Influence of using amorphous silicon stack as front heterojunction structure on performance of interdigitated back contact-heterojunction solar cell (IBC-HJ)

Abstract

Interdigitated back contact-heterojunction (IBC-HJ) solar cells can have a conversion efficiency of over 25%. However, the front surface passivation and structure have a great influence on the properties of the IBC-HJ solar cell. In this paper, detailed numerical simulations have been performed to investigate the potential of front surface field (FSF) offered by stack of n-type doped and intrinsic amorphous silicon (a-Si) layers on the front surface of IBC-HJ solar cells. Simulations results clearly indicate that the electric field of FSF should be strong enough to repel minority carries and cumulate major carriers near the front surface. However, the over-strong electric field tends to drive electrons into a-Si layer, leading to severe recombination loss. The n-type doped amorphous silicon (n-a-Si) layer has been optimized in terms of doping level and thickness. The optimized intrinsic amorphous silicon (i-a-Si) layer should be as thin as possible with an energy band gap (E_g) larger than 1.4 eV. In addition, the simulations concerning interface defects strongly suggest that FSF is essential when the front surface is not passivated perfectly. Without FSF, the IBC-HJ solar cells may become more sensitive to interface defect density.

Keywords

amorphous silicon / front surface field / simulations / interdigitated back contact-heterojunction solar cells

Introduction

Interdigitated back contact-heterojunction (IBC-HJ) solar cells have attracted much attention due to their inherent advantages [ 1]. Very recently, Kaneka Corporation has achieved the highest conversion efficiency of 26.33% in the world in a practical size (180 cm²) crystalline silicon solar cell [ 2]. Placing all the electrodes on the rear surface eliminates shadow loss, leading to a high J_sc and an independent optimization of the front surface [ 3]. Series resistance can be reduced by utilizing a high proportion of metallization area on the rear side. The avoidance of lateral current in the doped layer also contributes to the reduction of the internal series resistance. No series resistance degradation is observed under high intensity light, which makes it suitable for application in concentrated photovoltaics. In addition, in terms of module fabrication, IBC-HJ solar cells exhibit simplicity for interconnection and improved packing factor.

IBC-HJ structure requires a high minority carrier lifetime and exceptionally low surface recombination velocity because most electron-hole pairs are generated near the front surface and have to diffuse through the bulk of substrate to reach the rear surface in order to be collected by electrodes. Therefore, the effective front surface passivation is deemed to be critical to achieve high conversion efficiency. Traditionally, thermal silicon oxidation or SiN_x has been used as effective passivation layer [ 4, 5].

Compared with SiO₂ and SiN_x, the amorphous silicon (a-Si) deposited by plasma enhanced chemical vapor deposition (PECVD) has the advantage of low temperature (~200℃) deposition. The excellent passivating ability of intrinsic amorphous silicon (i-a-Si) has been shown by the great success of HIT solar cells [ 6].

In the present work, the application of stack of doped and intrinsic amorphous silicon layers has been proposed as FSF which integrates the heterojunction structure to IBC-HJ solar cells with the assistance of TCAD simulation tools. The preliminary experiments on such stack have shown that the passivation of crystalline silicon wafer is promising. The minority carrier lifetime of more than 2 ms demonstrates the excellent surface passivation capability of this FSF structure. In the simulation, the layer of n-type doped amorphous silicon (n-a-Si) has been optimized in terms of thickness and doping level. The intrinsic amorphous silicon (i-a-Si) layer has also been optimized with respect to thickness as well as energy band gap (E_g). In addition, the influence of interface defect density on crystalline Si (c-Si) substrate and i-a-Si layer is also investigated.

Simulation method

The simulations have been performed using the commercial device simulation package TCAD tools. A schematic cross-section of the IBC-HJ solar cell structure analyzed in this work is shown in Fig. 1.

