An Occurrence of a Modulated Serpentine Related to the Greenalite

American Mineralogist, Volume 74, pages 637441, 1989

An occurrenceof a modulatedserpentine related to the greenalite-caryopiliteseries

SrnpnnN GuccnNnprpr Department of Geological Sciences,University of Illinois at Chicago,Chicago, Illinois 60680, U.S.A. SruncBs W. Berr.nv Department of Geology and Geophysics,University of Wisconsin, Madison, Wisconsin 53706, U.S.A.

Ansruct Transmission electron microscope (rnvr) and X-ray powder-diffraction studies indicate that material originally described as the Zn- and Mn-rich serpentine miheral baumite contains predominantly submicroscopic coherent intergrowths of 7-A and l4-A phases. The 7-A phasesinclude at least two polytypes (group A and either group B, C, or D) of lizardite and a modulated l: I layer silicate similar to those of the greenalite-caryopilite series.The l4-A phaseincludes a dominant chloite-Ibb structure.A chrysotile-like phase is present also, although it is rare. Semiquantitative chemical analysesindicate that all phases are Zn and Mn rich, but crystal-chemicalarguments are used to suggestthat the greenalite-caryopilite-like phase is relatively Al poor. The modulated l:l layer silicate difers structurally from greenaliteand caryopilite by having island-like domains of about 30 A (vs. 21.3-23.3 A for greenalite and 16.7-17.2 A for caryopilite). Accompanying veinlets appear to be lizardite-1Z altering to chlorite-Ibb and chloile-Iba, with these phaseschemically distinct from those more directly associatedwith the modulated l: I layer-silicatephase.

Ixtnonucrrox over a foot in length, baumite comprised no more than 25o/obyvolume. The chamositeand baumite brecciahosts Frondel and Ito (1975) describedbaumite, a new ser- white willemite (crystallites 1 x I cm in size) and white pentine mineral, and "brunsvigite," a variety of chlorite, granular calcite, along with franklinite, acmite (:aegir- from specimensfound at the Buckwheat dump at Frank- ine), and stilpnomelane. The stilpnomelane is brown, lin, New Jersey.They also reported a "pennine" from a probably Fe rich, and varies in texture from fine to coarse franklinite ore specimen.These minerals contain unusu- grained. Locally, stilpnomelane may be enclosedby frie- ally large amounts of MnO (5.5 to 12.3 wto/o)and ZnO delite, which occurs as late-stagecrystals in fractures. (4.75 to 9.6 wto/o).Baumite occursas fine-grained,dense, Baumite is of special interest becauseit is possibly a black masses"up to a foot in diameter" and is brownish modulated structure as a result of its reported low Al yellow in thin section.The "brunsvigite" occursas green- content and its high tetrahedral Si and octahedral Fe * ish-black radial crystals in low-temperature hydrother- Mn + Mg * Zn. Samplesstudied (Harvard Museum no. mal veinlets crossingthe baumite. The "brunsvigite" and 114072:type material labeled "baumite" and an unnum- "pennine" describedby Frondel and Ito are more prop- bered specimen labeled "brunsvigite" from J. L. Baum) erly designatedmanganoan zincian chamosite and clino- did not contain any of the clinochlore crusts.The present chlore, respectively,based on the reported chemical anal- paper concludesthat both the baumite and the chamosite yses.Both chloritescontain substantialamounts (13.1 to are submicroscopic, coherent intergrowths of 7-A and l4-A 14.0 wto/o)of alumina, and baumite apparently contains phases plus small amounts of other materials. It is lesseramounts (6.60 wto/o).Bayliss (1981) argued that noted, furthermore, that the material does contain a new baumite should be redefinedas a manganoanferroan liz- mineral similar to those of the greenalite-caryopilitese- ardite-lT, after he confirmed that the powder pattern as ries. However, this mineral cannot be documented ade- presentedby Frondel and Ito could be indexed on a 1T quately to give it new speciesstatus becauseit is inti- cell. This suggestion,however, was not approved by the mately mixed with other phases. IMA (E. H. Nickel, personalcommunication, 1988). Unpublished work (P. Dunn, personalcommunication, ExpnnrrvruNTAl, METHoDS AND INTERpRETATIoN 1988) indicated that a clinochlore does occur associated with baumite, but this material has not been studied. Baumite Dunn re-examined the original material and found X-ray study. Over 40 Debye-Scherrerpatterns were ob- baumite to occur as a breccia cement in a low-tempera- tained from different submillimeter-sized particles of ture cavity or vein filling. Although the original masswas baumite. Most particles were cut from thin sectionsso as 0003-004x/89/0506-0637$02.00 637 638 GUGGENHEIMAND BAILEY: MODULATED SERPENTINE

