Exploring new phenomena in ice, salty ice, and ice clathrates under extreme conditions

• Livia Eleonora Bove (CNRS & Université P&M Curie, EPFL)

Room: NO Building, Room C 60 Compressed water is overspread on Earth at depth and in the extra-terrestrial space, both interstellar and on outer planets and moons (ice bodies) [1]. Under the conditions experienced in these celestial bodies water displays an incredibly rich phase diagram, including seventeen known crystalline phases and three amorphous states, and predicted exotic properties like plasticity [2], ionization [3], and superionicity [4]. In this talk I will review some of our recent experimental results on pure ice, salty (LiCl, LiBr, NaCl) ices, and ice (H2, CH4) clathrates under extreme conditions including: new dynamical effects in ice VII [5], salty ice crystallization under high pressure [6,7], hydrogen bond symmetrisation in salt-doped ices at Mbar pressures [8,9], methane ultra fast diffusion and locking in methane hydrates [10,11]. [1] De Pater, I., and Lissauer, J.J. Planetary Sciences, Cambridge University Press (2004). [2] Aragones, L., and Vega, C., J. Chem. Phys. 130, 244504 (2009). [3] Wang, Y., Liu, H., et al. Nat. Comm. 563 1566 (2011). [4] Cavazzoni, C., et al., Science 283, 44-46 (1999). [5] Polidori A., Ranieri U., et al., under submission. [6] Klotz, S., Bove, L. E. et al. Nat. Mat. 8, 405 (2009) [7] Klotz, S., Bove, L. E. et al., Nat. Sci. Rep. 6, 32040 (2016). [8] Bove, L. E. , Gaal, R. et al., PNAS 112, 8216 (2015); [9] Bronstein, Y., et al. , Phys. Rev. B 93 024104 (2016). [10] Ranieri, U.L., et al.., Nat. Com., in press (2017). [11] Schaack, S., Ranieri, U.L., submitted (2017)

Born-Oppenheimer molecular-dynamics study of ice polymorphs (Phases of Ice)

• Niall English (University College Dublin, Chemical Engineering)

Room: NO Building, Room C 60 We report ab-initio molecular dynamics, using state-of-the-art non-local dispersion, to study, inter alia, the structural, hydrogen-bonding, vibrational and Raman properties of ices Ih, VII and XVII at appropriate temperature and pressure ranges for their stability. In many senses, probing of ice Ih serves as a validation tool for the functional, with reasonable agreement with experiment. In the case of ice VII, we are motivated to elucidate any possible hints of intramolecular strain that may serve as precursor events for proton hopping to unfold. We determine such equilibrium properties to be in reasonable agreement with experiment, although we do not detect any water-dissociation and proton-hopping events per se, owing to still-large water-dissociation free-energy barriers. The observed anomalous self-diffusivity of ice VII in the region of 10 GPa at ~400 K has been suggested to arise from a change in proton-hopping mechanism involving a transition from ionic-defect-driven diffusivity to that dominated by diffusion of rotational defects. In the case of ice XVII, we probe the stability of the cage structure, and dissect the vibrational properties, obtaining reasonable agreement with recently-reported experiments.

New hydrates in the Ne-, He- and D2-water system via a high-pressure route (Amorphous Ice)

• Andrzej Falenty (GZG Abt. Kristallographie Uni Göttingen)

Room: NO Building, Room C 60 Formation of gas hydrates from ice is generally a slow process due to the intrinsically sluggish solid-state mass transfer (Salamatin et al. 2015). Small size Ne, H2 and He are an exception to this rule by being able to penetrate into the lattice of the ordinary ice Ih with ease. Upon further compression ice Ih presaturated with He-gas was found to transform rapidly into gas-filled ice II. Interestingly, H2 and Ne were found also to form a cubic sII clathrate hydrate at intermediate pressures. In planetary science, such ices filled with Ne, H2 and He are seen as potent materials selectively trapping and fractionating these primordial gases. Moreover, in the context of fundamental clathrate research Ne- and H2-hydrates have played the unorthodox role as starting point for the formation of ice XVI, the empty clathrate structure sII (Falenty et al. 2014). Here we report on the gas loading into ice Ih spheres. The pressure in the system was increased stepwise to 320MPa for D2 and 400MPa for He and Ne. All guest types were found to rapidly penetrate into ice yet with marked differences. Pressurization of ice with deuterium lead to a rapid formation of sII D2-hydrate initiated already at 110 MPa. The sII structure remained stable at least up to 320MPa. Ice loaded with Ne remained stable up to 210MPa, i.e. to approximately the stability boundary between pure Ih and unfilled ice II. After crossing this threshold, the sample adopted the sII Ne-hydrate structure and remains stable at least up to 350MPa. Further compression up to ~400MPa resulted in a new Ne-hydrate, topologically similar to the H2-filled chiral C0-structure and the recently discovered C0 structure of CO2 (Amos et al. 2017). Upon decompression to 350MPa C0 structure reverted back to sII Ne-hydrate. Upon recovery at liq. N2 temperature to ambient pressure, cage occupancies in Ne- and D2-hydrates rapidly readjust by decreasing the number of guests in all cages. A similar pattern can be seen upon guest-removal under vacuum where larger cages empty much faster than the smaller ones; curiously D2 molecules at comparable conditions seems to leave the sII structure slower than the larger Ne atoms. Smaller cavities seem to be much more difficult to empty and may require the presence of water vacancies for the cage-to-cage hopping. The He-water system has proven to be the most peculiar one as the compression up to 300MPa did not yield any structural change. He-loaded ice Ih remained stable in the whole investigated pressure range. Still, for the first time we succeeded to form unequivocally sII He-clathrate by filling ice XVI at 120K with 150MPa of He gas; both, the crystal structure and the He-uptake and release, was studied as a function of pressure. Amos et al.(2017) J.Phys.Chem.Lett. 8, 4295-4299 Falenty, Hansen & Kuhs (2014) Nature 516, 231-233 Salamatin, Falenty, Hansen & Kuhs (2015) Energy & Fuels 29, 5681-5691

Gas selectivity in mixed gas hydrates formed from pressurized ice. (Phases of Ice)

• Arnaud DESMEDT (CNRS)

Room: NO Building, Room C 60 In addition to their various applications in the environment, energy and technology fields, gas hydrates are supposed to be involved in the formation of planetesimals, comets and other planets, such as Titan or Enceladus, both being Saturn’s moons. For instance, CO gas hydrate was one of the hydrates especially highlighted in the first work outlining this hypothetical hydrate formation in astrophysical conditions. Moreover, recent theoretical works suggest that the nitrogen depletion observed on the Jupiter family comet 67P/Churyumov-Gerasimenko might be due to N2 encapsulation within hydrates and so assume that 67P agglomerated from these crystalline compounds. The preferential trapping of carbon monoxide with respect to nitrogen inside mixed gas hydrate allow to determine the temperature range of the cometary grains formation in the protosolar nebula. The present work is dedicated to the experimental investigation of such a preferential encapsulation by means of Raman scattering in various mixed gas (CO, CO2, N2) hydrates. Beyond the astrophysical interest, the molecular selectivity in mixed gas hydrate plays an important role in many areas such as process engineering or geoscience. The investigation is based on the study of the gas selectivity in gas hydrates formed with high driving force, i.e. by ice exposure to different gaseous mixtures (in various pressure and temperature conditions). These studies have been carried out on the pure gas hydrates (CO, N2 and CO2) as well as on the corresponding mixed gas hydrate. Such a series of gaseous molecules offer the opportunity to explore the impact of various physical-chemistry parameters onto the gas selectivity by forming hydrates. Many factors, such as steric hindrance or thermodynamic conditions of formation, govern this selectivity. CO2 possesses a strong water affinity while CO and N2 do not. CO and N2 molecules have similar steric hindrance and close hydrate formation P-T conditions. CO is the only considered guest molecules possessing a dipolar moment. Raman spectroscopy results have put forward a clear preferred encapsulation of carbon monoxide molecules compared to nitrogen molecules during the mixed gas hydrate formation. Similar preferential encapsulation is observed in the case of CO2 molecules with respect to CO and N2 molecules. Moreover, neutron powder diffraction data have shown phase mixtures involving various structures (so-called type I and type II) and structural metastability. This whole set of results will be discussed in conjunction with DFT calculations, kinetic formation conditions and structural metastability properties evidenced in these gas hydrates.

Why is Ice Less-dense and Slippery? (Phases of Ice)

• Changqing Sun (NTU, SIngapore)

Room: NO Building, Room C 60 As the source and central part of all lives, water is most abundant yet least known. This presentation shares the recent progress [1-6]: (i) correlation of the length scale, structure order, and mass density of molecular packing in water ice; (ii) potential paths for O:H-O bond at relaxation; and anomalies of water ice under (iii) compression; (iv) molecular under-coordination; and (v) thermal excitation. Hydrogen bond (O:H-O) possesses memory and extreme deformation recoverability, which resolves mysteries of: (i) density of ice [7]; (ii) slipperiness of ice [8]; (iii) Mpemba paradox - hot water freezes faster [9]; (iv) Regelation [10] – ice melts under compression and freezes again when the pressure is relieved. Understanding may extend to fields such as water - bio-molecular interaction, water purification, energy management, etc. 1. Sun, C.Q. and Y. Sun, The Attribute of Water: Single Notion, Multiple Myths. Springer Ser. Chem. Phys. Vol. 113. 2016, Heidelberg: Springer-Verlag. 494 pp. 2. Huang, Y.L., et al., Hydrogen-bond relaxation dynamics: Resolving mysteries of water ice. Coordination Chemistry Reviews, 2015. 285: p. 109-165. 3. Zhang, X., et al., Water’s phase diagram: from the notion of thermodynamics to hydrogen-bond cooperativity. Progress in Solid State Chemistry, 2015. 43: p. 71-81. 4. Sun, C.Q., X. Zhang, and W.T. Zheng, Hidden force opposing ice compression. Chem Sci, 2012. 3: p. 1455-1460. 5. Sun, C.Q., et al., Density and phonon-stiffness anomalies of water and ice in the full temperature range. Journal of Physical Chemistry Letters, 2013. 4: p. 3238-3244. 6. Sun, C.Q., et al., Density, Elasticity, and Stability Anomalies of Water Molecules with Fewer than Four Neighbors. Journal of Physical Chemistry Letters, 2013. 4: p. 2565-2570. 7. Sun, Q., Raman spectroscopic study of the effects of dissolved NaCl on water structure. Vibrational Spectroscopy, 2012. 62: p. 110-114. 8. Zhang, X., et al., From ice supperlubricity to quantum friction: Electronic repulsivity and phononic elasticity. Friction, 2015. 3(4): p. 294-319. 9. Zhang, X., et al., Hydrogen-bond memory and water-skin supersolidity resolving the Mpemba paradox. Physical Chemistry Chemical Physics, 2014. 16(42): p. 22995-23002. 10. Zhang, X., et al., Ice Regelation: Hydrogen-bond extraordinary recoverability and water quasisolid-phase-boundary dispersivity. Scientific Reports, 2015. 5: p. 13655.

Experimental evidence for a second hydrogen ordered phase of ice VI (Phases of Ice)

• Tobias M. Gasser (University of Innsbruck)

Room: NO Building, Room C 60 In the last decade five new ice phases were experimentally prepared. Two of them(1, 2) are empty clathrate hydrates and three of them(3, 4) represent hydrogen ordered counterparts of previously known disordered ice phases. In our experiments, we investigated hydrogen ordering in ice VI samples upon cooling at pressures up to 1.8 GPa. Using calorimetry, dielectric relaxation spectroscopy, Raman spectroscopy, and powder X-ray diffraction we provide evidence for the existence of a second hydrogen ordered phase related to ice VI, that we call ice β-XV. This phase is more ordered than ice XV by 14% and directly transforms to ice XV above 103K and to ice VI above 129K. That is, upon heating an order->order->disorder transition is experienced. The new phase is thus thermodynamically more stable than ice XV requiring a new stability region in the phase diagram of water. Raman spectroscopy indicates ice XV and ice β-XV to be different in terms of symmetry and space group. The activation energies, measured by dielectric spectroscopy, are 45 kJ mol-1 in ice β-XV compared to 18 kJ mol-1 in ice XV. Powder X-ray data show the oxygen network to be the one of ice VI. The ordering of hydrogen atoms induces a significant peak shift to lower d-spacings at d=0.265 nm in ice β-XV, whereas for ice XV shifts to higher d-spacings are found. The reproduction of these effects with fully deuterated samples would be desirable to refine the crystal structure of ice β-XV. Unfortunately, a pronounced isotope effect precludes such experiments and leads us to the temporary name ice β-XV. We suggest to change this name to ice XVIII as soon as the crystal structure is known. Our experiments represent a unique realization of a second electric ordering in an ice phase, including the identification of two triple points between three hydrogen ordered phases. 1. del Rosso L, et al. (2016) Refined Structure of Metastable Ice XVII from Neutron Diffraction Measurements. J. Phys. Chem. C 120(47):26955-26959. 2. Falenty A, Hansen TC, & Kuhs WF (2014) Formation and properties of ice XVI obtained by emptying a type sII clathrate hydrate. Nature 516(7530):231-+. 3. Salzmann CG, Radaelli PG, Hallbrucker A, Mayer E, & Finney JL (2006) The preparation and structures of hydrogen ordered phases of ice. Science 311(5768):1758-1761. 4. Salzmann CG, Radaelli PG, Mayer E, & Finney JL (2009) Ice XV: A New Thermodynamically Stable Phase of Ice. Phys. Rev. Lett. 103(10):105701.

