The JSC decreased with further increase of silver thickness, for example, a 15 nm thin silver film decreased the JSC to 7.09 mA/cm2 (see Supporting Information Table S5), which is much lower than the bare UMG-silicon solar cell. This variation is likely because with increasing thickness, it becomes difficult to cause the dewetting of the deposited silver film.40 To further differentiate between the light absorption effects of the silver nanoparticles and the antireflection effects of the spin on glass film, we fabricated and measured the properties of four different solar cells: (i) a bare UMG-silicon cell, (ii) a UMG-silicon cell with a spin on glass layer, (iii) a UMG-silicon cell with a spin on glass layer, covered on top with a silverbased double plasmonic layer and (iv) a UMG-silicon cell with a silver-based double plasmonic layer covered on top with a spin on glass film (Figure 4a). Figure 4b,c shows the J−V curves and efficiency measurements for the four cells. Comparing across the obtained data, it is clear that the surface
passivation from the spin on glass film enhanced both the VOC (∼0.53) and FF (∼0.63) for the UMG-silicon cell (also see Supporting Information Table S6). Further, the absorptance enhancement caused by the spin-on-glass film increased JSC to 19.17 mA/cm2, resulting in an overall efficiency increase of 2.4%. The addition of the top silver-based double plasmonic layer (cell iii) and the bottom silver-based double plasmonic layer (cell iv) increased the JSC to 25.07 and 28.28, mA/cm2 respectively. The change of VOC (∼0.53) and FF (∼0.63) among the three cells (ii, iii, and iv) was negligible. Thus the top and bottom silver-based double plasmonic layers increased the absolute efficiency of the cells by ∼2.1 and ∼3.2%, respectively. Note that the total efficiency improvements caused by the bottom silver-based double plasmonic layer (3.20%), spin on glass (2.40%), and top silver-based double plasmonic layer (2.10%) add up to essentially the overall efficiency enhancement caused by the quadruple plasmonic layer on UMG-silicon (7.50%). This trend also agrees with the calculated power absorption profiles of SiO2 on silicon (Figure 4d), top silver-based double plasmonic layer on silicon (Figure 4e), and bottom silver-based double plasmonic layer on silicon (Figure 4f). The larger increase in the magnitude of JSC caused by the bottom double plasmonic layer when compared with the top double plasmonic layer can be explained by the difference in their respective environments. The bottom silver double 1965
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
Letter
Figure 5. (a,b) J−V curves (a), as well as FF and efficiency plots (b) for bare UMG-silicon, gold/silver quadruple plasmonic layers, and silver/silver quadruple plasmonic layer. Note the thickness of UMG-silicon is 180 μm. (c,d) Field-intensity profiles calculated by FDTD model under an incident light of 400 nm along the x−z plane of gold/silver quadruple plasmonic layer (c) and silver/silver quadruple plasmonic layer (d) on silicon. White dotted lines show surfaces of silicon (near 0 nm) and SiO2 (near 100 nm).
electromagnetic field intensity maps for the Au/Ag cell (Figure 5c) and the Ag/Ag cell (Figure 5d) clearly show the difference in light absorption intensities between the silver and gold nanoparticles. The power absorption profiles (see Supporting Information Figure S9) also support this hypothesis. The results from this work illustrate that multiple plasmonic layers composed of double or quadruple metallic plasmonic nanoparticle layers can be used to fabricate ultrathin solar cells with nearly perfect light absorption. Our quadruple plasmonic layers were fabricated on ultrathin (12 μm), low-cost UMGsilicon substrates and exhibit significantly enhanced JSC, VOC, and FF over bulk cells, resulting in overall solar cell efficiencies of almost 11%. The developed cells fabricated using scalable manufacturing techniques are flexible and use relatively impure, ultrathin silicon, thereby potentially addressing both the material and installation costs for silicon based solar cells. Systematic studies on the optical and electrical properties of the quadruple plasmonic layers elucidate the light absorption effects of each individual component within the developed module. FDTD calculations help explain the mechanisms behind the enhanced absorption caused by each double plasmonic layer and the spin on glass. The results reported here can provide useful design considerations for future work on upgraded metallurgical grade silicon cells, as well as, a range of other classes of solar cells.
