Petrology of Contact Metamorphosed Argillite from the Rove Formation

American Mineralogist, Volume 66, pages 70-86, l98l

Petrologyof contact metamorphosedargillite from the Rove Formation, Gunflint Trail. Minnesota

THeonone C. Lesorre. J. J. PAprKE.D. T. VeNnaaN' Department of Earth and Space Sciences State University of New York at Stony Brook Stony Brook, New York I 1794

exo G. B. Monnv M innesot a Geolo gic al Survey St. Paul, Minnesota 55108

Abstract

The contactmetamorphic efects of the Duluth Complex on the argillaceousRove Forma- tion, northeasternMinnesota, are confined within a few hundred metersof the contact. The argillite has recrystallizedfrom the low-grade assemblagechlorite + muscovite+ quartz to assemblagescharacterized by cordierite + biotite + microcline + muscovite+ quartz. Cor- dierite coexistswith biotite over a wide range in Mgl(Mg + Fe) from 0.90 to lessthan 0.50. Andalusiteis restrictedto iron-rich rocks, although it occursin one unusual magnesium-rich assemblage.Locally, corundum + microcliae and cordierite * hyperstheneassemblages occur at the contact with the Duluth Complex. The pressure,P"oria, of metamorphismis estimated to have been 1500bars. The temperaturenear the contact ranged from -500o to 600oC,based on the partitioning of Na and K betweencoexisting muscovite and microcline. The equilibrium constant for the continuous reaction muscovite + phlogopite * quartz : cordierite * microcline + HrO is calibrated and used to determine "/n,o attending metamorphism.Results indicate that P",o rangesfrom P"",,oto jP"",,o.Calculation of the composition of COH fluid for a graphite-bearingsample iadicates that the remainder of the fluid is mostly COr. The Rove Formation doesnot show the extensivemineralogic eflectsof contact metamorphismthat the iron-formation shows,and the differencesreflect the large thermal and compositionalstability field for coexistingcordierite + biotite in pelitic rocks and the relatively small stability fields of the iron-rich minerals in the iron-formation.

Introduction to changesin metamorphic grade, and rocks as far The -l Gy Duluth Complex in northeasternMin- away as 10 km (along strike) show effectsof contact nesotaintruded an -2 Gy sequenceof sedimentary metamorphism.The zonesof metamorphismrecog- rocks that includes banded iron-formation (Gun- nized by French (1968) and Floran and Papike flint-Biwabik) and argillite and greywacke (Rove- (1978) are low-grade zonescharacterized by green- Virginia Formations).The contact metamorphic ef- alite or minnesotaite,intermediate-grade zones de- fects of the Duluth Complex on the iron-formation fined by grunerite-and fayalite-bearingassemblages, have been extensivelystudied by French (1968), and a high-grade zone characteitzed by ferrohyper- Bonnichsen(1968, 1969),and Morey et al. (1972) stheneor inverted pigeonite. (Biwabik Iron Formation), and by Simmons et al. The Rove Formation consistspredominantly of (1974)and Floran and Papike(1975, 1978) (Gunflint carbonaceousargillite and quartz-rich greywacke, Iron Formation). Theseiron-rich rocks are sensitive and the petrologyof theserocks has been determined in order to comparethe responseof the more K- and Al-rich rocks to that of the iron-formation under rPresent address: Los Alamos Scientific Laboratorv. Los low-pressuremetamorphic conditions. In this r€port, Alamos,New Mexico87545. the mineral assemblagesand mineral compositionsin 0003-o0/.x / 8 | / 0 I 02-0070$02.00 70 LABOTKA ET AL.: ROVE FORMATION, MINNESOTA 7l the Rove Formation are presentedso that the distri- The Rove Formation consistsof a lower argillite bution of elementsamong coexisting minerals in low- unit, a middle transition unit, and an upper thin-bed- pressure,pelitic schistscan be characterized.We use ded greywackeunit. The lower argillite unit is ap- the mineral reactionsthat are inferred to have oc- proximately 150m thick and consistsof thin-bedded, curred to estirnatethe conditions of metamorphism. fissile, fine-grained greywacke, silty argillite, and graphitic argillite. Irregular limestonelenses and cal- Geologicsetting careousconcretions also occur in the lower part of The Rove Formation is part of a sequenceof Pro- the unit. Bedsof greywackeare abundantin the tran- terozoic (-2Gl sedimentaryrocks assignedto the sition unit, and greywackedominates the upper unit. Animikie Group. The group unconformably overlies The complete thickness of the two upper units is 2.7 Gy Archean rocks and consistsof two units, the about 900 m, but the upper parts are generally trun- Gunflint lron-Formation and the overlying Rove cated by the Duluth Complex in the region around Formation. The stratigraphy and sedimentary pe- Gunflint Lake. trography of the Rove Formation have been de- The Duluth Complex,a compositemafic intrusion, scribedby Morey (1969),and the generalgeology of truncatesthe rocksof the Animikie Group with slight northeasternMinnesota is shown in Figure l. angulardiscordance. Consequently, the complexis in

Fig. l. Generalgeology of the Gunflint Lake area,northeast Minnesota. Sample localities are indicatedby the lettersKT (Kakakebic Trail), CR (CrossRiver), GT (Gunflint Trail), LL (Loon Lake), GP (Gunflint Palisades),ML (Mayhew Lake), SRL (South Round Lake), RL (RoseLake), and DL (Daniel's Lake). LIIBOTKA ET AL.: ROVE FORMATION. MINNESOTA contact with progressivelyhigher units of the Ani- be usedas phasediagrams if the chemical potentials mikie Group from west to east. Gabbroic sills and of HrO and each additional componentare constant dikes assignedto the Logan Intrusions (Weiblen et for all assemblagesin each projection (Thompson, al., 1972)pervasively invade the Animikie Group in 1957). the vicinity of Gunflint Lake. The Logan Intrusions grade appearto be both older and younger than rocks of Low the Duluth Complex and are part of the sameintru- The rocksfrom the Rove Formation are character- sive complex. ized by detrital texturesexcept near the Duluth Com- plex and adjacentto somesills. The clastic grains in Methods siltstonesand greywackeshave angular to sub- The Rove Formation is not well exposed;outcrops rounded shapes,and the matrix consists of fine- generallyonly occur beneaththe more resistantLo- (<0.25 nm) to very fine-grained (<0.0J mm) mi- gan sills or beneaththe Duluth Complex adjacentto caceousminerals. The micaceousminerals chlorite. the contact. Sampleswere collected at sill contacts muscovite,and biotite have recrystallizedfrom the and in sequencesaway from the sills to determine the presumably original clay matrix. The textures de- regionalthermal environmentand perturbationsim- pend stronglyon the rock type. The graphite-rich ar- posedby the sills. gillite near the baseof the formation contains angu- The rocks were examinedin thin section and the lar quartz and plagioclasegrains, very fine-grained identification of some minerals was aided by X-ray biotite and chlorite, and comparatively large (>0.5 diffractometry.Selected samples were analyzedwith mm) muscovitegrains that have an idioblastic form an ARL-EMXautomated microprobe using simple sili- and appearto be recrystallizeddetrital muscovite.In cate standards.Data were reduced on-line by the many samples,chlorite and muscoviteare finely in- method of Bence and Albee (1968) and the conec- tergrownin the matrix. Albite grains are irregular in tion factors of Albee and Ray (1970).Precision was shapeand contain fine-grainedinclusions of epidote monitoredby analyzingminerals of known composi- ("saussurite").In some rocks, epidote occurs as dis- tion and is -2Vo for major elements and -ll%o for tinct grains. minor elements.Unknown mineralswere analyzedin The observedmineral assemblagesinclude quartz two or three areason a standard thin section, and + muscovite+ chlorite * biotite * plagioclaseand mineral compositionshave rangeswithin the stated quartz * muscovite+ biotite * microcline + plagio- precisionover the area of a thin section. clase.Rocks that were collectedfrom RoseLake and The structural state of potassium-feldsparhas not farther east,near Lake Superior,do not contain bio- been determined,and the use of the term "micro- tite. The diagnosticassemblage quartz * muscovite cline" doesnot imply that K-feldspar has the triclinic + chlorite + biotite + microcline + plagioclaseoc- structure. curs in sample4ls from South Round Lake. Graph- ite is very abundant in rocks from the lower part of Mineral assemblagesin pelitic rocks the formation. near the western end of Gunflint Samplelocations are shown on Figure l, and the Lake. All assemblagescontain ilmenite or sphene mineral assemblagesin analyzed samplesare given and commonly pyrrhotite which is locally altered to in Table l. Rocks that exhibit only incipient recrys- marcasite. The chlorite-free assemblagequartz * tallization and retain detrital textures are termed muscovite+ biotite * cordierite + plagioclaseoccurs "low grade." Rocks that have recrystallizedto fine- in the Rove adjacent to the sill contact at South grained hornfels are termed "medium grade," and Round Lake. thosethat showextensive interaction with the Duluth Figure 2 illustrates the compositionsof minerals Complex are called "high grade." that occur in muscovite + quartz-bearing assem- The assemblagesin quartz-bearing rocks can be blagesfrom low-graderocks of the Rove Formation. representedin the system KrO-FeO-MgO-AlrOr- Chlorite + biotite coexistover a rangein Mgl(Mg + HrO. HrO is assumedto be a boundary-valuecom- Fe) from 0.65 to 0.50, biotite is slightly more iron- ponent, and becauseeither muscovite + quartz or rich than coexistingchlorite, and microcline coexists microcline + quartz assemblagesoccur, mineral with magnesium-richchlorite. The low-gradesample compositionsmay be projectedonto the plane FeO- 4le contains cordierite + biotite and indicates that MgO-AlrO, from either of thesepoints to show the near the sill chlorite has broken down at the compo- effectsof Fe-Mg partitioning. Theseprojections may sition Mg/(Mg + Fe) : 0.50. LABOTKA ET AL: ROVE FORMATION. MINNESOTA 73

