THE AMERICAN MINERALOGIST. VOL 49, NOVEMBER DECEMBER, 196'1
THE BREAKDOWN OF POTASSIUM FELDSPAR, KAISigOS AT HIGH TEMPERATURES AND HIGH PRESSURES' Y6rAno SBr
KAlSirOs I HzO = KAlSi3O8' HrO Under dry conditions, horvever, potassium {eldspar is apparently stable even at 1000'C' and 60 kilobars. hqtnooucrroN There are four major groupsof feldsparminerals in nature: potassium feldspar (KAlSiaO8), sodium feldspar (NaAISiaOs), caicium-feldspar (CaAlzSizOr)and barium feidspar (BaAlrSizO)' Among these, sodium feldspar, calcium-feldspar and barium feldspar are easilv broken down into other phasesor phaseassemblages either at high solid pressureor at high water pressures. It is well known that sodium feidsparis transformedinto the assem- blagejadeite plus quartz at high pressureby the following chemicalreac- tion (Birch and Le Comte 1960;Robertson el al., 1957)'. albite jadeite ---. _Lqvartz' NaAlSirOs NaAlSizOo SiOz At higher pressures,albite transforms into the assemblageof jadeite and coesite b1' the reaction (Boyd and England, 1960; MacDonald, 1e56):
albite jadeite , coesite - NaAlSiaOs NaAlSizOe SiO, Pistorius et al. (1962) and Newton and Kennedv (1963) showedthat calcium-feldsparwould be transformedinto lawsoniteat high water pres- suresby the following chemicalreaction:
anorthite 2 water . lar'r'sonite caAlrSiros* 2 Hro caAlzsizos'2Hzo Barium feldspar also undergoes mineralogical change into cvmrite when subjected to high water pressures (Seki and Kennedy,1964a).
1Publication No. 306, Institute of Geophysicsand Planetar5'Physics, University of California,Los Angeles,California. 2 Departmentof Earth Sciences,Saitama University, Urawa, Japan. 1688 BREAKDOWNOF K FELDSPAR 1689
Early reconnaissanceexperiments showed that orthoclaseundergoes a phase changeat water pressuressomewhat in excessof 10 kilobars. In this paper we present results of experimentalwork on the stability of potash feldspar at high temperatures and high confining pressureas well as at high partial pressuresof water.
ExpnnrnBwrAl TECI{NTQUES Runs were made in both piston-anvil and piston-cylinderdevices. The behavior of a piston-anvil device (simple squeezer)has been dis- cussedseveral times (Griggs and Kennedy, 1956;Pistorius et al., t962). In "wet" runs water is added to the starting materialsin order to make confining pressuresand water pressuresas equal as possible. In "dry" runs, however,the starting materialswere kept atleast 24 hours at 110o C. in an air bath in order to eliminate all but the most tightly absorbed water from the starting materials.The sampleswere kept at the required temperature and pressure in the piston-anvil device for 2 hours above 500oC., 4 hours at 400o,and 8 hours or more below 400oC. Our piston-cylinderapparatus has been describedpreviously by New- ton and Kennedy (1963).In wet runs in this apparatus,the starting ma- terial was sealedwith water in platinum capsules,but in dry runs the starting material was kept at 110oC. in an air bath for 24 hours beforeit was sealedin the platinum capsules.The sampleswere kept at the re- quired temperature and pressurein the piston-cylinderapparatus for 1 hour above 700oC., for 2 hours at 500-700' C., and, below 500oC., for 18 hours. Recently, Newton and Kennedy (1963) have shown that results ob- tained by simple squeezeron the stability relations between lawsonite and anorthite plus water are much difierent from those obtained from the piston-cylinder apparatus. However, in the present system the results ob- tained from the two methodsare almost identical. Mineral phasesin our experimental products were identified using both a Philips c-ray diffractometer and a petrographic microscope. Even though identification of the new mineral phases, made at high pressures,has been by both Philips r-ray diffractometerand petrographic microscope,we relied almost wholly on the r-ray data for the positioning of the various phase boundaries shown in this paper. A few minute high index particles could always be found in the sample bv careful micro- scopicexamination even though the sample had never been to excessively high pressuresor near the position of the suspectedphase transition. This minute amount of high index material could either result from contamina- tion or possibly very local high stresseson the sample. Thus, microscopic work could not be relied upon exclusively to give the lower pressureposi- tions of the phase boundary. 1690 Y. SEKI AND G. C. KLNN?:,DY
.U € oooo s o 4No o o400 to Tr4No rrln .c --4 q €9cc4 n | _do E
a q-N O. oNO € N]otDN4O.\O€n LQ i€$S N oO N4rO N €4NO\$Ooo\OO@ € 4€NO\ cO N \O4.+ I NNNO\C^O.cOF\bn 'o ro-+d -; -; d*; *j -j -j.o-j j-; -i -i -i ^i oi
a z | 'n\OO€ | O 6-O c)O c)O | * 'D- | l4-NFF-+OO€€l -Nl z
o ooo \o on@ + a.:o.€ ,io.i-e € I NNO!tr N N+4 N @oNc^ rOO,9?-€ r + F N €4V q NINo' E€ 4 p lan.tq 'I X\O4 '| I . d d ,-: Fa \o4+ro .O OOO S SF|ON ts-NUNN N
|-l
O lO-4€OT,OONC)Ol F ,€NF4ONc[\c:t -N+l z
z o oO4 o.s.:6d .=-9 a t&rh4i€oN6r0610rx, z I N €@+ I N'ONo. l-OcON 14 I o O O€O O OtsaON f-NN N z E Q o\ o\o\3: | $ 533 g SFs€ F] lghK*g l il io
Q a a z eea\Y2!! q, D 9S€ g 3$K3S85S83S € o € 4CONO\ co N \O4.+r O NNNO\O\6€N\bAa n €; €;;=r*; -; -; -;-j-; d -id-;.i.i-i.iaiosi-i -i rJ o'] x N F d N ON€ O\ OO\€O+N:D€€4 @ tr cp i\h!+r co d @ N4D O O\+rNq\!'QO\\Oo64 + a € h@No\ co N \o4+r o NNNo\O\6-N6D4 4 6 'c;;+-; d d d-; -; -i d-; -; .i .i -i J.i -i 6i -i 6i BR]],AKDOWN OF K-FELDSPAR 1691 -i .; bot N F tr - dn da'd6 "("("dg- *"< a9 NO\ON ;i *.,T > ^ i€rO ri E(4!> . .os06 N/ ! d .FJ € erd\ oi b E i\r-: tr ! E-A 'o ll ll ll ll u'E " ila saa d a,E x*** o o oa - h 'O:.=+ 'O +.r-:.= cN .+r qllTnl OrN O irrNrrl I **** o Q oo .loron e o,.! \o.:.: o\ o r o-ON I O. O N O a N r r r 6 r , !-nN N d ! o-oE o' | | ,r I d d dd -NNNNN-N--i Hi * ANO\O N N * o\oo a +ro loN I 3 N:ddlol I N I O:O =anOn:3; O,C\€ C\ 6S \O\O .3.-l r\Or N i4 | O\ 4 0O+1O\Oio+r\OnNOOO : r4\Odr\Od F \o *iarsoo\o€€Nr o :4oa+ro !i +.eNNNNN*iooA o\ o.@@rr€ N NNNNNNNNNNNH i i*HiiH r co tl o tt \O@oO4 *r !i r \O\Oh O \D\O\Odr€* NN*N E \OOio -= d . .\o s o rt co\oo\o : OO F N N NO N - O OdON dHrA OO.OcO{| 44n + + oN94 € - HO N €D + ts4 Results from r-ray examination were much more unequivocal.Runs spacedno more than f kb apart near the transition boundary normally showedeither very little new phaseor a very large amount of the new high index phase.For this reasonwe have relied almost exclusivelyon the r-rav data in the selection of our phase boundary. SranrrNG Merrnrers Two kinds of starting materials were used: (1) Pure sanidine crystallized from KAlSiaOs glass in the simple s0ueezer. Tl.vtx,2. X-nav Pownrn Dar.q or Na:runar- Onruocla.sn Usen as e Srenrtl+c Marrnrer- hkr I hkr d(A) o20 I t.)l 2.416 4 T11 A 242 2.320 3 201 417 i l.) 060 2.274 t3 tn 3.92 8 152 2.114 I 130\ 3.75 130J 30 061 2 .050 T31 3.627 6 JJJ 1.958 3 221 J,JI/ 4 043 1.794 34 tt2 3.461 10 3-34 1 621 3 202 3.269 53 002 3.221 100 131 2.982 18 041 2.903 1a 132 2.75r 9 241 2 568 18 310\ 2.540 6 31oJ The KAlSiaOsglass was preparedby Dr. Schairerof the Geophysical Laboratory and suppliedto us through the courtesyof Dr. F. R. Boyd. X-ray powder data for the sanidineare shown under Column A of Table 