- Email: [email protected]
Fabrication and characterization of nanorod solar cells with an ultrathin a-Si:H absorber layer
Journal of Non-Crystalline Solids 358 (2012) 2209–2213
Contents lists available at SciVerse ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol
Fabrication and characterization of nanorod solar cells with an ultrathin a-Si:H absorber layer Yinghuan Kuang ⁎, Karine H.M. van der Werf, Z. Silvester Houweling, Marcel Di Vece, Ruud E.I. Schropp Faculty of Science, Debye Institute for Nanomaterials Science, Nanophotonics—Physics of Devices, Utrecht University, P.O. Box 80.000, 3508 TA Utrecht, The Netherlands
a r t i c l e
i n f o
Article history: Received 12 August 2011 Received in revised form 4 November 2011 Available online 5 December 2011 Keywords: Nanorods; Thin film solar cells; Extremely thin absorber layer; Amorphous hydrogenated silicon; ZnO
a b s t r a c t In this paper, we present a three-dimensional nanorod solar cell design. As the backbone of the nanorod device, density-controlled zinc oxide (ZnO) nanorods were synthesized by a simple aqueous solution growth technique at 80 °C on ZnO thin film pre-coated glass substrate. The as-prepared ZnO nanorods were coated by an amorphous hydrogenated silicon (a-Si:H) light absorber layer to form a nanorod solar cell. The light management, current–voltage characteristics and corresponding external quantum efficiency of the solar cells were investigated. An energy conversion efficiency of 3.9% was achieved for the nanorod solar cells with an a-Si:H absorber layer thickness of 75 nm, which is significantly higher than the 2.6% and the 3.0% obtained for cells with the same a-Si:H absorber layer thickness on planar ZnO and on textured SnO2:F counterparts, respectively. A short-circuit current density of 11.6 mA/cm 2 and correspondingly, a broad external quantum efficiency profile were achieved for the nanorod device. An absorbed light fraction higher than 80% in the wavelength range of 375–675 nm was also demonstrated for the nanorod solar cells, including a peak value of ~ 90% at 520–530 nm. © 2011 Elsevier B.V. All rights reserved.
1. Introduction In the past few years, thin film silicon solar cells have attracted great attention in the photovoltaic community because of the lower production cost compared to their crystalline silicon counterparts [1-5]. In thin film devices, the thickness of the absorber layer is governed by the well-known trade-off between light absorption and minority carrier collection [6-8]. The absorber layer must be optically thick to capture a sufficient fraction of incident photons but must also be sufficiently thin to enable minority carrier collection lengths larger than the thickness of the absorber layer. In order to obtain a large collection length, high quality material is a prerequisite. Hydrogenated amorphous silicon (a-Si:H) and nanocrystalline silicon (ncSi:H) are among the most developed thin film photovoltaic materials, but suffer from small minority carrier diffusion length. One solution to relieve the trade-off between light absorption and carrier collection is the usage of an extremely thin absorber layer (ETA) on structured substrate, which can keep a significant amount of optical absorption but remarkably reduce charge carriers recombination [911]. An alternative approach is enhancing light trapping by employing a surface texture such as that of natively textured commercially available SnO2:F or as that made by hydrochloric acid etching of ZnO:Al [12,13] for superstrate p-i-n structures, or that of structured Ag back electrodes in substrate n-i-p structures [14]. In these types
⁎ Corresponding author. E-mail address: [email protected] (Y. Kuang). 0022-3093/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2011.11.021
of structures, the incident light is scattered into off-normal angles; therefore the optical path length in the absorber layer is increased and light absorption is enhanced. Recently, the application of nanostructures such as nanorod [7-10,15-17] and nanowire [18-28] in thin film solar cells has gained considerable attention as a method to eliminate the trade-off between light absorption and carrier collection. The orthogonalization of light absorption and carrier collection paths plays an important role in nanorod/nanowire systems. An ultrathin cell built on nanorods is thick enough for sufficient light absorption in the axial direction while it is sufficiently thin in the radial direction to guarantee efficient carrier extraction and collection. To obtain a conformal layer coating in nanorod/nanowire solar cells, the morphology of the nanorod/nanowire array is critical. Approaches widely employed for the preparation of such nanostructured materials are for example reactive ion etching [27,29,30] and vapor–liquid–solid methods [19,21,23,24]. However, complicated fabrication processes, area limitation, sophisticated apparatus requirements, and energy consumption in these approaches greatly hinder the practical application in solar cells. In contrast, in this work we present a simple and scalable method. ZnO nanorods are selected as the base structure since they can very easily be synthesized by hydrothermal growth at a low temperature. The morphology of these ZnO nanorods can be precisely tuned by adjusting the growth conditions such as the precursor solution concentration and the growth temperature [31]. As the solar cell absorber layer we employ a-Si:H since this is a well-developed photovoltaic material, and because it is non-toxic and source materials are abundantly available.
2210
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
2. Experimental A flat ZnO layer was sputtered onto a glass substrate as a seed layer to favor nucleation and epitaxial growth of ZnO nanorods. A mixture of 0.0005 mol zinc acetate dihydrate (Zn(CH3COO)2·2H2O, SigmaAldrich) with an equal amount of 0.0005 mol hexamethylenetetramine (HMT, Sigma-Aldrich) was dissolved in 1 liter de-ionized water as the reactive solution. The solution was magnetically stirred for 15 minutes at room temperature. After that substrates were immersed in the solution, holding the pre-coated ZnO seed layer downwards. The growth temperature was kept constant at 80 °C during 3 hours. After growth, the substrates were thoroughly rinsed by de-ionized water and dried by blowing nitrogen. The as-grown nanorods were overcoated with an Ag layer and a ZnO:2%Al transparent conductive oxide (TCO) layer by sputtering. Then, the deposition of a n-i-p layer stack was carried out in a multi-chamber deposition system using hot-wire chemical vapor deposition (HWCVD). The deposition of the solar cells was described elsewhere [15]. An array of 4 mm × 4 mm squares of a transparent conducting indium tin oxide (ITO) layer was sputtered through a mask to define individual cell pads. Gold top-grid contacts were evaporated onto the square ITO pads using another mask, leaving an active cell area of 0.13 cm 2 for each cell. For reference, solar cells on flat ZnO coated glass and on commercial standard Asahi U-type substrates with randomly textured SnO2:F were also fabricated. Thicknesses of the layers applied in the studied cells are shown in Table 1. The photovoltaic performance of all solar cells was measured using a solar simulator creating one sun illumination (AM1.5, 100 mW/cm 2). The corresponding external collection efficiency (ECE) was also investigated using monochromatic light from a Xenon lamp. The morphology of ZnO nanorods, the thickness of all applied layers, and the microstructure of the completed nanorod solar cells were characterized by high resolution scanning electron microscopy (HRSEM) with a Philips XL30 scanning field emission gun electron microscope. Phase identification of as-prepared ZnO nanorods was carried out by X-ray diffraction (XRD) employing a powder diffraction setup equipped with a Philips PW 1729 X-ray generator using monochromatic Cu Kα radiation at a voltage of 40 kV with a current of 20 mA. The diffuse reflection at Ag coated substrates, the transmission and the front surface reflection of the completed cells were measured by a Perkin Elmer Lambda 2S double beam spectrophotometer equipped with an integrating sphere. Angular resolved scattering measurements were performed using a home-made set-up with a HeNe laser source (632.8 nm) and a photodiode detector under variable angles. The incident beam is directed under normal incidence onto the sample after passing through a polarizing filter and a chopper.
