American Mineralogist, Volume 80, pages 1286-1292, 1995
Lawsonite: Upper pressure stability and formation of higher density hydrous phases
M,lx W. ScHuror
BayerischesGeoinstitut, Universitiit Bayreuth, D-9 5440 Bayreuth, Germany
Ansrucr
The high-pressurephase relationships in a HrO-saturated synthetic CaO-AlrO'-SiOrHrO (CASH) systemwere studied by multi-anvil experiments.The most extreme pressure
condition under which pure lawsonite [CaAlrSirO,(OH)r.HrO] exists is 120 kbar at 960
'C. This maximum stability is located at the intersection of the lawsonite breakdown
reaction with the topaz-OH + stishovite : phase "egg" reaction. At lower pressuresand
higher temperatureslawsonite decomposesto grossular + topaz-OH + stishovite + H2O,
whereas at lower pressuresand lower temperatureslawsonite first reacts to grossular +
phaseegg * topaz-OH + HrO and at still lower temperaturesto glossular * phaseegg *
diaspore + HrO. The two latter reactions have positive dP/dT slopes,with lawsonite on
the low-pressure side, and thus delimit the occurrenceof lawsonite toward higher pressures.The occrurenceof topaz-OH (10.7 wto/oHrO) is limited through a reaction to phase
'C to
eggQ.5 wto/oHrO) * diaspore; the phaseboundary extendsfrom I l0 kbar and 720
oC.
and
a
AlSiOr(OH)
of
have
a
composition
Phaseegg is inferred to
130 kbar and 920
monoclinic unit cell similar to that proposed by Eggletonet al. (1978).
The high-pressurebreakdown of lawsonite in CASH does not result in an anhydrous
assemblage.Lawsonite is known to occur experimentally in basaltic and andesitic compositions to at least 77 kbar; however, it is unknown whether topaz-OH and phase egg
appear in natural multicomponent systems,in particular those saturated in kyanite and
HrO.
saltic and andesitic compositions to pressuresof more
than 77 kbar; relatively cold thermal regimesin subducLawsonite [CaAlrSirOr(OH)r.HrO] is a relatively dense, tion zones permit lawsonite, formed at blueschist-facies
hydrous mineral with a wide P-l'stability field in syn- conditions, to persist in subducted crust to a depth of
thetic [CaO-AlrO3-SiO2-HrO (CASH)] and natural sys- more than 240km. Lawsonite does not commonly occur
tems. In addition to two OH groups,the lawsonite struc- in eclogitesfrom depths equivalent to > 15 kbar. This is
ture hosts an HrO molecule, and Al is completely in probably the result of thermal relaxation during the exoctahedral coordination (Rumanova and Skipetrova, humation process.Steep geothermal gradients are com1959). This results in a density of 3.09 g,/cm3,which is mon during the descentof oceaniccrust, however. When
higher than the 2.76 g/cm3density of its anhydrouschem- the descentcomes to a halt and exhumation begins,therical equivalent, plagioclase.In the CASH system, law- mal relaxation of the anomalous cold thermal structures
sonite forms from laumontite at approximately 3 kbar also begins. Consequently,most exhumation paths com(Nitsch, 1968)and persiststo more than92 kbar (Schmidt mence with a prograde temperature evolution (e.g., Enand Poli, 1994). Pawley 099$ found lawsonite to be gland and Thompson, 1984).The occulrenceof lawsonite
stable at 120 kbar and extrapolated the stability limit to in eclogitic xenoliths in a kimberlite pipe (Watson and
higher pressures.The maximum thermal stability of law- Morton, 1969), i.e., in a locality where extremely fast
sonite is 1040 'C at a pressureof 92 kbar (Schmidt and exhumation occurred, proves that eclogitesthat ascended
Poli, 1994). Pawley (1994) determined the maximum through relatively slow tectonic processesdo not necesthermal stability to ca. 1080 'C at 94 kbar. The reasons sarily preservethe most extreme high-pressure,low-temperature conditions. During the exhumation process,cold
for this small differencewill be discussedbelow.
In natural rocks subductedto high-pressure,low-tem- eclogitesare likely to follow initially prograde temperaperature conditions, lawsonite is common in blueschist ture paths, and consequently lawsonite reacts to form
terrains (mostly 3-15 kbar, seeEvans and Brown, 1986, zoisite. The latter reaction is documented in many ecloand referencestherein) but also occurs in basaltic xeno- gitic terrains (Newton, 1986, and referencestherein).
