American Mineralogist, Volume 66, pages 576-585, 1981
The "10A phase"in the systemMgO-SiOz-HzO
JoN F. BaUER'AND CHARLESB. ScLAR
Department of Geological Sciences
Lehigh University
Bethlehem, Pennsylvania I 80 I 5
Abstract
The l0A phaseis a unique pressure-dependent
hydrous phyllosilicatein the systemMgOSiOr-H2Owhich is stableat pressuresbetween32 and 95 kbar and temperaturesup to 535'C.
Although it was previously characterized as a 2rh-octahedralmember of the mica family, reexaminationof this phasehas resultedin a new chemicaland structuralmodel representedby
the formula [(H3O)2xrIMgyISil"Orr(OH)r].The results of X-ray powder difraction analysis,
thermal analysis,and experimentsin which the stoichiometryof oxide starting mixtures was
carefully controlled indicate that the structureof the l0A phaseis that of a fully trioctahedral
2:l phyllosilicate;it is similar to talc, but contains additional chemically-bound"water" in
l2-coordinatedsites[(HrO)]IIMg]ISi["O2.(OH[1.The resultsof infrared spectroscopy,however,indicate that this "water" is best representedas oxonium ions (H.O*) formed by interaction of the l2-coordinatedinterlayer H2O moleculeswith certain octahedral-layerhydroxyl
groups accordiagto the relation:
HrO + OH- --+ H3O* + 02or
[(H,O)fIrMg]rSilvO,o(OH)41- [(H,O)]"Mg]'SilvOrdOH),1
The presenceof the interlayer HrO* resultsin a basal spacing(d*,) of 9.96A as compared
with 9.35A for the correspondingbasal spacing(doo,)of talc. It also allows for strongerbonding betweenthe l2-coordinatedaqueousspeciesand the silicatenetwork, which accountsfor
the high thermal stability of the phaserelative to other crystalline hydrates.Although the
proposedstructureof this phaseis similar to that of a l0A mica, the l0A phaseis not strictly
isostructuralwith the mica familv.
Introduction
The existenceof a new pressure-dependenthydrous magnesiumsilicate,which is stableat pressures
between32 and 95 kbar and temperaturesbetween
375 and 535"C, was first reported by Sclar et al.
(1965 a,b). Preliminary X-ray powder diffraction
analysisof this phase showedthat it was a 2:l
phyllosilicatecrystallographicallysimilar to talc, but
with a l0A basal-planespacinganalogousto that of
mica. The deducedchemical formula for this phase
was reported as [(HrO)]" (Mgr!,)urSii"O,o(OH)ol
where the Roman numeral superscriptsindicate the
coordination number with oxygen of each cation in
I Presentaddress:Johns-Manville
Researchand Development
Center,P. O. Box 5108,Denver, Colorado 80217.
00,03-004x
/ 8t /0506-0576$02.
00
the lattice and the symbol ! denotesa vacant lattice
site. Sclar et al. (1965 a,b) concludedthat the structure and chemistryof this phasewere indicative of a
2r/z-octahedral,alkali-free, non-aluminous mica.
Somesupport for this model was provided by the
resultsof Seifert and Schreyer(1965)and Franz and
Althaus (1974) who synthesized,respectively,what
appearedto be the K*- and Na*-bearing analogues
of this phase hydrotherrrally at pressureslessthan 3
kbar (Fig. l). The potassium analogue was later
found to be part of an extensivesolid-solutionseries
which included the aluminous mica phlogopite
[K]"MgJ'(AlrSL)'"Oro(OH).1 (Seifert and Schreyer,
l97l) and its ferrous iron analogue (Kwak, l97l).
Both Seifertand Schreyer(1971)and Kwak (1971)
found that there was continuous solid solution between the silicon-rich end member [Kf"
576
577
BAUER AND SCL.tlR: 104 PHASE IN SYSTEM MgO-SiOz-HzO
F E - B E A R I N GA N A L O G U E S
\
lxl
-f
oro(otL
"# 1*.r,rf
+
II
tt'-l
Y
- * (x.o)fl tt.u or)%ifl o.o(ox)o*- * r'1"fl(naguo'1rsif;o.o0H)o
o.o(ott)o*
xf; (rvrsur.,)ssirlE
(seitert and SchreYer,1965)
I
+
" ' r oi p H r s e "
.'
(Franz and Althaus,1974)
(Sctar et a1.,1965a,b)
b,^l
/
r.F (Ar2si6
)E%o(oH)
*I
PHLOGOPITE
/
/
FE. BEARING ANALOGUES
Fig. l. potential crystallo-chemical relationships of the l0A phase to various alkali-bearing micas. Solid arrows represent
solid solutions
expeimentally verified solid solutionsbetweenthe two indicated end members.Dashed arrows repr€senthypothetical
and Schreyer'
to
Seifert
(1965a,b).
refers
al.,
Sclar
et
of
model
phase
on
the
mica
based
[S]
l0A
the
and
members
mica
end
between
(1971);[K] refersto Kwak (1971).