plasmonic layer is surrounded by higher refractive index materials, silicon (refractive index, n ∼ 3.5) and SiO2 (n ∼ 1.5), which prevent significant dispersion of light.41 In comparison, the top silver-based double plasmonic layer is surrounded by air (n ∼ 1) and SiO2 (also see Supporting Information Figure S8). The larger mismatch in refractive indices leads to increased scattering at the SiO2/Si interface, which results in a larger increase of JSC. Besides silver, other metallic nanoparticles can also be used to fabricate effective plasmonic light absorpting layers. Gold nanoparticles have been previously considered as good plasmonic materials.42−44 We compared the effects of utilizing silver and gold nanoparticle layers within the quadruple plasmonic layer on the overall efficiency of the UMG-silicon solar cells. Figure 5a,b shows the J−V curves, FF, and overall efficiency plots for two different UMG-silicon cells with quadruple plasmonic layers. The first cell (denoted as the Au/Ag cell) has a bottom double plasmonic layer composed of gold nanoparticles, a spin on glass layer, and a top double plasmonic layer composed of silver nanoparticles. As silver nanoparticles are known to have higher light absorption efficiency than gold nanoparticles in the short wavelength regime,28 we used silver nanoparticles for the top double plasmonic layer. This is because the top plasmonic layer typically absorbs the short wavelength light. For the second cell (denoted as the Ag/Ag cell), both the bottom and top double plasmonic layers are composed of silver. The JSC for Au/Ag cell and the Ag/Ag cell were measured to be 24.82 and 34.23 mA/ cm2, respectively (see Supporting Information Table S7). The fill factors for both the cells were 0.66 and 0.63, respectively. Although Au/Ag cell showed a slightly higher value for the fill factor (probably due to the antioxidation properties of gold), the Ag/Ag cell had a higher value for the short circuit current due to significantly higher light absorption. The calculated
■
ASSOCIATED CONTENT
S Supporting Information *
Experimental section, supporting figures, and tables. This material is available free of charge via the Internet at http:// pubs.acs.org. 1966
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
Nano Letters
■
Letter
(25) Li, X.; Choy, W. C.; Huo, L.; Xie, F.; Sha, W. E.; Ding, B.; Guo, X.; Li, Y.; Hou, J.; You, J.; Yang, Y. Adv. Mater. 2012, 24, 3046−3052. (26) Zschech, D.; Kim, D. H.; Milenin, A. P.; Scholz, R.; Hillebrand, R.; Hawker, C. J.; Russell, T. P.; Steinhart, M.; Gosele, U. Nano Lett. 2007, 7, 1516−1520. (27) Lee, D. H.; Shin, D. O.; Lee, W. J.; Kim, S. O. Adv. Mater. 2008, 20, 2480−2485. (28) Shin, D. O.; Jeong, J. R.; Han, T. H.; Koo, C. M.; Park, H. J.; Lim, Y. T.; Kim, S. O. J. Mater. Chem. 2010, 20, 7241−7247. (29) Cheng, J. Y.; Mayes, A. M.; Ross, C. A. Nat. Mater. 2004, 3, 823−828. (30) Park, M.; Harrison, C.; Chaikin, P. M.; Register, R. A.; Adamson, D. H. Science 1997, 276, 1401−1404. (31) Park, S.; Lee, D. H.; Xu, J.; Kim, B.; Hong, S. W.; Jeong, U.; Xu, T.; Russell, T. P. Science 2009, 323, 1030−1033. (32) Spinelli, P.; Verschuuren, M. A.; Polman, A. Nat. Commun. 2012, 3, 692(1)−692(5). (33) Parker, A. R.; Townley, H. E. Nat. Nanotechnol. 2007, 2, 347− 353. (34) Zhou, W.; Tao, M.; Chen, L.; Yang, H. J. Appl. Phys. 2007, 102, 103105(1)−103105(9). (35) Baffou, G.; Quidant, R.; Girard, C. Appl. Phys. Lett. 2009, 94, 153109(1)−153109(3). (36) Nelson, J. The physics of solar cells; Imperial College Press: London, 2003. (37) Das, U. K.; Burrows, M. Z.; Lu, M.; Bowden, S.; Birkmire, R. W. Appl. Phys. Lett. 2008, 92, 063504(1)−063504(3). (38) Green, M. A. Prog. Photovoltaics 1999, 7, 327−330. (39) Cid, M.; Stem, N.; Brunetti, C.; Beloto, A. F.; Ramos, C. A. S. Surf. Coat. Technol. 1998, 106, 117−120. (40) Maier, S. A.; Atwater, H. A. J. Appl. Phys. 2005, 98, 011101(1)− 011101(10). (41) Primak, W. Appl. Opt. 1971, 10, 759−763. (42) Maier, S. A.; Kik, P. G.; Atwater, H. A. Appl. Phys. Lett. 2002, 81, 1714−1716. (43) Wei, Q. H.; Su, K. H.; Durant, S.; Zhang, X. Nano Lett. 2004, 4, 1067−1071. (44) Derkacs, D.; Lim, S. H.; Matheu, P.; Mar, W.; Yu, E. T. Appl. Phys. Lett. 2006, 89, 093103(1)−093103(3).
