Color confinement in high-energy quark knockout

[nightvidcole]

TL;DR

If a quark is knocked out of a hadron at ultrahigh energies, how does the glue field respond?

At low energies, color is confined because attempting to remove a quark from a hadron will cause a response in the glue field that is often described as "snapping", or more formally, quark-antiquark pair production. However, how does this work at ultrahigh energies, let's say around 10^21 or 10^22 eV - still well below GUT energies? If a 1-10 ZeV electroweak-interacting particle is incident on a hadron and knocks out a quark at high momentum transfer, relativity dictates that the glue field can only respond within a very small distance of the quark's trajectory, due strictly to causality and special relativity. Any response further away from the quark will never be able to "catch up" and pull energy away from the quark since that quark will have departed to a very large distance by the time a light-speed signal can reach it. Given that the QCD coupling constant is suppressed at short length scales, what allows color confinement to operate in this ultrahigh-energy regime? Has anyone run numerical simulations to see if you still get full hadronization even at these ultrahigh energy quark knockouts?

 

[Vanadium 50]

Let's start with a simpler question. If I do this to a magnet and the north and south sides go flying away, what keeps the poles' field lines attached?

 

[nightvidcole]

Those fields come from a current density - but the analogous thing isn't true for QCD. The dual of the magnetic charge is electric charge - which can be isolated. But whatever the dual of color charge is - it does not at all behave like electric monopoles that can exist independently.

 

[Vanadium 50]

But color charges do behave like magnetic poles, in that they are confined - just like magnetic field lines are closed.

If you try and concoct a situation where the collision is "too fast for the color lines to reconnect", it also is too fast for the magnetic lines to reconnect. Since the latter doesn't happen, neither does the former.