Clay Minerals (1999)34, 185-191
Fundamental particles" an informal history
P. H. N A D E A U
Statoil, N-4035 Stavanger, Norway
(Received 27 May" 1997; revised 3 December 1997)
A B STRACT: An informal overview is given of the development of the fundamental particle
model, where, in a series of papers published from 1982-1987, researchers from the Macaulay
Institute presented a radical model for the interpretation of the crystal structure, chemistry, and
genesis of interstratified clays. The model reconciled electron microscopy and X-ray diffraction
(XRD) data from a variety of sedimentary clay specimens, and proposed that these minerals were
composed of nano-crystalline 'fundamental particles' whose adsorptive interfaces were responsible
for the expanding smectite layers observed by XRD. The term 'interparticle diffraction' was used to
describe this phenomenon. Experiments were reported which proved the model, by producing
randomly interstratified illite-smectites from combinations of ordered illite-smectite and smectite
clay dispersions. The model was extended to propose that fundamental particles were the primary
units of crystallization, and that changes in the particle population were responsible for the
commonly observed evolution in XRD character of these minerals with increasing depth of burial
and temperature.
The advancement of science is often marked by
specific turning-points, which establish new perceptions of the coherence of nature (Bronowski, 1973).
This paper presents an informal view of what is, in
the author's opinion, such a turning-point within the
field of clay mineralogy, and the background from
which it emerged. The research mainly concerns
interstratified or mixed-layered clays which constitute an important class of clay minerals. The results
also have broader implications within mineralogy
and geochemistry.
During the 1970s several research efforts were
underway to apply advancements in modelling onedimensional XRD structures (Reynolds, 1967) of
interstratified illite-smectites (I-S) to regional
geologic mapping of low-grade thermal alteration
of sedimentary rocks. These efforts resulted in part
from the success the methodology had shown in US
Gulf Coast petroleum exploration wells (Perry &
Hower, 1970). Cretaceous marine sediments in the
western interior of North America were prime
research candidates due to their petrologic composition, the numerous associated bentonites, and their
involvement in Larimide tectonic/igneous activity
(Nadeau & Reynolds, 1981).
The mechanism by which smectite was converted
to illite with increasing depth/temperature was
generally accepted as K+ fixation and smectite
layer collapse (Hower et al., 1976). Research results
identified certain problems with this mechanism
(Nadeau, 1980), including:
(1) Chemical data indicated decreased Fe and Mg
octahedral compositions of silicate layers which
were inconsistent with smectite collapse. (2) The
one-dimensional I-S layer sequence was observed to
change from randomly interstratified (Ro) for
smectite rich composition, to IS ordered (R1), IIS
(R2), and IIIS (R3) with progressively increasing
illite layers. These changes in layer sequences,
particularly the significant number of R2 observations (Fig. 1), were incompatible with a simple
collapse mechanism. (3) The three-dimensional
structure of the silicate layer configuration was
observed to change from turbostratic, or rotationally
disordered, to a highly ordered configuration, which
again was difficult to reconcile with a smectite
lattice conservation mechanism.
A modified mechanism which involved beidellite
as an intermediate between montmorillonite and
illite was postulated (Fig. 2), and an application for
9 1999 The Mineralogical Society
186
P. H. Nadeau
All samples
IlliUc clays R-->I
aoo l
160
140
25O
120
200
100
80
150
60
100
40
50
20
0
R0
R1
R2
R3
R1
R2
R3
F~c. 1. Distribution of illite-smectite layer sequence types based on XRD character for Mancos Shale and
bentonites from all samples (left) and illitic clays (right). Random interstratification - R0, Ordered
interstratifications = Rb R2, and R3. Note that R2 type interstratification is approximately twice as abundant
as R3 type (Nadeau, 1980).
a post-doctoral research position at the Macaulay
Institute was sent out while completing m y
dissertation at Dartmouth College.
In 1980 I undertook to apply my thesis results to
petroleum exploration within Cretaceous sedimenIllite
oo.y
..... ......... .............
