Insights into the role of the interface defects density and the bandgap of the back surface field for efficient p-type silicon heterojunction solar cells

Abstract

An n-type silicon wafer appears to be an excellent base material for high efficiency silicon heterojunction solar cells with high annual energy output. The further development of viable p-wafer cells would open many industrial photovoltaic options. Here, the role of the interface defects density and the bandgap of the back surface field of p-type silicon heterojunction solar cells have been investigated in detail to provide guidelines for achieving high performance. The result indicates that the output characteristics of the heterojunction solar cells are sensitive to the density of interface defects (D_it) at both sides of the c-Si surface. However, the output parameters primarily affected by D_it at the front and rear side are not identical. Back surface field (BSF), which aims to reduce photo-generated carriers' recombination at the rear interface, has also been optimized by adjusting the bandgap to improve the collection efficiency. Finally, based on experimental results, we propose that the conversion efficiency of p-type silicon heterojunction solar cells could be increased beyond 23% by efficiently regulating the bifacial D_it and the bandgap of the BSF layer.

---

1. Introduction

Silicon heterojunction (SHJ) solar cells constructed with ultrathin hydrogenated amorphous silicon (a-Si:H) on a crystalline silicon (c-Si) absorber layer are promising candidates for high-efficiency, low-cost solar cells. Compared with traditional monocrystalline silicon solar cells, SHJ solar cells generally exhibit higher open-circuit voltages (V_oc) because wide-band-gap a-Si:H is used as the emitter layer and the metal contact is separated from the wafer by a-Si:H films and indium tin oxide (ITO), which reduces the amount of interface recombination.^1–3 Moreover, SHJ solar cells could be fabricated with low-temperature or photolithographic-free processes, since passivation of the c-Si wafer surfaces and formation of the emitter and back surface field (BSF) are all performed with thin intrinsic and doped a-Si:H layers, usually deposited by low-temperature plasma-enhanced chemical vapor deposition.^4,5

So far, the highest efficiency silicon heterojunction solar cells were obtained using n-doped, rather than p-doped, c-Si wafers.⁶ However, n-type wafers are still less readily available on the market than p-type wafers, which are widely used for the production of conventional (diffused-emitter) silicon solar cells. The possibility to produce high-efficiency heterojunction solar cells from p-type wafers is, thus, of great practical interest. The p-type SHJ solar cells have already been demonstrated by several groups but with efficiencies significantly lower than those achieved on n-type wafers.^3,5,6 For p-type silicon heterojunction solar cells, the performance is also more sensitive to the density of a-Si:H/c-Si interface defects, not only because the p–n junction is across two different materials, but also owing to the asymmetry between the conduction and valence band offsets.⁷ Yet up to now, there still have unified understanding about the different influences caused by the D_it at the front and rear sides of the p-type SHJ solar cells. Thus, it is necessary to understand the role of the D_it and the bandgap alignment for p-type SHJ solar cells in order to further improve the performance and accelerating its practical application.

Numerical analysis describe the basic phenomena present in photovoltaic devices, allowing intuitive examination of each parameter in solar cells and thus identifying the optimal conditions for operating. In this work, we make an insight into the relationship of the D_it at the front and rear sides of the crystalline silicon by an extended Shockley–Read–Hall formalism.⁸ Our results present that the D_it at front side is primarily affected the V_oc of the p-type SHJ solar cells, while the D_it at the rear side could act upon the FF and J_sc of the device. Also, in order to achieve a good function of photo-generated carriers' collection, the bandgap of the BSF should be adjusted in the range of 1.7–1.8 eV to achieve the efficiency more than 23%.