1. (2) Natural orthoclaselwas obtainedfrom the mineral collectionof the GeologyDepartment of UCLA. X-ray powder data of the orthoclaseare shownin Table 2. NazO and KzO contentsof the orthoclaseare asfollows: Na2O: 4.407o,KrO : ll.l7 7a l "Orthoclase," as defined by Laves (1951), is a mineral that deviates from a truly monoclinic sanidine but that appears to be monoclinic because of triclinic domains too small to be resolved by the microscope or even by *-ray diffraction. B II]JAKDOW N OF K-F ELDSPAR 1693 ExppnrMriNter RBsurrs (1) Resultsobtained by the simple squeezer-usingsanidine as a start- ing material under high water pressuresare summarizedin Fig. 1. Sanidine*H2O changed into a hexagonalphase at very high water pl essures.X-ra1. diffraction patterns, unit-cell dimensions, unit-cell volume, optical propertiesand densitv of this nervphase are represented x KAlSrsOE.HzO aa SANIDINE+a TEMPERATURECC} .!'rc. 1. Diagram showing the synthesis fields of sanidine and KAISi:Os.H:O deter- mined by the sinrple squeezer O sanidine + \Yater --. KAlSiaOs.HzO A sanidine f lvater . sanidine * water X-ray powder data often show the presence of small amount of sanidine associated with KAISLO8.HTO crystals. The presence of the sanidine must be due to either deficiency of water in these runs or to incomolete chemical reaction. under Column A in Tabie 3. Refractiveindices \vere measured by the im- mersionmethod. Density was measuredby meansof suspensionmethod using methyleneiodide liquid. The measuredvalue of density agreeswell with the value calculated from unit-cell dimensionsand the assumed chemical composition of KAlSiaOa'HzO. The *-ray powder data and other physical properties correspondedclosely with those of cymrite (BaAlSiaOs(OH),(Col. B, Table 3), a naturalll' occurring barium analog of our mineral. Thus the new phasel the assumedchemical compositionof which is t694 Y. S]],KI AND G. C. KENNEDY Tearr 3. X-nav Powoen Dera, UNrr-CrrL DTMENSToNSAND Orr{ER Prrvsrcar, Pnoprnuss ol KAlSirOa.H:OSyn:rnrsrzeo lnou KAlSirOsGr.nss Prus Warrn., ANDor CyMRrrn,BaAISirOr(OH) A B (35 KAlSi3Os.HrO kb,640o C.) Cymrite (Smith el d.1.,1949) Measured Calculated d(A) d(A) 001 7.67 5 7.69 7.7 S 010\ 4.61 100] 30 4.6r 46 WW r01\ 3.96 30 3.96 3.95 011J VS o02 3.85 33 3.84 102\ 2.957 100 3 956 2.95 0r2) VS 110 z ool 89 2 667 2.67 S 003 z,Jl VW 111 2.518 10 2.518 253 vw 200\ 2 .308 15 2.308 232 0201 W 103 2.241 + 2.24r 2.24 M 201\ 2.211 2.212 2.21 02rl M 2.11 VW 202 1.990 w 004 1.924 | 920 MW IIJ 1.848 |.849 M 211 1.783 VW |.746 1.746 I. / IJ 1.703 MW 2r2l tJ 1.591 MW tt4 7 1.560 w oosl 300f w ffi0J IUJ 1.468 w 213 1.452 VW 220\ 1 334 1.334 r.341 1lsJ VW 033 | 1.32r 1.324 303J VW BREAKDOW N OF K.FELDSPAR 1695 'l l^etr, 3- (conti,nueil) A B KAlSi3O8.HrO(35 kb,640. C.) Cymrite (Smith eJaJ., L949) Calculated d (A) d(A) l,u,_l '_ 006 | 283 8 1.283 r.283 I w 222 1.260 3 1.260 1lrJ 312l 1 2t6 3 1.216 3041 u4! 1.203 223 1.t84 o:s.33(s) A a:5.32A c:7.70(O) A c:7 .67fr Hexagonal Hexagonal Unit-cell volume: 189. 8 A3 Unit-cell volume: 188.0A3 ':j-:::f+o ,.,: | 6??