American Mineralogist, Volume 58,pages 73-80' 1973

X-RayDiffractionAspectsof Montmorillonites
RooNev TBrrBNnonsr, Department ol Geology and Mineralogy,
The Ohio State (Jniuersitv. Columbus, Ohio 43210
o"o

HBnulN E. RossnsoN, Department of Geology, State [Jniaersity of New York

at Binghamton, Binghnmton, New York 13901

Abstract

The breadth and position of the basal X-ray diffraction maxima were analyzed for nine
montmorillonites and hectorite. Some samples were size-fractionated and all were subsequently
treated with ethylene glycol and/or heated to 450"C. None of the samples gave a rational
series of orders, a result believed to be due to the thinness of the particles among other factors'
Tho apparent particle thicknessesdecreasedwith sedimentation size fraction and decreasedfol-
lowin! ireatmint of the clay suspensionswith an ultrasonic probe. The identity of the inter-
layer cation made no change in the particle thicknesses.Wyoming montmorillonites contained
on the average more silicate layers per particle (N) than did the others' Surface glycol ap-
peared to .ont.ibutu less to the diffraction pattern than the glycol bounded on both sides by
silicate layers. With one exception the glycol-treated clays gave N values smaller than those
obtained from the same clays that had been heat-treated. This result may be caused by incom-
plete penetration of the glycol molecules between silicate layers and consequent interstratifica-
tiott, or to some other factor that resulted in relatively broad reflections from the glycol-treated
clays. The heat-treated clays are believed, therefore, to provide a more reliable estimate of N
than the organic-treated clays.
Theoretical analyses were made of the extent of irrationality of the basal orders of diffrac-
tion as a function of particle thinness and layer structure factor. The overali agreement of the
experimental and theoretical data was not impressive. The apparent lack of agreement was
paitially due to the chemical variations among the clays and may reflect individual difterences
which were not accounted for in the theoretical computations, e.g., size distribution or inter-
stratification.

Introduction than particle thinnesssuch as layer disorder or in-

terstratification.
Ross (1968) and Reynolds(1968) have shown The main objectiveof this study was to determine
theoreticallythat the apparentspacingsof finite par- if a relationship existed between particle thickness
ticles deviate from the true values as the particles and basal X-ray spacingsfor montmorillonites,and
becomethin. The truo spacingsare thosethat would to provide an overall view of the group which might
be exhibited by a large crystal containingmany unit suggestfuture lines of investigation.Montmorillon-
cells. The variations in spacing are such that a se- ites were chosen for the following reasons: ( 1) A
quenceof basal reflectionsmay be irrational owing seriesof size-fractionatedsamplesthat were charac-
to particle thinness.The direction of the spacingshift terized by other methodswere available (Roberson,
and its magnitudedependson the natureof the crystal Weir andWoods,1963). (2) Montmorilloniteslikely
surface (Tettenhorstand Reynolds, 1971) among were co,mposedof particles containing fewer layers
other factors. (unit cells) than the substancesstudied previously'
Ross(1968)comparedexperimental datawith the- Smaller particleswould provide larger spacingshifts
oreticalpredictionsfor graphite,and Reynolds(1968) which might prove easierto measureexperimentally'
did likewise for illite. The comparisonsshowedthat (3) Montmorilloniteshave large repeatdistanceses-
spacingdeviationscould be reasonablyaccountedfor pecially when complexed with organic compounds
by thin particles. These investigatorsreiterated that and, therefore, basal spacingsthat occur at srnall
irrational spacingscould result from factors other diftraction angles.Spacingshifts are relatively more
 74 TETTENHORST AND ROBERSON

TSE I. rcWRILLONIIES SMIED

low layer charges,the Otay and Chambers clays
Si2e fracrion (F) DeaigDatlon neferencet
have relativelyhigh layer charges,and the remain-
Fayette (Na)
3
M
ing montmorillonites(Camp Berteaux,Polkville,and
o.05-o. I M
SantaRita) havemediumlayer charges.The Mowry
< o.05 _1 M
clay is the same as that named "Crook" by Rober-
Fryette (O) < o.05 CaF. M

