American Mineralogist, Volume 63, pages 316-325, 1978

Biaxialitvin'isometric' and'dimetric' crvstals

EucrNn E. Foono'nNo BRropoRn A. Mrlls

Department of Geology, Stanford Uniuersity S tanfo rd, C al ifornia 94 305

Abstract

In an attempt to explain the anomalousoptical propertiesof certain cubic, tetragonal,and hexagonal(trigonal) species,variations in2V, composition,and space-groupsymme- try have been determinedfor eight different inctuding garnet (And-Gross), rutile, jeremejevite [A16850lu(F,OH)r],apatite, kleinite [Hg,N(Cl,SO,).xH,O], beryl, , and elbaite. The space-groupsymmetries of the specimens,determined by single-crystalX-ray techniques,showed no deviationsfrom previously-observedcrystallographic symmetry, yet significantdeviations from isometricand uniaxial optical characterwere observed.All speci- mensexcept rutile were found to be compositionallyzoned, but no systematiccorrelation was found betweenelemental zoning and 2/measured along the sametraverses The anomalous optical propertiesare thought to be due to strain induced by chemicalsubstitutions and/or defectsoccurring during crystal growth, by rapid temperatureor pressurequenches, or by mechanicaldeformation.

Introduction but jeremejevite,quartz, and have cen- Numerous observationsof anomalously biaxial trosymmetricstructures. These eight mineralswere cubic, tetragonal,and hexagonal(including trigonal) chosenbecause of the range of crystallographicsym- substanceshave been reported over the past half- metry, , and varying chemistry that century. Various hypotheses have been offered to they display. explain the biaxiality of these materials,including The proposedcauses of observedanomalous opti- internal stresses(Deer et al., 1966),twinning (Win- cal charactercan be consideredconceptually as being chelland Winchell,1951; Turner, 1975a, b), and me- primary or secondary in nature, depending on chanical deformation (Johannsen, 1918; Turner. whether they originated during or after crystal 1975a).These proposedcauses of biaxiality need not growth. A primary causefor an anomalousoptical affectthe space-groupsymmetry as determinedby X- axialangle, in an otherwiseisotropic or uniaxialmin- ray diffraction. Thus, cubic, tetragonal,or hexagonal eral, involvesthe generationof internal stressespro- mineralsdisplaying anomalous optics commonly dis- duced by a range of chemicalsubstitutions or by play cubic, tetragonal, or hexagonal space-group defectsoccurring during crystal growth. The compo- symmetry. sitionally-inducedstresses could be especiallylarge From the many mineralsexhibiting anomalous op- near contacts between zones of sharply contrasting tical properties,the following have been chosen for composition,and may resultin a2V in theseregions further study: (l) garnet (cubic); (2) rutile (tetrag- togetherwith related changesin refractiveindices. A onal); (3) jeremejevite, A16BbOlu(F,OH)s(hexa- possiblesecondary cause of anomalousbiaxiality in gonal); (4) apatite (hexagonal); (5) kleinite, compositionally-zonedcrystals involves rapid pres- HgrN(Cl,SO,).xH,O (hexagonal);(6) beryl (hexa- sure and temperaturechanges, subsequent to crystal- gonal), includingvarieties goshenite, morganite, and lization of the mineral considered,which produce emerald; (7) quartz (trigonal), variety amethyst;and internal stressesand anomalousoptical properties. (8) tourmaline(trigonal), variety schorl-elbaite.All The effectof internal stressesproduced by any of the above mechanismson the magnitudeof anomalous I Presentaddress: Stop 905, U.S. Geological Survey, Denver 2lns and refractive indices is at present unknown. Federal Center, Lakewood, Colorado 80225. The purposeof this study is to examinethe rangein

0003-004x/78/0304-03 I 6502.00 3 I 6 FOORD AND MILLS BIAXIALITY 317 anomalousoptical properties of theseeight minerals, Table L Minerals analyzedby electronmicroprobe; elements and most likely causedby chemicalzonation, primary standards growth features,or secondarypressure-temperature Analyzed elements (staodatds) effects. Tourmaline Fe, Mn, Ca, Ti, Si, A1, Mg, I, Na

Experimentaldetails Zn (pure netal) All opticalobservations were made with a monoc- Beryl ular Zeisspolarizing microscope utilizing a 3200'K emerald ar rn,r,,r:l.h,nhirc\