(Mgr!,)"'Si["Oro(OH)n]and an ideal magnesium-rich
end member [K]'IIMgJ'(MgSi')'"O'.(OH)J. It appeared,therefore,that the l0A hydrous magnesium
silicate phase of Sclar et al. (hercafter referred to as
the l0A phase)might also be part of such a solid-solution seriesat very high pressures.The existenceof
such phasesmight be of great importance with respectto understanding(l) the storageand releaseof
water in the earth'smantle and (2) the reactionof anhydrous silicate minerals with water at pressuresin
excessof 30 kbar.
Recent high-pressurestudiesin the systemMgOSiO,-H,O (Yamamoto and Akimot o, 1977) indicated
that the chemicalcompositionof the l0A phasemay
be diferent from that initially reported by ScIar et al.
Yamamoto and Akimoto gavethe compositionof the
l0A phaseas (3MgO'4SiO, '2Il2O), although this
conclusionwas not supportedwith crystallochemical
data. Their results suggestthat either (l) the l0A
phaseis indeed an oxonium-bearingnon-aluminous
phlogopitic mica with the composition [(H'O)]"
which
Si,,3u3)r'Oro(OH)oJ,
Mg, ,.oil .ru)'t(Mg.r.o
approachesthat of the magnesium-richend member
of the non-aluminousphlogopiteseries(Figure l), or
(2) the l0A phaseis a hydrated Variety of talc [Mg]r
Si1"O,"(OH)rlwhosedetailedstructureis asyet undefined. The experimentalresults of Bauer and Sclar
(lg7g) indicate that the l0A phase 16 n limited
range of solid solution, which supports the latter
model. In this paper,we report the resultsof a study
of the composition and structure of the l0A phase
basedon X-ray powder diffraction data, infrared absorption spectroscopy,and thermal analysis.
Experimentaland analyticalmethods
All the high-pressure experiments in this study
were carried out in a modified "belt" apparatus
(Sclar et al., 1963) at pressuresbetween 35 and 60
kbar and temperaturesbetween425 and 500'C. The
experimental P-T path employed was thermal
quenching under pressure followed by pressure release.The reactantswere sealedin welded gold capsules,and reactiontimes rangedfrom 1.5to 24 hours;
most experimentalruns had a duration of l-2 hours.
The reactantsusedin theseexperiments5's1shighpurity MgO (periclase)and SiO, (dehydrated silicic
578
BAUER AND SCI-4R: IOA PHASE IN SYSTEM MgO-SiO2-H2O
o
\,
o
4
o
J
E
.9
(t
3
Temperoture(oC)
Fig. 2. Thermogravimetric curv€s showing progressive
dehydrationofthe l0A phascand ofsynthetic talc preparedfrom
the same oxide components.Heating rate for all sampleswas
200oC per hour rn vacuo. Curve (l) = l0A phase;Curve (2) :
synthetic talc.
acid) plus a fixed amount of high-purity water (45
mole 7oof the total oxide sum) which was previously
distilled, deionized, and freshly boiled. With these
starting materials, there was minimal contamination
of the high-pressureproductsby magnesiumcarbonate which was present in experimental products obtained from mixtures of brucite and silicic acid and
brucite and silica gel. The latter mixtures were used
in earlier experimentalstudies(Sclar er al., l965a,b;
Yamamoto and Akimoto, 19'17).
The high-pressureproductswere analyzedby optical microscopy and by X-ray powder diffraction
methods including both Debye-Scherrertechniques
Hzo
Serpent ine
a
,lto i pto..
I
j Tolc
5i02
Coesile
ai.,*9o
Fig. 3. The ternary system MgO-SiO2-H2O showing the
compositionof the l0A phase,as proposedin this paper,and that
of other selected phases. Note that the composition of the l0A
phase lies along the talc-H2O join (dashed line).
and difractometry. Single-crystal X-ray diffraction
analysiswas precludedbecausethe largestindividual
crystallites in the fine-grained platy aggregatesof the
l0A phase were smaller than 5 pm. X-ray powder
diffraction patterns of the l0A phase were indexed
and the correspondinglattice parameterscalculated
with the aid of the least-squaresrefinementprogram
of Appleman and Evans (1973).Infrared absorption
spectra were obtained in a dry N, atmosphereby
means of the KBr pressed-pellet technique using a
Perkin-Elmer 283 infrared spectrophotometer.
Thermogravimetric analyses were obtained with a
Sartoriusmicrobalancein a vacuum of 5 x l0-a Torr
using a heating rate of 200oCper hour.
Resultsand discussion
Composition
The chemical composition and stoichiometry of
the l0A phase was ditermined through a seriesof
high-pressureexperimentsin which the atomic ratio
of MglSi in the starting mixtures was carefully controlled. The purposeof theseexperimentswasto synthesize the l0A phase from m.ixtures representing a
range of MglSi ratios and to determinethe number
and identity of the coexisting phases. By incrementally varying this ratio irl successiveruns, a
unique bulk compositionwas found which yielded a
single-phaseproduct of the l0A phase.This unique
bulk compositionhas a Mg/Si ratio of 0.75 which is
equivalent to that of talc. This Mg/Si ratio of the
l0A phasewas also found to be independentof both
temperatureand pressure.
