THE AMERICAN
MINERALOGIST,
VOL. 50, JULY_AUGUST'
1965
AN X-RAY STUDY OF AN ETHYLENE GLYCOLCOMPLEX
MONTMORILLONITE
RosBnr C. RovNor.os, JR., U. S. Army Cold RegionsResearchand
Engineer'ingLaborotory and Department of Geology,Dartmouth
College,H anozter,!{ ew H amPshire.
Assrt.{cr
oriented aggregates of ethylene glycol-montmorillonite were studied by :r-ray difiractometer methods. Basal reflections, through the 00 14, provided the basis for structural
analysis by Fourier and trial and error methods. On the basis of the intensity data, it is concluded that glycol molecules form a staggeted, two-layered complexl u'ater molecules and
exchangeable cations lie close to but not coincident with a plane that separates the two
glycol layers. The glycol molecules in each layer are disposed in c-axis face centered array.
They are oriented so that the plane of symmetry of the aliphatic chain parallels the c-axis.
The proposed structure is a tlvo-layer modification of a previously described glycoi-vermiculite structure.
INrnolucrroN
Bradiey (1945) first applied the technique of one-dimensionalFourier
synthesisto the study of organo-montmorillonitecomplexes.His work
disclosedthat ethylene glycol moleculesform a two-layer sheet between
the mica portions of the montmorillonite lattice. MacEwan (1948) studied the thicknesses,by r-ray methods, of the organic portions of a wide
variety of organo-montmorillonite and organo-halloysite complexes.He
concluded that the glycol moleculesin each layer of the montmorillonite
compiex were oriented so that the plane ol the zig-zag of the aliphatic
chain lay parallel to the surface oxygen network of the interlamellar clay
surface.MacEwan (1948) visualized the organic layer as a two-dimensional Iiquid without strict crystallographic regularity. Mackenzie
(1948) shorvedthat water moleculesmay proxy, to a certain extent' for
glycol in a typical complex.Later studiesby Brown (1950)indicated that
the exchangeablecations lie along the center of the complex,i.a', in the
plane of symmetry.
This model of the glycol-montmoriilonite complex has not been refined
or modified by recent work. Most oriented montmorillonite aggregates
produce no more than eight or nine basal reflections. Fourier syntheses
based upon so few terms do not provide the requisite resolution for
unique solutionsof the positionsand orientationsof the glycol molecules.
In a recent paper, Bradley et al. (1963) have proposeda different structure for the glycol-vermiculite complex. A single crystal of vermiculite
produced a sufficiently intense diffraction pattern to enable the recording
of 16 ordersof reflection.The d(001) for the single-layervermiculite complex is lessthan d(001) for the two-layer montmorillonite complex;there-
ETHYLEND GLYCOL-MONTMORILLONITE
COMPLDX
991
fore, the d-spacingsof the higher order terms are small enough to allow
the resolution of the individual scattering centers within the glycol molecules. The structure proposed by Bradley et al. (1963) consists of glycol
moleculesoriented so that the plane ol the zig-zag of the aliphatic chain
is perpendicularto the basaloxygensheetand parallel to the b-axis.
Experimental work described below indicates that a two-layer modification of a similar structure is probable for glycol-montmorillonite. Attempts to obtain closeagreementbetween observedand calculated F factors, using MacEwan's model, have been unsuccessful,
Since the present paper was written, the author's attention has been
drawn to an earlierpaper by Brindley (1956) on the structure of the glycol-allevardite complex. Brindley suggested that the glycol molecules
form a two-layer complex in which the planes containing the aliphatic
chains lie parallel to the c-axis. The details of his Fourier projection in the
interlamellar region are almost identical to those shown in Fig. 2 (of the
present paper) for glycol-montmorillonitel apparently glycol-allevardite
and glycol-montmorillonite complexeshave similar or identical interlam e l l a rs t r u c t u r e s .
