---
title: Spin-Valley Resonance in Multivalley Systems
url: https://www.emergentmind.com/topics/spin-valley-resonance
type: topic
---

# Spin-Valley Resonance in Multivalley Systems

Spin-valley resonance denotes a class of resonant phenomena in multivalley systems where spin and valley cease to be independent labels and the relevant transitions are set by their hybridization. In the narrow silicon-quantum-dot sense, it appears when a Zeeman scale approaches a valley splitting and interface-enabled spin-orbit coupling mixes states such as $\ket{v_1,\uparrow}$ and $\ket{v_2,\downarrow}$, so that a microwave electric field can drive an otherwise weak spin transition through the excited valley state [1708.02903]. In a broader usage, the term also covers coherent oscillations between states differing in both spin and valley in Si/SiGe double dots [2111.14847], valley-dependent spin resonance in transition-metal dichalcogenide bilayers [1303.3932], and blockade-lifting or transport-resolved resonances among entangled spin-valley states in carbon nanotubes and other Dirac materials [1611.02203].

## 1. Conceptual scope and terminology

The term is not used uniformly across material platforms. In silicon quantum dots it usually refers to resonant phenomena near a spin-valley anticrossing, often summarized by the condition $E_Z \sim E_{\mathrm{VS}}$, where a ground-valley spin state and an excited-valley opposite-spin state become nearly degenerate and are mixed by interface spin-orbit coupling [1702.06210]. In that regime, resonance is a property of hybridized eigenstates rather than of pure spin or pure valley transitions.

In carbon nanotubes, the corresponding resonant degrees of freedom are often valley-spin doublets rather than isolated spins. The relevant states are already entangled by spin-orbit coupling, orbital magnetic response, and disorder-induced valley mixing, so blockade-lifting spectroscopy and electric-dipole driving probe a mixed spin-valley manifold from the outset [1408.0951]. In ambipolar nanotube double dots, the experimentally relevant resonances connect blocked and unblocked many-body states that differ in their combined spin-valley content, and the observable is the lifting of Pauli blockade rather than a conventional ESR line [1611.02203].

A further terminological broadening occurs in layered and Dirac materials. In AB-stacked TMDC bilayers, the central effect is a valley-dependent spin resonance in which spin dynamics are conditional on valley because spin, valley, and layer pseudospin are locked together [1303.3932]. In monolayer MoS$_2$ and graphene transport structures, “spin-valley resonance” can also mean spin-valley-resolved resonant tunneling or Fabry–Pérot-type transmission rather than ESR-like coherent precession [2605.20790]. This broader usage preserves the idea of resonance in a coupled spin-valley sector, but the resonant object is transport rather than a driven two-level transition.

A useful negative definition is also necessary. Direct measurements of valley composition and valley interference in silicon donors strongly constrain spin-valley models, but they do not by themselves demonstrate spin-valley resonance. In particular, the donor study of a subsurface arsenic impurity explicitly does not show a driven spin transition, a spin-valley anticrossing, a direct measurement of spin-orbit matrix elements, or relaxation hot-spot spectroscopy [1403.4648].

## 2. Microscopic ingredients and resonance conditions

Across platforms, spin-valley resonance requires three ingredients: a low-lying valley degree of freedom, a mechanism that couples spin and valley, and a drive or tuning knob that brings the coupled states into or near resonance. In silicon heterostructures the two lowest valleys are typically the $\pm z$-like valleys, split by interface coupling into states $v_1$ and $v_2$. The minimal four-state basis is then $\{\ket{v_1,\downarrow},\ket{v_1,\uparrow},\ket{v_2,\downarrow},\ket{v_2,\uparrow}\}$, with the key inter-valley spin-flip matrix element
\[
C_{v_1v_2}=\bra{v_2,\uparrow}H_{\rm SOC}\ket{v_1,\downarrow}.
\]
This produces an anticrossing between $\ket{v_1,\uparrow}$ and $\ket{v_2,\downarrow}$ at
\[
B_{\rm V}=\frac{\Delta}{g\mu_B},
\]
with an avoided-crossing gap $2|C_{v_1v_2}|$ [1708.02903].

In the perturbative regime below the anticrossing, the electrically driven Rabi frequency in a silicon corner dot takes the form
\[
hf = 2eg\mu_B B\,\delta V_g \frac{|C_{v_1v_2}|\,|D_{v_1v_2}|}{\Delta^2},
\]
where $D_{v_1v_2}$ is the inter-valley dipole matrix element. This formula makes explicit that weak intrinsic silicon SOC can nevertheless yield practical EDSR when a nearby valley excitation supplies the intermediate state [1708.02903].

