In this work, the solar-grade n-type c-Si wafers with (100) orientation are used in the experiment. The resistivity is , and the thickness is . All wafers are first cleaned by standard RCA, then dipped in HF to remove the native oxide layer formed at the surface of the wafers. Both intrinsic and doped a-Si: H thin films are deposited by multi-chamber plasma-enhanced chemical vapor deposition (PECVD). And the i-a-Si: H layers with the thicknesses of 3 nm, 5 nm, and 10 nm are prepared to study the surface chemical passivation performance.

The n⁺-a-Si: H(O/C) thin films are deposited on quartz substrates using a gas mixture comprising silane (SiH₄), hydrogen (H₂), phosphorane (PH₃), and methane (CH₄) or nitrogen dioxide (NO₂) under the pressure of 70 Pa. The plasma power is 10 W, and the substrate temperature is 250 °C.

The effective minority carrier lifetime (τ_eff) measurements are performed on the samples with i-a-Si:H layers on the front and the back sides of the silicon wafers, respectively, which are realized by using the quasi-steady-state photoconductance (QSSPC) method (WCT-120, Sinton Instruments, USA). The optical properties are tested by a UV-3600 Plus UV–VIS–NIR spectrophotometer. The doping profiles of the doped a-Si thin films are measured by electrochemical capacitance–voltage (ECV) profiling (CVP21, WEP, Germany).

As shown in Fig. 1, the effective minority carrier lifetime τ_eff decays dramatically with the thickness reduction. And so does the implied V_oc (i-V_oc). At the low-injection level, τ_eff can be expressed as 1 The surface recombination velocity (S, cm/s) is used to represent the chemical passivation. The bulk lifetime τ_bulk is impacted by the Auger recombination and Shockley–Read–Hall (SRH) recombination. W is the thickness of the sample. The τ_bulk is simplified to be infinite, and S is calculated to be 10 cm/s, 100 cm/s, and 500 cm/s for the thicknesses of 10 nm, 5 nm, and 3 nm.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. Minority carrier lifetimes of samples with different i-a-Si: H layer thicknesses.

In IBC-SHJ, J_sc and V_oc increase dramatically when the front surface chemical passivation ameliorates,^[9] which means that the thicker i-a-Si: H is beneficial for the chemical passivation. But thicker i-a-Si: H layer would induce optical loss in the front surface. The 5 nm thick i-a-Si: H is chosen to provide chemical passivation in the simulation.

As shown in Fig. 2, the optical band gaps of five n⁺-a-Si: H materials, including microcrystalline silicon ( -Si), amorphous silicon (a-Si), amorphous silicon carbide (a-SiC), and amorphous silicon oxide (a-SiO_x), increase from 1.52 eV to 2.30 eV, which are calculated by Tauc’s equation 2 with 3 where is the absorption coefficient, is the incident photon energy, B is a proportional constant,^[22] t is the transmittance. r is the reflectivity, and d is the thin film thickness.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. Tauc’s plots of optical absorption spectra of different n+-a-Si: H materials.

The transmission and reflectance spectra of different P doped Si-based alloy materials are tested by the UV-3600 Plus UV–VIS–NIR spectrophotometer. The band gap width of different n⁺-a-Si: H materials is in effect modulated by altering the a-Si morphology or impurity compound.

The simulator Sentaurus TCAD is extensively used for modeling solar cells based on the calculation of Possion equation and carrier continuity equation. The unit cell of the IBC-SHJ solar cells studied in this case is shown in Fig. 3. The substrate is an N type c-Si wafer with thickness of . On the front textured surface, the FSF layer, a stack of 5 nm i-a-Si: H and 10 nm n⁺-a-Si: H, is deposited. Then a 75 nm SiN_x layer is deposited as anti reflection (AR) coating. The P doped amorphous silicon (n⁺-a-Si: H) and B doped amorphous silicon (p⁺-a-Si: H) are prepared on the back side in an interdigital pattern. The lateral dimensions of the n-strips and p-strips are and , respectively. And the gap is set to be . The other parameters are listed in Table 1 and the state density distributions of the three types of a-Si are based on ^[23]. In this work, the optical modeling is performed by means of ray tracing method combined with optical generation model and complex refractive index model. The drift–diffusion model, thermionic emission model, and trap-assist tunneling model are applied to describe carriers’ transportation. The SRH recombination model, surface recombination model, and Auger recombination model are used to describe the recombination process in the cell. In addition, the energy band gap narrowing model, mobility model, and quantum potential model are also taken into consideration.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. Schematic structure of the simulated IBC-SHJ solar cells.

