DE102013215662B3 - Glass-ceramic, thermoelectric component comprising the glass-ceramic and use of the glass-ceramic - Google Patents

Abstract

Glass ceramic, the glass ceramic containing the following components (in% by weight on an oxide basis): Bi2O3 20–70, SrO 15–40, CoO 15–50, CaO 0–10, Al2O3 0–15, B2O3 0–10, SiO2 0 –20, P2O5 0–20, and wherein the glass ceramic contains a main crystal phase from the ternary phase system Bi2O3-SrO-CoO.

Description

The invention relates to a glass ceramic, the use of the glass ceramic and a thermoelectric component comprising the glass ceramic.

wobei S der Seebeck-Koeffizient ist, σ die elektrische Leitfähigkeit, κ die Wärmeleitfähigkeit des Materials und T die absolute Temperatur in Kelvin.Thermoelectric materials have been increasingly explored recently as they are considered to be potential materials that generate waste heat generated from numerous sources, such as waste heat. As computers, electronic devices, the human body, motor vehicles, etc., can be converted into electrical energy, using the Seebeck effect. Furthermore, they can be used conversely for cooling by utilizing the Peltier effect. The performance of thermoelectric materials is characterized by the dimensionless figure of merit ZT:

where S is the Seebeck coefficient, σ the electrical conductivity, κ the thermal conductivity of the material and T the absolute temperature in Kelvin.

The best thermoelectric materials should have the thermal properties of glass and the electrical properties of a perfect single crystal. Semiconducting materials offer the best trade-off between Seebeck coefficient and electrical conductivity.

Numerous materials are known in the prior art, which should have the highest possible quality factor ZT. For decades, the maximum quality factors were limited to values up to 1. In 1993, Hicks and Dresselhaus (Phys. Rev. B, 47, 12727 (1993)) published a theoretical article in which the importance of electrical conductivity with a dimension of less than 3 is a key to improving the quality factor of thermoelectric materials has been described.

As a result, numerous materials were designed, either dimensioned to a smaller dimension by a microlithographic process or near phase separation. In 1997, I. Terasaki (I.Terasaki, Phys. Rev. B, 56, R12685 (1997)) reported Na _x Co ₂ O ₄ as an effective oxidic thermoelectric material. Thereafter, further oxide thermoelectric materials were discovered.

The largest surviving quality factors ZT in oxides should be above 2. Thus, Ohta et al. (Ohta et al., Nature Materials 6, 129 (2007)) that for SrTiO _3, a high-density two-dimensional electron gas contained in a SrTiO ₃ layer of unit cell thickness has a good quality factor. The quality factor ZT is estimated to be 2.4 for a two-dimensional electron gas of this structure, which corresponds to a quality factor of 0.24 for the complete three-dimensional system having the two-dimensional electron gas as the active region.

Today, about 60% of all energy resources are wasted as waste heat. If it were possible to absorb this excess heat, it would be possible to reduce the consumption of fossil fuels, the main cause of current high CO ₂ emissions. Energy could thus be used more effectively and help to find a solution to the problem of global warming. A good opportunity for energy recovery is the direct conversion of waste heat into electrical energy. So-called thermoelectric materials, which can accomplish this, use the Seebeck effect discovered in 1821 by Thomas Johann Seebeck. A thermoelectric module consists of two different thermoelectric materials, an n-doped (the conductivity is based on free negative electrons) and a p-doped (conductive due to positive holes). These materials are electrically interconnected in series and thermally connected in parallel. For example, materials commonly used in such modules are PbTe and SiGe. PbTe is currently in use in some experimental vehicles, while SiGe is used in aerospace applications.

Significant applications of thermoelectricity include the automotive industry or large industrial devices such as smelting furnaces that release a great deal of energy as waste heat. SiGe is not suitable for such applications because germanium is a very rare and expensive material that is not suitable for mass production. PbTe is used in some experimental vehicles, but will be banned in passenger cars (except the battery) from 2018 due to the toxic, polluting properties of lead. Also Tellurium is also due to its rarity and its price not for mass production. For high temperature applications (eg smelting furnaces) PbTe is unsuitable as it can only be used at temperatures up to 600 ° C.