A Mg/Si ratio of 0.75 for the l0A phasewas also
strongly suggestedby the results of simple thennal
dehydration experimentscarried out at room pressure.Sclar (1967)showedthat the 10A phasemay be
partially dehydrated at temperaturesof 500-600'C
within 24 hours and reportedthat this resultedin the
collapse of the basal-planespacing from l0A to
9.35A which is equivalent to that of talc. These experimentswere repeatedin this study with the objective of carefully checking the thermal products for
phases other than talc. X-ray powder diffraction
analysis by the Debye-Scherrer method did not reveal any phasesin the systemMgO-SiOr-HrO other
than talc which supports the conclusion that the Mgl
Si ratio of the l0A phaseis 0.75.Additional suppon
for this conclusion is forthcoming from the reaction
talc + H2O + l0A phasewhich was experimentally
demonstratedat elevatedtemperature and pressure
(Sclar er al., 1965b; Sclar, 1967). Debye-Scherrer
BAUER AND SCLAR: lOA PHASE IN SYSTEM M|O-SiOZ-HZO
Table l. X-ray powder diffraction data for the l0A phase
t/t
dot
hkl
calc
(' A )
"(i)
9.97
100
001
9 .95-l
4.91
I5
002
4.9lA
4-59
50
020
:. oso*
10
lr2
3.320
BO
003
3.319
2.903
10
113
2.903
2.640
25
130
2.643
2 .6lA
50
204
131
2.616
2 . €.I4
2.502
t5
tt3
t3L
2.543
2.500
3.650
579
similar in shapeto their magnesium-and hydrogenbearing analogues;they showedweight lossesdue to
structurally bound water that correspond exactly to
those of their hydromagnesiancounterparts.
The results of the high-pressuresynthesisexperiments and the thermogravimetric data indicate
that the l0A phase has a fixed composition within
the system MgO-SiOr-HrO. In terms of oxide
ratios, this composition may be expressed as
3MgO:4SiOr:2H,O which is equivalentto the composition inferred by Yamamoto and Akimoto (1977).
relationship of the l0A phase to
Th.
"o-po.itional
other selectedphasesin this systemis shown in Figure 3.
Structure
The crystal structureof the l0A phasewas defined
principally by X-ray powder diffraction analysis of
2 .42-l
t32
40
2.433
high-pressureproducts.The d-valuesand relative in2 -294
040
10
2.291
2.295
22r
tensitiesof the X-ray diffraction lines for this phase
2.165
l3l
10
2.164
were constant regardlessof the pressureor tempert.992
oo5
20
1.996
ature of synthesisor of the bulk composition of the
r-136
L34
synthesismix and the resulting presenceor absence
1.735
24r
l5
t.135
of other phases.This indicatesthat the fundamental
L.134
150
crystal structure of the l0A phase as well as its
I .669
I52
L.668
24L
15
r.668
stoichiometry is not only unique, but very specific
r.665
t3s
and non-variablein the systemMgO-SiO'-H'O.
r .532
060
40
r .532
X-ray powder diffraction data for the l0A phase
are given in Table l. Comparisonof the unit-cell dimensionsof the l0A phase with those of synthetic
phlogopite (Table 2) reveals the similarity of the
patterns
*obtained
powder
Debye-Scherrer
from
only
.t-Joi" or ine toA phasewith that of a l0A mica.
The unit-cell dimensionsand chemical composition
patterns of products from these runs show that no of the l0A phase correlate well with a monoclinic
crystallinephasesother than talc and the l0A phase 2:l trioctahedral phyllosilicate structure (lM polyare present. It thus appearsthat in both hydration type).The b/aratio of 1.731is very closeto the theoand dehydration reactionsinvolving the l0A phase,
the Mg/Si ratio remainsconstant.
The total water content of the l0A phasewas de- Table 2. Computed lattice parametersand densitiesfor the l0A
Phaseand PhlogoPite
termined by (l) vacuum thermogravinetric analysis
(TGA) of two l0 mg samplesusing pure alumina as a
Phlogopite *
l-0A Phase
referencematerial (Fig. 2) and (2) simple dehydration experiments based on the difference in
5. 3l4A
a
5.316 I .0I0A
a
weight loss after 24 hrs at I lOoc and 24 hrs at
9.204A
1
b
.010A
9.r91
b
1000"C on three discrete15-20 mg samples.The to1
0. 314A
i
.015A
10.118
c
tal weight loss due to the releaseof structurally
s
g. go
s
Loo.25 r.t5"
I
+
bound water was9.0+0.2percent(TGA) and 9.2+0.5
2
.185 g,/cm
D
2.712 q/cm3
percent by simple dehydration methods. Thermogravimetric curvesof nickelian talc and of nicketan
*
and Wones (1972) ,
10A phase(Ni*' in place of Mg*'z)and of deuterated * * BDaast ea d f roonm l MH aPz eo nI Y t Y P e .