A layer of n-a-Si is deposited following the deposition of i-a-Si on the front surface of the crystalline silicon (c-Si) substrate (5 W·cm (N_base=10¹⁵ cm^-3)) with a thickness of 300 mm. The stack of n-type doped and intrinsic a-Si layers acts as FSF. At the rear side of the device, the substrate is selectively doped as p-type emitter and n-type back surface field (BSF) whose lateral width is 1000 mm and 150 mm, respectively. Both donors and acceptors have a Gaussian distribution with a doping peak at the rear surface of the substrate. The junction depth of both the emitter and the BSF are 0.4 mm. The gap between the emitter and the BSF is set as 50 mm. On the outside of the emitter and BSF, the anode and cathode are formed by Aluminum contacts. Other default parameters and defect details of the a-Si used in the calculation are listed in Tables 1 and 2, respectively. The default structure parameters used in the simulations are listed in Table 1. The doping concentration of the emitter and the BSF refers to the Gaussian peak concentration. The parameters of the amorphous silicon are taken from Ref. [ 7]. The mid-gap states in amorphous silicon consist of the exponential band tail states near the valance and conduction band edge, and the Gaussian dangling band states in the middle of the band gap, as tabulated in Table 2 [ 8]. The position of the Gaussian distribution peak for acceptor-like states is measured from the conduction band edge, while the position of the Gaussian distribution peak for donor-like states is measured from the valance band edge. The parameters of the crystalline silicon are taken from the material library of the simulation tool used, while the parameters of amorphous silicon are provided by the Pennsylvania State University, which have been obtained through fitting specific device output characteristics and have been used by various groups in their computer simulation.

The Ray tracing model has been used to simulate the propagation of light in the solar cell. The photo generation rate is given by using

(1) G={\eta }_{\mathrm{0}}{\displaystyle \frac{P\lambda }{hc}}\alpha e^{-\alpha y},

whereh₀ is the internal quantum efficiency, P is the light intensity factor, l is the wavelength, h is the Planck’s constant, c is the speed of light, a is the absorption coefficient, and y is the relative distance from the ray to the specific grid point. The optical parameters of a-Si, c-Si and aluminum on the rear side are taken from SOPRA database which is a collection of complex refractive index for various materials sometimes as a function of temperature and composition.

Regarding the recombination mechanisms, the Shockley-Read-Hall (SRH) recombination is calculated using the concentration dependent lifetime model [ 9, 10]. The SRH recombination rate is given by [ 11]

(2) R_{\mathrm{SRH}}={\displaystyle \frac{np-n_i^{\mathrm{2}}}{{\tau }_p\{n+n_i\exp ({\displaystyle \frac{E_t}{kT}}\left)\right\}+{\tau }_n\{p+n_i\exp (-{\displaystyle \frac{E_t}{kT}}\left)\right\}}},

where {\tau }_p={\displaystyle \frac{{\tau }_{p\mathrm{0}}}{\mathrm{1}+{\displaystyle \frac{N_{\mathrm{acceptor}}}{\mathrm{5.0}\times {\mathrm{10}}^{\mathrm{16}}}}}},{\tau }_n={\displaystyle \frac{{\tau }_{n\mathrm{0}}}{\mathrm{1}+{\displaystyle \frac{N_{\mathrm{donor}}}{\mathrm{5.0}\times {\mathrm{10}}^{\mathrm{16}}}}}} ,τ_p0,and τ_n0 are the hole and electron life time, N_acceptor and N_donor are the concentration of acceptors and donors, n_i is the intrinsic carrier concentration, E_t is the trap level energy with respect to Fermi level, k is the Boltzman constant, and T is the absolute temperature.

The Auger recombination model are also considered. The Auger recombination rate is defined by [ 12]

(3) R_{\mathrm{Auger}}=C_{\mathrm{n}}(p{n}^{\mathrm{2}}-n{n}_i^{\mathrm{2}})+C_{\mathrm{p}}(n{p}^{\mathrm{2}}-p{n}_i^{\mathrm{2}}),

where C_n=2.8×10^–31cm⁶/s and C_p=9.9×10^–32 cm⁶/s are the Auger coefficient for electron and hole, respectively.

In addition, the Fermi-Dirac carrier statistics and concentration dependent mobility are considered in order to make simulations closer to real conditions [ 13]. The Carriers’ transport across the interface between the amorphous silicon and the crystalline silicon is modeled by the thermionic emission and thermionic field emission [ 14]. An AM 1.5G solar spectrum is used for the optical generation to simulate the J-V curve under standard one-sun illumination condition at an intensity of 100 MW/cm².

Results and discussion

Optimization of n-a-Si layer

First, a series of doping level, varying from 1×10¹⁵ cm^-3 to 1×10²¹ cm^-3, of the n-a-Si layer with different thicknesses (5 nm, 10 nm, 15 nm, 20 nm) are simulated systematically.

