The world appears to be on the cusp of a quantum revolution. Quantum computing, quantum communication, and the quantum internet all promise to enrich society. Quantum computing is touted to deliver faster drug discovery, better weather prediction, new sustainable materials, and a host of other benefits. Quantum communication should offer higher levels of security, speed, and bandwidth in getting information from point A to point B. And the combination of these technologies is expected to lead to a super-secure, super-capable quantum version of the internet.
Listen to the hyperbole and you might even believe all of this is just around the corner, that you will be sitting in front of your quantum laptop in just a few years’ time wondering what the fuss was all about. Well, you won’t: For all but the most challenging tasks, regular classical computers will remain perfectly adequate.
But at the same time, some of these quantum technologies are more mature than many people realize; the quantum internet is much more than a pipedream. The key element underpinning all but the most basic version of a quantum internet will be a reliable and deployable source of single photons. More specifically, a source that generates bright, indistinguishable photons one at a time, on demand, while at the same time exhibiting properties such as low cost, simplicity, compactness, and easy fabrication. Despite intense research since the first source was engineered around 60 years ago, however, this perfect single-photon source remains elusive. Why?
“The ideal is what’s called a deterministic single-photon source, where you push a button and it emits one and only one photon,” explains Alan Migdall, who leads the quantum optics group at the US National Institute of Standards and Technology (NIST). “Well, nature doesn’t generally work like that.”
Two broad approaches are being taken to fight against nature’s stubbornness and deliver the perfect single-photon source.”The first is developing pair sources to “herald” single photons, and the second is engineering isolated quantum systems that emit single photons.
The first successful single-photon sources by the first approach were developed in the 1960s, and they were based on the property that a cascade transition—the stepwise de-excitation of an excited atom—will emit two photons at different frequencies. By inducing this transition and spectrally filtering the light, researchers were able to observe one photon. The fact of this photon’s existence heralded the other’s existence.
Since this breakthrough, heralded photons have become the workhorse of quantum optics. Typically, they are created by spontaneous, parametric down-conversion sources that, unlike atomic cascade sources, are highly directional and therefore make it easier to catch photons. This type of source is a laser, producing photons that are sent through a nonlinear crystal. From time to time, a photon from this source will split into two separate photons, and the detection of one photon heralds the existence of the other. However, the key phrase here is “time to time”—the source is inherently probabilistic: “If you send in your pump pulse, you’re not always guaranteed of producing a pair; most of the light goes in and doesn’t even interact,” explains Migdall.
An obvious way to get around this difficulty is to generate more photons by increasing the laser pump power. But doing so risks generating two or more photon pairs at a time, which means the source is no longer a producer of single photons.
Instead, a more refined approach that Migdall and many others have pursued is to build a series of down-conversion sources all pumped by the same pulse—a technique known as multiplexing. Each source will have a relatively low probability of producing a pair of photons to avoid generating multiple pairs. But if heralding detectors are placed on every single source, and there are, say, 20 sources, the probability is high that one of them will generate a photon within each pulse.
As Migdall explains, “With multiplexing, you’re taking this probabilistic process of converting pump photons into output pairs and engineering a larger structure that becomes more deterministic.” While this brings heralded photon sources a big step closer to becoming ideal on-demand single-photon sources, it also makes them significantly more complex, requiring a large amount of tightly packed sources and detectors, and a really fast, reconfigurable switch network. What’s more, despite all this work, there remains a small probability of detecting more than one photon from the same source at the same time. The perfect single-photon source would never produce more than one photon at a time, a property known as anti-bunching.
The alternative approach to generating single photons is, on face value, simpler and more favorable when it comes to anti-bunching. In theory, the purest isolated quantum system is a single atom. Electrons surrounding the atomic nucleus can become excited into a higher energy state by absorbing photons or by colliding with other electrons, but they do not stay in these states for very long. They soon return to their ground states, emitting a photon with the same energy as was absorbed.
Integrated photonic chip made out of silicon nitride that has many photon-pair generating waveguides. Photo credit: QET_Labs, Bristol, Alex Clark.
This process is, in principle, deterministic, allowing the production of single anti-bunched photons on demand. “But those single photons are usually emitted in a random direction,” explains Alex Clark, co-director of the Quantum Engineering Technology Labs at the University of Bristol. “So, the difficulty comes down to, ‘How do I put my detector into the right place, such that I always receive the photon that was generated?’” Solid-state systems clear this hurdle by effectively engineering the atom’s environment, embedding it in a tailored waveguide or cavity to direct or funnel the produced photons.
Today, various quantum emitters of this type are being pursued, including solid-state systems involving single atoms and molecules, but also crystal defects like nitrogen-vacancy and other color centers in diamond, and quantum dots (tiny particles of a semiconducting material)—basically anything that confines electrons at the nanoscale and therefore emits single photons.
“The added difficulty here is that when I embed an atom or a defect or a molecule in the solid state, it starts to interact with all the vibrations around it,” adds Clark. “This means that the photons that I create don’t have what we call any coherence [also called indistinguishability]. And if a photon doesn’t have any coherence, then it can’t do all of the operations that people want to build a single-photon source for.”
Conquering this incoherence hurdle requires researchers to cool their systems to cryogenic temperatures, adding complexity and expense to the simple task of creating an efficient on-demand, single-photon source. And even then, the photons produced are not ideal.
The outputs of two quantum dots will depend on the dots’ shapes, sizes, and their local environments, making the production of indistinguishable photons from two or more sources a significant, though not impossible, challenge—as was demonstrated in 2022 when the University of Basel’s Richard Warburton and collaborators produced twin photons from two different quantum dots that were 93-percent identical.
