Method of producing thermoelectric material

FIELD OF THE INVENTION

This invention relates to materials having thermoelectric properties and to thermoelectric devices.

BACKGROUND OF THE INVENTION

A thermoelectric device can be used to obtain electrical energy from a thermal gradient (for example, a thermoelectric generator using the Seebeck effect), or to generate a thermal gradient from electrical energy (for example, a thermoelectric refrigerator using the Peltier effect). The discussion below is directed to the Seebeck effect, but the general concepts also apply to applications of the Peltier effect.

A typical thermoelectric device is built up from several unicouples, which are typically pairs of thermally conductive p-type (P) and n-type (N) semiconductors. These unicouples are connected electrically in series and thermally in parallel. Theoretically, the maximum efficiency of the conversion of heat energy to electrical energy is given by:

$ζ_{\max} = \frac{(T_{H} - T_{C})}{T_{H}} \frac{\sqrt{1 + {ZT}_{ave}} - 1}{\sqrt{1 + {ZT}_{ave}} + T_{C} / T_{H}}$

where T_ave=(T_H+T_C)/2 is the average temperature of thermal gradient having a hot temperature (T_H) end and a cold temperature (T_C) end, and Z is a figure of merit, defined as Z=S²σ/κ. The figure of merit Z depends on the macroscopic transport parameters of the materials, namely the Seebeck coefficient (S), electrical conductivity (σ), and thermal conductivity (κ). A large figure of merit is provided by a thermoelectric material having a large Seebeck coefficient, high electrical conductivity, and low thermal conductivity.

The Seebeck coefficient is further defined as the ratio of the open-circuit voltage to the temperature difference between the hot and cold junctions of a circuit exhibiting the Seebeck effect, or S=V/(T_H−T_C). Since Z varies with temperature, a useful dimensionless figure of merit can be defined as ZT.

By the end of the 1950s, the best bulk thermoelectric materials were found to be alloys of bismuth telluride and antimony, which gave a room temperature ZT˜1. Workers in the thermoelectric field have been attempting to improve the figure of merit over the past 40 years without much success. Increasing ZT is difficult because the three parameters S, σ, and k are all related to the free carrier concentration and are usually not independent. For example, doping typically increases the semiconductor's electrical conductivity, but decreases its Seebeck coefficient and increases the thermal conductivity. Efforts to reduce the lattice thermal conductivity by alloying also reduce the electrical conductivity by providing an extra scattering mechanism.

Dresselhaus and coworkers at MIT theoretically demonstrated that quantum confinement of electrons and phonons within nanowires of a thermoelectric material can increase the value of ZT. 1-D nanowires in particular could reach ZT≈2-5 if the nanowire diameter lies in the range of 5-10 nanometers. Certain structures have been investigated, for example such as described in Heremans, J. P. et al., “Thermoelectric Power of Bismuth Nanocomposites”; Phys. Rev. Lett.; 2002, 88, 216801; Venkatasubramanian, R. et al., “Thin-film thermoelectric devices with high room temperature figures of merit”; Nature; 2001, 413, 597-602; Harman, T. C. et al., “Thermoelectric quantum dot superlattices with high ZT”; Electron. Mater.; 2000, 29, L1-L4; Rabin, O. et al., “Anomalously high thermoelectric figure of merit in Bi_1-xSb_xnanowires by carrier pocket alignment”; APL; 2001, 79, 81-83; and Dresselhaus, M. S. et al., “Low-dimensional thermoelectric materials”; PSS; 1999, 41, 679-682. However, these approaches do not provide a simple approach to making large-scale, low-cost thermoelectric devices. Conventional semiconductor device fabrication methods are unsuitable for manufacturing bulk samples, and are often expensive.

In automobiles, about 70 percent of energy derived from fuel is lost to waste heat and engine cooling. Only a small proportion of energy provided by fuel combustion is used, and a large amount of thermal energy is thrown away. Recovery of waste thermal energy is a big challenge in automotive industries due to the increasing energy crisis. Thermoelectric conversion of thermal energy to electrical energy could be an effective way to obtain electrical energy from otherwise wasted heat production. However, direct thermal to electric conversion (DTEC) technology currently faces two major challenges: low conversion efficiency and insufficient power density. Hence, improved materials and devices having high thermoelectric conversion efficiency are urgently required.

In response to the need for high thermoelectric conversion efficiency materials, Zhang et al. have investigated thermoelectric materials comprising two or more components, at least one of which is a thermoelectric material (U.S. Pat. No. 7,309,830). However, a given thermoelectric material system can have a wide range of compositions that may, or may not, exhibit high ZT values, and as such, Banerjee et al. have developed a process for determining an optimum range of compositions for a nanocomposite thermoelectric material system (U.S. Pat. No. 7,734,428).

In addition to the above, other factors such as grain size and grain boundary properties have been postulated to affect the properties of thermoelectric materials. However, as of yet no process has been developed to determine if there is and/or which optimum range of such factors can provide a thermoelectric material with an improved ZT. Therefore, a process to model, calculate and/or determine an optimum range of grain related properties in which a thermoelectric material exhibits high ZT values would be desirable.

SUMMARY OF THE INVENTION

A process for manufacturing a thermoelectric material having a plurality of grains and grain boundaries is provided. The process includes determining a material composition to be investigated for the thermoelectric material and then determining a range of values of grain size and/or grain boundary barrier height obtainable for the material composition using current state of the art manufacturing techniques. Thereafter, a plurality of Seebeck coefficients for the material composition as a function of the range of values for the grain size and/or grain boundary barrier height are calculated. In addition, a plurality of electrical resistivity values and a plurality of thermal conductivity values for the material composition and as a function of the range of values for the grain size and/or grain boundary barrier height are calculated.

Once such plurality of values have been determined, a range of figure of merit values for the material composition as a function of the calculated Seebeck coefficients, calculated electrical resistivity values, and calculated thermal conductivity values are determined/calculated. Based on the range of figure of merit values, a generally maximum range thereof for the material composition is determined, a thermoelectric material having the determined material composition and an average grain size and grain boundary barrier height corresponding to the maximum range of figure of merit values is manufactured.

The material composition of the thermoelectric material can be a bulk thermoelectric material composition, or in the alternative, a nanocomposite thermoelectric material composition. The range of values of grain size can be between 5 and 100 nanometers while the range of values of grain boundary barrier height can be between 10 and 300 milli-electron volts.

In some instances, the nanocomposite thermoelectric material can have a first matrix phase, an inter-grain phonon scattering second phase and a plurality of third phase grain boundaries. The first matrix phase has an average grain size within a range of 5-100 nm, preferably within 5-50 nm and more preferably within 5-25 nm. The inter-grain scattering second phase can be a plurality of oxide nanoparticles that have an average diameter within a range of 2-100 nm, preferably 4-50 nm and more preferably 6-14 nm. Also, the plurality of third phase grain boundaries have an average width within a range of 2-75 nm, preferably 5-70 nm, more preferably 10-70 nm, even more preferably 15-65 nm and still yet more preferably 27-61

The grain size of the manufactured thermoelectric material can be obtained by consolidating a plurality of nanoparticles having a mean diameter generally equal to less than the grain size and the grain boundary barrier height can be obtained by doping of the thermoelectric material, altering a surface of a plurality of nanoparticles used to manufacture the thermooelectric material, and the like.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGS. 1a-1d are schematic illustrations of: (a) the grain structure of a prior art thermoelectric material; (b) a schematic illustration of a thermoelectric material having altered grain boundaries according to an embodiment of the present invention; (c) a schematic illustration of a prior art nanocomposite thermoelectric material; and (d) a schematic illustration of a nanocomposite thermoelectric material having modified grain boundaries according to an embodiment of the present invention;

FIG. 2 is a graph illustrating calculated electron mean free path, electron wavelength, and carrier percentage occupation as a function of dimensionless electron energy;

FIG. 3 is a schematic illustration of a model for treatment of grain boundary barrier height according to an embodiment of the present invention;

FIG. 4 is a graph illustrating effect of grain boundary properties on electrical conductivity;

FIG. 5 is a graph illustrating effect of grain size on lattice (Kl) and electronic (Ke) thermal conductivity;

FIG. 6 is a graph illustrating effect of grain boundary properties on total thermal conductivity;

FIG. 7 is a graph illustrating effect of grain boundary properties on Seebeck coefficient;

FIG. 8 is a graph illustrating calculated normalized Seebeck coefficient as a function of electron energy;

FIG. 9 is a graph illustrating effect of grain boundary properties on dimensionless figure of merit (ZT);

FIG. 10 is a graph illustrating Seebeck coefficient for a bulk (Bulk) and nanocomposite (NC) thermoelectric material as a function of grain boundary barrier height;

FIG. 11 is a graph illustrating electrical conductivity and electronic thermal conductivity for a bulk (Bulk) and nanocomposite (NC) thermoelectric material as a function of grain boundary barrier height;

FIG. 12 is a graph illustrating ZT as a function of grain boundary barrier height;

FIG. 13 is a flowchart for a process according to an embodiment of the present invention;

FIG. 14 is a flowchart for a process according to an embodiment of the present invention;

FIG. 15 is a flowchart for a process according to an embodiment of the present invention;

FIG. 16 is a flowchart for a process according to an embodiment of the present invention;

FIG. 17 is a flowchart for a process according to an embodiment of the present invention;

FIG. 18 is a schematic illustration of a computer with a processor according to an embodiment of the present invention;

FIG. 19 is a series of graphs illustrating: (A) figure of merit ZT versus temperature for a bismuth antimony telluride matrix with zinc oxide nanoparticles and zinc antimony modified grain boundaries (BAZT—filled square data points) and a bismuth antimony telluride matrix with no zinc oxide nanoparticles or modified grain boundaries (BAT—open square data points); (B) electrical conductivity versus temperature for the BAZT and BAT materials; (C) Seebeck coefficient versus temperature for the BAZT and BAT materials; (D) power factor versus temperature for the BAZT and BAT materials; and (E) thermal conductivity versus temperature for the BAZT and BAT materials;

FIG. 20 is a transmission electron microscopy (TEM) image illustrating: (A) a bismuth antinomy telluride matrix with zinc antinomy modified grain boundaries; and (B) zinc oxide nanoparticles within the bismuth antinomy telluride matrix; and

FIG. 21 is a graph illustrating the x-ray diffraction (XRD) results of the BATZ material.

DETAILED DESCRIPTION OF THE INVENTION

The present invention discloses a process for determining an optimum range of compositions for a thermoelectric material system, within which the material system may exhibit generally high figure of merit values. As such, the process has utility for improving the efficiency of experimental design and production of thermoelectric materials.

The process for determining an optimum range of compositions for a thermoelectric material system considers a variety of relevant factors, parameters and the like in order to determine which material systems should be considered and/or which range of compositions should be studied in more detail. A thermoelectric material exhibiting a dimensionless high figure of merit (ZT) needs to possess a high Seebeck coefficient (S) for high voltage generation, a low electrical resistivity (ρ) to minimize Ohmic losses and a low thermal conductivity (κ) to minimize heat conduction.

