Talk:Thermoelectric effect

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Isn't there a Seebeck votage term missing from the joule heating in this equation ?[edit]

Something like

I think, has to be replaced by

With for Thomson-coefficient. Any comments on this? — Preceding unsigned comment added by (talk) 11:23, 16 June 2015 (UTC)

Very badly written article[edit]

This is a very badly written article about a very interesting topic, a major re-write is long overdue:

PLease re-organize this article from basic principles (thermodynamics) ---> details.

Put all the integrals at the END of the article, where they actually mean something, so that a person interested in this topic can read the article in a LINEAR fashion, without having to crypto-analyze it.

Martin —Preceding unsigned comment added by (talk) 23:28, 22 July 2008 (UTC)

I second the motion. (And "Charge-carrier diffusion", in particular, is a repetitious yet self-contradictory muddle.) (talk) 07:06, 2 August 2008 (UTC)

This article conflicts with itself[edit]

Joule heating has not been reversed, but is theoretically possible under the laws of thermodynamics ...and... The first term ρ J² is simply the Joule heating, which is not reversible

Planning major re-write[edit]

I've been working on a paper regarding thermoelectricity for a class, and it's taken me a while to get the factors straightened out in my own head. At this point, I think I can present the information fairly clearly. The peer review page doesn't seem to be there still, but I'll definitely be looking over the material here for suggestions on directions to take. -- Beakdan 16:52, 15 November 2007 (UTC)

Dead External Link[edit]

I've commented out the link to "Two Utah teens invented a durable, clean, efficient air conditioner based on this principle" because the page has been removed. It is however available in the google cache, so I'm not sure if whoever added that would like to do something about it to keep it around, or not. -Scott Dial 02:08, 3 September 2005 (UTC)

The google cache is now gone too.


How is the EMF generated and where does the engery for this come from?

It's described in the article. The hot and cold charged particles diffuse through the metal, but they move at different rates, leading to a net movement of charge from one end to the other. The energy comes from the applied heat difference. - Omegatron 03:14, August 2, 2005 (UTC)
Maybe the wording of this could be improved; Also is it required that there be two junctions? I find this point confusing, at one point it states that the EMF is due to the junction, where later on in the article it states that the EMF is due to the flow of heat, presumably if it is due to the flow of heat, no junctions are needed, not to mention two. mickpc
Also what is the relationship between the electropotential of the different metals (or semiconductors) and the EMF produced? mickpc
When a single bar of metal has a heat difference between the two ends, a voltage will appear between the two ends as well. If you connect two bars of the same metal in a loop, you will get no net current because the voltages generated in each will cancel out. If you connect two dissimilar metals in a loop and keep a temperature difference between the two junctions, the net speed of particle diffusion in each will be different and a net current will flow. That's how I understand it, anyway. - Omegatron 14:33, August 3, 2005 (UTC)
Any ideas on the difference between say the primary thermocouple material and the conducting wire so that the conducting wire does not effectivly cancel the EMF of the primary thermocouple? mickpc
Ok I have read my text again and it specifically says that the EMF is due to the junction not the tempriture difference in the main thermocouple material. So which is right? mickpc
It is an extremely common misconception that the voltage is either a function of the junction between disimilar metals or is developed across the junction. It is NOT a junction effect at all and there is typically no voltage across the juntion. It is an effect along the length of the conductor (a voltage gradient) that is a funtion of the thermal gradient (The Thomson/Seebeck effect). The only reason you need dissimilar materials is that otherwise the voltage gradients along two conductors would be identical and you would not be able to measure or untilize any voltage differential a the ends of the conductor pair. Perhaps someone can take the time to explain that better and add an illustration. Here is one reference that provides some explanation of what I am trying to explain,
I will try to find another good reference. (talk) —Preceding comment was added at 01:22, 9 May 2008 (UTC)

Year of discovery[edit]

The first line of the description is incorrect. It was first discovered by Thomas Johann Seebeck in 1821. Not by Peltier in 1884. Seebeck was not even alive anymore in 1884. He died in 1881!

I saw two pages that said 1821, but this page [1] says 1823. I intend to add more to this article. Omegatron


Here are some references you might find interesting.

The original reference for Seebeck's discovery is:

T. J. Seebeck, Magnetische Polarisation der Metalle und Erze durch Temperatur-Differenz, Abhandlungen der Königlichen Akademie der Wissenschaften zu Berlin (im Jahre 1825 gedruckt), S. 265-373, 1821.

This has illustrations just after the article.

This volume of the "Abhandlungen" was printed in 1825, but is for 1822-1823. The content of the article was presented before the Akademie in October, 1821.

