American Mineralogist, Volume 66, pages 596-600, I98l

Quartz with rhombohedralcleavage from Madagascar

Orro W. FLORKE,HnrNz G. Mmrre, JtlncnN Wnrcnr,nt Institutfiir Mineralogie,Ruhr-Universitrit Bochum, Postfach, D-4630 Bochum, Germany

AND HOLGER KULKE DEMINEX, DorotheenstraBe1, D-4300Essen, Germany

Abstrsct

Anhedral quartz slabsfrom a vein depositin Madagascarshowing smooth cleavageplanes parallel to positive rhombohedra r {l0ll) are described.Only two of the three equivalent pairs ofthe form {l0Tl} show cleavage,although ofdiflerent quality (smoothness).The best developedplanes lie parallel to Brazil-twin lamellae;the lesswell developedplanes do not. We presumesmall tectonically induced thermal shocksto be the releasingmechanism for rhombohedral cleavagefractures. In combination with hydrolytic weakening and directed stress,large smoothcleavage planes were formed. Sincethe r-facesof quartz are F-faces,they have a minimum of free surfaceenergy. We assumethis to be the reasonfor cleavageparallel to positive rhombohedralfaces.

Introduction surfaceswith pearly luster over areasof l5 to 20 cm'z. Faint, unevenlyspaced, parallel striationsindicated a Many authors describe qtrartz that, beyond its secondand a third direction of parting which were well-known conchoidalfracturing, under specialcon- not as well developedas the first. Thin sectionscut ditions shows cleavage of decreasing quality normal to the principal parting plane showedparallel (smoothnessor perfection)on r {l0Tl}, z {0lTl}, lamellae of varying width. The alternating arrange- and m {10T0}, respectively.In most caseshowever ment of these lamellae resembled albite twinning. no distinction has been made betweenr and z. The Therefore the authors concluded that this peculiar cleavageparallel to ru usually is not well developed. parting might be explained as a pseudomorph of Rose (1844) mentioned that cleavage of quartz single-crystalquartz after triclinic feldspar,though it may occur to an equal extent parallel to the positive was admitted that pseudomorphic replacement by and negativerhombohedra. The l9th century litera- single crystalsinstead of fine-grainedpolycrystalline ture on quartz cleavage was reviewed by Hintze material is very unusual. Jahns (1953) stated that (1915), and the literature up to 1960 by Frondel there existsno evidencethat large crystalsin pegma- (1962). All authors reported that pronounced platy tites grow by replacementof other minerals. cleavage or parting occasionally occurs in vein The production of cleavagein quartz through ther- qtarlz, mostly in the inner parts of vein fillings. mal shockby Kenngott (18M49) and under directed Rogers (1934) mentioned huge anhedral blocks of stressby Judd (1892) was mentioned by Hintze vein quartz from California pegmatites.Since parting (1915).Borg and Maxwell (1956)subjected qluartz wasnot observedparallel to all three equivalentpairs sand to a directed stressof about 2 kbar at 300'C. of rhombohedral faces, Rogers rejected the term under a confining pressureof about I kbar trans- cleavageand concludedthat parting might be due to mitted by alkaline hydrous solutions.They observed polysyntheticDauphin6 1ryinning, mechanicallypro- a remarkableincrease of fracturing parallel to r and duced during the high-low transformation of the z, but none parallel to lrt. The percentageof grains qvafiz. with deformation lamellaedecreased simultaneously. Murdoch and Webb (1938) describedquartz with The fractures occurred rarely in parallel sets and oc- nearly perfect platy parting in pegmatitic veins from casionallyconformed to more than one crystallogra- another California localitv. The slabs showed flat phic plane. Bloss (1957)crushed quartz crystalsboth 0003-004x / 8| /0506-0s96$02.00 596 FLORKE ET AL.: QUARTZ WITH RHOMBOHEDRAL CLEAVAGE at room temperature and between 650 and 700"C. tions perpendicularto the two recognizablecleavage He obtained a pronounced cleavageparallel to the planeswere measured to be 87!2". rhombohedra r and z and to a much lesser extent A systemofshort and lesslinsa1, healed fissures M parallel to the prism m. also exists (Fig. 2). Angles of intersection between theseflssures and R, or R,, are about 130to l40o in Geologicalsetting and mineralogy the same sectionsas above. The M-fissures.which