Figure 2 demonstrates the influence of the doping level of n-a-Si layers with different thicknesses on the efficiency of IBC-HJ solar cells. It can be clearly seen that the efficiency sharply increases when the doping level is close to 1×10¹⁸ cm^-3 for all investigated thicknesses of n-a-Si layers. The efficiency then decreases gradually as the doping concentration further increases. In addition, at a high doping level (e.g.≥5×10¹⁸ cm^-3), the cell efficiency exhibits a decrease when the thickness of the n-a-Si layer varies from 5 nm to 20 nm. Interestingly, an increase is observed at a lower doping level (e.g.<1×10¹⁸ cm^-3). This discrepancy can be explained by considering the electric field formed between the n-a-Si layer and the c-Si substrate.

Similar to the HIT (heterojunction with intrinsic thin layer) solar cells, when a stack of doped/intrinsic amorphous silicon layers is deposited on the c-Si substrate, a heterojunction was formed and the charge carriers could transport through this junction by diffusion or tunneling. Assuming a high doping level (≥5×10¹⁸ cm⁻³) in the n-a-Si layer, a high-low junction is formed between the heavily doped n-a-Si layer and the moderately doped c-Si substrate with a vertical build-in electric field pointing from the front surface into the substrate serving to repel holes to the rear surface and draw electrons to the front surface. In such a way, the electrons and the holes are separated and the recombination is lowered. In Fig. 3(a) and (c), the distribution of electric field strength is shown by different colors, while the total current densities are represented by different arrows whose directions and lengths symbolize the directions and magnitudes of total current densities. It can be clearly seen that near the front surface, the total current flows laterally from above the BSF region to above the emitter region resulting from the distribution gradients of charge carriers and the lateral electric field formed by the p-type emitter, the n-type substrate and the BSF. However, the total current is a combination of the electron current and the hole current. In order to analyze the composition of this lateral current, a cut line is made near the front surface above the middle of the gap between the emitter region and the BSF region. The current densities of electrons and holes are plotted separately along this cut line. Figure 3(b) shows that a high density of electron current is present near the front surface of the c-Si substrate, while the hole current density maintains at a low level, which is caused by the vertical electric field formed between the highly doped n-a-Si layer and the moderately doped c-Si substrate. As abundant electrons accumulate and transport near the front surface, the conductivity of this accumulation layer increases. Hence, the series resistance loss is reduced and FF gets improved. Similar phenomenon has also been observed in diffused FSF of back-junction back-contact solar cells where majority carriers take advantage of the high conductivity of the front diffused n⁺ layer, resulting in a reduced resistance loss [ 14].

However, Fig. 3(c) and (d) clearly show that when the doping level is as high as 10²¹ cm^-3, the vertical electric field gets so strong that a considerably high proportion of lateral electron current is drawn into the n-a-Si layer. In contrast to the c-Si substrate, the n-a-Si layer is a highly defective layer. The SRH recombination is the dominant recombination mechanism in material with a high density of states present within the band gap. According to Eq. (2), it is clear that the defects within the n-a-Si layer with a density as high as that defined in Table 2 must induce a serious recombination loss. This explains the comparative decrease in efficiency when the doping level is higher than 5×10¹⁸ cm^-3.

Once the doping level is high enough (≥5×10¹⁸ cm^-3), the electric filed is independent on the thickness of the n-a-Si layer. In this case, the excessive thickness of the n-a-Si layer will cause more recombination loss and optical loss. Figure 4 shows the internal quantum efficiency (IQE) data of solar cells with n-a-Si layers of various thicknesses, and an identical doping level of 10²¹ cm^-3. The spectral response in short wavelength drops dramatically with the increase in the thickness of the n-a-Si layer, which confirms the increasing recombination loss and optical loss within the n-a-Si layer. But with the certain surface recombination, the IQE reduction in short wavelength can be ascribed to the light absorption of amorphous silicon, including the intrinsic layer (i-a-Si) and the doped layer (n-a-Si).

Figure 5 shows the distribution of electric filed and current density near the front surface of n-a-Si layers with a low doping level and different thicknesses. For a doping level of 10¹⁵ cm^-3, the electric field is reversed with direction pointing from the c-Si substrate to the front surface. Therefore, the electrons are repelled, while holes are accumulated near the front surface, which similarly reduces the surface recombination slightly. At this doping level, the strength of electric field depends on the thickness of the n-a-Si layer. When the thickness is 20 nm, a stronger electric field can be seen clearly in the i-a-Si layer and the n-a-Si layer. However, when the thickness drops to 5 nm, the electric field strength weakens comparatively. In other words, at such a doping level, a thicker n-a-Si layer means more FSF effect.