The Akademie publications of the time were usually published some years after the initial work of an author. Very often one sees that the author presented his work to meeting of the Akademie (sitzugberichte = meeting reports) well before actual publication. Often slightly different versions would appear in other journals, notices etc. before the Akademie version went to print. This can be a bit confusing when working up citations.

The same material, with updates, is available online:

T. J. Seebeck Ueber die Magnetische Polarisation der Metalle und Erze durch Temperatur-Differenz, Annalen der Physik, (Poggendorff) Bd. 82, Heft 1, S. 1-20, Heft 2, S. 133-160, Heft 3, S. 253-286, 1826.

This is a three part article appearing in successive issues of volume 82. (The S in this citation is from the German Seite for page, Bd is from Band for volume, and heft means issue.)

The above article is from the Annalen right after Ludwig Wilhelm Gilbert (1769-1824) died and Johann Christian Poggendorff (1796-1877) took over as editor. Poggendorff renamed the journal Annalen der Physic und Chemie. In 1900, under Paul Karl Ludwig Drude (1863-1906), the name was changed back to Annalen der Physik. The numbering of the issues reflect the changes of editorship and this can be very confusing. In this case Seebeck's article appears before Poggendorff had his "neue reihe" where the issues were renumbered. This one missed the "mark of Poggendorff." Note that the online version uses a numbering for the entire existence of the publication and in this case it coincides.

Older journals are a pain to cite!

Zorpoid — Preceding unsigned comment added by Zorpoid (talkcontribs) 20:07, 3 October 2016 (UTC)

Direction of electron flow[edit]

"Electrons flow from the hot end to the cold end."

I am not so sure this is true. I think it depends on the type of conductors, so I am taking it out for now. Feel free to put it back if you know it is true. Omegatron

It is true in metals. Apparently not in semiconductors, in which "positive particles" - holes - are free to flow. I have expanded on it a lot now, though. (Great... now I'm talking to myself...) - Omegatron
I have a similiar question so I'm just going to post it here: In the graphic, an electron would flow from the minus pole, through the p-doped semiconductor, through the n-doped one, back to the plus pole. Now, since the conduction bands of the p-doped semiconductor are on a higher energy level than those of the n-type, shouldn't the electron EMIT energy, making the upper side the warm side? Also, following the same logic, the cold side should be at at the place, where an electron moves from the n-doped semiconductor to the p-doped one. Since such a junction doesn't exist in the picture, there shouldn't be any "cold side" at all. Am I wrong? — Preceding unsigned comment added by (talk) 16:04, 6 July 2014 (UTC)
The questions can be answered by this page: I think this is authoritative enough. Also, in the discussion of the Seebeck constant on the so-entitled Wikipedia page, it is explained that the material nature (makeup) determines whether the overall flow is from hot end to cold end, or vice versa. It is a very complicated and delicate measurement, and the effect depends on microscopic balance between the relative movement of positive and negative charge carriers. That page is a very good resource for this page. To answer the anonymous question above, the Seebeck effect, which causes the potential difference in the material, does NOT require a junction at all! Please read the article about the Seebeck constant for further clarification. So, in short, yes, you are wrong. A million Peltier cooling devices also say so :-). JoGusto (talk) 11:34, 20 August 2014 (UTC)

Redirect thermoelectricity to here[edit]

Just to make sure people know, I redirected thermoelectricity to here, although there might be other forms besides the peltier-seebeck kind?

I changed this back to a non-redirect, since there is also vacuum tube thermoelectricity. - Omegatron

Diagram check[edit]

Please double-check the polarity of my voltmeter in the diagram. I think it is right, but i am not terribly qualified. - Omegatron

What came first?[edit]

Did Seebeck discover the voltage difference first or the current loop/compass first? - Omegatron 17:09, Apr 2, 2004 (UTC)

I believe Seebeck THOUGHT he had discovered a thermo-magnetic effect, since he used two dissimilar metal wires joined at each end and forming a loop, in close proximity to a compass. Therefore I would suggest, from my limited reading on the subject, it was the current effect that he initially noticed.
I took the liberty to remove the word "is" from immediately before an expression because the subject of the expression was stated immediately before the expression, that is;
"If a current density J is passed through a homogeneous conductor, heat production per unit volume is q =......",
changed to;
"If a current density J is passed through a homogeneous conductor, heat production per unit volume; q =......"
I do not like the term 'thermopower' as an alternative for the Seebeck constant, since the Seebeck constant is measured in Volts and is in reference to the potential so I would assume is an open circuit quantity, no power would therefore be produced. I do realize that this is a term in general use and I would like to make it clear that I am not critisising the article, just putting an opinion forward for discussion. —Preceding unsigned comment added by (talk) 22:05, 1 January 2008 (UTC)