Table l. Analyzed samples,Rove Formation

Sample Quartz Muscovite Microcline Biotite Chlorite Cordierite Amphibole Hypersthene Al2SiO5l Plagioclase Others2

LOV GRADE

Cross River 3X X X i,g 43d X XX x i,t 43t X X X ep,sph

South Rounci Lake 4Ie X x X X i,s 4ls X x X rrs

Daniels Lake 42c X

MEDIUM GRADE

Cunllint Trail 22a X XX Xi 23 X cpx,sph,s ?6X XX X grirs

Mayhew Lake 35a X X X lrS 35t X X XX X X lrS 35q X x XX X X I's

Loon Lake 39a X X X X X lrS 396 X X X X X lrs )9c X X x X X X 39d X X X X X x lrs 39\ X X X I's 39m X X 39r X X x I's

Cunllint Palisades 8rs

HIGH GRADE

Kakakebic Trail l6a X Y X 8,r l5b XX X X lr8 l6i X X X X trSrS t6j X XX X X r,g l8a X XX X l8b X x X X l, s,cPX l8c X cor,gah,r,s l8e x X

lAnd = Andalusite, Sil = Sillimanite 2i=ilmenite,r:rutilersph=sphene,g=graphite,s=sulphide,cpx=diopside,cor=cordundum,gah=gahnite,eP=ePidote

Medium grade At both Loon Lake and Mayhew Lake, closely At the localitiesof Mayhew Lake, Loon Lake, and spacedsequences of samplestaken progressivelyfar- Gunflint Trail adjacent to the Duluth Complex, ther from the contact show that the grain sizes of rocks from the Rove Formation are completely re- muscovite and biotite decreasefrom I mm to less crystallizedto fine-grainedhornfels. All aspectsof a than 0.25 mm within a distance of 20 m from the sedimentaryprotolith, exceptcompositional layering, contact. havebeen obliterated. The most common mineral as- Cordierite occursas ellipsoidal poikiloblasts, -0.5 semblageis quartz * muscovite + microcline * bio- mm in diameter,that contain numerousquaftz and tite + cordierite + plagioclase.The grain sizesof most feldsparinclusions. In many samplescordierite is al- mineralsrange from <0.10 mm to 0.25mm, but mus- tered to a yellowish chlorite. Plagioclase,microcline, covite occursas porphyroblastsup to I mm in size. and quartz are fine-grained(-0.1 mm), equigranular, LABOTKA ET AL; ROVE FORMATION. MINNESOTA

AlzSiOs and cannot easily be distinguished in thin section from one another. Pyrrhotite and ilmenite are additional minerals in all rocks, and graphite is locally present. Becausemicrocline occurs in most assemblages, the compositionsof coexisting minerals are representedin Figure 3 in a projection from KAISi.O, and SiOr. Muscovite compositions contain minor amountsof Mg and Fe2* (Table 2), but the range is too small to show on Figure 3; thus, muscovite is plotted at the apex of the triangle. The assemblage muscovite + cordierite + biotite occurs in the two suitesof samplescollected adjacent to sills at May- hew Lake (35) and Loon Lake (39). Samples col- lectedadjacent to the sills are labeled a, and samples collectedprogressively farther from the sills are labeled b, c, and so on. Samplesfrom the Loon Lake locality are coarser-grainedand closer to the high- Microc/rne grade region at Kakakebic Trail than the samples Fig. 2. Compositions of coexisting minerals in low-grade from the Mayhew Lake locality. muscovite + quartz-bearing rocks. The compositionsof biotite and cordierite in the

Andolusite Muscovite

Antiteoarho-l$t -t--f--' roo_ _ _ !_ Jt- /s Mso Microc/ine Amphibole Fig. 3. Compositionsof coexistingminerals in medium-grade microcline * quartz-bearing rocks. Distribution of Na and K in muscovite+ microcline + andalusiteassemblage 48f also shown.Points labeled by letter only are samples39. Biotite + cummingtonite assemblagesdo not contain microcline. I"./IBOTKA ET AL: ROVE iORMATION, MINNESOTA 75

Table 2. Representativemineral compositions

MUSCOVtTE BIOTITE CHLORITE CORDIERITE

35t )5q J9a 39r 4EI l6b 351 359 )9a 19r 48f 3 t5b J5t 359 39a 39t 481

SiO2 46.4 47.t 46.0 47.7 47.2 47.1 35.8 37.2 35.5 )4.7 3E.I 36,1 41.2 26.3 49.6 4t.E 47.9 49.3 4E.E J0.2 Tio2 0. I 0. I 0.0 0.5 o.3 0.5 1,5 5,O 3.8 4,0 3.E 2.3 0.8 0.4 0.0 0,0 0.I 0,0 0.0 0.0 AI2O3 )5.4 36,7 34.8 35.6 37.6 34.9 t9.3 15.2 18.5 t9.0 17.9 19.9 15.9 22.5 )4.2 32.7 )4.6 32.9 )4.1 35,1 MgO. 0.2 0,6 2.4 0.7 1.0 0.8 9.7 lJ,2 6.7 7.4 9.4 9.5 23.1 t6.3 9.6 6.6 6.6 7.3 7.2 12.3 FeOr 2.2 0.9 3.O 0.8 0.7 0.1 17.l t6.4 22.5 2l.l 19.l 19.2 2.8 22.1 6.8 ll.3 10.9 10.0 10.2 2.2 MnO 0.0 0.0 0.0 0.0 0,0 0.0 0,t 0.1 0.0 0.0 0.0 0.0 0.0 0.2 0.0 0.0 0.1 0.1 0.1 0.1 CaO 0,0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0,0 0,0 0.I 0.0 0.0 0.0 0.0 0,0 Na2O 0.) 0.3 0.2 0.6 O.2 0.3 0.0 0.I 0.0 0.0 0.0 0.0 0.0 0.1 0.1 o.2 0.0 0.1 0,1 0.0 K2o^ 9.E 10.I 9.4 10.3 E.4 10.7 9.2 9.3 9.9 9.5 9.1 9.4 9.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 H2Ot 5.6 3.5 4.2 3.E 4.6 5.6 7.3 3.5 1.O 4.J 2.6 3.6 6.J ll.9 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 t00.4 99.6 t00.2 99.7 100.5 99,9