Uptoh (Na) o.r-o.5

son, Weir and Woods (1968). The Otay, Hector,
U^ M

o.05-o. r
Polkville and Santa Rita clays were obtained from
u- M

< o.05 ut
Wards,fnc., and none of thesewas subjectedto an
M

Chanbers (Na) < o.l w,

ion exchangepretreatment.Otay, Hector and Vol-
hry (Na) o.r-o.5 M M
clay are presumedto be mainly sodium-saturated
Canp Bertdu (Na) < 2 naturally. The Polkville and Santa Rita clays are
otay < 2 o presumedto be mainly calcium-saturated. The size
Polkville < 2 P GK fractions indicated in Table 1 for Otay, Polkville,
Hector < 2 H Hector, Santa Rita, and Volclay are approximate.
Seta Rita < 2 SR API The fraction used for theseX-ray studieswas ob-
Volcley < 2 tained by taking the materialthat remainedin sus-
pensionfollowing settling in water for 24 ho.urs.
t CheDical analyses, Btructural forDulas dd othex &tail6 Dey b

found in the foltowing! (mw) RobeisoD, weir and r{oods (1969); Experimental
(W) welr (1965); (GK) crin and Krlbicki (1961); (Apr) herican X-ray
Petrol€q Inrtitute project 49 (l9SO).
The X-ray data were taken on a Philips high-angle
diffractometer equipped with a geiger detector. The goni-
pronouncedat small diffraction angles.(4) No large_ ometer was aligned carefully and calibrated with a sample
scalestudy of particle thicknessof montmorillonites of powdered semi-conductor-puresilicon metal. The posi-
appeaxsto have been carried out. tions of the (001) reflections from a permanently-mounted
The montmorillonitesinvestigatedare listed in Ta- sinall muscovite flake were calibrated with the silicon re-
flections. The low angle peaks from the muscovite were
blb 1 along with the designationsused in this report. used as an external standard to
calibrate the positions of
The main interlayer cation is indicated for those the clay maxima throughout the diftraction analyses.The
clays subjectedto ion exchangeprior to this X-ray clay maxima were scannedautomatically at a rate of one-
study. All of the clays are dioctahedral aluminous quarter degree 20/min A significant portion of the back-
montmorillonitesexcept hectorite. The Hector, Vol_ ground was included on either side of the peaks. A correc-
tion suggestedby Warren and Bodenstein(1965) was made
clay, Mowry, Upton, and Fayetteclayshaverelatively for steeply sloping background
on non-Wyoming mont-
morillonites to determine the position of the (002) peak
TAELE 2. ABSOLI'TEI{A](IU['U AND AVERAGEDAUIATIONS FROM of the ethylene glycol-treatedclays and the (001) peak of
UEAN POSITION AND BREADTEVAI,UES FRCII{REPLIC,ATE
SIPERII,IENTSIN DEGREEST.IilOTHETA
the samples heated to 450"C. The correction was never
more than 0.03' two theta.
Other X-ray conditions were: Ni-filtered CuKa radia-
GLYCOL-TREATED
tion, 1' divergence and anti-scatter slits, 0.006,, receiving
Order Poaltlona Breadths slit, 30 inches per hour chart speed, and a time constant of
.001 0.03 ( M , Fz) 0.05 (Fz)
4. Peak positions and breadths were measured to the nearest
0.01' two theta. A strip of scotch tape was placed over
oo2 .03 (l,l' F2) .03 (F2, B, SR)
the slot in the scatter shield to provide an enclosed chamber.
003 .04 (u) .04 (F3, It, P) The linear range of the detector was determined to be
005 .05 (Fz) .04 (o) about 480 cps for the lowest tube potential used so the
Average scale factor settings were fixed to provide a maximum on-
0.0r2 0.015
scale reading of 400 cps. The power settings (Kv and ma)
were adjusted for all maxima to keep them on scale but
HEATEDTO 450'C
as possible, in order to reduce errors in measure-
PositloEs Breadths
il"lT*"
001 0.04 (caF1) 0.06 (sR)
Clays
002 .05 (CaFr) .07 (E)
The diffraction conditions and the clay specimen dimen-
003 .05 (CaFr) .06 (H, P)
sions were designed to minimize errors in positions and
Average 0.015 0.021 breadths of X-ray maxima inherent in the method of anal-
ysis, Montmorillonite specimens were prepared by evap-
 X-RAY DIFFRACTION OF MONTMORILLONITES 75

