The Structure Of Liquid Water
Novel Insights From Materials Research
Potential Relevance To Homeopathy

Keywords: Water, Structure of water, Epitaxy, Succusion, Nanobubbles, Colloids.

This paper provides an interdisciplinary base of information on the structure of liquid water.
It begins with a synthesis built on the information base on the structure5 of noncrystalline,
inorganic, covalently-bonded condensed liquid phases, such as SiO2, S, Se, P, and H2O, which
exists in the materials science literature. The data for water are analyzed through the prism of
well-established algorithms in materials research: the connection of properties to structure;
the pressure-temperature (P-T) phase diagrams; the phenomenon of epitaxy; the phenomenon
of liquid-liquid phase separation; the stability of two phase colloids; and, the recently
discovered effects of weak magnetic and electric fields on the structure of simple inorganic
oxides. A thorough combing of the literature of the condensed matter properties reflecting
structural features of essentially pure water obtained via the normal processes of preparing
homeopathic remedies, provides another rich data base.

The examination of these data through the standard materials science paradigms leads to the
following conclusion: Many different structures of liquid water must exist within the range of
observations and processes encountered near ambient conditions. A typical sample of water
in these experimental ranges no doubt consists of a statistical-mechanical-determined
assemblage of monomers and oligomers (clusters) of various sizes up to at least several
hundred H2O units. The importance of the structural similarity of SiO2 and OH2 is very
relevant to the structure of the latter as well as to the probability of epitaxy in controlling at
least the region contiguous to the silicate glass surfaces of many common containers.

The most distinctive feature of bonding in liquid water is not only the “well-known hydrogen
bonds, but the necessary presence of a wide range” of van der Waals bonds between and
among the various oligomeric (cluster) structural units. It is this range of very weak bonds
that could account for the remarkable ease of changing the structure of water, which in turn
could help explain the half-dozen well-known anomalies in its properties. In its subtler form,
such weak bonds would also allow for the changes of structure caused by electric and

1
Evan Pugh Professor of the Solid State, Emeritus, and Founding Director of the Materials Research
Laboratory at Penn State (rroy@psu.edu).
2
Professor Emeritus and former Department Chair of Materials Science, Stanford University.
3
Professor of Medicine, Psychiatry, Family and Community Medicine, and Public Health, Director of Research,
Program in Integrative Medicine, University of Arizona (ibell@u.arizona.edu).
4
Assistant Professor, Materials Research Institute, Penn State(rickhoover@psu.edu)
5
The term structure is used as in all materials research to designate the 3D arrangement of atoms or molecules,
not the chemical usage of the term describing the structure of a single molecule or oligomer.

©2005 Matrice Technology Limited Materials Research Innovations 9-4: 1433-075X 

This paper does not deal in any way with, and has no bearing whatsoever on, the clinical
efficacy of any homeopathic remedy. However, it does definitively demolish the objection
against homeopathy, when such is based on the wholly incorrect claim that since there is no
difference in composition between a remedy and the pure water used, there can be no
differences at all between them. We show the untenability of this claim against the central
paradigm of materials science that it is structure (not composition) that (largely) controls
properties, and structures can easily be changed in inorganic phases without any change of
composition. The burden of proof on critics of homeopathy is to establish that the structure of
the processed remedy is not different from the original solvent.

The principal conclusions of this paper concern only the plausibility of the biological action of
ultradiluted water remedies, they are based on some very old (e.g. homeopathy) and some
very new (e.g. metallic and nanobubble colloids) observations which have been rejected on
invalid grounds or due to ignorance of the materials research literature and its theoretical
basis. This constitutes an excellent example of the common error in rejecting new scientific
discoveries by using the absence of evidence as evidence for absence. 

Introduction 

The “structure of liquid water” receives some 8 million hits on Google and the “structure of
water” over twice as many. Any contribution that can be made to this vast body of knowledge
is sure to be marginal. This paper does not report any such incremental advance with
ultraprecise measurements about the structure of oligomers, femtosecond spectroscopy of
bond breakage or phase transitions in glassy water. Instead, it examines the literature to
establish only one proposition, that pure, thermodynamically stable or metastble liquid water
can have more than one 3-D condensed matter structure. While we assemble here various sets
of relevant data and lines of argumentation, by a coincidence, at the same time as this paper
was first presented orally (April 2004), Kawamoto et al. published their paper providing the
experimental proof of this assertion [1, 2]. Of course solid crystalline water has been known
to exist in nearly ten different structures, and workers such as Angell and DeBenedetti and
Stanley have given us an extraordinarily precise and interesting picture of certain metastable
liquid waters, or metastable solid glasses of water with different properties and structures [3,
4]. These observations mimic the same phenomena known for generations in H2O’s close
relative, SiO2. This paper brings together a very wide range of disparate observations on
water (and other liquids which share one or more structural or bonding parameters) to support
the case that water can indeed have its properties and hence its structure changed rather easily
in non-linear ways without any change of composition.