The relationship between ZT, S, ρ, and κ can be expressed as:

ZT=S²T/κρ Eqn 1

and/or as:

$\begin{matrix} ZT = \frac{S^{2} T}{ρ (k_{el} + k_{ph})} & Eqn 2 \end{matrix}$

where κ_eland κ_phare the electronic and phonon contribution to the overall thermal conductivity k.

Typically, S, ρ, and κ are interdependent with an increase of the Seebeck coefficient resulting in an increase in electrical resistivity, whereas a decrease in the electrical resistivity results in an increase of the thermal conductivity. At least one approach for obtaining high figure of merit values has investigated the insertion of nanoparticles within a thermoelectric material (U.S. Pat. No. 7,309,830). Materials using this approach can result in phonons being scattered by the nanoparticles, thereby reducing the lattice thermal conductivity while leaving the electrical resistivity and Seebeck coefficient for the thermoelectric host matrix unchanged.

Elemental substitutions, also known as atomic substitutions, in potential thermoelectric materials have imperfections on the order of 1 angstrom (Å). Thus alloying additions can result in the scattering of short-wavelength phonons much more effectively than mid- and long-wavelength phonons. Therefore, mid- and long-wavelength phonons dominate the heat conduction in alloys and thermoelectric materials that have been doped with other elements not originally within the starting material. In the alternative, the inclusion of additions such as nanoparticles in the size range of phonon wavelengths introduces another scattering mechanism that affects mid- and/or long-wavelength phonons, thereby providing an opportunity to reduce the thermal conductivity of such materials below the alloy limit. However, which nanoparticles with respect to their composition, size and size distribution, and which host matrix the nanoparticles should be added to has heretofore been a difficult task to predict. In response to the difficulty in predicting successful thermoelectric material systems, a process to perform just this task has been developed by Banerjee et al. (U.S. Pat. No. 7,734,428).

An embodiment of the current process includes determining a material composition to be investigated for the thermoelectric material and determining a range of values for a grain related property that is obtainable for the material composition using state of the art manufacturing techniques. Once the material composition and the range of values for the grain related property have been determined, a plurality of Seebeck coefficients for the material composition as a function of the range of values can be calculated. In addition, a plurality of electrical resistivity values and a plurality of thermal conductivity values for the material composition as a function of the range of values for the grain related property can also be calculated.

It is appreciated that once the plurality of Seebeck coefficients, electrical resistivity values, and thermal conductivity values have been determined, a range of figure of merit values as a function thereof can be calculated and a generally maximum range of figure of merit values can be determined, such values being a function of the range of values of the grain related property. Naturally, once the maximum range of figure of merit values has been determined, a thermoelectric material having the determined material composition and the grain related property(ies) corresponding to the maximum range of figure of merit values is manufactured.

In the alternative to the above outlined embodiment, a plurality of material compositions can be investigated in this manner and a single material composition or a limited range of compositions having a potential and desired ZT are manufactured.

The grain related property can include any grain related property known to those skilled in the art, illustratively including grain size, grain boundary barrier height, and the like. For the purposes of the present invention, the term “grain size” is defined as the average mean diameter of grains within a thermoelectric material obtained through any method and/or technique known to those skilled in the art. For example and for illustrative purposes only, a grain size can be determined by taking a statistical average of a plurality of grain diameters from a metallographic cross-section of the material with a single grain diameter obtained by averaging at least two linear and orthogonal measurements across a given grain.

Also for the purposes of the present invention, the term “grain boundary barrier height” is defined as the energy potential of a grain boundary that will scatter an electron having less energy than the potential and allow an electron having more energy than the potential to pass therethrough.

The material composition to be investigated can be a bulk thermoelectric material composition, or in the alternative, a nanocomposite thermoelectric material composition. It is appreciated that the term “bulk thermoelectric material” refers to a polycrystalline material without the presence of second phase particles such as nanoparticles of an insulating type material. In the alternative, the term “nanocomposite thermoelectric material” refers to a bulk thermoelectric material having second phase particles such as nanoparticle insulating material inclusions, e.g. nanoparticle inclusions such as silicon oxide, zinc oxide, and the like.

The range of values for grain size of the material composition to be investigated can be between 5 and 100 nanometers (nm) while the range of values of grain boundary barrier height can be between 10 and 300 milli-electron volts (meV). In addition, the grain size of the manufactured thermoelectric material can be obtained by consolidating a plurality of nanoparticles having a mean diameter that is less than or generally equal to the final grain size of the material. The grain boundary barrier height of the manufactured thermoelectric material can be altered and/or obtained by doping of the material and/or altering a surface of the plurality of nanoparticles used to manufacture the thermoelectric material. In some instances, the surface of the plurality of nanoparticles is altered by applying a coating thereon before the nanoparticles are consolidated to produce the thermoelectric material.

Not being bound by theory, it is appreciated that a grain boundary is a result of and/or forms from a crystallographic misalignment between adjacent grains. In addition, the misalignment results in a residual electric charge across the grain boundary which can produce an electrostatic potential commonly referred to as an interfacial barrier and/or grain boundary barrier height which can be measured using AC impedance. As a first approximation, the magnitude of this interfacial barrier, also known as the grain boundary barrier height, can be calculated from the expression:

$\begin{matrix} E_{b} = \frac{{eN}_{t}^{2}}{8 ɛ N_{D}} & Eqn 3 \end{matrix}$

where N_tis the number density of traps, ε is the permittivity and N_Dis the doping concentration. The trap density is generally unknown and can vary widely, however assuming a generally high doping level and reported values for N_tin the range of 10⁻¹¹-10⁻¹³cm⁻³, an E_bof around 25 meV can be calculated.

The process can provide a thermoelectric material as schematically illustrated in FIGS. 1a and 1d. In particular, FIG. 1b illustrates a bulk thermoelectric material 100′ having altered and/or engineered grain boundaries 110′ when compared to the grain boundaries 110 of the material 100 illustrated in FIG. 1a. In addition, FIG. 1d illustrates a nanocomposite thermoelectric material 200′ having altered and/or engineered grain boundaries 210′ in addition to nanoparticle inclusions 205 when compared to the material 200 having the nanoparticles 205 and grain boundaries 210.

The grain size of the thermoelectric material 100 and/or 200 can be altered and/or engineered, e.g. by using nanoparticles with a desired average size to consolidate and manufacture the material. In addition, local electronic and thermal properties of the grain boundaries can be altered and/or engineered by controlling the interfacial composition between the grains, that is the interfacial composition of the grain boundaries. For example and for illustrative purposes only, a second phase can be engineered to be present at the interface between the grains such as Pb_0.75Sn_0.25Se coatings on Pb_0.75Sn_0.25Te; CoSb₃coatings on La_0.9CoFe₃Sb₁₂; and alkali-metal salt coatings on (Bi_0.2Sb_0.8)₂Te₃. In fact, results from CoSb₃/La_0.9CoFe₃Sb₁₂and coated (Bi_0.2Sb_0.8)₂Te₃materials have shown moderate improvements in the figure of merit ranging from 15-30%.

In order to incorporate a grain related property into a modeling and/or manufacturing process, the scattering behavior of electrons, holes and/or phonons within a material can be useful. Not being bound by theory, a theoretical simulation can be based on the Boltzmann equation with relaxation time approximation. For example, a modified Callaway model with respect to the lattice of a thermoelectric material can be incorporated with scattering of phonons through grain boundaries, defects, nanoparticles, and the like provided by Equation 4 below:

τ_c⁻¹=τ_B⁻¹+τ_U⁻¹+τ_N⁻¹+τ_A⁻¹+τ_D⁻ Eqn 4

where τ corresponds to scattering time and the subscripts B, U, N, A and D correspond to boundary, Umpklamp, normal, alloy, and nanoparticle, respectively, related scattering.

With respect to carriers, that is electrons and holes, Equation 5 can be used where Op, DOp, and DAp represent optical phonon, deformation potential of optical phonon, and deformation potential of acoustic phonon related scattering.

τ_ξ⁻¹=τ_Op⁻¹+τ_DOp⁻¹+τ_DAp⁻¹ Eqn 5

In addition to scattering time, the total electrical conductivity can be expressed as a summation of the contributions from both electron and hole bands, while the overall Seebeck coefficient can be obtained through weighting each band's contribution using a normalized electrical conductivity. In order to obtain the electronic thermal conductivity, the electronic thermal conductivity from the Lorentz number (L) can be obtained using Equations 6-8 below. In particular, Equation 6 is an expression of the total electrical conductivity (σ), Equation 7 is an expression of the overall Seebeck coefficient, and Equation 8 is an expression for the electronic thermal conductivity. It is appreciated that the bipolar thermal conductivity contribution to the electronic thermal conductivity must also be considered and that this type of conduction occurs when carriers moving between different bands carry heat via the Peltier effect and as such can still transport heat even if the net electric current is zero.

$\begin{matrix} σ = \sum_{i}^{e, h} σ_{i} & Eqn 6 \\ S = \sum_{i}^{e, h} \frac{S_{i} σ_{i}}{σ} & Eqn 7 \\ k_{e} = {(\frac{k_{B}}{e})}^{2} (\sum_{i}^{e, h} L_{i} + L_{b}) σ T & Eqn 8 \end{matrix}$

In addition to the above, the nature of grain boundary scattering exhibited by carriers can be estimated from the electron wavelength and electron mean free path (MFP) and the cumulative distribution function of the electron occupation number versus electron energy can provide the percentage of electrons that have energy less than a certain value. In particular, Equations 9-13 afford for the electron MFP, electron wavelength, and carrier percentage occupation as a function of dimensionless electron energy shown in FIG. 2 where the electron MFP can be calculated using the expression 1=ντ where ν and τ are provided by Equations 11 and 4, respectively, g is the density of state function and a is equal to the inverse of the bandgap for the material (1/E_g), in some instances referred to as the parabolicity factor.