In this study two hydrous phaseswere encountered,
lithsinkimberlites(WatsonandMorton, 1969)thatoriginated from a depth equivalent of more than 25 kbar which are known from experimental studiesbut not from
(Helmstedt and Schulze, 1988). Poli and Schmidt (1995) the Earth's surface. Topaz-OH was first described by
'Wunder
et al. (1993a)and is a phasewith topaz structure
have shown experimentally that lawsonite occurs in ba1286
/1112-1286$02.00
0003-004x/95
IxrnonucrroN
SCHMIDT: UPPER PRESSURE STABILITY OF LAWSONITE
in which F is completely replacedby OH. Phaseeggwas
first synthesized by Eggleton et al. (1978) at approximately 1000'C and at pressuresgreaterthan 100 kbar.
Eggletonet al. (1978)did not name this phase.The name
phaseeggis assignedin this study to give credit to Eggleton et al. (1978). Although the HrO content of phaseegg
as inferred from this study is different from the original
study, X-ray difraction shows that both phasesare otherwise identical.
This contribution is intended to examine the upper stability of lawsonite in a synthetic CASH system and to
clarify whether other hydrous phases are formed from
lawsonite at its pressurebreakdown.
ExpnnrnrsxrAl
coNDrrtoNs
The experimentswere performed in a split sphereand
a split-cylinder-type multi-anvil equipped with WC cubes
of 32 mm edge length. Mg-octahedra (95 wto/oMgO, 5
wto/oCrrOr) of an edge length (M) of 14 mm were used
on a truncation edgelength (TEL) of 8 mm. Pyrophyllite
gasketswith a 2.5 x 5.0 mm cross section, cardboard of
0.4 mm thickness,and Teflon tape of 0.13 mm thickness
were directly glued onto the WC cubes.The 14 M/8 TEL
setup was calibrated against coesite-stishovite(yagi and
Akimoto, 1976)and a-B spinel (Katsura and lto, 1989)
and can be usedat pressuresranging from 90 to I 60 kbar.
The octahedrawere drilled and furnished with a zirconia
sleevefor thermal insulation and a steppedLaCrO, heater with a wall thicknessof 0.3 mm in the upper and lower
3.1 mm and with a wall thicknessof 0.5 mm in the central 2.7 mm. A MgO spacer separatedthe capsule from
the furnace, and Mo disks or rings were placed between
the LaCrO, heaterand the WC cube (for details seeRubie
et al., 1993). Capsuleswere made from 1.6 mm outer
diameter Pt tubing. The starting material was composed
ofsyntheticlawsonite + zoisite * grossular+ cristobalite
+ AI(OH)3, and of natural kyanite. Syntheseswere performed as describedin Schmidt and Poli (1994). Fe and
other trace elements in the natural kyanite were below
the detection limit of an electron microprobe. The bulk
composition of the starting material is located in the
quadrangle lawsonite-topaz-OH-phase egg-HrO (Fig. l).
Precisely l2wto/o HrO was presentin the starting material
bound in lawsonite, zoisite, and mostly in stoichiometric
A(OH)3. The starting material was filled into half capsules,which were then welded and pressedinto cylinders
with lengths varying from 1.4 to 2.6 mm. Temperatures
were measured with welded Pt-PteoRhr' thermocouples
(S-type).No pressurecorrection for the emf was applied.
The standard assembly includes only one axial thermocouple, however a few experimentswere performed with
a secondaxial thermocouple in the geometrical center of
the steppedheater.At 700-950 oC,temperaturegradients
over the length of the capsule were between 20 and 40
oC, the highest temperatur€ always
being located in the
geometric center of the furnace.The pressureprecision is
estimated to +4o/o(Walker, l99l), however the accuracy
is probably significantly better becauseP spacingsof 5
1287
CaO
Alp3
crn
Fig. 1. Startingmaterialcomposition.Square: bulk composition.Abbreviationsfor Figs. 1-3 areas follows:and: andalusite,coes: coesite,crn : corundum,dsp: diaspore,egg:
phaseegg,grs: grossular,
ky : kyanite,lws : lawsonite,pi :
phase"pi," sil : sillimanite,stish: stishovite,toz: topaz-OH,
v : HrO, zo : zoisite.
kbar yielded consistent results. The experimental products were measuredwith a STOE powder diffractometer,
equipped with a Co tube and a monochromator.
