Interstellar glycine & spectroscopic simulations

The simplest amino acid, glycine (NH2CH2COOH), has long been a target for detection in interstellar and star-forming regions; however, it has not yet been securely identified in any such region. Detection of glycine in the interstellar medium or in protostellar envelopes would indicate that star-forming regions may be seeded for life at very early times, before even the formation of a protoplanetary disk.

Experiments involving the irradiation of organic, mixed ices with UV photons or charged particles indicate that glycine and related molecules may be formed through photo-chemical processes, and glycine was recently identified in cometary samples returned to Earth from comet Wild 2; this is encouraging, as interstellar and cometary ices are similar in many respects, and cometary ices are thought to be (at least) partially composed of ice material that was originally formed in the interstellar medium.

In spite of the lack of an interstellar detection of glycine, molecules of similar structure and complexity — such as acetic acid (CH3COOH) and amino acetonitrile (NH2CH2CN) — have been firmly identified. This led me to investigate both whether glycine may indeed be formed within or upon dust-grain ices in star-forming cores known as hot cores, and whether it would be detectable following likely sublimation into the gas phase.

Glycine chemistry simulations

Building on my previous hot-core chemical models, I constructed a new reaction set for glycine and related molecules, based around the addition of radicals within the dust-grain ice mantles. I included four mechanisms for glycine formation:

  1. H + NHCH2COOH → NH2CH2COOH
  2. NH2 + CH2COOH → NH2CH2COOH
  3. NH2CH2 + COOH → NH2CH2COOH
  4. NH2CH2CO + OH → NH2CH2COOH
Reaction (1) is the trivial case of recombination following photo-dissociation or H-abstraction from glycine itself, although it may also follow NH addition to an acetic acid-derived radical. Reaction (2) would require radicals derived from ammonia and acetic acid, respectively, which could be obtained via abstraction or photo-dissociation. Reaction (3) is that suggested by David Woon (2002), which he suggested could occur as the final step following the hydrogenation of HCN and the addition of CO and OH on the grains. In the model, this full mechanism may occur, but is found to be much less important than the formation of the required radicals from methylamine (NH2CH3) and formic acid (HCOOH) formed on the grains by other means. Reaction (4) would be fed by radicals derived from glycinal (NH2CH2CHO), the amino aldehyde associated with glycine, which may be formed by the addition of HCO and NH2CH2 radicals. Various pathways exist to form the precursors for each reaction, and the diffusion rates and abundances of each precursor will be temperature dependent.

GLY_fig
The figure, right, shows the time/temperature-dependent abundances of a selection of complex molecules, using my most recent hot-core chemical model. The molecules shown are: glycine (NH2CH2COOH), glycinal (NH2CH2CHO), amino acetonitrile (NH2CH2CN), propionaldehyde (propanal, C2H5CHO), and propionic (propanoic) acid (C2H5COOH). Solid lines show gas-phase abundances; dotted lines of the same colour show the same species in the solid phase.

Glycine abundance (shown in black) is seen to peak at around 200 K, as a result of its sublimation from the grains. A medium time-scale model run is shown here; the peak gas-phase fractional abundance of glycine ranges from ~8 x 10-11 for the shortest warm-up timescale to 8 x 10-9 for the longest.

Glycine is found to form in/upon the ice mantles in the broad temperature range of ~40 — 120 K. Each of reactions (2) — (4) play a significant role in formation, with the precise influence dependent on time-scale. A plausible gas-phase mechanism for glycine formation in hot cores, involving the reaction of protonated hydroxylamine (NH2OH2+) and acetic acid (CH3COOH) was also tested; its influence was found to be minimal compared to formation within/upon the ice mantles, producing a fractional abundance of only ~10-12.

Spectroscopic simulations

The peak abundances calculated for glycine are quite high; however, the high temperature at which it is predicted to sublime from the grains (~200 K, which is dependent on my binding energy estimate for glycine on water ice) indicates that it would be released in a much more compact region than most hot-core molecules.

