Talk:Statistical mechanics

From Wikipedia, the free encyclopedia

	Physics Portal This article is within the scope of WikiProject Physics, which collaborates on articles related to physics.
B	This article has been rated as B-Class on the assessment scale.
Top	This article is on a subject of top importance within physics.
Help with this template Comments: edit – history – watch – refresh The discussion of 'fundamental postulate' is incorrect. It refers only to the formulation of Boltzmann, and not to the very different formulation of Josiah Willard Gibbs. In understanding the difference between the Gibbs and Boltzmann approaches, you actually have to read Gibbs' book Elementary Principles in Statistical Mechanics, and not later summaries, most of which appear to pass through Paul and Tatiana Ehrenfest's 1912 article "The Conceptual Foundations of the Statistical Approach in Mechanics", now available from Dover Press as a mostly-English translation. I am reasonably sure that I am not the first to have noted that the Ehrenfest presentation of Gibbs' book does not do as well as might have been desired by Gibbs, but that needs to be researched. A few issues are treated in my textbook "Elementary Lectures in Statistical Mechanics, Springer-Verlag). In particular, in the actual book by Gibbs, which is reasonably a reliable source on what Gibbs wrote: 1) Gibbs used a different fundamental postulate, _not_ the principle of equal a priori probabilities, because Gibbs viewed the canonical ensemble as primary and the microcanonical ensemble as secondary. In modern notation, Gibbs viewed Wj = exp(- beta Ej) as fundamental. 2) The notion that statistical mechanics is only applicable to large systems is not found in Gibbs' book. Indeed, he deliberately compares his treatment with a treatment that he does not identify as Boltzmann's, showing the differences in the predictions of teh two models. 3) Gibbs certainly does not speak of the H-Theorem. After all, in Gibbsian statistical mecahnics the entropy is a constant of the motion.

This article is within the scope of WikiProject Mathematics.
Mathematics grading:	Start Class	High Importance	Field: Applied mathematics
It's only "high" in maths because its a fairly small but much used branch, esp. in Physics. It's only "start" in maths because a non-mathmatician would be left pretty clueless. Tompw 11:23, 5 October 2006 (UTC)
Comments: The discussion of 'fundamental postulate' is incorrect. It refers only to the formulation of Boltzmann, and not to the very different formulation of Josiah Willard Gibbs. In understanding the difference between the Gibbs and Boltzmann approaches, you actually have to read Gibbs' book Elementary Principles in Statistical Mechanics, and not later summaries, most of which appear to pass through Paul and Tatiana Ehrenfest's 1912 article "The Conceptual Foundations of the Statistical Approach in Mechanics", now available from Dover Press as a mostly-English translation. I am reasonably sure that I am not the first to have noted that the Ehrenfest presentation of Gibbs' book does not do as well as might have been desired by Gibbs, but that needs to be researched. A few issues are treated in my textbook "Elementary Lectures in Statistical Mechanics, Springer-Verlag). In particular, in the actual book by Gibbs, which is reasonably a reliable source on what Gibbs wrote: 1) Gibbs used a different fundamental postulate, _not_ the principle of equal a priori probabilities, because Gibbs viewed the canonical ensemble as primary and the microcanonical ensemble as secondary. In modern notation, Gibbs viewed Wj = exp(- beta Ej) as fundamental. 2) The notion that statistical mechanics is only applicable to large systems is not found in Gibbs' book. Indeed, he deliberately compares his treatment with a treatment that he does not identify as Boltzmann's, showing the differences in the predictions of teh two models. 3) Gibbs certainly does not speak of the H-Theorem. After all, in Gibbsian statistical mecahnics the entropy is a constant of the motion.

Contents 1 Out-of-place quotation ? 2 Link 3 Structure of statistical mechanics articles 4 Factorising Z: is it correct? 5 Article could do with *extensive* slimming 6 Fundamental postulate 7 Merge Statistical thermodynamics 7.1 Statistical mechanics 7.2 Statistical thermodynamics 7.3 Statistical Physics 8 Helmholtz free energy - F or A? 9 FYI 10 "List of notable textbooks in statistical mechanics" proposed for deletion 11 What is SiSj

[edit] Out-of-place quotation ?