A way out of this situation could allow certain oxides. These are non-toxic, environmentally harmless materials. The starting materials are available at very low cost. Another advantage is the good thermal and chemical stability of these materials. The disadvantage is the fact that their quality factor (ZT) is less good than other materials, such as PbTe. Since the thermoelectric Efficiency of a module increases linearly with temperature, however, oxidic thermoelectrics can be used at high temperatures.

ausgedrückt, wobei S der Seebeck-Koeffizient (eine Materialkonstante), σ die elektrische Leitfähigkeit, κ die Wärmeleitfähigkeit und T die Temperatur ist. Obwohl Oxide niedrigere ZT-Werte als manche anderen Materialien aufweisen, können sie wegen ihrer hohen thermischen Stabilität mit diesen konkurrieren. Eine bedeutende Erhöhung der Wirksamkeit wird auch durch die Herstellung Nanostrukturen aufweisender Thermoelektrika erhalten, wie dies in den theoretischen Arbeiten von M. Dresselhaus und L. D. Hicks gezeigt wurde.The quality of a thermoelectric material is determined by the dimensionless figure of merit

where S is the Seebeck coefficient (a material constant), σ is the electrical conductivity, κ is the thermal conductivity, and T is the temperature. Although oxides have lower ZT values than some other materials, they can compete with them because of their high thermal stability. A significant increase in efficacy is also obtained by the production of nanostructured thermoelectrics, as shown in the theoretical work of M. Dresselhaus and LD Hicks.

The great interest in the oxides of the transition metals as thermoelectrics arose from the discovery of NaCo ₂ O ₄ by Ichiro Terasaki in 1997. The best known thermoelectric oxide materials to date are polycrystalline ZnO (n-doped semiconductor, ZT = 0.65) and Ca ₃ Co ₄ O ₉ single crystals (p-doped semiconductor, ZT = 0.87).

Thermoelectric oxide materials are nowadays usually produced as ceramics, the advantage of which is a high electrical conductivity resulting from their purity. The disadvantage here is the fact that ceramic is porous and thus less solid. Also, it is very difficult to create nanostructures inside the material because grain growth is difficult to control during sintering processes in the ceramic manufacturing process. As a result, even a ceramic made of nanocrystalline powder often shows crystal sizes of several micrometers. In this regard, glass-ceramic is more advantageous and from the writings EP 1 174 933 A2 . EP 2 468 693 A2 . WO 2012/134759 A1 and DE 10 2008 057 615 B3 known.

The thermoelectric materials known in the prior art are unsatisfactory because they are sometimes extremely difficult to produce, are very expensive, contain toxic components and / or are mechanically unstable.

The object of the invention is therefore to provide an improved thermoelectric material with the best possible properties, which can be reproducibly produced in a simple and cost-effective manner possible.

This object is achieved according to the invention by a glass ceramic which contains the following constituents (in% by weight based on oxide): Bi ₂ O ₃ 20-70, SrO 15-40, CoO 15-50, CaO 0-10, Al ₂ O ₃ 0-15, B ₂ O ₃ 0-10, SiO ₂ 0-20, P ₂ O ₅ 0-20, wherein the glass-ceramic contains a main crystal phase of the ternary phase system Bi ₂ O ₃ -SrO-CoO.

The object of the invention is completely solved in this way. In fact, according to the invention, it has been recognized that a glass ceramic inherently has the necessary properties for producing a thermoelectric material having a high quality factor ZT. Since the glass matrix is electrically non-conductive, an electrical conductivity is achieved in an electrically conductive crystal phase portion, which is not a volume conductivity, since the glass matrix is indeed electrically non-conductive. In this way, according to the invention, a thermoelectric material is specified which, when the crystal phase fraction is electrically conductive by doping, has an electrical conductivity of lower dimension, so that an improved quality factor ZT results. According to the invention, it has therefore been recognized that a glass ceramic inherently provides the properties for a thermoelectric material with a high quality factor, provided that the doping results in electrical conductivity of the crystal phase portion.

In the following, preferred embodiments of the glass-ceramic will be described. The glass-ceramic preferably contains at least one of the following constituents (in% by weight based on oxide): Bi ₂ O ₃ 40-55, SrO 16-28, CoO 18-30 Al ₂ O ₃ 0-6, B ₂ O ₃ 0-4.

The glass-ceramic preferably contains as main crystal phase Bi ₈ Sr ₈ Co ₄ O ₂₅ , Bi ₂ Sr ₂ Co ₂ O ₉ or BiSrCo ₂ O _x , with x of 6 to 7.

The main crystal phase of the glass-ceramic is preferably free of calcium, except for unavoidable traces.

The glass-ceramic is preferably made of a glass by ceramization of the glass to obtain the glass-ceramic.

The ZT value of the glass-ceramic is preferably in the range of 0.005 to 3.

