WO2005098981A1

WO2005098981A1 - Thermoelectric applications of composites of ceramics and carbon nanotubes

Info

Publication number: WO2005098981A1
Application number: PCT/US2004/007187
Authority: WO
Inventors: Guodong Zhan; Joshua D. Kuntz; Amiya K. Mukherjee
Original assignee: The Regents Of The University Of Califonria
Priority date: 2004-03-08
Filing date: 2004-03-08
Publication date: 2005-10-20

Abstract

Composites formed by consolidating powder mixtures of ceramic materials and carbon nanotubes are discovered to have favorable properties as thermoelectric materials for use in devices for generating electric power by the Seebeck effect or for causing heat flow by the Peltier or Thomson effects.

Description

THERMOELECTRIC APPLICATIONS OF COMPOSITES OF CERAMICS AND CARBON NANOTUBES

STATEMENT AS TO RIGHTS TO INVENTIONS MADE UNDER FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0001] This invention was made with Government support by Grant No.

G-DAAD 19-00-1-0185, awarded by the United States Army. The Government has certain rights in this invention.

BACKGROUND OF THE INVENTION

1. Field of the Invention

[0002] This invention resides in the field of thermoelectric materials and devices, and in particular, in methods and materials for improving the energy conversion efficiencies of devices that utilize thermoelectric effects.

2. Description of the Prior Art

[0003] Thermoelectric power generation arises from the phenomenon known as the

Seebeck effect, first reported by TJ. Seebeck, in R.. Acad. Sci.: 265 (1822). The Seebeck effect is the spontaneous development of an open-circuit emf in a circuit formed by joining two dissimilar conductors when the junctions of the two conductors are maintained at different temperatures. The emf arises because the higher energies of the electrons at the hot junction produce a net diffusion of electrons (or charge carriers) toward the cold junction, creating a charge imbalance that results in an electric potential between the hot and cold junctions. The magnitude of the electric potential is roughly proportional to the temperature differential, and the proportionality factor is termed the Seebeck coefficient. In practical applications of the Seebeck effect, semiconductors are used as the conducting media, and the circuit is formed by pairing n-type and p-type semiconductors and imposing a temperature gradient across them in parallel while electrically connecting them in series. The temperature gradient produces net charges of opposing polarities and hence a voltage between electπcally joined ends of the semiconductors.

[0004] Related to the discovery of Seebeck were the later discoveries of Peltier and Thomson. The Peltier effect, first reported by J.C. Peltier in Ann. Chim. Phys.(2d Ser.) 56: 371 (1834), is the reversible evolution or absorption of heat that occurs when an electric current is passed through a junction between dissimilar metals. The Thomson effect, first reported by W. Thomson in Philos. Trans. R. Soc: 1 (1855), is the reversible evolution or absorption of heat that occurs whenever an electric current traverses a region where a temperature gradient exists along a single homogeneous conductor. The Seebeck, Peltier, and Thomson effects have certain principles in common, but each utilizes these principles in a separate way and finds independent applications.

[0005] Of the three effects, the one that has the greatest economic potential but is the least developed for commercial use is the Seebeck effect. The economic potential of the Seebeck effect for generating thermoelectric power is considerable in view of the differences between thermoelectric power generation and conventional power generation. Unlike conventional power generation, for example, thermoelectric power does not require moving parts. Devices and generators of thermoelectric power thus offer greater reliability and service life by the avoidance of mechanical wear, plus a lower operating cost due to less need for maintenance. The lack of a need for operator attendance also permits thermoelectric devices to be used in hostile environments such as those at high temperatures. A still further advantage is the ability of these devices to make productive use of waste heat from industrial operations or of heat that is readily available from natural sources.

[0006] The efficiency of the conversion of thermal energy to electrical energy is a function of several variables and has two components, the Carnot efficiency, which is the ratio of the temperature differential to the absolute temperature at the hot junction, and a dimensionless parameter known as the thermoelectric figure of merit, which is represented by ZT. Here again, T is the absolute temperature at the hot junction, and Z = α²/(κp), where α is the Seebeck coefficient, K is the thermal conductivity, and p is the electrical resistivity. Achieving a thermoelectric generator of high efficiency, therefore, requires both a large temperature differential and a conducting medium that has a large Z value.

