So as to "kick off" the discussion here, I'll just summarise a few of the emails exchanged about ice nucleation so far.
Bruce Moffett asked about the Bergeron-Findeisen process in mixed-pahse clouds. Zev Levin explained that it involves air that has both liquid and ice present tending to have a humidity close to water saturation, so that there is great supersaturation with respect to ice. The latter causes rapid growth of ice crystals by diffusion of vapour, and any subsaturation with respect to water causes evaporation of the liquid.
Zev comments on the difficulty of identifying the IN in any given heterogeneous crystal sampled from the real atmosphere. How does one know that the IN particle(s) found when the ice of the crystal is evaporated away was the cause of the particle's heterogeneous ice nucleation ? He suggests inspecting the change in freezing temperature caused by heating the residual particles to try to infer whether any were biogenic.
Bruce Moffett asked about whether the requirement for IN to be larger than a critical size applies to both contact and bulk-liquid modes. Viruses are small, so this is an issue.
I replied that the critical size (0.1 microns) probably does apply to both contact and immersion/condensation-freezing modes, because Raymond Shaw observed a constant "temperature shift" between contact and bulk-liquid freezing modes of only 4 to 5 degC (Shaw et al. 2005), a paper I emailed to everyone. I also emailed to everyone my own JAS (2008) paper in which I cite Vrbka's molecular dynamcis paper, which indicates that the liquid surface has special dynamics that favour ice nucleation.
This critical size is apparent from field studies (Chen et al. 1998; Prenni et al. 2007) in which the insoluble particles at the centres of heterogeneously nucelated ice particles have been analysed. Such observations are summarised by Phillips et al. (2008, JAS).
Zev pointed out that he has observed a smaller shift than this for bacteria in his paper fromthe 1980s with Yankovsky (1983): a shift of 2 degC. Zev also commented that contact-freezing in natural clouds is inhibited by their size when greater than 0.1 microns, owing to other phoretic forces being more important than Brownian motion. I amplified this assertion with results from my cloud model in which both outside-in and inside-out contact-freezing are less important than the other modes of heterogeneous ice nucleation. However, inside-out contact-freezing generates many more crystals than outside-in contact-freezing, when I do the number budget of crystals initiated for tropical deep convection. Thermophoresis (sueprsaturaation-dependent) and the transience of regions of strong evaporation are the reasons for outside-in contact-freezing being so infrequent.
Bruce pointed out the absurdity of a constant temperature shift for all conditions, which would imply a freezing temperature greater than 0 degC for some very active bacteria.
Zev warned that the temperature shift seen by Shaw cannot be universal for all types of IN and all temperatures, and that Shaw's experiments were done with unrealistically large IN particles (>> 10 microns).
I also agreed that it is not universal, hypothesizing that any such temperature shift must diminish as the freezing temperature approaches 0 degC. This would explain why Zev observed a smaller temperature shift for bacteria. But equally, Shaw's more recent unpublished studies show no dependence of his observed temperature shift (4 to 5 K) on composition and size of the IN particle. This supports that notion that Shaw's results may apply to natural IN particles but perhaps only at the colder freezing temperatures he studied (e.g. about -20 degC).
So, for bacteria their contact-freezing is probably at a freezing temperature about 2 or 3 K warmer than in the bulk-liquid modes (immersion-/condensation-freezing).
Finally, Paul DeMott has estimated that the active IN concentration from biological particles (e.g. bacteria) is about 0.01 per Litre in the atmosphere at -10 degc. This he inferred from the data in Brent Christner's Science paper about the IN activity of particles in frozen rime collected at the Storm Peak Lab.