The glass ceramic is preferably used for converting thermal energy into electrical energy, in particular engine heat or exhaust heat of automobiles or waste heat from industrial installations into electrical energy.

A thermoelectric component preferably comprises a glass ceramic according to the invention.

The crystal phase content in the glass-ceramic should preferably be 10 to 80% by volume, more preferably 20 to 70% by volume.

The glass ceramic may have a p-type doping or an n-type doping.

A doping of the crystal phase on the oxygen side with fluorine leads to an n-type conductivity.

Exemplary embodiments (BSCO-1 and BSCO-2):

A glass-ceramic of the type Bi ₈ Sr ₈ Co ₄ O ₂₅ / BiSrCo ₂ O _x (BSCO-1) was obtained. After examining the properties of this material, a second glass ceramic was melted with the aim of obtaining a BiSrCo ₂ O _x (BSCO-2) glass-ceramic. For this purpose, a glass ceramic according to the invention was melted, cast and quenched between two copper plates. For both samples, complete crystallization took place after casting, whereby crystallization could only be controlled to a limited extent. Images taken with a scanning electron microscope show the high density of the glass-ceramic. X-ray diffraction patterns show the impurity of the materials resulting from the early crystallization. The thermoelectric properties of these first glass ceramics were determined and gave a figure of _merit (the so-called ZT value) of ZT _BSCO-1 , = 0.008 and ZT _BSCO-2 = 0.017 at 700 ° C.

Glass-ceramic is a promising class of new thermoelectric materials. The additional glass phase surrounding the thermoelectric phase lowers the thermal conductivity and increases the stability of this completely nonporous material. It is important that the proportion of the thermoelectric phase remains high, so that the crystals are interconnected. The glass phase should only serve to prevent pores, reduce the thermal conductivity of the material and protect it against external influences. Another major advantage of glass ceramics is that they can grow crystals of a few nanometers in size, which positively influence the thermoelectric properties.

The samples were prepared by a melt process. For BSCO-1, CoO, CaCO ₃ , Bi ₂ O ₃ and SrCO _{3 were} mixed in an atomic ratio of 2: 1: 1: 1. In addition, some Al ₂ O _{3 was} added and melted in a platinum-iridium crucible. The mixture was melted for 30 minutes at 1500 ° C and then poured onto a copper plate and quenched with a second copper plate.

For BSCO-2, CoO, Bi ₂ O ₃ and SrCO _{3 were} mixed in a molar ratio of 2: 1: 2 and some additional B ₂ O _{3 was added} to the mixture. This was melted in a platinum-iridium crucible for 30 minutes at 1500 ° C. Then it was cast on a copper plate and quenched with a second copper plate.

To identify the material structure, some SEM images were taken using the Zeiss LEO 1530 Gemini system. The X-ray diffraction patterns (XRD) were obtained with the device X'Pert Pro MPD from PANalytical. The thermal properties of the materials were measured by differential scanning calorimetry (DSC) with a DSC 404 Pegasus from Netzsch Gerätebau GmbH.

For the measurements of the temperature-dependent electrical conductivities and the Seebeck coefficients, a few 10 × 2 × 2 mm ³ rods were prepared from these materials. Some 1 mm thick disks of 10 mm diameter were made to measure the thermal conductivity.

To determine the Seebeck coefficient and electrical conductivity of BSCO-1, an Ozawa Science RZ2001i was used. The measurements were carried out in an O ₂ atmosphere in a temperature range from 50 ° C up to 700 ° C. The errors due to the measurement accuracy of the thermocouples, the sample geometry and the average of the 5 data points obtained in each measurement should be 5%.

The device Netzsch LFA 457 MicroFlash was used to measure the thermal diffusivity α. This value, together with the density ρ of the sample and its specific heat capacity cp, enables the calculation of the thermal conductivity κ by means of the formula: κ = α · cp · ρ

Because of the various measurements required to calculate the technical conductivity, the error is approximately 7.5%.

A Linseis LSR-3 was used to measure the Seebeck coefficient and electrical conductivity of BSCO-2. In contrast to BSCO-1, these measurements were carried out in air. We assume that the error is 5-10%. For the measurements of the thermal diffusivity a Netzsch LFA 457 MicroFlash was also used.

