@fifiifinflfifieésfisw£31;v..-'; .1 ~ "“3-’3‘-1'-2‘¥'<'.':f:""‘ . ' ~ ‘ . A 1. c...-— ”V," .— ., - Iva-Cw K] Rb RATIOS 0F SGME PRECAMIRIAN GRANUUTES: lMPUCATION TOWARD. Rb DEPLETION‘ IN THE LOWER CRUST . Thesis for the Degree of M. S. MECHIGAN STATE UNlVERSITY JERRY D. LEWIS 1971 L I B R A R Y Michigan State University ‘H'P't‘ I ? amgme BY ‘3' ”DAB 8 SUNS' , 300K BINDERY WC ,. LIBRARY BIN? '5 . ’MNBPD; ABSTRACT K/Rb RATIOS OF SOME PRECAMBRIAN GRANULITES: IMPLICATION TOWARD Rb DEPLETION IN THE LOWER CRUST BY Jerry D. Lewis A study of the K/Rb ratios of 62 Precambrian shield granulite rocks was conducted to determine whether: (1) these rocks fall along an established trend for normal crustal rocks; (2) the K/Rb ratio remains constant or decreases as regional metamorphism proceeds; (3) these shield rocks are depleted in Rb with respect to K. Analytical determination of K was done by atomic absorption Spectroscopy and the Rb by isotope dilution mass spectrometry. The analytical values obtained were compared statistically by covariance analysis techniques. It was found that: (l) the trend established by the population tested created a linear plot significantly above higher K/Rb ratios in the previously established trend for normal crustal rocks. (2) the K/Rb ratio increases as regional metamorphism and/or anatectic melting proceeds; these rocks are depleted in Rb with respect to K along the main trend. Jerry D. Lewis The equation established for normal crustal rocks does not describe the Precambrian granulites studied here. Statistical analysis of the K and Rb data therefore sug- gest that the following equation be used: loglo(ppm Rb) = 0.7062 loglO(K%) + 0.7362. The squaring of the product moment correlation coefficient reveals that Rb is associated with K 80 per cent of the time in these rocks. The association in crustal rocks was found to be 56 per cent and is probably considerably lower for pegmatitic-hydrothermal rocks. K/Rb RATIOS OF SOME PRECAMBRIAN GRANULITES: IMPLICATION TOWARD Rb DEPLETION IN THE LOWER CRUST BY 9.. QT Jerry D? Lewis A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geology 1971 ACKNOWLEDGMENTS Appreciation is extended to Dr. C. M. Spooner for his suggestions and patience during the research and preparation stages of this thesis; Dr. T. A. Vogel for suggestions, proofreading, and assistance with XRF and A.A. units; Dr. H. B. Stonehouse for suggestions offered during the proofreading. Further appreciation is extended to Dr. B. G. Ellis of the Soils Department for assistance with the flame photometer; Dr. R. E. Ehrlich for help with statistical interpretation; the many other pe0ple of the Soils and Geology departments who offered helpful advice and assistance. ii TABLE OF CONTENTS Chapter I. INTRODUCTION . . . . . . . . Outline of the Problem . . . . II. MINERALOGY . . . . . . . . . Mineralogic Associations . . . III. PREVIOUS WORK . . . . . . . . Metamorphic Studies . . . . . IV. DATA COLLECTION AND INTERPRETATION . Comparison of Granulites to Other Crustal Rocks . . . . . . Modal Mineral Composition . . . V. SUMMARY AND CONCLUSION . . . . . Conclusion . . . . . . . . BIBLIOGRAPHY . . . . . . . . . . APPENDIX A. Analytical Procedures . . . . . iii Page 10 ll 13 16 16 22 30 31 32 37 LIST OF TABLES Table Page 1. Properties of the alkali metals Na, K, and Rb . . . . . . . . . . . . . 6 2. Typical K/Rb values of non-pegmatite forming minerals . . . . . . . . . . 9 3. Analytical K(%) values . . . . . . . . . l7 4. Comparison of analytical K and a calculated crustal K . . . . . . . . . 21 5. Estimated modal mineral composition (in %) . . . . . . . . . . . . . . 25 iv Figure 1. LIST OF FIGURES Plots of Rb vs. K in micas and coexisting feldspars . . . . K/Rb fractionation trends in igneous and associated rocks . . . . Plot of K/Rb values Scatter diagram for Scatter diagram for Scatter diagram for granulites . amphibolites . crustal rocks Page 14 19 23 24 25 CHAPTER I INTRODUCTION In recent years many geochemists suggest that there is a depletion with crustal depth of K, Rb, U, Th, and other lithophile elements due to fractionation processes which take place during high grade regional metamorphism or with the generation of magmas. The extent of depletion may be a function of temperature, pressure, time, and/or nature of the melt phase since the primary generation of the crust from the mantle. Investigations of this problem have been inconclu- sive to date, investigators stating that more trace elemental data are needed for a better evaluation of crustal geochemical processes. Outline of the Problem In this study, pyroxene granulites, possibly of igneous origin (Spooner, 1969) are analysed by atomic absorption methods for K. These K values are compared with previously determined (Spooner, 1969) Rb values (by isotope dilution) to test a theory of crustal evolution first set forth by Poldervaart (1955) and later expanded by Hurley (1968) and revised by Russel and Ozima (1971). The generalized idea set forth is that there is a strong concentration of Rh in the continental crust. It is considered that "outgassing" or diffusion of the alka- lies took place very early in the earth's history, perhaps during the formation of the crust. The K and Rb were trapped in the residual fluids of peridotites, carried upward into the base of the crust and mixed with sial con- stituents by remelting and anatexis. More Rb was carried off with respect to K because of the affinity of the plagioclase lattice for the K, rather than Rb. If the above statements are true it may be expected that a depletion of Rb with respect to K will be found in the lower crustal granulites. On the basis of this