Lanonarony Pnocnounps
Ethylene glycol-montmorillonite complexeswere prepared from Clay
Spur Bentonite (A.P.L std. No. 26). The montmorillonite was dispersed
in 0.01 N sodium pyrophosphateand settled to obtain a (2 p(e.s.d.)
fraction. The clay fraction was warmed in 0.05 N HCI (to remove iron
oxidesand carbonate),washed,and titrated with Ca(OH)2.The calcium
saturated clay was washed to dispersion.
Oriented specimensIor r-ray studies were prepared as follows. Portions
of the dispersedmontmorillonite were centrifuged on to glassslides. The
method is similar to one describedby Kinter and Diamond (1956).Four
thin montmorillonite aggregates,1?X1 inches, were prepared in this
fashion and dried at 110" C. No special precautions were taken to ensure
complete water removal. The dried aggregates were then steeped in
ethylene glycol (at 110' C.) until the transparent appearanceof the specimen indicated glycol imbibition. The clay aggregateswere peeledfrom the
slides and carefully stacked, one on another, to produce a laminated
specimen(P6zent and M6ring, 1954).The purposeof this procedureis to
produce highly oriented specimensthat exceedthe requirements of infinite thickness for CuK" radiation.
X-ray diffraction studies were performed using a General Electric
XRD-5 spectrometer equipped with a copper tube. The radiation detector system utilizes a GE No. 6 proportional counter tube. This detector
tube is linear to counting rates far in excessof any observed, therefore,
992
ROBERTC, REYNOLDS,TR.
no correction was required for coincidence losses.Diffraction patterns
were obtained using a 0.4" beam slit for the region between 20:5" and
20: 12", a 1obeam slit between20: 10" and 2d: 50o, and a 3o beam slit
between20--40" and20:90o. The use of theseslits in conjunction with a
1$ inch sample ensuresthat the beam width never exceedsthe length of
the sample (Klug and Alexander, 1954).
For the first eight orders, peak heights were scaled directly from the
charts and consideredto be proportional to peak intensities. The higher
orders exhibited some peak broadening (probably due to imperfect
CuKal - CuKa2 resolution); intensities for these reflectionsare based
on peak areas.All peak intensities were put on the same basis by measuring several peaks, at the limits of each 20 range, by the different methods
and slits and calculating appropriate intensity conversionfactors. Reproducibility studies on difierent specimensindicate a standard deviation of
approximately 57o for the stronger peaks, and approximately 10/6 for
t h e w e a k e ro n e s( e . g . , 0 01 1 , 0 0 1 2 , 0 0 1 3 ) . F o r t h e v e r y w e a k p e a k s( 0 0 7
and 00 14) the standard deviation may be as great as 30/6. A d(001)
value of 16.8610.01 was calculatedfrom the sevenhighest orders.
Mrrnols
ol CALcULATToN
The method of structure analysisusedhere consistsof the adjustment
of trial structures toward optimum agreement between calculated F factors (F") and observedF factors ( ] F" l) A GeneralElectric 235 computer
was used (1) to generate F" values for different models of the montmorillonite-glycol complex, (2) to compute Fourier electron density projections, and (3) to compute Fourier difierencesyntheses(Cochran, 1951).
Final adjustments of the various model parameterswere made solely on
t h e b a s i so f c l o s e sat g r e e m e nbt e t w e e nI E " I a n d l n " l . f n . r e f i n e m e npt r o cessconsistedof minimizins the function
V>lF"-F"1'
This method of structure analysis is much more time-consuming than
straightforward interpretation of Fourier projections, but the ultimate
attainable resolutionis higher (Hughes, 1941).Furthermore, the method
provides direct information on the degree to which various portions of
the structure are resolvable. For example, if the trial structure contains
atoms at a specific position in the unit cell, and the Fourier curve of F"
shows no electron maximum at that coordinate, then the investigator
will not expect to observe the position of these atoms on the Fourier
curve of Fo. Conversely, the method allows the rapid distinction between
real electron maxima and termination-of-seriesripples on the projections
(Cochran, 1951). Peaks on the Fourier projection of F" where no atoms
ET H Il LEN E GLY COL-MON TMORI LLON I TE COMP LEX
993
occur in the trial structure show that thesemaxima are spuriousand can
be disregarded on Fo curves. The usual method for identifying termination-of-series ripples is to reduce F factors by an artificial temperature
factor of the form s-A sin'd(Klug and Alexander,1954).This method does
indeed remove spurious ripples, but it also seriously compromises the
resolutionof valid details of the structure (Cochran. 1951).