2. Numerical modeling

In order to give an accurately insight into the sensitive parameters for p-type SHJ solar cells, analytical calculations are unavoidable tool to model the behavior of a full cell architecture especially under equilibrium conditions. The Poisson and the continuity equations for electrons and holes are solved considering Shockley–Read–Hall recombination statistics for the a-Si:H and for the defect layer at the interface between a-Si:H and c-Si. In practical manufacturing process, D_it can be suppressed by wet chemical cleaning of the c-Si surface followed by H-terminated and the growth of a-Si:H, which saturates the silicon dangling bonds with atomic hydrogen.^9–11 Garín et al.¹² suggested that the Shockley–Read–Hall formalism, which has been widely used to depict c-Si surface passivation schemes by SiO₂ and a-SiN_x:H, can model the i-a-Si:H/c-Si interface recombination, in which the i-a-Si:H thin film dangling bond states in the proximity of abrupt i-a-Si:H/c-Si interface can be considered to be an effective interface and included in D_it. Nevertheless, the charge effect of doped a-Si:H layer defects on the interface recombination was eliminated, and D_it was oversimplified by a single level defect state at mid-gap. Shu et al.¹³ extend the Shockley–Read–Hall approach to more accurately model the interface recombination through a continuously distributed D_it, with the effect of doped a-Si:H defect states also taken into account. Herein, we further extend this comprehensive Shockley–Read–Hall model in this work by inducing a defect layer to describe injection level dependent surface recombination velocity (S_it).

The complete architecture of ITO/n-a-Si:H/i-a-Si:H/p-c-Si/i-a-Si:H/p⁺-a-Si:H/Ag and corresponding bandgap schematic diagram in this numerical analysis are exhibited in Fig. 1a and b. We performed numerical simulations of p-type SHJ solar cells with efficient tool named AFORS-HET.¹⁴ The basic parameters in this numerical modeling, which extract from different state-of-the-art SHJ solar cells experimental results,^6,15–18 are shown as Table 1. The absorber of the solar cell is standard 300 μm thick p-type c-Si, with an acceptor density of 1.5 × 10¹⁶ cm⁻³. It is characterized by a single donor-type defect centered at mid-gap with a density of 10¹⁰ cm⁻³ and the capture cross section for holes and for electrons is equal to 10⁻¹⁴ cm². The density of states of the n-a-Si:H emitter layer and i-a-Si:H passivation layer, including two Gaussian distributions of deep defects and the valence and conduction band tails states are exhibited in Fig. 1c and d. An 80 nm ITO is used to be the transparent conductive oxide layer and a thin c-Si layer (3 nm) with average distribution continuous donor and acceptor defect states in the band gap has been inserted between the i-a-Si:H and c-Si to substitute the interface defects. The density of interface defects translates into a volume defects density by:

d × S × ρ_volume = S × ρ_interface (1)

where d is the thickness of the volume defects layer, S is the surface area of the c-Si, ρ_volume and ρ_interface are the defect states density at the volume and surface of c-Si respectively. The surface recombination velocities of the electrons and holes are both set to 1.0 × 10⁷ cm s⁻¹. The solar AM1.5 radiations are adopted as the illuminating source with the power density of 100 mW cm⁻² and the ITO/n-a-Si:H contact is fixed as flat band in calculations.

Fig. 1 (a) Schematic of the SHJ solar cells for numerical analysis. (b) Schematics of energy diagram of p-type SHJ solar cell at equilibrium condition. The dashed lines are Fermi energy (E_F), conduction band offset (ΔE_c) and valence band offset (ΔE_v), respectively. Distribution of gap defects in (c) n-a-Si:H and (d) i-a-Si:H layer include two Gaussian distributions of deep defects and the band tails states of the valence band and conduction band.

Table 1 Main parameter values adopted for SHJ solar cells in this study

Parameters	a-Si:H(n)	a-Si:H(i)	a-Si:H(p^+)	c-Si(p)
Layer thickness (nm)	10 nm	5 nm	5 nm	300 μm
Electron affinity (eV)	3.9	3.9	2.9	4.05
Mobility gap (eV)	1.80	1.80	Variable	1.12
Optical band gap (eV)	1.72	1.72	1.72	1.12
Effective density of states Nc/Nv (cm^−3)	10^20/10^20	10^20/10^20	10^20/10^20	2.8 × 10^19/2.6 × 10^19
Doping concentration (cm^−3)	5 × 10^20	Nd = 100	9 × 10^20	1.5 × 10^16
Electron mobility (cm^2 V^−1 s)	5	5	5	1041
Hole mobility (cm^2 V^−1 s)	1	1	1	412.9
Standard deviation in Gaussian states (eV)	0.21	0.21	0.21	—
Electron capture cross (cm^2)	7 × 10^−16	7 × 10^−16	3 × 10^−15	1 × 10^−14
Hole capture cross (cm^2)	7 × 10^−16	7 × 10^−16	3 × 10^−14	1 × 10^−14
Urbach energies Etail,V and Etail,C (eV)	0.094/0.068	0.094/0.035	0.12/0.08	—

