# Phase space volume and relativity

+ 8 like - 0 dislike
391 views

Much of statistical mechanics is derived from Liouville's theorem, which can be stated as "the phase space volume occupied by an ensemble of isolated systems is conserved over time." (I'm mostly interested in classical systems for the moment.)

It's clear that special relativity doesn't change this, since relativity just adds a different set of invariances to the Hamiltonian. So under special relativity, the phase space volume of an ensemble must remain constant over time, as long as we're consistent in using a given inertial reference frame. However, this doesn't tell us how the phase space volume would change if we boosted to a different reference frame. So I'd like to know the following:

1. Can anything useful be said about how the phase space volume occupied by an ensemble of systems changes under a Lorentz boost? I guess this would entail taking a different time-slice through the system, as explained below. I would suspect that phase space volume doesn't behave very nicely under Lorentz transforms, since phase space is defined in terms of 3-position and 3-momentum.

2. If phase space volume doesn't behave nicely under Lorentz boosts, is there a generalisation (e.g. in terms of 4-position and 4-momentum) that does? And if so, where can I go to read more about it?

3. Optional bonus questions: what effect do general relativity and quantum mechanics have on all of the above?

Below are some diagrams which should make it clearer what I'm asking for. (a) shows a space-time diagram of a system consisting of several classical particles. We take a simultaneous time-slice though the system (red line) and note the 3-position and 3-momentum of each particle. This gives us a point in a phase space (diagram (b)), which I've drawn as 2 dimensional for convenience. An ensemble of similar systems (meaning systems with the same dynamics but different initial conditions) can be thought of as occupying a region of this phase space. Liouville's theorem tells us that if we do the same thing at different values of $t$ (for the same ensemble of systems), the shape of this region may change but its volume will be the same.

(c) shows a Lorentz-transformed version of the same system as (a). We can take a simultaneous time-slice through the system in the new $(x',t')$ reference frame, but it will not be parallel to the one in the $(x, t)$ frame. The particles' 3-positions and 3-momenta will also be different. We can plot the system's position in the new phase space generated by doing the same procedure in the new reference frame (d). We can also plot the region of phase space occupied by the ensemble. Doing the same thing at different values of $t'$ will produce different regions with the same volume as each other. However, my question is about whether volume of the region plotted in (d) must equal the volume of the region in (b). Progress

In this document (J. Goodman, Topics in High-Energy Astrophysics, 2012, p.12-13; unfortunately Goodman gives no further reference) there is a proof that infinitesimal phase space volumes are Lorentz invariant. It looks legit, but the author assumes that every particle's $x$ position is within the same small interval $[x, x+dx]$, which means that he doesn't have to take account of the fact that you take a different time slice when you change the reference frame. Additionally, I'm using an integrated version of Liouville's theorem, in which the ensemble has a finite rather than infinitesimal phase space volume, and it isn't immediately clear to me whether this makes a difference. So this seems to suggest that the phase space volumes sketched in figures (b) and (d) above will be equal, but I'm still not sure and would like to know where I can find the full proof, if it exists.

Further Progress

I'm awarding a 100 point bounty to Qmechanic for his helpful answer, which expresses Goodman's argument (discussed above) in more formal language. However, I don't think the question has been answered yet. This is because the argument doesn't just assume that the system occupies an infinitesimal volume of phase space. On its own this would be fine, since you can just integrate over phase space to get the finite phase-space volume version. The problem is that this argument also assumes the system occupies only an infinitesimal volume of real space. In general, a system that occupies an infinitesimal volume of phase space can occupy a finite area of actual space (think of a system composed of multiple, spatially separated, classical particles, as illustrated above) and so this argument seems to cover only a restricted subset of the cases I'm interested in. Goodman's argument is fine if you're considering non-interacting particles or fluids in equilibrium (which makes sense given its origins in astrophysics) but I'm also interested in multi-particle systems that may or may not be in thermal equilibrium. It is my strong intuition that the argument can be extended to deal with all cases, and I'll happily award an additional 100 point bounty to anyone who can show how to do so.