3. Results Fig. 1 shows XRD patterns of the ZnO seed layer and the as-grown ZnO nanorods on the seed layer prepared at 80 °C using a precursor solution concentration of 0.0005 mol/L. High quality single-crystal ZnO nanorods were formed, which was supported by the very strong and narrow ZnO peak of the (002) plane. The inset HRSEM image in Fig. 1 shows the morphology of as-prepared nanorods, which have
Table 1 Thicknesses of the applied layers in nanometer of the three types of fabricated solar cells. No.
Ag
TCO
i-Layer
ITO
F75 T75 NR75
100 100 20
100 100 38
75 75 75
80 80 35
Fig. 1. XRD pattern of as-prepared ZnO nanorods on a ZnO coated glass grown at 80 °C during 3 hours under precursor solution concentration of 0.0005 mol/L. Inset HRSEM top-view (tilted 45°) image shows the morphology of as-grown ZnO nanorods.
an average diameter of about 100 nm, a length of 400 nm and a site-density of about 7 × 10 8/cm 2. Fig. 2(a) represents a schematic diagram of the nanorod three dimensional (nano-3D) solar cell design. Fig. 2(b) shows a HRSEM sectional view of the completed nano-3D solar cells with an intrinsic (i-) layer thickness of about 75 nm (NR75). The average thickness of all applied layers in the nano-3D cells was determined by measuring the growth in diameter of the coated nanorods in HRSEM images after each individual layer deposition. Cells on flat ZnO thin film coated glass (F75) and on an Asahi U-type textured substrate (T75) with the same 75 nm thick a-Si:H absorber layer are prepared for comparison with the nano-3D cells. Table 1 shows the layer thickness in the three types of studied cells. The layer thicknesses in F75 and T75 are determined by the deposition rate which is already known for such morphologies, while the average thickness of all the applied layers in the nano-3D cells was determined by measuring the growth in diameter of the coated nanorods in HRSEM images after each individual layer deposition. It should be pointed out that the Ag, TCO and ITO layers were simultaneously deposited on the nanorod substrate, the textured substrate and the flat substrate, whereas the n-i-p stack on the nanorod substrate was deposited in a separate run. Obviously the conformal layer coating on the nanorods is much more difficult than that on the flat and the textured substrates. Besides, on the nanorod substrate there is a much higher surface area to be coated in the same projected unit of area compared with that on the flat and the textured substrates. Therefore the thickness of Ag, TCO and ITO layers coated on ZnO nanorods in NR75 is much thinner than that in F75 and T75. The current density versus voltage (J–V) characteristics of the investigated three sets of devices are shown in Fig. 3(a). An efficiency (η) of 3.9% is achieved for the nanorod cell NR75, which is significantly higher than that of the flat cell F75 (2.6%) and the textured cell T75 (3.0%). NR75 yields a short-circuit current density (Jsc) of 11.6 mA/ cm 2, which is much higher than that of F75 (4.9 mA/cm 2) and T75 (6.1 mA/cm 2). However, NR75 demonstrates lower open-circuit voltage (Voc) and fill factor (FF) as compared to the two reference cells. The corresponding ECE is shown in Fig. 3(b). Over the studied wavelength range of 350–800 nm NR75 demonstrates a much broader ECE curve than that of F75 and T75. At wavelengths shorter than 400 nm there is very small difference between the three cells. Beyond the wavelength of 400 nm NR75 exhibits significantly higher ECE compared to F75 and T75, especially in the 500–700 nm region. The maximum ECE of F75 and T75 is at the wavelength of around 480 nm, whereas in NR75 the ECE peak is around 530 nm. Higher ECE in NR75 than that in F75 and T75 is consistent with higher Jsc as shown in Fig. 3(a).
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
2211
Fig. 2. ZnO-nanorod/a-Si:H solar cell with a 75 nm thick i-layer thickness. (a) 3D schematic diagram (not to scale). (b) HRSEM cross-section view of a completed nanorod cell.
The reflection at the front surface of the completed cells was measured to better understand the light management within the devices. Fig. 4(a) shows a photograph of NR75 and F75. As can be easily seen by naked eye, F75 has a mirror-like reflection (on this picture the reflection of the camera can be identified). In contrast, NR75 shows a dark blue surface with little reflection. The reflection spectra are shown in Fig. 4(b). Compared to F75 and even compared to T75, NR75 exhibits a much lower surface reflection throughout the spectral range. The positions of the minimal reflection are consistent with that of the maximum ECE as shown in Fig. 3(b). We can calculate the absorption (A) by A = 1 − R − T, where R is the reflection and T is the transmission. Almost no transmission is observed in the light wavelength range of 350–800 nm in all the three types of completed devices. Therefore the maximum light absorption fraction in the
nanorod cell is around 90% at the wavelength of 530 nm, which is much higher than that in the flat and the textured cells. In order to investigate the light scattering ability of the three kinds of substrates, their diffuse reflection has been determined. Fig. 5(a) shows the diffuse reflection of the Ag coated substrates for the studied cells. Almost no diffuse reflection was observed for the flat substrate. The nanorod substrate exhibits significantly higher diffuse reflection than the textured counterpart for wavelengths larger than 520 nm, where optical absorption in amorphous Si decreases and the effect of light trapping increases. The improvement of light scattering in the long wavelength region, where amorphous Si has a very weak absorption, is crucial for improving the photovoltaic performance of the devices. Fig. 5(b) shows the angular resolved scattering of the three kinds of Ag coated substrates. The amount of light scattered as a function of angle was measured by a movable photodiode detector. Compared to the flat and the textured substrates, more light is scattered into large off-normal angles for the nanorod substrate.
