DE102016223548B4 - Simultaneous precision method and device for measuring thermoelectric properties and contact resistance - Google Patents

Abstract

Apparatus for the simultaneous measurement of thermoelectric properties of thermoelectric materials and contact resistors comprising: a first block (2a) and a second block (2b) different therefrom, arranged such that between them the sample (1) to be analyzed is introduced and via a first Contact material (3a) with the first block (2a) and a second contact material (3b) with the second block (2b) can be brought into surface contact, - at least one first probe (4) formed as a thermocouple of two thin lines of thermocouple Alloys, in each case on the first block (2a) and the second block (2b), via which a temperature difference □ T and an electrical voltage V which occurs between the blocks (2a, 2b) can be determined, - a measuring system for Determining the values of the electrical voltage V between the blocks (2a, 2b) measured by a line pair of the at least one first probe (4) in high temporal Resolution, at least one second probe on each of the first block (2a) and the second block (2b), designed as an electrical supply line or as a pair of leads, via which a direct current through the sample or a heating current for heating one of the blocks in one or both or from one or both blocks (2a, 2b) or through one or both blocks, - at least two thermalisable potential probes (6, 7), which are at a spatial distance I, measured along the direction of current flow through the sample, during the A thermal anchorage (8) comprising a first temperature reservoir (8a) and a second temperature reservoir (8b), which are thermally connected to one another via a thermally conductive mechanical guide (8c), wherein the potential probes ( 6, 7) with the thermally conductive mechanical guide (8c) are in thermal contact.

Description

The present invention relates to a device for the simultaneous measurement of thermoelectric properties of thermoelectric materials and the electrical and thermal contact resistance and a method for the simultaneous determination of these properties by means of a corresponding device.

Current applications of thermoelectric materials include the use of thermal generators (TEGs) to convert automotive exhaust heat into usable electric power, and increasingly in industrial waste heat utilization, the use of TEG as autonomous energy sources in space missions in radionuclide thermogenerators, as alternative biothermal Energy sources in pacemakers and the electrical supply of self-sufficient micro-sensors and actuators by so-called energy harvesting.

The complete characterization of thermoelectric materials requires the measurement of several physical quantities: Seebeck coefficient, electrical conductivity, thermal conductivity. From these, the thermoelectric efficiency Z, the relevant figure of merit of a thermoelectric material, is calculated. It can also be measured directly using the Harman method. The state of the art is the determination of the aforementioned measured variables in a plurality of, typically three different measuring systems, such as, for example KA Borup et al. in Energy Environ. Sci., 2015, 8, 423. This is on the one hand error-prone, but in particular also time-consuming, which significantly slows down the development and optimization of thermoelectric materials. In addition, especially with newly developed materials usually only a small number of samples, usually only a single, are available, which is used for the analysis of all parameters / are. Thus, there is a strong need to develop measuring systems that can measure all important thermoelectric quantities simultaneously on a sample (see also J. de Boor and V. Schmidt in Adv. Mater., 2010, 22, 4303). , Recently, approaches have been taken by, among others D. Vasilevski et al. (J. of Electronic Materials, 2014: System for Simultaneous Harman-Based Mearning of All Thermoelectric Properties, from 240 to 720K, by Use of a Novel Calibration Procedure ) that demonstrate the potential of a commercial application based essentially on the classical loffe method. Simultaneous methods for measuring all thermoelectric material properties require the contacting of the sample between electrodes for the supply of the sample stream, which are used simultaneously as a heat reservoir or to impose a temperature difference. Thus, the loffe method and derived methods are all affected by uncertainties resulting from the thermal contact resistance and the position inaccuracy of the potential probes or waiving the difficulty attaching potential probes to the sample of the electrical contact resistance in their accuracy.

This results in the need for a preferably ideal, that is mechanically stable and resistance-free contacting of the sample in the sample holder, however, the production of suitable solders or protective layers for contacting proves to be practically difficult and expensive and thus a critical condition for precise measurements. It has already been possible to develop promising contacting approaches for a few important exemplary materials (see for example J. de Boor et al., Journal of Alloys and Compounds, 2015, 632, 348) However, significant efforts are still needed to expand the range of available classes of materials.

There is a realistic expectation that contacts with very good properties can be found for important sample materials. On the other hand, it is also probable that the indispensable mechanical stability of the contact can not be combined with sufficiently low contact resistance for all target systems. This is also to be assumed for the typical application situation of a measuring system in material development: When a new material system is to be examined, the user first has to face the task of applying a contact as a prerequisite for the simultaneous measurement. Extensive attempts of the user to optimize the contact can not be assumed in the usual laboratory operation. As a rule, therefore, a non-negligible disturbance of the measurement result due to the unknown thermal and electrical contact resistance must be assumed. In addition, the achievable with a few attempts contacts are also not secure stable. If necessary, they change during the measurement due to thermo-mechanical stress or limited temperature resistance.

Known methods (such as length variation, thermal potentiometer) for determining the contact resistance are out of the question for a routine measurement. In practice, usually only a single sample is available, or identical properties in repeated production are not guaranteed. The sample should not be changed for the measurement, but should be available for further measurements or further examinations. In particular, easily removable contacts are preferred.

JP 2013/214642 A discloses an apparatus for determining thermoelectric properties of a sample. The sample is clamped between two electrodes, wherein the electrodes are designed such that the sample can be fixed to them at the same time.

So far there is no simultaneous method for the measurement of all essential thermoelectric properties (Seebeck coefficient, electrical and thermal conductivity), which is not subject to considerable uncertainty due to the unknown contact resistance. The object of the present invention is to provide a corresponding method and a device for carrying out a corresponding method.

• - einen ersten Block (2a) und einen hiervon verschiedenen zweiten Block (2b), die derart angeordnet sind, dass zwischen diese die zu analysierende Probe (1) eingebracht und über ein erstes Kontaktmaterial (3a) mit dem ersten Block (2a) und ein zweites Kontaktmaterial (3b) mit dem zweiten Block (2b) flächig in Kontakt gebracht werden kann,
• - wenigstens eine erste Sonde (4), ausgebildet als Thermoelement aus zwei dünnen Leitungen aus Thermoelement-Legierungen, jeweils auf dem ersten Block (2a) und dem zweiten Block (2b), über welche eine Temperaturdifferenz □T und eine elektrische Spannung V, welche zwischen den Blöcken (2a, 2b) auftreten, bestimmt werden können,
• - eine Messanlage zur Bestimmung der von der wenigstens einen ersten Sonde (4) gemessenen Werte in hoher zeitlicher Auflösung,
• - wenigstens eine zweite Sonde jeweils auf dem ersten Block (2a) und dem zweiten Block (2b), ausgebildet als elektrische Zuleitung oder Zuleitungspaar, über welche ein Gleichstrom durch die Probe oder ein Heizstrom zum Erwärmen eines der Blöcke in einen oder beide oder von einem oder beiden Blöcke (2a, 2b) oder durch einen oder beide Blöcke geleitet werden kann,
• - wenigstens zwei thermalisierbare Potentialsonden (6,7), welche in einem räumlichen Abstand I_P, gemessen längs der Stromflussrichtung durch die Probe, während der Messung mit der Probe in Kontakt gebracht werden können,
• - eine thermische Ankerung (8) umfassend ein erstes Temperaturreservoir (8a) und ein zweites Temperaturreservoir (8b), welche über eine wärmeleitende mechanische Führung (8c) thermisch miteinander verbunden sind, wobei die Potentialsonden (6, 7) mit der wärmeleitenden mechanischen Führung (8c) in thermischem Kontakt sind.

The object underlying the present invention is achieved by a device for the simultaneous measurement of thermoelectric properties of thermoelectric materials and contact resistances, comprising:

• - a first block ( 2a ) and a second block ( 2 B ) arranged such that between them the sample to be analyzed ( 1 ) and via a first contact material ( 3a ) with the first block ( 2a ) and a second contact material ( 3b ) with the second block ( 2 B ) can be brought into contact,
• at least one first probe ( 4 ) formed as a thermocouple of two thin lines of thermocouple alloys, each on the first block ( 2a ) and the second block ( 2 B ), over which a temperature difference □ T and an electrical voltage V, which between the blocks ( 2a . 2 B ), can be determined
• a measuring system for determining the at least one first probe ( 4 ) measured values in high temporal resolution,
• at least one second probe in each case on the first block ( 2a ) and the second block ( 2 B ), formed as an electrical supply line or pair of leads, via which a direct current through the sample or a heating current for heating one of the blocks in one or both or of one or both blocks ( 2a . 2 B ) or through one or both blocks,
• at least two thermalisable potential probes (6, 7) which can be brought into contact with the sample at a spatial distance I _P , measured along the direction of current flow through the sample, during the measurement,
• - a thermal anchorage ( 8th ) comprising a first temperature reservoir ( 8a ) and a second temperature reservoir ( 8b ), which via a thermally conductive mechanical guide ( 8c ) are thermally connected to each other, wherein the potential probes ( 6 . 7 ) with the heat-conducting mechanical guide ( 8c ) are in thermal contact.

A device according to the invention which enables the simultaneous determination of the essential properties of thermoelectric materials thus comprises a first block ( 2a ) and a second block ( 2 B ). Between these two the sample to be analyzed ( 1 ) are introduced. The contact between the blocks and the sample is made by contact materials ( 3a . 3b ). The sample ( 1 ) has a length l (distance between the two-sided contacts).

1. a) in Kontakt bringen einer ersten Kontaktfläche der Probe (1) über ein erstes Kontaktmaterial (3a) mit einem ersten Block (2a) und
2. b) in Kontakt bringen einer zweiten Kontaktfläche der Probe (1) über ein zweites Kontaktmaterial (3b) mit einem zweiten Block (2b), wobei erster Block (2a) und zweiter Block (2b) ein Metall und/oder eine metallische Beschichtung aufweisen, wobei Metall und/oder metallische Beschichtung einen geringen Seebeck-Koeffizienten, typischerweise < 5 µV/K, aufweisen,
3. c) Aufbringen mindestens jeweils einer Sonde (4), ausgebildet als Thermoelement aus zwei dünnen Leitungen aus Thermoelement-Legierungen, jeweils auf dem ersten Block (2a) und auf dem zweiten Block (2b), derart, dass sie über den jeweiligen Block (2a, 2b) elektrisch leitend mit dem jeweiligen Kontaktmaterial (3a, 3b) verbunden sind,
4. d) in Kontakt bringen von Potentialsonden (6, 7) mit der Probe (1), wobei der Abstand zwischen den Sonden I_P beträgt und I_P < l ist, wobei die Potentialsonden auf der Probe (1) und mindestens jeweils eine Leitung der Sonden auf den Blöcken (2a, 2b) im Wesentlichen aus demselben Material bestehen und die Aufbringung derart erfolgt, dass während der nachfolgenden Messung kein elektrischer Strom und keine Wärme zu oder aus der Probe über die Sonden zu- oder abfließen kann,
5. e) Bestimmung des effektiven Abstandes der elektrischen Potentialsonden auf der Probe (1),
6. f) Bestimmung der Wärmeleitfähigkeit k und der thermoelektrischen Güte ZT der Probe (1)
7. g) Bestimmung des Kontakt-Wärmewiderstandes R_C ^th des Kontaktmaterials (3a, 3b) und
8. h) Bestimmung der elektrischen Leitfähigkeit.