10A phase(preparedfrom DrO instead of H"O), all
t Based on the composltaon [3Mgo:4sio2:2H2o]'
of which were synthesizedat high pressure,are very
2-489
to
004
2 .4A9
BAUER AND SCI,AR: lOA PHASE IN SYSTEM MgO-S\O2-H2O
retical value of (3)'/2for an ideal monoclinic layer sil_ dration of the l0A phase proceedsin two discrete
icate, and all indexed reflectionshave even valuesof steps,which occur at intervalsof approximately 475(h + k) in accord with the systematicextinction laws 675" and 675-875",respectively.The weight loss associatedwith each of these thermal intervals is the
same, namely, 4.5Voby weight. Both the nickelian
and the deuteratedisostructuralanaloguesof the l0A
phasealso show a well-resolvedtwo-stepthermal dehydration pattern involving weight lossesof the same
quently than their lM counterpartsin trioctahedral magnitudein the correspondingthermal intervals.Xphyllosilicatestructures,so it is probablethat the lat_ ray powder difraction data obtained from
samples
subjectedto prolongedheat treatmentat 500oCshow
that the first dehydration step correspondsto a collapse of the basal spacing of the l0A phase from
9.96A to 9.35A and also resultsin a marked decrease
in all observed d-values for reflections other than
thors were not indexed and the correspondingcell those from
re{hk0} planes.The d-valuesof
parameterswere not given, preliminary assignment flections show a relatively small decrease{hk0}
which is
of their reported reflectionsbased on the indexing probablydue to the small decrease
in D from 9.l9lA
procedure used in this study indicates a unit cell (l0A phase)to 9.l7lA (talc). The resulting
diffracwhich is approximately3Volarger in the c-axis direc_ tion pattern is identical with that of pure magnesian
tion and 0.5Volarger in the D-axisdirection than unit_ talc. The additional 4.5 tttt.Voloss occurs only when
cell dimensionsof the l0A phasegiven in Table 2. In the sample is dehydrated at temperaturesgreater
addition, Yamamoto and Akimoto report a variable than 650o.A comparisonof curves(l) and (2) in Figvre 2, shows that the second dehydration step is
equivalent in temperature range and magnitude of
weight lossto the single-stepdehydrationof talq The
high-temperaturedehydration stepof the tOA phase,
sure or temperatureof synthesis.No such variation therefore, involves decompositionof the octahedral
in d*, was found in the present study. Apparent layer hydroxyl groupsasin the simpledehydration of
basal spacingsof lessthan 9.96A may be attributed talc. This was confirmedby X-ray powder difraction
to composite reflections resulting from talc-l0A analysisof samplesof both talc and the l0A phase
phase random interstratifications, whereas those which were subjectedto prolongedheat treatmentat
greater than 9.96A may be due to the incorporation 1000'; both products were composedsolely of
ctiof weakly-boundexcessHrO layersbetweensomeof noenstatiteand cristobalite.
the phyllosilicatesheets.Unfortunately, neither TGA
Theseresultsstrongly indicate that the l0A phase
nor other thermal dehydrationdata were provided by is a monohydrate with the structure of talc whose
Yamamoto and Akimoto so that the distinction be_ formula may be representedstoichiometrically as
tween relatively weakly-bound interlayer HrO and MgrSi.O,o(OH)r.HrO. The marked effect of the
relatively strongly-boundHrO could not be made. In "hydrate HrO" on the magnitude of the c parameter
the presentstudy, all synthesisexperinents were car_ suggestsa structural model in which H-O-H groups
ried out with a fixed mole fraction of water in the are stereochemicallyarrangedparallel to (001) of the
starting mixture of XH,o : .45; however, the l0A
phyllosilicate structure. It also appears that the
phasesynthesizedin twelve earlier experi_mental
runs "hydrate HrO" is more weakly bound than the "hyby C. B. Sclar in which Xr,o2.45 has a constantX_ droxyl HrO" of the octahedrallayer, inasrruch as the
ray powder diffraction pattern with d-valuesand rel_ former is releasedfrom the structureat lower temperative intensities in excellent egreementwith those atures.
given in Table L
The structural constraintsimposed by the X-ray
X-ray powder diffraction data were also obtained difraction and thermogravimetricdata can best be
from sampleswhich were heated at severaltemper_ accountedfor by assigningthe "hydrate HrO", as inaturesbetween100and 1000.C in order to correlate dividual HrO molecules,to quasi-hexagonalinterstructural changeswith the thermogravimetricdata. layer sitesof l2-fold coordination within the opposAs shown by curve (l) in Figure 2, thermal dehy_ ing tetrahedral layers of a fully trioctahedral 2: I
BAUER AND SCLAR: IOA PHASE IN SYSTEM M9O-SiOZ-HZO
phyllosilicate lattice. These l2-coordinated sites are
structurally equivalent to those occupied by alkali
ions in a 10A mica. The characteristic"packing
radius" of an HrO molecule is nearly identical with
that of the K* ion (-1.35A), so thereappearto be no
stereochemicalrestrictions on such an assignment.