At sharp Si/SiO$_2$ interfaces, the spin-valley coupling can be formulated microscopically in terms of inter-valley spin-flip tunneling,
\[
B(\mathbf{k})=\chi_D H_D(\mathbf{k})+\chi_R H_R(\mathbf{k}),
\]
which yields valley-dependent Rashba and Dresselhaus couplings and therefore opposite electric-field tunability of the $g$-factor for lower- and upper-valley qubits [1505.01213]. In that framework, inter-valley spin-flip tunneling dominates over intra-valley spin-flip processes, so the natural microscopic route to spin-valley resonance is intervalley rather than purely intravalley.

In Si/SiGe dots with micromagnets, the nearby regime is often detected first as valley-dependent ESR rather than as a resolved anticrossing. The two valleys $v_-$ and $v_+$ have different ESR frequencies,
\[
f_{v_-}\neq f_{v_+},
\]
and the difference is anisotropic, linear in $B_{\mathrm{ext}}$, and strongly sensitive to interface steps. The spin-valley-relevant part of the splitting is governed by differences in valley-dependent Rashba and Dresselhaus parameters, while the micro-magnet gradient contributes only when the two valleys already have different spatial dipole moments [1702.06210]. This places interface disorder at the center of the resonance problem: it does not merely broaden a pre-existing anticrossing, but can change its position, magnitude, and even sign-related anisotropies.

## 3. Silicon donors and quantum dots

Silicon donor systems supply the clearest example of how valley structure constrains spin-valley resonance even when resonance is not directly driven. For an arsenic donor $2.85\pm0.45$ nm below an interface, scanning-tunneling spectroscopy directly imaged the real-space and reciprocal-space valley interference of the donor ground state, extracted a $z$-valley population of approximately $38\pm2\%$ compared with the bulk value $33.3\%$, and concluded that the donor differs by only about $\sim 5\%$ from a bulk donor in valley population [1403.4648]. The same work states that for depths $>3$ nm the valley-perturbation-induced enhancement of $T_1^{-1}$ is negligible. For spin-valley resonance, this means that mere proximity to a clean interface at $\sim 3$ nm is not sufficient to generate strong ground-state valley admixture in donor qubits, even though the measured valley interference still renders donor-donor exchange highly sensitive to atomic placement.

Quantum dots exhibit the complementary regime in which the valley degree of freedom is electrically and magnetically accessible. In a Si/SiGe dot with integrated Co micro-magnets, the measured valley splitting is about $29~\mu\mathrm{eV}$ experimentally, with a best-fit theoretical value $34.4~\mu\mathrm{eV}$, and the two valley states show different ESR frequencies whose difference has a $180^\circ$ periodic anisotropy [1702.06210]. Those experiments were performed above the spin-valley anticrossing point, but they established the preconditions for resonance: valley-dependent Zeeman splittings, intrinsic interface-enabled SOI, and disorder-sensitive valley-orbit hybridization.

A direct realization of electrically driven spin-valley resonance was then obtained in a silicon-on-insulator corner dot. There, the valley splitting was extracted as $\Delta=36~\mu\text{eV}$, the anticrossing field was $B_V=0.314\pm0.001$ T, and the observed resonance lines included a main spin line, an intervalley spin line, and a zero-frequency signature of the anticrossing itself [1708.02903]. The mechanism was explicitly spin-valley-orbit: the valley state supplied the near-degenerate intermediate level, SOC mixed spin and valley, and an RF electric field coupled the valley/orbital sector.

Silicon can also exhibit coherent spin-valley dynamics without any ac magnetic or ac electric field. In an undoped Si/SiGe double quantum dot near the $(4,0)\leftrightarrow(3,1)$ transition, pulsing to a spin-valley anticrossing produced coherent oscillations between states differing in both spin and valley, with directly measured frequencies $f_1=8.2~\mathrm{MHz}$ and $f_2=15.2~\mathrm{MHz}$ for the two dots, and corresponding $S$–$T_-$ oscillation frequencies $f_1^{ST}=6~\mathrm{MHz}$ and $f_2^{ST}=11.3~\mathrm{MHz}$ [2111.14847]. In that experiment the coupling was always on; resonance was activated by detuning the system into the avoided crossing.

Recent planar Si-MOS spectroscopy near a valley anticrossing sharpened the same picture. In a four-state system with two valleys and two spins, one- and two-photon resonances were observed near the valley anticrossing, and the Rabi frequency increased from $5.46\pm0.01~\mathrm{MHz}$ far from the anticrossing to $34.0\pm0.3~\mathrm{MHz}$ near it, while $f_{\mathrm{Rabi}}/V_{MW}$ increased by more than a factor of 20 [2604.22547]. The enhancement was attributed to an electric-dipole transition activated by admixing of the upper level through inter-valley spin coupling, making explicit how spin-valley resonance can be converted into a fast electrical control channel.