After the experiment, the amount of free fluid phasein
the capsule was not determined. In a sample weighing
approximately 4 mg, a total of 0.5 mg HrO was present
in the capsuleat the beginning of an experiment. At the
end of the experiments, depending on the resulting assemblage,0.1-0.4 mg of the HrO was bound in hydrous
phases,thus the remaining amount of free fluid phasewas
too small to measureby weight differencebecausepuncturing of a capsule generally results in small losses of
material from the capsule.Samplesthat obviously failed
becausethe thermocouple ceramics penetrated the capsule resulted in the anhydrous assemblagesgrossular *
stishovite * corundum or grossular + stishovite * kyanite. In some samples,compositions of the experimental
products were determined by a Cameca SX50 microprobe operatedat 12 kV and 20 nA.
In all samplesexceptexperiment lwma2 l, reaction was
always complete, resulting in a three-phase * HrO assemblage.Lwma2l, which was first analyzed in a longitudinal section by electron microprobe, showed minor
grossularinstead of lawsonite in the outer portion from
the middle of the capsule.This is interpreted as an oversteppingof the phaseboundary as a result of the thermal
gradient present in the capsule. The experiments contained lawsonite and grossular in the starting material.
However, they do not representtrue reversalsbecausethe
starting material did not contain topaz-OH or phaseegg
and SiO, was not presentas stishovite. However, because
reaction was always complete, the results are believed to
representequilibrium. The lack of apparentkinetic problems is probably due to the high reactivity of the fluid-
l 288
SCHMIDT: UPPER PRESSURE STABILITY OF LAWSONITE
TABLE
1, Experimental
conditions
andresultsof H2o-saturated
experiments
Expt.
lwma20
lwma23
lwma26
lwma27
lwma32
lwma30
lwma28
lwma21
lwma24
lwma36
lwma35
lwma29
lwma31
PTt
(kbar)
150
140
133
125
125
120
120
120
120
110
110
110
105
fc)
850
700
875
800
900
720
800
950
985
675
750
800
700
(min)
b5
340
210
270
't20
oo
175
720
260
oo
o4
188
120
1rto
Expt.products
grs, egg, dsp
grs, egg, dsp
grs, egg, dsp
grs, egg, osp
grs, egg, roz
grs, egg, osp
grs, egg, osp
lws, toz, stish,grs(-)grs, toz, stish
lws, dsp, stish
lws, toz, stish
lws, toz, stish
lws, toz, stish
F
5
o
130
'120
6 rro
100
90
80
8@
900
lemperalure (oC)
Fig. 2. Experimentallydeterminedphasediagram for the
lawsonite-out
reactionsin the HrO-saturated
systemfrom 90 to
refersto thepresence
I 50 kbar.The shading(solid/open)
oflawsonite,thesymbolshapeto theappearance
oftopaz-OH(circles),
phaseegg(diamonds),
Lawand diaspore+ stishovite(square).
sonite-outreactionat 90 kbar as determinedby Schmidtand
saturatedCASH systemat the P-Z conditions investigat- Poli (1994).The numbersare describedin the text. Abbreviaed. By contrast, the low nucleation energy ofthis system tionsasin Fig. l.
at high pressuresis a problem for the synthesisof large
crystals.Experimental products have an equilibrium tex- kyanite + HrO, as determined experimentally by Wunture with 120'phase angles,but the averagegrain size of der et al. (1993a).At higherpressurestopaz-OH reactsto
the experimental products is typically between 0.5 and 2 form AlSiOr(OH), a phasedenoted here as phaseeggthat
pm, and only a few crystalshave significantly larger sizes, was first synthesizedby Eggletonet al. (1978).
up to l0 pm. This averagegrain size did not increasewith
Eggleton et al. (1978) estimated the HrO content in
the duration of experiments, indicating that after initial their formula AI5Si5O,,(OH)by X-ray diffraction. Micronucleation and formation of a texturally equilibrated as- probe measurementsin severalsamplesresultedin totals
semblage,growth processesdid not play an important of 92-93 wto/o(the same samplesyielded totals of 99.1role.
