Kara Lamb¹, Maximilien Bolot¹, Benjamin Clouser¹, Laszlo Sarkozy², Steven Wagner³, Volker Ebert⁶, Erik Kerstel⁴, Harald Saathoff⁵, Ottmar Möhler⁵, Elisabeth Moyer²

1. Department of Physics, University of Chicago

2. Department of the Geophysical Sciences, University of Chicago

3. Center of Smart Interfaces, Technische Universität Darmstadt, Darmstadt, Germany

4. J. Fourier Laboratoire Interdisciplinaire de Physique, University of Grenoble, Grenoble, France

5. Institute for Meterology and Climate Research, KIT, Karlsruhe, Germany

6. Physikalisch-Technische Bundesanstalt, Braunschweig, Germany

The isotopic composition of water vapor is widely used in atmospheric science as a tracer for convective and microphysical processes. However, isotopic fractionation factors between water vapor and ice have not previously been directly measured at the temperatures characteristic of much of the upper troposphere (< 235 K). Fractionation factor values used in models are extrapolated from measurements made at warmer temperatures (Merlivat and Nief, 1967, which has recently been called into question by Ellehøj et al, 2013). We report here on results from the 2012-2013 IsoCloud campaigns at the AIDA Aerosol and Cloud Chamber, which investigated isotopic fractionation in atmospherically relevant conditions by simulating cirrus clouds at temperatures and pressures characteristic of the upper troposphere. The IsoCloud campaign involved a series of adiabatic expansion experiments at temperatures between 190 and 233 K, during which multiple in-situ and extractive instruments measured water vapor, total water, and water vapor isotopic composition, as well as ice particle number density. We use measurements from the Chicago Water Isotope Spectrometer (ChiWIS) to determine the degree of preferential condensation of HDO over H₂O over this temperature range. Results are consistent with the extrapolated relationship of Merlivat and Nief in 1967, and cannot be reconciled with Ellehøj et al, 2013. Ice growth experiments at a variety of supersaturations also allow us to make the first direct tests of kinetic isotope effects resulting from diffusion-limited vapor deposition at these temperatures. IsoCloud experiments confirm that isotopic fractionation is reduced at high supersaturations, with the degree of reduction highly consistent with the model suggested by Jouzel and Merlivat,1984. We conclude that the isotopic physics of ice formation at cold temperatures are well-understood. These results are consequential for atmospheric science: we demonstrate that the range of proposed values of the fractionation factor and the magnitude of kinetic effects would strongly affect the interpretation of data from aircraft and satellite measurements in the upper troposphere / lower stratosphere.