The appropriate question is on the longevity of particular structures, and the statistical
distribution of particular structures, and the statistical distribution of such as a function of
temperature. Relevant data which bear on this question, but obviously provide no quantitative
answers, are the facts that the concentration of the different clusters or fragments resembling
“dense ice” which must be present in all water samples at say 3 or 4º C, is very much higher
than that present at room temperature. And these clusters do not disappear because of various
bond breakage phenomena. They are of course in thermodynamically stable equilibrium. (i.e.
last forever).

Deconstructing the terminological confusion around the term “structure of water”
The sections above have adduced evidence from, and hence have been written in, the
“language” of materials science. Strangely, however, in spite of some 17 million hits on
Google for “structure of water,” materials scientists rarely study this most common material.
The structure of water has been largely the province of chemists, and the reader must
understand the differences in language and approach between these two communities. The
vast majority of papers on the “structure of water” in the chemical and biochemical literature
start (and most often end) with statements and claims about what molecules exist in the water,
on the basis of particular, increasingly specialized, tools. The prominence of hydrogen-
bonding in the molecules is regularly commented on.

The very first (cited from July 25, 2004) reference listed on the Google list is (in our opinion)
one of the very best and most comprehensive and most valuable reviews of this topic ever
devised. It is a website by Martin Chaplin, of London’s Southbank University, which
contains an enormous, complex, and well-organized review of the entire field
(http://www.lsbu.ac.uk/water/) [38]. Navigating through data from several dozens of papers,
each only a click away, it is fair to say that Chaplin presents others’ data on some hundreds of
“structures of water” molecules. A small selection is assembled in Figs. 8 and 9 just to
illustrate the ambiguity in the chemical literature associated with the term “structure of water”.



Fig. 8 The enormous variety of structures of the molecules in which almost certainly the chemical entity H2O
can exist. The well known H2O monomer with its precisely defined tetrahedral angle is shown on the top left and
below it a series of dimers, trimers, tetramers which can be constructed on paper from the relatively rigid H2O
molecule, and so on. Moderate sized molecules are on the right. See Chaplin 2004 (q.v.) for individual
references for any particular structure pictured above [38].



Fig. 9 This figure shows some of the larger polyhedra which are presumed to exist, largely on the calculation of
likely structure of tetrahedrally bonded units. For refs. see Chaplin [38]. The relationship of the images of
individual molecules, and how they are related to each other, in 3-D space, in liquid water, are rarely treated, the
emphasis being on which units are present.

Probably several hundred thousands of papers discuss the structure of the monomeric H2O
molecule itself, and an equal number discuss some selection of the other molecules. The
question is: Is it legitimate to use the term “structure of water” in presenting such images? It
would certainly be more precise to call it the “structure of the water molecule(s).” But of
course the rest of Chaplin’s references also address exactly the same subject and deal not with
H2O but (H2O) trimers, oligomers and polymers where x varies from 2 to say 250. Clearly
H2O but (H2O) trimers, oligomers and polymers where x varies from 2 to say 250. Clearly
water molecules appear in a whole range of sizes. The structure of a condensed phase,
however, must surely also describe how these units are packed together. A very large number
of similar papers in the chemical literature on the structure of water report the presence of
some particular complex oligomer or polymer detected by a particular experimental method
under particular circumstances. But these papers do not specify how these molecules are
arranged in space, nor do they address what other molecules may be present. Moreover, a
very large number of these papers deal with water vapor, not liquid water, a distinction easily
lost in the reading.

A very large number of additional excellent and detailed papers have appeared which present
evidence for the presence of specific molecular arrangements. An interesting cluster of these
appeared recently in Science. Miyazaki et al. (Science, May 21, 2004) show infrared
spectroscopic evidence for oligomers of different shape and sizes from n=4-27 in (H2O)n [41].
Shin et al. (May 21, 2004) present intriguing IR data near the 3.7µ O-H stretching band in
oligomers from 6-27, around the “magic number” of n=21 [42]. From neither of these papers
can one tell whether the authors believe that water—all waters under undelimited
conditions—contain 100% of these molecules, or a majority. Nor is there any comment on
how such clusters are distributed in space, or whether different size clusters are themselves
formed into separate regions of the nano-heterogeneous bulk water.

Some six months later, the October 22 and October 29 issues of Science carry several
exquisitely detailed papers on water from senior authors. They discuss the energetics and
dynamics of electron binding and transport in various cluster sizes, some of it in vapor
samples. These processes are extremely rapid in the tens of femtoseconds. The papers do not
consider any models with a distribution of cluster sizes, nor do they show how reproducible
the data are with different water samples, even allegedly pure’ ones, or prepared by different
means. Wernet, et al. using XRD and Raman spectroscopy, supported the view favoring only
ring and chain molecules, while J.D. Smith et al. used their total electron yield near-edge X-
ray absorption time structure (TEY-NEXAFS) technique to come to very different
conclusions that the water and ice H-bondings are very similar, and that the usually accepted
1-5 kcal/mole for the H-bond strength is consistent with their data [43, 44].

Somewhat analogous, albeit much less precise, measurements were made on nearest neighbor
arrangements, 30-40 years earlier on (silicate) glasses by the then state of the art tools: optical,
XRD, IR and Raman spectroscopy and EXAFS (Extended X-ray Absorption Fine Structure).
None of these even hinted at the subsequently established nano-heterogeneities as the real
structure of many glasses. Of course, the kinetics of bond making and breaking, are radically
different. As discussed earlier, this complicates, but does not eliminate, the need to consider
the model of nanoheterogeneity for the generalized structure of bulk water.

Clearly the origin of some of the inherent confusion in the field is based on the materials
scientists’ and the chemists’ use of the same term to mean different things. Chemists use
“structure” to describe the structure of the molecules or structural building blocks.’ Materials
Scientists use “structure” to describe the 3-D structural architecture of the material. The
former describe the size and shape of the bricks or cement blocks; the latter describe the shape
and size of the walls and the room and how the bricks and blocks are arranged within it.