Structural Disparities Between Chalcedony and Macrocrystalline

American Mineralogist, Volume 79, pages 452-460, 1994

Structural disparities betweenchalcedony and macrocrystallinequ^rtz

Pprnn J. HBaNrv Department of Geological and GeophysicalSciences, Princeton University, Princeton, New Jersey08544, U.S.A. D.c,vro R. Vrnr,rN Department of Earth and Planetary Sciences,The Johns Hopkins University, Baltimore, Maryland 21218, U.S.A.

Departmentof Mineralsciences, -""1il1T;*r;"r:;rr,ion, washington,DC 20560,U.s.A.

Arsrnlcr Examination of length-fastchalcedony by transmission electron microscopy reveals significant structural differencesbetween microcrystalline fibrous quartz and ideal a qvarlz' Individual crystalsthat make up chalcedonyfibers are elongatealong I I0] and are twisted about the fiber axis. Streaksin electron diffraction pattems of single fibers indicate per- vasivedisorder among the {l0l} and possiblythe {011} planes,and in dark-fieldimages the texture ofthis structural disorder appearsdendritic. Becausethe streak systemsasso- ciated with { 101} and with {01 I } intersectat the expectedlocations of primary diffractions associatedwith the mineral "moganite", demonstrations of the presenceof "moganite" require complementary characteization techniques,such as X-ray powder diffraction. In addition, genuine superperiodicitiesof three, four, and five times the primary translations in quartz are discernedthrough electron diffraction ofchalcedony crystals,and violations of systematic absencesassociated with the 3, screw symmetry are apparent in X-ray diffraction experiments.

IttttooucrroN erties of chalcedony and bulk quartz regainedthe atten- .rytjullographers' Experiments by Frondel (1982, That signilicanr structural differencesdistinguish chal- li:l:f and Fldrke and his associates(Graetsch et al' 1985; cedonyfrom bulk quarrzhas beenknown f". *;ii;;;;" f.85) et al, l99l) have highlightedonce again the subtle century.The lower refractiveindex and ro*". i.nritv-o? ftilte microcrystallinefibrous quartz and librous quartz led Fuchs (1834) and others ,.;;;";; distinctionsbetween quartz' In this paper,we expand upon that chalcedonyis an intimate mixture of crystatt'ine'anO macrocrystalline studies,using transmission electron micros- amorphous qvartz. Michel-L6vy and Munier-dil; these earlier X-ray powder diffraction techniques' (1892)noted that, in contrastto prismaticq"uar, frb;o;, topy and quartz frequently is elongatein a direction that is normal ExpnnrvrnrvrAl PROCEDURES to the optic axis and displays biaxial behavior, wirl. 2v ranging up to 30'. Furthermore, chalcedony fibers sys- Experimentswere performed on clear seam-fillingchal- tematicallytwistaboutthefiberaxis(Lacroix, 1900;Ber- cedony nodules and on agatesthat were colorlessand nauer,l92l; Frondel, 1978),and chalcedonyismore sol- translucent.Specimens examined are listed in Table l. uble in HrO than is quartz (Pelto, 1956; Fournier and Despite the wide geographicdistribution of the agates Rowe. 1966). examined,all samplesexhibited similar microstructures Nevertheless,early X-ray diffraction studies of chal- and diffraction characteristics. For TEM examination, cedony revealed the presence of only pure a quartz specimenswere sectionedboth parallel and perpendicular (Washburn and Navias, 1922), and the differencesbe- to the radial quartz fibers at a thickness of -30 rrm and tween fibrous and prismaticqtafiz soon were overshad- thinned further by Ar ion milling. Thin foils then were owed by their apparent structural identity. The discovery coated lightly with amorphous C. Electron microscopy by SEM of multiple bubble cavitiesinterspersed among was performedwith a Philips 420 microscopeequipped the fibers (Folk and Weaver, 1952)further addedto the with T or ST objectivelenses and operatedat 120 keV. conviction that chalcedonicquartz differs from drusy va- X-ray analysesin the TEM were obtained with an EDAX rieties only in terms of grain size; the presenceof molec- energy-dispersivespectrometer (EDS) and a Princeton ular HrO in microporescould accountfor the lower den- Gamma-TechSystem IV analyzer. sity and refractive index, as well as the light scattering For X-ray powder diffraction experiments,samples were observedin chalcedony(Pelto, 1956). ground in porcelainmortars under acetoneand smeared Only recently have the differencesin the physical prop- onto low-background quartz plates or pressedinto pack 0003-004x/94l0506-0452$02.00 452 HEANEY ET AL.: CHALCEDONY AND QUARTZ DISPARITIES 453

Taele 1, Specimensused in this studv

Locality ldent.no.