RBsurrs
Table 1 summarizes the proposed model for the structure of glycolmontmorillonite. One-half of a unit cell is shown. The composition of the
T.clr,r 1. PnolosnoSrnucrunr ron Ernvr,rnr Gr.ycor-MoNruopJt.I.oNrrE
Atomic Coordinates
inA
Atoms
Temperature
Factor B, in A2
0
1.54Al
0 .1 6F e
0 . 3 3M g
1.68
1.06
40
2oIJ
1.68
2.70
3Si
1.68
3.27
60
1. 6 8
6.12
1.70HrC-HO
11.0
7. O 7
1.70cHr-oH
11.0
7 .94
0.80r{ro
0 . 2 0C a
1.68
mica portion of the structure is taken from Kerr et el,. (1950). The coordinates of the atoms in the mica portion of the lattice are modi6.ed
from those listed by P6.zeratand M6ring (1954). Neutral atom scattering
factors (Ibers, 1962) were used in all calculations, and the random powder Lorentz-polarization factor
7 I cosz20
2 sin2 0 cos d
was used in reducing all intensity data to I F.l factors.
Figure 1 shows a projection of the structure on to the Dc plane. The
glycol molecules have been positioned solely to rationalize volume re-
994
ROBDRTC. RDYNOLDS,JR.
quirements with relative electron densities.The structure of the glycol
moleculesin the interlamellar spaceis simply a tivo-layer modification of
the glycol-vermiculitestructure (Bradley et al ., 1963).The glycol molecules are oriented with their plane of symmetry parallel to the Dcplane
(or along a 110 analog) in c-axisface-centeredarray. The secondlayer of
glycol is displacedin the 6 direction,leadingto a displacementof approximately I an oxygen diameter between opposite oxygen networks of the
S C A L E
l-rc. 1. Proposed structure of the glycol layer projected on to a plane parallel to bc.
Water-calcium positions may be at either position shown at the center of the figure. Glycol
molecules at the distance a/2 in front of and behind the plane of the projection are indicated by dashed aliphatic chains.
clay surface.An alternative situation involves displacementsin both the
a and b directions.The r-ray data do not allow the selectionof the most
appropriate of these stacking possibilities; quite arbitrarily, the former is
assumed here.
Water molecules and exchangeablecations (up to the exchange capacity) occupy spacesbetween the two glycol sheets.These positions are
close to, but not exactly coincident with the center of the complex.
Actually, the available space would allow the H2O, Ca to be anywhere
betweenthe two positionsshown on Fig. 1, but the r-ray intensitiesagree
better with a distribution of sitesup and down against the OH groups of
adjacent glycol molecules(Table 3). Figure 1 showsthat the limiting di-
],:.THYLENDGLYCOL-MONTMORILLONITE
COMPLEX
995
mension of the voids between glycol sheetsis approximately 2A. Thus
calcium ions fit well, but water moleculesthat act as rigid sphereswith a
radius similar to oxygen (i.e., -1.4 A) .atr not be accommodated.However, it should be noted that slight adjustments in the relative displacement of the glvcol sheets cause proportionately large increasesin the
limiting dimensionof the voids. Furthermore,no accounthas been taken
of the efiects of polarization and hydrogen bonding, which would cause
the water moleculesto depart from spherical symmetry. In addition,
only 3.4 out of four possibleglycol positionsappear to be occupied,indicating that somewhat more void spaceis available than that shown on
Fig. 1. The present data, and the resulting analysis, are not sufficiently
preciseto evaluate these possibilities.Nevertheless,it is concludedhere
that the geometric limitations of the model do not preclude the occupancy of available voids by water molecules.There are two H2O, Ca positions per unit cell. Proper adjustment of the calculated F factors requires
that ions (or molecules)with scatteringfactors of -10 (e.g.,FI2O,
Na+, or
i Cu'*; occupy 2.4 such positions. This is in excessof the available positions. But, if the complex is assumedto contain 1.6 water moleculesand
0.4 calcium ions per unit cell, then the required scattering power is approached without violating cation exchange capacity or steric limitations. Because all available spaces appear to be filled, the water molecules must occupy most of the positions, for if all of the positions were
filled with calcium or sodium, the cation exchangecapacity of the montmorilionite would be greatly exceeded.