To determine whether glycine would be detectable in a hot core, I constructed a spectroscopic modelling code, written in IDL. The code takes an observationally-determined temperature/density/radius profile for a hot core and maps all the chemical abundances onto this framework according to their temperature dependences. This produces a spherically-symmetric three-dimensional grid of absolute abundances of all species at all points in the source. Combining this with the temperature structure allows the emission and absorption characteristics of each molecule to be calculated, using published line lists (CDMS, JPL, Splatalogue), so that the overall emission and absorption coefficients are known at all points (assuming local thermodynamic equilibrium). Radiative transfer is then calculated along lines of sight through the hot core, in frequency bins covering a range of mm/sub-mm wavelengths.

This process produces a data cube (X x Y x freq.) that may be convolved with a telescope beam to determine the strength of line emission at any chosen frequency, allowing simulated spectra to be constructed that may be directly compared with observed spectra. The method naturally takes account of beam-dilution effects and line blending for all molecules that are present in both the line lists and the chemical model.

The southern-hemisphere hot-core source NGC 6334 IRS 1 was chosen for simulation, due to its proximity (~1.7 kpc) and its strong, fairly narrow (~5 km/s), molecular rotational emission lines. The observational temperature/density profile of van der Tak (2000) was used, and a fixed line width of 5 km/s was chosen for all positions in the model. Radiative transfer calculations were conducted in the approximate 218 — 250 GHz range. The results were convolved with telescope beams of size appropriate to the James Clerk Maxwell telescope (12m), the IRAM 30m telescope, and ALMA.

MAP_fig1
MAP_fig2
The figure, left, shows the predicted unconvolved line emission from methyl formate (HCOOCH3) at 221.979 GHz. On the right is the predicted emission from glycine at a similar frequency of 218.105 GHz. Note the significant size difference of the emission regions; ~1.5 arcsec vs. ~0.4 arcsec radii. Telescope beams that cannot resolve the smaller of these would suffer a weaker glycine emission signal, due to beam dilution. While beam dilution would also likely occur for methyl formate emission, using a single-dish telescope, the emission is still sufficiently strong to allow detection. Methyl formate is indeed detected strongly toward the chosen source (Bisschop et al. 2007).

gly_specs
This figure shows simulated spectra that include all molecules present in the chemical model for which spectroscopic data are available; these spectra are convolved with beams of varying size. The top panel shows an unconvolved spectrum; while this eliminates beam dilution, there is still clear line blending (the main emitting molecule is indicated at each peak). Glycine emission at 241.373 GHz is highlighted in red, and has a peak line intensity of ~ 3K. It is also relatively unblended. The middle panel shows the spectrum as convolved with a beam size appropriate to ALMA (0.4 arcsec). The glycine line is still present and unblended. At ~ 2K, it could plausibly be detected using ALMA. The lower panel shows the emission convolved with a beam appropriate to the JCMT (20.3 arcsec). The emission is very much too weak to be detected with the JCMT over realistic observing times.

Similar line convolution for emission from methanol (CH3OH) and methyl formate (HCOOCH3), both typical hot core molecules, produced integrated line intensities for those species that fell within ~50 % of the values observed by Bisschop et al. (2007) for the simulated source (NGC 6334 IRS 1). This shows that the underlying spectroscopic model is plausible, and in this case very accurate.

Searching for glycine

Based on these combined chemical/physical and spectroscopic simulations, it appears that glycine may exist, and be detectable, in some hot cores.

It is essential that future searches for glycine choose nearby hot-core sources whose putative glycine emission region may be resolved with ALMA. Effective searches will also require a careful analysis of which lines may be both strong and unblended with lines from other molecules (using methods outlined here). The use of ALMA will reduce the amount of line blending to some degree as larger-scale emission from more typical hot-core molecules may be resolved out, using an appropriate array configuration.

For more information please see the publication, Garrod (2013).