Crystal Phase Quantum Dots (CPQDs)
- Crystal Phase Quantum Dots are nanoscale semiconductor structures defined by alternating crystal phases (e.g., ZB/WZ) that create confinement via phase-dependent band offsets.
- They are fabricated by controlled phase switching during nanowire growth or by introducing stacking faults, achieving atomically sharp interfaces and tunable carrier localization.
- CPQDs exhibit distinct optical emissions, spin responses, and transport properties, offering promising applications in quantum electronics and photonic integration.
Crystal phase quantum dots (CPQDs) are quantum-dot-like nanostructures formed not by changing alloy composition, but by switching the crystal phase of a single semiconductor material. In the present literature, this typically means alternating zinc blende (ZB) and wurtzite (WZ) segments in a nanowire, or exploiting stacking faults that insert a zincblende-like region into a wurtzite host. The confinement potential is generated by phase-dependent band offsets, often in a type-II crystal-phase heterostructure with atomically sharp interfaces and low lattice mismatch. Direct realizations include WZ/ZB axial heterostructures in InP and InAs nanowires, electric-field-switched GaAs nanowires, and basal-plane stacking faults in ultrathin GaN nanowires (Yu et al., 9 Jul 2025, Wu et al., 2019, Nilsson et al., 2015, Corfdir et al., 2016).
1. Concept and heterostructure physics
The defining feature of a CPQD is that the quantum dot is phase-defined. In GaAs nanowires, the relevant polytypes are ZB, the cubic and thermodynamically stable phase, and WZ, the hexagonal and metastable phase. A short segment of one phase embedded in the other forms a crystal-phase heterostructure whose confinement arises from ZB/WZ band offsets rather than alloying or external patterning. The same principle underlies InP and InAs CPQDs, where WZ/ZB axial sections are deliberately arranged to confine carriers (Yu et al., 9 Jul 2025).
In InP nanowires, the WZ/ZB interface is explicitly described as type-II because ZB InP has a smaller band gap than WZ InP, with the conduction and valence band edges lower by about $129$ meV and $45$ meV, respectively. This makes electrons and holes occupy different phase regions and gives the dot a spatially indirect character. In InAs nanowires, the transport literature models the WZ segments as square potential barriers in the conduction band, so that a ZB segment enclosed by two WZ barriers becomes the dot region (Wu et al., 2019, Nilsson et al., 2015).
A distinct but related realization occurs in GaN. An basal-plane stacking fault inserts a thin zincblende-like region into wurtzite GaN, creating a crystal-phase heterostructure without alloy disorder or conventional interface roughness. In thicker nanowires this structure behaves as a quantum well for the bound exciton, whereas in ultrathin nanowires the same defect acquires zero-dimensional character because strong radial confinement is added to the axial confinement of the stacking fault (Corfdir et al., 2016).
A common source of confusion is the use of “quantum dot” for any nanophotonic emitter. A telecom-wavelength InAs/InP single-photon source in a photonic crystal cavity, for example, can be highly relevant to CPQD-adjacent photonics while still not being a CPQD. The droplet-epitaxy MOVPE InAs/InP dots used in that platform are explicitly not crystal-phase quantum dots in the WZ/ZB phase-engineered sense (Phillips et al., 2023).
2. Structural realizations and fabrication control
In InP, CPQDs are realized in single tapered nanowires by alternating WZ and ZB axial sections. High-angle annular dark-field scanning transmission electron microscopy identifies these sections directly: the nanowires are typically about $700$–$800$ nm long with diameters around $30$–$80$ nm, ZB segments are generally predominant, and WZ insertions are only a few to more than ten monolayers thick. Several axial motifs are observed, including long ZB segments containing a rotational twin and WZ or ZB segments sandwiched between opposite-phase regions, all of which can act as CPQDs (Wu et al., 2019).
