<<

American Mineralogist, Volume 65, pages 302-305, 1980

Garnet-fassaitecalc- nodule from La soufriire. st. vincent

JosEpH D. DEvINE AND HARALoUn SlcunDSSoN Graduate School of Oceanography, (Jniversity of Rhode Island Kingston, Rhode Island 02881

Abstract

The mineralogyof an unusualcalc-silicate nodule brought to the surfaoeby the 1902erup- tion of La Soufrierevolcano on St. Vincent (LesserAntilles arc) is described.It consistsof the crystalloblasticmineral assemblagefassaite + grandite + + t calcite.The pyroxeneis more aluminous(Al2O3 : 19.66weight percent)than any previously describedfrom terrestrialrocks, with about 45 mole percent Ca-Al Tschermak's molecule,37 mole percentdiopside, and 15 mole percent Ca-Fe3* Tschermak'smolecule in solution. Unit-cell data have been calculatedon the basisof X-ray powder diflractom- etrydata for thefassaite (a:9.104, D: 8.828,c:5.2MA, F: 106.95.;V:426.49A3) and the grandite (grossularrr-andraditerr;a: ll.905A V: 1687Ai).We infer a paragenesisre- lated to skarn-typemetamorphism of calcareoussediment.

Introduction the border betweencalcite and wollastonite is more The 1902 eruption of La Soufridre volcano on St. serrate, giving the appearancethat calcite has re- Vincent in the Lesser Antilles arc brought a variety placed the wollastonite. Small irregular patches of of nodules to the surface, including calc-silicate calcite locally extend from intergranular layers into metamorphic xenoliths (Lacroix, 1949; Sandrea, wollastonite grains. Other silicate have 1949) and cumulus nodules (Lewis, 1964, l973a,b). sharp borders where they are in contact with calcite We describe here a fist-sized calc-silicate nodule layers. (+V37472), found in the pyroclastic flow deposits of Apart from small silicate inclusions and the 1902 eruption, which consists of the mineral as- calcite,grain sizesare in the range0.2-0.8 mm. Table semblage fassaite pyroxene + grandite garnet * wol- I showsthe mode of one thin section of the nodule lastonite + anorthite t calcite. It is the only known which is believed to be representativeof its assem- occurrence of a garnet-bearing calc-silicate xenolith blage. associated with this volcano. The nodule also bears a The fassaite pyroxene in the nodule is richer in fassaite pyroxene with about 45 mole percent of Ca- alumina than any specimenobserved so far in terres- Al Tschermak's molecule in solid solution. trial rocks. Table I showsthe mean composition of 19 electron microprobe analysesof the pyroxene. The mean AlrO, content is 19.66 weight percent The texture of the nodule is mainly crystalloblastic (range 17.16Io 20.94weight percent).Knopf and Lee but in places becomessubpoikloblastic with garnet (1956) describeda similar skarn-typepyroxene con- or pyroxeneenclosing smaller anhedralgrains of the taining 15.75 weight percent AlrO, (the most alu- other silicate minerals. Becausethe silicatesin the minous pyroxenereported previously from terrestrial nodule generally appear to be interlocking, se- rocks). quencesof crystallization cannot be reliably inferred The pyroxenegrains are dark greenin hand speci- on the basis of texture..Calcite occurs as thin (-10 men and light golden brown in thin section.Anoma- micron) layers separatingsilicate grains and is never lous (Berlin blue) interferencecolors appearnear ex- observedto be completely enclosedby any silicate tinction under crossednicols. The axial angle (+) is mineral. Bordersbetween silicate minerals are gener- about 60-70". Typical pyroxene cleavageis moder- ally sharp,often planar, and show no sign ofreaction ately well developed.Neither twins nor exsolutionla- between grains. Where thin layers of calcite occur mellae were observed. between wollastonite and other . however. The pyroxene showsa nearly stoichiometricCaO 0/J0.3--OO4)K/ 80/0304-0302$02.00 DEVINE AND SIGURDSSON: GIRNTET-FASSAITE NODULE content, implying that all M2 sites are filled by Ca. Table I Composition of minerals in the V37472 nodule We have used the method of Papike et al. (1974) to Fass ai te crandite I'iollastonite Anorthite Calcite calculate relative proportions of Fe2* and Fe'* from our microprobe data. The results show that over 25 sio2 38.09i86)l 38.s1(26) 49 68(6s) 43 62(17) 036 ri02 0 72(I0.) 0.i8(7) 0 01(ll 00 002 percent of tetrahedral sites are filled by Al and that A1203 19.b6(85) 16. 84(s i) 1 4s(96) 36 31(20) 065 over half of the Ml sites are occupied by trivalent FeO*r 8 82(57) 7 .46(73) 0.0 0.0 00 Nlno 0 14 (4) 0.28(2.) 0 r7(2) 0.04(0) 0.49 species (Al and Fe'*). ugo 7.26143) 0.s2(3) 0.2r(s) 0.0 0.35