Furthermore, the present invention relates to a method for the simultaneous determination of thermoelectric properties of a sample ( 1 ) having thermoelectric properties by means of a device according to the claims 1 to 6 wherein the sample has two spatially separated contact surfaces and the distance between these contact surfaces corresponds to the length 1, comprising:

1. a) bringing into contact a first contact surface of the sample ( 1 ) via a first contact material ( 3a ) with a first block ( 2a ) and
2. b) contacting a second contact surface of the sample ( 1 ) via a second contact material ( 3b ) with a second block ( 2 B ), the first block ( 2a ) and second block ( 2 B ) have a metal and / or a metallic coating, wherein metal and / or metallic coating have a low Seebeck coefficient, typically <5 μV / K,
3. c) applying at least one probe each ( 4 ) formed as a thermocouple of two thin lines of thermocouple alloys, each on the first block ( 2a ) and on the second block ( 2 B ), in such a way that they pass over the respective block ( 2a . 2 B ) electrically conductive with the respective contact material ( 3a . 3b ) are connected,
4. d) contacting potential probes ( 6 . 7 ) with the sample ( 1 ), where the distance between the probes is I _P and I _P <1, with the potential probes on the sample ( 1 ) and at least one line of the probes on the blocks ( 2a . 2 B ) consist essentially of the same material and the application is carried out such that during the subsequent measurement no electrical current and no heat to or from the sample via the probes can flow in or out,
5. e) Determining the effective distance of the electric potential probes on the sample ( 1 )
6. f) determination of the thermal conductivity k and the thermoelectric quality ZT of the sample ( 1 )
7. g) determination of the contact thermal resistance R _C ^{th of} the contact material ( 3a . 3b ) and
8. h) Determination of electrical conductivity.

The process according to the invention can be carried out at room temperature (usually 20 ° C. to 25 ° C.). However, it is also possible to determine the temperature dependence of the relative or absolute contact heat resistance, in which the inventive method at temperatures of 4 K to 1200 ° C, preferably from 80 K to 600 ° C, particularly preferably from room temperature to 300 ° C, performs. The process according to the invention preferably takes place at temperatures of from 4 K to 1500 K, in particular from 80 K to 900 K, preferably from 300 K to 600 K.

Hereinafter, both embodiments are explained, wherein features of one embodiment can also be combined with features of the other embodiment and all features with each other, unless explicitly described otherwise.

First contact ( 3a ) and first contact material ( 3a ) and first contact ( 3a ) are used interchangeably in the present application. While "contact" and "contact" focus on physical properties (contact between block and sample), "contact material" focuses on the material that gives these properties. Understand this is the same. The same applies to second contact ( 3b ) and second contact material ( 3b ) as well as second contacting ( 3b ). If first and second contact are covered by a description, only "contacts" or "contact materials" or "contacts" are used. Similarly, the term "blocks" always includes both the first block ( 2a ) as well as the second block ( 2 B ), unless explicitly stated otherwise.

The core of the invention is thus a thermoelectric measuring method for the simultaneous determination of all essential thermoelectric material properties with increased accuracy by integrated determination and correction of measurement errors, which are caused by a non-negligible electrical and thermal contact resistance and a device suitable for this purpose. By integration of additional potential probes in known measuring arrangements, preferably via spring force against the sample ( 1 ), a method for determining the electrical contact resistance can be integrated into the measurement, which makes its costly microscopic-visual determination dispensable via a special thermoelectric determination of the effective probe distance. At the same time, a transient method for determining the thermal contact resistance is included by providing the measuring system with fast time-resolved signal recording. As a result, the falsification of the measured values of electrical and thermal conductivity can be eliminated by the contact resistance; At the same time, the measurement of the thermoelectric efficiency according to the Harman method and the measurement of the Seebeck coefficient can be carried out unadulterated by the thermal contact resistance. The additional measurements for the elimination of the contact resistance can be carried out accompanying the thermoelectric property measurement at each temperature point of the temperature-dependent measurement series of these quantities. Thus, the consideration of the influence of contact can be done correctly even with a temperature-dependent or temporal change of the contact resistance or the sample properties.

With the availability of the correction in terms of the influence of the contacts, a measurement of novel and unknown materials is already rewarding, if no costly development of a good contact has yet occurred. This is typically the case in the development of new thermoelectric materials. By tolerating even moderately good contacts, which can often be found with acceptable effort for a new material, results for the measurement a nearly universal suitability for a variety of material classes. Only one single sample is required for each measurement, which is easy to install in the measuring apparatus and is available for further investigation / use after the measurement without mechanical processing / alteration (beyond the contacting).

The sample is a material with thermoelectric properties. Thus, in principle, the measurement of temperatures and voltages, in particular of thermal voltages, of interest. To make this possible, both the first block ( 2a ) as well as the second block ( 2 B ) at least one first probe ( 4 ), with which the temperature and / or the voltage can be determined. For the temperature measurement, the sensor consists, for example, of two lines which form a thermocouple and are usefully considered together as a probe. These can be mounted in a common point on the respective block or, more advantageously, for the sake of the desired isotherm at the measuring point, in the interior of the same. As a potential probe, ie for measuring an electrical potential, the probe ( 4 ) may be formed as a thin metal wire alone.

In the determination of the thermal conductivity, the thermoelectric quality and the contact thermal resistance has been shown that high-resolution time voltage measurements must be present in order to detect system-inherent errors and to be able to correct them. Therefore, the device according to the invention comprises a measuring arrangement for determining the pair with the at least one first probe ( 4 ) of each block obtained in high temporal resolution. The measuring arrangement thus detects functional elements for determining the temperature and / or the voltage. In particular, the voltage that occurs between the first and second block must be determinable with high temporal resolution.

Furthermore, the first ( 2a ) and the second block ( 2 B ) in each case at least one second probe, designed as an electrical supply line or supply line pair, via which a direct current through the sample ( 1 ) or a heating current for heating one of the blocks into one or both or from one or both blocks ( 2a . 2 B ) or through one or both blocks. To get current through the blocks through the sample ( 1 ) is in particular a power supply on the first block ( 2a ) and the second block ( 2 B ) appropriate. This may be a line of the aforementioned lead pair of the second probe. This then allows power from the first block ( 2a ) via the first contact material ( 3a ) through the sample ( 1 ) through the second contact material ( 3b ) to the second block ( 2 B ) flow. Here the current then emerges through the corresponding power supply.

To set a temperature difference between the first block ( 2a ) and the second block ( 2 B ) can either one or both a heating device ( 9 ) contain. This can be arranged inside the blocks. This may be, for example, a resistor that heats up when a current flows through. However, any other heating elements are conceivable.

A temperature control of the first and / or the second block can also be effected by an external heating device, which is able to heat the respective block homogeneously.

Furthermore, the device according to the invention has at least two thermalizable potential probes ( 6 . 7 ), which are at a spatial distance I _P from each other, measured in the direction of the current or heat flow through the sample, and during the measurement with the sample ( 1 ) or can be brought into contact. The distance between these potential probes ( 6 . 7 ) is smaller than the length l of the sample ( 1 ), measured in the direction of the current or heat flow.

The potential probes are in thermal contact with an anchor ( 8th ). The anchor comprises a first temperature reservoir ( 8a ), a second temperature reservoir ( 8b ) and a thermally conductive mechanical guide ( 8c ), which thermally interconnects first and second temperature reservoirs.

If the device has an external heating device, then this is preferably also suitable, not only the first and / or the second block (FIG. 2a . 2 B ), but also the temperature reservoirs ( 8a . 8b ) to temper. For this purpose, several heaters may be present. Likewise, the temperature reservoirs can also be tempered by internal heating elements.

In order to determine the effective distance of the electric potential probes, it is necessary to distinguish between the first block ( 2a ) and the second block ( 2 B ) a temperature difference can be adjusted. For this, the first block ( 2a ) brought to a first temperature T ₁ . The second block in this case has a second temperature T ₂ , wherein T ₁ and T ₂ are different from each other. The thermal anchor ( 8th ), with which the potential probes ( 6 . 7 ) are thermally connected, in this case has the same temperature difference. The first temperature reservoir ( 8a ) thus also has the temperature T ₁ , the second temperature reservoir ( 8b ), The second temperature T _2. The probe closer to the first temperature reservoir on the guide ( 8c ) is placed on the sample ( 1 ) then in the direction of the first block ( 2a ) applied. The probe ( 7 ), which are spatially closer to the second temperature reservoir ( 8b ) is determined accordingly on the sample ( 1 ) spatially in the direction of the second block ( 2 B ) applied. This will ensure that the temperatures of the probes, as far as they are in contact with the sample ( 1 ), have a similar temperature gradient as in the interior of the sample ( 1 ) prevails.

Preferably, the sample comprises ( 1 ) not only a thermoelectric material, but consists essentially of this. The purity of the sample is not relevant to the method according to the invention. In principle, the sample should be as homogeneous as possible, so that the ratio of electrical to thermal conductivity is not locally subject to large fluctuations. If the electrical conductivity is not too low, however, this is basically already ensured by the Wiedemann-Franz law. In addition, this problem also arises in measuring methods known in the prior art.

The first block ( 2a ) and the second block ( 2 B ) consist essentially of the same Material or have the same coating. Contacts ( 3a . 3b ), Blocks ( 2a . 2 B ) and at least one of the probe lines on each of the blocks preferably have the same Seebeck coefficient, preferably a very small value. As far as it is very small against the Seebeck coefficient of the sample, it need not be known exactly. While the materials of the blocks ( 2a . 2 B ) and the probe lines are selected accordingly, the Seebeck coefficient of the contacts ( 3a . 3b ) often unknown. With metallic contacts ( 3a . 3b ), however, a sufficiently low Seebeck coefficient of usually less than 5 μV / K is to be expected and realized. As material for blocks ( 2a . 2 B ), Contacts ( 3a . 3b ) and at least one of the probes on each block is selected, in particular, such a material which has a Seebeck coefficient of 10 μV / K or less. The Seebeck coefficient S is preferably in the range from 0 μV / K to 10 μV / K, in particular from 0 μV / K to 8 μV / K or from 0 μV / K to 5 μV / K. The Seebeck coefficient can assume positive or negative values. The values mentioned in the present application for each embodiment are therefore absolute values.