By analogywith the mica structure,a new structuralchemicalformula for the l0A phasecan be written as
[(HrO)*"Mgy'Si1"O,o(OH)r].In a strict sense,however, this talc-hydrate formula does not representa
true mica inasmuch as the octahedral-tetrahedral
network carriesno net charge.
The model for the l0A phasepresentedabove satisfies both the crystallographic and chemical constraints and is in reasonableaccord with the results
from thermogravimetric analysis.It does not, however, provide a mechanismby which the l2-coordinated HrO speciescan be chemically bound within
the structure except for the possibility of hydrogen
bonding to the oxygensof the tetrahedral (Siror)'znetwork. Such a simple model is not readily compatible with thermal experiments,which show that releaseof this HrO from the structurerequirestemperaturesgreaterthan about 400'C; this is much greater
than the temperatures(l0G-200'C) which are normally required to dehydrate phases containing
simple hydrogen-bondedwater of hydration.
Infrared absorptionspectraofthe l0A phasewere
obtained in order to more fully characterize this hydrous species.Pure talc wasusedas a referencephase
becauseit is structurally similar to the l0A phaseand
its infrared absorptionspectrumis well characterized
(Wilkins and Ito, 1967; RusselTet al., 1970). Inasmuchas infrared absorptionspectraprovide qualitative information on bond strength, the coordination number of cations,and structural configurations
within a phyllosilicatelattice, this analytical method
was used as a check on severalof the deduced features in the structural model presentedabove.
RepresentativeIR absorption spectra of the tOA
phaseand of synthetic talc are given in Figure 4 (a
and b). As expected,the two patternsare very similar
in accord with their close structural relationship.
Correlation of observedabsorptionfrequenciesin the
spectra of the l0A phase and synthetic talc with
those observedby Wilkins and Ito (1967) and by
Russell et al. (1970) shows that all absorptionsinvolving the tetrahedral layer and all absorptionsinvolving interactionsbetweenthe tetrahedral and octahedral layers (mixed vibrations) are virtually
identical in position and have similar relative intensities (see Farmer, L974, for a summary of as-
581
signedvibrational modesin the talc structure). This
indicates that at least the tetrahedral layers in the
l0A phasestructuredo not deviatesignificantlyfrom
the relatively simple (SirO,)S network of talc. Absorptions involving the octahedral layer of talc are
principally those of the hydroxyl group, which has
three prominant vibrational modes.All three are infrared active and give rise to absorptionsat 3677,
669, arrd 445 cm-'. Theseabsorptionsalso appear in
the spectrumof the l0A phase;however,they have a
much lower intensity than the correspondingabsorptions in talc basedon spectraobtainedwith the same
concentrationof eachmaterial in the beam path.
The only absorptionswhich are potentially assignable to the l2-coordinated hydrous speciesin the
l0A phase are given in Table 3. These are absorptions which are present in the spectra of the l0A
phase but are absent from the correspondingtalc
spectra.As shown in Table 3, three of theseabsorptions may be assignedto vibrational modesinvolving
the oxonium (HrO*) ion as observedin solid acid
hydrates (Ferriso and Hornig, 1955;Savoieand Giguere, 1964; Fournier et al., 1969) and in various
mineral phases(White and Burns, 1963;Wilkins et
al., 1974; Bokij and Arkhipenko, 1977)' Of these,
only one absorptionfrequency(1700-l'720cm-') appearsto be due solelyto HrO*; the other two overlap
absorptions which might arise from hydrocarbon
contaminants, molecular HrO, or physically adsorbed hydrogen-bondedHrO networks. Addilional
evidencefor the existenceof HrO* in the l0A phase
was obtained from the infrared spectraof samplesof
the l0A phase synthesizedat high pressure from
MgO-SiO,-D,O rather than MgO-SiO'-H,O mixtures. Spectraof these deuteratedsamples(Fig. ac)
show a predictable decreasein all octahedral layer
hydroxyl absorption frequencies(see Russell e/ a/.,
1970),and in addition, exhibit a new discreteabsorption at 2445 cm-'which correlateswell with one assigned fundamental frequency of the DrO* ion
(Fournier et al., 1969).Furthermore,this absorption
band falls outside the reported range of absorptions
attributable to surface adsorbed, hydrogen-bonded
D,O groups (Yuklnevich, 1963).
