Compressibility of grunerite Lr Zn,LNcr* HaNs Ausn.urs, Ar ...

American Mineralogist, Volume 77, pages480-483, 1992

Compressibility of grunerite

Lr Zn,LNcr* HaNs Ausn.urs, Ar,r Kurocr.uo SrrrnN S. H,lrNrn

Institute of Mineralogy, University of Marburg, 3550 Marburg, Germany

Ansrnlcr

The unit-cell parametersof natural grunerite were measuredin a diamond anvil cell at
nine different pressuresup to 5.1 GPa. The bulk modulus Ko and its pressurederivative
Ko' were calculatedby fitting the pressure-volumedata to a Birch-Murnaghan equation of
state:Ko : 50(l) GPa and K0' : l3(l). The linear compressibilitiesB", Bu,and p,are
4.97(6),3.50(4)a,nd 3.62(5)x l0-3 GPa-', respectivelyy,ieldingthe ratioB":Bu:p,:1.418(l):
1.000(2):1.032(5G). runerite exhibits stronganisotropiccompressionin the low-pressure
region (P < I GPa). With increasingpressure,the compressionbecomesmore isotropic.
This behavior and the larger volume compressibility compared to that of Ca-rich clino-
pyroxenesin the pressurerangestudied are proposedto be due to the strong compression
of the empty A cavities in the crystal structureat low pressures.

INrnoouc:ttor,g Aloo*)for the eight tetrahedralpositions, and (Orr,OH, rr,

The crystalstructureof the iron magnesiummanganese Foor,Clo.of,o) r the anions (Chen et al., 1984).A single
amphibole grunerite (Finger, 1969) consistsof corner- crystal with dimensions of approximately 0.15 x 0. I x
linked tetrahedral double chainsconnectedby ribbons of 0.05 mm3 was selectedoptically and with the X-ray pre-
edge-sharingoctahedra. Four crystallographically non- cessionmethod to ensurediffraction quality.

equivalentpolyhedralpositions,designatedas Ml, M2, High pressureX-ray diffraction technique
M3, and M4 may be distinguishedM. g and Fe occupyall
four positions, Fe generally preferring M4 (Hafner and The measurementswere made by use of a modified
Ghose, l97l). A large empty cavity (A site) exists be- miniature Merrill-Bassett diamond anvil cell with four
tween the back-to-back tetrahedral double chain. With screwss, imilarto that of Mao and Bell (1980).An Inconel
this structure in mind, it is interestingto compare com- 750X gasketpreindentedto 0.085 mm with a hole of 0.3
pressibilities,e.g.,with thoseof pyroxenes,in which there mm wasused.A 4:l methanolto ethanolmixture served
are only two types of polyhedra, Ml and M2, and no asa hydrostatic pressuretransmitting medium (Piermari-
large A-type cavities. The purpose of this study was to ni et al., 1973),and a ruby chip of about 0.025 mm was
comparethe compressionbetweenamphibole and pyrox- included in the cell aspressurecalibrant. The hydrostatic
eneand to evaluatecorrelationsbetweenthe compression pressurewas measuredbefore and after each X-ray dif-
of the bulk volume and local field parametersat structure fraction experiment by use of the pressure-inducedshift
sitesmeasuredwith 57Fegamma resonanceat high pres- of the Rl fluorescenceline of ruby. The error in pressure
sures(Zhanget al., 1989a),usingthe sameamphiboleof calibration wasestimatedto be 0.05 GPa. Data measure-
relatively simple chemical composition. ment was carried out on a Stoefour circle diffractometer

ExprnrprnNTAl DETATLS

Sample !oo ta
o 12
The grunerite sample is from a metamorphic rock of I 0123 45 012 34 5
greenschistto low amphibolite faciesin the Gongchang- I
ling iron mine, China. The mineral associatedwith garnet
and magnetite occursas veins in the iron ore or in am- 345
phibolites.The yellowishgreenerystalsappearin the form
of fibrous needles,which have dimensions of 1 mm or
more in length and 0.I mm or lessin width. They often
contain small fractures.The compositionis (Naoorfuo,-
Fel.jrMg,ouFelf"Aloo,f)or the sevenM positions,(Sirrr-

P (GPd P (GPa) P (GPa)

Fig. l. Compression of the unit-cell axes; a/ao, b/bo, all.d
* Presentaddress:Max-Planck-Institut ftir Chemie. Postfach c/co.The symbols ao, bo,co representthe cell parametersmea-

3060, W6500 Mainz, Germany. suredat ambientconditions.