orating dilute suspensionsonto glass slides. The mounts were

made as thin as possible to avoid errors caused by sample
transparency with specimens whose absorption coefficient
is low. Most mounts were about 0.001" thick or less as
measured by a micrometer. The length of the samples
parallel to the X-ray beam was about 25 mm. A specimen
length of 25 mm allows for an error in line positions owing
to the flat specimen aberration of not more than about
0.02"20 over the entire range of diftraction angles we
studied. The length and thickness were sufficient to provide |r)
data that were reasonably reproducible (Table 2) within 9

the limits of the power settings for the X-ray tube.

One set of slides was treated with excess liquid ethylene
glycol and maintained in an enclosed chamber containing
additional free liquid glycol, all of which was placed in an
oven at 60"C. for at least several days. An initial attempt
to obtain montmorillonite-glycol data by employing the
vapor pressure complexing method gave non-reproducible
results and was discontinued. The first, second, third, and
fifth order basal reflections were scanned. The first order r5 (A")
reflection was then rescanned as a check on any possible
deterioration of the complexes during X-ray analysis. Some Ftc. 1 Apparent spacingsof (001) for montmorillonite-
specimens were X-rayed periodically up to two months ethylene glycol complexes versus particle thickness. Open
after the initial complex was made with no apparent circles represent experimental data, curves represent com-
change in the diffraction patterns. Ocassionally higher puted data. Solid curve is for particles terminating on basal
orders of diffraction were scanned for the well-oriented o*yg.n clay surfaces; dotted curve is for particles termi-
samples, usually the Wyoming clays and the very fine- nating on glycol molecules.
grained fractions of the others. The samples were scanned
while they were slightly moist with glycol.
mean taken on three separatespe€imens.All other
Another set of slides was heated overnight to 450'C.
Drierite was placed in the X-ray chamber to deter rehydra- data representthe averageof two runs on separate
tion during the scan. The first, second and third basal mounts. Maximum individual and overall average
orders were scanned and the first order was rerun as a deviationsfrom the final mean position and breadth
check. It was hoped to use the same slides for the heating values from the replicate analysesare in Table 2.
experiments that were glycol-treated. Tttis was not possible
The averagedeviationsfrom the mean values were
as these slides when heated gave non-reproducible results
probably becauseof retention of variable amounts of glycol' about the same as the experimentalelTor in estab-
This was so regardless of many attempts to rid the clay of
the sorbed glycol by such means as standing under ambient
conditions for long periods, passing dry air over them,
or preheating to modest temperatures prior to firing at
450'C. Heating to temperaturesbelow 450'C. gave poorer
reproducibility probably becauseof incomplete dehydration.
One such specimen (hectorite heated to 200"C.) is included
for comparison in Figure 4. Heating to temperatures greater
than 450'C. was not investigated because of possible
dehydroxylation and the consequent expansion perpen-
dicular to the layers (Bradley and Grim, 1951) which would o
complicate interpretation of the results. Unfortunately some !
N

of the fractionated clays which were glycol-treated were

not available for the heating studies.

