GaAs–GaP Superlattice Nanowires

• GaAs–GaP superlattice nanowires are one-dimensional heterostructures featuring periodic GaAs and GaP segments that engineer electronic valley folding and phonon dispersion.
• They are fabricated via Au-assisted chemical beam epitaxy to form sharp, strain-relaxed interfaces that accommodate a 3.7% lattice mismatch.
• Nonmonotonic thermal conductivity measurements reveal a coherent–incoherent phonon transport crossover, enabling tunable thermal management in nanoscale devices.

Searching arXiv for papers on GaAs-GaP superlattice nanowires, phonons, and electronic structure. GaAs–GaP superlattice nanowires are axial III–V heterostructures in which GaAs and GaP are stacked periodically within a nanowire geometry, producing a one-dimensional superlattice whose structural periodicity, lattice mismatch, and finite cross section jointly reshape both electronic and vibrational spectra. In the recent literature, these systems are treated not merely as nanoscale heterostructures but as coupled phononic and optoelectronic platforms: the superlattice period controls acoustic and optical phonon dispersion, thermal conductivity exhibits a nonmonotonic dependence on period, and the underlying electronic structure is governed by valley folding, strain response, confinement, and surface relaxation rather than by a single bulk-like conduction minimum (Sivan et al., 2023, Arya et al., 13 Aug 2025, Santos et al., 2018).

1. Material system, architecture, and epitaxial context

GaAs/GaP superlattice nanowires have been realized as axial GaAs/GaP superlattices embedded inside GaP nanowires and as axial GaAs–GaP superlattice nanowires grown by Au-assisted chemical beam epitaxy (CBE) on GaAs (111)B substrates. The nanowire architecture reported in the experimental studies is segmented rather than uniform: a GaAs stem is followed by a GaP segment, then by a GaAs/GaP superlattice segment, and, in many samples, by a top GaP segment. One study gives representative as-grown nanowire dimensions of $3.58 \pm 0.24~\mu\text{m}$ in length and $40 \pm 5~\text{nm}$ in diameter, while the thermal-transport study reports a core diameter in the SL region of 30–50 nm together with a surrounding shell of about 20 nm thickness, composed mainly of GaP (Sivan et al., 2023, Arya et al., 13 Aug 2025).

The superlattice period is defined as

$L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$

with experimentally studied periods including 4.8, 6.0, and 10.0 nm in the Raman/Brillouin study and 4.8 to 23.3 nm in the thermal-transport study. Reported examples include $2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$ for a 4.8 nm period, $3.0 + 3.0\ \text{nm}$ for 6.0 nm, and two 10.0 nm designs, namely $4.2 + 5.8\ \text{nm}$ and $5.0 + 5.0\ \text{nm}$. Most thermal-transport samples contain 100 repetitions, whereas the 14.6 nm and 23.3 nm samples contain 30 repetitions and do not include the top GaP segment (Arya et al., 13 Aug 2025).

A central materials motivation is the about 3.7% lattice mismatch between GaAs and GaP. The nanowire geometry is therefore important not only as a dimensional reduction but as an epitaxial strategy: the literature emphasizes that nanowires can relax mismatch strain radially, permitting epitaxial GaAs/GaP interfaces without misfit dislocations in a system that is difficult in conventional planar heterostructures (Sivan et al., 2023).

2. Crystal phase, orientation, and interface quality

The crystallographic description differs somewhat across the electronic and phononic literature, but several orientations and phases recur systematically. The binary electronic-structure study treats both zinc-blende (ZB) and wurtzite (WZ) GaAs and GaP, with nanowires along $[111]$ in ZB and $[0001]$ in WZ, using hexagonal cross sections bounded by six side facets. In contrast, the phonon-engineering study models wurtzite GaAs/WZ GaP superlattices stacked along the [0001] crystal axis, and its Raman analysis of the reference GaP nanowire identifies WZ-specific modes such as $E_2^H$, supporting the relevance of a wurtzite nanowire phase (Santos et al., 2018, Sivan et al., 2023).

Interface quality is a recurring experimental theme. High-resolution transmission electron microscopy (HRTEM/TEM) resolves the alternating GaAs and GaP layers and shows sharp interfaces between them. The transport paper states explicitly that the nanowires possess sharp interfaces between GaAs and GaP layers, as confirmed by HRTEM. At the same time, the metrology is limited in scope: the experiments do not quantitatively report numerical interface roughness, quantified intermixing width, dislocation density, or detailed defect density. Theoretical treatments sometimes assume sharp but gradual interfaces over two unit cells as a model, but that is presented as a computational assumption rather than as a direct experimental fit (Arya et al., 13 Aug 2025).