A major advance in deterministic fabrication is electric-field-assisted phase switching during vapor-liquid-solid growth of GaAs nanowires. In that method, Au-catalyzed GaAs nanowires are grown by metalorganic chemical vapor deposition inside an in situ TEM on custom silicon micro-substrates. The crystal phase is governed by the catalyst droplet contact angle, with a critical contact angle near for the ZB/WZ transition in GaAs. The reported observations are at $129$0 V with ZB growth and $129$1 at $129$2 V with WZ growth. Because the electric field deforms the conducting Au-Ga droplet quasi-instantaneously, the phase can switch within the time of a single monolayer, enabling atomically sharp interfaces, isolated single-monolayer faults, WZ dots in ZB nanowires, ZB dots in WZ nanowires, and multiple CPQDs with controlled spacing. The same work explicitly demonstrates two ZB dots embedded in a WZ nanowire as a structure suitable for a multiple-quantum-dot / qubit platform (Yu et al., 9 Jul 2025).
The geometry of the droplet is central to that control scheme. The electric field competes with surface tension, reshaping the droplet and reducing the contact angle; the effect is independent of field polarity. A nominal critical field of about $129$3 is reported for one nanowire at a droplet-electrode distance of about $129$4 nm, although the authors emphasize that the actual field is geometry-dependent and enhanced at the droplet tip. Finite-element simulations reproduce droplet shapes to within less than $129$5 in height and contact angle and show stronger deformation for smaller nanowire radii, suggesting that narrower nanowires are easier to switch phase in (Yu et al., 9 Jul 2025).
In InAs, structural control has also been extended beyond phase switching alone. One route begins with a WZ–ZB–WZ axial heterostructure and then reduces the nanowire cross-section by isotropic radial wet etching. The QDs in that work have dot lengths of about $129$6–$129$7 nm, barrier lengths of about $129$8–$129$9 nm, and pre-etch diameters of about $45$0–$45$1 nm. Etching with citric acid + H$45$2O$45$3 mixed at $45$4, at an average radial etch rate of $45$5 nm/s, and using etch times of $45$6 s and $45$7 s, yields etched QD diameters of $45$8–$45$9 nm. This does not define the dot axially—the WZ barriers already do that—but it strengthens radial confinement and drives the system toward strongly three-dimensional confinement (Aspegren et al., 28 Nov 2025).
For transport devices in InAs, a further fabrication contribution is the use of SEM electron-channeling contrast imaging to determine WZ and ZB segment lengths directly after electrical measurements. The reported accuracy is about 0 nm for SEM/ECCI, compared with about 1 nm from TEM, and this makes it possible to correlate transport observables with the actual crystal-phase geometry of the measured device (Nilsson et al., 2015).
3. Optical signatures of dot formation
The optical phenomenology of CPQDs reflects both phase-defined confinement and the dimensionality of the bound state. In InP nanowires, micro-photoluminescence and magneto-photoluminescence at 2 K show a broad higher-energy emission from the nanowire itself together with sharp lower-energy peaks from embedded CPQDs. These dot emission energies lie roughly in the range 3–4 eV and vary from dot to dot because the lengths of the WZ and ZB segments change the confinement energies. Under higher excitation power, the relevant charged complexes are identified as the negatively charged exciton 5 and the negatively charged biexciton 6, and the 7 linewidth is reported to be around 8–9eV, indicating high-quality quantum confinement (Wu et al., 2019).
In ultrathin GaN nanowires, the decisive optical evidence for CPQD formation is a dimensional crossover. At $700$0 K, as-grown nanowires show donor-bound excitons $700$1 at $700$2 eV and $700$3-BSF-bound excitons at about $700$4 eV. After thinning the nanowires from an average base diameter $700$5 nm to $700$6 nm, the donor-bound line shifts only from $700$7 eV to $700$8 eV, whereas the $700$9-bound exciton undergoes a much larger blueshift of $800$0 meV. The interpretation given is dielectric confinement: the BSF-bound exciton is located in the thinner upper parts of the tapered wire, where radial confinement is strongest (Corfdir et al., 2016).