## 4. Carbon nanotubes

In carbon nanotubes, spin-valley resonance is inseparable from the basic structure of the low-energy spectrum. Spin-orbit coupling splits each shell into Kramers doublets, disorder mixes $K$ and $K'$, and the orbital magnetic moment strongly couples the valley degree of freedom to the component of magnetic field along the nanotube. A convenient microscopic Hamiltonian is
\[
H(z) = -\frac{1}{2}\Delta_{\mathrm{SO}}\,\tau_3\,\hat{\mathbf n}(z)\!\cdot\!\boldsymbol{\sigma}
-\frac{1}{2}\Delta_{KK'}(z)\,[\tau_1\cos\varphi(z)+\tau_2\sin\varphi(z)]
+\frac{1}{2}g_s\mu_B\,\mathbf B\!\cdot\!\boldsymbol{\sigma}
+\frac{1}{2}g_{\mathrm{orb}}\mu_B\,\mathbf B\!\cdot\!\hat{\mathbf n}(z)\,\tau_3,
\]
so the driven object is already a valley-spin Kramers qubit rather than a bare spin [1408.0951].

The transport prerequisite for observing these resonances is valley-spin blockade. In low-disorder nanotube double dots, blockade was demonstrated using both valley and spin selection rules rather than spin selection alone, and single-electron spin resonance was detected through the lifting of that blockade [1210.2622]. Because the low-energy eigenstates are valley-spin states, the observed resonance is most naturally interpreted as resonance of a coupled spin-valley pseudospin. The use of the nanotube bandgap to obtain a large level spacing was crucial: it kept the relevant triplet-like $(2,0)$ states outside the transport window and thereby restored robust blockade.

A fully microscopic theory of electrically driven resonance in bent disordered nanotubes showed how gate-induced motion through a spatially varying spin-valley Hamiltonian generates an oscillating effective field in the Kramers-qubit subspace [1408.0951]. Two mechanisms were identified. The first is bend-mediated: motion changes the local nanotube axis and therefore the projection of the orbital magnetic moment. The second is disorder-mediated: motion changes the local valley-mixing amplitude $\Delta_{KK'}(z)$ and phase $\varphi(z)$. The resulting Rabi frequency is determined by the component of the effective drive field transverse to the static effective field, so the angular dependence is highly anisotropic and sensitive to whether the inhomogeneity is geometric or disorder-related.

In ambipolar nanotube double dots, the resonant dynamics become a genuinely multilevel spin-valley problem. For the $(1e,1h)$ configuration, the blocked states $B1,B2$ and nonblocked states $N1,N2$ differ in combined spin-valley content, and periodic electric driving can lift blockade through transitions such as $B2\rightarrow N1$ that correspond predominantly to simultaneous valley and spin inversion [1611.02203]. A notable result is that the dynamics are strongly renormalized by coupling to the higher-lying “vacuum” charge state $(0e,0h)$, which shifts transition energies, modifies effective $g$ factors, and produces multiphoton structure.

A conceptually earlier but still relevant precursor is disorder-mediated electron valley resonance. There the focus was the valley sector alone: atomic disorder generated intervalley mixing, ac electric motion made that mixing time dependent, and the resulting two-level problem supported valley Rabi oscillations with a period on the nanosecond timescale [1010.4338]. That mechanism anticipated a central CNT lesson of later spin-valley work: inhomogeneity that would normally be treated as an imperfection can become the transduction element that makes electrically driven valley or spin-valley resonance possible.

## 5. TMDCs, acoustically driven resonance, and transport-resolved variants

In TMDC bilayers, the relevant coupling is among spin, valley, and layer pseudospin. Near the $K$ valleys, the low-energy hole Hamiltonian
\[
H_v = - \lambda \tau_z \sigma_z s_z + t_\perp \sigma_x + B_x s_x + B_z s_z + E_z \sigma_z
\]
implies a projected doublet Hamiltonian
\[
H_D =
\begin{pmatrix}
-\tau_z E_z \cos 2\alpha + B_z & B_x \sin 2\alpha \\
B_x \sin 2\alpha & \tau_z E_z \cos 2\alpha - B_z
\end{pmatrix},
\]
so the spin splitting is valley dependent and can be driven either magnetically or electrically [1303.3932]. The key result is a valley-dependent spin resonance rather than a spin-to-valley conversion: the resonance frequency differs in $K$ and $-K$, and under suitable static $B_x$ both $B_z^{\rm ac}$ and $E_z^{\rm ac}$ drive the same transition with amplitude proportional to $|B_z^{\rm ac}-\tau_z E_z^{\rm ac}\cos 2\alpha|$. This directly enables valley-controlled spin gates.

