It seems that
ALICE measurements shows that the ratio of O2/H2O is much higher than that inferred from the ROSINA measurements.
Putting another nail into the coffin of the Yao & Giapis mechanism.
It should be mentioned that Yao & Giapis have a reply to Heritier et al:
Reply to “On the origin of molecular oxygen in cometary comae”
Yao, Y. & Giapis, K. P.
https://www.nature.com/articles/s41467-018-04943-w#article
They seem to concede that the surface mechanism is a non-starter. However, in what looks like a pretty lame attempt to keep their mechanism alive, they postulate that water ions impinging on the spacecraft, in particular the solar panels, could cause O
2 to end up being detected by Rosina.
As I say, this looks pretty lame, and they probably should have just conceded without reply.
As Tusenfem notes above, ALICE also detected the oxygen. Y & G may claim that what they are seeing is influenced by a shed load of the stuff generated at the spacecraft (as in a reply to an email from me). That seems highly suspect to me. The ALICE team would have to answer for themselves, but the lead author of the paper linked by Tusenfem has another in preparation, for which there is only a conference abstract at the moment:
Stellar Occultation by Comet 67P/Churyumov-Gerasimenko Observed with the R-Alice Ultraviolet Spectrograph
Keeney, B. A. et al.
http://adsabs.harvard.edu/abs/2017DPS....4950908K
The significance of that is the sentence:
We find that both the H2O column density and the relative O2/H2O abundance decrease with increasing impact parameter, in accordance with expectations.
The impact parameter is the distance between the line of sight and the nucleus. This was increasing as the two observations were made. It is therefore very difficult to see how the environment around the spacecraft could have shown a decreasing amount of O
2 relative to water, in ten minutes, if the ALICE team were being fooled by O
2 produced in the close vicinity of the spacecraft. For all the world, it looks like they were indeed observing the molecules close to the nucleus.
ETA:
I'm guessing that by ""in accordance with expectations", above, Keeney et al expect O
2 to become ionised more rapidly than H
2O.