Scientists from the US have derived a metric to analyse the cost of power generation using thermoelectric technology. The metric shows that thermoelectric devices have greater potential in large-scale power generation than previously thought.
Thermoelectric generators convert heat into electricity by a physical phenomenon called the Seebeck effect. They are compact, robust, and have no moving parts. This means they are useful for low-maintenance applications, such as in spacecrafts. However, their device efficiencies – the power produced for a certain heat flow – are low, so they are not used for large-scale power generation.
Thermoelectric research has typically focussed on the dimensionless figure-of-merit ZT, a metric that relates directly to device efficiency. Shannon Yee from the University of California Berkeley, part of the team that performed the analysis, explains that for power generation to be commercially viable, what really matters is a technology’s capital cost per watt and not just its device efficiency: ‘we found that the design that minimises a system’s cost per watt value is generally very different to the design that maximises efficiency.’
The team developed a comprehensive metric that takes into account the device architecture, manufacturing costs and materials costs, including the cost of the components, called heat exchangers, that harvest heat on the hot side and reject it on the cold side. They found that for many realistic designs it is actually the device components rather than the thermoelectric material that dominate the overall system cost.
Ali Shakouri, an expert in thermoelectric technology at the University of Purdue, US, was enthused by the work pointing out that we do not have to wait for major breakthroughs in thermoelectric materials to realise the benefits of thermoelectric energy conversion systems. Another leading figure in thermoelectric research, Gang Chen from the Massachusetts Institute of Technology, in the US, agrees: ‘the field has been focusing on ZT. This paper directs attention to cost per watt, which is what really matters in real world applications.’
The analysis identifies several new research directions. One area in particular, says Yee, is the need to improve the plates that sandwich the thermoelectric material. ‘If the costs of certain critical system components could be reduced, even some existing thermoelectric materials could become cost competitive in a design that minimises the cost per watt value.’