Comparison of their chemical compositions shows, to first order, a good agreement between the cometary and interstellar abundances. However, a complex O-bearing organic molecule, ethylene glycol (CH$_{2}$OH)$_{2}$, seems to depart from this correlati
on because it was not easily detected in the interstellar medium although it proved to be rather abundant with respect to other O-bearing species in comet Hale-Bopp. Ethylene glycol thus appears, together with the related molecules glycolaldehyde CH$_{2}$OHCHO and ethanol CH$_{3}$CH$_{2}$OH, as a key species in the comparison of interstellar and cometary ices as well as in any discussion on the formation of cometary matter. We focus here on the analysis of ethylene glycol in the nearest and best studied hot core-like region, Orion-KL. We use ALMA interferometric data because high spatial resolution observations allow us to reduce the line confusion problem with respect to single-dish observations since different molecules are expected to exhibit different spatial distributions. Furthermore, a large spectral bandwidth is needed because many individual transitions are required to securely detect large organic molecules. Confusion and continuum subtraction are major issues and have been handled with care. We have detected the aGg conformer of ethylene glycol in Orion-KL. The emission is compact and peaks towards the Hot Core close to the main continuum peak, about 2 to the south-west; this distribution is notably different from other O-bearing species. Assuming optically thin lines and local thermodynamic equilibrium, we derive a rotational temperature of 145 K and a column density of 4.6 10$^{15}$ cm$^{-2}$. The limit on the column density of the gGg conformer is five times lower.
We used several data sets from the Plateau de Bure Interferometer to map the dimethyl ether emission in Orion-KL with different arcsec spatial resolutions and different energy levels to compare with our previous methyl formate maps. Our data show rem
arkable similarity between the dimethyl ether (CH3OCH3) and the methyl formate (HCOOCH3) distributions even on a small scale (1.8x0.8 or about 500 AU). This long suspected similarity, seen from both observational and theoretical arguments, is demonstrated with unprecedented confidence, with a correlation coefficient of maps of 0.8. A common precursor is the simplest explanation of our correlation. Comparisons with previous laboratory work and chemical models suggest the major role of grain surface chemistry and a recent release, probably with little processing, of mantle molecules by shocks. In this case the CH3O radical produced from methanol ice would be the common precursor (whereas ethanol, C2H5OH, is produced from the radical CH2OH). The alternative gas phase scheme, where protonated methanol CH3OH2+ is the common precursor to produce methyl formate and dimethyl ether through reactions with HCOOH and CH3OH, is also compatible with our data. Our observations cannot yet definitely allow a choice between the different chemical processes, but the tight correlation between the distributions of HCOOCH3 and CH3OCH3 strongly contrasts with the different behavior we observe for the distributions of ethanol and formic acid. This provides a very significant constraint on models.