Results
Experimental
The breadths of all maxima were measured at
half-maximum intensity and the two theta positions
t5(A')
were determinedas the midpoint of the chord drawn
at half-maximum intensity. The fractionated Upton Frc. 2. Apparent spacings of (002) for montmorillonite
and Fayette glycol-treatedclay data represent the ethylene glycol complexes versus particle thickness.
 TETTENHORST AND ROBERSON

lishing the two theta values.The glycol-treatedclays thicknessof the unit layer. The extent of the irra-
gave somewhat more reproducible results. The tionality of the basal reflectionswas determinedby
method chosento representthe data in the figures subtracting these latter unit layer thickness values
was based on the following considerations. from those apparent layer thicknessesdetermined
Three facts were apparent in the data: ( 1) Re- from other reflectio'nswhose positionswere affected
gardlessof pretreatmentnone of the montmorillon- by the number of layers per particle.
ites showed an integral or rational series of basal The experimentaldata are plotted as open circles
reflections. (Complexes with other organic com- in Figures 1-4. Figures L-3 representa summaryof
pounds,e.g., tetraethyleneglycol dimethyl ether, also the clay-glycoldata and Figure 4 representsthe data
gaveirrational seriesof orderswith thesemontmoril- from the heating experiments.The ordinate in the
lonites.) (2) The basal repeat distanceswere not figuresrepresentsa d-spacingdifference,e.g.,in Fig-
identicalfrom clay to clay. (3) The widths of the ute 2
maxima were not identical from clay to clay, nor
were they identical for a given clay. With respectto Ad2-s: 2 X d(002)- 5 X d(005).
the last observationno consistentrelationship was The horizontalcoordinateof the data points in the
found between breadths and diffraction angle, i.e., figuresis particlethicknessderivedfrom the Scherrer
cos d or tan d. equationu,singa constantof 0.9. The widths of the
The layer structurefactor is rather constantin the diffraction maxima from a small muscoviteflake
region of the (005) montmorillonite-glycol maxi- wereusedas a measureof the instrumentalbroaden-
mum and the (003) collapsedmontmorillonitemax- ing (b). The observeddiffractionwidths (B) were
imum. This suggestedthat thesetwo reflectionswere converted into pure diffraction breadths (p) by
best suited among those studiedto servethe follow- meansof the relationshipB'9= B2 - b'. The particle
ing dual purpose: (1) The breadth"sof thesereflec- thicknessesfor the clay-glycoldata (t5) were ob-
tions were a measureof the particle thickness,i.e., tained from the measuredbreadthsof (005), the
the mean distance perpendicularto the layers and, particle thicknesse's of the fired clays (13) were de-
thereby,the numberof layersper particle (N). (2) rived from (003).
The positions of thesemaxima were relatively unaf- The basal repeat distancesof the glycol-treated
fected by the thinnessof the particles,overall chem- claysrangedfuon 76.72A for Otay to 17.03A for
ical composition,or the specific nature of the par- Mowry. The basal spacingswere somewhatlessfor
ticle's surfaceand, therefore,were a measureof the the high layer-chargeclays complexedwith ethylene
glycol. The Ca-Fayetteclay gavea basal repeat of
16.88A which is closeto that determinedfor a Ca-
Wyomingclay (16.86A) by Reynolds(1965). The
effect of the layer charge was not evident on the
heatedclays.All of the Na-clayshad collapsedspac-
ings of about9.6 A.
Larger-valenceinterlayer cations decreasedthe
basalspacingsof both the glycol-treatedand heated
clays. For example, the glycol-treatedNa-Fayette
o
and Ca-Fayette clays in the same size fraction had
!
o spacingsof 16.95 and 16.88 A respectively.The
Fayettemontmorillonitesheatedto 450"C had spac-
ings of 9.60 A for Na--clay and 9.47 A for the
Ca-clay.
Additional observations from the experimental
data were made. ( 1) The particle thicknessesof
the fractionatedUpton and Fayette clays (e.g., Fig.
1) decreasedwith sedimentationsize fraction. A
r5(A1 larger differencewas indicated betweenthe particle
Frc. 3. Apparent spacingsof (003) for montmorillonite- thicknessesof very fine and fine Upton (Ur and Uz)
ethylene glycol complexes versus particle thickness. than betweenvery fine and fine Fayette(Fr and Fz).
 X-RAY DIFFRACTION OF MONTMORILLONITES 77