sup- posed depletion, several ideas may be suggested for the evolution of the crust. Following the earth's accretion, a primary stage of differentiation began which concentrated the lithophile and radioactive elements into the upper mantle (Ringwood, 1969). The upper mantle probably consisted of peridotite or pyrolite. Later differentiates of these rocks produced a low density (P=3.0) protocrust consisting of some combination of gabbro, anorthosite, and granulite. From these materials the upper crust gradually evolved, with much of the present lower crustal material probably being similar in chemical composition to its parent protocrust. This differentiation apparently occurred over a long period of time in the earth's history, and continues today. Rb-Sr age determinations suggest that this conti- nental crust development commenced about 3400 x 106 years ago (Hurley, 1968). Some protocontinents (Goodwin, 1968) could have first formed from the accumulation of crystalline material on and around heat resistant (refractory) anorthosites. The association of granulites and anorthosites has been noted in several Precambrian shield areas (Windley, 1969; 1970). These protocontinents could have eventually joined together in some way to form more stable cratons (Goodwin, 1968). During their formative period, when high heat flow and high temperatures prevailed, metasomatic processes could have carried the alkali metals, K and Rb in particu- lar, upward and concentrated them in the intermediate and upper crustal zones. Where the temperature and pressure were extreme, hydrous minerals were dehydrated and some anatectic melting accomplished, driving off some of the Rb into a liquid or vapor phase. In areas of less intense heat and pressure, as near the surface of the protocrust, hydrous minerals reformed, thus creating a reservoir for Rb (and K). Erosion of exposed surfaces created sediments rich in Rb, which eventually were recycled to form gneisses and migmatites, many of which have been associated with high to medium grade metamorphic terrains. In general, then, future enrichment of Rh in the upper crust resulted from sedimentary reworking and/or selective melting of the lower crust. The study of the origin of continents is still in the primitive stages, as is the interpretation of various petrologic indicators. However, it does appear that the distribution of K and Rb in crustal rocks of various ages carry much important information and may be one of the keys to understanding the development of the crust from the upper mantle as well as later crustal processes. CHAPTER II MINERALOGY The introduction of the idea of various element pairs as petrologic indicators has been received with high optimism. Elements which have similar geochemical proper- ties and behavior, such as the alkali metals: K and Rb, have proved to be most reliable in the interpretation of various petrologic processes when a deviation from estab- lished behavior patterns can be found. Table 1 gives some of the physical data which shows the close association of the elements to one another. The element K has an 8, 10, and 12 coordination with oxygen in silicates and fits well into lattices such as feldspar and sheet structures that require this type of coordination. The Rb has a characteristic 12 coordination and fits well into the crystal lattice of the silicate sheet structures. In a melt the K/Rb ratio would be controlled mainly by the amount of plagioclase present plus the relative amounts of K and Rb in the magma (Aldrich, gt_al., 1965). Discrimination against Rb by metamorphic feldspars has been TABLE 1.--Properties of the alkali metals Na, K, and Rb. Sodium Potassium Rubidium Atomic weight 22.99 39.10 85.47 Density (20°C) 0.97 0.86 1.53 Melting point (°C) 97.5 62.3 38.5 Boiling point (°C) 883 760 700 Electronegativitya 0.9 0.8 0.8 Radii (in i) M b 1.53 2.025 2.16 M+ C 0.97 1.33 1.47 M+ d 0.95 1.33 1.48 Bond length (in i) 3.72 4.54 ’ 4.95 Source: Handbook of Chemistry and Physics, 1965-66: aPauling (1948) bHeier and Adams (1964) cAhrens (1952) dPauling radii noted by several investigators (Lambert and Heier, 1968; Jakes and White, 1969). Early investigators thought that there was no increase in Rb with differentiation (Erlank, 1968). How- ever, it has since been found that generally, Rb increases systematically with differentiation and reaches a maximum concentration in residual fluids such as those associated with a pegmatitic stage of crystallization (Taubeneck, 1965; Heier and Adams, 1964; Shaw, 1968; Philpotts and Schnetzler, 1970). It is believed that a high water pres- sure and Al content would induce phyllosilicates to form thus increasing the Rb content. A concentration of Rb in various phyllosilicates is examined in a literature review by Heier and Adams (1964). The range of Rb values for each of 3 types of mica were: ranging from (low) Biotite-- Muscovite--Lepidolite (high) Biotite (x—lOOppm); Muscovite(x-1000ppm); Lepidolite (x-10,000ppm), where x is variable. This tendency of the phyllosilicates to concentrate Rb has been noted by many investigators for regional metamorphism and partial melting (the reader is referred to Figure l). The author points out the noticeable differences in the biotite and lepidolite and the wide spread of values for muscovite in the following figure. Jakes and White (1969) suggest that the K/Rb ratio in metamorphic rocks is proportionately related to the ppm Rb Source: Heier and Adams, 1964. '- l1.”L=.':OU FIGURE l.