Amorphous silicon oxide window layers for high-efficiency silicon heterojunction solar cells

Abstract

In amorphous/crystalline silicon heterojunction solar cells, optical losses can be mitigated by replacing the amorphous silicon films by wider bandgap amorphous silicon oxide layers. In this article, we use stacks of intrinsic amorphous silicon and amorphous silicon oxide as front intrinsic buffer layers and show that this increases the short-circuit current density by up to 0.43 mA/cm2 due to less reflection and a higher transparency at short wavelengths. Additionally, high open-circuit voltages can be maintained, thanks to good interface passivation. However, we find that the gain in current is more than offset by losses in fill factor. Aided by device simulations, we link these losses to impeded carrier collection fundamentally caused by the increased valence band offset at the amorphous/crystalline interface. Despite this, carrier extraction can be improved by raising the temperature; we find that cells with amorphous silicon oxide window layers show an even lower temperature coefficient than reference heterojunction solar cells (−0.1%/°C relative drop in efficiency, compared to −0.3%/°C). Hence, even though cells with oxide layers do not outperform cells with the standard design at room temperature, at higher temperatures—which are closer to the real working conditions encountered in the field—they show superior performance in both experiment and simulation.

Low-temperature plasma-deposited silicon epitaxial films: Growth and properties

“…A KAI-M PlasmaBox TM reactor (Oerlikon/Tokyo Electron) in parallel-plate configuration powered at very high frequency (VHF, 40.68 MHz) with an inter-electrode distance of 12.5 mm was used for the depositions. 24,25 The base pressure of the reactor was typically about 4 Â 10 À7 mbar. The three substrates were always codeposited for given conditions.…”

Multiprobe Characterization of Inversion Charge for Self-Consistent Parameterization of HIT Cells

“…Finally, it is shown that the developed theory is consistent to the more widely used measurements, such as dark I-V and light I-V. The general nature of the characterization approach can be extended not only to Si heterojunction with new emitter stack designs, such as MoO x [42] and SiO x [43], but to other heterojunction thin-film solar cell designs as well, which are susceptible to S-type light I-V curve. This self-consistent determination of the HIT cell parameters is an 3.9 3 .9 4 .05 3.9 3 .9 Conduction Band Tail States N t a il = 10 1 9 cm −3 · eV −1 N t a il = 10 1 9 cm −3 · eV −1 − N t a il = 10 1 9 cm −3 · eV −1 N t a il = 10 1 9 cm −3 · eV −1 E t a il = 0.05 eV E t a il = 0.019 eV E t a il = 0.019 eV E t a il = 0.05 eV C h = 10 −1 6 cm −2 C h = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 C h = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 C h = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 Valence band tail states N t a il = 10 1 9 cm −3 · eV −1 N t a il = 10 1 9 cm −3 · eV −1 − N t a il = 10 1 9 cm −3 · eV −1 N t a il = 10 1 9 cm −3 · eV −1 E t a il = 0.1 eV E t a il = 0.049 eV E t a il = 0.049eV E t a il = 0.1eV C h = 10 −1 6 cm −2 C h = 10 −1 6 cm −2 C h = 10 −1 6 cm −2 C h = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 C n = 10 −1 6 cm −2 Donor-like defects N t 1 = 10 1 6 cm −3 N t 1 = 5 × 10 1 5 cm −3 − N t 1 = 5 × 10 1 5 cm −3 N t 1 = 10 1 6 cm −3 E t 1 − E i = 0.31 eV E t 1 − E i = 0.21 eV E t 1 − E i = 0.21 eV E t 1 − E i = 0.31eV σ t 1 = 0.08eV σ t 1 = 0.08 eV σ t 1 = 0.08 eV σ t 1 = 0.08eV τ p 1 = 10 −9 sτ n 1 = 10 −9 s τ p 1 = 2 × 10 −9 s τ p 1 = 2 × 10 −9 s τ p 1 = 10 −9 s N t 2 = 10 1 8 cm −3 τ n 1 = 2 × 10 −9 s τ n 1 = 2 × 10 −9 s τ n 1 = 10 −9 s E t 2 − E i = −0.26 eV N t 2 = 10 1 8 cm −3 σ t 2 = 0.15 eV E t 2 − E i = −0.26 eV τ p 2 = 10 −9 sτ n 2 = 10 −9 s σ t 2 = 0.15eV τ p 2 = 10 −9 s τ n 2 = 10 −9 s Acceptor-like defects N t 1 = 10 1 6 cm −3 N t 1 = 5 × 10 1 5 cm −3 − N t 1 = 5 × 10 1 5 cm −3 N t 1 = 10 1 6 cm −3 E t 1 − E i = −0.21 eV E t 1 − E i = 0.11 eV E t 1 − E i = 0.11 eV E t 1 − E i = −0.21 eV σ t 1 = 0.08 eV σ t 1 = 0.08 eV σ t 1 = 0.08 eV σ t 1 = 0.08 eV τ p 1 = 10 −9 sτ n 1 = 10 −9 s τ p 1 = 2 × 10 −9 s τ p 1 = 2 × 10 −9 s τ p 1 = 10 −9 s τ n 1 = 2 × 10 −9 s τ n 1 = 2 × 10 −9 s τ n 1 = 10 −9 s Contact properties…”

Optimization of the window layer in large area silicon heterojunction solar cells

“…As is known, extinction coefficient and absorption coefficient are positively correlated. So, the incorporation of oxygen in these lms is the origination of the decreasing absorption in short wavelength region, 19,23 this veries the result of k presented in Fig. In the wavelength region from 700 nm to 1200 nm, extinction coefficients are zero, indicating that a-SiO x :H lms have no absorptance in this wavelength region.…”

“…14,15 R. Biron et al observed an enhancement of 0.6 mA cm À2 in J SC in the shortwavelength region due to the enlargement of optical band gap of silicon oxide based emitter, 16 and other studies have obtained similar results. 19 The optical properties of n-type a-SiO x :H lms and the effect of n-type a-SiO x :H lms on SHJ solar cell performance were investigated in details. For the passivation layer, intrinsic a-Si:H lm was le to keep an excellent passivation effect.…”