Background Art A thermoelectric module based on the Peltier effect which states that when. electric current flows across a function. of two dissimilar materials, heat absorption or heat release takes place at the junction in accordance with a direction of the current has been widely used, for example, in a small-size refrigerate, an electronic cooling device, a semiconductor process chiller, etc. and is still widening its application fields because its thermoelectric elements have the advantages of rapid response and accuracy, noiselessness, simplicity and so forth.

Such a thermoelectric module usually comprises a combination of an n-type thermoelectric semiconductor element and a p-type thermoelectric semiconductor element, and a thermoelectric material as material of the thermoelectric semiconductor elements is divided into three kinds on the basis of a using temperature, that is, a material for use at a low temperature below a room temperature, a material for use at the room temperature and a material for use at a high temperature above the room temperature. Of these kinds of materials, the

material for use at the room temperature takes the greatest share in usage, and a Bi2Te3 compound or a solid solution alloy prepared by alloying the Bi2Te3 compound and Sb2Te3 or Sb2Se3 is mainly used as a p-type thermoelectric material of a thermoelectric module for use at the room temperature.

Usually, a process for producing the thermoelectric material includes a zone melting growth method, a powder sintering method using milling, a hot pressing method and the like. Among these methods, the zone melting growth method can produce a thermoelectric material having excellent thermoelectric properties, and thus is used as a major process for producing a p-type Bi2Te3-Sb2Te3 or Bi2Te3-Sb2Te3-Sb2Se3 based thermoelectric material. This is due to the fact that for the sake of effective use of specific anisotropy in the thermoelectric properties in making a thermo-module using the thermoelectric material, it is required to control a growth direction of the thermoelectric material by directional solidification and then to employ thermoelectric properties in a direction of a higher figure of merit.

In general, the figure of merit (Z) indicating the quality of the thermoelectric material is represented by the following expression : Z = a2 #/k where a (A V/h T) is the Seebeck coefficient, a is electric conductivity and K is thermal conductivity.

The thermoelectric properties of the Bi2Te3-Sb2Te3 and Bi2Te3-Sb2Te3-Sb2Se3 based alloys are mainly influenced by carrier concentration of a solidified solid phase after composition of the alloy is fixed. The carrier concentration of the alloy is determined by the amount of anti-structural defects by which the.

carriers are created. Since the amount of anti-structural defects is proportional to a degree of deviation of solid phase composition from stoichiometric composition, the solidified solid phase composition has a decisive influence on the thermoelectric properties of the alloy. The Bi2Te3-Sb2Te3-Sb2Se3 alloy in which the concentration of Bi2Te3-Sb2Te3 or Sb2Se3 is 5 % or less forms a complete solid solution, and the composition of the solid phase precipitated upon solidification of the alloy is determined by concentration of a liquid phase from which the solid phase is precipitated.

It is commonly known that the alloy maximizing the figure of merit has composition comprising 60 to 100 % Bi2Te3, 15 to 35 % Sb2Te3 and 0 to 10 % Sb2Se3 and having lower lattice thermal conductivity. Taking into account that the alloy solidified within these composition ranges contains the carriers in higher concentration than a proper value, excess Te is added to the stoichiometric composition prior to the crystal growth in order to optimize the solid phase composition during the zone melting crystal growth.

A conventional process for producing the thermal electric material by the zone melting method generally comprises the steps of adding excess Te less than 5 % by weight to an alloy having the above-mentioned composition to form an ingot of overall homogeneous composition, introducing the ingot into a quartz tube and passing a zone melting furnace through outside of the tube to grow a crystal.

When the ingot is grown in this manner, the thermoelectric properties of the ingot vary with growth speed, which is shown in FIG. 1 representing variation of the thermoelectric properties with the growth speed when a ingot is grown by the conventional zone melting method.

As seen from FIG. 1, the thermoelectric properties along a

longitudinal direction of the ingot exhibit persistency to a certain extent in a case of high-speed growth of 0.5 mm/min or more whereas the figure of merit of the ingot along the longitudinal direction is increased up to its maximum value and then is decreased again in a case of low-speed growth of 0.2 mm/min or less. At this time, the maximum value of the figure of merit in the low-speed growth is larger than the figure of merit in the high-speed growth. That is to say, the high-speed growth causes the thermoelectric material to have a relatively low, but uniform figure of merit over the ingot while the low- speed growth provides the thermoelectric material which has an excellent figure of merit in only a part of the whole ingot.