--Plots of Rh versus K in micas and some coexist- ing feldspars. Tielines connect coexisting minerals. Legend: A feldspar O biotite O muscovite fl lepidolite modal content of amphibole. They found gabbroic rocks with more than 50 per cent amphibole to have a K/Rb ratio of 1000, whereas a dioritic rock with less than 15 per cent amphibole had a low K/Rb ratio of 295. Values for K/Rb ratios from two to four times lower were found.in coexist- ing pyroxenes and plagioclases and lower yet by 10 for coexisting biotite (refer to Table 2). TABLE 2.--Typical K/Rb values of non—pegmatite forming minerals. Mineral K/Rb Range Source Amphibole 550-3720 Jakes & White, 1969 Pyroxene 243-670 Hart & Aldrich, 1967 _ Jakes & White, 1969 Biotite 35 470 Hart & Aldrich, 1967 Muscovite 70-424 Shaw, 1968 K-spar 108-620 Lange, 1967 . _ Murthy & Griffin, 1969 Plagioclase 280 4450 Shaw, 1968 Garnet Less than 70, Jakes & White, 1969 Examination of the above table gives an indication of the extent that various minerals reject or accept Rb in their structures. For coexisting pyroxenes and garnets it was found that K/Rb values were lower for the garnets. In a garnet from a granulite rock studied by Griffin and Murthy (1968) they found a minor amount of K20 (0.025%) and a measurable 10 amount of Rb (Sppm) thus implying a low K/Rb value for the rock. Mineralogic Associations Granulites have a unique mineral association of pyroxene, potash feldspar, quartz, plagioclase, and garnet, and usually lack the hydrous minerals biotite and horn- blende. However, in some cases of retrograde metamorphism the hydrous minerals may be formed. Generally, granulites are found only in very deeply, eroded, highly metamorphosed terrains. The parentage of some of these rocks are believed to be meta-igneous while others are believed to be meta- sedimentary in origin. CHAPTER III PREVIOUS WORK The first preliminary investigation into the use of the K/Rb ratio as a petrologic indicator was done by Ahrens and coworkers (1952). Since this work several papers (Ahrens, 1964; Taylor, 1956, 1965, 1970; Turekian, 1963 and Abbott, 1967 amongst others) and reviews (Heier and Adams, 1964; Erlank, 1968; Shaw, 1963) have been pub- _lished dealing with the use of the K/Rb ratio as an indi- cator. The first comprehensive summary of the alkali elements (Heier and Adams, 1964) established the general distribution and concentration of these elements and that the ratio may have potential as an indicator. The next review came four years later in an excel- lent summary by Erlank (1968). In a paper presented at a symposium in Paris, he presented five conclusions which he drew from the literature, as follows: 1. The K/Rb coherence is not as close as previously thought. 11 12 2. There appears to be no systematic difference between continental and oceanic areas with respect to K/Rb ratios as high values (K/Rb approx. 1000) are reported for rocks from continental areas (i.e., anorthosites). 3. High ratios (K/Rb of 1000 or more) have so far been found only in low K rocks (K less than 0.2%), notably tholeiitic basic rocks, anorthosites and peridotites from St. Pauls Rocks. 4. No firm conclusions may be drawn as to the vari- ation of the K/Rb ratio with respect to igneous differentiation. 5. Wide spread in K/Rb of material presumed to origi- nate in the earth's mantle (basalts, ultramafic rocks and eclogites) does not allow a clear dis- tinction to be made as to whether the mantle has a K/Rb ratio similar to chrondritic or achrondritic meteorites. In addition he suggests that insufficient attention has been given to the influence of mineralogy on the vari- ation of the K/Rb ratio in igneous rocks and metamorphic rocks. Very shortly after Erlank's (1968) presentation Shaw (1968) followed with a review and statistical study of 21 suites of igneous to quasi-igneous rocks drawn from 13 the previous literature. From this work three basic trends were established. These trends are displayed in Figure 2 below. The three trends are: 160-300 1. Continental and oceanic rock types; K/Rb 2. Oceanic tholeiites and achondrites; K/Rb = 1000+ 3. Pegmatitic-hydrothermal rocks; K/Rb = 160- The main trend established for normal crustal rocks had average K/Rb value of 229 with limits of 160 and 300. The few papers published since Shaw's (1968) review have agreed reasonably well with the ratios estab- lished for the various rock types. Metamorphic Studies In Australia, Lambert and Heier (1968) studied the relationship of K to Rb in the Precambrian shield rocks to determine if a depletion of Rb existed. From their work they found that Rb was not only depleted but also K, Th, and U. These elements were concentrated into the upper crust due to pressure, and temperature effects of regional metamorphism. They assume the high to medium pressure granulites reflect chemical trends with depth in the conti- nental crust as a result of this metamorphism (including anatectic melting). In an investigation of a metamorphic terrain in the Adirondacks, Whitney (1969) reached the same conclusion l4 I I I IO. - / / / /’ ,IPH L0 . . ‘ MA"! K‘ // / "RENO 0.1“ / l / d // / ,ra’ /' ;% 01' 'v“ 0.01 - .. ”No.1 1.3.. to 100 1000 ppm lb FIGURE 2.