American Mineralogist, Volume 94, pages 1653–1671, 2009

New insights into smectite illitization: A zoned K-bentonite revisited

Douglas K. Mccarty,1,* Boris a. saKharov,2 anD victor a. Drits2

1
Chevron ETC, 3901 Briarpark, Houston, Texas 77063, U.S.A.
2
Geological Institute of the Russian Academy of Science, Pyzhevsky per. D.7, 119017 Moscow, Russia

aBstract
The illitization reaction in a thick K-bentonite bed located in upper Cretaceous marine shale in the
Montana disturbed belt was studied by X-ray diffraction, chemical analysis, and thermal gravimetric
analysis. Modeling of the experimental XRD patterns from oriented clay specimens in air-dried and
glycolated states shows that at each sample location in the bentonite bed a mixture of R0 illite-smectite
(I-S) and R1 I-S coexist. Each of these phases in all samples consists of the same or similar content
of illite and expandable layers independent on location in the bed. In particular, the illite content in
the R0 I-S and the R1 I-S from the <0.5 µm fractions is equal to 30 and 62%, respectively. The main
difference between the samples at different locations is the different weight concentrations of the
coexisting I-S phases. The R1 I-S content decreases progressively from the lower and upper contacts
of the bed to its center. The reverse trend was observed for the R0 I-S. The layer unit-cell parameter b
increases from samples located near the middle of the bed toward samples near the bed margins. The
DTG patterns of the samples contain two endothermic maxima at about 640 and 470 °C, corresponding
to cis-vacant (cv) illite and trans-vacant (tv) smectite layers coexisting in the R1 I-S and R0 I-S.
Analysis of the crystal-chemical features of the R1 I-S and R0 I-S shows that, in the middle of
the bed, both phases are characterized by the lowest octahedral Mg and the highest tetrahedral Al
contents. In the structural formula of the R1 I-S, the tetrahedral Al content is significantly higher than
the (K+Na) content independent of sample location. In contrast, tetrahedral Al in the R0 I-S located
near the bed boundaries is lower compared with (K+Na) content.
To account for the crystal-chemical features of the coexisting I-S, a first assumption is that the
initial volcanic ash was altered into tv smectite having a homogeneous Al-rich composition through-
out the bed. Second, along with K, the active role in illitization was controlled by Mg. Mineralogical
zonation of the K-bentonite is explained by the progressive migration of K from the margins toward
the bed center with the associated decrease of K cations in the pore fluids. However, the decrease in
K concentration was accompanied by a successive increase in R0 I-S content, but not a progressive
decrease in illite layer content in a single I-S phase. The main role of Mg was to redistribute octahedral
and tetrahedral Al in the 2:1 layers of the R0 I-S and R1 I-S in such a way that the amount of Al in
the tetrahedral sheets increased at the expense of the substitution of Mg for Al in the octahedral sheet
of the 2:1 layers in the initial smectite.
These results demonstrate a new insight into mineralogical sequences of intermediate members of
smectite illitization. Instead of a statistically homogeneous and continuous reaction associated with the
increase of illite layers in I-S and the simultaneous increase of order of the layer stacking sequence,
the illitization reaction in the thick K-bentonite consists of the formation of a physical mixture of
two I-S having a contrasting content of layer types and their distribution. Factors responsible for the
formation of the coexisting R0 I-S and R1 I-S are discussed.
Keywords: Illitization, K-bentonite, multispecimen fitting, X-ray diffraction, smectite, cis-vacant,
trans-vacant, TGA DTG

introDuction the formation of I-S in different geologic environments have

The illitization of smectite consists of the formation of illite been proposed (e.g., Schroeder 1992). However, as was noted
from smectite through intermediate mixed-layered illite-smectite by Altaner and Ylagan (1997) smectite illitization can be ap-
(I-S) structures. This transformation is one of the most important proximated by two main mechanisms: solid-state transformation
mineralogical reactions occurring in sedimentary basins during (SST) and dissolution-precipitation (DP). The first mechanism
burial diagenesis, as well as in active and fossil hydrothermal sys- describes illitization in the solid state as a result of layer-by-layer
tems, in conditions of contact metamorphism, and metasomatic replacement (Shutov et al. 1969; Hower et al. 1976). The second
alteration of bentonites (Środoń 1999). Several mechanisms for mechanism involves dissolution of smectite followed by precipi-
tation of a newly formed I-S (Ahn and Peacor 1985; Nadeau et al.
* E-mail: dmccarty@chevron.com 1984; Inoue 1986; Środoń et al. 1999, 2000). The main structural,
0003-004X/09/1112–1653$05.00/DOI: 10.2138/am.2009.3260 1653
1654 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

chemical, and morphological features characteristic of SST and the I-S in these clay assemblages has a consistent illite smectite
DP mechanisms have been summarized by Altaner and Ylagan ratio of 65% I and 35% S layers that does not change within the
(1997) and Cuadros and Altaner (1998a, 1998b). studied depth interval. This corresponds to early diagenesis and
Until recently the commonly accepted view was that smec- the layers are randomly interstratified despite a significantly high
tite illitization is a statistically homogeneous and continuous illite layer content, WI = 65%, for R0 structures. In light of these
reaction associated with a steady increase of illite interlayers results, the origin and transformation of the coexisting phases
in I-S and a simultaneous increase in the order of the stacking during early diagenesis have been reconsidered for the Callovo-
sequence. In particular, illite and expandable smectite interlayers Oxfordian (Claret et al. 2004) as well as for the Oligocene Frio
are interstratified at random when the content of illite layers, WI, (McCarty et al. 2008) formations. The coexistence of smectite
is <50%. Further increase of the illite layer content (WI > 50%) and I-S in shallow sections of sedimentary basins is consistent
is accompanied by formation of more ordered I-S structures in with TEM observations showing that this phase association is
which interstratification of the layer types obeys short-range dominant in smectite-rich samples (Dong 2005).
order factors, R, which increase from 1 to 3 (Moore and Reynolds A new pattern for the evolution of clay minerals at different
1997; Środoń 1999). stages of diagenesis was discovered when the multispecimen
Despite the numerous investigations of smectite-to-illite approach was applied to XRD patterns of clay fractions from
transformation and extensive discussion of the reaction mecha- samples taken from different depths of the Carter well (onshore,
nisms, the mineralogical sequence of the intermediate I-S mem- Texas, Gulf Coast shale). In contrast to the illitization reaction
bers describing smectite illitization have not been determined described by Hower et al. (1976) as a series of pure I-S with dif-
unambiguously. One of the possible reasons is that in most pub- ferent contents and distributions of the layer types, new results
lications the interpretation of the experimental XRD patterns is show that smectite illitization is not homogeneous. The coexis-
confined to the analysis of the positions of basal reflections often tence of different phases at different stages of burial diagenesis
ignoring their intensity and profile (Środoń 1980, 1981, 1984; indicates heterogeneous reactions and new reaction mechanisms
Inoue et al. 1988; Moore and Reynolds 1997). This approach is (lanson et al. 2009).
restricted to a two-component system and usually confined by Up to now the multispecimen method has not been applied to
the analysis of XRD patterns from ethylene glycol (EG) solvated study the structural features and probability parameters of clay
I-S samples. Therefore, the interpretation may not correspond to minerals and their evolution during smectite illitization in thick
the actual structure (Drits 1997, 2003; lanson et al. 2005, 2009; K-bentonite beds, which are ideal for the study of this reaction
McCarty et al. 2008). because they do not contain detrital illite or I-S. In most publica-
The reliability in the interpretation of diffraction effects tions smectite illitization in K-bentonites is described as a steady
from mixed-layered phases has been significantly improved reaction through an intermediate single-phase I-S (Cuadros
by the development of the multispecimen method (Drits et al. and Altaner 1998a, 1998b). The only exception is the study
1997; Sakharov et al. 1999) that was successfully applied in of the structure and origin of clay minerals at the Cretaceous/
other studies (lindgreen et al. 2000, 2002; Drits et al. 2002a, Tertiary boundary layer, Stevn Klint (Denmark). Modeling of
2002b, 2004, 2007; Claret et al. 2004; Ferrage et al. 2005a, the experimental XRD patterns showed that a sample from the
2005b, 2007; McCarty et al. 2004, 2008; lanson et al. 2005). boundary clay layer consists of two phases: a highly smectitic
This method requires that, for the same sample, XRD patterns I-S containing 95% smectite layers and a low-smectitic I-S
should be recorded after different treatments (e.g., Ca- and/or consisting of 50% smectite and 50% illite (aluminoceladonite)
Mg-saturated specimens recorded in both air-dry and EG states). layers (Drits et al. 2004). Previous work assumed that such XRD
Using computer modeling for each such pattern, an optimal patterns correspond to a pure smectite (Rampino and Reynolds
agreement between the experimental and calculated intensity 1983; Kastner et al. 1984).
distribution is obtained for the same structural model by a trial- In this paper, the multispecimen method will be applied to
and-error procedure. This new approach shows that the com- study the structural and probability parameters of mixed-layered
monly accepted description of smectite illitization needs to be I-S phases from a thick, “zoned” K-bentonite bed located in Up-
revised. For example, lindgreen et al. (2002) used this method per Cretaceous marine shale in the Montana Disturbed Belt.
to determine the actual structure and diagenetic transformations
of clay minerals from North Sea and Denmark chalk formations. Geology, clay mineralogy, and previous work
The experimental XRD patterns were similar to those attributed Several K-bentonite beds of various thicknesses were en-
in many publications to pure smectite or highly smectitic I-S. closed in Upper Cretaceous marine shale that was subjected to
However, modeling of these patterns showed that the samples metamorphism by burial beneath Precambrian and Paleozoic
consist of a physical mixture of high-smectitic (HS) and high- thrust sheets in the Montana disturbed belt during the laramide
illitic (HI) randomly interstratified I-S varieties containing 15 Orogeny (Mudge 1970). The Cretaceous rocks in the disturbed
and 50% illite layers, respectively. belt have undergone low-grade metamorphism at T > 100 °C. The
Similarly, modeling of experimental XRD patterns showed K-bentonites formed by the metasomatic migration of K into the
that clay fractions of samples from the Callovo-Oxfordian forma- initially smectite-rich bentonite beds, resulting in the formation of
tion of the Paris basin (Claret et al. 2004) and from the Oligocene I-S. Altaner et al. (1984) carried out a comprehensive investiga-
Frio Formation of the offshore Gulf of Mexico (McCarty et al. tion of 34 K-bentonites from the Marias River Formation in the
2008) consist of a physical mixture of smectite and random (R0) Montana Disturbed Belt and proposed a model for K-bentonite
I-S. One of the remarkable features revealed by modeling is that formation. They found that a 2.5 m thick K-bentonite is zoned
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1655