The glycol-vermiculitestructure (Bradley et at.,1963)showsthe filling
of available sites near the van der Waals surfaceof the clay's oxygen network. These positions are apparently vacant in the present structure although steric considerations make them suitable for occupancy by exchangeable cations.
Table 2 shows a comparison between calculated F factors and intensities and observed F factors and intensities. I F. I values have been normalized on the basisof a minimum reliability factor; the intensitieshave
beenadjusted so that the strongestpeak (001) is near 100.
The reliability factor (R) has been computed for the proposed model.
This value is compared with other possible models (Table 3). The reliability factor is given by
o:
' l u :-
l"l
ZIF"I
The data of Table 3 show that variations of MacEwan's basic model are
in poor agreementwith observedF factors. The agreementis substantially better for the proposal structure. Furthermore, the data provide
ROBERT
Tl.;rl-r2.
C. RDYNOLDS,
aNo Cer-cularrn F Fecrons lxn IxtrxsrltEs
Olsrnvro
Ernvr,rNr Gr-vcor--MoxrMoRrr.loNrrE Coltpr-nx
63.4
24.6
36.2
20.9
70.9
4 7. 2
-L6'
100
'
J./
-.t-)..1
-20.8
(
r7t
+4s.7
6-5
38.7
.1.) - I
3.6
23.8
24.2
24.8
r,/.o
loR Tr{E
I.
F
lF.l
001
002
003
004
005
006
007
008
009
00 10
00 1 1
00 1 2
00 13
00 t4
JR.
- 6.8
-40 6
-34.9
- 6.0
-L)/.
'
+26.1
-l-11
'.)
+18.6
3.5
0.65
4.O
t.4
o.032
0.47
o.27
N.D.'
0.09
007
0.06
0.03
100
3.43
3.48
0.66
4.9r
1.35
o.021
0.s49
0.308
0.0070
0.090
0.085
0.056
0.031
1Not detected;IFoI assumed
value(Buerger,
to be one-halfthe minimumobservable
1960,p. 587).
cations andf or water at 7.9+ L
some basis for locating the exchang-eable
instead of in the 180oplane (8.43 A). The data of Table 3 indicate that
the proposed model is to be preferred over MacEwan's. Reliability factors for structural models based on the single crystal Lorentz fa"ctor
andf or lower temperature factors for the organic layer gave much higher
values for R; therefore,they have not been consideredhere.
Figure 2 shows a one-dimensionalFourier electron density projection,
and the final differencesynthesis(p"-p"). The good agreementbetween
Fo and F" is shown by the low amplitude of the difference synthesis
maxima and minima. At no position does the value for po-pc equal or
exceedtwo electronsper A, and the average deviation of po-p" about
zero is +0.78 electronsp.t A. These data refer to a unit cell containing
-500 electrons, whose formula is 2 {AhSinOto(OH)r'1.7(CHtr(OH),
. 0 . 8 H , O' 0 . 2 C a 1 .