---

3. Results and discussion

Role of the bifacial D_it for p-type SHJ solar cells

Recombination through defect levels in semiconductors is usually described by the Shockley–Read–Hall theory in this work. The surface recombination rate (N_s) of solar cells can be given by,¹⁹Where σ_n, σ_p and n_s, p_s are the capture cross-sections and concentrations of electrons and holes respectively. v_th is the thermal velocity. n_l and p_l are the statistical factor for electrons and holes respectively. Therefore, the S_it could be given as,where Δn_s is the excess carrier density at the surface. For SHJ solar cells, the V_oc is related to the interface recombination via:²⁰kT is the thermal energy, ϕ_B is the effective barrier, q is the elementary charge, N_v is the effective density of states in the valence band. According to eqn (3) and (4), it can be concluded that (n_sp_s − n_i²) and D_it affect the performance of the solar cells by determining the V_oc. So we assumed different D_it at the front side of c-Si surface to depict the corresponding influences on device, thus providing the threshold of the D_it to maintain high performance p-type SHJ solar cells.

Fig. 2a shows the band diagrams with different values of D_it at the front side of the heterojunction under illumination. The interface charge (Q_it) by trapping carrier is related to D_it (E) and the difference in occupancy by electrons and holes via¹³

where the D^A_it(E) and D^D_it(E) are the continuous distributed acceptor-type and donor-type D_it(E) within c-Si bandgap. The f_D(E) and f_A(E) are the occupancies of the donor and acceptor defects states, respectively, expressed by

Fig. 2 Band diagrams and quasi-Fermi levels of the heterojunction in (a) front and (b) rear side of the p-type SHJ solar cells as the D_it increased from 10⁹ cm⁻² eV⁻¹ to 10¹³ cm⁻² eV⁻¹. ψ₁, ψ₂, and ψ₃ are the equilibrium band bending in c-Si when the D_it is 10⁹, 10¹², and 10¹³ cm⁻² eV⁻¹, respectively. (c) The photo-generated holes and electrons current density (J_n and J_p) as a function of position in the SHJ solar cells when the D_it at rear side is 10¹³cm⁻² eV⁻¹. (d) The EQE spectrum of the p-type SHJ solar cells with different D_it at rear side of c-Si.

Since the minority carriers in p-type are electrons, which have larger capture cross sections than holes, the high D_it will increase the number of Q_it by trapping carriers. If Q_it can become comparable with the charges in the a-Si:H and c-Si, and hence the bands need to realign to keep the system charge balance.²¹ As a result, the band bending (ψ) in c-Si is reduced (ψ₁, ψ₂, and ψ₃ are the equilibrium band bending in c-Si when the D_it are 10⁹, 10¹², and 10¹³ cm⁻² eV⁻¹, respectively). Meanwhile, the recombination possibility at the a-Si:H/c-Si heterointerface will increase with the reduction of the band bending, due to the declining of the driven force in c-Si for separating photo-generated carriers.²²

The influences of D_it on the rear-side heterojunction band bending have been exhibited in Fig. 2b. As described above, the number of trapped holes will be much larger than that of electrons, and thus creating positively Q_it. Similar with the front side, when the magnitude of Q_it can be closed to the ionized donor in the depletion region (e.g., 10¹³ cm⁻² eV⁻¹), c-Si energy band at rear interface even can bend downward. Fig. 2c shows that the photo-generated holes and electrons current density (J_n and J_p) as a function of position in the p-type SHJ solar cells when the D_it of the rear hetero-interface is 10¹³ cm⁻² eV⁻¹. The negative partial of J_n (circled in green dash) means the opposite diffused electrons from the front side of c-Si to the rear side of it owing to the band bend downward.²³ Apparently, the negative J_n is controversial to the operating electron current of the solar cells, and will increase the recombination probability with the majority carries (holes) at the rear p-c-Si/p⁺-a-Si:H interface. Fig. 2d shows the external quantum efficiency (EQE) spectrum of the solar cells. The response at long wavelength has obviously decreased when D_it increased to 10¹³ cm⁻² eV⁻¹, which is accordance with the opinion that the recombination at rear c-Si/a-Si:H interface become more seriously.