Norms calculated accordi.g to the conventions of Cao 24.88(3I) 3s 44(60) 46 47(62) 19 71(s9) s2 92 Na2O 0.03(1) o 04(1) 0 02(l) o 0s(2) o.o Yoder and Tilley (1962, p. 366-367) show that less K20 0.01(r) 0.0 0.0 00 00 than haH of the pyroxene's normative components lie Total 99 60(95) 99 87 98.01 99 13 54-79 within the pyroxene quadrilateral. Tschermak's #anals l9 3C 3I molecule accounts for the bulk of the norm with Ilodal e, 73.6 22 05 about 45 mole percent Ca-Al Tschermak's com- percent Nlicroprobe analyses were acconplished using ndneral and glass standards; ponent (CAT) and 15 mole Ca-Fe'* Tscher- data ieduction was acconplished using Bence and Albee (1968) alpha correc- (Fingcr mak's component (FTS). The corrponent is tlons applied autonatically by a slavcd PDP-11 niniconPuter and lladidracos , It 72) . percent lParentheslzc,l about 37 mole of the norm. figures represent the estinated staDdard deviation (e s d) As a first approximation this pyroxene may be in terro of ]east units cited for the value to their imcdiate left, thus 38.09(86) indicates an e s d of 0.86 wt %. viewed as a solution between diopside and CAT. de 2Feo* = all calculateal as Feo Neufville and Schairer (1962) report that the experi- mental limit of CAT solution in diopside is about 20 Hence, we infer that the garnet's composition may be weight percent at I atm pressure. The solubility limit expressed as grossularrr-andraditer, (mole). There is of CAT in diopside may be increased in two ways. very little component, most of the Mg in the Firstly, Hijikata (1973) has demonstrated experimen- nodule having partitioned into coexisting fassaite tally that complete solid solution along the join can (Dffil"o- : 0.07). X-ray powder diffractometer data be attained at a pressure of I I kbar (1020"C). Sec- were used to calculate the unit-cell dimension of the ondly, Wood (1975) has shown that addition of Fe'* grandite (a: ll.905A; unit-cell volume : 1687A'). to a pyroxene solid solution increases CAT solubility. Nearly pure wollastonite (Table l) occurs as clear We suggest that addition of Fe'* as FTS in the Sou- vitreous elongated grains showing good cleavage and fridre pyroxene has also promoted increased CAT sol- as anhedral mineral inclusions in garnet or pyroxene. ubility. Increased CAT solubility in the Soufridre py- Wollastonite is the only mineral to show complex in- roxene due to high pressure is considered unlikely tergrowths with intergranular calcite. It is uncertain because the required pressure is geologically unrea- whether calcite is actually replacing wollastonite or if sonable for the inferred source region. the opposite is true, but optically invisible inter- Unit-cell parameters of the pyroxene were calcu- growths of extremely small calcite grains and wollas- lated using X-ray powder diffractometer data (a : tonite could conceivably account for the low totals in 9.705,b: 8.828,c: 5.2A4A,F : 106.95";unit-cell microprobe analyses of wollastonite. volume :426.49A'\. The data of two runs were aver- The anorthite (>Annn) in the nodule is colorless in aged for the calculation. A Si calibration standard thin section; K is present in only trace amounts was used; no unit-cell refinement program was ap- (Table l). Anorthite is only a minor constituent and plied. The unit-cell dimensions differ from those ex- is most frequently observed enclosed by garnet. Al- pected for ideal solid solution between CAT and bite twins and one other unidentified twin set are diopside: the c axis is compressed and the B angle en- present. The low alkali content of the anorthite (and larged. The unit-cell volu:me is also smaller than that of the nodule in general) is a curious feature of this expected from ideal solid solution (by about one per- xenolith. cent), implying a small nr:gative volume of mixing. Calcite appears as thin intergranular layers in the The grandite in the nodule are golden nodule and as a possible alteration product of wollas- brown in hand specimen and light honey brown in tonite, but is only a minor constituent of the nodule. thin section. They are isotropic but show anomalous It is low in Mg (Table l). interference along fraclures which cross-cut the Paragenesis grains. The composition tiTable l) shows that nearly all X positions are filled by Ca; this also implies that By analogy with other well-documented skarn- nearly all iron is Fe'* arLd confined to Y positions. type occurrences of coexisting grandite garnet and 3M DEVINE AND SIGURDSSON:GARNET-FASSAITENODULE

