Illinois reports quantum chip storing light for over a microsecond

The University of Illinois Urbana-Champaign described a chip-based quantum memory that stores telecom-band light for more than a microsecond in a university-authored report published on Phys.org on October 9. The experiment addresses a timing problem for quantum computers and communication networks: preserving information carried by light while other operations catch up.

The work, led by physics professor Elizabeth Goldschmidt and reported in Nano Letters, remains experimental. The university says a practical version needs better retrieval efficiency and longer storage times. Its potential for scalable manufacturing is a research ambition, rather than evidence of a commercially deployed memory.

The October account brings attention to results already disclosed earlier this year. Priyash Barya, Daren Chen and their colleagues submitted their original preprint on May 12 and revised it on May 15. October 9 dates the university report, not the first disclosure of the experiment.

Why quantum chips need to store light

Photons can transport quantum information, but a system's operations do not necessarily finish together. Temporarily storing that information provides a way to accommodate slower operations. The challenge is to preserve the information during the wait, rather than simply extend the time light spends travelling.

“Long delays are an outstanding problem in quantum information processing,” said Chen, a physics graduate student and co-first author, in the university report. The prospective applications are in chip-based quantum computing and communication systems, where a memory would form one component of a larger system.

What the Illinois memory demonstrated

The device uses a nanoscale waveguide made from thin-film lithium niobate containing erbium atoms. The university identifies strong light confinement, low optical loss and compatibility with wafer-scale integration as reasons for choosing the material. Those properties underpin the researchers' argument that the platform could support scalable fabrication.

A tunable laser prepares evenly spaced spectral features in the atoms, creating an atomic frequency comb. According to the university's explanation, this allows the atoms to absorb incoming light and release it after a predetermined interval. The preparation changes the ensemble's spectral response; it does not physically rearrange the atoms.

In their author-posted study abstract, the Illinois researchers report storing single-photon-level optical pulses at telecom wavelengths for more than one microsecond. They support the quantum-storage claim with measurements showing preserved phase coherence and noise below the single-photon level when the light is retrieved.

The study also reports storing up to 20 temporal modes and an acceptance bandwidth up to 2.2 GHz. These are separate reported capabilities: the abstract does not establish that every maximum was achieved in the same configuration. Nor does the temporal-mode count establish storage of 20 individually verified photons or entangled qubits.

A separate experiment shows why efficiency matters

Independent context comes from a separate preprint submitted on May 13 by Xiao-Jie Wang, Yong-Teng Wang, Zi-Wei Zhao, Yong-Min Li and Tian-Shu Yang. That team also investigated an erbium-doped thin-film lithium-niobate memory at telecom wavelengths. Its results are a separate experiment, not a replication or endorsement of the Illinois work.

The Wang and Yang team reports storage lasting 400 nanoseconds, efficiency of 1.95% and storage of four temporal modes. It also reports 96.8% fidelity for time-bin qubits encoded in single-photon-level coherent pulses, exceeding the classical measure-and-prepare limit identified in its paper.

The high fidelity and low efficiency illustrate different measures of a memory's performance: how faithfully information survives and how efficiently stored light can be recovered. Although the Illinois study reports a longer storage duration, the two abstracts do not establish an overall performance ranking under matched experimental conditions.

What remains before practical quantum memory

The Illinois team plans to improve preparation of its frequency comb and test different erbium isotopes that are less susceptible to noise. These steps accompany the university's stated need for longer storage and better retrieval efficiency. Goldschmidt's group is also exploring other quantum-photonic devices using spectral tailoring of atomic ensembles.

“This is not a one-off bespoke device,” Goldschmidt said in the university report, arguing that the platform could move a specialised laboratory technique towards commercial-scale fabrication. The report does not announce a product, customer deployment, manufacturing yield or timetable for practical availability.