(tungsten)light source,an 8X ocular, and a 40X gosheni Le-morganite Na, Cs (natural beryls) (N.A. objective 0.85). A calibratedsingle-line mi- Utah red beryl Fo an.i"r.l L6n.rira\ crometerocular (6X) was used for measurementof Mn, Zn, Sn (pure metals) t1 (svnEneEacr1v-, isogyreseparation. For all mineralsexamined opti- -l cally, doubly-polished(3-micron polish) plates rang- Si, A1, K (natural otthocfase) Mg (synEhetic Mgo) ing from 0.03 to 1.00 mm thick were used.Single Na, Cs (natural beryls) crystalswere cut normal to the c axis (hexagonalor Jerenej ev iie F6 rnrf,'r.l ham.iii.l trigonal and tetragonal). The garnet (grossular- Si (synthetic SiO^) andradite) was examined in a doubly-polished A1 (syntheEic A1^0^) petrographicthin sectionof 0.03mm thickness.Mea- I (syntheric LII) surementsof 2V weremade by Tobi's method(Tobi, KfeiniEe al /cwnrhcri..:l.mFl\ r956). ilo q fn,i"7.1.inn.hrrl

X-ray precessionphotographs were obtained for ApatiEe Ca, Sl, P, Ce, Y, I (naEural apatite) tourmaline,beryl, jeremejevite, quartz, and kleinite. Mn, I'e (pure metals) For tourmaline,beryl, and quartzno deviationsfrom Garne t previously-reportedsymmetry were observed.The gros sular-andrad ite Ca, Si, A1, Fe, Mn, Ti, Mg spacegroups for jeremejeviteand kleinite are cur- (nacuraf garnets) Rutile rently being redetermined.No violations of hexa- I gonal symmetryin eitherwere observed. Quartz J An ARL-EMX-SM microprobe was used for quantitativeelement determinations. In all casesac- celeratingvoltage was l5 kV, samplecurrent was 0.04 Grande) has been examinedin most detail (Foord, to 0.06 microamperes,and the spot sizewas about 2 1976). microns.Analyzed elements and standardsfor each "Pocket" tourmaline crystals show concentric mineralare listedin Table l. Raw data werereduced growth zoning with or without additional sectorzon- using the MAGIC IV computer program (Colby, ing and twinning on {101l}. The basal portions 1968).Microprobe scanningtraverses were made at (roots) of thesecrystals are usually very homoge- points separatedby 0.01 to 0.5 mm with l0-second neous over a large region (mm scale) and are per- counting times.2Z determinationswere made along fectly uniaxial. Extinction positionsare sharp, and the sametraverses at intervalsof 0.I to 0.5 mm. All basal sectionsshow minimum birefringenceat all measurementswere taken at 23o + 2"C. positionsof rotation of the microscopestage. In eu- hedral "pocket" crystals,chemical zonation becomes Data significant,as doesvarying degrees of biaxiality.This biaxiality is generally most pronounced at color Tourmaline boundaries(pink-green, pink-colorless, or others) Of the eight mineralsexamined (Table 2), tourma- where the steepest chemical gradients occur. Al- line receivedby far the most attention. "Pocket" though there is an apparent correlation between2V tourmalinesfrom the Mesa Grande, Pala, and Ra- variationand elementalzonation within tourmaline, mona districtsin San Diego County, California; from there is no consistentrelationship betweenany one the Newry and Paris districts, Oxford County, elementand 2V variation (Figs. 1a and 2a). Secon- Maine; and from the state of Minas Gerais, Brazil, dary or later episodesof tourmaline growth (pencils) have been examined to evaluatethe relationshipbe- result in material which is also strongly composition- tween compositionalzonation and optical character. ally zoned.These crystals, although flawlessor nearly Tourmaline from the Himalaya dike system (Mesa so, show large variations in 2 Z and optic-figurequal- FOORD AND MILLS BIAXIALITY

Table 2. Compositions,space groups, and anomalousoptical propertiesof mineralsexamined

