COMPOSITIONAL VARIATIONS of PLAGIOCLASE FELDSPAR from a BASALTIC LAVA FLOW Jnnnv M

THE AMERICAN MINERALOGIST, VOL.51, MAY-JUNE. 1966

COMPOSITIONAL VARIATIONS OF PLAGIOCLASE FELDSPAR FROM A BASALTIC LAVA FLOW Jnnnv M. Honlnn, Departmentof Geology,Teras Western College, El Paso,Teras.

Assrnecr

A detailed study of the Rock Creek basalt flow reveals significant relationships between chemical composition of the plagioclase feldspar and texture. The An content of the groundmass plagioclase varies inversely with size and directly with the amount of associated glass. The rate of cooling, alter extrusion, determined the composition of the groundmass plagioclase. The plagioclase phenocrysts, averaging 20 percent more calcic than the corresponding groundmass plagioclase, are intratelluric.

INrnooucrroN At the present time no detailed information is available concerning the variation in composition of any one mineral or the compositional variation of a mineral present in two distinct sizes (i.e., phenocryst and groundmasscrystals) in a lava flow. In order to obtain such information a study has been completed concerningthe variation in composition of groundmass and phenocryst plagioclase from a thick porphyritic flow (250 to 470 feet), the Rock Creek flow, of the Columbia River Basalt in west-centralIdaho (Bond, 1963). Sampleswere collected from five measuredsections of the Rock Creek flow and the phenocryst and groundmassplagioclase in each sample were concentratedseparately with a Frantz Isodynamicseparator after crush- ing and screening.The composition of the plagioclase,expressed as mole per cent An content,was determinedby the glassfusion method (Foster, 1955).In addition, the textural relationshipsin eachsample were studied with thin sections,with specialemphasis on determinationof the sizeof the groundmass plagioclaseand the amount of associatedglass so as to obtain an index of the rate of coolingfor each sample.The factors that determine the texture of any igneousrock include the rate of cooling,bulk chemical composition, composition with regard to hyperfusibles,and pressure.The similarity in averagecomposition of the plagioclasephenocrysts from all samples suggeststhat the first-formed crystals of plagio- claserepresent crystallization from the samemelt. There is no evidenceto indicate that the bulk chemical composition changed during crystallization. The content of hyperfusibles and the effect of pressureduring crystallization are hard to appraise, but it seemslikely that the influence of pressurewas small becauseof the crvstallizationat the surface.There- fore, the rate of cooling,after extrusion,was the primary factor in deter- mining the final texture. However, in small areas variations in volatile constituents and pressure may have influenced the rate of cooling and 807 808 JERRY M. HOFFER

Tarr,r 1. CouposrrroN.tr, .a.NoToxrunnl Dera ol Pr-a.crocr-asr FRoM THE Rocr Cnaor Fr-ow

Sample An Content An Content Average Size (width in mm) Per Cent Nunber Pheno.rJSt Groundmass of Groundmass Crvstals Glass

w-1 600 o12 7 w-3 61.5 54.5 0.04 48 w-4 61 5 530 004 45 635 485 0.07 27 w6 59 .5 51 0 0.04 .57 w-7 670 540 007 7 w-8 675 395 0.09 t7 w9 645 500 0.07 23 w-10 620 s2.0 0 0.1 .3-5 W-11 600 ,13.5 009 28 w-12 625 56.0 0.06 53 SKC-1 620 ,t0 0 009 5 SKC-2 61 5 370 008 8 SKC-3 63 .5 360 0.09 8 SKC.4 620 400 0.07 17 SKC_6 61 5 56..5 0.02 62 630 570 002 67 SKC-8 62.0 420 0.07 17 SKC-9 630 505 003 SKC-IO 61 0 440 004 32 sKc-11 730 480 004 20 SK-1 635 4.3 5 006 8 SK-2 620 4r.5 008 10 SK,3 620 400 008 6 SK-4 630 490 0.04 35 SK-5 62.0 475 0.07 8 630 53.0 0.o2 38 SK-8 62.O 5l 0 0.03 SK-9 63.0 47,5 006 12 sK-10 635 54.0 003 SK-11 52.O 0.04 35 SK-12 620 530 007 SK-13 620 51 0 0.03 43 R-1 635 430 0.13 640 ,+5.0 0.13 R-3 640 41 5 0.13 R-4 640 445 0.12 R-5 650 52.0 0.11 R-6 635 0.10 R7 60 .5 45.0 009 R-8 620 .56.5 0.03 52 R-9 620 505 0.04 30 R-10 61 .5 51.5 003 R-11 61 5 505 004 28 s-1 620 ,15.0 008 7 s-2 640 435 009 10 5-.1 670 430 0.09 J s-4 650 460 010 5 660 .52 0 007 38 s-6 630 .5.5.0 006 s-8 635 4q5 009 7 s-9 625 48.0 006 12 s-10 660 550 0 0.5 s-12 62.5 003 20 s-13 62.5 .50.0 0.02 s-14 62.5 520 003 17 s-l t 62.5 .5S0 o02 )l PLAGIOCLASE COMPOSITIONAL VARIATION hencethe final texture. The averagewidth of the groundmassplagioclase rvasdetermined by micrometerocular measurements on two thin sections per sample.The percentageof associatedglass was estimatedfrom two thin sectionsper sampleby comparisonwith standardpercentage charts. To check the relative accuracyof theseabove determinations,each thin sectionwas ranked in order of increasingsize of the groundmassplagio- claseand the amount of associatedglass by projectionfrom a 35 mm slide projector. In all samples,excellent agreement was obtained by both methods.The resultsof thesedeterminations are summarizedin Table 1.