The results obtained from infrared spectroscopy
indicate that chemical bonding in the l0A phase is
somewhatmore complexthan would be predictedby
the relatively simple talc-hydrate model. In particular, the possiblepresenceof H'O* in this structure
suggestsa model which is more compatible with an
oxonium-bearingmica (Figure l). As noted above,
the mica model cannotbe rejectedsolely on the basis
s82
BAUER AND SCLAR: IOA PHASE IN SYSTEM MgO-S\O2-H2O
o,
(.)
o
o(,
c
o
.:
E
tr
o
4000
3000
2000
1600
1200
800
Frequency(.--l)
c,
U
o
CL
o
o
tr
o
4000
3000
2000
1600
1200
Frequency (.--l
800
400
)
Fig. 4' Representative
infrared absorptionspectraof experimentalproducts.Spectrawereobtainedin a dry N, atmospherefrom
samplesdispersed
in KBr pressedpellets:(a) l0A phase;(b) Synthetictalc; (c) l0A phaseheatedin air at 550ocfoi 24 hours;
(d) Deuterated
l0A phase.
of the MglSi ratio. It does,however,have the following shortcomings:
(l) The H,O content of a mica phase with the
composition [(HrO)f', (Mgr.uo ! .ru)''(Mg.r.o
Si,.636)rvO2o(OH).]
is not equivalent to that observed
for the l0A phase;completedehydration of the l0A
phaseresults in a weight loss of 9.0Vo,not ll.SVo as
predicted by this model.
(2) Dehydration of the l0A phaseresultssolely in the
formation of talc plus free HrO; dehydration of the
correspondingmica phaseto talc plus HrO would result in the production of free protons, which would
create a charge imbalance in either of the product
phases, and would, therefore, present a chemical
problem which is not easily resolved.
(3) The presenceofoctahedral vacancies,as required
by the mica model, would result in different chemical
environmentsfor the individual octahedrallayer hydroxyl groups which should thus give rise to unique
well-definedIR absorption bands satellitic to and at
BAT]ER AND SCDLR: j,OL PHASE IN SYSTEM MgO-SiO2-H2O
slightly lower frequencies than the principal O-H
stretchingfrequency(3675cm-') (seeKodama et al.,
1974);no evidence for such bands was found in the
spectrumof the l0A phase.
The structural model for the 10A phasewhich best
fits the sum of the analytical data is illustrated in Figure 5. Chemically,this model may be representedas
follows:
583
(a)
[(H3O)]uMgJISil"O,,(OH),I
Note that this is stoichiometrically equivalent to the
talc hydrate formula presentedabove; however, allowancehasbeenmadefor the following interaction:
HrO+OH-+HrO*+02octahedral
12octahedral
12layer
coordinated
layer
coordinated
Such an interaction appears likely if the octahedral
layer hydroxyl groupsare orientednormal to thLea-b
plane and lie directly beneath the l2-fold sites as
they do in other trioctahedral phyllosilicates such as
micas (Serratosaand Bradley, 1958).Becauseof this
unique orientation of the O-H groups' the K-H distancein phlogopite,for example,is closeto the K-O
distance in the l2-coordination "sphere," nam€l],
3.10A vs. 2.964, respectively(Rayner, 1974).However, the K-H distance in phlogopite is largely controlled by repulsive forces between the positively
charged potassium ion and the positive end of the
hydroxyl dipole (Bassett,1960),which resultsin both
a larger c and a Iarger b (due to shortening of the
OH-OH shared edges)than would be expected if
there were either an attractive interaction or no inter-
O Oxygen
g Hydroxyl
I Mognesium
Q oxonium
6 Silicon
o Silicon
(b)
Table 3. Unique infrared absorptions characteristic of the l0A
phase
!requency
(cm
3400-3440
-t
)
Possible
o-H
Assignment
stretch
o-H stretch
+
2v4 (HrO )
(H2o"'-HOH)
(HOH" .'O-Sa)
3200-3250
2r2
lE2o)
29rO-2930
c-H
vr'
scret.h+ (conLaninaLion)
v3 (H3o )
11 00-r'7 2a
v,
+
(H.O )
1600-1640
v2
(H2O)
Fig. 5. Idealized crystallographic model of the 10A phase
showing the arrangement ofthe tetrahedral and octahedral layers
and the position of the interlayer l2-coordinated H3O+ species:
(a) (100) projection (modifed from Grim, 1968); (b) (0Ol)
projcction (modifed from Bailey, 1967).
action between the l2-coordinated speciesand the
hydroxyl hydrogens.Inasmuch as H2O, HrO*, and
K* are nearly the same size, the smaller c and b parameters of the l0A phase relative to those of
phlogopite suggeststhat the (hydrous species)xll-OH
BAUER AND SCLAR: lOA PHASE IN SYSTEM MgO-SiO2-H2O
bond in the 10A phaseinvolvesa net attractiveinteraction. Correspondingly,this supportsthe view that
the l2-coordinatedhydrous speciesin the l0A phase
is (H,O)*.
