† Corresponding author. E-mail:
Project supported by the National Key Research and Deveopment Program of China (Grant No. 2018YFB1500402), the National Natural Science Foundation of China (Grant Nos. 61674084 and 61874167), the Fundamental Research Funds for Central Universities, China, the Natural Science Foundation of Tianjin City, China (Grant No. 17JCYBJC41400), the Open Fund of the Key Laboratory of Optical Information Science & Technology of Ministry of Education of China (Grant No. 2017KFKT014), the 111 Project, China (Grant No. B16027), the International Cooperation Base, China (Grant No. 2016D01025), and Tianjin International Joint Research and Development Center, China.
Hydrogenated amorphous silicon oxide (a-SiOx:H) is an attractive passivation material to suppress epitaxial growth and reduce the parasitic absorption loss in silicon heterojunction (SHJ) solar cells. In this paper, a-SiOx:H layers on different orientated c-Si substrates are fabricated. An optimal effective lifetime (τeff) of 4743 μs and corresponding implied open-circuit voltage (i Voc) of 724 mV are obtained on 〈100〉-orientated c-Si wafers. While τeff of 2429 μs and i Voc of 699 mV are achieved on 〈111〉-orientated substrate. The FTIR and XPS results indicate that the a-SiOx:H network consists of SiOx (Si-rich), Si– OH, Si– O– SiHx, SiO2≡Si– Si, and O3 ≡S– Si. A passivation evolution mechanism is proposed to explain the different passivation results on different c-Si wafers. By modulating the a-SiOx:H layer, the planar silicon heterojunction solar cell can achieve an efficiency of 18.15%.
Nowadays, silicon heterojunction (SHJ) solar cell has received much attention, due to its advantages of high performance, low cost, simple manufacturing, non-toxic, etc. Meanwhile, the word-record conversion efficiency of 26.63% has been achieved on SHJ.[1] As is well known, the remarkable passivation materials such as a-Si:H play a crucial role in SHJ devices, of which the open-circuit voltage (Voc) value has reached up to 750 mV.[2–4] However, the SHJ solar cells suffer the parasitic absorption of a-Si:H layers, and thus reducing the short-circuit current density (Jsc) in solar cells. In addition, it is difficult to fabricate a-Si:H film with low optical gap, which can easily cause an epitaxial growth on c-Si substrate.[5] To avoid such a kind of loss, a-SiOx:H layers offer an optimistic solution.
The development of a-SiOx:H was derived from the technology in which a-Si:H is alloyed with oxygen, thus an increase in the optical gap was observed with increasing oxygen content (co), which leads the parasitic absorption losses to decrease.[6] On one hand, the alloyed oxygen can improve passivation quality of a-Si:H, which is mainly attributed to the manipulated defect density and suppressed epitaxial growth on c-Si substrate.[2,7–9] On the other hand, the exceeded oxygen might lead to inferior passivation quality in the presence of interconnected voids.[10] Thus, the pivotal to acquiring superior passivation quality is an appropriate value co. Therefore, the passivation quality is a rather complex quantity, influenced not only by the intrinsic properties of the layer, but also by the characteristics of the substrate. In this paper, a-SiOx:H passivation layers with different thicknesses are investigated on 〈100〉- and 〈111〉-orientated c-Si wafers respectively, aiming to achieve better passivation quality on different substrates. Furthermore, the passivation mechanism and the microstructure evolution of a-SiOx:H are analyzed by x-ray photoelectron spectroscopy (XPS) and reflect Fourier transform infrared (Re-FTIR) spectroscopy. Finally, the a-SiOx:H layers are applied to bifacial-polished silicon substrate for SHJ device fabrication.
The 〈100〉- and 〈111〉-orientated bifacial-polished floating-zone n-type c-Si wafers with resistivity values in a range of 1 Ω ⋅cm– 10 Ω ⋅cm and a bulk lifetime of > 1000 μs were used. The c-Si wafers were cleaned by the standard RCA process followed by the cleaning with a 5% hydrofluoric acid (HF) solution to remove the native silicon oxide layer.[11] The a-SiOx:H film was deposited on the bifacial side of the cleaned 〈111〉-orientated substrate by changing CO2 flow, mixture gas pressure, and RF initial power. Then, the optimal deposition parameters were applied to the bifacial side of the cleaned 〈100〉 and 〈111〉 c-Si wafers with various film thicknesses.