FORMULA PROPORTIONS3

si 3.u 3.10 3,05 3.ti 3,07 i.t4 2.75 2,76 2.69 2.65 2.79 2.69 2,92 2.7| 4.97 5.00 4.90 5,04 4.95 4.94 Al 2.79 2.81 2.72 2.75 2.89 2.74 1.75 1.13 1.66 t.7t 1.55 t.75 t.31 2.7) 4.01 3.96 4. t7 1.97 4.0t 4.05 Ti 0.01 0,01 0.00 0.02 0.01 0.03 0.09 0.28 0.22 0.23 0.21 0.13 0.04 0.03 0.00 0.00 0.00 0.00 0.00 0.00 MB 0.02 0.06 0.24 0,o7 0.09 0,08 t.t2 1.46 0.75 0.E4 1,02 t.05 2.44 2.50 1.43 l.0t l.0t l. I I 1.09 1.80 Fe O.l2 0.O5 0. l7 0.04 0.04 0,01 l.l0.l.0l t.4J t.35 Lt7 1.20 0.17 l.9l o.57 0.97 0.93 0.85 0.87 0. lt Mn 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.01 Ca 0,00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0r 0.00 0.00 0.00 0.00 0.00 Na 0.04 0.04 o,o2 0.07 0,03 0.04 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0,04 0.00 0.01 0.02 0,00 K 0.E3 0.84 0.80 0.t6 0.70 0.91 0.90 0.t7 0.96 0.93 0.85 0.89 0.90 0.00 0.00 0.00 0.00 0.00 0.00 0.00 assemblagequartz * microcline + muscovite + cor- for Mayhew Lake samplesin which biotite has an dierite + biotite range from relatively Mg-rich com- Mgl(Mg + Fe) ratio of 0.35 and cordierite a ratio of positionsfor Loon Lake samples,in which biotite has about 0.50.Because this assemblagequartz * micro- an Mg/(Mg * Fe) ratio of about 0.50 and cordierite cline + muscovite+ cordierite + biotite is invariant a ratio of about 0.65, to more Fe-rich compositions under constaat T, P, and &u,o conditions, the ob-

Table 2. (continued)

HYPERSTHENE MICROCLINE PLAGIOCLASE

Samplel l6b l6b )5t 35q )9a 39r 4Ef l6b 35f 35q )9a

SiO2 5t.2 65.1 66.8 64.E 65.8 65.9 65.5 68.7 60.9 66.7 62.0 67.6 69.9 Ti02 n) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.I 0.0 0.0 0.0 Al2o3 3.4 19.3 18.8 19.8 18.7 19.7 t8.7 l9.l 25.8 2t.5 25.5 22.6 21.0 Meo 18.0 0.0 0.0 0.0 0.0 0.0 0.5 0.0 0.0 0.0 0.0 0.0 0.0 Fe-o2 26.7 0.2 0.2 0.2 0.2 0.4 0.0 0.2 0.2 0.3 0.2 0.I 0.2 MnO 0.J 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 CaO o.2 0.2 0.0 0.0 0.0 0.0 0.1 0.3 7.3 2.9 5.5 2.7 0.7 Na20 0.0 1.7 2.t 1.9 1.8 t.J 1.8 ll.5 7.4 10.8 8.8 10.9 t0.7 KeO 0.0 14.0 14.2 14.I l3.E 14.8 t3.2 0.1 0.2 0.3 0.2 0.2 0.2 ejo3 Total aa( 100.7102. I 100.8101.3 t02.1 99.8 99.9 I01.8 102.6102.2 104.1 102.7

FORMULA PROPORTIONS'

Si t.94 2.97 3.00 2.95 3.01 ) q7 ? nl 3.00 2.67 2.E8 2.70 2.86 2.97 AI 0.15 1.04 0.99 l 06 0.99 1.04 l.0l 0.99 t.13 1.09 1.3t l.13 1.05 T,i 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Mg 1.02 0.00 0.00 0.00 0.00 0.00 0,03 0.00 0.00 0.00 0.00 0.00 0.00 Fe 0.83 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.00 Mn 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0l 0.0I 0.00 0.00 0.00 0.00 0.00 0.01 0.34 0.t4 0.26 0.t2 0.0) Na 0.00 0.l5 0.18 0.17 0.16 0.12 0.16 0.98 0.62 0.90 0.74 0.89 0.88 K 0.00 0.81 0.82 0.E2 0.79 0.85 0.78 0.00 0.01 0.02 0.01 0.01 0.01

I all Fe as FeO 2 tl2o by difference r based on total positive charges of 12 (hypersthene), 15 (microcline and plagioclase), 22 (muscovite and biotite), 24 (chlorite), and 36 (cordierite) L/IBOTKA ET AL.: ROVE FORMAT'ION, MINNESOTA servedrange in mineral compositionsindicates that between coexisting muscovite and microcline (Fig. the rocks containing this assemblageequilibrated 4). Microcline is more Na-rich than coexistingmus- over a rangein metamorphicgrade. The positions of covite, but the partition coefficient Ko is not con- the three-phasefields on the AFM diagram in Figure stant. Ko is a function of temperatureand composi- 3 indicatethe relativegrades of metamorphismunder tion, but becausethe range in Na/K for the which the various muscovite + cordierite + biotite coexistingphases is limited to K-rich compositions, assemblagesformed. Displacement of the field is Ko is essentiallya function of temperature.Because causedby the continuousdehydration reaction Ko approaches1.0 as temperatureincreases, samples from Mayhew Lake equilibrated at lower temper- muscovite+ biotite + quartz : cordierite atures than the samplesfrom Loon Lake. On the + microcline + HrO basis of compositionsof coexisting cordierite and An increasein 7 or a decreasein ,rH,odrives this re- biotite, the lower-temperatureMayhew Lake samples action to the right, and becausebiotite is more iron- formed at higher grade than the higher-temperature rich than coexistingcordierite, the three-phasefield Loon Lake samples;thus variationsin fluid composi- cordierite* biotite * muscoviteis displacedto more tion stronglyinfluenced the equilibrium muscovite+ iron-rich compositions.Hence, the samplesfrom phlogopite + quartz : cordierite * microcline * Mayhew Lake (35f,q) formed under higher-grade HrO. conditionsthan the samplesfrom Loon Lake (39a-r). The only andalusite-bearingrock (sample48f) was The relative differencein grade betweenthe Loon collectedfrom the Gunflint Palisadeslocality (Fig. l) Lake and Mayhew Lake localities is reversedfrom and containsthe assemblagequartz + andalusite + the anticipated difference based on the grain sizes cordierite + biotite * muscovite + microcline + il- and the distance from the high-grade Kakakebic menite + graphite. Andalusite occurs as relatively Trail locality. An independentestimate of the rela- large (l-2 mm) crystals with chiastolite habit in a tive grade is basedon the partitioning of Na and K graphite-richmatrix. The assemblageis unusual be-