Fig. l. Lattice fringes in optical (rur,r) images show 7-A and 14-A phasescommonly intergrown as aligned plates on the (001) plane. ChrysotileJike grains (note ringJike feature)were rare. The upper portion ofthis photograph,therefore, is more representative of the sample than the lower portion. to include only microscopically homogeneousmaterial. series or a physical mixture of 7 A and 14 A phases. The particles were mounted on glass fibers, centered in Support for both interpretations can be found in the ac- 114.6-mm-diameter film cameras,and exposedto filtered companying rrrvr study. A few difraction patterns show FeK" X-radiation. even-order00/ reflectionsas doublets,apparently because All patterns contain a strong 7-A diffraction line and a of compositional diferences between the 7-A and l4-A medium-intensity l4-A Hne. The ratio of the two inten- phaseswithin someparticles. Nearly all patternsalso show sities is approximately constant in patterns from most diffraction lines of strong to medium intensity at 1.54 A particles, although in some cases,the 7-A hne is en- and 1.60 A, which can be indexed as douoof chlorite and hanced.The non-00/ lines show considerablesimilarities douoof a greenalite-caryopilite-like phase, respectively. between diffraction pattems also, but with some differ- Heating particles at 500 "C increasesthe intensity of the encesin the weaker lines. Two patterns contained a line l4-A hne and decreasesall other lines. at about 9.5 A that is interpreted as indicating the pres- Transrnissionelechon microscopy(TEM) study. Sam- enceofa talc-like phase. ples of baumite were prepared by standard ion-thinning The l4-A hne indicates that chlorite is a major con- techniquesusing Cu grid supports. Also, powder mounts stituent of baumite, but the structural type could not be were made by dispersingfinely abradedmaterial on holey identified becauseofthe overlap with the patterns from C film supported on a nylon grid. Material was examined one or more 7-A phases.It was observed that the even in a rnol roocx microscope with a Tracor Northern Bps orders ofthe l4-A spacing (001 /: 2n) are consistently attachment. stronger than the odd orders. Calculations show that no The ion-thinned grids showedthe variable morphology distribution of heavy and light atoms between the octa- and textures of the mineral assemblagesof the samples. hedral interlayer sheetand the octahedralsheet ofthe 2: I Although most of the grains have platy morphology, some layer of a chlorite structure can account for the ob- sample areas exhibit chrysotileJike rolls (Fig. l). Most servedintensity ratios. For example,increasing the heavy platy grains are intergrowths of 7-A and, l4-A phases atom content ofthe interlayer decreasesthe 001 intensity aligned parallel to the (001) planes. In one case,a 2l-A relative to 002, but increasesthat of 003. The observed repeat was observedin the diffraction pattern, suggesting intensities require either an interstratified 7- and l4-A a regularity in intergrowth at the unit-cell level. This phase, GUGGENHEIM AND BAILEY: MODULATED SERPENTINE 639 however, was extremely rare. Unlike the X-ray study, no phasewith a 9.5-A spacingwas noted by rnvr. Electron-diffraction patterns ofgrains showing a platy morphology and containing c* characteistically show the superpositionof two nets. One net could be attributed to a l4-A chlorite and the other to a 7-A greenalite-caryopilite-like serpentine. Like greenalite-caryopilite (cf. Guggenheimetal.,1982, Fig. 5C), this 7-A phasehas a modulation that producessatellite reflectionsadjacent to the reflections normally observedin serpentinesof ideal l: I structure, thereby making identification definitive. In zt,t all caseswhere the 00/ diffraction patterns were compos- ites, the l4-A reflectionseither superimposedirectly over the reflections of the 7-A phase or fall exactly halfway between.Therefore, any differencesin chemical composition betweenthe 7-A phaseand the l4-A phaseare not observed by comparing c-axis spacings.Reflections on 0kl-type nets with k + 3n are generally streaked along l the row lines because of random layer-stacking se- I v* quences. A For one grain, however, the diffraction pattern showed Crl sharp k + 3n reflections superimposed on the streaks. The modulated 7-A phaseis dominant in this grain, and Fig. 2. Electron-diffractionpattern showing the superposi- it is apparent that the lateral dimensions of this phaseare tion of two hol-typenets, one 7-A (lizarditeof groupB, C, or : largerthan thoseofthe l4-A phase.The l4-A 00/ reflec- D) andone 14-A(chlorite- Ibb) phase.Both phases have B 90'' : : tions could be removed from the pattern by tilting the c chlorite,I lizardite. sample stage,thereby allowing the identification of the polytype of the Z-A modulated phase. The positions of the sharp k + 3n reflections require a oneJayer c repeat only be characterizedas a member of groups B, C, or D with B : 90o,and this uniquely identifies the 7-A mod- (Bailey, 1988), with group A serpentineeliminated as a ulated phase as related to the 1f polytype of group C. possibility. This is the structure on which the modulated greenalite One observed h)l pattern is an exception in that the superlatticeis based(Guggenheim et al., 1982).The over- chlorite can be indexed on a lattice with B : 97oand not lap of reflectionsprevented the accuratedetermination of on a lattice with B : 90'. Overlap on this pattern with lateral cell dimensions a* and D*, but such determina- reflections from the intergrown 7-A phase prohibits the tions were possible from the hk}-type patterns (see be- use ofthe ftOl intensities to characterizethe chlorite fur- low). ther. The 7-A phase can be identified from the lattice A second set of superimposednets containing 00/ re- geometry as a group A (1M, 2M,, ot 3T) serpentine' flections shows spacingssimilar to those describedabove Finely divided material examined on holey C films and also contains lateral directions of both a 7-A and commonly showed hl