mineralogically and chemically, such that the <0.5 µm clay frac- treated to remove carbonate and Fe-oxide cements with a Na-acetate buffer and
tions of the samples at the contacts with the surrounding black Na-dithionite, respectively (Jackson 1985). The Na+ saturated clays were thoroughly
exchanged with Ca2+, with dialysis to remove excess salt (e.g., Moore and Reynolds
shale contain an ordered I-S with 65 to 70% of illite layers and 1997). Oriented aggregates were made by evaporation onto glass slides to provide
R = 1. In contrast, the middle of the bed contains R0 I-S with an a sample ~4 cm long with at least 10 mg clay per cm2 (Moore and Reynolds 1997).
average of about 50% illite layers. The authors concluded that Diffraction scans were collected with a Thermo Xtra diffractometer with a θ-θ
the decrease of the illite layer content in the I-S demonstrates goniometer and a 250 mm radius, equipped with a solid-state Si detector in the
air-dried (AD) state, and after ethylene glycol (EG) treatment by vapor solvation
an incomplete reaction caused by a deficient K supply, and the
in a heated chamber (60 °C). The scans were made from 2 to 52 °2θ with a 0.02
K2O content in the clay fractions changes in accordance with the °2θ step increment and counting rate of 4 s per step or longer using CuKα radiation
variation of the illite layer content in the I-S. transmitted through a 1.00 mm divergence and 1.80 mm antiscatter slit. Detector
Altaner et al. (1984) assumed that K for the I-S formation slits were 2.00 and 0.3 mm.
in the K-bentonite is derived externally from the long-range Random powder XRD patterns were also collected with the same diffraction
system over the I-S 060 angular region using side-loading sample mounts with a
breakdown of K-feldspar and muscovite in the shale host rock, 10 wt% internal standard of NIST SRM 640c Si powder mixed in with the sample
which migrated into the bentonite from the margins by diffusion to calibrate 2θ offset. Scans were made from 54 to 64 °2θ, with a 0.005° step
through the pore fluids. The time of the conversion of the initial increment, and a count time of 30 s per step. The d(060, 331) peak values were
bentonite into K-bentonite was estimated by the determination determined for each sample by three different methods. The first was made after
calibration by picking the peak maximum and using an average CuKα λ value of
of the I-S K/Ar age. I-S age at the upper contact is about 56 m.y.
1.5418 Å. The second used a Savitzky-Golay smoothing function in the “Jade”
and about 54 m.y. at the lower contact. The I-S age at the center computer program (MDI, Inc.). The third used a λ value of 1.54059 Å, and the
of the bed is 3 to 6 m.y. younger than at the contacts, indicating Jade program was used to remove the Kα2 contribution with a Rachinger algorithm
a slow rate of K migration into the bentonite bed. Based on these before smoothing.
data, Altaner et al. (1984) proposed the following mechanism
Modeling of the XRD patterns
of K-bentonite formation. Initially volcanic ash was altered to
To simulate XRD patterns from oriented clay specimens, the multispecimen
smectite under near-surface conditions. This was followed by a method initially developed by Drits and Sakharov (1976) and later by Drits et al.
thermal event due to burial beneath laramide thrust sheets, which (1997a) and Sakharov et al. (1999) was used. For each I-S, the investigated struc-
transformed smectite into I-S. Mineralogical zonation in the 2.5 tural model includes the number of layer types (not limited to two), the nature and
m thick bentonites is caused by formation of I-S with illite layer the proportion of the different layer types, and the statistical description of their
stacking sequences. Smectite layers and interlayers containing two or one layers
contents that are progressively limited by the availability of K.
of EG or H2O molecules are denoted S1 and S2 and correspond to 2EG and 1EG in
the EG state and to 2W and 1W in the AD state, respectively.
Samples layer sequences in I-S varieties can obey laws that differ by the short-range
Detailed analysis was performed on nine regularly spaced order factor, R, equal to the number of the nearest preceding layers that influence
samples from a ~2.7 m thick bentonite bed in the Cretaceous the occurrence probability for a final layer of a given type. In the case of R = 0,
layer types are interstratified at random, and R = 1 if the occurrence of a layer of
Marias River Formation along the South Fork of the Sun River a given type depends on the nearest predecessor. To describe a layer sequence in
in the Montana Disturbed Belt, located several miles east of two component I-S with R = 0, the relative content of each interstratified layer
Gibson Reservoir. Major element chemical analysis by XRF type, Wi, should be known. In the case of R = 1, four additional junction probability
was made on all bulk samples. Quantitative phase analysis by parameters, PII, PIS, PSI, and PSS are required. The value of Pij defines the probability
for layer type j to follow layer type i (i, j = I, S). The occurrence probability of
XRD was carried out on the bulk samples following the method
layer pair, ij, layer triplet, ijf, etc., are calculated by the equations Wij = WiPij, Wijf
of Środoń et al. (2001) and shows that the samples consist of = WiPijPjf (i, j, f = I, S). The Wi and Pij are interrelated and only two independent
quartz, plagioclase feldspar, calcite, kaolinite, chlorite, and I-S parameters are needed to describe any layer sub-sequence. In particular, when WI >
(Table 1). K-feldspar was not detected in any samples. A sharp WS and R = 1 parameters WS and PSS should be known. When PSS = 0, a maximum
lower stratigraphic contact exists between the K-bentonite and possible degree of order (MPDO) in the distribution of the layer types occurs. Such
an I-S can be denoted as R1 I-S with MPDO. However for a given WS, a set of I-S
the host shale, whereas the upper contact is less sharp perhaps varieties with R = 1 and some tendency to order in the layer type distribution can
due to depositional or thrusting processes. exist with WS > PSS > 0. The condition PSS = WS corresponds to I-S with R = 0. In
contrast, a set of I-S with R = 1 and a tendency to segregation of the interstratified
ExpEriMEntal MEthoDs layers occur when 1 > PSS > WS. The value of PSS = 1 corresponds to a physical
mixture of illite and smectite. For a three-component model consisting, for example,
Sample preparation and X-ray diffraction pattern
of illite and two S1 and S2 smectite layers and WI > WS1 > WS2, six independent
recording probability parameters were used for R = 1: WI , WS1, PS1S1, PS1S2, PS2S1, PS2S2 (see
The <1.0 and <0.5 µm equivalent spherical diameter size fractions were Drits and Tchoubar 1990 for details).
separated from samples by standard centrifugation methods. The samples were Because the different sample treatments may change the thickness and the

Table 1. Whole-rock chemistry from XRF and mineralogy from XRD (wt%)
Sample Distance from SiO2 Al2O3 CaO MgO Na2O K2O Fe2O3 TiO2 P2O5 LOI Total Qtz Plag Calc Kaol Chl I-S
lower contact (cm)
S001 12.7 55.3 19.47 4.10 2.07 0.40 3.87 2.17 0.17 0.05 12.18 99.77 16 4 5 3 2 71
S002 25.4 64.9 16.61 1.83 1.82 0.44 3.11 1.94 0.16 0.04 9.20 100.05 30 3 1 2 2 62
S003 50.8 70.3 14.09 1.08 1.64 0.31 2.62 1.57 0.13 0.03 8.37 100.15 41 2 0.5 2 1 53
S004 101.6 70.8 13.07 1.44 1.56 0.37 2 1.89 0.13 0.04 8.45 99.76 38 3 1 3 5 49
S005 114.3 68.7 14.85 1.50 1.55 0.40 2.22 1.49 0.15 0.04 9.18 100.10 37 3 0.6 3 4 53
S006 139.7 73.6 13.42 1.49 1.56 0.43 2.08 1.97 0.15 0.04 5.05 99.80 46 7 0 4 4 39
S007 152.4 73.0 12.23 0.38 1.21 1.21 1.72 1.89 0.17 0.05 7.93 99.81 43 4 1 4 3 45
S008 165.1 69.8 13.67 1.88 1.57 0.45 2.28 1.84 0.16 0.04 8.23 99.92 39 5 1 4 2 47
S009 177.8 71.9 12.72 1.50 1.54 0.42 2.03 1.88 0.14 0.04 7.80 99.94 41 4 0.8 3 4 48
Note: Qtz = Quartz, Plag = Plagioclase feldspar, Calc = Calcite, Kaol = Kaolinite, Chl = Chlorite, I-S = Illite-smectite.
1656 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

scattering power of swelling interlayers, but not the distribution of the layer types, Relationship between K content in the I-S structural
a consistent structure model is obtained when the stacking sequences of these formula and in illite fundamental particles in the I-S
layer types are almost identical for different XRD patterns of the same sample. In
addition, the relative content of the different phases determined by simulation of Each I-S structure consists of a set of illite fundamental particles separated
these patterns must be identical or close. from each other by smectite interlayers. A statistically weighted or mean thickness
Instrumental and experimental factors such as goniometer radius, length, and of illite fundamental particles is equal to several 2:1 layers bound by interlayer K.
thickness of the oriented specimens were measured and introduced without further The equation: mKn = (n – 1)nK is related to the mean number of 2:1 layers in illite
adjustment (Drits and Tchoubar 1990). The optimal value of sigma star (σ*) was fundamental particles, n, the number of K cations, nK, in the interlayers of illite
found to be 12°, and the mass absorption coefficient (µ*) was set at 45, as recom- fundamental particles, and the number of K cations, mK, in the structural formula
mended by Moore and Reynolds (1997). For structural models of AD and EG I-S of the I-S. The mean number of 2:1 layers in illite fundamental particles containing
containing Ca in the interlayers, the z coordinates and site occupancies for 2:1 (n – 1) illite interlayers is related to the mean number of layers, N, in the coherent
layers and interlayers of Moore and Reynolds (1997) are used. The thicknesses scattering domains of the I-S by the equation (Drits et al. 1997b):
of the coherent scattering domains (CSD’s) are log-normally distributed and the N
parameters of this distribution are determined using a mean thickness of CSD’s n= ;
and the regression given by Drits et al. (1997b), with mean and maximum thick- ( N − 1)W + 1
 exp 
ness of CSD’s as variable parameters. Possible variations in the distance between
parallel CSD’s were taken into account by introducing into the program a special therefore:
parameter H, which is equal to the mean distance between adjacent CSD values
mK N (6)
(Sakharov 2005). nK =
The quality of the agreement between the experimental and calculated XRD ( N − 1)(1− W ) .exp
patterns is estimated by a profile factor Rp:

Here Wexp is the amount of the expandable layers in the I-S. Equations 5 and 6 will
∑I −I exp
i
calc
i be used to interpret the experimental data obtained for the samples by combining
Rp = i XRD and chemical analyses.
∑I exp
i
.
i Thermal analysis
Here Iiexp and Iicalc are the measured and calculated intensities at each i step Dehydration and dehydroxylation reactions and weight loss vs. sample heating
increment. experiments were performed on all the samples using a “Mettler Toledo” TGA/
DSC 1 thermogravimetric analysis device with a heated capillary tube from 25 to
Relationship between weight concentrations of EG 1000 °C at a heating rate of 5 °C/min. To record the atomic mass signal of evolved
solvated and dehydrated I-S structures gas during heating, the tube was connected to a “ThermoOnix” quadropole mass
spectrometer. Reagent grade N2 was used as a carrier gas flowing at 50 ml/min
let us consider an I-S structure for which its weight concentration, CIS, as well
during the heating cycle. This equipment allows the weight loss owing to the
as occurrence probabilities or relative amounts of illite, WI , and smectite, WS, layers
release of different molecules (H2O, OH, CO, CO2, etc.) to be estimated at a given
are known from the experimental XRD pattern modeling. The weight concentrations
temperature. The TG curves were transformed into DTG patterns.
of illite and smectite layers in the I-S are denoted as CIIS and CSIS, therefore,

CIS = CIIS + CSIS. (1) Chemical analysis

Elemental oxide concentration for bulk rock was determined by X-ray fluo-
A ratio of the weight concentrations of smectite and illite layers in the I-S can rescence spectrometry, and the Ca-saturated <0.5 µm size fractions of the samples
be determined from the equation: by lithium metaborate/tetraborate fusion ICP analysis.