The strength of the electric field between the lightly doped n-a-Si layer, with doping levels ranging from 10¹⁵ cm^-3 to 10¹⁷ cm^-3, and the c-Si substrate weakens with the increasing doping level of the n-a-Si layer. As mentioned above, the weakening of the strength of the electric field of FSF inevitably results in the drop in efficiency. In addition, the efficiency decrease is partly caused by the band gap narrowing effect [ 15]. The band gap of the n-a-Si layer drops with the increase in the doping level. This effect causes a longer wavelength light to be absorbed by the n-a-Si layer. Obviously, the thicker the n-a-Si layer, the more significant the optical loss caused by over doping, which explains the faster decrease rate of efficiency of IBC-HJ solar cells with the 20 nm n-a-Si layer.

Optimization of i-a-Si layer

In the heterojunction solar cells, a layer of i-a-Si is usually inserted between the doped a-Si layer and the c-Si substrate. This i-a-Si layer helps to reduce the dangling bonds on the surface of the c-Si substrate and then impede the recombination of photon generated carriers near the front surface, which greatly improve the open-circuit voltage (V_oc) and efficiency (E_ff) of solar cells [ 16]. In the simulations, a layer of i-a-Si is added for the same reason. The effect of the thickness of the i-a-Si layer on the performance of solar cells has been studied.

In order to calculate the performance of the cell, the I-V characteristic of the device is simulated under a light beam with standard AM 1.5 spectrum. Figure 6 shows the influences of the thickness of the i-a-Si layer on the performance of IBC-HJ solar cells. The increase in the thickness of the i-a-Si layer can barely affect FF, but it deteriorates the J_sc and V_oc of the IBC-HJ solar cell.

Figure 7 shows the IQE data of solar cells employing i-a-Si layers of different thicknesses. The spectral response in short wavelength drops dramatically as the thickness of the i-a-Si layer increases. This result can be attributed to the optical loss as well as the recombination loss in the thick i-a-Si layer larger than that in the thin i-a-Si layer. I-a-Si is a defective layer full of recombination centers, the light absorbed by the i-a-Si layer can hardly contribute to the photocurrent of solar cells. Moreover, the i-a-Si layer is sandwiched between the doped n-a-Si layer and the c-Si substrate. The electric field in the high-low junction crosses through the i-a-Si layer, which gets weakened by the increase of the thickness of the i-a-Si layer, resulting in the deterioration of the performance of solar cells.

Therefore, the i-a-Si layer should be as thin as possible to avoid the great optical loss and recombination loss.

Due to the difference of the energy band gap between a-Si and c-Si, a heterojunction structure forms as long as a-Si is deposited on the surface of c-Si. Figure 8 shows the influences of the E_g of the i-a-Si layer on the performance of solar cells. It can be seen clearly that in order to obtain a high efficiency, the E_g should be larger than 1.4 eV. This result shows the necessity and intrinsic advantage of the heterojunction acting as the FSF of IBC-HJ solar cells.

The thermionic emission and the field emission transport model have been applied along the interface of the heterojunction of i-a-Si/c-Si. The hole current density resulting from the thermionic emission and the field emission from the c-Si substrate to the i-a-Si layer is defined as

(3) \overrightarrow{J_p}=q{v}_p(\mathrm{1}+\delta )p\exp \left({\displaystyle \frac{-\mathrm{\Delta }{E}_{\mathrm{V}}}{kT}}\right),

where v_p is the hole thermal velocity, δ represents the contribution due to field emission, p is the concentration of holes in c-Si near the interface, ΔE_V is the valance band change from c-Si to i-a-Si. It is assumed that \overrightarrow{J_{p\mathrm{1}}} and \overrightarrow{J_{p\mathrm{2}}} are the densities of hole currents due to the thermionic emission and the field emission at the interface of i-a-Si/c-Si when E_g of i-a-Si is 1.2 eV and 1.4 eV, respectively. According to Eq. (3), {\displaystyle \frac{\overrightarrow{J_{p\mathrm{1}}}}{\overrightarrow{J_{p\mathrm{2}}}}}=\exp \left({\displaystyle \frac{\mathrm{\Delta }{E}_{\mathrm{V}\mathrm{2}}-\mathrm{\Delta }{E}_{\mathrm{V}\mathrm{1}}}{kT}}\right)\approx \mathrm{2980} . It is obvious that the hole barrier formed when E_g of i-a-Si is 1.4 eV effectively hampers the hole emission into the i-a-Si layer, as compared with the case when E_g of i-a-Si is 1.2 eV. Therefore, there is no significant improvement in the performance of solar cells when a hole barrier forms as the energy band gap of i-a-Si increases to 2 eV.

