Spin Mapping in Donor-Bound GaAs

• Spin mapping is an ultrafast technique that converts a photon's polarization from the Poincaré sphere into a corresponding electron spin state on the Bloch sphere in GaAs.
• It employs a stimulated Raman process using a picosecond laser pulse and polarization-resolved transitions in a Λ-system to achieve precise spin-state preparation and readout.
• Experiments reveal that repeated pulse sequences allow additive spin coherence, offering potential for multi-pulse encoding in quantum-information protocols.

Searching arXiv for the cited paper and closely related work on donor-bound electron spin preparation/readout in GaAs. Spin mapping, in semiconductor spin optics, is the ultrafast preparation and readout of localized electron-spin coherence by direct transfer of an optical polarization state onto a spin state. In donor-bound electrons in GaAs, a single picosecond laser pulse with prescribed polarization prepares a corresponding coherent superposition of $|\uparrow\rangle$ and $|\downarrow\rangle$, establishing a bijective correspondence between the Poincaré-sphere representation of photon polarization and the Bloch-sphere representation of electron spin. In the same system, ultrafast optical Kerr detection accesses the resulting coherence and enables spin-state tomography, so the optical interface acts both as a preparation map and as a measurement basis selector (Denega et al., 2011).

1. Physical platform and qubit structure

The demonstrated platform is an epitaxial GaAs film, $10\,\mu\text{m}$ thick, grown along $[001]$ and doped with Si donors at low concentration $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$. At $4.2\,\text{K}$, each donor binds one electron, forming a donor-bound state $D^0$. The qubit is the spin-$\tfrac{1}{2}$ ground-state doublet

$|\uparrow\rangle,\quad |\downarrow\rangle,$

split by a Zeeman energy $E_Z$ in a magnetic field $|\downarrow\rangle$0 applied along $|\downarrow\rangle$1, with $|\downarrow\rangle$2.

Optical access is provided by the donor-bound trion manifold $|\downarrow\rangle$3, whose lowest relevant excited state is denoted $|\downarrow\rangle$4. In Voigt geometry, the selection rules are polarization-resolved: the transition $|\downarrow\rangle$5 couples only to $|\downarrow\rangle$6-polarized light, while $|\downarrow\rangle$7 couples only to $|\downarrow\rangle$8-polarized light. This realizes an effective three-level $|\downarrow\rangle$9-system in which orthogonal optical polarizations address the two spin basis states independently.

The donor-bound electrons are attractive as solid-state qubits because they combine optical addressability with long spin lifetimes. The longitudinal relaxation time is $10\,\mu\text{m}$0; in ensembles the dephasing time is $10\,\mu\text{m}$1; and previous work cited in the study reported $10\,\mu\text{m}$2. A plausible implication is that the platform occupies an intermediate regime between atomic-like optical selection rules and semiconductor-compatible integration.

2. Polarization-to-spin correspondence

The central mapping is expressed between a general preparation-pulse polarization state and the prepared electron-spin state. In the $10\,\mu\text{m}$3 basis, the optical polarization is written as

$10\,\mu\text{m}$4

with $10\,\mu\text{m}$5, $10\,\mu\text{m}$6. The experimentally established spin state is

$10\,\mu\text{m}$7

or equivalently

$10\,\mu\text{m}$8

This map swaps the amplitudes $10\,\mu\text{m}$9 and $[001]$0, conjugates the relative phase, and introduces a minus sign. Its significance is geometric as well as algebraic: every polarization state reachable on the Poincaré sphere corresponds to a unique spin state on the Bloch sphere, and probe polarization can be chosen to analyze arbitrary spin orientations. Equatorial optical states map to equatorial spin superpositions with phase control, while the poles map to the spin basis states up to phase.

Using the usual spherical parameterization

$[001]$1

the mapped spin state can be rewritten in Bloch-sphere form with

$[001]$2

up to a global phase. This shows that the mapping is a fixed inversion between the two spheres rather than a trivial identity.

3. Stimulated Raman preparation and ultrafast timescales

The preparation mechanism is a stimulated Raman process in the $[001]$3-system. A single short pulse with spectral width $[001]$4 exceeds the Zeeman splitting $[001]$5, so it simultaneously drives both optical legs of the $[001]$6-system. In the coherent-population-trapping picture, the pulse polarization defines effective Rabi frequencies $[001]$7 and $[001]$8, and the dark-state structure

$[001]$9

directly yields the observed mapping relation (Denega et al., 2011).

The preparation is ultrafast relative to dissipative processes. The pulse duration is on the picosecond scale, whereas spontaneous emission from $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$0 to $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$1 is $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$2, and ensemble dephasing is $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$3. The preparation therefore occurs orders of magnitude faster than both spontaneous emission and initial dephasing. The study further states that the ratio of coherence time to operation time exceeds $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$4.