I would note here that Paul's estimate is broadly consistent with my estimate of biological IN contributing at least a few percent to the total number of active IN at any temperature from Phillips et al. (2008, JAS) (total number of all active IN is about 0.1 - 1 per Litre at -10 degC inferred from the measurements at Storm Peak of 6 per Litre at -30 degC).
cheers,
Vaughan Phillips.[justify]
Bruce Moffett asked about the Bergeron-Findeisen process in mixed-pahse clouds. Zev Levin explained that it involves air that has both liquid and ice present tending to have a humidity close to water saturation, so that there is great supersaturation with respect to ice. The latter causes rapid growth of ice crystals by diffusion of vapour, and any subsaturation with respect to water causes evaporation of the liquid.
Zev comments on the difficulty of identifying the IN in any given heterogeneous crystal sampled from the real atmosphere. How does one know that the IN particle(s) found when the ice of the crystal is evaporated away was the cause of the particle's heterogeneous ice nucleation ? He suggests inspecting the change in freezing temperature caused by heating the residual particles to try to infer whether any were biogenic.
Bruce Moffett asked about whether the requirement for IN to be larger than a critical size applies to both contact and bulk-liquid modes. Viruses are small, so this is an issue.
I replied that the critical size (0.1 microns) probably does apply to both contact and immersion/condensation-freezing modes, because Raymond Shaw observed a constant "temperature shift" between contact and bulk-liquid freezing modes of only 4 to 5 degC (Shaw et al. 2005), a paper I emailed to everyone. I also emailed to everyone my own JAS (2008) paper in which I cite Vrbka's molecular dynamcis paper, which indicates that the liquid surface has special dynamics that favour ice nucleation.
This critical size is apparent from field studies (Chen et al. 1998; Prenni et al. 2007) in which the insoluble particles at the centres of heterogeneously nucelated ice particles have been analysed. Such observations are summarised by Phillips et al. (2008, JAS).
Zev pointed out that he has observed a smaller shift than this for bacteria in his paper fromthe 1980s with Yankovsky (1983): a shift of 2 degC. Zev also commented that contact-freezing in natural clouds is inhibited by their size when greater than 0.1 microns, owing to other phoretic forces being more important than Brownian motion. I amplified this assertion with results from my cloud model in which both outside-in and inside-out contact-freezing are less important than the other modes of heterogeneous ice nucleation. However, inside-out contact-freezing generates many more crystals than outside-in contact-freezing, when I do the number budget of crystals initiated for tropical deep convection. Thermophoresis (sueprsaturaation-dependent) and the transience of regions of strong evaporation are the reasons for outside-in contact-freezing being so infrequent.
Bruce pointed out the absurdity of a constant temperature shift for all conditions, which would imply a freezing temperature greater than 0 degC for some very active bacteria.
Zev warned that the temperature shift seen by Shaw cannot be universal for all types of IN and all temperatures, and that Shaw's experiments were done with unrealistically large IN particles (>> 10 microns).
I also agreed that it is not universal, hypothesizing that any such temperature shift must diminish as the freezing temperature approaches 0 degC. This would explain why Zev observed a smaller temperature shift for bacteria. But equally, Shaw's more recent unpublished studies show no dependence of his observed temperature shift (4 to 5 K) on composition and size of the IN particle. This supports that notion that Shaw's results may apply to natural IN particles but perhaps only at the colder freezing temperatures he studied (e.g. about -20 degC).
So, for bacteria their contact-freezing is probably at a freezing temperature about 2 or 3 K warmer than in the bulk-liquid modes (immersion-/condensation-freezing).
Finally, Paul DeMott has estimated that the active IN concentration from biological particles (e.g. bacteria) is about 0.01 per Litre in the atmosphere at -10 degc. This he inferred from the data in Brent Christner's Science paper about the IN activity of particles in frozen rime collected at the Storm Peak Lab.
I would note here that Paul's estimate is broadly consistent with my estimate of biological IN contributing at least a few percent to the total number of active IN at any temperature from Phillips et al. (2008, JAS) (total number of all active IN is about 0.1 - 1 per Litre at -10 degC inferred from the measurements at Storm Peak of 6 per Litre at -30 degC).
cheers,
Vaughan Phillips.[justify]