27

  1. Is every phase transition associated with a symmetry breaking? If yes, what is the symmetry that a gaseous phase have but the liquid phase does not?

  2. What is the extra symmetry that normal $\bf He$ has but superfluid $\bf He$ does not? Is the symmetry breaking, in this case, a gauge symmetry breaking?

Update Unlike gases, liquids have short-range order. Does it not mean that during the gas-to-liquid transition, the short-range order of liquids breaks the translation symmetry? At least locally?

SRS
  • 26,333
  • 1- the order parameter is related to the difference of the densities in the two phases, though I don't remember the details. 2- It is the U(1) global symmetry (not really a gauge symmetry, since there is no gauge field). – Adam Mar 26 '14 at 13:41
  • @Adam - I was guessing short-range order which the liquid has but the gas does not. But don't know whether it is write or wrong. – SRS Mar 26 '14 at 13:58
  • 3
    for the liquid-gas transition, short range correlation is not important (though it has of course important physical consequences). The order parameter is something like $\rho_A-\rho_B$ where $\rho_A$ is the density in the current phase, and $\rho_B$ the other. It plays the same role than the magnetisation in the Ising model (the symmetry is $Z_2$). – Adam Mar 26 '14 at 14:23
  • @Adam - I understand the idea of order parameter. But I'm sorry I didn't get the answer I was looking for. Actually I wanted to know precisely which symmetry is broken during gas to liquid transition (like rotational and translation symmetries are broken during liquid to solid transition). I hope I'm not being vague. Same question I asked for normal He to superfluid He transition, in the first part of the second question. – SRS Mar 26 '14 at 14:29
  • the thing is, liquid and gas are mostly the same, and that's why the broken symmetry is rather subtle (it is not something you measure in the lab easily). Same thing with helium. – Adam Mar 26 '14 at 15:05
  • There is no symmetry breaking in the liquid/gas phase transition. As @Adam said, it is characterized by a jump in the density. – Yvan Velenik Mar 26 '14 at 17:04
  • 1
    @YvanVelenik - Then the conclusion is: All phase transitions are not associated with symmetry breaking. – SRS Mar 26 '14 at 17:32
  • 1
    @Roopam : yes, of course they are not always related to symmetry breaking. – Yvan Velenik Mar 26 '14 at 20:27
  • @YvanVelenik: you can describe the liquid-gas phase transition by an Ising model (for both 1st and 2nd order transitions), so there is in fact an order parameter and a broken symmetry. But it's not a symmetry of the physical system itself (rotation, translation, etc.). – Adam Mar 27 '14 at 19:37
  • @Adam: It's just one rough model of the liquid/gas transition, and the symmetry is accidental. You can actually consider infinitely many perturbations of the Ising lattice gas, which we'll be just as good approximations to the real fluids, and in which there is no symmetry. There is no reason to expect that there is any, even hidden, symmetry for the real liquid/gas transition. – Yvan Velenik Mar 28 '14 at 07:49
  • @YvanVelenik: the critical point of the liquid-gas transition is described by the Ising universality class, characterized by the dimension of space, and the symmetry of the order parameter. Close to this transition, all these perturbations will be irrelevant and the symmetry will be emergent (but none the less real). – Adam Mar 28 '14 at 13:32
  • @Adam: your argument does not apply to the first order phase transition between the liquid and gas phases, which I am discussing. – Yvan Velenik Mar 28 '14 at 13:36
  • @YvanVelenik: Yes, it does not work for the 1st order phase transition (this corrected one of my previous comment), but that does not change the fact that there can be a broken symmetry for a liquid-gas transition (at least a particular point). – Adam Mar 28 '14 at 13:42
  • @Adam: But the question was "Is every phase transition associated with a symmetry breaking? If yes, what is the symmetry that a gaseous phase have but the liquid phase does not?", and the answer is negative. – Yvan Velenik Mar 28 '14 at 13:44
  • 1
    @YvanVelenik: I never said that all transition is associated with a symmetry breaking, which is definitely not the case. I'm just arguing that saying that there is no symmetry breaking in the liquid-gas transition is too crude an answer. – Adam Mar 28 '14 at 13:55
  • @Adam: But, stricto sensu, at the critical point, there is no transition from gas to liquid phases (the two are undistinguishable at that point). So, I keep with what I said: there is no symmetry breaking at the liquid/gas phase transition. – Yvan Velenik Mar 28 '14 at 14:02
  • @YvanVelenik: I'm not so sure. Close enough to the critical point, we expect this effective symmetry to describe the system, so there could be a more or less small range along the 1st order transition line with effectively a $Z_2$ symmetry, though how wide is this range will be model (or species) dependent. – Adam Mar 28 '14 at 14:30

4 Answers4

14

Let me answer your first question: Phase transitions do not necessarily imply a symmetry breaking. This is clear in the example your are mentioning : The liquid-gas transition is characterized by a first order phase transition but there is no symmetry breaking. Indeed, liquid and gas share the same symmetry (translation and rotation invariance) and may be continuously connected in the high temperature/pressure regime. In quantum systems at zero-temperature, one may also encounter transition in between quantum spin-liquid states for which there is also no symmetry breaking. Yet another example is the case of the 2D XY model where there is a continuous phase transition but there is no symmetry breaking (Kosterlitz-Thouless transition).