CSIS  C − C IS  M IS N IS M ISW IS rEsults

=  IS IS I  = SIS SIS = SIS SIS . (2)
C IS
 C I  M N M I WI Qualitative description of the experimental XRD patterns
I I I

The experimental XRD patterns recorded for nine oriented,

Here MSIS and MIIS are the molecular weights of smectite and illite layers, NSIS/NIIS is Ca-saturated, and EG solvated specimens from the <1.0 and
the ratio of the smectite and illite layers in the I-S and WSIS/WIIS = NSIS/NIIS; NSIS and
NIIS are the numbers of smectite and illite layers in the mean CSD of the I-S.
<0.5 µm size fractions contain two characteristic 2θ regions, one
The weight concentration of illite layers in the I-S can be obtained from from 2 to 8 °2θ and the other from 8 to 30 °2θ angles, respec-
Equation 2. tively. Each of these regions shows particular diffraction features
C IS M IISWIIS that reflect the evolution of these samples and the zonation of the
(3)
C IIS = K-bentonite bed. To illustrate let us consider the main diffraction
( M IISWIIS + M SISWSIS ) features observed for the <1.0 µm fractions (Fig. 1). The low
let us denote the weight concentration of the I-S in which smectite interlayers 2θ region of sample S001 (13 cm) contains two well-resolved
are dehydrated as CdhIS. By analogy with Equation 2 one can write the equation: reflections with d-values equal to ~31 and 13.4 Å, respectively.
In samples S002 (25 cm) and S003 (51 cm), the 13.4 Å peak
(C IS
dh
− C IIS )
=
M SdhWSIS practically disappears and its trace is only a slight bend, and the
. (4)
C IIS M IISWIIS 31 Å peak becomes less resolved at the lower 2θ angle side. Si-
multaneously, a small hump or weak peak with d = 16.8 Å appears
Here MSdh is the molecular weight of the dehydrated smectite layers.
Replacing the term CIIS in Equation 4 by Equation 3, the weight concentration
in the low 2θ regions of samples S002 and S003. The transition
of the dehydrated I-S is obtained: from sample S003 to sample S006 (140 cm) is accompanied by a
successive increase in intensity of the 16.8 Å reflection, whereas
IS
C IS ( M IISWIIS + M SdhWSIS ) the 31.0 Å peak intensity decreases and loses resolution at the
C =dh
. (5)
low 2θ angle side. In contrast, the diffraction effects change in
M IISWIIS + M SISWSIS
the low 2θ angle regions in samples S006 to S009 (178 cm) and
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1657

whereas sample S001 is represented by a single I-S. The high

intensity of the 31 Å super-reflection in the XRD pattern of
sample S001 is a result of the high degree of ordering in the
illite and smectite layer interstratification. The positions of the
other basal reflections in the XRD pattern are located between
the d(00l) values of rectorite with d(001) = 26.86 Å and of illite
with d(001) = 9.98 Å, which means that sample S001 consists of
I-S with WI > WS and R = 1 (Drits et al. 1994). This or a similar
I-S phase exists in all the samples because the XRD patterns
always contain the super-reflection at the same or similar position
as that in the XRD pattern of sample S001. The regular variation
of the super-reflection intensity shows that the content of this
high illitic I-S decreases from samples S001–S006 and increases
again from samples S006–S009.
The second phase may be either a high smectitic I-S or pure
smectite. Sample S006 contains the maximum amount of this
high smectitic phase because its XRD pattern has the most intense
16.8 Å peak with respect to that of the 31 Å super-reflection.

Quantitative description of the experimental XRD patterns

To simulate the experimental XRD patterns, a set of trial-and-
error steps was used to get satisfactory agreement between the
XRD patterns for each specimen in both AD and EG states with
the same structural model. Based on the qualitative analysis of
the XRD patterns, each size fraction of the sample was considered
to consist of a physical mixture of two mixed-layered I-S phases
having contrasting contents of smectite layers. To be valid, for
FigurE 1. XRD patterns from oriented <1.0 µm Ca2+-saturated EG both AD and EG states, the model should have almost the same
sample preparations. Samples S001 to S009 correspond to the distance
proportions of the coexisting I-S varieties, each of which should
from lower shale contact.
contain the same amount of illite and expandable layers with the
same mode in their interstratification. In each I-S, the number of
are reversed with respect to the features described for the first S1 and S2 expandable layers may be different in AD and EG states,
sample sequence. That is, the intensity of the 31 and 16.8 Å peaks but their total number should be the same. For each size fraction
increase and decrease, respectively (Fig. 1). of the samples (except S001), the best possible fit between the
In the high 2θ regions of the sample sequence, the number experimental and calculated XRD patterns was achieved for a
of basal reflections is the same and the main difference between model in which one phase corresponds to a highly smectitic
the XRD patterns is in migrations and profile variations of these I-S with random interstratification, R0 I-S, whereas in the low
reflections. For example, in the transition from samples S001 smectitic I-S the layer types are interstratified with a tendency
to S006, a migration of the reflections with d-values near 9.215 to order (R1 I-S). The experimental and simulated XRD patterns
and 5.277 Å can be observed to higher and lower 2θ angles, for EG and AD specimens of the <1.0 µm fractions as well as
respectively. The d-values of the corresponding reflections are for EG specimens of the <0.5 µm fractions are shown in Figures
equal to 9.165 and 5.291 Å for sample S002; 9.165 and 5.306 Å 2, 3, and 4, respectively. Tables 2–5 contain the structural and
for S003; 9.071 and 5.389 Å for S004; 9.118 and 5.438 Å for probability parameters for both phases. Tables 6–7 contain the
S005; and 9.025 and 5.471 Å for sample S006. According to weight concentrations of each phase. For the <1.0 µm fractions,
Méring’s rules, the observed diffraction features are related to the RP factors vary from 5.5 to 7.6% for the EG and from 5.0
an increase in smectite layer content. Note, however, that start- to 7.5% for AD structural models (Table 6). For the EG < 0.5
ing from sample S004, the profiles of these reflections change µm fractions, these values vary from 4 to 6.7% (Table 7). Note
significantly and become broader and asymmetrical. that the <1.0 µm fractions contain noticeable amounts of kao-
As with the low 2θ regions, changes in the high 2θ regions linite, chlorite, and quartz. Reflections of these impurities were
from samples S006 to S009 are inverse with respect to those ignored and the simulation determined only the relative weight
described for samples S001 to S006, the 9.025 and 5.471 Å concentrations of the coexisting R0 I-S and R1 I-S phases. The
reflections of sample S006 changes direction and shifts to lower <0.5 µm fractions were impurity-free.
and higher 2θ angles, respectively. As a result, for sample S009, The <1.0 µm size fraction. The primary requirement of the
the d-values of the corresponding reflections are equal to 9.07 and multispecimen approach is fulfilled because the amount of the
5.356 Å, and their profiles become narrower and symmetrical. coexisting phases determined by modeling the XRD patterns
Based on this qualitative description of the diffraction for both the AD and EG states is almost the same (Table 6). The
features, which is similar for both the <1.0 and <0.5 µm size weight concentration of the R0 I-S phase increases progressively
fractions, samples S002 to S009 consist of two discrete phases, from 0% in sample S001 to 46% in sample S006 and then de-
1658 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

b b

a a

FigurE 2. (a) Measured and modeled (shaded curves) XRD patterns

from oriented <1.0 µm EG Ca2+-saturated sample preparations. Sample FigurE 3. (a) Measured and modeled (shaded curves) XRD patterns from
number increases with distance from lower shale contact. Phase content oriented <1.0 µm AD Ca2+-saturated sample preparations. Sample number
and fitting parameters are shown in Tables 2, 3, and 6. (b) Measured increases with distance from lower shale contact. Phase content and fitting
and modeled (shaded curves) XRD patterns from oriented <1.0 µm parameters are shown in Tables 2, 3, and 6. (b) Measured and modeled (shaded
EG Ca2+-saturated sample preparations. Sample number increases with curves) XRD patterns from oriented <1.0 µm AD Ca2+-saturated sample
distance from lower shale contact. Phase content and fitting parameters preparations. Sample number increases with distance from lower shale contact.
are shown in Tables 2, 3, and 6. Phase content and fitting parameters are shown in Tables 2, 3, and 6.
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1659

a b

FigurE 4. (a) Measured and modeled (shaded curves) XRD patterns from oriented <0.5 µm EG Ca2+-saturated sample preparations. Sample
number increases with distance from lower shale contact. Phase content and fitting parameters are shown in Tables 4, 5, and 7. (b) Measured and
modeled (shaded curves) XRD patterns from oriented <0.5 µm EG Ca2+-saturated sample preparations. Sample number increases with distance
from the lower shale contact. Phase content and fitting parameters are shown in Tables 4, 5, and 7.

Table 2. Structural and probability parameters obtained for the R1 I-S (<1.0 µm size fraction)
Sample and hI hS1 hS2 WI WS1 WS2 PII PIS1 PIS2 PS1I PS1S1 PS1S2 PS2I PS2S1 PS2S2 D1 N H
treatment
S001 EG 9.98 16.85 12.90 0.66 0.33 0.01 0.535 0.45 0.015 0.9 0.1 0 1 0 0 0.294 10.0 10.0
AD 9.98 14.85 12.50 0.66 0.30 0.04 0.535 0.409 0.056 0.9 0.1 0 0.925 0 0.075 8.0 3.5
S002 EG 9.98 16.80 0.63 0.37 0.518 0.482 0.82 0.18 0.486 10.0 10.0
AD 9.98 14.95 12.50 0.63 0.33 0.04 0.518 0.43 0.052 0.82 0.16 0.02 0.82 0.16 0.02 10.0 7.0
S003 EG 9.98 16.80 0.62 0.38 0.51 0.49 0.8 0.2 0.526 10.0 11.0
AD 9.98 15.15 12.50 0.62 0.36 0.02 0.51 0.465 0.025 0.8 0.189 0.011 0.8 0.189 0.011 8.0 6.0
S004 EG 9.98 16.80 0.62 0.38 0.497 0.503 0.82 0.18 0.474 10.0 11.0
AD 9.98 14.90 12.50 0.62 0.34 0.04 0.497 0.45 0.053 0.82 0.161 0.019 0.82 0.161 0.019 8.0 6.0
S005 EG 9.98 16.80 0.62 0.38 0.522 0.478 0.78 0.22 0.579 10.0 10.0
AD 9.98 14.90 12.50 0.62 0.33 0.05 0.522 0.415 0.063 0.78 0.192 0.028 0.78 0.192 0.028 8.0 4.0
S006 EG 9.98 16.80 0.62 0.38 0.522 0.478 0.78 0.22 0.579 10.0 10.0
AD 9.98 14.95 12.50 0.62 0.26 0.12 0.522 0.327 0.151 0.78 0.15 0.07 0.78 0.15 0.07 8.0 4.0
S007 EG 9.98 16.80 0.62 0.38 0.522 0.478 0.78 0.22 0.579 10.0 10.0
AD 9.98 14.85 0.62 0.34 0.04 0.522 0.428 0.05 0.78 0.197 0.023 0.78 0.197 0.023 8.0 5.0
S008 EG 9.98 16.80 0.62 0.38 0.497 0.503 0.82 0.18 0.474 10.0 9.0
AD 9.98 14.90 12.50 0.62 0.34 0.04 0.497 0.45 0.82 0.161 0.019 0.82 0.161 0.019 8.0 5.0
S009 EG 9.98 16.75 0.62 0.38 0.497 0.503 0.82 0.18 0.474 10.0 10.0
AD 9.98 14.85 12.50 0.62 0.34 0.04 0.497 0.45 0.82 0.161 0.019 0.82 0.161 0.019 8.0 5.0
Notes: hI, hS1, hS2 are the thicknesses of illite (I) and expandable (S1 and S2) layers, S1 corresponds to 2EG and 2W, S2 to 1EG and 1W expandable layers, respectively.
WI, WS1, WS2 are contents of I, S1, and S2 layers. N is the mean number of layers in CSD's. H is the parameter of CSD's coherency (the mean distance between parallel
CSD's in megacrystals in angstroms).

