The carbon footprint of the Internet already exceeds that of air travel. Oscar Céspedes, Associate Professor of Physics at the University of Leeds provides an informative article on unsustainability of the computing world and the need to find low-carbon solutions. He shows that there are options to continue our progress and minimise the ecological impact.
Low-carbon computing is needed to avoid a technological collapse
Human society has come to rely on superior gadgets being produced every year. Each year, new phones or laptops are faster, sleeker and have even more capabilities. However, electronics are rarely recycled, and the carbon footprint of the internet already exceeds that of air travel. The internet also relies on “rare earths” and heavy and noble metals, leading to pollution and socio-economic problems associated with mining these scarce and often noxious materials.
Maintaining this information network across the world requires more energy than the whole of the UK – and energy demand from data centres is expected to treble over the next decade. Three google searches are equivalent to a 60 watt bulb being switched on for a minute and produce 0.6g of CO₂. And computers already generate so much wasted heat that companies like Facebook are now placing new data centres in the Arctic to keep them cool.
Given all this, the capability to fulfil our computing needs while producing negligible heat waste and reducing our power needs would have momentous consequences. Indeed, to avoid a technological and societal collapse, we’ll need a massive shift towards low-carbon computing.
Even if we ignore the environmental impact, we are still faced with the problem that current computing technologies are inching closer to their fundamental limits in size and processing speed. More transistors are being packed on to computer chips and they will eventually, within the next decade or so, reach the smallest possible size at which we can reliably transmit charge or information. This is why progress depends on developing new fabrics using eco-friendly materials and low power electronics.
To justify the huge industrial and public investment associated with replacing production lines, infrastructure and working methods, this new technology would have to be not only better for the environment, but simply much better in all aspects. Imagine computers that operate orders of magnitude faster, do not need fans to cool down, can operate for longer on a single battery charge or can store vast amounts of information with near immediate access.
For now, conventional hard disks (magnetic domain storage) and silicon-based technologies still dominate. Scientists have come up with lots of alternatives, but many, such as molecular electronics, have so far failed to deliver fully on their original promise. Others, such as quantum computers or superconducting devices, are highly specialised or lie in the long-term future.
There are other options to continue our progress and minimise the ecological impact. Within my own field of “spintronics”, we explore how to transmit information through measuring an electron’s “spin” rather than its movement or charge. This loses very little energy when transmitting electricity.
Spintronics is currently used mostly in sensing and computing, for example in the read-heads of hard disks (without it we couldn’t store/read so much information in so little space). In the future, it could be used to convert wasted heat back into electricity, or in car sensors or biomedicine.
In all of this, carbon-based molecules can play a critical role. These environmentally friendly materials can carry a charge or store information. But they can also be applied as a very thin layer of molecules on top of a thin layer of metal which means they share electrons (hybridise), thus changing the properties of both materials. This can be used to, for instance, generate magnetism in non-magnetic materials such as copper. This will allow technological progress while reducing both the power needed and the use of rare earth metals, which are expensive and environmentally damaging.
Many of these technologies may not make it out of the lab in our lifetimes. Rather than becoming common in personal devices, they’ll remain a plaything of scientists, confined to experiments. However, even if only some of them were to be successful, it would lead to a step change on how we interact with our virtual environment and what it can do for our environment, health, politics, or transport. The future of the Homo technologicus will depend on it.