Smectite
FIG. 2. Proposed diagenetic trend for the conversion of
smectite to illite in Mancos shale and bentonites
(Nadeau, 1980).
tary units. I had identified prospective areas based
on relationships between the degree of thermal
alteration of the sediments as indicated by the clay
mineralogy from surface samples, and occurrences
of oil and gas accumulations in nearby basins. As
an exploration geologist with royalty interests in my
project areas, I was later able to convey those
interests for research support.
Studies at the Macaulay began in October, 1982,
and were originally anticipated to last six months.
The initial project was based mainly on selected
bentonite samples. With some additional support
from the British Technology Group and Shell
Research Laboratories in the latter stages, the
project expanded in scope to cover ultimately a
period of over three years, and produced no less
than 15 publications.
MODEL
DEVELOPMENT
The actual work got off to a hesitant start.
Discussions with V.C. Farmer suggested the
unlikely role of beidellite as an intermediary to
illite. [ was not easily put off this line of research. I
was sufficiently persuaded, however, to include a
Fundamental particles
study of beidellite in the research program. A
sample was kindly provided by G. Lagaly, which
later resulted in a well-received journal article
(Nadeau et al., 1985a). Research results on layer
charge, Greene-Kelly determinations, and cation
exchange capacity (CEC) were going slowly. It was
not until early 1983 when characterization of the
samples by transmission electron microscopy
(TEM) was underway, that I began to appreciate
the magnitude of the problem. J.M. Tait kept
producing micrographs of exceedingly thin particles
from R1 and R2 type illitic bentonites. From my
XRD background, I knew that MacEwan crystallites
were composed of 5 to 15 silicate layers
(MacEwan, 1956; Reynolds, 1980). The problem
was, no such particles were observed in my
specimens. Electron diffraction indicated that
larger particles were aggregates of the smaller
thin particles, which commonly were only a few
silicate layers in thickness. These observations
constrained the illite layers within the thin particles,
which required that smectite layers must occur at
expandable interfaces between the particles. In this
regard the illitic bentonite clays were similar to
diagenetic illitic clays from North Sea Magnus
Field reservoir sandstones, which had recently been
studied at the Macaulay Institute (McHardy et al.,
1982).
To my mind, describing particles as thin was not
sufficient. So the arduous task of quantitatively
measuring the thickness, area, and perimeter
distributions of these particles was undertaken. At
this time we included a relatively well-crystallized
diagenetic illite from the Rotleigende sandstone and
the Wyoming bentonite into the sample suite to
represent end-members of illite and smectite. The
entire research direction was changing, focusing
more and more on these nanometer-scale particles.
A major problem that had to be overcome was to
demonstrate clearly that these TEM particle
observations were significant, representative and
related to the XRD character of the specimens. This
placed strict requirements on sample preparation as
well as analytical methods. If the one-dimensional
XRD I-S layer sequence was determined by
interparticle coherency of diffracted radiation
within specimen aggregates, which was an hypothesis from the Magnus work, then experiments were
required to validate the phenomenon and establish
the scientific theory.
It was reasoned that the particle population must
determine the layer sequence observed by XRD. If
187
we could combine experimentally different particle
populations in completely dispersed suspensions in
a quantifiable way, the layer sequence examined by
XRD of the altered aggregated specimens should
also change systematically. An RI I-S bentonite
clay separate with -50% illite layers by XRD was
selected for the first experiments, where TEM
indicated a relatively uniform population of
particles 2 nm thick, along with a clay separate
from the Wyoming bentonite, 100% smectite layers
by XRD and predominantly composed of particles
1 nm thick by TEM. The results yielded an XRD
Ro I-S series with progressively increasing smectite
layers with increasing proportion of Wyoming
bentonite clay. The layer sequences had been
entirely rearranged, a result entirely consistent
with the concepts of fundamental particles and
interparticle diffraction. While reviewing XRD
patterns produced from these experiments, M.J.
Wilson noted the significance of the results and
prophetically added that it would take at least ten
years before such a radical model would be
generally accepted. The interstratification of clay
minerals and their evolution could now be
interpreted in an entirely new deterministic way,
which rationalized the probabilistic nature of their
XRD character.