4. Discussion
Fig. 3. Photovoltaic performance of the three kinds of studied solar cells. (a) Best J–V measurements of the flat cell F75, the texture cell T75 and the nanorod solar cell NR75. The legend shows the cell parameters. (b) The corresponding spectral response curves.
For the nano-3D solar cell design, the morphology of the base structure is a critical component. To obtain a conformal layer coating and to avoid shunting of the completed devices, the length of the nanorods is one of the most important design considerations. The length used in this work is in the range of 300–500 nm. In our other experiments we observed that long nanorods (>500 nm) easily lead to shunts in the completed devices. The site density of the nanorods is another critical parameter. It is difficult to form cell structures with a rod-like shape if the density is too high, whereas a too low site density leads to a reduced light absorbing volume and a reduced diffuse scattering between individual nanorods, since significant amount of flat area would be presented in this case. From a practical point of view, the preparation process of the nanostructured rod forest must be controllable and adjustable. In our design we employ ZnO nanorods as the backbones of the nano-3D solar cells, since they can be prepared with the simple technique of hydrothermal growth on cheap substrates such as glass and plastic. The morphology can be tuned by simply changing growth conditions, such as precursor solution concentrations and growth time [31]. The nano-3D cell has a lower Voc than that of the flat and the texture reference cells, as shown in Fig. 3(a). A similar phenomenon is also reported in the literature [7-10,15]. In the nanorod system, the surface and the interface areas are significantly increased with respect to volume. For this reason, the cells contain larger amounts of defects per projected unit of area than conventional flat or textured cells. These defects trap the charge carriers and subsequently cause an increase in recombination. Bulk recombination and surface recombination in a semiconductor limit the Voc [32]. For the planar and the textured structure, bulk recombination is dominating, whereas for the nanorod geometry, surface and interface recombinations are
2212
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
Fig. 4. Reflection at the front surface of the completed cells. (a) A photograph of NR75 and F75. (b) Reflection spectra of F75, T75 and NR75.
dominant. This demonstrates that limitation of surface and interface states is crucial for further optimization of these nanorod cells. The most interesting benefit of the nano-3D cells is the high current density due to strong light trapping. Multiple optical enhancement effects are introduced in the nano-3D system. The first effect is the concept of orthogonalization of light absorption and carrier collection paths. Red light penetrates deeply in the absorber layer. In order to utilize this part of light spectrum, the typical thickness of a-Si:H absorber layer should be 300–500 nm. For the flat and the texture cell, the 75 nm thick i-layer is much too thin. In contrast, in the nanorod system the absorber
layer in the axial direction is thick enough to capture deeply penetrating photons, while in the radial direction it is thin enough for efficient carrier extraction and collection. The improved usage of long wavelengths visible light paves the way for enhanced current density and consequently, improved efficiency of the devices. The second effect is the antireflection effect, as shown in Fig. 4. The rough surface of the nano-3D cells reduces the reflection and thus decreases the light loss at the front surface. The nanorod arrays act as a light trap and internal multiple reflection between neighboring nanorods contributes to a multiplication of the absorption path in the absorber layer. The third effect is the highly diffusive scattering at the back reflector, as shown in Fig. 5. The optical path length can be significantly increased when light is scattered into large off-normal angle. This is similar with a textured front TCO layer in a superstrate structure. Another possible additional effect is plasmonic light scattering, since Ag is a commonly used plasmonic material. This effect is reported in Refs. [33] and [34] in which a similarly scaled back reflector structure is employed. As opposed to Ag nanoparticles which are widely used in plasmonic cells, the ZnO/Ag core-shell nanorod structure used in this work might couple incident light into guided modes and thus increase the absorption in the absorber layer. The existence of a plasmonic effect in the ZnO/Ag core-shell structure still needs to be further investigated and proved in our future work. 5. Conclusions In summary, a ZnO nanorod 3D solar cell structure, exhibiting an impressive current density and energy conversion efficiency, is presented. An efficiency of 3.9% has been obtained for the nano-3D solar cell with a 75 nm thick a-Si:H absorber layer. We demonstrate that the nanorod solar cell design shows light trapping that is enhanced beyond that of a randomly textured cell. Combined effects of orthogonalization of light absorption and carrier collection paths, anti-reflection at the front surface, and highly diffusive scattering at the back reflector result in a higher light absorption in the nanorod devices, while plasmonic effects cannot be ruled out. The photovoltaic performance of the nano-3D solar cells can significantly be improved by further optimization of the morphology of ZnO nanorods and the thicknesses of the applied layers. This design has broad potential applications because of the simple and scalable preparation process. It is also applicable to low-cost substrate solar cells using other semiconductor materials such as thin film micro- or nanocrystalline Si, CdTe, and CuInxGa1 − xSe2. Acknowledgments
Fig. 5. Light scattering of the three kinds of substrates for cells. Flat: glass/ZnO-film/Ag, texture: Asahi SnO2:F/Ag, nanorod: glass/ZnO-film/ZnO-nanorod/Ag. (a) Diffuse reflection. (b) Angular resolved scattering.