To measure the τeff of the immersed HF silicon substrate and the passivation wafer, Sinton Instruments WCT-120 QSSPC tester was used in transient or quasi-steady state analysis mode of generalized (1/1) or (64/1) at an injection level of 1 × 1015 cm−3.[12] The optical properties of the thin film sample were analyzed by measuring the transmission and reflection with a Varian–Cary 5000 spectrometer, including the Tauc optical bandgap and refractive index. For the FTIR measurements, a Thermo Scientific Nicolet iS10 spectrometer was used in reflection mode in a spectral range of 400 cm−1– 4000 cm−1. The FTIR spectrum was universally used to obtain the information about the a-Si:H[13–16] and a-SiOx:H[17, 18] layers. In this paper, we mainly acquired the hydrogen and oxygen content from the FTIR spectrum measurement according to different peaks' wavenumbers for different absorbance values. For the Si– H bond, the stretching modes at 2000 cm−1– 2260 cm−1,[14,19,20] among which the stretching modes at 2085 cm−1– 2100 cm−1 originated from the stretching in (SiH2.)n or micro voids,[14] and stretching modes at 2140 cm−1– 2260 cm−1 came from the stretching in H-Si(SixOy) (x = 0,1; y = 1,2,3).[17–20] For the Si– O– Si bond, the stretching modes at 940 cm−1– 1150 cm−1 were from the stretching in Si– O– Si– SiyOz (y = 1; z = 0,1,2).[18–20] The simulation fitting peaks were mainly split into several peaks at 900 cm−1– 1200 cm−1 and 1850 cm−1– 2300 cm−1. In addition, the oxygen content (co) was estimated from the XPS spectrum in the passivated layer according to the peaks of Si+x (x = 1, 2, 3, 4) shown in the XPS spectrum.[21–24] In terms of the chemical bonding and energetic data, we could further obtain the growth mechanism of a-SiOx:H from XPS spectra with various passivation layer thickness, and build a growth model of passivation layer.[25]
Finally, the optimized passivation condition was applied to the bifacial side of the cleaned substrate for fabricating the planar SHJ solar cell. The structure is schematically depicted in Fig.
In RF plasma, SiH4 can be dissociated into SiH3 and H (formula (
Thus, the oxygen in a-SiOx:H can be reasonably attributed to the OH bonded with other complex groups, which incorporates with dangling Si bonds on the c-Si surface, reacts with SiH3, H and itself to form novel complex groups as follows:[27–29]
From formulas (
The XO is defined as the ratio of CO2 flow rate to SiH4 flow rate as follows:
From formulas (
In this study, we deposit a-SiOx:H films with various thicknesses on 〈111〉- and 〈100〉-orientated c-Si substrates, respectively. Then τeff and i Voc of corresponding samples were measured and are shown in Fig.
Figure
For the 〈111〉-orientated c-Si substrate, the τeff and iVoc of the passivated samples show a trend of the extension with the inctrease of passivation thickness, indicating that the quantity of the precursors (OH and H) can fluctuate with the continuous introducing of CO2, which further affects the content of O and H in the passivation material film. The passivation performances of a-SiOx:H on the 〈111〉- and 〈100〉-orientated c-Si substrates present different rules, which probably relate to their corresponding c-Si surface microstructures.
In order to analyze the process of a-SiOx:H on the 〈100〉- and 〈111〉-orientated c-Si substrates, the FTIR spectroscopy is used to analyze the H and O contents in the film. For studying the thinner a-SiOx:H, its Fourier transform infrared spectrum is measured by the reflected light signal. The result demonstrates that the Si–Si, Si–H, and Si–O–Si bonds may exist
in the a-SiOx:H film. The feature signals of Si–H and Si–O in 800 cm–1–1200 cm–1 and 1900 cm−1–2300 cm−1 are corresponding to FTIR spectra on the 〈100〉- and 〈111〉-orientated c-Si substrates (see Figs.
