When Your Quantum Spin Memory Lives on a Chip
26 Apr 2026, Yanjiang
There’s something quietly revolutionary about the idea of reading a quantum spin state with nothing more than an electrical signal. No bulky optics. No painstaking alignment of lasers and lenses. Just a current, a voltage, and the faint whisper of an electron’s magnetic personality imprinted on a thin film of silicon carbide.
A team led by Robert Cernansky at Ulm University — working with Alexander Zappacosta, Ben Haylock, Paul Fisher, and Naoya Morioka — has demonstrated exactly this: electrical readout and coherent control of silicon vacancy spin ensembles in a silicon carbide-on-insulator (SiCOI) platform. Their work appears in a preprint (arXiv:2511.22485) that extends the capabilities of this promising material system toward scalable quantum technologies.
What makes this result noteworthy is not just that it works — it’s that it works across an unusually broad range of excitation wavelengths, from 780 to 990 nanometers. This is the first demonstration of spin state readout well beyond the zero phonon line of the V2 silicon vacancy, a finding that opens practical doors for integrating quantum spin defects with existing photonic infrastructure.
The Platform and the Problem
Silicon carbide (SiC) has long been a material of interest for quantum technologies. It’s commercially available, compatible with CMOS fabrication processes, and hosts a variety of color centers — atomic-scale defects whose spin states can be initialized, manipulated, and read out. Among these, the negatively charged silicon vacancy (V⁻_Si) has emerged as a particularly attractive candidate, with spin properties that persist at room temperature and optical transitions in the near-infrared.
But there’s a catch. Most readout schemes for these defects rely on optical detection: you shine a laser on the sample, collect the emitted fluorescence, and infer the spin state from the intensity. This works beautifully in the lab, but it’s difficult to scale. Every defect needs its own collection optics. Every measurement requires careful alignment. It’s not a recipe for the kind of densely integrated quantum devices that many envision.
The alternative is photoelectrical detection of magnetic resonance (PDMR). Instead of collecting photons, you measure the photocurrent generated when the defect absorbs light. The spin state affects this current, giving you an electrical readout that requires no collection optics at all. It’s a technique that has been demonstrated in bulk SiC, but adapting it to thin-film platforms — the kind needed for integrated photonics — has been an open challenge.
What the Team Found
Cernansky and colleagues fabricated thin-film SiCOI devices and characterized the spin properties of silicon vacancy ensembles containing approximately 540 defects. They demonstrated coherent control of the spin states using microwave pulses and read out the results electrically via PDMR across a wide range of excitation wavelengths.
The key numbers are these: the measured T₂ coherence time — a measure of how long the spin information survives — was approximately 7 microseconds, comparable to what is observed in bulk SiC. This is significant because it suggests that the thin-film processing required to create the SiCOI platform does not degrade the spin coherence properties that make silicon vacancies attractive in the first place.
The team also provided a direct comparison between optical and electrical readout in both bulk and thin-film platforms, confirming that PDMR works reliably in the SiCOI architecture. This is not merely a proof of principle; it’s a practical demonstration that a scalable readout technique can be married to a scalable material platform.
Why This Matters
The combination of electrical spin readout with a thin-film, CMOS-compatible platform addresses a bottleneck in quantum technology development. Many of the most promising spin qubit systems — nitrogen-vacancy centers in diamond, silicon vacancies in SiC — are studied in bulk crystals or nanoscale samples that are difficult to integrate into larger systems. Moving to thin-film platforms that can be fabricated using standard semiconductor processes is a necessary step toward devices that combine quantum functionality with classical control electronics on the same chip.
The broad wavelength range is equally important. Not every application has the luxury of choosing its excitation wavelength. A quantum sensor operating in a biological environment, for example, might need to work at wavelengths that minimize tissue damage or autofluorescence. A quantum repeater node might need to interface with a specific telecom-band laser. Demonstrating spin readout from 780 to 990 nm suggests that SiCOI-based devices can be flexible in their operational requirements.
The Road Ahead
This work extends the capabilities of SiCOI toward electronic and spin-based devices for scalable quantum technologies. The next steps will likely involve moving from ensembles of approximately 540 defects to single defects — a nontrivial challenge, since the electrical readout signal from a single spin is orders of magnitude smaller. The team’s results suggest that the basic physics works; engineering the sensitivity to reach the single-defect level is the next frontier.
There is also the question of integration with photonic components. SiCOI is already being explored as a platform for waveguides, resonators, and other integrated photonic structures. Demonstrating spin readout in this platform opens the possibility of devices that combine optical manipulation, electrical readout, and photonic interconnects — all on a single chip.
For now, the result stands as a practical step forward: a demonstration that two promising technologies — electrical spin readout and thin-film SiC — work together as hoped. The path from here to a fully integrated quantum device is long, but the foundation is now a little more solid.
Yanjiang is an online editor of Loom Science
References
- Alexander Zappacosta et al., Wavelength-Dependent Electrical Readout of Spin Ensembles in a Thin-Film SiC-on-Insulator Platform, arXiv:2511.22485