Y. Kuang acknowledges the financial support from China Scholarship Council (CSC) under contract no. 2009615001.
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
References [1] D.E. Carlson, C.R. Wronski, Appl. Phys. Lett. 28 (1976) 671–673. [2] J. Meier, E. Vallat-Sauvain, S. Dubail, U. Kroll, J. Dubail, S. Golay, L. Feitknecht, P. Torres, S. Fay, D. Fischer, A. Shah, Sol. Energy Mater. Sol. Cells 66 (2001) 73–84. [3] J. Müller, O. Kluth, S. Wieder, H. Siekmann, G. Schöpe, W. Reetz, O. Vetterl, D. Lundszien, A. Lambertz, F. Finger, B. Rech, H. Wagner, Sol. Energy Mater. Sol. Cells 66 (2001) 275–281. [4] D. Derkacs, S.H. Lim, P. Matheu, W. Mar, E.T. Yu, Appl. Phys. Lett. 89 (2006) 093103. [5] K.F. Feenstra, R.E.I. Schropp, W.F. Van der Weg, J. Appl. Phys. 85 (1999) 6843–6852. [6] H.A. Atwater, A. Polman, Nat. Mater. 9 (2010) 205–213. [7] B.M. Kayes, H.A. Atwater, N.S. Lewis, J. Appl. Phys. 97 (2005) 114302. [8] J.M. Spurgeon, H.A. Atwater, N.S. Lewis, J. Phys, Chem. C 112 (2008) 6186–6193. [9] D. Kieven, T. Dittrich, A. Belaidi, J. Tornow, K. Schwarzburg, N. Allsop, M. Lux-Steiner, Appl. Phys. Lett. 92 (2008) 153107. [10] M. Krunks, E. Kärber, A. Katerski, K. Otto, I. OjaAcik, T. Dedova, A. Mere, Sol. Energy Mater. Sol. Cells 94 (2010) 1191–1195. [11] I. Kaiser, K. Ernst, Ch.-H. Fischer, R. Könenkamp, C. Rost, I. Sieber, M.Ch. Lux-Steiner, Sol. Energy Mater. Sol. Cells 67 (2001) 89–96. [12] J.K. Rath, Y. Liu, M.M. de Jong, J. de Wild, J.A. Schuttauf, M. Brinza, R.E.I. Schropp, Thin Solid Films 518 (2010) 129–135. [13] O. Kluth, B. Rech, L. Houben, S. Wieder, G. Schöpe, C. Beneking, H. Wagner, A. Löffl, H.W. Schock, Thin Solid Films 351 (1999) 247–253. [14] R.H. Franken, R.L. Stolk, H. Li, C.H.M. van der Werf, J.K. Rath, R.E.I. Schropp, J. Appl. Phys. 102 (2007) 014503. [15] Y. Kuang, K.H.M. van der Werf, Z.S. Houweling, R.E.I. Schropp, Appl. Phys. Lett. 98 (2011) 113111. [16] A. Belaidi, T. Dittrich, D. Kieven, J. Tornow, K. Schwarzburg, M. Lux-Steiner, Phys. Stat. Sol. (RRL) 2 (2008) 172–174. [17] M. Krunks, A. Katerski, T. Dedova, I.O. Acik, A. Mere, Sol. Energy Mater. Sol. Cells 92 (2008) 1016–1019. [18] V. Sivakov, G. Andrä, A. Gawlik, A. Berger, J. Plentz, F. Falk, S.H. Christiansen, Nano Lett. 9 (2009) 1549–1554.
2213
[19] L. Tsakalakos, J. Balch, J. Fronheiser, B.A. Korevaar, O. Sulima, J. Rand, Appl. Phys. Lett. 91 (2007) 233117. [20] M. Law, L.E. Greene, J.C. Johnson, R. Saykally, P. Yang, Nat. Mater. 4 (2005) 455–459. [21] Th. Stelzner, M. Pietsch, G. Andrä, F. Falk, E. Ose, S. Christiansen, Nanotechnology 19 (2008) 295203. [22] K. Peng, Y. Xu, Y. Wu, Y. Yan, S.T. Lee, J. Zhu, Small 1 (2005) 1062–1067. [23] M.D. Kelzenberg, S.W. Boettcher, J.A. Petykiewicz, D.B. Turner-Evans, M.C. Putnam, E.L. Warren, J.M. Spurgeon, R.M. Briggs, N.S. Lewis, H.A. Atwater, Nat. Mater. 9 (2010) 239–244. [24] B. Tian, X. Zheng, T.J. Kempa, Y. Fang, N. Yu, G. Yu, J. Huang, C.M. Lieber, Nature 449 (2007) 885–890. [25] E.C. Garnett, P. Yang, J. Am. Chem. Soc. 130 (2008) 9224–9225. [26] B.D. Yuhas, P. Yang, J. Am. Chem. Soc. 131 (2009) 3756–3761. [27] E. Garnett, P. Yang, Nano Lett. 10 (2010) 1082–1087. [28] O.L. Muskens, J.G. Rivas, R.E. Algra, E.P.A.M. Bakkers, A. Lagendijk, Nano Lett. 8 (2008) 2638–2642. [29] M.J. Naughton, K. Kempa, Z.F. Ren, Y. Gao, J. Rybczynski, N. Argenti, W. Gao, Y. Wang, Y. Peng, J.R. Naughton, G. McMahon, T. Paudel, Y.C. Lan, M.J. Burns, A. Shepard, M. Clary, C. Ballif, F.J. Haug, T. Söderström, O. Cubero, C. Eminian, Phys. Stat. Sol. (RRL) 4 (2010) 181–183. [30] J. Zhu, Z. Yu, G.F. Burkhard, C.M. Hsu, S.T. Connor, Y. Xu, Q. Wang, M. McGehee, S. Fan, Y. Cui, Nano Lett. 9 (2009) 279–282. [31] Y. Kuang, K.H.M. van der Werf, Z.S. Houweling, M. Di Vece, R.E.I. Schropp, Proceedings of 37th IEEE Photovoltaic Specialists Conference, June 19–24 2011, Seattle, U.S.A. [32] M.A. Green, Solar Cells: Operating Principles, Technology, and System Applications, Prentice-Hall, New Jersey, 1982 p. 93. [33] V.E. Ferry, M.A. Verschuuren, H.B.T. Li, R.E.I. Schropp, H.A. Atwater, A. Polman, Appl. Phys. Lett. 95 (2009) 183503. [34] V.E. Ferry, M.A. Verschuuren, H.B.T. Li, E. Verhagen, R.J. Walters, R.E.I. Schropp, H.A. Atwater, A. Polman, Opt. Express 18 (2010) 237–245.