Blocks ( 2a . 2 B ), Contacts ( 3a . 3b ) and probes preferably have an electrical conductivity of at least 10 ⁴ S / cm, preferably from 10 ^{4 to} 10 ⁹ S / cm, in particular from 10 ⁵ to 10 ⁸ S / cm or from 10 ⁵ to 10 ⁷ S / cm.

Particularly preferably, the blocks have ( 2a . 2 B ) a thermal conductivity of at least 10 W / (m · K), in particular of at least 100W / (m · K). The contacts ( 3a . 3b ) are usually part of the unknown sample installation. Significant as the thermal conductivity is the specific thermal contact resistance, or heat transfer coefficient is preferably in the range of 10 ³ to 10 ⁵ W / m ² K. The probes should be as low heat-conducting as possible, which can be made possible by a small diameter of the probe leads. The thickness of the probes depends on the individual structure of the device and the sample to be determined.

Metallic materials in the sense of the present invention are those elements which are located in the periodic table of the elements on the left and below a dividing line between boron and astatine. In this case, alloys and intrametallic phases, which in particular the aforementioned properties, namely low Seebeck coefficient, high electrical conductivity and high thermal conductivity, comprise thereof.

Particularly suitable and preferred materials for blocks ( 2a . 2 B ) and probes are Cu, Ag, Au, Pt, Fe, Ni, Al, Sn, Zn, Pb or mixtures of these. Suitable materials include Bi, Sb, as far as their mechanical properties are sufficient, but also special thermocouple alloys such as Konstantan, Chromel, platinum-rhodium. Is in the present application of a material for the blocks ( 2a . 2 B ), the blocks may consist essentially of this material or have a coating of this material, which then the electrical properties of the blocks ( 2a . 2 B ) defined for the inventive method, provided that the volume of the block is non-conductive. The thermal properties of the block are primarily determined by the material that makes up the bulk fraction of the block.

The contact material ( 3a . 3b ), as well as the material of the first and / or second block, is a metallic material having the aforementioned properties in terms of Seebeck coefficient, electrical conductivity and thermal conductivity. Particularly preferably, the contact material is liquid at room temperature and normal pressure or its melting temperature is in the range between 80 K and 600 ° C. This allows easy handling during the measurement. The sample can be placed between the block ( 2a . 2 B ) can be easily inserted and removed without having to sample ( 1 ) and / or block ( 2a . 2 B ) are subjected to a special mechanical and thermal stress. Thus, the mechanical stress of the sample is low, which is especially true for thin samples ( 1 ) is relevant.

The sample ( 1 ) comprises a thermoelectric material and consists in particular of it. The Seebeck coefficient of the sample differs from that of at least one of the probes per block to allow for measurement. The Seebeck coefficient of the sample is therefore preferably greater than 10 μV / K, preferably in the range from 25 μV / K to 500 μV / K, in particular in the range from 50 μV / K to 300 μV / K or from 70 μV / K up to 250 μV / K, especially from 100 μV / K to 200 μV / K.

Preferably, the thermoelectric material of the sample ( 1 ) selected from skutterudites, clathrates, half-Heusler compounds, Zintl compounds, quaternary chalcogenides, tellurides, in particular bismuth telluride, PbTe, SnTe and their mixed crystals and nanomaterials based thereon, for example LAST (lead silver antimony telluride), TAST (tin antimony silver telluride), BTST ( Bismuth silver silver telluride), and so on; Silicides, in particular of the elements magnesium, manganese, iron, chromium and other transition metals and their mixed crystals, in particular including stannides and germanides; Sulfides, in particular of titanium, tin and so on, and antimonides, in particular Zn ₄ Sb ₄ , ZnSb, MgAgSb, and many others.

The electrical conductivity of the sample ( 1 ) is preferably 10 ⁴ S / cm or less, in particular 10 ¹ to 10 ⁴ S / cm, preferably from 10 ² to 10 ³ S / cm. The thermal conductivity of the sample is preferably 20 W / (m · K) or less, in particular in the range of 0 to 15 W / (m · K), preferably from 0.2 to 10 W / (m · K), particularly preferably from 0 , 5 to 5 W / (m · K).

By combining the loffe method and its known variants with potential probes, which in particular by means of mechanical spring force to the sample ( 1 ), with a thermoelectric position determination of the effective probe distance and with a transient determination of the thermal contact resistance, a precision method is obtained, which is no longer falsified by the influence of the contact resistance. Thus, the main obstacle to a much wider dissemination of the simultaneous thermoelectric measurement method is eliminated as previously and at the same time a measurement method with low preparation of the sample created, excellent for serial use in routine laboratory operation and thus also for provision as a ready-made laboratory equipment, including in the commercial Sales, is suitable.

The use of peaks impressed by spring force on the sample as potential probes ( 6 . 7 ) avoids the usually difficult - and in the choice of the probe material and the process parameters material specific - attachment of thin probe wires by welding. The inherent isothermic nature of each probe (6,7) during the thermoelectric measurement of the probe distance can be realized by thermal anchoring according to the position-related intermediate temperature between the two block temperatures. In addition, in the neighborhood of the blocks ( 2a . 2 B ) of the sample holder, which are in electrical and thermal contact with the sample on both sides, thermal anchor points ( 8a . 8b ), for example in the form of thermostatically controlled blocks ( 8a . 8b ) intended.

For the thermoelectric determination of the probe positions (6,7), which is carried out under non-isothermal, but stationary conditions, the bilateral reservoirs of the thermal anchorage ( 8th ) adjusted to the block temperatures. For example, these reservoirs can be realized by cup-shaped umbrella heaters that block the blocks ( 2a . 2 B ) of the sample holder are surrounded and thermally separated from the blocks by insulating material.

Parallel to the sample ( 1 ) is between the thermal anchor points in the reservoirs ( 8a . 8b ) a thermally conductive mechanical guide ( 8c ), for example in the form of a profile connecting rod, mounted and thermally anchored both sides thermally in the reservoirs, resulting in the leadership ( 8c ) forms a temperature gradient corresponding to the temperature distribution in the sample. The fixation of the leadership ( 8c ) in the blocks ( 8a . 8b ) is preferably mechanically detachable and positionally displaceable according to the different possible sample lengths. On this tour ( 8c ) are also positionally displaceable and thermally anchored, narrow terminals or similar devices in which the potential probe tips ( 6 . 7 ) are guided flexibly and resiliently perpendicular to the sample surface. In this way, the temperature of the guide of each probe tip corresponds almost ideally to the undisturbed temperature in the sample ( 1 ) at the touchdown position of the probe tip, and parasitic heat flows from / to the contact points of the probes ( 6 . 7 ) are avoided during the measurement for position determination.

During measurements for electrical conductivity, any offsets due to thermal voltages do not play a disturbing role, since the electrical measurements are usually carried out as alternating current measurements.

The two pairs of potential probes used for these measurements in steps c) and d) on the first and second blocks ( 2a . 2 B ) and the sample ( 1 ), consist essentially of the same material. The potential probes are metallic probes.

The method according to the invention makes it possible to determine the effective distance of the potential probes ( 6 . 7 ) on the sample ( 1 ), the determination of the thermal conductivity and the thermoelectric quality and the determination of the contact heat resistance of the contact material and the determination of the electrical conductivity. By determining the effective probe distance, the precise determination of the electrical contact resistance is possible, and thus its separation and correction of the electrical conductivity. Thus, all the essential properties of a thermoelectric sample can be performed simultaneously in one device in one process. According to the invention, the steps e), f), g) and h) of the process according to the invention can be carried out in any order. Care must be taken to ensure that possibly switched on previously heat or electricity flows must be switched off if necessary.

• e1) Einstellen einer Temperaturdifferenz ΔT zwischen dem ersten Block und dem zweiten Block,
• e2) Bestimmung der Thermospannung V₁ zwischen dem ersten Block und dem zweiten Block mittels der auf den Blöcken jeweils aufgebrachten Sonden,
• e3) anschließend oder gleichzeitig mit e2) und unter denselben Temperaturbedingungen wie in Schritt e2), Bestimmung einer Thermospannung V_P mittels der auf der Probe aufgebrachten elektrischen Potentialsonden, und
• e4) Bildung des Verhältnisses V_P/V₁ zu Bestimmung der effektiven Sondenposition.

Step e) of the method according to the invention preferably comprises the following steps:

• e1) setting a temperature difference ΔT between the first block and the second block,
• e2) Determining the thermal voltage V ₁ between the first block and the second block by means of the probes applied to the blocks,
• e3) subsequently or simultaneously with e2) and under the same temperature conditions as in step e2), determination of a thermal voltage V _P by means of the electric potential probes applied to the sample, and
• e4) Formation of the ratio V _P / V ₁ to determine the effective probe position.

To perform the necessary measurements, each on the first ( 2a ) and the second block ( 2 B ) as well as on the sample ( 1 ) applied electrical potential probes of substantially the same material. These probes allow the thermoelectric voltage between the first ( 2a ) and the second block ( 2 B ) as well as within the sample ( 1 ). For this purpose there is at least one probe on the first block and at least one probe on the second block. The thermoelectric voltage V ₁ in step e2) of the method according to the invention is then determined via this.

The arrangement for determining the effective probe position corresponds in principle to a typical 4-probe arrangement for measuring the electrical conductivity in prismatic geometry, wherein the electrical probe leads are guided on the sample largely isothermally to the contact points. A sample with the Seebeck coefficient S, with the electrical resistance R and the thermal conductivity K is formed by metallic contacting, for example by liquid metal contacts, soldering, brazing, diffusion welding or the like (contacts ( 3a . 3b ) in 1 enlarged), between two metal blocks ( 2a . 2 B ) attached. Metals have several orders of magnitude higher electrical conductivity than typical thermoelectric materials; the blocks ( 2a . 2 B ) therefore do not significantly contribute to the ohmic potential drop when current is passed through the blocks across the sample. Between the blocks ( 2a . 2 B ) can be adjusted via integrated heaters or by external heating a temperature difference. To measure the electrical conductivity, blocks ( 2a . 2 B ) and sample ( 1 ) are kept isothermal, if this is done with a DC method. However, the electrical conductivity is preferably measured by means of conventional alternating current methods. This eliminates the influence of interfering thermal voltages. The feeding of the measuring current can be done in one embodiment via thin wire-shaped metallic power supply lines to the blocks. Between the potential probes on the blocks with the designation V ₁ , the voltage V ₁ is formed when the current flows. The measuring voltage V _p for determining the conductivity of the sample is measured between metallic potential probes, consisting of the same material as the probes V ₁ , which are mounted at a distance I _P from each other on the sample. If one also wishes to separate the electrical contact resistance, which is preferred in the present case, the measuring current is supplied via additional lines which are different from the lines V ₁ at each of the two blocks. In this case, all thermoelectric quantities of the sample ( 1 ). If the supply of the measuring current via the probes V ₁ , only probe distance and conductivity can be determined. Therefore, it is preferred to provide separate lines for the supply of electricity.