Therefore, if the required figure of merit is not so good, yield may be enhanced by growing the ingot at high-speed of 5 mm/min or more, but in order to produce a thermoelectric module requiring excellent thermoelectric properties, the low-speed growth of the ingot must be selected even though it is inevitably resigned to lowering of the. yield.

The variation of the thermoelectric properties along the longitudinal direction of the ingot in the case of the low-speed growth is due to a partition coefficient of a solute Te element.

The partition coefficient refers to a concentration ratio between two solute elements being present in an equilibrium state. The partition coefficient of the solute Te element is less than 1, which results in insoluble Te incapable of remaining in the solid phase solidified with the progress of the zone melting. This insoluble Te is discharged from a planar solid-liquid interface into a molten zone, and thus liquid phase composition and corresponding solidified solid phase composition are changed.

At an initial stage of solidification, the concentration of Te in the liquid phase is lower than an optimal value, and so

the concentration of Te in the solidified solid phase is also lower than a proper value. As the solidification goes on, the concentration of Te in the liquid phase continues to be increased to the optimal value and the solidified solid phase composition is correspondingly changed to have an optimal concentration value of Te. In this way, the solid phase having optimal thermoelectric properties can be precipitated. The concentration of Te in the liquid phase, however, is continuously increased even thereafter, and thus the concentration of Te in the solidified solid phase exceeds its optimal value. As the result of this, the figure of merit of the solid phase is continuously decreased after this point of time. In other words, the reason why the thermoelectric properties of the precipitated solid phase are not uniform is, that Te in the liquid phase forms a concentration gradient within the molten zone in the process of growth.

On the contrary, when the growth speed is as fast as 0.5 ; mm/min or more, a solid-liquid interface is not planar as is in the case of the low-speed growth, but grows in the form of a dendrite. In this case, the insoluble Te is not completely discharged into the liquid phase, but is accumulated between the dendrites, and thus constant concentration of Te is secured in the course of solidification even though the solid phase is not homogeneous.

Therefore, the present inventors have repeatedly studied on the process for producing the thermoelectric material by means of the conventional zone melting method and finally accomplished the present invention using an ingot having partially different compositions.

Disclosure of the invention Accordingly, the present invention has been made

considering the above-stated problems, and it is an object of the present invention to provide a process for producing a thermoelectric material, in which an ingot having partially different compositions is used in making the thermoelectric material by means of a zone melting method to obtain an ingot having excellent thermoelectric properties at enhanced yield.

To achieve the above-mentioned object, there is provided a process for producing a p-type Bi2Te3 based thermoelectric material at enhanced yield using an ingot of Bi2Te3 based alloy composition in a zone melting single crystal growth method in accordance with an aspect of the present invention, wherein: the ingot comprises a first ingot and a second ingot having different compositions from each other; the composition of the second ingot is the same as that of a solid phase precipitated upon solidification of a distal. end of the first ingot ; and the zone melting growth is proceeded from the first ingot to the second ingot.

In accordance with another aspect of the present invention, there is also provided a process for producing a p-type Bi2Te3 based thermoelectric material at enhanced yield using an ingot of Bi2Te3 based alloy composition in a zone melting growth method, wherein: the ingot comprises a first ingot and a second ingot having different compositions from each other; the first ingot comprises a Bi2Te3 based alloy containing excess Te of 0.5 to 30 % by weight ; the second ingot comprises a Bi2Te3 based alloy having deficient Te in amount of 0 to 3 % by weight ; and the single crystal is grown in a direction from the first ingot to the second ingot.

Growth speed of the single crystal is preferably 0.5

mm/min or less.

Brief Description of the drawings The above objects, and other features and advantages of the present invention will become more apparent from the following detailed description in conjunction with the drawings, in which: FIG. 1 is a graph comparing figures of merit (Z) of an ingot in cases of conventional low-speed and high-speed growths of the ingot; FIG. 2 is a cross sectional view schematically showing a process of growing a crystal by means of a zone melting method using an ingot according to the present invention together with changes in composition during the crystal growth; FIG. 3 is a graph showing variation of the figure of merit with amount of excess Te; and FIG. 4 is a graph comparing the figure of merit of an embodiment according to the present invention with that of a comparative example.

Best Mode for Carrying Out the Invention Hereinafter, preferred embodiments of a process for producing a p-type Bi2Te3 based thermoelectric material according to the present invention will be described with reference to the accompanying drawings. Since these embodiments are given only for the purpose of description, it will be apparent by those skilled in the art that the present invention is not limited to these embodiments.