[0007] The search for thermoelectric materials of high efficiency first resulted in investigations of bismuth telluride and lead telluride. These were followed by the development of Si_xGeι-_x alloys, particularly for high-temperature use in space vehicles. Optimizations of these alloys were made by doping the alloys as n-type and p-type semiconductors and varying the doping levels and alloy compositions, these optimizations were unsuccessful however in obtaining ZT values greater than approximately 1, which is not high enough to support commercially viable applications.

[0008] The possibility of obtaining thermoelectric materials of greater efficiency arose with the development of nanotechnology and further by the discovery of carbon nanotubes in 1991. Nanostructured materials in general are promising candidates for thermoelectric materials in view of the quantum confinement effect of the charge carriers of nanostructures and the size effect of the heat carriers. The quantum confinement of the charge carriers has the potential of increasing the Seebeck coefficient and the electrical conductivity, while the enhanced boundary scattering due to the small size of the heat carriers has the potential of reducing the thermal conductivity. As the formula for Z above indicates, increasing the electrical conductivity (i.e., lowering the electrical resistivity) and reducing the thermal conductivity both result in an increase in the thermoelectric figure of merit. Among nanostructures, carbon nanotubes have generated special interest due to their unusual properties, including both their thermal and electrical behavior. While the results of several investigations have been published, these have been directed to individual carbon nanotubes or to "ropes," "mats" or films of carbon nanotubes, in attempts to understand the thermoelectric properties and behavior of these materials. Reports have been published by Hone, J., et al., "Thermoelectric power of single- walled carbon nanotubes," Phys. Rev. Lett. 80(5): 1042-1045 (1998), Bradley, K., et al., "Is the intrinsic thermoelectric power of carbon nanotubes positive?," Phys. Rev. Lett. 85(20): 4361-4364 (2000), Sumanasekera, G.U., et al., "Giant thermopower effects from molecular polyisosorption on carbon nanotubes," Phys. Rev. Lett. 89(16): 166801-1-4 (2002), Vavro, J., et al., "Thermoelectric power of -doped single-wall carbon nanotubes and the role of phonon draft," Phys. Rev. Lett. 90(6): 065503- 1-4 (2003), Choi, E.S., et al., "Magnetothermopower of single wall carbon nanotube network," Synthetic Metals 103:2504-2505 (1999), Small, J.P., et al., "Modulation of thermoelectric power of individual carbon nanotubes," Phys. Rev. Lett. 91(25): 256801-4 (2003), and Laguno, M.C., et al., "Observation of thermopower oscillations in the Coulomb blockade regime in a semiconducting carbon nanotube," Nano Letters 4(1): 45-49(2004). Despite these investigations, an effective way to utilize carbon nanotubes in a thermoelectric application has been elusive. [0009] Of further relevance to this invention is the literature on electric field-assisted sintering, also known as spark plasma sintering, plasma-activated sintering, and field-assisted sintering. This process is disclosed in the literature for use on metals and ceramics, for consolidating polymers, for joining metals, for crystal growth, and for promoting chemical reactions. The densification of alumina powder by spark plasma sintering is disclosed by Wang, S.W., et al., J. Mater. Res. 15(4) (April 2000): 982-987.

[0010] All citations appearing in this specification, including published papers, patents and Internet websites, are hereby incorporated herein by reference in their entirety for all purposes legally capable of being served thereby. SUMMARY OF THE INVENTION

[0011] It has now been discovered that composites of carbon nanotubes in a ceramic matrix when connected in a thermoelectric circuit function as thermoelectric materials with a high degree of thermoelectric efficiency owing to a high electrical conductivity and a low thermal conductivity. These qualities render these composites highly effective as materials for use in thermoelectric power generation under the Seebeck effect, and in the transport of thermal energy, either by absorption or evolution, under the Peltier and Thomson effects. It has also been discovered that these composites exhibit a Seebeck coefficient whose rate of increase with increasing temperature is approximately linear. A still further discovery is that composites of this invention, without further doping, function as p-type conductors. Accordingly, the present invention resides in methods of generating thermoelectric power, of transporting thermal energy by either heating or cooling, of both generating thermoelectric power and transporting thermal energy, and utilizing these materials as thermoelectric media and in thermoelectric devices in which these composites serve as thermoelectric conductors. Thus, in applications where an electric current is generated by a temperature differential in accordance with the Seebeck effect, a ceramic composite in accordance with this invention is used in place of the conductors of the prior art. In applications in which heat is drawn from or supplied to a surface or body in accordance with a device operating under the Peltier effect or the Thomson effect, a ceramic composite in accordance with this invention is used as the electrical and thermal energy transport medium. Furthermore, in applications that utilize the Seebeck effect and that contain p-type and n-type conductors that are paired and electrically and thermally connected in a manner that will result in thermoelectric power generation when temperature differentials are imposed, a ceramic composite in accordance with this invention serves as at least one of the two conductors, and, when only one, preferably the p-type conductor.