--K/Rb fractionation trends in igneous and asso- ciated rocks (after Shaw, 1968). Legend: 0T oceanic tholeiitic basalts MT main trend PH pegmatitic hydrothermal 15 as Lambert and Heier (1968) regarding a regional depletion of Rb. He found the K/Rb ratio to be controlled by modal composition. Biotite-poor granites had K/Rb values of 431, the host gneisses 227. The paragneisses from granulite zones seem to have higher K/Rb ratios than those from amphibolite terrains possibly indicating a depletion of Rb °at depth due to more extreme pressure and temperature con- ditions. CHAPTER IV DATA COLLECTION AND INTERPRETATION This study is an extension of a previous investi- gation into variations in initial Sr-87/Sr-86 ratios in high metamorphic grade terrains (Spooner, 1969). In this study, samples from various Precambrian shield areas belonging to the granulite facies were analysed for K. Samples were analysed by atomic absorption and x-ray fluorescence (refer to Appendix A for analytical procedures). Flame photometer was also used but the data was disregarded due to instrumental drift and atomizer plugging. Data collected for each run are tabulated in Table 3. Plotting of the analytical values developed a trend roughly paralleling the 300 ratio line on a log-log plot. This trend lies significantly above a trend developed by Shaw (1968) for normal crustal rocks (refer to Figure 3). Comparison of Granulites to Crustal Rocks To compare granulite data with upper crustal rocks Shaw's (1968) equation will be used. The equation is as follows: 16 17 TABLE 3.--Ana1ytical K(%) values. Sample F.P. F.P. A.A. A.A. XRF no. run #1 run #2 run #1 run #2 R7340 1.19 1.51 1.74 1.70 R7341 2.00 1.83 1.80 R7344 3.05 1.60 2.42 2.32 R7345 3.11 3.66 3.65 R7346 2.62 3.73 3.75 R7347 0.80 1.09 1.12 1.12 R7244 0.28 0.32 0.32 R7329 2.37 3.46 3.09 R7071 2.79 3.26 2.50 R7085 0.91 1.10 1.09 1.15 R7178 0.83 1.04 1.03 R7180 0.77 1.31 1.34 R7019 3.62 4.36 4.22 4.37 R7027 1.10 1.24 1.18 R7038 0.57 0.81 0.79 R7039 1.04 1.17 1.03 R7040 0.62 0.95 0.82 R7042 2.31 2.80 2.76 R7049 0.96 0.92 0.79 R7050 1.50 2.10 2.00 R7051 1.44 2.04 2.00 R7052 3.11 3.16 3.16 3.23 R7053 2.43 2.39 2.15 R7054 0.91 1.22 1.23 R7055 2.51 2.00 1.95 R7057 2.17 2.13 1.98 R7225 3.45 3.40 2.40 R7228 0.69 0.69 0.70 0.75 R7229 5.26 4.11 4.00 R7230 0.56 0.64 0.62 R7233 5.44 4.63 4.48 G-l ave. 4.61 3.60 4.58 4.50 4.53 18 .¢ .m .D .xuor 3oz .mxma unfiocH mumcmo .ofiumuso .uuomummz .< .m .2 .xno» 362 ..u: «ammo mowumfid :usom .wcmmsw .sxscmx macsH .Emum>maamm mHBaH .mumum mmucmz mowumd .mosmmb .maaoxo a amoxom BUHHmm .mwsmNcma .uwuuwm Honmq can .92 mumm D> <1 1| C) >< -+ C) (D moaum‘m desensme .fiammmz D “DZNUNA .mmsam> AM\M mo uoamul.m mmDon 19 Al Ema 68— cc— on —.c 4 4 8.6 .? a a. 000 rl J Fug c O O O. 0. 4r! . o 9 D, .’ O ‘5 Q 1.: <1! 0 \ \bec! .. . :2 Gd 4, «4.4.3... +7, + . _ c.9— 20 %K = 0.0369 (ppm Rb)°'897 If the equation describes the granulites, then it can be assumed that they are of the same population as normal crustal rocks. If this is true, when Rb is substi- tuted into the equation 50 per cent of the K values should be greater than zero, and 50 per cent less than zero when the calculated KC is subtracted from the analytical Ka: Ka-Kc=0. It was found that 53 of the 62 samples were greater than zero, i.e., above Shaw's (1968) line (Table 4). The use of binomial tables gives a value of p less than .00001 indicating that these granulites do not fall in this trend established for normal crustal rocks. With this high level of significance it becomes apparent that Shaw's line does not describe granulite rocks. An equation is proposed here based on statistical data obtained from the population studied. The equation as developed: loglo(ppm Rb) = 0.7062 loglo(K%) + 0.7362 is based on average K and Rb Values of 2.38 per cent and 75ppm respectively. The incorporation of scatter diagrams was made to test for random distribution of K and Rb. Examination of the following diagram (Figure 4) clearly shows that there is no randomness in the values plotted, but rather a 21. TABLE 4.--Comparison of analytical K and a calculated crustal K. WELL ML 2W 051m 41110 7 11 0 AN 0 0040 165.00 4.83 3.56 1.27 296 7012 2 AN 0 0049 132.00 4.00 3.01 1.07 361 _' - 1303-50 .1-35 0.21 .1-06 144 701920403! 7 040 371.50 4.29 7,44 .3.15 116 7020 I AN 0 0049 207.00 4.00 4.41 .0.41 193 1 000 30-00 1-21 0.26 0-25 310 703BROKOS! s 0K1 19.00 0.00 '0.52 0.26 421 703900405 4 one 19.00 1.10 0.52 0.50 579 " 0452 0117 1001 704200x03! 1 OK? 43.00 2.70 1.00 1.70 647 7849204050 1 0K3 19.00 0.86 0.52 0.34 453 1 66.00 2-05 41.58 0.47 1:111 7051434011 71.00 2.02 1.69 0.33 205 7052H340Ll 00.00 3;16 2.05 1.11 359 [2101101) 70. 0 2-27 .1-04 0.43 291 705446401! 46.00 1.23 1.14 0.09 267 70554340LI 01.00 1.90 1.90 0.05 244 ____20210310L1 63. 0 2.06 1.52 0.54 327 225P4RE '7 51.70 2.90 1.27 1.63 561 72202425 H7 2.55 0.70 0.09 0.61 2745 2229P10F "I 79.40 4.06 1.07 2.19 511 7230PARF ’7 42.90 0.63 1.07 -0.44 147 '7233L480RCSERRl? 222.50 4.56 4.71 -0.15 205 7205 pALLAyARM 1. 134-00 3-25 2-99 0.26 243 7214 PALL‘VARH 7 6.30 0.20 0.19 0.01 317 7215 PALLAVARM 7 6.50 0.20 0.20 0.09 ‘52 A n 1 12.50 0-47 0.36 0.11 370 7217 PALL‘VARH v 3.90 0.20 0.13 0.07 513 7216 PALL4VARM 7 11.00 0.37 0.32 0.05 335 A 1 11.50 0.51 .0-33 0.16 441 220 PALL”VARN 7 4.30 0.26 0.14 0.13 609 7221 DALL‘VARM 7 20.20 0.52 0.55 -0.03 256 A 1 140-00 1-20 3-22 .2-02 .02 I41 PALLAVARM 7 7.60 0.40 0.23 0.10 532 7242 PALLIVARH 7 210.00 1.57 4.47 ~2.8° 79 7144PALLAVARAH 9-50 0.32 0.25 0.04 317 7061 9297 2007 1 28.00 1220 0.73 0.47 429 7062 9557 P027 5 24.00 0.99 0.64 0.35 413 7070 BEST P097 0 140.00 4.23 3.11 1.12 302 7071HESY D007 09 03.00 2.90 1.52 1.35 450 7003 WEST P097 0 65.00 3.75 1.90 1.77 441 70654557 ”0P1 04 15.00 1.10 0.42 0.65 733 7090 HEST P007 1 57.00 3.12 1.39 1.73 947 7112 HEST 2007 1 90.00 4.00 2.05 1.95 455 1013 HEST P007 1 125.00 0.10 2.01 3.37 494 7123 CRANE M7 N 165.00 4.65 3.60 1.09 282 7125 cRANh 07 N 151.00 4.42 3.32 1.10 293 7126 0040 HT N 138.00 4.24 3.07 1.17 307 7127 CRANt 07 N 126.00 4.33 2.03 1.50 344 7178HADRAS 74.70 1.04 0.41 0.63 707 7100010046 40-50 1-33 1-02 0.31 .320 7321 CAKES 077) 93.00 2.24 2.15 0.09 241 7822 LAKES ~17) 135.03 3.97 3,01 0.96 294 ____;125 {AXE 0171 121-00 3-73 2122 1.01 30! 