Talk:Thermal expansion

I just updated the page with the strain definition, but if you look at the formula it has an empty \frac tag. If you remove that the equation gets small. Anyone know how to fix that? - EndingPop 03:00, 12 April 2006 (UTC)Reply

Dear Writer,

epsilon_thermal cannot change inversely with temperature and be proportional to it at the same time (epsilon = alpha*delta_T)! Please, correct.

Regards, YS

—Preceding unsigned comment added by 132.207.82.124 (talk) August 14, 2006

There is a "hello!!!!!" at the beginning of the page which doesn't seem to exist in the source text. Can somebody fix this?

—Preceding unsigned comment added by 130.207.197.190 (talk) April 9, 2007

i just waaanna aaskk if thoose given with statement are in correct dimension of matter?

—Preceding unsigned comment added by 210.1.72.5 (talk) September 15, 2006

if:

${\displaystyle {\frac {}{}}\epsilon _{thermal}=\alpha \Delta T}$ 

is true, and:

${\displaystyle {\frac {}{}}\alpha }$  is the coefficient of thermal expansion in inverse kelvins

is true, thus:

${\displaystyle \epsilon _{thermal}\propto {\frac {1}{\Delta T}}}$ 

must be true, then ${\displaystyle \epsilon _{thermal}}$  must be inversely related to ${\displaystyle \Delta T}$ .

Da2ce7 15:55, 30 September 2006 (UTC)Reply

Sorry, but you're wrong. The proportionality is independent of the units on the coefficients. Think of it this way, if the temperature doubles, the thermal strain doubles. Therefore: ${\displaystyle \epsilon _{thermal}\propto \Delta T}$  - EndingPop 17:47, 2 October 2006 (UTC)Reply

"Polymers expand as much as 10 times more than metals, which expand more than ceramics."

its no where near that much, plastics 60-90, metals 10-30, so 4 times is nearer, comparing a rare high expansion plastic to a rare low expansion metal is misleading. —Preceding unsigned comment added by Asplace (talk • contribs) 13:59, 5 September 2007 (UTC)Reply

"Thermal expansion generally increases with bond energy"

isn't this the wrong way round? Asplace 15:52, 25 May 2007 (UTC)Reply

Do bones expand when heated? Like, does thermal expansion apply to everything? —Preceding unsigned comment added by 169.229.78.148 (talk) 07:10, 4 September 2007 (UTC)Reply

they must, all things do, even if its hidden by other effects or is very small, but bone is a complex and variable material, and people are very temperature stable (2C i'd guess), and the way bones/muscles work will mean a few percent size change won't be a problem anyway. —Preceding unsigned comment added by Asplace (talk • contribs) 15:02, 28 September 2007 (UTC)Reply

but it does make you wonder, if its taken into account when they estimate the size of a person from their cold skeleton? —Preceding unsigned comment added by Asplace (talk • contribs) 15:09, 28 September 2007 (UTC)Reply

Saying 'all things do' (expand when heated) is not correct. There are materials which contract when heated over very large temperature ranges Cubic Zirconium Tungstate is a good example 70.171.44.124 (talk) 01:05, 24 September 2013 (UTC)BGriffinReply

Hi Asplace,

Well done on the rewrite. It motivated me to go through the article editing for style and content. You will notice that I changed some of your text in the process. Feel free to change some back if you disagree; we can discuss as needed. --Slashme 08:25, 28 September 2007 (UTC)Reply

added a couple more things, i ran out of time before, and did change back the first sentence, its a fine point, but its the first sentence that needs to be most 'spot on', by which i mean clear and pedantically correct.

was worried about no citations but some of this basic stuff is very often not really stated directly.

i left the whole section about 'state equation' but do think its a bit over the top and basically BS, in that its a lot of frightening maths for a very small point, how do you feel about loosing it or simplifying and dropping it down the bottom? Asplace 14:36, 28 September 2007 (UTC)Reply

I think the 'Bet you didn't know that.' comment at the end of the article is inappropriate. I personally am aware of that, and I think it is beyond the tone of the good material on Wikipedia. Also, near the start of the article it says that the particles of a substance become 'active' when they are heated. The activity was the same all along: thermal motion. The thermal motion, average knetic energy, specifically just increases.