The energy band gap of a-Si can be tuned from 1.4 eV to 1.9 eV depending on the deposition conditions and thicknesses. Cody et al. reported that the optical band gap of hydrogenated a-Si is controlled by the amount of disorder, and that hydrogen affects the band gap through its effect on disorder [ 17]. Yaser Abdulraheem et al. observed that the optical band gap of a-Si increases as the thickness decreases, which can be attributed to the nano-clusters present being close to the interface of a-Si/c-Si and quantum confinement effects [ 18].

Influences of interface states

The heterojunction structure of FSF features an abrupt discontinuity of crystal network at the c-Si surface to the a-Si layer, which results in a large density of defects in the band gap due to a high density of dangling bonds. The recombination loss of charge carriers at the interface is mainly controlled by these interface states. In order to evaluate the effect of such interface recombination, a layer of defective silicon with a thickness of 1 nm is introduced at the interface between the i-a-Si layer and the c-Si substrate. The defect distribution in this thin layer is defined as Gaussian with donor-like defects located at 0.5 eV above the valance band edge and acceptor-like defects located at 0.5 eV below the conduction band edge. The interface state density, N_ss, is determined from the thickness of the defective silicon layer, d_int, and the total density of defects in Gaussian distribution, N_it, by N_ss=d_intN_it. The capture cross sections of electrons and holes are set as 10^-17 cm².

It can be seen from Fig. 9 that the performances of solar cells deteriorate dramatically when N_ss is higher than 10¹² cm^-2·eV^-1 due to the great recombination loss at the front interface. Unlike the traditional solar cells having electrodes on the front surface, front recombination is extraordinarily important for the IBC-HJ solar cell since photon-generation takes place mostly near the front surface, while most carriers are collected on the rear surface. The suppression of front surface recombination is the prerequisite to achieve high efficiency for IBC-HJ solar cells with a high level of interface state density.

To investigate the necessity of the FSF layer, the same simulations on a similar solar cell structure without the n-a-Si layer have been performed. In other words, only a 5 nm i-a-Si layer is deposited on the c-Si substrate.

Figure10 presents the dependence of performances of solar cells upon N_ss. Once the n-a-Si layer is eliminated, solar cells become more sensitive to N_ss. The efficiency of the solar cell starts to decay as long as N_ss is more than 10¹⁰ cm⁻²·eV⁻¹ which is much lower than that of the solar cells with FSF. Provided that the surface of the c-Si substrate gets well passivated, the n-a-Si layer and FSF are not crucial parts any more. For instance, with N_ss lower than 10¹⁰ cm⁻²·eV⁻¹, the efficiency of solar cells without FSF are actually higher than that of solar cells with FSF. This conclusion is in agreement with previous researches. Various research groups have reported IBC-HJ solar cell without FSF, passivating front surface solely with a passivation layer such as SiN_x layer or stack of amorphous silicon/silicon nitride layer [ 19]. The outstanding achievement of a conversion efficiency of 24.4% is also based on IBC-HJ structure without FSF.

Consequently, it is necessary to improve front surface passivation using FSF when the front surface of the c-Si substrate could not be passivated well. Only when the interface defect density is kept lower than 10¹⁰ cm⁻²·eV⁻¹ could FSF be eliminated.