Rrsur,rs ANDDrscussroN The compositionof the groundmassplagiociase ranges from An36to An5eand shows no s1'stematicspatial variation vertically or laterallv

(l) .. u, bo y=O.2445-O.OO38x tro o.t5 - E ctl r= 0.68 ?o '-(L o.to fo 33 o.o5

3eE q)c o.oo N= cn9 (9 MolePer CentAn Content

Fro. 1. Variation of mole per cent An content of the groundmass plagioclase samples with size (width) of the groundmass plagioclase (y: equation of the line and r : Iinear corrclation coefficient). Each point represents a sample. through the flow. Marked correlationsexist, however, between the An content and cr1'stalsize of the groundmassplagioclase (Fig. 1) and between the An content of the groundmassplagioclase and the amount of associatedglass (Fig.2). As the An contentof the groundmassplagioclase increases,the sizeof the crystalsdecreases (r: -0.68), and the amount of associatedglass increases (r: +0.75). In addition, a marked degreeof negative correlation (r: -0.i9) exists between the proportion of associated glassand the sizeoI the groundmassplagioclase crystals. The plagioclasephenocrysts range in compositionfrom An5e5 to An73, with 98 per cent of the samplesin the interval An5e5 to Anozr. Textural relationships and composition (the phenocrysts average20 per cent more 810 IERRV M. HOFFER calcic than the corresponding groundmass plagioclase)suggest that the plagioclasephenocrysts are intratelluric and crystallized at depth under conditions of slow cooling, before the crystallization of the groundmass plagioclase. The relationship of the size of the groundmass plagioclase and proportion of associatedglass indicates that there were variationsin the rate of cooling after extrusion of the magma. In samplesrepresenting conditions of relatively slow cooling the proportion of glassis small and the average size of the groundmass plagioclaseis relatively large. Con- versely, small groundmassplagioclase crystals, formed under condition of rapid cooling, are found in samples containing a large amount of glass. Rock Creek samplesrepresenting rapid cooiing, inferred from the small 60 y=42.OOO1.O.2332x. E55 :so o4so E

35 o l5 30 45 60 75 PerCent Gloss

Frc. 2. Variation of mole per cent An content of the groundmass plagioclase samples rnith per cent glass (y:equation of the line and r:linear correlation coefficient). Each point represents a sample.

crystal size and abundant glass, possessgroundmass plagioclaseof high An content (An66to An6e)whereas those that cooled more slowly contain groundmassplagioclase of low An content (Anaeto Anro). With regard to the plagioclasephenocrysts, no significant relationship was found to exist between the composition and the rate of cooling, as indicated by the proportion of glass. The relationship between compositional variations of the groundmass and phenocryst plagioclaseand cooling history, as indicated by textural features, during crystallization of the Rock Creek flow can best be ex- plained by referenceto the plagioclasephase diagram in Fig. 3 (Bowen, 1913).In this simplifiedscheme aII componentsof the Rock Creekmagma are neglectedexcept plagioclase.The temperaturesindicated on the phase diagram do not imply the temperatures of crystallization of the magma of the Rock Creek flow. PLAGIOCLASE COMPOSI TION A L VARI AT ION 811

t540

() o t420

(l) 5 E, q) CL E t300

o 20 40 60 80 too An Content

Frc. 3. The phase relationships of the plagioclase feldspars (A:composition of the sim- plified hypothetical Rock Creek magma. B:composition of the most calcic plagioclase phenocryst (Ana), C:composition of the most sodic plagioclase phenocryst (Anrr.u), D:composition of the most calcic groundmass plagioclase (An5e), E:composition of the most sodic groundmass plagioclase (Aq6), and F:composition of the hypothetical Rock Creek magma after extrusion to the surface).

