Thermoelectric generators are devices which transform a temperature difference to electrical energy or vice-versa. One way of increasing the efficiency of a thermoelectric generator is to improve the capability of a material to maintain a difference in temperature, which can be achieved by creating a material that consists of many small crystallites. A good choice of different analytic methods to determine the critical parameters of a nanostructured material is therefore essential.
Thermoelectric generators are devices which transform a temperature difference to electrical energy or vice-versa. They are usually used for recovering waste heat from exhaust systems of cars, to power space probes, to charge electrical devices while camping, or for low-noise Freon-free refrigerators. The low efficiency of thermoelectric converters, between 5 and 8%, still limits its usage.
One way of increasing the efficiency of a thermoelectric generator is to improve the capability of a material to maintain a difference in temperature, that means to reduce the heat conductivity. To achieve this goal without changing the material's composition one can introduce more grain boundaries, that means to create a material that consists of many small crystallites or grains varying in orientation.
Let’s take silicon as an example. The pure single-crystalline material without any grain boundaries has a thermal conductivity of 156 W/m-K. This is a comparably large value, which means that if you start to warm up one side of a piece of silicon, the other side will become hot very fast and the entire piece of material will have nearly the same temperature all over. With no difference in temperature in the material, no electricity can be generated.
Through nanostructuration methods, i.e. by creating a piece of silicon from many tiny grains, it is possible to create a very high density of grain boundaries where the heat conduction waves are scattered. Such a nanostructured piece of silicon can have a thermal conductivity between 7 and 22 W/m-K.
Another important property of a thermoelectric material is good electrical conductivity so that the electrical current flowing through the device does not cause thermal losses. By nanostructuring a material such as silicon with grain sizes larger than 10 nanometers, the thermal conductivity is decreased without any significant effect on the electrical conductivity. This works because the mean free path of waves that carry heat is much longer than the waves that carry electrical properties. The material acts like a nearly perfect crystal with excellent conductivity on the length scale of waves that carry electricity, while the waves that carry heat are strongly inhibited. For more details on the subject, please refer to an excellent review on nanostructured silicon by Dr. Gabi Schierning.
Since silicon is a very cheap material, such optimization of its properties makes it the material of choice for high-temperature thermoelectric applications. Other materials that have good thermoelectric properties, such as LAST (AgPb18SbTe20) and YbMnSb11, contain noxious or expensive substances, and they are less resistant to high temperatures and corrosive environments.
Optimizing the nanostructuration of a material and its heat transport properties is therefore a highly promising approach towards competitive materials. A good choice of different analytic methods to determine the critical parameters of a nanostructured material is therefore essential. In the following table I summarize some characterization methods for this application with their advantages and disadvantages. A combination of some of these methods is usually an ideal solution.
|Transmission Electron Microscopy (TEM)||Grain sizes; Purity at grain boundaries; Microstructural features||Very local information about elemental distribution and structure down to single atoms||Invasive; Information about a very small part of the sample|
|X-Ray Diffraction (XRD)||Crystal structure; Chemical composition; Average nanocrystallite size||Average over the whole material; Cheap Cu-α in-house instruments; Clean data from sychrotron||Not trivial data analysis|
|Pair Distribution Function (PDF)||Local structure; Real space information on bond length||Very precise information (ex: Si-Si and Si-O bonds have very distinct bond length)||Synchrotron radiation required|
|Small Angle Neutron Scattering (SANS)||Average crystallite size||Average information over relatively large samples||Large amount of sample and neutrons source required|
|Prompt Gamma Activation Analysis (PGAA)||In-depth information about a sample's chemical composition||Information about chemical elements which are not detectable with other methods (e.g. H)||Neutron source required|
|Density of Phonon States with Time of Flight (TOF)||Material's purity; Effects of nanostructuration on heat transport waves||Capable of precisely showing the presence of Hydrogen||Neutrons source required|
|Heat capacity||Similar information as TOF + Electronic contribution||Cheaper and easier access than neutrons||Measurement performed at low temperatures (2 K)|
|Ressonante Ultrasond Spectroscopy (RUS)||Elastic constants of a material||Equipment can be constructed in-house by knowledge workers||Not trivial data analysis|
|Direct measurements of thermoelectric properties||Thermal conductivity; Seebeck coefficient; Electrical conductivity||Complete information about thermoelectric properties||Restrict temperature ranges depending on equipment|
The biggest hurdle for measurements at neutron or synchrotron sources are the required expertise and difficult access for the general public. Bond Consulting has experts for experimental design and data analysis in this area and have a large network within this highly specialized and tight-knit scientific community. This guarantees fast, cost-effective and reliable results that lead to high-end materials and products. In addition, we can quickly perform analysis with SEM, TEM or other methods at the institutions we collaborate with.