creases to 21% in sample S008, but increases slightly again to of illite and expandable layers independent of its location and
28% in sample S009 with the increasing distance from the lower weight concentration. Indeed, the contents of smectite and illite
bentonite-shale contact (Table 6; Fig. 5). layers in the R0 I-S in samples S004, S005, S006, and S007
The variation in the weight concentration of the R1 I-S phase are equal to 75 and 25%; in samples S003, S008, and S009 the
has the opposite trend. The remarkable feature of the coexisting smectite and illite layer proportions are 70 and 30%, and only in
phases is that each of them contains the same or similar amounts the R0 I-S of sample S002 is the illite smectite layer ratio equal
1660 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

to 65:35% (Table 3; Fig. 5). In the AD state, the R1 I-S consists of 14.85–15.15 Å (WS1)
Similarly, the illite layer content in the R1 I-S structures is and 12.50 Å (WS2) expandable layers containing two and one
equal to 62% for all samples except samples S001 and S002. layers of H2O molecules in their interlayer, respectively (Table
Sample S001 consists of a single I-S phase in which 66% of 2). Interestingly, the number of 12.5 Å layers in all R1 I-S phases
illite layers are interstratified with 33% S1 and 1% S2 expand- varies in a narrow interval from 2–5% and only in the R1 I-S
able 16.85 and 12.9 Å layers containing two and one layers of sample S006 does this type of layer increase to 12%. After
ethylene glycol molecules in their interlayers, respectively. The glycolation, all expandable layers swell to 16.75–16.85 Å and
interstratification of the layer types in this structure deviates only 1% of the 12.90 Å layers exist in the R1 I-S in sample S001
from the maximum possible degree of order (MPDO) at R = 1 as determined by modeling. The R0 I-S smectite layers contain
because PS1S1 = 0.1 (Table 2). two layers of H2O molecules and two layers of glycol molecules
The degree of disorder in the distribution of the layer types at in their interlayers in both the AD and EG states (Table 3).
R = 1 and Wi > Ws may be estimated by the equation (Drits and The <0.5 µm size fraction. The experimental XRD patterns
Tchoubar 1990): D1 = PSS/WS, where WS is the total content of of both the <1.0 and <0.5 µm size fractions of each sample in the
smectite layers. In the two-component EG R1 I-S structure WS = AD or EG states are similar. Therefore modeling of the experi-
WS1 and in the three-component AD R1 I-S WS = WS1 + WS2, PSS mental XRD patterns was carried out only for the EG < 0.5 µm
is the junction probability for layer type S to follow layer type fractions. Table 7 shows that the weight concentration of the R0
S. PSS = PS1S1 for the EG R1 I-S and PSS = PS1S1 + PS1S2 for the AD I-S phase increases from 13% in sample S002 to 55% in sample
R1 I-S. When the layer types are distributed at random, PSS = WS S006 and decreases to 30 and 32% in samples S008 and S009.
and the parameter of disorder D1 = 1; when PSS = 0, the disorder As with the <1.0 µm fraction, the I and S layer contents are
parameter is also equal to 0 and the layer types are interstratified equal to 30 and 70%, respectively, independent of their weight
with MPDO. The values of D1 show that the R1 I-S in sample concentrations (Table 5). Sample S005 has a slightly different
S001 has the lowest degree of disorder (D1 = 0.294), whereas the I-S layer ratio of 75:25% (Table 5; Fig. 6).
degree of disorder in the R1 I-S varieties progressively increase In the R1 I-S structures, the illite layer content is similar
from samples S002, S004, S008, and S009 (D1 = 0.474–0.486) in all the samples and equal to 62% in samples S003 through
to samples S005, S006, and S007 (D = 0.579) (Table 2). S008, 64% in sample S002, 65% in sample S001, and 60%
in sample S009 (Table 4; Fig. 6). The values of PS1S1 and D1
Table 3. Structural and probability parameters obtained for the R0 progressively increase in the R1 I-S structure from samples
I-S (<1.0 µm fraction) S001–S009 (Table 4).
Sample and treatment hS hI WS WI N H
S002 EG 16.73 9.98 0.65 0.35 10.0 10.0 DTG pattern decomposition
AD 15.00 9.98 0.65 0.35 8.0 2.0 The DTG patterns of the <0.5 µm size fractions contain one
S003 EG 16.73 9.98 0.70 0.30 10.0 5.0
AD 15.00 9.98 0.70 0.30 8.0 5.0 relatively narrow high-temperature peak at about 640 °C and
S004 EG 16.85 9.98 0.75 0.25 10.0 8.0 another much broader and lower temperature peak at about
AD 14.90 9.98 0.75 0.25 8.0 5.0
S005 EG 16.85 9.98 0.75 0.25 10.0 1.0
480 °C (Fig. 7). Each DTG pattern was decomposed into two
AD 14.90 9.98 0.75 0.25 8.0 3.0 peaks using the Jade computer program, and the profiles of the
S006 EG 16.77 9.98 0.75 0.25 10.0 5.0 observed and simulated DTG patterns almost coincide (Fig. 8).
AD 15.00 9.98 0.75 0.25 8.0 5.0
S007 EG 16.75 9.98 0.75 0.25 10.0 5.0 This quantitative treatment of the DTG patterns shows that from
AD 14.90 9.98 0.75 0.25 8.0 2.0 samples S001 to S005, the relative areas under the high- and
S008 EG 16.75 9.98 0.70 0.30 10.0 3.0 low-temperature peaks progressively decrease and increase,
AD 14.90 9.98 0.70 0.30 10.0 3.0
S009 EG 16.75 9.98 0.70 0.30 10.0 10 respectively. In particular, the areas under T1 = 640–646 °C, are
AD 14.85 9.98 0.70 0.30 8.0 3.0 equal to 74% for sample S001 and to 51% for sample S005. The
Notes: hI, hS are the thicknesses of illite (I) and expandable (S) layers, S corresponds reverse trend is observed from sample S006 at 52% to sample
to 2EG and 2W expandable layers, respectively. WI, WS are contents of I and S layers.
N is the mean number of layers in CSD's. H is the parameter of CSD's coherency S009 at 62% (Table 8).
(the mean distance between parallel CSD's in megacrystals in angstroms).

Table 4. Structural and probability parameters obtained for the R1 I-S (<0.5 µm size fraction)
Sample and treatment hI hS1 hS2 WI WS1 WS2 PII PIS1 PIS2 PS1I PS1S1 PS1S2 PS2I PS2S1 PS2S2 D1 N H
S001 EG 9.98 16.85 12.90 0.65 0.33 0.02 0.482 0.487 0.031 0.96 0.04 0 1 0 0 0.114 8.0 13.0
S002 EG 9.98 16.85 0.64 0.36 0.522 0.478 0.85 0.15 0.417 10.0 9.0
S003 EG 9.98 16.85 0.62 0.38 0.497 0.503 0.82 0.18 0.474 10.0 9.0
S004 EG 9.98 16.85 12.90 0.62 0.36 0.02 0.492 0.476 0.032 0.82 0.18 0 1 0 0 0.474 9.0 10.0
S005 EG 9.98 16.80 12.90 0.62 0.36 0.02 0.515 0.453 0.032 0.78 0.22 0 1 0 0 0.579 10.0 10.0
S006 EG 9.98 16.80 0.62 0.38 0.522 0.478 0.78 0.22 0.579 9.0 10.0
S007 EG 9.98 16.80 0.62 0.38 0.510 0.490 0.80 0.20 0.526 9.0 10.0
S008 EG 9.98 16.80 0.62 0.38 0.510 0.490 0.80 0.20 0.526 10.0 9.5
S009 EG 9.98 16.80 0.60 0.40 0.493 0.507 0.76 0.24 0.600 8.0 9.5
Notes: hI, hS1, hS2 are the thicknesses of illite (I) and expandable (S1 and S2) layers, S1 corresponds to 2EG and S2 to 1EG expandable layers, respectively. WI, WS1, WS2
are contents of I, S1 and S2 layers . N is the mean number of layers in CSD's. H is the parameter of CSD's coherency (the mean distance betw0een parallel CSD's in
megacrystals in angstroms).
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1661

Table 5. Structural and probability parameters obtained for the R0 2:1 layers of both I-S phases. All Fe cations are considered to
I-S (<0.5 µm fraction) be trivalent. Because the samples were thoroughly saturated by
Sample and treatment hS1 hS2 hI WS1 WS2 WI N H Ca and dialysis was used to remove excess electrolyte, all Na
S002 EG 16.73 12.90 9.98 0.65 0.05 0.30 8.0 8.0
S003 EG 16.73 12.90 9.98 0.60 0.10 0.30 8.0 5.0
and K cations are assumed to occupy the illite interlayers in both
S004 EG 16.80 9.98 0.70 0.30 8.0 6.0 coexisting I-S phases.
S005 EG 16.80 9.98 0.75 0.25 8.0 3.0 Table 10 contains the structural formulae, and Figure 9
S006 EG 16.75 9.98 0.70 0.30 9.0 3.0
S007 EG 16.75 9.98 0.70 0.30 8.0 4.0 shows the variation of the cation composition as a function of
S008 EG 16.75 9.98 0.70 0.30 8.0 3.0 distance from the lower bentonite shale contact. In particular,
S009 EG 16.75 9.98 0.70 0.30 8.0 3.0 the transition from samples S001 to S005 is accompanied by a
Notes: hI, hS1, hS2 are the thicknesses of illite (I) and expandable (S1 and S2) layers;
S1 corresponds to 2EG and S2 to 1EG expandable layers, respectively. WI, WS1, WS2
significant decrease in octahedral Mg and interlayer (K+Na) and
are contents of I, S1 and S2 layers. N is the mean number of layers in CSD's. H is a significant increase in interlayer Ca. This trend is reversed with
the parameter of CSD's coherency (the mean distance between parallel CSD's further distance from the lower shale contact going from samples
in megacrystals in angstroms).
S005 to S009. The tetrahedral Al content and charge in average
is about the same in all samples (Fig. 9).
Table 6. Phase composition of the samples in weight percent (<1.0
µm fraction)
Dependence of d(060) and b unit-cell parameter on sample
Sample and treatment R0 I-S R1 I-S Rp location
S001 EG 0 100 7.6 Table 11 and Figure 10 show an increase of the d(060) and
AD 0 100 7.4
S002 EG 10 90 7.0
the b lattice parameter values from samples located near the
AD 8 92 7.5 middle of the bed (samples S004–S006) to those located at the
S003 EG 12 88 6.1 bed margin (samples S001–S003 and S008–S009).
AD 8 92 6.5
S004 EG 34 66 6.3
AD 30 70 6.7 Discussion
S005 EG 40 60 6.9 These results make it necessary to consider several aspects of
AD 42 58 6.7
S006 EG 46 54 6.7 smectite illitization in K-bentonite, related to the specific struc-
AD 46 54 6.6 tural and crystal-chemical features of the coexisting I-S phases
S007 EG 28 72 7.5
AD 24 76 7.5
and to the main factors responsible for the formation of these I-S
S008 EG 23 77 6.1 phases, as well as for the observed mineralogical zonation. The
AD 19 81 7.0 mineralogical sequence of intermediate members of the smectite
S009 EG 29 71 5.5
AD 26 74 5.0 illitization described in this paper will be compared with those
Notes: Rp is the profile factor characterizing the overall fit quality. EG = ethylene described in the literature.
glycol, AD = air-dried samples.
Relationships between the probability parameters
determined for the AD and EG R1 I-S structures
Table 7. Phase composition of the samples for the EG and dehy- Table 2 shows that in the <1.0 µm fraction the Wj and Pij
drated states (in wt%; <0.5 µm fraction) modeling parameters are different for the R1 I-S structures of
Sample and treatment R0 I-S R1 I-S Rp the same sample from EG and AD states. The reason is that the
S001 EG 0 100 6.7
S002 EG 13 87 6.0
AD R1 I-S structures contain two types of expandable layers,
DH 12.0 88 1W and 2W, having different thicknesses, whereas most of the
S003 EG 20 80 5.6 expandable layers in the EG R1 I-S swell homogeneously to
DH 18.5 81.5
S004 EG 45 55 4.5 2EG. However, because the distribution of illite and expandable
DH 42.7 57.3 layers in the R1 I-S should be identical in both states, the three
S005 EG 43 57 5.2 component AD R1 I-S must be reduced to a two-component
DH 40.7 59.3
S006 EG 55 45 5.6 EG R1 I-S when two types of expandable layers in the AD state
DH 52.7 47.3 are replaced by one expandable layer type in the EG state. For
S007 EG 42 58 4.0
DH 39.7 60.3
example, let us consider R1 I-S in which a WI of illite layers is
S008 EG 30 70 5.5 interstratified with a WS1 of 14.9 Å and WS2 of 12.50 Å smectite
DH 28.1 71.9 S1 and S2 layers, respectively. If in the EG I-S all smectite layers
S009 EG 32 68 4.6
DH 30.0 70.0 swell to 16.80 Å, then the following relationships should be valid
Notes: Rp is the profile factor characterizing the overall fit quality. EG = ethylene for three component I-S:
glycol, DH = dehydrated.
WSI = WSPSI = WS1I + WS2I = WS1PS1I + WS2PS3I (7)

Average structural formulae of the samples WIS = WIPIS = WIS1 + WIS2 = WIPIS1 + WIPIS2 =
Table 9 contains elemental oxides determined for the Ca- WI (PIS1 + PIS2) (8)
saturated <0.5 µm size fractions. The average structural for-
mula for each sample was calculated per O10(OH)2 because the WSS = WSPSS = WS1S1 + WS1S2 + WS2S1 + WS2S2 =
<0.5 µm fractions are impurity-free and thus consist of only the WS1 (PS1S1 + PS1S2) + WS2 ( PS2S1 + PS2S2) (9)
1662 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

a b

FigurE 5. R1 and R0 I-S % illite layer content (a), and weight fractions (b) compared with distance from the lower bentonite-shale contact
based on XRD computer simulation from the EG <1.0 µm size fractions.

a b

FigurE 6. R1 and R0 I-S % illite layer content (a), and weight fraction (b) compared with distance from the lower bentonite-shale contact
based on XRD computer simulation from the EG <0.5 µm size fraction samples.