I think one route to an answer might be to note that, in special relativity, classical particles cannot interact unless they actually collide. This probably means that in the times in between collisions they can be thought of as occupying their own individual phase spaces after all, and perhaps with the aid of the proof that's been presented and some careful accounting about what happens during collisions, it will all work out OK. But of course then it has to be extended to deal with (classical) fields, which could be tricky. One gets the feeling there ought to be a nice simple way to derive it from Lorentz invariance of the Lagrangian, without worrying about what type of system we're dealing with. (I'll award a 200 point bounty to a satisfactory answer that takes this latter approach!)

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
afaik Liouville's theorem can be formulated relativistically, but you may need additional constraints (things like a foliation into space-like leaves, an asymptotically flat space-time, a time-like Killing-field - haven't investigated what's necessary in this particular case); Minkowski space should satisfy these constraints, so I do believe an appropriate generalization of Liouville's theorem does hold; I'm interested in an answer as well, so if no one provides one, I might hunt down some references myself...

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Christoph
I don't think special relativity changes this much at all--classical mechanics is already invariant under a Galilean boost, after all. If you're talking about general relativity, there is a perfectly good Hamiltonian formulation of that, even when coupled to matter.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Jerry Schirmer
If you like this question you may also enjoy reading this post.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Qmechanic
@JerrySchirmer I've edited the question to make it clear that I know special relativity can't change or invalidate Liouville's theorem in itself. Hopefully this will make it clearer that what I'm asking ("how does the phase space volume change under a Lorentz boost?") is a different question.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
A very respectable answer to your Question is definitely contained in the link given by Qmechanic. I find it helpful to keep in mind that Lorentz $in$variance depends on both the dynamics and the state being invariant under Lorentz boosts. A classical thermal state over a Lorentz invariant dynamics is not Lorentz invariant, for example. Anything can be made manifestly Lorentz $co$variant with enough work.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Peter Morgan
@PeterMorgan I'm truly sorry, but there's no way I can understand the terminology in the nLab post, so reiterating it doesn't really help. I'm not a mathematical physicist (nor aiming to become one). Rather, I'm an non-physicist with an interest in statistical mechanics, and I'm interested in the answer to this question because of its implications in my field. I hope that some day you or someone else will find time to answer the question using language of a similar level to that used in the question.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
@PeterMorgan having said that, I can vaguely understand the notion of defining the phase space as the [topological] space of all trajectories. But (i) this doesn't seem to provide any obvious way to define the phase space volume of an ensemble (unless there's also some way to define a measure), and (ii) if you lose the notion of taking a time slice then you also lose the ability to talk about conservation of phase space volume conservation over time. So while covariant phase space is an invariant generalisation of the notion of phase space, it doesn't seem to be the one I'm looking for.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel

+ 4 like - 0 dislike

I) Let us here prove the invariance of the two $3$-forms

$$\tag{1} p^0 dq^1 \wedge dq^2 \wedge dq^3\qquad \text{and}\qquad \frac{dp_1 \wedge dp_2 \wedge dp_3}{p^0}$$

under (restricted) Poincare transformations. As a consequence, the volume-form $dq^1 \wedge dq^2 \wedge dq^3 \wedge dp_1 \wedge dp_2 \wedge dp_3$ is also an invariant.

Let $c=1$. Here we will assume:

1. The metric is $\eta_{\mu\nu}={\rm diag}(-1,+1,+1,+1)$.

2. The $q^{\mu}=(t, {\bf q})$ and $p^{\mu}=(p^0, {\bf p})$ transform under Poincare transformations as an affine and a linear $4$-vector, respectively.

3. All particles have the same rest-mass $m_0\geq 0$. In particular, $$\tag{2} p^0 ~=~ \sqrt{{\bf p}^2+m_0^2 }.$$

4. The momentum is kinetic ${\bf p}~=~p^0{\bf v}$.

Since the two $3$-forms (1) are clearly invariant under translation and rotations, it is enough to consider a Lorentz-boost along the $q^1$-axis. This follows because

$$\tag{3} p^{0}dq^1, \quad dq^2, \quad dq^3, \quad \frac{dp_1\wedge dp_2 \wedge dp_3}{p^0},$$

