Quantum frequency conversion to telecom wavelengths in two experiments

Short answer: Quantum frequency conversion shifts the wavelength of a single photon without destroying its quantum properties, so that light from an atom or a quantum dot can travel through telecom fibre. Two papers, from 2012 and 2025, report conversions to 1313 nm and 1552 nm with efficiencies of over 30 percent and about 25 percent respectively. These are laboratory demonstrations.

Why convert single-photon wavelengths?

The 2025 Physical Review Applied abstract describes quantum frequency conversion as a widely used technique to interface atomic systems with the telecom band so that photons can propagate over longer distances in fibre. Many quantum emitters give off visible or near-infrared light, which is lost faster in fibre than telecom wavelengths. Conversion is one way the authors propose to bridge that gap.

What did the 2012 quantum dot experiment show?

The 2012 Physical Review Letters paper demonstrates efficient quantum frequency conversion of visible single photons at 711 nm, emitted by a quantum dot, to the telecom wavelength 1313 nm, with an efficiency above 30 percent. The authors analyse the first- and second-order coherence before and after conversion and say this clearly shows that important properties, such as the coherence time and photon antibunching, are fully conserved during the frequency translation. Antibunching is the signature that the source emits one photon at a time. They say their findings underline the potential of on-demand single photon sources combined with conversion for new quantum technology applications, which is their outlook. A companion paper in the same issue, listed as a related article on the journal page, studied two-photon interference using conversion of photons from a different quantum dot.

What did the 2025 experiment show?

The authors demonstrate difference-frequency conversion from 606 nm to 1552 nm of microsecond-long weak coherent pulses at the single-photon level. The conversion is compatible with quantum memories. They use a single-step difference-frequency-generation process, with a continuous-wave pump at 994 nm in a periodically poled lithium niobate waveguide, together with ultranarrow spectral filtering down to a bandwidth of 12.5 MHz. They convert pulses of up to 13.6 microseconds with a device efficiency of about 25 percent and a signal-to-noise ratio above 460 for pulses containing one photon on average. The authors say this signal-to-noise ratio is large enough for high-fidelity conversion of qubits emitted from an emissive quantum memory, and for an interface with quantum processing nodes based on narrow-linewidth, cavity-enhanced trapped ions. Part of the memory material's name did not render in the page text, so it is not named here. The statements about fidelity and interfacing are the authors' projections.

What about the third source?

A 2021 Applied Physics Letters paper on a bright source of telecom single photons based on quantum frequency conversion was on the source list. Its abstract could not be read, so nothing in this post is taken from it beyond its title.

How to read this

The two experiments use different emitters, wavelengths and pulse types. The 2025 work used weak coherent pulses and not single photons from an emitter, so its signal-to-noise figure is not a measurement of single-photon quality. All numbers are the authors' own. This post is based on abstracts and visible page text, and full texts were not checked. It makes no claim about networks in use.

FAQ

Does conversion keep the photon's quantum properties?

In the 2012 experiment, the authors report that coherence time and antibunching were fully conserved.

What efficiency was reported?

Above 30 percent in 2012, and about 25 percent in 2025.

Why use a periodically poled waveguide?

It hosts the difference-frequency process in the 2025 setup, per the abstract.

Was this done with true single photons in 2025?

The abstract describes weak coherent pulses at the single-photon level.