Impurity Photovoltaic Effect in Multijunction Solar Cells

Available online at www.sciencedirect.com Procedia Technology 7 ( 2013 ) 166 172 The 2013 Iberoamerican Conference on Electronics Engineering and Computer Science Impurity Photovoltaic Effect in Multijunction Solar Cells Md. Shahriar Parvez Khan a, *, Esmat Farzana a,b a Ahsanullah University of Science and Technology, 141 & 142, Love Road, Tejgaon Industrial Area, Dhaka-1208, Bangladesh b Bangladesh University of Engineering and Technology (BUET), Dhaka-1000, Bangladesh Abstract A study has been carried out to investigate the Impurity Photovoltaic (IPV) effect on the performance of a double junction solar cell, l. In this work, impurity has been introduced in a double junction solar cell - separately and simultaneously in the two sub-cells which have been found to improve the cell performance. However, the effect of impurity becomes prominent when it is applied to the sub-cell with lower current. Furthermore, impact of impurity concentration has been obtained on short-circuit current density, open-circuit voltage and overall efficiency which have been explained through compensation by deep -level impurity dopant, conversion of conductivity type, carrier lifetime and total equilibrium carrier density. This improvement demonstrates the potential of the combination of IPV effect in multijunction solar cell. 2012 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Global Science and 2013 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Technology Forum Pte Ltd Selection and peer-review under responsibility of CIIECC 2013 Keywords: Multijunction Solar Cell; Impurity Photovoltaic Effect; Double Junction; Efficiency;Carrier Lifetime. 1. Introduction The rising cost of electricity from fossil fuels is intensively favouring energy production with renewable resources s uch as Solar Photovoltaic (PV). Increasing improvement in solar cell structure along with environmental concern are the primary factors for the impetus in solar energy production in recent years. But compared to the conventional ones, solar cell efficiency is still unfavourable to grid electricity. Hence, to become a feasible source, the performance of solar cell needs to be improved. Due to band gap constraints, a single-junction solar cell has limited wavelength response and low efficiency. Several methods have been adopted for efficiency improvement in solar cells of which Multijunction Solar Cells reported by Gray et al. [1], Huang et al. [2] Green et al. [3], Cotal et al. [4], Lu et al. [5] and Impurity Photovoltaic (IPV) effects shown by Wolf [6], Karazhanov [7], Shockley and Quiesser [8], Keevers and Green [9], Yuan et al. [10], Yuan et al. [11], Schmeits and Mani [12] are two notable ones. Multijunction cells, comprising of several materials with different band gaps, utilize a broader range of solar spectrum and hence, extract more energy from sunlight which is impossible to do with a single junction. * Corresponding author. Tel.: +88-0191-111-8868; fax: +88-02-8870417-18. E-mailaddress: khan.shahriar.parvez@gmail.com. 2212-0173 2013 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Selection and peer-review under responsibility of CIIECC 2013 doi: 10.1016/j.protcy.2013.04.021

Md. Shahriar Parvez Khan and Esmat Farzana / Procedia Technology 7 ( 2013 ) 166 172 167 The IPV effect also enhances the performance of a solar cell in a similar manner, i.e. via introduction of impurity which creates a sub-gap energy level. The cell then can absorb additional lower energy photon, leading to enhanced spectral response equivalent to the sub-gap. Introduced first by Wolf [6], however, this concept was brought to question by Shockley and Quiesser [8] because the improved carrier generation was supposed to be at the cost of increased recombination in the cell, the impurities acting as recombination centres. Later, it was found that the deep defect level contributes to degrade cell efficiency only if its energy level is located near the middle of the band gap, causing electron-hole recombination through the defect. However, if the energy level is carefully chosen far from the gap center, then IPV effect can significantly improve cell efficiency, as shown by Karazhanov [7]. Although IPV effect enhances cell performance by increasing the short circuit current density but, at the same time, it slightly decreases