() oo3 e: r . JJJ] .: i.i;;;1*o'oot Uniaxial negative Uniaxial negative Elongation positive Calculated density: 2 fg Measured density: Measured density: 2 58+0.02 3.413+0.005 KAlSiaOs'H:O;isbelieved to have beenformed by the followinqchem- ical leaction: potash feldspar*water hexagonal new phase KAlSisOs H:O --. KAlSi3O8.HrO molecular volume l82f\3 3043 19043 \__ / I total volume 2t2L3 19043 The volume relation shows that a high partial pressureof water should promote the formation of the new phase. Unfortunately we did not establish by direct analysis that our new phasewas of the assumedcomposition, potash feldsparf water. However, anhydrous runs at 1300' C. and 60 kilobars failed to make this phase directly from orthoclase,presumptuous evidence that the phase cannot be made from pure orthoclase.The exactstructural identity of our phase 1696 Y. SEKI AND G. C. KENNIT'DY and the barium analogueis further striking evidencethat it is indeed a h1-drate.Perhaps the strongestevidence, however, is the closeagreement of our measureddensity, 2.58+0.02 and the calculateddensity of 2.6O. This agreementof the two densities,within to/6, is compelling. If our new phase had either none or two waters of hydration, disagreement should be approximatell' 8a/6between the measured and calculated den- sitl'. The boundary shown in Fig. 1, for the formation of the denseh,vdrate of orthoclase,corresponds closely to the boundary reported b1'Kennedl' (1961) for the transition of orthoclaseto orthoclaseII. In the earlier work Kennedy (1961)added small amountsof water to orthoclaseto flux the reaction.He did not suspectthat the new formed densephase was a hydrate and consequentll'labeledthis phaseOrthoclase II' The relation betr,veenthe new phaseand cymrite will be discussedlater in some detail. (2) Figure 2 showsthe stability of sanidineand its hvdrate determined in the piston-cylinderapparatus, using pure sanidine as a starting ma- terial with added water. The boundary between sanidineand hexagonalphase KAISiBO8'HrO previousll' determined b1' means of the simple squeezer(Fig. 1) shows close agreementwith that determined b.v the piston-cylinder method (Fig. 2). We could not find any evidenceshowing the formation of micro- cline from sanidine in any of our experiments. Petrographicstudies of potash-feldsparsin metamorphic and igneous rocks show that order-disordertransition in monociinic and triclinic potash-feldspartakes placeverv closeto the granulite-amphibolitefacies transition (Heier, 1957).Laves (1952)and Tuttle and Bowen (1958)esti- mated the transition temperaturebetween monoclinic and triclinic potas- sium feldsparto be 650 700' C. and 650' C. respectivel.v. Attempts to synthesizemicrocline or to convert sanidineto microciine have beenunsuccessful (Goldsmith and Laves, 1954)' Recentll' Tomisaka (1962), placed natural monoclinic potash feldspar and water in the simple squeezerfor 5 to 1000hours (400 hours in aver- age) and found some evidenceshowing the transformation of the ortho- claseinto microcline.The transition temperaturesat various water pres- sures as estimated from his data, are as follows: +70' C. at 4 kilobars 460" C. at 2.4 kilobars 450' C. at 1.5 kilobars 400oC. at 1 kilobar Any appreciabletransformation from monoclinicform into t.riclinicform BREA I(DOW N OF I{-F]JLDSPA R r697 4l = KArSiigo!.Heo Il| =.E alt o ul E o- E ul k SAiltDtXE + H2O = TEMPERATURE TC) Frc 2. Diagram showing stability fields of sanidine and KAISi:Oa'HzO determined by piston-cylinder high pressure apparatus. O Sanidine f HzO + KAli3Os'HrO A Sanidine * H:O - Sanidine f HuO I KAISiIOE'H:O + Sanidine * HzO did not take place when the starting material was dry. This implies that the soiution-recrystallizationaction of water or aqueoussolution is very' important in the order-disordertransition in potash-feldspar. (3) Completely dry sanidinewas held at 60 kb and 1300oC. for two hours in the piston-cylinderapparatus and showedno change.This con- trasts sharpll' with the fact that dry albite is easilytransformed into the assemblagejadeitef quartz or coesiteat the same physical