The present invention relate to a process for producing a p-type Bi2Te3 based thermoelectric material, in which a high quality ingot is obtained at enhanced yield in making the thermoelectric material by means of a zone melting method.

Taking consideration into the above-mentioned principle,

the present inventors have succeeded in producing a thermoelectric material having a constant figure of merit even in a case of low-speed growth as does a case of high-speed growth and simultaneously having excellent thermoelectric properties, and thus solved the problem of low yield in the conventional zone melting single crystal growth method.

To this end, the present inventors have given attention to the fact that non-uniform thermoelectric properties are caused by a concentration gradient of Te formed in a liquid phase within a molten zone with the progress of crystal growth. That is, an attempt has been made to prevent the formation of the concentration gradient of Te in a solidified solid phase by maintaining liquid phase composition within the molten zone to be constant when the crystal is grown by means of the zone melting..

Now, this concept will be described in detail.

The present invention is characterized in that an ingot having different compositions between a beginning part and an end part being subjected to the zone melting is used instead of an ingot having overall uniform composition used in the conventional zone melting method.

So to speak, the zone melting is performed using two ingots of a first ingot and a second ingot having different compositions from each other.

In the course of crystal growth, insoluble Te continues to be discharged from the solidified solid phase into the liquid phase of the molten zone to gradually increase concentration of a solute Te in the liquid phase. Thus, the concentration of Te in the liquid phase reaches an optimal value at which a solid phase having the highest figure of merit can be precipitated.

An ingot having the same composition as that of a solid phase precipitated upon solidification of a distal end of the first

ingot is used as the second ingot, in a case of which Te is not discharged from the solidified solid phase into the molten zone any more, and so composition of the liquid phase within the molten zone becomes constant. If so, the solid phase precipitated thereafter also keeps its composition and thus its thermoelectric properties constant. In this way, an ascent curve of figure of merit starting from the beginning part does not descend even after it reaches a maximum value and continuously maintains the maximum value of figure of merit.

FIG. 2 is a view showing a schematic process for producing the thermoelectric material using the first and second ingots as stated above together with changes in compositions of the ingots.

In the process for producing the thermoelectric material according to the present invention, the first ingot 1 and the second ingot 2 are charged into a quartz tube (not shown) in sequence, and then a heating furnace 3 is passed through outside of the ingots 1, 2 in a direction from. the first ingot 1 to the second ingot 2 as shown in FIG. 2. FIG. 2 also depicts a composition change curve 4 of the solid phase precipitated at solid-liquid interfaces of the ingots and a composition change curve 5 of the liquid phase as well as composition change profiles 6 within the molten zone.

As growth of the first ingot 1 progresses, the solute Te is accumulated and the composition of the liquid phase at the solid-liquid interface turns composition from which a solid phase having a maximum figure of merit can be precipitated as described above. From beyond this section exhibiting the maximum figure of merit, the second ingot 2 having the same composition as that of the distal end of the first ingot is used in the zone melting. At this time, the solid phase precipitated from the second ingot 2 maintains constant composition 7 along a longitudinal direction of the ingot because no more Te is

discharged into the molten zone and Te within the molten zone is not used in precipitation of the solid phase.

Also, if composition of the first ingot is greater in amount of excess Te than that of a commonly used ingot, the section exhibiting the maximum figure of merit appears so earlier as to obtain overall higher yield. This is because speed at which the insoluble Te incapable of being melted in the solidified solid phase is discharged into the liquid phase within the molten zone and is accumulated between dendrites is increased with increasing the amount of excess Te, which can be also recognized from FIG. 3. That is, when the excess Te of 3 % and the excess Te of 5 % by weight are added to the ingot, respectively and then the ingot is grown at the same growth speed of 0.1 mm/min, a peak value corresponding to the maximum figure of merit is the same in both cases, but a peak point in the case of adding the excess Te of 5 % by weight precedes that in the case of adding the excess Te of 3 % by weight. Therefore, it is preferred that length of the first ingot is made shorter as the amount of excess Te in the first ingot is increased.

Composition of the second ingot 2 can be derived from the first ingot.

As the best way to determine the composition of the second ingot 2, it is preferable to find out the composition from a specimen prepared by growing the first ingot 1. First, the first ingot 1 having proper alloy composition is grown at low growth speed in a preliminary test to produce a thermoelectric material having the same thermoelectric properties as in FIG. 1.

A section exhibiting the highest figure of merit is taken out of the resulting thermoelectric material, composition of this section is analyzed and then the analyzed composition of the section is determined as the composition of the second ingot 2.