$\begin{matrix} g = \frac{\int_{0}^{E / k_{b} T} g (E) d E}{\int_{0}^{\infty} g (E) d E} & Eqn 9 \\ g (E) = v^{2} D (E) (- \frac{\partial f}{\partial E}) & Eqn 10 \\ v = \frac{{(2 E (1 + α E) / m_{c}^{*})}^{0.5}}{(1 + 2 α E)} & Eqn 11 \\ D (E) = \frac{\sqrt{2} {(m_{d}^{*})}^{1.5}}{π^{2} ℏ^{3}} E (1 + 2 α E) (1 + α E) & Eqn 12 \\ λ = \frac{2 π ℏ}{m_{c}^{*} v} = \frac{2 π ℏ (1 + 2 α E)}{\sqrt{2 m_{c}^{*} E (1 + α E)}} & Eqn 13 \end{matrix}$

As shown in FIG. 2, the majority of electrons have a MFP less than 30 nanometers which is comparable and/or of the same order of magnitude of a grain size of between 20 to a few hundred nanometers. As such, FIG. 2 confirms that a majority of carriers will experience multiple scattering due to the grain boundaries and “memory” of a previous collision by a carrier will be retained by the carrier when it reaches another portion of the grain interface, i.e. each scattering point on a grain boundary is not independent from others. In addition, since the dominant electron wavelength is approximately 28 nm and an assumed grain boundary width of 1 to 2 nm is much smaller, there is no possibility of diffuse scattering of electrons. Finally, the electron MFP of less than 30 nanometers is comparable to a hole wavelength of 28 nm which implies that the Boltzmann equation is at the edge of its validity using this process. As such, it is appreciated that other expressions with respect to relaxation time approximation can be used for the basis of a theoretical simulation that incorporates grain boundary effect in determining various properties of a thermoelectric material and still fall within the scope of the present invention.

Turning now to the actual effect of grain boundary properties on thermoelectric characteristics, FIG. 3 provides a model of a grain having a grain boundary with a width w and a grain boundary potential barrier of E_b. In addition, the grain size has a dimension of L which naturally separates the grain boundary on opposing sides of the grain. As illustrated in the figure, if an electron has an energy of E, the electron will pass through the grain boundary barrier height if E>E_band will scatter if E<E_b.

Not being bound by theory, assuming T(E) is a transmission probability of an electron passing through a grain boundary barrier height and there are N grain boundaries, the MFP of the electron due to scattering by the grain boundary can be expressed as Equation 14 when N is assumed to be infinity.

$\begin{matrix} λ_{grainboundary} = \sum_{n = 1}^{N \to \infty} {T (E)}^{n} (1 - T (E)) nL = \frac{T (E) L}{1 - T (E)} & Eqn 14 \end{matrix}$

which further provides a relaxation time of:

τ_B=λ_{grainboundary}/ν Eqn 15

where ν is given by:

$\begin{matrix} τ_{B} = \frac{L}{v} (1 + \frac{4 \frac{E}{E_{b}} \langle 1 - \frac{E}{E_{B}} \rangle}{\sinh^{2} \langle \sqrt{\frac{2 m_{c}^{*} E_{B} w^{2}}{ℏ^{2}} ❘ 1 - \frac{E}{E_{B}}} \rangle}) & Eqn 16 \end{matrix}$

In order to better understand the effect of grain related properties on the thermoelectric material behavior, and based on the model shown in FIG. 3, the effect of grain size on electrical conductivity was examined with results shown in FIG. 4. The width w of the grain boundaries was assumed to be constant at 2 nm while the grain boundary barrier height was varied from 20 meV to 300 meV. In addition, the electrical conductivity of a bulk thermoelectric material and a nanocomposite thermoelectric material was investigated and is shown in the graph. In the case of the nanocomposite thermoelectric material, SiO₂nanoparticles of 3 nm diameter were used for the calculations. It is appreciated that FIG. 4 illustrates that with increasing grain size, the electrical conductivity increases, which can be explained due to decreasing probability of scattering events. In addition, with the inclusion of ceramic nanoparticles within the material, significantly lower electrical conductivities were observed. Finally, varying the grain boundary barrier height significantly affects the conductivities of both the bulk thermoelectric material and the nanocomposite thermoelectric material. It is appreciated that this effect is stronger at smaller grain sizes simply due to the fact that smaller grains increase the number of scattering events and thus reduce a carrier MFP.

Turning now to FIG. 5, a graph illustrating the lattice and electronic thermal conductivity as a function of grain size is shown. Similar to FIG. 4, grain boundary scattering clearly affects both lattice and electronic thermal conductivity with the most significant effect occurring for grain sizes below 25 nm. In addition, FIG. 6 illustrates that total thermal conductivity illustrates a similar behavior to electrical conductivity which provides evidence that reduction in grain size for a thermoelectric material can be an effective way of reducing the material's thermal conductivity.

Regarding the Seebeck coefficient for a thermoelectric material, FIG. 7 illustrates a complicated relationship between the Seebeck coefficient, grain size, and grain boundary barrier height. In particular, and for bulk thermoelectric material, the highest Seebeck coefficient occurred for a grain boundary barrier height of 60 meV while for a nanocomposite thermoelectric material, the highest Seebeck coefficient was observed for a grain boundary barrier height of 20 meV. It is appreciated that the difference between the two materials and the associated Seebeck coefficient can be the result of filtering of low energy electrons within the grains of the nanocomposite thermoelectric material. In addition, FIG. 8 provides a typical normalized Seebeck coefficient distribution as a function of electron energy. As shown by this figure, a maximum value or maximum range of values for the Seebeck coefficient does not result from electron energies that are too low or too high. Stated differently, there is an intermediate value or range of values for electron energy that provides a desired Seebeck coefficient. In addition, low energy electrons pose a negative impact to the Seebeck coefficient.

Based on these figures and their teachings, it is clear that small grains with high grain boundary barrier potentials, for example E_b=300 meV, have the least effect on the Seebeck coefficient since such high potential barriers can filter even high energy electrons. On the other hand, FIG. 7 illustrates that the Seebeck coefficient behavior flipped or was inverted for the nanocomposite material versus the bulk material when the grain boundary barrier height was 20 meV and 60 meV, respectively. Not being bound by theory, this is postulated to be due to the Seebeck coefficient distribution being different for the two materials, and depending on the location of the peak of normalized Seebeck coefficient as shown in FIG. 8, either 20 meV or 60 meV can be more effective in increasing the Seebeck effect.

Regarding the dimensionless figure of merit ZT, FIG. 9 provides a comparison of ZT as a function of grain size, bulk thermoelectric material, nanocomposite thermoelectric material, and grain boundary barrier height. From this figure, it is appreciated that grain sizes below 25 nanometers can provide a dramatic improvement in the performance of bulk thermoelectric material and nanocomposite thermoelectric material. In addition, the grain boundary barrier height can significantly alter the ZT for a particular material. For example, at lower grain sizes, the ZT for the bulk and nanocomposite materials overlap, thereby suggesting that the benefit of adding second phase nanoparticles to a thermoelectric material can be diminished in cases where the grain boundary barrier potential is high due to impurities, doping, and the like.

The effect of ceramic nanoparticle inclusions within a bulk thermoelectric material on grain boundary barrier height can also be of interest with FIG. 10 illustrating a graphical representation of Seebeck coefficient as a function of grain boundary barrier height for a bulk thermoelectric material (Bulk) and a nanocomposite thermoelectric material (NC). As shown in FIG. 10, and for which a grain size of 30 nm was assumed, smaller grain boundary barrier heights are preferred with potentials over 100 meV virtually having no effect on the Seebeck coefficient of the material. In addition, FIG. 11 shows or illustrates the same behavior with respect to electrical conductivity and electronic thermal conductivity and FIG. 12 provides a graph illustrating the effect of grain boundary barrier height on ZT for bulk thermoelectric material and nanocomposite thermoelectric material.

It is appreciated that FIG. 12 could lead to the conclusion that lower grain boundary barrier heights are desired in all cases in order to achieve an increase in ZT for any thermoelectric material. However, such a conclusion can be false, for example when the grain size is also considered as discussed above in relation to FIG. 9.

Turning now to FIG. 13, a process for manufacturing a thermoelectric material is shown generally at reference numeral 10. The process 10 includes determining a thermoelectric material composition to be investigated at step 100, followed by determining a range of ZT values for the thermoelectric material composition at step 110. Given the range of determined ZT values determined at step 110, step 120 determines a range of maximum ZT values and a thermoelectric material is manufactured at step 130.

FIG. 14 illustrates a flowchart for a process according to another embodiment of the invention at reference numeral 20. The process 20 includes determining a thermoelectric material composition (C_i) to be investigated at step 200, followed by determining a range of matrix grain sizes that are obtainable for C_iat step 210 and determining a range of grain boundary barrier energy heights obtainable for C_iusing current state of the art manufacturing techniques at step 220. The determined thermoelectric material composition may or may not include inter-grain phonon scattering particles. In addition, step 210 or step 220 may or may not include determining a range of particle sizes obtainable for inter-grain phonon scattering particles using current state of the art manufacturing techniques.

At step 230, ZT values are calculated for C_ias a function of the obtainable range of matrix grain sizes and obtainable range of grain boundary heights that were determined in steps 210 and 220. Thereafter, a matrix grain size and grain boundary height is selected as a function of the calculated ZT values for C_iat step 240. In some instances, step 240 can include selecting a desired average grain size for the matrix phase and a desired particle size for inter-grain phonon scattering particles as a function of the calculated ZT values. Finally, a thermoelectric material is manufactured at step 250, the thermoelectric material having the composition C_ifrom step 200, the selected matrix grain size, grain boundary height and/or inter-grain phonon scattering particles with an average particle size that is equivalent to the desired particle size for the inter-grain phonon scattering second phase from step 240. For the purposes of the present invention, the term equivalent is defined to be within 10%, i.e. the average particle size of the inter-grain phonon scattering particles is within +/−10% of the desired particle size.

Another flowchart illustrating a process according to another embodiment of the present invention is shown generally at reference numeral 30 in FIG. 15. The process 30 includes determining a thermoelectric material composition (C_i) to be investigated at step 300 followed by determining a range of matrix grain sizes and a range of grain boundary heights that are obtainable for a thermoelectric material having composition C_iat steps 302 and 304, respectively. It is appreciated that the composition C_ican include a solid solution material, i.e. a material having a relatively uniform composition without the addition of second phase particles, second phase precipitates, and the like. In the alternative, the composition C_ican be a precipitation type material such as a precipitation type alloy in which second phase precipitates/particles are formed within the alloy/material during processing. In another alternative, the composition C_ican be a composite type material in which second phase particles are present and/or added to a matrix material. The matrix material can be a metal, alloy, semiconductor, ceramic, e.g. an oxide, nitride, etc., and the like. Also, the second phase particles can be a metal, alloy, semiconductor, ceramic, e.g. an oxide, nitride, etc., and the like.

The process 30 also includes calculating Seebeck coefficients for the C_icomposition as a function of the obtainable range of matrix grain sizes and obtainable grain boundary heights at step 306. At step 308, the process 30 includes calculating electrical resistivity values for C_ias a function of the obtainable range of matrix grain sizes and grain boundary heights. At step 310, the calculation of thermal conductivity values for C_ias a function of the obtainable range of matrix grain sizes and grain boundary heights is performed. Next, the calculation of ZT values for C_ias a function of the calculated Seebeck coefficients, electrical resistivity values, and thermal conductivity values is performed at step 312.

Once the ZT values have been calculated, a maximum range of the calculated ZT values for C_iis determined at step 314 and step 316 includes determining a matrix grain size and a grain boundary height that is within the maximum determined range from step 314. At step 318, a thermoelectric material having the composition C_iis manufactured, the material having the determined matrix grain size and grain boundary height from step 316. It is appreciated that in order to manufacture the thermoelectric material at step 318, material corresponding to the composition C_iis provided and processed to produce the material.