This distinction is important because much of the reported physics relies on periodicity remaining meaningful. The observation of clear layer alternation and sharp interfaces supports the use of the superlattice concept, while the lack of full roughness statistics leaves open how far idealized folding and coherence arguments remain valid under realistic interface morphology.

3. Electronic structure: valley folding, strain, confinement, and pseudodirect character

The most detailed electronic framework available for GaAs–GaP superlattice nanowires is indirect: it comes from a study of bulk GaAs and GaP and of thin GaAs and GaP nanowires in ZB and WZ phases rather than from an explicit GaAs/GaP heterostructure calculation. Its core result is that the nanowire band edges can be understood in terms of bulk conduction valleys that are folded onto the nanowire axial Brillouin zone and then shifted by biaxial strain, confinement, and especially surface relaxation (Santos et al., 2018).

In the binary reference systems, the calculated equilibrium gaps are direct for GaAs, ZB $40 \pm 5~\text{nm}$0, direct for GaAs, WZ $40 \pm 5~\text{nm}$1, indirect for GaP, ZB $40 \pm 5~\text{nm}$2, and direct in their calculation for GaP, WZ $40 \pm 5~\text{nm}$3. For ZB GaP, the conduction-band minimum lies near X, specifically near

$40 \pm 5~\text{nm}$4

along the $40 \pm 5~\text{nm}$5 direction, while the $40 \pm 5~\text{nm}$6-point direct gap is $40 \pm 5~\text{nm}$7, about $40 \pm 5~\text{nm}$8 above the indirect minimum. The authors also discuss the literature expectation that WZ GaP is pseudodirect, linked to the proximity of the ZB-derived $40 \pm 5~\text{nm}$9 and $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$0 states, even though their explicit WZ calculation is described as direct.

For the thin nanowires studied in that work, the gap character reverses relative to bulk in several cases. The reported results are indirect for ZB GaAs [111] nanowires, direct for ZB GaP [111] nanowires, indirect for WZ GaAs [0001] nanowires, and indirect for WZ GaP [0001] nanowires. The two competing nanowire conduction minima are labeled $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$1 and $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$2, and their ordering determines whether the gap is direct, indirect, or effectively pseudodirect.

The folding argument is geometrically explicit. By enlarging the bulk unit cell in the planes perpendicular to the wire axis, bulk dispersions from off-axis lines are folded into the nanowire axial $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$3-A direction. In this framework, the nanowire valleys do not emerge independently; rather, they are folded descendants of bulk valleys. This matters directly for superlattice nanowires because an axial GaAs/GaP compositional modulation adds another source of zone folding. A plausible implication is that momentum-space “directness” at the superlattice zone center need not imply strong optical activity if the low-energy state retains folded $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$4- or $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$5-derived character.

The same work also analyzes biaxial strain applied in the plane perpendicular to the growth direction, using the conventions

$L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$6

For ZB GaAs, increasing in-plane compression drives the conduction-band minimum sequence

$L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$7

with a direct-to-indirect transition when the in-plane lattice parameter is reduced by about 3.2%. For ZB GaP, the material remains indirect across the strain range spanning GaP to GaAs lattice constants, but the minimum switches from near $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$8 to $L = L_{\mathrm{GaAs}} + L_{\mathrm{GaP}},$9 when stretched by about 1.4%. These trends suggest that in a GaAs/GaP superlattice nanowire, valley ordering should be treated as strain-selective rather than as a rigid shift of a single conduction edge.

A major corrective to a common simplification also emerges here: the authors conclude that the change in direct or indirect gap character in thin nanowires is due mainly to surface relaxation effects, not confinement. Confinement opens the gaps, but full lateral relaxation can reverse the valley ordering.

4. Phonon spectra and superlattice-induced mode engineering

The most direct spectroscopic evidence for GaAs/GaP superlattice behavior comes from single-nanowire micro-Raman spectroscopy and Brillouin light scattering (BLS), supported by ab initio density functional perturbation theory (DFPT) calculations. The phononic result is that the superlattice behaves as a distinct vibrational system whose spectra are not reducible to those of bulk GaAs, bulk GaP, or a non-superlattice GaP nanowire (Sivan et al., 2023).