For the AD R1 I-S phases (except the I-S in sample S001), However, comparison of the weight concentrations of these
PS1S1 = PS2S1 and PS1S2 = PS2S2 (Table 2), therefore WSS = WSPSS phases and of the relative areas at the different temperatures
= (WS1 + WS2)( PS1S1 + PS2S2) and as a result PSS = PS1S1 + PS2S2. shows strong disagreement, especially for samples S001, S002,
Here Wij = WiPij is the occurrence probability for an ij layer pair and S003 (Table 8). In contrast, better agreement exists when the
and i,j = I, S1, S2 for the AD three-component, and i,j = I,S for relative areas of the maxima in the DTG patterns are compared
the EG two-component I-S. The validity of Equations 7–9 can with the weighted average amount of illite and smectite layers
be checked for each pair of AD and EG R1 I-S models using the in the R1 I-S and the R0 I-S of each sample. As shown in Table
probability parameters given in Table 2. 8 the values of the illite and smectite layer contents are close to
the corresponding values for the areas under the high- and low-
Origin of the high- and low-temperature maxima in the temperature peaks, respectively.
DTG patterns The two maxima in the DTG patterns at about 470 and about
Two possible interpretations about the origin of the coexisting 640 °C (Table 8; Fig. 8) correspond to the dehydroxylation of
DTG maxima can be considered. One is based on the assump- trans-vacant (tv) and cis-vacant (cv) 2:1 layers (Drits et al. 1995,
tion that the maxima at different temperatures observed in the 1998). This means that in the R1 I-S and R0 I-S, most of the illite
DTG patterns correspond to the R1 I-S and R0 I-S, respectively. and smectite layers are cv and tv, respectively. The coexistence
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1663

FigurE 7. DTG patterns from <0.5 µm Ca2+-saturated samples. FigurE 8. DTG pattern peak decomposition examples showing
relative areas of high- (T1) and low- (T2) temperature peaks, and
composite curve (shaded line) for samples S001 and S005.

Table 8. Comparison of the relative areas under the high- and low-temperature DTG peaks, the total illite and smectite layer content in the
coexisting I-S phases calculated as a weighted average, and the R0 and R1 I-S weight concentration
Dehydroxylation temperature (°C) DTG Profile fit quality Total illite and smectite Phase content (wt%)
layer content (wt%)
Sample T1 T2 Peak area T1 Peak area T2 R (%) Illite Smectite R1 I-S R0 I-S
S001 646 465 74 26 7.54 65 35 100 0
S002 640 468 70 30 7.96 61 39 90 10
S003 640 471 67 33 5.25 58 42 88 12
S004 634 467 56 44 10.33 58 42 66 34
S005 640 467 51 49 9.55 48 52 60 40
S006 634 467 52 48 10.70 47 53 54 46
S007 635 475 57 43 9.23 53 47 72 28
S008 640 469 61 39 8.90 55 45 77 23
S009 636 477 62 38 5.55 52 48 71 29

Table 9. Major element chemistry for the Ca-saturated <0.5 µm size fractions (values in wt%)
Sample Distance from lower contact (cm) SiO2 Al2O3 CaO MgO Na2O K2O Fe2O3 TiO2 P2O5 LOI Total
S001 12.7 53.53 25.40 1.27 3.11 0.24 4.86 2.09 0.06 0.04 7.82 98.42
S002 25.4 54.57 25.31 1.42 3.14 0.13 4.56 2.14 0.06 0.02 7.38 98.73
S003 50.8 55.07 25.27 1.49 3.21 0.21 4.03 2.16 0.06 0.01 7.64 99.15
S004 101.6 54.93 24.93 2.22 2.96 0.08 3.64 1.93 0.07 0.02 8.66 99.44
S005 114.3 54.31 25.76 2.21 2.64 0.09 3.48 1.62 0.07 0.02 8.92 99.12
S006 139.7 55.27 25.53 2.03 2.74 0.14 3.23 1.70 0.10 0.02 8.44 99.20
S007 152.4 54.78 24.55 2.06 2.98 0.11 3.67 1.98 0.11 0.02 8.49 98.75
S008 165.1 54.13 24.86 1.54 3.60 0.08 3.87 1.88 0.08 0.02 8.01 98.07
S009 177.8 55.60 24.77 1.61 3.74 0.11 3.75 2.00 0.07 0.01 8.48 100.14

of tv and cv layers in I-S was discovered by Reynolds (1993) and in I-S indicates a compositional control over the distribution of
often occurs in bentonite I-S (McCarty and Reynolds 1995, 2001; octahedral cations over trans- and cis-sites (Drits et al. 2006). In
Altaner and Ylagan 1997; Cuadros and Altaner 1998a, 1998b). particular, compositional restrictions exist to control the occur-
Crystal-chemical analysis of dioctahedral 2:1 layer silicates rence of cv illite layers. These authors showed that cv 1M illite
consisting of tv and cv layers and interstratified cv and tv layers and cv illite fundamental particles in I-S have a high amount of
1664 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

Table 10. Mean I-S structural formula for <0.5 µm size fraction
Sample Distance from Ca K Na Al(VI) Mg Fe Sum Al(IV), Oct Total nK CR0
dh/C dh
R1

lower contact (cm) Tet Charge Charge Charge

S001 12.7 0.091 0.417 0.031 1.610 0.312 0.106 2.028 0.402 0.229 0.631 0.79
S002 25.4 0.101 0.387 0.017 1.608 0.311 0.107 2.026 0.374 0.232 0.606 0.79 0.136
S003 50.8 0.106 0.340 0.027 1.609 0.316 0.107 2.033 0.360 0.218 0.578 0.78 0.220
S004 101.6 0.158 0.309 0.010 1.611 0.294 0.097 2.001 0.345 0.291 0.636 0.83 0.730
S005 114.3 0.158 0.297 0.012 1.658 0.263 0.081 2.003 0.371 0.254 0.625 0.84 0.672
S006 139.7 0.144 0.273 0.018 1.655 0.271 0.085 2.010 0.339 0.241 0.579 0.81 1.101
S007 152.4 0.148 0.313 0.014 1.605 0.297 0.100 2.002 0.332 0.291 0.623 0.84 0.656
S008 165.1 0.111 0.331 0.010 1.599 0.360 0.095 2.055 0.367 0.196 0.563 0.79 0.395
S009 177.8 0.114 0.315 0.014 1.584 0.367 0.099 2.050 0.339 0.217 0.556 0.81 0.435
Notes: nK = the amount of K+Na cations in illite interlayers of both R1 and R0 I-S phases. CR0
dh/C dh = the ratio of weight concentrations of the R0 and R1 I-S dehydrated
R1

phases.

a b

FigurE 9. Cation content per formula unit (a) and layer charge distribution (b) for the <0.5 µm size fractions based on the mean I-S structural
formulae.

Table 11. The values of d(060) and b parameters for the samples 2006). Based on these data it must be concluded that the 2:1 lay-
compared with distance from the lower shale contact ers that form illite interlayers should be Al-rich in both I-S phases.
Sample Distance from d(060) b (Å) For dioctahedral smectites compositional control is a reasonable
lower contact (cm) 1 2 3 {d(060)mean×6}
assumption. Although tv smectite layers prevail in beidellite, they
S001 12.7 1.5018 1.5017 1.5013 9.010(1)
S002 25.4 1.5018 1.5017 1.5014 9.010(1) also occur in montmorillonite (Drits et al. 2006).
S003 50.8 1.5018 1.5017 1.5016 9.010(1) The stability of cv illite layers in I-S in low-temperature
S004 101.6 1.5008 1.5008 1.5009 9.005(1)
S005 114.3 1.5005 1.5003 1.5005 9.003(1)
natural environments, along with the high content of Al, is prob-
S006 139.7 1.5003 1.5005 1.5001 9.002(1) ably related to the specific structural features of cv 1M illite.
S007 152.4 1.5010 1.5010 1.5009 9.006(1) The interlayer environment in the cv 1M illite is similar to that
S008 165.1 1.5015 1.5015 1.5012 9.008(1)
S009 177.8 1.5014 1.5013 1.5010 9.007(1) in 2M1 muscovite which provides a small difference between
Note: 1: at λ = 1.5418 Å, step 0.005 °2θ; 2: at λ = 1.5418 Å, step 0.005 °2θ, individual K-O basal bond lengths and minimizes the repulsion
smoothing; 3: at λ = 1.54059 Å, step 0.005 °2θ, contribution of Kα2 was removed, of oxygen atoms on the adjacent basal surfaces that form the
smoothing.
illite interlayers (Drits et al. 2006).

Al in the octahedral and tetrahedral sheets of the 2:1 layers. In Structural formulae for the coexisting R0 I-S and R1 I-S
particular, the amount of Aloct and Altet are greater than 1.55 and phases
0.35 atoms per O10(OH)2, respectively, when the content of cv To determine an individual structural formula for the R0 I-S
layers in I-S containing both layer types exceeds 60% (Drits et al. and R1 I-S that coexist in the same sample, one has to know
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1665

a b

FigurE 10. Average octahedral Mg content for the <0.5 µm fractions (a), and b parameters for <1.0 µm fraction compared with distance from
the lower shale contact based on random powder XRD scans at step increments of 0.005 °2θ calibrated with an internal standard (b).

its mole concentration relative to the average formula of the normalized Cdh R0
and CdhR1 values and their ratio for the dehydrated
sample. R0 I-S and RI I-S phases in each sample are presented in Tables
The first step is to determine the weight concentrations of 7 and 10.
the R0 I-S and R1 I-S in their dehydrated states. The weight The structural formulae for the dehydrated R0 I-S and R1 I-S
concentrations of the I-S phases are used for each EG specimen were obtained based on the average structural formula of each
of the sample from the experimental XRD pattern modeling sample and the following assumptions. First, for each sample,
because the smectite layers swell homogeneously (Table 7). the ratio of the molecular weights of the structural formulae of
The corresponding concentrations for the R0 I-S and R1 I-S are the coexisting R1 I-S and R0 I-S is equal to that of the weight
denoted as CR0 and CR1, respectively, and: concentrations of these phases in their dehydrated state. Second,
the sum of octahedral cations per O10(OH)2 in each of the coexist-
CR0 + CR1 = 1, CR0 = CSR0 + CIR0, CR1 = CIR1 + CSR1. ing R1 I-S and R0 I-S phases is the same as that in the average
formula of the sample. Third, the amount of (K+Na) cations in
Here CIR1, CSR1 and CIR0, CSR0 are the weight concentrations of illite interlayers in each of the coexisting phases is the same. This
illite and smectite layers in the R1 I-S and R0 I-S, respectively. assumption is in agreement with the results of Środoń (1999)
The weight concentrations of the R0 I-S and R1 I-S in their showing that the amount of K and Na cations in illite interlayers
dehydrated states were found using Equation 5 (Table 7). In the of I-S is fixed and does not depend on the I-S expandability. This
R1 I-S phase of each sample, the WIR1 and WSR1 values are gener- condition is fulfilled if the amounts of (K+Na) cations in illite
ally equal to 62 and 38%, respectively, independent of location interlayers of the R0 I-S and R1 I-S are equal; that is nKR1 = nKR0.
(Table 2). The formulae of illite and glycolated smectite layers As a first approximation the nKR1 and nKR0 values were assumed
in the I-S phases have been deduced based on the results of XRD to be equal to the number of (K+Na) cations, nKR1 = 0.79, in the
modeling. The illite interlayers contain 0.80 atoms of K per illite fundamental particles of the single R1 I-S phase present in
O10(OH)2, whereas the composition of the glycolated smectite sample S001. This nKR1 value, obtained by Equation 6, and the
interlayers is Ca0.2, and 3.4(CH2OH) 1.3H2O. In addition, the values of N and Wexp from the XRD modeling of the R1 I-S in
main source of the illite negative layer charge is assumed to be sample S001, were used to calculate the mKR1 values for the R1 I-S
located in the tetrahedral sheets and is provided by 0.5 atom of phases of the rest samples using the same Equation 6 and the N
Al per O10(OH)2. Therefore, the formulae of illite and smectite and Wexp corresponding to each R1 I-S structure (Table 4).
layers in their dehydrated and EG states can be presented as: From the equality nKR1 = nKR0 and Equation 6 the ratio, p, of the
mKR1 and mKR0 corresponding to the numbers of (K+Na) cations in
K0.8(Al1.7Fe0.1Mg0.3)(Si3.5Al0.5)O10(OH)2, the structural formulae of R1 I-S and R0 I-S is
Ca0.2(Al1.5Fe0.1Mg0.4)Si4.0O10(OH)2, and
Ca0.2(Al1.5Fe0.1Mg0.4)Si4.0O10(OH)2 3.4(CH2OH) 1.3H2O. mKR1 N R 0 ⋅( N R1 − 1) ⋅WIR1
p= = . (10)
These formulae were used to calculate the molecular mKR 0 N R1 ⋅( N R 0 − 1) ⋅WIR 0
weights of the corresponding layers. Under such conditions This equation was used to calculate the mKR0 value for each R0
Cdh
R1
= 0.887CR1. Using a similar approach Cdh
R0
= 0.807CR0. The I-S phase using the value of mKR1 calculated by Equation 6 and
1666 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