open circuit voltage. On the other hand, multijunction cells have better open circuit voltages compared to their single junction counterparts but their short circuit currents are almost reduced by a factor equal to the number of junctions, as reported by Gray et al. [1] and Huang et al. [2]. According to them, the short circuit current is limited by the junction or cell producing the lowest current. So, current matching between component cells in a multijunction setup is always an important criterion which, often, determines the optimum design. IPV effect can be utilized to solve this problem of multijunction cell. If impurity is introduced in the cell with the lowest current in a multijunction setup, it can increase the short circuit current of that particular cell. Decrement in open circuit voltage due to impurity should be negligible compared to overall improvement of short circuit current of the cell and so, the overall performance of the cell should improve. In the literature to date, several multijunctions cells, composed mostly of III-V materials have been reported by Gray et al. [1], Huang et al. [2], Green et al. [3], Cotal et al. [4] and Lu et al. [5]. IPV effect has also drawn significant interest and has been studied in detail by Keevers and Green [9], Yuan et al. [10], Yuan et al. [11], and Schmeits and Mani [12]. But, to the best of knowledge, there has been no attempt yet to combine these two approaches. Here, IPV effect in a two junction cell has been investigated and found to be improving cell performance. Open circuit voltage, short circuit current density and efficiency- have been demonstrated to investigate the enhancement through simulation software wxamps described by Liu et al. [13] and Liu et al. [14]. 2. Proposed Cell Structure Fig. 1 shows the simulated structure of the double junction cell along with the respective layer type, thickness and doping concentration. In the III-V compound InGaAs, fraction of In is 0.2 and of Ga is 0.8. Top and bottom surface reflectivity have been kept respectively 0 and 1. Basic parameters as reported by Streetman and Banerjee [15] and Levinshtein et al. [16] and used for the cell are listed in Table 1. n type GaP emitter layer, 0.005 μm, p type GaP IPV layer, 0.25 μm, p + type GaP base layer, 0.01 μm, p type GaP tunnelling layer, 0.001 μm, n type GaP tunnelling layer, 0.001 μm, n type InGaAs emitter layer, 1.0 μm, p type InGaAs IPV layer, 40 μm, p + type InGaAs base layer, 4.0 μm, Fig. 1: Structure of simulated cell

168 Md. Shahriar Parvez Khan and Esmat Farzana / Procedia Technology 7 ( 2013 ) 166 172 T able 1: Basic parameters used for the cell under study Parameter and unit Values for GaP Values For InGaAs Energy gap, E g (ev) 2.26 1.139 Dielectric constant 11.1 13.23 Electron affinity (ev) 3.8 4.236 Effective density of states in conduction band, N C (cm -3 ) Effective density of states in valence band, N V (cm -3 ) Electron mobility, μ n (cm 2 V -1 s -1 ) 250 6928 Hole mobility, μ p (cm 2 V -1 s -1 ) 150 350 3. Methodology and Formulation The model applied to IPV effect is modified Shockley-Read-Hall (SRH) model as shown by Shockley and Read [17] and Hall [18]. For a solar cell with idealized light trapping, the net recombination rate U via impurity is given by Keevers and Green [9] and Yuan et al. [10] as: In the expressions stated above, and are the electron and hole concentrations, and the lifetimes for electrons and holes, and the optical emission rates from the impurity for electrons and holes, and the electron and hole concentrations when the Fermi level coincides with the impurity level, the impurity concentrations, the impurity energy level, and the electron and hole thermal capture cross sections, the thermal velocity, and the conduction and valence band edges, and the effective densities of states in conduction and valence bands, and the electron and hole optical emission cross sections of the impurity, respectively. In Eq. 3, the band-to-band absorption coefficient, and the impurity absorption coefficients for electron and hole photoemission from the IPV impurity, the absorption coefficient for free-carrier absorption, and the occupation probability of impurity level as given by Hsieh and Card [19], respectively. The occupancy is can further be approximated to for p type impurity in n type layer as shown by Keevers and Green [9] and for the opposite.