These two articles seem fairly intimately related. In some cases they seem to mention the same / similar points. I don't see a clear rationale for keeping them separate. Thoughts? Dhollm (talk) 11:10, 17 August 2010 (UTC)Reply

Agreed. Wizard191 (talk) 14:12, 17 August 2010 (UTC)Reply

Strong agree. It is clear this subject would benefit from being centralized.--DragonFly31 (talk) 19:48, 22 August 2010 (UTC)Reply

Thanks for the input, I've performed the merge. David Hollman (Talk) 14:09, 24 August 2010 (UTC)Reply

  Done

Thank you! NCdave (talk) 22:20, 4 June 2013 (UTC)Reply

A strong rationale for keeping separate articles is to underline the distinction betweem them, one is a phenomenon while the other is a physical quantity like in the similar situation of heat transfer and heat transfer coefficient.--82.137.10.164 (talk) 11:32, 10 October 2017 (UTC)Reply

The articles were merged seven years ago. If you want to propose forking them then start a new thread. Meters (talk) 18:20, 10 October 2017 (UTC)Reply

The format of this page was pretty messed up and looking at the history there were a couple cases of vandalism a while back which were unnoticed at the time. I have restored content which was deleted and added some subst:unsignedIP templates in various places as well. I also added a few headings which were not original just to make it readable. Hopefully this all makes the page more usable.

Dhollm (talk) 20:54, 22 August 2010 (UTC)Reply

This section has a table with a sorting function which doesn't work exactly right and I don't have any idea how to fix it. If you sort by linear expansion you'll see gasoline with 317 get sorted into the set of materials with COE of about 3 rather than with other high-COE materials. And what is odder is that it won't do that the first time. Or at least it didn't for me. The first sorting looks OK, with gasoline down with ethanol. To see what I mean, I'd invite the usual editors of this page to play with the sorting functions, then refresh the page and try it again. I'm using Mozilla Firefox, if that has any bearing on it. ???? Trilobitealive (talk) 04:35, 12 January 2011 (UTC)Reply

If it makes you feel better, I have the same problem. The first time I hit the sort its fine, after that it's all mixed up. You might want to post the question over at Help talk:Table, although I'm not 100% sure this is the best place to ask. Wizard191 (talk) 14:26, 12 January 2011 (UTC)Reply

Due to the merger of two articles, this article was left with a mixture of notations for the thermal expansion coefficients. One article used ${\displaystyle \alpha }$  ${\displaystyle \beta }$  for the linear and volume coefficients of thermal expansion, respectively. The other article used ${\displaystyle \alpha _{L}}$ , ${\displaystyle \alpha _{A}}$ , and ${\displaystyle _{V}}$  for the linear, area and volume coefficients of thermal expansion, respectively. The current article uses a mixture of both. We should pick just one convention, although we could mention the other. Cardamon (talk) 22:54, 20 January 2011 (UTC)Reply

Does the formula for length expansion assume the material is not under any stress or does it merely assume that the material is "free" to expand? (This question is relevant to how elementary physics is taught. For example, in the thermodynamics section of an elementary physics video ("The Ultimate Physics2 Tutor") there is the example of finding the work done by a column that has a weight on top of it when the column is heated. It is solved using the linear expansion formula without any discussion of stresses or strains due to the weight on top or the weight of the column itself.)

Tashiro (talk) 16:08, 15 March 2011 (UTC)Reply

A material has to be free to expand in order for the dimensions of the material to change. If the material is not free, internal compressive stresses will result. Think of it like a gas inside of a bottle that wants to expand as it is heated. Because the bottle has put a dimensional constraint on the gas, the gas cannot expand. Instead of expanding, the gas pressure inside the bottle will increase. Solids are very similar in that the material will want to expand when heated and will apply a force to anything trying to prevent that expansion. Hence, the weight on top of the column rises when the column is heated. NedWit (talk) 23:30, 24 June 2013 (UTC)Reply

If the material is constrained so it cannot expand, then you can calculate the stress by using the thermal expansion equations and imposing compatibility with the boundary conditions that constrain it. - EndingPop (talk) 13:54, 1 July 2013 (UTC)Reply

Typical data are for a material free of any stress except normal atmospheric pressure (or possibly other value if noted). If a material is under stress is modified a little as the elastic constants usually depend on temperature and so there is an additional change of the elastic strain superimposed to thermal expansion.--Ulrich67 (talk) 21:13, 28 October 2014 (UTC)Reply