Time-resolved photoluminescence then distinguishes the quantum-well and quantum-dot regimes. In as-grown GaN nanowires, the $800$1 transition decays exponentially with a radiative lifetime of about $800$2 ns at $800$3 K and about $800$4 ns at $800$5 K, consistent with a two-dimensional exciton in a BSF quantum well. In ultrathin wires with $800$6 nm, the decay is initially nonexponential during the first $800$7 ns and then becomes exponential with a long component of about $800$8 ns; similar behavior is observed for $800$9 nm. The crucial observation is that this long decay component is independent of temperature up to $30$0 K, which the authors identify as the fingerprint of zero-dimensional confinement (Corfdir et al., 2016).
The GaN work also provides a compact scaling form for the radiative decay rate,
$30$1
where the coherence area shrinks with nanowire radius. Self-consistent eight-band $30$2 calculations for diameters from $30$3 to $30$4 nm show that although the oscillator strength per unit area increases as the radius decreases, the total radiative strength $30$5 decreases, so the lifetime lengthens as the wire becomes thinner. The same study fits thermal quenching with
$30$6
and reports $30$7 meV for the BSF-bound exciton, matching the energy separation between the BSF-bound state and the free exciton in fault-free regions (Corfdir et al., 2016).
These optical results establish that CPQDs are not restricted to one microscopic implementation. In one case the dot is an axial WZ/ZB phase sequence in InP; in another it is a single stacking fault in an ultrathin GaN wire. What is common is that the observed spectrum, linewidth, lifetime, and temperature dependence are governed by crystal-phase-defined confinement rather than alloy-defined heterostructures (Wu et al., 2019, Corfdir et al., 2016).
4. Spin, Zeeman response, and diamagnetic anisotropy
One of the most detailed parameterizations of CPQD spin physics is the study of WZ/ZB InP nanowire dots in magnetic field. For the charged-exciton spectrum, the transition energies are written as
$30$8
where $30$9 is the zero-field transition energy, $80$0 and $80$1 are the electron and hole $80$2-factors, $80$3, and $80$4 is the exciton diamagnetic coefficient. In this representation, the bright-state splitting depends on the sum of the electron and hole $80$5-factors, while the dark-state splitting depends on their difference (Wu et al., 2019).
For an in-plane field, the extracted $80$6-factors show a clear tensor anisotropy described by
$80$7
with $80$8 the nanowire growth axis and $80$9 a facet-related in-plane direction. The fitted values are
0
1
The hole anisotropy is much stronger than the electron anisotropy, and the authors interpret this as a structural fingerprint of the CPQD geometry (Wu et al., 2019).
The diamagnetic response is likewise anisotropic. The standard form
2
is used together with
3
so that 4 measures the average exciton extent transverse to the field. Experimentally, 5 is about 6 at 7 and decreases to about 8 at 9, again demonstrating anisotropic confinement (Wu et al., 2019).
The interpretive framework is the spin-correlated orbital current model. In this picture, the effective 0-factor contains a spin contribution that is approximately isotropic and an orbital contribution that depends strongly on the nanostructure geometry. For a WZ/ZB dot, the orbital current encloses different effective areas depending on field orientation, so the orbital moment and the resulting 1-tensor component change with angle. The same geometric logic is used to explain the diamagnetic anisotropy (Wu et al., 2019).
These extracted parameters are presented as a practical basis for band-gap engineering, spin-state control, spin-precession control, 2-tensor modulation, and potentially spin resonance in crystal-phase low-dimensional structures. A plausible implication is that CPQDs offer an experimentally parameterized route to spin engineering in nanowires where the relevant control variable is the phase sequence itself rather than only chemical composition.
5. Single-electron transport and strong-confinement electrostatics
The transport literature establishes that CPQDs can be electrically functional few-electron devices. In InAs nanowires, two thin WZ segments inserted into an otherwise ZB wire create a WZ–ZB–WZ structure in which the WZ sections act as tunnel barriers and the enclosed ZB segment is the dot. Growth control is achieved by changing the AsH3/TMIn ratio: a lower ratio favors WZ and a higher ratio favors ZB. The resulting devices display lower conductance than pure ZB references, a much larger pinch-off voltage, and a difference in pinch-off voltage of about 4 V at 5 K between a CPQD device and a ZB reference device (Nilsson et al., 2015).
