Structural Transformations in Amorphous Ice and Supercooled Water and Their Relevance to the Phase Diagram of Water Alan K Soper
Total Page:16
File Type:pdf, Size:1020Kb
Structural transformations in amorphous ice and supercooled water and their relevance to the phase diagram of water Alan K Soper To cite this version: Alan K Soper. Structural transformations in amorphous ice and supercooled water and their relevance to the phase diagram of water. Molecular Physics, Taylor & Francis, 2008, 106 (16-18), pp.2053-2076. 10.1080/00268970802116146. hal-00513204 HAL Id: hal-00513204 https://hal.archives-ouvertes.fr/hal-00513204 Submitted on 1 Sep 2010 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Molecular Physics For Peer Review Only Structural transformations in amorphous ice and supercooled water and their relevance to the phase diagram of water Journal: Molecular Physics Manuscript ID: TMPH-2008-0088 Manuscript Type: Invited Article Date Submitted by the 17-Mar-2008 Author: Complete List of Authors: Soper, Alan supercooled water, amorphous ice, structure, second critical point, Keywords: computer simulation URL: http://mc.manuscriptcentral.com/tandf/tmph Page 1 of 36 Molecular Physics 1 2 Structural transformations in amorphous ice and supercooled water and their 3 relevance to the phase diagram of water 4 5 A.K.Soper∗ 6 ISIS Facility, STFC Rutherford Appleton Laboratory, 7 Harwell Science and Innovation Campus, Didcot, Oxon, OX11 0QX, UK 8 (Dated: March 17, 2008) 9 Arguably the most important liquid in our existence, water continues to attract enormous efforts 10 to understand its underlying structure, dynamics and thermodynamics. These properties become 11 increasingly complex and controversial as we progress into and below the “no man’s land” where 12 bulk water can only exist as crystalline ice. Various, so far unconfirmed, scenarios are painted 13 for this region, including the second critical point scenario, the singularity-free scenario, and most Science 319 14 recently the “critical point-free” scenario [C.A. Angell, , , 582 (2008)]. In this article the structural aspects of water (as opposed to its dynamic and thermodynamic properties) in its 15 supercooled, amorphous and confined states are explored, to the extent that these are known and 16 related toFor the water phasePeer diagram. AnReview important issue to emerge Only is the extent to which structural 17 measurements on a disordered material can tell us about its phase. 18 19 20 I. INTRODUCTION competing scenarios for understanding the properties of 21 supercooled and glassy water. One of these was the pro- 22 Figure 1 shows a simplified phase diagram of water in posal that at a temperature of around 220K and pressure 23 the ambient and low temperature region. In particular of 0.1GPa, water exhibited a second critical point below 24 it is seen that, starting from ambient pressure, the melt- which it separated into one of two distinct forms, high 25 ing point falls to about 251K at a pressure of 0.2GPa, density liquid (HDL) and low density liquid (LDL), with 26 then rises to higher temperatures with further increases a first order transformation between them. These liquids 27 in pressure. Below the equilibrium melting line lies the were a continuous extension of their much lower temper- 28 homogeneous nucleation line - this is the lowest temper- ature amorphous counterparts, HDA and LDA respec- 29 ature to which the liquid can be cooled before sponta- tively. The increase in density and entropy fluctuations 30 neous crystallisation sets in: the bulk liquid cannot exist as one approached this critical point could be used ex- 31 below this line. At ambient pressure it occurs at 231K. plain the increased compressibility and heat capacity of 32 At much lower temperatures, below 150K, water can be water on cooling water below its stable melting temper- 33 made to form an amorphous solid, either by condensing ature. The consequences of such a second critical point 34 water vapour onto a cold substrate, or by hyperquenching would extend up into the accessible supercooled and low temperature stable liquid regions. Hence there is for ex- 35 droplets of the liquid, or by pressurising crystalline ice to ample a marked but continuous structural transforma- 36 around 1GPa. The material made by the latter process tion between HDL and LDL in the low temperature sta- 37 is actually a high density form of amorphous ice, called ble liquid region [3]. However the difficulty of observing 38 HDA: when the pressure is released it converts spon- this second critical point in the bulk liquid, or even get- 39 taneously to the low density form, LDA, which, struc- ting close to it, means it is ‘virtual’ in the same sense that 40 turally at least, appears very similar to the amorphous the focal point of a concave lens is virtual, namely you 41 ices formed by vapour deposition and hyperquenching. might be able to see the effects of the postulated second 42 Subsequently an even denser form of amorphous ice was found, VHDA, formed by annealing HDA under pressure critical point in the stable liquid region, even though you 43 can’t observe it directly. 