I would like to show that these are functions of T, but

are all ugly. - Omegatron 20:25, Jul 5, 2004 (UTC)

Phonon drag picture[edit]

I forgot which one it was but, why did you take off the change carrier diffusion/phonon drag picture? lol, I would not have read that paragraph or two if the picture didn't catch my attention. -- Mac Davis 11:13, 9 Jan 2005 (UTC)

File:Seebeck effect.png
The Seebeck effect in a bar of metal, showing the flow of hot electrons only.
The flame and the metal bar? I don't think it is accurate, after thinking about it. The first explanation I heard said that hot electrons would diffuse through the metal bar, which made sense to me, and explained the basics of how heat could generate electric current. So I drew that picture. Then I read more into it, and the cold electrons diffuse in the opposite direction in roughly equal quantities. The imbalance between the two is what actually causes the net electric current, and is caused by the phonon drag and scattering. So, since the picture only shows hot electrons diffusing, it's not accurate. I could make a more accurate one if you could describe what it should look like. I don't really understand phonons and have no idea how to draw phonon drag. :-) - Omegatron 18:42, Jan 9, 2005 (UTC)


Is pyroelectricity the same phenomenon as the seebeck effect?

I don't think so. One is in metals and one is in crystals. But I am no expert. - Omegatron 00:15, Feb 12, 2005 (UTC)
I think Omegatron is talking about piezoelectricity. The article states materials with one usually have the other. This article is about two dissimilar materials (metals) creating current between hot and cold ends, but pyroelectricity is about one material developing an electric potential when heated. Is that what you guys get from the pyroelectricity article?

Why not to make it in three pages?[edit]

I think that this page can be put in three:

Like you can see all those pages are already here, and making Peltier-Seebeck effect page in three part will just ask to remove the redirect and put part of article in. In fact main matter of having on page insted of three is that links to the others languages are falses, and that make bots from others' wiki copying bad links and put mess in link between the wikis. So I think that is a good thing to make this page become three. Please say what you think about it? Oliviosu 17:10, 10 July 2005 (UTC)

I combined the three articles into one a while ago, since they are all variations on the same effect. I think they should stay together. - Omegatron 19:41, July 10, 2005 (UTC)
Not true. "The three thermoelectric phenomena are the Seebeck, Peltier and Thomson effects. Of these only the Seebeck effect converts thermal energy to electric energy and results in the thermocouple voltage used in thermometry." CRC Press. Measurement, Instrumentation, and Sensors Handbook.
There's also no mention of absolute Seebeck effect. —Preceding unsigned comment added by Quincy8Boy (talkcontribs) 14:03, 2 October 2007 (UTC)

Links to the others wikis[edit]

For this time I set this page as if it was Peltier effect page by deleting links to Thermoelectricity that are aleredy in Thermoelectricity page, I also set nl language link to the nl peltier page insted of nl seebeck page because all links to other wiki are to the peltier effect pages and making that allow bot to work without putting mess. Oliviosu 17:28, 10 July 2005 (UTC)

Slashdot Warning[edit]

An article [2] on the Peltier effect was posted to recently. There are no links to this page, but I think that should account for the reoccurring vandalism.

Wow! My first chance to see a revert happen right before my eyes. (I pulled up the history and edit pages and it had already been reverted).--Lzygenius 05:36, 21 July 2005 (UTC)
May it be suggested that another revert and lock be put on the article until things settle down? Maybe a programmic lock on stuff that gets referred off of Slashdot for at least a week? --STrRedWolf 05:41, 21 July 2005 (UTC)
Yeah, I've protected it now. There's no problem reverting all the slashdot-kids, but having it open is just adding too much nonsense to the page history at the moment. Shanes 05:44, 21 July 2005 (UTC)
Maybe there should be a {{slashdotted}} template. - Omegatron 06:09, July 21, 2005 (UTC)
I don't particulary like that template myself. To me it looks more likely to be encouraging slashdot-kids to live up to the reputation of setting their mark on every site being linked to. "Yeah, we slashdotted it!!!!112!". That's why I was reluctant to protect the page myself in the first place. I hate giving in to vandals. Anyway, now that the current and protected version is exactly the same as the version before the slashdot-link, is it really neccesary to warn people about the acuracy of it? Shanes 07:05, 21 July 2005 (UTC)
I agree with Shanes. It seems to play into the vandal mentality. Better to just quietly resist, I think. --InformationalAnarchist 07:21, 21 July 2005 (UTC)
Yeah. Well, anyway, I'll unprotect it now. The worst should be over now. Shanes 08:29, 21 July 2005 (UTC)
Me too, I think there should be enough responsible Wikipedians directed here by the slashdot article at the same time as the vandals. I for one come and check Wikipedia articles that are linked to by Slashdot. Usually someone beats me to it and the vandalism is already reverted :-) I did once suggest protecting Slashdot-linked articles, but was shouted down and for good reason I think. PhilHibbs | talk 09:01, 21 July 2005 (UTC)
I agree that protecting the page is a bad idea. We want to convert them, if possible, and they won't be converted if they can't edit the page. :-) I think the current wording of the slashdot template could be changed. - Omegatron 14:05, July 21, 2005 (UTC)