The conditions according to the invention enable an exact determination of the divider ratio V _p / V ₁ . This can be used both for the correct determination of the electrical conductivity and for measuring the electrical contact resistance itself.

The distance between the potential probes on the sample ( 1 ) I is _P. This refers to the distance in the direction of the block distance, ie the distance between the probes in the direction of the distance between the blocks ( 2a . 2 B ).

If the sample has a straight prismatic geometry and is contacted at its base surfaces over the entire surface and homogeneously, ie if a one-dimensional model of heat and current flow is valid, the ratio V _P / V ₁ corresponds to the aspect ratio I _P / I and the effective distance of the electrical Potential probes on the sample. Preferably, therefore, the sample has the shape of a straight prism. In this preferred embodiment, step e4) comprises forming the ratio V _P / V ₁ and therefrom the aspect ratio I _P / I, which corresponds to the effective distance of the electrical potential probes on the sample.

In the one-dimensional case, the effective probe position in the sense of the present invention is the location coordinate at which the potential value real measured by the probe is present in the potential field undisturbed by the measurement; For complicated geometries, the effective probe position corresponds to a mostly curved equipotential surface.

On the test ( 1 ) itself, at least two potential probes are applied, through which the measurement of a thermal voltage V _{P is} possible. The application of the probes to the sample takes place in such a way that during the measurement no electrical current or heat to or from the sample ( 1 ) can flow in or out via the probes. If heat or electric current were flowing in or out via the probes, this would have a not insignificant influence on the result of the measurement. Therefore, it is essential for the process that this is precisely what is prevented. The sample ( 1 ) includes thermoelectric material. If electric current were to flow through the probes into the sample, the current along the probes would cause an electrical voltage drop, which would falsify the measured thermal voltage. In addition, due to this current flow, a temperature difference would be set in the thermoelectric sample (Peltier effect). This would influence the temperature difference set in step e1) of the method according to the invention, so that it would no longer be possible to determine the effective positions. The same applies if heat enters the sample through the probes.

After connecting the probes to the sample or blocks, a temperature difference ΔT between the first block and the second block is set. For this purpose, the first block is brought to a temperature T ₁ and the second block to a temperature T ₂ , wherein T ₁ and T ₂ are different from each other. Through the contact material ( 3a ) heat is transferred from the first block ( 2a ) into the sample ( 1 ). As the temperature on the opposite side of the sample ( 1 ) is different from this first temperature, a temperature gradient sets in the interior of the sample. Due to the thermoelectric properties of the sample ( 1 ) creates an electric field in the interior of the same. The associated electrical potential is proportional to the temperature field in the sample at small temperature differences ( 1 ). The temperature difference across the sample should therefore not exceed about 10K.

Between the probes, which are located on the first and the second block, a thermoelectric voltage V ₁ can now be determined. Likewise, if possible simultaneously, but necessarily under the same temperature conditions, a thermal voltage V _{P can be determined} with the probes which are applied to the sample. The thermoelectric voltage V ₁ includes the total temperature difference across the sample between the two contact points as well as the properties of the contact material and the material of the first and second block, but make only a negligible contribution to the thermoelectric voltage according to the selected materials of the arrangement. The thermal voltage V _P is only dependent on the property of the sample and the temperature difference between the effective probe positions.

For a precise measurement of the conductivity, which takes place in step h) of the method according to the invention, the correct potentiometric divider ratio of the sample section enclosed by the probes is decisive - which, according to the invention, a step e) is determined. This is determined by the effective potential values at the probes when applying a unit voltage under isothermal conditions. Under ideal conditions theoretically their exact position information could also be determined in this way. The application of a defined and known voltage across a conductive sample ( 1 ) is practically impossible because of the ever-occurring contact resistance; the current flow leads to a voltage drop at the contact resistance, which is actually in the sample ( 1 ) reduces overall voltage over the voltage applied from the outside. A purely electrical determination of the effective probe positions from the potentiometer ratio of the block and sample probe pairs is thus basically not possible with current flow through the sample; only the measurement of the currentless generated thermoelectric voltages makes this possible. For highly conductive samples - this is typical for thermoelectric materials - the electrical contact resistance can certainly reach the order of magnitude of the sample resistance.

The potential field of the thermal voltage (electrical voltage associated with a temperature difference caused by the Seebeck effect) is formed in the sample, regardless of the electrical contact resistance, since it is generated electrically de-energized, solely by imposing the temperature difference. Via contact and block materials, which show a vanishing Seebeck coefficient, the total thermal voltage across the sample can be measured without interference between the contact surfaces, ie the current entry locations of the conductivity measurement; the potentiometer ratio of the thermoelectric voltage thus provides the exact effective probe positions in the temperature field.

The temperature drop across the always existing thermal contact resistance at the current entry location causes no distortion, as long as the associated thermoelectric voltage - because of the very low Seebeck coefficient of contact - can be neglected. With metallic contacting this is very well met; In the case of oxide coatings on pressure contacts, on the other hand, considerable interference voltages can result. Thus, by exploiting the thermoelectric properties of a sample whose electrical properties are to be measured by the widely used 4-point method, one can obtain the correct potentiometric divider ratio, that is, the effective probe positions (for prismatic / one-dimensional arrays: the effective probe spacing) Voltage ratios at four points (one probe outside the sample at the current entry / exit points and two probes at a sufficient potential distance from each other on the surface of the sample between the entry points) determine, wherein the assembly is in a de-energized state in a temperature difference.

In addition, the same probes can be used to measure current flow through the sample under isothermal conditions or with an alternating current method. This results in changed Potentiometerverhältnisse compared to the previous case, since now the electrical contact resistance makes a not negligible contribution. From the difference between the two measurements, the electrical contact resistance can be separated. At the same time, the electrical conductivity of the sample is obtained, free from adulteration by the contact resistance or by uncertainty of the probe position. In addition to the exclusion of the electrical contact resistance as a source of interference in the measurement of electrical conductivity, the principle can be used specifically for the measurement of the contact resistance, for example in the context of Kontaktierungsentwicklungen.

If one introduces the relative electrical contact resistance r = R _K / R as the ratio of the electrical contact resistance to the sample resistance and assumes a negligible Ohmic voltage drop across the blocks and power leads (this is given if one does not use the potential probe on the block itself but a separate power line for block leads), as are the Potentiometerverhältnisse the thermal voltage measurement (V _P / V _{1) □ T,} and the electrical measurement (V _P / V _{1) I} each other in the relation (1 + r) (V _P / V ₁ ) _{□ T} = (V _P / V ₁ ) _I , that is to say in the electrical measurement a potentiometer ratio increased by the relative contact resistance is obtained compared to the currentless thermoelectric voltage measurement. If one measures both ratios, one can calculate the relative contact resistance.

• f1) gegebenenfalls Beenden der Wärmezufuhr aus Schritt e1),
• f2) Erreichen eines thermischen Gleichgewichtszustandes,
• f3) Änderung der Stromstärke zum Zeitpunkt t₀, und
• f4) zeitlich aufgelöste Messung der Spannung über einen Zeitraum unmittelbar nach Abschalten des Stromes, wobei der Zeitraum wenigstens einer Halbwertszeit des raschen Abklingens und mit einer Abtastrate erfolgt, die es ermöglicht, innerhalb der ersten Halbwertszeit wenigstens 10 bis 100 Messwerte (Spannungswerte) aufzuzeichnen, oder mittels einer anderen Methodik, die es gestattet, auf den Anfangswert der relaxierenden Messspannung unmittelbar nach Abschalten des Stromes zurückzuschließen, sowie einen möglichst frühen Zeitpunkt anzugeben, zu dem die anfängliche rasche Relaxation praktisch restlos abgeklungen ist,
• f5) anschließende zeitaufgelöste Messung der Spannung über einen Zeitraum, in welchem sich das System im Wesentlichen auf den neuen Gleichgewichtszustand einstellt, mit einer Datenrate von mindestens 5 bis 20 Werten pro Halbwertszeit des langsamen Abklingens beim Einlaufen in den neuen Gleichgewichtszustand, oder mittels einer anderen Methodik, die es gestattet, auf den Anfangswert der zweiten, langsameren Relaxation bezogen auf den Zeitpunkt der Änderung der Stromstärke to sowie auf deren Endwert nach ihrem vollständigen Abklingen zurückzuschließen,
• f6) numerische Anpassung der Daten aus den Schritten f4) und f5) und zeitliche Rückextrapolation des Spannungsverlaufes der zweiten, langsameren Relaxation aus Schritt f5) zum Erhalt einer Basislinie bis zum Zeitpunkt to sowie zeitliche Extrapolation des Spannungsverlaufes der zweiten, langsameren Relaxation zum Erhalt des Endwertes nach dem vollständigen Abklingen,
• f7) Bilden der Differenz aus dieser Basislinie und den experimentellen Daten aus Schritt f4), wobei die Differenz des Spannungswertes bei to zur Basislinie bei to der Amplitude A₂ entspricht und die Differenz der Basislinie bei to zum Endwert der Amplitude A₁ entspricht, und Bildung des Verhältnisses A₂/A₁, welches den relativen systematischen Fehler einer Wärmewiderstandsmessung an der Probe verursacht durch den Wärmewiderstand des Kontaktes angibt,
• f8) Bestimmung der Wärmeleitfähigkeit k und der thermoelektrischen Güte ZT unter Berücksichtigung des relativen systematischen Fehlers aus Schritt f7).

Preferably, step f) of the method according to the invention comprises the following steps:

• f1) optionally terminating the heat supply from step e1),
• f2) reaching a thermal state of equilibrium,
• f3) change of current at time t ₀ , and
• f4) time-resolved measurement of the voltage over a period of time immediately after switching off the current, wherein the period of at least one half-time of the rapid decay and at a sampling rate, which makes it possible to record at least 10 to 100 measured values (voltage values) within the first half-life, or by means of another methodology which makes it possible to deduce the initial value of the relaxing measurement voltage immediately after the current has been switched off, and to indicate the earliest possible time at which the initial rapid relaxation has subsided almost completely,
• f5) subsequent time-resolved measurement of the voltage over a period in which the system substantially conforms to the new equilibrium state, with a data rate of at least 5 to 20 values per half-time of slow decay when entering the new equilibrium state, or by another methodology which allows one to deduce the initial value of the second, slower relaxation with respect to the time of change of the current t0 and its final value after its complete decay,
• f6) numerical adaptation of the data from the steps f4) and f5) and temporal re-extrapolation of the voltage curve of the second, slower relaxation from step f5) to obtain a baseline until the time to and temporal extrapolation of the voltage curve of the second, slower relaxation to obtain the final value after the complete decay,
• f7) forming the difference from said baseline and the experimental data from step f4), wherein the difference of the voltage value at to the baseline at t corresponds to the amplitude A ₂ and the difference of the baseline at to the final value corresponds to the amplitude A ₁ , and formation the ratio A ₂ / A ₁ , which indicates the relative systematic error of a thermal resistance measurement on the sample caused by the thermal resistance of the contact,
• f8) determination of the thermal conductivity k and the thermoelectric quality ZT taking into account the relative systematic error from step f7).