...differ from those on engineeringtoolbox.com See ethanol and gasoline for eg. 89.176.34.187 (talk) 14:26, 21 November 2011 (UTC)Reply

The german wikipedia gives for Ethanol: γ(20 °C) = 1.4 × 10^-3 K^-1; the source is the CRC Handbook of Chemistry and Physics. Regards, Herbmuell (talk) 13:01, 2 January 2016 (UTC)Reply

The usual formula with ΔT is obviously inconsistent (of course, it is approximate, but why should it be inconsistent?). So, when length of a rod is described as a function of temperature, is this function better approximated as linear or as exponential function?--Ilevanat (talk) 01:06, 21 December 2011 (UTC)Reply

The shape of the thermal expansion vs. temperature curve strongly depends on the material and temperature range. There is no general function, but linear approximation works best, at least in a narrow T range. Any strong function like exponent, while giving a better agreement in some very specific cases, will lead to a much stronger disagreement in most other cases. Materialscientist (talk) 01:37, 21 December 2011 (UTC)Reply

The usual textbook formula with ΔT is typically applied to the "initial length" measured at the initial temperature of the interval ΔT. Heating a 1m rod from 0°C to 20°C, using an exaggerated expansion coefficient of 0.02 (for clarity), gives the "final length" of 1,4m. But if the rod is first heated from 0 to 10°C, and the "intermediate-final" length of 1.2m is then used as "initial" to describe heating from 10°C to 20°C, the "final length" at 20°C will be 1.44m (this is what I called "the obvious inconsistency" of the formula). And, of course, with sub-division of the interval into small steps one gets the exponential expansion using the same formula.

A consistent linear formula would need only a minor modification (I have seen it in some textbooks): e.g. the "initial length" measured always at 0°C, and ΔT replaced by the "final temperature" t(°C). Then in the above example one arrives at 1.4m regardless of the interval sub-division.

I know that these differences are far smaller (probably irrelevant in practical applications) for the usual coefficients of about 0,00001. But why the obviously inconsistent formula is preferred? Am I missing something that makes its usage more appropriate?--Ilevanat (talk) 17:03, 21 December 2011 (UTC)Reply

Sorry, for a quick comment - got to run. Experimentalists measure length of a rod vs. temperature that gives L(T) function, which is unique for every material and can't be approximated by any general analytical function. To simplify life of engineers, they use the universal approximation that any smooth physical function can be expanded into series by a small parameter, which is temperature change in this case. This gives aΔT, with a note of the temperature range where coefficient a is valid. I see no reason to choose exponent, or 0 °C (why not 20 °C then? Note that many materials are aimed for use above or below room temperature). Materialscientist (talk) 00:58, 23 December 2011 (UTC)Reply

Again, thanks for the response. I used 0°C only as an example; the appropriate "initial temperature" should be somewhere in the middle of the range for which the given value of the coefficient may be used in that simple approximation. But perhaps I was not sufficiently clear: it is the "initial length" that should always be set at that "initial temperature" if a truely linear approximation is used for L(T). I would not mind the use of ΔT, if it did not imply that the "initial length" can be taken at any temperature that is selected as the starting point of the interval ΔT. After all, if a "smooth physical function is expanded into series by a small parameter", the crucial element is the starting point arround which it is expanded.

To summarize: the "initial temperature" should be specified along with the expansion coefficient value, and ΔT should always be taken from that temperature on. This cannot be too complicated from the engeneering point of view, and this would have a clear advantage from the educational point of view: the formula would not be obviously inconsistent.

Otherwise, I would have to tell my students: "The engineers are to stupid to calculate an expansion from 20°C to 40°C if they have to use the initial length at 10°C"--Ilevanat (talk) 02:06, 25 December 2011 (UTC)Reply

Sorry, I don't understand. Absolute length is not a material property; thermal coefficient is a material property and gives a relative expansion ΔL/L₀. Temperature range for the coefficient a is merely the range where its value is valid. There is no use providing anything else. A properly educated engineer, if needed to calculate expansion of his rod from T1 to T2, measures L₀ at T1 (i.e. L(T1)) and calculates expansion as a×(T2 - T1)/L₀(T1). Materialscientist (talk) 02:52, 25 December 2011 (UTC)Reply

Thank you again! You made a very good point here, and I think I might change my approach and stick to that: the engineer takes a length measurement at the present temperature, and makes a single estimate for the expected future temperature using the simplest approximate formula, and that is as good as it gets (unless he wants to study detailed L(T) dependance for that material). (After all, thermal expansion is not a major theme in my lectures, and whoever might need to use it in later applications should better comply with the typical practice.)