oo03-oo4x/ 92/0506-0480$02.00 480

ZHANG ET AL.: GRUNERITE COMPRESSIBILITY 481

TABLE1. Unit-cellparametersof grunerite between ambient pressure and 5 GPa asinp

Pa b (A)
(A)
(GPa) (A) (A) f) (4") 9.337
9.289
0.0 9.5390(5) 18 . 3 2 1 )( 1 s.3345(4) 10 1 . 8 2 ( 1 ) 912.5(1) 9.234
o.4 101.95(5) 903.9(5) 9.202
9.49s(5) 18.288(2) 5.320(2) 102.26(41 894.5(4) 9.153
102.42(4) 887.5(4) 9j24
1.1 9.451(4) 18.248()1 5.307(1) 102.63(5) 875.8(6) 9.109
102.73(6) 868.1(7) 9.081
1.7 9.422(41 18 . 2 1 ( 1 ) 5.296(2) 1't0022..7887((s8)) 863.8(6) 9.058
859.7(9)
2.7 9.380(5) 18.139(2) s.275(21 103.02(6) 853.4(7)

3.5 9.3sqs) 18.089(2) 5.259(2)

3.9 9.339(5) 18.061(2) 5.251(2)

4.4 9.315(7) 18.047(3) s.246(3)

5.1 9.297(71 17.994121 s.236(3)

/Vote.'Standarddeviationsin the last decimalplace are given in parentheses.

using MoKa radiation monochromated by a graphite than those of a powdered sample of the same specimen
crysral(I : 0.7107A).
measuredin a lever-arm NBS diamond anvil cell up to
The unit-cell parameterswere measuredat nine differ- l0 GPa (Zhanget al., 1989a).As shownin Table l, the
ent pressuresbetweenambient pressureand 5.1 GPa. For B angleincreaseswith increasing pressure.To facilitate
this, reflectionswere selectedas follows: Tllre20rangesof the comparisonwith the resultsof Ca-rich clinopyroxenes
the powder lines from Be disks supporting the diamond (Zhang et a1., 1989b), the cell parameterswere recalcu-
anvils were determined first. Subsequently,those reflec- lated for spacegroup I2/m (Suenoet al., 1913);they are
tions that are not in the 20 rangesof the powder lines presentedin Table 2. Herc 0 decreaseswith increasing
from Be were centered. The unit-cell parameterswere pressure,showing the same trend as that of Ca-rich cli-
determined from 15 to 19 reflections,2d rangingfrom 15 nopyroxenes.No discontinuity in the unit-cell parame-
to 42. Each reflectionwas centeredin eight positions ters over the entire rangeofpressure studiedwas observ-
(King and Finger, 1979) to reducezero and crystal-cen-
tering errors. The influence of counting times on the re- able within the experimentalerror.
sults of refinement was estimated through experiments.
The unit-cell parametersobtained by centeringa reflec- The data reveal an anisotropic compressionof the unit
tion in 20 s per stepexhibited halfthe standarddeviation cell. The linear compressibilities8", po, and 0. are shown
of those measuredin I s per step.The resultspresented in Table 3, and axial compressionsare plotted in Figure
here were measuredusing a counting time of 20 s per l. It is clear from Table 3 that the a axis is the most
step. The cell parametersat ambient pressurewere de-
termined by useof an automated Siemens-D500powder compressibledirection, whereasb and c have nearly the
diffractometer.
samecompressibilities.
The unit-cell parameterswere calculatedby means of The unit-cell volume at 5.1 GPa is 6.50/osmallerthan
the procedureofRalph and Finger(1982).Therewasno
significant deviation betweenthe results without mono- at ambient pressure.In Figure 2 the least-squaresfit of
clinic constraints and with the constraints. The results the pressure-volumedata to a Birch-Murnaghan equation
presentedhere were refined with the constraints. of stateis presented.It yields an isothermalbulk modulus
K. of 50(l) GPa, and its pressurederivative Ko' yields
RBsur-rs one of l3(l). Assumptionof Ko' : 4 yieldsKo: 63(l)
GPa. This, however, doesnot seemappropriate for gru-
The unit-cell parametersofgrunerite are listed in Table nerite, sincethe pressure-volumedata could not be fitted
l. They are generally I order of magnitude more precise properly.

The axial ratios a/b and a/c decreasewith increasing
pressure,whereasthe ratio b/c is nearly invariable. The

relationshipsare shown in Figure 3.

TABLE2. Unit-celpl arametersof gruneriteat pressuresassumingspacegroupl2lm*

Pabcp asinB

(GPa) (A) (A) (A) f) (A") (A)

0.0 9.929 18.321 5.3345 109.90 912.5 9.337
o.4 9.876
18.288 5.320 109.85 903.9 9.289

1.1 9.807 18.248 5.3C7 109.66 894.5 9.234

1.7 9.765 18.211 5.296 109.56 887.5 9.202

2.7 9.704 18.139 5.275 109.40 875.8 9.153

3.5 9.667 18.089 5.259 109.32 868.1 9.'t24

3.9 9.649 18.061 5.251 109.28 863.8 9.109

4.4 9.619 18.047 5.246 109.25 859.7 9.081

5.1 9.587 17.994 5.236 109.13 853.4 9.058

* Calculatedaccordingto the ,2/rn cellwith transformationmatrix 101/010/00-1 basedon the data of the CZm cell listed in Table 1. This was carried
out for comparisonwith pyroxene.