the values of NR1 and NR0 from Tables 4 and 5. octahedral cations in the R0 I-S from samples located near the
let us denote the mole contributions of the formulae of the boundaries is much lower than that in this phase from the other
coexisting R1 I-S and R0 I-S to the average sample formula as samples. However, the amount of the Aloct and Altet in the struc-
WmR1 and WmR0, then the mKR1, mKR0, and mK in the average formula tural formulae of the R0 I-S is lower than 1.55 and 0.35 atoms,
are related by the equations: which is consistent with the low amount of cv illite layers in
this phase (Table 13).
WmR1mKR1 + (1− WmR1 ) ⋅ mKR 0 = mK Because each of the coexisting I-S phases contains a similar
illite layer content, the fixed (K+Na) content is also similar. In
WmR1 pmKR 0 + (1− WmR1 ) ⋅ mKR 0 = mK particular, in the structural formulae of the R1 I-S and R0 I-S, the
number of (K+Na) cations varies irregularly from 0.43 to 0.47
mK − mKR 0 mK − mKR 0 and from 0.19 to 0.22, respectively (Tables 12–13). Therefore,
WmR1 = = .
m ⋅( p − 1)
R0
K
mKR1 − mKR 0 the regular zoned variation of (K+Na) values in the average for-
mulae as a function of the sample depth (Fig. 9) is related to the
The mKR1, mKR0, WmR1, and WmR0 values were used as the starting variation of the relative weight concentrations of the coexisting
parameters and their slight variation has led to the structural I-S phases. The lower the R1 I-S content, such as in the transi-
formulae of the R1 I-S and R0 I-S of each sample, which satis- tion from samples S001 and S009 toward samples S005–S006,
fies the main three conditions mentioned above (Tables 12 and the lower the (K+Na) content in the average formulae (Table
13). In particular, the number of (K+Na) cations, nK, in the il- 10; Figs. 5–6).
lite interlayers of the fundamental particles of the I-S changes In the R1 I-S formulae, tetrahedral Al is significantly higher
from 0.78 to 0.84 per O10(OH)2 (Table 10). Figures 11 and 12 than the (K+Na) content (except sample S002), independent of
show how the cation composition and layer charge of these sample location (Table 12, and charge plot in Fig. 11). Therefore
phases depends on the sample location with respect to the lower it can be assumed that the tetrahedral sheets that form not only
bentonite-shale contact. illite, but also smectite interlayers have a significant negative
charge due to substitution of Al for Si. A similar regularity is
Crystal-chemical features of the R1 and R0 I-S phases observed for the R0 I-S located in the middle of the bed (samples
One of the crystal-chemical features of both I-S varieties is S005–S006). In contrast, tetrahedral Al content in the R0 I-S
the relationship between the Mg content and sample location phases located near the bed boundaries (samples S002, S003,
(Figs. 11 and 12). The Mg cation content in samples S004–S007 S007, S008, and S009) is lower compared with (K+Na) content
is significantly lower than in samples S001–S003, S008, and (Table 13; Fig. 12). In these R0 I-S phases, substitution of Al for
S009 (Tables 12 and 13; Figs. 11 and 12). Note that the higher Si probably took place only in the tetrahedral sheets that form
the octahedral Mg in the structural formulae of both I-S phases, illite interlayers. Tetrahedral sheets of the adjacent 2:1 layers that
the higher the sum of octahedral cations in these two phases. In form smectite interlayers are probably Al-free, and the source
particular, the octahedral cation contents per O10(OH)2 are equal of the total negative layer charge originates in octahedral sheets
to 2.026–2.033 for samples S001–S003 and 2.050–2.055 for due to substitution of R2+ for R3+ cations.
samples S008 and S009. This means that the octahedral sheets The amount of Ca in the structural formulae for the R1 I-S
of the R1 and R0 I-S phases contain small trioctahedral clusters. increases systematically from the margins of the bed toward its
The presence of such clusters along with the regular variation in center. Because no trend exists in the smectite layer content in
the Mg content are in agreement with the noticeable increase of this phase and Ca is the only cation occupying these interlayers
the d(060) and b lattice parameter values in the transition from (Table 12; Fig. 11), the 2:1 layers that form smectite interlayers in
sample S004 to samples S001–S003 and from sample S007 to the R1 I-S located near the middle of the bed have a significantly
samples S008–S009 (Table 11; Fig. 10). higher negative charge than that of 2:1 layers in the R1 I-S near
In the average I-S formulae, tetrahedral Al varies irregularly the bed boundaries. The amount of Ca in the formulae of the R0
from sample to sample within a narrow interval (except sample I-S decreases slightly from samples S002–S005 [0.171–0.176
S001) from 0.33 to 0.37 cations per O10(OH)2 (Table 10; Fig. 9). atoms per O10(OH)2] to samples S006–S009 [0.158–0.138 atoms
In contrast, in the structural formulae of the R1 I-S and R0 I-S per O10(OH)2] (Table 13; Fig. 12).
corresponding to samples S004–S006, tetrahedral Al is higher Two assumptions are needed to explain the crystal-chemical
than that in samples S002–S003 and S007–S009, which are features of the coexisting I-S phases. First, the initial volcanic
located near the bed boundaries (Figs. 11–12; Tables 12–13). ash was altered into tv smectite having a homogeneous Al-rich
Thus, in the middle of the bentonite bed, both coexisting I-S composition throughout the bed. Second, along with K the ac-
phases are characterized by the lowest octahedral Mg content tive role in formation of the I-S phases was due to Mg cations,
and the highest tetrahedral Al cation content. As mentioned, and both K and Mg decreased toward the bed center. In this
the crystal-chemical analysis of the I-S consisting of tv and cv case a relatively high content of K and Mg is responsible for
layers shows that the probability of cv layers is >60% when the the crystal-chemical features of the coexisting I-S phases that
amounts of Aloct and Altet are higher than 1.55 and 0.35 atoms formed near the bed margins. The high concentration of K cre-
per O10(OH)2, respectively (Drits et al. 2006). These criteria ated favorable conditions for the formation of R1 I-S, compared
and the high Aloct and Altet content in the structural formulae of with R0 I-S. The main reaction mechanism involving Mg was
the R1 I-S (Table 12) are in agreement with the 62% of cv illite probably the redistribution of octahedral and tetrahedral Al in
layers composing this phase (Table 4). The amount of trivalent the newly formed 2:1 layers of the R1 I-S and R0 I-S in such
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1667

Table 12. R1 I-S structural formula for the <0.5 µm size fraction
Sample Distance from Ca K Na Al(VI) Mg Fe Sum Al(IV), tet Oct Total
lower contact (cm) charge charge charge
S001 12.7 0.091 0.417 0.031 1.610 0.312 0.106 2.028 0.402 0.229 0.631
S002 25.4 0.083 0.445 0.010 1.655 0.270 0.101 2.026 0.429 0.191 0.620
S003 50.8 0.076 0.425 0.010 1.690 0.242 0.101 2.033 0.443 0.144 0.587
S004 101.6 0.140 0.460 0.000 1.690 0.210 0.101 2.001 0.532 0.208 0.740
S005 114.3 0.140 0.468 0.000 1.690 0.210 0.103 2.003 0.546 0.202 0.748
S006 139.7 0.140 0.448 0.000 1.695 0.225 0.090 2.010 0.532 0.196 0.728
S007 152.4 0.135 0.447 0.015 1.690 0.222 0.090 2.002 0.515 0.217 0.732
S008 165.1 0.090 0.433 0.010 1.690 0.270 0.095 2.055 0.517 0.107 0.623
S009 177.8 0.090 0.416 0.010 1.690 0.270 0.090 2.050 0.485 0.122 0.606

Table 13. R0 I-S structural formula for the <0.5 µm size fraction
Sample Depth (cm) Ca K Na Al(VI) Mg Fe Sum Al(IV), tet charge Oct charge Total charge
S001 12.7
S002 25.4 0.172 0.165 0.043 1.428 0.468 0.130 2.026 0.163 0.389 0.552
S003 50.8 0.176 0.137 0.067 1.417 0.493 0.123 2.033 0.162 0.394 0.556
S004 101.6 0.171 0.201 0.018 1.554 0.353 0.094 2.001 0.211 0.350 0.562
S005 114.3 0.172 0.164 0.021 1.634 0.304 0.065 2.003 0.235 0.295 0.529
S006 139.7 0.146 0.190 0.026 1.636 0.292 0.082 2.010 0.247 0.261 0.509
S007 152.4 0.158 0.208 0.014 1.538 0.357 0.107 2.002 0.187 0.350 0.537
S008 165.1 0.138 0.197 0.011 1.481 0.479 0.095 2.055 0.170 0.315 0.484
S009 177.8 0.142 0.194 0.019 1.456 0.484 0.110 2.050 0.164 0.333 0.496

a b

FigurE 11. Cation content per formula unit (a) and layer charge distribution (b) for the <0.5 µm size fractions based on the R1 I-S structural
formulae (see text for details).

a way that the amount of Al in the tetrahedral sheets that form charge of the 2:1 layers that form the smectite interlayers is
illite interlayers increased at the expense of the substitution of located in the tetrahedral sheets due to substitution of Al for Si.
Mg for Al in the octahedral sheet of the 2:1 layers in the initial The reason for this type of smectite layer charge distribution is
smectite. Increased Al in tetrahedral sheets provided the charge the low Mg cation content available for the reaction. Therefore,
for K fixation in illite interlayers. Simultaneously, the relatively tv 2:1 layers that form smectite interlayers in the R0 I-S located
high substitution of Mg for Al in the octahedral sheets of adjacent near the bed center probably have a cation composition that is
2:1 layers that form smectite interlayers created a negative charge similar to that of tv 2:1 layers of the initial smectite. This means
to compensate the positive charge from interlayer Ca. that the initial smectite had 2:1 layers in which the negative
The samples located near the middle of the bed have a higher charge was located in both octahedral and tetrahedral sheets.
R0 I-S content than the R1 I-S because of the lower available K Under such conditions, the main source of Al in the tetrahedral
content (Figs. 5–6). The significant difference between the R0 sheets that form illite interlayers in the R1 I-S and R0 I-S should
I-S located near the margins and near the center of the bed is originate from the dissolution of some of the 2:1 layers of the
that in the I-S phases from the center part of the bed, a negative initial smectite, whereas most of the layers of the initial material
1668 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

a b

FigurE 12. Cation content per formula unit (a) and layer charge distribution (b) for the <0.5 µm size fractions based on the R0 I-S structural
formulae (see text for details).