Referring to FIG. 16, another process according to an embodiment of the present invention is shown generally at reference numeral 40. The process 40 includes determining a thermoelectric material composition C_ito be investigated at step 400, followed by determining a range of matrix grain sizes and a range of grain boundary heights for the composition C_ito be investigated at steps 410 and 420, respectively. At step 430, ZT values for C_iare calculated as a function of the determined ranges of matrix grain sizes and grain boundary heights from steps 410 and 420. It is appreciated that the ZT values can be calculated as a function of Seebeck coefficients, electrical resistivity values, and thermal conductivity values for C_ithat are also a function of the range of matrix grain sizes and grain boundary heights determined in steps 410 and 420, respectively.

A desired range of the calculated ZT values for C_iis determined at step 440, e.g. a desired maximum range of the calculated ZT values can be determined. Based on the determined desired range of the calculated ZT values, the range of matrix grain sizes and grain boundary heights that fall within this range are determined at step 450. Then, matrix grain sizes and grain boundary heights that are obtainable using current state of the art production techniques and that fall within the determined desired range of calculated ZT values for C_iare determined at step 460.

A powder that has the C_icomposition and an average particle diameter that is equivalent to an obtainable matrix grain size that falls within the grain sizes of step 460 is provided. It is appreciated that the powder can be a homogeneous powder, i.e. a powder made of individual powder particles that have a generally uniform composition with each other or, in the alternative, a non-homogeneous composition, i.e. the powders being made from first component powder particles and second component powder particles. As stated above, the first component powder particles can be in the form of be a metal, alloy, semiconductor, ceramic, e.g. an oxide, nitride, etc., and the like. Also, the second component particles can be in the form be a metal, alloy, semiconductor, ceramic, e.g. an oxide, nitride, etc., and the like. In addition, third component particles, fourth component particles, etc. can be included in the powder having the C_icomposition at step 470, the third, fourth, etc. particles being of the form of be a metal, alloy, semiconductor, ceramic, e.g. an oxide, nitride, etc., and the like. Also, the powders, particles, etc., may or may not be in the form of nanoparticles such a nano-spheres, nano-rods, nano-discs nano-ellipsoids, and the like.

The powder having the C_icomposition is processed to produce a component. For example, a process for processing the powders is shown in FIG. 17 at reference numeral 60. The process 60 includes taking or obtaining the powders of the thermoelectric material at step 600, the powders having desired chemical and/or physical properties/characteristics. For example, the powders can include a mixture of first component particles P₁, second component particles P₂, etc. as shown at 602. In addition, the powders can have a desired particle diameter or desired particle size as shown at step 604 and/or a desired surface modification as shown at step 606. It is appreciated that the desired particle diameter is determined from a process disclosed herein, e.g. the processes illustrated in FIGS. 13-16.

It is also appreciated that the particle diameter of the first component particles P₁can be the same or different than the particle diameter of the second component particles P₂. Finally, it is appreciated that the terms “particle diameter” and “matrix grain size” refer to an average particle diameter and average matrix grain size, respectively, as is known to those skilled in the art. The variation in particle diameters and matrix grain sizes can have a half width at half maximum height (HWHM) to modal diameter ratio between 0.4-0.6 for a differential distribution as is known to those skilled in the art. In other instances, the HWHM/modal diameter ratio is between 0.3-0.4 or, in the alternative, between 0.2-0.3. In the alternative, the HWHM/modal diameter ratio is between 0.6-0.7 or, in the alternative, between 0.7-0.8.

The process 60 further includes compaction of powders to produce a thermoelectric component at step 610. The step 610 can be a compaction by any means or method known to those skilled in the art, illustratively including sintering, hot isostatic pressing (HIP), cold isostatic pressing (CIP), die pressing, continuous particle or powder processing (CPP), etc. as shown at 612. In some instances, the component is further treated at step 620 in order to obtain a desired matrix grain size and/or desired grain boundary barrier height. For example and for illustrative purposes only, the treatment at step 620 can include a thermal treatment, a mechanical treatment, and/or a thermal-mechanical treatment. Other treatments such as exposure to electromagnetic radiation, nuclear radiation, and the like can be included.

Regarding the calculation of the various Seebeck coefficients, electrical resistivity values, thermal conductivity values, ZT values, and the like, FIG. 18 provides a schematic illustration of a computer at reference numeral 50. The computer 50 can include a CPU 500 with a plurality of components 502, 504, 506, . . . 516. For example and for illustrative purposes only, the CPU 500 can include a processor 502, memory 504, database 506, and the like. It is appreciated that the various components 502-516 are in communication with each other and with the CPU 500 as is known to those skilled in the art.

The various calculations can include algorithms with respect to Equations 1-16, assumptions, electric constants and/or physical constants known to those skilled in the art. For example, the temperature T referenced in Equation 1 can be assumed to be room temperature. In the alternative, the calculations can assume a plurality of temperatures, e.g. temperatures ranging from 0° C. to 200° C. at increments of 1° C., increments of 5° C., increments of 10° C., and the like.

In order to illustrate a sample calculation, but not limit the scope of the instant disclosure in any way, an example of a ZT calculation is provided below.

The figure of merit (ZT) for a given grain size and grain barrier height combination “j”, is given by:

ZT=S²·T·C/(K_e+K_l) Eqn 17

where S is the Seebeck coefficient for the given grain size and grain boundary barrier height, C is the electrical conductivity, K_eis the electron contribution to the overall thermal conductivity and K_lis the phonon contribution to the overall thermal conductivity—all for the given grain size and grain boundary barrier height. It is appreciated that the index “j” for the given grain size and grain boundary energy barrier height is not shown for convenience, unless needed for clarity. Stated differently, each of the expressions discussed are for the given grain size and grain boundary energy barrier height “j”.

In order to properly define S, C, K_phand K_e, a series of constants known to those skilled in the art are required and provided below. It should be appreciated that the values for the constants listed below are presented without units, as used in computer code for the calculation of ZT as a function of grain size and grain boundary energy height. However, it should also be appreciated that the units for the constants provided below would be known to one skilled in the art and be in accordance with units that afford calculation of the Seebeck coefficient in microvolts per meter (μV/m), electrical conductivity in siemens per meter (S/m) and thermal conductivity in watts per meter kelvin (W/mK).

General constants used in an example calculation include:

π=3.14
κ_B=1.38×10⁻²³=Boltzmann constant
h_c=1.054×10⁻³⁴=Planck's constant
e_v=1.6×10⁻¹⁹=1 electron volt
e=1.6×10⁻¹⁹=electron charge
m_e=9.1×10⁻³¹=electron effective mass
ε₀=8.85×10⁻¹²=permittivity of air
e₁=69.8·ε_o·4π=high frequency dielectric constant Eqn 18
e₀=400·ε_o·4π=static dielectric constant Eqn 19

Also, density of states constants and expressions related to effective mass include:

N_v=12=number of valleys in the electron bandstructure
E_g=0.13·e_v−1.08×10⁻⁴·T·e_v=electron bandgap Eqn 20
β_L=E_g/(k_BT)=bandgap in k_BT Eqn 21
β=k_BT/E_g=inverse of bandgap in k_BT Eqn 22
e_g=E_g/e_v=bandgap in electron volt Eqn 23
m_h1=0.0308·m_e=hole effective mass Eqn 24
m_h2=0.441·m_e=hole effective mass Eqn 25
m_h3=0.0862·m_e=hole effective mass Eqn 26
m_e1=0.0213·m_e=electron effective mass Eqn 27
m_e2=0.319·m_e=electron effective mass Eqn 28
m_e3=0.0813·m_e=electron effective mass Eqn 29
md_e=N_v^2/3·(m_e1·m_e2·m_e3)^1/3=density of state electron effective mass Eqn 30
md_e1=(m_e1·m_e2·m_e3)^1/3=density of state electron effective mass Eqn 31
md_h=N_v^2/3·(m_e1·me_e2·m_e3)^1/3=density of state hole effective mass Eqn 32
md_h1=(m_e1·m_e2·m_e3)^1/3=density of state hole effective mass Eqn 33
Mc_e=3/(1/m_e1+1/m_e2+1/m_e3)=Total effective mass of electron Eqn 34
Mc_h=3/(1/m_h1+1/m_h2+1/m_h3)=Total effective mass of hole Eqn 35

Fermi energy expressions include:

$\begin{matrix} zF = 1.68000000000002000000 \times 10^{- 9} \cdot T^{3} - 2.32525000000003000000 \times 10^{- 6} \cdot T^{2} + 1.18722500000002000000 \times 10^{- 3} \cdot T - 2.36675820000002000000 \times 10^{- 1} & Eqn 36 \\ E_{f} = zF \cdot e_{v} = Fermi energy in e_{v} & Eqn 37 \\ x_{f} = E_{f} / (k_{B} \cdot T) = Fermi energy in k_{B} \cdot T & Eqn 38 \\ F_{e} = (E_{f} - E_{g} / 2) / (k_{B} \cdot T) = energy of electron & Eqn 39 \\ F_{h} = - (E_{g} / 2 + E_{f}) / (k_{B} \cdot T) = energy of hole & Eqn 40 \end{matrix}$

with Fermi functions:

$\begin{matrix} f_{e} = 1 / (1 + \exp (z - F_{e})) & Eqn 41 \\ d 1_{e} = \frac{d f_{e}}{d z} = \frac{d (1 / (1 + \exp (z - F_{e}))}{d z} & Eqn 42 \\ f_{h} = 1 / (1 + \exp (z - F_{h})) & Eqn 43 \\ d 1_{h} = \frac{d f_{h}}{d z} = \frac{d (1 / (1 + \exp (z - F_{h}))}{d z} & Eqn 44 \end{matrix}$

and conductivity effective mass expressions:

$\begin{matrix} L 1_{e} = \int_{- F_{e}}^{\infty} {(z + β z^{2})}^{1.5} \frac{d (1 / (1 + \exp (z - F_{e}))}{d z} d z & Eqn 45 \\ L 2_{e} = \int_{- F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot \frac{d (1 / (1 + \exp (z - F_{e}))}{d z}}{(1 + 2 β z)} d z & Eqn 46 \\ L 1_{h} = \int_{- F_{h}}^{\infty} {(z + β z^{2})}^{1.5} \cdot \frac{d (1 / (1 + \exp (z - F_{h}))}{d z} d z & Eqn 47 \\ L 2_{h} = \int_{- F_{h}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot \frac{d (1 / (1 + \exp (z - F_{h}))}{d z}}{(1 + 2 β z)} d z & Eqn 48 \\ \begin{matrix} m c_{e} = M c_{e} \cdot L 1_{e} / L 2_{e} \\ = conductivity effective mass of electron \end{matrix} & Eqn 49 \\ \begin{matrix} m c_{h} = M c_{h} \cdot L 1_{h} / L 2_{h} \\ = conductivity effective mass of hole \end{matrix} & Eqn 50 \end{matrix}$