One of us (HK) collected anhedral fragments of have healed after fracturing, representcharacteristic large single crystalswith almost perfect platy cleav- accumulations of aligned fluid inclusions. The R- age from qvartz veins in NE Madagascar.The hills, fracturesare unhealed. bordering D'Antongil Baie (Fig. l) are part of an old Universal-stagemeasurements proved the R-frac- cratonic block of Gondwana. Granites, migmatites tures to be the tracesof rhombohedralfaces and the and gneisseswere cut by a great number of quartz M-fissures to be atgned almost parallel to prism veins and basalticdikes during the Late Cretaceous. faces. At that time the separation India of from Madagas- Experimentalresults car causedintense shearing and mylonitization of the basementrocks (Hottin, 1969). X-ray measurements,according to the orientation Deep lateritic weatheringobscures the attitude and method of Keppler (1975),proved that the observed thicknessof the quartz veins. Furthermore, frequent cleavageplanes are rhombohedralfaces. It is difficult landslides have produced accumulations of quartz to distinguish between r- and z-rhombohedrausing blocks at the baseof the slopescontaining the origi- X-ray methods.Although the intensitiesof the reflec- nal outcrop. tions l0Tl and 0lTl are different (International Ta- Our samplescome from quartz veins of about 2 m blesfor X-ray Crystalography,1969, Vol. I, p. 465) thickness, presumably crystallized from hydro- their exact determination is nearly impossible, be- thermal solutions at the expenseof the quartz con- cause both reflections are strongly affected by ex- tent ofthe shearedhost rocks.The quartz variesfrom tinction effects.Literature data (International Tables, milky or slightly smoky to absolutelytransparent in 1962,Yol. III, p. 79; Le Page et al., 1976)are con- various parts of the vein. Cleavageoccurs especially tradictory on the question of which reflection is the in the inner parts of the vein fi[ing. The remainder more intense. exhibits conchoidal fracturing only. The best devel- An unequivocal experimental indication of oped cleavageswere observedin blocks of approxi- whether the rhombohedralcleavage planes are posi- mately I m' volume composedof alnost clear quartz tive r- or negative z-planes (Bravais-setof crystal- with only minor trains of inclusions and cloudy lographic axes,see Dana Frondel, 1962)can be ob- schlieren(Fig. 2). tained by hydrothermal growth. Prisms measuring 5 x 5 x 15rrm were cut from platy fragmentswith two natural cleavageplanes forming two faces of these Sampledescription prisms. These served as seeds.Growth conditions Thin sections of various thicknessesproved the were: seedtemperature 380oC, nutrient temperature samplesto be single crystals; however, the micro- 400"C, pressure1.5 kbar, duration 14 days;nutrient: structure is characterized by twin lamellae following broken quartz of the same sample, solvent: I m the Brazil law. Neighboring lamellae show differ- NarCO, in H, O. After the runs the seedcrystals were encesin extinction position of 2o. The lamellaethem- terminatedonly by real facesm, r and z, with r prom- selves extinguish homogeneously.Specimens with inently developed parallel to the original cleavage macroscopicallyplanar cleavagesshow systemsof planes of the seed.The facesm and r were morpho- long linear fractures parallel to the best developed logically dominant, whereasz-faces were subordinate cleavageplanes. We refer to this fracture system as due to their higher growth rate. The product crystals R,. The twin lamellae boundaries are frequently exhibited an unusual pseudoorthorhombic-prismatic identical with the R,-fractures.On the universalstage habit becauseof the prominent developmentof the a secondsystem of linear fractures,here termed R,r, rhombohedral faces. In agreement with literature running parallel to a secondpair of rhombohedral values,the measuredvelocities of growth VG were: faces r, could be measured.These fractures are sig- VG- < VG, << YG" (cf. Jost, 1955).Our growth ex- nificantly shorter than the R, type. The anglesof in- perimentsat different temperatures(pH ll, pressure tersectionbetween fracture systemsR, and R,, in sec- 1.5to 3 kbar) show that in the temperaturerange up FLORKE ET AL.: QUARTZ WITH RHOMBOHEDRAL CLEAVAGE