"Ludwig Boltzmann, who spent much of his life studying statistical mechanics, died in 1906, by his own hand. Paul Ehrenfest, carrying on the work, died similarly in 1933. Now it is our turn to study statistical mechanics". -- David L. Goodman "States of Matter"

It is a little weird to start an article with a quotation, like that. — Miguel 07:33, 2004 May 28 (UTC)

I know, but it is such a great quote from a standard reference, I thought it was worth including. Michael L. Kaufman 22:29, May 28, 2004 (UTC)

It is pretty interesting, maybe inclusion at the end of the article. Edsanville 05:40, 19 Aug 2004 (UTC)

[edit] Link

Perhaps the page should link to Maxwell-Boltzmann distribution (AC)

[edit] Structure of statistical mechanics articles

I'd like to make some more contributions to the Stat. Mech. subject, so I was trying to figure out the present structure. I've made up the following "summary" and I suppose it ought to go on the Stat. Mech. page, but I don't know where. Any suggestions or modifications? Paul Reiser 21:32, 11 Dec 2004 (UTC)

		Maxwell Boltzmann		Bose-Einstein		Fermi-Dirac
Particle				Boson		Fermion
Statistics		Maxwell-Boltzmann statistics Derivation of the partition function Gibbs paradox		Bose-Einstein statistics		Fermi-Dirac statistics
Gas		Ideal gas		Bose gas Bose-Einstein condensate Planck's law of black body radiation		Fermi gas Fermion condensate

General:

• Partition_function_(statistical_mechanics)
• Statistical_ensemble
  • Microcanonical_ensemble
  • Canonical_ensemble
  • Grand_canonical_ensemble
• Entropy
  • Principle of maximum entropy

[edit] Factorising Z: is it correct?

Is sum of the products equal to product of sums, or I am missing something here?(Igny 02:55, 8 September 2005 (UTC))

If only it were that easy :-) It is generally false, unless all the energies are completely independent (completely uncoupled).

Consider an example:

Zf = Zf (pr(f1) + pr(f2)),

Zg = Zg ((pr(g1) + pr(g2)).

(Zf)(Zg) = Zf (pr(f1) + pr(f2)) Zg ((pr(g1) + pr(g2))

= Zf Zg (pr(f1)pr(g1) + pr(f1)pr(g2) + pr(f2)pr(g1) + pr(f2)pr(g2))

Does this equal Zfg? -- Only if the probabilities for f and g are completely independent, so pr(f1)pr(g1) = pr(f1,g1) etc; ie pr(f1|g1) = pr(f1), etc.

Assuming you can neglect correlations between f and g may or may not be a reasonable first guess. Jheald 22:29, 6 November 2005 (UTC)

Line added to article, to reflect this Jheald 22:46, 6 November 2005 (UTC)

You're right, this is linked to the idea of independent degrees of freedom There are still quite some cases where it works, obvious examples are the monoatomic ideal gas and paramagnetic systems. Actually, any system that has N quadratic degrees of freedom can be reduced into a set of N or less independent and quadratic degrees of freedom. For example, this is what happens with the diatomic ideal gas, and even with some quite complex systems like phonons in a solid. The degrees of freedom (physics and chemistry) article will explain this when I'll have finished with it. Actually, in all the examples I give here, the partition function breaks down into a product involving only individual independent dofs. ThorinMuglindir 23:02, 6 November 2005 (UTC)

Assuming the interactions really are quadratic, of course, so can be diagonalised; and also that we really are talking about equilibrium/maximum entropy distributions, so we can ignore any preparation effects. Jheald 23:13, 6 November 2005 (UTC)