From this point, we prepared a manuscript for
Clay Minerals along the lines of a cautionary note,
but outlining the implications of the work. The
reviews could not have been more divergent, with
one very positive and the other very negative. It
was a very courageous editorial decision by D.J.
Morgan that allowed this communication to appear
in print (Nadeau et al., 1984a). The extension of the
model to R 2 and R3 I-S, and chlorite interstratifications proceeded rapidly with the experience we had
already gained. An important achievement was a
publication in Science (Nadeau et al., 1984b) which
demonstrated the generality of the new model. An
alternative dissolution/precipitation mechanism for
the diagenetic conversion of smectite to illite was
further elaborated in Mineralogical Magazine
(Nadeau et al., 1985).
We had not yet addressed the issue of internal vs.
external silicate layer charge of fundamental
particles, an important point perhaps anticipated
by G. Brown at a Royal Society Discussion
Meeting in London (Brown, 1984). Later, during a
meeting of the Clay Minerals Group held at the
Royal Entomological Society, G. Brown made a
presentation which was highly critical of the new
188
P. Hi Nadeau
model, including the failure to stay within accepted
standards of clay mineral nomenclature. I made a
diplomatic rebuttal to his comments, and noted a
helpful criticism he provided to a figure from a
manuscript which dealt with the layer charge issue
then in press (Nadeau & Bain, 1986). The modified
figure is presented here with due acknowledgement
(Fig. 3).
INDUSTRIAL
reluctant to pursue such actions, but a presentation
by R.H. Ottewill also at the Royal Society meeting
(Lubetkin et al., 1984) convinced me otherwise. It
is important to note here that certain comments
which appeared in the published proceedings were
not disclosed at the actual meeting. Meanwhile, a
broadly based application was filed by the British
Technology Group (BTG), prior to the first Clay
Minerals' paper in 1984, resulting in eight patents
awarded in the major industrialized nations.
The BTG also provided additional support for the
research programme, which resulted in collaboration with M.J. Adams, then at the University of
Aberystwyth in Wales. The BTG was mainly
interested in using nanometer scale fundamental
APPLICATIONS
Shortly after experimental confirmation of the
model, R.C. Mackenzie suggested that a patent
brief be filed with the National Research and
Development Corporation. At first we were
Interstratification
Silicate
layer
sequence
Model
layers
Smectite
E+ n H 2 0 +-s
Fundamental
Particle
Model
particles
Smectite : 1 nm
E+ n H20 +-~
Smectite : 1 nm
Smectite
E+ n H 2 0 +-s
Illite : 2 nm
Illite
Smectite
Illite
Illite : 3 nm
Illite
Smectite
Illite
Illite : 4 nm
Illite
Illite
E+ n H20 ~ 0
FIG. 3. Silicate layer sequence of illite-smectite based on interstratification within MacEwan crystallites (left) and
the fundamental particle model (right). Two to one (2:1) silicate layers are represented by two tetrahedral sheets
(black) joined by one octahedral sheet (stippled). Exchangeable cations, water, or organic molecules such as
ethylene glycol are indicated by E + nH20 • ~ , and planes of non-exchangeable cations within fundamental
illite particles or interstratified illite layers are represented by 'k's. Note that a single 2:1 silicate layer is identical
to an elementary smectite particle according to the fundamental particle model, but not to a smectite layer in the
interstratification model where the octahedral sheets are taken as the layer boundaries (modified after Nadeau &
Bain, 1986; and G. Brown, pers. comm.).
Fundamental particles
particle technology to produce petroleum cracking
and refining catalysts. The Group was well aware of
the highly lucrative zeolite catalyst patents held by
a major oil corporation.
The work supported by the BTG did not live up
to their expectations. It did, however, result in a
contribution published in the journal Applied Clay
Science, which also described applications in films,
membranes, composites and ceramics (Nadeau,
1987a). The collaboration did allow me to
characterize the distribution of fundamental clay
particles in a more extensive sample set. This work
would later prove very valuable. Industrial interest
in fundamental particles is currently on the rise
again, particularly in the field of clay-polymer
nanocomposite materials.
