Nanosecond-scale spectral diffusion in the single photon emission of a GaN quantum dot

Abstract

Autocorrelation measurements are used to reveal the spectral diffusion time scale in the single photon emission of a GaN interface fluctuation quantum dot. Typical characteristic diffusion times of such QDs are revealed to be of nanosecond order. The excitation power dependence of the diffusion rate is also investigated, whereby an increase in the diffusion rate with increasing excitation power is observed. This result provides information on experimental conditions that will be required for the generation of indistinguishable photons.

Spectral diffusion time scales in InGaN/GaN quantum dots

“…4(b), we present the power dependence of the spectral diffusion time scales under excitation of 375 nm. The data reveal a more or less linear increase in the spectral diffusion rate with increasing power, in good agreement with studies on other materials, 23,43,45 and supporting the notion that the spectral diffusion dynamics are governed by single carrier processes (rather than multiple carrier processes such as Auger-assisted carrier escape from traps).…”

“…We note that this spectral diffusion time scale of 260 ns is longer than values reported in the literature for some other typical semiconductor QDs, such as $10-20 ns in GaN/ AlGaN, 23 $5 ns in CdSe/ZnSe, 41,43 and comparable to some InGaAs/ GaAs QDs which exhibit time scales of 90-1900 ns. 44 Although a direct comparison between different materials can be difficult, we note that the spectral diffusion times measured from these InGaN quantum dots are an order of magnitude longer than those measured in GaN QDs, 23 despite the fact that the excitation power density is over an order of magnitude greater in this case. 21,23 The relatively long spectral diffusion times measured here could be due to the existence of fairly deep charge traps (such that carrier escape and subsequent migration occur on long time scales).…”

“…44 Although a direct comparison between different materials can be difficult, we note that the spectral diffusion times measured from these InGaN quantum dots are an order of magnitude longer than those measured in GaN QDs, 23 despite the fact that the excitation power density is over an order of magnitude greater in this case. 21,23 The relatively long spectral diffusion times measured here could be due to the existence of fairly deep charge traps (such that carrier escape and subsequent migration occur on long time scales). This may be somehow linked to the fragmentation of the quantum well that forms along with the QDs during the growth process or due to compositional fluctuations in the ternary material.…”

“…The data are fitted with a typical model of an antibunched correlation with a single exponential bunching envelope. 23 As we become more spectrally selective with our filter, a pronounced bunching can be seen in the autocorrelation, characterized by g 2 ð Þ s ð Þ > 1. The bunching appears because the data points for the autocorrelation histograms are acquired by the measurement of time differences between detected photons, and we are less likely to measure two photons with longer time differences because of an increasing chance that the spectral diffusion may have shifted the emission out of the measurement window at longer times.…”

“…However, III-nitride QDs are currently known to suffer from severe spectral diffusion caused by strong Coulomb interactions between QD-confined charges and itinerant charges in the surrounding environment. [22][23][24][25][26][27] This spectral diffusion will severely limit (and possibly negate) the possibility of indistinguishable photon generation [28][29][30][31] from these structures, and thus demands a detailed investigation, particularly with regard to the time scales on which it occurs.…”