Md. Shahriar Parvez Khan and Esmat Farzana / Procedia Technology 7 ( 2013 ) 166 172 169 4. Choice of Impurity Keevers and Green [9] suggest that, a midgap impurity maximizes amount of subgap spectrum but has poor photogenerative role. On the other hand, a shallow impurity provides much less access to subgap spectrum, but allows strong excitation processes. As a compromise, a non-midgap but deep-level impurity is taken which provides sufficient access to the subgap spectrum with reasonable photogeneration. The impurity type is chosen according to compensation of dopant in the base layer as done by Keevers and Green [9] and Yuan et al. [10]. Since p type IPV and base layers have been used in the simulated cell of Fig. 1, n type impurity has been used. For GaP, Ge (doped in Ga) can act as an n type impurity which will form an impurity energy level 0.204 ev below the conduction band as reported by Levinshtein et al. [16]. and in this case have been and respectively. For the particular composition of InGaAs used in this setup, Sze and Ng [20] indicate that the n type impurity (Ge) will form an energy level approximately 0.1 ev below the conduction band. and in this case have been and respectively. Electron and hole thermal capture cross sections for both GaP and InGaAs have been considered and respectively following Jayson et al. [21]. For electron and hole optical emission cross sections of the impurity in both of the materials, the value of has been considered following Mizuta and Kukimoto [22]. Both of the optical emission cross sections are assumed zero above bandgap energies in accordance with Keevers and Green [9]. 5. Results and Discussions Table 2- Performance Comparison of Solar Cell in terms of V oc, J sc and Fill Factor Solar Cell V oc (V) J sc (ma/cm 2 ) FF (%) No impurity 0.862 40.20 86.07 Impurity in GaP 0.86 40.07 86.63 Impurity in InGaAs 0.863 42.48 86.04 Impurity in both layers 0.862 42.35 86.70 - open circuit voltage (Voc), short circuit current density (Jsc) and Fill-Factor (FF) obtained for the two junction cell without impurity, impurity used separately in two sub-cells and impurity used simultaneously in both sub-cells. Fig. 2(b) shows I-V characteristics for the conventional two junction cell with no impurity and the best case when impurity is used in IPV layers of both sub-cells. All the simulations were performed at 300K and under the illumination of AM 1.5G, 100mW/cm 2. Simulations were performed without introducing any impurity in the cell, with impurity in the IPV layers of GaP and InGaAs separately and then, simultaneously. (a) (b) Fig. 2: Performance comparison of the cell with different impurity conditions in terms of- (a) Efficiency, (b) I-V characteristic

170 Md. Shahriar Parvez Khan and Esmat Farzana / Procedia Technology 7 ( 2013 ) 166 172 For impurity introduced with a concentration of Nt minor improvement of efficiency by 0.08% compared to conventional cell with no impurity. However, for impurity in InGaAs with the same concentration, Jsc 42.48 ma/cm 2 and 31.54% respectively which is a significant improvement in short circuit current density of 2.28 ma/cm 2 and efficiency of 1.71%. When impurity is introduced in both the layers with the same concentration at a time, the best efficiency is obtain with a value of 31.66% which is 1.83% more from the case of conventional no impurity cell. The fill -factor is also highest in this case with 86.7 %. It was said before that the short circuit current in a multijunction cell is determined by the sub-cell producing the lowest current. It is apparent from the result that in this case, InGaAs is that sub-cell. That is why impurity introduced in GaP does not result in any significant improvement. 