The fabrication and characterization of single element p-type SiGe/Si superlattice coolers are described. Superlattice structures were used to enhance the device performance by reducing the thermal conductivity between the hot and the cold junctions, and by providing selective emission of hot carriers through thermionic emission. The structure of the samples consisted of a 3 μm thick symmetrically strained Si0.7Ge0.3/Si superlattice grown on a buffer layer designed so that the in-plane lattice constant is approximately that of relaxed Si0.9Ge0.1. Cooling up to 2.7 K at 25 °C and 7.2 K at 150 °C were measured. These p-type coolers can be combined with n-type devices that were demonstrated in our previous work. This is similar to conventional multi element thermoelectric devices, and it will enable us to achieve large cooling capacities with relatively small currents.

Based on the versatile combination of PbQ- and Bi2Q3-type (Q = S, Se, Te) fragments, we explored new compounds in the Pb/Bi/Se ternary system. The new class of compounds, Pb5Bi6Se14, Pb5Pb12Se23, and PbBi8Se13 are homologues with different combination of alternating Bi2Se3- and PbSe-type layers. α- and β-Pb6Bi2Se9 were obtained in different synthetic conditions and the former is isostructural to heyrovskyite (Pb6Bi2S9) while the latter is a NaCl-type cubic phase. Pb5Bi6Se14 shows a power factor of 11.2 μW/cm·K2 with electrical conductivity of 657 S/cm and thermopower of -131 μV/K at 271 K. The most significant characteristic of this material is the extremely low thermal conductivity of less than 1.0 W/m·K at room temperature. On the basis of these properties, a preliminary doping study for Pb5Bi6Se14 with Sn, Sb, and SbBr3 as dopants was undertaken and the results are presented in this report.

The key challenges in discovering new high ZT thermoelectrics are understanding how the nearly contradictory requirements of high electrical conductivity, high thermopower and low thermal conductivity can be achieved in a single material and based on this identifying suitable compounds. First principles calculations provide a material specific microscopic window into the relevant properties and their origins. We illustrate the utility of the approach by presenting specific examples of compounds belonging to the class of skutterudites that are or are not good thermoelectrics along with the microscopic reasons. Based on our computational exploration we make a suggestion for achieving higher values of ZT at room temperature in bulk materials, namely n-type La(Ru,Rh)4Sb12 with high La-filling.

We have re-examined solid-state thermionic emission cooling from first principles and report two key results. First, electrical and heat currents over a semiconductor – semiconductor thermionic barrier are determined by the chemical potential measured from the conduction band edge, not the energy band offset between the two materials as is sometimes assumed. Second, we show the upper limit to the performance of thermionic emission cooling is equivalent to the performance of an optimized thermoelectric device made from the same material. An overview of this theory will be presented and instrumentation being developed to experimentally verify the theory will be discussed.

Bi2Te3 based low dimensional structures are interesting material systems to increase the thermoelectric figure of merit ZT by either the expected reduction of the thermal conductivity or by a possible power factor enhancement due to quantum confinement. Due to low lattice mismatch Bi2(Te1-xSex)3, PbSe1-xTex, as well as Pb1-xSrxTe, and BaF2 are suitable for Bi2Te3 based low dimensional structures. Especially due to their significantly enhanced band gap lead chalcogenide compounds like Pb1-xSrxTe (Pb0.87Sr0.13Te: 0.6 eV) are well-suited barrier materials in MQW structures. Alternatively the insulator BaF2 can be used for that purpose.

Here we report mainly on results of different superlattice structures mentioned above grown by molecular beam epitaxy (MBE) on BaF2(111). The structural properties of these layers were investigated by X-ray diffractometry (XRD), scanning electron microscopy (SEM) and secondary ion mass spectroscopy (SIMS). Structural performance and thermoelectric properties of different Bi2Te3 based superlattices were reported and compared with regard to their superlattice parameters.

Predicted and observed reductions in thermal conductivity of materials with superlattice structure have prompted interesting research into the possibility of using these materials as higher efficiency thermoelectrics. Synthesis of superlattice materials is challenging however, as the structure itself is generally not very stable at high temperatures as it is prone to interdiffusion. Presented here is the successful synthesis and characterization of a superlattice composed of two materials with the skutterudite structure, IrSb3 and CoSb3.

Bismuth nanowires for thermoelectric applications have been made by electrochemical deposition from an aqueous Bi solution into nanoporous mica substrates, and Anopore (Al2O3) filters. The nanoporous substrates are created by acid etching damage tracks that are produced in the mica using heavy-ion irradiation. In this work, further improvements in the fabrication process are made by investigating the electrochemical growth at the 1-d to 2-d transition region at the nanochannel-substrate surface interface as a function of deposition time and electrochemistry. Issues related to doping of the Bi wires with Te along with the problem of electrochemical growth of contacts to the nanowires as compared to vacuum deposition of contacts will be discussed.

Results on the synthesis and characterization of the solid solutions CsBi4-xSbxTe6, CsBi4Te6-ySey, as well as doping experiments on CsBi4Te6 are reported. We report X-ray structural investigations showing that the Sb or Se atoms in these compounds are not uniformly distributed in the lattice but show preferential occupation of specific crystallographic sites. Thermoelectric properties of selected systems are presented.