Anisotropic impurity effect for ice crystal growth in supercooled water

• Yoshinori Furukawa (Institute of Low Temperature Science, Hokkaido University)

Room: NO Building, Room C 60 Antifreeze protein (AFP) and antifreeze glycoprotein (AFGP) molecules adsorb on ice-water interfaces and control the ice crystal growth. However, there still remain many missing pieces of the puzzle for the growth control mechanism. The basic observations from the viewpoint of crystal growth, for example, in-situ observations of morphologies, precise measurements of growth rates, understandings of the adsorption states of these molecules on ice-water interface, are important to clarify the mechanism of crystal growth control. We carried out various experiments on free growth of ice crystals in supercooled water both on the ground and under the microgravity condition without any influence of convection, which is realized in the International Space Station. Based on these experiments, we discuss about the anisotropic effect of AFGP molecules as an impurity for ice crystal growth [1]. In situ observation of ice crystal growth in a supercooled water including AFGP molecules labeled by the fluorescent molecules indicated that growth of the prismatic (and/or pyramidal) faces was inhibited by the adsorption of AFGP molecules. Growth inhibition for these faces is explained by the pinning effect of adsorbed AFGP molecules, which is the well-known impurity effect for the crystal growth. In contrast, long-term continuous measurements of growth rates performed in ISS clarified that the growth of basal face is promoted by the effect of AFGP molecules. Namely, the adsorbed AFGP molecules may work as the inhibitor for prismatic and pyramidal faces but as the promotor for basal faces. These results indicate the anisotropic impurity effect depending on the crystallographic orientation of interfaces. On the contrary, we found that the growth rates of basal faces were periodically changed as a function of growth time. The growth enhancement and oscillation observed on the ice basal faces are the previously unknown impurity effects for crystal growth of not only the ice crystal but also the other crystals. A preliminary explanation for this anisotropy will be given on the basis of the preferential adsorption of AFGP molecules on the prismatic and/or pyramidal faces [2,3]. Our findings will lead to a better understanding of a novel process for ice growth kinetics, and shed light on the role that crystal growth kinetics has in the onset of the mysterious antifreeze effect in living organisms, namely, how this protein can prevent freezing of living organisms under sub-zero temperature condition. References: [1] Y. Furukawa et al.: Scientific Reports, 7:43157, (2017), doi:10.1038/srep43157. [2] S. Zepeda, E. Yokoyama, Y. Uda, C. Katagiri and Y. Furukawa, Cryst. Growth Des. 8, 3666-3672, (2008), doi:10.1021/cg800269w. [3] Y. Uda, et al.: J. Phys. Chem. B, 111, 14355-14361(2007), doi:10.1021/jp075429s.

Snowflake growth in supersaturated atmosphere using a three-dimensional phase-field model (Ice Crystal Growth)

• gilles demange (GPM, University of Rouen)

Room: NO Building, Room C 60 Simulating ice crystal growth is a major issue for meteorology and aircraft safety [1]. Notwithstanding, very few models currently succeed in reproducing the diversity of snowflake forms in three dimensions, and the link between model parameters and thermodynamic quantities is far from being established. Here, we present a 3D modified phase field model that describes the subtlety of the ice-vapor phase transition leading to snow crystal growth, through anisotropic water molecules attachment and condensation, surface diffusion, and strongly anisotropic surface tension [2]. We demonstrate that this model is capable of reproducing properly the morphology and growth kinetics of snowflakes in supersaturated atmosphere. Aside from that, we show that the growth dynamics of snow crystals simulated thereby, satisfies the selection theory, consistently with previous experimental observations [3]. Finally, we present a morphology diagram in experimentally accessible units, that links the manifold of simulated snowflake shapes to atmospheric parameters. [1] BRUNEL, M., DEMANGE, G., FROMAGER, M., et al. Instrumentation for ice crystal characterization in laboratory using interferometric out-of-focus imaging. Review of Scientific Instruments, 2017, vol. 88, p. 083108. [2] DEMANGE, G., ZAPOLSKY, H., PATTE, R., et al. A phase field model for snow crystal growth in three dimensions. NPJ Computational Materials, 2017, vol. 3, p. 1. [3] DEMANGE, G., ZAPOLSKY, H., PATTE, R., et al. Growth kinetics and morphology of snowflakes in supersaturated atmosphere using a three-dimensional phase-field model. Physical Review E, 2017, vol. 96, p. 022803.

Time dependent change of tip velocities of a dendritic ice crystal in growing from supercooled water (Ice Crystal Growth)

• Etsuro Yokoyama (Computer Centre, Gakushuin University)

Room: NO Building, Room C 60 We measured the growth rates of secondary branches of a dendritic ice crystal growing from supercooled heavy water in both microgravity and 1g conditions. We have analyzed recorded sequences of images of microgravity experiments that were carried out 134 times in the International Space Station(ISS) and measured the growth rates of both the tip of a primary dendrite and the basal faces of an ice crystal growing from supercooled heavy water in the range of supercooling from 0.03 to 2.0 K and found that the steady state primary tip velocity agrees with the dendritic growth theories [1,2] of dimensionless supercooling for Δ > 0.002 (=0.16 K) when the growth on the basal face is not zero [3]. In contrast to steady-state primary tips motion, the measured data show that the tip growth rates of secondary branches are not constant with time and almost half of that of the primary tip in the initial stages and subsequently decrease with time for the small undercooling region < 1.0 K. The time dependency of tip growth rates of secondary branches is different for microgravity and 1g conditions. For the larger undercooling, the time dependency can depend on the position on the main stem. Furthermore, we extracted the contour images of a dendritic ice crystal from images and examined the pattern matching of extraction contours around the tips of both primary and secondary branches. The dendrite tip of the ice crystal is composed of two distinct tip radii, R1 and R2. The tip radius R1 is parallel to the basal plane, while the tip radius R2 is perpendicular to the basal plane. The contour images corresponding to R1 show that the radii do not change with time and the radii tips of secondary branches agree with that of the primary tip. As a result, we expect that all radii R1 of dendritic tips have the same value and determined by supercooling, nevertheless the tip velocities of secondary branches are different from the primary tip velocity. Finally, we calculate a critical value of the interfacial stability parameter [2,4] using the data obtained from both the ISS experiments and 1g experiments and show these two sets of the values are fairly separated. The critical value in 1g decreases with increasing dimensionless supercooling, while the value in microgravity is almost constant because of no natural convection. This result is almost the same as the analysis for succinonitrile experiments having four fold symmetry [4] and agrees with the Interfacial Wave (IFW) Theory [2]. We also discuss the occurrence of side branches and the asymmetric with respect to the main stem. [1] J. S. Langer, H. J. Müller-Krumbhaar, J. Cryst. Growth, 42(1977)11/ Acta Metall., 26(1978) 1681. [2] J. J. Xu, Phys. Rev. A 43(1991)930. [3] E. Yokoyama et al., J. Phys. Chem., 115 (2011)8739. [4] J. J. Xu, D. S. Yu, J. Cryst. Growth, 198(1999)43.

Upcaling microscopic crystal growth dynamics in snow (Ice Crystal Growth)

• Quirine Krol (WSL Institute for Snow and Avalanche Research SLF)

Room: NO Building, Room C 60 While the growth of single ice-crystals in the atmosphere is reasonably well understood, the microstructure dynamics of snow on the ground as the collective growth of aggregated crystals, is still lacking a unified treatment based on first principles of ice crystal growth. One of the key-challenges is the lack of models for the microstructure evolution of the bicontinuous ice matrix. To this end, we developed a rigorous microstructure upscaling scheme which takes common pore-scale (diffusion) principles of crystal growth to predict the volume-averaged evolution of the interface morphology. We derived a coupled set of evolution equations for the (volume averaged) ice volume fraction, specific surface area, Gaussian curvature and first and second moment of the mean curvature distribution, and demonstrate their correctness by a comparison to the evolution of different non-spherical grains under a prescribed growth law. In a second step, we use the model as a benchmark tool without a-priori assumptions for a comparison to experiments of snow microstructure evolution via interface tracking on 4D X-ray tomography data. The benchmarking allows us to quantify uncertainties in local estimates of crystal growth velocities. Finally, we demonstrate how the rigorous model facilitates a statistical assessment of common growth laws by combining 4D microstructure data with finite element numerics for heat transfer. The results e.g. show that a decay of the specific surface area cannot be explained by diffusion-limited growth models for the interface velocity. In addition, the evolution of microstructural parameters confirm a morphological asymmetry between growth and sublimation sites. Further conceptual insight is therefore needed to improve the link between volume averaged microstructural evolution in snow and ice crystal growth mechanisms.

Hypothetical Ultralow-density Ice Polymorphs (Ice Crystal Growth)

• Masakazu Matsumoto (Research Institute for Interdisciplinary Science, Okayama University)

Room: NO Building, Room C 60 Among 17 known ice polymorphs, only two are less dense than normal ice, and they are made very recently by a tricky procedure, vacuum pumping the small guest molecules from clathrate hydrate. The discovery of low-density ice polymorphs opens the door to the survey on ices under negative pressures. Negative pressure is very hard to be accessed experimentally, and theoretical predictions are indispensable. How many stable phases are possible for ices under negative pressures? Are two ices really the most stable phases at some thermodynamic conditions? How low can the density of ice be? Exhaustive evaluation of sparse ice structures and theoretical considerations provide unexpected answers to these questions.

Ice crystal formation on soot particles (Ice Crystal Growth)

• Fabian Mahrt (Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, 8092, Switzerland)

Room: NO Building, Room C 60 Soot particles are primary particles produced by incomplete combustion of both biomass and/or fossil fuels and thus constitute a major anthropogenic pollutant. They are generally complex internal mixtures of black carbon (BC) and organic matter (OM) (Bond et al., 2013; Petzold et al., 2013). While these properties strongly depend on the emission sources, they can be altered during atmospheric cycling, including interactions of the primary particles with other atmospheric matter and/or gases. Physically, soot particles usually form fractal-like aggregates composed of the primary carbonaceous spherules. Such aggregates include a variety of cavities and pores that are formed between primary carbonaceous spherules or as a result of random packing of the fractal chains. Overall, estimates of the net radiative forcing associated with BC are still associated with large uncertainties, even though being one of the most important climate forcing agents (Bond et al., 2013). While BC generally absorbs solar radiation, thus contributing to a warming effect on global climate it can also act as a heterogeneous ice nucleating particle (INP) and impact cloud-radiation interactions, potentially cooling the climate (Lohmann, 2002). Prediction of the ice nucleation behavior of soot particles requires consideration of both, the chemical and physical properties of the particles. Here, we present a systematic laboratory-based investigation of the ice formation behavior of a set of soot particles. Different commercial soot samples are used, along with a soot produced by a propane diffusion burner, using a Combustion Aerosol Standard Generator (miniCAST, JING AG). miniCAST burners are frequently used as proxy for atmospheric soot particles. Ice nucleation ability of these soot types is tested on size-selected particles covering the size range between 100 and 400 nm and a temperature range from 253 K to 218 K, using the Horizontal Ice Nucleation Chamber (HINC, Lacher et al. 2017), a Continuous Flow Diffusion Chamber (CFDC). We complement our ice nucleation experiments with auxiliary measurements including BET-surface determination and thermogravimetric analysis, in order to characterize the physio-chemical properties of the tested aerosol particles. In addition, analysis of Transmission Electron Microscope (TEM) images allows for a detailed characterization of the morphology associated with our soot samples. This supports investigating the influence of any cavities on our ice nucleation results, considering a Pore Condensation Freezing (PCF) mechanism for ice nucleation onto soot particles. During PCF liquid water is taken up within pores below water saturation due to the inverse Kelvin effect, which can subsequently freeze heterogeneously in presence of an active site (for T > 235 K) or homogeneously (for T < 235 K) causing ice to nucleate on the soot aggregate. Results show different activation behavior of the soot over the temperature range investigated. While the commercial soot samples show freezing well below water saturation in the cirrus regime, CAST-brown soot needs conditions above water saturation to show any activation. For the mixed-phase cloud conditions all soot types show droplet activation for high supersaturations w.r.t water.

Hydrochloric acid droplets on ice crystal surfaces embedded in ice by vapor growth (Ice Crystal Growth)

• Ken Nagashima (Institute of Low Temperature Science, Hokkaido University, Japan)

Room: NO Building, Room C 60 Ice crystal surfaces are covered with quasi-liquid layers (QLLs) and it leads great influence on the global environment whether ice surfaces are dry or wet. For example, since QLLs enhance various chemical reactions in ice clouds, the formation of QLLs by atmospheric gases has been studied intensively. In particular, chemical reactions on ice surfaces in the presence of hydrogen chloride (HCl) gas play important roles in ozone-depleting reactions [1]. However, such studies were performed using spectroscopy techniques, which did not have enough spatial and temporal resolution. After that previous method would not reveal the dynamic behavior of individual QLLs because we revealed by advanced optical microscopy that the QLLs are not static complete wetting but dynamic partial wetting on ice surfaces [2]. Therefore, we show direct visualization of QLLs on ice basal faces in the presence of 0.1%HCl gas by advanced optical microscopy. We found that the HCl gas induced the appearances of liquid layers with a droplet shape in the temperature range of -15.0 ~ -1.5 degC, where no QLL appears in the absence of HCl gas [3]. However, note that since the lowest temperature of our experimental setup is -15.0 degC, the lowest temperature for the appearance of the droplet QLLs is still unknown. We concluded that the droplets were not bulk pure water but HCl solution because we observed various features of the HCl-induced droplets as described below, which were not observed in the absence of HCl gas. Under undersaturated condition, the HCl droplets were irregularly moving and broke into smaller ones [3]. Because the HCl droplets could be assumed to be equilibrium with ice crystals, the HCl concentration is adjusted by melting and freezing of ice crystal so that its melting point equals to the ice temperature. Therefore, these phenomena are closely related to the evaporation of the HCl droplets and subsequent melting of ice crystals caused by the increase in the HCl concentration. Under supersaturated condition, the HCl droplets were quickly embedded into ice crystals because growth of ice crystals preferential started from the droplets surface and spread in the horizontal direction. After that, when we gradually decreased the water vapor pressure to undersaturated condition, the embedded droplets reappeared on ice crystal surfaces by the evaporation of ice. The HCl droplets were equilibrium with ice crystals around the droplets because reappearance droplets were same size as embedded droplets. Therefore, when the temperature of ice crystals were -10 degC, the HCl concentration of the embedded droplets were estimated at 7 wt.% by the melting point of HCl solution. The new incorporation mechanism of HCl gas into ice crystals suggested the possibility that natural ice crystals also include soluble atmospheric gases. For example, this mechanism might explain a large amount of chloride ions stored in the polar stratospheric clouds that cause ozone depletion. [1] McNeill et al. (2006) PNAS 103, 9422. [2] Sazaki et al. (2012) PNAS 109, 1052. [3] Nagashima et al. (2016) Cryst. Growth Des. 16, 2225.

