Magnifying the molecules of tomorrow

Materials are technology enablers. If you want to build new types of cars, planes and renewable-energy devices, then you need new materials that have better properties than the ones we use today. It’s our task as materials scientists to develop these materials.

Part of our job is to consider the whole lifecycle of materials; how they age during use and can form defects that could lead to failure. Think of an airplane for example – you do not want to have any material defects. Materials scientists aim to develop materials that are failure proof.

In order to do so, we need to work with the microstructure of the materials. We need to characterize and develop the material’s internal properties. This is why we use microscopes that can magnify materials up to one million times.

Renewable energy is a huge topic in materials research and systems engineering at the moment. This is the major challenge for our society in the next couple of decades.

We have to move away from fossil power to sources that are abundant – such as wind or the light from the sun. In the past couple of years, for example, Germany has been hitting the break-even point where more electricity is produced from renewables than from conventional sources. So we’re on a very good track. But what we urgently need are ways to store this energy in different forms so we can use it when and where we need it.

If you look at global power consumption, we are rapidly approaching the point where 20% of the electricity we consume is for running our computers and servers. So can you imagine the impact if we would find new ways to make computers more energy efficient? This would be a massive contribution towards solving global energy problems. We are only now starting to enter the age of computers with lower power consumption.

Today, the most popular storage devices are lithium-ion batteries. But they are a problem in that the earth has limited lithium resources. This also makes them quite expensive.

I believe part of the solution lies in hydrogen. It’s the smallest atom we know and can diffuse any metal without even needing an electron. To handle hydrogen – with all its properties and promises – is one of the most fascinating challenges of the modern age.

To answer this question we need to look at Moore’s Law.

Gordon Moore predicted that computing speed and the number of transistors in a computer’s CPU would double every one and a half to two years. If you plot the data, you see that he has been absolutely correct for the past 60 years. In this way, Moore’s Law has been an amazing guide in telling us how the computer world would develop over the past years and decades.

It is our task as materials scientists to work on the materials for the next era of computing.

Think about this. From a laptop to a supercomputer, nothing has fundamentally changed in the internal operating principles of the computer for some 60 years. We have memory and data storage, both of which are transformed onto a silicon microchip and then processed by the CPU.

Now, one of the areas materials scientists are working on is the development of devices where the data is stored, processed and rendered from a single source – a single chip. We call this neuromorphic computing. It’s one of the most dynamic areas in materials development.

The most interesting area in neuromorphic computing is the work we are doing with oxides. Titanium oxide and other oxides are the new silicon. They offer energy consumption properties at least ten times lower than what supercomputers are using today.

The next decades will be a very interesting time for the computers. I believe we will have computers so small that we will be able to use them wherever we need them. Thanks to materials development, we are entering an era of information technology systems that are highly specialized in performing specific tasks.

The first challenges between people and computers were chess competitions. As chess is a game where you can actually calculate the end result, there is no way we can beat computers anymore.

Now the question is how much more intelligence can a computer develop? Can a computer develop more capabilities in areas where humans have been trained by evolution?

From my perspective, this may actually be a dangerous vision. But we are certainly approaching this point, and I think that computer technology will be able to reach these dimensions within the next two to three decades.