So, what was all the above fuss about (if you should care to know)? You know the educators, we conjure hypothetical and frequently unrealistic problems for students to practice. So, for example, imagine that the engineer measured 1m length at 0°C and estimated 1.20m at 10°C (the example above). And when the rod heated, he indeed measures such length at 10°C. But now they ask him to estimate the length at 20°C: should it be 1.40m or 1.44m? I guess the best way to deal with that is to avoid such "problems" as unrealistic.

As for the other aspect: in physics we teach that same equations have same solutions. So most of my students know that the differential equation (from this article and elsewhere)

${\displaystyle \alpha _{L}={\frac {1}{L}}\,{\frac {dL}{dT}}}$ 

has the exponential function for the solution. For small temperature intervals we can use only the first two terms in Taylor series ${\displaystyle f(a)+{\frac {f'(a)}{1!}}(x-a)+{\frac {f''(a)}{2!}}(x-a)^{2}+{\frac {f^{(3)}(a)}{3!}}(x-a)^{3}+\cdots .}$ 

where both the function value f(a) (the "initial length") and its first derivative (the expansion coefficient) are taken at the same "initial temperature". But such details can be dismissed with the claim of small coefficients and simple approximation.--Ilevanat (talk) 02:05, 26 December 2011 (UTC)Reply

Rubber certainty does contract upon cooling. As a machinist, the mould I made for a finished rubber product had to be sized quite larger because the rubber shrinks substantially upon cooling.

I agree with the discussion below about a locked in stretched state being released to shrink upon heating. A common example is heat shrink tubing. That is not thermal expansion since it does not regain it's lost diameter upon cooling.

It's quite anomalous for a substance to expand upon cooling. Water is the most commonly known that does. Water expands upon cooling only in a very narrow range of temperature just before freezing. When water expands as it forms ice, that is different because that is due to the nature of the open structure of it's crystalline state. — Preceding unsigned comment added by Ronald H Levine (talk • contribs) 00:54, 4 February 2014 (UTC)Reply

Rubber has a negative thermal expansion coefficient. There is/was a special word for these materials, which escapes me. Usually now termed NTE materials. Most elastomers do have NTE.220.244.84.181 (talk) —Preceding undated comment added 23:28, 10 April 2012 (UTC).Reply

Yeah, that's what I thought, that's what the Berkley Physics Dept. says, and that's what this book says.

The Wikipedia article, however, gives positive values: 77 (linear) and 231 (volumetric), which might have come from here.

Berkeley is a great school, but their physics demo for undergrads is not peer reviewed research. Furthermore, the demo does not measure thermal contraction directly; it merely measures change in tension. The cited book [above] that agrees with the Berkeley demo has a copyright of 1986. 40 years ago I learned the same in engineering. It seems that we were taught wrong. So, the solution is to have 2 entries for rubber: “rubber, not under tension: 77 231” [and] “rubber under tension: see https://en.m.wikipedia.org/wiki/Gough%E2%80%93Joule_effect “ Having 2 entries will save the next wiki patrons a lot of time and will allow them to choose the correct coefficient based on their use for the rubber. — Preceding unsigned comment added by 74.77.148.238 (talk) 12:06, 7 July 2018 (UTC)Reply

This book says, "The coefficient of thermal expansion of a rubber compound depends largely on the type and amount of fillers incorporated in the crude rubber," this book says, "The coefficient of thermal expansion for elastomers is approximately 4.8 x 10^-4/K, similar to a hydrocarbon liquid," and this book says, "The coefficient of thermal expansion of rubber compounds varies from 0.00066 for a pure gum mix down to approximately 0.00048 depending on the amount of the loading."

