Solar Energy Materials & Solar Cells

Transcription

1 Solar Energy Materials & Solar Cells 112 (2013) Contents lists available at SciVerse ScienceDirect Solar Energy Materials & Solar Cells journal homepage: Letter Broadband near-field effects for improved thin film Si solar cells on randomly textured substrates Lucia V. Mercaldo n, Paola Delli Veneri, Iurie Usatii, Tiziana Polichetti ENEA Portici Research Center, P.le E. Fermi 1, Portici (Napoli), Italy article info Article history: Received 12 October 2012 Received in revised form 8 January 2013 Accepted 14 January 2013 Keywords: Thin film solar cells a-si:h/mc-si:h tandem cells Silicon oxide n-layer Back-reflector Near-field effects abstract We show the superior performance in terms of light management of n-doped mixed phase silicon oxide in direct contact with Ag with respect to the established ZnO/Ag back reflector in thin film Si tandem solar cells on randomly textured substrates. Based on coexisting larger quantum efficiency and amplification of the optical Raman mode of Si, we propose that, in spite of a simple cell design, near-field concentrated light radiated at metal nanoprotrusions on the back interface is harvested. The effect is a remarkable 10% increase of the generated current density in the bottom cell, holding promise for significant conversion efficiency gain in matched devices. & 2013 Elsevier B.V. All rights reserved. 1. Introduction Thin film silicon solar cell technology, with its potential to reduce material and fabrication costs, is considered as a promising approach for a large-scale deployment of photovoltaics. When using thin photon absorbers, however, light trapping is crucial to realize high efficiency solar cells. In prospect, advanced threedimensional designs and photon management strategies based on properly engineered photonic and plasmonic nanostructures have then gained tremendous interest [1 6]. Presently, for a broadband optical enhancement the cells rely on light scattering and diffraction at randomly textured interfaces, for which different morphologies and scattering characteristics are being explored [2,7 12], combined with a back reflector. The pyramidal morphology in particular, also employed in large-area module production, has demonstrated outstanding light trapping capabilities until now leading to certified world-record conversion efficiencies [13]. Fig. 1 shows a common architecture for the micromorph tandem case. This tandem cell, one of the more efficient thin film Si cell concepts, is a series of a high-gap (1.75 ev) amorphous silicon (a-si:h) top cell and low-gap (1.1 ev) microcrystalline silicon (mc-si:h) bottom cell, with the two materials almost perfectly matched to efficiently share the spectral content of the sunlight [12 14]. The back reflector also plays an important role, as it has to bring back the un-absorbed light efficiently where it can still be n Corresponding author. Tel.: þ address: lucia.mercaldo@enea.it (L.V. Mercaldo). used. One of the conventional solutions makes use of a rear metal layer like Ag. In this case a low refractive index buffer layer, like doped ZnO, is generally inserted between Si and Ag to avoid losses in reflected light due to the excitation of surface plasmon polaritons (SPPs), travelling waves of surface charge density at the metal/dielectric interface. These losses fall in the spectral region where optimum reflectance is desired and the role of the spacer is to shift the SPP dispersion relation, moving the plasmon frequencies toward higher energies (essentially already absorbed in the front part of the cell) [15]. In previous work, phosphorousdoped mixed phase silicon oxide (n-sio x :H), thanks to a sufficiently low refractive index, has shown similar capabilities and it has been then proposed as an advanced n-layer for thin film Si solar cells [16,17]. This material class, with independently tunable electrical and optical properties, is demonstrating interesting light management capabilities, and various versions are now being successfully used as superior doped layers and/or reflecting layers in multijunction thin-film Si solar cells [18 20]. Here we have taken the concept introduced in Refs. [16,17] further to show the superior performance in terms of light management of the n-sio x :H/Ag combination with respect to the already established ZnO/Ag back reflector. A remarkable 10% increase of the current density generated in the bottom cell of micromorph devices is reported. Based on a concurrent Raman amplification effect, we propose that the thin n-sio x :H layer, by keeping the Si absorber layer sufficiently close to the Ag contact, permits also the absorption of near-field concentrated light re-radiated by the nanotextured metal surface due to localized phenomena [21 24]. With the appropriate materials, even with random /$ - see front matter & 2013 Elsevier B.V. All rights reserved.