[0012] Prefened composites of this invention are those formed from nano-sized ceramic particles and from single-wall carbon nanotubes. It is further preferred that the composites are formed by densification of mixtures of these materials, particularly by spark plasma sintering to achieve densities approaching 100% theoretical density. In addition to their unique thermoelectric properties, these highly dense materials exhibit high fracture toughness and mechanical properties in general. These and other features, advantages and objects of this invention will be apparent from the description that follows.

DETAILED DESCRIPTION OF THE INVENTION AND PREFERRED EMBODIMENTS

[0013] Any of the wide variety of known ceramic materials can be used in the composites of this invention, although metal oxide ceramics are preferred. Examples of metal oxide ceramics are alumina, yttria, magnesium oxide, zirconia, magnesia spinel, titania, calcium aluminate, cerium oxide, chromium oxide, and hafnium oxide. Further examples are combinations that include non-metal oxides such as silica. Still further examples are metallic oxides that also contain elements in addition to metals and oxygen, such as SiAlON and A1ON. Combinations of any of these metal oxides can also be used. A metal oxide that is currently of particular interest is alumina, either α-alumina, γ-alumina, or a mixture of both. When the metal oxide includes alumina and the composite is formed by consolidating a powder mixture that includes γ-alumina powder, the γ-alumina will often convert to α-alumina during the consolidation.

[0014] Carbon nanotubes are polymers of pure carbon. Both single-wall and multi-wall carbon nanotubes can be used in the practice of this invention, although single- wall carbon nanotubes are preferred. Single-wall and multi-wall carbon nanotubes are known in the art and the subject of a considerable body of published literature. Examples of literature describing carbon nanotubes are Dresselhaus, M.S., et al., Science ofFullerenes and Carbon Nanotubes, Academic Press, San Diego (1996), and Ajayan, P.M., et al., "Nanometre- Size Tubes of Carbon," Rep. Prog. Phys. 60 (1997): 1025-1062. A single-wall carbon nanotube is a single graphene sheet rolled into a seamless cylinder with either open or closed ends. When closed, the ends are capped either by half fullerenes or by more complex structures such as pentagonal lattices. The average diameter of a single-wall carbon nanotube typically ranges of 0.5 to 100 nm, and most often, 0.5 to 10 nm, 0.5 to 5 nm, or 0.7 to 2 nm. The aspect ratio, i.e., length to diameter, typically ranges from about 25 to about 1,000,000, and most often from about 100 to about 1,000. A nanotube of 1 nm diameter may thus have a length of from about 100 to about 1,000 nm. Nanotubes frequently exist as "ropes," which are bundles of 3 to 500 single-wall nanotubes held together along their lengths by van der Waals forces. Individual nanotubes often branch off from a rope to join nanotubes of other ropes. Multi- walled carbon nanotubes are two or more concentric cylinders of graphene sheets of successively larger diameter, forming a layered composite tube bonded together by van der Waals forces, with a distance of approximately 0.34 nm between layers. Further descriptions of carbon nanotubes are given by Peigney, A., et al., "Carbon nanotubes in novel ceramic matrix nanocomposites," Ceram. Inter. 26 (2000) 677-683.

[0015] Carbon nanotubes can be prepared by arc discharge between carbon electrodes in an inert gas atmosphere. This process results in a mixture of single-wall and multi-wall nanotubes, although the formation of single-wall nanotubes can be favored by the use of transition metal catalysts such as iron or cobalt. Single-wall nanotubes can also be prepared by laser ablation, as disclosed by Thess, A., et al., "Crystalline Ropes of Metallic Carbon Nanotubes," Science 273 (1996): 483-487, and by Witanachi, S., et al., "Role of Temporal Delay in Dual-Laser Ablated Plumes," J. Vac. Sci. Technol. A 3 (1995): 1171-1174. A further method of producing single-wall nanotubes is the HiPco process disclosed by Nikolaev, P., et al., "Gas-phase catalytic growth of single- walled carbon nanotubes from carbon monoxide," Chem. Phys. Lett. 313, 91-97 (1999), and by Bronikowski M. J., et al., "Gas-phase production of carbon single-walled nanotubes from carbon monoxide via the HiPco process: A parametric study," J. Vac. Sci. Technol. 19, 1800-1805 (2001).