The factorization does not work at all in liquids. As a general statement, one may separate treatments of separable systems, in which the factorization works, and non-separable systems, in which the factorization is utterly wrong.Phillies 20:36, 21 November 2006 (UTC)

All prior comments about this deeply flawed sub-section are quite correct, though much could be saved by simply introducing a heading "Application to Independent Subsystems" or better yet "Application to Simple Ideal Gases". Even then, there are many issues of pedagogy and rigor that must be addressed: 1) Where is the Z_N = Z(single)^N? 2) Where is the division by N!? Both of these are far more fundamental and important than the approximate decomposition of a molecule into independent internal degrees of freedom. 3) The mixing of classical integration results with quantum mechanically rigorous summations needs to be indicated and maybe even motivated. The vibrational results are summations for harmonic oscillators and the rotational results are integrations for high-temperature rigid rotors. 4) Neither the Z_t nor the Z_c expressions are in any way correct for an ideal gas or any other system. Partition functions--even single-particle partition functions--are intrinsically dimensionless, as the original definition in the article clearly indicates. In actuality, the *correct* single-particle Z_t is actually the product of the article's Z_t and its Z_c. The article confuses the configurational INTEGRAL with a partition function. 5) What are all these theta's? There is no motivation for or definition of characteristic temperatures anywhere in the article. 6) If one is intent on introducing these various expressions for the Z's, then at least indicate the genuine triumphs of the whole undertaking: the Sacur-Tetrode equation for the third-law entropies of monatomic ideal gases and the direct calculation of third-law entropies for stiff diatomics like NO or CO, both without any consideration of the character of the substance in condensed phases. Here are clear examples where statistical mechanics unequivocally trumps classical thermodynamics!

All in all, this article has a strong chemical smell (pun intended). Though myself a chemist, I have taught statistical mechanics in chemistry, chemical engineering, materials science and physics at both the uncergraduate and graduate levels. Each course requires a very different presentation. (And remember J. W. Gibbs was the very first U.S. Ph.D. engineer!) No physicist would develop statistical mechanics in this fashion. I therefore suspect many of the earlier critiques were from physicists. A little more of their point of view would help a great deal.

One point that I often make in class is that quantum mechanics developed in a very real sense from statistical mechanics. Both Planck (black body radiation) and Einstein (solid heat capacities) had more faith in Boltzmann and Gibbs--however reluctantly in Planck's case--than they did in Newton. A good portion of the "old quantum mechanics" rested on preserving the viability of statistical mechanics, even at the cost of jettisoning ideas in classical mechanics. (Of course it didn't hurt that other successful early applications, like the photoelectric effect and the electronic structure of hydrogen, helped cement quantum ideas independently.) Sure these seminal statistical mechanical applications deserve mention here, particularly since they are also examples of results for independent subsystems.68.41.166.61 21:02, 14 January 2007 (UTC) For whatever reason, my previous post (68:41:166:61) did not show my name. Wmadden 21:06, 14 January 2007 (UTC)

[edit] Article could do with *extensive* slimming

Since this article was first begun, there is now a whole *category* of articles on statistical mechanics.

Can I suggest therefore, that as much of the detail as possible is handed off to other articles, out of this one; leaving this article just as a general introductory tour/overview?

A much shorter, more focussed, less extensive, less duplicative article would be better. Jheald 23:05, 6 November 2005 (UTC)

the probem is that a lot of articles link here. many probably reference some expression in an ensemble or another, or the concept of statistical ensemble. Doing this would suppose to go through the links and send them to the appropriate article. ThorinMuglindir 23:58, 6 November 2005 (UTC)

[edit] Fundamental postulate

I am not entirely certain of this, but I wanted to add a paragraph to the "fundamental postulate" section which mentions the subtle difference between W and Ω :

In Boltzmann's original formulation for entropy (S = k log W), the quantity W is defined to be the number of microstates which are consistent with the macrostate. In the nomenclature of the day, W was refered to as the "number of complexions". This postulate is necessary as it provides the conditions under which the equivalence W = Ω holds.