Table 1.

Table 1.

Table 1. Default parameters used in the simulation. . Material Eg/eV χ/eV Doping concentration/cm−3 Si 1.12 4.1 2 × 1015 i-a-Si: H 1.70 3.8 2 × 1015 p+-a-Si: H (back) 1.65 3.8 ECV data n+-a-Si: H (back) 1.72 3.7 ECV data n+-a-Si: H (front) 1.52–2.30 3.7 1 × 1017–1 × 1020

Table 1. Default parameters used in the simulation. .

The substrate doping is set to be 2 × 10¹⁵ cm⁻³ to match the wafer’s resistivity. The effective doping concentrations of n⁺-a-Si: H and p⁺-a-Si: H are imported according to the ECV measurement results, as shown in Fig. 4. The doping concentration of the n⁺ Si-based alloy materials is set from 1 × 10¹⁷ cm⁻³ to 1 × 10²⁰ cm⁻³, under the assumption of complete ionization.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. Doping concentrations of n+-a-Si: H and p+-a-Si: H thin films on the back of IBC-SHJ solar cells.

The energy band model is based on the following equation:^[24] 4 where E_g(0) is the band gap energy at 0 K, T is the thermal temperature, and α and β are the material parameters. The electron affinity χ is the difference between the lowest energy in the conduction band and the vacuum level, as described by 5 where χ₀ and C_Bgn2Chi are adjustable parameters, and the default of C_Bgn2Chi is 0.5 in Sentaurus. In this simulation, the electron affinity is assumed to be constant for various band gaps.^[22] Five kinds of n⁺ Si-based alloy materials with band gaps of 1.52 eV, 1.72 eV, 1.92 eV, 2.10 eV, and 2.30 eV are simulated. The band gaps are determined from Fig. 2.

The energy band diagram of n⁺-a-Si: H/i-a-Si: H/c-Si/i-a-Si: H/p⁺-a-Si: H in thermal equilibrium is shown in Fig. 5. The doping concentration of n⁺-a-Si: H is 6 × 10¹⁹ cm⁻³. It is shown that the application of n⁺-a-Si: H with different band gaps on the front surface will neither cause the bend of the energy band, nor the change of the field passivation.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. Energy band diagram of n+-a-Si: H/i-a-Si: H/c-Si/i-a-Si: H/p+-a-Si: H with different band gaps in thermal equilibrium.

The reason is that the constant doping concentration of n⁺-a-Si: H defines the position of the Fermi level. And the bending does not vary with the constant Fermi level position. Thus, the bending at the i-a-Si: H/c-Si interface is not influenced by the band gap of n⁺-a-Si: H. Although the interface band bending does not change, the band structure on the front surface is still different. Figure 6 shows the integral photo-generated carrier in the entire substrate section of five n⁺-a-Si: H materials under AM1.5 illumination. The number of photo-generated carriers increases with the enhancement of the band gap, which results from the fact that with the band gap widening, the photons in the short-wavelength range are not absorbed by the passivation stack, and thus can be effectively utilized by the substrate.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. Integral photo-generated carriers in the entire substrate section with different band gaps under AM1.5 illumination.

It can be seen from the external quantum efficiency (EQE) curve of the IBC-SHJ solar cells in Fig. 7 that due to the reduction of the absorption loss of photons by the passivation stack, the EQE is improved under the same field passivation. Compared with the case of band gap 1.52 eV, the weighted average EQE of 2.30 eV is increased by 12% according to the equation 6 where N_ph is the photon-generated carriers collected in the terminal, and N_in is the total input photons. Therefore, increasing the bandgap width of the n⁺-a-Si: H material can effectively improve the output photon current density.

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. EQE data of IBC-SHJ with n+-a-Si: H thin films of different band gaps.

The energy band diagram of n⁺-a-Si: H/i-a-Si: H/c-Si/i-a-Si: H/p⁺-a-Si: H with different doping concentration under thermal equilibrium is shown in Fig. 8. The band gap of n⁺-a-Si: H is set to be 2.30 eV. As is seen, the band bending at the i-a-Si:H/c-Si interface changes with the doping concentration. With the increasing doping concentration, the energy band bends downward more seriously, which creates a higher barrier for holes. That is to say, the field passivation becomes stronger, and more holes can be driven back to the substrate.

Fig. 8.

