Talk:Extremal principles in non-equilibrium thermodynamics

Starting a stub here

No doubt everyone has seen debate over Swenson... Nerdseeksblonde (talk) 20:39, 8 October 2009 (UTC)[reply]

this already looks too general if any contributes much, but should be ok for now.

you've already got a bunch of authors and a rather general topic. Nerdseeksblonde (talk) 23:01, 8 October 2009 (UTC)[reply]

Proposing some changes

I am suggesting various changes as is obvious. There is reason to recognise the importance of the dissipation function, not only because of its historical importance, but also as indicated by Ziegler 1983 in his conclusion: "... the dissipation function is the key to irreversible thermodynamics". I think mention of Swenson is unnecessary and even misleading, and I think it would be better not to give too much emphasis to Dewar, whose work is only formal and theoretical as distinct from physical.Chjoaygame (talk) 23:03, 8 October 2009 (UTC)[reply]

The dissipation function is one of two terms in the Onsager 1931 formulation, the other term being the entropy production. Under certain speficic conditions, one is just half the other. It will be hard to give a serious account without mentioning both. For example, some of Paltridge's empirical work is in terms of entropy production and some of it in terms of the dissipation function. As noted above, Ziegler says the dissipation function is the key. They are both more or less equally important for the present purpose, so far as I can work out.Chjoaygame (talk) 23:47, 8 October 2009 (UTC)[reply]

OK, it sounds like you have this thought out let's see how it works out. I guess by dissipation they mean thermalization so you don't have non-equillibrium distributions to worry about ( my example I suggested earlier, take a 2 or 3 level system and create somekind of classic population inversion best described by a negative T and see how its equillibration or thermalization relies on micro issue. I even saw metastable states mentioned in Dewar IIRC, the mico matters but I guess if you have regions where T makes sense, you have an easier time of it. I seem to recall that fluctuation-dissipation theorems were taught in stat mech when I was a kid but that still seems to be a little peripheral here. Nerdseeksblonde (talk) 12:54, 9 October 2009 (UTC)[reply]

The entropy production is a sum of terms like Ji Xi. The dissipation functions of Rayleigh and Onsager are sums of terms like Rik Ji Jk /2. Gyarmati pointed out the usefulness of the alternative dissipation function as a sum of terms like Lik Xi Xk /2. I think it would be hard to present the subject without a fair amount of talk about the dissipation functions. As a further problem, as Martyushev and Seleznev 2006 note, diverse approaches were followed by scattered workers. As far as I can work out, the subject is about fluctuations that grow and fluctuations that fade, in a word, stability. Stable structures are more reproducible than unstable ones. I think when people speak of 'order' they are often referring to reproducible structure, and in this case reproducible dynamical structure.Chjoaygame (talk) 20:50, 9 October 2009 (UTC)[reply]

If I get a chance I'll give it some thought but see how it goes. Order can be subjective or based on conditional probability ( given there is a thing here, what is probability of one at x, 2x, 3x etc). Certainly there can be a stability issue, but I think the issue here is that the system is non-deterministic, it fluctuates, but it is more likely to go along a path with increasing entropy than decreasing. I don't think this is like the stability of an orbit or damped system, and the question of where it ends up with small pertubring excitations. A thermo system will fluctuate, presumably around the highest entropy value and can make arbitrary deviations with decreasing probability but it should spend most of the time near the equlibtirum configuration. Apparently though you are considering linear cases and local temperatures can be defined- at a given location you don't have non-thermal distributions or different parts at different temperatures. Presumably these are important limitations but not equivalent AFAIK but I could be wrong, I'll have to dig through this some more when I can. Nerdseeksblonde (talk) 21:00, 9 October 2009 (UTC)[reply]

The linear response ideas seem to be covered here, http://en.wikipedia.org/wiki/Fluctuation_dissipation_theorem , but this is not the same as the Fluctuation Theorem . After looking some more here, http://www.rsbs.anu.edu.au/ResearchGroups/EBG/profiles/Roderick_Dewar/Martyushev%20and%20Seleznev%202006%20Phys%20Rep.pdf it does seem as you point out that all of these have relied on local equillibrium that presumably is obtained faster than remote so definitions aren't a problem in the cases considered and the authors point out that far from being a universal law these are limited generalities. They mention things like chemical reactions but it would also eliminate population inversions etc. I thought you said something about highly dissipative, but presumably these fluxes are small enough to be have linear relationships with some forces. Nerdseeksblonde (talk) 00:01, 10 October 2009 (UTC)[reply]