What do you Wikipedians think? Good to add, or should be completely neglected? --HappyCamper 22:29, 24 December 2005 (UTC)

Good to add, but only as an historic footnote. 81.83.108.233 19:06, 6 January 2006 (UTC)

DEFINETELY NOT. Statistical thermodynamics is a branch of thermodyanamics that tries to intepretate some theermodynamic phenomenas with statistical manner. Statistical mechanics tries out to intepretate some phenomenas in PHYSICS CONDENSSED MATTHER(NOT THERMODYNAMICS)with statistical ways.

This is not the fundamental postulate of Gibbs' statistical mechanics. I refer you to his book. In Gibbs' statistical mechanics, the canonical ensemble is primary, and all states of the system are not equally likely; instead, the log of the statistical weight is proportional to E/kT, where E is the energy of the state. Furthermore, Gibbs uses a different formula for the entropy, and is emphatic that you cannot refer to the number of microstates of a classical system, because the states are labelled by real variables, so the number of states is continuously (uncountably) infinite.Phillies 20:43, 21 November 2006 (UTC)

[edit] Merge Statistical thermodynamics

• Yes - merge it into this article. PAR 03:32, 19 July 2006 (UTC)
• Concur - 04:13, 19 July 2006 (UTC)
• Disagree - If Britannica can have separate articles for each of these terms, I don’t see why Wikipedia can’t as well. There is much overlap, yet each term has enough peculiarity such to allow full books and textbooks to exist in their own standing. If you compare the table of contents in Gibbs’ Statistical Mechanics with the table of contents in Schrödinger’s Statistical Thermodynamics, you see that they have related but separate agendas. --Sadi Carnot 21:03, 19 July 2006 (UTC)

'Comment' : Sadi, I did as you suggested, but to me it looks like the two books are covering essentially the same subject (with differences in emphasis). I've never been able to discern any significant difference in the usage or meaning of the two terms, although "statistical mechanics" is far more common. I think (but am not sure) that the difference in usage is simple one of preference, not of substance. The OED has an entry for "statistical mechanics" (Attributes to Gibbs) but none for "statistical thermodynamics". I can't read Brittanica online. Could you summarize what they think the difference is? Nonsuch 21:34, 19 July 2006 (UTC)

More commentary -- Hill wrote one book called '"Statistical mechanics" and another called "Introduction to Statistical thermodynamics". But in the introduction to the latter he says "Such an explanation falls rather within the province of statistical mechanics or statistical thermodynamics, terms which we regard in this book as synonymous". Nonsuch 21:45, 19 July 2006 (UTC)

• Comment - At the moment, the statistical thermodynamics article is much more gentle and introductory, for a more general reader; while the statistical mechanics article goes straight for the maths, and piles it on. I suspect that distinction is also foreshadowed in the titles: statistical thermodynamics (I think) is more what a generalist might call the subject, seeing it as part of themodynamics as a whole; whereas (I suspect) people who would call it statistical mechanics are be more likely to be specialists, emphasising it as a subject in its own right. It seems to me the articles at the moment perform rather different functions, and it might be useful to keep a very light article just motivating what the concept is, in parallel with a proper thorough heavy-duty treatment. Jheald 22:03, 19 July 2006 (UTC).

  I'm not very keen on merging the articles either, but from my perspective, I've always thought of statistical thermodynamics as being a subdiscipline of statistical mechanics. --HappyCamper 22:06, 19 July 2006 (UTC)

As a graduate physics student, I was taught thermodynamics by Dr. Ta-You Wu. I remember he asked a thermodynamics question in class and I volunteered an answer that was essentially a statistical mechanics answer. He about ripped my head off - he explained that the two approaches to the same problem are entirely distinct. Thermodynamics as a theory is practically complete. (Einstein said thermodynamics "is the only physical theory of universal content that … will never be overthrown.") Statistical mechanics attempts to derive and explain the basic assumptions of thermodynamics, in order to produce a theory which explains thermodynamics, and yields results which extend beyond thermodynamics. I am not familiar with the term "statistical thermodynamics", but it sounds like, as mentioned above, that it might be a subdiscipline of statistical mechanics dedicated only to explaining thermodynamics. PAR 22:57, 19 July 2006 (UTC)