[0016] Preferred starting materials for the composites of this invention are powder mixtures of the ceramic material and the carbon nanotubes. The particle size of the ceramic material can vary while still achieving the benefits of the presence of the carbon nanotubes when consolidated and densified. The ceramic particles can thus be micron-sized, sub-micron- sized, or nano-sized. The term "micron-sized" refers to particles having diameters that are greater than 1 micron, "sub-micron-sized" refers to particles whose diameters are within the range of 100 nm to 1,000 nm, typically 150 nm or above, and "nano-sized" refers to particles whose diameters are less than 100 nm, particularly 50 nm or below. As the particle size of the starting ceramic material decreases, the final density of the densified composite product increases. Accordingly, particles that are less than 500 nm in diameter on the average are preferred, and those that are less than 100 nm in diameter are particularly preferred.

[0017] The carbon nanotubes may be single-wall nanotubes, multi-wall carbon nanotubes, or mixtures of single-wall and multi-wall carbon nanotubes. It is preferred however that the mixtures, and the final product as well, be free of all carbon nanotubes except single- wall carbon nanotubes, or if double- wall or multi-wall carbon nanotubes are present, that the carbon nanotubes in the mixture be predominantly single-wall carbon nanotubes. The term "predominantly" in this context is used herein to mean that either no multi-wall carbon nanotubes are present or that the amount of carbon nanotubes having more than a single wall is so small relative to the amount of single- wall nanotubes that the multi-wall nanotubes do not obliterate or significantly reduce the beneficial properties of the composite that are attributable to the single-wall nanotubes.

[0018] The relative amounts of ceramic material and carbon nanotubes can vary, although the mechanical properties and possibly the performance characteristics may vary with the proportions of the carbon nanotubes. In most cases, best results will be achieved with composites in which the carbon nanotubes constitute from about 1% to about 50%, preferably from about 2% to about 30%, and most preferably from about 3% to about 20%, by volume of the composite. The volumes used in determining the volume percents referred to herein are calculated from the weight percents of the bulk starting materials and their theoretical densities.

[0019] The carbon nanotubes can be dispersed through the ceramic powder by conventional means to form a homogeneously dispersed powder mixture, although a preferred method of preparing such a mixture is by suspending the materials together in a common liquid suspending medium. Any readily removable, low-viscosity, inert suspending liquid, such as a low molecular weight alcohol (ethanol, for example), can be used. Carbon nanotubes are available from commercial suppliers in a paper-like form, and can be dispersed in ethanol and other liquid suspending agents with the assistance of ultrasound.

[0020] Once the components are combined, a preferred means of mixing to achieve a uniform dispersion is ball-milling using conventional rotary mills with the assistance of tumbling balls. The sizes of the balls, the number of balls used per unit volume of powder, the rotation speed of the mill, the temperature at which the milling is performed, and the length of time that milling is continued can all vary widely. Best results will generally be achieved with a milling time ranging from about 4 hours to about 50 hours. The degree of mixing may also be affected by the "charge ratio," which is the ratio of the mass of the balls to the mass of the powder. A charge ratio of from about 20 to about 100 will generally provide proper mixing.

[0021 ] The qualities of the product can be enhanced by mechanical activation of the ceramic particles prior to consolidating them into a composite. Mechanical activation is likewise achieved by methods known in the art and is typically performed in centrifugal, oscillating or planetary mills that apply centrifugal, oscillating and/or planetary action to the powder mixture with the assistance of grinding balls. The grinding balls produce impacts of up to 20 g (20 times the acceleration due to gravity). Variables such as the sizes of the milling balls, the number of milling balls used per unit amount of powder, the temperature at which the milling is performed, the length of time that milling is continued, and the energy level of the mill such as the rotational speed or the frequency of impacts, can vary widely. The number and size of the milling balls relative to the amount of powder is typically expressed as the "charge ratio," which is defined as the ratio of the mass of the milling balls to the mass of the powder. A charge ratio of at least about 1, preferably about 1 to about 10, and most preferably about 2 to about 5, will generally provide the best results. Preferred milling frequencies are at least about 3, and preferably about 3 to 30 cycles per second or, assuming two impacts per cycle, at least about 6, and preferably about 6 to about 60 impacts per second.