This new model is in accord with the existencein
the l0A phaseof two typesof structurally bound sitespecifichydrous species,namely, oxonium ions and
hydroxyl ions, both of which can be recognizedin
the infrared absorptionspectrum.It also accountsfor
the observedreduction in intensity of all absorptions
assignedto vibrations of the octahedral layer hydroxyl groups relative to those in talc. which has
twice as many hydroxyl groups per formula unit. In
addition, this model helps to explain the relatively
high thermal stabitty of the l0A phaseas compared
to other crystallinehydratesby allowing for (a) electrostatic attraction between the l2-coordinated hydrous species(H.O*) and the octahedral-tetrahedral
network and (b) hydrogen bonding betweenthis speciesand the oxygensofthe surfacesofthe adjoining
tetrahedrallayers.
Thermal dehydrationof the l0A phaseis thus consideredto procecdas follows:
ofthe l2-coordinatedhydrous speciesto form bonds
with the octahedral layer oxygensprovides definite
restrictions on ion interchangeabilityin the l2-fold
site. Inasmuchas micasreadily acceptAl3* into their
structure whereastalc acceptsonly a few weight per
cent (seeFawcett and Yoder, 1966),the lack of solid
solubility of aluminum in the l0A phaseis more consistent with a structural model based on a talc-like
lattice rather than on a mica lattice. The earlier conclusion of Bauer and Sclar(1979)that solid solubility
in the l0A phase appears to be limited to substitutions in the octahedral layer of ions similar in
size and charge to Mg2*, such as Co2* and Ni2*, is,
therefore,completelycompatiblewith this model.
Conclusions
The l0A phase is a unique high-pressurephyllosilicatephaseh the ternary systemMgo-Sior-Hro.
It has a specific fixed composition, independent of
the temperatureand pressureof synthesiswhich may
be expressed stoichiometrically as 3MgO :
4SOr:2HrO.The crystalstructureof the l0A phaseis
that of a fully trioctahedral 2: I phyllosilicate with a
9.96A basal spacing.The octahedraland tetrahedral
(H,O),Mg.SLOz,(OH),+
network of this phaseis similar to that of talc; how10A phase
ever, in the l0A phase a discreteHrO molecule ocMg.Si,Oro(OH)4
+ 2HrO
(l) cupies each l2-coordinated interlayer site. Infrared
talc (-4.5 rtrt.VoHzO)
absorption spectroscopy and thermogravimetric
analysisindicate that each l2-coodinatedHrO moleMguSi.Oro(OH)o
cule is extensively bonded to both the tetrahedral
talc
and the underlying octahedrallayer of the structure.
6MgSiO, + 2SiO,
+ 2}J2O
e,)
Interaction of each HrO molecule with the octaheenstatite
cristo(-4.7 vrt.VoHrO\
dral layer hydroxyl group resultsin the formation of
balite
HrO* and O'- from the HrO and OH-. It is conor silica
cluded that the preferred formula for the l0A phase
glass
is [(H,O)*"MgI'SiI"O,,(OH)]. Although they are
structurally similar, the l0A phaseis not a true isoNote that the theoretical weight loss values shown structural member of the mica family and does not
above are in excellent agreementwith the results of appear to form solid solutions with the isostructural
- KMg, rD 'SLO,'(OH),].
thermogravimetricanalysis(Fig. 2). Becausethe first series[KMg3AlSi3O,o(OH),
step in this processinitially involves re-protonation
Acknowledgments
ofoxygen ions in the octahedrallayer, it doesnot require that, upon initial loss of water (oxonium ion
This paper is basedon a doctoral dissertationpredestruction), residual protons remain in the inter- sentedto the faculty of the Departmentof Geological
layer site (or anywhereelsein the lattice).
Sciencesat Lehigh University by Jon F. Bauer in
This model also provides an explanation for the partial fulfillnent of the requirementsfor the Ph.D.
lack of incorporation of (K*)o' or (Al3*)rv".'' into degree.
the lattice of the l0A phase in high-pressuresynThis researchwas supportedin part by Grant No.
thesis runs (Bauer and Sclar, 1979);that is, neither 1975-75 to JFB from the Geological Society of
potassium-nor aluminum-bearinganaloguesof this America and a Grant-in-Aid to JFB from Sigma Xi.
phasecould be prepared.It appearsthat the ability Thermogravimetric analyseswere obtained at Bat-
BAUER AND SCLAR: lOA PHASE IN SYSTEM MgO-SiO2-H2O
telle-Columbus Laboratories under Grant No. oel-124-a673to CBS.
ARo-D-3
References
Appleman,D. E. and Evans,H. T'' Jr. (1973)Indexing and leastrquut.t refinementof powder diffraction data' U. S' National
Technical Information Service,PB-216-188.
Bailey, S. W. (1967) Polytypism in layer silicates.In Layer Silicates,AGI Short Course Lecture Notes' American Geological
Institut€, Washington.
in mica alteraBassett,W. A. (1960)Role of hydroxyl orientation
'll'
M9456'
tion. Bulletin GeologicalSocietyof America,
Bauer,J. F. and Sclar,C. B' (1979)Crystal chemistryand stability
of a high-pressurehydrous 10A layer silicate in the syst€m
MgO-SiO2-HrO. In K' D. Tirnmerhausand M' S. Barber,Eds',
High-Prcssure Science and Technology (Proceedings Sixth
AIRAPT Conference),Vol. 2, p. 144-152.Plenum Press,New
York.