The system of blocks ( 2a . 2 B ), Contact materials ( 3a . 3b ) and sample ( 1 ) is in a thermal equilibrium state at the beginning of the measurement in step f2). This thermal equilibrium state is characterized by a temporally constant temperature difference between the blocks; a slight residual drift in block temperatures may persist, depending on the strength of the flowing sample stream. If a measurement is previously performed as described in step e), it is necessary to wait until the equilibrium state has set. By changing the current, the system is also brought out of this state of equilibrium. After a certain period, the system then reaches a new equilibrium state.

The system can according to the invention in a preferred embodiment step f2) in a state of equilibrium in which no current flows. The change then occurs by supplying a DC current across the blocks by means of a first pair of wires so that the current flows from one block to the other block, the current flowing through the contact material, the sample and again the contact material.

It is also possible that the equilibrium state already takes place in step f2) while supplying a current. In this likewise preferred embodiment, step f2) therefore comprises supplying a DC current across the blocks by means of a first pair of wires so that the current flows from one block to the other block, the current flowing through the contact material, the sample and again the contact material, and Maintaining the flow of current until an equilibrium state is reached. Upon receipt of the equilibrium state, the voltage V _{1 is} measured across the other second pair of lines (voltage probes).

The change in the current intensity in this case can mean the complete switching off of the current at the time to, which is likewise preferred according to the invention. However, it is also possible that the current is increased or decreased at the time to, that is, a stronger or weaker current is applied.

The change in the current takes place abruptly, that is, the change is not continuous, but there is an immediate change to the new value or the immediate shutdown of the current.

Step f4) of the method according to the invention relates to the time-resolved measurement of the voltage over a short period immediately after the current intensity has changed. This is characterized by an initial faster drop in voltage compared to the following phase of uniform decay according to a simple exponential law analogous to the capacitor discharge. The voltage measurement should preferably be made over the duration of at least one half-life of the rapid decay and at such a high sampling rate that at least 10 to 100 voltage values can be recorded within the first half-life. This half-life is dependent on the size of the contact resistance and is typically less than 1 s.

The determination is also possible by means of other metodics which allow to deduce the initial value of the relaxing measurement voltage immediately after the current has been switched off and to specify an earliest possible time at which the initial rapid relaxation has subsided almost completely. This is typically the case after about 5 half-lives of this relaxation.

Step f5) of the method according to the invention relates to the time-resolved measurement of the voltage following step f4). This takes place over a period of time in which the system has approximately reached the new equilibrium state and with a data rate of at least 5 to 20 values per half-time of slow decay in the case of temperature equalization or re-entry into the new equilibrium state). This half-life is dependent on the size and properties of the sample and the blocks and is typically about 5 to 60 seconds in laboratory geometry with mm-sized blocks and samples, much more in large systems; in miniature versions or microsystems, it is smaller.

Alternatively, the measurement may be made by another methodology which allows one to deduce the initial value of the second, slower relaxation with respect to the time of change of the current magnitude t ₀ and its final value after its complete decay.

In step f6) of the method according to the invention, the numerical adaptation of the experimental data takes place. It depends on the data rate. At a sufficiently high data rate and signal quality, starting from the decay of the initial faster decay over a period of less than about 0.2, half-lives of the slower decay are adjusted to a linear function. A more precise fit, starting after the initial faster decay has subsided over a time of about 3 to 5 half-lives of slow decay, is given to a functional set V ₁ (t) = A ₁ * exp (-t / τ ₁ ). This is followed by a temporal extrapolation of this function of the measured voltage drop to obtain a baseline up to the time t ₀ .

The final value referred to in step f6) is zero when the current is turned off as a possibility of change in step f3), with a slight residual offset after. after switching on or switching over, the end value is the respective set value.

This method thus enables a correct determination of thermal conductivity k and thermoelectric quality ZT, which system-inherent errors of the Harman method are eliminated. The thermoelectric quality ZT depends on the thermal conductivity of the material on which the measurement is taken (sample ( 1 )), its measured value is determined by the thermal resistance of non-ideal contacts ( 3a . 3b ) falsified.

To the thermoelectric sample ( 1 ), this is between two blocks ( 2a . 2 B ) and connected via a contact material with these blocks. The sample has two opposing contact surfaces, these are in each case via contact materials ( 3a . 3b ) with the blocks ( 2a . 2 B ) connected. The blocks ( 2a . 2 B ) heating elements ( 4a . 4b ), but this is not necessary. Furthermore, the blocks ( 2a . 2 B ) each have at least two electrical connection lines (probes). If two probes are present, one probe can be used to inject the current and the other to measure the voltage in the Harman measurement. When measuring the Seebeck coefficient, both connection lines can be used as voltage probes. This measurement is currentless. Preferably, however, at least three electrical connection lines are present on each block: a power supply and two voltage probes, which consist of different materials. If a block contains a heater, an additional heating current supply is required. At least one of the blocks is adiabatically shielded from the environment. However, both blocks ( 2a . 2 B ) be adiabatically shielded.

Particularly good results are obtained when the heat capacity of the blocks ( 2a , 2b) is greater than the heat capacity of the sample ( 1 ). In particular, the ratio of the heat capacity of the blocks to the heat capacity of the sample ( 1 ) in the range from 2: 1 to 50: 1, in particular from 15: 1 to 10: 1. Usually, a sample to be examined has a spatial extent of 5 mm to 6 mm, possibly of 2 to 10 mm. Accordingly, a block in a size of about 15 mm or larger is necessary.

By way of example, an embodiment in which the change in the current intensity is achieved by switching off and the system is in an equilibrium state after switching on a direct current in step f2) will now be described below.

To determine the thermoelectric quality, a stream is passed from one block to the other block. The current thus flows from a block ( 2a ) by the contact material ( 3a ) into the sample ( 1 ), from this through the second contact material ( 3b ) in the other block ( 2 B ). The sample ( 1 ) comprises a thermoelectric material, so that sets with the flow of current due to the associated Peltier effect, a temperature difference in the interior of the sample. After usually a few minutes, an equilibrium state is reached. In this there is a drop in temperature through the sample. The principle of Harman measurement requires that at least one of the blocks is adiabatically separated from the environment, that is, its heat content can only change by the heat flow through the sample. Within the contact material as well as within the blocks ( 2a . 2 B ) there is a constant temperature. The temperature difference inside the sample ( 1 ) is caused by the Peltier effect.

If a sudden change in the current flow is made by switching off the current at time t = t ₀ , the system is in a non-equilibrium state. In this case, the voltage initially falls rapidly, then in a slower exponential course. In the usual conventional evaluation, this slower exponential curve was extrapolated to a time t = t ₀ .

In 2 ) shows the time profile of the total voltage across a thermoelectric sample when switching on and then switching off a direct current through the sample. The current flow is accompanied by an ohmic voltage drop, which is connected to the electrical resistance of the sample. The Peltier heat transported by the stream introduces a temperature difference between the blocks which stabilizes after some time when the back-flowing Fourier heat compensates for the Peltier heat flow. This is described as equilibrium state according to the principle of Harman measurement. After switching off the current, the temperature compensation can be tracked by the drop in the thermoelectric voltage.

It has been found, however, that within the first moments immediately after the abrupt change and thus in the present exemplary embodiment, after the current flow has been switched off, there is a state in which there is a temperature change within the contact material which is influenced by the incipient drop in temperature across the sample while the temperature of the blocks does not change significantly yet. However, this is not taken into account by the usual extrapolation, so that the falsifying influence of the thermal resistance of the contact material on the measurement was not taken into account by methods of the prior art.

In 3 is a temperature profile over the blocks, contacts and the sample ( 1 ). The solid line shows the equilibrium state after switching on a current, as described in step f2) of the method according to the invention. The broken line shows the non-equilibrium state after turning off the current (step f3) after a short time, neglecting the slower change of the block temperature.

It has now been found that within a first short phase after the shutdown of the current, the drop in the thermal voltage does not follow the later simple exponential law. This deviation is measurable, so that according to the invention measurements with a high measuring rate are carried out over a period which is short against the half-life of the decay of the temperature difference between the blocks. This means that as soon as possible after switching off the current, a first measurement is made and then after a short period of time (much shorter than the half-life of the first rapid voltage drop) a further measurement is made and so on. Usually, the measurement is already started before switching off; For Harman measurement, it is absolutely necessary to know the measuring voltage before switching off or switching the current. After a short time, which is usually small compared to the half-life of the subsequent decay, the proportion of the thermoelectric voltage attributable to the contact material is no longer measurable, and the measurement of the voltage in the usual way over a period of about 3 to 5 half-lives the slow decay continued (step f5) of the method according to the invention). Here the exponential decay of the voltage over time is then displayed.

From the extrapolation of the measured voltage drop towards time to, an amplitude A ₁ can be determined. According to the temperature control of the sample environment and other peculiarities of the design of the measuring arrangement, it may be necessary to superimpose a weakly inclined linear function on this basically exponential drop in order to improve the accuracy of the method, namely if the entire sample holder is exposed to a slight drift in the middle temperature ; This is also known in the conventional method and not relevant to the invention. From the measurement immediately after the current has been switched off in step f4) of the method according to the invention, an amplitude A ₂ can be determined as a difference to the value A ₁ , wherein if appropriate step f 5) also comprises the extrapolation of the measured values towards the time t o. A corresponding measurement course is in 4 shown.

The ratio A ₂ / A ₁ corresponds to the relative systematic error of the thermal resistance measurement on the sample, so that now the thermal conductivity and the thermoelectric quality can be determined taking into account this error.