PHYSICAL
DIMENSIONS
Measuring the physical dimensions of hundreds of
nanometer-scale clay particles may not seem like a
rewarding way to spend one's valuable research
time, but I realized that these data were the first to
provide independent and quantitative determinations
of interstratified clay mineral layer sequences. We
appreciated early on that the most important
parameter to define was the mean particle
thickness, which fortunately did not take very
many observations. Statistically speaking, 30 or so
were generally good enough. The problem was
defining the distribution, particularly as the mean
thickness increased and the distribution broadened.
The main data set (Nadeau, 1985) showed rather
high standard deviations. At first we thought this
might be due to the limited number of observations.
Then D.D. Eberl sent us a set of sericite samples
for TEM measurements, from which four selected
samples proved very useful for filling in the picture
for thicker particle populations. The sericite TEM
data, though limited, were incorporated into an
overall study (Eberl et al., 1987), and gave
remarkably good correlations with other measured
parameters by XRD, infrared spectroscopy and
chemical analysis.
When the sericite TEM data were added to the
main data set for illitic clays, I realized that the
high standard deviations were in fact a function of
the mean. The data also showed that the mean
particle area was correlated with the mean
thickness. This meant that the mean particle
thickness or area, which were relatively easy to
determine, could be used to calculate the overall
189
particle population distributions. This was as good a
proof as I could obtain to demonstrate that the
particles being measured were the primary units of
crystal growth processes. The results appeared as a
note in Clay Minerals (Nadeau, 1987b), and this
was the last communication 1 was to write at the
Macaulay, as the research project drew to a close.
PARADIGM
SHIFT
The fundamental particle concept had been established, and the model would now have to survive or
perish depending on its merits. Our anticipation that
the radical new model would be in for a rough ride
was confirmed by some comments in the literature
(e.g. Ahn & Buseck, 1990). Close examination of
these critical comments, however, revealed that
they were mainly based on opinions in that the
evidence presented, often anecdotal high-resolution
TEM lattice-fringe and structure images, could be
readily interpreted in terms of fundamental particle
theory (Wilson, 1990), particularly if the complex
petrographic features of epitaxial diagenetic illite
growth on pre-existing detrital illite and micas in
shales are taken into account.
Meanwhile at Dartmouth College, R.C. Reynolds
and his students were busy evaluating the new
model. Detailed work on the stoichiometry of the
smectite to illite reaction in sediments adjacent to
igneous intrusions indicated that more illite was
formed than would be anticipated from the initial
smectite content (Lynch, 1985), demonstrating the
importance of diagenetic illite neoformation in
pelitic sediments. The fundamental particle data
set was used to calculate one-dimensional layer
sequences in a newly created computer code called
' F P M O D ' , and compared with conventional
Markovian stacking arrangements within MacEwan
crystallites (Tellier, 1988). The results indicated
that, in general, the calculated diffraction profiles of
I-S as fundamental particle aggregates realistically
simulate XRD experimental observation, and that
for long-range ordered R2 and R3 I-S, the calculated
diffraction profiles of fundamental particle aggregates more realistically simulate experimental
observations than calculated diffraction patterns
based on Markovian sequences in MacEwan
crystallites. The work on particle population
distributions was proving very useful indeed,
particularly considering the 102~ or so silicate
layers examined during a typical clay XRD
analysis. D.D. Eberl and his colleagues were also
190
P. H. Nadeau
making significant headway with the new model,
establishing a continuum between diagenetic clays,
hydrothermal sericites and metamorphic minerals
based on log-normal crystal Size distributions (Eberl
et al., 1990).