Anornalousoptical Mineral Conposition Space group n?^narfi ac

Garnet

ornqqrr'l ar- CarFerSi,O' andradite Ca3Al2Si -LA,5d , 2/ 3012 | ^-. Rutile Tio2 P4/nwn DlretTusence I c. zv Jeremejevite At6B5Ols(F'0H) 3 Ca, (Poo) (F, oH) , P6 '/n Kleinite HgrN(Cl , S0O)-cHrO Beryl goshenite - Be-A1^Si-0.^ J Z b 16 norganite enerald Utah red beryl :: Quartz amethyst sio2 P3.2 or PS^2 rl t4 Tounnaline schorl - Na(Fe, Mg) (OH,F) ,A1UBSSi602 7 4 RSn elbaite Na(Li ,Al (0H,F) ) 3Al6B3Si6o27 4

ity. However, strain shadowsand patchy extinction crystalsshowing strain shadows and mosaicas well as are often noticeable, especially near small cracks. lamellartextures show no observablerelationship be- Occasionally, the extinction pattern resemblesthe tween the azimuth of the optical axial plane and grid twinning seenin microcline.Crystals in which crystallographicsymmetry (Figs. lb, 2b). The lack of domains of differingcomposition are in sharpcontact correlation is probably related to the domain size with one another (e.g. dendritic growth) frequently being examined and the presenceof irregularly dis- show microcline-likegrid twinning. In portions of tributed internal stresses.Primary developmentof crystals(tourmaline as well as others) where the la- internal stressesand consequentdevelopment of pri- meflar pattern is present, 2V is clearly decreased. mary 2V may be due to growth imperfections and Source-imagedistortion techniques(Wagner et al., growth characteristicsof the crystals,e.g., screwdis- 197l) showed the existenceof domain structure in locations and growth hillocks (spirals). These fea- color-zoned from Brazil. Mosaic tex- tures are very commonly observedon the pinacoid tures are often noticeablein crystalsexamined from and pedion terminations of pocket crystals of all districts.Those "pocket" crystalswhich are obvi- tourmaline and beryl. Differing elasticproperties as a ously deformed(e.g. bent) often show strain shadows function of the composition in thermally shocked and strain birefringencealong with biaxial character. crystals at high pressures may produce internal Most crystalsare not visibly deformed but may still stressesand consequentlya secondary2V.This is true be biaxial. for all other gem-pocketpegmatite minerals, and may Azimuth orientation of the optical axial plane be true for minerals formed in other lower temper- within tourmaline has been observedto be generally ature and pressureenvironments. A second mecha- very irregular. Figures lb and 2b show orientation nism involving rapid drop in pressure(pressure diagramsof the azimuthsof the opticalaxial planes quench)may also producea secondary2V in a pre- for two crystalscut normal to c. Other crystalsexam- viously uniaxial crystal or even changethe value of ined in the same manner show essentiallythe same 2V in a biaxial crystal.It is most likely, however,in features.In somecrystals (possibly strain-free or with the case of tourmaline and probably other pocket minimum strain), the azimuth orientationsof the mineralsfrom pegmatites,that the2V seenis primary optical axial planes often are nearly constant with in origin. Annealing of strain produced by thermal relation to the crystallographic symmetry. Other shock or pressu[e changes during crystallization FOORD AND MILLS, BIAXIALITY

COLORLE SS DARK ELUE frodu16 filling

JIU^ .--..4

90 WEIGHT PERCENT70 OXIDE

35

30

i3 25 !3

?: 20

t5 :FEU to

o5 4ii ii i'

DISTANCE(mm ) A B Fig. la Weight percentoxide and 2V varialioncurves within a'pocket'crystalof schorl-elbaitefrom the TourmalineQueen Mine. Pala,California. MgO content decreasesfrom about 0.04weight percent in the core to lessthan 0.01weight percent in the rinr Sectioncut normal to c