• Comment: I’m quite sure it would take a few weeks worth of reading to dig out exactly what the technical differences are for each term. For the moment, I would suggest putting see also headers at the top of each article, so that as time moves on editors and readers can add what they know to each, based on their unique knowledges. Also, to note, a Google search shows about 24 million hits for SM and 9 million hits for ST. Lastly, if this helps, here are introductions to the ‘02 entries from Britannica for each term:

Did you remember to quote the google searches? I make it 10 million SM, 350,000 ST, a 30:1 ratio. Nonsuch 00:20, 20 July 2006 (UTC)

[edit] Statistical mechanics

branch of physics that combines the principles and procedures of statistics with the laws of both classical and quantum mechanics. It aims to predict and explain the measurable properties of macroscopic systems on the basis of the properties and behaviour of the microscopic constituents of those systems. Statistical mechanics, for example, interprets thermal energy as the energy of atomic particles in disordered states and temperature as a quantitative measure of how energy is shared among such particles. Statistical mechanics draws heavily on the laws of probability, so that it does not concentrate on the behaviour of every individual particle in a macroscopic substance but on the average behaviour of a large number of particles of the same kind.

The mathematical structure of statistical mechanics was established by the U.S. physicist J. Willard Gibbs in his book Elementary Principles in Statistical Mechanics (1902), but two earlier physicists, James Clerk Maxwell of Great Britain and Ludwig E. Boltzmann of Austria, are generally credited with having developed the fundamental principles of the field with their work on thermodynamics. Over the years the methods of statistical mechanics have been applied to such phenomena as Brownian motion (i.e., the random movement of minute particles suspended in a liquid or gas) and electric conduction in solids. They also have been used in relating computer simulations of molecular dynamics to the properties of a wide range of fluids and solids.

[edit] Statistical thermodynamics

Thermodynamics is the study of the various properties of macroscopic systems that are in equilibrium and, particularly, the relations between these various properties. Having been developed in the 1800s before the atomic theory of matter was generally accepted, classical thermodynamics is not based on any atomic or molecular theory, and its results are independent of any atomic or molecular models. This character of classical thermodynamics is both a strength and a weakness: classical thermodynamic results will never need to be modified as scientific knowledge of atomic and molecular structure improves or changes, but classical thermodynamics gives no insight into the physical properties or behaviour of physical systems at the molecular level.

With the development of atomic and molecular theories in the late 1800s and early 1900s, thermodynamics was given a molecular interpretation. This field is called statistical thermodynamics, because it relates average values of molecular properties to macroscopic thermodynamic properties such as temperature and pressure. The goal of statistical thermodynamics is to understand and to interpret the measurable macroscopic properties of materials in terms of the properties of their constituent particles and the interactions between them. Statistical thermodynamics can thus be thought of as a bridge between the macroscopic and the microscopic properties of systems. It provides a molecular interpretation of thermodynamic quantities such as work, heat, and entropy.

Research in statistical thermodynamics varies from mathematically sophisticated discussions of general theories to semiempirical calculations involving simple, but nevertheless useful, molecular models. An example of the first type of research is the investigation of the question of whether statistical thermodynamics, as it is formulated today, is capable of predicting the existence of a first-order phase transition. General questions like this are by their nature mathematically involved and require rigorous methods. For many scientists, however, statistical thermodynamics merely serves as a tool with which to calculate the properties of physical systems of interest.