[0022] Consolidation of the powder mixture into a continuous mass is preferably performed by uniaxial compression. The composite will further benefit if densified to a high density as it is being consolidated. Optimal densities are those that approach full theoretical density, which is the density of the material as determined by volume-averaging the densities of each of its components. The term "relative density" is used herein to denote the actual density expressed as a percent of the theoretical density. Preferred composites thus have relative densities of 90% or above, more preferably at least 95%, still more preferably at least 98%, and most preferably at least 99%.

[0023] Uniaxial compression is preferably performed in combination with electric field- assisted sintering. One method of performing this type of sintering is by passing a pulsewise DC electric cunent through the dry powder mixture or through a consolidated mass of the mixture while applying pressure. A description of electric field-assisted sintering and of apparatus in which this process can be performed is presented by Wang, S.W., et al., referenced above. While the conditions may vary, best results will generally be obtained with a densification pressure exceeding 10 MPa, preferably from about 10 MPa to about 200 MPa, and most preferably from about 40 MPa to about 100 MPa. The prefened current is a pulsed DC cunent of from about 250 A/cm² to about 10,000 A/cm², most preferably from about 500 A/cm² to about 2,500 A/cm². The duration of the pulsed cunent will generally range from about 1 minute to about 30 minutes, and preferably from about 1.5 minutes to about 5 minutes. Prefened temperatures are within the range of from about 800°C to about 1,500°C, and most preferably from about 900°C to about 1,400°C. The compression and sintering are preferably performed under vacuum. Prefened vacuum levels for the densification are below 10 Ton, and most preferably below 1 Ton.

[0024] Thermoelectric devices and power generators in accordance with this invention can assume configurations cunently known in the art except for the substitution of the composites of the present invention for the conducting media. The composites can thus be doped with conventional doping materials to form n-type or p-type semiconductors, and n-type and p-type semiconductors can be paired and connected electrically and thermally to serve the needs of the device, which is either to generate power from a temperature gradient or to heat or cool due to the imposition of an electric potential. For power generation applications, the n-type and p-type semiconductors will be thermally connected in parallel together with externally applied temperature gradient across both, and electrically connected in series to form an electric circuit. Examples of such configurations are shown in Fleurial et al., United States Patent No. US 6,660,926 B2 (General Motors Corporation; California Institute of Technology), December 9, 2003. For heating or cooling applications, thermal and electrical connections will be made in the same general manner, except with an external voltage source to impose a voltage across the semiconductors. Examples of such configurations are shown or described in Zimmerman, United States Patent No. 3,248,889, (North American Philips Company), May 3, 1966; Nagakubo et al., United States Patent No. 5,515,682 (Fujitsu Limited), May 14, 1996; Shimizu, United States Patent No. 5,960,142 (NEC Corporation), September 28, 1999; and Strnad, United States Patent No. US 6,399,872 B2 (National Semiconductor Corporation), June 4, 2002. The composites of the present invention can serve as one or both of the semiconductors, appropriately doped, or in the case where the composite serves as the p-type semiconductor, the composite can serve this function without doping in view of its spontaneous p-type behavior.

[0025] The following examples are offered for purposes of illustration and are not intended to limit the scope of the invention. EXAMPLES

[0026] Gas-condensed synthesized γ-alumina powder with average particle sizes of 15 nm and nanocrystalline 3Y-TZP (3 mol% yttria-stabilized tetragonal zirconia polycrystals) with an average particle size of 28 nm were obtained from Nanophase Technologies Corporation (Darien, illinois, USA). Purified single-wall carbon nanotubes, produced by the HiPco process with more than 90% of the catalyst particles removed, were obtained from Carbon Nanotechnologies (Houston, Texas, USA). The carbon nanotubes were dispersed in ethanol and the assistance of ultrasound. The alumina nanopowder was then combined the 3Y-TZP nanopowder and the mixture was mixed by ball-milling for 24 hours using zirconia milling media. The ball-milled mixture was then added to the nanotube dispersion, and the combined dispersion was sieved with a 200-mesh sieve, then ball-milled for 24 hours (in ethanol) using zirconia milling media, then dried to form a dry powder mixture. The proportions used were such that the final mixture consisted of 10% single-wall carbon nanotubes, 20% 3Y-TZP, and 70% alumina, all by volume.