Conclusions

The FSF of the IBC-HJ solar cell can be replaced by the deposition of an i-a-Si layer and an n-a-Si, instead of diffusion junction. With the replacement, a high conversion efficiency can be obtained. The systematic simulations show that the doping level of n-a-Si should be high enough to form a strong electric field between the c-Si substrate and the a-Si layers. On the other hand, an over-strong electric field will draw electrons into the n-a-Si layer, leading to a serious recombination loss of charge carriers. The optimized doping level for the n-a-Si layer is 5×10¹⁸ cm^-3. Both doped and intrinsic a-Si layers should be as thin as possible to avoid the recombination loss and optical loss. From energy band diagrams, it can be concluded that the E_g of i-a-Si has to be larger than 1.4 eV to avoid pulling holes into defective a-Si layers and deteriorating the FSF effect. In addition, the influence of interface states is revealed. Simulations indicate that FSF is crucial for the substrate with a high interface state density (higher than 10¹⁰ cm^-2·eV^-1). For well passivated substrate, the FSF is not essential any more. The simulations demonstrates the possibility and potential of FSF formed by stack of i-a-Si/n-a-Si on the front surface of IBC-HJ solar cells.

References

[1] Saint-Cast P, Padilla M, Kimmerle A, Reichel C. An analytical model for interdigitated back contact-heterojunction solar cells. IEEE Journal of Photovoltaics, 2014, 4(1):114–121

[2] World’s highest conversion efficiency of 26.33% achieved in a crystalline silion solar cell. 2016,

[3] Kerschaver E V, Beaucarne G. Back-contact solar cells: a review. Progress in Photovoltaics Research & Application, 2006, 14(2):107–123

[4] Aberle A G. Surface passivation of crystalline silicon solar cells: a review. Progress in Photovoltaics Research & Application,  2000,  8(5):473–487

[5] Peibst R, Harder N, Merkle A, Neubert T, Kirstein S, Schmidt J. High-efficiency RISE IBC-HJ solar cells: influence of rear side passivation on pn junction meander recombination. In: Process 28th Europe Photovoltaic Solar Energy Conference Exhibition. Paris, France, 2013

[6] Taguchi M, Yano A, Tohoda S, Matsuyama K, Nakamura Y, Nishiwaki T. Record efficiency HIT solar cell on thin silicon wafer. IEEE Journal of Photovoltaics,  2014,  4(1):96–99

[7] Van Sark W, Korte L, Raco F. Physics and Technology Amorphous-Crystalline Heterostrcutre Silicon Solar Cells. Berlin: Springer-Verlag Berlin and Heidelberg GmbH & Co. KG, 2013

[8] Kanevce A, Metzger WK. Role of amorphous silicon and tunneling in heterojunction with intrinsic thin layer (HIT) solar cells. Journal of Applied Physics. 2009, 105(9): 094507–094507-7

[9] Shockley W, Read WT. Statistics of the recombinations of holes and electrons. Physical Review,  1952,  87(5):835–842

[10] Hall R N. Electron-hole recombination in germanium. Physical Review,1952, 87(2):387

[11] Fossum J G, Lee D S. A physical model for the dependence of carrier lifetime on doping density in nondegenerate silicon. Solid-State Electronics,  1982,  25(8):741–747

[12] Dziewior J, Schmid W. Auger coefficients for highly doped and highly excited silicon. Applied Physics Letters,  1977,  31(5):346–348

[13] Slotboom J W, Graaff H C D. Measurements of bandgap narrowing in Si bipolar transistors. Solid-State Electronics,1976,  19(10):857–862

[14] Wagner J, Del Alamo J A. Band-gap narrowing in heavily doped silicon: acomparison of optical and electrical data. Journal of Applied Physics,  1988,  63(2):425–429

[15] Tsunomura Y, Yoshimine Y, Taguchi M, Baba T, Kinoshita T, Kanno H. Twenty-two percent efficiency HIT solar cell. Solar Energy Materials & Solar Cells,  2009,  93(6–7):670–673

[16] Cody G, Tiedje T, Abeles B, Brooks B, Goldstein Y. Disorder and the optical-absorption edge of hydrogenated amorphous silicon. Le Journal De Physique Colloques,  1981,  92(20):1480–1483

[17] Abdulraheem Y, Gordon I, Bearda T, Meddeb H, Poortmans J. Optical bandgap of ultra-thin amorphous silicon films deposited on crystalline silicon by PECVD. AIP Advances,2014, 4:057122

[18] Fahrner W R. Amorphous Silicon/Crystalline Silicon Heterojunction Solar Cells. Berlin: Springer Berlin Heidelberg,  2013,  90(12):G1117–G1121

[19] Haase F, Eidelloth S, Horbelt R, Bothe K, Rojas E G, Brendel R. Loss analysis of back-contact back-junction thin-film monocrystalline silicon solar cells. Journal of Applied Physics,  2011,  110(12):002874–002877

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg