:_:_:.:___E::__3:52:_E_:___::_:_‘ mm gran o I: 2:4; ; 1; iv in 3.“..3jtkfg 9 1 t I I 2 *2...» :0“ t A ;’ ‘~. m V .q a a fl . ‘ O '- 1‘! .t‘ Fn": D. on 2 ' 134:. ; 6.3.3.!Qs A}! . .. o’u\l :. S 1‘ ‘33: .. o. A r 1' . : A ‘l U; 1.“ 3'713 2:227 )1 d": I ‘n'ka' LIB: 43y Michig; .1 State Univcrsity ABSTRACT SOME EVIDENCE ON THE VALIDITY OF THE EXSOLUTION MODEL FOR THE FORMATION OF ANTIPERTHITES By Richard J. Wharton In order for antiperthites to form by an exsolution process. Ca, Na, K, A1, and Si must all be re-arranged throughout the crystal. Alternatively, if an anti- perthite forms by “nucleation“. the ions need only be re-arranged at the sites of K—feldspar growth. Regard- less of the mode of formation. a volume of potassium feldspar must replace a volume of plagioclase. The validity of the exsolution model was tested by determining potassium solubility in calcium-bearing plagioclase. "Clean“ plagioclase was heated both dry and hydrothermally in the presence of K-feldspar. The result was the growth of K—feldspar blebs on the plagio- clase with concomitant potassium and sodium gradients increasing. respectively, towards and away from the blebs. Very little potassium was found dispersed within the plagioclase. These results imply that insufficient potassium is held within a host plagioclase to form an antiperthite by exsolution. SOME EVIDENCE ON THE VALIDITY OF THE EXSOLUTION MODEL FOR THE FORMATION OF ANTIPERTHITES By \ Richard J. wharton A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geology 1972 4 «[197 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to: Dr. T. A. Vogel. Without his invaluable assistance and interest throughout the study, it would not have been com- pleted. Dr. R. Ehrlich. for reading the manuscripts and contri- buting valuable constructive remarks. Dr. C. Spooner, for his assistance on the theoretical aspects of the problem. Dr. J. Fawcett and the Geology Department of the University of Toronto, for their cooperation and use of experimental apparatus. ii TABLE OF CONTENTS LIST OF TABLES..................................... LIST OF FIGURES.................................... LIST OF PLATES..................................... INTRODUCTION....................................... EXPERIMENTAL PROCEDURES............................ ONE ATMOOOOCOOCCOCOOOOOIOCOIOOII.0'.0.0.0.000. HYDROTHERMAL.................................. DATA............................................... RESULTS OF lATM. AND 2 KBAR EXPERIMENTS....... NATURAL ANTIPERTHITES......................... INTERPRETATION OF DATA.................¢........... DISCUSSION......................................... SUMMARY............................................ LIST OF REFERENCES......................o.......... GENERAL REFERENCES................................. iii Page iv vi VGU‘UNNH 17 21 23 25 26 Table 1. 2. LIST OF TABLES Weight Percentages of K-feldspar. 1 ATM.’ Dry.OOOOOCOCOOOOOOOCOOOOOOI0.0.0000... Weight Percentages of K—feldspar, 2 Kb0.0000000000000COOOOOOOCOC.OOCOOOOOOOOOOOO Page Figure 1. 2. 3. h. 5. 6. 7. 8. N-S E-W N-S N-S N-S N-S N-S probe probe probe probe probe probe probe Model for trace trace trace trace trace trace trace LIST OF FIGURES of of of of of of of grain grain grain grain grain grain grain 16..................... 20..................... 6...................... 8...................... 9...................... 10..................... 15000000000000.0000...- the concentration of potassium and the grOWth Of K-feldspar blebso o o o o o o o o o o o o o 13 13 1“ 1h 15 15 16 19 LIST OF PLATES Plate Page 1. Grain 16 (lATM. Dry) 5u/cm.................. 8 2. Grain. 20 (1ATM. Dry) 2u/cm.................. 8 3. Grain 6 (Hydrothermal, 2 Kb.) 6.6u/cm....... 9 #. Grain 8 (Hydrothermal. 2 Kb.) 8.2u/cm....... 9 5. Grain 9 (Hydrothermal, 2 Kb.) 6.0u/cm....... 1o 6. Grain 10 (Hydrothermal, 2 Kb.) 6.0u/cm...... 10 7. Grain 15 (Natural antiperthite. hydrothermal. 1.33 Kb.) 2u/cm. Electron Photograph........ 11 80 Grain 15 Na‘phOtographooooooooooooooooooooco 11 90 Grain 15 K-photograph........oc.o........... 12 vi INTRODUCTION There are two major models for the formation of antiperthites. the exsolution.model and the "nucleation” model. The exsolution model, the classic model accepted by most (Barth. 1969), consists of a ternary feldspar unmixing to form potash feldspar intergrowths within a calcium-bearing plagioclase matrix or host during cooling. The alternate model, “nucleation“, is based on an outside source of potassium, nucleating on plagioclase surfaces or discontinuities within the plagioclase. The only difference in the two models is the source of potassium. Both models. of course. require nucleation to initiate the growth of the K-feldspar. Therefore, the nucleation for each model is controlled by high-energy discontinu- ities within the crystal. The textural pattern produced by either the exsolution or ”nucleation" model is. there- fore, identical. Perthites. which consist of albite in a K-feldspar matrix. are easily homogenized at temperatures slightly below the solidus. Thus, the exsolution model for the formation of perthites is supported. If a similar exsolu- tion model for the origin of antiperthites is preposed. one test of the model is to determine if antiperthites can be homogenized under reasonable geologic conditions; this is the purpose of this research. EXPERIMENTAL PROCEDURES Three different experimental procedures were used to help alleviate any problems of insufficient or excess energies or error in procedure. First. a ”clean" plagio- clase and K-feldspar were heated together to just below the solidus at atmospheric pressure and held for 72 hours. A second run of the same components was conducted at high temperature and two Kbars for 48 hours. Finally, a natural antiperthite was run at low temperature and 1.33 Kbars for 92 hours. Any variation in potassium content or potassium gradient due to a change in the experimental environment can thus be noted. For the 1 atm. and the high water pressure runs. the starting materials were two natural feldspars. Bedford Perthite (a source of K: 90.67% Cr. 9.33% Ab) and 1-87-11 plagioclase (10.24% Can 48.4% Ab, 49.2% An, 2.4% 0r) mixed together in varying proportions. In the moderate water pressure run, the starting material was a natural antiperthite. ONE ATM. -Both starting materials (Redford Perthite and 1-87—11 plagioclase) were finely ground to ~140 mesh and homogen- ized by heating to 1050°C for three days to eliminate any pre-existing crystallographic differences involving sodium and potassium. Homogenization was substantiated by both x-ray and microprobe techniques. 2 3 The two feldspar components were then mixed in various weight proportions to insure a potassium saturated condition in the plagioclase. Table 1. Weight Percentages of K—feldspar. 1 ATM., Dry SAMPLE WT. % SAMPLE WT. % NUMBER K-FELDSPAR NUMBER K-FELDSPAR 1 0.5 6 10.0 2 . 1.0 7 15.0 3 2.0 8 25.0 4 3.0 9 90.0 5 5.0 10 95.0 Each mixture was then placed in an open boat and heated to 10501200 for 72 hours. At the end of 72 hours. the boats were removed to room temperature; cooling time was about 10 minutes. The charges were then mounted for microprobe examination. HYDROTHERMAL The experiment was then repeated hydrothermally: first with the same starting materials through the facil- ities at the University of Toronto and secondly with a natural antiperthite using a similar apparatus at Michi- gan State University. As in the high temperature experiment, the starting materials for the high pressure run (Bedford Perthite and 1-87—11 plagioclase) were heated dry to 1050°C for 72 hours to eliminate sodium-potassium and sodium—calcium crystallographic differences. The feldspars were then 4 mixed in varying proportions as shown in Table 2. Table 2. Weight Percentages of K-feldspar, 2 kb. SAMPLE WT. % BOMB N0. SYSTEM NUMBER K-FELDSPAR ggg 1-87-11 0.0 11 closed 3' 2.0 11 closed 3” 2.0 11 closed 4' 3.0 13 open 8" 25.0 13 closed A' 50.0 13 closed 9“ 90.0 13 closed '==homogenized feldspar "==charges used in 1 ATM. experiment The method of loading the charges was that of Fawcett (personal communication. 1970). The sealed charges were then placed in the cold seal pressure vessel (Tuttle. 1949) with spacer rods keeping their position constant. Pressure was achieved in each bomb through a hydraulic pump and two large-volume water reservoirs maintained at 1000 and 2000 bars. The purpose of the two reservoirs was to minimize both the pumping time and the effects of minor leaks in the pressure system. Pressure and temperature were checked at eight hour intervals. Pressure showed no measurable fluctuation. reading a constant 2 Kb. pressure. Temperature varied 677.811.706 on bomb no. 13 and 677.0t2.o°c on bomb no. 11. At the end of 48 hours. the bombs were quenched with cold water and compressed air. reaching room temperature and 1 ATM. pressure in approximately 10 minutes. The charges were removed. dried for 10_minutes at 5 110°C and mounted for microprobe analysis. The hydrothermal run was then repeated using a natural antiperthite. The instrument used was the same with the exception of the large water reservoirs. which were omitted. The charges were maintained at 1.33 Kbi500 p.s.i. and 500.01200 for 92 hours. They were then quenched and mounted. All micrOprobe work was done on an ARL Model EMXSM Electron Microprobe with a line voltage of 15 Kvolts and a current of .02 Mamps. Both N-S and E-W traverses were done with point counts being taken at critical points along the grain (grain boundaries. to and across K—feldspar blebs. or any anomalies in the calcium. sodium or potassium content found during the scan). In all probe work. care was taken to count sodium first. as it tends to vaporize under the beam. DATA Approximately the same results were obtained from the two runs with “clean“ materials. the atmospheric pressure run and the 2 Kbar experiment. The run at 1.33 Kbars and low temperature involving a natural anti— perthite gave slightly different results. but not contra- dictory to the previous two experiments. The purpose of these three runs of considerably different conditions was. in fact. two fold: first. to determine whether potassium could be held in solution within a plagioclase 6 grain under reasonable conditions. and secondly to see if the results were confined to one system. i.e. to one particular energy environment. The importance of the data lies not in the absolute percentages of potassium or sodium in a given grain, but in the relative amounts. both at bleb sites and that which is being driven through the grain. ESULTS OF 1 ATM. AND 2 KBAR EXPERIMENTS One of the most important results of these exper- iments is the development of K-feldspar blebs (small areas of actual potassium feldspar). primarily on or near grain boundaries. and the corresponding potassium gradients into. and sodium gradients away from. the K-feldspar locations. Plates 1 and 2 are two plagioclase grains after heating to 105000 at 1 ATM. Plates 3. 4. 5 and 6 are plagioclase grains heated to 677°C at 2 Kbars pressure. 0n the gradient. potassium increases from a low of 2.2 weight percent K-feldspar to approximately 2.8 weight percent next to the K-feldspar bleb. In one case. grain 8. a second area of high K-feldspar concen- tration appears to be forming on the opposite grain boundary. with a 4.15 weight percent K-feldspar composition. There is also a concomitant gradient with sodium. however with a negative slope. This same type of potassium concentration along grain boundaries can be seen in the traces of grains 9 and 16. and. to a lesser degree. in 6 grain under reasonable conditions. and secondly to see if the results were confined to one system. i.e. to one particular energy environment. The importance of the data lies not in the absolute percentages of potassium or sodium in a given grain. but in the relative amounts. both at bleb sites and that which is being driven through the grain. RESULTS OF 1 ATM. AND 2 KBAR EXPERIMENTS One of the most important results of these exper- iments is the development of K-feldspar blebs (small areas of actual potassium feldspar). primarily on or near grain boundaries. and the corresponding potassium gradients into. and sodium gradients away from. the K-feldspar locations. Plates 1 and 2 are two plagioclase grains after heating to 105000 at 1 ATM. Plates 3. 4. 5 and 6 are plagioclase grains heated to 677°C at 2 Kbars pressure. 0n the gradient. potassium increases from a low of 2.2 weight percent K-feldspar to approximately 2.8 weight percent next to the K-feldspar bleb. In one case. grain 8. a second area of high K-feldspar concen- tration appears to be forming on the opposite grain boundary. with a 4.15 weight percent K-feldspar composition. There is also a concomitant gradient with sodium. however with a negative slope. This same type of potassium concentration along grain boundaries can be seen in the traces of grains 9 and 16. and. to a lesser degree. in 6 grain under reasonable conditions. and secondly to see if the results were confined to one system. i.e. to one particular energy environment. The importance of the data lies not in the absolute percentages of potassium or sodium in a given grain. but in the relative amounts. both at bleb sites and that which is being driven through the grain. RESULTS OF 1 ATM. AND 2 KBAR EXPERIMENTS One of the most important results of these exper- iments is the development of K-feldspar blebs (small areas of actual potassium feldspar). primarily on or near grain boundaries. and the corresponding potassium gradients into. and sodium gradients away from. the K-feldspar locations. Plates 1 and 2 are two plagioclase grains after heating to 1050°C at 1 ATM. Plates 3. 4. 5 and 6 are plagioclase grains heated to 677°C at 2 Kbars pressure. 0n the gradient. potassium increases from a low of 2.2 weight percent K-feldspar to approximately 2.8 weight percent next to the K—feldspar bleb. In one case. grain 8. a second area of high K-feldspar concen- tration appears to be forming on the opposite grain boundary. with a 4.15 weight percent K-feldspar composition. There is also a concomitant gradient with sodium. however with a negative Slope. This same type of potassium concentration along grain boundaries can be seen in the traces of grains 9 and 16. and. to a lesser degree. in 7 grain 20. Grain 10 has no high concentrations of K-feldspar but does show potassium and sodium gradients. The plagioclase grains before heating were "clean”. that is. they contained no K-feldspar blebs. Grain 10 apparently had no suitable nucleation Site. and. there- fore. no K-feldspar bleb formed. The result is a slight gradient in potassium with an inverse sodium gradient. It should be noted that no more than 2.8 weight percent K-feldspar is found in this grain. This also appears to be the maximum amount in other hydrothermal grains when the actual K-feldspar site is discounted. The plagioclase heated to 105000. under atmospheric pressure. contains more K-feldspar (4.8 weight percent) than the hydrothermally treated plagioclase (about 2.8 weight percent). Volume diffusion increases with in- creasing temperature (Bailey. 1971) and these results apparently reflect the increase in diffusivity of potas- sium with temperature. NATURAL ANTIPERTHITES Several natural antiperthite grains were studied under the E.M.P.3 grain 16 is a typical example (see Plate 7). The natural antiperthites were subjected to a lower energy level (500°C. 1.33 Kbars) for a longer period of time (96 hours). and. as is illustrated in the probe trace (Figure 7). there are no observable gradients of either sodium or potassium. It seems .4 4 4 .1 .4 J dr- IIIIdVY'l TT LL11 . l r Plate 1. Grain 16 (1 ATM. Dry) 5u/cm . Plate 2. Grain 20 (1 ATM. Dry) 2u/cm. Plate 3. Grain 6 (Hydrothermal. 2 Kb.) 6.6u/cm. Plate 4. Grain 8 (Hydrothermal, 2 Kb.) 8.2u/cm. 10 Plate 5. Grain 9 (Hydrothermal. 2 Kb.) 6.0u/cm. Plate 6. Grain 10 (Hydrothermal. 2 Kb.) o.0u/om. 11 Plate 7. Grain 15 (Natural anti- perthite hydrothermal. 1 .33 Kb. 2u/cm. Electron Photograph. Plate 8. Grain 15 Na-photograph 12 . .' I . s . ‘4‘ ye¥533335’@9;9”$*”‘*mgfi“fifiiei-7 ."*‘. '_ ‘~.~‘Ao. ' {Cw-firs: Plate 9. Grain 15 K-photograph 13 Grain 16 N-S "H 40 1 j s a .. 1 K ’57 f -55 30-: —55 °\0 a; .1 3}. f} 3 ‘ f/Ca. sl ‘7 f 1 - ‘\.‘ J.“ 3‘ 2.0‘1 1 / - k. 47 j P I ‘- 45 3 43 [.01 s i . ‘ I l 1 ' ' Figure 1. N-S probe trace of grain 16. A “N 2 a 1: 9" (A A r% é- ‘1 J 4" {1" ”w b N" 35’1‘111’1" AC'A‘ J ”fl”; "M Na. "i (r N 1 +1 y .1 w Y 1 b .r 1m? ‘5' ‘1:— ' V- d ‘- 5:" w .5 «£151. is," {7 “1 :o 0 J0 «H ‘b 5’ “is 5 3 ‘ :2 . 3;; { £5. 6 4%: :L'; 7' n 1’: w :r or a... n- “A- ————— Figure 2. E-W probe trace of grain 20. 14 "60 .b ‘ L55 6.6 )1 Grain 6 N' S K ’5‘! WWW) M / . .1 Na in : r““WW WW “ Wk :2: l 1 1 1' Y 1 I I l I t 1 Y r v I v r ‘0 1 I ma Jo 1 *50 1 ~18 k'JIO4f ‘W 9'. [-5 lo 25‘ So 3.)" 10 va’ 1 l 1.1-1 ll 1 xx \\ ‘V. Figure 3. N—S probe trace of grain 6. Grain 8 ”OT N- S T54 Iaai 8 )4 'Sz 70: (WA I’L/Nd .50 80-: W» (VAX WA" '45 w 7 3 VA VI)?» _ «as 0 Q '0‘. s '- ‘ 0\ (L I -‘ - V) 1 11 b k 6.0? Cd. (3.. 33 $01 -4Z 4.0.3 40 5°": .1 NMFWW P58 1 I“ I - 36 20; N \ / Figure 4. N-S probe trace of grain 8. IIT'rrvv 11 .a.- , .1.-. .h., -1 , .......... 3? 1 $2 ' A- ' $3 2 Ab’JF-y O/O Figure 5. N-S probe trace of grain 9. Grain 1O N-S £0? .—6.L T54 1 ”ATHWWW ;' J 5,4 ' 56 5.04 r f. NAM), ) M I/C. .50 i A *5). F MN \. ‘8 1.0-a: A. ~ WA 46 ~ ~ ‘2: S“1 K ' w s“ ~; r MWWWWW o\°zo—§ ' . ~38 ‘1 A f A) ’36 10.: K .34 Figure 6. N-S probe trace of grain 10. 16 loci 2 “31‘s MK __._,_._. 5) . Na 2; .‘M f) .11 ("1.2; fl iN’ W6) \h \C‘v‘r HWWMVMAWM) ~ (1' (W WW '4 (‘H‘Jy Nam) .1 AT“ 5‘3 s r «b a j A k V j . 0 0 W Jr \ M! '\l “AW ”InNW1'WHWQ N , 4° Figure 7. N-S probe trace of grain 15. 16 \\ ma; “'31:“ /K 55 ‘_ 1‘ ,/M Nd‘ . . AH; “A “E“ 'A ”k :43 @215!“ WMV WWR WW, {x}! {ff} \‘WN ‘Wnlrk J'il M“ “It“ ‘03 >‘c W h «5 °\° E '3! k .4?" 00‘: 1 “‘3“?ka WWI ,J'xwyn~w’\n‘,w”~¢hh'~k~4° // I \\ >55 L \ K 56 Figure 7. N-S probe trace of grain 15 16 we; / l \ Gain 15 N-S 4 q 4 4 4 .. 4 4 ‘1 l'. Ill In /N& " ‘ '\ ' p I ! . l - ”-M l I L43 N ' . 4 6 l ', U ‘n‘ a sg't .11“ §‘€%WVW “A W fl ‘F?*““£ :‘c 1 ( V 4.1:. ‘ 0" °\0 3 N ~41 0 0" 'fiW~*W/‘W\ ’RWMNWHWMMW * 4 ° .1 ' ,58 E “5.. I \ § Figure 7. N-S probe trace of grain 15. 1? evident that. for any migration of sodium or potassium to occur. there must be a high energy environment. Even then. the actual weight percent held in ”solution” appears to be very low. INTERPRETATION OF DATA These experiments were designed to test the validity of the exsolution model. If the exsolution model were viable it would be expected that, under conditions near the solidus (in the high temperature experiment, some grain boundaries were fused), some evidence of solid solution between K-feldspar and plagioclase would be present. This is not the case and the present author considers that the exsolution model for antiperthite formation is not valid. A surprising result of these experiments was that antiperthitic plagioclase was produced; there are potas- sium feldspar concentrations on high energy surfaces (grain boundaries. crystal defects. etc.). Even though there was sufficient energy within the system to mobilize potassium and sodium. as evidenced by their respective compositional gradients. very little potassium was found dispersed in the plagioclase matrix. If the exsolution model were viable. the potassium diffusing into the plagioclase to form the K-feldspar bleb should have homogenized in large proportions within the host. Under hydrothermal conditions. 2.8 weight 18 percent K-feldspar is the maximum held within plagioclase. Grain 6 is a good example of this. The potassium can be seen to increase from approximately 2.2 weight percent K—feldspar on the far edge to approximately 2.8 weight percent next to the bleb. In conjunction with this. grains 8. 9 and 10 support the hypothesis of low potassium solubility. In some cases (grains 8 and 9 for example) it appears that there is a concentration of potassium into domains. This is interpreted as being a nucleation phenomena since they are always associated with grain boundaries. In both of these examples there is a local enrichment of potassium up to 4.0 weight percent K-feldspar. Grain 10. on the other hand. shows no concentration of potassium. Apparently there was no suitable nucleation site available. and no K—feldspar bleb formed, merely gradients of sodium and potassium. This data is not inconsistant with a solute-vacancy complex model. a model which has been well supported by experimental evidence in ceramic systems (Westbrook, 196?). This model. in summary, is that at high temperatures. there will be a supersaturation of vacancies and a certain fraction of these are theoretically and experimentally associated with a vacancy-impurity complex (Westbrook. 1967). These vacancies and the associated vacancy-impurity complex will decay during the quench toward high energy surfaces. Thus. as the crystal tends to relieve itself of the vacancies, there will be a net flow of vacancies 19 /////// a- ---dp.----.k- \ \ \ I let-SPA& : K‘O 'GRow‘rH : Nd- , / : , 1 [(+0 ' é—¢———K,u \ + I —E+Kv___!9 ' Plagioclase : Nd; .—Jii_, : (k 6 K+o W f/ s \ Figure 8. Model for the concentration of solute-vacancy complexed potassium and the growth of K-feldspar blebs. 20 and impurities toward thses surfaces. This process will favor a dissociation of the vacancy-solute complex near the grain boundary. thus producing a build-up of free solute atoms near these high energy surfaces. As this model applies to the antiperthite system. the potassium is the solute in complex with the vacancies. Due to the existing solid solution of the alkali feldspars. the K-feldspar would act as an unlimited sink for sodium. Contrary to this. plagioclase would be a limited sink for potassium. The result is a free movement of sodium into the K-feldspar and a concomitant restricted diffusion of potassium into the plagioclase. Decay of potassium- vacancy complexes could occur on any high energy surface (grain boundaries. either in contact or as free surfaces. twin planes, crystal defects. etc.) with the growth of vpotassium feldspar proceeding outwards. Indeed, a large number of such surfaces may actually be a driving force for the concentration of potassium within the plagioclase. As stated above, plagioclase is greatly restricted to the amount of potassium. and the data indicate that the limit is approximately three weight percent K-feldspar. Regardless of the exactness of this model, it is clear that the K-feldspar forms along high energy surfaces if such surfaces are available, and that the host plagio— clase is incapable of holding sufficient quantities of potassium in "solution” to form exsolution antiperthites. DISCUSSION Both models for the formation of an antiperthite. the exsolution model and the “nucleation” model. involve the replacement of a volume of plagioclase feldspar by a volume of potassium feldspar. Because of the two feldspar solid solutions. there is little problem with the mobility and replacement of Na and K or Na and Ca. Thus it is logical to assume that these ions will be randomly distributed throughout the original feldspar in a high-temperature disordered state. Because of this random distribution. the feldspar may be considered homogeneous. With the exsolution model. three simultaneous dif- fusions must take place: (1) diffusion of potassium from matrix to potential bleb site. (2) diffusion of calcium and sodium away from bleb site. (3) re-propor- tionment of aluminum and silicon to the proper ratios. both inside and outside the bleb. The first two. the diffusion of potassium and the diffusion of sodium and calcium are of no major consequence. involving little crystal structure distortion. The re-proportionment of aluminum and silicon presents a greater problem. since the Al-O and 31-0 bonds are the strongest found in feldspar. In either model these tetrahedral bonds are broken: in the exsolution model. the readjustment takes place through- out the plagioclase grain with many Al-Si-O bonds being brOken. whereas. in the ”nucleation" model the only 21 22 disruption is at the K-feldspar location. As stated by Vogel (1970). ”regardless of the origin of the potash feldspar. the growth of the potash feldspar within the plagioclase grain involves the geplacement of a volume of plagioclase by a volume of potash feldspar". Several researchers (Vogel. Smith and Goodspeed. 1968; Griffin. 1969; Vogel. 1970) have cited evidence that this replacement is not a result of exsolution and its thermodynamic explanation. but rather “nucleation” energies and kinetics. Griffin (1969) referred to the occurrence of zoning (halos and rims) around the antiperthitic blebs as support of the "nucleation” theory. These zones would not be present if the blebs formed by exsolution. but rather. would become homogenized as exsolution took place. A study by Vogel. et. a1. (1968) suggested a twofold hypo- thesis: '1. The exsolution of microeline from plagioclase is permitted by the structure of plagioclase between An33 to An50.” or ”2. The development of antiperthite is dependent upon the feldspar phase that crystallized first.“ Although the present study involved plagioclase within the An33_50 range. the author believes the micro- probe results clearly support the latter hypothesis. Several researchers (Sen. 1959) distinguish between exsolution antiperthites and “replacement antiperthites”. However. as stated. earlier. regardless of the means of formation of the antiperthite. a volume of plagioclase 23 must be replaced by a vglgme of K-feldspar. Secondly. the only real distinction between the two antiperthites of Sen appears to be size. If there exists an excess of potassium migrating through the host plagioclase. it will grow on previously nucleated blebs. rather than forming new smaller blebs. The only factors that govern the size of the antiperthite formed are. therefore. the pressure- temperature environment. time. and the amount of potassium available for diffusion. It does not appear feasible to assume that larger amounts of potassium will be held in solid solution to exsolve into a coarse grained anti- perthite at higher temperatures as conditions in the present study were just below the solidus at two consid- erably different pressures. SUMMARY If the exsolution model is valid. then. at temper- atures near the solidus. a homogeneous potassium-rich plagioclase should be produced. This model was tested under both dry and hydrothermal conditions; the maximum amount of K-feldspar dissolved was less than 5 weight percent. Whether or not this small amount of potassium is actually in solid solution is a matter of conjecture. It may be that the potassium is being driven through the plagioclase along sub-micron defects. or in the case of the high temperature run. driven through the crystal lattice by volume diffusion (Bailey. 1971). 24 The formation of potassium and-sodium diffusion gradients as well as K-feldspar blebs within the plagio-. clase during these experiments supports a “nucleation” model for the formation of antiperthites. This data is consistent with Westbrook's (1967) solute-vacancy complex model for the diffusion of sodium and potassium and the formation of antiperthites. LIST OF REFERENCES 1. 2. 3. u. 5. 7. 8. 9. LIST OF REFERENCES Bailey. A. (1971). Comparison of Low-Temperature with High Temperature Diffusion of Sodium in Albite. Geochim. et Cosmochim. Acta. v. 35. pp. 1073-1081. Barth. T.F.W. (1969). Feldsnars. John Wiley and Sons. Inc. New York. Fawcett. Jeff (1970). Personal Communication. Griffin. W.L. (1969). Replacement Antiperthites in Gneisses of the Babbit-Embarrass Area, Minnesota. U.S.A. Lithos. v. 2. pp. 171-186. Sen. S.K. (1959). Potassium Content of Natural Plagio- clases and the Origin of Antiperthites. J. Geol. v. 67. pp. 479-495- Tuttle. O.F. (1949). Two Pressure Vessels for Silicate- Water StUdieSo G901. SOC. Am. B11110 V. 60' PP. 1727“17290 Vogel. T.A. (1970). The Origin of Some Antiperthites- A Model Based On Nucleation. Am. Min. v. 55. PP- 1390-1395. Vogel. T.A.. Bennett L. Smith. and Robert M. Goodspeed. (1968). The Origin of Antiperthites From Some Charnockite Rocks in the New Jersey PreCambrian. Am. Min. v. 53. pp. 1696-1708. Westbrook. J.H. (1967). Impurity Effects at Grain Boundaries in Ceramics. Science Of Ceramics. Proc.: Third Conference of the British Ceramic Society and the Nederlandse Kera- mische Vereniging. 5-8 July, 1965. Academic Press. New York. PP. 263-284. 25 2. 3. 5. 7. GENERAL REFERENCES Carstens. Harald (1967). Exsolution in Ternary Feldspars. I. On the Formation of Anti- perthites. Contrib. Mineral. Petrology. V0 in! PP' 27-35. Corlett. M. and P.H. Ribbe (1966). Electron Probe Micro-analysis of Minor Elements in Plagioclase Feldspars. Schweizer. Mineral. Petrogr. Mitt. V. 47,1313. 317‘3220 Goldsmith. J.R. (1952). Diffusion in Plagioclase Feldspars. J. Geol. v. 60. pp. 288-291. Hubbard. Fred H. (1967). Exsolution Myrmekite: A Proposed Solid—State Transformation Model. Geologiska Foreningensi Stockholm Forhandlingar. v. 89. pp. 410-922. Lindsley. D.R.. Douglas Smith and S.E. Haggerty (1970). Petrography and Mineral Chemistry of a Differentiated Flow of Picture Gorge Basalt Near Spray. Oregon. Carnegie Institute Yearbook. 1969-1970. pp. 264. Mueller. R.F. (1966). Mobility of the Elements in Metamorphism. J. Geol. v. 75. pp. 565-582. Waldbaum. D.R. (1968). High-Temperature Thermodynamic Properties of Alkali Feldspars. Contr. Mineral. and Petrol. v. 17. pp. 71-77. 26 ”'TITIiszl'IILflflfljjfi’[MMAMQI‘IQHMMITI'ES