K AlSisOe

A B

o ; o.2 o o 6 =J \< o : O.l

NoAlSisOe o.l o.? Plogioclasc CoAlzSizOe) (No/K)Microcrnc

Fig. a. (A) Compositionsof coexistingfeldspars, and (B) distribution of Na and K betweencoexisting microcline and muscovite. L./IBOTKA ET AL.: ROVE FORMATION. MINNESOTA causethe compositionof biotite is very Mg-rich, and Guidotti et al. (1975a)find no dependenceof Ko on becausethe coexistingcordierite is more Fe-rich than composition for coexisting cordierite + biotite + biotite. The compositionsof theseminerals are given muscovite (no microcline) in very magnesium-rich in Table 2 and shown on Figure 3. This reversal in pelitic schists,and Guidotti et al. (1975b)also find no Fe-Mg partitioning suggeststhat the assemblage relationbetween'IAl and Mgl(Mg + Fe) in biotite. may not have been in equilibrium, but there is no Figure 5b showsan inversecorrelation betweenMg/ textural evidencefor disequilibrium, and the compo- (Mg + Fe) and '"Al which is a gaugeof the amount sitional relationshipbetween cordierite and biotite is of substitution of cations other than Fe and Mg in constantover the area of a thin section.Mg-rich as- biotite. Herein lies the reasonfor the compositional semblagesin pelitic schistsare not well documented dependenceof Ko. If the Al content of biotite is con- in the literature for comparison,and the assemblage stant, then the partition coefficient may be directly in sample48f is tentatively acceptedas an equilib- related to the equilibrium constantfor the exchange rium assemblage.Because plagioclase is absent,this reaction assemblageindicates that in the calcium-freesystem aMg-biotite + : aFe-biotite the reaction muscovite+ albite * quartz : andalu- BFe-cordierite + site + microcline * HrO has occurred. In the me- BMg-cordierite dium-grade rocks at Loon Lake muscovite coexists where a and B are the numbersof exchangeablecat- with plagioclaseas sodicas An,r; thereforeandalusite ions in cordieriteand biotite, respectively.The equi- and microclinecoexist only in very Na-rich, Ca-poor librium constant is (a."-o,",,,")'(a*"-.",u)B / (a*s-biotit )"(aF"- rocks. ."d)8. If the Fe-Mg solutions are ideal, then the equi- The partitioning of Fe and Mg betweencoexisting librium constant equals (Ko)"u (Albee, 1972) and K" cordierite and biotite in medium-grade and high- is independent of composition. If, however, some of grade samplesis illustrated in Figure 5a. In all sam- the Fe in biotite is associated with octahedral Al, ples exceptnumber 48f biotite is more Fe-rich than then this Fe cannot exchange with Mg in cordierite, coexistingcordierite, but the magnitude of the distri- and Ko is not related in a simple way to the equilib- bution coefficientseems to depend on composition. rium constant for the exchange reaction. Instead of

B. t. o o. .9 (D E u- r6b o a C) + o. ctl *"&':: LL = o. a a ql c'l 35f q o.? oQ'l t.o t.2 1.3 t.4 t.5 rvA Iog (ttttgZ Fe ) g;ot ite I Fig. 5. (A) Distribution of Fe and Mg betweencoexisting cordierite and biotite in medium-gradeand high-graderocks. Cordierite is generallymore Mg-rich than biotite exccptfor the Mg-rich sample48f. The lines Ko : 1.00and Ko : 1.78are shownfor reference.(B) Inversecorrelation between Mg/(Mg + Fe) and IVAI content in biotite. LABOTKA ET AL: ROVE FORMATION. MINNESOTA two componentsrelated by the substitutionof Fe for Mg, biotite from the Rove Formation contains additional componentsthat relate an increasein Al and Ti to an increasein Fe. In this case,XBl" includes Fe from both the annite and the other components,and K, will be different from the equilibrium constant for the exchangereaction in which only the mole fraction of annite is considered. The assemblagequaftz + biotite * cummingtonite * plagioclaseoccurs in two samplesfrom Loon Lake (39) and one from Mayhew Lake (35). The amphibole is very fine-grainedand is intergrown with the quaftz + plagioclasematrix. Microcline does not coexistwith cummingtonite,but the compositionsof coexistingbiotite and amphibole are shown on Fig- - .- ure 4 to illustrate the partitioning of Fe and Mg be- FeO,/ .< \ Mqo tweenthem. Hyperslhene Fig. 6. Distribution of Fe, Mg, and Al amongminerals in high- High grade graderocks. Samples l8c and l6b do not contain quartz. Samplcs l6i not microcline. Rocks from the Kakakebic Trail show the greatest l6a and do contain degreeof recrystallization,and some show extensive interaction with the Duluth Complex during em- and suggestthat the silica-rich country rock has re- placement.Here the Duluth Complex intruded the acted with the silica-poor Duluth Complex. Assem- lower, carbonaceouspart of the Rove Formation so blagesin two of the analyzedrocks are cordierite + that graphite has recrystallizgdto 0.25 to 0.50 mm biotite * hypersthene* K-feldspar + plagioclase+ grains. Cordierite forms spherical porphyroblasts itnenite + graphite + pyrrhotite (l6b) and cordierite with no inclusionsand unusualtwin patterns.Plagio- + sillimanite * corundum + K-feldspar * Zn spinel clase grains are -0.25 mm in diameter and have (gahnite) (l8c). The compositionsof these quartz- polygonal forms. free assemblagesare also shown in Figure 6. Pelitic Severalsamples show a variety of textures within schistshave interacted with the gabbro through the small areasand contain unusual assemblagesthat in- reactions dicateassimilation and reactionof the Rove with the gabbro.There are regionsin sample l6b that contain (a) biotite + SiO, : cordierite + hypersthene tabular, igneousplagioclase, and adjacentregions in + K-feldspar + HrO which the plagioclaseforms a polygonal aggregate. (b) muscovite* quartz : sillimanite The shapesof hypersthenegrains range from sub- + K-feldspar + HrO hedral with few inclusionsto very irregular poikiloblasts. and Pelitic rocks near the contact with the Duluth (c) muscovite: corundum * K-feldspar * HrO. Complex contain the muscovite-freeassemblage q\aftz + biotite * cordierite + K-feldspar + plagio- The activity of SiO, in sample 18c was buffered by clase+ rutile + graphite * pyrrhotite. The composi- the reaction tions of coexistingcordierite * biotite in K-feldspar- sillimanite : corundum + SiOr. bearing rocks are shown in Figure 6. Most of the Rove Formation that crops out on In sample l6b the ratio f ,,o/a"io, was determinedby Kakakebic Trail occurs as inclusions in the contact the compositionsof coexistingbiotite + cordierite + zone with the gabbro. The assemblagein these hy- hypersthene+ K-feldspar through reaction (a). brid rocks is quartz + biotite * K-feldspar + hypersthene+ plagioclase* ilmenite + graphite + pyrrho- AKFM facies series tite. The compositionsof coexisting biotite and In well-documentedstudies of contact- or low- hyperstheneare shown on Figure 6. Many pelitic pressure regional metamorphism of pelitic schists rocks adjacentto the gabbro are deficient in quartz (SantaRosa, Nevada, Compton, 1960;Steinach, Ba- LABOTKA ET AL.: ROVE FORMATION, MINNESOTA varia, Okrusch,1969; Alsace, Bosma, 1964; Buchan, reactionsif reaction I goesto completionover the en- Read, 1952;Abukuma, Miyashiro, 1958)biotite + tire rangein Mgl(Mg * Fe) beforereaction 4 occurs. andalusiteis a very common assemblage,and La- Figure 7 predicts a variety of assemblagesthat botka (1978)demonstrated for the Panamint Moun- ought to have formed in rocks with compositionson tains that rockswith Mgl(Mg + Fe) in the range 0.35 the plane chlorite * muscovite* quartz but are not to 0.56 contain andalusite.The rarity of andalusite- observedtvrz. chlorite * cordierite * microcline and bearingassemblages in the Rove Formation was puz- chlorite + cordierite + biotite. Low-grade assem- zling in light of thesestudies. However, in the Rove blages in pelitic rocks include chlorite, chlorite * Formation muscovite+ quartz + cordierite is stable biotite, or rarely chlorite * biotite * microcline, and over a large range in Mg/(Mg + Fe) from 0.09 to thesereact to chlorite-freecordierite * biotite assem- 0.50,and the biotite that coexistswith the most iron- blages with no apparent intermediate steps. Either rich cordieritehas Mg/(Mg + Fe) : 0.35.Microcline the intermediatereactions that limit the stability of also coexistswith this biotite * cordierite + mus- chlorite occur over a very narrow temperatureinter- covite assemblage,and becausethe four minerals are val, or the intermediatereactions were overstepped relatedby the continuousreaction becauseof the rapid temperature rise during emplacement of the gabbro complex. The total mass muscovite+ biotite * quartz : cordierite + transferrecorded in theserocks is the sum of the in- microcline + HrO termediatereactions: chlorite + muscovite f quartz muscovitecannot coexist with magnesium-richbio- --+ biotite + cordierite * microcline + HrO. The co- tite. Thus, andalusitein pelitic schistsis restrictedto efficientsin this mass-balancerelation dependon the iron-rich compositions. These phase relations are bulk-rock composition. similar to thosereported by Tilley (1924)for the contact metamorphismof pelitic schistsin the Comrie IntensiveParameters area,Scotland. Pre s sur e and temperature The facies series for progressivecontact metamorphismof the Rove Formation is shown in Figure Pressure,temperature, and activities of volatile 7a-b. This seriesprogresses from the low-grade as- componentsmay be calculated from the composi- semblagemuscovite + quartz + chlorite + biotite + tions of coexistingphases in the assemblagequartz * microcline to the medium-grade assemblagemus- biotite * muscovite+ cordierite + microcline * il- covite + quartz + biotite * cordierite * microcline. menite + graphite * pyrrhotite which occurs at the The reactionsequence is basedon the observedorder Gunflint Palisades,Loon Lake, and Mayhew Lake of increasingFe content in coexistingminerals: cor- localities. Pressureestimates can be made from the dierite, chlorite, biotite. The unusual Mg-rich assem- estimated thickness of overburden at the time of blage in sample 48f is neglected.The order of the metamorphism.The Duluth Complex was emplaced generalizedreactions assumesthat H'O is evolved near the surface,and the aggregatestratigraphic with increasinggrade: thicknessof the gabbroic rocks and overlying vol- canicrocks near Lake Superiorwas 15krrr. However, (l) chlorite + microcline : biotite * muscovite * Weiblen and Morey (personalcommunication, 1979) qaartz + HrO infer that both the plutonic and extrusiverocks thin (2) + andalusite + quartz : cordierite + chlorite northward to near zero thicknessin the vicinity of HrO Gunflint Lake. Simmons et al. (1974) estimated a (3) muscovite + chlorite + quartz : andalusite + minimum pressureof 2.0 kbar during metamorphism + H,O biotite of the underlying Gunflint Iron Formation, a pres- (4) chlorite * muscovite* quartz: cordierite + mi- surethat is basedon the compositionof iron-rich pi- crocline+ HrO geoniteproduced during metamorphism. (5) * microcline + quartz : cordierite * chlorite A calcsilicaterock (sample23) which occursat the biotite * muscovite+ HrO Gunflint Trail locality containsthe assemblagediop- (6) chlorite + muscovitel quartz: cordierite + an- side + anorthite * sphene+ sphalerite + pyrite + dalusite*biotite+HrO chalcopyrite,and an additional estimate of the at- (7) biotite * muscovite* quartz : cordierite + mi- tending pressuremay be basedon the FeS content of + HrO crocline sphalerite.Scott (1973,1975) calibrated the composi- Figure 7c illustrates a slightly different sequenceof tion of sphaleritethat coexistswith pyrite and pyr- 80 LABOTKA ET AL: ROVE FORMATION. MINNESOTA