Fig. 3 shows the cells' performances as a function of the D_it at front and rear sides of the c-Si, respectively. Fig. 3a shows that V_oc significantly decreases about 300 mV as the D_it at the front i-a-Si:H/p-c-Si interface increases from 10⁹ cm⁻² eV⁻¹ to 10¹³ cm⁻² eV⁻¹ because of the increased interface recombination. While the V_oc is not seriously affected as the increasing D_it of the rear side. However, it could be found that the fill factor (FF), short-circuit current density (J_sc), and conversion efficiency (Eff) have been seriously decreased when the D_it of the rear side increase to 10¹³ cm⁻² eV⁻¹.

Fig. 3 The V_oc, J_sc, FF and Eff of p-type solar cells as functions of the D_it at the front side and rear side of the solar cells.

The reason for this phenomenon could be ascribed to the relation between the D_it and the trapped charges at hetero-interface. As concluded from eqn (3) and (4), the D_it could directly impact on the V_oc of the solar cells by changing band bending the recombination velocity of the front-side hetero-interface. Compared with the influences of front D_it on the device, the D_it of rear side has relatively small impact on the V_oc. As most of the photons are absorbed at the front side of c-Si, the entire photo-generated carriers are separated (and transporting) basically rely on the built-in field established by front n-a-Si:H and p-type c-Si.^24,25 Heterojunction established at the rear side is not the decisive factor for the V_oc of the cells. The main function of the p⁺-a-Si:H BSF is to enhance carriers' selective holes-collection and transportation. Hence, the changed energy band structure of the rear p-c-Si/p⁺-a-Si:H junction has a relatively small impact on V_oc compare with the one of n-a-Si:H/p-c-Si junction. However, the FF and J_sc are more sensitive to the rear-side D_it because of the band bending downward, which caused by high D_it (as shown in Fig. 2b), could induce the opposite electron current and creating a barrier for majority carrier (holes) transportation, thus increasing the series resistance of the p-type SHJ solar cells.

Analytical BSF's bandgap of p-type SHJ solar cells

The BSF layer should be included to obtain good cell performance. The offsets of the conduction band and valence band at heterojunction are:

ΔE_c = χ_c-Si − χ_a-Si:H (8)

ΔE_v = E_ga-Si:H − χ_c-Si + χ_a-Si:H − E_gc-Si (9)

where χ_c-Si, χ_a-Si:H, E_gc-Si and E_ga-Si:H are the electron affinity and bandgap of c-Si and a-Si:H. The electron affinity of c-Si and a-Si:H is 4.05 eV and 3.9 eV, respectively. In practical experiment, the bandgap of the BSF could be adjusted by changing certain parameters of the p⁺-a-Si:H layer, such as the hydrogen content, doping concentration and crystallization.^26,27 Fig. 4 shows the performance of the p-type SHJ solar cells as a function of BSF's bandgap. The V_oc of the solar cells is improved significantly as the bandgap increased from 1.2 to 1.9 eV. The J_sc and FF of the solar cells are almost unaffected with the increasing of the bandgap of the BSF below 1.8 eV. While a pronounced decreasing could be found for J_sc and FF if further increasing the bandgap of the BSF to 1.9 eV.

Fig. 4 The V_oc, J_sc, FF and Eff of p-type solar cells as functions of the BSF's bandgap.

The band diagrams of the rear side p-c-Si/p⁺-a-Si:H structure are shown in Fig. 5, considering the BSF bandgap of 1.3 eV and 1.8 eV, respectively. The wide bandgap (1.8 eV) of BSF with a large conduction band offset could achieve good surface field for the rear-side diffusing electrons of the device.^26–28 However, the valence band offset also increased, and hence establishing a large barrier for holes to flow through the rear contact. However, in contrast, the low bandgap of the BSF layer resulted in band bend downward and small conduction band offset. As illustrated for Fig. 2b, this is detrimental for electrons diffused forward and will increase the recombination possibility at rear surface.