would be expectedin pegmatitepockets if slow cool- axial planes for the same crystal and the traverse ing occurred subsequently;considerable evidence ex- along which probe data were obtained.No definite ists (Jahnsand Burnham, 1969)to support this hy- relationshipbetween Na contentor Cs content and pothesis. Cracking, also commonly observed in 2V appearsto exist. Viewed down c, gosheniteand pocket minerals,may also reducestress produced by morganitefrom San Diego County and Brazil (total external strain effects. of l0 specimens)all show undulose,anisotropic rims with growth zones,whereas the coresare much more Beryl uniform and nearly perfectlyisotropic. Loci of strain Beryl,varieties goshenite and morganitefrom the are defined by irregular areas of extinction which OceanView (ElizabethR) mine at Pala and the Little become evident upon rotation of the microscope Three mine at Ramona,shows strong compositional stage.On the other hand, the unusualberyl described zoning of the alkaliesand anomalousbiaxial charac- by Schaller et al. (1962) from Arizona contains the ter. Figure3a is a plot of 2 Z and Na * Cs contentus. highest-knowncontent of elementsother than Be, Si, distancefor a beryl from the Ocean View mine. Fig- Al, and O (16.87weight percent) bui is homogeneous ure 3b shows the azimuth orientationsof the ootical in compositionand displaysno biaxiality.From this 320 FOORD AND MILLS BIAXIALITY

FRACTURE FILLING (ochroile) A specimenof the unusualred beryl from the Wah Wah Mountains, Utah, was examinedin detail for compositionalzoning and optical properties.There is no apparent correlation between 2V and composi- tional zonation. The core displayspatchy areasshow- * ing strain centersand mosaic textures with high 2V *6 values(as much as l5o + 2"), while the rim displays J BLUE fairly sharpextinction, and only portionsare slightly biaxial. The core areamay representthe combination {* *\ of primary and secondaryinduced 2V while the rim +** may representonly primary 2V effects.Areas of la- mellartwinning (within core and rim) areuniaxial or very slightlybiaxial. Synthetic emeralds with CrrO, contents ranging from 0.1 to 2.7 weight percentare both uniaxialand biaxial.Figure 4a is a plot of CrzOecontent and 2V u.r.distance in a crystal cut normal to c. The CrrO, b?.,' content is clearlynot responsiblefor the anomalous Fig lb. Azimuth orientations of optical axial plane optical character. The emeralds examined were measurementswithin the tourmaline shown in Fig. la. A-B indicalesmicroprobe and 2V traverseline. grown by flux-melt techniquesat different times and in different batches.One crystal showsgood uniaxial observation we infer that anomalous optics are not materialin the core with biaxialmaterial on the rim, due to liberal substitutionof elementsalone. Zoned while a second shows no uniaxial material at all. crystalsof beryl from other than pegmatiticenviron- While the first crystalshows no strain birefringencein mentsmay or may not be anomalouslybiaxial (pri- the core(most portions),the secondcontains numer- mary or secondary2V). ous loci of strain shadows. Differing growth rates may possibly account for the differencein the strain E z o birefringenceand biaxiality observed.Where strain v!l F =E 2 coL birefringenceis at a minimum, the 2V is at a mini- Rr{ o U ,Eb mum regardlessof the changesin chemistry. Thus, t .Ei -'on ..6 | geometry i two different crystals with identical and 2V Po o _oE' magnitude of chemical zonation may show com- pletely different 2V traversecurves. Figure 4b shows 42 the orientationsof the opticalaxial planeswithin the 4l crystalshown in Figure4a. 405

I WEIGHT COLORLESS PERCENT I 3 oxtDE o7

o9 o6 o3

DISTANCE A B Fig 2a. Weight percentoxide and 2V variation curveswithin a lmm. 'pencil'tourmaline crystal from the San Diego Mine, Mesa Grande, California. SiO, content is nearly constantat about 38 5 Fig.2b. Azimuth orientations of optical axial plane weight percent,while MgO and TiO, contentsare alwaysless than measurementswithin the tourmaline shown in Fig. 2a A-B 0.02 weight percent Section cut normal to r'. indicatesmicroprobe and 2V traverseline. FOORD AND MILLS BIAXIALITY JZI

l2' CoRE + go

2V 40 oo

NorO

WEIGHT PERCENT OXIDE

CsrO b, \ooo.o_ - _d

ol 34 8 mm DISTANCE A B

Fig 3a. Weight percentNarO, CsrO, and2V variation curveswithrn a'pocket'crystal of goshenite-morganitefrom the OceanView Mine, Pala. California. Sectioncut normal to c.