The first melt (BSCO-1) was very fluid, which is why it can be assumed that no crystallization nuclei were present in the melt. Nevertheless, complete crystallization of the melt occurred after casting. For this reason, the crystallization could not be controlled, causing an impurity of the material. As the X-ray diffraction pattern of BSCO-1 shows, the main phase consists of Bi ₈ Sr ₈ Co ₄ O ₂₅ (JCPDS 85-1501), which has good thermoelectric properties. In addition to this phase, three other crystal phases also formed, including CoO (JCPDS 48-1719), which is known to be a good insulator with a bandgap of approximately 2.5 ± 0.3 eV. The presence of CoO will therefore affect the thermoelectric properties of the material. The other two phases are Sr _0.5 CoBi _0.5 O _x (JCPDS 49-0691) and CaSrBi ₂ O ₅ (JCPDS 49-1539). It can be assumed that Sr _0.5 CoBi _0.5 O _x also has good thermoelectric properties, while CaSrBi ₂ O ₅ is detrimental to the thermoelectric properties of the material.

1 shows the X-ray diffraction pattern of BSCO-1. 2 shows the X-ray diffraction pattern of BSCO-2.

The second melt (BSCO-2) was carried out with the intention to stabilize the melt accordingly, in order to be able to obtain a proper glass first. To do this, the composition had to be mixed in some proportion to obtain a mixture lying in a glass formation region of the ternary state diagram. Estimates of such areas have been made. Furthermore, efforts were made to suppress the secondary phases, excluding Al ₂ O ₃ and adding in its place some B ₂ O ₃ which, together with Bi ₂ O _{3, has} good glass-forming activity. Finally, the phase Bi ₂ Sr ₂ Co ₂ O _{9 should be} obtained, which is why CaCO _{3 was also} excluded and its proportion was added to the SrCO ₃ .

Also, this melt was very thin liquid and crystallized after casting also completely out, which is why the crystallization could not be controlled here, which in the X-ray diffraction pattern of 2 existing impurities is recognizable. Main phase is again ST _0.5 COBi _0.5 O _x (JCPDS 49-0691). The other phases are SrBi ₄ O ₇ (JCPDS 46-0752), CoO (JCPDS 48-1719) and Sr ₃ B ₂ O ₆ (JCPDS 31-1343).

The 3 and 4 show the SEM images of the two samples, in which again the impurities of the materials are visible. Different phases are visible with different intensities. It is also obvious that both materials have no pores and high density. Phases 4 and 5 of BSCO-1 contain Bi ₈ Sr ₈ Co ₄ O ₂₅ , Sr _0.5 CoBi _0.5 O _x and some CaSrBi ₂ O ₅ . CaSrBi ₂ O ₅ and CoO are present in phases 1 and 3 and phases 2 and 6 comprise Sr _0.5 CoBi _0.5 O _x , CaSrBi ₂ O ₅ and CoO. It is apparent from this that the bright phases contain most of the thermoelectric phase. However, much of the material consists of other phases which suppress the good thermoelectric properties of the material.

Also BSCO-2 contains different phases. However, the different crystalline regions are much larger than BSCO-1. Phases 1 and 2 are made of pure CoO, the phases 3 and 4 of pure Sr _0.5 CoBi _0.5 O _x and the layers 5 and 6 made of SrBi ₄ O. ₇

As can be seen, in this sample, the content of Sr _0.5 CoBi _0.5 O _{x is} much higher than that of BSCO-1. This is the reason for the strong increase in the Seebeck coefficient. All tested points of the BSCO-1 sample have some alumina content, indicating that alumina is also present as an amorphous phase.

The temperature-dependent electrical conductivities of the two samples are in

5 shown. As can be seen, both conductivities increase with temperature, indicating a semiconductor, with the slope being much smaller in BSCO-2 than in the other case. BSCO-1 reaches a value of up to 16 S / cm at 700 ° C while BSCO-2 reaches only about 5 S / cm. The reason for the different gradients could be the different oxidation states. The electrical conductivity at high temperatures thus decreased after changing the melting process. These samples have one by one (BSCO-1) or two (BSCO-2) orders of magnitude smaller electrical conductivity than the known from conventional thermoelectric ceramic oxide values.

6 shows the temperature dependence of the Seebeck coefficient. As you can see, the Seebeck coefficient of BSCO-1 is nearly constant in this temperature range. It reaches a value of about 86 μV / K at 700 ° C. However, the Seebeck coefficient of BSCO-2 increases from about 100 μV / K at 100 ° C to up to 233 μV / K at 700 ° C. It was thus possible to significantly increase the Seebeck coefficient from one melt to the other.

The resulting performance factors of the two melts are in 7 shown. The power factor was increased in BSCO-2, despite its lower electrical conductivity.