Moiave Desert, CA 87415 U.S NationalMuseum Yellowstone Nat'l Park 83871 U.S NationalMuseum Brazil 117434 U.S NationalMuseum Mexico c6537 U.S. NationalMuseum Australia 131877 U.S NationalMuseum HillsboroughBay, FL 47899 U.S. NationalMuseum Locality unknown HAR C Harvard Mineralogical Museum mounts. The data were obtained on a Scintagcomputer- automated powder diffractometer with CuKa radiation and an intrinsic-Ge solid-state detector. Counting times rangedfrom 2 to l0 s per 0.03'step for the 2d range3-80". A pack mount of specimen 47899 was scannedfrom 12.2 to 12.3 20 (^: 1.15A) usingdoubly monochromatized synchrotron radiation at Beamline X7A, Brookhaven National Laboratory. Rrsur,rs Absenceof opal Despite past assertionsthat the quartz in chalcedonyis intergrownwith opal (Corrensand Nagelschmidt,1933; Donnay, 1936; Jones 1952),careful X-ray powder diffraction studiesof chalcedony(e.C., Novdk, 1947;Midg- ley, l95l) haverevealed little in the way of opalinephases. Graetschet al. (1985)reported fibrous opal-C within the thin rims surrounding agatenodules from the Rio Grande do Sul in Brazil, and powder neutron diffraction experiments of chalcedonyfrom the silicified corals of Hills- boroughBay, Florida, revealeda minor amorphoussilica component(Heaney and Post,unpublished manuscript). Nevertheless,in the courseof our TEM examination,we O.5mm observed no opaline silica residing either between or Fig. 1. Opticalmicrograph ofchalcedony. Fibers contained within chalcedony fibers. Rather, chalcedonyfibers were within the samebundle go to extinctionsimultaneously. Coarse uniformly crystallineand tightly packed. fibers(C) are distinguished from fine-grainedareas (F) by longer twist periodicity,higher birefringence, and absence offine band- Optical characterization fibers of lng. As observed with the petrographic microscope,all the fibrous quartz analyzed during these experiments was length-fast,indicating elongationnormal to c. Thin sec- visible as the specimenwas rotated 90'about the goni- tions of the variousspecimens displayed textures that are ometer axis. Such electron diffraction effects indicate a typical for chalcedony(see Frondel, 1962):sheafs offibers fairly random orientation of the crystallitesthat consti- emanated from a few points along the rims of the nod- tute the fibers,as is consistentwith observationsmade ules,suggesting initial precipitationat discretenucleation by Sunagawaand Ohta (1976). In these fine-grained centers.Optical examinationindicated that a strongori- regions,crystals appeared equant and measured<0. 1- entationalrelationship exists among the fibersthat make 0.5 pm across.Preferred orientation among these crystals up these sheafs. The fibers within the bundles went to wasdiscernible only whenthe diffractingarea was limited extinction simultaneously(Fig. l), and the extinction to < 5 pm in diameter.As was noted alsoby Miehe et al. bandsthat result from twisting ofthe optic c axis about (1984),grains that do sharea common crystallographic the fiber axis (Runzelbdnderung)were fairly continuous orientation appearto be aligned(Fig. a). Presumably,these acrossfiber bundle boundaries(Fig. 2). strings of crystallites make up the fine-grainedfibers that Despite the optical indications of crystallographiccon- are observedin the optical microscope. tinuity within thesesheafs, selected-area electron diffrac- In many chalcedonyspecimens, these zones of fine- tion (SAED) of regionswithin the fiber bundles produced grained quartz fibers alternatewith bands ofcoarse fibers. ring patterns (Fig. 3), especiallyin fine-grainedareas. All In someinstances, this bandingis highly periodic (Wang Debye-Scherrerrings were present, and they remained and Merino, 1990),but frequentlythe distancebetween 454 HEANEY ET AL.: CHALCEDOI{Y AND QUARTZ DISPARITIES