AUTHOR INFORMATION
Corresponding Authors
*E-mail: (A.T.) [email protected]. *E-mail: (A.B.) [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank A. Blosse for the silicon raw materials. We acknowledge financial support from the National Science Foundation (Grant CBET-1066447). A.T. acknowledges financial support from the Air Force Office of Scientific Research (AFOSR) for financial support under grant FA955011-1-0017. This work used the Lurie Nanofabrication Facility at University of Michigan, a member of the National Nanotechnology Infrastructure Network funded by the National Science Foundation.
■
REFERENCES
(1) Bagnall, D. M.; Boreland, M. Energy Policy 2008, 36, 4390−4396. (2) Ginley, D.; Green, M. A.; Collins, R. MRS Bull. 2008, 33, 355− 364. (3) Surek, T. J. Cryst. Growth 2005, 275, 292−304. (4) Lewis, N. S. Science 2007, 315, 798−801. (5) Buonassisi, T.; Istratov, A. A.; Marcus, M. A.; Lai, B.; Cai, Z. H.; Heald, S. M.; Weber, E. R. Nat. Mater. 2005, 4, 676−679. (6) Boukai, A.; Haney, P.; Katzenmeyer, A.; Gallatin, G. M.; Talin, A. A.; Yang, P. D. Chem. Phys. Lett. 2011, 501, 153−158. (7) Shir, D.; Yoon, J.; Chanda, D.; Ryu, J. H.; Rogers, J. A. Nano Lett. 2010, 10, 3041−3046. (8) Yoon, J.; Baca, A. J.; Park, S. I.; Elvikis, P.; Geddes, J. B.; Li, L. F.; Kim, R. H.; Xiao, J. L.; Wang, S. D.; Kim, T. H.; Motala, M. J.; Ahn, B. Y.; Duoss, E. B.; Lewis, J. A.; Nuzzo, R. G.; Ferreira, P. M.; Huang, Y. G.; Rockett, A.; Rogers, J. A. Nat. Mater. 2008, 7, 907−915. (9) Kwon, J. Y.; Lee, D. H.; Chitambar, M.; Maldonado, S.; Tuteja, A.; Boukai, A. Nano Lett. 2012, 12, 5143−5147. (10) De Wolf, S.; Szlufcik, J.; Delannoy, Y.; Perichaud, I.; Hassler, C.; Einhaus, R. Sol. Energy Mater. Sol. Cells 2002, 72, 49−58. (11) Kwapil, W.; Wagner, M.; Schubert, M. C.; Warta, W. J. Appl. Phys. 2010, 108, 023708(1)−023708(6). (12) Yu, K. J.; Gao, L.; Park, J. S.; Lee, Y. R.; Corcoran, C. J.; Nuzzo, R. G.; Chanda, D.; Roger, J. A. Adv. Energy Mater. 2013, 3, 1401− 1406. (13) Jeong, S.; McGehee, M. D.; Cui, Y. Nat. Commun. 2013, 4, 2950(1)−2950(7). (14) Zhao, J.; Green, M. A. IEEE Trans. Electron Devices 1991, 38, 1925−1934. (15) Richards, B. S. Prog. Photovoltaics 2004, 12, 253−281. (16) Tian, B. Z.; Zheng, X. L.; Kempa, T. J.; Fang, Y.; Yu, N. F.; Yu, G. H.; Huang, J. L.; Lieber, C. M. Nature 2007, 449, 885−890. (17) Hu, L.; Chen, G. Nano Lett. 2007, 7, 3249−3252. (18) Zhu, J.; Yu, Z.; Burkhard, G. F.; Hsu, C. M.; Connor, S. T.; Xu, Y.; Wang, Q.; McGehee, M.; Fan, S.; Cui, Y. Nano Lett. 2009, 9, 279− 282. (19) Inomata, Y.; Fukui, K.; Shirasawa, K. Sol. Energy Mater. Sol. Cells 1997, 48, 237−242. (20) Blakers, A. W.; Wang, A.; Milne, A. M.; Zhao, J.; Green, M. A. Appl. Phys. Lett. 1989, 55, 1363−1365. (21) Haase, C.; Stiebig, H. Appl. Phys. Lett. 2007, 91, 061116(1)− 06116(3). (22) Barnes, W. L.; Dereux, A.; Ebbesen, T. W. Nature 2003, 424, 824−830. (23) Atwater, H. A.; Polman, A. Nat. Mater. 2010, 9, 205−213. (24) Ding, I.-K.; Zhu, J.; Cai, W.; Moon, S.; Cai, N.; Wang, P.; Zakeeruddin, S. M.; Gratzel, M.; Brongersma, M. L.; Cui, Y.; McGehee, M. D. Adv. Energy Mater. 2011, 1, 52−57. 1967
dx.doi.org/10.1021/nl4048064 | Nano Lett. 2014, 14, 1961−1967