D

E

gronrtes

1O81m vetn t wtthc ,

@

o

o O lO 2okm t-t-t

Fig. 1. Sketch of deposit, to 300oCthe free surfaceenergy difference between r that rhombohedral face r which representsthe best and z is so large that z showsmuch higher growth developedcleavage plane. rates than r. With increasingtemperature, however, Untwinned, single crystal slabs(3-5 cm diameter) the growth rates of r and z approacheach other and of transparent quartz fron diferent localities (Ar- at 500"C there is almostno differencebetween them. kansas,Goiaz, Madagascar)were slowly heated(1" / Twinning was analyzed by etching cuts per- min) in air (l atm) to l00oC and subsequently pendicularto the c-axis(45Vo }JF,20"C, 30 min), re- quenched in water to room temperature.After this vealing polysyntheticBrazil twinrningparallel to only procedure a few linear fractures parallel to the

'.1'K' r{1: 1, \ ''-"# Rrt if )*:.-o-', *, .tI

*,

Fig. 2. Photomicrograph of linear rhombohedral cleavage fracture systems R1 and R1y and of healed fissures M; thin section cut perpendicular to the two macroscopic cleavage systems. FLORKE ET AL.: QUART:Z WITH RHOMBOHEDRAL CLEAVAGE rhombohedra were observed.When the maximum tendency to cleave. Hartman's method, though ex- temperaturereached is increased,the fractures be- cellent,is a qualitative evaluationofsurface energies. camemore frequent but lesslinear. The reversepro- Especiallyin complicatedstructures with a complex cedure-shock heating and slow seeling to room bonding character,the results derived from it must temperature-resulted in the sameobservations. be viewed with caution. Completely different fractures were produced by The review of literature and our own experiments shock heating to temperaturesabout the low-high- show that experimentally produced cleavage in transformation of quartz. In this case the slabs qtartz, either by directed stressor by thermal shock cracked into a fabric of hexagonal, contiguous without a hydrothermal confining pressure,is poor prisms, elongatedparallel to the c-axis. Even under compared to the observednatural cleavage.There- hydrothermal conditions, as well as comparatively fore we assume,without experimental evidence at low ratesof heating and cooling the quartz shattered present, that the coupled action of a directional during the low-high-transformation. stress,with the resolvedshear stressRSS exceeding the cleavagestress o" l0Tl, and a confining hydro- hydrolytic Discussionand conclusions thermal pressurehigh enough to initiate weakening (cf. Griggs, 1967) are necessaryto pro- The quartz cleavageparallel to the positive rhom- duce cleavagein quartz. bohedron is in all casesobserved by us to be of dif- We presume small tectonically induced thermal ferent quality on the three equivalent pairs of the shocks(cf. Gay et al.,1978) under hydrothermal con- form {l0Tl}. One pair always showsthe best devel- ditions in comparatively small volumes of the vein- opment of cleavageplanes and fractures(R). A sec- bearing rocks to be the releasing mechanism for ond pair is of poorer cleavagequality (R,,). The third cleavagein giant quartz crystals of suitable purity possiblesystem Rr' was never observedin our sam- and suitable orientation relative to the stressdirec- ples. At standardtemperature and pressure(25"C, I tions. From our experimental results we assumea atm), and in the absenceof water, suitably oriented rise of temperatureon the order of about l00oC to be stressescaused samples with R-fractures to dis- necessary.We considerthe temperaturenot to have integrate into thin flakes parallel to these fracture exceeded300"C. Due to the decreasingsurface en- planes,but they do not show any tendencyto form ergy differencesof r and z above300'C the tendency new ones. for cleavageshould not show significant differences. Under the as yet unknown natural