T.lsre 3. A CowanrsoN or Rellq.nrlrrv Facrons lon Snvpnel Srnucllrrnar,Molels
0.4 Ca per
Unit Cell,
at 1800
Aliphatic Chain Flat (MacEwan, 1948)
Aliphatic Chain Upright
0 .2ll
0 . 116
1'6 IIrO, 0.4 Ca 1.6I{:O' 0.4 Ca
per Unit Cell at per Unit Cell at
1700
1800
0. 198
O.126
0.042
ETH VLEN E GLY COL-M ON TMORI LLON I TE COMP LEX
997
The po curve of Fig. 2 shows several apparent discrepanciesfrom the
parametersof the proposed model. These may be listed as follows, (1)
the electronic maxima are not coincident with the proposed atomic coordinates,(2) the secondplane of the giycol moleculesshowsa lower electronic density than the first plane, and (3) the H2O-Ca position is not
resolved on the po curve. These discrepanciesoccur to nearly the same
degree on a Fourier projection based on F, values. Consequently,the
discrepanciesprobably arise from termination-of-seriesripples or imper-
N
3ro
F
(n
z
oe
lrJ
o
z40
o
E
F
uJ 20
J
LI
r . 5 4A l
o16Fe
o . 3 3M g
40
2OH
| 70 H2C-HO
1,70CH2-OH
o.e'H2O
lrrt
c s i n pi n A
Fro.2. One-dimensional
Fouriersynthesis
of electrondensityparallelto C sinBfu6).
The curvelabelledpo-p" shorvs
thelast difference
synthesis.
fect resolution of the individual electronic maxima. This conclusion is. of
course,predictable from the low amplitude of the po-pc curve.
DlscussroN oF MoDEL Pan.rlmrnns
The structural model developed above was determined empirically;
each parameter was fixed solely on the basis of best agreement between
l-l,t-r^
I F.l and ll " I . Sucha procedurecanleadtofortuitous agreementfor models
in which one gross error compensatesfor another. No rigorous and absolute verifiiation can be accomplished here. Nevertheless, the varidity of
the proposed model is strengthened by evidence that the model parameters are within or close to limits that have been establishedby other
workers on similar materials.
The temperature factors used are the most uncertain of the model
parameters.Mathieson and Walker (1951) found a value of B:I.2 A2
998
ROBERTC. RI'YNOLDS,JR.
for the water-vermiculite complex. This value is somewhat different from
the value used here (B:1.68 A'). Atomic scatteringfactors for silicates,
listed by Bragg and West (1929), show a decreasewith increasingsin
0/),,that roughly approximatesa B value of 1.4 A'z.This latter value is
c l o s et o t h e o n e u s e dh e r e a n d s u g g e s t st h a t B : 1 . 6 8 A ' - a y n o t b e u n realistic.
Reasonableagreement between calculated and observed F-factors requires the application of a very high temperature coefficient(B:11 A')
to the glycol scattering factors. This value greatly exceedscommonly
used factors for organic crystals. Indeed, the magnitude of B used here
calls for some explanation other than that of thermal vibration.
The high value for B can be explained by a totally different mechanism. If the glycol molecules(Fig. 1) are rotated slightly (clockwise)
about an axis that parallels o and includes the center of gravity of each
molecule, then the methylene and hydroxyl groups would be displaced
from coplanarity. Each of the two planesshown in Fig. 1 would be split
into a pair of closely spacedplanes, producing an effect similar to that of
a single plane affected by high amplitude thermal vibration. It may be
that the rotations are random and are distributed so as to produce the
effect of "fuoze\" thermal motion. However this alternative would randomize the positions of the water-calcium atoms; the non-anomalous B
value for the water-calcium atoms seemsto preclude this possibility. Consequently, it is suggestedthat the high value for the glycol temperature
factor doesnot indicate thermal vibration, but instead, i-ndicatesa slight
but consistent rotation of the glycol molecules; the rotation is in the
plane of the projection of Fig. 1.
Bradley (1954, p. 329) has pointed out that no rigorous analysis has
been made of the intensities of diffraction from oriented aggregatesunder
the conditions of a focusing spectrometer. He found that the most reasonableresults were obtained when the random powder (Debye-Scherrer)
Lorentz factor was applied to intensity data. Schoen (1962) also found
that the random powder Lorentz factor applied to oriented aggregates.