are all invariant under boost along the $q^1$-axis. Only the invariance of the first item $p^{0}dq^1$ on the list (2) is not completely obvious or well-known, so let us concentrate on that one. The derivation essentially follows Ref. 1. Consider an arbitrary fixed point $({\bf q}_{(0)},{\bf p}_{(0)})$ in phase space at $t=t_{(0)}=\overline{t}_{(0)}$. Because of translation symmetry, we may assume that the point ${\bf q}_{(0)}=\overline{\bf q}_{(0)}$ is a common origin for the two coordinate systems (one barred and one un-barred) of the Lorentz transformation at $t=t_{(0)}=\overline{t}_{(0)}$. Let us define

$$\tag{4} {\bf x}(\Delta t)~:=~{\bf q}(t)-{\bf q}_{(0)}, \qquad \Delta t~:= t-t_{(0)},$$

and $$\tag{4'} \overline{\bf x}(\overline{\Delta t})~:=~\overline{\bf q}(t)-\overline{\bf q}_{(0)}, \qquad \overline{\Delta t}~:= \overline{t}-\overline{t}_{(0)}.$$

We imagine that we observe an infinitesimally small space-time (and energy-momentum) region around the fixed point $(q^{\mu}_{(0)},p^{\nu}_{(0)})$. Since we are only interested in first-order variations in positions, it is enough to work to zero-order variations in momentas. In other words, we can imagine all particles travel with the same constant energy-momentum $p^{\mu}=p^{\mu}_{(0)}$ (and velocity ${\bf v}={\bf v}_{(0)}$). Then

$$\tag{5} {\bf x}(\Delta t)~=~{\bf v}\Delta t+ {\bf x}_0 ,\qquad {\bf v}~=~\frac{{\bf p}}{p^0}, \qquad {\bf x}_0~=~d{\bf q},$$

and $$\tag{5'} \overline{\bf x}(\overline{\Delta t})~=~\overline{\bf v}\overline{\Delta t}+ \overline{\bf x}_0,\qquad \overline{\bf v}~=~\frac{\overline{\bf p}}{\overline{p}^0}, \qquad \overline{\bf x}_0~=~d\overline{\bf q}.$$

The Lorentz transformation reads $$\overline{\Delta t}~=~\gamma(\Delta t-\beta x^1(\Delta t)), \qquad \overline{x}^1(\overline{\Delta t})~=~\gamma(x^1(\Delta t)-\beta \Delta t),$$ $$\tag{6} \qquad \overline{x}^2(\overline{\Delta t})~=~x^2(\Delta t), \qquad \overline{x}^3(\overline{\Delta t})~=~x^3(\Delta t),$$

and

$$\tag{7}p^0~=~\gamma(\overline{p}^0+\beta \overline{p}^1) , \qquad p^1~=~\gamma(\overline{p}^1+\beta \overline{p}^0), \qquad p^2~=~\overline{p}^2, \qquad p^3~=~\overline{p}^3 .$$

Eqs. (5) and (5') can only both hold if the following well-known relativistic formulas hold

$$\tag{8} v^1~=~\frac{\beta+\overline{v}^1}{1+ \beta\overline{v}^1}, \quad v^2~=~\overline{v}^2,\quad v^3~=~\overline{v}^3, \quad\text{(rel. velocity addition)}$$

and

$$\tag{9} x^1_0~=~\frac{\overline{x}^1_0}{\gamma(1+ \beta\overline{v}^1)}, \quad x^2_0~=~\overline{x}^2_0, \quad x^2_0~=~\overline{x}^2_0, \quad \text{(length contraction)}.$$

On the other hand the first eq. in (7) yields

$$\tag{10} \frac{p^0}{\overline{p}^0}~=~\gamma(1+ \beta\overline{v}^1).$$

Combining the above equations yields the invariance of the first item $p^{0}dq^1=\overline{p}^{0}d\overline{q}^1$ on the list (2).

1. Part I discusses the local Poincare invariance. An integrated version therefore also exists (with appropriate change of integration regions under Poincare transformations).

2. Part I concerns systems consisting of particles of a single kind only. The generalization to mixtures is e.g. partially discussed in Ref. 2.

3. Perhaps surprisingly, a similar proof as in part I shows that the symplectic $2$-form $$\tag{11} \omega ~=~\sum_{i=1}^3 dp_i \wedge dq^i$$
is not invariant under restricted Poincare transformations.