A atzsios

T2 T5

T. Micr T5 T2 T,

Fc-Chloritc Mg-Chlorilc

Micr

Fc-Chloritc O.5 Me-Chlorilr

F Bi

Micr Fig. 7' Generalizedfacies series for the low-pressure,contact metamorphism of the Rove Formation. (A) Faciestypes for muscovite + quartz-bearingrocks. Low-grade rocks are representedby Tr, (although the three-phasefield biotite + chlorite + microclineoccurs at more Mg-rich compositions)and medium-graderocks are representedby T5. High-graderocks are not represented.(B) PseudobinaryZ- Xdiagram on the join chlorite + muscovite+ qtrafiz for the faciesseries shown in A. (C) T-X diagramfor an alternatefacies series in which the reactionchlorite + microcline : biotite + muscovite+ quartz + H2O goesto completion over the entire rangein Mgl(Mg + Fe) beforechlorite beginsto break down to cordierite + biotite. rhotite for pressures(independent of temperatureex- sure. The composition is ZnorrFeorrSand indicates cept at high temperatures)down to 2.5 kbar. that the maximum pressurewas less than 2.5 kbar. Pyrrhotite does not occur in sample 23, so that the Metamorphismof the Rove Formation probably oc- FeS content in sphaleriteindicates a maximum pres- curred at a pressureof 1.5 kbar. This value is the LABOTKA ET AL: ROVE FORMATION, MINNESOTA 8l minimum pressuresuggested by Simmons et al. and the calibration of Powell and Powell (1977) has (1974)minus 0.5 kbar, becauseof the higher strati- been used for severallocalities that contain assem- graphicposition of the Loon and Mayhew Lake sam- blages with coexisting feldspars (Table 3). These ples relative to the Gunflint Iron Formation. feldspar temperaturesare generally consistentwith Temperaturesmay be calculatedfrom the compo- metamorphicgrade, but their absolute values seem sitions of coexistingplagioclase and alkali feldspar, low in comparisonwith estimatesfrom the stability

Table 3 Temperaturescalculated from coexistingK-Na-Ca phases

SA[,IPLE PLACIOCI,ASE r\TICROCLINE N4I.lSCOVITE Kil Tz Tt

Xab Xun Xo. Xab X"n Xo. X,.nrsc Xp.. o- o-

l6b 0.540 0.350 0.0I0 0.156 0.008 0.836 )q/ 16j 0.550 0.340 0.010 0.t80 0.010 0.810 566 18a 0.752 0.2)9 0.009 0.147 0.001 0.857 490

lza 0.897 0.096 0.007 0.t9) 0.000 0.E07 497 26 0.837 0.151 0.012 0.261 0.0l t 0.728 35f 0.857 0.t29 0.0r4 0.r82 0.001 0.817 0.954 0.046 4.61 )TJ 548r10 )5q 0.733 0.253 0.014 0.169 0.00I 0.E30 0.976 0.024 8.27 555 484t10 39a 0.870 0.119 0.011 0.158 0.001 0.831 0.924 0.076 2.45 482 620+5 )9t 0.\42 0.000 0.ii58 0.r+() 0.060 2.59 6Ii+7 39c 0.128 0.000 0.872 0.94t 0.059 627+7 '1 39d 0.150 0.000 0.850 0.955 0.045 7L 568t I 0 39r 0.954 0.033 0.0I3 0.I 20 0.000 0.8E0 0.961 0.0J9 3.58 4r8 570+ I0 4)d 0.900 0.072 0.028 0.081 0.002 0.917 379 48f 0.115 0.000 0.884 0.971 0.029 l' 20 544+12

_ mlca mlcr. I x grylslq r L) mica micr (Xpar Xor )

2 Temperature calcu.latedfrom coexisting feldspars (Powell and Powell, 1977).

micr- 2 micr -(xo. )- l$30 * 0.093p+ 2x;t- (1340 + o.0lep) I 2 ,.-Jnicrl^*P^= micr micr Rrn . (*o. ) (-\.et + 1.54X"6 ) I| ptag I Ixau ) 3 Temperature calculated from coexisting muscovite and microcline (Thompson, 1974).