On the other hand, MacEwan et at. (1961) maintain that a factor should
be used that is intermediate between the random powder and the single
crystal Lorentz factors. The writer has found that the use of the random
powder Lorentz factor was necessaryto obtain agreement between observed and calculated intensities for any acceptable model of the glycolmontmorillonite structure. As a further test, a comparison was made
between calculated intensities and observed intensities from a dry sample of sodium Clay Spur montmorillonite. Calculated F factors were corrected for temperature (using B:1.68 A') and converted to intensities
using both forms of the Lorentz-polarization factor. The results are
ETH YLENE GLYCOL-MONTMORI LLON ITE COMPLEX
shown in Table 4. The agreementis good for intensities based on the
random powder Lorentz factor. Intensities computed on the basisof the
single crystal Lorentz factor show large deviations from observed intensities.It is concludedthat r-ray diffractometerwork on oriented clay
aggregatesrequires the application of a factor that is verv closeto the
random powder Lorentz-polarizationfactor.
MacEwan (1948)proposeda "flat" orientationf or the gl.vcolmolecules
becausethe interlamellar clear spacejust accommodatestwo molecular
layers. The a-b area allows 1.75 moleculesper layer per unit cell or 3.5
glycol molecules per unit cell. This- figure is similar to the minimum
amount of glycol requiredfor a 17.0A d(001) spacing(Mackenzie,1948).
T.tsln 4. A CoupenrsoN
BrrwrnN OesrnvrnINrrNsrrroseNoTnosnCercur-arun
UsrNc rnr Two Fonlrs ol rrrn Lonrxtz-PolmrzattoN
Facron. Dnv Sooruu
Cr,lv Spun Mowrltonrlloxrrr;
d(001) :9. 7A
Random Powder
L-P
Iob"
I*t"
001
o02
003
004
005
100
270
54.8
Not Detected
8.2
100
2 7. 3
47.r
0.0002
8.6
Single Crystal
L-P
Ic.tc
100
5 4 .6
141.3
0.0010
12.8
The glycol orientation proposedhere also is consistentwith the available
interlamellar volume. The methylene groups of the gll.col moleculesseat
into holesin the clay oxygen surfacenetwork (Bradley et al., 1963).Two
layers of glycol, offset as shown in Fig. 1, provide the correct thickness
for a d(001) value of 16.9or 17.0 A, and the a-b areais sufficientfor 1.70
moleculesper unit cell per layer or 3.4 moleculesper unit cell. Therefore,
either model is satisfactoryfrom a steric or stoichiometricpoint of view.
The selectionof the more appropriate of the two must be basedon other
consideratiorrs,
e.g.,r-ray intensity data.
CoNcrusroNs
The modei proposedherefor the structure of the ethylene-glycol-montmorillonite complex gives calculated F factors that agree well with observed F factors. The random powder Lorentz factor has been shown to
be applicableunder the experimentalconditionsused. Temperature factors appear to be closeto values establishedby other workers for similar
systems,or can be explainedby reasonablestructural arrangements.The
structure is consistentwith steric and stoichiornetricrequirements.Con-
1000
ROBERT C. REYNOLDS, JR.
sidering all of the evidence, it is concluded that the organic layer of the
glycol-montmorillonite complex contains 3.4 glycol molecules per unit
cell, arranged in two layers; the glycol moleculesare oriented so that
their plane of symmetry parallels the b-c plane of Lhe crystal; water
moleculesand exchangeablecations occupy positions just above and below an o-b plane that bisectsthe organiclayer.
AcrNowr-BnGMENTS
The author wishesto expresshis thanks to Dr. E. H. Mclaren whose
comments concerning an original manuscript aided the author in a reinterpretation of some aspectsof the structure. Dr. J. B. Lyons critically
read the manuscript and contributed materially to its final form.