(2) The particle thicknessof the very fine Na- and

Ca-Fayetteclays were virtually identical. (3) The
medium size fractions of Fayette and Upton (F3
and Us) were additionallypretreatedfor 5-10 min-
utes with an ultrasonicprobe. The ad and breadth
valuesobtainedfrom the probe-treated samples,both
heatedand glycol-treated, were similar to thoseval-
ues obtained from the non-probe-treatedvery fine
sizefractions,F1 and U1,respectively.(4) The data ro
from the glycol-treatedclayswasmore scatteredthhn E
that from the heated clays. (5) All of the heated
clays except Santa Rita had larger N values than
their glycol-treatedcounterparts.(6) The particle
thicknessesof the heated clays fell into two groups.
The Wyoming clays were thicker and had about 13-
14 layers per particle, all othershad N = 8 except
CampBerteaux(N = 10).Mering (1946) deduced
the sameN for this latter clav.

Theoretical
Ftc. 4. Apparent spacings of (001 ) for montmorillonites
Known variations in chemicalcompositionof the
heated to 450"C versus particle thickness. Solid curve is for
claysplus the scatterin the diffraction data, as shown an all-aluminum octahedral composition; dashed curve is for
in Figures1-4, precludedan attemptto accountfor an all-iron octahedral composition.
all of the experimentaldata points by means of a
single theoretical curve. Nevertheless,it seemedde-
sirableto computeand plot a theoreticalrelationship was dioctahedral with aluminum as the cation and
(irrationality of basal orders versusN) basedon a the interlayer cation was sodium and its position
minimum number of assumptionsto which the ex- was fixed midway between adjacent silicate layers.
perimentaldata could be referredand discussed. A The model used to cornputethe theoreticalrelation-
computer program was written to calculate the ship for the heated clays was identical to that just
X-ray diftracted intensity from montmorillonites described except that water and glycol molecules
along the c* direction. The program allowed for were omitted from the interlayer region and the
variations in layer structure factor (true d-spacing, true d-spacingwas taken as 9.60 A. Two models
chemicalcomposition,atomic positions,temperature were used for the clay-glycol system; one model
factors) and N. Computationswere madefor CuK" representedby the dotted curves (Figures 1-3)
radiation and ranged fuom 4-32" 20 in O.OI" 20 assumeddouble layers of glycol molecules (plus
increments.The po,wderLorentz-polarizationfactor water and cations) betweenadjacentsilicatelayers
was used throughout and the origin of all of the with a singlelayer of glycol (plus waterand cations)
layer repeat units was the octahedralsheet. The at the ends of the particles;the secondmodel repre-
powder Lp factor was employed to be consistent sentedby the solid curves in Figures 1-3 was ident-
with the findings of other investigators(e.g., Reyn- ical to the first model except that the single glycol
olds, 1965). However,use of the singlecrystal Lp layer and water moleculesat the ends of the par-
factor did not change the positions of computed ticles were eliminated.We wish to thank Dr. R. C.
maxima by more than 0.01' 2A. The breadthsand Reynolds,Jr., for suggestingthe latter model to us.
positionsof maxima were determinedfrom computed Less glycol overall for the case where surface
patternsin the samemanner as was done for the ex- glycol was presentshifted the dotted curves toward
perimentaldata. the solid curves but had the undesirableeftect of
The structures and temperaturefactors assumed changing the structure factors considerably from
for the clay-glycol complexeswere those proposed those given by Reynolds (1965). The model with
by Reynolds(1965), with the following exceptions: no surfaceglycol did not producethis latter effect.
The true d-spacingwas 16.90 A, the silicatelayer Severalparameterswere varied to ascertaintheir
78 TETTENHORST AND ROBERSON