Laboratory Experiments on Heterogeneous Nucleation and Growth Rates of pure H2O ice and hydrates of HCl and HNO3 in the range 165-195 K . (Nucleation)

• Michel J. ROSSI (Paul Scherrer Institute (PSI))
• Riccardo Iannarelli (EPFL)

Room: NO Building, Room C 60 A multidiagnostic Stirred Flow Reactor has been used to investigate heterogeneous nucleation of pure H2O ice from water vapor depositing on a silicon substrate mounted in a cryostat at controlled temperatures. The gas phase has been monitored using residual gas mass spectrometry, the cryogenic deposit was investigated using FTIR spectroscopy in transmission in the range 650 to 4000 cm-1 across typically 0.5 to 2 micrometer thick ice films. Subsequent to ice deposition controlled amounts of HCl and HNO3 were deposited on top of the previously deposited thin ice film all the while monitoring both FTIR absorption as well as the partial pressures of H2O and HCl or HNO3 using m/e 18, 36 and 46 amu, respectively. Ice deposition on the Si substrate of the cryostat required a supersaturation ratio r (with respect to pure ice) varying from 2.22 to 16.45 with decreasing temperature at a typical growth rate of 2 monolayers (ML) s-1. The data lie on a straight line when plotted as ln(1/r) vs. 1/T with an activation energy of 20.2 ± 2.6 kJ mol-1 (Arrhenius plot). The underlying idea of the rate model for ice nucleation is that 1/r is proportional to the encounter probability (or rate) of individual H2O molecules adsorbed on the thin (hydroxylated) SiO2 (quartz) surface of the Si window. Low temperature leads to a high surface coverage of adsorbed H2O and to a large encounter probability to form a viable ice embryo at the SiO2 interface. However, the surface self-diffusion of adsorbed H2O is an activated process whose rate slows down with decreasing temperature. The measured activation energy of 20.2 kJ mol-1 corresponds to the bond energy of a single Hydrogen bond between a H2O molecule adsorbed onto a surface OH-group of SiO2. In contrast, when depositing HCl from the gas phase onto the thin H2O ice film we observe barrierless, that is spontaneous nucleation of HCl hexahydrate (HCl•6H2O or HH) at T < 173 K monitored by FTIR absorption in situ. With increasing HCl deposition the growth rate of HH, typically 2 ML s-1, slows down owing to the decrease of the surface diffusion rate of adsorbed HCl across the existing HH film and comes to a halt after approximately 500 ML of deposited HCl. When depositing HNO3 on pure thin ice films we essentially observe barrierless spontaneous formation of alpha-NAT (HNO3•3H2O) or NAD (HNO3•2H2O) at typical growth rates of 0.6 ML s-1 akin to the HCl/ice system. In some cases a very small supersaturation is required to initiate crystal growth. The key message from this work is that spontaneous, that is barrierless, nucleation enables the formation of HCl and HNO3 hydrates (type II Polar Stratospheric Clouds or PSC’s) in the presence of stratospheric ice (type I PSC’s).

Amorphous ices and liquid states

• Katrin Amann-Winkel (Stockholm University)

Room: NO Building, Room C 60 The amorphous forms of water play an important role in the understanding of water´s anomalous properties. Computer simulations suggests that the anomalous behaviour of ambient and supercooled water could be explained by a two state model of water [1]. Since the discovery of two distinct amorphous states of ice with different density (high- and low density amorphous ice, HDA and LDA) it has been lively debated whether and how this phenomenon of polyamorphism at high pressures is connected to the occurrence of two distinct liquid phases (HDL and LDL) [1]. Alternatively, amorphous ices have been suggested to be of nanocrystalline nature, unrelated to liquids. The glass transition in both amorphous states, LDA and HDA, was investigated using different experimental techniques [2] and discussed controversially over the last years [3]. Our recent results using X-ray diffraction as well as X-ray correlation spectroscopy (XPCS) support the previous findings of HDA undergoing a glass-liquid transition at ambient pressure around 110 K and are consistent with the hypothesis of a liquid-liquid transition between HDL and LDL [4]. [1] P. Gallo, K. Amann-Winkel et al., Water a Tale of Two Liquids, Chem. Rev. 116, 7463-7500 (2016) [2] K. Amann-Winkel, C. Gainaru et al., Water’s second glass transition, PNAS 110, 17720 (2013) [3] K. Amann-Winkel et al., Water´s controversial glass transition, Rev. Mod. Phys. 88, 0110002 (2016) [4] F. Perakis, K. Amann-Winkel et al., Diffusive dynamics during the high-to-low density transition in amorphous ice, PNAS 114, 8193 (2017)

Calorimetric studies on doped high-density amorphous ices (Amorphous Ice)

• Violeta Fuentes Landete (Institute of Physical Chemistry, University of Innsbruck, Austria)

Room: NO Building, Room C 60 Studies on doped crystalline ice phases have been done over the past decades. Dopants have allowed unlocking the kinetically hindered, but thermodynamically favored transitions to hydrogen ordered ices, allowing the discovery of ices XI[1], XIII[2], XIV[2] and XV[3]. The role of these dopants is to enhance the microscopic dynamics up to 100.000[4], as reveal by the dielectric studies. This enhancement is the key to induce hydrogen-ordering transitions at T < 150K. Given the claim of a high similarity between amorphous ices and crystalline ices (e.g., high-density amorphous ice and ice VI[5] or ice VII[6]), it is of interest to study which impact these dopants might have on hydrogen-ordering in amorphous ice. In the present contribution we discuss the influence of dopants on equilibrated high-density amorphous ice (e-HDA) by differential scanning calorimetry and compare the phenomenology (i.e., appearance of latent heat associated with hydrogen ordering, impact on glass transition step) with those cases usually reported for crystalline ices. None of the large variety of studied dopants has an influence like that seen earlier on crystalline ice, thereby refuting the claim of crystalline-like nature of e-HDA. 1. S. Kawada, J. Phys. Soc. Jpn. 32 (5), 1442 (1972). 2. C. G. Salzmann, P. G. Radaelli, A. Hallbrucker, E. Mayer and J. L. Finney, Science 311, 1758 (2006). 3. C. G. Salzmann, P. G. Radaelli, E. Mayer and J. L. Finney, Phys. Rev. Lett. 103, 105701 (2009). 4. K. W. Koester, A. Raidt, V. Fuentes-Landete, C. Gainaru, T. Loerting and R. Boehmer, Phys. Rev. B 94 (18), 184306 (2016). 5. A. I. Kolesnikov, V. V. Sinitsyn, E. G. Ponyatovsky, I. Natkaniec and L. S. Smirnov, Physica B: Condensed Matter 213 (Supplement C), 474-476 (1995). 6. A. M. Saitta, T. Strässle and S. Klotz, EPL (Europhysics Letters) 74 (3), 445 (2006).

Acid-Promoted Crystallization of Amorphous Solid Water (Amorphous Ice)

• Du Hyeong Lee (Seoul National University)

Room: NO Building, Room C 60 Amorphous solid water (ASW) has attracted much attention in the ice research community because of its unique properties as a metastable solid form of water and possible existence in interstellar clouds. ASW is kinetically metastable and undergoes a phase transition to crystalline ice (CI) upon heating; these phenomena have been studied in depth in recent years. [1-3] In this presentation, we discuss the effect of excess protons (hydroniums) on the crystallization of ASW. We prepared an ASW film sample by water vapor deposition on a Pt(111) single crystal surface in an ultra-high vacuum environment. Excess protons were generated by the ionization of adsorbed hydrogen chloride (HCl) gas on the ASW surface. The crystallization of an ASW film was monitored by temperature-programmed desorption and reflection-absorption infrared spectroscopy measurements. In contrast to a normal behavior that solute species retard the crystallization of ASW due to an entropic effect, the excess protons released from HCl molecules facilitated the crystallization of ASW. This phenomenon was observed when HCl was supplied not only at the surface of an ASW film but also in the interior region. In both cases, the crystallization was initiated at the location where the acid was provided to the sample. The presence of excess protons lowered the apparent activation energy of the crystallization of ASW. A possible promotion mechanism for the crystallization of ASW is discussed in relation to the properties of excess protons in ASW. [1] Safarik, D. J.; Mullins, C. B. The nucleation rate of crystalline ice in amorphous solid water. J. Chem. Phys. (2004) 121, 6003-6010. [2] Backus, E. H.; Grecea, M. L.; Kleyn, A. W.; Bonn, M. Surface crystallization of amorphous solid water. Phys. Rev. Lett. (2004) 92, 236101. [3] Yuan, C; Smith, R. S.; Kay B. D. Surface and bulk crystallization of amorphous solid water films: Confirmation of “top-down” crystallization. Surf. Sci. (2016) 652, 350-354.

Study of crystallite size distributions in ices and gas hydrates using a novel X-ray diffraction procedure (Phases of Ice)

• Werner F Kuhs (University of Göttingen)

Room: NO Building, Room C 60 Crystallite size distributions (CSDs) provide important insights into the formation processes of materials, in particular the balance between nucleation and growth rates. Both equilibrium and non-equilibrium physical properties of the resulting polycrystalline gas hydrate assemblages will depend on the CSDs, in particular their mechanical properties as well as their interactions with any surrounding fluid phase or other transport properties alike. We have developed over the last years a fast X-ray diffraction based procedure to determine CSDs with unprecedented precision and have applied this method to the study of ices and gas hydrates (Chaouachi et al. 2017) amongst other examples from material science. Various gas hydrates (GH) were analyzed during formation and coarsening in a sedimentary matrix. The GH crystallites have a size of a few µm when formed (confirmed also by synchrotron X-ray tomography), followed by a coarsening process which mainly takes place at the surface of GH aggregates. Important conclusions can be drawn from the time-dependent analysis of CSDs: (1) Coarsening by normal grain growth proceeds several orders of magnitude slower than in normal ice Ih at similar temperatures; this points to very slow grain boundary migration rates seemingly related to the complexity of topological reconstruction of the crystalline network across a disordered grain boundary. (2) The persisting small crystallites together with their known high resistance against deformation by dislocation motion must lead to grain size sensitive creep, most likely governed by grain boundary sliding. (3) The CSDs of GHs formed in the laboratory appear to have distinctly smaller sizes compared to natural GHs. In consequence, laboratory-based studies of GH can only be safely related to the natural situation once the mutual CSDs are characterized. Starting from crystallites of a few µm in size, obtained under common laboratory conditions from liquid water and gas, it would take hundreds to thousands of years to reach crystal sizes encountered in some NGHs, at least when assuming a normal grain growth mechanism. Anomalous grain growth may intervene and increase crystal sizes, in particular under overburden load and local stresses. Such an anomalous grain growth was indeed observed in natural air hydrates found in Greenland deep ice cores. Finally, we like to mention that our FXR-CSD method is particularly powerful for small crystallites (µm to tens of µm in size) and will allow quantitative studies of in-situ processes occurring during snow coarsening or dynamic recrystallization in ice. M. Chaouachi, S.H. Neher, A. Falenty & W.F. Kuhs (2017) Crystal Growth & Design 17, 2458-2472

Characterizing Key Features in the Formation Mechanisms of Ice

• Peter Kusalik (University of Calgary)

Room: NO Building, Room C 60 Ice crystallization is ubiquitous in nature and has numerous important roles, technological applications and implications. Yet, the ordering processes associated with the nucleation and growth of ice crystals have proven difficult to study directly with experiments, in part due to their stochastic nature of the underlying molecular processes. Consequently, a complete molecular-level picture of the details of ice formation has yet to emerge. Molecular simulations have afforded an excellent opportunity to investigate crystal nucleation and growth of ice at a molecular level, since they are able to probe directly the microscopic environment of a crystal during its nucleation or the interface of a growing crystal. In this presentation I will begin with a brief review of some key issues around simulations of crystallization, considering the attributes and limitation of various models and methods. I will describe some of the approaches that have been developed and utilized for the simulation of the formation and growth of crystals, both in the homogeneous and heterogeneous contexts. While specific results for ice will be a focus, the “simpler” case of gas clathrate hydrates will also be examined and compared. I will present results that will demonstrate that the process of crystallization is characterized by collective phenomena involving many molecules, where the organization can be seen to occur in stages. The nature of the structural topology and fluctuations that characterize a solid/liquid interface will be examined. I will show that defects can play key roles in observed behaviours. The lifetimes and transitions of specific structures will also be briefly discussed. Finally, I will consider how rugged funnel-shaped potential energy landscapes can provide a lens for understanding aspects of the phenomenological differences in the crystal nucleation of ice and gas hydrates.

Increased ice nucleation efficiency of mineral dusts in dilute ammonium sulfate solutions (Nucleation)

• Claudia Marcolli (ETH Zurich, Institute for Atmospheric and Climate Science)

Room: NO Building, Room C 60 Mineral dust has lifetimes of several days in the atmosphere and can be transported over long distances. During transport, mineral dust particles may acquire a coating when they come in contact with reactive gases and semivolatile species or when they undergo cloud processing (e.g. Tang et al., 2016). Coatings can influence the ice nucleation (IN) efficiency of mineral dust in different ways. Water-soluble coatings reduce the critical supersaturation required for cloud droplet activation. Coatings may chemically change the mineral surfaces. Moreover, coatings open up a pathway for immersion freezing below water saturation in concentrated solutions. In many cases, the freezing temperature of mineral dust in solutions can be predicted by the water-activity based IN theory (Zobrist et al., 2008). Here we show examples where this approach to predict freezing temperatures in solutions failed because of specific interactions between the solutes and the ice-nucleating surface. Our focus was on microcline, a K-feldspar known to be highly IN active (e.g. Atkinson et al., 2013). We performed immersion freezing experiments of emulsified droplets with a differential scanning calorimeter. At low concentrations of NH4+-containing salts, namely (NH4)2SO4, NH4HSO4, NH4NO3, and NH4Cl, the ice freezing temperature was increased by up to almost 5 K above the value in pure water, whereas at high concentrations the ice freezing temperature was decreased below the predictions from the water activity-based IN theory for all investigated solutes ((NH4)2SO4, NH4HSO4, NH4NO3, NH4Cl, Na2SO4, H2SO4, K2SO4 and KCl). An increase in IN efficiency was also present in the case of dilute NH3 solutions. An increase of the IN efficiency in very dilute NH3 and NH4+-containing solutions followed by a decrease with increasing concentration was also observed for sanidine (a K-feldspar) and andesine (a Na/Ca-feldspar), as well as for mica and kaolinite. This indicates the presence of specific chemical interactions between solutes and the feldspar surface which are not captured by the water activity-based IN theory. We hypothesize that the hydrogen bonding of NH3 molecules with surface -OH groups could be the reason for the enhanced freezing temperatures in dilute ammonia and ammonium containing solutions as they could form an ice-like overlayer providing hydrogen bonding groups for ice to nucleate on top of it. This enhanced IN efficiency might be of relevance for freezing in condensation mode when ammonium sulfate coatings on mineral dust particles dilute during cloud droplet activation. Atkinson, J. D., Murray, B. J., Woodhouse, M. T., Whale, T. F., Baustian, K. J., Carslaw, K. S., Dobbie, S., O’Sullivan, D., and Malkin, T. L, Nature, 498, 355–358, 2013. Tang, M., Cziczo, D. J., and Grassian, V. H, Chem. Rev., 116, 4205–4259, 2016. Zobrist, B., Marcolli, C., Peter, T., and Koop, T, J. Phys. Chem. A, 112, 3965–3975, 2008.