100.5 wt0/0for grossular) implying an HrO content of
roughly 7-8 wto/o,which is consistentwith the most simExpnnrtvrcNTAr- REsuLTs
ple formula AISiO3(OH) (7.5 wto/oH,O). The X-ray powThe experiments resulted in phase assemblagescon- der diffraction of this study resulted in almost identical
taining lawsonite, diaspore,grossular,phaseegg,stishov- peak positions ofphase eggas determined by Eggletonet
ite, topaz-OH, and HrO. Experimental results are given al. (1978). From the best quality diffractogram (experiin Table I and Figure 2. The reaction positions in Figure ment lwma27), l4 reflectionswere selectedthat fulfill the
2 are constrained by the experiments and drawn in ac- criteria of Orville (1967). Cell parameters obtained by
fitting a monoclinic unit cell as proposed by Eggletonet
cordancewith Schreinemakerrules.
Between95 and 120 kbar the lawsonite breakdown re- al. (1978)are as follows:a: 9.820(7),b: 18.236(7),and
action in a HrO-saturated CASH system is
c : 5 . 3 7 2 @A) , p : 1 0 3 . 8 4 ( 7 )a" n d V : 9 3 4 . 0 + 1 . 0A ' .
Eggletonet al. (1978) determined a unit cell haingT2 O
3lawsonite: I grossular+ 2topaz-OH
atoms, thus for the composition AISiO.(OH) Z would be
(l)
* I stishovite* 4HrO.
18, resultingin a density of 3.84 g/cm3.
This reaction has a negative P-I slope of approximately
The breakdown of topaz-OH to phase egg occurs by
- 3 'Clkbar. The highest temperature stability of chemi- two reactions. With increasingpressure,these are
cally pure lawsoniteis 1040'C (Schmidtand Poli, 1994),
(2)
I topaz-OH + I stishovite:2egg
located at the low-pressureend of Reaction I where Reaction I intersectsthe kyanite + HrO : topaz-OH (Wun- and
der et al., 1993a)and coesite : stishovite (Yagi and Ak(3)
Itopaz-OH: legg+ ldiaspore.
imoto, 1976)phaseboundaries(invariant points 4 and 3,
Fig. 2). Whether the lawsonite breakdown reaction inter- At P-7 conditions lying between Reactions 2 and 3 (Fig.
sectsthe kyanite + HrO : topaz-OH hydration reaction 2),topaz-OH and phaseeggcoexist. Both reactions have
slightly above, at, or below the coesite : stishovite tran- positive P-7" slopes and intersect at approximately 700
sition is beyond the experimental resolution of a multi- "C, 110 kbar. At 900'C they differ in pressureby < l0
anvil apparatus.The two possible topologies are shown kbar.
in Figure 3.
The most extremepressurecondition under which pure
Topaz-OH [AI,SiO4(OH),, 10.7 wto/oHrO] forms from lawsonite exists is 120 kbar. 960 'C. at the intersection
Nofe; Abbreviations of experimental products are as tollows: grs
grossular, egg : phase egg, dsp : diaspore, toz : topaz-OH, lws
lawsonite,stish : stishovite.
'The minus sign indicatesthat a very minor amount was present.
t289
SCHMIDT: UPPER PRESSURE STABILITY OF LAWSONITE
160
150
140
130
I zv
110
100
(U
-o
.:<
90
o
at
o
o
o-
80
70
60
50
40
30
20
10
0
300
400
500
600
700
800
900
1000
1100
1200
temperature(oC)
Fig. 3. Topology of the CASH system to 150 kbar with reactions delimiting the stability fields of lawsonite and zoisite in
CASH and ofdiaspore, phase"pi," topaz-OH, and phaseeggin
ASH. Thick lines delineate the lawsonite and zoisite stability
fields as experimentally determined by Schmidt and Poli (1994)
and in this study. Quartz : coesitefrom Bohlen and Boettcher
(1982), coesite: stishovite from Yagi and Akimoto (1976), topaz-OH : kyanite * HrO from Wunder et al. (1993a), and reactions involving phase "pi" from Wunder et al. (1993b). Experimental brackets for the equilibrium rliaspore-corundum from
Grevel et al. (1994, squares),Schmidt and Poli (1994, triangles),
and Vidal et al. (1994, diamonds); open symbols : diaspore,
solid symbols : corundum. The arrowheads at 120-150 kbar
indicate diaspore stability; the phase boundary is extrapolated
becausea high-temperatureconstraint at these pressuresis not
available. Solid circles are invariant points with P-Z locations
well defined by experiments; exact P-I locations of invariant
points representedby open circles remain uncertain. Arrows indicate in which direction(s) the latter invariant points might be
displaced. Note that the high-pressurelawsonite delimiting reactions are hydration reactionswith increasingtemperature.The
inset shows the second possible topology at conditions around
90 kbar, 1040 "C. Abbrewiationsas in Fig. 1.
