Turning a temperature difference into a voltage, or a voltage into cooling. It powers space probes such as Voyager from the heat of decaying plutonium, and cools small camping fridges. Layered crystals are among the best materials for it, because heat travels through them badly while electricity still travels well.
Going deeper
A temperature difference drives carriers from the hot end to the cold one, leaving a voltage – and the reverse process cools. The trouble is that the ingredients of a good thermoelectric fight each other: doping that raises conductivity lowers the Seebeck coefficient, so the useful combination peaks in between.
Three quantities that will not cooperate
The efficiency of a thermoelectric material is captured by the dimensionless figure of merit ZT = S2σT/κ. A large Seebeck coefficient S means a large voltage per degree; a large electrical conductivity σ means the current can be drawn without losses; a small κ means the temperature difference is not short-circuited by heat flow.
The three are coupled through . Adding carriers raises σ but lowers S, because a wider spread of occupied states makes the average energy carried by a charge less lopsided; the product S2σ, the power factor, therefore peaks at intermediate doping, typically at heavily doped densities. Meanwhile electrons carry heat as well as charge, so raising σ raises the electronic part of κ too.
What layered and low-dimensional materials offer
A 1993 proposal argued that confining carriers in thin wells sharpens the and could raise the power factor beyond bulk limits. That idea drove much of the field, although in practice the gains have come as much from the other end: interfaces, heavy elements and soft, anharmonic bonding scatter and cut the lattice thermal conductivity.
Layered crystals are natural candidates because heat crosses their weak interlayer bonds badly while charge still moves well within a layer. SnSe is the standout: a layered crystal with extremely anharmonic bonding and very low thermal conductivity, reported with ZT near 2.6 at about 923 K along its softest in-plane direction. Bi2Te3, the workhorse near room temperature, is layered as well.
Reading claims with care
ZT is assembled from measurements made on different instruments, often on different pieces of sample, and the errors multiply: S and σ from one bar, κ from thermal diffusivity, density and heat capacity. Small systematic errors in each can inflate ZT substantially, and crystals make it worse if the three quantities are not measured along the same direction – one reason the highest SnSe values were debated.
A record ZT at one temperature is also not a device. What matters for a module is the average figure of merit across the working range, contacts that survive thermal cycling, mechanical strength, and cost. For specifically, thermoelectrics are hard to use as heat engines at all; their role is more often as a probe, since thermoelectric signals reveal , or as part of photothermoelectric detectors.
For specialists
Conversion between heat and electricity, measured by the figure of merit ZT = S2σT/κ. Low-dimensional and layered materials help by partly decoupling those terms: confinement was predicted to raise the power factor, while heavy elements, anharmonic bonding and interfaces suppress lattice thermal conductivity. SnSe reaches ZT ≈ 2.6 near 923 K along its soft in-plane direction.