Regarding scattering terms, the polar optical phonon scattering of electrons (τ_poe) and holes (τ_poh) can be determined from the following constants and expressions:

ε_o=8.85×10⁻¹²=permittivity of air
e₁=69.8=high frequency dielectric constant
e₀=400=static dielectric constant
K₀=0.1e_v=optical phono energy
N₀=1/(exp(K₀/(κB·T))−1)=phonon Plank function Eqn 51
cop_e=(4π·ε_o·h_c²/(3e²·(1/e₁−1/e₀)·N₀))·(2/(md_c·K₀·(1+K₀/E_g)))^0.5 Eqn 52
cpo_h=(4π·ε_o·h_c²/(3e²·(1/e₁−1/e₀)·N₀))·(2/(md_h·K₀·(1+K₀/E_g)))^0.5 Eqn 53
b_e=(1+2βz)/(z+βz²) Eqn 54
τ_z3=1/b_e Eqn 55
with:
τ_poe=cpo_e·τ_z3=lifetime for optical phonon-electron scattering Eqn 56
and:
τ_poh=cpo_h·τ_z3==lifetime for optical phonon-hole scattering Eqn 57

Deformation potential scattering of electrons (τ_dae) and holes (τ_dah) by acoustic phonons can be determined using the following constants and expressions:

K_a=1.1=fitting constant
C₁=7.1×10¹⁰=combination electric constant
E_ac=3.5e_v=deformation potential for acoustic phonon Eqn 58
B_a=8βz(1+βz)·K_a/(3(1+2βz)²) Eqn 59
A_a=βz(1−K_a)/(1+2βz) Eqn 60
Cda_e=2πh_c⁴·C₁/(E_ac²·(2·md_c·κ_B·T)^1.5) Eqn 61
Cda_h=2πh_c⁴·C₁/(E_ac²·(2·md_h·κ_B·T)^1.5) Eqn 62
τ_z1=1/(((z+βz²)^0.5)·(1+2βz)·((1−A_a)²−B_a)) Eqn 63
with:
τ_dae=Cda_e·τ_z1=lifetime for acoustic phonon-electron scattering Eqn 64
and:
τ_dah=Cda_h·τ_z1=lifetime for acoustic phonon-hole scattering Eqn 65

Deformation potential scattering of electrons (τ_doe) and holes (τ_doh) by optical phonons can be determined using the following constants and expressions:

r_ho=7.86×10³=density
a=10.45×10⁻¹⁰=lattice constant
K₀=1.1=fitting constant
E_oc=60e_v=deformation potential for optical phonon Eqn 66
B_o=8βz(1+βz)·K_o/(3(1+2βz)²) Eqn 67
A_o=βz(1−K_o)/(1+2βz) Eqn 68
Cdo_e=2h_c²·a²·K₀²·r_ho/(πE_oc²·(2·md_e·κ_β·T)^1.5) Eqn 69
Cdo_h=2h_c²·a²·K₀²·r_ho/(πE_oc²·(2·md_h·κ_B·T)^1.5) Eqn 70
τ_z2=1/(((z+βz²)^0.5)·(1+2βz)·((1−A_o)²−B_o)) Eqn 71
with:
τ_doe=Cdo_e·τ_z2=lifetime for deformation potential-electron scattering Eqn 72
and:
τ_doh=Cdo_h·τ_z2=lifetime for deformation potential-hole scattering Eqn 73

In the event that the thermoelectric material includes nanoparticles, the scattering of elections (τ_ie) and holes (τ_ih) by the nanoparticles can be determined using the following constants and expressions:

$\begin{matrix} λ_{e} = Real Portion of : ((((2 e^{2} \cdot {(2 \cdot {md}_{e})}^{1.5} / (e_{0} π \cdot h_{c}^{3})) \cdot {(k_{B} T)}^{0.5} \cdot {(F_{e} + F_{e}^{2} β)}^{0.5} \cdot {(1 + 2 F_{e} β))}^{- 0.5}) / 10^{- 9}) & Eqn 74 \\ λ_{h} = Real Portion of : ((((2 e^{2} \cdot {(2 \cdot {md}_{h})}^{1.5} / (e_{0} π \cdot h_{c}^{3})) \cdot {(k_{B} T)}^{0.5} \cdot {(F_{h} + F_{h}^{2} β)}^{0.5} \cdot {(1 + 2 F_{h} β))}^{- 0.5}) / 10^{- 9}) & Eqn 75 \\ WF = 300 = work function of an inclusion,  e . g . a ceramic inclusion \\ EE = 4.76 = electron affinity of host material \\ {Vb}_{e} = WF - EE + F_{e} \cdot (k_{B} T) / e_{v} & Eqn 76 \\ {Vb}_{h} = WF - EE + F_{h} \cdot (k_{B} T) / e_{v} & Eqn 77 \\ W_{e} = {Vb}_{e} \cdot R / \exp (- R / λ_{e}) & Eqn 78 \\ W_{h} = {Vb}_{h} \cdot R / \exp (- R / λ_{h}) & Eqn 79 \\ W = 10 = fitting constant \\ U = 0.3 = volume % of nanoparticles \\ R = 1.5 = radius of nanoparticles in nm \\ {Vr}_{e} = W_{e} \cdot (1 / 1) \cdot \exp (- 1 / λ_{e}) \cdot e_{v} : V_{e} = {Vr}_{e} / e_{v} & Eqn 80 \\ {Vr}_{h} = W_{h} \cdot (1 / 1) \cdot \exp (- 1 / λ_{h}) \cdot e_{v} : V_{h} = {Vr}_{h} / e_{v} & Eqn 81 \\ a_{e} = \int \int_{R, 0}^{\infty, W} (q \cdot \sin (\frac{1 t}{R}) v_{re}) d ld t & Eqn 82 \\ α_{e} = (1 / R^{4}) \cdot a_{e} & Eqn 83 \\ a_{h} = \int \int_{R, 0}^{\infty, W} (1 \cdot \sin (\frac{1 t}{R}) v_{rh}) d ld t & Eqn 84 \\ α_{h} = (1 / R^{4}) \cdot a_{h} & Eqn 85 \end{matrix}$

with:
τ_ie=(κ_BT)^1.5·(z^1.5·(1+βz)^1.5·4R·(2md_e1)^0.5)/((1+2βz)·U·3λ_e)=lifetime for nanoparticles-electron scattering Eqn 86
and:
τ_ih=(κ_BT)^1.5·(z^1.5·(1+βz)^1.5·4R·(2md_h1)^0.5)/((1+2βz)·U·3λ_h)=lifetime for nanoparticles-electron scattering Eqn 87

Grain boundaries can naturally be a source of scattering, and the scattering of electrons (τ_be2) and holes (τ_bh2) can be determined using the following constants and expressions:

e_b=0.3(values range from 0.003-0.3 with 0.3 corresponding to a very strong electrical conductivity)=Barrier height
d₁=30=grain boundary constant
E_b=−e_b·e_v=grain boundary barrier energy height=‘E_b’ in FIG. 3 Eqn 88
DL=d₁·10⁻⁹=grain size=‘L’ in FIG. 3 Eqn 89
Gw=5×10⁻⁹=grain boundary width=‘W’=FIG. 3
E=κ_BTz Eqn 90
GN=4(E/E_b)·(1−E/E_b) Eqn 91
GD_e=((2md_e·E_b·Gw²/h_c²)·(1−E/E_b))^0.5 Eqn 92
Z_e=(exp(GD_e)−exp(−GD_e))/2 Eqn 93
Z_e1=GD_e Eqn 94
with:
τ_be2=DL·(md_e/(2·E))^0.5·(1+GN/(Z_e1)²)=lifetime for grain boundary-electron scattering Eqn 95
and:
GN=4·(E/E_b)·(1−E/E_b) Eqn 96
GD_h=((2·md_h·E_b·Gw²/h_c²)·(1−E/E_b))^0.5 Eqn 97
Z_h=(exp(GD_h)−exp(−GD_h))/2 Eqn 98
Z_h1=GD_h Eqn 99
with:
τ_bh2=DL·(md_h/(2E))^0.5·(1+GN/(Z_h1)²)=lifetime for grain boundary-hole scattering Eqn 100

Though not required, interfacial surface roughness scattering of an inclusion particle can be considered with such scattering of electrons (τ_ifre) and holes (τ_ifrh) determined by the following constants and expressions:

d=1=roughness height in nm
c₁=2.4=correlation length in nm
λ=c₁×10⁻⁹ Eqn 101
δ=d×10⁻⁹ Eqn 102
ε₀=8.85×10⁻¹²

For holes:

$\begin{matrix} h_{WL} = 11.4 \times 10^{- 9} = hole constant \\ {Ns}_{h} = 10^{18} = carrier concentration \\ k_{h} = 2 π / h_{WL} = counting constant for phonon energy & Eqn 103 \\ K_{h} = 2 k_{h} & Eqn 104 \\ ɛ_{s} = 400 = high frequency dielectric constant \\ b_{h} = {(6 * 2 {md}_{h} \cdot e^{2} \cdot N_{Sh} / (h_{c}^{2} {λɛ}_{s} \cdot ɛ_{o}))}^{0.333} & Eqn 105 \\ {fz}_{h} = {(0.5 \cdot b_{h}^{3} \cdot z^{2})}^{0.5} \cdot \exp (- 0.5 b_{h} z) & Eqn 106 \\ {fzq}_{h} = {(0.5 \cdot b_{h}^{3} \cdot z_{q}^{2})}^{0.5} \cdot \exp (- 0.5 b_{h} z_{q}) & Eqn 107 \\ {Fq}_{h} = \int \int_{0, 0}^{10^{9}, 10^{9}} (({fz}_{h}^{2} \cdot \exp (- z \cdot q) \cdot {fzq}_{h}^{2} \cdot \exp (z_{q} q) d z \cdot d z_{q} & Eqn 108 \\ {qs}_{h} = (e^{2} \cdot {md}_{h} / (h_{c}^{2} \cdot 2 π \cdot ɛ_{s} \cdot ɛ_{o})) \cdot {Fq}_{h} & Eqn 109 \\ J_{h} = \int_{0}^{K_{h}} \frac{(- q^{2} λ^{4} / 4) \cdot q^{4}}{2 {k_{h}^{3} (q + {qs}_{h})}^{2} \sqrt{{(1 - (\frac{q}{2 k_{h}}))}^{2}}} d q & Eqn 110 \\ w_{ifrh} = ({(e^{2} \cdot {Ns}_{h} \cdot δ \cdot λ / (2 ɛ_{s} \cdot ɛ_{o}))}^{2} \cdot {md}_{h} \cdot J_{h} / h_{c}^{3}) & Eqn 111 \\ τ_{ifrh} = 1 / {wifr}_{h} = lifetime for roughness - hole scattering & Eqn 112 \end{matrix}$

and for electrons:

$\begin{matrix} e_{WL} = 11.4 \times 10^{- 9} = electron constant \\ {Ns}_{e} = 10^{18} = carrier concentration \\ k_{e} = 2 π / e_{WL} & Eqn 113 \\ K_{e} = 2 k_{e} & Eqn 114 \\ ɛ_{s} = 400 = high frequency dielectric constant \\ b_{e} = {(6 \cdot 2 {md}_{e} \cdot e^{2} \cdot {Ns}_{e} / (h_{c}^{2} \cdot ɛ_{s} \cdot ɛ_{o}))}^{0.333} & Eqn 115 \\ {fz}_{e} = {(0.5 \cdot b_{e}^{3} \cdot z^{2})}^{0.5} \cdot \exp (- 0.5 b_{e} z) & Eqn 116 \\ {fzq}_{e} = {(0.5 \cdot b_{e}^{3} \cdot z_{q}^{2})}^{0.5} \cdot \exp (- 0.5 b_{e} z_{q}) & Eqn 117 \\ {Fq}_{e} = \int \int_{0, 0}^{10^{9}, 10^{9}} (({fz}_{e}^{2} \cdot \exp (- z \cdot q) \cdot {fzq}_{e}^{2} \cdot \exp (z_{q} q) d z \cdot d z_{q} & Eqn 118 \\ {qs}_{e} = (e^{2} \cdot {md}_{e} / (h_{c}^{2} \cdot 2 π \cdot ɛ_{s} \cdot ɛ_{o})) \cdot {Fq}_{e} & Eqn 119 \\ J_{e} = \int_{0}^{K_{e}} \frac{\exp (- q^{2} λ^{4} / 4) \cdot q^{4}}{2 {k_{e}^{3} (q + {qs}_{e})}^{2} \sqrt{{(1 - (\frac{q}{2 k_{e}}))}^{2}}} d q & Eqn 120 \\ w_{ifre} = ({(e^{2} \cdot {Ns}_{e} \cdot δ \cdot λ / (2 ɛ_{s} \cdot ɛ_{o}))}^{2} \cdot {md}_{e} \cdot J_{e} / h_{c}^{3}) & Eqn 121 \\ τ_{ifre} = 1 / w_{ifre} = lifetime for roughness - electron scattering & Eqn 122 \end{matrix}$

Thus the total scattering for electrons (τ_ze) and holes (τ_zh) can be obtained from:

τ_ze=1/(1/τ_doe+1/τ_dae+1/τ_poe+1/τ_be2+1/Γ_ie+1/τ_ifre)=total lifetime for electron scattering Eqn 123
τ_zh=1/(1/τ_doh+1/τ_dah+1/τ_poh+1/τ_bh2+1/Γ_ie+1/τ_ifrh)=total lifetime for hole scattering Eqn 124

Once the scattering terms have been determined, electrical conductivity (C_e, C_h), Seebeck coefficient (S_e, S_h) can be determined for the given grain size and grain boundary energy barrier height using the following expressions. In particular, C_eand C_hcan be determined from:

$\begin{matrix} c_{e} = (e^{2} / (3 π^{2} \cdot h_{c}^{3} \cdot m_{ce})) \cdot {(2 \cdot m_{de} k_{B} T)}^{1.5} & Eqn 125 \\ C_{e} = c_{e} \cdot \int_{- F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{ze} \cdot \frac{d (1 / (1 + \exp (z - F_{e}))}{d z}}{(1 + 2 β z^{2})} d z & Eqn 126 \\ c_{h} = (e^{2} / (3 π^{2} \cdot h_{c}^{3} \cdot m_{ch})) \cdot {(2 \cdot m_{dh} k_{B} T)}^{1.5} & Eqn 127 \\ C_{h} = c_{h} \cdot \int_{- F_{h}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{zh} \cdot \frac{d (1 / (1 + \exp (z - F_{h}))}{d z}}{(1 + 2 β z)} d z & Eqn 128 \end{matrix}$

and the total electrical conductivity (C) is simply:

C=(C_e+C_h) Eqn 129

Also, S_eand S_hcan be determined by:

$\begin{matrix} Se = - (k_{B} / e) \cdot \frac{\int_{F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{ze} \cdot \frac{d (\frac{1}{1 + \exp (z - F_{e})})}{ⅆ z} \cdot (z - F_{e})}{(1 + 2 β z)} d z}{\int_{- F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{ze} \cdot \frac{d (\frac{1}{1 + \exp (z - F_{e})})}{d z}}{(1 + 2 β z)} d z} & Eqn 130 \\ S_{h} = - (k_{B} / e) \cdot \frac{\int_{F_{h}}^{\infty} \frac{\begin{matrix} {(z + β z^{2})}^{1.5} \cdot τ_{zh} \cdot \\ \frac{d (1 / (1 + \exp (z - F_{h}))}{d z} \cdot (F_{h} - z) \end{matrix}}{(1 + 2 β z)} d z}{\int_{- F_{h}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{zh} \cdot \frac{d (1 / (1 + \exp (z - F_{h}))}{d z}}{(1 + 2 β z)} d z} & Eqn 131 \end{matrix}$

and the total Seebeck coefficient (S) is simply:

S=(S_e.C_e+S_h·C_h)/C Eqn 132

It is appreciated that the power factor (P) is given by:

P=S²·C Eqn 133

Also, it should be appreciated that the preceding constants and expressions afford for the calculation of electrical conductivity and Seebeck coefficient as a function of grain size, grain boundary width and/or grain boundary barrier height. Naturally, certain constants would be assumed and varied as needed. For example, the temperature (T) for the calculations used to produce FIGS. 3 and 7 was 300° C., however, the temperature could be varied to determine the effect of temperature on C and S. Also, the size/radius, volume percent, fitting constants, etc., of any inclusions, nanoparticles, etc., can be varied to determine their effect on calculated results.

Regarding thermal conductivity of the material, it is appreciate that additional expressions and scattering terms are needed. For example, Lorentz numbers for electrons (L_e) and holes (L_h) are useful and can be obtained from the following expressions:

$\begin{matrix} L_{e} = {(\begin{matrix} (\frac{\frac{C_{e}}{C} \int_{- F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{ze} \cdot z^{2} \cdot \frac{d (\frac{1}{1 + \exp (z - F_{e})})}{d z}}{(1 + 2 β z)} d z}{\int_{- F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{ze} \cdot \frac{d (\frac{1}{1 + \exp (z - F_{e})})}{d z}}{(1 + 2 β z)} d z}) - \\ (\frac{\int_{- F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot t_{ze} \cdot z \cdot \frac{d (1 / (1 + \exp (z - F_{e}))}{d z}}{(1 + 2 β z)} d z}{\int_{- F_{e}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{ze} \cdot \frac{d (1 / (1 + \exp (z - F_{e}))}{d z}}{(1 + 2 β z)} d z}) \end{matrix})}^{2} & Eqn 134 \\ and \\ L_{h} = {(\begin{matrix} (\frac{\frac{C_{h}}{C} \int_{- F_{h}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{zh} \cdot z^{2} \cdot \frac{d (\frac{1}{1 + \exp (z - F_{e})})}{d z}}{(1 + 2 β z)} d z}{\int_{- F_{h}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{zh} \cdot \frac{d (\frac{1}{1 + \exp (z - F_{e})})}{d z}}{(1 + 2 β z)} d z}) - \\ (\frac{\int_{- F_{h}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot t_{zh} \cdot z \cdot \frac{d (1 / (1 + \exp (z - F_{e}))}{d z}}{(1 + 2 β z)} d z}{\int_{- F_{h}}^{\infty} \frac{{(z + β z^{2})}^{1.5} \cdot τ_{zh} \cdot \frac{d (1 / (1 + \exp (z - F_{e}))}{d z}}{(1 + 2 β z)} d z}) \end{matrix})}^{2} & Eqn 135 \\ and \\ \begin{matrix} L_{b} = {(e / k_{B})}^{2} \cdot C_{e} \cdot C_{h} \cdot {(S_{e} - S_{h})}^{2} / C^{2} \\ = bipolar Lorentz number \end{matrix} & Eqn 136 \end{matrix}$

which afford for the electronic thermal conductivity (K_e) to be calculated from:

K_e=T·C·(k_B²/e²)·(L_e+L_h+L_b) Eqn 137

As such, the electronic thermal conductivity as a function of grain size can be calculated and plotted as shown in FIG. 5.

Regarding lattice thermal conductivity, the following constants and expressions are useful:

$\begin{matrix} K_{0} = 0.1 e_{v} = phonon energy & Eqn 138 \\ L_{EO} = 1 \times 10^{- 3} = grain size parameter of sample B_{u} = 10.9 \times 10^{- 18} = Umpklamp scattering constant n = 5.2 = fitting parameter  B_{n} = 42 \times 10^{- 24} = normal scattering constant  A = 1.9 \times 10^{- 42} = alloy scattering constant \\ E_{d} = 25 e_{v} v = 3000 = sound velocity θ_{d} = 164.9 = Debye temperature & Eqn 139 \\ Q = θ_{d} / T U = 0.3 = volume % of particles r = 1.5 = radius of particles in nm & Eqn 140 \\ A_{1} = r^{2} / 1^{2} & Eqn 141 \\ B_{1} = 1^{2} / r m = - 0.925 = mass constant k = 0.3 = spring constant q = 80 = fitting parament & Eqn 142 \\ r_{1} = (α^{2} / 4) \cdot m^{2} + 3 \cdot (α^{8}) \cdot (k^{2}) \cdot {(\sin (α \cdot q_/ 2))}^{4} / ({(α \cdot q_/ 2)}^{4}) & Eqn 143 \\ r_{2} = 3.14 \cdot (\cos (4 r \cdot q) - 1 + (4 r \cdot q) \cdot \sin (4 r \cdot q) + 32 {(r \cdot q)}^{4} - 8 {(r \cdot q)}^{2}) / (16 {(r \cdot q)}^{6}) & Eqn 144 \\ σ_{_ray} = {(r \cdot q)}^{4} \cdot r_{1} \cdot r_{2} & Eqn 145 \\ α = 1 / \sqrt 2 = trigonometric ratio q_= 0.1 = wave vector parameter \\ X = (\sqrt{1 + m)} \cdot \sqrt{1 + k)}) - 1 & Eqn 146 \\ σ_{_ng} = 2 \cdot (1 - \sin (2 \cdot r \cdot q \cdot X) / (r \cdot q \cdot X) + \sin (r \cdot q \cdot X) \cdot \sin (r \cdot q \cdot X) / ({(r \cdot q \cdot X)}^{2})) & Eqn 147 \\ σ_{_total} = 1 / ((1 / σ_{_ray}) + (1 / σ_{_ng})) & Eqn 148 \\ η_{1} = \frac{U (3 \times 10^{27})}{4 π \int_{0}^{\infty} \frac{r^{3} r^{(A 1 - 1)} \exp (\frac{- r}{B_{1}})}{B_{1}^{A 1} Γ (A_{1})} d r} & Eqn 149 \end{matrix}$

which afford for grain boundary scattering (t_B), Umpklamp scattering (t_u), normal scattering (t_N), alloy scattering (t_A) and nanoparticle scattering (t_D) to be determined from the following expressions:

$\begin{matrix} \begin{matrix} t_{_B} = v / L_{EO} \\ = total lifetime for phonon and boundary scattering \end{matrix} & Eqn 150 \\ \begin{matrix} t_{_U} = B_{u} \cdot s^{2} \cdot k_{B}^{2} \cdot T^{3} \cdot \exp (- θ_{d} / (n \cdot T)) / h_{c}^{2} \\ = total lifetime for phonon - and UmKlapp \\ scattering \end{matrix} & Eqn 151 \\ \begin{matrix} t_{_N} = B_{n} \cdot s^{2} \cdot k_{B}^{2} \cdot T^{5} / h_{c}^{2} \\ = total lifetime for phonon - and Normal \\ scattering \end{matrix} & Eqn 152 \\ \begin{matrix} t_{_A} = A \cdot s^{4} \cdot k_{B}^{4} \cdot T^{4} / h_{c}^{4} \\ = total lifetime for phonon - and alloy scattering \end{matrix} & Eqn 153 \\ t_{_D} = v η_{1} (1 / ({π (A_{1} \cdot B_{1})}^{2})) \cdot (1 \times 10^{- 18}) \cdot \int_{0}^{\infty} π^{2} r^{4} σ_{_total} \cdot r^{(A 1 - 1)} \cdot \exp \frac{(- r / B_{1})}{B_{1}^{A 1} Γ (A_{1})} d r = total lifetime for phonon - and nanoparticle scattering & Eqn 154 \end{matrix}$

For grain boundary scattering:

D_L=30×10⁻⁹=grain size
FT₁=0.8=fitting parameter
FT₂=10¹⁵=fitting parameter
ε=0.1=fitting parameter
t_ref=(D_L/v)·FT₁=total lifetime for phonon- and reflection type scattering Eqn 155
t_diff=(D_L/v)·(κ_B·θ_d/K₀)·1/η=total lifetime for phonon- and diffraction type scattering Eqn 156
t_ray=(v/D_L)³·(θ_d·h_c/(T·K₀))⁴·FT₂=total lifetime for phonon- and Rayleigh type scattering Eqn 157
and
t_Bn=1/t_ref+1/t_diff+1/t_ray Eqn 158

If surface roughness scattering is to be considered, the following constants and expressions can be included in a calculation:

n=1
W_L=3·10⁻⁹=grain boundary width
k=2π/W_L Eqn 159
dw_AB=9.69×10¹²=used/calculated for BiTe/SiO₂
h_cw=0.0024·e_v
u₁=1730=constant for BiTe
λ=c₁·10⁻⁹ Eqn 160
δ=d·10⁻⁹ Eqn 161
w_w1=2=1^stgrain boundary thickness value/parameter
L₁=w_w1·10⁻⁹ Eqn 162
q_z1=n·π/L₁ Eqn 163
w_w2=5=2^ndgrain boundary thickness value/parameter
L₂₌w_w2·10⁻⁹ Eqn 164
q_z2=n·π/L₂ Eqn 165
w_w3=10=3^rdgrain boundary thickness value/parameter
L₃=w_w3·10⁻⁹ Eqn 166
q_z3=n·π/L₃ Eqn 167
Z_p=∫₀^6.28((1−cos(θ))(cos(θ))²exp(−k²λ²(sin(θ/2))²)dθ Eqn 168

And interface roughness scattering:

α₁=(1−(q_z1²−k²)/(q_z1²+k²)·sin(2q_z1δ)/(2q_z1δ))·(δ/L₁) Eqn 169
τ_{ifr_1}=(dw_AB)²·(h_cw/h_c)·λ²·α₁²·Z_p/(2u₁²)=lifetime for roughness and phonon scattering Eqn 170
t_c=t_U+t_N+t_A+t_D+t_Bn+τ_{ifr_1}=total lifetime of phonon scattering Eqn 171
t_c=1/t_c Eqn 172
t_N=1/t_N Eqn 173

and the Lorentz numbers can be calculated from:

$\begin{matrix} L_{1} = \int_{0}^{Q} (t_{c} \cdot s^{4} \cdot \frac{\exp (s)}{{(\exp (s) - 1)}^{2}} d s & Eqn 174 \\ L_{2} = \int_{0}^{Q} (t_{c} \cdot s^{4} \cdot \frac{\exp (s)}{{t_{N} (\exp (s) - 1)}^{2}} d s & Eqn 175 \\ L_{3} = \int_{0}^{Q} (1 - t_{c} / t_{N}) \cdot s^{4} \cdot \frac{\exp (s)}{{t_{N} (\exp (s) - 1)}^{2}} d s & Eqn 176 \end{matrix}$

which affords for the lattice contribution to the thermal conductivity to be determined from the following expression:

K₁=(κ_B⁴·T³/(2π·v·h_c³))·(L₁+L₂²/L3) Eqn 177

As such, the lattice thermal conductivity (K₁) can be calculated as a function of grain size and plotted as shown in FIG. 5. Furthermore, the total thermal conductivity (K_e+K₁) can be determined/calculated and plotted as shown in FIG. 6.

Finally, the figure of merit as a function of grain size can be determined as a function of grain size and plotted as shown in FIG. 9. Furthermore, and as shown in FIG. 9, ZT for various values of grain boundary barrier height and/or inclusion of a nanocomposite particle can be calculated and plotted.

$\begin{matrix} ZT = S^{2} \cdot T \cdot C / (K_{1} + K_{e}) = \frac{S^{2} T}{ρ (k_{el} + k_{p h})} & Eqn 178 \end{matrix}$

Naturally, the calculation of ZT would be repeated for various values of matrix grain size, grain boundary barrier height, grain boundary width, nanoparticle, nanoparticle amount and/or nanoparticle size that were desired to be examined. Also, such calculations would be performed by a computer such as the one illustratively shown in FIG. 18.

In this manner, researchers can estimate which matrix grain size, grain boundary barrier height, grain boundary width, nanoparticle, nanoparticle amount and/or nanoparticle size systems are more likely to exhibit relatively high ZT values and/or which compositions or range of compositions within a particular system may provide the highest ZT values. This range of compositions with the associated high ZT values can also be compared with other material properties such as mechanical property data, chemical property data and the like, in order to choose an optimum thermoelectric material composition for a given application. As such, the process provides a valuable tool to guide experimental design of thermoelectric materials.

For example, the process disclosed herein was used to design and produce a novel thermoelectric material with an improved ZT as discussed below. In particular, the process guided utilization of phonon scattering via inclusion of nanoparticles into a thermoelectric matrix to reduce phonon thermal conductivity, and also hybridization of the matrix with grain boundary modification to improve the carrier mobility, and therein the power factor.

A unique nanocomposite (referred to as BATZ) was created of a bismuth antimony telluride matrix with both zinc antimony grain boundary modifications and inter-grain phonon scattering zinc oxide nanoparticles. The power factor augmentation, in conjunction with reduction of thermal conductivity, resulted in an 83% improvement to the figure of merit ZT compared to an analogous or equivalent sample without zinc-nanostructures (referred to as BAT). In addition, it is appreciated that for the purposes of the present invention, the term “analogous” and “analogous material” refers to a material having generally the same non-oxide matrix composition and crystallite or grain size as the modified nanocomposite disclosed and discussed below and in FIGS. 19-21.

The BATZ material was made by means of a wet-chemistry synthesis that first yielded an admixture of bismuth antimony telluride nanoparticles and zinc oxide nanoparticles. This nanoparticle mixture was then consolidated, by hot pressing, to form a BATZ nanocomposite. As indicated below, the BAT nanocomposite was formed in an analogous manner, excluding the presence of zinc oxide nanoparticles and thereby precluding the formation of complex zinc-nanostructures responsible for improving the ZT from 0.6 to 1.1 (at 100° C.) as shown in FIG. 19A.

Nanoparticle Synthesis:

The synthesis of the BAT and BATZ nanoparticles was conducted as follows. A reagent solution of sodium telluride hydride was made in the following manner. Water (103 mL) and tellurium powder (5.91 g) were added to a flask degassed with inert gas, rapidly stirred and then cooled in an ice water bath. Sodium borohydride (6.32 g) was then added, in portions, and the reaction was allowed to stir for at least 12 hours until all of the tellurium powder has dissolved. The product solution was filtered through a fitted glass filter, still excluding oxygen, to collect a merlot-colored filter cake product solution. The filter cake was then washed with water (15 mL), through the fritted glass filter, and combined with an initially collected quantity of sodium tellurium hydride solution.

A solution of water and 28% ammonium hydroxide (6.5 mL and 5.5 mL, respectively) was prepared, and a combination of potassium antimony tartrate (9.02 g) and bismuth citrate (1.54 g) were dissolved completely in the diluted ammonium hydroxide solution. The antimony and bismuth salts were dissolved in portions; rigorously dissolving each portion before adding more of the salts. The freshly prepared aqueous solution of antimony and bismuth salts was then added to a reaction flask that had previously been degassed with inert gas and charged with water (480 mL). For the BATZ synthesis, a finely dispersed aqueous suspension of zinc oxide nanoparticles was added to the reaction solution (2.27 g of zinc oxide nanoparticles in 68 mL water).

The collected sodium tellurium hydride solution was then added dropwise to the rapidly stirring reaction solution containing the dissolved bismuth and antimony salts and zinc oxide nanoparticle suspension. After addition of the sodium tellurium hydride solution was complete, the reaction was allowed to stir for an additional 20 minutes. The product was then collected using centrifugation and washed under an inert atmosphere in a Soxhlet apparatus with a solution of water, methanol, and 28% ammonium hydroxide (35/165/0.8 respectively by volume). A final rinsing with methanol was administered and the methanol-slurry of nanoparticle product was dried under an inert gas flow and then ground to a fine powder, in a glovebox.

Nanocomposite Sintering:

Sintering of the composite nanoparticle powders was performed using graphite punch and dies and a hot press. All samples were first baked at 400° C. for 20 minutes and then sintered at 400° C. and 100 MPa for 4 hours under an argon atmosphere.