Deposition Nucleation or Pore Condensation and Freezing? The role of pores on ice nucleation (Nucleation)

• Robert O. David (Institute for Atmospheric and Climate Science, ETH Zürich)

Room: NO Building, Room C 60 Ice crystals in clouds play an important role in initiating precipitation and thus cloud lifetime which acts to moderate the Earth’s radiative balance. As such, understanding the mechanisms responsible for ice formation is necessary to quantify the impacts of anthropogenic climate change. Pore condensation and freezing (PCF), is a proposed mechanism for ice formation below water saturation. During PCF, water condenses in pores or surface imperfections (herein referred to as pores) below water saturation, as predicted by the inverse-Kelvin effect. Thus, the ability of a pore to take up water strongly depends on the pore width and the contact angle of the pore wall. Once liquid water is present, ice can nucleate heterogeneously if an ice active site exists on the pore surface, or homogeneously if the temperature is below 235 K. However, the pore width must also be capable of accommodating the critical ice embryo for ice nucleation to occur. Therefore, the PCF mechanism is limited to particles with a critical pore size narrow enough to fill with water and wide enough to accommodate a critical ice embryo for a given temperature. To investigate the PCF mechanism, spherical mesoporous silica particles with well-defined pore diameters of 3 to 4 nm are evaluated for their ice nucleation ability in the Zurich Ice Nucleation Chamber (ZINC). Within ZINC, the particles are exposed to temperatures ranging between 223 and 238 K and varying supersaturations with respect to ice. To further test the applicability of the PCF mechanism, the particles are functionalized with methyl and hydroxyl groups to alter the contact angle of the particles. We find that the contact angle of the particles helps determine the relative humidity at which pore condensation occurs and subsequently, ice nucleation. These results, combined with experiments performed with nonporous spherical silica, strongly support the proposed PCF mechanism.

Deliquescence and Ice Nucleation at near- and sub- Eutectic Temperatures (Nucleation)

• Erik S. Thomson (University of Gothenburg)

Room: NO Building, Room C 60 The phase state and role of aerosol particles in the radiative budget of the planet are important sources of uncertainty for climate modeling and prediction. The phase behavior of atmospheric particles depends on both environmental conditions and particle properties and can influence surface and bulk processes in both physical and chemical contexts. In the atmosphere soluble particulate is primarily considered to occur in one of two phase states. Either as dry particulate at low relative humidity (RH) conditions or as dissolved solution droplets at higher RH; and the limiting deliquescence RH (DRH) when the soluble material dissolves from one state to another is considered to have sudden discontinuous phase transition like character. A more realistic model of the meta-stable continuous deliquescence of particles has implications for liquid availability at RH smaller than DRH. Furthermore, it raises the question of soluble particles acting as ice nucleating particles in low temperature environments. Here we discuss theoretical developments and experimental observations related to sub-eutectic deliquescence and ice nucleation. Implications for atmospheric processes such as mixed-phase cloud longevity will be discussed. ACKNOWLEDGEMENTS This work was supported by the Swedish Research Council, the Swedish Foundation for International Cooperation in Research and Higher Education (IB2016-6612), and the US National Science Foundation (AGS-1461305, AGS-1461347).

Enhanced heterogeneous ice nucleation by special surface geometry (Nucleation)

• Tianshu Li (George Washington University)

Room: NO Building, Room C 60 Surface roughness has been long known to affect nucleation, but its role in ice formation remains controversial and poorly understood. Experiments showed ice nucleation is significantly promoted by surface irregularities on hematite, mica, and potassium-rich feldspar, but is relatively insensitive to the roughness of superhydrophobic surfaces. By employing large-scale, accelerated molecular simulations, here we show that heterogeneous ice nucleation is strongly affected by surface geometry1. An atomically sharp, concave wedge is found to significantly enhance ice nucleation rate by many orders of magnitude over flat surface. However, in contrast to classical nucleation theory, the enhancement is found to occur only under special surface geometries. These include the 70- and 110-degree wedges that match the two intersecting {111} planes of cubic ice lattice simultaneously, thus significantly enhancing the formation of cubic ice. Remarkably, significant enhancement of ice nucleation also occurs when lattice match does not exist, as in a 45-degree wedge. In this geometry, the 45-degree wedge is found to facilitate the formation of special topological defects of ice that subsequently catalyze the growth of regular ice. Therefore, our study not only highlights the active role of defects in nucleation, but also suggests the traditional concept of lattice match between a nucleation center and crystalline lattice should be extended to include a broader match with metastable, non-crystalline structural motifs. 1. Bi, Y., Cao, B. & Li, T. Enhanced heterogeneous ice nucleation by special surface geometry. Nat Commun 8, 15372 (2017).

Ice Formation on Organic Crystals: Molecular Simulations Tell their Stories (Nucleation)

• Gabriele Cesare Sosso (University of Warwick)

Room: NO Building, Room C 60 The formation of ice is one of the most ubiquitous examples of crystal nucleation and growth, affecting our everyday life as well as technologies such as cryotherapy [1] or fossil fuel extraction [2]. However, pure water freezes only when cooled about 30 K below its melting point [3]. Thus, ice on earth forms mostly heterogeneously, facilitated by substrates which lower the free energy cost needed to nucleate a sufficiently large (or critical) ice nucleus. Experiments can assess whether a specific material is good or bad at promoting ice formation. For instance, both inorganic and biological particles have been found capable of boosting the ice nucleation rate at few degrees only below the freezing point of water [4], with great impact on the formation of clouds in the atmosphere. In particular, recent evidence suggests that some materials such as pyroelectric systems or polar crystals have the potential to induce local electric fields within the water network [5,6], with a spectacular enhancement of the kinetics of crystal nucleation and growth. However, the microscopic details of ice formation on this class of most promising ice nucleating agents remain largely unknown. Here, we have investigated by means of atomistic simulations the formation of ice on two organic crystals: cholesterol (CHL) and metaldehyde (MDHE) - a polar crystal [7]. The low Miller index surfaces of these crystals are characterized by the presence of hydroxyl (-OH, MDHE) and etheric (R-O-R’, MDHE) groups. The presence of these functional groups influences the structure and the dynamics of the water network in proximity of the water-crystal interface, consolidating the hydrogen bond network in different, albeit similarly effective, ways. This evidence directly translates into a huge boost of the ice nucleation rate, which we quantify by means of enhanced sampling (forward flux sampling [8]) simulations that also allow us to unravel the molecular-level details of the nucleation mechanism on both CHL and MDHE These results offer a comprehensive, microscopic picture of ice formation enhanced by organic crystals, paving the way to long-awaited molecular simulations of ice formation in biological matter. While the vast majority of the literature has been focusing on hydroxylated surfaces, we demonstrate the importance of functional groups such as the etheric R-O-R’ groups in the context of heterogeneous ice nucleation. [1] R. Y. Tam, C. N. Rowley et al., J. Am. Chem. Soc. 131, 15745 (2009).
[2] C. A. Koh, Chem. Soc. Rev. 31, 157 (2002).
[3] J. A. Sellberg, C. Huang et al., Nature 510, 381 (2014). [4] B. J. Murray, D. O’ Sullivan et al., Chem. Soc. Rev. 41, 6519 (2012). [5] D. Ehre, E. Lavert et al., Science 327, 672 (2010). [6] A. Belitzky, E. Mishuk et al., J. Phys. Chem. Lett. 7, 43 (2016). [7] N. Fukuta, Nature 199, 475 (1963). [8] R. J. Allen, C. Valeriani et al., Journal of Physics: Condensed Matter 21, 463102 (2009).

The Structural Origins of Dynamical Heterogeneity in Water Pinpoint the Source of Ice-Nucleation (Nucleation)

• Martin Fitzner (Department of Physics and Thomas Young Centre, University College London)

Room: NO Building, Room C 60 The dynamical properties of liquid water play a role of great importance in the context of hydrophobic interactions, where the mobility of water molecules affects e.g. both the thermodynamics and the kinetics of ionic and proton transfer, as well as the mass diffusion of molecular species in aqueous environment. Not least, the ubiquitous phenomena of the glass transition and, as we argue in this work, of ice nucleation are intimately connected to the dynamics of supercooled water. For most atomic and molecular liquids it has been shown that their dynamics are heterogeneous in nature: different regions of the system are characterized by different degrees of mobility, yielding macroscopic effects such as e.g. the breakdown of the Stokes-Einstein relation. This dynamical heterogeneity is at the heart of phenomena of practical relevance such as the formation of ice in clouds, ruling the water cycle on Earth, and low-temperature science including cryobiology. However, obtaining experimental insight into the microscopic details of water dynamics within its supercooled regime is a most challenging task. For instance, how exactly the dynamics of water molecules at low temperatures influences solvation processes is still largely unknown. Most importantly, it is presently unclear whether there exists a specific correlation between the heterogeneous dynamics of supercooled water and the molecular-level details of its structure. In this work [1], we assessed the structural origins of dynamical heterogeneity in supercooled water by means of molecular dynamics simulations, using coarse-grained and fully atomistic water models. We described the emergence of heterogeneous dynamics in terms of domains of slow and fast moving water molecules identified by iso-configurational analysis. We found that slow moving regions are on average less defective in terms of the hydrogen bond network, which translates into more "tetrahedrality" in agreement with the existing literature. Most importantly we could pinpoint a specific hallmark of these slow-moving regions, namely the abundance of five, seven, and most prominently six-membered rings. Interestingly, most of the water molecules belonging to these rings cannot yet be labelled as ice-like per common topological criteria. We argue that the substantial presence of six-membered rings within slow-moving domains has the potential to facilitate ice nucleation. This is because the building blocks of ice (six-membered hydrogen bonded rings of water molecules) are already present in the slow-moving domains, hence a spatial re-arrangement of this topological feature can very well result in the onset of crystallization. This in turn means that a change of dynamical order parameters could precede the structural ordering – a fresh point of view on ordering phenomena such as ice nucleation. Our results provide a new link between two of the most fundamental aspects of water and other tetrahedral liquids. [1] M Fitzner, GC Sosso, SJ Cox and A Michaelides, in preparation

BIRCH TREES - SOURCES OF ICE-NUCLEATING MACROMOLECULES (Ice Binding Proteins)

• Hinrich Grothe (TU Wien)

Room: NO Building, Room C 60 Aerosol-cloud interaction is an important part in understanding the Earth radiation balance and thus global climate. The IPCC indicates the level of understanding of this issue as rather low. This is particularly true for mixed phase clouds and ice clouds. In the recent past a broad variety of aerosols has been test for ice nucleation ability. However, there is still little knowledge about the impact of biological ice nuclei. In particular, biological macromolecules have for a long time been neglected by atmospheric scientists. However, plants are known by biologists to produce macromolecular ice nuclei as a part of their low-temperature survival strategy. In the past, it has been shown by us that birch pollen exhibit ice nucleation active macromolecules at their surface. These molecules can be washed off from the pollen grains and nucleate ice independently. Only recently, we found the same ice nuclei also on secondary and primary wood and on leafs of birch trees. The question remains if these biological ice nuclei can be dispersed through the atmosphere and can impact cloud glaciation processes. Here we apply different spectroscopic and microscopic techniques to gain insights into the nucleation mechanisms and into the chemical signatures of the molecular ice nuclei.

Ice surface chemistry studied by core level spectroscopy

• Hendrik Bluhm (Lawrence Berkeley National Laboratory)

Room: NO Building, Room C 60 Ice surfaces play a major role in many environmental processes and heterogeneous chemical reactions in the atmosphere. The properties of ice surfaces under environmental and atmospheric conditions, in particular the presence of a liquid-like layer at the ice surface at temperatures close to the melting point, are still far from being understood. For instance, reported thicknesses of the liquid-like layer at temperatures close to the melting point vary widely. This might partly be due to the influence of adsorbates, which most likely alter the onset temperature for pre-melting. The interaction of trace gases with ice has also attracted considerable attention over the past decades due its relevance for atmospheric and polar chemistry. Core level spectroscopies such as X-ray photoelectron spectroscopy (XPS) and near-edge X-ray absorption fine structure (NEXAFS) spectroscopy are promising tools to investigate simultaneously the surface chemistry and premelting transition at ice surfaces. While NEXAFS is highly sensitive to small changes in the structural order at the ice surface, XPS is well suited for the characterization of the chemical composition of the ice surface, including the presence of adsorbates and reaction products. In this presentation we will discuss the current status of XPS and NEXAFS measurements of ice surfaces and give an outlook on future opportunities for core-level spectroscopy-based research of the heterogeneous chemistry of ice.

CHARACTERIZATION OF SNOW, FIRN AND ICE

• Ian Baker (Dartmouth College)

Room: NO Building, Room C 60 In this presentation we give an overview of techniques used to characterize the microstructures of snow, firn (multi-year snow) and ice found in both cold regions and in polar ice sheets. These techniques include: transmission electron microscopy, synchrotron-based X-ray topography, cold-stage scanning electron microscopy coupled with energy dispersive X-ray spectroscopy, electron channeling patterns, and electron backscatter patterns; cold stage confocal scanning optical microscopy coupled with Raman spectroscopy; and micro X-ray computed tomography. The capabilities and information obtainable along with the limitations and challenges of each technique will be discussed. Examples of each technique will be presented and future prospects discussed.

What we can learn about dislocations in ice from EBSD analyses (Mechanics & Microstructure)

• Maurine Montagnat (IGE - CNRS - UGA)

Room: NO Building, Room C 60 For some years now, Electron Backscattering Diffraction (EBSD) analysis has been applied to characterise, with high spatial and angular resolution the texture and microstructure of natural ice samples [Obbard et al. J. Glaciol. 2006], or artificially deformed samples [e.g. Piazolo et al. J. Microscopy 2008, Weikusat et al. J. Microscopy 2011]. Contrary to classical optical measurements performed on ice (Rigsby stage, Automatic Ice Texture Analyzer...), EBSD provides the full crystallographic orientation at every analysis point with a spatial resolution down to 0.5 micron. When the resolution and the indexation ratio are large enough, full crystallographic orientations can be used to evaluate the lattice misorientations [Weikusat et al. 2011, Chauve et al. 2017 Phil. Trans. Roy. Soc. A], and therefore access information about Geometrically Necessary Dislocations (GNDs) that remain within the sample after deformation. Piazolo et al. [2008] first used the “boundary trace analysis” to derive the type of dislocations consistent with EBSD data forming a given subgrain boundary. From that, GNDs in laboratory deformed samples appeared to be mostly basal, with an edge characteristic when forming the very common tilt or kink bands, or with a screw or mixed characteristics when forming more continuous substructures. More recently, Piazolo et al. [Acta Mater. 2015] and Chauve et al. [EPSL 2017] applied the Weighted Burgers Vector technique [Wheeler et al. J. Microscopy 2009] to well-indexed, high resolution EBSD maps, providing a more precise characterization of GNDs. The WBV technique derives a 2D equivalent of the Nye tensor from EBSD data collected from the sample surface. Although it does not provide the full misorientation tensor (the Nye tensor), this tool is well adapted to the limitation of the 2D orientation data extracted from EBSD measurements and enables to discriminate between the Burgers vectors of sampled GNDs. Thanks to detailed WBV analyses performed on laboratory deformed ice samples (in compression and torsion), we were able, for the first time, to show that a non-negligible amount of GNDs have a Burgers vector component along the c-axis [Chauve et al. EPSL 2017]. This type of dislocations was rarely evidenced before, and has been assumed to play a minor role in the ductile deformation of ice [Hondoh 2000, Hokkaido Univ. Press]. Although GNDs do not directly relate to the dislocations responsible for local plastic gliding, the observed contribution of up to 30% of non-basal c-component GNDs points to a non-negligible role of such dislocations in ice viscoplastic flow, offering a more complex picture of the key plasticity processes responsible for ice deformation. This presentation will detail the experimental observations and the WBV tool used to evidence the c-component dislocations in the laboratory deformed samples. We will then discuss the importance of such observations for defining constitutive mechanical laws for ice deformation modelling up to the ice sheet scale.