Jeremejeuite uniaxialcore is surroundedby a zoneof biaxialmate- rial with 2V rangingfrom l5o to 33o(+ 1'), which in Specimensof jeremejevitefrom the type locality in turn is mantledby a thin zoneof lesser2Z(0" to 5') the USSR and from South Africa have been ex- and finally by a uniaxial rim. The South African mained in detail.This study is part of a more compre- material, on the other hand, has a uniaxial center hensiveexamination involving redetermination of the with biaxial outer zonesonly, and the maximum 2Z optical propertiesand composition as well as X-ray was 22" t lo. Sector extinction patterns are similar characterization (Erd and Foord, in preparation). to those given by Strunz and Wilk (1974), and are Azimuth orientationsof the optical axial planesin consistent with hexagonal symmetry. Within each the various optical domains are extremely regular sector and outer growth zonesextinction is undula- and relatedto the six-fold axis of the mineral (Strunz tory, while the central core in both casesis generally and Wilk, 1974).X-ray precessionand Laue photo- perfectlyisotropic (vieweddown c). The Russianma- graphsshow no deviationsfrom hexagonalsymmetry terial has a portion of the core which is not uniaxial in different portions of material from both localities. (is biaxial) and is homogeneousin chemicalcomposi- A maximum 2V of 33" a lo has beendetermined for tion. There is no apparent correlation between 2Z the Russian material. The material from both local- and chemical zonation (Fe, Si, F) in either sample. ities is of gem quality and has virtually no flaws or The apparentcombination of first-orderand second- inclusions.Boundaries between the optical domains order prisms may be due to differencesin original are very sharp.We note that in the Russianmaterial a growth velocities,or to recrystallizationphenomena. 322 FOORD AND MILLS' BIAXIALITY

GROWTH ZON€ 0o). However,random areaswithin the apatitecrys- tal have 2V valuesas high as 4o (t l'). Mn-bearing carbonate-apatitefrom the Laacher See,with a uni- axial core (opticallynegative) and biaxial rim (opti- cally positive)was describedin Rosenbusch(1927, p. rlc+-,d 242). Heating causedthe disappearanceofthe biaxial opticalcharacter, most likely by the annealingout of x internalstresses. A- + -B (anrelhysl) x+ 1( Quctrtz * t Concentric and sector-zonedamethyst crystals are biaxial (Hurlbut and Walker, 1976).Vein-filling amethyst from Thunder Bay, Ontario, Canada, showssimilar properties, with 2 Z valuesas high as6o * I ". X-ray precessionphotographs show hexagonal symmetry. There is no apparent correlation between iron (FeO < 0.12) and (AlrO, ;;; aluminum < 0.05) content,2V, or amethystinecolor. Schistsand Fig 3b Azimuth orientations of optical axial plane gneissesmay contain strainedquartz which often is meilsurementswithin the goshenite-morganjteshown in Fig 3a ,,1-Bindicates microprobe and 2/ traverseline. anomalouslybiaxial. Kleinite Apatite KIeinite, Hg,N(Cl,SOo).xH,O, from Terlingua, Compositionally-zonedapatite from the Himalaya Texas, and McDermitt (Cordero), Nevada, shows dike system,showing a pink centerin sharpcontact braxialcharacter. Terlingua material containing both with a light bluerim, is both uniaxialand biaxial.The SO, and Cl is uniaxial or only slightly negative, weight percentMnO risesabruptly from about 0.08 whereasthe McDermitt material,which is nearlythe to 0.65, while that of CaO decreasesfrom 55.7 to pure Cl end-member,has a high 2V (75-85") and is 55.3, in traversingfrom the pink core to blue rim. opticallypositive. Two polymorphsof kleiniteoccur There is no changein 2Z acrossthis contact(2V : in the materialfrom Terlingua(USN M 86641) with a

I I I RIM snrnecoNracl CORE suanecorracr I RIM l I 2V 8:

2.O 'o-q WEIGH T -o_-o- -o--o--o--o-_o - -o---o''o--o--o- -o I ^---- -^- -o-{- -o --o- PERCENT I.O o-o.Gocooo-" ' Cr^O- z c o.o o | 2mm A DISTANCE B Fig 4a. Weight percentCrrO, and 2V varialion curveswithin a crystal of syntheticemerald cut normal toc FOORD AND MILLS. BIAXIALITY