Contents lists available at SciVerse ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol
Fabrication and characterization of nanorod solar cells with an ultrathin a-Si:H absorber layer Yinghuan Kuang ⁎, Karine H.M. van der Werf, Z. Silvester Houweling, Marcel Di Vece, Ruud E.I. Schropp Faculty of Science, Debye Institute for Nanomaterials Science, Nanophotonics—Physics of Devices, Utrecht University, P.O. Box 80.000, 3508 TA Utrecht, The Netherlands
a r t i c l e
i n f o
Article history: Received 12 August 2011 Received in revised form 4 November 2011 Available online 5 December 2011 Keywords: Nanorods; Thin film solar cells; Extremely thin absorber layer; Amorphous hydrogenated silicon; ZnO
a b s t r a c t In this paper, we present a three-dimensional nanorod solar cell design. As the backbone of the nanorod device, density-controlled zinc oxide (ZnO) nanorods were synthesized by a simple aqueous solution growth technique at 80 °C on ZnO thin film pre-coated glass substrate. The as-prepared ZnO nanorods were coated by an amorphous hydrogenated silicon (a-Si:H) light absorber layer to form a nanorod solar cell. The light management, current–voltage characteristics and corresponding external quantum efficiency of the solar cells were investigated. An energy conversion efficiency of 3.9% was achieved for the nanorod solar cells with an a-Si:H absorber layer thickness of 75 nm, which is significantly higher than the 2.6% and the 3.0% obtained for cells with the same a-Si:H absorber layer thickness on planar ZnO and on textured SnO2:F counterparts, respectively. A short-circuit current density of 11.6 mA/cm 2 and correspondingly, a broad external quantum efficiency profile were achieved for the nanorod device. An absorbed light fraction higher than 80% in the wavelength range of 375–675 nm was also demonstrated for the nanorod solar cells, including a peak value of ~ 90% at 520–530 nm. © 2011 Elsevier B.V. All rights reserved.
1. Introduction In the past few years, thin film silicon solar cells have attracted great attention in the photovoltaic community because of the lower production cost compared to their crystalline silicon counterparts [1-5]. In thin film devices, the thickness of the absorber layer is governed by the well-known trade-off between light absorption and minority carrier collection [6-8]. The absorber layer must be optically thick to capture a sufficient fraction of incident photons but must also be sufficiently thin to enable minority carrier collection lengths larger than the thickness of the absorber layer. In order to obtain a large collection length, high quality material is a prerequisite. Hydrogenated amorphous silicon (a-Si:H) and nanocrystalline silicon (ncSi:H) are among the most developed thin film photovoltaic materials, but suffer from small minority carrier diffusion length. One solution to relieve the trade-off between light absorption and carrier collection is the usage of an extremely thin absorber layer (ETA) on structured substrate, which can keep a significant amount of optical absorption but remarkably reduce charge carriers recombination [911]. An alternative approach is enhancing light trapping by employing a surface texture such as that of natively textured commercially available SnO2:F or as that made by hydrochloric acid etching of ZnO:Al [12,13] for superstrate p-i-n structures, or that of structured Ag back electrodes in substrate n-i-p structures [14]. In these types
⁎ Corresponding author. E-mail address: [email protected] (Y. Kuang). 0022-3093/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2011.11.021
of structures, the incident light is scattered into off-normal angles; therefore the optical path length in the absorber layer is increased and light absorption is enhanced. Recently, the application of nanostructures such as nanorod [7-10,15-17] and nanowire [18-28] in thin film solar cells has gained considerable attention as a method to eliminate the trade-off between light absorption and carrier collection. The orthogonalization of light absorption and carrier collection paths plays an important role in nanorod/nanowire systems. An ultrathin cell built on nanorods is thick enough for sufficient light absorption in the axial direction while it is sufficiently thin in the radial direction to guarantee efficient carrier extraction and collection. To obtain a conformal layer coating in nanorod/nanowire solar cells, the morphology of the nanorod/nanowire array is critical. Approaches widely employed for the preparation of such nanostructured materials are for example reactive ion etching [27,29,30] and vapor–liquid–solid methods [19,21,23,24]. However, complicated fabrication processes, area limitation, sophisticated apparatus requirements, and energy consumption in these approaches greatly hinder the practical application in solar cells. In contrast, in this work we present a simple and scalable method. ZnO nanorods are selected as the base structure since they can very easily be synthesized by hydrothermal growth at a low temperature. The morphology of these ZnO nanorods can be precisely tuned by adjusting the growth conditions such as the precursor solution concentration and the growth temperature [31]. As the solar cell absorber layer we employ a-Si:H since this is a well-developed photovoltaic material, and because it is non-toxic and source materials are abundantly available.
2210
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
2. Experimental A flat ZnO layer was sputtered onto a glass substrate as a seed layer to favor nucleation and epitaxial growth of ZnO nanorods. A mixture of 0.0005 mol zinc acetate dihydrate (Zn(CH3COO)2·2H2O, SigmaAldrich) with an equal amount of 0.0005 mol hexamethylenetetramine (HMT, Sigma-Aldrich) was dissolved in 1 liter de-ionized water as the reactive solution. The solution was magnetically stirred for 15 minutes at room temperature. After that substrates were immersed in the solution, holding the pre-coated ZnO seed layer downwards. The growth temperature was kept constant at 80 °C during 3 hours. After growth, the substrates were thoroughly rinsed by de-ionized water and dried by blowing nitrogen. The as-grown nanorods were overcoated with an Ag layer and a ZnO:2%Al transparent conductive oxide (TCO) layer by sputtering. Then, the deposition of a n-i-p layer stack was carried out in a multi-chamber deposition system using hot-wire chemical vapor deposition (HWCVD). The deposition of the solar cells was described elsewhere [15]. An array of 4 mm × 4 mm squares of a transparent conducting indium tin oxide (ITO) layer was sputtered through a mask to define individual cell pads. Gold top-grid contacts were evaporated onto the square ITO pads using another mask, leaving an active cell area of 0.13 cm 2 for each cell. For reference, solar cells on flat ZnO coated glass and on commercial standard Asahi U-type substrates with randomly textured SnO2:F were also fabricated. Thicknesses of the layers applied in the studied cells are shown in Table 1. The photovoltaic performance of all solar cells was measured using a solar simulator creating one sun illumination (AM1.5, 100 mW/cm 2). The corresponding external collection efficiency (ECE) was also investigated using monochromatic light from a Xenon lamp. The morphology of ZnO nanorods, the thickness of all applied layers, and the microstructure of the completed nanorod solar cells were characterized by high resolution scanning electron microscopy (HRSEM) with a Philips XL30 scanning field emission gun electron microscope. Phase identification of as-prepared ZnO nanorods was carried out by X-ray diffraction (XRD) employing a powder diffraction setup equipped with a Philips PW 1729 X-ray generator using monochromatic Cu Kα radiation at a voltage of 40 kV with a current of 20 mA. The diffuse reflection at Ag coated substrates, the transmission and the front surface reflection of the completed cells were measured by a Perkin Elmer Lambda 2S double beam spectrophotometer equipped with an integrating sphere. Angular resolved scattering measurements were performed using a home-made set-up with a HeNe laser source (632.8 nm) and a photodiode detector under variable angles. The incident beam is directed under normal incidence onto the sample after passing through a polarizing filter and a chopper.
