Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012;7(9):e44382.
doi: 10.1371/journal.pone.0044382. Epub 2012 Sep 26.

Molecular Probe Dynamics Reveals Suppression of Ice-Like Regions in Strongly Confined Supercooled Water

Free PMC article

Molecular Probe Dynamics Reveals Suppression of Ice-Like Regions in Strongly Confined Supercooled Water

Debamalya Banerjee et al. PLoS One. .
Free PMC article


The structure of the hydrogen bond network is a key element for understanding water's thermodynamic and kinetic anomalies. While ambient water is strongly believed to be a uniform, continuous hydrogen-bonded liquid, there is growing consensus that supercooled water is better described in terms of distinct domains with either a low-density ice-like structure or a high-density disordered one. We evidenced two distinct rotational mobilities of probe molecules in interstitial supercooled water of polycrystalline ice [Banerjee D, et al. (2009) ESR evidence for 2 coexisting liquid phases in deeply supercooled bulk water. Proc Natl Acad Sci USA 106: 11448-11453]. Here we show that, by increasing the confinement of interstitial water, the mobility of probe molecules, surprisingly, increases. We argue that loose confinement allows the presence of ice-like regions in supercooled water, whereas a tighter confinement yields the suppression of this ordered fraction and leads to higher fluidity. Compelling evidence of the presence of ice-like regions is provided by the probe orientational entropy barrier which is set, through hydrogen bonding, by the configuration of the surrounding water molecules and yields a direct measure of the configurational entropy of the same. We find that, under loose confinement of supercooled water, the entropy barrier surmounted by the slower probe fraction exceeds that of equilibrium water by the melting entropy of ice, whereas no increase of the barrier is observed under stronger confinement. The lower limit of metastability of supercooled water is discussed.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.


Figure 1
Figure 1. Two idealized ice/water mixtures with different polycrystallinity at .
The scale of the pictures is the same. The two panels refer to the QRW (a) and the SC protocols (b), see text for details. The width of the liquid veins in the two mixtures is very similar and of the order of dozens of nanometers, whereas the size of the ice grains is formula image times larger in SC mixtures. Augmenting of the polycrystallinity increases the water fraction formula image and reduces its confinement due to the additional paths and intersections. According to ref. and the present study, ice-like patches (blue) with slow (S) mobility are included in the QRW liquid fraction. The patches are suppressed in the SC mixtures, leaving only the less ordered liquid fraction (light blue) with fast (F) mobility. The shape of the patches is unknown.
Figure 2
Figure 2. Structure and selected ESR lineshapes of the spin probe TEMPOL in water.
Left: slowly cooled bulk water (SC protocol). Right: quenched and slowly re-heated bulk water at the indicated temperature (QRW protocol). The QRW sample contains ice with higher polycrystallinity. Note that the SC sample at 210 K exhibits narrower lines than QRW sample at formula image K, i.e. TEMPOL is rotating faster in SC water. The H-bonding of TEMPOL with water is shown in the top of the the right panel. Owing to the very weak ESR signal from TEMPOL in the SC sample at formula image K, a small spurious signal from the quartz capillary used is observed at formula image G.
Figure 3
Figure 3. Rotational correlation time of TEMPOL in SC and QRW water.
Part of the data are in the “no man's land” (formula image). The two parallel dashed lines with slope formula image kJ/mol are the Arrhenius best-fit of the correlation times of TEMPOL in equilibrium water, SC water (blue) and in the low-mobility S fraction of the QRW water (red). The inset plots the data including the sub-formula image region. Note: i) the change of regime at formula image K close to formula image, ii) the absence of any abrupt change at formula image and, in SC water, at both formula image and formula image.
Figure 4
Figure 4. The entropy barrier to TEMPOL reorientation rising in supercooled QRW water.
The dashed lines mark the entropy of melting formula image J Kformula image molformula image (black) and the best-fit value of the barrier of the slow fraction of TEMPOL formula image J Kformula image molformula image (red). The solid black line is a measure of the number of water configurations lost on cooling from formula image to formula image, formula image, where formula image is the excess entropy of the liquid over the crystal . The blue circle corresponds to formula image J Kformula image molformula image .

Similar articles

See all similar articles

Cited by 3 articles


    1. Stillinger FH (1980) Water Revisited. Science 209: 451–457. - PubMed
    1. Mishima O, Stanley HE (1998) The relation between liquid, supercooled and glassy water. Nature 396: 329–335.
    1. Debenedetti PG (2003) Supercooled and glassy water. J. Phys.: Condens. Matter 15: R1669–R1726.
    1. Angell CA (2008) Insights into Phases of Liquid Water from Study of Its Unusual Glass-Forming Properties. Science 319: 582–587. - PubMed
    1. Soper AK (2008) Structural transformations in amorphous ice and supercooled water and their relevance to the phase diagram of water. Mol Phys 106: 2053–2076.

Publication types

Grant support

The work was financially supported by the Indian National Science Academy and the University of Pisa. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.