Considering that numerous "Reliable Sources" disagree, not only on the value of the thermal expansion coefficient, but even on its sign, I'm going to change the article to simply say "disputed" for the values. NCdave (talk) 22:07, 4 June 2013 (UTC)Reply

Rubbers do not have a negative thermal expansion coefficient. The effect where a rubber contracts when heated under load is called the Gough-Joule effect. This phenomenon only occurs if the rubber is stretched when heat is applied. If the rubber is not stretched, it will still expand when heated. NedWit (talk) 23:19, 24 June 2013 (UTC)Reply

Rubbers, both natural and synthetic, absolutely definitively have positive linear thermal expansion coefficients, regardless of fillers, plastisizers etc.. The only thing that can happen, which may look like the rubber is shrinking, when heated, is the evaporation of plasitsizers. Also, rubber material that has been stretched to the limits of its entropy elasticity may shrink in the dimension transverse to the stretching direction, when heated. but these are exceptions! 79.235.241.19 (talk) 18:01, 26 November 2013 (UTC)Reply

Note to article authors - rubber has a positive thermal expansion coefficient as described by NetWit above. If you have concerns about types of rubber, consider providing a range of values or adding categories for filled and unfilled rubber, etc. — Preceding unsigned comment added by 156.65.14.229 (talk) 15:53, 21 April 2014 (UTC)Reply

In order for a material to have a negative thermal expansion coefficient it must have a fairly ordered, or low energy structure that has lots of empty space. Then as it is heated and gets disordered it is capable of bending (due to movement around a low energy site) and filling in that empty space. A material like rubber does not match this.Black.jeff (talk) 22:51, 13 September 2014 (UTC)Reply

Here is the source of confusion. "Entropic retraction" aka "Entropic elasticity" aka Gough-Joule effect is the tendency of rubber to stiffen, or increase in modulus. This occurs because rubber stiffness is due mostly to entropy, and the contribution of entropy to free energy increases with temperature. It has nothing to do with the volume of the rubber. Regardless of stretching, rubber will change volume just like any other hydrocarbon liquid. The values should be changed to something like what is given above, this book says, "The coefficient of thermal expansion of rubber compounds varies from 0.00066 for a pure gum mix down to approximately 0.00048 depending on the amount of the loading."Polymerman (talk) 13:08, 3 May 2017 (UTC)Reply

Most isotropic materials increase volume (length in all directions) when heated, as interatomic/intermolecular distances increase. Fibrous, anisotropic materials tend to decrease length along the fiber direction when heated. Straight, parallel fibers have low entropy. Therefore, heating tends to kink and misalign the fibers, shortening them along the original fiber direction, and widening them in perpendicular directions. The confusion about rubber is due to its exceptionally high elasticity, ductility, and elongation. Rubber can generally be assumed to have randomly oriented polymer chains when in a relaxed state, displaying isotropic behavior (e.g. positive thermal expansion). As it is stretched to high elongation, the molecular chains principally align with the direction of tension, similar to a fibrous material. At some degree of elongation, the coefficient of linear expansion will reduce to zero and become negative, but only along the direction of tension. AweStrike (talk) 22:53, 24 October 2022 (UTC)Reply

Vandalism on this page on 11/2/12. Don't know how to fix it. — Preceding unsigned comment added by 128.111.179.98 (talk) 02:54, 3 November 2012 (UTC)Reply

I'm not sure what you see. I saw "boobies" in the lead, which was a remainder of the last revert. I added ?action=purge to the url string and hit <enter> to purge the page and the boobies disappeared. Materialscientist (talk) 03:02, 3 November 2012 (UTC)Reply

Typically in literature (actually I have never seen β for volumetric expansion before):

Solid

x_ΔT = x · (1 + α (ΔT) + β (ΔT)² + …)

α: 1st greek letter, 1st term in the Taylor series
β: 2nd greek letter, 2nd term in the Taylor series

Liquid/Gas

V_ΔT = V · (1 + γ (ΔT) + …)

γ: 3rd greek letter, γ ≈ 3 α — Preceding unsigned comment added by 87.171.50.251 (talk) 13:19, 14 July 2013 (UTC)Reply

'Examples and Applications', toward the end, lists examples for which engineering has had to talk thermal expansion into account. Two in the list need improvement or removal:

-'Rubber tires' This does not provide sufficient explanation to be helpful.

- Attributing poor cold automobile engine performance to poor fit not yet corrected by thermal expansion or engine components is highly questionable, unreferenced, and counter-intuitive to old physics thermal expansion demonstration of the ring and ball. Viscosity of oil and other fluids and significant quenching of combustion heat through the cold cylinder walls likely has far more to do with cold performance.

I have removed that.

(First of all, the proper term is "clearances", not "spacings"; that betrays a lack of knowledge of the subject).

The assertion about "large spacings" is not true -- in fact, racing engines typically have larger clearances than road engines.