References:

1. J. Goodman, Topics in High-Energy Astrophysics, 2012, p.12-13.
2. S. R. de Groot, W. A. van Leeuwen and Ch. G. van Weert, Relativistic kinetic theory, 1980.
This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Qmechanic
answered Jan 2, 2013 by (2,860 points)
Many thanks for the answer. It will take me some time to digest it, because I'm not familiar with the wedge-product notation. (This is a good excuse to learn it I suppose.) But without fully understanding it, I'm suspicious of the bit where you say "We imagine that we observe an infinitesimally small space-time (and energy-momentum) region around the fixed point (...), where we can imagine all particles travel with the same constant velocity (and momentum)," as I'm explicitly interested in macroscopically-sized systems, where this assumption can't be made.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
The above answer discusses the local invariance. An integrated version therefore also exists (with appropriate change of integration regions under Poincare transformations).

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Qmechanic
I guess that's the bit that really isn't clear to me. If it was only local in the sense of applying to a local region of phase space then that would obviously be true, but this proof is local in space-time as well as in phase space, and I don't have any sense of how you can integrate over both those things simultaneously, or of what the result would look like in terms of the diagrams in my post. I'll think about it some more, but any insight you can give would help.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
Guh, orbiting stars are a bad example, as this is meant to be SR, so let's make it two point masses joined by a spring instead. In this case one could try to consider an infinitesimal space-time region surrounding just one of the masses, but the problem is that it doesn't really have a phase space by itself, because it's interacting with the spring. In cases like this, as far as I can see, you can't consider phase space without considering the positions and momenta of both masses simultaneously (as well as the unavoidable wave degrees of freedom in the spring).

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
(Note: my second comment was written before I saw your first one.)

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
@twistor59 thanks, it's useful, but it doesn't consider interacting particles. (Their "bundle of identical particles" could just as easily be interpreted as a statistical ensemble over a single particle.) Hence, like Goodman's proof, it only deals with a special case, and because of this it skips over the issue I'm stuck on.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
@Nathaniel - created a chat room to understand the question better without angering the StackExchange gods

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user twistor59
+ 0 like - 0 dislike

Here is a nice physical intuition that might finally lead me to being able to answer this question. I'm posting it as answer to my own question, but I won't accept it until I'm sure it works.

Let's consider a constantly uniformly accelerating reference frame. The planes of simultaneity in this reference frame are the dark grey lines passing through the origin in this diagram (from Wikipedia): Obviously this is not an inertial reference frame. If we can show that phase space volume is nevertheless conserved in this reference frame then we should be able to show that it is conserved under Lorenz boosts, simply by considering the non-inertial reference frame that uniformly accelerates from one inertial frame to another.

My physical intuition is that phase space volume should be conserved in a uniformly accelerating reference frame, because of Einstein's equivalence principle: transforming to a uniform reference frame is just the same as adding an extra force to every particle, proportional to its mass. In an inertial frame with this extra force added, phase space volume should still be conserved, because all we've done is change the Hamiltonian.

This seems nice, but I'm not sure of all the details yet. In particular, the Rindler space-time diagram above has a horizon at $(0,0)$, and I'm not sure what implications this might have in terms of phase space volume. I might end up asking a question or two about this later on.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
answered Feb 15, 2014 by (495 points)
Actually I think this doesn't work - it seems that the equivalence principle is only local, even in Minkowski space time - adding a constant acceleration to every object seems not to be the same as adding a constant acceleration to the observer.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel
+ 0 like - 0 dislike

Phase space is Lorentz invariant. You can prove this by writing an integral over d^4 x d^4 p and doing a Lorentz transformation, but there's a nice short proof in Padmanabhan's book "Gravitation", p.26 : For an observer moving with four-velocity ui, the proper three-volume element is given by d3V = u0d3x which is a scalar invariant. To prove this, note that the quantity d4V = dx dy dz dt is a scalar. Multiplying this by 1 = u0(dτ/dt) and noting that dτ is invariant, we conclude that d3V = u0d3x is an invariant.

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Judy
answered Feb 25, 2014 by (0 points)
I don't have easy access to that book, and can't easily understand your notation (do you know LaTeX? If so you can use it on this site. Just use dollar signs as usual - that will make it much easier to read), but are you sure this isn't just for a single point particle as in Qmechanic's answer?

This post imported from StackExchange Physics at 2014-04-08 05:11 (UCT), posted by SE-user Nathaniel