conditions. Potassiumfeldspar must be stableunder the rangeof temperaturesup to 1300' C. and pressuresup to 60 kilobars when it is kept in the absenceof water. (4) Figure 3 showsthe phaserelations of naturai orthoclasededuced from the experimentalwork with wet samplesin the simple squeezer' Natural orthoclase was transformed into the assembiages sanidine + jadeite+ quartz, sanidine{ jadeite{ coesite,and KAISiTOs'HzO* ja- deite*coesite with increasingwater pressure. Columns C and D in Table 4 representr-ray powder diffraction pat- terns of jadeites associatedwith sanidine*coesite and KAlSiaOa'HrO f coesitein the breakdown products of natural orthoclaseunder high r,vaterpressure. The x-ray data of these jadeitesare similar to those of Y. SEKI AND G, C. KL,NNT'T.)Y SAXIDIT€+JAD€ITE+ v tu QUAetZ qelE!3IgE ec, lirc. 3. Diagram sho{'ing the stability field of natural orthoclase, sanidine*jadeite *quartz, sanidine*jadeitef coesite and KAlsiaos.Hrefjadeitelcoesite determined by the simple squeezer using natural orthoclase pius rvater as starting material. ---. f orthoclase f water orthoclase + water ---. A orthoclase f water sanidine f jadeite * quartz * water - fl orthoclase + water sanidine f jadeite f coesite - O orthoclase * water KAlSi3Os.HzO * jadeite f coesite X-ray porvder data of some of these runs shorv the presence of sanidine crystals as u,ell as KAlsLo8'Hro. r'his is believed to be either due to deficiency of water or to incomplete chemical reaction. natural jadeite (Col. E, Table 4) except d-value of the former are gen- erally a little smallerthan those of the iatter. (5) Natural orthoclasewas transformedinto the assemblagessanidine *jadeitef quartz, ar'd sanidinefjadeitelcoesite without water in the piston-cylinderapparatus (Fig. a). X-ray powder data of sanidinesassociated with jadeite and quartz or coesitesynthesized from natural orthoclaseare shown unclercolumns c and D in Table 1. Jadeites, synthesized without water, appearing in the breakdown products of natural orthoclaseshow almost exactly the same r-ray dif- fraction patterns as thoseof jadeitesformed by the transformationof the orthoclaseunder high water pressures.(Cols. A, B, Table 4). The stable association of sanidine with jadeite and quartz or coesite is believedto show that sanidineis stableeven under verv high pressures, providing no water is present. BREAKDOW N OF K-FELDSPAR oo4oo o oooo4 o hr4c)n o 4oo HNNTc){IOONNOiN o C\ 4 O4\Oic) € Nr\OcO :ti .1O\€4Od6ll'iq O sc\dd\O q\O qqql4O,\O€ \O oor4 n aanrl \ 9aa d \o+1 eoN N NNNNN N Neii i F o @ o oN+r\o\o co @oo I oO - €\O + N C) iONi N o r f; x r -.9 .io+oo\-o\.€NoN +if,-o 6€+ 4F N t@rHD r, 4CC€ € -N4 -or!O\ @ D++N- O OO\O\; N <4rl .ci.$-;aa; a -i .i .i ^i -i -i 6i JJ' J VJJ N ts oNrNl | | rco z O@i€l I o :; 6i !! oq ts N r - o a.,l o o l.t o: ll d =+.- .- ^-= aov a*E'4"-KRS= EE = :N E Ho\ oo 4+ E I q q i q I I . . d-Y . . . d | 6 6 6 | E €+ C/)JN N Na(, U) A A o Nl ON€€J l* 14 ! 'i) F a- E.Eo n rJ E N-.iHoo ON.; q?: N D.;\c€q NN E XO\ € olqlN\o14 . . 't6.1 6-: EAE d €+OUN N a a NA 'ocolll6 $o€+ql ooooo -E-3xgbE ==B*= lN trOo. co h c)O :c) dO dNdH N NOiiOidol OO dO d N NNN H Hc)c)NiNOONNS e 5 HI olal HI TETPERATURECCI Fro. 4. Diagram representing the stability frelds of sanidinefjadeitefquartz- and sanidine*jadeitel coesite-assemblages derived from natural orthoclase under completely 'l'he dry conditions. piston-cylinder high pressure apparatus was used. O Natural orthoclase .