2 164 L.V. Mercaldo et al. / Solar Energy Materials & Solar Cells 112 (2013) Fig. 1. (a) Schematic representation of the micromorph solar cell architecture in superstrate (p i n) configuration with randomly nanostructured front and back contacts (TCO stands for transparent conducting oxide); (b,c) AFM topographic images (image size 5 mm 5 mm and z axis 300 nm/div) of the nanoscale texture of the Asahi U substrate used in this work (b) and of the back surface of the entire Si stack (c); (d) representative profile along a randomly chosen line on the map in (c), showing the characteristic dimensions of the back contact protrusions toward the Si layers. morphologies localized phenomena could then be exploited with remarkable effect in the solar cell efficiency. Complementary Raman measurements have been carried out on the devices with a Renishaw invia Reflex Raman spectrometer. 2. Experimental details Identical superstrate-type micromorph solar cells, including n-sio x :H or standard mc-si:h n-layers (35 nm thick) in the bottom cell, were deposited by plasma-enhanced chemical vapor deposition in a lab-scale reactor on cm 2 commercial Asahi U-type glass, where randomly textured SnO 2 :F acts as transparent conducting electrode. Details on cell fabrication parameters are reported elsewhere [25,26]. Based on the previous work [17], two appropriate mixed phase n-sio x :H layers with slightly different stoichiometry were selected. The refractive index (at 800 nm) and electrical conductivity, evaluated on 100 nm thick films deposited on glass, are n¼2.6, s¼0.2 (O cm) 1 in one case and n¼2.4, s¼ (O cm) 1 in the other. TnSiO1 and TnSiO2 are the corresponding tandem solar cells employing these materials in the bottom cells, while TnSi is the tandem device with standard mc-si:h n-layer [n¼3.3 at 800 nm and s¼0.1 (O cm) 1 ]. For the back reflecting contact two classical options were considered: a single evaporated Ag layer or sputtered Al doped ZnO (80 nm thick) followed by evaporated Ag, realized on p i n/p i n Si stacks grown in the same run to allow direct comparison. In order to enlarge the spectral range where the back reflector effects are observable, reduced absorber layer thicknesses were adopted: 140 nm for the a-si:h top cell and 720 nm for the mc-si:h bottom cell, for a total device thickness of 1 mm (including all the Sibased layers). For each design, 1 1cm 2 cells were defined at the back contact realization stage by using a metal mask. External quantum efficiencies EQEs of the a-si:h top and mc-si:h bottom cells were measured under red and blue bias light illumination, respectively. The corresponding short-circuit current densities J sctop and J scbot were calculated from the EQE curves by convolution with the photon flux of the global air mass 1.5 (AM1.5 g) solar spectrum. The current density voltage J(V) characteristics, from which the open-circuit voltages V oc and the fill factors FF were determined, were measured using a dual lamp solar simulator in standard test conditions (25 1C, AM1.5 g, 1000 W/m 2 ). 3. Results and discussion The devices, whose structure is schematized in Fig. 1(a), are characterized by nanostructured front and back contacts. Topographic images of the rough surfaces acquired by Atomic force microscopy (AFM) are shown in Fig. 1(b) and (c). Since each subsequent layer is conformally deposited, the underlying pyramidal structure of the substrate shown in Fig. 1(b) (measured root mean square roughness of 32 nm) is transferred to the other interfaces with similar features [Fig. 1(c)]. The metal back contact, the negative of Fig. 1(c), is then characterized by random surface protrusions extending toward the Si stack with heights up to 100 nm and wide apex divergence angles, as visualized by the representative profile in Fig. 1(d). The performance metrics of the different cells are summarized in Table 1. For all the cells open circuit voltage V oc and fill factor FF are close to state-of the-art values, while the thin absorber layers limit the efficiencies Z in between 8.2% and 9.4%, due to the reduced short circuit current densities. With a top cell current of 10 ma/cm 2, the desired current matching between the two