Running on empty

C&I Issue 4, 2016

It’s an often quoted fact that the average car loses around 70% of energy generated through fuel consumption as heat. Thermoelectric generators (TEGs) improve fuel efficiency by directly converting energy from the hot engine exhaust into electricity, used to power accessories such as communication and navigation systems, and air conditioners and heaters.

A thermoelectric material generates electricity when there is a temperature difference between one end of the material and the other. The electrons in the hot end diffuse to the cool end, producing an electric current. A good thermoelectric material must be good at conducting electricity but bad at conducting heat. If it was good at conducting heat, the whole material will rapidly reach the same temperature and the electrons would stop flowing.

Thermoelectric materials have been used for many years as generators for deep space probes, where they convert the heat from radioactive decay into electricity. In 2013, for example, Voyager 1 became the first man-made object to exit the solar system and enter interstellar space after being continuously powered by a thermoelectric generator for 36 years.

Temperatures in spacecraft generators are typically high – around 1000°C – hence the use of thermoelectric materials such as silicon germanium. Back on earth, temperatures in car exhausts are much lower, usually 400-700°C – a diesel generator exhaust is around 500°C – so different materials are required. Advantages of thermoelectric generators include no moving parts, quiet operation, reliability and minimal maintenance; however, their high cost and low efficiency restricts them to niche applications, such as air-conditioned car seats, wine coolers and medical refrigerators.

Various international consortia have shown that thermoelectric modules can contribute to energy efficiency in vehicles. Prototypes have already created up to 600W of electrical power from the waste heat in the exhaust system of a car. ‘There were almost 60m motor vehicles registered in Germany at the beginning of 2013,’ says Kilian Bartholomé from the Fraunhofer Institute for Physical Measurement Techniques (IPM) in Freiburg, Germany. ‘If all of these were equipped with small thermoelectric power plants in the exhaust systems, energy of the order of the amount produced annually by a nuclear power plant could theoretically be saved today. That corresponds to a saving of several million tons of carbon dioxide.’ In view of the continually stricter environmental regulations, automobile manufacturers are very interested, he adds.

Jeffrey Snyder of the California Institute of Technology (Caltech), US, specialises in thermoelectric materials and devices. He says that ‘if we could double their efficiency, then thermoelectric modules incorporated into an automobile engine’s exhaust system could generate enough power to replace the alternator, which would increase the car’s gas mileage’.

There are many technical challenges to integrating thermoelectrics into car engines. The US Department of Energy (DOE) lists developing materials that are good conductors of electricity but bad conductors of heat; scaling up materials developed in the laboratory; availability of cost-competitive materials; cost-effective manufacturing technology; and integrating thermoelectric generators with other vehicle components.

Promising materials

Thermoelectric materials are classified by structure and composition. Some of the main classifications are chalcogenide, such as, bismuth telluride and lead telluride; clathrate; skutterudite; half-Heusler; silicide; and oxide. Commercial thermoelectric modules for low-temperature use are mainly made of bismuth telluride and its solid solutions with antimony or selenium; lead telluride has better thermoelectric properties at higher temperatures (500–600°C) but has environmental issues.

In 2014, an interdisciplinary team led by inorganic chemist Mercouri Kanatzidis at Northwestern University, US, discovered that the crystal form of tin selenide could be the world’s most efficient thermoelectric material (Nature, doi:10.1038/nature13184).

Unlike most thermoelectric materials, tin selenide has a simple lattice structure, like an accordion, which provides the key to its exceptional properties. The team found that the bonds between some atoms in the compound are very weak, meaning that atoms vibrate very slowly. Together with the accordion-like structure, this makes the material a poor heat conductor. It conducts heat so poorly that even moderate thermoelectric power and electrical conductivity are enough to provide high thermoelectric performance at high temperature.

The efficiency of waste heat conversion in thermoelectrics is reflected by its ‘figure of merit’, called ZT. The ZT metric represents a ratio of electrical conductivity and thermoelectric power in the numerator, which needs to be high, and thermal conductivity in the denominator, which needs to be low. Tin selenide exhibits a ZT of 2.6, the highest reported to date at around 650°C.

‘The inefficiency of current thermoelectric materials has limited their commercial use,’ says Kanatzidis. ‘We expect a tin selenide system implemented in thermoelectric devices to be more efficient than other systems in converting waste heat to useful electricity.’

More recently, the team has been working to increase the efficiency of their materials even further by doping tin selenide with sodium. They found the doped material produces a significantly greater amount of electricity than the undoped material, given the same amount of heat input. They also were pleased to see that adding sodium did not affect the already very low thermal conductivity of the material.

Tin selenide is very unusual, not only because of its exceedingly low thermal conductivity, but also because it has many conduction bands of energy levels partly filled with mobile electrons, explains Kanatzidis. Most semiconducting materials, such as silicon, have only one conduction band to work with for doping - whereby ‘dopants’ donate electrons to the band to boost electrical conductivity. The team showed that they could use sodium to access the multiple channels in tin selenide and send electrons quickly through the material, driving up the heat conversion efficiency.

‘The secret to our material is that multiband doping produces enhanced electrical properties,’ says Kanatzidis. ‘By doping multiple bands, we are able to multiply the positive effect. To increase the efficiency, we need the electrons to be as mobile as possible. Tin selenide provides us with a superhighway - it has at least four fast-moving lanes for hole carriers instead of one congested lane.’