conditions,the At higher temperatures,and especially above the huge vein quartz crystalsunderwent an in situ cleav- low-high transformation temperature, we expect ing process without the formation of deformation from our experimentalresults that evenunder hydro- (Boehm) lamellae or deformation bands. The cleav- thermal conditions quafiz should shatterrather than agedescribed here must be correlatedwith the struc- cleave. ture of quartz.It requiresbreaking of the strong Si- The cleavagestrength o" of the positive rhombohe- O-bonds in the tetrahedra-framework.Fairbairn dral facesr at low temperatures(far below the low- (1939)calculated the number of Si-O-bonds per face high-transformation) is, from our undershlding of unit and showedr and z to have a minimum, corre- the quartz structure,less than it is for any other form spondingwith a minimum of coherenceof the corre- and especially for the negative rhombohedron z. sponding lattice planes. Though he did not quan- Thus the explosivetransformation of the storedstress titatively discriminatebetween r and z, he assumeda energyinto surfaceenergy serves to form large, even decreaseofcoherence in the sequence: smoothcleavage planes. Additional hydrolytic weak- ening moreover specificallylowers the ideal and real o : c: ffi, z, r. {1120}, {0001}, cleavageenergy for r by Si-O-bond breaking and ac- Fl6rke (1955,1959) described the positiverhombohe- tivation of dislocationsin the structural framework. dron as the only F-form, accordingto the PBC the- The proposedmechanism of cleavageaccounts for ory of Hartman and Perdoc (1955). In order to ex- both fracture and cleavagesystems R' and Rrr, but it plain its morphologicalprominence, it must have the does not explain the di-fferencein cleavagequatty lowest free surfaceenergy, 1.e.,a minillum of loose and the absenceof the symmetricallyequivalent sys- Si-O-bonds after cleavage.According to Hartman tem R,,,. We assumethis to be due to the poly- (1959), however, the z- and rz-faces should be F- synthetic Brazil twinning parallel to the best rhom- faces as well, but these do not show a comparable bohedralplane of cleavage.Brazil twinning along the 600 FLORKE ET AL.: QUARTZ WITH RHOMBOHEDRAL CLEAVAGE systemR,, doesnot occur. Twin boundariesare two- tobalit. Berichte der D€utschenKeramischen Gesellschaft, 32, dimensional areas of structural weakness.This 369-381. causescleavage to be more easily achievedand ex- Fl6rke, O. W. (1959) Regelungserscheinungenbei der para- von SiO2-Kristallen. Zeitschrift fiir plains morphen Umwandlung the better cleavage of R, compared to R,,. Kristallographie, ll2, 126-135. Therefore we assumethe polysynthetic Brazil-twin- Frondel, C. (1945)Secondary dauphind twinning in quartz. Amer- ning to have been formed before cleavage. Poly- ican Mineralogist,30, 47-460. synthetic Brazil-twinning has been frequently ob- Frondel, C. (1962)The systemof mineralogy.7. Ed. Vol. IIL Sil- servedin vein quartz (Frondel, 1962,p.87, Figs.62, ica minerals,Wiley, New York, p. 104. Gay, N. C., Comins,N. R. and Simpson,C. (1978)The composi- 63). tion of spherules and shatter cone surfaces from the Vredefort The absenceof the third cleavagesystem R,,, is structure,South Africa. Earth and PlanetaryScience Letters, 41, due to the orientation of the qvartz with regardto the 372-380. directed stress.According to Schmid'slaw the shear Griggs, D. (1967)Hydrolytic weakeningof quanz and other sili- stressr reachesa maximum at r-o* : o/2, if the angle cates.Geophysical Journal of Royal Astronomical Society, 14, l9-3l. betweenstress o and gliding plane happensto be 45o. Hartman, P. (1959)La morphologiestructurale du quartz. Bulletin The interfacial angle between two symmetrically Soci€t€ Frangaisede Mineralogie et de Cristallographie, 82, equivalent rhombohedron faces r is 87". If the di- 33s-340. rected