are superimposed.The patterns I were not sufrciently de- fined to permit detailed interpretation. The overlap of row lines is not due to twinning, as Gandolfi patterns of the same fragments show powder lines that require in- , dexing on two differently shapedunit cells, one with P : 90'and one with P :97" (or 104'for a 7-A phase).Gan- dolfi patterns of20 ditrerent particlesare fairly consistent. V; * The observed 00/ intensities of the chamosite can only be reproduced by having a mixture of 7-A and l4-A phases,similar to baumite. The evidencefrom the preces- patterns phases rr sion is that the two are in an oriented intergrowth, at least within the small volumes of the crys- * I tal that were studied. A 7-4, serpentine with B : 104' (lM polytype) gives I 20/ powder lines at d values similar to those of 201,I : 2n lines of a chlorite with : 97". Likewise, 1?n ,a\i* B a serpentine polytype has the same relationship compared to a chlorite with : 90". This complicates identifica- Fig. 3. Electron-difractionhk) net of the greenalite-cary- P the polytypes present. phases : opilite-like phaseshowing satellite reflections around ortho- tion of the The with B 90' hexagonalk: 3n reflections.The satellitereflections form rings are dominant, however, and this suggeststhat an original aroundthe subcell(k : 3n) reflections,suggesting considerable lT lizardite (B : 90) has been partly transformed to positional(rotational) disorder of the tetrahedralislands (see chlorite, primarily a lbb chlonte with B : 90'. A powder Guggenheimet al., 1982,for detailson a structuralmodel). line at d : 2.45 A, however, can only be indexed as 203 of a chlorite with 0 : 97' (most likely chlorite type lba). The lbb and Iba chlorites are commonly intermixed in Semiquantitative chemical analysisof grains supported natural occurrences(Bailey and Brown, 1962,p.848). by the nylon-grid were made by EDS.Each grain was iden- The 060 reflections of the two phasesin chamosite are tified by the nature of its hk) pattemon the basisof either coincidentin a very intensepowder line at d: 1.56 A. modulated or nonmodulated nets. Initial spectracollect- This line is not presentin any of the baumite patterns, in ed from empty grids to establishbase-line data indicated which douo: |.54 A for the chlorite phase.Likewise, the considerable Cu contamination caused by the sample 060 reflectionof the modulatedT-Lphase of baumiteat holder and microscope column, but no Zn contamina- d : 1.60 A is not presentin any of the chamositepatterns. tion. Zn appearsto be presentin all grains examined, but Therefore, the chlorites within baumite and chamosite considerable overlap with the Cu spectrum makes Zn are different, as are the 7-A phases.Although rrrvr study identification difrcult. Spectra (not illustrated) showed of the Z-A phase in chamosite has not been made, it is high background, but all analysesindicated the presence anticipated that it is a Fe-, Mn-, and Zn-ich lizardite of considerableamounts of Si, Fe, Mn, and Mg and lesser without structural modulation. amounts of Zn and Al. Spectra varied considerably between grains and from point locations within the bound- Drscussrou aries of individual grains. Becauseof the variations in Baumite was described originally as a simple serpen- grain thicknesses,high background, and considerableCu tine. Results here show that the material called baumite contamination, phasescould not be distinguished from contains a modulated serpentine(l:l layer silicate) struc- these analysesalone. turally similar to the greenalite-caryopilite series. Al- though the diffraction patterns of these three minerals Chamosite (greenalite,caryopilite, and the phasedescribed here) are X-ray study. The ubiquitous presenceof Mn- and Zn- nearly identical in configuration, the phasedescribed here rich chlorite intergrown with a modulated 7-A phase in differs in the measured satellite separation, E of 30 A. the original baumite mass may suggestthat this chlorite This contrasts sharply with the measured separation of is identical to the Mn- and Zn-rich chamositein the low- greenalite(21.3-23.3 A) and caryopilite(16.7-17.2 A) as temperaturehydrothermal veinlets that crossthrough the given by Guggenheimet al. (1982).These values repre- baumite mass. Although the chamosite crystals measure sent a measureofthe lateral extent to which congruence up to 2 mm in size, they are very poorly crystallized and between misfit sheetsoftetrahedra and octahedracan be appear to be pseudomorphs.Precession photographs ap- maintained before strain must be relieved by a structural proximating a single-crystalpattern could be obtained only perturbation. from very small fragments of the macroscopic crystals. The relatively large S value indicates that misfit be- Larger fragments produce powder patterns only. tween the sheetsof tetrahedra and octahedra is not as Although the precessionpatterns are poor, they do show greatas in the more Fe-rich (radius r :0.78 A) or Mn- 201and 40/ row lines in which two nearly alignedpatterns rich (r : 0.83 A) samplesof the study by Guggenheimet GUGCENHEIM AND BAILEY: MODULATED SERPENTINE 64r

