The materials science problem

2 mins read

How the lessons from GaN and Oxide materials will enable the future of AI, according to Dr. Rodney Pelzel, Chief Technology Officer, IQE.

Artificial Intelligence (AI) has taken the world by storm over the last year, and this has served as an inflection point for many semiconductor sectors (e.g., data centre infrastructure).  It is important to realise that AI is only at the beginning, and a new generation of semiconductors, compound semiconductors (CSs), will be combining Smart Devices and AI to create Intelligent AI that will drive the world of the future.  CS technology underpins the advanced sensing, data transmission, and power usage required by AI.    

Intelligent AI will also require advancements in computing.  Not only will this employ leading-edge computer algorithms, but it will also need improved hardware.  Quantum computing is one possible solution to this hardware dilemma.  However, there are fundamental challenges that exist today to adoption of quantum. Namely, the technology is limited by physics and the materials available, including the challenge of reducing the qubit error rate.  

Overcoming these challenges will require new processes and materials, the development of which could leverage capabilities that IQE has demonstrated as part of its CS business.

From Power to Quantum

Some of the next generation materials and processes that are being widely adopted now for power, including gallium nitride (GaN) and oxide materials, are likely applicable to quantum. These will be combined with new materials such as rare earth elements, superconductors, and 2D materials to create materials specific to quantum applications.

In order to meet the challenges for quantum development, it is important to see the problems for what they are and what they are not.  The challenges are not semiconductor manufacturing problems, at least not yet. Instead, they are materials science problems. 

The good news is that these are the types of problems that have already been successfully worked by IQE for many years in developing CS materials for communications, advanced sensing, and power electronics.  Fundamental materials innovation will be the key for quantum to step out of its niche position and become more mainstream, and when/where appropriate IQE will be there.

Manufacturing Process

For a materials solutions provider like IQE, quantum computing and AI are unique in they are both inputs and outputs of IQE’s technology development and manufacturing processes.  As previously noted, IQE’s technology and manufacturing pedigree could enable the production of quantum materials. 

At the same time, as a semiconductor manufacturer, IQE and companies like it can adopt and apply quantum computing to its processes and products to enable their development and manufacture.  Quantum could play roles in design, simulation, testing, optimization, and “smart manufacturing.”

The most obvious place to apply quantum computing would be in the use of AI for Fault Detection and Classification (FDC), which is already well-fitted for automation because it’s a process that relies heavily on data to make predictions about the performance of products and reactors.

Quantum computing could also be used in this context and the technology could then help with feed forward-looking predictions for the manufacturing process.

Trust, but Verify

Quantum computing holds tremendous promise. However, as noted, before we begin scaling up for this next-gen computing, we need to have a firmer grasp of what we’re scaling. While the science of quantum computing is of course fascinating, as business leaders, we operate in a practical space, and we must find the best practical solutions. This may or may not be quantum computing.

If solutions are found for quantum’s fundamental, materials challenges, scaling will be relatively straightforward if the underlying technology is solid. This may sound odd when talking about quantum computing, and it is not to say that manufacturing for the technology will be simple and problem-free, but I am confident that any issues that we, as manufacturers, encounter at that stage, will be able to be reasonably addressed.

The exact timing of when we will reach that stage remains to be seen, but there will likely be a steady development of the associated materials technology and manufacturing processes throughout the remaining years of this decade.