series-connected subcells is more or less achieved for the devices with ZnO, while with single Ag reflector TnSi is bottom limited and TnSiO1 and TnSiO2 are top limited. The share of the various wavelengths to the generated current for all the cells is shown in Fig. 2. The spectral response of the top cell, that absorbs mostly the blue and green part of the spectrum between 350 and 800 nm, is very similar in all cases, as the top cells were designed to be identical. Even if the same absorber layer thickness was considered, the contribution from the bottom cell, absorbing from 500 to 1100 nm, is instead very different due to the different management of the long wavelengths light. The known beneficial effect of the ZnO spacer in the standard cell (TnSi) is evident. A similar role is played by n-sio x :H [17], but interestingly a superior response is achieved in the absence of ZnO, while the EQE of all the cells with ZnO is similar independently of the type of n-layer. This effect has been observed for all the cells realized on the substrate and confirmed by analogous devices realized in different deposition runs. Clearly

3 L.V. Mercaldo et al. / Solar Energy Materials & Solar Cells 112 (2013) the indications are that the n-sio x :H/Ag combination in the back reflector is a better arrangement than ZnO/Ag. In terms of bottom cell short-circuit current density J scbot, TnSiO1/Ag and TnSiO2/Ag outperform TnSi/ZnO/Ag by 10% and 8%, respectively. Notably a general improvement can be obtained without having to finely tune the n-sio x :H properties, that is advantageous in module production perspective. To investigate purely optical effects of the different back reflectors we have performed micro-raman measurements on the cells with appropriate probe light. The measurements were carried out in backscattering configuration by using a 785 nm diode laser and focusing the beam through the substrate (glassþtco) on the Si stack. At this wavelength, with an absorption coefficient a cm 1 for mc-si:h (collection length 1/ 2a much larger than cell thickness) and negligible for a-si:h, the entire p i n/p i n Si stack contributes to the Raman response, including the deepest portion of the cells. For each device, the measurements have been performed in two regions: next to the cell pad (region I) and within the cell area in correspondence of the back contact pad (region II), as illustrated in Fig 3(a). Figs. 3(b) (d) show the first-order optical mode of Si at 520 cm 1 (raw spectra with no background correction) as observed in the two regions of the different tandem cells, both in case of Ag back contact (red lines) and ZnO/Ag back contact (black lines). The lineshape is a convolution of different contributions from the ordered and disordered phases within the different Si layers [25]. Differences are found in the intensity of the peak. The measurement in region I serves as a control: the same intensity observed from the Si stacks deposited in the same run and dedicated to the Ag and ZnO/Ag back contacts (in each separate panel of Fig. 3) rules out thickness disuniformities that would affect the Raman signal. In all cases, when moving from regions I to II, enhanced intensity is expected due to the Raman scattering from the extra back-reflected light. This is indeed observed. In particular, for cell TnSi the same increased intensity is detected with both Ag and ZnO/Ag back reflector [Fig. 3(b)]. However, for cells TnSiO1 and TnSiO2 [Fig. 3(c,d)] an even larger intensity is measured with the Ag back reflector, which is the case where enhanced EQE is also detected. Based on the discussed experiments on the cells, we gather that some optical effect enhances both the Raman signal and the EQE when a n-sio x :H/Ag reflector is used. Changes in the light trapping properties of the different back reflectors would affect the EQE. On the other hand, due to the geometry of the experiment and the optics of the instrument, out-of-focus diffuse reflected light might partially affect the background but not the intensity of a Raman mode. The possible role of different reflectance R into Si due to interference effects has also been considered with simulations for the n-sio x :H/Ag and n-sio x :H/ZnO/Ag combinations in the approximation of flat interfaces, using the dispersion of the optical constants determined by spectroscopic ellipsometry on single films deposited on glass or tabulated data [27]. This analysis shows slightly larger R in the presence of ZnO at the wavelengths of interest (above 600 nm). Therefore, enhanced EQE and amplified Table 1 Photovoltaic parameters of the micromorph solar cells. Cell J sctop (ma/cm 2 ) J scbot (ma/cm 2 ) V oc (V) FF (%) Z (%) With Ag TnSi TnSiO TnSiO With ZnO/Ag TnSi TnSiO TnSiO Fig. 2. External quantum efficiency (EQE) for the cells with Ag (solid lines) and ZnO/Ag (dashed lines) in case of standard mc-si:h n-layer (TnSi) and n-sio x :H layers (TnSiO1 and TnSiO2). Raman mode in absence of ZnO cannot be ascribed to improved reflectance. On the other hand, the intensity of a Raman mode may vary due to a change of the local electric field distribution, concentrating to some extent light in the Raman sensitive layer. Besides SPPs, with randomly textured Ag surfaces like those in our solar cells, localized phenomena might come into play [21 24]. Localized surface plasmons can be excited at the metal nanoprotrusions, acting as source of strong local field (light) enhancement. At the same time geometric singularities cause the electrostatic lightning-rod effect: enhanced surface charge density, and then crowding of the electric field lines develops around regions with large curvatures leading to a weakly frequencydependent near-field light concentration. It is known that, due to such enhanced fields, rough metal surfaces as well as metal nanoparticles can lead to a strong amplification of the Raman intensity of molecules adsorbed on the metal (surface enhanced raman scattering or SERS) [22,23]. Although SERS is typically done on surfaces, similar effects in the Raman signal, attributed to an increased electromagnetic field, have been already reported for p i n a-si:h solar cells with Ag nanoparticles integrated in the back side [28]. We propose that also the strong Raman signal observed in Fig. 3(c) and (d) with Ag back reflector might be an effect of enhanced Raman scattering similar to SERS in the deepest portion of the Si layers, thanks to concentrated light in the vicinity of the rough metal interface. Based on the results of the Raman experiment, we suggest that the improved photovoltaic performance of TnSiO1/Ag and TnSiO2/Ag originates from localized phenomena at the Ag interface. The crucial difference with the other cells that allows these effects to come into play is the simultaneous decoupling and close proximity of silicon absorber layer and metal contact obtained thanks to the n- SiO x :H optoelectronic properties. Once removed (at least partially) the SPP-related losses, superior spectral response of solar cells is expected if the re-radiated light, of plasmonic and/or geometrical origin, is concentrated to some extent within the intrinsic absorber layer [1]. As such fields are typically confined in close proximity of the metal surface, appropriate design is needed to give it a chance to reach the intrinsic mc-si:h layer where it can be utilized in the conversion process. We propose that this is the case of the cells TnSiO1/Ag and TnSiO2/Ag, where the intrinsic mc-si:h layer and the Ag contact are separated only by 35 nm of n-sio x :H. When ZnO is added to the structure, the near-field light is instead lost in the spacer and the EQE of the bottom subcell equals the corresponding one of the standard cell (TnSi/ZnO/Ag). On the other hand, for TnSi/