Fundamental particle theory contributed to major
simplifications in one-dimensional (00/) computer
simulations of I-S XRD maxima. This step was
achieved by reducing the number of input parameters, mainly by using the mean particle thickness
to define the entire population distribution, and
thereby the sequence of layers. A powerful basis was
now established from which to achieve the next
level of XRD simulation complexity, three-dimensional structural hkl determinations of interstratified
clays. After several years of intense and dedicated
work by the leading scientist in this area, the results
finally became available. They showed that both the
overall 00l and hkl XRD characteristics can be
modelled based on fundamental particle concepts
(Reynolds, 1992; 1993). These studies found that the
hkl maxima arise from within the fundamental
particles, and that turbostratic dislocations occur at
the expandable interfaces between fundamental
particles. Experimental XRD data from rock
fragments were also consistent with I-S consisting
of fundamental particle aggregates in nature.
Although the debate may continue in the literature
for some time to come as to the origin of these
particles (e.g. Dong & Peacor, 1996), the way
forward is now clear. As it was when it was first
proposed over ten years ago, the fundamental
particle model is the only model which can
uniformly explain the layer sequence and geological
evolution of layer sequences of interstratified I-S
clays.
As a young research worker, one of the most
illuminating experiences is to be part of a scientific
revolution, even one related to a comparatively
narrow field such as clay mineralogy. It provides a
perspective of science, its foundations, workings,
progress, limitations and human aspects, not
achieved from any other vantage point. Although
I cannot say it necessarily benefits one's career, at
least in the short term, the history of science reveals
that this is often part of the price demanded for
major advancements of knowledge. This results, at
least in part, because the emergence of new theories
is often accompanied by periods of "pronounced
professional insecurity" as described in The
Structure of Scientific Revolutions (Kuhn, 1970),
which introduced the term "paradigm shift".
Currently, our profession appears to have arrived
at what might be termed a 'Copernican' accommodation with fundamental particle theory. We use
the concept, for example, to calculate clay mineral
structures, and to explain their evolution and
genesis, but still hold to the reality of interstratified
clays in the form of MacEwan crystallites. To what
extent this represents a metastable situation and
how it might be resolved in the future is a task at
hand.
ACKNOWLEDGMENTS
The author thanks T.S. West, Director, and R.C.
Mackenzie, Head of Pedology, when this work was
initiated at the Macaulay Institute. The open environment they provided to visiting scientists, the experienced research staff with m u l t i - d i s c i p l i n a r y
capabilities in clay mineral studies, and the opportunity
to pursue new directions within a defined programme,
were all critical factors in the emergence of the
concepts outlined here. M.J. Wilson provided constructive comments which helped to prepare this manuscript
for publication.
REFERENCES
Ahn J.H. & Buseck P.R. (1990) Layer stacking
sequences and structural disorder in mixed-layer
illite/smectite: Image simulations and HRTEM
imaging. Am. Miner. 75, 267-275.
Bronowski J. (1973) The Ascent of Man. Little Brown
Co., Boston.
Brown G. (1984) Crystal structures of clay minerals and
related structures. Phil Trans. Royal Soc. Lond.
A311, 221-240.
Dong H. & Peacor D.R. (1996) TEM observations of
coherent stacking relations in smectite, I-S and illite
of shales: Evidence for MacEwan crystallites and
dominance of 2M1 polytypism. Clays Clay Miner.
44, 257-275.
Eberl D.D., Srodofi J., Kralik M., Taylor B.E. &
Peterman Z.E. (1990) Ostwald ripening of clays
and metamorphic minerals. Science, 248, 474-477.
EberI D.D., Srodofi, J., Lee M., Nadeau P.H. & Northrop
H.R. (1987) Sericites from the Silverton Caldera,
Colorado: Correlation among structure, composition,
origin, and particle thickness. Am. Miner. 72,
914 934.
Hower J., Eslinger E.V., Hower M. & Perry E.A. (1976)
Mechanism of burial metamorphism of argillaceous
sediments: I. Mineralogical and chemical evidence.
GeoL Soc. Amer. Bull. 87, 725-737.
Kuhn T.S. (1970) The Structure of Scientific
Revolutions. Univ. Chicago Press, Chicago.
Fundamental particles
Lubetkin S.D., Middleton S.R. & Ottewill R.H. (1984)
Some properties of clay-water dispersions. Phil.