[0027] Thus formed, the powder mixture was placed on a graphite die of inner diameter 19 mm and cold-pressed at 200 MPa. The cold-pressed powder mixture was then sintered on a Dr. Sinter 1050 Spark Plasma Sintering System (Sumitomo Coal Mining Company, Japan) under vacuum. Electric field-assisted sintering was then performed at an applied pressure of 80 MPa with a pulsed DC cunent of about 5,000 A maximum and a maximum voltage of 10 V. The pulses had a period of about 3 ms and followed a pattern of 12 cycles on and 2 cycles off. Once the pressure was applied, the samples were heated to 600°C in 2 minutes and then heated further at rates of 550°C/min to 600°C/min to 1,150 °C where they were held for 3 minutes. The temperature was monitored with an optical pyrometer focused on a depression in the graphite die measuring 2 mm in diameter and 5 mm in depth.

[0028] The final densities of the sintered compacts were measured by the Archimedes method using deionized water as the immersion medium. The density of the carbon nanotubes used as a starting material was 1.8 g/cm'^*. Fracture toughness measurements were performed on a Wilson Tukon hardness tester with a diamond Vickers indenter. Bulk specimens were sectioned and mounted in epoxy, then polished through 0.25 micron diamond paste. The indentation parameters for fracture toughness (Kι ) were a 2.5 kg load with a dwell time of 15 seconds. The fracture toughness was calculated by the Anstis equation as disclosed by Anstis, G.R., et al.„ "A critical evaluation of indentation techniques for measuring fracture toughness: I. Direct Crack Measurement," J. Am. Chem. Soc. 64(9): 533- 538 (1981).

[0029] Specimens for thermoelectric measurements were prepared by cutting the sintered compacts into rectangular bars whose dimensions were 14 mm x 5 mm x 1-3 mm. The heads of two Pt-Ptl3%Rh thermocouples were embedded at the two ends respectively of each specimen and fixed with platinum wires. Thermoelectric properties of the specimens were then measured by use of an automatic thermoelectric measuring apparatus (RZ2001K, Ozawa Scientific Co., Japan). Electrical conductivity measurements were performed at 373-673 K by a DC four-probe technique in air. For thermopower measurement, the temperature gradient in the specimen was generated by passing cool air in an alumina protection tube placed near one end of the specimen. The temperature difference between the two ends was controlled at 2-15 K by varying the flow rate of air. Measurements of thermopower as a function of temperature gradient indicated a linear dependence and the Seebeck coefficient were calculated from the slope of the thermopower-temperature line. The values thus determined were as follows:

Temperature Coefficient 345 K 28.5 μV/K 440 K 34.8 μV/K 542 K 42.3 μV/K 644 K 50.4 μV/K

[0030] The dependence of the Seebeck coefficient on temperature indicated by the data in this table is approximately linear. The Seebeck coefficient at the lowest temperature (28.5 μV/K) is comparable to that of aligned ropes of single-wall carbon nanotubes (approximately 27 μK/V at 300 K), as reported by Hone, J., et al., Appl. Phys. Lett. 77(5): 666-8 (2000). The fracture toughness of the material was 8.5 MPam" , and the room- temperature conductivity was 3,638 S/m.

[0031] The foregoing is offered primarily for purposes of illustration and explanation. Further variations, modifications and substitutions that, even though not disclosed herein, still fall within the scope of the invention may readily occur to those skilled in the art.

Claims

WHAT IS CLAIMED IS: 1. A method for generating thermoelectric energy, said method comprising imposing a temperature differential across an electrically conducting element in an electric circuit, in which said element is a composite comprising carbon nanotubes dispersed in a matrix of ceramic material, thereby causing electric cunent resulting from said temperature differential to flow through a load in said circuit.

2. A method for transporting heat between first and second media, said method comprising passing an electric cunent between first and second sites on a thermoelectric element disposed between said first and second media, said first and second sites in thermal contact with said first and second media, respectively, said thermoelectric element being a composite comprising carbon nanotubes dispersed in a matrix of ceramic material.

