Earlier quoted context omitted.
As far as we know, molecules and atoms escaping into space comes due to the solar wind. Somehow the solar wind activates these atoms and carries them off into the distant vacuum, ad infinitum presumably though (I would imagine) limited by the heliopause. [edit, benefiting from convo: mechanisms on atmospheric escape, to varying degrees of verification) • https://en.wikipedia.org/wiki/Atmospheric_escape • https://en.w…
With high enough temperature some molecules may achieve velocities enough to escape. The question is - how hot was it really? The initial collision happened at least with escape velocity, so there was roughly enough energy for volatiles an non-volatiles to escape. But non-volatiles condensed presumably pretty quickly (but were still hot) in comparison to volatiles.
https://en.wikipedia.org/wiki/Atmospheric_escape
https://en.wikipedia.org/wiki/Hydrodynamic_escape
As a layperson I see our current epistemological state as long on models, and short on empirical verification (because we're talking about a difficult phenomenon to verify).
I think I mostly wanted to offer counterpoint to the original comment that 'this atmosphere can't stay long' i.e. even under elevated temperatures.
(I'll probably update my prev comment with those wikipedia links.)