At low temperature and small source-drain bias, these InAs devices show regular Coulomb oscillations over a gate-voltage range of 6–7 V. Near depletion the peak spacing is somewhat irregular, but then more than a hundred peaks can appear with nearly constant spacing. Standard Coulomb-blockade relations are used,
8
together with
9
For three longer-dot devices the extracted values are approximately $129$00 meV with $129$01 aF, $129$02 meV with $129$03 aF, and $129$04 meV with $129$05 aF. For the shortest reported dot, $129$06 nm long, $129$07 meV, $129$08 aF, and $129$09 meV; at more negative gate voltages, $129$10 increases to about $129$11–$129$12 meV (Nilsson et al., 2015).
A central quantitative result is a lower bound of about $129$13 meV for the ZB–WZ conduction-band offset. This is inferred by counting the number of Coulomb oscillations from near depletion until the oscillations vanish, estimating the electron density from the dot volume, and relating the density to the Fermi level with a nonparabolic conduction-band expression. The reported lower-bound barrier heights are $129$14 meV, $129$15 meV, and $129$16 meV for three devices (Nilsson et al., 2015).
The later strong-confinement work extends this picture by adding post-growth radial etching to the crystal-phase barriers. The relevant electrostatic relations are
$129$17
With this approach, the maximum observed charging energy exceeds $129$18 meV. The strongest device has $129$19 meV at a diameter $129$20 nm; other etched devices span $129$21 meV to $129$22 meV for diameters from $129$23 nm down to $129$24 nm, while unetched reference dots at $129$25–$129$26 nm remain at about $129$27–$129$28 meV (Aspegren et al., 28 Nov 2025).
The onset of electron filling shifts systematically with diameter. Reported examples are $129$29 V for the thinnest dot D1, $129$30 V for D6, and $129$31 V for the unetched reference D8. The interpretation is stronger radial confinement: thinner wires push the subband energies upward, so a more positive gate bias is needed for first occupation. Finite-element simulations using COMSOL and density-gradient quantum confinement reproduce the increase of $129$32 with decreasing diameter, but also show that the increase is weaker than pure geometric scaling would suggest because stray capacitances from metallic contacts, unetched nanowire segments, semiconductor leads, and electrostatic screening become non-negligible in the smallest-diameter regime (Aspegren et al., 28 Nov 2025).
This identifies two limits of CPQD electrostatics. For large diameters, gate-induced depletion of the leads reduces $129$33 and produces a saturation-like behavior in $129$34 for $129$35 nm. For very small diameters, further radial shrinking yields diminishing returns because stray capacitances do not scale away proportionally. The same study reports extracted $129$36-factors roughly in the range $129$37 to $129$38, with no clear suppression of $129$39 with reduced diameter, and explicitly connects these strongly confined CPQDs to the interpretation of strong spin-orbit interaction and possible polarization-charge effects at WZ/ZB interfaces (Aspegren et al., 28 Nov 2025).
6. Nanophotonic interfaces and adjacent photonic-crystal platforms
CPQDs are frequently discussed as candidates for single-photon emitters, photon cascades, quantum-state coupling and entanglement, and broader quantum-information and spintronic applications. In the provided literature, however, most mature photonic-crystal demonstrations use other quantum-dot platforms. These results are therefore best understood as adjacent nanophotonic benchmarks rather than direct CPQD demonstrations.