effect on computed positions and breadthsof maxima. there is a lack of impressiveagreementof the ex-
( I ) Large variations in the scattering power of the perimental data, which in some casesare highly
atoms caused discernible effects in the ad vs / plots. scatteredand show no trend (e.9., Figure 1), with
To illustrate, a dashed curve is plotted in Figure 4 the theoreticalcurves.The theory was basedon an
for the heated clays, which represents the results ob- analysisof two factorsonly, the layer structurefac-
tained from an all-Fe octahedral composition rather tor and N. Some of the disagreementis appargnt
than an all-Al octahedral composition. These since the curves were drawn for an "ideal diocta-
two compositions representextremesin the scattering hedral aluminummontmorillonite"and none of the
efficiency of the octahedral sheet and approximately claysanalyzedhasthat compositionprecisely.Never-
bracket any composition from a naturally occurring theless,other factorsnot includedin the theoretical
clay. The position of the dashed curve in Figure 4 analysis,such as size distribution,interstratification,
indicated that the first-order reflection from the and detailedstructuralaspects(e.9., inverted silica
collapsed clays was more nearly rational (for a tetrahedra)are likely to be more or less important
given N) with more scattering matter in the octa- than particle thicknessin determiningthe breadths
hedral sites. The identity or position of the interlayer and irrationality of basal reflections of a given
cation or the Si to Al ratio in the tetrahedral sites montmorillonite.
caused no measurable differences. The effect of an None of the montmorillonitesstudied, organic-
all-Fe octahedral composition on the clay-glycol treatedor heated,gave a rational sequenceof basal
complexes was also computed but the results were orders.This resultis likely due to the thinnessof the
omitted from Figures 1-3 for clarity. Regardless of particles, among other factors. A "particle size
the clay-glycol model, the eftect of an all-Fe com- effect" was clearly indicatedby the results of the
position made the first and second orders more fractionatedUpton and Fayetteclayswhosebreadths
rational and the third order more irrational. Com- increasedand thicknesses decreased with a decrease
pared with the effect of this substitution on Ad1_3for in sizefraction.
the collapsed clays, the shift in the theoretical Ad Roberson,Weir and Woods (1968) showedthat
vs I curves was smaller for the first and third order clay suspensions subjectedto a concentratedenergy
clay-glycol reflections and larger for the second input with an ultrasonic probe exhibited 'thinner
order clay-glycol reflection. (2) The temperature particlesthan beforesuchtreatment.A 5-10 minute
factor of all of the atoms in the outer half (includ- treatmentof the mediumFayetteand Upton samples
ing the octahedral atom) of the surface silicate layers (Fs and U3) gave increasedbreadthsof diffraction
on both ends of the particles was increased. To maximaand reducedparticlethicknesses which were
illustrate this effect, a temperature factor of 100.0 similar to the non-probe-treatedvery fine fractions
for the outer half of the end silicate layers had about of these clays (F1 and Ur) for both the glycol-
the same result as changing the octahedral composi- treated and heated specimens.The probe-treated
tion from all-Al to all-Fe, i.e., from the solid to the and very fine fractions alwaysgaveenhancedintensi-
dashed line in Figure 4. This temperature factor ties for the basalorderscomparedto the intensities
increase also caused the computed breadths to in- obtained from coarser fractions, a result noted by
crease, more for small N values. (3) Additional Roberson,Weir and Woods (1968) and attributed
scattering matter in the interlayer or on the surfaces by these investigatorsto better particle orientation
of the "collapsed" silicate particles, e.g., water in the former. The better oriented specimenspro-
molecules, suggested by the hectorite heated to duced the broaderreflections,a result likely due to
20OoC., raised the computed curve, i.e., increased thinner particles.
Ad1-3. (4) The orientation of the glycol molecules Non-rigorous chemical treatment or different
made little difference in the computed ad and methods of sample preparation did not affect the
breadth values. experimentalbreadthsand consequentN values.A
changein the identity of the interlayer cation, i.e.,
Discussion
Ca for Na, did not changeN for the Fayette clay.
An objective of this study was to determine Roberson,Weir and Woods (1968) and othershave
whether or not the'apparent basal spacingsof mont- shown,however,that the microaggregate morphology
morillonites were affected by particle thickness. Al- of suchclayswas definitelyafiectedby this exchange,
though the results indicate a particle size effect, and other studieshave shown that rheologicalprop-
 X-RAY DIFFRACTION OF MONTMORILLONITES

