Sparrow Quantum Reports Record Single-Photon Flux

Many quantum-light experiments have historically relied on chance: you excite a material, wait, and occasionally a photon appears. However, Sparrow Quantum and Ruhr University Bochum have developed a source that delivers photons on demand, exceeding 500 million usable photons per second into optical fiber. This achievement, which the team claims is the highest single-photon flux measured to date, is detailed in a preprint on arXiv.

Let’s explore why photon supply is crucial for a specific approach to building quantum computers.

The Lottery Problem and Its Escalation with Scale

Photonic quantum systems encode information using light. The advantage is clear: photons can travel through optical fiber, which means the same hardware for a processor can also connect a quantum network without conversion.

The persistent challenge has been the photon source itself. Ideal sources produce photons that are identical clones, arriving precisely when needed, rather than randomly. For much of this field’s history, sources were probabilistic: excite a material, hope for a photon-generation event, and discard failures.

Though a probabilistic single-photon source is merely inconvenient, scaling this to ten photons becomes a significant hurdle.

Why?

In multi-photon experiments, all photons must arrive simultaneously. Losses are multiplicative, not additive. A source suitable for a single photon might be useless for ten, as the probability of ten independent events all succeeding at once drops sharply. Consequently, the source fundamentally limits the entire system built upon it. Optimizing the source at the point of generation provides benefits throughout the entire setup.

Bright Enough for a Standard Power Meter

One remarkable aspect is the sheer optical power: this stream carries over 100 picowatts. This may sound modest, but quantum light is typically so faint that individual photons must be counted using detectors that can cost more than a car. This new source is bright enough to be measured by an ordinary optical power meter.

The source operates at 1 gigahertz with over 50% fiber efficiency, maintaining high single-photon purity and two-photon indistinguishability throughout. These last two qualities refer to how pure each photon is and how identical any two of them are, respectively.

What’s particularly commendable is their reporting methodology. Source figures are often quoted after spectral filtering, which removes undesirable light that would skew average performance. Sparrow, however, did not filter. They measured the source’s full emission and reported the lower bound instead of the best-case scenario. 

By capturing the entire output, the power measured at the fiber directly reflects the source’s efficiency, eliminating the need for detector calibrations or “fudge factors” that make published numbers difficult to compare. This transparency could allow the source to also serve as a metrology tool, a photon-flux standard for calibrating the very detectors others rely on.

They also pushed the source to its limits. At one gigahertz, there’s no room to fit another pulse between existing ones. Such intense operation typically degrades photon quality. Yet, in this case, it didn’t, and crucially, this applies to the unfiltered emission.

Experiments Awaiting This Breakthrough

Now for the impact. If this stream is split across ten channels using time-space demultiplexing, each channel still receives tens of millions of photons per second. This rate provides enough coincidence for ten-photon experiments to become practical rather than merely aspirational, and sufficient for interference involving 10 to 20 photons, surpassing the capabilities of conventional commercial sources.

Many theoretical protocols have been awaiting precisely this kind of development. Linear optical quantum computing, quantum-enhanced machine learning, quantum key distribution, and quantum networking (all designed and analyzed on paper), have rarely been implemented because generating so many indistinguishable photons simultaneously was impractical outside of a few specialized labs. Juan C. Loredo, Sparrow’s VP of Innovation, who used earlier versions of these sources at the University of Vienna, plainly expressed the past frustrations:

“As an end-user, I always knew what the source was stopping me from doing. (…) You design the protocol you want, then you cut it down to what the photon rate will support, and then you wait days for enough data. This is the first time I have looked at a photon flux and thought the limit is somewhere else now.”

What This Is, and What It Is Not

It’s important to clarify the scope, as a bright source alone does not constitute a quantum computer.

This preprint details an engineering achievement in single-photon source development. Peter Lodahl, Sparrow’s founder and Chief Quantum Officer, emphasized that it represents serious engineering, not a redefinition of the technology. It demonstrates that the emitter can be driven as hard as physics allows while maintaining photon quality. It is not a demonstration of a 20-photon computation, nor is it a claim of quantum advantage over classical machines.

The missing ingredient is entanglement. Though more photons are a start, larger photonic systems require these photons to be linked into a shared quantum state. Generating and controlling entanglement at scale is the next significant challenge. As Lodahl directly stated, making individual photons behave took years; making them work together with the same reliability is the next frontier.

So, what is confirmed? A source that delivers photons on demand, bright enough to measure with off-the-shelf equipment, and honest in its reported numbers. What remains aspirational is everything that could eventually be built once these photons can reliably interact.