-. sanidine f jadeite f coesite O Natural orthoclase -'.. sanidine f jadeite I qtartz Numbers shorv the ratio of I (310) jadeite x 100 I (310) jadeite + I (002) sanidine DrscussroN Solid,solulion zsi,th Lhe phase KAlShOs.H2O. We have alreadr.noted the hexagonalphase of KAlSi3O8.HzOshows very similar r-ray diffraction patterns to thoseof hexagonalcymrite (BaAISiaOs(OH)). This indicates the phase KAISi3O8.HrO and cymrite have almost the same internal framework, and have an isostructural relationship. It is to be expectedthat the phaseKAlSiaOs.HzO and cymrite form a completeor partial solid solution becausethese two phasesare isostruc- tural and the ionicradii of Ba (1.43A) and K (1.33A) ur" not greally different. BREAKDOWNOF K-FELDSPAR 1701 Further, although we have written the formula for the new phase as KAlSi3Os.H2O, we might perhaps have better written it as KHAISi3Os(OH), a structural analog of cl,mrite BaAlSirOs(OH), the substitutionis K+*H+ for Ba2+. We have succeededin synthesizinga hexagonalphase which is be- Iieved to have an intermediate chemical composition between KAlSi3Os oC. .HzO and BaAlSiaOs(OH)at 620 and 30 kilobars water pressureby the simple squeezer.The starting material was a 1: 1 molecular mixture of sanidine and cymrite which had been synthesizedfrom KAlSi3Osglass and BaAlSiaOeoxide mixture respectively (Seki and Kennedy 1964a). Resulting grain size was approximately .05 mm. As far as could be de- termined by r-ray powder data and bv microscopic examination, the product was homogeneousand sanidinedid not appear. X-ray powder data, unit-cell dimensions and optical properties of this hexagonalphase are shown in Table 6 and are intermediate betweenthe phaseKAlSi3O8 "HzO and cymrite. Thus the following ionic substitution is presumedto have occurredto form a solid solution betweenthe phaseKAlSi3O8.HrO and cvmrite: (.Ba)2+:=(K*) + (H+) Tanro 5. X-ney Poworn Dara or. Coosrms. Coesite formed by breakdown of natural Synthetic coesite orthoclasel (28 kb, 810o C.) (this paper) (Boyd and England, 1960) dtAr d(A) 6.19 4 619 3 4.36 +.J/ 2 3.43 26 3.436 52 Jadeite2 3 099 100 2.761 2 2.765 18 2.696 3 2.698 1l 2.333 3 2 337 3 2.293 6 2.295 6 2.182 3 2.186 A 2.026 .5 2.033 6 KAlSisOs.HrO' 1.849 .) .833 3 1.839 J .795 4 1.794 4 1.788 1.787 4 KAlSi30s.I{rO' 1.698 10 1.654 1.655 6 1 Associated with jadeite and KAISirOs.HzO. 'zHidden by x-ray difiractions of jadeite and KAISirOg.HzO. L7O2 Y. SEKI AND G. C, KENNIiDY TeBr,r 6. X-nev Pomnn Dera, UNtr-Crrr. DruBNsroxs aNl Op:rrc.Lt Pnoprntrns ol e Hoxecoltal Prresr Maln lnotr e 1:1 Mrxrunr ol Sar'trlrnn aNn Cvlrnrr:a er 620o rtNo 30 Krr,osens Wnrtt Pnrssunn d(A) 001 I 7.67 t2 010\ 4.62 27 1oOJ 101 \ 3.96 50 01lJ I 40 002 3.84 102\ z .955 100 0r2J 110 2.670 100 111 2.520 27 2.31O 35 2 240 8 2.2t5 27 t12 2 r90 18 202\> 10 022J 2.980 004 1.923 25 113 1 847 23 120\ 7 2roi 1.745 r2l\ 1-704 t2 21r) a:534 ft c:7.68A ,r:1.570+0.003 The formation of a hexagonalcrystal of the chemical compositionof KAISiBO8.HzO from sanidinerequires very high water pressures(Figs. 1, 2). However, f ar less water pressureis necessarvto form cymrite from monoclinic barium feldspar (celsian) (Seki and Kennedy, 1964a"). Monoclinic potash-feldspar and barium-feldspar form a partial solid solution through the intermediate hyalophane series. The phase KAISi3Os'H2Ocannot be expectedto be found in nature becauseof the very high water pressurerequired for its formation even at low temperatures(Fig. 5). However, cymrite has beenfound in nature (Smith et al., 1949a,b). We believe it is not improbable to find, though perhapsrarely, an intermediatemember betweencymrite and KAlSieOa 'HrO. BREAKDOW N OF K-FDLDSPAR 1703 EH .ga NS x76 *t' ! it Z,A :bo .q !? e^ A ,tl i'a 9j 'J >bo ar _L vE D. (q)r)3unss3udu3rvn 170+ Y, SEKI AND G. C KENNT:,DY We unsuccessfullyattempted to synthesizea hexagonalphase from albite glassplus water and albite crystal plus water by the simplesqueezer and by the piston-cylinder apparatus.Only the assembiagesof jadeite of lquartz or jadeite{coesite have been obtained at temperatures 500" C. and water pressuresto 40 kilobars. As was noted, natural orthoclase with some sodium usualiv decom- posedinto the assemblageKAlSiaOa.H2O, jadeite, and qtatlz or coesite at very high water pressures. Thus there is no hexagonal phase of NaAlSisOa.HzOwhich forms a solid soiution serieswith the hexagonal phaseof KAlSi3O8.HzO.Further study on the stability relationsbetween nepheline, analcite, albite and jadeite however are necessar)-' The swbstitwtionof K for No in jad.ei.teslructure. Sanidine alone is stable even at verl'high pressuresand temperatures.At high water pressures, however, it can be easil.vtransformed into the hexagonal phase of KAlSi3O8.HzO.The phase KAlsizoo, with a pyroxene structure that is pure potassium jadeite was not found even at 62 kb and 1000oC' We have studied the high pressurebreakdown products of the natural orthoclase(Fig. 3). Figures4 and 6 show the intensity of r-ray peaksof jadeite increasewith increasingdistance into the jadeite P-T field. How- ever, the d-spacingof the jadeite showsno shift with changeof pressure or temperature. There is an equilibrium partition coefficient of sodium between the jadeite molecule and between orthoclase. At successively higher pressuresthe sodium is squeezedout of the orthoclase molecule and into the jadeite with coesite or qlrartz appearing as an additional product. We thus interpret the increasedamounts of jadeite with in- creasingpressure as the result of this equilibrium partition changing with pressure.We do not seeany evidenceto suggestthat appreciableamounts of the potassium molecule are going into the jadeite structure inasmuch as our 29 values of all r-ray peaks remain independent of pressuresand temperatures. Figure 6 shows r-ray diffraction charts of some high pressurebreak- down products of naturai orthoclase. These charts were taken by a Philips r-ray diffractometer carefully controlled to keep all conditions the same. IIowever, it is not impossiblethat at high pressuresome Na (ionic radius:0.98 A; i.r laaeite structure ma1' have been partly substituted for by the larger ion of K (ionic radius:1.33 A) which came from asso- ciatedpotash feldspar. AcrNowrsocMENTS We are greatly indebted to Dr. R. C' Newton and Dr. J. R. Goldsmith of The tlniversity of Chicagofor their valuableadvice, help and criticism BREA RDOWN OF K-FDLDSPAR 1705 rooo. G 900. c 40 xb 40 Kb 6 ry, Jd 70e c 25 tG (3ro s T 2eCuKo 28CuKc 20CuKq Frc. 6. X-ray powder data of high pressure breakdown products of natural orthoclase. S: Sanidine; Jd: Jadeite; Qz: Qtaftz; Co: Coesite.AII of these three runs were done under completely dry conditions by the piston-cylinder apparatus. These r-ray data were taken under the same conditions of r-ray diffractometer (Fig. 5). in carrying out this researchand to Dr. L. H. Adams of the Institute of Geophysics and Planetary Physics, University of California, Los Angeles for his many helpful suggestionsin taking the r-ray data. Dr. F. R. Boyd of the GeophysicalLaboratory kindly suppliedus with KAlSi3Osglass. Dr. A. Miyashiro and Dr. H. Hagiwara of the Universitr' of Tokyo helped us to get the chemicaldata of natural orthoclase.Mr. Terrence J. Thomas and Mrs. E. Szentvari assistedus greatly in the progressof this work. To