327 flAKES u)v) 53.05 1.25 1.30 00.09 236 73291NDIA LAKE 101.00 3.28 2.32 0.96 325 ' U 57-“0 1-72 1-39 0.33 302 7341KUNUK9 60.00 1.62 1.62 0.20 260 7344KUNUKU 71.00 2.37 1.69 0.66 334 ZSASKUMUKU 120. 0 3-06 2.70 0.96 _ 305 'KU 130.00______3111______3.011 0111, .211 7347KUNUK” 16.00 1.11 0.44 0.67 694 22 significant linear trend indicating a very high association of one variable upon or with the other. The squaring of the product moment correlation gives the association of Rb with K as 80 per cent. Each unit of Figure 4 represents one standard deviation. Similar operations performed on amphiboles and normal crustal rocks (Figures 5 and 6 respectively) lead to the observation that the Rb is associated with K 80 per cent of the time in the amphibolites and that in normal crustal rocks this association is not as prominent, being only 56 per cent. This association would probably be significantly lower for pegmatitic-hydrothermal rocks due to the very low K concentration in the parent residual fluids. Both elements appear to be dependent upon the bulk modal composition of the rock, or melt and the presiding pressure and temperature variables for their respective concentrations. A brief look at the estimated modal compo- sition is given below. Modal Mineral Composition The following is a brief description of the modal mineral composition and its possible effect on the K/Rb ratio. The estimates as given by Spooner (1969) are listed in Table 5. Rakosi and Okollo, Uganda, Africa (R7011-7049): in these samples two suites are represented. One, an 23 .Amc0flumw>mo oumocmum v\H nay Emummap Hmuumom muwascmuoun.v mmeHm an x x x n m a o a. m. n. x x ..1_4.4.4...u-...--_.-..4-...............-4.....-.._-. “ In. a a a a ~ ON. a ~ _ a .UH. dd .v-Cv-O H". Inc-I‘d (UN to.» won-a «an n . «I «a C x 0‘ «amount: F. ,4 N 9-0 0—0 .o-Oo-oo-o [Hunt-o .o-o—o—o .u-oo-ao-o n . >- 24 .Amsowunw>mo ouuocmum «\H Gav EMHmMfiU Houumom ouHHonwnmE an > H >1. 25 .4msofiumw>mo onmocmum «\4 say Emummwo Hmuuwom xoou Hmumouo HmEHozlu.m mmDUHm .Ammmac .46 pm .uoamma “Ammmav ummmsmom a xcmaum “Achmav sumo “Achmac smaofim a umxmm «Ahmmav uponna "muonsom x x an x n N a o a. - n6 X X IldulIWIJCJuWIWIiiluIIJIJIJufiIIIOIONIIIFOWJ~IOJIOOI~II.\ a U”. 9-. 9-7 H d Nd NOON a nvn Nvomn an dd I c a 0 ‘x 4 «fan '4 a NdNNh' H ddavJauo 0| houndue—oflun—nuofiuan—s'54—”— n )- 26 TABLE 5.--Estimated modal mineral composition (in %). No. Qtz Biot Plag K-spar Pyr Amph Garn Mag K/Rb 7011 35 1 -- 60 -- —- 2 -- 296 7012 60 -- -- 40 -- -- -- -- 361 7018 60 -- -- 40 -- -- -- -— 144 7019 116 7020 60 -- -- 40 -- -- 2 -- 193 7027 318 7038 421 7039 579 7040 468 7042 30 20 15 lO 3 -- -- -- 647 7049 453 7050 18 8 48 13 13 -- -- —- 311 7051 25 5 45 25 10 -- -- -- 285 7052 45 -- 35 20 -- -- -- -- 359 7053 35 10 30 25 -— -- -- -- 291 7054 30 12 50 -- 10 -- -- -- 267 7055 30 15 30 —- -— 8 -- -- 244 7057 35 15 25 lo -- 10 -- -- 327 7225 60 -- -- 40 —- -— -- -- 561 7228 2745 7229 511 7230 147 7233 205 7205 35 l 5 50 -- -- -- 2 243 7214 5 -- 30 -- 40 20 -- 5 317 7215 5 -- 30 -- 40 20 -- 5 442 7216 5 -- 30 -- 20 15 25 5 370 7217 10 -- 30 -- 10 50 -- -- 513 7218 10 -- 20 10 40 20 -- 5 335 7219 10 -- 30 10 15 30 -- 5 441 7220 15 -- 20 5 10 45 -- 5 609 7221 10 -- 20 15 15 20 -- 5 256 27 TABLE 5.—-Continued. No. Qtz Biot Plag K-spar Pyr Amph Garn Mag K/Rb 7240 10 20 20 15 30 -- -- 2 82 7241 532 7242 10 20 20 20 30 -- -- -- 75 7244 337 7061 15 10 -- 15 -- 60 -- -- 429 7062 15 10 —- 15 -- 60 -- -- 413 7070 50 3 —- 40 -- -- -- -- 302 7071 35 5 15 45 -- -- —- -- 460 7083 30 10 50 -- 10 -- —- 441 7085 30 3 20 40 -- 10 -- -- 733 7090 40 2 15 45 —- -- -- -- 547 7112 30 -- -- 40 20 -- -- 5 455 7113 30 -- -- 4O 20 -- -- 5 494 7123 31 2 22 39 -- 5 —— -- 282 7125 30 2 18 47 -- 4 —- -- 293 7126 23 -- 22 41 11 -- -- 307 7127 20 -- 24 46 8 -- -- 344 7178 35 -- -- 40 20 5 -- —- 707 7180 20 5 10 20 -- 20 -- 328 7321 15 -- -- 55 35 -- -- 241 7321 15 -- -- 55 35 -- -- 241 7322 20 2 20 30 10 15 -- -- 294 7326 20 -- 20 40 10 -- -- -- 308 7327 20 10 20 20 15 15 -- -- 236 7329 20 -- 20 40 10 -- -— -- 325 7340 20 10 -- 50 15 -- -- 5 302 7341 20 15 -- 50 10 -- -- -- 268 7344 20 15 -- 40 8 -- 10 5 334 7345 20 20 -- 40 5 -- 10 5 305 7346 271 7347 694 After Spooner (1969). 28 aplitic granulite, rich in potash feldspar, was believed to have formed in a high temperature environment with some partial melting. The Rb was probably transported by solution into surrounding amphibolite rocks. The second, an acid granulite was formed at a high water pressure and low temperature thus allowing biotite to form and capture some of the Rb, thus resulting in generally lower K/Rb ratios. K/Rb values for both suites range from 116-646. Msagli, Tanzania, Africa (R7050-7057): These rocks may be of a low granulite facies as suggested by the Ca plagioclase and amphibole content. A high biotite content probably was responsible for low K/Rb values (244-359). Labor Serrit (R7233) and Pare Mt. (R7225-7230), Tanzania, Africa: The few samples represented do not allow for any suggestions to be made. Madras State and Pallavaram, India (R7205-7244): The suites contain large amounts of amphibole and plagio- clase. The high K/Rb values obtained, 75-707, may be a reflection of these minerals. Westport, Ontario, Canada (R7061-7113): The suite contains abundant potash feldspar and has high K/Rb ratios of 302-733. Crane Mt. (R7123-7127) and Indian Lake (R7321-7327), New York, U. S. A.: The suites contained varying amounts of plagioclase, potash feldspars, and biotite. The 29 corresponding K/Rb ratios, 235-344, seem to reflect the modal composition. Kanuku, Guyana, South America (R7340-7347): This suite contains a large amount of biotite and potash feld- spar and has generally low K/Rb ratios of 271-694. In the above