Bandgap Engineering On Demand in GaAsN Nanowires by Post-Growth HydrogennImplantation

TL;DR

The paper demonstrates post-growth bandgap engineering in GaAsN nanowires on Si by hydrogen implantation, achieving a reversible and tunable shift up to $460\,\mathrm{meV}$ toward the GaAs bandgap via N–H passivation. This approach leverages the relaxed strain in NW heterostructures to incorporate high N concentrations ($\leq 4.2\%$) and enables uniform, full hydrogenation along polytypic ZB/WZ NWs, with substantial PL intensity enhancement. Thermal annealing reverses the passivation, allowing continuous tuning between GaAsN and GaAs, while local laser annealing enables site- and energy-controlled bandgap patterns, paving the way for on-demand quantum dots/rings and energy-matched photonic devices. Overall, the work introduces a scalable, post-growth, hydrogen-based method for versatile bandgap engineering in NWs with potential telecom and solar-energy applications, without extensive device fabrication.

Abstract

Bandgap engineering in semiconductors is required for the development of photonic and optoelectronic devices with optimized absorption and emission energies. This is usually achieved by changing the chemical or structural composition during growth or by dynamically applying strain. Here, the bandgap in GaAsN nanowires grown on Si is increased post-growth by up to 460 meV in a reversible, tunable, and non-destructive manner through H implantation. Such a bandgap tunability is unattained in epilayers and enabled by relaxed strain requirements in nanowire heterostructures, which enables N concentrations of up to 4.2% in core-shell GaAs/GaAsN/GaAs nanowires resulting in a GaAsN bandgap as low as 0.97 eV. Using micro-photoluminescence measurements on individual nanowires, it is shown that the high bandgap energy of GaAs at 1.42 eV is restored by hydrogenation through formation of N-H complexes. By carefully optimizing the hydrogenation conditions, the photoluminescence efficiency increases by an order of magnitude. Moreover, by controlled thermal annealing, the large shift of the bandgap is not only made reversible, but also continuously tuned by breaking up N-H complexes in the hydrogenated GaAsN. Finally, local bandgap tuning by laser annealing is demonstrated, opening up new possibilities for developing novel, locally and energy-controlled quantum structures in GaAsN nanowires.

Figures (5)

• Figure 1: Bandgap engineering of the GaAsN NW shell. (a) Bright-field STEM image of the core-multishell GaAs/GaAsN/GaAs NW cross-section with 1.6% N, a darker contrast is observed in the GaAsN shell compared to the GaAs core and outershell. (b) illustrates how the energy of the GaAsN bandgap (orange) shifts to the bandgap energy of GaAs (purple) upon H implantation of a single core/shell/shell GaAs/GaAsN/GaAs NW lying on a substrate. (c) shows RT µ-PL spectra measured in a point on single, transferred NWs before (shaded orange) and after (purple solid line) hydrogenation. The N content in the GaAsN shell of these NWs is 0, 0.6%, 1.6% and 4.2% (spectra from top to bottom). After hydrogenation, and the passivation of N-atoms by H, the bandgap of all samples is upshifted to the value of GaAs, by 140 meV, 250 meV and 460 meV respectively. The H-dose here is $d_H = 1.2-1.4 H_0$, where $H_0$ corresponds to $10^{19}$ H-ions/cm$^2$. Normalization factors discussed in Figure \ref{['fig02']} and reported in the SI 2. Notice that the small post-hydrogenation redshift in the 0% sample is not statistically relevant, as can be seen in the SI 2 and 4 and in panel (d) here. (d) and (e) show the energy at the maximum of the emission band and FWHM of the GaAsN shell before (orange squares) and after (purple dots) H implantation as a function of different N concentrations. The values are the average from a collection of several measurements on different NWs shown in SI 2. The error bars are given by the standard deviation. The dotted line in (d) shows the trend of the bandgap energy as a function of the N concentration according to the BAC and the area shaded orange shows the energy range of the telecommunications O-band.
• Figure 2: Effect of an increasing H-dose on the RT PL emission of the GaAs/GaAsN/GaAs core-shell NW with 1.6% N. (a) shows µ-PL spectra of NWs hydrogenated with the indicated H-dose, where H$_0$ corresponds to $10^{19}$ H-ions/cm$^2$. The interplay between the emission intensity at low and high bandgap energy is measured as a ratio of the integrated PL intensity over the respective band for increasing H-doses and shown in (b). (c) shows the PL intensity increase at RT as a function of H-dose calculated as the ratio of the integrated emission intensity over all bands of the hydrogenated and the untreated NW. The data points in panels (b) and (c) are averages of at least 3 different points from different NWs and the error bars the uncertainties on the measured PL intensity.
• Figure 3: Bandgap tuning along a NW with 1.6% N to a uniform GaAs-like energy in all regions of the NW. (a) and (b) show a map of RT µ-PL spectra as a function of the position along the axis of the same NW before and after hydrogenation. The emission energy of GaAs and GaAsN is marked by arrows. After hydrogenation only emission at GaAs-like energy is visible. All µ-PL spectra are individually normalized to 1 to highlight variations in the bandgap energy. The normalization factors are shown in (d) as the maximum PL intensity at the respective position measured in counts per second. An SEM image of the NW is shown in (c). Quantitative information regarding the maximum emission intensity, the local bandgap energy and FWHM along the NW axis is summarized in (d)-(f). For the pristine NWs the GaAs (black squares) and GaAsN-shell (orange circles) emission are analyzed separately. Data points along the hydrogenated NW are marked with purple triangles.
• Figure 4: Tunability of the bandgap and reversal of H implantation by thermal annealing. (a) $\upmu$-PL spectra showing the bandgap emission in the same point on a hydrogenated NW after annealing at different temperatures as indicated. The N concentration is 1.6%. The emission of the NW before (orange) and after (purple) hydrogenation are shown in the top row and the gradual recovery of the low energy bandgap of GaAsN after annealing at increasing temperatures in the spectra below. (b) Effects of implanted H on the pure GaAs reference samples. The µ-PL spectrum of a point on a pristine GaAs NW is shown in the top row, after hydrogenation in the second row and after mild annealing and the removing the of inserted H states in the bottom row. The spectra are normalized to the the maximum PL intensity of the pristine sample; normalization factors are given for all spectra and they reflect well those in Figure \ref{['fig02']} (c) and in the SI for both samples. At high annealing temperatures, a general decrease of signal is observed showing due to crystal quality deterioration.
• Figure 5: Two approaches for local bandgap engineering through hydrogenation. (a) illustrates the concept for achieving local bandgap tuning by a spatially selective hydrogenation approach, where part of the NW is masked by a H-opaque mask, below which GaAsN stays unpassivated. In the unmasked parts of the NW, the bandgap is shifted to the high GaAs-like energy (purple), confining the carriers in the low bandgap of the unpassivated GaAsN (orange). (b) shows a reverse approach, consisting in bandgap tuning by local laser annealing a hydrogenated high bandgap NW, locally removing the H below the laser spot, leading to low bandgap unpassivated GaAsN painted orange in the part illuminated by the laser beam. (c) shows a proof of concept for the second approach by laser annealing in the PL emission of a NW measured before hydrogenation, after hydrogenation and after laser annealing at coordinates of 2 $\upmu$m along the NW axis is shown from left to right. Owing to laser annealing, the bandgap is indeed shifted to low energy on one side of the NW, while the high energy bandgap of the hydrogenated GaAsN is conserved on the other side of the NW.