It is surprisingly difficult to freeze water. Almost all ice that forms under "mild" conditions (temperatures > -40 degrees C) requires the presence of a nucleating agent--a solid particle that facilitates the freezing process--such as clay mineral dust, soot or bacteria. In a computer simulation, the presence of such ice nucleating agents does not necessarily alleviate the difficulties associated with forming ice on accessible timescales. Nevertheless, in this work we present results from molecular dynamics simulations in which we systematically compare homogeneous and heterogeneous ice nucleation, using the atmospherically important clay mineral kaolinite as our model ice nucleating agent. From our simulations, we do indeed find that kaolinite is an excellent ice nucleating agent but that contrary to conventional thought, non-basal faces of ice can nucleate at the basal face of kaolinite. We see that in the liquid phase, the kaolinite surface has a drastic effect on the density profile of water, with water forming a dense, tightly bound first contact layer. Monitoring the time evolution of the water density reveals that changes away from the interface may play an important role in the nucleation mechanism. The findings from this work suggest that heterogeneous ice nucleating agents may not only enhance the ice nucleation rate, but also alter the macroscopic structure of the ice crystals that form.

Glycolaldehyde is a simple monosaccharide sugar linked to prebiotic chemistry. Recently it was detected in a molecular core in the star-forming region G31.41+0.31 at a reasonably high abundance. We investigate the formation of glycolaldehyde at 10 K to determine whether it can form efficiently under typical dense core conditions. Using an astrochemical model, we test five different reaction mechanisms that have been proposed in the astrophysical literature, finding that a gas-phase formation route is unlikely. Of the grain-surface formation routes, only two are efficient enough at very low temperatures to produce sufficient glycolaldehyde to match the observational estimates, with the mechanism culminating in CH 3 OH + HCO being favoured. However, when we consider the feasibility of these mechanisms from a reaction chemistry perspective, the second grain-surface route looks more promising, H 3 CO + HCO.

Glycolaldehyde, the simplest monosaccharide sugar, has recently been detected in low-and highmass star-forming cores. Following on from our previous investigation into glycolaldehyde formation (Woods et al. 2012), we now consider a further mechanism for the formation of glycolaldehyde that involves the dimerisation of the formyl radical, HCO. Quantum mechanical investigation of the HCO dimerisation process upon an ice surface is predicted to be barrierless and therefore fast. In an astrophysical context, we show that this mechanism can be very efficient in star-forming cores. It is limited by the availability of the formyl radical, but models suggest that only very small amounts of CO are required to be converted to HCO to meet the observational constraints.