The thermal conductivities of the two samples are not very different. Only in the temperature range between 100 ° C and 250 ° C and at 700 ° C can some significant differences be noted. For sample BSCO-2, there is a decrease with increasing temperature, as one might expect for a (p-doped) semiconductor. A value of about 1.53 W / (m · K) at 700 ° C is obtained. BSCO-1 shows a nearly constant behavior in the temperature range up to 650 ° C. Only at 350 ° C is a slight increase in thermal conductivity visible. By contrast, the value at 700 ° C is much lower than one would expect from the previous curve. This corresponds to the measurements of the specific heat capacity cp, which showed a low point at 700 ° C. The thermal conductivity decreases continuously with increasing temperature. We obtain a value of 1.28 W / (m · K) for the thermal conductivity at 700 ° C.

The corresponding quality numbers ZT are in 8th shown. For BSCO-1, a value of 0.008 at 700 ° C was obtained. At the same temperature, BSCO-2 reached 0.017, which is twice the value of BSCO-1. It was thus possible to improve the material by changing the composition of the melt. The results at 700 ° C are the most interesting as oxide thermoelectric ceramics find applications at high temperatures. With respect to this point, the electrical conductivity decreased while the Seebeck coefficient increased. The thermal conductivity also increased a bit. Despite the lower electrical conductivity, the power factor increased, resulting in a higher ZT value.

The ZT values of these glass-ceramics are 1 to 2 orders of magnitude lower than for ordinary bulk oxide ceramics. To increase this value, the melt must be stabilized in such a way as to give a proper glass after casting, which should be possible by slightly changing the compositions. Only with a glass can a controlled crystallization be obtained by special heat treatment of the glass. As a result, impurity resulting from other phases could also be suppressed and thus a single-phase glass ceramic obtained. Ultimately, the proportion of the main phase would then be much higher than for the present samples, which would result in higher electrical conductivities and higher Seebeck coefficients.

Both melts crystallized completely after casting and therefore have a significant amount of unwanted secondary phases because of the uncontrolled crystallization. The main phases of the samples were Bi ₈ Sr ₈ Co ₄ O ₂₅ / Sr _0.5 CoBi _0.5 O _x, and the thermoelectric phase ratio was higher within BSCO-2 than BSCO-1. The second material could be significantly improved compared to the first, resulting in a ZT of 0.017 at 700 ° C. To further increase this value, the electrical conductivity and Seebeck coefficient must be improved by reducing the unwanted secondary phases of the sample so that a single-phase glass-ceramic can be obtained.

In addition, it should be noted that these samples have high temperature stability. They can be thermally cycled in the entire temperature range up to 700 ° C. For the sake of clarity, however, the values measured during cooling were not shown in the figures.

Claims (8)

Glass-ceramic, characterized in that the glass-ceramic contains the following constituents (in% by weight based on oxide): Bi ₂ O ₃ 20-70, SrO 15-40, CoO 15-50, CaO 0-10, Al ₂ O ₃ 0-15, B ₂ O ₃ 0-10, SiO ₂ 0-20, P ₂ O ₅ 0-20, wherein the glass-ceramic contains a main crystal phase of the ternary phase system Bi ₂ O ₃ -SrO-CoO. Glass-ceramic according to Claim 1, characterized in that the glass-ceramic contains at least one of the following constituents of the following composition range (in% by weight based on oxide): Bi ₂ O ₃ 40-55, SrO 16-28, CoO 18-30 Al ₂ O ₃ 0-6, B ₂ O ₃ 0-4. Glass-ceramic according to claim 1 or 2, characterized in that the glass-ceramic contains as main crystal phase Bi ₈ Sr ₈ Co ₄ O ₂₅ , Bi ₂ Sr ₂ Co ₂ O ₉ or BiSrCo ₂ O _x , with x of 6 to 7. Glass-ceramic according to one of the preceding claims, characterized in that the main crystal phase is free of calcium except for unavoidable traces. Glass-ceramic according to one of the preceding claims, characterized in that the glass-ceramic is produced from a glass by means of ceramization of the glass to obtain the glass-ceramic. Glass-ceramic according to one of the preceding claims, characterized in that the ZT value is in the range of 0.005 to 3. Use of the glass-ceramic according to one of the preceding claims for the conversion of thermal energy into electrical energy, in particular engine heat or exhaust heat of automobiles or of waste heat from industrial installations into electrical energy. Thermoelectric component comprising a glass ceramic according to one of the preceding claims 1 to 6.

Images