Seebeck coefficient measurement device

Abstract

Device for measuring Seebeck coefficients, characterized in that the sample to be measured is fixed in the recesses of two metal blocks such that the remainder of the recess is largely filled by a movable block of the same metal and thickness, and the Seebeck coefficient is measured in the sample perpendicular to the metal block. Figure to be published with the abstract: Figure 1

Description

Title of the invention: Device for measuring Seebeck coefficients

[0001] Description.

[0002] The Seebeck coefficient (thermal electromotive force) is an intrinsic property of all electrically conductive materials. It depends, among other things, on the crystalline structure, impurities, defects, imposed mechanical stresses, magnetization and temperature of a material. Measuring the Seebeck coefficient allows for monitoring manufacturing and aging processes.

[0003] The electromotive force that occurs between two points of a material is an intensive measurement quantity, i.e. its value does not depend on the geometry of the sample to be measured, but only on the temperature difference between two points. The electromotive force is an integral on the path from a to [3 of the Seebeck coefficient of the material at the point considered, multiplied by the infinitesimal temperature element dT. Point a is at temperature T1, point [3 is at temperature T2

[0004] In prior art Seebeck coefficient measurements, one method involves using two thermocouples AB and A'B', the positive branch (A and A') and the negative branch (B and B') of the thermocouples being made of identical metals or alloys. The thermocouples are placed on the sample to be measured and a temperature difference is created in the sample by heating or cooling one contact point (or both: for example, by cooling one contact point and heating the other). With a voltmeter having as high an internal resistance as possible, three (or four for testing, depending on the method) electromotive forces are then measured between the wires A and A', A and B, A' and B' (and possibly B and B' for testing). The temperature of the junction between the thermocouple wires and the power supply wires of the measuring device (usually copper) is also determined.The Seebeck coefficient S is defined as the infinitesimal electrical potential difference that arises along a homogeneous conductor that simultaneously exhibits an infinitesimal temperature difference. In principle, the temperature difference that exists along the conductor can be determined by the electrical voltages measured on the respective thermocouples, taking into account the temperature of the contact point. The conversion is carried out, for standardized thermocouples, using corresponding tables or polynomials. The electrical potential is determined by measuring the electrical voltage on the measuring path A-sample-A' (or B-sample-B' for control).

[0005] This type of measurement is based on certain assumptions that are only very rarely respected, which impairs the accuracy of the measurement result. The connection between the thermocouple wires (A and B and A' and B') is made by welding, which creates a transition zone with a concentration gradient between the starting materials. Typically, a spherical body with a diameter of 0.3 to 1.5 mm is obtained. The exact composition of the contact point with the sample is unknown. In order for the potential measured by the measuring section A-sample-A' not to be influenced by the concentration gradient present in the welding point, the temperature difference in the welding point must be zero.

[0006] Common thermocouples such as copper-constantan (type T) or chromel-alumel (type K) generate thermoelectric voltages of about 40 pV per Kelvin of temperature difference. If a measuring accuracy of + / - 40 nV / K is aimed for determining the Seebeck coefficient, this assumes that the temperature inhomogeneity in the area of ​​the contact points is less than 1 / 1000 K. Such constant temperature distributions can only be achieved in very rare cases.

[0007] A second known method for determining the Seebeck coefficient consists of placing a heated copper electrode on the sample. The contact area is heated by this electrode, so that a temperature gradient is formed between this first electrode and a second electrode [5]. Usually, the major part of the sample is at a constant temperature and the temperature gradient is limited to a small area around the contact point of the first electrode. In the known configuration, the temperature of the first electrode is measured by forming a thermocouple with another metal (see fig. 3 of Coste Leborgne...), the welding point of which is located as close as possible to the contact between the electrode and the sample.Since in this case a constant heat flux must be generated from the electrode to the sample to maintain the temperature gradient, it is obvious that the two temperatures of the contact point and the welding point differ. To improve the accuracy of this method, numerically calculated corrections are therefore indicated [5], but they depend among other things on the thermal conductivity of the sample and its thickness (cf. Proceeding of Kléber workshop Lyon 2002).

[0008] A second known method for determining the Seebeck coefficient is to place a heated copper electrode on the sample. The contact area is heated by this electrode, so that a temperature gradient is formed between this first electrode and a second electrode. Usually, most of the sample is at a constant temperature and the temperature gradient is limited to a small area around the contact point of the first electrode. In the known configuration, the temperature of the first electrode is measured by forming a thermocouple with another metal, the welding point of which is located as close as possible to the contact between the electrode and the sample. Since in this case a constant heat flow must be generated from the electrode to the sample to maintain the temperature gradient, it is obvious that the two temperatures of the contact point and the welding point differ. To improve the accuracy of this method, numerically calculated corrections are therefore indicated, but they depend, among other things, on the thermal conductivity of the sample.