From the article on 20050904, Superconductors have zero thermopower, and can be used to make thermocouples. Should that read, "... and can't be used to make thermocouples." It seems to me that with no voltage drop, there can be no temperature drop either. If it means that they can be used in conjunction with other materials to make thermocouples, then I still think it's a bit confusing.

I'm pretty sure cryogenic coolers can be made from high-Tc superconductors [3], though I do not understand the mechanism at present. —Preceding unsigned comment added by Parthi.santhanam (talkcontribs) 20:58, 27 July 2008 (UTC)

Elite Language[edit]

can somebody translate the article to common english? the language used is CRAP. the author of the article should ask him/herself: for what kind of audience am i writing?! answer: OH, I AM WRITING IT FOR EVERYONE ON THIS PLANET FROM THE MOST DUMB TO THE MOST INTELLIGENT! noone wants to decipher an article before one is able to read it.

Examples? — Omegatron 20:31, September 11, 2005 (UTC)
The language seems to require a basic high school science education. I do not think that is unreasonable.
I agreee, the language should be changed
Examples? — Omegatron 18:32, 2 March 2007 (UTC)

Seebeck-Peltier-Thomson: Who Discovered What[edit]

The Wikipedia article describing the Seebeck effect (see above) suggests that it was Thomas Johann Seebeck who discovered that "a voltage existed between two ends of a metal bar when a temperature gradient existed in the bar".

However, it is more likely the case that Seebeck never actually knew that this fact was true. The explanation for this is as follows:

Textbooks often propose that there are only three fundamental thermoelectric effects, however it is in fact possible to describe four.

The four thermoelectric effects, listed in chronological order of their discovery, are:

Effect 1 - If two different conductors are joined and the two junctions are maintained at different temperatures, an electromotive force is developed in the circuit.

Effect 2 - If a current flows in a circuit consisting of two different conductors then one of the junctions is heated and the other is cooled.

Effect 3 - When a temperature difference exists between two points in a single electrical conductor an electrical potential is established between the points.

Effect 4 - If a current passes through a conductor in which a temperature gradient exists, this current causes a flow of heat from one part to the other.

These effects are obviously very closely related. Indeed, each of them represents a reversible effect whereby effects 1 and 2 are the reverse of each other and, similarly, effects 3 and 4 are the reverse of each other.

Thomas Johann Seebeck first identified Effect 1 in 1821. He spent the rest of his scientific career measuring the size of this effect for different pairs of dissimilar conductors in contact with each other. Seebeck died in 1831.

In 1834 Jean Charles Athanase Peltier first identified Effect 2, the reverse of Effect 1. Peltier died in 1845.

Significantly later (around 1854-1855), William Thomson first deduced and demonstrated BOTH of the effects numbered 3 and 4.

Starling and Woodall describe part of Thomson's contribution thus (from "Physics", Longmans, 1950):

"He [Thomson] suggested that there must be other electromotive forces in the circuit and that these exist in the metals themselves, acting between the parts of any one metal at different temperatures. This was found to be correct. Thus if two points in the metal differ in temperature by the amount dT, the electromotive force in this element of the metal is s.dT. The quantity s is called the Thomson coefficient. It is taken to be positive when directed from points of lower to points of higher temperature."

As a result of the above, the four thermoelectric effects are correctly attributed the following names:

Effect 1 is the Seebeck effect.

Effect 2 is the Peltier effect - and is correctly identified as the reverse of the Seebeck effect.

Effects 3 and 4 together comprise both "directions" of the Thomson effect.

Some recent sources restrict the definition of the Thomson effect to that of Effect 4 only, and this may be either the cause or the result of a further tendency to prefer that the definition of the Seebeck effect may be satisfied by that of Effect 3 (with the possible consequence that Effect 1 is rendered anonymous).

Again, it is clear that the relationship between Effect 1 and Effect 3 must be a very close one.