In the reference GaP nanowire, Raman in $2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$0 shows peaks at about 352.2 cm$2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$1 $2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$2, 362.2 cm$2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$3 $2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$4, 371.0 cm$2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$5 (surface optical), and 400.7 cm$2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$6 (LO). In a 4.8 nm-period superlattice nanowire, the Raman spectrum changes substantially. In the GaAs-like window between about 240–300 cm$2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$7, the main peaks are 261.3, 276.4, and 287.6 cm$2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$8; in the GaP-like window between about 340–400 cm$2.0\ \text{nm GaAs} + 2.8\ \text{nm GaP}$9, the main peaks are 352.8, 358.3, 368.1, 376.7, 388.2, and 391.98 cm$3.0 + 3.0\ \text{nm}$0. These lines are interpreted as superlattice phonons created by the enlarged periodicity.

The mechanism is described as zone folding / back folding. Because the superlattice enlarges the real-space unit cell, the Brillouin zone shrinks, and phonons that would lie away from $3.0 + 3.0\ \text{nm}$1 in the bulk are folded back to the new zone center, where some become Raman active. The calculations further show that the number of Raman-active modes increases with period, since a longer period means more atoms per superlattice unit cell and therefore more phonon branches. The reported calculated Raman-active frequencies for 3.83 nm, 5.12 nm, and 6.39 nm periods document both branch multiplication and period-dependent frequency shifts.

BLS provides the corresponding low-frequency picture. The reference nanowire shows peaks at 25.6, 30.8, 37.1, and 101.3 GHz. By contrast, the superlattice nanowires show a richer acoustic spectrum: for 10 nm, peaks at 25.4, 29.2, 36.1, 48.3, 68.1, 90.7, 95.43, and 99 GHz; for 6 nm, peaks at 26.4, 36.6, 49.6, 72.1, and 95.6 GHz; and for 4.8 nm, peaks at 26.4, 36.1, 40.5, 50.1, 68.5, 86.1, and 97 GHz. The scattering wave vector is written as

$3.0 + 3.0\ \text{nm}$2

and for 532 nm excitation the paper estimates

$3.0 + 3.0\ \text{nm}$3

The experimental and DFPT results therefore support a consistent phononic picture: the GaAs/GaP superlattice acts as a periodic medium with modified optical and acoustic dispersion, increased mode count, and tunable phonon frequencies. The work does not claim a complete nanowire-specific mode assignment, since the calculations are for a bulk superlattice rather than a finite-diameter nanowire; surface modes and confinement-specific acoustic modes are explicitly listed as omissions.

5. Thermal transport and the coherent–incoherent crossover

Direct thermal-transport measurements on individual GaAs–GaP superlattice nanowires were carried out with the suspended thermal bridge method, using two suspended SiN$3.0 + 3.0\ \text{nm}$4 platforms with Pt serpentine resistors acting as heater and thermometer, operated in a Janis ST-500 probe station under $3.0 + 3.0\ \text{nm}$5–$3.0 + 3.0\ \text{nm}$6 mbar. The platform gap was approximately $3.0 + 3.0\ \text{nm}$7. The nanowire thermal conductance is extracted from the measured temperature-rise slopes according to

$3.0 + 3.0\ \text{nm}$8

and the thermal conductivity is then written as

$3.0 + 3.0\ \text{nm}$9

The measurements cover 16–350 K, with period-dependent data shown explicitly at 140 K and 300 K (Arya et al., 13 Aug 2025).

The central observation is a nonmonotonic thermal conductivity as a function of superlattice period. As the period is decreased from 23.3 nm toward shorter values, $4.2 + 5.8\ \text{nm}$0 first decreases, reaches a minimum around 8 nm, and then increases again for still shorter periods such as 4.8 nm. This minimum is interpreted as the signature of a crossover from incoherent to coherent phonon transport. On the long-period side, interfaces act as largely independent scattering barriers, so increasing interface density lowers $4.2 + 5.8\ \text{nm}$1. On the short-period side, phonons retain phase coherence across multiple interfaces, the superlattice behaves as a periodic medium, and the transport is influenced by interference, minibands, possible forbidden gaps / stop bands, and modified group velocity.