suites it is very apparent that the presence of biotite decreases the K/Rb ratio by a signifi- cant amount. It also seems that such minerals as amphibole and plagioclase tend to raise the ratio thus suggesting that they do discriminate against the Rb. CHAPTER V SUMMARY The preceding data have established that the amounts of K and Rb present in granulites are dependent primarily on the pressure-temperature variations encoun- tered, and secondly on the modal composition of the melt. Extremes in temperature can destroy Rb storing minerals such as the phyllosilicates. The work of Lambert and Heier (1968) and Whitney (1969) have arrived at similar ideas. Work by Goodwin (1968) and Windley (1968) have substantiated the association of granulites with anortho- sites in early Precambrian terrains thus implicating the high temperatures necessary for metasomatic processes which could have carried the Rb (and minor K) by solution and vapor phase, into a developing upper crust. The concentration of Rb has been established rather well in that the sediments usually have a low K/Rb ratio. These sediments could further account for the lowered crustal values when mixed with higher ratio rocks by ana- tectic melting and/or regional metamorphism processes. 30 31 The Precambrian granulites studied in this problem clearly indicate that the lower crust is depleted in Rb with respect to K, and is continuing to be depleted by various geochemical processes. On the basis of a suggested Rb depletion, Hurley (1968) made an estimate of the total Rb and K content of the crust as a whole. For the lower crustal rocks an estimated K/Rb value of 300 was selected. This value is in very good agreement with the average K/Rb value found in this study. Conclusion In conclusion it was found that the Precambrian granulites analysed in this study established a trend sig- nificantly above and roughly parallel to the established trend for normal crustal rocks by Shaw (1968). The K/Rb ratio increases as regional metamorphism and/or partial melting processes proceed due to the removal of Rh in larger quantities than K. It appears that the Rb ion is more sensitive to the higher heat than K thus becom- ing a more mobile element. Rubidium is more receptive to a hydrous environment which is usually found closer to the surface of the upper crust. With this in mind along with the physical proper- ties of K and Rb, it is easily seen why the lower crust is depleted in Rb with respect to K. BIBLIOGRAPHY BIBLIOGRAPHY Abbott, M.J. (1967) K and Rb in a continental alkaline igneous rock suite. Geochim. Cosmochim. Acta 31, 1035-1041. Ahrens, L.H. (1964) The significance of the chemical bond for controlling the geochemical distribution of the elements: Part I. Physics and Chemistryiof the Earth 5, 1-54. Ahrens, L.H., et a1. (1952) Association of Rb and K and their abundance in common igneous rocks and meteor- ites. Geochim. Cosmochim. Acta 2, 229-242. Baker, I. and Ridley, W. (1970) Field evidence and K, Rb, Sr data bearing on the origin of the Mt. Taylor volcanic field, New Mexico, U.S.A. Earth Planet. Sci. Lett. 10, 106-114. Barbieri, M., et al. (1968) Rb and K relationship in some volcanoes of Central Italy. Chem. Geol. 3, 189-197. Brooks, C. (1966) The effect of mineral age discordancies on total-rock Rb-Sr isochrons of the Heemskirk Granite, Western Tasmania. Jour. Geophys. Res. 71, 5447-5458. Cann, J.R. (1970) Rb, Sr, Y, Zr and Nb in some ocean floor basaltic rocks. Earth Planet. Sci. Lett. 10, 7-11. Clifford, T.N., et a1. (1969) Petrochemistry and age of the Franzfontein granitic rocks of northern South-West Africa. Geochim. Cosmochim. Acta 33, 973—986. Cooper, J.A. (1963) The flame photometric determination of K in geological materials used for K-Ar dating. Geochim. Cosmochim. Acta 27, 525-546. Eade, K.E., et al. (1966) Composition of crystalline shield rocks and fractionating effects of regional meta- morphism. Nature 211, 1245-1259. 32 33 Engel, A.E.T. and Engel, C.G. (1960) Migration of elements during metamorphism in the Northwest Adirondack Mountains, New York. U.S.G.S. Prof. Paper 400B, 465-470. Erlank, A.J. (1968) The terrestrial abundance relationship between K and Rb. Symposium on The Origin_and Dis- tribution of the Elements. Paris, 1967. Ed. L.H. Ahrens, New York, Pergamon Press, 871-889. Erlank, A.J. and Hofmeyr, P.K. (1966) K/Rb and K/Cs ratios in Karoo dolerites from South Africa. Jour. Geo- phys. Res. 71, 5439-5445. Erlank, A.J. and Hofmeyr, P.K. (1968) K/Rb ratios in meso- zoic tholeiites from Antarctica, Brazil, India. Earth Planet. Sci. Lett. 4, 33-38. Fleischer, M. and Stevens, R.D. (1962) Summary of new data on rock samples G-1 and W-l. Geochim. Cosmochim. Acta 26, 525-543. Fleischer, M. (1965) Summary of new data on rock samples G-1 and W-l, 1962-1965. Geochim. Cosmochim. Acta 29, 1263-1283. Gast, P.W. (1965) Terrestrial ratio of K to Rb and the com- position of the Earth's mantle. Science 147, 858- 860. Goldschmidt, U.M. (1937) The principles of distribution of chemical elements in minerals and rocks. Jour. Chem. Soc. 1937, 655-693. Goodwin, A.M. (1968) Archaean protocontinental growth and early crustal history of the Canadian Shield. Intl. Geol. Congress, Prague, Czechoslovakia. Goodwin, A.M. (1968) Evolution of the Canadian Shield. Proc. Geol. Assoc. Canada 19, 1-13. Griffin, W.L. and Murthy, R.V. (1968) Abundances of K, Rb, Cs, and Ba in some ultramafic rocks and minerals. Earth Planet. Sci. Lett. 4, 497-501. Griffin, W.L., et a1. (1967) K/Rb in amphiboles and amphibolites from northeastern Minnesota. Earth Planet. Sci. Lett. 3, 367-370. Hart, S.R. (1969) Isotope geochemistry of crust-mantle processes. The Earth's Crust and Upper Mantle, P.J. Hart, ed., p. 50. 