Trans. Royal Soc. Lond. A311, 353 368.
Lynch F.L., III (1985) The stoichiometry of the smectite
to illite reaction in a contact metamorphic environment. AM thesis, Dartmouth College, USA.
MacEwan D.M.C. (1956) Fourier transform methods for
studying lamellar systems. I. A direct method for
analyzing interstratified mixtures. Kolloidzeitschrifi,
149, 96-108.
McHardy W.J., Wilson M.J. & Tait J.M. (1982) Electron
microscope and X-ray diffraction studies of filamentous illitic clay from sandstones of the Magnus Field.
Clay Miner. 17, 23-39.
Nadeau P.H. (1980) Burial and contact metamorphism
in the Mancos shale. PhD thesis, Dartmouth College,
USA.
Nadeau P.H. (1985) The physical dimensions of
fundamental clay particles. Clay Miner. 20,
499-514.
Nadeau P.H. (1987a) Clay particle engineering: A
potential new technology with diverse applications,
Appl. Clay Sci. 2, 83-93.
Nadeau P.H. (1987b) Relationship between the mean
area, volume and thickness for dispersed particles of
kaolinites and micaceous clays and their application
to surface area and ion exchange properties. Clay
Miner. 22, 351 356.
Nadeau P.H. & Bain D.C. (1986) Composition of some
smectites and diagenetic illitic clays and implications
for their origin. Clays' & Clay Miner. 34, 455-464.
Nadeau P.H. & Reynolds R.C. (1981) Burial and contact
metamorphism in the Mancos Shale. Clays Clay
Miner. 29, 249 259.
Nadeau P.H., Farmer V.C., McHardy W.J. & Bain D.C.
(1985a) C o m p o s i t i o n a l v a r i a t i o n s of the
Unterrupsroth beidellite. Am. Miner. 70, 1004-1010.
Nadeau P.H., Tait J.M., McHardy W.J. & Wilson M.J.
(1984a) Interstratified XRD characteristics of physical mixtures of elementary clay particles. Clay
191
Miner. 19, 923 925.
Nadeau P.H., Wilson M.J., McHardy W.J. & Tait J.M.
(1984b) lnterstratified clay as fundamental particles.
Science, 225, 923-925.
Nadeau P.H., Wilson M.J., McHardy W.J. & Tait J.M.
(1985b) The nature of some illitic clays from
bentonites and sandstones: Implications for the
conversion of smectite to illite during diagenesis.
Mineral. Mag. 49, 393-400.
Perry E.A. & Hower J. (1970) Burial diagenesis in Gulf
Coast pelitic sediments. Clays Clay Miner. 18,
165-177.
Reynolds R.C. (1967) Interstratified clay systems:
Calculation of the total one-dimensional diffraction
function. Am. Miner. 52, 661-672.
Reynolds R.C. (1980) Interstratified clay minerals. Pp.
249 303 in: Crystal Structures' of Clay Minerals and
their X-ray Identification (G.W. Brindley & G.
Brown, editors) Mineralogical Society, London.
Reynolds R.C. (1992) X-ray diffraction studies of illite/
smectite from rocks, <1 ~tm randomly oriented
powders, and <1 lain randomly oriented aggregates:
The absence of laboratory induced artifacts. Clays
Clay Miner. 40, 387-396.
Reynolds R.C. (1993) Three-dimensional X-ray powder
diffraction from disordered illite: Simulation and
interpretation of the diffraction patterns. Pp. 44-78
in: Computer Applications to X-ray Diffraction
Analysis of Clay Minerals (R.C. Reynolds & J.R.
Walker, editors) CMS workshop lectures 5, Clay
Minerals Society, Boulder, USA.
Tellier K.E. (1988) Calculated one-dimensional X-ray
diffraction patterns of illite/smectite as .fundamental
particle aggregates': Implications for the structure of
illite/smectite and the smectite to illite transition.
PhD thesis, Dartmouth College, USA.
Wilson M.J. (1990) Fundamental particles and interstratified clay minerals: Current perspectives. Miner.
Petrog. Acta, 33, 5 14.