For the First Time, Researchers Bridge Quantum Computers on a Single Chip
Maybe building a full-scale quantum computer is just a matter of linking a bunch of small ones.
Image: Sandia National Laboratories
Despite the rapid emergence of mobile devices and cloud computing, it's still intuitive to think of computers as relatively self-contained systems. And, as self-contained systems, they can be relied on to have many if not most of the same capabilities offline as they do online. They work just fine in isolation.
Quantum computing is poised to upend a great many of our assumptions about computers and what it even means to be such a thing. Rather than centralized machines, it may make more sense to imagine the future's powerful quantum computers as distributed networks of very tiny computers operating in a way similar to how parallel computation occurs in classical computing. A problem is broken apart many times and then sent across many different processors at once; at the other end, all of the results are put together into a final solution.
Physicists are routinely breaking distance records for quantum correlations, but connecting quantum computers at very short distances is important too when it comes to distributed quantum computing. To this end, researchers at Sandia National Laboratories have successfully bridged quantum computers at atomic scales. Their work, which is described in the current issue of Science, offers the possibility of arranging many small quantum computers into dense, powerful parallel networks.
More specifically, the problem is in linking together photons—light particles—and quantum emitters, which are atoms that spit out photons at different frequencies as they transition between energy states. These emitted photons represent information, and so the task is to bridge together different emitters in close proximity.
"Efficient interfaces between photons and quantum emitters are central to applications in quantum science, but are challenging to implement due to weak interactions between single photons and individual quantum emitters," the Sandia team writes in Science. "Despite advances in the control of microwave and optical fields ... the realization of integrated quantum devices where multiple qubits are coupled by optical photons remains an outstanding challenge."
Qubits are the units of quantum information. A qubit can be implemented in different ways, but the general idea is that it's something that can represent information not as 1s and-or 0s, but a boundless set of intermediate states. These represent the nodes, or computational units, being connected on the Sandia team's chip. The messengers that carry quantum information between them are photons.
The networked quantum emitters take the form of single silicon atoms embedded in a diamond matrix. This is done by replacing the normal carbon atoms of the diamond with silicon, which has the effect of repelling or crowding out other nearby carbon atoms. As a result, the transplanted silicon winds up with a small surrounding cavity, which protects it against interference from errant electrical currents. This tiny gap is what allows for the stability required for this sort of on-chip quantum networking.
"What we've done is implant the silicon atoms exactly where we want them," Sandia researcher Ryan Camacho offered in a statement. "We can create thousands of implanted locations, which all yield working quantum devices, because we plant the atoms well below the surface of the substrate and anneal them in place. Before this, researchers had to search for emitter atoms among about 1,000 randomly occurring defects—that is, non-carbon atoms—in a diamond substrate of a few microns to find even one that emitted strongly enough to be useful at the single photon level."
Building a large-scale, unified quantum machine remains a huge challenge, but we've made a lot of progress on making relatively small ones. The ability to network a large quantity of them in an integrated environment offers an alternative and perhaps more immediate approach to the quantum computing problem.