thesepeople our sincerethanks are due. Partial financial support is from our contract for Mineral Synthesis from the Geophysics Branch of the Office of Naval Research, Nonr 23s(28). RnrnnnNcns Brncrr, F. eNr P. LB Cortrn (1960) Temperature-pressure plane for albite composition. Am. f our. Sci. 258, 2W-217. Bowax, N. L. axn O. F. Turrln (1950) The system NaAISLOTKAISLOTHzO. Jour. Geol. 58, 489-517 . Bovo, F. R. axo J. L. ENcr,aNl (1960) The quartz-coesitetransition. Jour. Geophys.Res, 65,749-756. DoNNAy, G. aNo J. D. H. DowNev (1952) The symmetry change in the high-temperature alkali-feldspar series.Am. Jour. Sci. Bouen Vol. 115-132. EnNst, W. G. (1963) Polymorphism in alkali amphiboles. Am. Mineral. 48, 241-260. Gornsurrrr, J. R. .o.uo F. Lnvns (1954) The microcline-sanidine stability relations. Geo- chim. Cosmochim. Acta. 5, l-19. 1706 Y. StKI AND G. C. KI,NNIIDY GRrccs, I). T. aNr G. C. KnNmrnv (1956) A simple apparatus for high pressuresand temperatures. Am. J our. Sci. 254, 722-735. Htrun, K. S. (1957) Phase relations of potash feldspar in metamorphism. louy. Geol.65, 468-479. KtrNenv, G. C. (1961) Phase relations of some rocks and minerals at high temperatures and high pressures. Ad,z:ancesin Geoplrysi.cs7, 303-322, Academic Press. eNo R. C. NswroN (1963) SolidJiquid and solid-solidphase transitions in somepure metals at high temperatures and pressures. Sol,i.dsUntler Pressure. Paul & War- schauer ed McGraw-HilI, 163-178. Lrvrs, F (1951) Artificial preparation of microcline. Jour. GeoI.59, 511-512. ---- (1952) Phase relations of the alkali feldspars I, Introductory remarks. Jotr. Geol. 60. 43G450. AND J. R. Golnsurrn (1961) Polymorphism, order, disorder, difiusion and con- fusion in the feldspars. Ctr,ysillosy ConJerenciasFase.7, 7l-80. N{.ncDoxero, G. J. F. (1956a) Experimental determination of calcite-aragonite equi- librium relations at elevated temperatures and pressures.Am. Mi.neral.41,7M-756. (1956b) Quartz-coesite stability relations at high temperatures and pressures. Am. Jour. Sci. 254, 713-721. Nu,vroN, Il. C. nnn G. C. Kaxmrov (1963) Some equilibrium reactions in the join CaAI:Si:Os-HzO.Jour. Geophys.Ras.68, 2967-2983. Prsronrus, C. W. F. T., G. C. KrNNrnv ANDS. SouRrRAyaN(1962) Some reiations between the phasesanorthite, zoisite and lau'sonite at high temperatures and pressures z1z. -Iour. Sei.260,44-56. Romnrsox, E. C., F. Brncn aNo G. J F. X{ecDoNaro (1957) Experimental determina- tion of jadeite stability relations to 25,000 bars. Am. Jour. 9ci.255,115-137. Snxr, Y., X,I. Area axo C. Kero (1960) Jadeite and associateminerals of metagabbroic roclis in the Sibukawa district, Central Japan, Am. Mineral.45,668-679. aNo G. C. KoNNrry (1964a) Phase relations between cymrite, BaAISLOs(OH), and celsian, BaAlrSirOs.Am. Mineral.49, 1407-1426. G. C. KrNNnov (1964b) An experimental study on leucite-pseudoleuciteprob- lem Am. Mineral.49. 1267-1280. Slrrrn, W. C., F. A. BaNxtsrnn eNo M. H. Hri,v (1949a) Cymrite, a ne\.vbarium mineral from the Benallt manganesemine, I{hiw, Carnarvonshire, Mineral. futag.28r 676-681. - -, F. A. BeNxrsrnn ANDM. H. Hrv (1949b) A new barium mileral from the Benallt manganesemine, Rhiw, Carnarvonshire. Am. Mi,neral,.34,614. Tonrrsaxa, T. (1962) On order-disorder transformation and stability range of microcline under high water vapor pressure.Mi.n. four.3,26l-281. Turrrn, O. Ii. lNn N. L. BownN (1958) I'he origin of granite in the tight of experimental studies in the system NaAlSisOs-KAlSi2O8-SiO2-H2O.Geol. Soc. Am. trtem.74. Manuscripl receiled.,February 7, 1964; acceptedfor pu,blicol.ion,June 17, 1964.