44 at 120K. When the pressure is released VHDA converts This means its existence will remain controversial, 45 directly to LDA and does not stop at HDA on the way, since there is also the “singularity free” scenario, which 46 so there is a debate about whether VHDA is truly a new reports a similar phase diagram to that with the second 47 form of amorphous ice, or whether it is simply densified, critical point, but without any divergences in thermody- 48 annealed HDA. The space (shaded in Fig. 1) between the highest amorphous ice formation temperature, TX , namic quantities and with a rapid but continuous trans- 49 formation between HDL and LDL. Within this scenario 50 and the homogeneous nucleation line, TH , is often called “no man’s land” since experiments on the bulk liquid in the transition from HDA to LDA would also be rapid but 51 that region are impossible. continuous. 52 More recently, Angell [4, 5] has proposed, in analogy 53 Comprehensive and very readable reviews of the cur- to what happens in crystalline C , the hypothesis of an 54 rent state of thinking on supercooled water were pub- 60 lished in 2003.[1, 2] At that time there appeared to be two order-disorder transition occurring as deeply cooled wa- 55 ter is heated, this transition occurring also at 225K. Such 56 a transition is apparently observed in highly confined wa- 57 ter, where the freezing transition is suppressed. It does 58 ∗[email protected] not exclude the second critical point scenario, but it also 59 60 URL: http://mc.manuscriptcentral.com/tandf/tmph Molecular Physics Page 2 of 36 2 1 2 does not require it. How this most recent suggestion re- bulk water, which had generally been more or less under- 3 lates to the full phase diagram of cold water remains to stood for many years as a tetrahedrally coordinated net- 4 be determined. work, was thrown into disarray in 2004 [14] with the chal- 5 It is interesting that in 1997 Jeffery and Austin, [10], lenging, and so far unconfirmed, proposition that water in 6 published an equation of state that they claimed was sig- fact was more chain-like than network-like. Independent 7 nificantly more accurate than existing equations of state evidence in support of this idea is still lacking, and the 8 for water. Extrapolating this phase diagram into the su- interpretation of the x-ray absorption spectroscopy data 9 percooled region, they refer to a second critical point for that led to the original proposal is controversial and not fully understood. Nonetheless the proposal has led to a 10 water and the likelihood of a high density to low density substantial amount of research trying to corroborate or 11 transition which occurs immediately before freezing at understand the findings. 12 high enough pressures. These ideas were elaborated in In this situation it is not possible in a review of this 13 a subsequent paper, [12]. In the meantime Mishima and kind to cover all the topics related to water with any de- 14 Stanley in 1998 observed a discontinuity in the decom- gree of completeness. Instead the aim is to summarise as 15 pression induced melting line of ice IV and concluded this best as possible the primary structural aspects of ambient 16 could mean there wasFor a second criticalPeer point at aboutReview the Only and supercooled water and amorphous ice, to the extent 17 same pressure and temperature as Jeffery and Austin, that these are understood, and to identify discrepancies 18 namely 0.1GPa and 220K, [11]. They do not refer to between different accounts. Reference to computer sim- 19 the Jeffery and Austin equation of state, but the latter ulations of water will only be made where relevant, since 20 appears to derive from previous work, [13], so it is not clear how truly independent are the two results. Even so the computer simulation of water is itself a vast topic 21 it seems curious that a very accurate equation of state that could not possibly be given adequate coverage here. 22 seems to predict the same behaviour as the observed Although computer simulations of water and other mate- 23 ice IV melting curve.