A clear example is the telecom C-band single-photon source based on MOVPE-grown droplet-epitaxy InAs/InP QDs in an L3 photonic crystal cavity. The device operates in the weak-coupling regime and uses temperature and excitation-power tuning to bring the QD into resonance with the cavity mode. The reported bulk lifetime is $129$40 ns, the cavity-modified lifetime is $129$41 ns at resonance near $129$42 nm, and the inferred Purcell factor is about $129$43. The single-photon purity is $129$44 at $129$45 K and $129$46 at $129$47 K, supporting cryogen-free-adjacent operation. The paper explicitly states that this is not a CPQD platform, but it is relevant to the broader landscape of engineered telecom quantum-dot photonics (Phillips et al., 2023).
Waveguide quantum electrodynamics provides another adjacent reference point. In a GaAs photonic crystal membrane waveguide with embedded InAs quantum dots, the beta-factor is defined as
$129$48
and extracted experimentally using
$129$49
By temperature-tuning a single QD across the waveguide band edge, decay rates up to $129$50 are measured, corresponding to a Purcell factor of $129$51, and beta-factors up to $129$52 are extracted. This establishes the performance envelope of photonic crystal waveguides as high-efficiency on-chip single-photon channels (Thyrrestrup et al., 2010).
Collective waveguide-mediated optics has also been demonstrated with two InAs quantum dots in a photonic crystal waveguide. After strain-tuning one dot into resonance with the other, the photoluminescence statistics show a bunching peak $129$53, the normalized transmission dip becomes deeper than for either dot alone, and the resonant pair linewidth broadens to about $129$54 GHz. The interpretation is superradiant collective emission and enhanced few-photon nonlinearity in the common waveguide mode (Grim et al., 2022).
On-chip addressing of separated emitters has been realized with self-assembled InGaAs quantum dots in two L3 photonic crystal cavities connected by an in-plane waveguide. Proton implantation at $129$55 keV and $129$56 completely electrically isolates the two contact regions, allowing independent Stark tuning over more than $129$57 nm while preserving optical connectivity through the waveguide. Secondary-to-primary peak ratios of $129$58, $129$59, and $129$60 are reported depending on resonance overlap. This is not a CPQD experiment, but the architecture is directly relevant to any phase-defined emitter network that requires electrical independence and optical connectivity on one chip (Thon et al., 2011).
Theoretical proposals further extend the architectural picture. One scheme uses two nonidentical charged GaAs/AlGaAs quantum dots in two directly coupled photonic crystal nanocavities and classical laser fields to realize a controlled phase gate through virtual excitations only, so that neither the cavity modes nor the QDs are actually populated during the operation (Zhang et al., 2010). A related proposal uses two non-identical, spatially separated charged GaAs/AlGaAs quantum dots in a single-mode photonic crystal cavity to generate a conditional unconventional geometric phase by a state-dependent displacement of the cavity field along a closed path in phase space (Zhang et al., 2010). A plausible implication is that once CPQDs are integrated into comparable nanophotonic hardware, the relevant control abstractions—independent tuning, cavity-mediated coupling, virtual-excitation gates, and waveguide-mediated collective scattering—are already well developed in adjacent quantum-dot systems.
Visible-wavelength photonic crystal cavity integration provides a further materials contrast. In$129$61Ga$129$62As/GaP self-assembled dots in a GaP membrane show room-temperature visible-wavelength photoluminescence, biexponential decay with lifetimes of approximately $129$63–$129$64 ps, and cavity-enhanced outcoupling in a linear three-hole defect cavity with a fundamental resonance at $129$65 nm and $129$66. Narrow lines indicative of single quantum dot emission are observed, but photon statistics were not obtained because the signal-to-noise ratio was insufficient (Rivoire et al., 2012).
Taken together, these adjacent photonic-crystal results sharpen an important boundary in the terminology. A CPQD is defined by crystal-phase engineering of a chemically uniform semiconductor. A quantum dot in a photonic crystal cavity or waveguide is not, by itself, a CPQD. At the same time, these nanophotonic works delineate the device-level functionalities—Purcell enhancement, high $129$67-factor coupling, collective scattering, independent electrical tuning, and cavity-mediated gate proposals—that are likely to be central when CPQDs are deployed as active emitters in integrated photonic and quantum-information architectures.