: 13.5A and a : 40.5A, whereasall of the material GREEN CORE from McDermitt has a : 13.5 A. Color-zoningis observedin the kleinite from McDermitt (clear yel- low coreswith reddishorange rims). It is not known if the material from Terlingua is growth-zoned as ** only crystal fragmentswere obtained for study pur- X poses. In this near-surfacehot-spring environment (about I atm and 150"C), external mechanical stressescan be considerednegligible. However, inter- nal stressescaused by cooling or crystal-growthirreg- I mm. ularities may be present. Fig.4b. Azimuth orientations of optical axial plane measurementswithin the syntheticernerald shown in Fig.4a. A-B Rutile indicatesmicroprobe and 2Z traverseline. Rutile often shows polysynthetic twinning on {01l}. Suchmaterial examined in this study (locality automated diffractometer indicated no deviations unknown) shows 2Z valuesat the twin boundaries from Ia3d symmetry. IOAP : (ll0)l as high as 20'(+). Away from the twin boundaries, within chemically homogeneous Discussionand conclusions material,no biaxialityis observed.Compositionally- The optic angle is a function of the indices of zoned rutile has not been examined in this study. refraction, which in turn are functions of the struc- Winchell and Winchell(1951) ascribed the biaxiality ture and compositionof a mineral.Chemically-zoned to deformationand/or twinning. crystalsmay also exhibit correspondingzones of dif- fering thermal expansion coefficientsand compres- Garnet sibility constants, which may be correlated with Mineralsof the garnetgroup may showanomalous changes in optical properties. When pressure and birefringenceand optical character.Garnets showing temperature changesoccur, secondary2ts may be strong anomalous optical character are either induced. At present,we are unable to differentiate strained (mechanicalstress) and/or noticeably com- betweenprimary and secondary2V or the combina- positionally zoned (internal stress). Mechanical tion thereof.If chemicalzonation of sufficientmagni- strain can be ruled out in those crystalshaving alter- tude is presentover a small domain size(0.X mm), nating isotropic and biaxial zones, as for all other then the resultantchanges in the indicesof refraction minerals previouslydiscussed. Grossular-andradite causebiaxiality, whether or not secondarystrain is from Darwin, California, shows excellentbiaxial in- developed during subsequentpressure and temper- terferencefigures. Cores of homogeneous,isotropic ature changes. andradite (Androo)are mantled by successivegrowth Substitution of significantquantities of variousele- zonesof grossular-andraditeranging in composition ments into minerals resultsin changesof optical and from GrossuoAnduoto GrossoAnd,oo(H. S. Eastman, X-ray properties. A single color-zoned tourmaline StanfordUniversity, personal communication, I 976). crystal having cell dimensions(black-a = 15.97A, c Biaxial and isotropic zonesare intermixed, with the :7.19 A; pink-a : 15.85A, c:7.11 A) differing isotropic zones being Andro or greater. 2V values by one percentor more in the differentcolored zones from 30o (-) to 65'(+) havebeen measured, and the may result in covarying by as much as five percent birefringenceis about 0.01.Zones of extremelynar- ( 1.640-1.675).If the changein chemicalcomposition row lamellae (0.02 mm or less) yield lower-quality and resulting changesin X-ray and optical parame- optic figures. Recently, Takeuchi and Haga (1976) ters are distributed over a large domain, then the have documented ordering of Fe and Al in grossu- resultinginternal stressesand the index of refraction lar-andraditefrom Munam, Wanghedo,North Ko- differenceswill be minimal. However, where a sharp rea, which resultsin orthorhombic material with very break in composition occurs,the cell dimensionsdif- high 2L's. Zoned uvarovite from Outokumpu, Fin- fer over a smaller domain. The indicesof refraction land, is birefringent (G. E. Brown, Stanford Univer- also differ, causingan apparentbeta index and anom- sity, personalcommunication, 1976), but examina- alous 2V. Abiaxial charactercan thereby be induced tion of symmetry-equivalent reflections on an in a hexagonalsubstance. FOORD AND MILLS BIAXIALITY

Table 3 Proposedcauses for anomalous braxialityin some dimetric and isometriccrystals

Mineral Caus e Reference

Grossu1 ar- andradite order-disorder Takeuchi and Haga (1976) RutiIe twinning tntt,,n"o"t Jerenej evite Tecrystallization ?