1290
SCHMIDT: UPPER PRESSURE STABILITY OF LAWSONITE
50-120 kbar topaz-OH (p : 3.37 g/cm3)forms (Wunder
et al., 1993a).At pressureshigher than 90-120 kbar phase
e8B(p : 3.84 g/cm3)forms instead of topaz-OH. At temperaturesbelow 430-690'C stishoviteor coesitecoexists
with diaspore. The high-temperature behavior of phase
3 lawsonite : I grossular * I topaz-OH
egg and topaz-OH remain unknown; they could either
(4)
+2eEE+4H,O
melt or decomposeto stishovite * corundum + HrO.
and transforms at temperaturesbelow Reaction 3 into
Drscussrox oF pREtrrous EXpERTMENTALwoRK
3 lawsonite : I grossular + 3 egg
Pawley (1994) also determined the position of Reac(5)
+ldiaspore+4HrO.
tion I and estimated the maximum temperaturestability
Reactions 4 and 5 have positive slopesof approximately of lawsonite. Considering the relatively large uncertain40 and 25 "C/kbar, respectively, with lawsonite on the ties of Pawley'sexperiments,her experimental resultsare
high-temperature-low-pressure side. Thus, these reac- in fair agreementwith the present study. The results obtions delimit the stability of lawsonite toward higher tained in this study are considered to be more accurate
pressuresand lower temperatures.
becauseof the following differencesin the experimental
Diaspore is an experimental product in all experiments setup: (l) Pawley (1994) did not calibrate pressureabove
performed at pressuresabove the topaz-OH destabiliza- 90 kbar for the octahedra used on a 4 mm truncation
tion. As a result, the maximum pressurestability of dia- (TEL). Only a half-bracket at 140 kbar is reported. Thus,
spore exceeds 150 kbar (at 700-900 "C). Experiments Pawley's pressureshave a larger error than the present
conducted at temperaturesabove 900'C resulted in to- study, in which a complete calibration was undertaken
paz-OH and thus are not suitable to examine the thermal and a much larger octahedrawas used for the same presstability of diaspore becausethe bulk composition is lo- sure range (14 mm octahedra on 8 mm truncation edge
length). (2) Pawley (1994) used an inconel furnace, i.e., a
cated between topaz-OH and the SiO'-apex (Fig. l).
metal heaterwith a positive correlation betweenelectrical
Topor,ocy
resistancyand temperature.Such heatersalways result in
The topology of a H,O-saturated CaO-Al,O3-SiO,-HrO sharp hot spots, and consequentlyPawley (1994) found
system (Fig. 3) is constructed from the present experi- temperaturegradientsin excessof 100'C (at 1000'C) for
ments and from those of Schmidt and Poli tl99$ and a capsulelength of about 3 mm. Pawley points out that
Wunder et al. (1993a, 1993b).The breakdownreactions the "autoamplification" effect on the thermal gradient is
of lawsonite depend mostly on phaserelationships in the less pronounced for inconel furnaces than for rhenium
AlrO3-SiOr-HrO (ASH) subsystem.It is only at the in- furnaces,the latter resulting in the largestthermal graditersection of the lawsonite-out reactions with the pres- ents of all furnace materials commonly used in multisure-sensitivebreakdown reaction of zoisite to grossular anvil experiments.However, suchlargethermal gradients
+ kyanite + coesite + HrO that an invariant point (1, are not inherent to multi-anvil experiments. LaCrO, is
Fig. 3) is defined by three reactions between Ca-bearing one of the rare materials suitable for furnaces in which
phases.All other invariant points are defined by ASH electrical resistivity decreaseswith temperature.This bephase relationships, which determine the lawsonite-out havior causeswann zones to heat less than cold zones,
reactions and the topology ofthe CASH system at pres- and consequentlythe hot spot is spreadout. In addition,
suresabove 67 kbar. All lawsonite-out reactions at these the LaCrO, furnaces in the present experiments were
conditions are of the type lawsonite : grossular + 2 ASH steppedheaters,i.e., the electrical resistancyis lower in
phases + HrO. As the densities of the ASH phasesin- the central, thicker part of the furnace. This resulted in
oC over the
crease,lawsonite changesfrom the high-pressureside to temperaturegradients on the order of 20-40
the low-pressure side of its breakdown reactions. Very capsulelength of 2.6 mm (both experimental setupsusing
unexpectedly,Reactions 4 and 5 are hydration reactions 8 mm TEL).