The clearances that might affect performance (assuming reasonable clearances) would be piston to cylinder, and intake valve stem to valve guide. Piston-to-cylinder leakage ("blow-by") is mitigated by the piston rings; intake valve-stem leakage (diluting the fuel-air mix) would be minimal, and is mitigated by valve-stem seals on road cars (racing engines typically do without those seals, but would never be allowed to have significant wear there).

Engines with carburetors will not give full power output when cold, due to poor vaporization of the fuel or even condensation of the fuel-air mix, but that is not significant in fuel-injected engines.

BMJ-pdx (talk) 20:06, 30 June 2022 (UTC)Reply

My comments re intake valve-stem leakage apply to normally-aspirated engines, where leakage is of air from within the valve cover to the intake port. On a supercharged (including turbocharged) engine, the leakage would be of fuel-air mix, outward into the valve cover space. (There are a few caveats regarding "air" and "fuel-air mix", omitted for brevity.) BMJ-pdx (talk) 20:22, 30 June 2022 (UTC)Reply

.

'Thermal Expansion Coefficients for Various Materials' has some problems in the list:

- Either the linear coefficient, or volumetric coefficient or both are incorrect for Aluminum Nitride and Iron.

- Handling of significant digits looks bizarre. Different references might provide difference accuracies/precision, but if that is the case, the significant digits changes should at a minimum correspond with the footnote links. 70.171.44.124 (talk) 01:40, 24 September 2013 (UTC)BGriffinReply

As for the rubber tires, I have no idea what that has to do with expansion, except perhaps that it's able to stretch with the rims and steel belts. (I only assume we're talking car tires.) For the engine performance, I can say it is very much true. All parts are machined to fit when the engine is at its normal operating-temperature. When cold, there are large gaps in the piston rings, between bearings and their journals, valve seals, etc... It is true, that hydraulic lifters are also not operating properly until the engine oil begins flowing through them, and lubrication also needs to be at operating temperature for best performance However, even more detrimental is the increased blowby and localized friction from loose parts in point-contact with each other. (Which is why I always warm my car up in the winter before I dirve it.) Give me a little time, because this is easily referenced.

Another example is the SR-71 Blackbird. The plane leaks like a screen when it's sitting on the runway. Hydraulic oil, fuel, coolant, you name it. But when it gets to its normal operating temperature and pressure, everything seals up tight. Zaereth (talk) 21:19, 9 October 2014 (UTC)Reply

Any leakage due to larger clearances in a cold engine has an insignificant effect on performance, simply because the change in clearance is quite small. The best reason for warming an engine up is to ensure proper oil circulation. The most significant issue there is the splash (squirt/fling) lubrication of the cylinder walls, from a feed hole in the big end (a British term) of the connecting rods; a fast idle facilitates that. BMJ-pdx (talk) 20:34, 30 June 2022 (UTC)Reply

I might also add that thermal expansion often causes problems when different materials are used in conjunction with each other. For example, aluminum engine-blocks are wonderful, lightweight, and conduct heat like nobody's business. (In colder climates you can usually get by without a radiator fan.) But the aluminum heads expand more than the steel bolts holding them. Therefore, the bolts need to be "torqued to yield" (until they begin deforming), and then they slowly stretch over time. Eventually, the head gasket will blow out, and upon replacing it you'll need new bolts also. Zaereth (talk) 00:23, 10 October 2014 (UTC)Reply

Yes, aluminum blocks/heads conduct heat well, but to where? To their outside surfaces, but then the heat has to be dissipated, and dissipation there is minimal compared to what the cooling system does. BMJ-pdx (talk) 20:47, 30 June 2022 (UTC)Reply

I wouldn't torque head bolts (or nuts, if using studs) to yield unless called for by the manufacturer. Since an aluminum head will expand more than the head bolts/studs, clamping pressure will increase, not decrease.