The determination in the case of the shutdown of the current as a sudden change, after an equilibrium state was reached under current flow, takes place according to the following formula: $$ZT=\frac{U_{S.True}}{U_{OHm}}=\frac{\left(A_1+A_2\right)}{U_{OHm}}$$

$$ZT=Z{T}_{old}\frac{A_1+A_2}{A_1}\ast \frac{U_{ohm}+A_2}{U_{ohm}}$$

The incorrect ZT _old determined by the standard method can be corrected as follows $$Z{T}_{Old}=\frac{A_1}{U_{OHm}+A_2}$$

$$ZT=Z{T}_{O\mathrm{<figure-callout\; id="1"\; label="l\; d\; A"\; filenames="DE102016223548B4\_0008.png,DE102016223548B4\_0010.png"\; state="\{\{state\}\}">l\; </figure-callout>}d}\frac{A_{}}{}\frac{{}_1+A_2}{{\mathrm{<figure-callout\; id="1"\; label="A"\; filenames="DE102016223548B4\_0008.png,DE102016223548B4\_0010.png"\; state="\{\{state\}\}">A</figure-callout>}}_1}*\frac{U_{OHm}+A_2}{U_{OHm}}$$

The thermal conductivity usually calculated from the time constant τ ₁ is determined to be too small by the thermal contact resistances (series connection of the resistors from sample and contact). The corrected value κ results from the measured κ _meas by: $$\kappa ={\kappa }_{meas}*\frac{A_1+A_2}{{\mathrm{<figure-callout\; id="1"\; label="A"\; filenames="DE102016223548B4\_0008.png,DE102016223548B4\_0010.png"\; state="\{\{state\}\}">A</figure-callout>}}_1}$$

5 shows the transient thermal voltage between the blocks immediately after switching off the DC current through the sample. An enlarged view of the signal at the beginning is shown in the inset. The thermoelectric voltage represents the temperature difference across the sample (not the one between the blocks, because the metal contact area does not contribute to the thermoelectric voltage). The temperature compensation follows a simple exponential decay analogous to the capacitor discharge when the heat capacity of the blocks is large versus that of the sample. At the beginning of relaxation, a small but significant deviation from the simple exponential progression is observed. The thermoelectric voltage initially drops faster. This is due to the rapid formation of a temperature drop over the contact area after shutdown. In the previous equilibrium state no temperature difference drops over the contact, since the Peltier heat is not released in the block but directly on the thermoelectric material. If a temperature difference between the blocks is generated by the integrated heaters, one does not observe this initial, faster drop. The faster compensation at the beginning describes the beginning of temperature compensation over the sample in synchronism with the formation of the temperature difference across the contacts, while the block temperatures initially do not change. The back extrapolation of the later, slower exponential decay to the turn-off time describes a state where the equilibrium temperature difference of the blocks corresponds to the thermal potentiometer ratio divides between sample and contacts. The amplitude A _{2 of} the rapidly decaying signal component compared to the voltage A _{1 of} the slow relaxation thus represents the ratio of the thermal resistances of contacts and sample, provided that one of the blocks is adiabatically separated from the environment. Ideal for performing the measurement is an adiabatic isolation of both blocks; if it is only one-sided and if the other block exchanges a significant heat flow (this should be small and temporally constant) with the environment, the exponential fit of the slow temperature compensation must additionally contain a constant or weakly linear term. This improves the accuracy of determining the amplitude A ₁ , but is not critical to the gist of the invention. Another variant of the method is based on the adiabatic completion of one block while the other is regulated to constant temperature.

• g1) Zuführen eines Gleichstroms über die Blöcke, so dass der Strom I von einem Block zu dem anderen Block fließt, wobei der Strom durch das Kontaktmaterial (3a), das Substrat (1) und erneut das Kontaktmaterial (3b) fließt,
• g2) Bestimmung von A₂
• g3) Bestimmung des Temperaturunterschieds ΔT = A₂/S , wobei S die Differenz der Seebeck-Koeffizienten des thermoelektrischen Materials und des Fügepartners bezeichnet, und
• g4) Berechnung des Kontakt-Wärmewiderstandes R_C ^th nach ${\text{R}}_{\text{C}}{}^{\text{th}}=\frac{\mathrm{\Delta }\tau }{p}\left(\text{P}=\text{Peltierwärmestrom}\right)\text{.}$
  

In step g) of the process according to the invention, preference is first given to the determination of A ₂ as described in f3), f4), f5) and f6). Subsequently, the determination of ΔT. Thus, step g) of the method according to the invention preferably comprises the following steps:

• g1) supplying a DC current across the blocks so that the current I flows from one block to the other block, the current through the contact material ( 3a ), the substrate ( 1 ) and again the contact material ( 3b ) flows,
• g2) determination of A ₂
• g3) determining the temperature difference ΔT = A ₂ / S, where S denotes the difference of the Seebeck coefficients of the thermoelectric material and the joining partner, and
• g4) calculation of the contact thermal resistance R _C ^th after ${\text{R}}_{\text{C}}{}^{\text{th}}=\frac{\mathrm{\Delta }\tau }{p}\left(\text{P}=\text{Peltier heat flow}\right)\text{,}$
  

1. i) Zuführen eines Gleichstroms über die Blöcke, so dass ein elektrischer Strom I₀ von einem Block zu dem anderen Block fließt, wobei der Strom durch das Kontaktmaterial, das Substrat (1) und erneut das Kontaktmaterial fließt,
2. ii) Abschalten oder Umschalten des Stromes auf einen anderen konstanten Wert zum Zeitpunkt to, wobei die Änderung des Stromes mit I bezeichnet wird, und
3. iii) zeitlich aufgelöste Messung der Spannung über einen Zeitraum unmittelbar nach Abschalten des Stromes, wobei der Zeitraum wenigstens während einer Halbwertszeit des raschen Abklingens und mit einer Abtastrate erfolgt, die es ermöglicht, innerhalb der ersten Halbwertszeit wenigstens 10 bis 100 Messwerte (Spannungswerte) aufzuzeichnen, oder mittels einer anderen Methodik, die es gestattet, auf den Anfangswert der relaxierenden Messspannung unmittelbar nach Abschalten des Stromes v₁(t₀) zurück zu schließen,
4. iv) anschließende zeitaufgelöste Messung der Spannung über einen Zeitraum, beginnend nachdem die erste Relaxation weitgehend abgeklungen ist, in welchem sich das System im Wesentlichen auf den neuen Gleichgewichtszustand einstellt, mit einer Datenrate von mindestens 5 bis 20 Werten pro Halbwertszeit des langsamen Abklingens beim Einlaufen in den neuen Gleichgewichtszustand, oder mittels einer anderen Methodik, die es gestattet, auf den Anfangswert V_1,fit(t_o) der zweiten, langsameren Relaxation bezogen auf den Abschalt-/Umschaltzeitpunkt des Stromes to zurückzuschließen,
5. v) lineare oder exponentielle Extrapolation des gemessenen Spannungsabfalls in Schritt iii) zum Erhalt eines Anfangswertes V₁(t₀) zum Zeitpunkt to und lineare oder exponentielle Extrapolation des gemessenen Spannungsabfalls in Schritt iv) zum Erhalt einer Basislinie v_1,fit(t) bis zum Zeitpunkt to,
6. vi) Bestimmung des Temperaturunterschieds ΔT = (v₁(t₀)-v_1,fit(t₀))/S zwischen der in Schritt v) erhaltenen Basislinie zum Zeitpunkt to sowie des extrapolierten Messwertes aus Schritt c) zum Zeitpunkt to, wobei S die Differenz der Seebeck-Koeffizienten des thermoelektrischen Materials (Substrats (1)) und des Fügepartners (Kontaktmaterials) bezeichnet, und
7. vii) Berechnung des Kontakt-Wärmewiderstandes R_C ^th nach ${\text{Rc}}^{\text{th}}=\frac{\mathrm{\Delta }T}{p}$
   
   (P = Peltierwärmestrom).

The method may thus comprise the following steps:

1. i) supplying a DC current across the blocks so that an electric current I _{0 flows} from one block to the other block, the current through the contact material, the substrate ( 1 ) and again the contact material flows,
2. ii) switching off or switching the current to another constant value at time to, wherein the change of the current is denoted by I, and
3. iii) time-resolved measurement of the voltage over a period of time immediately after the current has been switched off, the time period being at least during a fast decay half-life and at a sampling rate allowing at least 10 to 100 measured values (voltage values) to be recorded within the first half-life; or by another method, which allows to the initial value of the relaxing measuring voltage immediately after switching off the current v ₁ (t ₀₎ to close back,
4. iv) subsequent time-resolved measurement of the voltage over a period of time after the first relaxation has largely subsided, in which the system substantially conforms to the new equilibrium state, with a data rate of at least 5 to 20 values per half-time of slow decay on entry into the new equilibrium state, or by means of another methodology that allows _{one to deduce} the initial value V _{1, fit} (t _o ) of the second, slower relaxation with respect to the switch-off / switchover instant of the current to,
5. v) linear or exponential extrapolation of the measured voltage drop in step iii) to obtain an initial value V ₁ (t ₀ ) at time to and linear or exponential extrapolation of the measured voltage drop in step iv) to obtain a baseline v _{1, fit} (t) to at the moment to,
6. vi) Determining the temperature difference ΔT = (v ₁ (t ₀ ) -v _{1, fit} (t ₀ )) / S between the baseline obtained in step v) at the time to and the extrapolated measured value from step c) at the time to, wherein S is the difference of the Seebeck coefficients of the thermoelectric material (substrate ( 1 )) and the joining partner (contact material), and
7. vii) calculation of the contact thermal resistance R _C ^th after ${\text{rc}}^{\text{th}}=\frac{\mathrm{\Delta }T}{p}$
   
   (P = peltier heat flux).

In principle, this measurement is not bound to room temperature as well as all other partial measurements described with the device and the method according to the invention. In particular, it is not affected by the effects of heat radiation. It can therefore be used in principle for the temperature-dependent determination of the thermal contact resistance. The method is tailored to the relevant case for the thermoelectric module development case of the surface connection of a thermoelectric material with a metallic contact bridge. The Seebeck coefficient of the material of the sample ( 1 ) is assumed to be known or can be determined with the described arrangement in the course of the measurement. Furthermore, it is assumed that the sample is provided with identical contacts either symmetrically on both sides is - this is particularly appropriate if the properties of the sample are not known - or is provided on one side with a contact of known properties.

If the measurement serves the purpose of investigating a contact between a given thermoelectric material (substrate) and a given metallic contact partner, the thermoelectric substrate (the sample ( 1 )), which typically consists of the technological material for which a contact is to be found, between two blocks ( 2a . 2 B ), which are made of the metal joining partner to be examined or solid and good thermal conductivity (for example, a brazed joint or a suitably shaped clamping connection) connected to the metallic joining partner, in turn, on the contact material to be examined ( 3a . 3b ) is connected to the substrate, and each with a temperature measuring point, optionally in addition with a thermocouple line (if not contained in the temperature sensor) and a power supply line are electrically connected.

With regard to the action to be taken in step g2), reference is made to the statements relating to steps f3), f4 (, f5) and f6), which apply here accordingly.