34 Hart, S.R. and Aldrich, L.T. (1967) Fractionation of K/Rb by amphiboles: Implication regarding mantle compo- sition. Science 155, 325-327. Hart, S.R., et a1. (1970) Ancient and modern volcanic rocks: A trace element model. Earth Planet. Sci. Hedge, C.E., et a1. (1970) Sr isotopes in some cenozoic lavas from Oregon and Washington. Earth Planet. Sci. Lett. 8, 434-438. Heier, K.S. (1964) Rb/Sr and Sr-87/Sr-86 ratios in deep crustal material. Nature 202, no. 4931, 477-478. Heier, K.S. (1970) Geochemical aspects of the Lofoten rocks. Norks. Geolggisk Tidsskrift 50, no. 3, 273-274. Heier, K.S. and Adams, A.S. (1964) The Geochemistry of the alkali metals. Physics and Chemistry of the Earth 5, 253-382. Hurley, P.M. (1968) Absolute abundance and distribution of Rb, K, and Sr in the earth. Geochim. Cosmochim. Acta 32, 273-283. Hurley, P.M. (1968) Correction to: Absolute abundance and distribution of Rb, K, and Sr in the earth. Geochim. Cosmochim. Acta 32, 1025-1030. Jakes, P. and White, A.J.R. (1970) K/Rb ratios of rocks from island arcs. Geochim. Cosmochim. Acta 34, 849-856. Krauskopf, K.B. (1967) Introduction To Geochemistry, McGraw-Hill, New York. Lambert, 1.8. and Heier, K.S. (1968) Geochemical investi- gations of deep-seated rocks in the Australian Shield. Lithos 1, 30-53. Lange, I.M., et a1. (1967) K/Rb ratios in coexisting K- feldspars and biotites from some New England granites and metasediments. Chem. Geol. 1, 317-322. Lanphere, M.A. (1968) Sr, Rb, K and Sr isotopic relation- ships in ultramafic rocks, Southeastern Alaska. Earth Planet. Sci. Lett. 4, 185-190. Poldervaart, A. (1955) Chemistry of the earth's crust. In Crust of the Earth--A Symposium (editor A. Polder- vaart). Geol. Soc. Amer. Spec. Paper 62, 119-144. 35 McDougal, I. and Compston, W. (1965) Sr isotope composition and K/Rb ratios in some rocks from Reunion and Rodriguez, Indian Ocean. Nature 207, 252-253. Murthy, V.R. and Griffin, W.L. (1970) K/Rb fractionation by plagioclase feldspars. Chem. Geol. 6, 265-271. Ozima, M., et a1. (1971) K/Rb ratios of dredged submarine basalts. Earth Planet Sci. Lett. Philpotts, J.A. and Schnetzler, C.C. (1970) Phenocryst- matrix partition coefficients for K, Rb, Sr, and Ba, with applications to anorthosite and basalt genesis. Geochim. Cosmochim. Acta 34, 307-322. Rhodes, J.M. (1969) On the chemistry of K feldspars in granitic rocks. Chem. Geol. 4, 373-392. Ringwood, A.E. (1969) Composition and evolution of the upper mantle. Amer. Geophys. Union Mono. 13, l-l7. Russel, R.D. and Ozima, M. (1971) The K/Rb ratio of the earth. Geochim. Cosmochim. Acta 35, 679-685. Shaw, D.M. (1968) A Review of K/Rb fractionation trends by covariance analysis. Geochim. Cosmochim. Acta 32, 573-601. Shaw, D.M., et al. (1967) An estimate of the chemical com- position of the Canadian Precambrian Shield. Can. Jour. Earth Sci. 4, 829-854. Spooner, C.M. (1968) Sr-87/Sr-86 initial ratios in charnockites and pyroxene granulites. M.I.T. Six- teenth Annual Progress Report, 57-66. Spooner, C.M. (1969) Sr-87/Sr-86 initial ratios and whole rock ages of pyroxene granulites. Ph.D. Disser- tation, Mass. Inst. Tech. Spooner, C.M. and Fairbairn, H.W. (1969) Rb, Sr initial ratios and ages of some charnockites and pyroxene granulites. Amer. Geophys. Union. Program 1969 Annual Meetipg. Steuber, A.M. and Murthy, V.R. (1966) K/Rb ratio in ultra- mafic rocks: Differentiation history of the upper mantle. Science 153, 740-741. Tatsumoto, M., et al. (1965) K, Rb, Sr, Th, U, and the ratios of Sr-87/Sr—86 in oceanic tholeiite basalt. Science 150, 886-888. 36 Taubeneck, W.H. (1965) An appraisal of some.K/Rb ratios in igneous rocks. Jour. Geophys. Res. 70, 475-478. Taylor, S.R., et a1. (1956) Some anomalous K/Rb ratios in igneous rocks and their petrological significance. Geochim. Cosmochim. Acta 10, 224-229. Taylor, S.R., et a1. (1969) Trace element abundances in andesites II, Saipan, Bougainville and Figi. Con- trib. Mineral. Petrol. 23, 1-26. Taylor, S.R. and White, A.J.R. (1965) Geochemistry of andesites and the growth of continents. Nature 208, 1271-73. Turekian, K.K. (1963) The use of trace element geochemistry in solving geologic problems. Studies in Analytical Geochemistgy, The Royal Society of Canada, Special Publications, no. 6, 3-24. White, A.J.R. (1966) Genesis of migmatites from the Palmer region of South Australia. Chem. Geol. 1, 165-200. Whitney, P.R. (1969) Variations of the K/Rb ratio in mag- matic paragneisses of the Northwest Adirondacks. Geochim. Cosmochim. Acta 33, 1203-1211. Windley, B.F. (1970) Anorthosites in the early crust of the earth and moon. Nature 226, 333-335. Windley, B.F. (1969) Anorthosites of southern West Green- land. Geological Survey of Greenland, 899-915. AAPG Memoir 12. 899-915. APPENDIX APPENDIX ANALYTICAL PROCEDURES The K analysis were determined using dilution methods on a Perkin-Elmer model 303 atomic absorption spectrophotometer utilizing a K hollow cathode lamp. A11 Rb determinations were made by isotope dilution methods, as outlined by Reesman (1968), using a 6-inch, 60° sector, solid-source, single filament, Nier-type mass spectrometer utilizing a Cary model 31 vibrating reed electrometer. Analytical precision was determined by running all samples in duplicate. Accuracy was determined by repeated analysis of G-l. Six samples run on a General Electric XRF unit, using a PET crystal and utilizing procedures outlined by the x-ray lab at the Michigan State University Department of Geology. The six samples agreed reasonably well with their atomic absorption counterparts. A Coleman flame photometer-galvanometer was used in the analysis of samples, but data was discarded due to inconsistency of readings and instrumental drift. 