6. Effect of Impurity Concentration on Solar Cell Performance To investigate the performance with variation of impurity concentration, simulation was done with impurity in both layers of the double junction cell but varying the concentration only in InGaAs since this sub-cell primarily determines effects of impurity. Fig. 3 shows Voc, Jsc consistent with those of single junction cell phenomena stated by Karazhanov [7], Keevers and Green [9] and Yuan et al. [10]. As shown in Fig. 1, InGaAs sub-cell p-type IPV layer doping was Na. Fro m Fig 3(a), it is evident that Jsc peaks just as the impurity almost compensates this background doping Na and then falls drastically. Fig 3(b) shows, Voc starts decreasing before this compensation occurs but not so abruptly. Fig 3(c) reveals that efficiency is held almost constant up to the point of compensation and then declines very fast. The reason behind this phenomenon is attributed to the dependence of minority carrier lifetime on Nt as shown by Karazhanov [23]: The three cases of Nt and Na are illustrated as below: At low trap density, Nt << Na, =, Na. As and Na are constant, Karazhanov [23] suggests that carrier lifetime becomes approximately independent of Nt. Hence, Jsc is almost constant for low Nt as seen from Fig 3(a). At high trap densities, however, free carriers are captured by the traps. When Nt = Na, all free holes generated from the acceptor level Na will be captured by the deep donors, resulting in an abrupt decrease of total free carrier density to a lower value which is of order of the intrinsic concentration. Hence, as given by respectively Karazhanov [7] and [23]. From Eq. 7, using a lower value of, a higher value of carrier lifetime is obtained leading to highest Jsc. This illustrates why Jsc goes on increasing having a peak value at Nt = Na. For Nt >> Na i.e. when donor Nt overcompensates the acceptor Na base IPV dopant, the conductivity in base IPV layer will be reversed from p type to n type and thus, holes become the minority carriers. Consequently, the free electron density will be increased and minority carrier lifetime is calculated for excess carrier Nt, as shown by Karazhanov [7]. Hence, carrier lifetime and consequently, Jsc is drastically reduced with Nt as predicted from Eq. 7 and obtained in Fig 3(a). On the other hand, Voc abruptly decreases after Nt = Na because of the increase of dark current caused by abrupt decrease of majority carrier concentration through compensation by deep-level impurity dopant and the increase of minority carrier density as indicated by Karazhanov [7]. Following these phenomena, the efficiency is relatively held constant when Nt << Na because Jsc is nearly constant although Voc has slightly decreased and declines rapidly after Nt = Na when both Voc and Jsc sharply degrades.

Md. Shahriar Parvez Khan and Esmat Farzana / Procedia Technology 7 ( 2013 ) 166 172 171 (a) (b) Fig. 3: Variation of cell performance with impurity concentration in terms of- (a) Short circuit current density J sc, (b) Open circuit voltage V oc, (c) (c) 7. Conclusion This work presents effect of impurity in double junction solar cells which, has been found to improve the cell performance. Non-midgap deep-level impurity has been incorporated separately and simultaneously in the two sub-cells of a double junction cell and improvements on short circuit current density and efficiency have been quantified. Moreover, it is obtained that impurity introduced in the sub-cell with the lower current has the dominant effect on improvement. The best conversion efficiency has been obtained with impurity in both sub-cells, which is as much as 1.83% higher than that of conventional double junction cell without impurity. The study suggests that significant short-circuit current improvements (up to 2.28 ma/cm 2 ) resulting from impurity incorporation can be effectively used to remove the problem of low shortcircuit current density in multijunction cells. Furthermore, change in cell performance in terms of open circuit voltage, sho rt