The new doped material produces high ZTs across a broad temperature range, from room temperature to 500°C (Science, doi:10.1126/science.aad3749).

So the average ZT of the doped material is much higher than straight tin selenide, resulting in higher conversion efficiency. ‘Now we have record-high ZTs across a broad range of temperatures,’ adds Kanatzidis. ‘The larger the temperature difference in a thermoelectric device, the greater the efficiency.’

Tin selenide is only one of many materials under investigation by researchers around the world. Others include skudderite, a mineral made up mainly of cobalt and nickel arsenide; tetrahedrite, a mineral consisting of a sulphide of antimony, iron, and copper; and oxides such as lanthanum strontium titanium oxide (LSTO).

Researchers at the University of Manchester, UK, are working with Leicestershire-based European Thermodynamics to develop LSTO materials as cheaper alternatives to tellurium-and selenium-based compounds. However, Robert Freer, a professor in the School of Materials at the University of Manchester, explains that current oxide thermoelectric materials are limited by their operating temperatures, which can be around 700°C. ‘This has been a problem, which has hampered efforts to improve efficiency by utilising heat energy waste for some time,’ he says.

The Manchester team reports that adding graphene to LSTO extends the operating range from 750°C down to room temperature (I. A. Kinloch et al, ACS Applied Materials & Interfaces, doi: 10.1021/acsami.5b03522). Graphene slows the transfer of heat through the material, Greer explains; this broadening of the temperature range was not seen with other dopants.

The team also found that the new nanocomposite material had lower thermal conductivity and significantly higher electrical conductivity and power output. The highest ZT was achieved when 0.6wt % graphene was added (ZT = 0.42 at room temperature and 0.36 at 750°C).

Reducing the thermal operating window to room temperature opens up huge numbers of possible applications, says Freer. ‘The new material will convert 3-5% of the heat into electricity. That is not much but, given that the average vehicle loses roughly 70% of the energy supplied to it by its fuel to waste heat and friction, recovering even a small percentage of this with thermoelectric technology would be worthwhile.’

He says the discovery highlights an alternative strategy to nanostructuring for developing high-performance, environmental friendly, low-cost thermoelectric materials.

Engineering design

Developing processes to make sufficient quantities of thermoelectric materials remains challenging. One tricky issue for industrial applications is to attain the efficiency values generated in the lab after mass production.

Kilian Bartholomé’s IPM team works with half-Heusler compounds, consisting of XYZ, where X can be a transition metal, a noble metal, or a rare-earth element; Y is a transition metal or a noble metal; and Z is a main group element. They possess good thermoelectric properties and withstand high temperatures. Working with a multidisplinary team as part of the government-sponsored thermoHEUSLER project, the researchers have synthesised alloys in kilogram quantities with a ZT value of 1.2.

Another issue, points out, Alex Cuenat, of the UK’s National Physical Laboratory (NPL), is that a material’s performance depends on its composition and it can be difficult to produce the same material consistently. The NPL is developing and testing materials to find the most stable one with the most consistent properties so they can then measure its efficiency very accurately; this would give industry a reference or calibration point for efficiency.

‘We have measured hundreds of materials but stability is an issue,’ says Cuenat. ‘The clear favourite is skudderite, an unusual but promising thermoelectric material.’

But even if these hurdles are overcome, challenges remain in engineering the chosen materials into devices. Thermoelectric modules are assembled from blocks a few millimetres each in size, consisting of two different types of thermoelectric materials, p-type and n-type, connected electrically in series and thermally in parallel.

Material properties are highly temperature-dependent but few applications have heat sources at one single temperature. One approach is to combine different materials in one device to give an average ZT over the application temperature range.

But designers also have to choose materials that are stable over this temperature range while allowing for the fact that the arrangement of the material within the module significantly affects thermal and electrical transport in the overall device.

For example, skutterudite is very dependent on small variations in position and needs to be arranged in a given shape, which can be time-consuming and expensive, says Cuenat. ‘The p- and n-type layers need to be arranged by hand, or possibly by robot. In future, perhaps a device will be invented that could spray or electrodeposit the layers.’

In a thermoelectric generator (TEG), the modules are connected in parallel with heat exchangers to facilitate the transfer of heat from the heat source to the module’s hot side and away from its cold side. Choosing the best materials for the other device components affects overall device performance. For example, the effectiveness of the heat exchangers directly affects the temperature drop - and thus voltage - across the thermoelectric material. Another critical aspect for efficiency is the design of electrical contacts. These need to withstand large temperature differences, yet at the same time keep the electrical resistance as small as possible.

Several major issues need to be resolved before TEGs for engines reach the market, writes Saniya LeBlanc an engineer at the George Washington University in Washington DC, US, in her 2014 review of TEGs (

The technology’s time-to-market will be influenced by ‘manufacturability’ combined with costs, she notes. However, ‘active’ R&D efforts along with emerging new prototypes indicate that solutions to the materials and systems challenges are ‘well under way’.

Meanwhile, Cuenat believes a ‘viable system’ is ten years away. ‘[Timing] will be driven mainly by energy efficiency regulation, when [auto] manufacturers have exhausted other options such as power and weight of engines,’ he believes.


Become an SCI Member to receive benefits and discounts

Join SCI