Arrx/RT - 6.489q - Gtt9.54lT')

Au x - (ur6156-1lp3;) - (uor-ps6) = RTlnKp + (W6*r-W6oy) + 2(W6'a-2wcms)X$fs%

+ 3(w6,'.-w6p"Xxfij.%)2 - 2(w636-2w6o.)X#.. - 3(w5,61-v/6abxxTicr)2

W6"6 = 6326.7 + 0.0925p - !.9??]I Waldbaum and Thompson (1969) Vcor = 7671,8 + O.ll2lP - 3.85657

VtcDa = 3082.1 + 0.0822p + 0.1598T Eu8ster et al' (1972) Wcms = 416).9 + 0.1259p + 0.39547

Errors represent a +10% error in the wt.% K2O in muscovlte

P = 1500 bars LABOTKA ET AL.: ROVE FORMATION. MINNESOTA of pigeonite in the iron-formation: 547oC from cline temperatures,and are closer to the high-grade samplel6b at the gabbro contactbased on coexisting regionthan samplesfrom Mayhew Lake (35), but the feldspars,and 800"C from Simmons er al. (1974) compositionsof coexistingbiotite and cordierite in basedon pigeonitestability. thesesamples indicate that the Mayhew Lake sam- Thompson (1974) calculated the Na-K exchange plesrecrystallized at 3 "higher grade." Thus, regional potential,Ap, for the reaction muscovite + albite : variation in the compositionof the fluid phasemust paragonite* orthoclase,and temperaturesmay also have affectedthe relation be calculated from the compositions of coexisting muscovite+ biotite * quartz : microcline * muscovite and microcline. He calculated two ex- cordierite + HrO. pressionsfor Ap. The first is derived from end-mem- ber dehydration curves, but predicts unrealistically Portions of the systemMgO-AlO,-KrO-SiOr-H,O low temperatures(--25oC) for the metamorphism have been experimentallyinvestigated by Schreyer, of the Rove Formation. The secondexchange poten- Yoder, Seifert, and others, and the data have been tial is derived from ion-exchangeexperiments and summarizedby Schreyerand Seifert(1969). Schreyer predicts temperaturesthat are similar to but gener- and Seifert(1967) reported the resultsof experiments ally higher than temperaturescalculated from co- for the reaction muscovite+ phlogopite + quartz : existing feldspars(Table 3). The muscovite-micro- cordierite * microcline + HrO which seemsto have cline temperaturesare consistentwith the geologic occurredat I kbar,550o+l5oc, 2kbar,600o+l5oC, settingof the samples.Sample 35f which is adjacent 3 kbar, 670o+15o,4 kbal 690otl5o, and 5 kbar, to the contact of a large sill gives a temperature of 7lOotlOoc (P",o : P'). Thesedata were fitted to a 548oC,and sample35q which"is5 m from the contact relation0 : (A/T) + B + (CP/T) * h.f",o, in which gives a temperatureof 484'C. Sample 39a which is C is the volume changeof solid phasesdivided by the very coarse-grainedgives a temperature of 620"C; gas constant.The least-squaresresult is this temperatureis greaterthan that from sample35f 0 : (1.126x 104T) - 19.96+ (0.128P/T) * ln./",o which is finer-grained and farther from the high- / grade area near Kakakebic Trail. The total varianc€, d: 2(predicted value - mea- The errors of +l0oC in the muscovite-microcline sured value)'z/(degreesof freedom), is 358, which temperatureare calculatedfrom a generoustl}Vo er- predicts temperatureswithin the error estimatedfor ror in the wt.VoKrO in muscovite.Realistic errors in the location of the reaction.This relation can then be the temperatureestimates are difficult to evaluatebe- extendedto include the effectsof solid solution: causeof the lack of agreementbetween the two-feld- 0 : (9.004x 104/T)- t59.7 + (1.022p/T) 4 tn IQo spar and muscovite-microclinevalues, and because the other calibration of the muscovite-microclineex- where changepotential by Thompson givesunrealis- Q97g ln I(-o : 8 ln afl" + 3 ln affJ9"*o tic results. The muscovite-microcline temperatures + 8 tn - 6 ln aHl,T"- 2 ln ar-'ii"" for samples39a,b,c are very closeto the dehydration /H,o of muscovite + quartz, which suggestthat they are for the stoichiometry too high and that the associatederrors must be 6 muscovite* 2 phlogopite + 15 quartz greater than those derived from analytical uncer- : 8 orthoclase* 3 cordierite + 8 HrO tainty. The errors in temperaturescalculated from the two-feldspar calibration of Stormer (1975) and The fugacity of HrO for a given temperature and Powell and Powell (1977)are probably also large be- pressuremay be determinedfrom the compositions causethe thermometeris relatively insensitiveto the of coexistingphases in the assemblagemuscovite * albite contentin plagioclaseat temperaturesless than microcline * cordierite + biotite + qvaftz, provided 600'C (Fig. 3 in Stormer,1975). The actualtemper- that the activity-compositionrelations for the miner- atures during metamorphismmay have been much als are known. lower than the muscovite-microclinevalue, but they The activity-composition relations of Waldbaum were probably higher than the two-feldsparvalue. and Thompson (1969) for alkali feldspars and of Eugsteret al. (1972)for white micas have been used, Volatile-p hase c omp o sition and cordieriteis consideredto behaveideally (4r"re_"o.a The medium-gradesamples from Loon Lake (39) : X""). The inversecorrelation of Mg and ''Al con- are coarse-grained,record higher muscovite-micro- tents in biotite from the Rove Formation (Fig. 5b) LIIBOTKA ET AL.: ROVE FORMATION, MINNESOTA 83

Table 4. /n,o in medium-gradeRove Formation

SA[4PLE TI CORDIERITE BIOTITE MLISCOVITE2 MICROCLINE2 fi 14 HZO HZO

mlca feld VInt VI- Xrr- RTInY RTInY bars rc XM8 Ti u Mg rv,5 musc mlcr

35f. 54E n 5l 0.22 0.)5 0.25 0.75 | .43 0.25 4.52 45.82 252 35q 484 .0.52 0.23 0.)6 0.22 0.84 t.J5'0.28 1.20 47.42 83 39a 620 0.56 0.20 0.30 0.30 1.oi l.17 0.34 t2.78 26.99 903 39b 6t3 0.58 0.09 0.44 0.l4 l. 19 r.14 0.40 7.82 L7.43 tt26 39c 627 0.56 0.20 0..36 0.26 1.02 l.16 0.14 /.)) 12.)6 954 )9d 568 0.50 0.t7 0.43 0.25 | .02 l. l1 0.14 4.32 20.54 422 )9r 570 0.56 0.13 0.44 0.18 1.05 | .20 0.15 t. zz 16.89 423 48f 544 0.9l 0.04 0.25 0.10 2.44 0.t7 0.81 t.76 15.)2 379

I Temperature derived from coexisting muscovite and microcline, Table 3. 2 RTlnYl = (l-Xl)2(Wcl+2(W6;2-W6 1)X1). X and W6 for muscovite and microcline are given in Table 3. ,tnr6calculatedfromtheequilibrium6muscovite+2phlogopite+lSquartz=3cordierite+8microcline+8H2O.