erties are affectedby a changein the cation com- and it is likely that none of the samplesregardlessof
position.The breadthsof the reflectionsusedherein pretreatmentwas completelyfree of this effect. The
are a measureof the thicknessof the "X-ray par- ordinateof the hectoriteheatedto 200'c (Figure4)
ticles" and this thicknessappearsto be unafiectedby relative to the theoreticalcurve is similar to those
cation exchange.Differentmethodsof sampleprep- of the other clays. Whether this result is significant
arationother than that describedabove,e.6'.,powder or fortuitous is unknown.
mounts from freeze-driedclays, gave the same N The experimentalB(005) values for the glycol
values as those reported here if the mounts were clayswere slightlylarger than the calculatedvalues,
thin. Sincecation exchange(Ca for Na) or powder more so for smallerN values,while the 9(003)
mounts prepared from freeze-driedclays produced valuesfor the heatedclayswere aboutegualto those
more random orientationof the clay particles,we computed. The experimental(001) and (002)
concludethat orientation did not affect the breadths breadthsfrom the glycol treated clays were defi-
of the reflections. nitely larger than,thosecomputed,e.9.,0.10' 20 for
Comparisonof the experimentaland theoretical N = 5 for (001), while the experimentalbreadths
clay-glycoldata suggests that the surfaceglycol may of the low orders from the heatedclays were slightly
make a rather small contribution to the diffraction less than the computed values. Also, the experi-
effects.This conclusionimplies that glycol molecules mental N values for the glycol-treatedclays were
must be sandwichedby two silicatelayers to con- smallerthan thosefor the comparableheatedclays
tribute significantlyto the diffraction pattern. While with one exception(SantaRita). Theseresultsmight
this conclusionis by no meansdefinite,the model be due to the glycol-treatedclays having larger
with no surfaceglycol shouldbe consideredin future amountof interstratification, or they couldbe caused
studiesof this system. by some other reflection-broadeningfactor that af-
The data in Figure 2 are pertinentwith respectto fected the experimentalglycol-clayresultsthat was
surfacecondition.We first attributedthe non-integral not accountedfor in the theoretical computations.
orders of the organic-treatedclays to interstratifica- It is suggested,therefore,that the N valuesobtained
tion which was a result of the partial glycol contents from the heated clays may be a somewhatmore
of variousinterlayerregions.This explanationproved reliable estimate than those obtained from the
insufficient, by itself, to account for the experi- organic-treatedclays. However, it is possiblethat
mental data. Computationsof expecteddiffraction thicker particles are created during the heating
effects were made for interstratified systemscom- process.
posed of abundant two-layer glycol spacingswith a Electron microscope observations (Roberson,
small percentageof zero andfor one layer glycol Weir and Woods,1968) on the ultrasonicallytreated
spacings and which allowed for variations in the clays reveal a significantnumber of apparentlyvery
scatteringefficiencyof the interlayermaterial. The thin par'ticles,presumablyone layer thick. We won-
calculationspredictednegativeAds-swhich was con- der, therefore, why the X-ray maxima from the
trary to the positive values observedexperimentally sonic-treatedclays were not broader than what we
for all of the clays.The model with no surfaceglycol, observed.Could it be that, although the sonic-
without any extra assumptionsconcerninginterstrat- treated clays contain a large number of exceedingly
ification, provided positive ,Ad2-5values and showed thin particles, the preponderanceof the X-ray in-
somewhatbetter agreementwith the other data. tensity is due to the presenceof a relatively small
The sample of hectorite heated to 200oC re- number of thicker particles?Or are the breadthsof
tained more water (as evidencedby its higher the basalreflectionsof montmorillonitesnot a meas-
d-spacing)than did the hectoriteheatedto 450oC ure of "mean particle thickness" but determinedby
and the former gave broader reflections.A smaller the amount of order obtained by piling layers in a
N value was therefore compu,tedfor this sample quasi-parallel fashion?The latter conclusion,if true,
which is not real (Figure 4). The increasedbreadth might preclude a relationship of apparent basal
of the reflectionswas due to the presenceof ran- spacingand particle thickness.
domly intercalatedwater moleculesand consequent
interstratification. This exampleservesto show the Acknowledgments
relative,though in this caserather small, effect of We wish to thank The Ohio State University Computer
interstratificationin reducing the apparentN values Center for granting us time and facilities. Appreciation is
80 TETTENHORST AND ROBERSON