3. Results Fig. 1 shows XRD patterns of the ZnO seed layer and the as-grown ZnO nanorods on the seed layer prepared at 80 °C using a precursor solution concentration of 0.0005 mol/L. High quality single-crystal ZnO nanorods were formed, which was supported by the very strong and narrow ZnO peak of the (002) plane. The inset HRSEM image in Fig. 1 shows the morphology of as-prepared nanorods, which have
Table 1 Thicknesses of the applied layers in nanometer of the three types of fabricated solar cells. No.
Ag
TCO
i-Layer
ITO
F75 T75 NR75
100 100 20
100 100 38
75 75 75
80 80 35
Fig. 1. XRD pattern of as-prepared ZnO nanorods on a ZnO coated glass grown at 80 °C during 3 hours under precursor solution concentration of 0.0005 mol/L. Inset HRSEM top-view (tilted 45°) image shows the morphology of as-grown ZnO nanorods.
an average diameter of about 100 nm, a length of 400 nm and a site-density of about 7 × 10 8/cm 2. Fig. 2(a) represents a schematic diagram of the nanorod three dimensional (nano-3D) solar cell design. Fig. 2(b) shows a HRSEM sectional view of the completed nano-3D solar cells with an intrinsic (i-) layer thickness of about 75 nm (NR75). The average thickness of all applied layers in the nano-3D cells was determined by measuring the growth in diameter of the coated nanorods in HRSEM images after each individual layer deposition. Cells on flat ZnO thin film coated glass (F75) and on an Asahi U-type textured substrate (T75) with the same 75 nm thick a-Si:H absorber layer are prepared for comparison with the nano-3D cells. Table 1 shows the layer thickness in the three types of studied cells. The layer thicknesses in F75 and T75 are determined by the deposition rate which is already known for such morphologies, while the average thickness of all the applied layers in the nano-3D cells was determined by measuring the growth in diameter of the coated nanorods in HRSEM images after each individual layer deposition. It should be pointed out that the Ag, TCO and ITO layers were simultaneously deposited on the nanorod substrate, the textured substrate and the flat substrate, whereas the n-i-p stack on the nanorod substrate was deposited in a separate run. Obviously the conformal layer coating on the nanorods is much more difficult than that on the flat and the textured substrates. Besides, on the nanorod substrate there is a much higher surface area to be coated in the same projected unit of area compared with that on the flat and the textured substrates. Therefore the thickness of Ag, TCO and ITO layers coated on ZnO nanorods in NR75 is much thinner than that in F75 and T75. The current density versus voltage (J–V) characteristics of the investigated three sets of devices are shown in Fig. 3(a). An efficiency (η) of 3.9% is achieved for the nanorod cell NR75, which is significantly higher than that of the flat cell F75 (2.6%) and the textured cell T75 (3.0%). NR75 yields a short-circuit current density (Jsc) of 11.6 mA/ cm 2, which is much higher than that of F75 (4.9 mA/cm 2) and T75 (6.1 mA/cm 2). However, NR75 demonstrates lower open-circuit voltage (Voc) and fill factor (FF) as compared to the two reference cells. The corresponding ECE is shown in Fig. 3(b). Over the studied wavelength range of 350–800 nm NR75 demonstrates a much broader ECE curve than that of F75 and T75. At wavelengths shorter than 400 nm there is very small difference between the three cells. Beyond the wavelength of 400 nm NR75 exhibits significantly higher ECE compared to F75 and T75, especially in the 500–700 nm region. The maximum ECE of F75 and T75 is at the wavelength of around 480 nm, whereas in NR75 the ECE peak is around 530 nm. Higher ECE in NR75 than that in F75 and T75 is consistent with higher Jsc as shown in Fig. 3(a).
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
2211
Fig. 2. ZnO-nanorod/a-Si:H solar cell with a 75 nm thick i-layer thickness. (a) 3D schematic diagram (not to scale). (b) HRSEM cross-section view of a completed nanorod cell.
The reflection at the front surface of the completed cells was measured to better understand the light management within the devices. Fig. 4(a) shows a photograph of NR75 and F75. As can be easily seen by naked eye, F75 has a mirror-like reflection (on this picture the reflection of the camera can be identified). In contrast, NR75 shows a dark blue surface with little reflection. The reflection spectra are shown in Fig. 4(b). Compared to F75 and even compared to T75, NR75 exhibits a much lower surface reflection throughout the spectral range. The positions of the minimal reflection are consistent with that of the maximum ECE as shown in Fig. 3(b). We can calculate the absorption (A) by A = 1 − R − T, where R is the reflection and T is the transmission. Almost no transmission is observed in the light wavelength range of 350–800 nm in all the three types of completed devices. Therefore the maximum light absorption fraction in the
nanorod cell is around 90% at the wavelength of 530 nm, which is much higher than that in the flat and the textured cells. In order to investigate the light scattering ability of the three kinds of substrates, their diffuse reflection has been determined. Fig. 5(a) shows the diffuse reflection of the Ag coated substrates for the studied cells. Almost no diffuse reflection was observed for the flat substrate. The nanorod substrate exhibits significantly higher diffuse reflection than the textured counterpart for wavelengths larger than 520 nm, where optical absorption in amorphous Si decreases and the effect of light trapping increases. The improvement of light scattering in the long wavelength region, where amorphous Si has a very weak absorption, is crucial for improving the photovoltaic performance of the devices. Fig. 5(b) shows the angular resolved scattering of the three kinds of Ag coated substrates. The amount of light scattered as a function of angle was measured by a movable photodiode detector. Compared to the flat and the textured substrates, more light is scattered into large off-normal angles for the nanorod substrate.