O.5mm Fig. 2. Extinctionbands in fibrousquartz are periodic along the fiberlength and roughly continuous across fiber boundaries. Fig. 3. A SAEDpattern from an areaoffine-grained fibrous This chevronextinction pattern is characteristicof chalcedony quartzabout 10 pm in diameter.Although some preferred ori- andresults from thecooperative rotation ofthe opticaxes about entationof the fibersin discernible,the overallringlike quality the fiber axesalong the lengthsof the fibers. ofthe patternsuggests that orientationsofthe quartzcrystals are poorlycorrelated within regionsof this size. coarse-grainedbands oscillatesrandomly. Optical examination of thick petrographicsections (-100 pm) reveals that these coarser-grainedareas have a distinctly variable.Frondel (1978) cited a rangefor the twist period higher birefringencethan the finer-grainedregions. The of 20-300 cm. fibers in the coarse-grainedzones can exceed 2 p.m rn Several explanations have been offered to explain fi- diameterand 5 pm in length,and the fibrousmorphology ber twisting in chalcedony.Graetsch et al. (1987) as- is evident even in bright-fieldimages (Fig. 5). serted that rotation about the fiber axis arises from lattice misfit betweenquartz-like modules and moganite- Fiber orientation and twisting like modulesin chalcedony.(Note: the name "moganite" Numerous X-ray diffraction studies have demonstrat- has not been acceptedby the IMA CNMMN.) Wang and ed that chalcedony fibers are elongate primarily along Merino (1990)proposed that tracesubstitutions of Al for I I 0] and more rarely along [2 l0] (seeFrondel, 1962).ln Si distort the lattice and induce a helical habit. By anal- addition, a SAED pattern superimposedupon the image ogy with the growth mechanisms operative in some of a coarsechalcedony fiber in Figure5 confirmsthe [ 10] whisker crystals,Frondel (1978) attributed chalcedony orientation of thesecrystals. The growth of chalcedony twisting to screwdislocations in which the Burgersvec- fibers normal to the c axis differs from the behavior of tors are parallel to the direction of elongation. More prismatic quartz, which typically is elongateparallel to c. specifically,Heaney ( I 993) observedthat a spiral growth Also in contrast to bulk qtarlz, chalcedonyfibers usu- mechanismactivated by screwdislocations with Burgers ally are twisted about the [110] fiber axis, giving rise to vectors equal to n/2fll}l accounts not only for fiber Runzelbdnderung,as mentioned above. Our TEM ex- twisting but also for the pervasiveplanar disorder that amination of chalcedonyrevealed that this twisting oc- characterizeschalcedony. curs within individual crystals that constitute the fibers; within fibrous quaftz it does not appear to arise from rotational misalignment Planar disorder of multiple singlecrystals along the fiber direction.In one Intersecting streak systems.SAED patterns of chalce- experiment,a traversewas performedalong the lengthof dony commonly reveal complex streak systems. These a singlecrystal within a coarsefiber whoseaxis of elon- streaks are particularly evident in diffraction patterns with gationwas parallelto the goniometeraxis. Consequently, zoneaxes parallel to (l-21) (Figs.6 andT) and (TIl ) (Fig. the pitch of the fiber twisting could be ascertainedby 8). ln their TEM study of Brazilian agates,Miehe et al. monitoring the relative orientationsof major zone axes (1984)noted that thesestreaks occur alongthe (101)* along the length of the crystal.This analysisplaced the directions, indicating extensiveplanar disorder parallel helicoid periodicity at about 550 pm per twist, though to {l0l}. On the basisof the orientationof thesecom- from optical observationthis periodicity clearly is quite position surfaces,they postulatedthat the disorderarises HEANEY ET AL.: CHALCEDOI.W AND QUARTZ DISPARITIES 455