Eggletonet al. (l 978) synthesizedphaseeggin a Bridgewith increasingtemperature, i.e., HrO is situated on the
low-temperature side of Reactions 4 and 5 (Figs. 2 and mann-anvil apparatusat "pressuresgreaterthan I 00 kbar
3). This indicates that phase egghas, besidesa relatively and at about 1000 "C." Becauseprecise synthesiscondihigh density, also a relatively low thirdlaw entropy. An- tions are not given and pressuresand temperatureswere
other hydration reaction with increasing temperature is not as well calibrated as can be presently achieved, this
constituted by the low-temperature(<700 .C), high-pres- first synthesisappearsto be in agreementwith the lower
pressure stability limit defined in this study. Phase egg
sure formation reaction of lawsonite (Fig. 3).
The principal featuresof the ASH topology (Fig. 3) are was also synthesizedby Fockenberget al. (1994) at 150
oC, and by Pawley (1994) at 140 kbar,740as follows: At relatively low pressuresof 20-60 kbar and kbar, 1000
'C,
phase 840 "C. Pawley (1994) recognizedthat the estimatedHrO
relatively low temperatures from 450 to 600
"pi" [AlrSirO'(OH)r, p: 3.23 dcm'] appears(Wunder contentof 1.6 wto/o(Eggletonet al., 1978)is probably too
et al., 1993b); the stability field of this phase entirely low. The differenceof Al,O, + SiO, totals from 100 wto/o
overlaps that of lawsonite. At intermediate pressuresof in microprobe analysesled her to conclude that the for-
of the lawsonite breakdown reaction with phaseeggforming Reaction 2 (invariant point 5, Fig. 2). At pressures
above Equilibrium 2 the lawsonite breakdown reaction
is
SCHMIDT:UPPERPRESSURE
STABILITY OF LAWSONITE
mula is probably AISiO3(OH). This is confirmed by the
presentstudy. However, a direct measurementof the HrOcontent ofphase egg is necessaryto obtain a precise determination of its stoichiometry.
A high-pressuretopology for the ASH systemwas also
presentedby Pawley (1994, her Fig. 6). Invariant point 8
(Fig. 3) [her invariant point (ky, coes,v)] is similar in that
arrangement,and stabilities of reactions around this invariant point are identical except that reaction slopesare
now constrained experimentally. The high-temperature
parts ofthe topologies are different becausethe reaction
diaspore : corundum + HrO was not considered by
Pawley (1994). The low-pressure part of the topologies
are different becausethe formation of phase "pi" from
topaz-OH (Wunder et al., 1993b) was intentionally not
consideredby Pawley(1994, p. 105).
THn occunnrNcE oF LAwsoNrrE,
rOp.Az.OH" AND PHASEEGG IN NATURE
The present study also indicates conditions of maximum persistenceof lawsonite in natural rock compositions. Lawsonite has been shown experimentally to be
presentin basalts,andesites,and probably greywackesto
pressuresof at least 77 kbar. Once formed at blueschist
conditions, lawsonite remains stable in natural rocks to
temperaturesabout 100-150 "C lower than in the synthetic CASH system (i.e., to 780 "C at 60 kbar in basalt
and to 900 "C at 77 kbar in andesite:Poli and Schmidt,
1995).Poli and Schmidt (1995)also showedthat in narural rock compositions lawsonite has only minor impurities (l-2 wto/o),whereasgrossularcontents in garnet increasewith pressurefrom l5 to 45 mol0/0.The maximum
stability of pure lawsonite is reached at 120 kbar (950
'C), equivalent to a depth of 350 km. In natural rocks,
impure lawsonite is expectedto decomposeat somewhat
lower pressures.