A factor in head gasket failure on iron block/aluminum head engines (e.g., the Lotus Twin Cam of the 1960s) is shear damage of the gasket due to the unequal expansion. That can be mitigated with teflon or graphite on the mating surfaces (counterintuitive, perhaps, for a seal), which facilitates slippage rather than shear damage within the gasket. BMJ-pdx (talk) 20:58, 30 June 2022 (UTC)Reply

Added lede to heat transfer, here. Prokaryotes (talk) 18:52, 5 April 2014 (UTC)Reply

The citated value for the volumetric expansion coefficients of Titanium (linear=8.6; volumentric=11–14) is wrong. It should be about twice the mentioned value, what is a big difference. This value also is not found in the cited reference [24]. Please correct ——————————————————————————————————————————————————————————————————————————————— — Preceding unsigned comment added by 134.102.240.41 (talk) 12:34, 23 December 2014 (UTC)Reply

Thanks for pointing this out. I can't find an explicit citation for the volumetric coefficient, but Titanium has an hcp structure, so I'll just multiply the linear coefficient by 3. RockMagnetist(talk) 16:51, 21 January 2015 (UTC)Reply

The value for quartz is probably way off. The Crystan Handbook lists it as 7.1/13.2, depending on which direction in the crystal. 131.215.211.195 (talk) 23:38, 13 October 2015 (UTC)Reply

Somebody please add discussion of distinction between secant and instantaneous coefficient of thermal expansion. 108.171.131.170 (talk) 18:14, 13 July 2016 (UTC) Chuck LundReply

Hi Chuck. Thanks for the comment. I agree with you, but am not well-enough versed in math to do it myself. As I recall (and somebody please correct me if I'm wrong), a secant (average) coefficient is one that is tied to a fixed reference point, whereas a tangent (instantaneous) coefficient is one that is tied to the temperatures immediately surrounding the temperature of interest. Beyond that, I can't offer much insight myself, but perhaps someone will come along who can. Zaereth (talk) 00:23, 14 July 2016 (UTC)Reply

Linear expansion coefficient for aluminium nitride is bigger than volumetric one. It seems erroneous. --194.190.225.135

I agree. According to this source, both 4.2 and 5.3 are for linear (along two different axes). This source lists the volumetric expansion as 10.1. Zaereth (talk) 18:16, 9 January 2017 (UTC)Reply

The section "Examples and applications" does not contain a single worked example of a common material or situation, which an average person can understand. Eg. "A steel rail expands by X cm over 100 metres for each 10C rise in temperature, because {maths}=X" would be invaluable to most people reading this. Bards (talk) 22:30, 23 November 2017 (UTC)Reply

Regarding the 'disputed' value for the coefficient of thermal expansion for rubber, there should be no controversy; Wood and Martin published their measurement in 1964 {L. A. Wood and G. Martin, Journal of Research of the National Bureau of Standars-A. Physics and Chemistry Vol 68A, No. 3 (1964). and Lawrence A. Wood and Gordon M. Martin, Rubber Chemistry and Technology Sep 1964, Vol. 37, No. 4 (September 1964) pp. 850-865.} For peroxide-cured natural rubber at 25 C, their value is +6.36 x 10^-4 deg. C^-1. It is slightly higher (6.56) for uncured natural rubber. Historically, thermal expansion played hob with early rubber elasticity experiments. Sufficiently large temperature increases caused the unstrained sample length to increase, complicating the very low strain measurements of tensile stress vs. temperature. Experimenters were confused by this and coined the (unnecessary) term "thermoelastic inversion". — Preceding unsigned comment added by Davidhanson471 (talk • contribs) 18:40, 1 November 2018 (UTC)Reply

There is no controversy (wide-spread public debate), just misunderstanding. It's like the myth that "by all physical laws a bee should not be able to fly". This comes from a misunderstanding of Reynolds number, which shows that if you created a bee as large as an airplane it most certainly couldn't fly, but at smaller scales it can. (On the opposite extreme is the myth that because of its small size the bee is somehow "swimming rather than flying", which is also incorrect.)

According to Natural Rubber: Biology, Cultivation and Technology, by M.R. Sethuraj, Ninan T Mathew (page 422), natural rubber (not neoprene, nitrile, EPDM, silicone, etc...) has a coefficient of thermal expansion ranging between 4x10^-4 and 7x10^-4, depending on the type and amount of fillers used. The greater the amount of fillers; the lower the thermal expansion. The myth is from the notion that rubber has negative thermal expansion, meaning it contracts when heated. However, this contraction only occurs when it is already under strain (stretched) prior to heating --not when it is fully relaxed-- so, because this is a function of potential energy rather than volume, this phenomenon is referred to as the "joule effect". (What happens is that the rubber becomes stiffer with increasing temp.)