However, it has been demonstrated that during his lifetime Thomas Johann Seebeck could not ever have been explicitly aware of Effect 3.

Furthermore, in the effect which Seebeck spent the greater part of his career measuring, when the junctions between the dissimilar metals are maintained at different temperatures a net electromotive force exists in the circuit which causes a current to flow around it. Such a circuit cannot be constructed with a single conductor, and therefore the definition of Effect 3 may not serve as an adequate explanation for the Seebeck effect.

It is not necessarily erroneous to say that there are only three thermoelectric effects, so long as it is understood that one of them, the Thomson effect, comprises both "directions" of the reversible effect; and that one of the others, the Seebeck effect, is the reverse of the Peltier effect and requires at least two dissimilar conductors to produce.

References: "Physics", Starling and Woodall, Longmans, 1950; "The Penguin Dictionary of Physics", Revised for the third edition by J P Cullerne B.Sc. D.Phil.

Keith P Walsh

Add this to the article! (With references, of course.) Very good info. — Omegatron 14:12, 7 December 2005 (UTC)
Ok, you literally added this section to the article. :-) That's not quite what I meant. I meant to modify the article's descriptions about each effect in light of your discoveries, with a reference for each fact changed. You're not going to append a paragraph onto the end of the article; you're going to edit the actual article to reflect your changes. Also, we don't include self-references like "The Wikipedia article above". — Omegatron 21:13, 22 January 2006 (UTC)
Thank you for your advice. The statement in the article that Seebeck "... found that a voltage existed between two ends of a metal bar when a temperature gradient dT existed in the bar." is inaccurate. Perhaps you might like to assume responsibility for correcting it yourself. --Keith P Walsh
Il you wish (and if you speak french), you can find a detailed description of the thermoelectric effects discoveries there : fr:Thermoélectricité (this article is a brief summary of my thesis introduction). And I can confirm that the statement that Seebeck "... found that a voltage existed between two ends of a metal bar when a temperature gradient dT existed in the bar." is inaccurate : Seebeck found a voltage at the jonctions of of two dissimilar metals. Sincerely, David Berardan 11:00, 27 January 2006 (UTC)

Good Article nomination has failed[edit]

The Good article nomination for Thermoelectric effect has failed, for the following reason(s):

  • The lead section should be expanded to summarise more of the article's content.
  • The article may be too technical for a general audience to understand.
Also, what's the relevance of the Patents section? It's not much use at the moment as a list with no context. I'd suggest removing the see-also section; relevant links are already in the text. And you might consider removing sub-section headings in 'external links' as it bloats the TOC somewhat unnecessarily. Worldtraveller 09:57, 14 June 2006 (UTC)
I agree about the patents - those aren't even particularly notable among the hundreds of TE-related patents. The Hansell patent isn't even for a TE device - it's a thermionic converter (?!). I'm taking that whole section out. If anybody wants to put it back in, please select only patents that are (a) relevant to the topic and (b) unusually significant, and add some commentary to justify it. Tarchon 20:08, 30 October 2007 (UTC)

Shame that this article is considered too technical for a general audience to understand! As a general reader I found it brilliant, clearly expressed and utterly fascinating. Maybe the world itself should be rewritten in a less complex form, so that some grumpy lazy people don't have to work at understanding it.

Peltier effect[edit]

I think this needs to point out in order to work the two metals need to have significatly different work functions. Otherwise the current itself will cause more heating than the cold end can compensate for.

So point it out. This is a wiki. — Omegatron 17:10, 9 December 2006 (UTC)
Uh... I did... I did this from memory and I was hoping someone with a reference available can confirm and thus make it more reliable. —The preceding unsigned comment was added by (talk) 18:27, 15 December 2006 (UTC).

Thermal differentials[edit]

Hello, I am new, but trying to be bold. I thought that "thermal" made a lot more sense than "heat" in the first sentence of the article since heat is not a quantity in the static sense, only as a flow. The old Caloric theory would say otherwise but it has of course been superceded. I think that the passage both makes more technical sense this way and the high school science educated reader can still understand it. Feel free to revert and discuss, I won't be offended! Wes Hermann 02:37, 5 December 2006 (UTC)

I've amended it to "temperature differences" - differentials and differences aren't the same thing, and TE is all about temperature differences. I think this is part of the language that other people were complaining about. "Differential" as a noun is a technical term from mathematics that's sometimes incorrectly used as a generic synonym for "difference". Tarchon 18:38, 30 October 2007 (UTC)

Joule Heating ≠ Thermoelectricity[edit]

