Time-resolved nanospectroscopy of III-V semiconductor nanowires

Abstract

We investigate ultrafast electron dynamics in individual GaAs/InGaAs core-shell nanowires using near-infrared pump-mid-infrared probe nanospectroscopy based on a scattering-type scanning near-field technique. Our results reveal a distinct blue shift in plasmon resonance frequency induced by photodoping. By extracting time-dependent electron densities and scattering rates, we gain insights into the effects of chemical doping and nanowire surface states on recombination dynamics and carrier mobility. Varying the pump power over two orders of magnitude reveals carrier recombination times in the range from a few ps at high power to 100 ps at low power, dominated by bimolecular recombination. Our findings highlight the potential of time-resolved nanoscopy for contactless probing of free carrier mobility and recombination dynamics on a local scale in individual semiconductor nanostructures or nanodevices.

Fig. 1. Schematic of the pump-probe s-SNOM setup integrated with a broadband nano-FTIR module. The MIR probe and external NIR pump beams are focused onto the AFM tip using a parabolic mirror.

Fig. 2. (a) AFM topography of dispersed NWs on Si(001) substrate. The white dashed rectangle highlights the NW selected for high-resolution mapping. (b) High-resolution s-SNOM amplitude map of the highly-doped NW on Si, acquired in ‘white light’ mode. The blue line marks the position of the longitudinal profile along the NW axis, shown alongside the map. (c) and (d) Near-field amplitude s(ω) and phase ϕ(ω) spectra of the highly-doped NW, measured at the center of the NW's top facet and normalized to the Si substrate response. The experimental data were fitted using a point-dipole model incorporating the frequency-dependent Drude–Lorentz permittivity, with fitting parameters for plasma frequency ω_pl, scattering rate γ, and weight factor c.

Fig. 3. (a) and (b) the near-field amplitude s(ω) and the phase ϕ(ω) spectra of the highly-doped NW, measured at the different delay times between nano-FTIR probe and NIR-pump (P_avg = 13 mW) and normalized to the response of the Si substrate. The plotted spectra correspond to a few representative time delays, and are fitted using a three-parameter point-dipole model based on the frequency-dependent Drude–Lorentz permittivity. (c) and (d) Color maps illustrating the evolution of s(ω) and ϕ(ω) upon NIR photoexcitation over the entire range of pump-probe delay times, with each line representing normalized near-field spectra at different delays. (e) and (f) Fitted values of the plasma resonance frequency ω_pl and the scattering rate γ_pl as functions of pump-probe delay time, showing decay times of t = 3.3 ps and t = 3 ps, respectively. (g) Pump-induced change in the spectrally integrated scattered intensity as a function of pump-probe delay time for the highly-doped NW, normalized to the unpumped baseline.

Fig. 4. (a) Temporal evolution of the total electron density n for the highly-doped, moderately-doped, and undoped NWs, showing an exponential decay with time constants of t = 2.8 ps, t = 4.3 ps, and t = 3.5 ps, respectively. (b) Scattering rate γ as a function of total electron density n for the highly-doped, moderately-doped, and undoped NWs. The solid lines approximate the linear density dependence of the total scattering rate upon photoexcitation. The dashed line connects the points corresponding to the minimal electron density for each NW type, highlighting chemical doping impact on the scattering rate.

Fig. 5. Pump-induced changes in the scattered intensity as a function of pump-probe delay time and pump power for highly-doped (a) and undoped (b) NWs. All pump-probe traces are normalized to the unpumped baseline. (c) 1/e decay times extracted from monoexponential fits of the pump-probe traces for each NW type as a function of photogenerated electron density Δn.