Fig. 4. A dark-fieldimage (g : 100)of fine-grainedagatoid quartzreveals that crystalswith similarorientations form a lin- eararrang€ment. These crystals are separated by low-anglegrain boundaries,and presumablythey makeup the fibersobserved with the opticalmicroscope. Fig. 5. The chalcedony fibers in this bright-field image of coarse-grainedchalcedony are oriented vertically. Superposition from random interstratificationof Brazil twin lamellae. ofa SAED pattern taken from the white circular region onto the In addition, Miehe et al. (1984) attributed apparentsu- image of the fibers confirms that the direction of elongation is perlatticediffractions at multiples of %(012)* (asseen in [1 10]*, which is parallelto [1 l0]. Fig. 6) to intersectionswithin the (101)* streaksystem. The three individuals that make up this streak system with strict periodicity, thereby giving rise to moganite. (Fig.9) arethe vectors (l0l)*. ([l l)*. and (0lT)*.The However, the bulk of a typical chalcedony fiber consists intersectionsof (0ll)* with the reciprocalspace plane of a highly aperiodic alternation of twin lamellae, pro- definedby the vectors(101)* and (lll)* occurat mul- ducing highly concentrated planar defects. These defect tiples of t/t(0121*.Therefore, the extra spotsseen in Fig- ure 6 alongthe direction (012)* may representthe out of plane (01T)* streak.Similarly, we commonly observed apparentsuperlattice diffractions at multiplesof %(01I )* (Fig. 8A), and theseextra spotsmay arisefrom the inter- sectionofthe (011)* streak,with the reciprocalspace planedefined by (l0l)* and (l l0)*. Presenceof moganite. The SAED patterns also reveal intensity maxima at h + Vz,k, I + t/z(Fig. 7). Miehe et al. (1984) attributed these superlatticediffractions to regionsin which the left- and right-handedtwin lamellae alternate uniformly at the scale of the unit cell, thereby doubling the primary translationsalong c and a,. The rigorous alternation along { I 0 I } of slicesof left- and right- handed quartz createsthe distinct silica polymorph called moganite,which was first observedas seamsin ignim- brite flows in Gran Canaria (Fl6rke er al., 1976, 1984). Electron diffraction patterns of moganite exhibit no streakingalong the (101)* directions,and the h + t/2,k, I + Vzreflections are quite sharp. Miehe and Graetsch (1992) offereda structuresolution for moganitewith space group symmetry12/a, based on X-ray powderdiffraction experlments. Graetschet al. (1987)interpreted chalcedony as a kind of structural hybrid betweenquartz and moganite. With- Fig. 6. SAED pattern with zone axis (T2l ) reveals streaking in some regions,a chalcedonyfiber may contain quartz along ( I 0 I )*. Apparent superlatticereflections occur at positions lamellae that are entirely left- or right-handed,and in correspondingto multiples of V:(012)*(arrows), and they arise other areas,lamellae of oppositechirality may oscillate from intersectionswithin the single(101)* streaksystem. 456 HEANEY ET AL.: CHALCEDONY AND QUARTZ DISPARITIES

-tL ITrr]* or tlTll*

Fig.1. A SAEDpattern with zoneaxis (Dl ) exhibitsstreaks alongsymmetrically equivalent [01]* and [Tl l]* directionsor [0ll]* and uTll directions.Distinguishing between these two possibilitiesrequires knowledge of rhe kinematicintensities of thedifraction spots.Additional moganitelike superlattice spots areshown with arrows.

structures are not restricted to fibrous microcrystalline silica; Wenk et al. (1988) describedsimilar featuresin sedimentarychert and fine-grainedquartz from the slick- ensidesurface ofa granite. Ambiguities in moganite identification. Our SAED studies (Figs. 6, 7, 8) of both agatoid and seam-filling chalcedonyclearly revealedthe existenceof the planar disorder observedby Miehe et al. (1984). In dark-field imagesofthe coarsefibers, these defects appear as featherlike or dendritic textures:traces of the planar defects Fig.8. SAEDpatterns with zoneaxis ( I I I ). (A) Faintstreak- run parallel to the fiber direction along the fiber cores, ing is observedalong (101)+, and apparentsuperlattice reflec- and additionaldefect traces branch out from thesecentral tions(arrows) occur at positionsthat correspondto multiplesof cores (Fig. 10). The distancebetween the tracesof the Yz(011)*.(B) Streaks are visible along both the (101)* and the planardefects is somewhatvariable, but the spacingstyp- (01I )* directions. ically rangefrom 50 to 100 A. lltre featherliketexture of thesechalcedony fibers strikingly resemblesthe chevron quartz lattice is higher than that of the quartz structure. pattern created by the intergrown laths of nearly pure As a result, diffraction patterns with zone axes that are moganite (Heaney and Post, 1992), perhapsindicating related by a sixfold rotation around c appeargeometri- that the crystallographicconstraints that dictate the habit cally identical,but the intensitiesofcorresponding spots of end-membermoganite also are operativeduring the differ. Diffraction patternsofquartz from zonesother than growth ofchalcedony. [001] or [uu0] cannot be indexed unambiguouslyin the Although we concur with Graetschet al. (1987) that absenceof intensity data, and intensitiesfrom electron most chalcedonycontains a disorderedmixture of quartz diffractionare unreliablebecause of dynamicalscattering and moganite,our investigationssuggest that electrondif- processes. fraction cannot revealwithout ambiguity the presenceof Consequently,determination of the streaksystems ob- moganitein associationwith planar disorder.The space servedin the SAED patternsof chalcedonyis problem- group for otquarlz is P3 ,21 (or P3r2l). Although the quartz atic. Although the streaksobserved in our diffractionpat- structure has only threefold rotational symmetry along terns could be indexedto the (l0l)* direction,as was the c axis, the lattice for the trigonal crystal systemcon- done by Miehe et al. (198a),those streaks could as easily tains a sixfold symmetryoperator along c. Thus, as with be indexedto the (01l)* direction(Fig. 7). This ambi- all trigonal crystals,the point group symmetry of the guity in reciprocalspace translates to an uncerlainty in HEANEY ET AL.: CHALCEDOI{Y AND QUARTZ DISPARITIES 457