Table 2. V37472nodule bulk composition position observed, but the relative compositional ho- mogeneity of the minerals and th€ crystalloblastic v37472 Nodul e texture of the nodule strongly suggest that the final nodule: bulk comp. "Average bulk comp. (all CaO Sedinent" assemblage observed is an equilib- of section recalculated (Poldervaart, rium assemblage. p1 ane as CaCOr) 1955) Calcite is considered to be a secondary alteration product for the following reasons. Firstly, si0 JI.O/ thermody- ? 38.89 44.5 Tio, 0 .69 0 .56 0.6 namic analysis of the metamorphic reactions possible A1.0" 18.45 15.03 10.9 in the sirnple system CaO-AlrOr-SiOr-CO, (data of Feo*l 7.85 6.39 Robie et al., 1978) reveals that prograde reactions be- FeO 0.9 tween calcite and silicates (e.9., garnet) oc- F"2os 4,0 cur at relatively modest temperatures relative to MnO 0.27 0.r7 0.3 IugO 4.76 3.88 2.6 those which must have prevailed in the pyroclastic CaO 29.06 t9.7 flow which carried the nodule down the flanks of the

Na2O 0 .04 0.03 1.1 volcano. If the extremely thin layers of calcite were KZO 0.0 0.0 1.9 present in the nodule at the time of its entrainment, coz (18.se) t3 .4 we feel that they would have reacted, with grandite Total 99.95 99.99 99 .90 for example, to produce a higher-grade metamorphic assemblage. There is no evidence that this occurred.

