What does it mean that there is no mathematical proof for confinement?

Question

I see this all the time* that there still doesn't exist a mathematical proof for confinement. What does this really mean and how would a sketch of a proof look like?

What I mean by that second question is: what are the steps one needs to prove in order to "mathematically prove confinement".

---

*See e.g. Scherer's "Introduction to Chiral Perturbation Theory" middle of page 7

This post imported from StackExchange Physics at 2014-08-29 16:47 (UCT), posted by SE-user Love Learning

Answer 1

The problem

In case you were not aware of this, finding a proof for confinement is one of the Millenium Problems by the Clay Mathematics Institute. You can find the (detailed) answer to your question in the official problem description by Arthur Jaffe and Edward Witten.

In short: proving confinement is essentially equivalent to showing that a quantum Yang-Mills theory exists and is equipped with a "mass gap". The latter manifests itself in the fact that the lowest state in the spectrum of the theory cannot have an arbitrarily low energy, but can be found at some energy $\Delta>0$. Proving this means to formulate the theory in the framework of axiomatic quantum field theory and deduce systematically all of its properties.

Mass gap implies confinement

In order to understand why proving that the theory has a mass gap is equal to proving confinement, we first have to understand what confinement is. In technical language it means that all observable states of finite energy are singlets under transformations of the global colour $\text{SU}(3)$. In simple terms this means that all observable particles are colour-neutral. Since quarks and gluons themselves carry colour charge, this implies that they cannot propagate freely, but occur only in bound states, namely hadrons.

Proving that the states in the theory cannot have arbitrarily low energies, i.e. there is a mass gap, means that there are no free particles. This in turn means that there cannot be free massless gluons which would have no lower bound on their energy. Hence, a mass gap implies confinement.

Motivation

The existence of confinement, while phenomenologically well-established, is not fully understood on a purely theoretical level. Confinement is a low energy phenomenon and is as such not accessible by perturbative QCD. There exist various low energy effective theories such as chiral perturbation theory which, while giving good phenomenological descriptions of hadron physics, do not teach us much about the underlying mechanism. Lattice QCD, albeit good for certain qualitative and quantitative predictions, also does not allows us to prove something on a fundamental level. Furthermore, there is the AdS/CFT correspondence, which allows us to describe theories which are similar to QCD in many respects, but a description of QCD itself is not accessible at this point. To conclude: there are many open questions to answer before we have a full understanding of QCD.

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Frederic Brünner

Comments

Could you try to explain what confinement really is (at a technical level, if you wish) and why it is essentially equivalent to solving the Millenium problem?

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Danu

---

@Danu: Thank you for the suggestion, I will edit my answer.

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Frederic Brünner

---

Very nice answer, thanks. I have a question about your sentence: "Since quarks and gluons themselves carry colour charge, this means that they cannot propagate freely, but occur only in bound states, so-called hadrons." Isn't this only true for energies below a certain level (below a certain critical temperature)? I thought it was generally believed that at very high energies quarks can occur as free particles, or have I misunderstood this?

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Hunter

---

Thank you! Yes, this is only true below the confinement scale. All the statements in my answer regarding confinement and mass gap are to be thought of in the context of low energy QCD.

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Frederic Brünner

---

If we assume that finite energy states are global-SU(3) singlets then it would imply that there is a mass gap. However, I don't understand how the converse is true. There may be a mass gap but, in principle, there may still exist finite energy states which are not global-SU(3) singlets. No?

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user user10001

---

No. As I mentioned in my answer, free gluons could have arbitrarily low energy, and as such contradict that the statement that there is a mass gap. Contrary to what you say, confinement does not imply a mass gap, as one can in principle construct confined states of zero energy (e.g. pions in the chiral limit).

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Frederic Brünner

---

Why a non-singlet bound state imply free gluons?

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user user10001

---

Non-singlet is equivalent to non-confining which in turn implies having free gluons. What do you mean by non-singlet bound states?

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Frederic Brünner

---

I mean a bound state which is not in singlet representation of global SU(3). Can't such states exist ?

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user user10001

---

Not in a confining theory.

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user Frederic Brünner

---

Yes, but I don't understand how the existence of mass gap can imply non-existence of such states.

This post imported from StackExchange Physics at 2014-08-29 16:48 (UCT), posted by SE-user user10001