4. Discussion
Fig. 3. Photovoltaic performance of the three kinds of studied solar cells. (a) Best J–V measurements of the flat cell F75, the texture cell T75 and the nanorod solar cell NR75. The legend shows the cell parameters. (b) The corresponding spectral response curves.
For the nano-3D solar cell design, the morphology of the base structure is a critical component. To obtain a conformal layer coating and to avoid shunting of the completed devices, the length of the nanorods is one of the most important design considerations. The length used in this work is in the range of 300–500 nm. In our other experiments we observed that long nanorods (>500 nm) easily lead to shunts in the completed devices. The site density of the nanorods is another critical parameter. It is difficult to form cell structures with a rod-like shape if the density is too high, whereas a too low site density leads to a reduced light absorbing volume and a reduced diffuse scattering between individual nanorods, since significant amount of flat area would be presented in this case. From a practical point of view, the preparation process of the nanostructured rod forest must be controllable and adjustable. In our design we employ ZnO nanorods as the backbones of the nano-3D solar cells, since they can be prepared with the simple technique of hydrothermal growth on cheap substrates such as glass and plastic. The morphology can be tuned by simply changing growth conditions, such as precursor solution concentrations and growth time [31]. The nano-3D cell has a lower Voc than that of the flat and the texture reference cells, as shown in Fig. 3(a). A similar phenomenon is also reported in the literature [7-10,15]. In the nanorod system, the surface and the interface areas are significantly increased with respect to volume. For this reason, the cells contain larger amounts of defects per projected unit of area than conventional flat or textured cells. These defects trap the charge carriers and subsequently cause an increase in recombination. Bulk recombination and surface recombination in a semiconductor limit the Voc [32]. For the planar and the textured structure, bulk recombination is dominating, whereas for the nanorod geometry, surface and interface recombinations are
2212
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
Fig. 4. Reflection at the front surface of the completed cells. (a) A photograph of NR75 and F75. (b) Reflection spectra of F75, T75 and NR75.
dominant. This demonstrates that limitation of surface and interface states is crucial for further optimization of these nanorod cells. The most interesting benefit of the nano-3D cells is the high current density due to strong light trapping. Multiple optical enhancement effects are introduced in the nano-3D system. The first effect is the concept of orthogonalization of light absorption and carrier collection paths. Red light penetrates deeply in the absorber layer. In order to utilize this part of light spectrum, the typical thickness of a-Si:H absorber layer should be 300–500 nm. For the flat and the texture cell, the 75 nm thick i-layer is much too thin. In contrast, in the nanorod system the absorber
layer in the axial direction is thick enough to capture deeply penetrating photons, while in the radial direction it is thin enough for efficient carrier extraction and collection. The improved usage of long wavelengths visible light paves the way for enhanced current density and consequently, improved efficiency of the devices. The second effect is the antireflection effect, as shown in Fig. 4. The rough surface of the nano-3D cells reduces the reflection and thus decreases the light loss at the front surface. The nanorod arrays act as a light trap and internal multiple reflection between neighboring nanorods contributes to a multiplication of the absorption path in the absorber layer. The third effect is the highly diffusive scattering at the back reflector, as shown in Fig. 5. The optical path length can be significantly increased when light is scattered into large off-normal angle. This is similar with a textured front TCO layer in a superstrate structure. Another possible additional effect is plasmonic light scattering, since Ag is a commonly used plasmonic material. This effect is reported in Refs. [33] and [34] in which a similarly scaled back reflector structure is employed. As opposed to Ag nanoparticles which are widely used in plasmonic cells, the ZnO/Ag core-shell nanorod structure used in this work might couple incident light into guided modes and thus increase the absorption in the absorber layer. The existence of a plasmonic effect in the ZnO/Ag core-shell structure still needs to be further investigated and proved in our future work. 5. Conclusions In summary, a ZnO nanorod 3D solar cell structure, exhibiting an impressive current density and energy conversion efficiency, is presented. An efficiency of 3.9% has been obtained for the nano-3D solar cell with a 75 nm thick a-Si:H absorber layer. We demonstrate that the nanorod solar cell design shows light trapping that is enhanced beyond that of a randomly textured cell. Combined effects of orthogonalization of light absorption and carrier collection paths, anti-reflection at the front surface, and highly diffusive scattering at the back reflector result in a higher light absorption in the nanorod devices, while plasmonic effects cannot be ruled out. The photovoltaic performance of the nano-3D solar cells can significantly be improved by further optimization of the morphology of ZnO nanorods and the thicknesses of the applied layers. This design has broad potential applications because of the simple and scalable preparation process. It is also applicable to low-cost substrate solar cells using other semiconductor materials such as thin film micro- or nanocrystalline Si, CdTe, and CuInxGa1 − xSe2. Acknowledgments
Fig. 5. Light scattering of the three kinds of substrates for cells. Flat: glass/ZnO-film/Ag, texture: Asahi SnO2:F/Ag, nanorod: glass/ZnO-film/ZnO-nanorod/Ag. (a) Diffuse reflection. (b) Angular resolved scattering.