Fig. 5 Band diagrams of the rear-side heterojunction when the bandgap of the BSFs are (a) 1.8 eV and (b) 1.3 eV, respectively.

Fig. 6 shows the EQE variation with bandgap to further explain the previously described results. With the bandgap increased from 1.3 to 1.72 eV, the spectrum response at longer wavelengths increased due to the suppressing of the recombination, which caused by enhancing back surface field effect. However, further increasing the bandgap will decrease the spectrum response owing to the large valence band offset blocking the majority holes' transportation. Thus, it can be concluded that there exists a trade-off between surface field effect and carrier transportation barrier, and it could be relaxed by carefully regulating the bandgap of the BSF layer. Also, the numerical calculation results indicate that the optimum bandgap of BSF should be controlled in the range of 1.7 to 1.8 eV for p-type SHJ solar cells in practical processing.

Fig. 6 The EQE curves of the p-type SHJ solar cells when the bandgap of the BSF layers variant from 1.3 eV to 1.8 eV.

Finally, as shown in Fig. 7, when the defect states density at both side of the c-Si is 10¹⁰ cm⁻² eV⁻¹, and the bandgap of the BSF layer is 1.72 eV (the thickness of a-Si:H emitter layer and intrinsic layer is set as 5 nm and 3 nm respectively. The doping concentration of n-a-Si:H and p⁺-a-Si:H layers are 6 × 10²¹ cm⁻³ and 9 × 10¹⁹ cm⁻³ respectively), an efficiency of the 23.27% could be achieved. The performance of the solar cells as a function of the c-Si thickness has also been study, which has been shown in Fig. S1 and S2.† It should be noted that this results are obtained only based on planar p-type substrate without incorporating any antireflection coating and light trapping architecture. Hence, for p-type SHJ solar cells, a more than 24% efficiency could be expected with fully designed light-management strategies in practical application. Based on the performance and cost analysis, it is observed that besides n-type solar cells, the simple and low-cost p-type SHJ solar cells with optimized D_it and BSF layer may also be potential candidate for high efficiency solar cell technology.

Fig. 7 Illuminated current density–voltage curve of the champion cell. All the optimized parameters are extracted from the practical experimental results.

4. Conclusions

In summary, the influence of interface defects at both side of c-Si and bandgap of BSF layer on p-type SHJ solar cell performance were investigated though numerical calculations. Hence, suggestive design parameters for p-type SHJ solar cell fabrication are proposed. The result shows that D_it at front and rear side of c-Si should be controlled lower than 10¹³ cm⁻³ to avoid significantly negative impact on the solar cells performance. Meanwhile, D_it at front side directly determine the values of V_oc. While D_it at the rear side of the solar cell seriously affects the J_sc and FF. Appropriate designation for the bandgap of the BSF (1.7–1.8 eV) can improve the efficiency of the p-type SHJ solar cells by relaxing the trade-off between surface field effect and carrier transportation barrier. Finally, a conversion efficiency of more than 23% could be achieved by adopting the parameters extracted from experimental results. This highlights that SHJ solar cells based on low-cost p-type c-Si may also be the market-leading products owing to the improvements on cost-to-performance ratio for c-Si photovoltaics.

Acknowledgements

The authors gratefully acknowledge the support from National Nature Science Foundation of China (Grant No. 61475063, 61505067 and 21576111), Program for New Century Excellent Talents in University (No. NCET-13-0824).