[0009]

[0010] A third known method for determining the absolute Seebeck coefficient uses, instead of thermocouples or moving electrodes, previously prepared sample holders, on which contacts and heating elements have been applied. This method is used in particular for measuring the Seebeck coefficient of thin films, the contacts, the heating elements and the thin film to be examined being generally created using sputtering techniques. A particular difficulty with this method lies in the fact that the contact materials in the vast majority of cases have thermoelectric properties different from those of the corresponding bulk materials, so that along the measuring paths for determining the temperature gradient and the electrical voltage, different Seebeck coefficients and temperature gradients come into play, which are not always known with the required accuracy.It is therefore not possible to relate the measurement to a material whose Seebeck coefficient is known. Even if the temperature measurements are not carried out with thermocouples, but with resistors, this source of error remains when measuring the electrical potential along the sample.

[0011] In order to exclude any source of error when measuring the Seebeck coefficient, two principles must be respected:

[0012] 1) In areas outside the measuring range, in which If temperature gradients occur, for example in power cables, the same material (usually copper or thermocouple alloys) must be used everywhere so that the potential differences that arise balance out.

[0013] 2) In areas where the same material cannot be used, for example in thermocouple welding points or contacts, there must be no temperature gradient.

[0014] The three prior art methods described above only imperfectly meet these requirements, each time for different reasons.

[0015] The three methods of the state of the art only imperfectly meet these requirements, for different reasons in each case.

[0016] [Fig.l] [Fig.l] is a schematic view of a measuring assembly.

[0017] The measuring setup illustrated in [Fig.l] below largely satisfies principles 1 and 2 and is particularly suitable for determining the Seebeck coefficient of thin layers. The sample to be measured is clamped against two metal blocks (IA and IB), using for clamping two double wedge elements (2A and 2B) whose wedge angles correspond exactly to the angles of the recess. The wedge-shaped elements are dimensioned so that the recess (3) is exactly filled by the sample and the wedge-shaped element.

[0018] In a preferred embodiment, the necessary temperature gradient is generated by creating a heat flow between the respective metal block (IA or IB) and a block (5A and 5B) through which water flows by means of a Peltier element (4A and 4B). It is convenient to bring the blocks (5A and 5B) to the same temperature with the same thermostat and to switch the electrical control of the Peltier elements so that one metal block is cooled and the other heated. Thus, the average temperature remains constant, the temperature gradient is symmetrical about the center of the sample and it is possible to measure without problems at room temperature. If the sign of the current direction through the Peltier elements changes, the temperature gradient along the sample changes accordingly.

[0019] In a preferred arrangement, copper is used for the metal blocks IA and IB and aluminum for the blocks 5A and 5B. The invention also relates to an arrangement in which other materials are used, for example those which are still resistant to higher temperatures or in corrosive atmospheres, as well as to an arrangement for higher temperatures in which the blocks 5A and 5B serving as a heat reservoir are heated and the heat released into the environment is used for regulation.

[0020] In a preferred variant of the invention, the temperature of the metal blocks IA and IB is measured using a thermocouple whose contact point is located as close as possible to the recess 3. In the preferred variant, the wires of a thermocouple T are clamped in the copper blocks by means of grub screws, so that the contact point for temperature measurement is located at the clamping point between the negative branch and the copper block. The measurement of the electrical potential along the sample is carried out by the positive (copper) branch of both thermocouples.

[0021] Since the Seebeck coefficient is temperature dependent and even changes sign depending on the temperature, measurements are typically carried out with temperature differences of 2 to 6 K between the metal blocks. By using suitable Peltier elements, however, the structure allows temperature differences of up to 80 K.

[0022] In a preferred variant, the thickness of the metal blocks IA and IB represents a quarter of the lateral dimensions.

Claims

Claims

1. Device for measuring Seebeck coefficients, characterized in that the sample to be measured is fixed in the recesses of two metal blocks in such a way that the remainder of the recess is largely filled by a movable block of the same metal and of the same thickness and that the Seebeck coefficient is measured in the sample perpendicular to the metal block.

2. Device according to claim 1, characterized in that the recess and the movable metal block are conical in two directions, which, in combination with a plate pressed against the metal block, allows the movable metal block to be clamped in the recess.

3. Device according to claims 1 and 2, characterized in that the plates pressed against the metal blocks are the outer faces of Peltier elements with which the metal blocks can be brought to a certain temperature.

4. Device according to claims 1 to 3, characterized in that the metal blocks can be connected to a voltmeter by conductive wires.

5. Device according to claims 1 to 4, characterized in that the metal blocks are connected to a temperature sensor.

Images