ai i; <101>*Streak System <011>*Streak System

Fig. 9. A schematicdiagram of the symmetricallydistinct ( I 0 1)* (left) and (0I 1)* (right) streaksystems in fibrousquartz. Superpositionofthese streak systems leads to intersectionsat I + t/2,k, I + t/z;h. k + t/2,I + t/z;and h + v2,k + t/2,I + vz. real space regarding the planes that actually are disor- dered.The defectsmay occur parallel to { l 0 I } or to {01 I } or to both setsof planes. This distinction is significant.In Figure 9, the (l0l)* and the (01I )* streaksystems are mappedout in separate reciprocal-spaceunit cells.Ifplanar defectsoccur among both the {l0l} and the {0ll} planes,these two streak (101)* systemsare excited simultaneously.When the Fig. 10. Dark-fieldimage (g : l0l) of agatefibers reveals streak system is superimposedupon the (0ll)+ streak the pervasiveplanar defects that give the crystalsa featherlike system, the streaksintersect at several points: h + V2,k, appearance. I + V2.,h, k + Vz,I + t/z;and h + t/2,k, + 1/2,I + t/2.The intersections of these streaks appear in electron diffraction patterns as points of heightenedintensity. Conse- of ttneh + t/2,k, / + Yzspots typically are elongate and quently, they may mask genuine superlattice spots, or somewhatdiffuse, in contrastto the sharp intensitiesas- they may be mistaken for genuine superlatticespots. Be- sociated with the true superlattice diffractions described causethe positions occupiedby the intersectionsof the below. Furthermore, Rietveld refinement of X-ray dif- (101)* and the (0ll)* streaksystems are exactlythose fraction patterns produced by most ofthe chalcedonyan- occupiedby the superlatticespots produced by moganite, alyzed in this study revealed relatively low concentra- the presenceof those diffraction maxima may not always tions of moganite,usually < l5 wto/0,as is typical for agates indicate the presenceof moganite. (Heaney and Post, 1992). Nevertheless,the h + Yz,k, I Mclaren and Pitkethly (1982) have correctlypointed + t/z diffraction maxima consistently appeared even in out that structural constraints favor an orientation for electron diffraction patterns from coarse chalcedony fi- Brazil twin boundaries along {l0l}; this composition bers, which contained much lessmoganite (< 5 wto/o)than plane requires the least amount of distortion from the the fine-grained regions. Consequently, we argue that ideal quartz structure at the interface. Nevertheless, identification of moganite in chalcedony cannot rely on structural continuity also can be maintained acrossBrazil TEM evidence alone and should be supplemented by composition planes oriented parallel to {0ll}. Indeed, X-ray powder diffraction techniques. Frondel (1962\ observedthat Brazil twin lamellae in am- in fibrous quartz ethyststypically occur parallel either to {101} or to { l0l} Superstructures and {0 I I }. Under the conditionsof rapid nonequilibrium In addition to the intensity maxima that result from crystallization that characterizethe growth of chalcedony the intersectionof( l0l)* and (01 I )* streaks,we detected fibers (Wang and Merino, 1990), the formation of Brazil faint but distinct diffraction spots along various recipro- composition planesparallel to the less favorable {011} cal lattice directions between the primary spots. These orientationmay not be uncommon. superlatticespots include t/t02l ,Vq14l, and % 032 (Fig. Our own experiments suggestthat in some casesthe I lA-l lC), and they correspondto multiples of the pri- apparent superlattice reflections at h + Vz,k, I + t/z and mary translations of a qluartz(e.g., three, four, and five related points can indeed arise from intersectionsofthe times). A doubled superperiodicity, as found in moganite (l0l)* and the (0ll)* streak systems.In Figure 88, (Miehe et al., 1984),may occur as well, but the inter- streaksalong both the (l0l)* and the (0ll)* directions secting streaksprevent a definitive determination. are clearlyobservable. In addition, the intensity maxima Severallines ofevidence suggestthat thesesuperlattice 458 HEANEY ET AL.: CHALCEDOI{Y AND QUARTZ DISPARITIES

