Overview
This study, co-authored by Jungmok Seo (then at the School of Electrical and Electronic Engineering, Yonsei University), who now leads the BLISS Lab at Yonsei University, used optical-pump terahertz-probe spectroscopy to follow carrier recombination along the transverse direction of silicon nanowires (SiNWs). Comparing bundled, individual and SiO2-encapsulated SiNWs, the authors found that Auger scattering counterbalances surface-trap-mediated decay, and the encapsulated SiNWs showed a carrier lifetime of around 1.1 ns at a pump fluence of 50 μJ/cm².
Background
Photoexcited carrier dynamics limit the performance of silicon-based photonic devices, and in SiNWs the carrier lifetime depends on direction relative to the wire axis, on wire-to-wire interconnection and on the many surface traps. Reported lifetimes had been attributed mainly to the longitudinal direction, and the transverse lifetime was barely understood. Surface encapsulation reduces trap centres but enhances multi-particle Auger scattering under strong pumping, and optical-pump optical-probe measurements also pick up photo-induced absorption from trapped carriers. A low-energy terahertz (THz) probe avoids exciting carriers from trap levels, which lie far above the probe photon energy.
Design and approach
SiNWs were made from p-type (100) silicon by metal-assisted chemical etching (MACE) combined with nanosphere lithography and transferred with polydimethylsiloxane stamps as a single vertically aligned layer onto a cellophane tape substrate, which is transparent in the THz range. The wires were about 60 μm long and about 500 nm in diameter, with areal densities of 30% for bundled and 3% for individual and encapsulated SiNWs. Encapsulated SiNWs received a 5-nm-thick SiO2 layer by thermal oxidation at 1000 °C for 4 min. A coupled rate-equation model containing bulk indirect recombination, saturable trap-mediated recombination and an Auger term was used to compare trap density with photoexcited carrier density.
How it was evaluated
Pump pulses of 50 fs at 1.55 eV from a 250 kHz Ti:sapphire regenerative amplifier excited the samples with polarization at 45° to the wire axis, at fluences up to 700 μJ/cm², and 0.5-3 THz probe pulses detected the pump-induced change in transmitted THz field at room temperature in ambient air. A 500-nm-thick bulk Si film served as the reference. Time constants were extracted by single-exponential fits for bundled and individual SiNWs and by bi-exponential fits for encapsulated SiNWs, and all transients were also fitted with the rate-equation model, which included the Auger term only for individual and encapsulated SiNWs.
Key results
The carrier lifetime of the bulk Si film was on the order of a microsecond, whereas decay in both bundled and individual SiNWs was much faster, which the authors attribute to defect trapping; the signal of individual SiNWs decayed faster than that of bundled SiNWs.
In bundled SiNWs the decay became slower with increasing pump fluence, with the slow decay appearing above 100 μJ/cm², which the authors explain by saturation of surface traps and a shift to bulk-like indirect recombination. Individual SiNWs instead showed a nearly fluence-independent decay, interpreted as Auger scattering counterbalancing trap-mediated recombination. For individual SiNWs, a longest free-carrier lifetime of 300 ps and a diffusion constant of 24 cm²/s gave a diffusion length of 0.85 μm and a surface recombination velocity of 0.4 × 10^5 cm/s, so carriers reach the surface at least once before recombining. In bundled SiNWs, Auger recombination played a minor role because carriers diffuse into adjacent interconnected wires.
Encapsulated SiNWs had a carrier lifetime of around 1.1 ns at 50 μJ/cm², much longer than that of individual SiNWs, and the diffusion length increased to approximately 1.6 μm, corresponding to at least three surface scattering events in a 500 nm diameter wire. The decay was a simple exponential at 50 μJ/cm² but its initial rate rose sharply at 700 μJ/cm², and the peak signal of the encapsulated SiNWs followed what the authors describe as a third-order fluence dependence, whereas it increased linearly for individual SiNWs. The authors attribute this new channel to multi-particle Auger-type scattering, and only the encapsulated SiNWs required a bi-exponential fit.
In the rate-equation analysis, the fitted initial trap density of the encapsulated SiNWs was about a quarter of that of the individual SiNWs, and the trap density was lower in bundled SiNWs than in individual SiNWs because of their interconnection. When the ratio of trap density to photoexcited carrier density fell below 2, traps saturated easily in the encapsulated SiNWs and Auger recombination dominated, whereas trap-mediated recombination remained dominant in individual SiNWs. The authors note that the absence of a clear Auger-induced fast decay in individual SiNWs reflects incomplete counterbalance of surface trapping rather than ineffective Auger scattering.
Related research
Direct Imaging of Free Carrier and Trap Carrier Motion in Silicon Nanowires by Spatially-Separated Femtosecond Pump–Probe Microscopy
The related paper used spatially separated femtosecond pump-probe microscopy to image photoexcited carriers spreading along the axis of a single silicon nanowire, reporting rapid ambipolar diffusion and surface recombination followed by slower trap carrier migration. The present study measures recombination across the wire direction in nanowire layers with terahertz probing.
Abstract-based introductionSources used in this summary
This summary is based on the full publisher PDF of the article, including figure legends; symbols and several fluence values missing from the PDF text layer were read from the page images, and the Fig. 4 legend gives the fit to the encapsulated-SiNW peak signal as proportional to F^1/3, whereas the text describes a third-order dependence.