lF.o* Secondly, it seems unlikely that silicate minerals = all iron calculated as Feo. growing at the expense of calcite in prograde meta- morphic reactions would leave behind only thin in- tergranular layers of remnant calcite and yet never entirely enclose any remnant carbonate grains. It fassaite pyroxene (Tilley, 1938; Shoji, 1975; Hucken- seems more likely, therefore, that the calcite now in holz et al., 1974), we conclude that the nodule under the nodule was precipitated from bicarbonate-laden consideration is a thermal contact-type skarn assem- meteoric in the time intervening between the blage, even though the original field relations are not 1902 eruption and the date of sample collection. in evidence. Specifically, no limestones or other cal- The paragenesis of the nodule can be interpreted careous are exposed on St. Vincent in the in light of phase relations in the simple system CaO- vicinity of La SoufriCre. We propose that the nodule AlrOr-FerOr-SiO, which has been investigated ex- is derived from calcareous sediments deposited on perimentally by Huckenholz et al. (1974). An iso- the submerged arc crest in the St. Vincent region baric (l aLm) T,X section of the grossular- prior to the build-up of La SoufriCre volcano from join (CarAlrSirO,r-CarFel*SirO,r) shows that com- the sea-floor. Such sedimentation is likely to have oc- plete garnet solid solution exists at temperatures be- curred during the Miocene hiatus (Martin-Kaye, low 798"C. At higher temperatures, however, garnet 1969) after cessation of activity in the Limestone Ca- solid solution breaks down to other more stable ribbees (eastern arc) and prior to the main phase of phase assemblages.One of these phase volumes con- Plio-Pleistocene volcanism in the present arc. The tains the mineral assemblagegarnet + fassaite * wol- most recent eruption of La Soufridre (April, 1979) lastonite (that which is observed in the nodule); it is brought xenoliths to the surface, which sup- stable over the temperature range 935 to I140'C. We ports the idea of calcareous sediments existing at suggest, therefore, that the Soufriire nodule repre- depth in the vicinity of the vent. sents calcareous which was first metamor- The calculated bulk composition of the nodule (on phosed to a grandite-bearing assemblage (Z < the basis of modal analysis) is shown in Table 2. Also 798'C); the assemblage was then heated to higher shown is the bulk composition with all CaO recalcu- temperatures, perhaps during entrainment, and the lated as CaCOr; this composition is similar to the grandite originally in the assemblage was metamor- "average sediment" ofPoldervaart (1955,p. 132) and phosed into the new assemblage grandite + fassaite presumably is close to the corrposition of the cal- + wollastonite (T > 935"C). Anorthite is interpreted careous sediment which was decarbonated by ther- as a remnant of some prograde assemblage and cal- mal . Metasomatic processes may cite as a secondary alteration product. have played a role in determining the final bulk com- The effect on the host magma of the possible as- D E VI N E A N D SI GU RDSSOIT..GIRffET- FA SSA I TE N O D ULE 305 similation of calcareoussediments and incorporation dr?s. Unpublished D. Phil. Thesis, Oxford University, Oxford' of CO, into the melt remains an open question.But England. - (1973a) Mineralogy of the ejected plutonic blocks of the escapeof assimilatedCO, from magmaswhich have Soufriere volcano, St. Vincent: , pyroxene, , stoped through calcareoussediments could obviously and magnetiteparagenesis. Contrib. Mineral. Parol', 38, 197- contribute to the explosivenessof eruptions. Volca- 220. noes intruding calcareouscountry rocks (e.9., Ve- - (1973b) Petrology of the ejected plutonic blocks of the suvius) might be profitably consideredin this light. Soufrierevolcano, St. Vincent, West Indies. J. Parol'' 14,81- tt2. Martin-Kaye, P. H. A. (1969) A summary of the geology of thc Acknowledgments LesserAntilles. OverseasGeol. Mineral Resources'10' 172-2M- (1974) We thank C. Klein and M. J. Gole for commentswhich im- Papike,J. J., K. L. Cameron and K. Baldwin quadrilateral proved the manuscript. This work was made possibleby NSF and : characterization of other than grant OCE-77-25689. components and estimates of ferric iron from microprobe data (abstr.) GeoL Soc.Am. Abslractswith Progratns,6, 1053-1054. Poldervaart,A. (1955) of the 's oatst' Geol. Soc- References Am. Spec.Pap.62. Bence,A. E. and A. L. Albee (1968)Empirical correction factors Robie, R. A., B. S. Hemingwayand J. R. Fisher (1978)Thermody- for the electron microanalysis of silicates and oxides. J. Geol., namic propertiesofminerals and relatedsubstances at 298'15K 76,382-403. and I bar (105 pascals)pressure and at higher temp€ratures. de Neufville, J. and J. F. Schairer(1962) The join diopside-Ca U.S. Geol. Surv.Bull. 1452. Tschermak's molecule at atmospheric pressure. Carnegie Inst. Sandrea,A. (1949)Les enclavescalcaries des lavesde Saint-Vin- l[ash. Year Book, 61, 56-59. cent (Petites Antilles). BulI. Soc.fr. Mineral' Cristallogt., 72' Finger, L. W. and C. G. Hadidiacos(1972) Electron microprobe 3't+387. automation. CarnegieInst. Wash. YearBook, 71,598-6N. Shoji, T. (19?5)Role of temperatureand CO2 pressurein the for- Hijikata, K. (1973) Phase relations in the system CaMgSi2O5- mation of skarn and its bearing on mineralizatiott. Econ' Geol., CaAl2SiO5 at high pressures and temperatures. "/. Fac. Scr'., 70,739-749. Hokkaido Univ., Ser.4, 16, 167-178. Tilley, C. E. (1938) Aluminous pyroxenes in metamorphosed Huckenholz,H. G., W. Lindhuber and J. Springer(1974) The join lime-stones.Geol. Mag., 75, 8l-86. CaSiO3-Al2O3-Fe2O3of the CaO-Al2O3-Fe2O3-SiO2quater- Wood, B. J. (1975)Mixing propertiesof tschermakiticclinopyrox- nary syst€m and its bearing on the formation of granditic gar- enes and their application to geobarometry' Exlended abstracts. nets and fassaiticpyroxenes. Neues Jahrb. Mineral. Abh., l2l, Int. Conf. Geothermometry Geobaromelry' The Pennsylvania t60-207. StateUniversity, University Park, Pennsylvania. Knopf, A. and D. E. Lee (1956)Fassaite from near Helena,Mon- Yoder, H. S. and C. E. Tilley (1962)Origin of basalt magmas.An tana.Am. Mineral., 49,73-7'1. experimental study of natural and synthetic systems' "/. Lacroix, A. (1949)Sur des enclavesendopolygdncs de Saint-Vin- Petrol., 3,342-532. cnnt.Bull. Soc.fr. Mineral. Cristallogr.,72,5'll-590. Lewis, J. F. (1964)Mineralogical and PetrologicalStudies of Plu- Manuscript received,June 7, 1979; tonk Blocks from the Soufriere Volcano, St. Vincent, West In- acceptedforpublication, September 14' 1979-