Monolayer MoS$_2$ supports a different variant: valley spin-acoustic resonance. Because broken inversion symmetry and SOC produce spin-split conduction subbands that cross at finite momentum, a Rayleigh surface acoustic wave can generate an oscillating strain-induced effective magnetic field that drives spin-flip transitions between the spin-split subbands near the crossing [2206.11551]. At zero magnetic field the resonance is valley degenerate; an external out-of-plane magnetic field replaces $2|\lambda_c|$ by $2(|\lambda_c|\pm\Delta_B)$ and splits the resonance into two valley-selective branches in both SAW absorption and acoustoelectric current. Here the resonance is still internal-state spectroscopy, but the drive is acoustic rather than electromagnetic.

A broader, transport-based usage appears in driven and barrier-defined 2D materials. In monolayer MoS$_2$ under off-resonant elliptically polarized light, a Floquet-engineered valley-dependent mass term,
\[
\Gamma_\tau=\frac{\Delta}{2}-\tau \frac{(ev_FA_0)^2}{\hbar\omega}(\xi^2-1),
\]
shifts the propagation thresholds and Fabry–Pérot phases of different spin-valley channels, allowing the same junction to act either as a broadband valley filter or as a resonance-selective spin-valley filter [2605.20790]. In graphene, a gated ferromagnetic/pristine/strained structure similarly produces flavor-dependent resonant tunneling and giant Goos–Hänchen shifts; the resonance is orbital and interference-based rather than ESR-like, but it is resolved by spin and valley and therefore belongs to the transport branch of the spin-valley resonance literature [1611.04722].

## 6. Experimental signatures, control implications, and limitations

Several signatures recur across the field. One is the appearance of valley-resolved ESR lines or valley-dependent spin splittings. Another is the avoided crossing or “hot spot” that occurs when Zeeman and valley scales align. In SiMOS and Si/SiGe quantum dots, the angular dependence of the intervalley coupling is strong enough that the resonance landscape depends sensitively on the magnetic-field direction, and SiMOS was found to exhibit an order of magnitude larger spin-valley coupling than Si/SiGe while preserving a similar angular pattern of maxima and minima [2604.16713]. This makes magnetic-field orientation a genuine control knob rather than a secondary detail.

A second recurring signature is enhanced electrical controllability near the anticrossing. In planar Si-MOS, valley admixture converted a nominally magnetic ESR transition into a strongly electric-dipole-active one, producing a large Rabi enhancement near the valley anticrossing [2604.22547]. In silicon corner dots, the same basic mechanism yields all-electric EDSR without a micromagnet, with the valley state providing the intermediate level that amplifies silicon’s otherwise weak SOC [1708.02903].

The cost of this enhanced controllability is usually stronger sensitivity to electrical noise. In a Si/SiGe dot driven by microwave electric fields, inter-valley spin transitions were reported to have a tenfold increase in sensitivity to electric fields and electrical noise compared with pure spin transitions, and the corresponding phase coherence was strongly reduced whenever the valley-orbit index changed [1608.06538]. A related but distinct mechanism appears even when the resonance is nominally spin-only: in a Si/SiGe dot with a micromagnet, a valley-dependent $g$-factor and valley-dependent Rabi frequency, combined with valley relaxation, generate a spin-dephasing channel that is not removed by spin echo [1603.02829].

Interface structure remains the dominant microscopic variable. In Si/SiGe dots with micro-magnets, monoatomic steps can change both the magnitude and sign of the valley-dependent spin splitting difference, and the gradient-field contribution becomes substantial only when the interface has already made the two valley wavefunctions inequivalent [1702.06210]. In donor-based silicon, by contrast, a donor about one effective Bohr radius below a clean interface remains almost bulk-like in valley composition, so interface-induced spin-valley admixture is weak for single-donor relaxation, while the bulk-like valley phase still keeps exchange coupling highly oscillatory and atomically placement-sensitive [1403.4648].

The field therefore presents a durable tradeoff. Spin-valley resonance can be an unwanted leakage and relaxation channel, a spectroscopic probe of interface physics, or a deliberate resource for all-electrical control. Which of these roles dominates depends on the valley splitting, the strength and anisotropy of intervalley SOC, the degree of disorder or interface roughness, and whether the relevant experiment measures coherent precession, blockade lifting, or resonant transport.

Source: https://www.emergentmind.com/topics/spin-valley-resonance