I am not so sure that it eliminates things like population inversions, because one can I think describe them by suitably extended macroscopic variables. I stand to be corrected about this. One finds discussions of negative temperatures which I think refer to population inversions, and one recalls Lamb's paper "Classical laser", Borenstein, M., Lamb, W.E. (1972), Physical Review A 5(3):1298-1311.Chjoaygame (talk) 09:09, 10 October 2009 (UTC)[reply]

I am sorry if somewhere that I cannot recall I wrote of "highly dissipative processes". My usual reaction is that if someone writes of something being "highly" so, most often I feel he is trying to hoodwink the reader. Are you referring to Prigogine's writing of "dissipative processes [that] are sufficiently dominant to exclude large deviations from statistical equilibrium"? It is not the absolute density of dissipation, but the density of it relative to other component flux densities, that I think Prigogine is talking about. I think the term "statistical equilibrium" may contain some subtleties.Chjoaygame (talk) 09:30, 10 October 2009 (UTC)[reply]

Well, as I understand the review above, it seems to suggest that different pieces of the system are in equillibrium so you can define local temperatures. A two level inverted system is just something to consider, but take a 3 level system where you may not have a distribution is is near equillibrium, this would seem to be excluded from all these principles based on review and it does simplify the analysis. I'm not sure where I got the "highly" term but that would suggest "large" fluxes of stuff quicky removing temperature differences and order, which is not where you expect linear or perturbation methods to apply. I guess this could mean highly damped to exclude some kind of oscillatory behabviour or it could be quick to thermalize locally, I dunno I'd have to review all this stuff. Nerdseeksblonde (talk) 12:02, 10 October 2009 (UTC)[reply]

Yes, one needs local temperature to define entropy by these methods, but I do not know much about these questions.Chjoaygame (talk) 21:34, 10 October 2009 (UTC)[reply]

consider drift-diffusion equation, the concentration gradient is not a real force

but at non-zero "T" it drives a current flow just as much as an electric field. I happen to note your section that says something about thermodynamic forces not being real, FWIW. Nerdseeksblonde (talk) 19:11, 11 October 2009 (UTC)[reply]

Yes it "drives" a current but is not an electromagnetic force. It is an apparent or what I call "metaphorical" force. I carefully avoided the use of the word "real". I think that word may obscure rather than clarify at this point. This is a hard one to make clear, as you point out. A like thing is that one does not find "heat" as such in the equation of state. I will think some more about this. It is an important but subtle matter.Chjoaygame (talk) 22:06, 11 October 2009 (UTC)[reply]

regrettable abuse? you may want to source adjectives like this.

I did note what appear to be editorial opinions but wasn't sure if they can be sourced. Certainly this area is controversial and don't want to argue about words right now but are these placeholder for expansion or sourceable characterizations? Nerdseeksblonde (talk) 23:22, 13 October 2009 (UTC)[reply]

Yes. Grandy doesn't actually use the adjective regrettable, but he does use the adverb unfortunately. But he does use the term mischaracterization, and he does say "The reader may wonder why we have gone into so much detail in debunking the order/disorder mischaracterization of entropy, particularly since it seems to have such a natural appeal. But that is just the point: It provides an easily comprehensible meaning to the nonscientist that is, and has been, widely abused. ... This is only a small sample of the misinformation generated by equating entropy with degrees of disorder. Unfortunately, this characterization continues in too many classrooms and texts, just where we would hope to correct it." Now, following your advice, I will change 'regrettable' to 'unfortunate'. Sorry I forgot to sign the just previous edit. Will correct.Chjoaygame (talk) 01:40, 14 October 2009 (UTC)[reply]

Oh, no I didn't need to sign the edit.Chjoaygame (talk) 01:51, 14 October 2009 (UTC)[reply]

Well, whatever the terms I just wasn't sure if they were placeholders for someone to expand or intended to convey something from another source. Nerdseeksblonde (talk) 01:55, 14 October 2009 (UTC)[reply]

Good point.Chjoaygame (talk) 04:09, 14 October 2009 (UTC)[reply]

time to delete the article section headed 'Other reviews'

It is time to delete the article section headed 'Other reviews'. The article by Juretic et al. contains some literature review material but does not really deserve the title of 'review article'; it is about photosynthetic models, and is not primarily a review article. The comment about Dewar's work is highly specialized and refers to essentially incomplete work and does not deserve space in a general Wikipedia article. The reference to Bejan and Lorente is inappropriate and has no place in an article about thermodynamics because their viewpoint is explicitly and essentially teleological and contrary to the principles of thermodynamics. The whole section is just a repository for more or less inappropriate odds and ends, and should be deleted entirely.Chjoaygame (talk) 15:35, 28 December 2010 (UTC)[reply]