Thermal models predict that relatively cold structures
are common in subduction zones with fast subduction
rates where relatively old crust is subducted:For the top
of the slab, Hsui and Toksiiz (1980) predict 650-800.C
at 400 km depth, Davies and Stevenson(1992)600-800
'C at 400 km depth, Ito and Sato (1992) 600-900
"C at
400 km depth, Furukawa (1993) ca. 650 "C at 200 km
depth, and Peacock(1993) ca. 800'C at 250 km depth.
These calculations show that temperatures are low enough
in cold subduction zonesto permit the occurrenceof lawsonite up to its maximum pressure stability. Nevertheless, a precise determination of lawsonite occurrencein
natural rock compositions is necessaryto elucidate this
point further.
In the synthetic system, hydrous breakdown products
of lawsonite are topaz-OH, phase egg, or both. TopazOH is expectedto appear in natural bulk compositions
saturatedin kyanite at relatively low-pressureconditions,
i.e., metapelites and some greywackes.In such compositions, the dehydration oflawsonite (l I wto/oHrO) would
result directly in the formation of topaz-OH (10.7 wto/o
HrO) at pressuresbetween 90 and 120 kbar. Direct ex-
t29l
perimental evidence for such a reaction is not available;
nevertheless,at pressuresabove 35 kbar, lawsonite decomposes in andesites (which have bulk compositions
similar to greywackes)to garnet * omphacite + kyanite
+ HrO. Thus, similar or more aluminous bulk compositions (e.g., metapelites,greywackes,or metagranitoids)
potentially could bear topaz-OH. However, it remains
uncertain if topaz-OH occurs in the complex multisystems that constitute natural rocks.
Phaseegg,having a composition located between kyanite (topaz-OH) and stishovite, could appear in compositions that are lessaluminous than those necessaryfor
the formation of topaz-OH. Becausephase egg has not
been reported from systemsmore complex than CASH,
its appearancein nature remains speculative. However,
the pressuredestabilization oflawsonite in natural rocks
is expectedto result in a high-density phase,which could
be the hydrous phaseeggor another new phase.The significance of the high-density, hydrous phasestopaz-OH
and phase egg for natural systems (if any) needs to be
clarified by additional experiments.
AcxNowr.nncMENTs
This project follows a previous study conducted in collaboration with
Stefano Poli, to whom I am grateful. I thank my colleagues at the BayerischesGeoinstitut, especiallyRoss Angel, Bob Linnen, Brent Poe, and
Dave Rubie, for their help and support. I also thank A. Pawley and J.G.
Liou for reviewing the paper
Rr,mnnNcrs crrED
Bohlen, S.R., and Boettcher, A.L. (1982) The quartz + coesitetransformation: A precisedetermination and the effectsofother components.
Journal of GeophysicalResearchB, 87, 7073-7078.
Davies, J.H., and Stevenson,D.J. (1992) Physicalmodel ofsource region
of subduction zone volcanics. Journal of Geophysical Research,97,
2037-2070.
Eggleton,R.A., Boland, J.N., and Ringwood, A-8. (1978) High presswe
synthesisof a new aluminium silicate: Al,SirO,,(OH). Geochemical
Journal. 12. L9l-194.
England, P.C., and Thompson, AB. (1984) Pressure-temperature-time
path of regional metamorphism: I. Heat transfer during the evolution
of regions of thickened continental crust Journal of Petrology, 25, 894928
Evans, B.W., and Brown, E.H. (1986) Blueschistsand eclogites.Geological Societyof America Memoir, 164,423 p.
Fockenberg,T., Schreyer,W., Skrok, V., and Wunder, B. (1994) Experimental studiesrelevant to ultrahigh-pressuremetamorphism (abs.),p.
12 l. International Mineralogical Association, 16th General Meeting,
Pisa, Italy.
Furukawa, Y. (l 993) Depth ofthe decouplingplate interfaceand thermal
structure under arcs.Joumal ofGeophysical Research,98, 2000!200 I 3.
Grevel, K.D., Fockenberg,T., Wunder, B., and Burchard, M. (1994) Experimental determination of the equilibrium curve 2 diaspore : corundum + H,O to high pressures and modified thermodynamic data
for diaspore.Terra Abstmcts, 6, supplement l, 20.
