Single-shot latched readout of a quantum dot qubit using barrier gate pulsing

Latching techniques are widely used to enhance readout of qubits. These methods require precise tuning of multiple tunnel rates, which can be challenging to achieve under realistic experimental conditions, such as when a qubit is coupled to a single reservoir. Here, we present a method for single-shot measurement of a quantum dot qubit with a single reservoir using a latched-readout scheme. Our approach involves pulsing a barrier gate to dynamically control qubit-to-reservoir tunnel rates, a method that is readily applicable to the latched readout of various spin-based qubits. We use this method to enable qubit state latching and to reduce the qubit reset time in measurements of coherent Larmor oscillations of a Si/SiGe quantum dot hybrid qubit.

💡 Research Summary

The authors address a longstanding challenge in single‑shot readout of semiconductor spin qubits that rely on spin‑to‑charge conversion followed by a “latched” charge state. Conventional latched readout requires two tunnel rates to be simultaneously tuned: a fast left‑dot‑to‑reservoir rate (ΓL ≫ T1⁻¹) to load the metastable charge state quickly, and a slow right‑dot‑to‑reservoir rate (ΓR ≪ measurement bandwidth ΔfBW) so that the metastable state persists long enough for detection. When only a single electron reservoir is available, as is desirable for scaling, these conditions become mutually exclusive because the inter‑dot tunnel couplings that are needed for fast qubit manipulation also increase ΓR via cotunneling, violating the second condition.

To solve this, the paper introduces a dynamic barrier‑gate pulsing scheme. The device is a Si/SiGe double quantum dot (DQD) forming a five‑electron hybrid qubit (QDHQ). Three barrier gates (B1, B2, B3) control the tunnel couplings; B3 is fully pinched off so that only the left reservoir (connected to dot P1) remains. A high‑frequency line is attached to the left barrier gate B1, allowing rapid modulation of the left‑dot‑to‑reservoir tunnel rate ΓL. The experimental sequence consists of six stages:

1. Initialization – B1 is set to a high voltage, opening ΓL so that the qubit relaxes into the (4,1) ground charge configuration.
2. Larmor pulse – A short (≈ 400 ps) π/2 pulse on the plunger gate P1 moves the system into the (3,2) charge region, creating a coherent superposition of logical |0⟩ and |1⟩.
3. Return to readout point – The system is pulsed back toward negative detuning; if the logical state is |1⟩ the electron occupies (3,2)g.
4. Latching – While still holding B1 high, the (3,2)g state rapidly tunnels to the metastable (4,2) charge configuration because ΓL ≫ ΓR. This “latch” stores the qubit information as a charge.
5. Readout – B1 is quickly lowered, suppressing ΓL and leaving ΓR dominated by cotunneling, which is much slower than the 1 kHz measurement bandwidth. The charge sensor (a nearby quantum dot under gate CS) records a current I_SD that distinguishes the (4,1) and (4,2) charge states.
6. Reset – After the measurement, B1 is pulsed high again to accelerate the (4,2)→(4,1) transition, thereby resetting the qubit for the next cycle.

Because the barrier‑gate pulse is applied only during the latching and reset phases, the two conflicting tunnel‑rate requirements are satisfied sequentially rather than simultaneously. The authors carefully compensate crosstalk from B1 onto P1 by applying a calibrated opposite‑sign pulse on P1, ensuring that the qubit’s detuning point remains fixed during the barrier‑gate operation.

Single‑shot traces show two well‑separated current levels corresponding to logical |0⟩ and |1⟩. A histogram of the minimum I_SD values during readout fits a sum of two Gaussians with a separation ΔI_SD that yields a signal‑to‑noise ratio (SNR) of 10.2. This translates to a charge sensitivity of 3 × 10⁻³ e / √Hz, comparable to state‑of‑the‑art charge‑sensor performance. Importantly, the reset pulse reduces the average time the qubit spends in the latched state from ~2 ms (without reset) to ~0.13 ms, a 15‑fold speed‑up, and restores the initialization probability to 98 % after 2 ms.

Using this readout, the authors demonstrate coherent Larmor oscillations of the QDHQ. By varying the Larmor‑pulse amplitude (V_Larmor) and wait time (t_Larmor), they observe both charge‑like oscillations (≈ 100 mV) and spin‑like oscillations (≈ 200 mV). Fast Fourier transform (FFT) analysis yields an inter‑dot tunnel coupling Δ₁ ≈ 750 MHz (minimum qubit splitting 2Δ₁/h ≈ 1.5 GHz) and a singlet‑triplet splitting E_ST/h ≈ 4.0 GHz, consistent with independent magneto‑spectroscopy measurements.

The discussion quantifies the impact of the reset scheme on experimental duty cycles. For a 1 ms measurement window, the required out‑tunneling time for the |1⟩ state to remain occupied with 99 % probability is ≈ 99.5 µs, which would imply an initialization time of ≈ 458 µs if the latch were not actively reset. With the barrier‑gate reset, this initialization time is reduced to ≈ 30 µs, dramatically improving the feasible repetition rate. The authors note that even with faster RF‑based readout (e.g., reflectometry or RF‑SET) that could shrink the measurement window to 1 µs, the reset advantage remains critical.

Finally, the paper emphasizes the broad applicability of the technique. Any double‑dot qubit that employs spin‑to‑charge conversion—such as singlet‑triplet, exchange‑only, or parity‑readout single‑spin qubits—can adopt the barrier‑gate pulsing method to achieve high‑fidelity single‑shot readout with a single reservoir. The approach simplifies device layout, reduces the number of required gates, and is compatible with scaling strategies that aim to minimize wiring and cross‑talk in large‑scale quantum processors.