More usually, this problem is overcome by producing photons from a single quantum dot and then using a series of delays to allow the photons to interact. “That’s fine, but the problem is, what if you want to do that for 20 or 100 photons?,” asks Migdall. “For this to be scalable, researchers need to be able to make 100 of them, and I haven’t seen it yet.”
Though both types of single-photon sources still face hurdles, each is being commercialized for various quantum technologies. “Nobody knows what the ultimate best photon source is going to be,” says Clark. “Many people who are working on integrated photonic chip-based photon sources, or in photonic quantum computing like US-based PsiQuantum, are going to be employing massive multiplexing [of heralded single-photon sources] to generate photons. But then other companies, like Quandela in France, are going to be generating photons using solid-state, quantum-dot systems.”
One of Quandela’s competitors is Sparrow Quantum. The company’s flagship product is a 3 mm × 3 mm integrated chip made from ultra-precise indium-arsenide/gallium-arsenide quantum dot structures, embedded in photonic crystal waveguides that are engineered to emit highly coherent single photons at specific wavelengths.
This packaged commercial chip is fundamentally based on the single-photon source Leonardo Midolo, of the Niels Bohr Institute, and international collaborators developed a few years ago to demonstrate secure quantum communication over an 18 km-long fiber located in Copenhagen.
The experiment utilized single photons to implement quantum key distribution (QKD), a secure communication method that enables two parties to produce a shared random secret key known only to them. “We don’t make any claims that single-photon sources are better at doing what we have demonstrated—you could in principle replace the sources with lasers and run it much faster and it will be as secure,” says Midolo. “What we were demonstrating is that the single-photon source is reaching that level of maturity where you can put it somewhere and let it run for days in a stable way, continuously, without human intervention.”
The QKD protocol Midolo and colleagues implemented is only secure if an eavesdropper has absolutely no chance of gaining access to the devices involved. “But sometimes the eavesdropper could be the one that sold you the devices,” explains Midolo. “A device-independent approach is potentially more secure because you only need to rely on the laws of quantum mechanics, and this is where entanglement plays a big role.”
Photon detections as functions of time for antibunched (top), random (middle), and bunched (bottom) light. Photo credit: Tommaso Pregnolato, Niels Bohr Institute.
In fact, quantum entanglement—where two or more particles’ identities are fundamentally dependent on one another, regardless of distance—is key not only to realizing truly secure quantum communication but also to unlocking the most exciting quantum applications. And entanglement between photons can only be achieved when the photons are indistinguishable (i.e., coherent), which is hard to achieve in setups where indistinguishable photons must be produced by multiple different quantum dots.
“This is a real research problem at the moment,” says Clark. “How do you either find a system that naturally gives you very similar photons every time you find this defect, or how do you engineer the system so that you can undo that inhomogeneity? Or is there a different way of making the quantum dots in the first place so that you can have them all emitting at the same wavelength?”
It appears whichever way researchers turn, they are faced with one of nature’s brick walls. A single-photon source might produce 100 percent indistinguishable photons, but retain a small probability of detecting more than one photon at the same time. Another might exhibit near-perfect anti-bunching, but a less than 100 percent chance of those photons being indistinguishable. Building a source that has high, or even perfect values, for these and other figures of merit seems like an extremely tall order. Yet some researchers feel it is not beyond the realms of possibility.
Fabrice Laussy of the Instituto de Ciencia de Materiales de Madrid (ICMM), Spain, is a French theoretical physicist who focuses on the foundational quantum elements of single-photon sources. “We all look at perfection as being not of this world, but actually it’s not true,” he explains. “In quantum mechanics, you can have perfection, but we use another word: Super.” A superconductor is something that conducts electricity perfectly with no loss whatsoever. This absolute mathematical perfection derives from an energy gap for single-particle excitation that the system simply cannot cross when below a given transition temperature.
“For a quantum single-photon source, what is interesting is that there is also a gap which is built in the design: The gap is the quantization of the number of photons,” says Laussy. “You go from zero photons to one photon by jumping over this gap.” A perfect single-photon source will exhibit perfect anti-bunching when this gap is opened in time so that there can never be two photons recorded during a given period.
Laussy first began to make progress in this direction by accident. While working at the University of Wolverhampton, he set his students the task of simulating a single-photon source by taking a stream of photons and, by hand, removing photons so that no two photons were in the same time window.
When plotting the density of two-photon events (the so-called Glauber correlation function) over time, he expected the students to come back with a simple flat line representing coherent light, punctuated by an abrupt dip to zero over the time window where they had manually removed photons to artificially create anti-bunching.
Instead, the students returned with the requisite abrupt dip but surrounded by oscillations on both sides. “I told them, ‘You made a mistake. It cannot be, so please check it,’” recalls Laussy. “But the next day they came back and said, ‘No, the result is the same.’”
These oscillations signaled to Laussy a new physics, where the emission of one photon makes it more likely that the next photon will be emitted after a multiple of the time gap. However, the simulation was completely artificial, with no realistic mechanism for opening such a gap.
To find such a mechanism, Laussy went all the way back to the first single-photon sources from the 1960s, which relied on cascade transitions within individual atoms. A cascade going from the highest state to the ground state would be an ideal mechanism, but it had to involve many levels and go in only one direction, not down a few energy states and then up a couple, for example. “How to do that?” asks Laussy. “A system that can do that is more complicated. At a fundamental level, it looks like it involves some concepts of topology, where we need to break some symmetry to go in one direction only.”
Laussy continues to explore different ways of creating such a system, perhaps offering a glimpse into how a truly perfect single-photon source might one day be engineered. But for now, the near-perfect sources that Migdall, Clark, Midolo, and many others around the world are working on will be more than adequate to keep the wheels of progress turning, inching us ever closer to the quantum revolution.
Benjamin Skuse is a science and technology writer with a passion for physics and mathematics whose work has appeared in major popular science outlets.