The power and detuning dependences indicate that the mapping is robust but not structureless. The Kerr amplitude scales approximately linearly with pump power up to at least $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$5, with $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$6. Increasing power somewhat reduces $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$7. A blue detuning of $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$8 from resonance produces a spin-precession phase shift of about $n_{\rm Si} \approx 3\times 10^{13}\,\text{cm}^{-3}$9, increases preparation efficiency, and reduces $4.2\,\text{K}$0. The involvement of higher trion levels $4.2\,\text{K}$1 enhances Raman efficiency but modifies effective selection rules and therefore the mapping phase. The method is reported to work best and most robustly for pulses resonant with, or slightly red-detuned from, the lowest trion state $4.2\,\text{K}$2.

4. Optical Kerr detection and spin-state tomography

Readout is performed through time-resolved optical Kerr rotation. A linearly polarized probe pulse is reflected from the sample after delay $4.2\,\text{K}$3 relative to the preparation pulse. The spin polarization of the $4.2\,\text{K}$4 ensemble modifies the complex refractive index differently for orthogonal probe polarizations, producing a polarization rotation or ellipticity change measured with a polarization bridge.

In this discrete-level system, the probe polarization selects the effective measurement axis on the Bloch sphere. The study reports that probe polarizations $4.2\,\text{K}$5 effectively measure spin orientation along $4.2\,\text{K}$6, respectively. By choosing probe polarizations among $4.2\,\text{K}$7, different spin projections can be accessed, which is why the readout constitutes spin-state tomography (Denega et al., 2011).

The Kerr signal is fit as

$4.2\,\text{K}$8

where $4.2\,\text{K}$9 is the transverse-spin amplitude, $D^0$0 the Larmor frequency, and $D^0$1 the precession phase. The measured electron $D^0$2-factor is

$D^0$3

A linear dependence of $D^0$4 on the preparation-pulse phase $D^0$5, together with probe-dependent offsets, confirms the mapping in Eq. (1). The observed Kerr signals beyond about $D^0$6 are attributed purely to $D^0$7 spin coherence rather than to electron-hole recombination.

5. Repeated preparation and additive behavior

A further test of the mapping uses two identical preparation pulses separated by a controlled delay. The first pulse prepares a spin coherence that precesses at the Larmor frequency. The second repeated-preparation pulse arrives after partial free evolution and prepares a new spin state whose relative phase can be chosen with respect to the precessing one.

When the second pulse is in phase with the pre-existing coherence, the Kerr amplitude is enhanced. When it is in counter-phase, the amplitude is suppressed. The resulting state is not described as a pure reset or a pure unitary manipulation; instead, it behaves as the vector sum of the Bloch vector generated by the first pulse and the Bloch vector prepared by the second pulse. Because partial dephasing has already occurred before the second pulse arrives, destructive interference is incomplete.

This additive character implies that multiple pulses can accumulate or subtract spin coherence in an ensemble. This suggests, without proving, that the preparation method can support multi-pulse encoding strategies in which successive optical inputs are stored as algebraic combinations of coherent spin excitations.

6. Quantum-information significance and limitations

The most direct significance of spin mapping is qubit initialization. A single picosecond optical pulse prepares an arbitrary pure spin state by choosing its polarization on the Poincaré sphere. Combined with polarization-selective Kerr readout, the same interface supports state preparation, basis change, and tomography within one optical control framework (Denega et al., 2011).

The method is also notable for its timescale hierarchy. Operation occurs on picosecond scales, whereas $D^0$8 is nanoseconds and previously reported $D^0$9 exceeds $\tfrac{1}{2}$0. This suggests substantial room for complex control sequences before coherence is lost. Because the mapping converts a photonic polarization qubit

$\tfrac{1}{2}$1

into a spin qubit

$\tfrac{1}{2}$2

it functions as a spin-photon interface relevant to quantum memory and hybrid-network architectures.

Several limitations are also explicit. The microscopic mechanism of the Kerr readout is described as less well understood than the preparation step, and a full theoretical framework for the detection process is not provided. Blue detuning and coupling to higher trion levels alter the mapping phase and increase dephasing. Preparation can induce some incoherent population of excited states, so the full three-level state is not purely coherent even when the spin component is. Finally, the experiments probe ensemble dephasing $\tfrac{1}{2}$3, not the much longer single-donor coherence times inferred from earlier work.

In this sense, spin mapping is both a geometric correspondence and an operational protocol: it is the experimentally validated transfer of optical polarization into localized-electron spin coherence, implemented through stimulated Raman dynamics and read out by ultrafast Kerr tomography in donor-bound GaAs.