Aside from that, nearly all polymers expand a little differently than other materials, where vibrational energy causes the atoms to increase the distances between them. In polymers, there tend to be so many molecular bonds as to hold the atoms firmly in place, so the distance between them is virtually independent of temperature, thus the thermal expansion is directly related to the Van der Waals volume of the atoms instead. Feel free to change the article, but remember we prefer secondary sources to primary ones like scientific studies. Zaereth (talk) 23:51, 5 November 2018 (UTC)Reply

Article currently reads in part Materials which contract with increasing temperature are unusual; this effect is limited in size, and only occurs within limited temperature ranges (see examples below). [1] But there are no examples given in the section that I can see. Mind you, the section is long enough that I might well have missed them.

Either way, room for improvement. Andrewa (talk) 20:09, 22 November 2019 (UTC)Reply

That has to do with an urban myth about rubber, most likely that arose from a general misunderstanding of the real physics. Rubber does contract when heated, but only when it's already stretched prior to heating. This actually has nothing to do with thermal expansion, but is called the "joule effect". In a relaxed state it expands like any other polymer. For more info, se the talk page section directly above this one. Zaereth (talk) 20:18, 22 November 2019 (UTC)Reply

There is one material I can think of that does expand and contract in a negative fashion (opposite to nearly all other materials), and that is water. But that weird phenomenon only happens between 32 and 39 degrees F (0 and 4 degrees C). But you're right, it doesn't mention that anywhere in the article. I'll see if I can find a source. Zaereth (talk) 20:46, 22 November 2019 (UTC)Reply

I was able to locate a rather good source, a book called Negative Thermal Expansion Materials, that gives some good explanations. First, water does not really have negative expansion in the usual sense. Rather, the anomalous behavior of water is due to a phase change from liquid to solid, or visa-versa. When it crystalizes it gains volume but loses density. This why ice floats on water while anything else, like melting gold in a crucible, the solid material sinks. (When water is supercooled, or cooled such that it becomes a glass, it doesn't display this behavior.) There is an alloy called Invar which has such low thermal expansion it is essentially zero, and possibly even sometimes slightly negative. The first truly negative thermal-expansion material was zirconium tungstate, and there are several others that have been developed, but these materials often only experience that negative expansion within a certain (usually very cold) temperature range.

In general, there are many common materials that display negative expansion within a certain (usually very cold) temp range, but only along one axis. These are usually non-cubic structures, such as graphite, cadmium, or arsenic. For example, graphite has a layered structure, and at room temperature the expansion is negative in the direction perpendicular to the layers while the bulk expansion remains positive. At higher temperatures the negative expansion becomes positive in the perpendicular direction. Arsenic also displays this behavior when between -281 and -2 degrees C. Selenium also does, but in the parallel direction. I hope that helps. Zaereth (talk) 23:12, 22 November 2019 (UTC)Reply

Since writing the above I've come across https://www.sciencedaily.com/releases/2019/11/191101143955.htm which reads in part The secret behind crystals that shrink when heated. Scientists have new experimental evidence and a predictive theory that solves a long-standing materials science mystery: why certain crystalline materials shrink when heated. Their work could have widespread application for matching material properties to specific applications in medicine, electronics, and other fields, and may even provide fresh insight into unconventional superconductors. The substance under research there is Scandium fluoride. Andrewa (talk) 14:56, 24 November 2019 (UTC)Reply

An editor has asked for a discussion to address the redirect Per kelvin. Please participate in the redirect discussion if you wish to do so. Utopes _{(talk / cont)} 23:45, 28 March 2020 (UTC)Reply

Which metal rod expands when heated? 41.115.12.42 (talk) 14:22, 30 January 2022 (UTC)Reply

Can any one help it? Sanjunepali2007 (talk) 12:01, 12 September 2023 (UTC)Reply

It's unclear what you are referring to. In general liquids expand uniformly in all directions, unless they are constrained by the geometry of their container to expand in one or two dimensions. Qflib, aka KeeYou Flib (talk) 17:28, 13 September 2023 (UTC)Reply

Describe it if you know? Sanjunepali2007 (talk) 12:03, 12 September 2023 (UTC)Reply

The page already has a description of this apparatus. Qflib, aka KeeYou Flib (talk) 17:28, 13 September 2023 (UTC)Reply

Thank's for help 2405:ACC0:1100:44A0:B5BB:D634:DCD4:3B14 (talk) 03:01, 14 September 2023 (UTC)Reply