Y. Kuang acknowledges the financial support from China Scholarship Council (CSC) under contract no. 2009615001.
Y. Kuang et al. / Journal of Non-Crystalline Solids 358 (2012) 2209–2213
References [1] D.E. Carlson, C.R. Wronski, Appl. Phys. Lett. 28 (1976) 671–673. [2] J. Meier, E. Vallat-Sauvain, S. Dubail, U. Kroll, J. Dubail, S. Golay, L. Feitknecht, P. Torres, S. Fay, D. Fischer, A. Shah, Sol. Energy Mater. Sol. Cells 66 (2001) 73–84. [3] J. Müller, O. Kluth, S. Wieder, H. Siekmann, G. Schöpe, W. Reetz, O. Vetterl, D. Lundszien, A. Lambertz, F. Finger, B. Rech, H. Wagner, Sol. Energy Mater. Sol. Cells 66 (2001) 275–281. [4] D. Derkacs, S.H. Lim, P. Matheu, W. Mar, E.T. Yu, Appl. Phys. Lett. 89 (2006) 093103. [5] K.F. Feenstra, R.E.I. Schropp, W.F. Van der Weg, J. Appl. Phys. 85 (1999) 6843–6852. [6] H.A. Atwater, A. Polman, Nat. Mater. 9 (2010) 205–213. [7] B.M. Kayes, H.A. Atwater, N.S. Lewis, J. Appl. Phys. 97 (2005) 114302. [8] J.M. Spurgeon, H.A. Atwater, N.S. Lewis, J. Phys, Chem. C 112 (2008) 6186–6193. [9] D. Kieven, T. Dittrich, A. Belaidi, J. Tornow, K. Schwarzburg, N. Allsop, M. Lux-Steiner, Appl. Phys. Lett. 92 (2008) 153107. [10] M. Krunks, E. Kärber, A. Katerski, K. Otto, I. OjaAcik, T. Dedova, A. Mere, Sol. Energy Mater. Sol. Cells 94 (2010) 1191–1195. [11] I. Kaiser, K. Ernst, Ch.-H. Fischer, R. Könenkamp, C. Rost, I. Sieber, M.Ch. Lux-Steiner, Sol. Energy Mater. Sol. Cells 67 (2001) 89–96. [12] J.K. Rath, Y. Liu, M.M. de Jong, J. de Wild, J.A. Schuttauf, M. Brinza, R.E.I. Schropp, Thin Solid Films 518 (2010) 129–135. [13] O. Kluth, B. Rech, L. Houben, S. Wieder, G. Schöpe, C. Beneking, H. Wagner, A. Löffl, H.W. Schock, Thin Solid Films 351 (1999) 247–253. [14] R.H. Franken, R.L. Stolk, H. Li, C.H.M. van der Werf, J.K. Rath, R.E.I. Schropp, J. Appl. Phys. 102 (2007) 014503. [15] Y. Kuang, K.H.M. van der Werf, Z.S. Houweling, R.E.I. Schropp, Appl. Phys. Lett. 98 (2011) 113111. [16] A. Belaidi, T. Dittrich, D. Kieven, J. Tornow, K. Schwarzburg, M. Lux-Steiner, Phys. Stat. Sol. (RRL) 2 (2008) 172–174. [17] M. Krunks, A. Katerski, T. Dedova, I.O. Acik, A. Mere, Sol. Energy Mater. Sol. Cells 92 (2008) 1016–1019. [18] V. Sivakov, G. Andrä, A. Gawlik, A. Berger, J. Plentz, F. Falk, S.H. Christiansen, Nano Lett. 9 (2009) 1549–1554.
2213
[19] L. Tsakalakos, J. Balch, J. Fronheiser, B.A. Korevaar, O. Sulima, J. Rand, Appl. Phys. Lett. 91 (2007) 233117. [20] M. Law, L.E. Greene, J.C. Johnson, R. Saykally, P. Yang, Nat. Mater. 4 (2005) 455–459. [21] Th. Stelzner, M. Pietsch, G. Andrä, F. Falk, E. Ose, S. Christiansen, Nanotechnology 19 (2008) 295203. [22] K. Peng, Y. Xu, Y. Wu, Y. Yan, S.T. Lee, J. Zhu, Small 1 (2005) 1062–1067. [23] M.D. Kelzenberg, S.W. Boettcher, J.A. Petykiewicz, D.B. Turner-Evans, M.C. Putnam, E.L. Warren, J.M. Spurgeon, R.M. Briggs, N.S. Lewis, H.A. Atwater, Nat. Mater. 9 (2010) 239–244. [24] B. Tian, X. Zheng, T.J. Kempa, Y. Fang, N. Yu, G. Yu, J. Huang, C.M. Lieber, Nature 449 (2007) 885–890. [25] E.C. Garnett, P. Yang, J. Am. Chem. Soc. 130 (2008) 9224–9225. [26] B.D. Yuhas, P. Yang, J. Am. Chem. Soc. 131 (2009) 3756–3761. [27] E. Garnett, P. Yang, Nano Lett. 10 (2010) 1082–1087. [28] O.L. Muskens, J.G. Rivas, R.E. Algra, E.P.A.M. Bakkers, A. Lagendijk, Nano Lett. 8 (2008) 2638–2642. [29] M.J. Naughton, K. Kempa, Z.F. Ren, Y. Gao, J. Rybczynski, N. Argenti, W. Gao, Y. Wang, Y. Peng, J.R. Naughton, G. McMahon, T. Paudel, Y.C. Lan, M.J. Burns, A. Shepard, M. Clary, C. Ballif, F.J. Haug, T. Söderström, O. Cubero, C. Eminian, Phys. Stat. Sol. (RRL) 4 (2010) 181–183. [30] J. Zhu, Z. Yu, G.F. Burkhard, C.M. Hsu, S.T. Connor, Y. Xu, Q. Wang, M. McGehee, S. Fan, Y. Cui, Nano Lett. 9 (2009) 279–282. [31] Y. Kuang, K.H.M. van der Werf, Z.S. Houweling, M. Di Vece, R.E.I. Schropp, Proceedings of 37th IEEE Photovoltaic Specialists Conference, June 19–24 2011, Seattle, U.S.A. [32] M.A. Green, Solar Cells: Operating Principles, Technology, and System Applications, Prentice-Hall, New Jersey, 1982 p. 93. [33] V.E. Ferry, M.A. Verschuuren, H.B.T. Li, R.E.I. Schropp, H.A. Atwater, A. Polman, Appl. Phys. Lett. 95 (2009) 183503. [34] V.E. Ferry, M.A. Verschuuren, H.B.T. Li, E. Verhagen, R.J. Walters, R.E.I. Schropp, H.A. Atwater, A. Polman, Opt. Express 18 (2010) 237–245.