Enhancing GaAs Nanowire Pin Junction Solar Cells via AlGaAs/GaAs Heterojunctions

Background

Gallium arsenide nanowires (GaAs NWs) have emerged as promising candidates for high‑efficiency photovoltaic conversion, owing to their 1.43 eV bandgap that outperforms silicon for spectral utilization [1–4]. Reported efficiencies for axial pn‑junction GaAs NW arrays reach 15.3 % [5]. Nonetheless, surface recombination remains a critical loss channel; thus, effective passivation strategies are essential [6, 7]. A well‑established passivation technique involves conformal AlGaAs shells, which introduce large conduction and valence band offsets that confine minority carriers and mitigate surface recombination [5, 8–10].

Beyond passivation, concentrating optical absorption within the intrinsic (i) region of pin junctions can further elevate efficiency [11–17]. Various methods have been explored, including junction‑position optimization, inclined NWs, plasmonic metal decoration, and high‑bandgap heavily doped layers [13–17]. While AlGaAs shells have been widely adopted for passivation, their role in steering photogenerated carriers toward the active region has received limited attention.

In this study, we systematically evaluate GaAs NW pin‑junction arrays that incorporate AlGaAs/GaAs heterojunctions, targeting both axial and radial geometries. By replacing the top p‑segment (axial) or outer shell (radial) with AlGaAs, we exploit its reduced absorption coefficient to concentrate photogeneration within the i‑region, thereby reducing recombination losses in heavily doped zones and protecting carriers from surface and contact recombination.

Methods

The device architecture is illustrated in Fig. 1. Each solar cell consists of a periodic NW array; for illustration, a single NW is shown. The Al_0.8Ga_0.2As layer serves as the top p‑type segment in axial pin junctions and as the outer p‑type shell in radial pin junctions; all remaining regions are GaAs. Both p‑ and n‑regions are doped at 10^18 cm⁻³. NW diameter and length are 180 nm and 1.2 µm, respectively, with a period of 360 nm, following the optimized D/P ratio reported in Ref. [18].

a Schematic of the GaAs NW axial pin junction solar cell and its AlGaAs/GaAs heterojunction counterpart. b Schematic of the GaAs NW radial pin junction solar cell and its AlGaAs/GaAs heterojunction counterpart.

Optical absorption was simulated using Lumerical FDTD Solutions, imposing periodic boundary conditions to model the array with a single NW. Complex refractive indices for GaAs and Al_0.8Ga_0.2As were taken from Ref. [19]. The photon generation rate G_ph was calculated via the divergence of the Poynting vector, assuming one electron–hole pair per absorbed photon:
G_ph = |∇·S| / (2ħω) = ε″|E|² / (2ħ). The AM 1.5 G spectrum was used to weight G_ph across the simulated spectrum.

Electrical characteristics were computed in Synopsys Sentaurus by coupling the optical generation profile to a self‑consistent solution of carrier continuity and Poisson equations. Doping‑dependent mobility, radiative, Auger, and Shockley–Reed–Hall (SRH) recombination were included. The AlGaAs/GaAs interface was modeled with a thermionic emission model (Eqs. 2–5) and assumed to be defect‑free, as appropriate for lattice‑matched epitaxy [21]. Surface recombination was considered only at air–NW interfaces; SRVs of 10³ and 10⁷ cm s⁻¹ were examined to represent well‑passivated and poorly passivated surfaces. Device parameters are summarized in Table 1.

Results and Discussion

Figure 2 compares the absorption spectra of AlGaAs/GaAs heterojunction NWs with their GaAs counterparts. Axial devices feature top p‑regions of 150 nm and bottom n‑regions of 200 nm; radial devices have 20 nm p‑shells and 20 nm inner n‑radius. The overall absorption is similar across structures, except for a slight drop in the radial heterojunctions near the GaAs bandgap. At ≈900 nm, the optical field concentrates along the sidewalls; in the radial heterojunction, the AlGaAs shell does not absorb efficiently, localizing carriers in the GaAs core.

a Absorption spectra for GaAs and AlGaAs/GaAs axial and radial heterostructures. b Cross‑sectional optical generation for the axial heterojunction. c Cross‑sectional optical generation for the radial heterojunction. d Cross‑sectional optical generation for the pure GaAs NW. e Intrinsic‑region absorption spectra for axial devices. f Intrinsic‑region absorption spectra for radial devices.

Electrical simulations reveal that, for low SRV conditions, axial devices using AlGaAs in the top p‑segment exhibit a 2.9 % absolute efficiency increase (from 11.6 % to 14.5 %) driven primarily by an 18.9 % rise in short‑circuit current (from 18.9 to 23.3 mA cm⁻²). Radial devices show a smaller but consistent improvement (10.8 % to 11.3 % efficiency, 22.6 to 23.8 mA cm⁻²). Under high SRV (10⁷ cm s⁻¹), axial devices experience a pronounced efficiency drop for both designs; nevertheless, AlGaAs structures retain a higher short‑circuit current due to suppressed recombination at the top contact. Radial AlGaAs/GaAs NWs maintain performance because the AlGaAs shell confines carriers away from the surface.

Current–voltage characteristics for axial (a) and radial (b) devices at SRVs of 10³ and 10⁷ cm s⁻¹.

Varying the p‑region volume elucidates the impact of heavily doped zones on efficiency. Figure 4a shows that, as the top p‑segment length increases from 50 to 200 nm, the generation hotspot shifts toward the bottom, concentrating carriers beneath the AlGaAs layer. In low SRV scenarios, the efficiency of AlGaAs/GaAs devices remains largely unchanged, whereas GaAs devices degrade linearly with increasing p‑length due to higher surface‑proximal photogeneration. At high SRV, AlGaAs/GaAs efficiency even rises with longer p‑regions, as carriers generated within AlGaAs are shielded from surface recombination. Consequently, a longer AlGaAs top segment can be employed without sacrificing performance, simplifying fabrication and contact processes for axial arrays.

a Generation profiles for axial devices with varying p‑lengths. b Efficiency versus p‑length for axial GaAs and AlGaAs/GaAs NWs.

Radial devices exhibit a similar trend with respect to p‑shell thickness. Figure 5a demonstrates that increasing the AlGaAs shell thickness confines more photogeneration within the GaAs core, mitigating surface recombination. Efficiency decreases for both designs as shell thickness grows, but the decline is markedly gentler for AlGaAs/GaAs devices. Under high SRV, AlGaAs/GaAs efficiency drops only marginally, whereas GaAs radial devices suffer a severe efficiency loss (from 10.8 % to 8.05 %) when the SRV rises from 10³ to 10⁷ cm s⁻¹. With a 30 nm p‑shell, AlGaAs/GaAs devices achieve 10.4 % efficiency, 8.42 % higher than the 1.98 % of pure GaAs radial NWs.

a Generation profiles for radial devices with varying p‑shell thicknesses. b Efficiency versus p‑shell thickness for radial GaAs and AlGaAs/GaAs NWs.

Conclusions

Coupled 3‑D optoelectronic simulations demonstrate that AlGaAs/GaAs heterojunctions effectively confine photogeneration to the active regions, diminish recombination losses in heavily doped zones, and create minority‑carrier barriers that protect against surface and contact recombination. For axial NWs, the use of AlGaAs allows a longer top p‑segment without performance penalty, easing device fabrication and electrical contact. Radial NWs benefit from superior tolerance to p‑shell thickness and surface recombination, maintaining high efficiencies even under aggressive SRV conditions. Overall, AlGaAs/GaAs heterojunctions represent a practical, high‑impact strategy for advancing GaAs nanowire solar cell performance.