extendedto Dr. Robert C. Reynolds,Jr., for severalhelpful glycol-montmorillonite complex. Amer. Mineral. 50,
discussions.Dr. Robert L. Crane and Mr. Thomas G. 990-1001.
Whitney gave valuable advice concerning the computer (1968) The effect of particle size on apparentlattice
programs. Dr. Robert R. Reeber kindly supplied the pow- spacings.Acta. Crystallogr. A24, 319-320.
dered silicon metal sample that was used to monitor the RorrnsoN, H. E., A. H. WErn, eNo R. D. Wooos (1968)
X-ray alignment. Volclay was obtained through the cour- Morphology of particles in size-fractionatedNa-mont-
tesy of Dr. Robert Doehler and The American Colloid morillonites. Clays Clay Minerals, 16, 239-247.
Company and is a Wyoming-type montmoriilonite. The Ross, M. (1968) X-ray diffraction effects by non-ideal
Camp Berteaux montmorillonite was obtained throueh the crystals of biotite, muscovite, montmorillonite, mixed-
courtesyof Dr. A. Weir. layer clays, graphite, and periclase. Z Kristallogr. 126,
80-97.
References TsrrrNuonsr, R., euo R. C. RpyNoros (19'71) Choice of
Bneornv, W. F., eNn R. E. Gnura (1951) High tempera- origin and its eftect on calculated X-ray spacings for
ture thermal effects of clay and related minerals. Amer. thin montmorillonite crystals. Amer. Mineral. 56, 1477-
Mineral. 3'6, 182-201. 1480.
Gnrrvr,R. E., eun G. Kursrcxr (1961) Montmorillonites: WARREN,B. E., ^rNoP. BoopNstrrN (1965) The diffraction
High temperature reactions and classificatjon. Amer. pattern of fine particle carbon blacks. Acta Crystallogr.
Mineral. 46, 1329-1369. IE,282-286.
MERINc, J. (1946) The hydration of montmorillonite. Wem, A. H. (1965) Potassium retention in montmorillo-
Trans. Faraday Soc. 42R,205-219. nites. C/ay Minerals 5, 17-22.
AND A. OsnnrrN (1967) Electron-optical study of
smectites.Clays Clay Minerals, 15,3-25. Manuscript receiued,March 28, 1972; acceptedfor
RnvNoI-os, R. C. (1965) An X-ray study of an ethylene publication,August 15, 1972.