The structure of crystalline inorganic matter which became a major pillar of the physics and
chemistry of solids was based on the discovery by von Laue and the Braggs, father and son, of
the diffraction of X-rays by the periodic array of atoms in crystalline solids. It remains the
sine qua non of characterization in contemporary materials research. The Braggs were
followed by the schools of V.M. Goldschmidt (including Barth, Lunde and Zachariasen in
Oslo), and Linus Pauling in California, who applied this new tool of X-ray diffraction (XRD)
to a very large number of the common (crystalline) solids in the world of inorganic science
and technology. Thus was born the extremely reliable science of crystal chemistry: the
relationship of structure to composition as a function of the most powerful intensive
thermodynamic variables: temperature, and pressure (see books by Goldschmidt; Pauling;
Evans; and Muller and Roy [5—8]). The term structure
relationship of structure to composition as a function of the most powerful intensive
thermodynamic variables: temperature, and pressure (see books by Goldschmidt; Pauling;
Evans; and Muller and Roy [5—8]). The term structure is unambiguously defined in crystal
chemistry as the position in 3-D space of each atom or ion typically with a precision
nowadays of say µ0.01 nm.

What immediately will catch the attention of an interested observer is the ratio, in inorganic
crystal chemistry books, of the space devoted to solids as compared to liquids. It approaches
100:1. And thereby hangs our tale. Why? Water as a liquid is the most common phase on
the surface of the earth, followed by ice. A very distant second is crystalline SiO2 as quartz
(one of the dozen structurally different forms of SiO2). The fact that we know the precise
details of the structure of each form of crystalline SiO2 while we have only the most
rudimentary understanding of liquid SiO2 is due to a fundamental lack in our arsenal of tools
for determining the structure of liquids. The fact that low viscosity liquids sustain a
continuous rapid movement of the atoms and/or molecules contained in them is not the
defining difficulty. The effective tool of XRD is totally lacking for all noncrystalline (i.e.
aperiodic) matter whether solid or liquid. The only tool which can now be used definitively
and directly (albeit partially) to show the structure of non-crystalline solids (e.g. glasses) is
transmission electron microscopy (TEM) and this cannot easily be directly used on liquids.
Thus it is not surprising that many scientists, due either to ignorance or powerlessness, hold
the naïve view that all liquids, like most crystalline matter, are more or less completely
homogeneous in structure down to the unit cell, atomic or molecular level, and they exhibit
structural characteristics in accord with the random network model, one of the two models
developed nearly a century ago, for glasses [9]. This model of the “structure of glass” starts
with that of the structurally homogeneous crystalline materials (i.e. those in which a structural
element, the unit cell, is repeated throughout the sample in all 3 dimensions), and moves the
atoms or ions from their normal sites, required for periodicity, by bending or stretching the
bonds. This so-called random network model taken from Zachariasen’s original paper is
shown in Fig. 1 [9]. 6


Fig. 1. The classical picture of the “Random Network Structure” as presented by Zachariasen in 1932, which
has become “established” as the structure of glass on the basis of model fitting on x-ray scattering data. The
key assumption (unrecognized by others for 7 or 8 decades) of this model is that the structure of all glasses is
“homogeneous” in the same ways as crystals are.

This now outdated image, based on no direct data from other methods, has dominated the
thinking of the physics and chemistry community ever since, and it became their “working
model”. Opposed to this “homogeneous structure” was the early “crystallite” theory (Prins
which posited that small 5—50 A° fragments of various crystalline structures floated in a
monomeric sea [10]. For over half a century, international conferences have periodically
revisited the question of homogeneous (random-network) or heterogeneous (crystallite)
structures for glass (frozen liquids). By the 1980s, the definitive relevant data came not from
XRD but from transmission electron microscopy (TEM) in common alkali boro- and alumino-
silicate glasses (Mazurin and Porai-Koshits; Fig. 2 is taken from their work) which showed
the heterogeneous “nano-structure” of many, many transparent glasses which have even 2 or 4
separate phases [11]!! (A phase is defined as a region of characteristic structure or
composition separated by a surface.)

Fig. 2 In sharp contrast with the hypothetical calculations based on Zachariasen’s random network theory, is
the direct TEM evidence. Shown are some examples of binary and ternary glasses, some quenched, some heat-
treated which clearly show actual phase-separation. One can confidently assume that in many if not most glasses
and in many liquids, structural (-composition) fluctuations must exist as precursors to such phase separation
(after Mazurin and Porai-koshits) [11].

The existence of the entire glass-ceramic industry depends on this incipient
nanoheterogeneity or actual phase separation in glass, and the myriad TEM images from
Corning (Beall and Pinckney) shows the true nanocomposites that result [12]. The existence
and high probability of nanoheterogeneity in most strongly bonded glass and liquid structures
are now established as the “standard model”.