Temperature dependent transport properties for the BAT and BATZ nanocomposites, between room temperature and 200° C., are shown in FIG. 19A-19E. The electrical conductivity of the BATZ sample is consistently higher over the entire temperature range measured as illustrated in FIG. 19B. For example, at 150° C., the BATZ material exhibited a 34% higher electrical conductivity versus the analogous BAT nanocomposite. Hall Effect measurements were conducted to probe the electrical conductivity improvement and the carrier concentration of the BAT control nanocomposite was insubstantially 6% greater than the quantity of charge carriers in the BATZ nanocomposite. However, the BATZ nanocomposite charge carrier mobility was found to be 67% higher than that of the BAT material and thereby was determined to dominate and be responsible for the measured increase in electrical conductivity. It is appreciated that the material property comparisons are consistent with the aforementioned zinc antimony grain boundary modification of the BATZ nanocomposite and it is fundamentally different than previous reports of grain boundary modification that relied on alkali metal salts or composites made from elemental chunks. It is also appreciated that zinc oxide is n-type, which means there may be an injection of minority carriers in the system due to the zinc oxide inclusions, but the BATZ nanocomposite is undoubtedly p-type.

The Seebeck coefficient was over 200 μV/K at temperatures below 150° C. for both BATZ and BAT as illustrated in FIG. 19C. The Seebeck coefficient of BATZ was lower than that of BAT below 150° C., but then tended to be higher above 150° C. Not being bound by theory, the smaller Seebeck coefficient at lower temperature is attributed to a higher carrier concentration in the heavily doped material, while the larger Seebeck coefficient of BATZ at higher temperature indicates suppression of minority carriers (electrons). As a result of better electrical properties, the BATZ material showed an overall improvement in the power-factor throughout the measured temperature range as illustrated in FIG. 19D.

The BATZ material thermal conductivity ranged from 0.4 to 0.6 W/mK. A maximum reduction of 41% in thermal conductivity (at 150° C.) was realized from adding zinc oxide nanoparticles as illustrated in FIG. 19E. The correlation of a stronger reduction in thermal conductivity at higher temperatures is consistent with the electrical measurement data. In addition, and even though higher electronic thermal conductivity is expected for the BATZ sample, the suppression of minority carriers contributed to the reduction of bipolar thermal conduction, which caused the overall dominance of phonon-driven thermal conductivity at temperatures higher than 150° C. Therefore, the inclusion of the zinc oxide nanoparticles was an effective means of phonon scattering, in such a thermoelectric system.

Transmission electron microscopy (TEM) imaging was conducted to correlate the structural origin of the electrical conductivity increase in the presence of a reduced thermal conductivity in the BATZ material while maintaining a generally constant Seebeck coefficient when compared to the BAT material. Zinc antimony formed from a reaction of the nanocomposite constituents during the sintering process, and precipitated at the boundaries between bismuth antimony telluride grains in the BATZ material as illustrated in FIG. 20A. Composition determination was conducted with TEM-EDS and Z-contrast TEM. The zinc antimony phase identification was further supported by its identification in the XRD spectrum as shown in FIG. 21. Zinc antimony is an inter-metallic semiconductor and also a thermoelectric material. Both polycrystalline and thin films morphologies of zinc antimony show an electrical conductivity on the order of 40,000 S/m at room temperature, about 50% than that of the BATZ sample. Again, not being bound by theory, formation of zinc antimony at the host-host grains is expected to reduce the barrier potential and consequently reduce hole scattering. By reducing the interfacial electrical resistance in the grain, it results in a higher electrical conductivity and consequently increases the power factor. The presence of zinc antimony at the grain boundary does not pose a thermal resistance as its thermal conductivity is 1.5-2.5 W/mK at room temperature, which is higher than the BATZ nanocomposite. Based on the physical size of the formed zinc antimony, it is not expected to contribute to the phonon scattering phenomenon either, with an average width of 44±17 nm and length over 100 nm as illustrated in FIG. 20A. It is appreciated that the average width of the grain boundary is also the average thickness of the grain boundary.

Two other crystalline phases were identified in the XRD spectrum shown in FIG. 21, with bismuth telluride and antimony oxide present in both the BAT and BATZ nanocomposites. Comparison of peak widths, for the normalized XRD spectra, showed that the average crystallite sizes were quite similar between the two different materials. In addition, the bismuth antimony telluride lattice of the BATZ nanocomposite was slightly expanded with the inclusion of zinc-nanostructures. Such alternations to the lattice of the nanocomposite matrix were commonly observed throughout the hot-press sintered nanocomposites and the antimony oxide is believed to have formed while processing nanoparticle powders by brief contact with air. Based on reference intensity ratio analysis of the XRD spectra, the quantities of bismuth antimony telluride oxidized to form antimony oxide for BAT and BATZ materials were within 3% of each other, and therefore presumed to have an equivalent influence on the properties in both sintered materials.

Zinc oxide nanoparticles were visible throughout the bismuth antimony telluride grains via TEM as shown in FIG. 20B. Characterization was, again, based on TEM-EDS and Z-contrast TEM, and the average diameter was measured to be 10±4 nm. Neither aggregation nor anomalous growth of the zinc oxide nanoparticles was observed. As expected based on the size of the zinc oxide nanoparticles, there was no indication of zinc oxide in the XRD spectrum of the BATZ nanocomposite and the absence of peaks in the XRD spectrum is an affirmation of consistent nanometer-dimensions for the occurrences of zinc oxide throughout the nanocomposite. Antimony oxide nanoparticles ranging 12 to 18 nm in diameter (in addition to the aforementioned large grains of antimony oxide reported by the XRD analysis) were also observed by TEM in the BATZ nanocomposite. These antimony oxide nanoparticles occurred too rarely to count for a meaningful average diameter calculation. In addition, and based on their scarcity, it is thought that the antinomy oxide nanoparticles do not appreciably influence the thermoelectric properties of the nanocomposite. Such antimony oxide nanoparticles were not observed by TEM in the BAT compaction and it is not believed that zinc oxide facilitates antimony oxide nanoparticle formation, but that antimony oxide nanoparticles were not observed with TEM due to their rarity.

The thermoelectric properties of the BATZ nanocomposite described were disentangled by the addition of phonon scattering zinc oxide nanoparticles and the formation of charge carrier mobility-enhancing zinc antimony grain boundaries. This effective decoupling of the electrical conductivity, Seebeck coefficient, and thermal conductivity, as shown here on multi-gram scale, is critical for the advancement of the field and its commercial viability. And in general, these two approaches to improving the ZT value, when combined in a single nanocomposite, offer a new hybrid methodology in thermoelectric material research.

As disclosed above, the plurality of material positions to be investigated can include a first component with a volume fraction of a second component ranging from 0.0 to 1.0. In some instances, the material compositions to be investigated can include the first component with a volume fraction of the second component ranging from 0.0 to 0.7. The plurality of thermal conductivity values are calculated as a function of the scattering cross section of the second component nanoparticles for the plurality of material compositions being investigated. In addition, the scattering cross section can be a function of the interfacial surface area of the second component nanoparticles for the plurality of material compositions being investigated. The function of the plurality of material compositions being investigated can include the size of the second component nanoparticles, the size distribution of the second component nanoparticles and an interfacial property of the second component nanoparticles. In some instances, an interfacial interaction property between the second component nanoparticles and the first component can be used.

It is appreciated that the thermoelectric device can be designed and developed using the process disclosed herein, the thermoelectric device having a first electrical contact, a second electrical contact, and a thermoelectric bulk material located within an electrical path between the first electrical contact and the second electrical contact. The thermoelectric bulk material can include a first powdered component having a particulate form, the first powdered component being electrically conducting, and a second powdered component having a particulate form, the second powdered component having an electrical conductivity substantially less than the first powdered component. The first and second powdered components can retain the particulate form in the bulk thermoelectric material and the thermoelectric bulk material can be a composite that has nanostructures of the first powdered component. The first component can be a metal or a semiconductor. The second component can be an electrical insulator in the form of a ceramic. It is appreciated that the process can also be used for semiconductor-metal and semiconductor-semiconductor thermoelectric material systems.

It is further appreciated that the bulk thermoelectric material can be an electrically conducting material such as a semiconductor or metal. In addition, the electrically conducting material can be an organic material, or an organic material such as an organic semiconductor.

In the temperature range between 300K to 500K, an n-type material such as Bi₂Te₃or Bi₂Se₃and/or the p-type material such as Bi₂Te₃or Sb₂Te₃can be used for the bulk thermoelectric material. For the temperature range between 500K to 700K, n-type materials such as PbTe or SnTe doped with Bi and/or p-type materials such as PbTe or SnTe can be used. In addition, materials such as ZnSb, SiGe, CoSb, CeFeCoSb, and alloys thereof can be used for the bulk thermoelectric material. Regarding nanocomposite thermoelectric materials, nanoparticles of insulating materials such as SiO₂, ZnO, Al₂O₃, LaCoO₄, NaCoO₄, SnO₂, (ZnO)_x(In₂O₅)_y, ZrO, Y-stabilized ZrO, ZrO₂, yttria stabilized ZrO₂(YSZ), La₂O₃stabilized YSZ, other oxide materials, carbon nanoparticles, electrically insulating polymer nanoparticles, fullerenes such as C₆₀.

The invention is not restricted to the illustrative examples described above. The examples are not intended as limitations on the scope of the invention. Methods, apparatus, compositions and the like described herein are exemplary and not intended as limitations on the scope of the invention. Changes therein and other uses will occur to those skilled in the art. The scope of the invention is defined by the scope of the claims.

Number	Name	Date	Kind
5929351	Kusakabe	Jul 1999	A
7195721	Gurin	Mar 2007	B2
7734428	Banerjee et al.	Jun 2010	B2
8419980	Rowe et al.	Apr 2013	B2
20080087314	Xiao et al.	Apr 2008	A1
20080202575	Ren	Aug 2008	A1
20100059723	Lee	Mar 2010	A1
20100108115	Lee et al.	May 2010	A1
20100171546	Kwok et al.	Jul 2010	A1
20110143489	Korevaar	Jun 2011	A1
20120114961	Lee	May 2012	A1
20120138873	Zhou	Jun 2012	A1
20120298924	Banerjee	Nov 2012	A1

Number	Date	Country
1187230	Mar 2002	EP
2002076452	Mar 2002	JP
WO 2011086464	Jul 2011	JP
2010041146	Apr 2010	WO

	Number	Date	Country
Parent	13548395	Jul 2012	US
Child	14303878		US
Parent	13117286	May 2011	US
Child	13548395		US

Method of producing thermoelectric material

Information

Patent Number

Date Filed

Date Issued

Inventors

Original Assignees

Examiners

Agents

CPC

Field of Search

US

CPC

International Classifications

Term Extension

Abstract

Description

Claims

CROSS REFERENCE TO RELATED APPLICATIONS

US Referenced Citations (13)

Foreign Referenced Citations (4)

Non-Patent Literature Citations (2)

Related Publications (1)

Continuation in Parts (2)

Entry
Popescu et al., Model of transport properties of thermoelectric nanocomposite materials, Physical Review B 79 (2009).
Moyzhes et al, Thermoelectric figure of merit of metal-semiconductor barrier structure based on energy relaxation length, Applied Physics Letters, vol. 73, No. 13, Sep. 28, 1998.