Investigating the Microscopic Location of Trace Elements in Glacier Ice (Mechanics & Microstructure)

• Sven Erik Avak (Laboratory of Environmental Chemistry, Paul Scherrer Institute, Switzerland)

Room: NO Building, Room C 60 Past changes in atmospheric pollution can be reconstructed from high-alpine ice core trace element records (Schwikowski et al., 2004). Eichler et al. (2001) suggested that the preservation of major ions with respect to meltwater percolation depends on their location in the ice crystal lattice. Species predominately segregated at grain surfaces during snow metamorphism were more efficient scavenged by percolating meltwater compared to others incorporated into the ice lattice. Here, we present for the first time a comprehensive study on the microscopic location of anthropogenic, dust and volcanic related trace elements in glacier ice using two different approaches. Moreover, we link the outcome of the study to the tendency of crystalline ice to embed dopants. On the one hand we assessed the microscopic location of trace elements indirectly by analyzing their concentration records in a high-alpine ice core from Grenzgletscher in the southern Swiss Alps using inductively coupled plasma mass spectrometry (ICP-MS). A 13 m w.eq. part of the ice core has been shown to be affected by an inflow of meltwater. Our results demonstrate for the first time that a variety of trace element records, such as Fe, Pb or Cu are not significantly altered by meltwater percolation, indicating a higher solubility of these species in ice. Other trace elements (e.g. Cd, Mn or Zn) are significantly depleted in the affected section of the ice core, presumably due to their initial segregation at grain boundaries. In a second approach we examine the outcome of the indirect analysis by direct spatial analysis of trace elements in Grenzgletscher ice samples. We developed a method based on cryocell laser ablation (LA) ICP-MS, which is the method of choice for the direct in situ chemical analysis of trace elements at a sub-millimeter resolution in glacier ice (Della Lunga et al., 2014, Sneed et al., 2015). The setup is currently tested to analyze concentrations differences of accessible trace elements (e.g. Al, Fe, Pb or Mn) within grains and along grain boundaries in ice core samples from Grenzgletscher. Results of both approaches will be presented. Based on these findings we will discuss possible driving forces causing the observed incorporation of trace elements (such as size and concentration levels) and address how the capability of the ice crystal to form solid solutions responses to size and concentration levels of solutes. This will not only help to evaluate the potential of trace elements as environmental proxies in glaciers partially affected by melting, but also involves the fundamental aspect of how dopants are incorporated into an ice lattice.

The skeletal layer of sea ice: X-ray microtomography and modeling (Ice Crystal Growth)

• Sønke Maus (Norwegian University of Science and Technology (NTNU))

Room: NO Building, Room C 60 When columnar sea ice forms on seawater by unidirectional freezing its interface is known to have a lammelar microstructure, consisting of vertically oriented plates with largely horizontal c-axis orientation, parallel within each grain. Away from the interface these plates thicken by lateral freezing, driven by (i) a decrease in temperature and (ii) intermittent convective exchange of brine against seawater. This convection is the main process that drives the desalination of sea ice, and takes place within a layer of a few centimeter thickness that often is called "skeletal layer" (1,2,3). The skeletal layer is not only relevant for convective desalination, but also shapes an important habitat for life near the sea ice bottom (4) and is critical for pollution issues like oil uptake by sea ice (5). As the skeletal layer has high porosity and is, due to its fragile nature, difficult to sample, relatively little is known about its 3-d structure and transport properties like permeability and thermal conductivity. A conceptual model that involves the bridging of plate-like crystals at some distance from the interface, setting the upper boundary of the skeletal layer, has been suggested half a century ago based on thin-section analysis (1,2), but still remains unvalidated. In the present study first different observations of the sea ice skeletal layer are reviewed, followed by an in-depth analysis of the skeletal layer from 3-d microtomographic imaging (XRT) of natural sea ice and laboratory-grown saltwater ice. The XRT images are analysed with emphasis on the process of necking of lamellar ice plates, changes in grain sizes, and the evolution of pore structure and permeability within the skeletal layer. The determined porosity-permeability relationship is used to model and predict the onset of convection and skeletal layer thickness and compare them to previous work (6,7). The present analysis highlights the importance of crystal structure for sea ice property evolution. [1] D. L. Anderson and W. F. Weeks, Trans. Amer. Geophys. Union, 1958, 39, 632. [2] A. Assur and W. F. Weeks, Int. Association of Scientifc Hydrology, 1963, 63, 95. [3] W. F. Weeks, On Sea Ice, University of Alaska Press, 2010. [4] D. Thomas and G. S. Dieckmann, Sea Ice: An Introduction to its Physics, Chemistry, Biology and Geology, Blackwell, 2003, 402 pp. [5] S. Maus, S. Leisinger, M. Matzl, M. Schneebeli and A. Wiegmann, in Proceedings - Port and Ocean Engineering under Arctic Conditions, Espoo, Finland., POAC, 10 pp. [6] J. S. Wettlaufer, M. G. Worster and H. E. Huppert, Geophys. Res. Letters, 1997, 24, 1251. [7] A. Wells, J. Wettlaufer and S. A. Orszag, Geophys. Res. Lett., 2011, 38.

Microscale distribution of major ions in snow (Mechanics & Microstructure)

• Jürg Trachsel (WSL Institute for Snow and Avalanche Research SLF)

Room: NO Building, Room C 60 Impurity records of an ice core are used as a proxy for past climate change. However, the processes leading to the embedding of chemical compounds are not completely understood: The compounds from the atmosphere are deposited on the ground during snowfall. If the snow does not completely melt, like on polar and alpine glaciers, the impurities will be preserved in the snowpack and later incorporated into the glacier ice. Especially the recrystallization of the snowpack during metamorphism processes can cause a redistribution of embedded compounds. We therefore investigated the transport of major ions of (NH4)2SO4, NaF and CaCl2 in the dynamic snow. In an elution experiment, chemically homogeneous ice droplets were metamorphosed and rinsed with zero-degree water to determine the redistribution of ions. Hereby, we concentrated on the accumulation of ions on the surface of the ice crystals and their inclusion in the ice. This experiment was repeated with natural snow from a field site above Davos, Switzerland. Further, from January to June 2017, we monthly sampled the vertical distribution of major ions in the natural snowpack at the same field site. First results show a strong separation of the ions during snow metamorphism. The concentration of sulfate and calcium on the outside of the ice crystals increased of up to 6 times with storage time. This project is a joint project between the Paul Scherrer Institute (PSI) and the WSL Institute for Snow and Avalanche Research (SLF).

Towards a molecular-level understanding of heterogeneous ice nucleation by direct surface deposition of water vapor

• Jan B. C. Pettersson (University of Gothenburg)

Room: Auditorium Ice nucleation research is currently booming and much of the attention is motivated by the need to improve the description of aerosol and cloud processes in the atmosphere. The formation of clouds has important effects on the water cycle on Earth and on the radiation budget of the atmosphere. The formation of liquid cloud droplets is comparatively well described by existing theory, while the formation of ice particles remains poorly understood. This is a major concern since it introduces uncertainties in our ability to model and project changes in climate. In the atmosphere water droplets tend to remain in a super-cooled state far below the melting point of water with temperatures of about 235 K required to homogeneously freeze pure cloud droplets. Alternatively, ice formation may be enhanced by heterogeneous nucleation that may occur, either when supercooled liquid water contacts an insoluble aerosol particle and subsequently freezes, or when water vapor is directly deposited as ice onto a particle surface. Here we focus on the latter mechanism that is commonly referred to as deposition freezing. This ice nucleation mode primarily contributes to nucleation at temperatures below approximately 243 K, and is more important than other heterogeneous ice nucleation modes at temperatures below 235 K. Deposition freezing is thus of particular relevance for clouds in the upper troposphere. Similar to other heterogeneous ice nucleation modes, deposition freezing remains poorly understood, and parameterizations currently used in cloud models are mainly based on empirical data from laboratory and field studies. Although we know that several different types of solid material may act as deposition freezing ice nuclei, we do not yet have a theoretical description that can be used to predict the outcome. Defects in the substrate, the type of bond that the adsorbing water molecules form with the substrate, and the type of atoms exposed on the substrate all appear to affect heterogeneous nucleation. This paper takes a molecular-level perspective and surface science-based approach to investigating deposition freezing phenomena. We summarize the current understanding of the ice formation process based on recent results obtained with several methods, including environmental molecular beam experiments, molecular dynamics simulations and kinetic modelling. Existing models are critically reviewed and outstanding questions are discussed.

Bilayer-by-Bilayer Surface Melting of Crystalline Ice (The surface and interface of ice)

• Ellen Backus (Max Planck Institute for Polymer Research)

Room: Auditorium Over 150 years ago, Faraday proposed the existence of a liquid-like layer at ice surfaces below the bulk melting temperature. This layer is important for surface chemistry and glacier sliding close to sub-freezing conditions. Since Faraday’s discovery, the properties of this water-like layer have been intensely debated, entailing considerable controversy. The experimentally reported onset temperature for quasi liquid layer (QLL) formation varies between 200 K and 271 K. Moreover, most experimental work shows that with increasing temperature, the QLL thickness gradually and continuously increases from the onset temperature up to the bulk melting point, with reported thicknesses varying from 2 nm to over 45 nm at 271 K. In contrast, early simulations showed that the QLL is formed in a more quantized, bilayer-by-bilayer manner. To elucidate the precise temperature variation of the QLL, and its nature, we investigate the surface melting of ice Ih by combining non-contact, surface-specific vibrational sum frequency generation (SFG) spectroscopy and spectra calculated from molecular dynamics simulations. In our SFG experiment an 800 nm and a 3 μm laser pulse are combined at the interface and the sum-frequency light is detected. Being a second-order nonlinear process, SFG is forbidden in centrosymmetric materials such as the proton disordered ice studied here. At the interface this symmetry is broken, thus allowing us to specifically probe the vibrational response of the interfacial region. The signal is strongly enhanced when the infrared laser pulse is resonant with a molecular vibration. Here we use the O-H stretch vibration of the interfacial water molecules to report on the (molten or non-molten) state of the interface. Using SFG, we probe the outermost water layers of distinct single crystalline ice faces at different temperatures. Macroscopic single-crystalline ice samples are grown from a melt using the seed extraction method. Afterwards the samples are oriented and cut to obtain a specific face. For the basal face, a stepwise, sudden shift in the SFG spectrum to higher frequency occurs around 257 K, which means that the hydrogen-bonded structure of the outermost water layers weaken at this temperature. The spectral calculations from the molecular dynamics simulations reproduce the experimental findings. Moreover, both the experimental and the calculated spectra show only a very weak change in the dangling OH bond. From the combined experimental and simulated surface-specific vibrational spectroscopy, we conclude that the thickness of the quasi liquid layer changes in a non-continuous, stepwise fashion around 257 K. Below this temperature, the first bilayer is already molten; the second bilayer melts at this transition temperature.

In situ observation of ice-water interfaces by advanced optical microscopy (The surface and interface of ice)

• Ken-ichiro Murata (Institute of Low Temperature Science, Hokkaido University)

Room: NO Building, Room C 60 Ice crystallization from supercooled water, more generally, crystallization from its own supercooled melt (the so-called melt growth), is one of the fundamental phase transitions seen everywhere in nature. Despite its ubiquity, the microscopic view of the melt growth is still far from completely understood, contrary to the crystal growth from vapor and solutions. It is well-known that the kinetics of the interface is responsible for the ordering process of crystallization, especially after initial nucleation, because of its strong first order nature. However, it still remains elusive how the liquid and the solid (crystal) phase, of which densities are no more different than several percent, are microscopically distinguished each other at their interface. The main difficulty comes from the significantly high crystal growth rate in melt, hampering direct and precise observations of the interface. Moreover, effects of the latent heat diffusion often obscure the molecular uptake mechanism at the growth front, revealing the microscopic information on the interface. Here we focus on two kinds of ice-water interfaces, ice/quasi-liquid layer (ice/QLL) and ice/bulk water interfaces. We performed in situ observations of these two interfaces by an advanced optical microscopy, whose resolution in the height direction reaches the order of an angstrom. We succeeded in making direct visualization of elementary steps at ice/QLL interfaces and bunching steps at ice/bulk water interfaces. These two exhibit characteristic behaviors different from each other although both systems are the same crystal-melt interface.

Ice growth and air/water interface motion of water droplets impinged on a horizontal cooling surface (The surface and interface of ice)

• Yoshimichi Hagiwara (Kyoto Institute of Technology)

Room: NO Building, Room C 60 The formation of ice layer by water droplets on surfaces causes many troubles, such as poor visibility through the windshields of aircraft, trains and automobiles; the breaking of power transmission lines; a deterioration of the aerodynamic performance of aircraft wings. Thus, many studies have been carried out. However, the details of heat transfer during the freezing have not yet been clarified. We have carried out experiments and numerical simulation to elucidate the detailed mechanism of droplet freezing. In this report, we have focused on time changes in the freezing fronts and air-water interfaces for impinged water droplets on a horizontal cooling surface. In this research, we carried out two-dimensional numerical simulation on the droplets. The time advancements of the air-water interfaces and ice-water interfaces for the droplets were predicted using Phase-field methods. The ice area was expressed with an immersed boundary method. The velocity fields of air and water were predicted by solving the Navier-Stokes equations. The temperature field in the whole domain was predicted by solving the energy equation. The surface tension force and the properties of ice, water and air at around 0 degrees were adopted. The temperatures of air and the cooling surfaces and the droplet impact velocity were varied. The Reynolds number and Weber number of the droplets were in the range of 35 - 129 and 1.6 - 22, respectively. The computational results showed that the height of the impinged droplets on the symmetrical axis started to oscillate as a result of the impact of the collision of droplets with the surfaces in all the cases that we investigated. The oscillations converged rapidly in all impinged water droplets that froze on the cooling surface. This is due to the fact that ice layer, developed along the cooling surface, reached the air-water interface of the droplet on the cooling surface and that the contact area of the droplet was fixed. The ice layer was downward convex during the growth of the layer. This is partly because the water temperature in the central part of droplets was slightly high as a result of the convection and conduction of the latent heat, and partly because the water temperature near the air-water interface was slightly low as a result of low temperature of the air near the cooling surface. When we also took account of the homogeneous ice nucleation, we introduced many ice nuclei into the sub-zero degree region inside the droplets. These nuclei melted by the latent except for the region adjacent to the air-water interface. The predicted oscillation frequency of the air-water interface during the freezing was much lower than the measured frequency, while the maximum amplitude and convergence time of the oscillation were reasonably predicted. This discrepancy was due to the difference in the freedom of oscillation in the two-dimensional computation and the three-dimensional measurements. Thus, our computational results were consistent and reasonable.