Fig. 12. EnlargedX-ray powderdiffraction pattern ofchalcedony(top) and prismatic quartz (bottom). Peaks labeled M are producedby intergrownmoganite- The peakat 16.3.20 cone- spondsto the (001 ) d valuefor quartzand violaresthe threefold screwsymmelry. The top scaleis givenin engstdms,the bottom scalein 20(degrees).

spotsdid not arise from intersectingstreaks. The superlattice spotswere circular, sharp,and of lesserintensity than the streakintersections. No streaksystems other than (l0l)* and (011)* were discernedduring our electron diffractionwork, and it appearsthat thesewould not give rise to the extra diffraction spotsobserved. Lastly, electron diffraction patterns with the samezone axis but from different chalcedonycrystals exhibited different superperiodicities.Specifically, the SAED patternsin Figure l lA and 11B havethe samezone axis,[212],but they exhibit different superlatticemaxima. No superlatticediffraction peakswere evident in the X-ray powder diffraction patterns taken for thesesamples, suggesting that the inten- sitiesofthe extra spotsin electrondiffraction patterns are stronglyenhanced by dynamicaldiffraction.

Violation of threefold screw symmetry The 3, or 3, screwaxes that characterizemacrocrystal- line quartz requiresystematic absences among 00/ reflections for which / + 3n. These absencesare commonly violated in electrondiffraction patternsof prismatic quartz becauseof dynamical scatteringprocesses, but they are obeyed in single-crystal and powder X-ray diffraction patterns. However, powder X-ray diffraction patterns of chalcedonydid reveal weak intensitiesfor values of 2d that correspondto the prohibited 00 I reflection (Fig. I 2). This peak typically displayed intensities that were three to four times that of the background, and its full width at half-maximum (FWHM) was approximately0.2 20 with CuKa radiation. That this symmetry violation is not spuriousis sup- ported by severalobservations: (l) Although doo,of moganite correspondsto -5.4 A, samplesof nearly pure moganite from Gran Canaria did not reveal this peak, Fig. I l. Genuine superlatticediffractions in chalcedony.Ex- presumably becausethe structure factor associatedwith th tra spotsoccur at (A) 021; (B) VaI4l , and (C) 1.40J2. this reflectionin moganiteis very low. In addition, the HEANEY ET AL.: CHALCEDONY AND QUARTZ DISPARITIES 459 peak could be discernedin weatheredchert samplesthat, The fibrosityand planardisorder that distinguishchal- accordingto Rietveld refinement,contain no moganite. cedony from prismatic quartz may provide a window into (2) The KB peakfcrr the (100)planes ofquartz is expected differencesbetween the mechanismsby which thesetwo to occur in the vicinity of this reflection.However, a syn- silica speciescrystallize. This possibility is addressedin chrotron X-ray radiation scanover this areaalso revealed a separatepaper (Heaney, 1993). the peak, and the synchrotronradiation usedin this ex- perimentwas highly monochromatic.(3) Althoughit was AcxNowr-nncMENTs oflow intensity,this peak was observedin scoresofdif- Our thanks go to Pete Dunn of the U.S. National Museum, to Carl fraction patternsfrom fibrousand nonfibrousmicrocrys- Francis and Bill Metropolis of the Harvard Mineralogical Museum, to and to EugeneSmelik for providing specimensthat were talline silica specimensin the collectionof the U.S. Na- Clifford Frondel, used for this study Joe Pluth performed the synchrotron data collection tional Museum of Natural History. Helpful reviews were provided by Alain Baronnet, Dave Vaniman, and The consistentviolation ofthe threefold screwopera- an anonymousreviewer. This researchwas supportedby NSF grantsEAR- tion indicatesthat the spacegroup symmetry for chalce- 8609277,EAR-8903630, and EAR-9206031.Electron microscopy was which was estab- dony is not trigonal but either monoclinic or triclinic. performed in the Johns Hopkins HRTEM laboratory, lished with partial support from NSF grant EAR-8300365, and the Gen- Facility of the Princeton Materials In- DrscussroN eral Motors Electron Microbeam strtute. This study underscoresone primary conclusion:chalcedony is not simply fine-grainedqvartz. In one sense, Rrrnnrxcns crrED this knowledgeextends back to prehistorictimes, when Tregoning,L.L. (1944) Reactivity ofaggregateconstituents quartz in fa- Bean, L., and hunter-gatherersspurned coarsely crystalline in alkaline solutions Journal of the American Concrete Institute, 16' vor of microcrystallinecherts and chalcedoniesfor fash- t I-J l. ioning their tools; the denselyinterwoven chalcedony fi- Bernauer, F. (1927) IJber Zickzackbiinderung(Runzelbiinderung) und berslent a toughnessto this material that is not matched veruandte Polarisationserscheinungenan Kristallen und Kristallaggre- Mineralogie, Geologie, und Paleontologie by prismatic quartz. Differencesin the chemical proper- gaten Neues Jahrbuch fiir Beilageband,55. 