The current I now sets at the material transition from the block ( 2a ) to the contact material ( 3a ) and from this to the thermoelectric substrate ( 1 ) a Peltier heat output of size I · S · T free (S - Seebeck coefficient of the thermoelectric substrate, if the contact is metallic with very low Seebeck coefficient, or the difference between the Seebeck coefficients of the substrate and the contact, if not, T - absolute temperature; I - current change when switching on, off or switching). If the Seebeck coefficient of the contact material is unknown and not small, the thermal contact resistance in the method according to the invention is determined quantitatively inaccurate. However, contacting attempts with the same contact material can be quantitatively compared among each other because the adulteration is a constant factor specific to the contact material. In the relevant case, the thermal resistance of the contact (contact material ( 3a . 3b )) small compared to the thermal resistance of the substrate ( 1 ), so that it can be assumed that in a state of equilibrium, which occurs after a very short time (about 1 s), the Peltier heat flows completely through the contact to the block. With high contact resistance, this approach has an inaccuracy that can be overcome by considering the parallel thermal connection between substrate and contact for Peltier heat dissipation. The outflow of the heat output P via the thermal contact resistance R _C ^th leads to the formation of a temperature difference ΔT = P · R _C ^th over the contact region, which leads to the formation of a thermal voltage U = S · ΔT across the substrate. However, this voltage can not be measured directly, as long as the current is flowing, because it is superimposed by the ohmic voltage drop.

When changing the current level (switching on, off or switching the current), the ohmic voltage drop changes almost instantly, while the temperature difference at the contact decays more slowly. Typically, this decay is traced over a few tenths of a second. With a sufficiently fast measuring electronics, it is possible to extrapolate back to the initial value immediately after changing the current.

The back extrapolation of the baseline at the switch-off time makes it possible, in the case of switching off the current, to separate the small voltage contribution from interfering offsets or to eliminate the ohmic voltage drop corresponding to the new current value when switching the current.

The thermal contact resistance is obtained from the ratio of the change in the temperature difference ΔT to the associated change in the heat flow: R _C ^th is equal to ΔT / P. P is the change in the Peltier heat flow, which corresponds to I · S · T (with the current change I). The Peltier heat flux is due to the knowledge of the measurement current with knowledge of the Seebeck coefficient of the substrate ( 1 ) known. If the Seebeck coefficient S is not known, this can be measured in the method according to the invention.

The arrangement has substantial similarities to the structure of the loffe method. A key difference, however, is that adiabatic separation or precise temperature control on the sample holder is not required for this partial measurement, which makes its design correspondingly simple if thermal and harman measurements are omitted.

For very small contact resistance, the associated thermal voltage is very small. Depending on the quality of the measuring electronics, the signal is correspondingly noise superimposed. By periodically repeating the switching on and off of the direct current, the relaxation process can be repeated and the measurement signal accumulated and thus filtered against stochastic noise.

The construction according to the invention makes it possible to carry out a Seebeck measurement (required for determining the contact thermal resistance according to g), if the Seebeck coefficient of the sample ( 1 ) is not known), if at least one of the blocks ( 2a . 2 B ) with a heater ( 4a ) is equipped or can be externally tempered. A thermal separation of the holder from the environment or a control of the heat exchange with the environment is not required for this partial measurement, for thermal and Harman measurements but indispensable.

The block temperatures are measured by temperature sensors, for example thermocouples. Thus, all measurements with the device are also temperature-dependent feasible. By utilizing the Peltier effect, a temperature difference across the contact areas is established by passing a direct current I through the sample. When using metallic contacts having a vanishing Seebeck coefficient, the thermoelectric voltage is V ₁ (or V ₂ ) between the blocks ( 2a . 2 B ), a measure of the temperature difference across the thermoelectric sample ( 1 ).

Due to the relatively low thermal resistance of the contacts involved (compared to the sample ( 1 )) and the low heat capacity involved (only a relatively narrow zone of the sample ( 1 ) near the contact has to be heated by the Peltier heat) the temperature difference in the contact area develops in a short time (typically within 1s). The temperature of the blocks ( 2a . 2 B ) remains largely unaffected by the Peltier heat in this short time. Do the blocks ( 2a . 2 B ) a small temperature difference (a few K), a measurement of the Seebeck coefficient S of the substrate can be carried out in a conventional manner, which in principle on a simultaneous determination of the temperature difference and the thermal voltage between the blocks ( 2a . 2 B ). Thus, the Peltier heat flux flowing over the contacts is known: I · S · T.

5 FIG. 12 shows the progression of the total voltage across the thermoelectric substrate after a short circuit (about 1 second) and a direct current cut-off through the sample, as already explained above.

• h1) Beheizen eines Blockes oder ungleichmäßiges Beheizen beider Blöcke (2a, 2b), so dass sich im Inneren der Probe (1) eine zeitlich konstante oder zeitveränderliche Temperaturdifferenz zwischen den beiden Kontaktflächen ausbildet, sowie Bestimmung der Temperaturdifferenz ΔT zwischen den Blöcken, und vorzugsweise gleichzeitige Bestimmung der Thermospannung V₁ zwischen den Blöcken (2a, 2b), welche sich durch die Temperaturdifferenz über der Probe (1) ausbildet, daraus Bestimmung des Seebeck-Koeffizienten in herkömmlicher Weise S_old = V₁/ ΔT;
• h2) Bestimmung der relativen Abweichung r von der nach h3) erhaltenen Temperaturdifferenz zu der wahren Temperaturdifferenz über der Probe, und
• h3) vorzugsweise Bestimmung des Seebeck-Koeffizienten, korrigiert durch die in Schritt h4) erhaltene Abweichung S_corr = (1+r) S_old.

Preferably, step h) of the method according to the invention comprises the following steps:

• h1) heating a block or non-uniform heating of both blocks ( 2a . 2 B ), so that inside the sample ( 1 ) forms a time-constant or time-variable temperature difference between the two contact surfaces, and determination of the temperature difference .DELTA.T between the blocks, and preferably simultaneous determination of the thermal voltage V ₁ between the blocks ( 2a . 2 B ), which is due to the temperature difference across the sample ( 1 ), from which determination of the Seebeck coefficient in a conventional manner S _old = V ₁ / ΔT;
• h2) determining the relative deviation r from the temperature difference obtained according to h3) to the true temperature difference across the sample, and
• h3) preferably determining the Seebeck coefficient, corrected by the deviation S _corr = (1 + r) S _old obtained in step h4).

Temperature sensors are used to determine the temperature difference. The temperature sensors can be represented, for example, as thermocouples, formed from two thin metallic lines with mutually different Seebeck coefficients.

According to step h1), a temperature difference in the interior of the sample and temperature differences over the contact areas are formed. It therefore prevails in the first block ( 2a ) a temperature T _1B and at the interface between the first contact material ( 3a ) and the sample a temperature T ₁ , in the second block ( 2 B ) a temperature T _2B and at the interface between the second contact material ( 3b ) and the sample a temperature T ₂ . It applies, depending on the direction of the temperature gradient, T _1B > T ₁ > T ₂ > T _2B or T _1B <T ₁ <T ₂ <T _2B . There is thus, caused by the Fourier heat flow from one block to another, a temperature profile in the interior of the sample and the contact areas in the direction of the block distance, ie from one contact surface to the other contact surface.

The conventionally calculated Seebeck coefficient S _old = V ₁ / ΔT is too small by the ratio of the temperature difference between the blocks to the temperature difference across the sample, so that a correction is necessary, which enable the method according to the invention and the device according to the invention. Step h2) is equivalent to the determination of the relative thermal contact resistance (based on the thermal resistance of the sample) r, for example by means of additional temperature sensors applied to the sample or thermographically or by a transient method in which a direct current is passed through the sample ,

According to the invention, the sample is now ( 1 ) over contact materials ( 3a . 3b ) with blocks ( 2a . 2 B ) in contact. The blocks ( 2a . 2 B ) contain the probes with which a measurement is then carried out. In contrast, in the prior art, the probes are directly on the sample ( 1 ) applied or this is tried. However, direct application is not possible such that the application of the sensors to the sample in the prior art falsifies the measurement because some of the heat from the sample flows into the contact point between the probe and the sample. This contact point has a spatial extent, so that it can not be considered ideal as a point, so that there is a parasitic heat flow, which creates an error in the determination of the Seebeck coefficient.

According to the invention, the probes are now incorporated in the block and the design of the arrangement ensures that the same heat flow flows through the contacts and the sample. Thus, the parasitic temperature difference can be determined and the Seebeck measurement correctable.

The first block ( 2a ) and / or the second block ( 2 B ) are tempered. This can preferably take place in that inside the block at least one heating element ( 4a ) is arranged. It is also possible that external heating devices are present, by means of which the first and / or the second block can be tempered.

By the application of a direct current, which of a block ( 2a ) through the sample to the other block ( 2 B ) flows, arises in the sample ( 1 ) due to the thermoelectric properties of the sample ( 1 ) a temperature difference. This arises along the current flow, so that the temperature difference between the two contact surfaces of the sample ( 1 ) forms with the block through the contact material. The thermoelectric voltage V ₁ can then be determined by the probe, which is a measure of the temperature difference across the sample ( 1 ). Furthermore, by the temperature sensors on the blocks, the temperature difference between the first block ( 2a ) and the second block ( 2 B ). From these two values, the relative deviation r of the thermal voltage can then be achieved. If one now determines the Seebeck coefficient, this is to be corrected by the relative deviation r, which was determined in step h4) of the method according to the invention.

In order to obtain particularly accurate measurements, the line cross section of the at least one probe, which is used as a power supply, in particular so large that result from the Joule heat of the stream no significant heat of reaction.

The core of the invention lies in the approach, the contact points for measuring the thermal voltage on the sample ( 1 ) so that the parasitic thermal resistance and the heat flow between the temperature sensor and the contact point, ie between the locations of the temperature and the thermoelectric voltage measurement, or, equivalently, the parasitic temperature drop between the two locations, are known or can be determined. From the relative temperature rejection results directly the systematic relative error of the Seebeck measurement, so that a computational correction can be made.

It has been described here that the parasitic thermal resistance of contacts in thermoelectric arrangements can be determined transiently under certain thermo-constructive boundary conditions. In particular, the following boundary conditions must be observed: In the sample ( 1 ) must be fed via the heat flow path between the temperature sensor and the location of the thermoelectric voltage measurement, an electrical DC current that is strong enough to set via the Peltier heat a measurable temperature difference across the sample. All wire cross-sections should be large enough that the joule heat of the stream does not produce any significant heat effects. The leads of conventional wire thermocouples are far too thin for this purpose.

If only the Seebeck coefficient is to be measured with the arrangement, at least one of the blocks ( 2a . 2 B ) be thermally decoupled from the environment so far that a sufficient temperature difference can be established over the sample; few Kelvin (K) are sufficient. When combined with thermal measurements, such as thermal conductivity or harman measurement, such conditions must be established that at least one of the heat reservoirs is thermally well insulated against the environment, so that upon relaxation of a temperature difference across the sample ( 1 ) only a negligible proportion of the heat stored in the reservoir flows into the environment.