VanillaSpinIce
  • 1,408
  • 4
    KT and spin liquids are nice example, but the liquid-gas case is more subtle. It can be mapped onto an Ising model, with the associated order parameter and symmetry broken phase. – Adam Mar 27 '14 at 19:38
  • 1
    Note that it is easy to construct models with no symmetry in which 1st order phase transitions occur (actually this is the generic situation! You have to work harder to construct models with symmetry breaking). There is a mathematical theory devoted to this problem: the Pirogov-Sinai theory. – Yvan Velenik Mar 28 '14 at 07:53
  • @Adam The liquid-gas maps to the Z2 Ising model(0T) with all spin up represents liquid, say, and all spin down to gas. I see no symmetry breaking during this transition. The finite temperature case is similar with the spin up sites decrease while spin down increase. No symmetry breaking occurs. – an offer can't refuse Apr 04 '14 at 06:07
  • As far as I know symmetry breaking happens when we pass through some critical point. So liquid and gas need to share the same symmetry. Is the supercritical liquid that have symmetries broken by the gas-liquid regime. – Nogueira Nov 08 '15 at 20:30
  • So Landau's paradigm cannot explain liquid-gas phase transition? Or generally Landau's paradigm cannot explain 1st order phase transition? – 346699 Feb 22 '17 at 23:20
2

@VanillaSpinIce I agree most part of the answer from VanillaSpinIce, instead "The liquid-gas transition is characterized by a first order phase transition but there is no symmetry breaking."

Below the critical point,when a gas-liquid phase transition happens, an interface form between the gas and the liguid(since they have different density), thus a discrete refleciton symmetry (between gas and liquid) is broken.

Wilson Ko
  • 116
-1

I think even in the case of liquid-gas transition there is symmetry breaking. The order parameter in this system is the density, and the symmetry that the gas density respects is translational symmetry. On the other hand, the liquid does not have translational symmetry. There might be some confusion when thinking of a liquid as an incompressible fluid. Doesn't that imply that the density of the liquid phase is uniform too? Actually one needs to think about the whole system, which includes the container.

Lets look at it from the perspective of degenerate ground states. For a gas, there is only one ground state, the state where all the particles fill the container uniformly. For a liquid, there are many ground states (assuming the liquid has no surface energy). In a zero gravity environment, you can have the liquid fill the bottom half of the container, top half, or be broken up into many droplets. All have the same energy, thus a degeneracy of the ground state. This implies a form of symmetry breaking. This is very similar to the Ising ferromagnetic transition.

Pecan Lim
  • 97
-2

The classical situation with no symmetry breaking is the case of the, so-called, isostructural transitions. The word "isostructural" is misleading, since what is meant is "isosymetric". However, historically the term emerged. There is a number of examples of such transiotions. One is the alpha-alpha' transitions in the hydrogen-metal systems, another is phase separations in fluids and polymer solutions, the coil-globule transition in polymers. Such a transition in a solid phase has been reported for SmS. In the case of the solid phase the crystal lattice changes its volume, but preserves its structure (this gave rise to its name).

Alexei Boulbitch
  • 1
  • This doesn't appear to actually answer the questions asked, though it is tangentially related. – Kyle Oman Apr 15 '14 at 14:16
  • This does not provide an answer to the question. To critique or request clarification from an author, leave a comment below their post - you can always comment on your own posts, and once you have sufficient reputation you will be able to comment on any post. – DavePhD Apr 15 '14 at 15:36
  • @ Kyle This answers the first part of the question. You simply cannot recognize the answer. It indicates the class of transition without the symmetry (or more generally, the structural) change, and gives examples of materials, where such transuitions are observed. Read it once more. I do not answer the second part of the question, where I am not a specialist. – Alexei Boulbitch May 02 '14 at 07:34
  • @DavePhD this indeed give the answer to the question. Read my comment to Kyle and my answer. In contrast the answers above along with discussion contain serious mistakes. – Alexei Boulbitch May 02 '14 at 07:39