From Fig.
From Fig.
The oxygen content values of a-SiOx:H on the 〈100〉- and 〈111〉-orientated c-Si substrates are valued by XPS technique. The binding energy scale is calibrated by assigning a value of 284.6 eV to the C 1s signal.[32] According to the relevant theoretical model,[25, 33] peaks of Si–Si and Si–O bands are observed obviously on the XPS spectra of a-SiOx:H films, respectively, deposited on 〈100〉-orientated Si 2p substrate (Fig.
From Figs.
The XPS peak of Si 2p can be decomposed into two peaks, that is, 2p3/2 and 2p1/2 peaks, and the area ratio of these two peaks is 2 : 1. According to the peak of 2p3/2, the chemical shifts of silicon oxides corresponding to Si+, Si2+, Si3+, and Si4+in their valence states are 0.86 eV, 1.5 eV, 2.4 eV, and 4.1 eV, respectively,[34] while the peak of 2p3/2 is at 98.86 eV. Figures
From Fig.
In summary, the components of a-SiOx:H film deposited on the 〈100〉- and 〈111〉-orientated c-Si substrates are identified by XPS analysis. The SiOx network model with Si-rich a-SiOx:H is comfirmed by XPS of Si 2p and O 1s which consists of Si–OH, Si–O–SiHx, and rich in oxygen vaccury and dangling bond complex components. Comparing with 〈100〉-orientated c-Si substrate, the passivation quality of the 〈111〉-orientated c-Si substrate is better subjecte to the subsequent arising of and with passivation layer thickness increasing. And a-SiOx:H is more suitable for passivating 〈100〉-orientated c-Si substrate.
We prepare two kinds of passivation layers (a-Si:H and a-SiOx:H) with different thicknesses, and apply them to SHJ solar cells. The results are shown in Fig.
For a-Si:H and a-SiOx:H passivation layer, as the thickness of the passivation layer increases, the Voc of SHJ solar cells becomes higher. Besides, as the thickness of the passivation layer is less than or equal to 4 nm, a-SiOx:H is better than a-Si:H passivation layer. However, as the thickness of the passivation layer is thicker than 4 nm, the performance of a-Si:H passivation is better than that of a-SiOx:H. The reason might be that the internal defect states of the a-SiOx:H film may increase with passivation layer thickening, leading the passivation effect to degrade. In addition, the FF and Jsc are observed to decease with passivation layer thickness increasing. SHJ solar cells with a-Si:H and a-SiOx:H passivation layers demonstrate conversion efficiencies of 16.88% and 18.15%, respectively. Due to much improved passivation effect of a-SiOx:H, the performance of SHJ cells with a-SiOx:H/c-Si structure is better than that of a-Si:H/c-Si.
With the mentioned model of a-SiOx:H affecting passivation performance of c-Si with 〈100〉- and 〈111〉-orientated substrates, we use the optimal a-SiOx:H on the SHJ solar cells to achieve excellent device with suitable thickness, which can achieve the excellent passivation performance with τeff = 2168 μs, thereby resulting in high Voc. The passivation layer modifies the open potential and enhances the red light response of the SHJ solar cell. As shown in Fig.
In this work, we investigate a-SiOx:H films deposited on the 〈100〉- and 〈111〉-orientated c-Si substrates with various passivation layer thicknesses, which exhibit different passivation qualities. According to the FTIR and XPS analysis, we conclude that Si-rich network model of a-SiOx:H consists of Si–OH, Si–O–SiHx, and other complex components and then the passivation evolution mechanism is proposed. When the optimal a-SiOx:H film is used as the passivation layer in the planar SHJ solar cell, an efficiency of 18.15% is achieved. This research may provide a useful guideline for further improving the performance of SHJ solar cells.
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] | |
[18] | |
[19] | |
[20] | |
[21] | |
[22] | |
[23] | |
[24] | |
[25] | |
[26] | |
[27] | |
[28] | |
[29] | |
[30] | |
[31] | |
[32] | |
[33] | |
[34] | |
[35] | |
[36] | |
[37] | |
[38] | |
[39] |