When the temperature of the hypothetical Rock Creek magma of composition A intersected the liquidus, the first plagioclaseto crystallize had a composition of An73, the composition of the most calcic plagioclase phenocryst (B in Fig.3). As cooling continued more sodic crystals of plagioclasecrystallized and some of the earlier more calcic crystals were converted to more sodic composition by reaction with the melt. The relatively small range of composition of the majority of the Rock Creek plagioclasephenocrysts suggeststhat, for the most part, close approach to equilibrium between the crystals and the melt was attained during 812 JDRRY M. HOFFI'R crystailization.Weak zoning in manv of the phenocrystsindicates, however, that equilibrium betweenthe melt and crystalswas not maintained throughout the crvstallizationof the plagioclasephenocrysts. At this stage, when compositionof the plagioclasephenocrysts was about An63(average An content of all phenocrystsamples), the magma and its phenocrysts,in the form of a crystal mush, were extruded to the surface.Cooling was immediately accelerated,and the phenocrystforma- tion halted. The compositionof the melt at this stagewas probably close to An26,point F on the liquidus curve of Fig. 3, and the groundmass plagioclasecrystals that formed initially, after or during extrusionof the magma, had an An content of approximatell'59 per cent (point D). As coolingproceeded, cr\rstals more and more sodic in compositioncrystal- Iized and some of the earlier more calcic crystals reactedwith the melt and becamemore sodic.The final compositionof the groundmassplagio- clasewas thereforedependent upon the rate of coolingafter extrusion.In groundmasssamples which reflecta rapid rate of cooling(small size of the groundmassplagioclase and larger amounts of associatedglass), the An content of the groundmassplagioclase is high. The high An content of thesegroundmass plagioclases is evidently the direct result of the rapid cooling.On the other hand, the samplesthat show evidenceof crystallization under conditionsof slower cooling (groundmassplagioclase crystals arerelatively largeand areassociated with small amountsof glass)contain groundmassplagioclase of low An content. Under these conditions of slow cooling the early-formed groundmasscrystals of composition ap- proximately An5ew€re able to react with the melt and were convertedin compositionto more sodic forms. Therefore,the slower the rate of cooling, the more sodicare the final groundmassplagioclase crystals.

CoNcr.usroNs In summary, the important featuresof the plagioclasefrom the Rock Creekflow may be stated.The compositionof the groundmassplagioclase crystals (An36to An5e)varies inversely with their size and directly with the amount of associatedgiass. Therefore, the variation of An content of the plagioclasegroundmass crystals is due primarily to the variation in the rate of cooling under conditions of fractional crystallization. The compositionof the plagioclasephenocrysts is more calcic than the corresponding groundmass(98 per cent in the interval An5e.5to Aners) and does not show any relationship to the amount of associatedglass. The first formed crystals, the plagioclasephenocrysts, crystallized at depth under conditions of slow cooling which allowed closeapproach to equilibrium. At a later stage the magma was extruded to the surface and PLAGIOCLASECOMPOSITIONAL VARIATION 813 variations in the acceleratedcooling rate produced the compositional variations found in the groundmassplagioclase crystals.

AcrNowrBocMENTS The author wishes to thank The Society of the Sigma Xi for financial support of this study. Thanks are alsoextended to R. K. Sorem,Washing- ton State University, who critically read the manuscript and offered helpful suggestionsduring the investigation.

RrprnnNcrs

BoNr, J. G. (1963) Geology of the Clearwater Embayment. Iil.aho Bureau Mines Geol. Pamph. 128. Bowrx, N. L. (1913)The melting phenomenaof the plagioclasefeldsparc. Am. Jour. Sci.35, 577 599. Fosrnn, VV. R. (1955) Simple method for the determination of plagioclase leldspar. Am. Mineral..40, 179-185.