In order to investigate the thermoelectric properties of Re-doped β-FeSi2 (Fe1-xRexSi2), Ir-doped β-FeSi2 (Fe1-xIrxSi2), and Pt-doped β-FeSi2 (Fe1-xPtxSi2), the electrical resistivity, the Seebeck coefficient, and the thermal conductivity of these samples have been measured in the temperature range between 300 and 1150 K. Fe1-xRexSi2 is p-type, while Fe1-xIrxSi2 and Fe1-xPt xSi2 are n-type over the measured temperature range. The solubility limits of dopant are estimated to be 0.2at% for Fe1-xRexSi2, 0.5at% for Fe1-xIrxSi2, and 1.9at% for Fe1-xPtxSi2. A maximum ZT value of 0.14 was obtained for Fe1-xPt xSi2 (x=0.03) at the temperature 847 K.

Half-Heusler alloys with the general formula TiNiSn1-XSbX are currently being investigated for their potential as thermoelectric (TE) materials. 1,2,3,4 These materials exhibit high thermopower (40–250μV/K) and low electrical resistivity values (0.1 - 8mΩ-cm) which yields a relatively large power factor (α2σT) of (0.2 - 1.0) W/m♦K at room temperature. The challenge is to reduce the relatively high thermal conductivity (≈ 10 W/m♦K) that is evident in these materials. The focus of this research is to investigate the effect of Sb-doping on the Sn site and Zr doping on the Ti site on the thermal conductivity of TiNiSn. Highly doped half-Heusler alloys have shown marked reduction in thermal conductivity to values on the order of 3.5 - 4.5 W/m♦K. Systematic determination of thermal conductivity in a variety of these doped materials as well as Sb and Zr doped TiNiSn are presented and discussed.

The small effective mass and high mobility of electrons in Bi, make Bi nanowires a promising system for thermoelectric applications. Dense arrays of 20–200 nm diameter Bi nanowires were fabricated by high pressure injection of the melt. Transport properties and Seebeck coefficient were investigated for Bi nanowires with various wire diameters as a function of temperature (1 K < T < 300 K) and magnetic fields (B < 0.6 T). We discuss the problem of the contact resistance of Bi nanowire arrays.

We report the synthesis and structural characterization of two Structure I clathrates in the KSi and Rb-Si systems. The alkali-Si clathrates are fully stoichiometric at the framework sites, i.e., devoid of framework vacancies. This is in sharp contrast to the analogous K-Ge, Rb-Ge and Rb-Sn, Cs-Sn systems, where vacancies are formed at one-third of the crystallographic 6c tetrahedral sites. This is rationalized in terms of Zintl-Klemm rules to remove the tetrahedral atom of its hypervalency. The contrasting behavior is understood in terms of weaker Tt-Tt (Tt – tetrelide, Si, Ge, Sn) bonding as one descends the periodic table, and results in poorly metallic conductivities for vacancy-free K7Si46 and Rb6Si46, but semiconducting behavior of K8Ge44. The observation suggests tuning of the electronic properties of Tt clathrates by substitution of (Si,Ge,Sn) on framework sites, for thermoelectric applications. We describe preliminary results designed to synthesize “mixed” Si-Ge clathrate structures. Thermal decomposition of K2SiGe results in formation of a Structure I clathrate with mixing of Si and Ge on framework sites. The lattice constant ao = 10.523(6) Å, is intermediate between those of K8Si46 and K8Ge44.

We report artificially atomic-scale ordered superlattice alloy systems, new scheme to pursue high-ZT materials. We have fabricated Bi/Sb superlattice alloys that are artificially ordered on the atomic scale using MBE, confirmed by the presence of XRD superlattice satellites. We have observed that the electronic structure can be modified from semimetal, through zero-gap, to semiconductor by changing the superlattice period and sublayer thicknesses using electrical resistivity, thermopower, and magneto-transport measurements. InSb/Bi superlattice alloys have also been prepared and studied using XRD and thermopower measurements, which shows that their thermoelectric transport properties can be modified in accordance with structural modification. This superlattice alloy scheme gives us one more tool to control and tune the electronic structure and consequently the thermoelectric properties.

We have deposited and investigated thin films of lanthanum hexaborate with 1% of the lanthanum replaced by cerium. In bulk single-crystalline form, this material has, due to the Kondo-mechanism, the highest known Seebeck coefficient at sub-K temperatures. Thus it is a good candidate for several thermoelectric applications at very low temperatures. We are studying the kinetic properties of thin films such as the conductivity and Seebeck coefficient as a function of temperature and the dependence of these properties on film thickness, substrate material and deposition conditions. The consequent theoretical performance limits on the device applications of these films are considered with a focus on detectors and refrigerators.

New experimental methods have been developed to electrochemically deposit p-type Sb-rich Bi2-xSbxTe3, Pb-doped and Bi-doped Bi2Te3, and PbTe thick films. Some of the deposited films were dense and had a smooth surface morphology. These films were deposited potentiostatically at room temperature in an acidic aqueous electrolyte. Experimental deposition of Bi2Te3 alloys into various thick nanoporous templates made out of anodized alumina has also been achieved. Miniaturized thermoelements for microdevices (25 μm tall, 60 μm diameter) were grown by plating through thick photoresist templates. The experimental techniques developed, as well as the transport properties of some of the films and filled templates, will be presented.