References

1. Y. Jia, J. Wei, K. Wang, A. Cao, Q. Shu, X. Gui, Y. Zhu, D. Zhuang, G. Zhang, B. Ma, L. Wang, W. Liu, Z. Wang, J. Luo and D. Wu, Adv. Mater., 2008, 20, 4594–4598.
2. Y. Jia, A. Cao, X. Bai, Z. Li, L. Zhang, N. Guo, J. Wei, K. Wang, H. Zhu, D. Wu and P. M. Ajayan, Nano Lett., 2011, 11, 1901–1905.
3. M. Lu, S. Bowden, U. Das and R. Birkmire, Appl. Phys. Lett., 2007, 91, 063507.
4. Y. Liu, J. Zhang, H. Wu, W. Cui, R. Wang, K. Ding, S.-T. Lee and B. Sun, Nano Energy, 2017, 34, 257–263.
5. F. Wang, R. Du, Q. Ren, C. Wei, Y. Zhao and X. Zhang, J. Mater. Chem. C, 2017, 5, 1751–1757.
6. K. Yoshikawa, H. Kawasaki, W. Yoshida, T. Irie, K. Konishi, K. Nakano, T. Uto, D. Adachi, M. Kanematsu, H. Uzu and K. Yamamoto, Nat. Energy, 2017, 2, 17032.
7. D. Macdonald and A. Cuevas, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 67, 075203.
8. N. Jensen, U. Rau, R. M. Hausner, S. Uppal, L. Oberbeck, R. B. Bergmann and J. H. Werner, J. Appl. Phys., 2000, 87, 2639–2645.
9. G. D. Cody, T. Tiedje, B. Abeles, B. Brooks and Y. Goldstein, Phys. Rev. Lett., 1981, 47, 1480–1483.
10. A. A. Langford, M. L. Fleet, B. P. Nelson, W. A. Lanford and N. Maley, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 45, 13367–13377.
11. P. Biswas, D. Paudel, R. Atta-Fynn, D. A. Drabold and S. R. Elliott, Phys. Rev. Appl., 2017, 7, 024013.
12. M. Garín, U. Rau, W. Brendle, I. Martín and R. Alcubilla, J. Appl. Phys., 2005, 98, 093711.
13. Z. Shu, U. Das, J. Allen, R. Birkmire and S. Hegedus, Progress in Photovoltaics: Research and Applications, 2015, 23, 78–93.
14. M. Schmidt, L. Korte, A. Laades, R. Stangl, C. Schubert, H. Angermann, E. Conrad and K. V. Maydell, Thin Solid Films, 2007, 515, 7475–7480.
15. S. Avasthi, S. Lee, Y.-L. Loo and J. C. Sturm, Adv. Mater., 2011, 23, 5762–5766.
16. M. Smeets, K. Bittkau, F. Lentz, A. Richter, K. Ding, R. Carius, U. Rau and U. W. Paetzold, Nanoscale, 2016, 8, 18726–18733.
17. M. Agarwal, A. Pawar, N. Wadibhasme and R. Dusane, Sol. Energy, 2017, 144, 417–423.
18. J. Yu, J. Bian, W. Duan, Y. Liu, J. Shi, F. Meng and Z. Liu, Sol. Energy Mater. Sol. Cells, 2016, 144, 359–363.
19. S. Olibet, E. Vallat-Sauvain, L. Fesquet, C. Monachon, A. Hessler-Wyser, J. Damon-Lacoste, S. De Wolf and C. Ballif, Phys. Status Solidi A, 2010, 207, 651–656.
20. M. Mikolášek, J. Racko and L. Harmatha, Appl. Surf. Sci., 2017, 395, 166–171.
21. R. Hussein, D. Borchert, G. Grabosch and W. R. Fahrner, Sol. Energy Mater. Sol. Cells, 2001, 69, 123–129.
22. S.-C. Shiu, J.-J. Chao, S.-C. Hung, C.-L. Yeh and C.-F. Lin, Chem. Mater., 2010, 22, 3108–3113.
23. Y. Zhang, C. Yu, M. Yang, Y. He, L. Zhang, J. Zhang, X. Xu, Y. Zhang, X. Song and H. Yan, RSC Adv., 2017, 7, 9258–9263.
24. L. Mazzarella, S. Kolb, S. Kirner, S. Calnan, L. Korte, B. Stannowski, B. Rech and R. Schlatmann, in 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), 2016, pp. 2955–2959.
25. K. Ding, X. Zhang, F. Xia, R. Wang, Y. Kuang, S. Duhm, J. Jie and X. Zhang, J. Mater. Chem. A, 2017, 5, 285–291.
26. M. H. Brodsky, M. Cardona and J. J. Cuomo, Phys. Rev. B: Condens. Matter Mater. Phys., 1977, 16, 3556–3571.
27. J. Kakalios, R. A. Street and W. B. Jackson, Phys. Rev. Lett., 1987, 59, 1037–1040.
28. L. Zhao, G. Wang, H. Diao and W. Wang, Phys. Status Solidi RRL, 2016, 10, 730–734.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ra04018k

---