stresso. bisectsthe interfacial angle between Hartman, P. and Perdok, W. G. (1955)On the relations between two equivalent rhombohedron faces r, and r, the structureand morphology of crystals,L Acta Crystallographica, 8.49-s2. shear stressr nearly reachesa maximum for two Hintze, C. (1915) Handbuch der Mineralogie, 1. Band, 2. Abt., planes at the same time. The analogousorientation Veith & Co., Leipzig. with regard to three equivalent rhombohedral faces Hoffer, A. (1961) Low quartz, on the geometry of its structure rr, r, and rr, is, accordingto Schmid'slaw, difficult to framework in terms of the directed bond. Zeitschrift ftir Kris- I 16,83-100. realize.For only ong single orientation will the di- tallographie, Hottin, G. (1969)Les terrainscristalline du Centreet du Nord-Est rectedstress be inclined at 45" to all three faces.The de Madagascar.Documentation du Bureau des G6ologiques, latter caseis lessprobable. Thus usually two rhom- No. 178,Servais g6ologiques, Tananarive. bohedral cleavagesystems are formed. InternationalTables for X-ray Crystallography,Vol. I (1962)Ky- Our understandingof rupture suggeststhat the dis- noch Press,Birmingham, England. Tables for X-ray Crystallography,Vol. III (1969) tinction betweenfracture, parting and cleavagesug- International Kynoch Press,Birmingham, Engiand. gestedby Fairbairn (1939) cannot be upheld. Even Jahns,R. H. (1935)The genesisof pegmatites.I. Occurrenceand the conchoidalfracture of quartz is controlled by its origin of giant crystals.American Mineralogist, 38, 563-598. structure (Hoffer, 196l). The term parting should be Jost, J. M. (1955) lst Quaterly ProgressReport, Army Signal preserved for fractures controlled by segregation, Corps, Contract DA-3G039-sc-64689. M., Neifeind, A. and Rohner, C. (1975)X-ray zoning of impurities or exsolution (Buckley, 1934; Keppler, U., Meier, crystal orientation determination for semiconductor single crys- White, 1979),whereas cleavage should be usedfor all tals. Kristall und Techrrik, 10,'19-84. planar fracturesof crystalswhich are intrinsically or Murdoch, J. and Webb, R. W. (1938) Notes on some minerals primarily controlled by the structure. from southernCalifornia. American Mineralogist, 23, 349-355, p.354. Acknowledgments Nicolas,A. and Poirier, J. P. (1976)Crystalline plasticity and solid stateflow in metamorphicrocks. Wiley, New York, p. 200. We gratefully acknowledge the thorough revision of the manu- Le Page, Y. and Donnay, G. (1976) Refinement of the crystal script by Walter V. Maresch Ph.D., and we would like to express structureof low quartz. Acta Crystallographica,32,2456-59. our most sincere thanks to Mrs. Helene Nowara and Mr. Dieter Rogers,A. F. (1934) Unique occurrenceof vein quartz in Mari- Dettmar for their valuable technical assistance. posa Country, California. Proceedingsof the Geological Society of America, 1934,327 -328. References Rose,G. (18,t4)Uber das Kristallisationssystemdes Quarzes.Ab- Bloss, F. D. (1957) Anisotropy of fracture in quartz. American handlungen der Akademischen Wissenschaften Berlin, Phys- Journal of Science,255,214-225. ikalisch MathematischeKlasse, 25.4. l 8214,217 -27 4, Tafel I-VI. Borg, I. Y. and Maxwell, J. C. (1956)Interpretation of fabrics of Strens,R. G. J. (1976)The physicsand chemistryof minerals and experimentally deformed sands. American Journal of Science, rocks. J. Wiley, London. 254, 7 r-81 . White, J. S. (1979)Boehmite exsolutionin corundum. American Buckley, H. E. (1934) On a cleavageinduced by impurity. Mineralogist,64, 1300-1302. Zeitschrift fUr Kristallographie(A), 88, 122-12'1. Fairbairn, H. W. (1939)Correlation of quartz deformationwith its crystal structure.American Mineralogist, 24, 351-368. Manuscript received, fune 16, 1980; Fl6rke, O. W. (1955) Strukturanomalienbei Tridymit und Cris- accepted for publication, January 5, I98I