37 38 Laboratory Procedure In the laboratory several techniques were followed consistently to minimize error in methodology. All glass- ware, plastic bottles, and teflon beakers were washed with soap and tap water, rinsed with doubly distilled water, then rinsed again with 2N HCl and doubly distilled water and finally with acetone to dry. Class A volumetric flasks were used throughout the study to prepare solutions. All solutions were then stored in plastic bottles (4 oz.) with taped screw-on caps to minimize evaporation losses. Preparation of Standards Fresh Baker analysed KCl (99.9% pure, and of Ana- lytical Reagent grade) was used to prepare K standards. An amount of approximately 2.1 grams of KCl was weighed out in a dry, clean weighing dish and then placed in an oven to dry overnight at 110°C. The sample was then removed and placed in a desiccator and allowed to cool to room tempera- ture before weighing. The exact amount of KCl needed to make a 1000 ppm K solution per liter (1000 ugm/ml) was then calculated using the following values for KCl: 73 II 39.102g Cl 35.4539 39 The amount needed, from the ratio: _K_ = 39.1029 = 1 KCl 74.5559 1.90678 An amount of 1.90699 was weighed out on a Mettler Electronic Scale precise to within 0.0002 grams. Glassine paper was used to contain the KCl during actual weighing. The KCl was then placed in a clean 1000 ml volumetric flask along with the residue on the glassine which was then filled to the mark with 0.1N HCl. From the 1000 ppm K standard 100 ml was pipetted off and placed into a clean 1000 m1 volumetric flask and then filled to the mark with 0.1N HCl thus giving a 100 ppm/ml K standard. Aliquots of 100, 50, 25, and 10 ml were pipetted from the 100 ppm K standard and placed into the corres- ponding flask and each diluted to the mark with 0.1N HCl. Each respective standard was then shaken well and trans- ferred to a labeled plastic bottle and sealed with elec- trical tape until needed. Preparation of Samples The sample was then placed into a clean teflon beaker, including the residue on the glassine which was washed with quartz distilled water. Approximately 10 ml of HF acid (48%i) was placed into the beaker with the sample and placed in a fume hood on a steam bath with the beaker 40 cover slightly open to allow evaporation. A heat lamp was directed onto the beaker to prevent a condensate on the beaker lid. After the sample dissolved (2 hours to 2 days) and went to dryness, it was taken back into solution by adding enough 0.1N HCl to dissolve (may need heating). After dissolving, the contents were quantitatively trans- ferred into a 100 m1 volumetric flask using a funnel, stir- ring rod and care. The dissolved residue was then allowed to trickle down the stirring rod with its end placed against the side of the funnel. The sides of the beaker and the stirring rod were washed and rewashed with quartz distilled water and transferred into the flask through the funnel. The funnel was then washed clean, into the flask, and the flask filled to the mark with quartz distilled water and shaken well. Ten ml of solution were then pipetted from the above into a clean 100 m1 flask and filled to the mark with 0.1N HCl and again shaken well. For the F.P. analysis, 10 m1 of this dilution was pipetted into each of five labeled bottles plus an addition of 1 ml of each respective stan- dard to a respective bottle such as the following: Bottle #1 10 ml soln Bottle #2 " " " + 1 ml of 1.0ppm std Bottle # 3 II II II II II II 2 . 5 H II Bottle # 4 II II II I! II II 5 . 0 H II Bottle # 5 " II II II II I! 10 . 0 I! II Bottles were then capped, shaken well and sealed with tape until ready to use. If solution was too concentrated a 41 10 ml aliquot was pipetted from the first dilution bottle and repeated as above. Flame Photometer A Coleman model flame photometer was used to determine total K. The scale was set on the galvanometer with a top end of 10 ppm equal to 100 per cent transmissability. Each sample was run 4 times with the atomizer being flushed after each run. Scale checks were made after every four samples. The values were averaged and plotted on linear coordinate graph paper using the addition method. Values of the unknown were then determined graphically and this value divided by weight of sample to determine %K per gram. Atomic Absorption A Perkin-Elmer model 303 atomic absorption spectro- photometer utilizing a potassium hollow cathode lamp was used. The scale was set with a range of 0 to 10 ppm. Each sample was run 4 times with the atomizer being flushed with distilled water after each run. Zero setting checks were made after each sample was run. Each sample and standard was spiked with 1000 ppm NaCl to eliminate any effects by Na in the sample readings. 42 Values were averaged and plotted on linear coordinate graph paper using normal determinative methods. Values of the unknown were determined graphically and this value divided by weight of sample to determine %K per gram. X-Ray Fluorescence Samples were prepared using standardized procedures outlined by the Department of Geology at M.S.U. for major light element sample preparation using the heavy absorber method. Prepared samples were run on a General Electric XRF unit, using counts per 100 sec. method. The instrument was standardized at 50.640 - 28 for K and the x-ray tube volt- age at 49 KV. Rocks G-l, W-l, BCR-l, AGV-l were utilized in setting up a determinative curve for K (%). Values were plotted on linear coordinate graph paper and the %K deter- mined for the prepared samples. AAAAAAAAAAAA “‘3‘ W W.“ (1) NEW?“ L 3 ”'TiiiuvLuuB 2