A laboratory study found that an individual crystal-phase quantum dot, or CPQD, inside an AlGaAs nanowire produced bright emission lines with little background light and unusually narrow widths. The two lines the study identifies as exciton and biexciton were 104 and 61 microelectronvolts wide, respectively. The paper gives an uncertainty of 4 microelectronvolts alongside that linewidth pair, without specifying separate uncertainties for the two lines.
The same emitter also showed the photon statistics expected of single-photon emission. Its zero-delay second-order intensity correlation, the measure used to assess whether the light arrived as isolated photons, was below 0.5, while a higher-resolution supplementary example was below 0.1. A cross-correlation between the exciton and biexciton signals also showed bunching consistent with a biexciton-exciton cascade, meaning a linked sequence of the two emissions. The main correlation dip was limited by the timing resolution of the setup.
A crystal switch inside the wire
The dots form in a core-shell AlGaAs nanowire system that contains two crystal phases, wurtzite and zincblende. Core diameters ranged from 7 to 20 nanometres, zincblende insertions appeared at 15 to 50 per micrometre, and a typical insertion was less than 3 nanometres in size. The optical work focused on core emission between 700 and 750 nanometres.
To examine the structure and composition, the researchers used HAADF-STEM and EDX. The nanowires were grown by Au-catalysed vapor-liquid-solid MBE on Si(111) under arsenic-rich conditions for 25 minutes at 510 degrees Celsius, with a nominal AlAs/GaAs ratio of 0.4. Optical measurements were made at 1.5 to 6 K using 532 nm continuous-wave excitation and individual-nanowire micro-photoluminescence, with approximately 0.02 nm spectral resolution.
The composition changed in the shell
EDX found that the aluminium-composition parameter used for the shell rose from 0.4 in the wurtzite phase to 0.6 in the zincblende insertions.
Other measurements pointed to a different picture for the core. At the nanowire tip, there was no intensity drop at the zincblende insertions nearest the catalyst droplet, while a farther insertion showed a shell-related drop. EDX values were 0.38 plus or minus 0.03 on the zincblende insertion and 0.34 plus or minus 0.01 on the adjacent wurtzite segment. The authors concluded that the aluminium fluctuations associated with the phase switch were confined to the shell.
Why the authors favor a type-I picture
The composition measurements feed into the paper's central interpretation. The authors link the clean emission to type-I band alignment and the low density of zincblende insertions. Here, type-I band alignment is the authors' name for their proposed picture of the dot's optical states. Crucially, the assignment is inferred from optical signatures rather than established by a direct measurement of the band offsets.
The timing results were part of that case. The fitted antibunching time was 0.90 nanoseconds, and the fast component of the measured lifetime was 1 nanosecond. The authors compared these values with the 3 to 10 nanosecond lifetimes cited in the paper for type-II CPQDs, and interpreted the shorter times as support for the type-I picture. That comparison does not by itself establish that every AlGaAs CPQD has the same band alignment.
More than one optical signature
Polarization-resolved photoluminescence found an exciton fine-structure splitting of 235 plus or minus 30 microelectronvolts, a separation between two polarization-resolved exciton lines. Magneto-photoluminescence in a 9-tesla vertical magnet showed Zeeman splitting into two lines and a diamagnetic shift. The authors describe these optical signatures as resembling those of typical type-I quantum dots.
Two other examples in the supplementary material also showed sharp emission lines of approximately 175 microelectronvolts and single-photon statistics. Their fitted antibunching times were 0.46 and 0.56 nanoseconds; one also showed a 1-nanosecond lifetime. The additional examples broaden the report's optical picture, but they remain illustrative measurements rather than a count of the platform's full performance.
A result still at laboratory scale
The scale of the evidence matters. The paper presents one main optical emitter and two supplementary optical examples; it does not report the total number of nanowires or emitters. The report also provides no population-level estimate of variability. The findings therefore describe what these measured structures did under the reported conditions, rather than a demonstrated performance level for all AlGaAs CPQDs.
The main photon-correlation measurement had an important technical limit. Its overall timing jitter was estimated at 505 picoseconds from detector and time-tagger specifications rather than measured directly, and the actual value may be higher. That limited the depth of the main zero-delay correlation dip. The time-resolved measurements used a 440 nm pulsed diode running at 20 MHz, with 200-picosecond pulses.
The document is arXiv:2608.28353v1, dated 28 August 2026, and is identified as an arXiv preprint. It reports laboratory-scale optical evidence, not deterministic quantum-dot arrays or a deployed quantum technology. The comparison with type-II CPQDs is contextual rather than a matched control, and the type-I assignment remains an optical interpretation rather than a direct band-offset measurement. The work leaves open how consistently the behavior can be reproduced across different nanowire diameters and growth conditions.
Paper data and sources
Original title: Crystal-phase quantum dots in AlGaAs nanowires
Authors: Rohan Radhakrishnan, Rodion Reznik, Gilles Patriarche et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-28
DOI: Not available
Original paper · Full text