al. (1982), presumablybecause of the Zn (r : 0 .74 A) and mation to chlorite is incomplete. The chlorite intergrown Mg (r : 0.72 A) content of this phase.This addirional with the modulated serpentineis diferent from the chlo- data point breaks the linear trend on a plot of S vs. D rite describedhere as chamosite. parameter(cf. Guggenheimet al., 1982,Fi9.7) and gives more credibility to the possibility that the variation in S AcrNowr,r'ocMENTS is step-like. We thank C. Frondel and C. Francis for providing sampleno. I I 4072 The diffraction data (b" : 9.60 A for the modulated from the Harvard Mineralogical Museum We thank P Dunn of the serpentinevs. D : 9.24 A for chlorite) imply a partition- Smithsonian Institution for unpublisheddata and for assistancein locat- ing "brunsvigite" through J. L. Baum, who kindly permitted us to use ing of the large (Mn, Fe) cations to the modulated ser- material from his private collection. We thank also G. Harris, Depart- pentine structure. The qualitative chemical analysis sug- ment of Geological Sciences,University of Illinois at Chrcago,for help gestsa significantZn contentin the modulated l:1 layer with the semiquantitative chemical analysisand the ResearchResources silicate,in accord with the original bulk analysis.It seems Center of the University of Illinois at Chicago for use of its electron- We F. Wicks, Royal Ontario Museum, for re- unlikely, however, that the Al content in modulated microscopefacility. thank the viewing the manuscript. We gratefully acknowledgethe support of the serpentine is as high as reported originally, presumably Petroleum ResearchFund of the American Chemical Society through becausethe incorporation of Al would minimize the mis- grants i7263-AC2 and 17966-AC2-Cand the National ScienceFounda- fit between component sheetsand that would lead to a rion undergrants EAR87-04681 and EAR86-14868. nonmodulated phase. Likewise, partitioning Al to the RnnBnrNcBscrrno chlorite phase would account for its more ideal (nonmodulated) structure. Bailey, S.W. (1988) X-ray diffraction identification of the polytypes of mica, serpentine,and chlorite. Clays and Clay Minerals, 36,193-213. Coucr,usrolls Bailey, S.W., and Brown, B E (1962) Chlorite polytypism. I. Regularand semi-random oneJayer structures.American Mineralogist, 47, 819- Bulk material called baumite from Franklin, New Jer- 850. sey, consistsof a mixture of a chlorite with a modulated Bayliss, P. (1981) Unit cell data of serpentinegroup minerals. Mineral- 7-A phase that is similar to the modulated greenalite- ogrcalMagazine, 44, 153-156. Frondel, C., and lto, J. (1975) Zinc-rich chlorites from Franklin, New caryopilite series but is different in composition and in Jerseywith a note on chlorite nomenclature.Neues Jahrbuch fiir Mi- scale of the modulation. In some places the two phases neralogie.Abhandlungen, 123, I I l-l 15. are intergrown coherently. Curled asbestiform particles Guggenheim,S., Bailey, S.W., Eggleton,R.A., and Wilkes, P. (1982) are present,as well as a lizardite and a talc-like phase. Structural aspectsof greenaliteand related minerals. Canadian Min- l-18. Chamosite ("brunsvigite") crystals in veinlets cutting eralogist,20, the bulk material appearto be pseudomorphsof lizardite. MaNuscnrpr RECETvEDJuNe 27, 1988 As both lizardite and chamosite are present, transfor- Mauuscnrpr AccEprEDJ.rNuenv 20, 1989