Calcit e nhrca tTanci fi ^n Boettcher and Wyllie (19671

Calcite rnechanical deformation Turner (1975a,b) Tournaline, beryl, apatite, k1einite, internal stress resulting from (Prinaty 2V) jeremejevite ?, arnethyst ?, garnet ? chenical zonation and/or This paper included growth defects

Tourmaline, beryl apatite, internal fron , stress resulting (Secondary 2ll) j erenej evite ? pressure and/or ternperature This paper changes Uvarovite G. E. Brown, personal comnunication, L976

Many of the crystalsexamined (Table 3) show possiblythe Darwin garnet, the anomalousoptical color zonation caused by compositionalzonation characteris believedto be due to non-topochemical (beryl, tourmaline,jeremejevite, apatite, and klein- causes. ite). Becausethe thermal expansioncoefficients and Materialswhich are optically biaxial may be un- compressibilityconstants are differentfor eachof the questionablyhexagonal, tetragonal, or cubic on an zones, subsequent pressure and/or temperature X-ray scale.Optical methodssample far larger vol- changeswill result in different internal stressesand umesol materialthegn X-ray diffractionmethods, and secondary2V's in differentzones. The releaseofthese hence the symmetry as determined by both proce- internal stressesmay result in the fracturingof the duresmay be the sameor different.For example,if interiorsof many concentrically-zoneduniaxial peg- an entirecut platefrom a biaxialtourmaline with no matite minerals,while the exteriorsremain relatively artificialstrain release due to crushingwere examined -free.In tourmaline, stressis further relieved by X-ray diffraction, then the symmetry as deter- by the developmentof microcline-likegrid twinning. mined by X-ray methodsshould be identicalto that This type of stress-relieffracturing should not be determinedoptically. In somecases, using the prop- confused with fracturing occurring from pressure erty ol anomalousbiaxiality, one can quickly deter- quenchingproposed by Jahnsand Burnham (1969). minethe presenceor absenceof compositionalhomo- Additionatly, secondary2V may be produced by geneityin materialsnot mechanicallystrained, even phasetransformations. Studies of calciteby Boett- though chemicalzonation may not be directly cor- cher and Wyllie (1967)and Turner (1975a,b)have relatableto the observedanomalous biaxiality. Very demonstratedthat mechanicalstrain and phasetrans- often color-zoningis absent,but anomalousbiaxial- formation (aragonite-)can inducean anoma- ity is observable(e.g. beryl). If a sectionis cut normal lous 2V in an otherwiseuniaxial mineral. to c and showsvariable2V, onemay suspectcompo- In the observationof anomalousoptics it is impor- sitional zonation of varyingmagnitude with or without tant to notethe distinctionbetween causes that areof accompanyingstrain (internal or external).Chem- crystallographicorigin and those that are not. Nu- ically-homogeneouscrystals containing primary merous mineralsof topologicallyhigher symmetry growth defectsmay yield anomalousoptical proper- have order-disorderrelationships which result in a ties. lowering of the space-groupsymmetry. Garnet These factors may affect the optical propertiesof (grossular-andradite)as describedby Takeuchiand trimetriccrystals as well. The observed2V for a given Haga (1976) is topologicallycubic, but becauseof trimetric crystal often will not agree with the 2V orderingin the octahedralM, sites(Fe-Al ordering), calculatedfrom indicesof refraction.Willard (1961) is topochemicallyorthorhombic with spacegroup has succinctly stated the problems encounteredbe- Fddd. Other common examples include potassium tweencalculated and observedoptical axial anglesfor leldspar(Al-Si order) and cordierite(Al-Si order). biaxial minerals. As statedby him (p. 202) "within For all of the casesdiscussed in this paper, except the limits of accuracyfor measuringindices of refrac- FOORD AND MILLS BIAXIALITY 325 tion, the 2V cannot be definitely establishedas a Countt'.Calilbrnia Ph D. thesis,Stanford University,Stanford, specificvalue." The lower the birefringenceof a min- California Hurlbut, C. S Jr. and D Walker (1976)Amethyst from the Ka- eral, the greater will be the possiblevariations in the lomo district, Zambia (Abstr.) 