92- | 43. ties ofchalcedony and quartz also have beenrecognized Correns,C.W., and Nagelschmidt,G. (1933) Uber Feserbauund optische by manufacturersof concrete;whereas quartz crystalscan Eigenschaftenvon Chalzedon.Z€itschrift fiir Kristallographie,85' I 99- serveas relatively inert componentsof a concretemix, zl5. chalcedonyis highly reactive and induces mechanical Donnay, J.D.H. (1936) La bir6fringence de forme dans la calc6doine Annales de la Soci6t6G6ologique de Belgique, 59,289-302. (Bean 1944). failure and Tregoning, Fl6rke, O.W., Jones,J.B, and Schmincke,H-U (1976)A new micro- Nevertheless,these differences and the optical discrep- crystalline silica from Gran Canaria Zeitschrift liir Kristallographie, ancies between chalcedony and quartz weire submerged 143,156-165. under the prodigious weight of X-ray povrdsl diffraction Fliirke, OW., Ktihler-Herbertz, B., Langer, K., and Tdnges, I. (1982) quartz of volcanic origin: Agates.Contribu- evidenceearly in this century.The nearly identicalpow- Water in microcrystalline tions to Mineralogy and Petrology, 80, 324-333. der patterns produced by quartz and chalcedony reveal Fl6rke, O.W., Fliirke, U., and Giese,U. (1984) Moganite: A new mrcro- much about their structural disparity: it must involve crystalline silica-mineral NeuesJahrbuch fiir Mineralogie Abhandlun- changesofsuch subtletythat they are apparentlynot de- gen, 149,325-336. (1991) tectableby conventionalX-ray diffraction techniques.One Fltirke.O.W.. Graetsch,H., Martin, B., Rtiller,K., and Wirth, R. of micro- and non-crystalline silica minerals, basedon is structural H, Nomenclature common culprit in such circumstances structureand microstructure.Neues Jahrbuch fiir Mineralogie Abhand- and much evidence suggeststhat H is incorporated into lungen,163, l9-42. the structure ofchalcedony. Florke et al. (1982) deter- Folk, R.L, and Weaver,C.E (1952)A study of the textureand compo- mined that agatescontain from I to 2 wt0/oHrO, which sition of chert. American Journal of Science,250, 498-5 10. (1966) Estimation of underground tem- is divided more or lessequally betweeninterstitial mo- Fournier. R.O.. and Rowe, J.J. peraturesfrom the silica content of water from hot springs and wet- lecularHrO and silanole-groupwater (SiOH). On the ba- steam wells. American Joumal of Science,264, 685-697. sis of infrared spectroscopystudies of chalcedony, Frondel, C. (1962) The systemofmineralogy, vol III, 334 p. Wiley, New Graetschet al. (1987)estimated that one-third of the si- York. - lanole is surficial and the remaining two-thirds is internal. ( 1978)Characters of quartz frbers.American Mineralogist, 63, l7- 27 We speculatethat the true superperiodicities and the -(1982) Structural hydroxyl in chalcedony(Type B quartz). Amer- violation of the threefold screw symmetry observed in ican Minerafogist, 67, 1248-1257. chalcedony arise from structural H. Simulated X-ray -(1985) Systematic compositional zoning in the quartz fibers of powder diffraction patterns for quartz in a triclinic setting agates.American Mineralogist, 70, 97 5-97 9. (l Zum Amorphismus festerKdrper. Annalen der Physik indicatethat the appearanceofa 00 I peakcan be induced Fuchs,J.H 834) -0.2 und Chemie.31. 577-583. by displacementof a singleO atom by A alongthe Graetsch,H., Fldrke,O.W., and Miehe,G. (1985)The natureof waterin z direction.Bonding H atomsmay be responsiblefor such chalcedonyand opal-C from Brazilian agategeodes. Physics and Chem- anionic displacements.Moreover, the superlatticedif- istry of Minerals,12, 300-306. -(1987) and fractions observedin SAED patterns but not in X-ray Srructural defects in microcrystalline silica. Physics Chemistryof Minerals,14, 249-257 patterns also may involve short-range structural distor- Heaney,P.J. (1993)A proposedmechanism for the growth ofchalcedony tions due to local ordering of H atoms. These specula- Contributions to Mineralogy and Petrology, ll5,66-74. tions await direct empirical validation. Heaney,P.J., and Post, J.E. (1992)The widespreaddistribution ofa novel 460 HEANEY ET AL.: CHALCEDO].IY AND QUARTZ DISPARITIES