I made some big changes in the intro, and the biggest one is that Joule heating is no longer identified as being a TE effect. I know it's related, and if you want to parse "thermoelectric" it seems like JH would be part of it (this is a semantic error that formerly plagued the thermoelectricity article), but in common scientific parlance, Joule heating is not usually considered a TE effect. It's extremely common to draw a dichotomy between thermoelectric and Joule/Ohmic behavior in the technical literature. As the article notes, reversibility is a key difference here. Tarchon 19:57, 30 October 2007 (UTC) doesn't work[edit]

The Link to related sites below contains a link which cannot be loaded:

Kind regards, Sebastian Morkisch -- S.Morkisch (talk) 21:22, 10 April 2008 (UTC)

Image Problem[edit]

I think the image "Thermoelectric Cooler Diagram.svg" is wrong, because it shows current coming out of the positive terminal of a battery and going into the negative terminal. Charges should be moving out from the negative and into the positive terminal. —Preceding unsigned comment added by (talk) 23:05, 14 May 2008 (UTC)

The direction of an electric current is defined as the direction in which positive charge carriers would flow. This is the opposite of the direction negative charge carriers (electrons) actually move. The direction of the current in the diagram is therefore correct (from the positive terminal, to the negative). Any introductory text in physics will confirm this. O. Prytz (talk) 20:04, 21 May 2008 (UTC)


Who is the admin of this page? I would like to discuss the future of the [*thermo*elec*] pages. -- Parthiban Santhanam (talk) 00:43, 28 July 2008 (UTC)

Commercial Power Generation Kit available - but ethical to post to main page?[edit]

Nextreme has come out with an evaluation power generation kit based on the Seebeck effect. Current page as of 2008-Dec-18 is: I considered posting this to the main page in a "Related Links" section, but thought I should put it here for discussion for 2 reasons:

  • It does not seem ethical to post a commercial offering, even though an evaluation kit, even if "reasonably" priced. Nextreme's kit is $295 What if everyone would do such a thing? Countering this question/point, perhaps some people *are* looking for such information; posting it here would help them quickly find it (that being the point of Wikipedia, right?), especially if there were several such listings to review and evaluation. Of course, a problem could be if this list gets to be too long. But then again, I have seen some "postings" of books in external links that seem to be more a promotion of the book and/or author than relevant information for the article in question.
  • The offering is probably time-bound, meaning that it will expire after a while, and what good is the link then? Well it is good in the sense that one knows what company is working in the particular topic of concern, but is that worth a "dead" link? —Preceding unsigned comment added by Dan Aquinas (talkcontribs) 22:41, 18 December 2008 (UTC)
Hi Dan! Two things: First, whether a link is improper advertising or not depends a lot on context. In particular, it would be really great if there was a paragraph or section describing the state of the commercial thermoelectric generator industry, with links to multiple companies (including this one), and the various different companies emphasized roughly in proportion to how important they are in the industry (as gauged by third parties). Obviously, this would take some work, but it would be a very useful addition. On the other hand, linking to just one company out of the whole industry might be (counter-intuitively) worse than linking to none, because it gives a misleading impression of how important that company is (see WP:NPOV) and comes across advertising.
The second thing is that this article is (or should be, in my opinion) about the thermoelectric effect as a phenomenon in physics. The articles thermopile, thermocouple, and thermoelectricity (for example) might be a better place to be discussing the devices based on the thermoelectric effect. --Steve (talk) 00:18, 19 December 2008 (UTC)


The article states the carnot efficiency is very low, between 0.3 and 0.6. Isn't this actually ranging from 'OK' to 'good' if you compare it to car engines or power stations? I've removed the very low comment. (talk) 11:17, 14 August 2009 (UTC)

COP is coefficient of performance for a heat pump, I think. It doesn't have to be between 0 and 1, it's usually higher than 1. Rather than saying high or low, it should be compared to a normal refrigerator at the same temperature, but I don't know that figure... --Steve (talk) 14:56, 14 August 2009 (UTC)
I don't think it is. Peltiers can of cource be used as heat pumps, but this is the efficiency as a method of generating electricity (or at least is says that at the top of its paragraph. If that's the case then like a heat engine doing work, it cannot be more than (talk) 14:22, 17 August 2009 (UTC)
I read the paper, and rewrote the paragraph accordingly. Is it clearer now? --Steve (talk) 14:53, 19 August 2009 (UTC)

Kelvin Probe[edit]