From these entries, I would gauge that Britannica views SM to be a sub-branch of physics and ST to be sub-branch of thermodynamics. As an example of a similar sort of debate on topic semantics you might want to see: Talk:Physical chemistry#Chemical physics vs. Physical chemistry. I hope this helps?--Sadi Carnot 00:14, 20 July 2006 (UTC)

There doesn’t seem to be a strong consensus to merge these articles. A bought another book yesterday on statistical thermodynamics (1965) by Leonard K. Nash so I will try to build up that article and simply put “otheruses” links at the top of each article to connect the two. I’m no expert on either of these topics, but there are dozens of books on each at the book store. As an example, it took me some time to figure out what the difference is between thermochemistry and chemical thermodynamics; for some time, “chemical thermodynamics” was a redirect to “thermochemistry” although they are technically not the same subject. I will proceed to make this change:--Sadi Carnot 15:14, 2 August 2006 (UTC)

[edit] Statistical Physics

In addition to Statistical Mechanics (conflict of interest confession: My textbook is Elementary Lectures in Statistical Mechanics) and Statistical Thermodynamics, there is also a large collection of texts that title themselves 'statistical physics' perhaps with modifiers. They appear to be a bit different in their coverages than stat mech or stat thermo texts. I can suggest several characteristic features of graduate-level statistical physics texts, not shared by all of them. Note that I am describing 'characteristic' features', not all found in all books, with this not meant to be a pejorative list (though you may draw conclusions as to why I wrote my own text.)

Limited treatment of interacting systems other than Ising models and renormalization group. Treatment of phase diagrams--or stat mech applied to nondilute systems--somewhat heavily focused on the approach to the critical point with less attention to liquids or solutions. Relatively strong treatment of au courant topics in research. More focus on Boltzmann than Gibbs, including the H theorem. Less tendency to see chemical equilibria, charging potentials, or topics studied by J.G. Kirkwood, who after all was the leading light of American statistical mechanics for something like the second quarter or a bit later of the 20th century. There is also a typical distinction between intended physics and chemistry audiences here: Physics audiences are more likely to hear about linear response theory; chemical audiences are more likely to hear about Mori-Zwanzig theory.

There is some tendency in both statistical physics and chemical thermodynamics texts to mix thermodynamic and statistico-mechanical arguments in reaching a conclusion. George Phillies 130.215.96.64 21:23, 9 January 2007 (UTC)

[edit] Helmholtz free energy - F or A?

Is there a discussion somewhere about which letter to use for Helmholtz free energy? --HappyCamper 19:51, 23 July 2006 (UTC)

See Helmholtz free energy. WP is following the current thinking from IUPAC to the extent of using the letter A; but not to the extent of deleting the word "free" from the phrase "free energy". Jheald 21:35, 23 July 2006 (UTC)

Hmm...because in this article, F is already being used. It does not seem to be used consistently on Wikipedia at least... --HappyCamper 23:41, 23 July 2006 (UTC)

Use of F in this article is an abberation. Generally the other articles have (now) been standardised to IUPAC notation. Persistence of the word "free" is partly a hang–over from when the original articles were created and partly due to tradition / conservatism / lack of awareness (I guess) among editors and contributors. In my view we _should_ be consistent, and retain "free" in discussions of "free energy" concepts, but not when referring to the variables "Gibbs energy" and "Helmholtz energy". — DIV (128.250.204.118 10:49, 29 December 2006 (UTC))

[edit] FYI

I am fairly sure that the standard format rule in Wikipedia is to not link headers in the main article. I would suggest that comeone correct this.--Sadi Carnot 17:12, 2 August 2006 (UTC)

[edit] "List of notable textbooks in statistical mechanics" proposed for deletion

• List of notable textbooks in statistical mechanics

People following this page might like to express their views in the AfD, one way or the other. Jheald 03:55, 8 October 2006 (UTC).

[edit] What is SiSj

The random walker section of the article references a <SiSj> in one of the equations, but what this is not explained. 69.143.208.183 22:32, 3 December 2006 (UTC)

Looks like it's stating that the S_i's are completely not correlated, but I'm not sure what S is. --HappyCamper 14:15, 4 December 2006 (UTC)

I believe S is the ith step.Fephisto 03:50, 13 December 2006 (UTC)