Confinement effects in premelting dynamics (The surface and interface of ice)

• John Wettlaufer (Yale University)

Room: NO Building, Room C 60 We examine the effects of confinement on the dynamics of premelted films driven by thermomolecular pressure gradients. Our approach is to modify a well-studied setting in which the thermomolecular pressure gradient is driven by a temperature gradient parallel to an interfacially premelted elastic wall. The modification treats the increase in viscosity associated with the thinning of films, studied in a wide variety of materials, using a power law and we examine the consequent evolution of the confining elastic wall. We treat (1) a range of interactions that are known to underlie interfacial premelting and (2) a constant temperature gradient wherein the thermomolecular pressure gradient is a constant. The difference between the cases with and without the proximity effect arises in the volume flux of premelted liquid. The proximity effect increases the viscosity as the film thickness decreases thereby requiring the thermomolecular pressure driven flux to be accommodated at higher temperatures where the premelted film thickness is the largest. Implications for experiment and observations of frost heave are discussed.

Water Mobility in the Interfacial Quasi-Liquid Layer of Ice/Clay Nanocomposites (The surface and interface of ice)

• Markus Mezger (Max Planck Institute for Polymer Research)

Room: NO Building, Room C 60 At ice/solid interfaces, a quasi-liquid premelting layer (qll) is formed at temperatures below the melting point of bulk water. This qll affects the properties of ice/clay nanocomposites found in ground ice and permafrost. One of the decisive parameters is the water mobility within the qll. Using quasi elastic neutron scattering, the translational diffusion constant of the qll was studied for model systems prepared from clay minerals with large surface to volume ratios. Measurements on a series of charged (vermiculite), hydrophilic (kaolin), and hydrophobic (talc) ice/clay nanocomposites unravel the influence of the solid surfaces. For all composites, the translational diffusion constants within the qll are strongly reduced compared to super cooled bulk water. Depending on their surface properties, significant differences were found for the studied clay minerals. This indicates that beside of confinement effect, intermolecular interactions between the water molecules and the solid surfaces play an important role for the water mobility in the qll.

Snow and Ice Chemistry:  Motivations and Key Questions

• Ted Hullar (University of California at Davis)

Room: NO Building, Room C 60

Fundamentals of acid-base chemistry at and in the quasi-liquid-layer (Chemistry)

• Thorsten Bartels-Rausch (Paul Scherrer Institut)

Room: NO Building, Room C 60 Earth’s surface snow plays an active part in atmospheric chemistry. Research over the past decades has provided an impressive observational basis of the resulting large scale effects, such as substantial modification of the composition and of the chemical reactivity of the lowermost atmosphere in polar regions. Here, I present details on the chemical mechanisms operating in environmental snow and ice derived from well controlled laboratory based experiments. The research is taking full advantage of the Near Ambient Pressure Photoelectron spectroscopy (NAPP) end station at PSI/SLS reveals the ability of atmospheric trace gases to modify the structure of ice at the upper few nanometers at the air-ice interface upon adsorption. Using near ambient pressure core level X-ray photoelectron spectroscopy, we directly detected depth profiles and dissociation degree of adsorbed acidic trace gases at 230 - 255 K at low surface coverage. Complementary X-ray absorption measurements (NEXAFS) show how the presence of acid induces changes to the hydrogen bonding network in the interfacial region. We interpret the data as * a Janus-type character of physisorbed molecular acid at the outermost ice surface and dissociation occurring upon solvation deeper in the interfacial region. * a non-uniform of chemical and physical properties of the hydrogen bonding network along the depth of the QLL Even with a focus of this presentation on adsorption of acidic trace gases (HCl, HNO3, formic acid, acetic acid) to ice and on the molecular structure of the hydrogen bonding network at the air-ice interface, the details on chemistry at extreme concentration and temperature conditions at interfaces might be of high relevance not only in environmental science but also in general chemistry, material science, catalysis, cryobiology, and astrophysics.

Photolysis rate differences of impurities in various frozen matrices, including nature-identical snow crystals (Chemistry)

• Ted Hullar (University of California at Davis)

Room: NO Building, Room C 60 Chemical compounds can be incorporated into snow crystals during formation, for example by co-deposition with water vapor, or after the crystal is formed, such as by vapor deposition onto the surface of the crystal. As fallen snow crystals consolidate in a snowpack, entrained compounds can remain at the ice-air interface of the snowpack (i.e., in the quasi-liquid layer (QLL) of the disordered interface), or be trapped in liquid-like regions (LLRs) at grain boundaries in the ice matrix. More than just a reservoir for these materials, snow can be an important location for a variety of chemical reactions, such as transformations of pollutants. Some recent work has suggested that some photochemical reactions proceed faster in surface QLLs than in either internal LLRs or in aqueous solution. This phenomenon has not been well studied. If photolysis at the air-ice is faster, it might have significant implications for our understanding of chemical fate in snow-covered regions. Laboratory studies have attempted to reproduce the physical reaction environment of snow by a variety of methods, including freezing aqueous solutions in molds, spraying aqueous solutions into liquid nitrogen to form ice grains, or grinding solute-containing ices into small pieces. However, natural snow crystals are quite complex, and while current snow surrogate preparation methods have yielded important insights, they do not accurately mimic the structure and physical properties of newly-fallen natural snow crystals. Here, we describe our work to study photolysis in various frozen matrices, including nature-identical snow samples. We built a snow-making machine, modeled on earlier efforts from other laboratories, which is based on the principle of nucleating supersaturated water vapor to form snow crystals. We incorporate chemicals into/onto the snow crystals either during the formation process or by treating the snow afterwards. Then, we illuminate this snow (or other samples) using the output from an arc lamp, filtered to approximate polar sunlight. We have measured photolysis rates for several compounds in both frozen and aqueous matrices. Our results have found similar photolysis rates in aqueous and frozen solutions for all tested compounds, suggesting chemical reactions in internal LLRs proceed similarly as in aqueous solution. However, reaction rates for some chemicals deposited to snow or ice surfaces are up to 20 times faster than in aqueous solution, while other compounds show no enhancement. We will discuss which compounds and freezing conditions lead to an enhancement.

Modeling Elementary Heterogeneous Atmospheric and Interstellar (Photo)chemical Processes on Ice and their Dynamics using Amorphous Solid Water (Chemistry)

• Patrick Ayotte (Université de Sherbrooke)

Room: NO Building, Room C 60 A thermodynamically reversible path was suggested to exist linking the low density forms of amorphous ice (LDA) and deeply supercooled liquid water (LDL), through the so-called no man’s land and finally onto normal liquid water.(1) Furthermore, at temperatures below its calorimetric glass transition temperature (Tg ~ 136K), transport kinetics are exceedingly slow in amorphous solid water (ASW). Therefore, it might provide a convenient model system to study elementary heterogeneous atmospheric and interstellar chemistry processes that occur on the quasi-liquid layer (QLL) that forms at the air-ice interface in the atmosphere at T<Tm. We will discuss how studying interfacial dynamics at cryogenic temperatures enables the decoupling of processes occurring onto the surface of ASW from those that take place within the bulk by strongly inhibiting the diffusive uptake kinetics. Using this strategy, we will show that ionic dissociation of simple acids [i.e., HF,(2) HCl,(3) HNO3(4)] remain facile down to temperatures as low as 20K at the surface of ASW. We will also demonstrate that heterogeneous nitrates photolysis can be enhanced up to 3-fold at the ASW surface hinting at a significant contribution from heterogeneous (photo)chemistry to the photochemical NOx fluxes that emanate from the sunlit snowpack to polar boundary layer.5 (1) O. Mishima and H.E. Stanley, Nature 396, 329–335 (1998). (2) P. Ayotte, M. Hébert, and P. Marchand, J. Chem. Phys. 123 184501 (2005); R. Iftimie, V. Thomas, S. Plessis, P. Marchand, and P. Ayotte, J. Am. Chem. Soc. 130, 5901-5907 (2008); P. Ayotte, S. Plessis and P. Marchand, Phys. Chem. Chem. Phys. 10, 4785-4792 (2008); P. Ayotte, Z. Rafiei, F. Porzio, and P. Marchand, J. Chem. Phys. 131, 124517 (2009); G. Marcotte and P. Ayotte, J. Chem. Phys. 134, 114522 (2011). (3) P. Ayotte, P. Marchand, J. L. Daschbach, R. S. Smith, and B. D. Kay, J. Phys. Chem. A 115, 6002-6014 (2011). (4) P. Marchand, G. Marcotte, and P. Ayotte, J. Phys. Chem. A 116, 12112-12122 (2012); G. Marcotte, P. Ayotte, A. Bendounan, F. Sirotti, C. Laffon and P. Parent, J. Phys. Chem. Lett. 4, 2643-2648 (2013). (5) Guillaume Marcotte, Patrick Marchand, Stéphanie Pronovost, Patrick Ayotte, Carine Laffon and Philippe Parent, J. Phys. Chem. A 119, 1996-2005 (2015).

Increased Acidity of Trifluoroacetic Acid in Amorphous Solid Water and Charge-Delocalized Hydroniums

• Heon Kang (Seoul National University)

Room: NO Building, Room C 60 We studied the ionic dissociation of trifluoroacetic acid (TFA) in amorphous solid water (ASW) using reflection absorption infrared spectroscopy, low energy sputtering, and H/D isotopic exchange experiment. TFA readily dissociated to hydronium and counter ions in ASW at 60 K, which indicates a significant increase of the acidity as compared to that in aqueous solution at room temperature. The acid dissociation in D2O-ASW produced a Zundel continuum band in the 1000–3000 cm-1 region and an accompanying lowered intensity of the O–D stretching band of D2O. The reduced intensity of D2O was several times larger than that expected for 1:1 stoichiometric proton transfer from TFA to water. Excess protons released from the acid migrated through as many as 20 water layers in ASW. These observations indicate that excess protons are highly mobile and dynamically delocalized in the hydrogen-bonded water chain. Such characteristics of excess protons may be related to the increased acidity of TFA and the appearance of the Zundel continuum absorption in IR spectra.

Elevated mercury and major ion concentrations in snowmelt runoff from a small Arctic Alaska watershed (Fundamentals of the Cryosphere)

• Thomas Douglas (Cold Regions Research and Engineering Laboratory)

Room: NO Building, Room C 60 Mercury is deposited to polar snow and ice surfaces during springtime atmospheric mercury depletion events (AMDEs). AMDEs require sunlight, frozen surfaces, and a reactive bromine source such as sea ice or halogen-rich snow. During these events gaseous elemental mercury (Hg0; GEM) is oxidized to reactive gaseous mercury (HgII; RGM) which can associate with particles to form particulate Hg (PHg). Hg concentrations above 1,000 ng/L have been measured in polar snow and ice, following AMDEs, far greater than that found in lower latitude snow. RGM can be photochemically reduced back to GEM and emitted to the atmosphere or retained by the snowpack and released in spring when snow melts. The majority of polar surface runoff, including AMDE-sourced Hg, occurs during this spring freshet, but the fate of the Hg in the snow is complex, with the Hg depositing in lakes, exporting to the ocean, sorbing to vegetation or soil, or accumulating in soil by microbes. In this study we tracked the fate of snowpack Hg in meltwater percolating through snowpack and snowmelt runoff near Utqiaġvik (formerly Barrow), Alaska. We worked in a small (2.5 ha) watershed near the Arctic Ocean coast where active AMDE chemistry and elevated Hg in snow and ice have been reported. In late winter prior to snowmelt (April) and during snowmelt runoff (May and June) in 2008 and 2009 we made over 10,000 snow depth measurements and 36 snow water equivalent (SWE) measurements in the watershed to calculate the end of winter SWE and Hg load. Gaseous elemental Hg in air and local meteorology were also measured. Snowpack, meltwater, and stream channel water were collected and analyzed for total Hg and major ion concentrations and stable oxygen and hydrogen isotopes. We calculated total snowpack and runoff water Hg fluxes. Results show a snowmelt “ionic pulse” of Hg and major ions during both melt seasons with the total dissolved Hg in meltwater in 2008 at 14.3 mg/ha and 8.1 mg/ha in 2009. These values are 5 to 7 times higher than reported from other Arctic watersheds. Record low summer 2007 precipitation may have led to 50% higher major ion and Hg concentrations during the 2008 melt compared to 2009. We calculated that 78% of snowpack Hg was exported with snowmelt runoff in 2008 and 41% in 2009. From these results we estimate roughly 25% of snowmelt Hg is attributable to AMDEs. We argue a projected warmer future Arctic with more ice sea ice leads, will provide greater halogen sources for AMDEs in the future, and this may increase Hg deposition, reduce Hg re-emission due to higher snow pack halogen concentrations and, ultimately, increase snowpack and meltwater Hg concentrations.