U fHzO calculated from the equilibrium muscdvite + quartz = andalusite,+ microcline + H2O (Chatterjee and Froese, 1975). o = 1500 bars.

suggeststhat all Mg occurs as the phlogopite com- and Froese,1975) for sample48f is also presentedto ponent. The activity of phlogopite in biotite is as- show the relative correspondencebetween the two sumedto equal the cube of the mole fraction of Mg equilibria. Figure 8 showsthat the calculatedvalues in the octahedralsite. of fr,o dependvery strongly on temperature,but not The fugacitiesof HrO calculated from the above on pressure.For example,an error in 7 of + l5'C for expressionfor eight medium-gradesamples are given sample35f resultsin an error of t50 bars in/",o but in Table 4. The temperaturesused in the calculation an error ir P""'oof 500 bars results in an error of only are those derived from the muscovite-microcline +22 bars in fr,o.Figure 8 also showsthat in sample equilibrium. Thesetemperatures are maximum tem- 48f for Pr,o- P"o,,othe biotite-cordierite equilibrium peratures,and the fugacities calculated from them occurs at higher temperaturestlan the muscovite- representmaximum possible values of fr,o during aluminositcate equilibrium. This topology violates metamorphismof thesesamples. Schreinemakers'rules and indicatesthat the relative Samples39a,b,c from Loon Lake give ,fx.o values error between the two equilibria is at least 25oC. close to (or exceeding)the values for pure HrO at Thus, even if the temperature is well known, the 1500bars (Burnham et al.,1969). The samplesfrom value of fu,o at P"ora: 1500 bars is only precise Mayhew Lake (35) and Gunflint Palisades(48f1 give within a range of 100 bars. For example,values of valuesmuch lower than the pure-HrO value. /n,o calculatedfor sample 48f at 544"C are 379 bars Figure 8 illustratesthe rangesof possiblevalues of and 483 bars from the cordierite-biotite and mus- P, T, andfr,o for samples35f, 39a, and 48f. These covite-andalusiteequilibria respectively (Table 4). valuesare determinedby the biotite-cordierite-mus- .The uncertainity is larger at greater rock pressures. covite-microcline'equilibrium,and the temperatures The total uncertainty in the maximum /",o for derived from the muscovite-microcline exchange sample48f resulting from error in the maximum tem- equilibrium are superimposedto show the error in perature and in calibration of the biotite-cordierite /",o resultingfrom a tl5o error. The equilibrium be- equilibrium is *100 bars. tween muscovite-microcline-andalusite(Chatterjee Sample48f contains graphite and if Pa"ia: P"oria LABOTKA ET AL: ROVE FORMATION. MINNESOTA

A 35f B 39o

o o -o (L

a + F

600 700 500 600 T ("C) Fig. 8. The equilibrium muscovite+ biotite + quartz : cordierite + microcline * H2O for samples35f (A), 39a (B), and 48f (C). Muscovite-microclinetemperatur€s are given to show the effectsof an error of tl}Vo on K2O in muscoviteon the calculated/uro. The equilibrium muscovite+ quartz : andalusite+ microcline + H2O for sample 48f is also shown in C to illustrate the relative error betweenthe equilibria for determining/""o. Broken lines in C correspondto the muscovite-andalusiteequilibria. and if C-O-H gas speciesare the only important tite-fayalite buffer are consistentwith the suggestions componentsin the fluid phase,then the composition of Floran and Papike (1978) and Simmons et al. of the fluid phasecan be calculated by the method (1974)for the compositionof the fluid phase during reviewedby Eugster(1977). The resultsfor the maxi- metamorphismof the underlying iron-formation. mum valuesof /",o are given in Table 5. At Po",o: The fluid phase may have communicated with a 1500bars the maximum fraction of HrO in the fluid large volume of rock during contact metamorphism; is 0.65+0.10,and most of the remainder of the fluid indeed,the assumptionthat Po",o: P*,,omay not be is CO,. The compositionof the fluid at 500oCis also valid for low-pressurecontact metamorphism. The given in Table 5 to show that, at lower valuesof /".o, effect of permeability (Po",o< P*,.) on the calcu- the ratio f .ro/f .o, is smaller and /o, is larger than at lation of the fluid composition is to increasa X^,o, higher valuesof _/",o.The maximum value of /o, for f and /",, and to decrease./o,.For example, "to/f "o,, sample48f is on the order of 10-23at 550oC,which is in sample48f, if the fluid pressurewas 1000bars (in- close to the value for the quartz-magnetite-fayalite steadof 1500bars) X',o : 0.69,f .',: 125bars, /.o, buffer. The graphite-freerocks from Mayhew Lake :287 barc,f o,: 4 x l0-'zobars, /", : 9.5 bars,and (samples35) were probably metamorphosedunder /.o: 1.7bars. more oxidizing conditions, and the coexisting fluid consistedalmost entirely of HrO and COr. Samples Table 5. Fluid compositionfor sample48f 39a,b,cfrom Loon Lake appearto have equilibrated with a relatively pure-HrO fluid. Thesesamples were fxzo lcoz Jcsq fco tn2 loz XH2o xcoz collectednear a fault zone that was active during the 3791 786 46 2,8 5.7 1.2 x lo-23 0.54 0.44 emplacementof the Duluth Complex, and the com- 4ut 292 2oo t,7 t2 4J x rc-24 0.76 0. 16 positionsof the fluids percolating through the fault zoo2 t25o 8 1.8 x l0-23 0.31 0.69 zone appear to have been very different from the I Range in the maximum values of I9.6 derived from maximum possible 'local' t" compositions ^ temierature oI 544oC of fluids. z Values calculated for a temperature of 500t to show effect of lower The relatively large amount of CO, in temperature on fluid composition the fluid p = 1500 bars phaseand the prevailing f o,near the quartz-magne- LABOTKA ET AL.: ROVE FORMATION. MINNESOTA 85