172 Md. Shahriar Parvez Khan and Esmat Farzana / Procedia Technology 7 ( 2013 ) 166 172 circuit current density and efficiency with the variation of impurity concentration have been found consistent with previous results reported in case of single junction cells. The results have been explained through dependence of carrier lifetime on trap concentration and conversion of conductivity resulting from compensation by deep-level impurity dopant. This is, to the best of knowledge, the first time that IPV effect has been studied in multijunction solar cell. Acknowledgements The authors would like to acknowledge the support provided by the Department of Electrical and Electronic Engineering, Ahsanullah University of Science and Technology (AUST) for conducting the research reported in the paper. The authors also express their gratitude, for providing open access to the software wxamps developed by Prof. Angus Rockett, Yiming Liu of University of Illinois at Urbana-Champaign and Prof. Stephen Fonash of Pennsylvania State University and to Mohammad Ziaur Rahman and Aminur Rahman, faculty of AUST, for providing fruitful suggestions. References [1] J. L. Gray, J. M. Schwarz, J. R. Wilcox, A. W. Haas, R. J. Schwartz, Peak efficiency of multijunction photovoltaic systems, In: Proc. 35th IEEE Photovolt. Spec. Conf., Honolulu, HI 2010, p. 002919 002923. [2] P. H. Huang, H. W. Wang, M. A. Tsai, F. I. Lai, S. Y. Kuo, Kuo, H. C. S. Chi, Optimum design of InGaP/GaAs/Ge triple-junction solar cells with subwavelength surface texture structure, In: Proc. 37th IEEE Photovolt. Spec. Conf., Seattle, WA 2011, p. 002071-002073. [3] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, E. D. Dunlop, Solar cell efficiency tables (Version 39), Prog. Photovolt: Res. Appl., 2012; 20:12 20. [4] H. Cotal H., C. Fetzer, J. Boisvert, G. Kinsey, R. King, P. Hebert, H. Yoon, N. Karam, III V multijunction solar cells for concentrating photovoltaics, Energy Environ. Sci.,2009, 2, 174 192. [5] X. Lu, M. B. Diaz, N. Kotulak, R. Hao, R. Opila, A. Barnett, Wide Band Gap Gallium Phosphide Solar Cells, IEEE Journal of Photovoltaics 2012, vol. 2, no. 2, p. 214-220. [6] M. Wolf, Limitations and possibilities for improvement of photo-voltaic solar energy converters, In: Proc. IRE, 1960, 48, p. 1246. [7] S. Zh. Karazhanov, Impurity photovoltaic effect in indium-doped silicon solar cells, J. Appl. Phys., 2001, 89 (7), p. 4030. [8] W. Shockley, H. J. Quiesser, Detailed Balance Limit of Efficiency of p-n Junction Solar Cells, J. Appl. Phys., 1961, 32, p. 510. [9] M. J. Keevers, M. A. Green, Efficiency improvements of silicon solar cells by the impurity photovoltaic Effect, J. Appl. Phys., 1994, 75, p. 4022. [10] J. Yuan, H. Shen, F. Zhong, X. Deng, Impurity photovoltaic effect in magnesium-doped silicon solar cells with two energy levels, Phys. Status Solidi A, 2012, 209, p. 1002-1006. [11] J. Yuan, Shen, H. Huang, X. Deng, Positive or negative gain: Role of thermal capture cross sections in impurity photovoltaic effect, J. Appl. Phys., 2011, 110, p. 104508. [12] M. Schmeits, A. A. Mani, Impurity photovoltaic effect in c-si solar cells. A numerical study, J. Appl. Phys., 1999, 85, p. 2207. [13] Y. Liu, Y. Sun, A. Rockett, A new simulation software of solar cells-wxamps, Solar Energy Materials and Solar Cells, 2012, 98, p. 124-128. [14] Y. Liu, D. Heinzel, A. Rockett, A Revised Version of the AMPS Simulation Code, In: Proc. 35th IEEE Photovolt. Spec. Conf., Honolulu, HI 2010, pp. 001943-001947. [15] B. G. Streetman, S. K. Banerjee, In: Solid State Electronic Devices, 6 th ed. Pearson, New Jersey, 2006 [16] Handbook Series on Semiconductor Parameters, vol.1, edited by M. Levinshtein, S. Rumyantsev, M. Shur, World Scientific, London, 1996. [17] W. Shockley, W. T. Read, Statistics of the Recombinations of Holes and Electrons, Phys. Rev., 1952, 87, p. 835. [18] R. N. Hall, Electron-Hole Recombination in Germanium, Phys. Rev., 1952, 87, p. 387. [19] Y. K. Hsieh, H. C. Card, Limitation to Shockley Read Hall model due to direct photoionization of the defect states, J. Appl. Phys., 1989, 65, p. 2409. [20] M. Sze, K. K. Ng, In: Physics of Semiconductor Devices, 3rd ed. Wiley, New York, 2007. [21] J. S. Jayson, R. Z. Bachrach, P. D. Dapkus, N. E. Schumaker, Evaluation of the Zn-0 Complex and Oxygen-Donor Electron-Capture Cross Sections in p-type GaP: Limits on the Quantum Efficiency of Red-Emitting (Zn, O)- Doped Material, Phys. Rev. B, 1972, 6, p. 2357-2372. [22] M. Mizuta and H. Kukimoto, Optical Cross Sections for the Zn-O Centerin GaP, Japan J. Appl. Phys., 1975, 4, p. 1617-1618. [23] S. Zh. Karazhanov, Mechanism for the anomalous degradation of silicon space solar cells, Appl. Phys. Lett., 2000, 76 (19), p. 2689.