Helmstedt, H., and Schulze,D.J. (1988) Eclogite-faciesultramafic xenoliths from Colorado Plateau diatreme breccias: Comparison with eclogites in crustal environments, evaluation ofthe subduction hlpothesis,
and implications for eclogite xenoliths from diamondiferous kimberlites. In D.C. Smith, Ed., Eclogitesand eclogite facies rocks, p. 387450. Elsevier Science,Amsterdarn.
Hsui, A.T., and Toksilz, M.N. (1980) The evolution of thermal structures
beneathsubduction zones.In Development in Geotectonics,14,325342.
Ito,8., and Sato, H. (1992) Etrect ofphase transformations on the dy-
r292
SCHMIDT: UPPER PRESSURE STABILITY OF LAWSONITE
namics of the descendingslab.In Y. Syonoand M.H. Manghnani, Eds.,
High-pressureresearch:Application to earth and planetary sciences,p.
257-262 Terra Scientific,Tokyo.
Katsua, T., and Ito, E. (1989) The systemMgrSiOo-FerSiOo
at high pressuresand temperatures:Precisedetermination of the stabilities of olivine, modified spinel and spinel. Journal ofGeophysical Research,94,
l 5 6 6 3 -l 5 6 7 0 .
Newton, R.C. (1986) Metamorphic temperaturesand pressuresof group
B and C eclogites.GeologicalSocietyof America Memoir, 164, 17-30.
Nitsch, K.H. (1968) Die Stabilitat von Lawsonit. Naturwissenschaften,
s5,388.
Orville, P.M. (1967)Unircell parametersof the microcline-lowalbite and
the sanidine-highalbite solid solution series.American Mineralogist,
52,55-86.
Pawley, A. (1994) The pressureand temperature stability limits of lawsonite: Implications for HrO recycling in subduction zones.Contributions to Mineralogy and Petrology, I 18, 99-108.
Peacock,S.M. (1993) Large-scalehydration ofthe lithosphereabove subducting slabs.Chemical Geology, 108, 49-59.
Poli, S., and Schmidt, M.W. (1995) HrO transport and releasein subduction zones:Experimental constraintson basaltic and andesiticsystems.
Journal ofGeophysical Research,in press.
Rubie, D.C., Karato, S., Yan, H., and O'Neill, H St.C. (1993) Low differential stressand controlled chemical environment in multianvil highpressureexperiments. Physics and Chemistry of Minerals, 20, 315322.
Rumanova, I.M., and Skipetrova, T.I. (1959) The crystal structure of
lawsonite. Soviet PhysicsDoklady, 4,20-23.
Schmidt, M.W., and Poli, S. (1994) The stability of lawsonite and zoisite
at high pressures:Experiments in CASH to 92 kbar and implications
for the presenceofhydrous phasesin subductedlithosphere.Earth and
PlanetaryScienceL€tters, 124, 105-1 18
Vidal, O., Theye, T., and Chopin, C. (1994) Experimental study of the
reactionsbetweenchloritoid and almandine at high pressureand various/, conditions. Contributions to Mineralogy and Petrology, 118,
256-270.
Walker, D. (1991) Lubrication, easketing,and precision in multianvil
experiments.American Mineralogist, 76, 1092- l 100.
Watson, K.D., and Morton, D.M. (1969) Eclogiteinclusionsin kimberlite
pipes at Garnet Ridge, northeastem Arizona. American Mineralogist,
54,267-285.
Wunder, B., Rubie, D.C., Ross, C.R., II, Medenbach,O., Seifert,F., and
Schreyer,W. (1993a) Synthesis,stability, and properties ofAl'SiOo(OlI)r:
A fully hy&ated analogueof topaz. American Mineralogist, 78,285297.
Wunder, B., Medenbach,O., Krause, W., and Schreyer,W. (1993b) Synthesis,properties and stability of AlrSirO,(OH), (phasePi), a hydrous
high-pressurephase in the system AlrO3-SiOr-H'O (ASH) European
Journal of Mineralogy, 5, 637-649.
Yagi, T., and Akimoto, S.I. (1976) Direct determination of coesite-stishovite transition by in-situ X-ray measurements.Tectonophysics,35,
259-270.
Mencg 15, 1995
MeNuscrrrr REcETVED
MeNuscnrrr AccEFTEDJutv 7, 1995