Component viability risks bursting the quantum bubble
To speed the arrival of quantum technologies, the incredibly
demanding compound semiconductor devices that lie at the heart of them
need to be produced on high-volume platforms.
BY DENISE POWELL AND WYN MEREDITH FROM THE COMPOUND SEMICONDUCTOR CENTRE, SAMUEL SHUTTS FROM CARDIFF UNIVERSITY, MOHAMED MISSOUS FROM THE INTEGRATED COMPOUND SEMICONDUCTORS, MOHSIN HAJI FROM THE NATIONAL PHYSICAL LABORATORY AND CHRIS MEADOWS FROM CSCONNECTED
There is no doubt that quantum technologies have the potential to revolutionise every sector we can think of. Their impact will include: highly accurate navigation, enabled by quantum gyroscopes; GNSS-free communications, underpinned by atomic clocks; ultra-secure communications, via quantum key distribution; improved manufacturing control and timely maintenance on infrastructure; the detection of anomalies in organs such as the brain and heart, through quantum magnetometers, alongside rapid drug and materials discovery, and financial modelling, enabled by quantum computers.
The possibilities for quantum technologies are so vast that this revolution is anticipated to be on par with that of AI, in terms of scale. In fact, these two headline-grabbing technologies are complementary, with the true magic underway when quantum systems are enabled by AI. This is not just fantasy: AI is already applied to data from quantum systems at the UK’s National Physical Laboratory, to ensure rapid analysis.
Unfortunately, for any nascent technology, promise is no guarantee of success. History attests that when a technology with great potential delivers encouraging results, substantial investment follows – but this optimism may well be short lived, with the bubbles breaking to induce a widespread cull that leaves those hanging under the spotlight having trying to salvage a future for their revolutionary technology. Today some firms are still recovering from the lidar aftermath, and reports suggest AI is next for re-evaluation.
And what of quantum? Why aren’t we seeing widespread deployment of this technology, on the back of investment totalling hundreds of millions of dollars? You might be thinking that the humble laser draws on quantum effects, so quantum is already well-embedded in our lives. That’s somewhat true, but misses the point that here we are considering what most refer to as ‘Quantum 2.0’ – that is, technologies that utilise superposition or entanglement, or as Einstein famously said, “spooky action at a distance”.
To delve deeper into the future of quantum, it's helpful to consider an example. One highly successful Quantum 2.0 product is the world's first commercially available chip-scale atomic clock, Microchip’s Microsemi SA.45s CSAC. According to the National Institute of Standards (NIST), this triumph is the culmination of more than 10 years of extensive R&D, costing several tens of millions of dollars, with support from both the Defence Advanced Research Project Agency (DARPA) and NIST. John Kitching, a key researcher at NIST involved in this development, rightly suggests that given that the market for chip-scale atomic clocks is worth around $200 million per annum, it’s difficult for industry to invest the amount needed for fundamental R&D in this area.
The Fabry-Pérot profile across a 894.6 nm VCSEL structure grown on a 100 mm GaAs substrate using Aixtron G3 series MOCVD system.
VCSELs: a hero in the making
A key component in miniature atomic clocks, as well as other quantum systems based on alkali metal vapour, is the VCSEL. Renowned for its low power operation, circular beam profile and intrinsic reliability characteristics, this class of laser is ideal for interrogating the alkali metal atomic species, typically rubidium or caesium, that are contained in a small cell within the atomic clock. As their principle of operation is coherent population trapping, the VCSEL must emit at highly precise wavelengths corresponding to atomic energy transitions. For example, the emission must coincide with D1 transitions at elevated temperatures, which occur at 795 nm and 894.6 nm for rubidium and caesium, respectively.
In addition to these highly stringent wavelength requirements, there are many other specifications that must be met, including single-mode operation with a narrow linewidth and high mode stability. Fulfilling them all is not easy, as in some cases, adhering to one narrow tolerance makes it harder to meet the demands of another. Given this state of affairs, it’s no surprise that it’s not a stroll in the park to realise the high levels of epitaxial material design, growth and fabrication demanded for VCSELs deployed in quantum technologies.
To illustrate this point, when VCSELs serve in quantum technologies, the ideal uniformity for the Fabry-Pérot dip across a wafer is below 1 nm, with the precise figure depending on the current tuning capability. In stark contrast, telecommunications applications can tolerate a non-uniformity of this parameter of several nanometres across a wafer.
Our organisations help to target these requirements. The Compound Semiconductor Centre (CSC), in partnership with IQE, has worked extensively on improving the centre-uniformity profile for 100 mm epiwafers produced with Aixtron series G3 MOCVD tools. This is the preferred substrate size for the VCSEL design, fabrication and test partner Integrated Compound Semiconductors, of Manchester, UK.
Fabrication is equally challenging, demanding tight control of the oxidation processes to ensure single-mode operation at the required optical output power. Naturally, these process challenges impact yield. Whilst standard VCSEL platforms boast yields typically in excess of 90 percent, those developed specifically for quantum applications are far lower, due to the cumulative effects of stringent specifications.