Ice Binding Proteins and Their Interaction With Ice Crystals

• Ido Braslavsky (The Hebrew University of Jerusalem)

Room: NO Building, Room C 60 We investigate the interactions of ice-binding proteins, IBPs, with ice surfaces. In particular, we investigate the dynamic nature of the protein&ice interaction using fluorescence microscopy techniques combined with temperature-controlled microfluidic devices. The results show that binding of IBP to ice is irreversible and that the freezing temperature depression is sensitive to the time allowed for the proteins to accumulate on ice surfaces. This time sensitivity changes dramatically between different types of IBPs. Our results relate the dynamics and level of activity of various types of IBPs to their ability to bind to specific ice orientations, in particular to the basal plane of the ice. These results contribute to the understanding of the mechanisms by which IBPs act that will be critical for the successful use of IBP in cryobiological applications. References: Ice-Binding Proteins and Their Function, M. Bar-Dolev, I. Braslavsky, and P.L. Davies, Ann. Rev. Biochem. 2016 Cryoprotective effect of an ice-binding protein derived from Antarctic bacteria M. Mangiagalli, et al. FEBS 2016 Putting life on ice: bacteria that bind to frozen water, M. Bar-Dolev et al. R. Soc. Interface 2016 Microfluidic Cold-Finger Device for the Investigation of Ice-Binding Proteins, L. Haleva et al. Biophysics J, 2016 When are antifreeze proteins in solution essential for ice growth inhibition? R. Drori et al. Langmuir, 2015 Experimental Correlation between Thermal Hysteresis Activity and the Distance between Antifreeze Proteins on an Ice Surface, R. Drori, P.L. Davies and I. Braslavsky, RSC Adv., 2015 Ice-Binding Proteins that Accumulate on Different Ice Crystal Planes Produces Distinct Thermal Hysteresis Dynamics, Drori, R., et al., R. Soc. Interface 2014. LabVIEW-operated Novel Nanoliter Osmometer for Ice Binding Protein Investigations, I. Braslavsky, and R Drori, Journal of Visualized Experiments 2013. Microfluidic experiments reveal that antifreeze proteins bound to ice crystals suffice to prevent their growth, Y Celik et al. PNAS 2013. New Insights into Ice Growth and Melting Modifications by Antifreeze Proteins, M. Bar-Dolev et al. R. Soc. Interface 2012 Superheating of ice crystals in antifreeze protein solutions, Y Celik et al. PNAS 2010. Supported by the European-Research-Council (ERC), the National-Science-Foundation (NSF), and the Israel-Science-Foundation (ISF).

Freezing-induced pH jump is proved to be an important degradation factor for proteins being frozen (Ice Binding Proteins)

• Dominik Heger (Masaryk Univerity)

Room: NO Building, Room C 60 The freezing of aqueous solutions leads to the separation of ice crystals and the remaining solutes into microscopic veins and pockets; these regions are known as the freeze concentrated solution (FCS). The interactions between the molecules present within this phase are of key importance for the stability or reactivity of compounds in both natural and human-controlled freezing. Pharmaceutic stabilization often involves two steps, namely, freezing and freeze-drying; as each of the procedures can have deleterious effects on active pharmaceutical ingredients (API), we examined them closely. We investigated the stability of proteins in terms of the enzymatic activity after exposure to repeated freezing and thawing. The most significant losses in enzyme structure and functionality were found in very fast cooling or when the apparent pH had changed markedly. Besides the simple freeze-concentration effect, the two main causes of acidity variation in FCS are as follows: a) preferential crystallization of one buffer component, letting the other one interact with API; b) uneven distribution of ions between the ice and FCS to be later neutralized by the flow of protons. The latter effect is also connected to the Workman-Reynolds freezing potential. The two effects are approximately additive, and we demonstrated that the pH jump induced by the former one can be neutralized by the latter. Moreover, we proved experimentally that when a pH jump was avoided, the fully preserved enzyme activity after reconstitution was sustained. Thus, we proposed a new protein stabilization strategy: adding a neutral salt to the buffer to minimize the pH after freezing. The method was named ionic cryoprotection. Next, we studied the effects of freeze-drying on the pH in FCS. The apparent pH change after freezing was substantial, but it was found to be much less pronounced after the lyophilization step. This is explainable by additional crystallization of its components. Thus, we consider misleading to assess the pH stress accompanying the lyophilization cycle solely by monitoring the pH of the lyophiles (as practised previously). The unexpected negative dependence between initial pH of the solution and the acidity of the lyophiles further demonstrates the remaining unpredictability accompanying the freezing process. The pH jump was assessed via the Hammett acidity function by measuring the amount of protonation in sulfonephthalein acid-base indicators. At this stage, sodium and potassium phosphate buffers and several common salts were examined in detail. In conclusion, we show that a freezing-induced pH change can be harmful to the compounds present in FCS but is easily avoidable if the salts are chosen rationally. The good correlation between the Hammett acidity function in the frozen and lyophilized states and the enzyme recovery activity suggests that the indicator method for the pH assessment exhibits sufficient functionality and can be applied also in other research fields dealing with frozen aqueous solutions.

Ice-binding proteins from a sea-ice diatom, their effect on ice growth and physical properties (Ice Binding Proteins)

• Maddalena Bayer-Giraldi (Alfred-Wegener-Institute Bremerhaven (Germany))

Room: NO Building, Room C 60 Ice-binding proteins (IBPs), produced by polar and cold-tolerant organisms, have the ability to bind to ice, affecting its growth. They are key elements in biological adaptation to cold environments, and no other particles, neither natural nor synthetic, show comparable effect in controlling ice growth. However, the details of the protein-ice interactions have not been clarified yet. Different IBP families affect ice in different ways, and the relevant common traits, as well as the differences of the ice binding mechanisms, are still under investigation. Here we present the IBP from the polar sea-ice diatom Fragilariopsis cylindrus (fcIBP). It belongs to a protein family defined by a domain (DUF 3494) extremely successful among polar microorganisms. We studied the effects of the fcIBPs on single crystal free growth using optical bright field and interferometric microscopy. We studied crystal morphology changes and growth rates dependent on supercooling and protein concentration. We saw differential effects of the protein on the growth of crystallographic planes, dependent on fcIBP dynamics in the different directions. We relate this to differential affinity and adsorption time of the protein to the basal and prismatic planes. Switching to a more macroscopic level, we analyzed the effect of the proteins to physical properties of polycrystalline ice. We show the strong inhibition of fcIBPs on grain growth. We observed the evolution of microstructure in fine-grained samples over several weeks, instead of the hours often used for annealing experiments with IBPs. Furthermore, we show that the effect of IBPs on the driving factors for ice deformation during creep, i.e. on internal dislocations due to incorporation within the lattice and on the mobility of grain boundaries due to pinning, make these proteins particularly interesting in studying the process od ice deformation.

Freezing by protein cages (Ice Binding Proteins)

• Alexander Bittner (CIC nanogune)

Room: NO Building, Room C 60 Assembled proteins, esp. cages and viruses, are essential ingredients of life [Calo 2016a]. Some assemblies appear to tolerate the contact to ice, or even the growth of ice. While the interaction of single proteins or simple biopolymers with ice is increasingly investigated [Hiranuma 2015], assemblies are new territory [Wilson 2015]. We pose two questions: 1. Molecular geometry: Is there evidence for the proposal of flat and large surfaces as efficient ice binders or nucleators? This idea [Davies 2014] is intriguing because it puts a purely physical mechanism above the usually assumed biochemical processes, although arguments such as hydrophobicity are also linked to the surface chemistry. 2. Cloud glaciation by biological ice nuclei: Can viruses induce freezing of water? This question is a natural extension from standard ice nucleating particles (mineral dust), over organic matter to "living" objects, which are indeed present in large numbers in sea water [Wilson 2015]. Since the surfaces even of small viruses are less curved and larger than the surface of a single protein, both questions are related. We are answering them by testing the freezing of supercooled water by selected proteins and protein assemblies. Our methods are microscopy, AFM [Calo 2016b], (nano)FTIR [Amenabar 2013] and environmental SEM [Alonso 2013a]; our proteins were the cages ferritin and apoferritin (each with 24 proteins), casein (aggregating to hundreds of nm), and a standard plant virus, Tobacco Mosaic Virus (TMV), which features a helical sheath of 2130 proteins [Alonso 2013b, Calo 2016b]. This virus is hydrophilic, and can be transported in clouds [Castello 1995]. We found a preference for ice growth induced by aggregated (assembled) proteins. Apoferritin and ferritin are especially effective. These assemblies are virus-like cages, constructed from 24 proteins, with a natural tendency to form aggregates (e.g. oligomers). Based on our studies on virus hydration [Alonso2013a], we now demonstrate that ice can grow on viruses without destruction. High-resolution AFM shows surface reconstructions of irregular lateral spacings (ca. 10 nm) that are caused by water adsorption, which is the precursor for ice formation. Literature: [Alonso 2013a] J.M. Alonso, F. Tatti, A. Chuvilin, K. Mam, T. Ondarçuhu, A.M. Bittner, Langmuir 29 (2013) 14580-14587 [Alonso 2013b] J.M. Alonso, M.L. Gorzny, A.M. Bittner, Trends Biotechnol. 31 (2013) 530-538 [Amenabar 2013] I. Amenabar, S. Poly, W. Nuansing, E.H. Hubrich, A. Govyadinov, F. Huth, R. Krutokhvostov, L. Zhang, M. Knez, J. Heberle, A.M. Bittner, R. Hillenbrand, Nature Commun. 4 (2013) 2890 [Calo 2016a] A. Calò, S. Eiben, M. Okuda, A.M. Bittner, Jap. J. Appl. Phys. 55 (2016) 03DA01 [Calo 2016b] A. Calò, A. Eleta-Lopez, P. Stoliar, D. de Sancho, S. Santos, A. Verdaguer, A.M. Bittner, Sci. Rep. 6 (2016) 21899 [Castello 1995] J.D. Castello et al., Phytopathology 85 (1995) 409 [Davies 2014] P.L. Davies, Trends in Biochem. Sci. 39 (2014) 548 [Hiranuma 2015] N. Hiranuma et al., Nature Geoscience 8 (2015) 273 [Wilson 2015] T.W. Wilson et al., Nature 525 (2015) 234

Investigation of Antifreeze Proteins on the Surface of Ice using Vibrational Sum-Frequency Generation Spectroscopy (Ice Binding Proteins)

• Konrad Meister (AMOLF)

Room: NO Building, Room C 60 Antifreeze proteins (AFPs) and Antifreeze Glycoproteins (AFGPs) collectively abbreviated as AF(G)Ps are a unique class of proteins that modify ice crystal growth and thereby ensure the survival of organism in freezing and subfreezing habitats. The molecular working mechanism behind AF(G)Ps freezing inhibition is not well understood, because, as yet, there are no experimental techniques that allow obtaining molecular details on how antifreeze proteins function directly at the surface of ice. Here we use surface specific heterodyne-detected vibrational sum-frequency generation spectroscopy (HD-VSFG) to study the properties of antifreeze proteins directly at the ice surface and at the molecular level.

Modelling mushy sea ice growth and brine convection

• Andrew Wells (University of Oxford)

Room: NO Building, Room C 60 More than 20 million square kilometres of the polar oceans freeze over each year to form sea ice. Sea ice is a mushy layer: a reactive, porous, multiphase material consisting of ice crystals bathed in liquid brine. The porosity of sea ice evolves in space and time, with important consequences for coupling to the oceans and sea ice ecosystems. Atmospheric cooling generates a density gradient in the interstitial brine, which can drive convection and rejection of dense brine from the sea ice to force ocean circulation and mixing. The resulting convective circulation also controls a nutrient supply to support biological activity, including sea-ice algae which grow within the liquid filled pores. This presentation will introduce the fundamentals of the multiphase behaviour of sea ice, with a particular focus on continuum phase-averaged models of sea ice thermodynamics using mushy-layer theory, and the generation of convective flows through the porous ice. I will present results from recent analysis of the transient growth of mushy layers. The cooling conditions provide a key control on the internal structure of the mushy layer during diffusive growth, with two regimes identified. A relatively porous mushy layer forms either for weak cooling, or early during growth when there are slow inefficient cooling rates. In contrast, strong and efficient cooling generates low porosity throughout much of the mushy layer, with a thin high-porosity boundary layer near to the mush-liquid interface. This localisation of porosity impacts the ice permeability, and provides a key control on the onset of convection in the mushy layer. I will conclude with a discussion of the nonlinear dynamics and the implications for the macroscopic ice structure, brine rejection and brine channel formation in growing sea ice.

Gas Nucleation Processes In Sea Ice Brines Inclusions (Fundamentals of the Cryosphere)

• Odile Crabeck (University of manitoba)

Room: NO Building, Room C 60 Past and recent literature have highlighted that sea ice might play a crucial role in controlling and contributing to the exchange of significant climatically active biogases between the ocean and the atmosphere in polar areas. However, the formation of air inclusions and the transport of gases within sea ice cover are still poorly understood. In it is pure form sea ice is a multi-phase system formed by the freezing of seawater, which consists of ice, salt precipitates, liquid brine, and air bubble (WMO, 1970). As the ice growth under cooling condition, each brine inclusion shrinks in size until the internal salinity of the inclusion is high enough to compensate further cooling. In these smallest inclusions, the concentration of salts as well as others solutes (e.g. gases, dissolved inorganic carbon) can increase to a point where the solution is supersaturated (concentration effect). Conversely, if the temperature of the ice cover increases under warming conditions, some of the pure ice along the inclusion walls melts, expanding the brine inclusions and diluting the brine, which lowers the brine salinity. Gases are incorporated in dissolved phase in the brine during ice growth and literature suggest that bubble nucleation further happened in supersaturated brine inclusions. Recent modelling study concluded that the partitioning between dissolved and gaseous phases can play a critical role in gas transport in sea ice, since the transport of gas in air bubble might be decoupled from brine motion. Since there is no analytical method to assess the partitioning of gases between dissolved and gaseous forms in brine, the brine saturation state is commonly used as proxy to estimate the fraction of gas residing in bubble. According to the reasoning therein, bubble formation should increase as temperature decreases due the concentration effect in winter (cooling). However recent experiences using both traditional (crushing and gas chromatography) and new sea ice analytical techniques (CT-Imaging and Raman spectrometry) show that nucleation is not a linear function of the saturation state and that bubble may shrink in size or disappear during cooling. Our results suggest that surface tension and brine internal pressure might control gas nucleation processes in brine. Results from this work are a valuable addition to limited studies available on gas content dynamics in sea ice, and provides new methodology for the interpretation and characterization of sea ice porosity as well as new knowledge for sea ice modelling.

