114
5 Nucleation of Gas Hydrates
5.1.3 Nucleation of Ice from Bulk Liquid Water
Given that liquid water already possesses important structural features of ice (the
coordination number of H 2 O only changes from about 4.0 to about 4.4 when ice melts
[5]), the deep subcoolings observed for liquid water are surprising. Water droplets of
5 µm radius emulsified in an oil using non-nucleating surfactants appear to subcool
to low temperatures similar to those found for similarly sized water droplets in cloud
chambers [3]. Just as the freezing point of water can be lowered by an application of
hydrostatic pressures, the temperature at which the nucleation probability becomes
high is also lowered by an application of hydrostatic pressures [3]. The deep subcoolings of water can therefore be further increased by increasing pressures [5], from
approximately −40 °C under the atmospheric pressure to −92 °C under 0.2 GPa [3].
Silver iodide (AgI) has lattice constants that closely match those of ice to within
a few percent [9]. Because of this very good lattice matching with ice, AgI has
long been regarded as an excellent nucleation promotor of ice. After decades of
commercial cloud seeding with AgI, however, the evidence that AgI really enhances
rainfall still remains inconclusive [9]. Perhaps surprisingly, ice grows as discrete
hexagonal islands on an AgI substrate, as opposed to a uniform film [9] which might
be expected of an epitaxial growth of a crystal on a lattice-matching substrate. This
rather apparent contradiction is not unique to AgI; BaF 2 is not an effective nucleating
agent despite its good lattice matching to ice [9].
It appears that the orientation of the hydrogen bonds (dipoles) of ice is an important
factor in the heterogeneous nucleation potency of an underlying substrate [4]. In
short, any substrate that orients dipoles at the surface of ice parallel to one another
appears a poor nucleation promoter, because it reduces the entropy and raises the
free energy of any nuclei growing on the substrate [4]. Then the basal crystal faces of
AgI or PbI would be poor ice nucleation agents and their activity is likely confined to
the prism faces [4]. A molecular dynamics simulation study found that structurally
identical substrates could both inhibit and promote ice formation, depending on the
interaction between the substrate surface and H 2 O molecules [10]. These points are
all highly relevant to nucleation of clathrate hydrates. For example, is AgI truly a
good nucleation agent of ice and is ice truly a good nucleation agent of clathrate
hydrates? At temperatures ice can be present, the subcoolings are already deep for
clathrate hydrate nucleation. Would the resulting large driving force be sufficient for
instant nucleation of clathrate hydrate, with or without the presence of ice? If so,
how can one tell whether ice is truly a good nucleation promotor of clathrate hydrate
or not? These points will be revisited in the next section.
In contrast, water-insoluble organic compounds with no structural similarities to
ice, such as steroids and cholesterols, have been found to effectively nucleate ice [11].
Formation of hexagonal ice crystals on cholesterols that have no lattice matching to
ice resulted in surprisingly small subcoolings of down to 1 K [12]. Testosterone (a
steroid) with no lattice matching to ice was also found to be an excellent nucleating
agent [9]. Bacteria Pseudomonas Syringaed were also found to greatly promote
nucleation of ice [9]. These are very surprising findings, to say the least, and more
5 Nucleation of Gas Hydrates
5.1.3 Nucleation of Ice from Bulk Liquid Water
Given that liquid water already possesses important structural features of ice (the
coordination number of H 2 O only changes from about 4.0 to about 4.4 when ice melts
[5]), the deep subcoolings observed for liquid water are surprising. Water droplets of
5 µm radius emulsified in an oil using non-nucleating surfactants appear to subcool
to low temperatures similar to those found for similarly sized water droplets in cloud
chambers [3]. Just as the freezing point of water can be lowered by an application of
hydrostatic pressures, the temperature at which the nucleation probability becomes
high is also lowered by an application of hydrostatic pressures [3]. The deep subcoolings of water can therefore be further increased by increasing pressures [5], from
approximately −40 °C under the atmospheric pressure to −92 °C under 0.2 GPa [3].
Silver iodide (AgI) has lattice constants that closely match those of ice to within
a few percent [9]. Because of this very good lattice matching with ice, AgI has
long been regarded as an excellent nucleation promotor of ice. After decades of
commercial cloud seeding with AgI, however, the evidence that AgI really enhances
rainfall still remains inconclusive [9]. Perhaps surprisingly, ice grows as discrete
hexagonal islands on an AgI substrate, as opposed to a uniform film [9] which might
be expected of an epitaxial growth of a crystal on a lattice-matching substrate. This
rather apparent contradiction is not unique to AgI; BaF 2 is not an effective nucleating
agent despite its good lattice matching to ice [9].
It appears that the orientation of the hydrogen bonds (dipoles) of ice is an important
factor in the heterogeneous nucleation potency of an underlying substrate [4]. In
short, any substrate that orients dipoles at the surface of ice parallel to one another
appears a poor nucleation promoter, because it reduces the entropy and raises the
free energy of any nuclei growing on the substrate [4]. Then the basal crystal faces of
AgI or PbI would be poor ice nucleation agents and their activity is likely confined to
the prism faces [4]. A molecular dynamics simulation study found that structurally
identical substrates could both inhibit and promote ice formation, depending on the
interaction between the substrate surface and H 2 O molecules [10]. These points are
all highly relevant to nucleation of clathrate hydrates. For example, is AgI truly a
good nucleation agent of ice and is ice truly a good nucleation agent of clathrate
hydrates? At temperatures ice can be present, the subcoolings are already deep for
clathrate hydrate nucleation. Would the resulting large driving force be sufficient for
instant nucleation of clathrate hydrate, with or without the presence of ice? If so,
how can one tell whether ice is truly a good nucleation promotor of clathrate hydrate
or not? These points will be revisited in the next section.
In contrast, water-insoluble organic compounds with no structural similarities to
ice, such as steroids and cholesterols, have been found to effectively nucleate ice [11].
Formation of hexagonal ice crystals on cholesterols that have no lattice matching to
ice resulted in surprisingly small subcoolings of down to 1 K [12]. Testosterone (a
steroid) with no lattice matching to ice was also found to be an excellent nucleating
agent [9]. Bacteria Pseudomonas Syringaed were also found to greatly promote
nucleation of ice [9]. These are very surprising findings, to say the least, and more