The temperature dependence supports the same interpretation. The peak temperatures of $4.2 + 5.8\ \text{nm}$2 are 161 K for a GaP reference nanowire, 173 K for a 10 nm superlattice nanowire, and 218 K for a 4.8 nm superlattice nanowire. At low temperature, from about 15 to 70 K, the measured conductivity roughly follows a $4.2 + 5.8\ \text{nm}$3 dependence. The persistence of the nonmonotonic period dependence up to room temperature is emphasized as a major result, because it indicates that surface boundary scattering and phonon-phonon scattering do not fully destroy the interference effect in these nanowires.

Two theoretical frameworks support the experiments. Ab initio lattice dynamics + BTE, performed with VASP, PAW, LDA, and almaBTE, reproduces a minimum in $4.2 + 5.8\ \text{nm}$4 at about 5 nm and shifts the calculated $4.2 + 5.8\ \text{nm}$5 maxima from 12 K in bulk GaP to 125 K for a 5.1 nm SL and 115 K for a 10.2 nm SL. Nonequilibrium molecular dynamics (NEMD) with LAMMPS and a bond-order (Tersoff-type) potential also shows a minimum, though at a shorter period than the BTE result. The exact crossover period therefore differs across experiment, BTE, and NEMD, but all three approaches support the same qualitative feature: a conductivity minimum separating incoherent and coherent transport regimes.

An important experimental limitation is that the measured $4.2 + 5.8\ \text{nm}$6 is an effective conductivity of the suspended nanowire assembly. It includes contributions from the superlattice segment, adjacent pure GaP segments in series, and contact resistances; no explicit de-embedding of contact resistance is reported.

6. Interpretive framework, misconceptions, and open problems

Three interpretive points organize the present understanding of GaAs–GaP superlattice nanowires. First, periodicity matters twice: it folds phonon dispersions directly in the superlattice, and, by analogy with the binary nanowire electronic-structure problem, it is expected to fold electronic valleys into the axial one-dimensional Brillouin zone. Second, strain is valley-selective. The binary GaAs/GaP results show that conduction valleys shift with different slopes under biaxial strain, so a superlattice should not be modeled reliably by a single “$4.2 + 5.8\ \text{nm}$7-like” conduction-band picture. Third, surface relaxation can be decisive in thin wires, which corrects the frequent assumption that quantum confinement alone determines the direct or indirect character of the gap (Santos et al., 2018, Sivan et al., 2023, Arya et al., 13 Aug 2025).

Several objective caveats remain. The phonon-engineering study demonstrates tunability of vibrational spectra and infers a reduced bandgap relative to the bulk constituents, with corrected gap estimates of 1.38 eV for a 5.11 nm SL, 1.36 eV for 6.39 nm, and 1.29 eV for 10.22 nm, compared with 1.46 eV for bulk WZ GaAs and 2.28 eV for bulk WZ GaP. However, that bandgap reduction is inferred from wavelength-dependent Raman resonance and DFT/LDA plus an estimated GW correction, not from direct absorption or photoluminescence. The thermal-transport study measures coherent–incoherent crossover signatures but does not isolate the intrinsic conductivity of the SL segment alone. The electronic-structure study supplies a transferable valley-folding framework but does not compute explicit GaAs/GaP heterointerfaces, band offsets, lineup potentials, minibands, or oscillator strengths.

There are also specific tensions internal to the literature. In the WZ GaP case, the electronic-structure paper describes its explicit WZ bulk result as direct, while also discussing the literature expectation of pseudodirect behavior. In strained WZ GaAs, the text emphasizes that WZ binaries remain direct over the studied strain range, yet a tabulated case at $4.2 + 5.8\ \text{nm}$8 is labeled indirect. These are not merely editorial details; they indicate that the classification of low-energy states in folded, strained, and finite one-dimensional III–V systems can depend sensitively on how valley character is identified.

Taken together, the literature defines GaAs–GaP superlattice nanowires as a coupled heteroepitaxial and spectral-engineering platform. Experimentally, they support sharp interfaces, period-controlled phonon spectra, and a thermal-conductivity minimum as a function of period. Theoretically, they are best understood through the combined action of band folding, back-folded phonons, biaxial strain, surface relaxation, and finite cross section. What remains to be developed is a full heterostructure theory that includes explicit interface chemistry, band offsets, miniband formation, localization, optical matrix elements, shell effects, and beyond-LDA energetics.