In order to take into account the influence of the thermal contact resistance on the thermoelectric measurements, the sample must be adiabatically terminated on its lateral surface (ie on all surfaces which are not connected to the blocks). The selected measurement geometry then sets due to the serial connection between contact and sample ( 1 ), without the possibility of lateral heat leakage, ensure that the ratio r in analogy to the voltage division on the potentiometer expresses the ratio of the contact thermal resistance to the thermal resistance of the sample. The measurement of the Seebeck coefficient S of the sample is done in a conventional manner by between the blocks a small drifting temperature difference of a few K is generated and the thermoelectric voltages V ₁ and V _{2 are} recorded, which in principle a simultaneous determination of the temperature difference of the blocks and Thermal voltage across the sample corresponds. The Seebeck coefficient determined in this conventional manner is to be corrected for the thermal resistance ratio; the larger the thermal contact resistance, the smaller the falsified, conventionally measured Seebeck coefficient.

The basic structure of a sample holder according to the invention for the simultaneous measurement of thermoelectric properties is shown in FIG 1 shown. In addition, additional thin probe leads are attached to the sample for the electrical contact resistance measurement ( 1 ) appropriate. If these are used in addition to similar probes on the blocks for the measurement of thermoelectric voltages in the currentless sample, the effective geometric distance of the sample probes is obtained from the comparison of both measurements ( 6 . 7 ), which must be known for the correct separation of the electrical contact resistance. This makes it possible to dispense with a complex and uncertain visual determination of the probe distance.

Analogous to the separation of the electrical contact resistance can be determined by determining the thermoelectric voltages inside and outside the sample ( 1 ) the thermal resistance between material and blocks ( 2a . 2 B ), if not probes (ie single leads), but thermocouple probes (lead pairs) in addition to the thermocouples present in the blocks also on the sample ( 1 ). This is known as a "thermal potentiometer", technically difficult to produce and fragile and requires a hardly realizable dynamic tracking of the probe temperatures during the measurement. This complex and uncertain procedure is circumvented with the structure according to the invention.

1 Thus, the schematic structure of an apparatus for simultaneous determination of all thermoelectric transport variables according to the invention (Seebeck coefficient, electrical conductivity, thermal conductivity, thermoelectric efficiency according to the Harman method), supplemented by stationary thermalizable probes for undisturbed thermoelectric measurement of the probe position. The heatable metal blocks ( 2a . 2 B ), which on both sides with the sample by metallic contacts ( 3a . 3b ) are firmly adiabatically insulated against the environment according to the measurement methodology according to Stecker / Teubner. Due to the availability of additional contacts on the sample ( 1 ) (at a distance of I _p ), which can be applied by spot welding or, more conveniently for laboratory operation, can be pressed by spring force, the electrical contact resistance can be determined and eliminated. In contrast to conventional prior art measurements, no dynamic tracking of the probe temperature is required.

The construction of the arrangement allows the redundant determination of the thermal contact resistance with two independent methods - accompanying the Harman measurement and with respect to a known Peltier heat flow, each from the dynamically separable parasitic temperature difference across the surface sample contacts - including the measurement arrangement with a sufficiently fast transient Measured value acquisition and a temporally extrapolating evaluation is provided. The redundant determination of thermal contact resistance reduces the uncertainty budget and helps detect systematic errors in thermal environment compliance.

The thus separated influence of the electrical and thermal contact resistance is used to correct the thermoelectric properties: From the conventionally measured electrical resistance of the sample is corrected subtracting the electrical contact resistance; from the conventionally measured thermal resistance of the sample, the thermal contact resistance is corrected in an analogous manner; the conventionally measured Seebeck coefficient is corrected by increasing the percentage of the relative thermal contact resistance, and the Harman measurement is determined with improved accuracy from a correct extrapolation of the relaxation of the thermovoltages, taking into account the rapid initial signal drop.

In addition to their role as correction quantities, the separated thermal and electrical contact resistance in the course of targeted development work on contacting development are also significant as independent parameters for contact characterization. The measuring system thus acquires an additional and previously unavailable utility value.

Claims (11)

Apparatus for the simultaneous measurement of thermoelectric properties of thermoelectric materials and contact resistors comprising: a first block (2a) and a second block (2b) different therefrom, arranged such that between them the sample (1) to be analyzed is introduced and via a first Contact material (3a) with the first block (2a) and a second contact material (3b) with the second block (2b) can be brought into surface contact, - at least one first probe (4) formed as a thermocouple of two thin lines of thermocouple Alloys, respectively on the first block (2a) and the second block (2b), via which a temperature difference □ T and an electrical Voltage V, which occurs between the blocks (2a, 2b), can be determined, - a measuring system for determining the values of the electrical voltage V between the blocks (2a, 2b) measured by a line pair of the at least one first probe (4) high temporal resolution, - at least a second probe respectively on the first block (2a) and the second block (2b), formed as an electrical supply line or as a pair of leads, via which a direct current through the sample or a heating current for heating one of the blocks in one or both or one or both blocks (2a, 2b) or one or both blocks, - at least two thermalizable potential probes (6, 7), which are located at a spatial distance I _P , measured along the current flow direction through the sample, during the measurement can be brought into contact with the sample, - a thermal anchor (8) comprising a first temperature reservoir (8a) and a second temperature reservoir ir (8b), which are thermally connected to one another via a thermally conductive mechanical guide (8c), wherein the potential probes (6, 7) are in thermal contact with the thermally conductive mechanical guide (8c). Device after Claim 1 , characterized in that the first block (2a) and the second block (2b) are adiabatically insulated from the environment. Device after Claim 1 or 2 , characterized in that the first block (2a) and / or the second block (2b) have a heating element (9) inside. Device according to one of Claims 1 to 3 , further comprising at least one heating device, with which the temperature of the first and / or the second block (2a, 2b) and / or the first and / or the second temperature reservoir (8a, 8b) can be adjusted. Device according to one of Claims 1 to 4 , further comprising means for adjusting the distance of the potential probes (6, 7) from the sample (1), wherein during the measurement the potential probes (6, 7) can be connected to the sample and released again. Device after Claim 5 , characterized in that the device brings the potential probes (6, 7) by spring force with the sample (1) in contact. Method for the simultaneous determination of thermoelectric properties of a sample (1) with thermoelectric properties by means of a device according to one of the Claims 1 to 6 in which the sample has two spatially separated contact surfaces and the distance between these contact surfaces corresponds to the length l, comprising: a) bringing a first contact surface of the sample (1) into contact with a first contact material (3a) with a first block (2a) and b) contacting a second contact surface of the sample (1) with a second block (2b) via a second contact material (3b), wherein the first block (2a) and the second block (2b) comprise a metal and / or a metal Coating, wherein metal and / or metallic coating have a low Seebeck coefficient, typically <5 μV / K, c) applying at least one probe (4) formed as a thermocouple of two thin lines of thermocouple alloys, respectively the first block (2a) and the second block (2b), such that they are electrically conductively connected to the respective contact material (3a, 3b) via the respective block (2a, 2b) nd, d) contacting potential probes (6, 7) with the sample (1), wherein the distance between the probes is I _P and I _P <I, the potential probes on the sample (1) and at least one each Conduction of the probes on the blocks (2a, 2b) consist essentially of the same material and the application is carried out such that during the subsequent measurement, no electric current and no heat to or from the sample via the probes or can flow, e) Determining the effective distance of the electric potential probes on the sample (1), f) Determining the thermal conductivity k and the thermoelectric quality ZT of the sample (1) g) Determining the contact thermal resistance R _C ^{th of} the contact material (3a, 3b) and h) Determination of electrical conductivity. Method according to Claim 7 characterized in that step e) comprises the steps of: e1) setting a temperature difference ΔT between the first block and the second block, e2) determining the thermal voltage V ₁ between the first block and the second block by means of the blocks respectively applied Probes, e3) subsequently or simultaneously with e2) and under the same temperature conditions as in step e2), determining a thermal voltage V _P by means of the electric potential probes applied to the sample, and e4) forming the ratio V _P / V ₁ for determining the effective probe position , Method according to Claim 7 or 8th , characterized in that step f) comprises the following steps: f1) optionally terminating the heat supply from step e1), f2) reaching a thermal equilibrium state, f3) changing the current at the time to and f4) time-resolved measurement of the voltage over a period immediately after switching off the current that makes it possible to the initial value f5) subsequent time-resolved measurement of the voltage over a period of time in which the system has substantially adjusted to the new state of equilibrium that it allows, on the voltage immediately after the current change and the approximate time to decay of the initial rapid voltage change; f6) numerical adaptation of the data from the initial value of this slower voltage change at the time of the current change as well as the time constant of the decay of this slower voltage change to the new equilibrium state and to the final value when equilibrium is reached Steps f4) and f5) and temporal extrapolation of the measured voltage drop from step f5) to obtain a base line up to the time t _0, f7) forming the difference between this baseline and experimental data from step f4), wherein the difference at t ₀ the amplitude A ₂ corresponds; Forming the difference A ₁ from this baseline and the final value in equilibrium from step f5) and forming the ratio A ₂ / A ₁ , which indicates the relative systematic error of a thermal resistance measurement on the sample caused by the thermal resistance of the contact, f8) determination of the thermal conductivity k according to the conventional dynamic absolute method (loffe method) or a comparable method and the thermoelectric quality ZT according to the Harman method, taking into account the relative systematic error from step f7). Method according to one of Claims 7 to 8th characterized in that step g) comprises the steps of: g1) supplying a DC current across the blocks so that the current I flows from one block to the other block, the current passing through the contact material, the substrate (1) and again the contact material flows, g2) determination of A ₂ , g3) determination of the temperature difference ΔT = A ₂ / S, where S denotes the difference of the Seebeck coefficients of the thermoelectric material and the joining partner, and g4) calculation of the contact thermal resistance R _C ^th to ${\text{R}}_{\text{C}}{}^{\text{th}}=\frac{\mathrm{\Delta }\tau }{p}\left(\text{P}=\text{Peltier heat flow}\right)\text{,}$

Method according to one of Claims 7 to 9 , characterized in that step h) comprises the following steps: h1) heating a block or non-uniform heating of both blocks (2a, 2b), so that inside the sample (1) forms a time-constant or time-variable temperature difference between the two contact surfaces and determining the temperature difference ΔT between the blocks, and preferably simultaneously determining the thermal voltage V ₁ between the blocks (2a, 2b), which is formed by the temperature difference across the sample (1); h2) determining the relative deviation from the temperature difference obtained according to h1) to the temperature difference across the sample, and h3) determining the Seebeck coefficient, corrected by the deviation S _corr = (1 + r) S _old obtained in step h2).

Images