482 ZHANG ET AL.: GRUNERITE COMPRESSIBILITY

1.OO

o.99 r. 'Il- r T'I ,,J
o.98 a tt, a ,
l_ 1l I I T lrlltllr
o o.97 1t
o12345 ,t I
P (ea) o12345
P {@a) o12345
P (GPa)

Fig.3. Axial ratiosa/b, a/c, andb/cvs.pressure.

o.95

cell volume in that pressurerange.The study of the struc-

o.94 ture of diopside at high pressures(Levien and Prewitt,
l98l) and that of amphiboleat high temperature(Sueno

et al., 1972, 1973)showedthat compressionand expan-

o123456 sion of single and double chain silicatesare mainly due
to compressionand expansionofthe A and M polyhedra.

P (GPa) Basedon this and related hyperfine data, it is reasonable
to conclude that at pressures< I GPa the large volume

Fig.2. Compressionof unit-cell volume. Zois the valuemea- compressibility of the unit cell must be accountedfor by
sured at ambient conditions. The solid line was calculatedwith
Ko : 50(l) GPa and &' : l3(l); the dashedline wascalculated a collapseofthe A cavities.
with Ko : 63(l) GPa and Ko',: 4. Preferential compressionof the A cavities exists only

at pressures< I GPa. At pressures> I GPa, the com-

pressionof M polyhedramay be the dominant factor gov-

DrscussroN erning the bulk unit-cell compression,similar to the com-
pressionmechanismin Ca-rich clinopyroxeneswherethe
It is interestingto note that the axial ratios a/b anda/c
decreasesignificantly with increasing pressurebetween unit cell is compressedby the same amount as the vol-
ambient pressureand approximately I GPa, whereasat umes of the Ml and M2 polyhedra (Levien and Prewitt,
pressurebetween I and 5 GPa they are nearly constant
(Fig. 3). This indicatesthat a is compressedmore strongly r9 8l ) .
relative to b and c in the low-pressureregion, while at
higher pressuresit is compressedapproximately the same The close relation betweenthe large linear compress-
as b and c. Thus, the apparently larger linear compress-
ibility 6" relative to B, and 6. betweenambient pressure ibility d, and the collapseof A cavities in the low-pressure
and 5 GPa is the result of greatercompressionof the a range explains why the largestlinear compressibility in
axis at low pressures.At pressuresP > I GPa the linear grunerite is parallel to the 4 axis while it is parallel to the
compressibilities approach each other. In other words,
the anisotropic compressionin the low-pressurerangebe- D axis in Ca-rich pyroxenes.Suenoet al. (1973) showed
comesmore or lessisotropic with increasingpressure. that a is the direction of largest thermal expansion in
tremolite becauseof the considerableexpansionof the A
The changein slopeof a/b and a/c tear I GPa may be cavity in the direction perpendicularto the b,cplane.This
correlatedwith those of the isomer shifts Dand quadru- observationis also consistentwith our analysis.From the
pole splittings AEo plotted vs. unit-cell volume (Figs. 5, discussionabove, it is to be expectedthat with greater
6 in Zhang and Hafner, 1992a).Betweenambient pres-
sureand I GPa, 6 and AEo, which are directly related to shrinking of the A cavities, the compressionmechanism
the relevant polyhedral volumes, remain approximately in grunerite will be more similar to that in pyroxenes.
constant. This means that the M polyhedra are not re-
sponsiblefor the observedbulk compressionof the unit- The significantly larger volume compressibility of gru-
nerite, 6 : 12.69(8) x l0 3 GPa ', compared with that
of Ca(Fe,Mg)Si,Ouclinopyroxenes(hedenbergite0: : 7.62
x l0-3 GPa-r) in the pressurerangestudied(Zhangand

Hafner, 1992b)may be attributed partly to the compres-
sion of the A cavity, although the changein geometrical

configuration of the M4 polyhedron may be important as

well. For further substantiation, precisestructure deter-

Traur 3. The linearcompressibilitioefsgrunerite(in unitsof minations of gruneriteat high pressureswill be necessary.
x10 3/GPa) Comparison of the unit-cell parametersofgrunerite be-

R P.iPo:P" tweenambient temperatureand 4.5 K (Ghoseet al., 1987)
with those at high pressuresand ambient temperature

4.e7(6) 3.50(4) 3.62(s) 12.69(8) 1.418(1):1.00(2):1.032(5(T) able l) showsthat the effectof decreasingtemperature

Note: P., Pb P" arc the linear compressibilities.The errors in the last from ambient temperatureto 4.5 K is about equal to the
decimalplace are given in parentheses.They were calculatedthrough an effectof increasingpressurefrom ambient pressureto 0.4
GPa at 293 K. Whether or not this is an inversepressure,
error propagationprocedureusing standard deviationsof Table 1.