The Kelvin Probe is probing the Work-function. Though there are a few books and articles that interpret the Seebecke-effect as the temperature dependence of the work function - this is just not correct. The Seebecke-effect is a bulk effect, while the work function depends on surface properties like adsorbed gas as well. So the Kelvin Probe will show something different that the thermal EMV. In addition the resolution of normal Kelvin Probes is just not sufficient for voltages in the µV range. So it's probably better not to mention the Kelvin Probe int this articel.--Ulrich67 (talk) 16:53, 3 January 2011 (UTC)

I agree with your point about work functions: Here's an article that I think explains pretty well why the Seebeck effect is not related to the temperature dependence of the work function. (As a side note, I think in semiconductors under certain conditions and assumptions they are approximately the same, but that's just an interesting fact with no deeper significance.) I don't really understand the article's text about kelvin probes, I can't even figure out whether or not it is contradicting anything you're saying. --Steve (talk) 02:17, 4 January 2011 (UTC)
I think this paragraph is not well written. For most readers this will only lead to confusion. The possibly relevant part that may have been intended with this paragraph would be that the Seebecke-voltage is something different from the temperature dependence of the work function. I see no reason to explain how the work-function is measured. The underlying physics actually does not need a second material, and hence no contact.--Ulrich67 (talk) 18:14, 4 January 2011 (UTC)


If power is applied to a peltier, upon disconnection there should be a potential difference between the terminals, but which polarity? Does it work like a battery, where positive voltage is applied to the positive terminal when charging? Or does current flow in the same direction after a power supply is exchanged for a load. The second coloured diagram in the article has no polarity marked. — Preceding unsigned comment added by (talk) 10:38, 13 December 2011 (UTC)

A Peltier connected to two thermally-insulated reservoirs functions like a rechargeable battery, and same polarities as a rechargeable battery. The direction of current flow is linked to the direction of heat flow. Either the current flow is always the same direction as heat flow (if the material has positive Seebeck coefficient), or it's always the opposite direction (if the material has negative Seebeck coefficient). (Maybe I have it backwards.) (If the heat is moving from hot to cold, power is generated; if the heat is moving from cold to hot, power is consumed.) :-) --Steve (talk) 14:13, 13 December 2011 (UTC)

Seebeck coefficient[edit]

The term Seebeck coefficient is used in the Seebeck effect section but never defined. I think I can infer what it means from the context, but I would be a lot more certain of my understanding if it was clearly defined in the section. --Kierkkadon talk/contribs 22:12, 29 January 2013 (UTC)

Unwanted thermoelectric effects[edit]

The article is lacking a discussion of unwanted thermelectric effects, for example in precision voltage measurements. How to overcome it, maybe by using materials or material combinations that have a small thermoelectric effect? Starblue (talk) 17:30, 13 October 2013 (UTC) Document with some good info from Keithley: Making Precision Low Voltage and Low Resistance Measurements Starblue (talk) 18:43, 13 October 2013 (UTC)

That's a good point, including problems with building wiring that utilizes dis-similar metals. SoftwareThing (talk) 22:38, 24 September 2018 (UTC)

I defy anyone to find the Thomson (Kelvin) Relations anywhere in Thomson's writings.[edit]

I defy anyone to find the Thomson (Kelvin) Relations in Thomson's writings. They may be based upon equations in "On The Dynamical Theory Of Heat," but look throughout his complete works and you will not find the Relations. I have been unable to find any discussions of them in late 19th and early 20th Century writings on thermoelectrics, either. They do not seem to emerge until around the time that Lars Onsager introduced his work on reciprocal relations in the 1930's. The Relations appear to grow out of 20th Century work in thermodynamics and discussion of the Thomson Relations takes off from there.

It should also be noted that the Relation linking the Peltier and Seebeck coefficients through absolute temperature (π = αT), is invalid. Solid state theory shows us that the Peltier coefficient is actually dependent upon two materials: 1) the junction-to-junction material establishing the energy level for transport (at the conduction or valence band in TE materials), and 2) the conductor at each junction (usually plated copper) which conducts electrons in proximity to its Fermi level. It is the transition between these levels which determines the amount of heat absorbed or released. That quantity is π•I where π is both the Peltier coefficient and the difference in energy (in Joules per Coulomb) between the Fermi level of the conductor at the junction, and the conduction or valence band of the junction-to-junction conductor. Because the Peltier coefficient actually reflects the properties of two materials, it cannot possibly be derived from the Seebeck coefficient of the junction-to-junction conductor (which has a value independent of any other materials) and the absolute temperature. The Relation is invalid. By the way, even though there are no junction conductors present along the length of the junction-to-junction conductor, the Fermi level remains relevant in understanding how the quantity of heat at any point, relates to the activity at the junctions. The temperature dependency of the Fermi level creates a virtual base line and the Peltier coefficient remains important in the mathematics of absorption, transport, and release throughout the length of a TE element.