either preserved their original composition, or it was modified the regular variation of the relative proportions of the high- and
only a little. If the initial smectite composition was constant low-smectitic I-S phases.
throughout the bed, then the coexisting I-S phases especially in As mentioned, the R1 I-S from all the samples have almost
the samples located near the bed margin would have formed by identical illite layer contents and differ from each other only by
a dissolution-precipitation mechanism. the mode of layer interstratification. This mode is also slightly
different in the R1 I-S varieties from the <1.0 and <0.5 µm frac-
Mineralogical zonation of the K-bentonite bed tions of the same sample. In particular, in the <1.0 µm fractions
Consistent with Altaner et al. (1984), the results from this the degree of disorder in the R1 I-S increases from samples
study show that K progressively migrated from the margins S001, S002, S008, and S009 located near the bed margins
of the bentonite bed toward its center with an associated de- (D1 = 0.47–0.49) to samples S005–S007, located near the middle
crease of the K concentration in the pore fluids. However, the of the bed (D1 = 0.58) (Table 2). In contrast, the degree of disorder
simultaneous decrease in the mean illite layer content in the I-S in the R1 I-S from the <0.5 µm fractions increases continuously
toward the center of the bed was not the result of an incomplete from the lower contact (sample S002, D1 = 0.42) to the upper
reaction caused by a deficient potassium supply as these authors part of the bed at sample S009 with D1 = 0.60 (Table 4). Thus,
assumed. Indeed, at each sample location in the bentonite two in the R1 I-S structures from the coarser fractions, a tendency to
I-S varieties with contrasting layer contents and distributions ordered interstratification exists that correlates positively with the
coexist. Moreover, each of these phases in all samples consists weight concentration of this phase. Such a correlation is absent
of the same or similar content of illite and smectite interlayers in the R1 I-S from the <0.5 µm fractions. This difference is not
independent on location in the bed. In particular, the illite layer related to a deficient potassium supply because the R1 I-S phases
contents in the R0 I-S and R1 I-S varieties from <0.5 µm frac- of both size fractions contain almost the same amount of illite
tions are equal to 30 and 62%, respectively, except the R1 I-S interlayers (Tables 2 and 4). This implies structural control for
from samples S001, S002, and S009 in which the illite layer the layer sequence in the I-S that depends on the thickness of
contents are equal to 65, 64, and 60%, respectively (Table 4). the newly formed I-S crystallites, the thinner the crystallites the
The main difference between the samples at different locations is weaker role of structural control responsible for formation of a
the different weight concentrations of the coexisting I-S phases. relatively stable phase having low potential energy.
Thus the decrease in K concentration in the pore fluids toward
the bed center was accompanied by a successive increase in the Mineralogical sequences of intermediate members of
content of the R0 I-S and a decrease in the content of the R1 smectite illitization
I-S, rather than a progressive decrease in the illite layer content Two different models describing the mineralogical sequences
of a single I-S phase, as assumed by Altaner et al. (1984). Thus, of early diagenetic smectite illitization have been suggested in the
a positive correlation between the total content of illite layers literature. One of them, based on the analysis of XRD patterns,
and the concentration of K cations in pore fluids is related to assumes that smectite illitization is a statistically homogeneous
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1669

and continuous reaction where a single I-S phase steadily increas- 2004) as well as highly smectitic R0 I-S and low-smectitic R1
es its content of illite layers along with a simultaneous increase in I-S or R1 I-S-Ch phases have formed together (lindgreen et al.
the order of their stacking sequence (Moore and Reynolds 1997; 2002, 2008). This does not mean, however, that K-bentonites
Środoń 1999). The other model, based on transmission electron always consist of two contrasting I-S phases. These XRD model-
microscopy (TEM) (Dong et al. 1997; Stixrude and Peacor 2002; ing studies showing coexisting I-S phases clearly demonstrate
Dong 2005), suggests the following mineralogical sequence that normal sample preparation procedures do not reconstitute
describing the transition from smectite to illite: the initial sequence of the layer types in each phase as suggested
by Stixrude and Peacor (2002).
smectite → smectite + R1 I-S → pure R1 I-S → Formation of a physical mixture of the R0 I-S and R1 I-S is not
R1 I-S + illite → illite in agreement with the interparticle model of smectite illitization
of Nadeau et al. (1984) and Nadeau (1998). According to this
The authors of this model assume that the series is not con- model the layer sequence is not within interstratified crystallites
tinuous, and in particular a large gap exists in the layer content and the illitization reaction includes the gradual dissolution
between smectite and R1 I-S. In addition, they emphasize that of individual smectite layers and the growth of thicker illite
“mixed layered illite-smectite” involves phases for which an fundamental particles. Therefore the interparticle model views
ordered R1 I-S is the unique phase, analogous to corrensite naturally occurring I-S as aggregates or stacks of smectite and
(Stixrude and Peacor 2002). The authors of the TEM model sug- illite fundamental particles. When WS > WI, the stacking sequence
gest that the differences in interpretation of XRD patterns and is R0, or if WS < WI and only illite fundamental particles having
TEM observations are caused by differences in sample prepa- different thickness exist, then the stacking sequence is R1 I-S
ration; TEM analysis preserves the initial texture of samples, with MPDO. This theory is in agreement with the concept of
whereas during sample preparation for XRD analysis the sample smectite illitization as a statistically homogeneous and con-
is subjected to treatments that cause separation of layers along tinuous reaction associated with a steady increase of the illite
the smectite interlayers. As a result the initial layer sequence may interlayers in I-S and the simultaneous increase in the order of
be destroyed and after sedimentation a new mutual arrangement the layer sequence. However, according to the recent results
of the layer sequence could be formed involving layers of pure (lindgreen et al. 2002; Claret et al. 2004; McCarty et al. 2008;
smectite in this new artificial layer arrangement. lanson et al. 2009), this commonly accepted model for smectite
Reynolds (1992) showed that the problem is not related to illitization should be revised. In addition, the interparticle theory
sample preparation, and this study shows that the discrepancy cannot account for the formation of a sequence of samples differ-
is a result of incorrect interpretation of the experimental XRD ing from each other by weight concentrations of coexisting I-S
patterns. Application of the multispecimen method in combina- phases, each of which has an identical or similar layer content
tion with modeling of the experimental XRD patterns from and mode of distribution. This argument is in agreement with
samples containing a variety of mixed layered phases has shown criticism of the interparticle theory given by Kasama et al. (2001)
that the commonly accepted model of a continuous series of and Stixrude and Peacor (2002).
interstratified I-S varieties needs to be revised. For example,
the clay fractions of samples taken at different depths from the Factors responsible for coexisting R1 I-S and R0 I-S
Callovo-Oxfordian formation of the Paris basin (Claret et al. phases
2004), from offshore (Aplin et al. 2006; McCarty et al. 2008) To explain the coexistence of these I-S phases, a set
and onshore from Gulf of Mexico mudstones (lanson et al. 2005, of considerations should be taken into account. First, the
2009) consist of a physical mixture of smectite and R0 I-S that physicochemical conditions at which the smectite illitization
contains an anomalous high content of illite layers (65%) for I-S took place were similar throughout the K-bentonite bed and were
with random interstratification. Thus, the coexistence of I-S with equally favorable for formation of the R1 I-S and R0 I-S because
different amounts of smectite or smectitic I-S presented in these each of these varieties contains the same or almost the same
XRD studies confirms the TEM model. However, contradictions amounts and distribution of the layer types. Second, the weight
still exist between the XRD and TEM models because mixtures concentrations of the coexisting I-S phases were determined
of smectite and R0 I-S containing 65% of illite layers determined by the average K content available for smectite illitization at a
by XRD differs substantially from mixtures of smectites and R1 particular location in the bentonite bed. Therefore, preferential
I-S with Wi >Ws as suggested by the TEM model. formation of either R1 I-S or R0 I-S was probably determined by
Moreover, the results obtained in this work show that during a certain concentration of K in the pore fluids during formation
smectite illitization of a thick K-bentonite, a mixture of two I-S of this particular phase. In other words, each of these I-S phases
phases having contrasting contents of illite interlayers and modes formed within a certain range of K concentration because of
of the layer distribution can form in the same sample. This model respective structural stability.
partially correlates with the TEM data because one of the two I-S The K migration from the margins to the center of the K-
phases is the R1 I-S, but the second component is R0 I-S, instead bentonite bed was likely accompanied not only by a decrease in
of pure smectite. Interestingly the results based on XRD model- the mean K content at each sample location, but also by a gradual
ing show that mudstones subjected to early diagenesis consist decrease in the K content within its mean concentration. As an
of a mixture of pure smectite and R0 I-S, with Wi > WS, whereas approximation, it may be assumed that at first the K concentration
during illitization of smectite having a volcanic origin, mixtures in the pore fluids at each given sample location was sufficiently
of high-smectitic and low-smectitic R0 I-S phases (Drits et al. high to form a more stable R1 I-S from a portion of the bentonite
1670 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION

material, but then the K concentration decreased to a low level Upper Jurassic shales from the North Sea and Denmark. Clay Minerals, 32,
351–371.
and formation of the R0 I-S phase became more favorable. Dur- ur- Drits, V.A., Środoń, J., and Eberl, D.D. (1997b) XRD measurements of mean crys-
ing formation of the I-S in samples located near the bed margins a tallite thickness of illite and illite/smectite: Reappraisal of the KÜbler index
high K concentration available for the reaction probably existed, and the Scherrer equation. Clays and Clay Minerals, 45, 461–475.
Drits, V.A., lindgreen, H., Salyn, A.l., Ylagan, R., and McCarty, D.K. (1998)
resulting in a higher content of the R1 I-S compared with that Semiquantitative determination of trans-vacant and cis-vacant 2:1 layers in
of the R0 I-S. illites and illite-smectites by thermal analysis and X-ray diffraction. American
In contrast, because the mean K content in the pore fluids Mineralogist, 83, 1188–1198.
Drits, V.A., lindgreen, H., Sakharov, B.A., Jakobsen, H.J., Salyn, A.l., and Dain-
decreased toward the bed center, the period of relatively high K yak, l.G. (2002a) Tobelitization of smectite during oil generation in oil-source
concentration was shorter further from the margins of the bed, shales. Application to North Sea illite-tobelite-smectite-vermiculite. Clays and
Clay Minerals, 50, 82–98.
and thus the content of the R1 I-S phase progressively decreased. Drits, V.A., Sakharov, B.A., Dainyak, l.G., Salyn, A.l., and lindgreen, H. (2002b)
If these considerations are valid, the formation of I-S with the Structural and chemical heterogeneity of illite-smectites from Upper Jurassic
observed layer content and mode of their interstratification mudstones of East Greenland related to volcanic and weathered parent rocks.
American Mineralogist, 87, 1590–1607.
is controlled by a specific range of K cations available for Drits, V.A., lindgreen, H., Sakharov, B.A., Jakobsen, H.J., and Zviagina, B.B.
reaction, which is restricted at least by a certain low level of the (2004) The detailed structure and origin of clay minerals at the Cretaceous/
concentration of K in solution. Tertiary boundary, Stevns Klint (Denmark). Clay Minerals, 39, 367–390.
Drits, V.A., McCarty, D.K., and Zviagina, B.B. (2006) Crystal-chemical factors
responsible for the distribution of octahedral cations over trans- and cis-sites
acKnowlEDgMEnts in dioctahedral 2:1 layer silicates. Clays and Clay Minerals, 54, 131–152.
The authors thank Chevron ETC for supporting this work. We thank Prince Drits, V.A., lindgreen, H.l., Sakharov, B.A., Jakobsen, H.J., Fallick, A.E., Salyn,
Ezeburio for sample preparation and data collection, and Arek Derkowski and A.l., Dainyak, l.G., Zviagina, B.B., and Barfod, D.N. (2007) Formation and
Edwin Zeelmaekers for their help and constructive comments and ideas. We thank transformation of mixed-layer minerals by Tertiary intrusives in Cretaceous
Gray Thompson for advice and help with sample collection. V.A.D. and B.A.S. mudstones, West Greenland. Clays and Clay Minerals, 55, 261–284.
thank the Russian Fund of Fundamental Research. The authors are also grateful Ferrage, E., Tournassat, C., Rinnert, E., Charlet, l., and lanson, B. (2005a) Experi-
for the constructive reviews by J. Kim, H. Dong, and P.A. Schroeder, which helped mental evidence for Ca-chloride ion pairs in the interlayer of montmorillonite.
to improve the manuscript. An XRD profile modeling approach. Clays and Clay Minerals, 53, 348–360.
Ferrage, E., lanson, B., Sakharov, B.A., and Drits, V.A. (2005b) Investigation
of smectite hydration properties by modeling experimental X-ray diffraction
rEFErEncEs citED patterns: Part I. Montmorillonite hydration properties. American Mineralo-
Ahn, J.H. and Peacor, D.R. (1985) Transmission electron microscopic study of gist, 90, 1358–1374.
diagenetic chlorite in Gulf Coast argillaceous sediments. Clays and Clay Ferrage, E., lanson, B., Sakharov, B.A., Geoffroy, E.J., and Drits, V.A. (2007)
Minerals, 33, 228–236. Investigation of dioctahedral smectite hydration properties by modeling of
Altaner, S.P. and Ylagan, R.F. (1997) Comparison of structural models of mixed- X-ray diffraction profiles: Influence of layer charge and charge location.
layer illite/smectite and reaction mechanisms of smectite illitization. Clays American Mineralogist, 92, 1731–1743.
and Clay Minerals, 45, 517–533. Hower, J., Eslinger, E., Hower, M., and Perry, E. (1976) Mechanism of burial meta-
Altaner, S.P., Hower, J., Whitney, G., and Aronson, J.l. (1984) Model for K- morphism of argillaceous sediments: I. Mineralogical and chemical evidence.
bentonite formation: Evidence from zoned K-bentonites in the disturbed belt, Geological Society of America Bulletin, 87, 725–737.
Montana. Geology, 12, 412–415. Inoue, A. (1986) Morphological change in a continuous smectite-to-illite conver-
Aplin, A.C., Matenaar, I.F., McCarty, D.K., and van der Pluijm, B.A. (2006) Influ- sion series by scanning and transmission electron microscopies. Journal of the
ence of mechanical compaction and clay mineral diagenesis on the microfabric College of Arts and Sciences, Chiba University, B-19, 23–33.
and pore-scale properties of deep-water Gulf of Mexico Mudstones. Clays and Inoue, A., Velde, B., Meunier, A., and Touchard, G. (1988) Mechanism of illite
Clay Minerals, 54, 500–514. formation during smectite-to-illite conversion in a hydrothermal system.
Claret, F., Sakharov, B.A., Drits, V.A., Velde, B., Meunier, A., Griffault, l., and American Mineralogist, 73, 1325–1334.
lanson, B. (2004) Clay minerals in the Meuse-Haute Marne underground Jackson, M.l. (1985) Soil chemical analysis—Advanced Course, 2nd edition, 11th
laboratory (France): Possible influence of organic matter on clay mineral printing. Self-published, Madison, Wisconsin.
evolution. Clays and Clay Minerals, 53, 515–532. Kasama, T., Murakami, T., Kohyama, N., and Watanabe, T. (2001) Experimental
Cuadros, J. and Altaner, S.P. (1998a) Characterization of mixed-layer illite-smectite mixtures of smectite and rectorite: Re-investigation of “fundamental particles”
from bentonites using microscopic, chemical, and X-ray methods; constraints and “interparticle diffraction.” American Mineralogist, 86, 105–114.
on the smectite-to-illite transformation mechanism. American Mineralogist, Kastner, M., Asaro, F., Michel, H.V., Alvarez, W., and Alvarez, W.l. (1984) The
83, 762–774. precursor of the Cretaceous-Tertiary boundary clays at Stevns Klint, Denmark,
——— (1998b) Compositional and structural features of the octahedral sheet in and DSDP Hole 465A. Science, 226, 137–143.
mixed-layer illite/smectite from bentonites. European Journal of Mineralogy, lanson, B., Sakharov, B.A., Claret, F., and Drits, V.A. (2005) Diagenetic evolution
10, 111–124. of clay minerals in Gulf Coast shales: New insights from X-ray diffraction
Dong, H. (2005) Interstratified illite-smectite: A review of contributions of TEM profile modeling. 13th International Clay Conference, Tokyo, Japan, 21–27
data to crystal chemical relations and reaction mechanisms. Clay Science, August, p. 39.
12, Supplement 1, 6–12. ——— (2009) Diagenetic smectite-to-illite transition in clayey sediments: A reap-
Dong, H., Peacor, D.R., and Freed, R.l. (1997) Phase relations among smectite, praisal of X-ray diffraction results using the multi-specimen method. American
R1 illite-smectite, and illite. American Mineralogist, 82, 379–391. Journal of Science, 309, 476–516.
Drits, V.A. (1997) Mixed-layer minerals. In S. Merlino, Ed., Modular Aspects of lindgreen, H.l., Drits, V.A., Sakharov, B.A., Salyn, A., Wrang, P., and Dainyak,
Minerals, 1, p. 153–190. European Notes in Mineralogy, Eötvös University l. (2000) Illite-smectite structural changes during metamorphism in black
Press, Budapest. Cambrian Alum shales from the Baltic area. American Mineralogist, 85,
——— (2003) Structural and chemical heterogeneity of clay minerals. Clay 1223–1238.
Minerals, 38, 403–432. lindgreen, H.l., Drits, V.A., Sakharov, B.A., Jakobsen, H., Salyn, A.l., Dainyak,
Drits, V.A. and Sakharov, B.A. (1976) X-ray Analysis of Mixed-layer Clay Miner- l.G., and Kroyer, H. (2002) The structure and diagenetic transformation of
als, 256 p. Nauka, Moscow (in Russian). illite-smectite and chlorite-smectite from North Sea Cretaceous-Tertiary chalk.
Drits, V.A. and Tchoubar, C. (1990) X-ray Diffraction by Disordered lamellar American Mineralogist, 87, 429–450.
Structures, p. 371. Springer-Verlag, Berlin. lindgreen, H.l., Drits, V.A., Jakobsen, H., and Sakharov, B.A. (2008) Clay mineral-
Drits, V.A., Varaxina, T.V., Sakharov, B.A., and Plançon, A. (1994) A simple tech- ogy of the central north sea upper Cretaceous-Tertiary chalk and the formation
nique for identification of one-dimensional powder X-ray diffraction patterns of clay-rich layers. Clays and Clay Minerals, 56, 693–710.
for mixed-layer illite-smectites and other interstratified minerals. Clays and McCarty, D.K. and Reynolds Jr., R.C. (1995) Rotationally disordered illite/smectite
Clay Minerals, 42, 382–390. in Paleozoic K-bentonites. Clays and Clay Minerals, 43, 271–284.
Drits, V.A., Besson, G., and Muller, F. (1995) Structural mechanism of dehydroxyla- ——— (2001) Three-dimensional crystal structures of illite-smectite minerals in
tion of cis-vacant 2:1 layer silicates. Clays and Clay Minerals, 43, 718–731. Paleozoic K-bentonites from the Appalachian Basin. Clays and Clay Miner-
Drits, V.A., Sakharov, B.A., lindgreen, H., and Salyn, A.l. (1997a) Sequential als, 49, 24–35
structure transformation of illite-smectite-vermiculite during diagenesis of McCarty, D.K., Drits, V.A., Sakharov, B.A., Zviagina, B.B., Ruffell, A., and Wach,
 McCARTY ET Al.: NEW INSIGHTS INTO SMECTITE IllITIZATION 1671

G. (2004) Heterogeneous mixed-layer clays from the Cretaceous greensand, fitting. Clays and Clay Minerals, 47, 555–566.
Isle of Wight, southern England. Clays and Clay Minerals, 52, 552–575. Schroeder, P.A. (1992) A multiple reaction mechanism (MRM) model for illitiza-
McCarty, D.K., Sakharov, B.A., and Drits, V.A. (2008) Early clay diagenesis in tion during burial diagenesis. In K. Nagasawa, Ed., Clay Minerals and Their
Gulf Coast sediments: new insights from XRD profile modeling. Clays and Natural Resources. 29th International Geological Congress Workshop WB-1,
Clay Minerals, 56, 359–379. Kyoto, Japan, 79–88.
Moore, D.M. and Reynolds Jr., R.C. (1997) X-ray Diffraction and the Identifica- Shutov, V.D., Drits, V.A., and Sakharov, B. (1969) On the mechanism of a post
tion and Analysis of Clay Minerals, 2nd edition. Oxford University Press, sedimentary transformation of montmorillonite into hydromica. Proceedings of
New York. the International Clay Conference, Tokyo, Israel University Press, 523–531.
Mudge, M.R. (1970) Origin of the disturbed belt in northwestern Montana. Geo- Środoń, J. (1980) Precise identification of illite/smectite interstratifications by X-ray
logical Society of America Bulletin, 81, 377–392. powder diffraction. Clays and Clay Minerals, 28, 401–411.
Nadeau, P.H. (1998) Fundamental particles and the advancement of geoscience: ——— (1981) X-ray identification of randomly interstratified illite-smectite in
Response to “Implications of TEM data for the concept of fundamental par- mixtures with discrete illite. Clay Minerals, 16, 297–304.
ticles.” Canadian Mineralogist, 36, 1409–1414. ——— (1984) X-ray powder diffraction identification of illitic materials. Clays
Nadeau, P.H., Wilson, M.J., McHardy, W.J., and Tait, J.M. (1984) Interstratified and Clay Minerals, 32, 337–349.
clays as fundamental particles. Science, 225, 923–935. ——— (1999a) Use of clay minerals in reconstructing geological processes: Recent
Rampino, M.R. and Reynolds, R.C. (1983) Clay mineralogy of the Cretaceous- advances and some perspectives. Clay Minerals, 34, 27–37.
Tertiary boundary clay. Science, 219, 495 – 498. ——— (1999b) Nature of mixed-layer clays and mechanisms of their formation and
Reynolds, R.C. (1992) X-ray diffraction studies of illite/smectite from rocks, <1 alteration. Annual Review of Earth and Planetary Sciences, 27, 19–53.
µm randomly oriented powders, and <1 µm oriented powder aggregates: Środoń, J., Eberl, D.D., and Drits, V.A. (2000) Evolution of fundamental particle-
The absence of laboratory-induced artifacts. Clays and Clay Minerals, 40, size during illitization of smectite and implications for reaction mechanism.
387–396. Clays and Clay Minerals, 48, 446–458.
——— (1993) Three-dimensional X-ray diffraction from disordered illite: Simula- Środoń, J., Drits, V.A., McCarty, D.K., Hsieh, J.C.C., and Eberl, D.D. (2001)
tion and interpretation of the diffraction patterns. p. 44–78. In R.C. Reynolds Quantitative X-ray diffraction analysis of clay-bearing rocks from random
and J. Walker, Eds., Computer Applications to X-ray Diffraction Methods. preparations. Clays and Clay Minerals, 49, 514–528.
Workshop lectures, 5, The Clay Minerals Society, Bloomington, Indiana. Stixrude, l. and Peacor, D.R. (2002) First-principles study of illite-smectite and
Sakharov, B.A. (2005) Improved model for simulation of experimental XRD implications for clay mineral systems. letters to Nature, 420, 165–168.
patterns for clays including mixed-layer clay minerals. The Clay Minerals
Society 42nd Annual Meeting, June 11–15, Burlington, Vermont, Program
and Abstracts, p. 95. Manuscript received March 28, 2009
Sakharov, B.A., lindgreen, H., Salyn, A.l., and Drits, V.A. (1999) Determination Manuscript accepted July 30, 2009
of illite-smectite structures using multispecimen X-ray diffraction profile Manuscript handled by paul schroeder