3. The method of claims 1 or 2 wherein said carbon nanotubes are predominantly single-wall carbon nanotubes.

4. The method of claims 1 or 2 wherein said carbon nanotubes consist essentially of single-wall carbon nanotubes.

5. The method of claims 1 or 2 wherein said ceramic material is a metal oxide selected from the group consisting of alumina, yttria, zirconia, magnesium oxide, magnesia spinel, zirconia, titania, cerium oxide, chromium oxide, hafnium oxide, and combinations thereof.

6. The method of claims 1 or 2 wherein said ceramic material is a metal oxide selected from the group consisting of alumina, yttria, zirconia, and combinations of alumina, yttria, and zirconia.

7. The method of claims 1 or 2 wherein said carbon nanotubes constitute from about 1% to about 50% of said composite by volume.

8. The method of claims 1 or 2 wherein said carbon nanotubes constitute from about 2% to about 30% of said composite by volume.

9. The method of claims 1 or 2 wherein said carbon nanotubes constitute from about 3% to about 20% of said composite by volume.

10. The method of claims 1 or 2 wherein said composite has a density of at least 90% relative to a volume-averaged theoretical density.

11. The method of claims 1 or 2 wherein said composite has a density of at least 95% relative to a volume-averaged theoretical density.

12. The method of claims 1 or 2 wherein said composite has a density of at least 99% relative to a volume-averaged theoretical density.

13. The method of claims 1 or 2 wherein said composite is the product of a process comprising consolidating a mixture of ceramic particles of less than 500 nm in diameter and carbon nanotubes into a continuous mass by uniaxially compressing said mixture while passing a pulsed electric cunent through said mixture.

14. The method of claims 1 or 2 wherein said process comprises uniaxially compressing said mixture at a pressure of from about 10 MPa to about 200 MPa and temperature of from about 800°C to about 1,500°C, and said sintering electric cunent is a pulsed direct cunent of from about 250 A/cm to about 10,000 A/cm .

15. In an application requiring the conduction of an electric cunent as the result of a temperature differential across a thermoelectric material, the improvement in which said thermoelectric material is a composite comprising carbon nanotubes dispersed in a matrix of ceramic material.

16. The improvement of claim 15 wherein said carbon nanotubes are predominantly single-wall carbon nanotubes.

17. The improvement of claim 15 wherein said carbon nanotubes consist essentially of single-wall carbon nanotubes.

18. The improvement of claim 15 wherein said ceramic material is a metal oxide selected from the group consisting of alumina, yttria, zirconia, magnesium oxide, magnesia spinel, zirconia, titania, cerium oxide, chromium oxide, hafnium oxide, and combinations thereof.

19. The improvement of claim 15 wherein said ceramic material is a metal oxide selected from the group consisting of alumina, yttria, zirconia, and combinations of alumina, yttria, and zirconia.

20. The improvement of claim 15 wherein said carbon nanotubes constitute from about 1% to about 50% of said composite by volume.

21. The improvement of claim 15 wherein said carbon nanotubes constitute from about 2% to about 30% of said composite by volume.

22. The improvement of claim 15 wherein said carbon nanotubes constitute from about 3% to about 20% of said composite by volume.

23. The improvement of claim 15 wherein said composite has a density of at least 90% relative to a volume-averaged theoretical density.

24. The improvement of claim 15 wherein said composite has a density of at least 95% relative to a volume-averaged theoretical density.

25. The improvement of claim 15 wherein said composite has a density of at least 99% relative to a volume-averaged theoretical density.

26. The improvement of claim 15 wherein said composite is the product of a process comprising consolidating a mixture of ceramic particles of less than 500 nm in diameter and carbon nanotubes into a continuous mass by uniaxially compressing said mixture while passing a pulsed electric cunent through said mixture.

27. The improvement of claim 15 wherein said process comprises uniaxially compressing said mixture at a pressure of from about 10 MPa to about 200 MPa and temperature of from about 800°C to about 1,500°C, and said sintering electric cunent is a pulsed direct cunent of from about 250 A/cm² to about 10,000 A/cm².

28. The improvement of claim 15 wherein said application is a thermoelectric device comprising p-type and n-type thermoelectric elements joined by an electric lead and means for imposing temperature differentials across each of said thermoelectric elements to cause cunent to flow through said electric lead, and said composite is at least one of said thermoelectric elements.