Rnrrnrrcns
Buolnv,
W. l'. (1945) Molecular associations between montmorillonite and some polyfunctional organic liquids. f ow. Am. Chem. 50c.67,975-981.
(1954) X-ray difiraction criteria of chloritic material. Proc. SecoruJNat. ConJ. on
Clays and,Clay Mi,nerals, Nat. Acad. Sci.-Not. Res.CounciJ Publ. 327,324-334.
E. J. WErss exo R. A. Rowr,.lxo (1963) A glycol-sodium vermiculite complex.
Proc. Tentlr Nat. Conf . on Clays and, Clay Minerals, Pergamon Press, New York, 117tt2.
Bnecc, W. L. ,LNo J. Wnsr (1929) A technique for the examination of crystal structures
with many parameters. Zeit. Krist.69, 118-148.
Bnrurr.rv, G. W. (1956) AJlevardite, a swelling doubleJayer mica mineral. Am. Mineral.
4t, 91-103.
BnowN, G. (1950) A Fourier investigation of montmorillonite. Clay Mineral, BuIl'. 4, L09111.
Bunnctn, M. J. (1960) Crystal-Structure Analgsis. John Wiley and Sons, fnc., New York.
Cocntor, W. (1951) The structure of pyrimidines and purines. V. The electron distribution in adenine hydrochloride. ,4cta Cryst,4,8l-92.
Huctts, E. W. (1941) The crystal structure of melamine. f our. Am. Chem. 5oc.63r 7737'
1 75 2 .
Innns, J. A. (1962) Atomic scattering factors, in IntrernationaL TablesJor X-Ray Crysta'l'l'ography III, K. Lonsdale, Ed., Kynoch Press, Birmingham,20l-2@.
Knnn, P. F., P. K. H,runroN, R. J. Prr,r,, G. V. Wurnr.nn, D. R. Lowrs, W. Bunr<rr.lror,
D. RnNo, G. L. T,lvron, R. C. MrrleNz, M. E. KrNo .lNn N. C. Scnrnr,tz (1950)
Analytical data on reference clay materials. Amer. Petr. Inst. Proj. 49, Preli.m. Repl.7.
KrNrrn, E. G. eNo S. DrauoNn (1956) A new method for preparation and treatment of
oriented-aggregate specimens of soil clays lor *-ray difiraction analysis. Soil,. Scd.8l,
r1t-120.
Kr.uc,H. P. enn L. E. Ar-exeNorn(1954)X-RayDi.fracti.on
Proceilures.
JohnWileyand
Sons, fnc.. New York.
Maturnsox, A. McL. eNo G. F. W,c.r,KEn(1954) Crystal structure of magnesium-vermiculite. A m. M i,neral. 39, 23 1-255.
MacEwax, D. M. C. (1948) Complexes of clays with organic compounds. Part I. Complex
formation between montmorillonite and halloysite and certain organic liquids. ?rozs.
Far ailay S oc. 44, 349-367 .
ETHYLENE GLYCOL-MONTMORILLONITE
COMPLDX
---
1OO1
A. R. Anrr, eNo G. BnowN (1961)Interstratified clay minerals,in The X-Ray ldentif.cation and, Crystal Structilres of Clay Mi.neral;, G. Brown, Ed., Mineral. Soc.,
London,393-455.
MacrrNzrn, R. C. (1948)Complexesof clayswith organiccompounds.Part IL lnvestigation of the ethyleneglycol-water-montmorillonitesystemusing the Karl FischerReagent. Tr ans.F arad.aySoc. 44, 368-375.
P6znnar, H. aNn J. Mfnrl{c (1954)Influence des substitutionsisomorphessur les parambtresde structure desphyllites. Cl,oyMin. Bull.2, 156-162.
Scnorx, R. (1962) Semi-quantitative analysis of chlorites by *-ray clifrraction.Am.
M in eraJ.47, 138+-1392.
publication,March3,1965.
Monuscriptrecehed,,
December
21,1964;accepted,Jor