Bokij, G. B. and Arkhipenko, D. K. (1977) Infrared spectraof
oxonium and ammonium ions in layer aluminosilicates'Physics
and Chemistry of Minerals, 1,233-242.
Farmer,V. C. (1974)The layer silicates.In V. C. Farmer, Ed'' The
Infrared Spectraof Minerals.MineralogicalSocietyMonograph
4,331-364,London.
Fawcett,J. J. and Yoder, H' S. (1966)Phaserelationshipsofchlorites. American Mineralogist, 51' 353-380.
Ferriso,C. and Hornig, P' (1955)The infrared spectraofoxonium
halidesand the structureofthe oxonium ion. Journal ofChemical Physics,23, 146+1468.
Fournier, M. G., Mascherpa,G., Rousselet,D., and Potier' J'
(1969) Assignment of oxonium ion vibrational frequencies'
C. R. Acad. Sci. Ser.C,269' 2'19-282.
Franz, G. and Althaus, E. (1974)Synthesisand thermal stability
sodium mica, NaMgr r(Si+Oro)(OH)z' Contriof 2Vz-octahedral
butions to Mineralogy and Petrology,46' 221-232'
Grim, R. E. (1968) Clay Mineralogy. SecondEdition' McGrawHill, New York.
Hazen,R. M. and Wones,D. R. (1972)The effectof cationic substitutions on the physical properties of trioctahedral micas'
American Mineralogist, 57' 103-129.
Kodama, H., Ross,J. G., Iyami, J. T., and Robert, J. L' (1974)Effect of layer charge location on potassium exchange and hydration of micas.American Mineralogist, 59' 491495'
Kwak, T. A. (1971)An experimentalstudy on Fe-Mg micastransitional between dioctahedral and trioctahedral compositions'
Neues Jahrbuch Mineralogie' Monatshefte, l9'l l, 326-335'
Rayner, J. H. (1974)The crystal structureof phlogopite' Mineralogical Magazine,39, 850-856.
Rusiell. J. D., Farmer,V. C., and Veldc' B. (1970)Replacementof
585
OH by OD in layer silicatesand identificationof the vibrations
of these groups in infra-red spectra.Mineralogical Magazine,
37. 869-879.
Savoie,R. and Giguere,P' A. (1964)Infrared study of the crystalline monohydrates of nitric, perchloric, and sulphuric acids'
Journal of Chemical Physics,41, 2698-2705'
Sclar,C. B. (1967)High pressuresilicatechemistry:synthesis,stability, and transitionsof crystallinephasesin the systemMgOSiO2-H2O.ProceedingsUSARO-Durham Conferenceon New
Materials from High PressureHigh Temperature Processes,
Watervliet.N.Y., 135-150.
Sclar.C. B., Carrison,L. C', and Schwartz,C' M' (1963)The calcite-aragonitetransition and the calibration of a "belt-type appur"tor; between15 and 25 kb, 700'C-l400oC' In A' A' Giardini and E. C' Lloyd, Eds., High Pressure Measurement'
Butterworths,Washington,286-300.
Sclar, C. B., Carrison, L. C', and Schwartz,C' M' (t965a) Highpressure synthesis arrd stability of a new hydronium-bearing
iayer silicate in the systemMgO-SiO2-H2O (abstr') Transactions of the American GeophysicalUnion, 46, 184'
Sclar,C. B., Carrison,L. C., and Schwartz,C' M' (1965b)The system MgO-SiO2-H2O at high pressures'25-130 kb' and 375l000oO(extendedabstr.),BasicScienceDivision, American Ce-
196-215.
Serratosa,J. M. and Bradley, W. F. (1958)Determination of the
orientation of OH bond axesin layer silicatesby infrared absorption. Journal ofPhysical Chemistry, 62, 1164-1167'
Whit;, J. L. and Burns, A. F. (1963) Infrared spectraof hydronium ion in micaceousminerals.Science,141' 800-801'
Wilkins, R. W. T. and Ito, J. (1967)Infrared spectraof some synthetic talcs.American Mineralogist' 52' 1649-1661'
Wilkins, R. W. T., Mateen,A., and West, G' W' (1974)The spec'
troscopicstudy of oxonium ions in minerals'American Mineralogist,59,81l-819.
Yariamoto, K. and Akimoto, S. (1977) The systemMgO-SiOrHzO at high pr€ssuresand temperatures-stabilityfield for hydr-oxyl chondrodite, hydroxyl clinohumite, and l0A phase'
American Journal of Science,277,288-112'
Yukhnevich, G. V. (1963)Advancesin the use of infrared spectroscopyfor the characterizationof O-H bonds' RussianChemical Reviews,32, 619-633.
Manuscript received, September 17' 1980;
acceptedlor publication, January 6' 1981'