Summary encesin the thermal propertiesand permeabilitiesof The contact metamorphic effects of the Duluth the two rock types. Complex on the argillaceousRove Formation are confinedto the region within a few hundred meters Acknowledgments of the contact. In this region the low-grade assem- Researchon the contact metamorphic effects of the Duluth blage chlorite * muscovite+ quartz reactedto form Complex was supportedby NSF grant EAR 7'123150.We thank R. Wintsch, C. Klein, and an anony- the medium-grade C. White for field assistance, assemblagecordierite * biotite + mous refereefor critical reviews,and R. Spencerand L. Koh for microcline.Locally adjacentto the Duluth Complex their technicalhelp on the assemblyof the manuscript. the high-gradeassemblages corundum + microcline and hypersthene* cordierite formed. References The compositionsof the coexistingminerals in the Albee, A. L. (1972)Metamorphism of pelitic schists:reaction rela- medium-gradeassemblage biotite + cordierite + mi- tions of chloritoid and staurolite.Geological Society of America croclineprovide estimatesof the maximum values of Bulletin. 83. 3249-3268. temperature and partial pressure of HrO at 1500 Albee, A. L. and Ray, L. (1970) Correction factors for electron phosphates, bars. These values indicate that medium-grade as- probemicroanalysis of silicates,oxides, carbonates, and sulfates.Analytical Chemistry,42, 1408-1414. semblagesformed in the range 500o to 600'C and Bence,A. E. and Albee, A. L. (1968)Empirical correctionfactors that for most samplesP",o = rhP"orio.The composi- for electronmicroanalysis of silicatesand oxides.Journal of Ge- tions of coexistingcordierite and biotite are sensitive ology, 76, 382403. to variationsin /,,,o. On the basisof texture, distance Bonnichsen,B (1968)General geology and petrologyofthe meta- Dunka River area,Minne- from high-gradearea, and partitioning of Na and K morphosedBiwabik Iron Formation, sota.Ph.D. Thesis,University of Minnesota,Minneapolis, Min- betweenmuscovite and microcline, sample 35f nesota. formed at a lower temperaturethan sample39a, but Bonnichsen,B. (1969)Metamorphic pyroxenes and amphibolesin sample35f containsa more dehydrated,higher-grade the Biwabik Iron Formation, Dunka River area, Minnesota. assemblagethan 39a. MineralogicalSociety of America SpecialPaper, 2, 217-241. (Vos- The contact metamorphic effects of the Duluth Bosma,W. (1974)The postsin the spottedslates of Stiege ges) and Vogtland (Saxony).Geologie en Mijnbouw, 43, 476- Complex on the Rove Formation contrast sharply 489. with the effectson the Gunflint Iron Formation. The Burnham,C. W., Holloway,J. R., and Davis,N. F. (1969)Ther- iron-formation preservesprograde reactions as far as modynamic properties of water to 1,000'C and 10,000 bars. l0 km from the contact.The differencesbetween the GeologicalSociety of America SpecialPaper 132. (1975) study of iron-formation and the Rove Formation are, in part, Chatterjee,N. D. and Froese,E. A thermodynamic the pseudobinaryjoin muscovite-paragonitein the systcm the result of the large number of iron silicatesthat KAlSi3Os-NaAlSi3Os-Al2O3-SiO2-H2O.American Mineral- may occur in rockspoor in aluminum and potassium. ogist,60, 985-993. The pelitic Rove Formation lacks a variety of assem- Compton,R. R. (1960)Contact metamorphism in the Santa Rosa blagesand recordsfew reactionsbecause in low-pres- Range, Nevada. Geological Society of America Bulletin, 71, sure, contact metamorphic aureolesthe assemblage 1383-1416. Eugster,H. P. (1977)Cornpositions and thermodynamicsof meta- biotite * cordierite * microcline is stable over large morphic solutions.In D. G. Frbser, Ed., Thermodynamicsin rangesin Mel(Mg * Fe), in Z, and in -/",o. Geology, p. 183-202. D. Reidel Publishing Company, Dor- The iron-formation appearsto have attained tem- drecht,The Netherlands. peratures as high as 800oC. Medium-grade Rove Eugster,H. P., Albee,A. L., Bence,A. E., Thompson,Jr., J. B., D. R. (1972)The two-phaseregion and excess Formation recordsmaximum temperatures and Waldbaum, of 600oC, mixing propertiesof paragonite-muscovitecrystalline solutions. but no reliable estimatesof metamorphic temper- Journal of Petrology,13, 147-179. ature of the high-graderegion can be made. There is Floran, R. J. and Papike, J. J. (1975) Petrology of the Gunflint no evidence of partial melting of the high-grade Iron Formation Ontario-Minnesota:the low grade rocks. Geo- Rove Formation, but the minimum melting temper- logicalSociety of AmericaBulletin, 86, I159-1190. R. J. and Papike,J. J. (1978)Mineralogy and petrology of of the assemblagequartz plagio- Floran, ature + biotite + the Gunflint Iron Formation, Minnesota-Ontario:correlation of clase* hyperstheneunder the condition P'zo ( P,oria compositionaland assemblagevariations at low to moderate is probablygreater than 800oC(Luth, 1967).Because grade.Journal of Petrology,19,215-288. all other factorsregarding the depth of emplacement French, B. M. (1968) Progressivecontact metamorphismof the and temperatureof the magma at emplacementare Biwabik Iron-Formation,Mesabi Range, Minnesota. Minnesota GeologicalSurvey Bulletin 45. equal, the differencesin temperaturesrecorded by Guidotti, C, V., Cheney,J. T., and Conatore,P. D. (1975a)Co- the iron-formation and Rove Formation at equiva- existingcordierite + biotite + chlorite from the Rumford Quad- lent distancesfrom the contactmust be due to differ- rangle,Maine. Geology, 3, l4'l-I48. 86 LABOTKA ET AL.: ROVE FORMATION. MINNESOTA

Guidotti,C. V., Cheney,J. T., and Conatore,P. D. (1975b)Inter- aluminum silicatesin the systemMgO-AI2O3-SiO2-H2O and relationshipbetween Mg/Fe ratio and octahedralAl content in K2O-MgO-AI2O3-SiO2-H2Oat high pressures.American Jour- biotite. American Mineralogist,60, 849-853. nal of Science.267, 37 l-388. Labotka,T. C. (1978)Geology ofthe TelescopePeak Quadrangle, Scott, S. D. (1973) Experimental calibration of the sphalerite California, and Regional Metamorphism in the Death Valley geobarometer.Economic Geology, 68, 466474. Area, California. Ph.D. Thesis, California Institute of Tech- Scott,S. D. (1975)Sulphide geothemometers and geobarometers. nology.Pasadena, Califomia. (abstr.)International Conferenc€ on Geothermometryand Geo- Luth, W. C. (1967) Studies in the system KAlSiOo-Mg2SiOo- barometry, Pennsylvania State University, University Park, SiOa-H2O: I, inferred phase relations and petrologic appli- Pennsylvania. cations.Journal of Petrology,8,372-416. Simmons,E. C., Lindsley,D. H., and Papike,J. J. (1974)Phase re- Miyashiro, A. (1958)Regional metamorphismin the Gosaisyc- lations and crystallization sequencein a contact-metamor- Takanuki district in the central Abukuma Plateau.Tokyo Uni- phosed rock from the Gunflint Iron Formation, Minnesota. versityFaculty Science Journal, ll,219-272. Journalof Petrology,15, 539-565. Morey, G. B. (1969)The geologyof the Middle PrecambrianRove Stormer, J. C. (1975) A practical two-feldsparth€rmometer. Formation in northeasternMinnesota. Minnesota Geological American Mineralogist,60, 667-6'1 4. SurveySpecial Publication 7. Thompson,A. B. (1974)Calculation of muscovite-paragonite-al- Morey, G. B., Papike,J. J., Smith, R. W., and Weiblen,P. W. kali feldsparphase relations. Contributions to Mineralogy and (1972)Observations on the contactmetamorphism of the Biwa- Petrology.4, l'13-194. bik lron-Formation, East Mesabi district, Minnesota. Geologi- Thompson,J. B. (1957)The graphical analysisof mineral assem- cal Societyof America Memoir, 135,225-264. blagesin pelitic schists.American Mineralogist, 42,842-858. Okrusch, M. (1969) Die Gneishornfelseum Steinachin der Tilley, C. E. (1924)Contact metamorphismin the Comrie area of Oberpfalz.Eine phasenpetrologischeAnalyse. Contributions to the PerthshireHighlands. Quarterly Journal of the Geological Mineralogy and Petrology,22, 32-72. Societyof London,81, 100-112. Powell,M. and Powell, R. (1977)Plagioclase-alkali feldspar geo- Waldbaum,D. R. and Thompson,Jr., J. B. (1969)Mixing proper- thermometryrevisited. Mineralogical Magazine, 41, 253-256. ties of sanidinecrystalline solutions:IV. Phasediagrams from Read, H. H. (1952) Metamorphism and migmatisation in the equationof state.American Mineralogist, 54, 1274-1298. Ythan Valley, Aberdeenshire.Transactions of the Edinburgh Weiblen,P. W., Mathez,E. A., and Morey, G. B. (1972)Logan in- GeologicalSociety, 15, 265-2'79. trusions.In P. D. Simsand G. B. Morey, Eds.,Geology of Min- Schreyer,W. and Seifert,F. (1967)Metastability of an osumilite nesota:A CentennialVolume, p.394-406. Minnesota Geologi- end-memberin the systemKrO-MgO-Al2Or-SiO2-HrO and its cal Survey,St. Paul, Minnesota. possiblebearing on the rarity of natural osumilites.Contribu- tions to Mineralogy and Petrology,14, 343-358. Manuscriptreceived, December 17, 1979; Schreyer,W. and Seifert,F. (1969)Compatibility relationsof the acceptedforpublication, September 19, 1980.