	Figure Option ViewDownloadNew Window
	Fig. 8. Energy band diagram of n+-a-Si: H/i-a-Si: H/c-Si/i-a-Si: H/p+-a-Si: H with different doping concentration (in cm−3) in thermal equilibrium.

The band bending at the interface can also be attributed to the change of the Fermi level. The higher the doping concentration is, the closer the Fermi level is to the conduction band, as shown by the following equation: 7 where E_F is the Fermi level, E_c is the conduction band, k is the Boltzmann constant, N_D is the donor concentration, and N_c is the effective state density of the conduction band. In general, since , the second term is negative.

The separation of electron–hole pairs on the front surface is due to the selection of the electric field. As shown in Fig. 9, the electric field becomes stronger because of the higher doping concentration of n⁺-a-Si: H, and thus generates a stronger driving force to the holes. The change of the electric field corresponds to the energy band diagram in Fig. 10. Since the electric field results from the energy band bending, the more serious the energy band bends, the stronger the field becomes.

Fig. 9.

	Figure Option ViewDownloadNew Window
	Fig. 9. Electric field on the front surface of different doping concentrations (in cm−3).

Fig. 10.

	Figure Option ViewDownloadNew Window
	Fig. 10. Electron and hole current densities near IBC-SHJ front surface with different doping concentrations (in cm−3).

Figure 10 shows the electron and the hole current densities near the front surface under the condition of different doping concentrations when the band gap is 2.30 eV. The electron current density increases rapidly with the increase of the doping concentration, which is contrary to the situation of the hole current density. The field effect strongly reduces the density of one type of carriers at the interface. The higher doping concentration results in the higher electric field, which can attract more majority carriers to the front surface. While it drives the minority carriers away from the front surface. Thus, the recombination of the carriers is reduced, which contributes to a higher carrier lifetime.

The mutual action between chemical and field passivations on the front surface is also considered. As shown in Fig. 11, the electric field intensity is extracted at in Fig. 9, the short circuit current J_sc, open circuit voltage V_oc, filling factor FF, and efficiency E_ff under different electric field passivation and chemical passivation are present. The surface recombination velocity, S varies from 10 cm/s to 500 cm/s to represent the chemical passivation quality from excellent to poor when ignoring the i-a-Si: H thickness.

Fig. 11.

	Figure Option ViewDownloadNew Window
	Fig. 11. The Jsc, Voc, FF, and Eff due to the total effect of chemical passivation and electric field passivation.

It can be seen that with the same chemical passivation, the J_sc, V_oc, FF, and E_ff increase with the increasing electric field on the front surface. E_ff decreases from 26.57% to 22.81% when the electric field reduces from 1.3 × 10⁵ V/cm to 1 × 10⁴ V/cm when S is 100 cm/s. And E_ff decreases from 26.41% to 16.13% when the electric field reduces from 1.3 × 10⁵ V/cm to 1 × 10⁴ V/cm with S of 500 cm/s. Under the condition of the same electric field passivation, the J_sc, V_oc, FF, and E_ff decay with the worsening chemical passivation. When the electric field passivation is strong ( ), the chemical passivation has no significant effect on the J_sc, V_oc, FF, and E_ff. When the electric field strength is 1.3 × 10⁵ V/cm, E_ff decreases from 26.57% to 26.41% with S increasing from 10 cm/s to 500 cm/s, resulting in only 0.16% reduction with inferior chemical passivation. Thus, the role of the field passivation cannot be ignored, especially in the case of thin intrinsic passivation layers and poor chemical passivation.

The S of 100 cm/s represents the poor chemical passivation condition (5 nm i-a-Si: H layer). The optical and electrical properties of n⁺-a-Si thin films are compared comprehensively. Figure 12 shows the changes of the J_sc, V_oc, FF, and E_ff with different doping concentrations and band gaps.

Fig. 12.

	Figure Option ViewDownloadNew Window
	Fig. 12. The (a) Jsc, (b) Voc, (c) FF, and (d) Eff of IBC-SHJ solar cells with different doping concentrations and band gaps.

It is manifested that J_sc enhances with the increasing band gap width and doping concentration. The broadening of the band gap can reduce the light absorption by amorphous silicon and increase the absorption in the substrate. The high doping concentration can enhance the field passivation of the cell, and thus significantly improve the V_oc. The FF increases with the increase of the doping concentration as the recombination is suppressed. The combined effect of these two factors enables the IBC-SHJ to achieve a conversion efficiency of over 26% in spite of the poor chemical passivation.