4 166 L.V. Mercaldo et al. / Solar Energy Materials & Solar Cells 112 (2013) Fig. 3. Raman analysis of the tandem solar cells with excitation at 785 nm: schematic of the experiment (a) and Raman spectra (with no background correction) around the first-order optical mode of Si measured on TnSi (b), TnSiO1 (c), and TnSiO2 (d) with Ag and ZnO/Ag back reflector next to the back contact pad (region I, thin lines) and within the cell area (region II, thick lines). The legend in panel (b) is valid also for (c) and (d). Ag, where Si and Ag are in direct contact as well, the near-field effects are not observed both in EQE and Raman measurements because hindered by the significant loss of light into SPPs excitation. In this case, even if a thinner ZnO layer is included with respect to the optimal value [15] here applied, still the need for two separate layers playing different roles (the n-layer and the dielectric spacer) would limit the possibility to bring the absorber layer close enough to the Ag nanostructures. Moreover, the optical properties of the different n-layers affect the extension of these fields. We stress that, independently of the physical origin of the discussed experiments, with our back reflector configuration the current generated in the bottom cell is significantly increased. Various mechanisms could contribute to this enhancement. Here one physical explanation valid for both the EQE and the Raman experiments has been suggested. If this picture is confirmed, even larger currents could be obtained in principle with use of a thinner n-doped layer, i.e. more concentrated light falling within the intrinsic absorber layer, but the optimal n-sio x :H thickness is delicate matter as this layer plays a multifunctional role. 4. Conclusions In conclusion, we have established that the n-sio x :H/Ag combination, with n-sio x :H functioning already as vital layer to form the p i n junction, is a better back reflector than standard ZnO/Ag in micromorph solar cells on commercial substrates with random texture. Enhanced spectral response of the bottom cell and amplified optical Raman mode of Si have been observed. Based on such concurrent effects, we have proposed that the n-sio x :H/Ag combination allows the utilization of a significant fraction of nearfield light radiated by the Ag nanoprotrusions of the back contact while simultaneously relieving the detrimental SPP-related losses. By demonstrating a 10% increase of the current density generated in the bottom cell, the present result sets a milestone on the route of design and development of an ideal back reflector for thin film Si solar cells, with straightforward application in production. Our approach opens the way for significant further efficiency improvements not only by device optimization (through careful tuning of n-sio x :H thickness and properties), but also by providing the means for exploiting the near-field light effects into Si with specifically designed metallic nanostructures. Acknowledgments A part of this work was carried out in the framework of the FP7 project Fast Track, financed by the EC under Contract number The authors would like to thank Etienne Moulin (Forschungszentrum Juelich, Germany) for helpful discussions. References [1] H.A. Atwater, A. Polman, Plasmonics for improved photovoltaic devices, Nature Materials 9 (2010) [2] C. Battaglia, J. Escarré, K. Söderström, M. Charriere, M. Despeisse, F.-J. Haug, C. Ballif, Nanomoulding of transparent zinc oxide electrodes for efficient light trapping in solar cells, Nature Photonics 5 (2011) [3] J. Bhattacharya, N. Chakravarty, S. Pattnaik, W.D. Slafer, R. Biswas, V.L. Dalal, A photonic plasmonic structure for enhancing light absorption in thin film solar cells, Applied Physics Letters 99 (2011)

5 L.V. Mercaldo et al. / Solar Energy Materials & Solar Cells 112 (2013) [4] M. Vanececk, O. Babchenko, A. Purkrt, J. Holovsky, N. Neykova, A. Poruba, Z. Remes, J. Meier, U. Kroll, Nanostructured three-dimensional thin film silicon solar cells with very high efficiency potential, Applied Physics Letters 98 (2011) [5] U.W. Paetzold, E. Moulin, D. Michaelis, W. Böttler, C. Wächter, V. Hagemann, M. Meier, R. Carius, U. Rau, Plasmonic reflection grating back contacts for microcrystalline silicon solar cells, Applied Physics Letters 99 (2011) [6] V. Ferry, M.A. Verschuuren, M.C. van Lare, R.E.I. Schropp, H.A. Atwater, A. Polman, Optimized spatial correlations for broadband light trapping nanopatterns in high efficiency ultrathin film a-si:h solar cells, Nano Letters 11 (2011) [7] C. Rockstuhl, S. Fahr, K. Bittkau, T. Beckers, R. Carius, F.