This and a lengthy proof showing that Thomson Effect was not adequately demonstrated(based on flaws in Thomson's proof, an alternate thesis not explored, and his failure to sufficiently examine the other effects), will be in my upcoming book, Rethinking Thermoelectric Fundamentals Within A Temperature Dependent Context (Michael Spry), intended for release in 2016.

  • For both the sebecke coefficient and the peltier coefficient there are values for single materials and contacts or material pairs. for the pairs you get the difference in value of the two materials involved. The normal tabulated values are used referenced to platinum as a reference material and not the absolute skale, as this absolute skale is difficult to measure. However even this does not interfere with (π = αT). But one has to take the true single material coefficients when it comes to connecting to the Thomson effect. Ulrich67 (talk) 11:30, 3 January 2016 (UTC)

  • Hello Michael Spry and Ulrich. I think you nailed it, in Thomson's "On the Dynamical Theory of Heat", page 320 eq f. It looks like dπ/dT + tau = S, where tau is the Thomson coefficient. This is from December 1851. That's the first Kelvin relation. Later in 1854 he rewrites what he's done, but using a different notation. This is found at page 135 of that same document, eq.(16), this time with sigma_1 - sigma_2 = π/T - dπ/dT. Which is essentially the same equation where sigma_1 - sigma_2 is the difference in the Thomson coefficient of two different metals. Note that this means that S = π/T. This is precisely the 2nd Kelvin relation. So, I can see both relations in that work. What do you think? — Preceding unsigned comment added by (talk) 13:48, 26 February 2019 (UTC)

Thomson Relations[edit]

Observing thermocouple data, for example in Wikipedia's "Thermocouple", Seebeck voltage is nearly proportional to the temperature difference between the junctions for a large range of temperatures. That is, Seebeck coefficient is nearly temperature independent. Then according to Thomson Relations, see Article, Thomson coefficient is nearly zero for such ranges. On the other hand the heat content of charge carriers is temperature dependent, in many cases nearly proportional to the temperature. Therefore, when a charge carrier flows along a temperature gradient, it must deliver heat to its vicinity or absorb heat from it, that is, Thomson coefficient cannot be zero even for ranges where Seebeck coefficient is temperature independent.

This observation seems to support the paragraph above, that Kelvin did not invent Thomson (Kelvin) relations. (Urila (talk) 15:34, 2 February 2017 (UTC))

  • I think I've found the two relations in Thomson's work from 1851/1854 as I point out above to Michael Spry. What do you think? — Preceding unsigned comment added by (talk) 13:50, 26 February 2019 (UTC)

SEEBECK PROCESS - alternate usage of term, in photography[edit]

See "Practical color photography" by Wall, E. J. Chapter XV[1] Where he attributes to J. T. Seebeck the discovery that the action of light on silver chloride under the influence of the spectral rays assumed the colors incident on it, which has been attributed to Alexandre-Edmond Becquerel in work he published more than 30 years later. Is Wall correct, does the document sent to Goethe exist to prove this beyond question?


  1. ^ Wall, E. J. (Edward John) (1922). Practical color photography. Getty Research Institute. Boston, Mass., American Photographic Publishing Co. p. 200.

Applications, subsection Peltier[edit]

The article could benefit from noting other common applications. In absorption spectroscopy, a laser emitting diode is cooled using the Peltier effect, with the current going to the TEC proportional to the amount of cooling, if any, needed to maintain a constant LED temperature. A feedback thermocouple is used to determine the current needed to drive the TEC to maintain constant LED temperature with great accuracy. Laser emitting diodes rely upon temperature and current to determine their output frequency, and very tightly-controlled temperature of the LED means high accuracy in spectroscopic measurements.

Spy satellite optical and IR imagery also use TE Cooling devices to keep their detector junction arrays cold since nuclear power is long-lived and cheap on satellites whereas cryogenic fluids must be replaced on spacecraft which gets expensive and takes the craft off line during the process. SoftwareThing (talk) 22:37, 24 September 2018 (UTC)

This is why it says "Main article: Thermoelectric cooling", where all of these applied things can be discussed at length. I don't know why half the subsection is obsessed with the niche application of DNA PCR -- rather we should be describing the peculiar characteristics of peltier coolers that make them generally good for some applications and generally not good for others. --Nanite (talk) 00:04, 25 September 2018 (UTC)
Maybe applications could be listed under "see other" since there are a pile of real world applications. You're right, we don't want to clog the main article with possible irrelevancies. SoftwareThing (talk) 16:15, 25 September 2018 (UTC)