From this, it is found that the composition of the second

ingot 2 may be determined by various compositions of the first ingot 1.

Usually, an ingot to which the excess Te of about 5 % by weight is added is used as the first ingot 1, but it does not matter if an excessive range of the excess Te is added to the first ingot within the scope of the present invention. Even so, only the section exhibiting the maximum value is lengthened because of earlier increase in the figure of merit of the thermoelectric material produced according to the present invention, but the thermoelectric material having a constant figure of merit can be still obtained thereafter due to the second ingot 2.

The first ingot, however, preferably has Bi. sub. 2 e. sub. 3 based alloy composition containing excess Te of 0.5 to 30 % by weight in view of overall yield. There is no effect of adding the excess Te if the excess Te. less than 0.5 % by weight is added to the first ingot 1, and a high quality ingot cannot be obtained from the first ingot 1 due to precipitation of a eutectic phase if the excess Te more than 30 % by weight is added to the first ingot 1.

Once the composition of the first ingot 1 is so determined, the composition of the second ingot 2 can be found out through the preliminary test on the basis of the first ingot 1. Taking consideration into the above-stated composition range of the first ingot 1, the second ingot 2 preferably has Bi. sub. 2 Te. sub. 3 based alloy composition having deficient Te in amount of 0 to 3 % by weight.

A better understanding of the present invention may be obtained in consideration of the following example which is set forth for illustration, but is not to be construed to limit the present invention.

Example In this example, a p-type Bi2Te3 based thermoelectric material was produced by the zone melting method.

An alloy used in this example was a BizTes-SbgTes based alloy.

First, high purity (99.99 %) Bi, Te and Sb granules (5 mm or less in size) were washed with 10 % nitric acid, acetone and distilled water to remove surface oxide layers. Next, appropriate amounts of Bi, Te and Sb were mixed and weighed so as to form overall 150 g of Bi2Te3-Sb2Te3 solid solutions, and then were charged into a quartz tube with excess Te of 5 % or less by weight. An inside wall of the quartz tube was covered with a carbon coating, and the quartz tube was evacuated to 10. sup.-5 torr and sealed. The Bi, Te and Sb within the quartz tube were melted at a temperature of 800 C for 2 hours using a rocking furnace so as to ensure composition homogeneity and then,- the melt was quenched to a room temperature. The resulting., 22.5Bi2Te3-77.5Sb2Te3 solid solution ingot was grown by the zone, melting method to obtain the first ingot.

An ingot having 22.5 % Bi2Te3-77.5 % Sb2Te3 alloy composition having deficient Te in amount of 0.2 % by weight was used as the second ingot on the basis of the result obtained from the above-mentioned preliminary test.

A more detailed description will be given with reference to FIG. 2.

The first and second ingot were charged into the quartz tube (not shown) in sequence and then the furnace 3 was gradually moved from the first ingot 1 to the second ingot 2 at a velocity of 0. 1 mm/min to grow a crystal. At this time, the first ingot 1 and the second ingot 2 can be charged into the quartz tube as they are without separate joining therebetween, but adjacent sections between each other were well finished to

prevent getting loose.

The figure of merit of the resulting thermoelectric material is represented in FIG. 4.

The thermoelectric material produced according to the present invention exhibits a substantially uniform figure of merit over the whole regions except for an initial region as seen from FIG. 4. Also, the figure of merit is as high as 3.2. times. 10-3 in. Length of the thermoelectric material in FIG.

4 designates a fraction to overall length of 1, and a part having a figure of merit of 3.0. times. 10-3/K or more required for a high quality thermoelectric material can be far longer than in a comparative thermoelectric material when the fractional length is converted into absolute length.

While the process for producing the p-type Bi2Te3 based thermoelectric material at enhanced yield according to the present invention has been illustrated and described under considering a preferred specific embodiment thereof, it will be easily understood by those skilled in the art that the present invention is not limited to the specific embodiment, and various changes, modifications and equivalents may be made without departing from the true scope of the present invention.

Industrial Applicability The process for producing the p-type thermoelectric material according to the present invention is an epoch-making improvement in the zone melting crystal growth method, and use of the first and second ingots having different compositions from each other can overcome limitation of low yield accompanied with production of the thermoelectric material by the zone melting method. Thus, the process for producing the thermoelectric material using the zone melting method, which has a problem in lowering of yield in spite of excellent thermoelectric properties, will be widely and effectively used in practice by the present invention.