silica polymorph in microcrystallinequartz varieties Science,255,44l- Nov6k, J. (1947) Les modifications fibreusesde la silice. Bulletin de la 443. Soci6t6frangaise de Min6ralogie, 70,288-299. Jones,F.T. (1952)Iris agate AmericanMineralogist, 37,578-587. Pelto,C.R. (1956)A study ofchalcedony.American Journal ofScience, Lacroix,A (1900)Suruneformedesiliceanhydreoptiquementn6gative. 254,32-50. ComptesRendusdel'Acad6miedesSciences,130,430-432. Sunagawa,I.,andOhta,E.(1976)Mechanismofformationofchalcedony. Mclaren, A.C., and Pitkethly, D.R. (1982) The twinning microstructure ScienceReports of the Tohoku University, 13, 13l-146. and $owth of amethyst quartz. Physicsand Chemistry of Minerals, 8, Wang, Y, and Merino, E. (1990) Self-organizationalorigin of agates: 128- 135. Banding,fiber twisting, composition, and dynamic crystallizationmod- Michel-L6vy, A., and Munier-Chalmas, C P E. (1892) M6moire sur di- el. Geochimica et Cosmochimica Acta, 54, 1627-1638. versesformes affect6espar le r6seau6l6mentaire du quartz. Bulletin de Washburn, E.W., and Navias, L. (1922) The relation ofchalcedony to the la Soci6t6frangaise de Min6ralogie, 15, I 59-190 other forms ofsilica. Proceedingsofthe National Academy ofSciences, Midgley, H.G. (1951) Chalcedonyand flint GeologicalMagazine, 88, 8, l-5. l'19-184 Wenk, H.-R, Shaffer,S.J., and Van Tendeloo,G (1988)Planar defects Miehe, G, and Graetsch,H. (1992) Crystal structure ofrnoganite: A new in low temperaturequartz. PhysicaStatus Solidi A, 107, 799-805. structure type for silica. EuropeanJoumal of Mineralogy, 4, 693-706. Miehe, G , Graetsch,H., and Fl6rke, O.W (1984) Crystal structure and gowthfabricoflength-fastchalcedony. PhysicsandChemistryofMin- MlNuscprm RECETVEDOcrosen 12,1992 erals,10, 197-199 MeNuscnrrr AccEprEDJeNueny 14,1994