Air cavity migration in ice as a case study for temperature gradient metamorphism of snow: time-lapse X-ray microtomography observation and phase-field simulations. (Fundamentals of the Cryosphere)

• Rémi Granger (Univ. Grenoble Alpes, CNRS, Grenoble INP, 3SR / Météo-France CNRS/ CNRM UMR 3589/ CEN)

Room: NO Building, Room C 60 Temperature gradient metamorphism of snow is a mechanism of snowpack transformation resulting from the coupling of elementary physical processes: heat conduction, vapour transport in the pore space and phase change at the ice/air interface. An important feature of this mechanism is that it produces a faceted microstructure with sufficiently strong gradient, as a consequence of the anisotropy of the kinetic coefficient of ice. To model this mechanism at the pore scale, the phase field method appears to be a relevant choice. Indeed, this method is well suited to describe interface evolution problems coupling phase change to other physics. In the case of snow, Kaempfer and Plapp (2009) developed such a model so that temperature gradient metamorphism could be addressed, with however an isotropic kinetic coefficient. Now, a desirable extension of this model is to add the ability to reproduce faceting. This asks for further model developments. For this, one needs suited experimental data to guide modelling and perform validations. Experiments using real snow are ideal to establish correlations between macroscopic conditions and properties. However, these present variability and complex geometry, which makes interpretation of comparison with microstructural model difficult. Instead, for microstructural modelling, it is practical at first to dispose of experimental data that imply all targeted processes but on a simpler case and, if possible, on which computationally light simulations can be performed. To that purpose, we submitted a monocrystalline ice block containing a spherical air cavity of 190 μm in diameter to an external temperature gradient of 45 K/m, parallel to the c-axis of the ice, with a mean temperature of -4°C using CellDyM cryogenic cell (Calonne et al 2015). We followed the evolution of the cavity using time-lapse X-ray microtomography for 4 days, scanning every 3 hours, with a resolution of about 7.4 μm. The cavity migrates in parallel to the gradient, towards the warm side. The speed of the bubble is constant to 0.00046 μm/s. Hexagonal faceting is observed on the sublimating part of the interface while the condensing part remains rounded. Then, simulation of this problem has been carried out using a COMSOL multiphysics implementation of Kaempfer and Plapp’s model (2009). First, the overall migration of the cavity could be retrieved, underlying the facts that the core physics is captured. Next, as expected, differences in terms of shapes appear clearly so that faceting can be studied. Finally, valuable information on temperature gradient metamorphism get out of these works, and constitutes the first steps for addressing the faceting problem in further modelling investigations. Kaempfer, T. U., & Plapp, M. (2009). Phase-field modeling of dry snow metamorphism. Physical Review E, 79(3), 031502. Calonne, N., Flin, F., Lesaffre, B., Dufour, A., Roulle, J., Puglièse, P., Philip, A., Lahoucine,F., Geindreau, C., Panel, J.M., Rolland du Roscoat, S. & Charrier, P. CellDyM: A room temperature operating cryogenic cell for the dynamic monitoring of snow metamorphism by time-lapse X-ray microtomography. GRL 2015. DOI : 10.1002/2015GL063541

Optical properties of ice and snow

• Stephen Warren (University of Washington)

Room: NO Building, Room C 60 Radiative properties of ice, and of ice-containing media such as snow and clouds, are determined by ice's refractive index and absorption coefficient (“optical constants”). The dominant absorption mechanisms are electronic in the ultraviolet and visible, molecular vibration in the near-infrared, (hindered) rotation in the thermal IR, and lattice translations in the far-IR. The molecular vibrations of water vapor are seen also in liquid water and ice; they are shifted somewhat in frequency by formation of the hydrogen bond. Cubic and hexagonal ice have nearly identical absorption spectra. The spectrum of liquid water is close to that of ice from the UV to the near-IR, but they diverge in the thermal-IR and become very different in the far-IR, microwave, and radiowave regions. The absorption coefficient of ice depends slightly on temperature (~1%/K) in all weakly-absorbing regions from the UV to the microwave. The temperature dependence increases rapidly with wavelength beyond 1 cm. The absorption coefficient varies by ten orders of magnitude from the near-UV (300 nm wavelength) to the near-IR (3 micrometers). The blue and near-UV absorption is so weak, with photon mean-free-path ~1000 m in pure ice, that it is essentially zero for many purposes, but its exact value does matter for computation of photochemical fluxes in snow and of ice thickness on the tropical ocean of “Snowball Earth”. Proceeding across the visible spectrum from blue to red, the absorption length decreases from ~1000 m to 2 m, explaining the blue color of crevasses and icebergs. But snow is white. The reflection of sunlight by snow is the result of successive refraction through small snow grains (~100 micrometers); the short total path length through ice experienced by a solar photon means that nearly all UV and visible photons survive to re-emerge from the snowpack. Indeed, the snow surface of Antarctica has been used as a calibration target for visible channels on satellites. Because the absorption of visible and UV radiation by ice is so weak, the absorption of sunlight at these wavelengths in natural snow is dominated by trace amounts of absorptive impurities such as black carbon (soot) and mineral dust. Although neither clouds nor snow absorb significantly at visible wavelengths, clouds can be detected over snow in reflected sunlight by satellite remote-sensing because they alter the angular reflectance pattern. Half the solar energy is in the near-IR, where ice is sufficiently absorptive that the near-IR flux-reflectance (albedo) becomes sensitive to the area-to-mass ratio (specific surface area, SSA). As snow ages the SSA decreases, darkening the snow. Beyond the solar spectrum, at thermal infrared wavelengths, ice is moderately absorptive, so snow is nearly a blackbody, with emissivity ~99%. But continuing on to longer wavelengths we come to the second region of weak absorption, the microwave and radiowave region (centimeters to meters), so that radiowaves can penetrate several kilometers of ice-sheet depth for sounding of ice thickness.

Serrated flow of snow: Strain-rate softening in uniaxial compression (Mechanics & Microstructure)

• Henning Löwe (WSL Institute for Snow and Avalanche Research SLF)

Room: NO Building, Room C 60 It is well known that the mechanical behavior of snow highly depends on strain-rate as a consequence of the visco-plastic nature of ice. Accordingly, the mechanical response of snow is commonly believed to undergo a single transition from ductile to brittle behavior at a critical strain rate. A recent analysis of snow compression has however revealed the relevance of another time-scale associated to healing effects (e.g. sintering) during compression. The competition of healing and visco-plastic effects gives rise to a rich mechanical behavior in an intermediate range of strain rates. To further elaborate these phenomena we have conducted deformation-controlled compression experiments of low-density snow in a micro-compression stage operated in an X-ray tomography scanner. By varying the imposed axial strain rate over three orders of magnitude, we find a transition regime where the stress response suddenly changes from a smooth to quasi-oscillatory behavior, reminiscent of the Portevin–Le Chatelier effect. This regime of serrated flow is accompanied by compaction bands that nucleate in the interior of the sample as revealed by the tomography analysis. Our experimental results can be well explained within a minimal, scalar model for a rate-dependent, elasto-plastic constitutive law with healing. The model shows a negative slope of the stress vs strain-rate characteristic (strain rate softening) and correctly predicts the range of instability as well as the amplitude and frequency of the serrations. This dynamical instability has an interesting practical consequence, since the results suggest a mechanism how an initially homogeneous snow layer may spontaneously develop a sub-layer structure under deformation.

The Effect of Ice Type on Ice Adhesion (Fundamentals of the Cryosphere)

• Sigrid Rønneberg (Norwegian University of Science and Technology)

Room: NO Building, Room C 60 Ice formation is unavoidable at low temperatures. Ice and frost cause not only inconvenience but also danger in the daily life of human beings, especially in cold regions. For example, ice accumulation on roads or on aircraft wings causes accidents and ice accumulation on wind turbines or overhead power lines can affect the distribution of electricity, power losses and mechanical and electrical failures. To avoid such situations, icing mitigation methods using both anti-icing and deicing systems have been developed over the last decades, with differing efficacy depending on their application field. There are three main pathways to achieve anti-icing surfaces, or icephobic surfaces. These three pathways are the removal of water before freezing, the delay of ice nucleation and the reduction of ice adhesion strength. The ideal icephobic surfaces have an ice adhesion so low that the ice formed on them would shed merely due to its own weight or a natural wind action. To achieve such a low ice adhesion, the fundamental mechanisms of ice adhesion need to be fully, or at least better, understood. This investigation is the first study on the impact of different types of ice on the ice adhesion strength for the same substrate and the same temperature of formation. Ice is likely to behave in different manners depending on the conditions during freezing, and the type of ice is therefore an important factor of ice adhesion strength. The different ice types were formed and tested at the same temperature of -10 oC. The three ice types studied were impact ice generated in a wind tunnel, hard rime ice created in a cold room with water droplets raining from above, and glaze ice frozen directly on to the bars with silicon molds. The ice was frozen on similar aluminum 6061-T6 bars, and ice adhesion strength was measured with the centrifuge adhesion test at the AMIL facilities. A total of 126 tests were performed. The mean ice adhesion strength was measured to 0.78 ± 0.10 MPa for hard rime ice, 0.53 ± 0.12 MPa for impact ice and 0.28 ± 0.08 MPa for glaze ice. The ice adhesion strength for glaze ice is therefore more than 60% lower than for hard rime ice. A significant correlation was found between apparent density of the ice and ice adhesion strength, indicating that the ice adhesion strength decreases when the density of the bulk ice increases. The interface porosity of ice seems also to influence the ice adhesion strength. The results indicate that ice adhesion models need to include a density factor of the ice in additions to the temperature to differentiate between different ice types. If correct, these observations may inspire a new strategy in icephobic surfaces.

The influence of subglacial water on glaciological friction (Fundamentals of the Cryosphere)

• Ian Hewitt (University of Oxford)

Room: NO Building, Room C 60 The movement of continental and mountain-scale ice masses is usually modelled as a problem of viscous creep. The viscosity arises from temperature- and stress-dependent migration of defects through the polycrystalline structure of the ice, and the rheology is usually described as a power law fluid for the purposes of large-scale ice-sheet models. Much of the ice motion arises from effective slip at the bed. This may arise from actual sliding at the ice-bed interface, from enhanced ice deformation close to the bed, or from deformation of the bed istelf (where it comprises deformable glacial sediments). The slip is described mathematically by a glaciological sliding or friction law. This presentation will review classical descriptions of the friction experienced by the flow of glacial ice over a hard bed, will highlight more recent work on the subject, and discuss opportunities for further understanding. The ice-rock interface is typically close to the pressure-dependent melting point and a thin film of water resides between ice and rock. At a small (~mm) scale, there is therefore little resistance to shear, and the macroscopic friction experienced by the glacier is thought to be primarily due to the requirement for ice to flow over or around roughness elements in the bed; effectively a type of form drag. Low stresses downstream of bumps in the bed can result in cavitation, with the ice peeling off from the bed to leave a water-filled cavity. Such cavitation on a large scale effectively reduces the ice-bed contact area and can result in a reduction of friction. We discuss conceptual models of this process and their implications in terms of the dependence of friction on ice speed, cavity pressure and volume. A crucial and unresolved question is the extent to which such cavities are hydraulically connected and can therefore evolve in response to meltwater that enters glaciers from the surface during the melt season. We will discuss observations of how glacier sliding varies in time, and hypotheses for the small-scale dynamics at the bed that give rise to these changes. We also discuss the role of bed deformation and the thermodynamic effects of regelation and viscous dissipation.

A statistical fracture model for Antarctic ice shelf regions (Fundamentals of the Cryosphere)

• Veronika Emetc (Research School of Earth Sciences, Australian National University)

Room: NO Building, Room C 60 Creating an effective model of the calving process in Antarctica is a challenging task in ice sheet modelling. The recent calving of Larsen B and Larsen C ice shelves has raised a number of questions as to the nature of the key factors that triggered these calving events and whether they are caused by natural ice sheet dynamics or climate change. To address this, a reliable calving model is required and previous studies and models have not been able to be utilised across the Antarctic continent. In fact, a comprehensive calving model needs to include components such as information about surface and basal crevasses (location and depth), advection of the crevasses downstream, the intersection of the surface and basal crevasses as well as estimation of the critical depth of a crevasse when calving occurs. The purpose of this research is to use observational data to construct a fracture model that can be applied to any selected ice shelf/glacier in Antarctica. The main aspects explored are the determination of the location of both surface and basal fractures as well as propagation of surface crevasses when meltwater is added. First, the basis of this project was to create a statistical model that can improve the predictability of the location of surface fractures. We used a logistic regression algorithm and a set of predictors taken from observations of glaciers/ice shelves. The modelled probability function varies from zero to one and in this way is similar to the previously proposed damage function. It can predict fractures for grounded ice as well as for floating ice shelves. Second, we used Linear Elastic Fracture Mechanics as a basis for modelling the depth of fractures. Then, based on the misfit between observed and modelled fracture depth we were able to perform an inversion for melt rate and, thus, obtain a depth of crevasses in better agreement with observations. Lastly, we explored the correlation between surface troughs and basal crevasses using altimetry data for over 45 glacier regions. Our findings confirm the previously suggested theory that these two processes are highly correlated. We use this results together with statistical algorithms to predict the location of basal crevasses. We found that statistical methods can provide valuable information for predicting zones of fractures. Our model shows a good agreement with the data about fractures taken from satellite imagery for any selected ice shelf in Antarctica. Furthermore, we found that different conditions of fracturing require different water levels inside crevasses in order to keep them open. Our inversion allows for predicting the depth of crevasses in a better agreement with the observed depth of crevasses obtained from radar data sets. The combination of these methods provides a good basis for future calving parameterisation.

Ice particle detection in clouds using a high-speed Particle Phase Discriminator (Fundamentals of the Cryosphere)

• Jörg Wieder (Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, 8092, Switzerland)

Room: NO Building, Room C 60 Mixed-phase clouds (MPCs), composed of both liquid droplets and ice crystals, play a crucial role in the global radiation budget. Yet, a reliable detection of liquid and ice cloud fraction remains difficult. This in turn hinders a complete understanding of the complex microphysical processes that occur within these clouds, and ultimately renders estimation of the radiative properties of MPCs difficult. Current devices are either only capable of counting the number of particles of different sizes without discriminating phase or cannot resolve to small enough scales (Baumgardner et al, 2011). However, detection of ice particle fraction down to particle sizes at the micrometer scale is critical. Here, we present a new instrument, the high-speed Particle Phase Discriminator (PPD-HS), significantly reducing shortcomings of earlier devices. PPD-HS sizes particles by passing a continuous laser beam (similarly to commercial Optical Particle Counters) and discriminates their phase by analyzing scattering patterns of forward scattered light. Compared to PPD-2K presented by Vochezer et al. (2016), our new instrument reduces the information collected per particle, used for phase discrimination, and thus allows for an increased detection rate of cloud particles. Phase discrimination is achieved by evaluating the spatial symmetry of the scattering pattern, allowing for classification of spherical cloud droplets and aspherical ice crystals. Here we present an in-depth analysis of instrumental performance. The phase discrimination potential of PPD-HS is determined through a suite of laboratory experiments, using particles with well-controlled properties. These include spherical and aspherical particles covering a size range from approximately 1 to 15 μm. Our results show that PPD-HS detects particles larger than approximately 1 µm in diameter and reliably discriminates ice crystals from other cloud particles for particle size > 3 µm, providing the capabilities to detect freshly nucleated ice in MPCs. In a second round of experiments, we coupled PPD-HS to the Horizontal Ice Nucleation Chamber (HINC), a Continuous Flow Diffusion Chamber (CFDC). HINC allows us to mimic formation of real cloud hydrometeors and assess the performance of PPD-HS to determine the hydrometeor phase under atmospherically relevant MPC conditions. Results from such experiments will be presented to assess the performance of PPD-HS in discriminating accurately the liquid and ice fractions formed in HINC. These will help us to evaluate whether PPD-HS could improve our ability to identify ice and water cloud fractions in the real atmosphere.