Reality, at least as we know it, only goes so deep. Look closely enough at any object, down to the level of molecules and atoms, and the world starts to play by its own rules. This is the realm of quantum physics: where waves of energy and particles are the same, and strange phenomena like teleportation are the norm.
These enigmatic traits could be the key to revolutionary new computers and electronic components. Instead of using silicon transistors, like a traditional computer or integrated circuit, quantum devices rely on subatomic particles as a means to route and process information, making them faster and more powerful than any other electronic hardware we can currently imagine.
Three new faculty from Cornell's School of Electrical and Computer Engineering are working to make quantum devices both practical and scalable. Assistant professor Karan Mehta, together with assistant professor Mohamed Ibrahim and associate professor Mark Wilde, are each going far beyond applied physics in their work, incorporating elements of circuit design, photonics, systems architecture, information theory and other fields to make quantum computers a reality.
Trapping ions
Mehta, for instance, studies a basic building block of quantum computers - a specialized component called a "trapped ion qubit." It's essentially a single atom suspended in a vacuum by electric fields and controlled with lasers. By using those lasers to manipulate the atoms' spin and charge, it's possible to "program" them to run simple algorithms.
As with any electronic component, however, these qubits have pros and cons, Mehta notes. One advantage is that each ion is suspended in space and isolated from other atoms, meaning it's exposed to very little interference or noise. But controlling these qubits is complicated, and as systems get larger and larger, other sources of noise can creep into the system, preventing it from working smoothly. Getting rid of that noise is a critical part of building a useful quantum computer, which would require thousands or even millions of qubits.
"When you have large numbers of ion qubits in a system, controlling them with millions of laser beams moving around in free space becomes very hard," Mehta says. "Whenever you add more qubits into the system, the complexity of the control apparatus will introduce more potential errors and noise."
In quantum computing, that noise can scramble the output of a machine. When minute vibrations, heat or anything else that randomly perturbs a trapped ion appears, the qubits lose a critical trait called superposition - a phenomenon where electrons exist in multiple states at once, letting programmers run different iterations of a problem at the same time. If there's any noise present, however, that superposition will collapse prematurely, creating errors in computation.
Mehta is trying to get around this limitation by using solid-state devices to manipulate and sense the state of each qubit. He thinks using pulses of light delivered to qubits and collected into chip-based control devices based on fiber optics may be the key to clean, low-noise quantum systems. Such systems could allow large scale systems, and also significantly reduce excess noise, making qubits more stable.
"From an engineering perspective, that can address the elephant in the room, which is the challenge of controlling these otherwise pristine quantum systems," he says. "The idea is to leverage the fundamental advantages of extremely clean, low noise quantum systems, together with scalable hardware."
Quantum systems on chip
Ibrahim is on board with that assessment. He's working on scalable chip-scale quantum systems in his lab utilizing today's advanced and miniscule integrated circuits (ICs).
Ibrahim is developing integrated quantum sensors using a specialized form of diamond crystals. Instead of pure carbon, these diamonds are seeded with atoms of nitrogen. When paired with a vacant site, each nitrogen atom introduces a nitrogen-vacancy (NV) center with unique new properties.
By exposing these crystals to a rising sweep of microwave energy and green light pulses, he says, they begin to glow fluorescent red with intensity depending on the spin states of the NV centers' electrons - and by recording the exact frequencies at which a dip in the fluorescence intensity occurs, Ibrahim can track the temperature and measure the intensity of magnetic and electric fields that are surrounding the sensor.
Although this is a well-known property, Ibrahim is working to combine all the elements involved into a single chip-scale miniaturized device, including on-chip microwave radio source and red-light detection circuits. These are co-packaged with a diamond crystal lattice and a green laser emitter.
Integrated circuits like these, he says, could have all sorts of different applications, from global navigation to sensing bioelectric signals in the heart and brain - but Ibrahim says he's also interested in building integrated controllers for quantum computers, where they might help to solve an age-old problem.
"Qubits need to be kept in a cryogenic fridge. In order to send signals between those ultra-cold environments and the classical computers that control the qubits, we currently use cables, which limit the scalability to thousands of qubits," he says. By using cryogenic ICs as an intermediary, operating at few Kelvins, it may be possible to build multi-qubit controllers that can scale to a larger number of qubits much more efficiently.
"However, we still need to communicate with intermediate cold temperature, which is currently done using conductive coaxial cables. Since these cables are also thermally conductive, we can actually lose energy along them on the order of a few milliwatts," he says.
Ibrahim is working on efficient transceivers that can solve this problem using either wireless communication or cables with very low heat conductivity, such as optical fibers. The utilization of ICs to develop new architectures to interface or directly control qubits would make it possible to increase their number, enabling the era of large-scale quantum computers.
Programming qubits
No matter how robust or efficient we can make a quantum computer, however, we won't get very far unless we figure out the most effective ways to use it - an area Wilde is actively studying. While his colleagues in the School of Electrical and Computer Engineering are developing new hardware and software to make these devices a reality, Wilde is turning his attention to quantum information theory, or the complex algorithms used to process information within that device.
Not surprisingly, he says, quantum computers are far less straightforward than classical silicon devices. A classical computer with two bits, each taking values zero and one, can generate four different combinations of those numbers (00, 01, 10 and 11), but can only calculate one at a time. A quantum computer, on the other hand, can explore all four possible answers at once - and as a result, requires entirely new methods of programming.
"The cleverness involved in devising a quantum algorithm is to make the bad possibilities for an answer go away; to eliminate them from the computation like pruning a tree, and then amplify the paths that will lead to a correct solution when you ultimately measure it," Wilde says.
Since noise in the quantum system will introduce errors during that pruning process, Wilde is working on ways to correct for those instances and ensure that noisy glitches don't skew the computer's output. One technique, he notes, is to make quantum algorithms as efficient as possible, reducing the amount of time they take to run and limiting the qubits' chances of being corrupted by noise as the computation occurs.
Although he's working on new ways of constructing robust quantum algorithms, Wilde's work isn't entirely focused on practical solutions. He's also trying to answer puzzles with a more philosophical bent.
"I want to understand the ultimate limits of communication," he says. "In every communication task, you're going to need to do some kind of computation on either end, and in every computation task, you're going to have to communicate between qubits inside the computer - so computation and communication are inevitably intertwined, and you can never separate them." With that in mind, he asks, what are the physical limits of those processes? And how far can we push them?
These questions aren't just abstract thought experiments; they're the bread and butter of the work that Wilde and his colleagues are currently doing. In time, the interdisciplinary research coming out of their labs may revolutionize computing and electrical engineering as a whole, opening an endless array of new possibilities based on quantum physics.