2nd Biennial MSA-FM Sym- calculated22. Studiessuch as that by Rinne (1926) posium on Crystal Chemistryand Paragenesisof the Gem Min- on topaz have shown the existenceof strain-induced erals,Tucson, Arizona. 2V changeswhich can be removed by heating. Iso- Isogami,M and I. Sunagawa(1975) X-ray topographrcstudy ofa gami and Sunagawa(1975) have correlatedgrowth topaz crystal. Ant Mineral , 60, 889-897 Jahns, R. H. and C W Burnham (1969) Experimentalstudies of defectswith optical anomaliesin topaz. pegmatitegenesis: I A model for the derivationand crystalliza- tion of granitic pegmatites Econ Geol.,61,843-864 Acknowledgments Johannsen,A (1918) Manual of PetrographicMethods, 2nd Ed. The authors wish to acknowledgethe following people for re- McGraw-Hill Book Company,New York viewingvarious preliminary versions ofthis paper and aiding in the Rinne, F. (t926) Bemerkungen uber optische Anomalien, ins- preparationofthe final version:W E Dibble, B. H W S DeJong, besonderedes brasilianerTopas. Z. Kristallogr,63,236. R C Erd, G D Eberlein,J G. Liou, G. R. Holzhausen,P. E. Rosenbusch,H (1927) Mikroskopische Physiographieder Mineral- Long, R. H. Jahns,and R C. Wilcox We thank G E Brown for ien und Gesteine E. Schweizerbart'scheVerlagsbuchhandlung, his suggestionsand commentsduring preparation,and Harvey S. Stuttgart Eastman for calling attention to the grossular-andraditegarnet Schaller,W. T, R. E Stevensand R H. Jahns(1962) An unusual lrom Darwin, California, and for the useof microprobe analytical beryl irom Arizona. Atn Mineral , 47, 672-699 results Edward and StanleyUstan of Fontana, California, kindly Strunz, H and H wilk (1974)Jeremejewit als Edelsteinaus SW- provided samples of synthetic emerald. Samples of Brazilian Afrika Z Dtsth Gerntnolog Gesellschatt,23, 142-150 tourmaline and Utah red beryl were provided by Edward R. Swo- Takeuchi, Y and N Haga (1976)Optical anomaly and structure boda of BeverlyHills, California Numerous specimensof tourma- of silicate garnets P/oc Japan Acad , 52, 228-231 line were obtained from EugeneB Rynersonof San Diego, Cali- Tobi, A. C (1956)A chart for measurementof optic axial angles fornia, and Linley K Hall of Lemon Grove, California Addrtional Ant Mineral . 41. 516-519 specimensof tourmaline and berylwere provided by Louis Spauld- Turner, F. J ( 1975a)Biaxial calcite: occurrence, optics, and associ- ing Jr. of Ramona, California, and specimensof goshenite-mor- ated minor strain phenomena Contrib Mineral. Petrogr, 50, ganite from the Ocean View Mine were provided by Roland Reed 245-255 of El Cajon, California Samples of jeremejevite and kleinite - (1975b)Biaxiality in relationto visibletwinning in experi- (Terlingua, Texas)were obtainedfrom the Smithsonianlnstitution mentally deformed calcite. Contrib M ineral Petrogr , 53, throughJ S White,Jr. 241-252 Wagner,C E , C O. Pollard,R A. Young and G Donnay(1971) References Texture variationsin color-zonedtourmaline crystals. Am Min- eral l14-132 Boettcher,A. L. and P J Wyllie (1967) Biaxial calcite inverted ,56, Willard, R J (1961) A re-examinationof some mathematical from aragoniteAtn Mineral,52,1527-1529. relations in anisotropic substances.Trans Am Microst Soc , Colby, J. W. ( 1968)Quantitative analysis of thin insulatingfilms 80. l9t-203 1n J Newkirk, G. Mallett and H Pfeiffer,F.ds.,Aduances in X- Winchefl, A. N. and H Winchell (1951)Elements o.f Optical Miner- ray Anolt'si.t,Vol I I Plenum Press,New York. alog.r Port ll Dest'riptions oJ Minerals (4th Ed.) John Wiley Deer, W A , R. A. Howie and J Zussman(1966) An Introduction and Sons.New York to the Rotk-l"orntittgMinerals Longmans,Green and Company, London. Foord, E E. (1976) and Petrogenesisof Layered Peg- Manust'ripl ret'eiued,Det'ernber 6, 1976: accepted nrutite-Aplite Dikes in the Mesa Grande District, San Diego lbr publication, May 12, 1977