ZHANG ET AL.: GRUNERITE COMPRESSIBILITY 483

temperaturebehavior of the unit cell, as shown for hed- Kine. H.E., and Finger, L.W. (1979) Diffracted beam crystal centering
enbergite(Zhang and Hafner, 1992b),needsto be inves-
tigatedin more detail.Technicallya pressureof about0.5 and its applicationto high pressurecrystallography.Joumal ofApplied
GPa can be generatedthrough differential thermal con-
traction of the sampleholder and sampleat low temper- Crystallography,12, 374-37 8.
atures.
Levien, L., and Prewitt, C.T. (1981) High-pressurestructurestudy ofdi-
AcxNowr,e ncMENTS
opside.American Mineralogist, 66, 315-323.
We thank R. Allmann, Univenity of Marburg, for stimulating discus-
sion, Chen Guanyuan, Chinese University of GeosciencesB, eijing, for Mao, H.K., and Bell, P.M. (1980) Design and operation of a diamond-
supplyingthe gruneritesample,and L.W. Finger, CarnegieInstitution of
Washington,Washington,DC, for providing a copy of his program for window, high pressurecell for the study ofsingle crystalsamplesloaded
the refinement of the lattice constants.This work was financially sup-
ported by the DeutscheForschungsgemeinschaaftnd Bundesministerium cryogenically.CarnegieInstitution ofWashington Year Book, 79,409-
fiir Forschungund Technologiegrant HAIMAR6.
41t.
RnrnnnNcns crrED
Piermarini, G.J., Block, S.,and Barnett,J.D. (1973)Hydrostatic limits in
Chen, Guanyuan, Li, Meihwa, Wang, Xuefang, Sun, Daisheng, Sun,
Chuanmin, Wang,Zufu, Su,Yuxuan, and Lin, Jiaxiang(1984)Genetic liquids and solidsto 100 kbar. Journal ofApplied Physics,44,5377-
mineralogy of Gongchanglingiron mine. Minerals and rocks, special
issue,vol. 6, p 74-1O8. Mineralogical,Petrological,and Geochemical 5382.
SocietyofSichuan Province, China, and ChengduCollegeofGeology
(in Chinese). Ralph, R.L., and Finger,L.W. (1982)A computerprogramfor refinement

Finger, L.W. (1969) The crystal structure and cation distribution of a of crystal orientation matrix and lattice constants from diffractometer
grunerite.Mineralogical Societyof America, SpecialPaper2, 95-100.
data with lattice symmetry constraints.Journal of Applied Crystallog-
Ghose,S.,Cox, D.E., and Dang, N.V. (1987)Magneticorder in grunerite,
Fe,SirOrr(OH),-A quasi-onedimensionalantiferromagnetwith a spin raphy, 15,537-539.
canting transition. Physicsand Chemistry of Minerals, 14, 36-44.
Sueno,S., Papike, J.J., Itewitt, C.T., and Brown, G.E. (1972) Crystal
Hafner, S.S.,and Ghose, S. (1971) Iron and magnesiumdistribution in
cummingtonites,(Fe,Mg)rSi"O,,(OH)r.Zeitschrift fiir Kristallographie, structureofhigh cummingtonite. Journal ofGeophysical Research,77,
133.301-326.
5767-5777.
- (1973)The high temperaturecrystalchemistryof tremolite. Amer-

ican Mineralogist, 58, 649-664.
ZhangL, and Hafner, S.S.(1992a)Gamma resonanceof 57Fein grunerite

at high pressuresA. merican Mineralogist, 77,474-479.

-(1992b) High-pressure57Fe7 resonanceand compressibility of

Ca(Fe,Mg)Si,Ouclinopyroxenes.American Mineralogist, 77, 462-47 3.

Zhang, L., Turk P -G., Ahsbahs,H., and Hafner, S.S.(1989a)"Fe hy-
perfine and volume compressibilitiesofchain silicatesbetween2 and

I I GPa. Twenty-eighthIGC Abstractswith Progams, 2, 296.

Zhang,L., Ahsbahs,H., Turk, P.-G., and Hafner, S.S.(1989b)A pressure

induced phasetransition in p),roxene.High PressureResearch,5, 732-

734.

MnNuscnrvr REcETvEADpnn 8, l99l
Mem;scnrp.r AccEprEDDrcpr.,rssn3. l99l