-J. Haug, T. Söderström, C. Ballif, F. Lederer, Comparison and optimization of randomly textured surfaces in thin-film solar cells, Optics Express 18 (2010) A335 A342. [8] O. Isabella, J. Krč, M. Zeman, Modulated surface textures for enhanced light trapping in thin-film silicon solar cells, Applied Physics Letters 97 (2010) [9] H. Sai, H. Jia, M.J. Kondo, Impact of front and rear texture of thin-film microcrystalline silicon solar cells on their light trapping properties, Journal of Applied Physics 108 (2010) [10] J.I. Owen, J. Hüpkes, H. Zhu, E. Bunte, S.E. Pust, Novel etch process to tune crater size on magnetron sputtered ZnO:Al, Physica Status Solidi A 208 (2011) [11] K. Soderstrom, G. Bugnon, F.-J. Haug, S. Nicolay, C. Ballif, Experimental study of flat light-scattering substrates in thin-film silicon solar cells, Solar Energy Materials and Solar Cells 101 (2012) [12] M. Boccard, C. Battaglia, S. Hänni, K. Söderström, J. Escarre!, S. Nicolay, F. Meillaud, M. Despeisse, C. Ballif, Multiscale transparent electrode architecture for efficient light management and carrier collection in solar cells, Nano Letters 12 (2012) [13] J. Bailat, et al., Recent developments of high-efficiency micro-morph tandem solar cells in KAI-M PECVD reactors, in: Proceedings of the 5th World Conference on Photovoltaic Energy Conversion, Valencia, Spain, 2010, pp [14] F. Meillaud, A. Shah, C. Droz, E. Vallat-Sauvain, C. Miazza, Efficiency limits for single-junction and tandem solar cells, Solar Energy Materials and Solar Cells 90 (2006) [15] F.-J. Haug, T. Söderström, O. Cubero, V. Terrazzoni-Daudrix, C. Ballif, Plasmonic absorption in textured silver back reflectors of thin film solar cells, Journal of Applied Physics 104 (2008) [16] P. Delli Veneri, L.V. Mercaldo, I. Usatii, Silicon oxide based n-doped layer for improved performance of thin film silicon solar cells, Applied Physics Letters 97 (2010) [17] P. Delli Veneri, L.V. Mercaldo, I. Usatii, Improved micromorph solar cells by means of mixed-phase n-doped silicon oxide layers, Progress in Photovoltaics: Research and Applications (2011), online early access. [18] A. Lambertz, T. Grundler, F. Finger, Hydrogenated amorphous silicon oxide containing a microcrystalline silicon phase and usage as an intermediate reflector in thin-film silicon solar cells, Journal of Applied Physics 109 (2011) [19] P. Cuony, D.T.L. Alexander, I. Perez-Wurfl, M. Despeisse, G. Bugnon, M. Boccard, T. Söderström, A. Hessler-Wyser, C. Hébert, C. Ballif, Silicon Filaments in Silicon Oxide for Next-Generation Photovoltaics, Advanced Materials 24 (2012) [20] B. Yan, G. Yue, L. Sivec, J. Yang, S. Guha, C.-S. Jiang, Innovative dual function nc-sio x :H layer leading to a 416% efficient multi-junction thin-film silicon solar cell, Applied Physics Letters 99 (2011) [21] J.A. Schuller, E.S. Barnard, W. Cai, Y.C. Jun, J.S. White, M.L. Brongersma, Plasmonics for extreme light concentration and manipulation, Nature Materials 9 (2010) [22] T.K. Sau, A.L. Rogach, F. Jäckel, T.A. Klar, J. Feldmann, Properties and applications of colloidal nonspherical noble metal nanoparticles, Advanced Materials 22 (2010) [23] S.A. Maier, H.A. Atwater, Plasmonics: Localization and guiding of electromagnetic energy in metal/dielectric structures, Journal of Applied Physics 98 (2005) [24] S.I. Bozhevolnyi, J. Beermann, V. Coello, Direct observation of localized second-harmonic enhancement in random metal nanostructures, Physical Review Letters 90 (2003) [25] P. Delli Veneri, L.V. Mercaldo, P. Tassini, C. Privato, Correlation between structural properties and performances of microcrystalline silicon solar cells, Thin Solid Films 487 (2005) [26] P. Delli Veneri, L.V. Mercaldo, C. Privato, Deposition pressure effects on material structure and performance of micromorph tandem solar cells, Renewable Energy 33 (2008) [27] The simulation tool is the IMD package extension of the XOP driver program, freely available through the European Synchrotron Radiation Facility website. The dispersion curves have been determined experimentally for all the materials besides Ag, for which we have used tabulated data from the Handbook of Optical Constants of Solids, vol. 1, edited by E. D. Palik, Academic Press, Orlando, [28] E. Moulin, J. Sukmanowski, M. Schulte, A. Gordijn, F.X. Royer, H. Stiebig, Thinfilm silicon solar cells with integrated silver nanoparticles, Thin Solid Films 516 (2008)