Hello to all :
This question is originated from the idea that every material (except for Superconductors) has a Seebeck Coefficient (S) different than zero, and from the idea that in every case when different metals or semiconductor materials are joined together, a Seebeck Effect at any scale is observed on the pair of materials.
So my question is:
When we measure the S of a given material, we place the sample between two plates at different temperatures, so we can stablish a Temperature Gradient and an unidirectional Heat Flux across the material. Then, we vary the Thermal Power imput, so we can vary the Temperature Gradient and obtain a set of Output Seebeck Voltages.
Then, we measure this Output Voltage and we plot a Graph of Delta(T) vs. Delta(V) as a linear x vs. y Graph. Finally we state that the slope of this Graph (positive for the N-Type materials and negative for P-Type) (because what we are measuring in our apparatus is the net Gradient (DeltaV/DeltaT), the formula in the Seebeck Coefficient has and additional (-) sign which turns S to a negative for N-Type semiconductors and positive por P-Type) is the Seebeck Coeff. for the material.
This measurement is always considered as if it was the absolute S coefficient of the material. But, What about the junction between the probe electrodes and our sample ? Since there is a Seebeck Voltage being generated at the junction too. Hence, our lecture from the voltmeter should be the Seebeck Coefficient of the junction: Se,s = Se - Ss
Where:
Se,s : is the Seebeck Coefficient of the junction between the sample and the electrode.
Se : is the Seebeck Coefficient of the Electrode.
Ss : is the Seebeck Coefficient of the Sample.
What do you think ?
Is this error virtually zero in practice, as much as we can ignore the effect of the Seebeck Effect of the junction electrodes/sample ?
How can we understand the fact that when we use these methods, we never talk about the contribution of he probe electrodes into the measured S Coefficient ?
Kind Regards !
Dear Thomas :
First, thank you very much for your answer, it's really motivating to read nice comments like these, which seeks to push someone who it's just in the process of gaining experience towards a positive direction.
Let me say first Sir, that you got the point correctly. Your idea is actually really good, and I understood the process correctly, thank you for the detailed explanation.
I will try to accompany my experiments with these kind of tests, and then, as you recommended, compare the results with my previous results, that way I will be able to spot the effect of the junction of my leads on my current measurements.
I believe that disruptive advancements in science come out when one puts to the limits its own findings and results, with even more constrained and more strict experimental conditions, always taking your experimental hypothesis to the limit. But sadly, sometimes to do this is not easy thing, since very often the time constraints and other's expectations for our results in our work seems much more important. For example, when we have to finish our thesis and we are squeezed in time, or when we hand in the results report to close the project in time, or the pressure to submit as many papers as one can.
Hopefully we can think more on the lines of scrutinize more our hypothesis than just with a result-based mind.
Send you kind Regards ! :)
Dear Franklin,
indeed, the experiments to carry out thermoelectric maesurements, must be carried out very carefully. Especially the different contacts to a thermocouple must be carried out identically respective material, the purity of the material, the temperature of the material and the thickness of metallic layers. This is often a very time consuming work. Some reference materials like gold exist. You can use superconducting wires (S=0) but it is a PhD work to find the right coefficients for a certain material. We determined shortly S values using Pt layers as contact material. You would assume, that Pt is a well known material, but for layers in the nanometer region it is not. The known models don't take into account surface effects which reduce S. So, you cannot say, a contact is a contact. All relationships must be very carefully studied.
With Regards
R. Mitdank
Dear Thomas and dear Rüdiger :
The samples I'm trying to make the measurements for are Bi2Te3-based alloys
Yes, it's a delicate consideration of the experiment.
Rüdiger Mitdank, what do you think about the last Thomas' comments ?
Best Regards both !,
Thank you for your help and advices,
Dear Thomas Anthony Troszak ,
if you use extensions of the same material and connect them under the same conditons, the additional thermovoltages rule out. But you must ever be sure, what you exactly do. Let me give an example used in practice:
Thermocouples are used in vacuum container to determine temperatures. Lets say, you want to bake out a recipient or you want to heat a sample holder to a certain temperature. You fix your thermocouple to the corresponding point the temperature T you want to control. It can be that the couple is not long enough. Furthermore, you have a feedthrough you must connect your wires with it. Then you have almost different metals. On the outside you must use a special plug which contains a contact of the identical material combination. The material combinations are standardized (example type K). The outer contact lies at room temperature. With a control unit which measures the T you control the heater. Of course, if you have more extensions, the contacts between different materials can be different (oxidation, lubrication). Therefore, it is important, if you want exact results, to control the measurement of T.
First test: The thermovoltage must be zero, if both contacts have the same T (room temperature). If there are any contact potentials then you see them.
Second test: You put your "hot contact" into a reference liquid (boiling point must be known). Then you must get the right temperature difference.
With Regards
R. Mitdank
Dear Rüdger M. :
Thank you for the insights ! your example is very explanatory, then one way to consider this effects, or to 'check' for them is via these two tests.
I think it should be more straightforward to conduct the first test than the second. Furthermore, I think should be 'doable' to control both contacts to the same temperature and measure the small potential drop (or gain) due to the contact, if there is some.
However, I will think on how to implement the second test to check again for a temperature measurement deviation expected because of the contact due to the use of extensions on the leads.
Regards !
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science