American Mineralogist, Volume 66, pages 763-769, I98I

Diffusive and displacivetransformationin plagioclase

and ternary feldsparseriesl
HeRnnRr KRoLL AND HANs-UrnrcH BeuneuBR
Institut fiir Mineralogie, Gievenbecker lile g 6I
D-4400 Miinster, Federal Republicof Germany

Abstract

The temperature of the displacive triclinic/monoclhic transformatiotr, Zai"pr,of synthetic

plagioclasesand ternary feldsparswas determinedfrom high-temperatureX-ray powder pho-
tographs. To.o,increaseswith increasingAb + An content in ternary feldsparsand with in-
creasingAn content in the plagioclaseseriessuchthat plagioclasewith more than = 15 molTo
An cannot becomemonoclinic before melting. The temperaturesof the displacive and di-ffu-
sive transformationof plagioclasescoincide.K-exchangeexperimentsare usedto discussthe
Al,Si topochemistryof the samples.

Introduction Experinental
The first detailed and thorough accountof the dis-
Ternaryfeldspars
placive transformation of plagioclasesand ternary
feldsparswas made by Laves (1952)and MacKenzie To investigatemeaningfully the displacive trans-
(1952). They determined the triclinic/monoclinic formation of feldspars the samples under consid-
transition temperatures from optical data or line eration have to be topochemically monoclinic.
splittings in X-ray powder diagrams.A study of the Becausethe temperaturesof the diffusive transfor-
transformation requires that the samplesunder in- mation are not known in the ternary feldspar field, it
vestigationbe grown with monoclinic Al,Si distribu- is impossibleto know a priori the topochemistry of
tion. Otherwise, a displacive transformation cannot the sampleswhich are to be hydrothermally synthe-
occur, and the extrapolation of extinction angles or sizedfrom gelsor glasses.Another problem in prepa-
line splittingsto zero resultsin a "simulated transfor- ration is causedby the large unmixing gap.
mation". Kroll el a/. (1980) and Kroll (in prepara- Both difficulties can be overcome by the K-ex-
tion) describe the characteristicsthat distinguish a changemethod of Viswanathan(1971).Severalsam-
"simulated transformation" from a true triclinic/ ples of low plagioclasesapproximately in the range
monoclinic change. Their investigations on alkali Aq to Anro were chosenfrom the Codett and Eber-
feldsparsare extendedin this paper to plagioclases hard (1967) collection and K-exchanged in molten
and ternary feldspars. The reader is referred to their KCI (850"C, 2-l2h). They were then dry heated at
papers for an explanation of nomenclatureand the ll00'C for l0 days. All samplesbetween Ano and
concept of treating the structural and metrical An=r, attained monoclinic cell dinensions at room
changesthat accompany the displacive and ditrusive temperature.Obviously, the (K,Ca)-microclinesinto
processes. which the plagioclaseswere convertedby the K-ex-
change had become topochemically monoclinic
(K,Ca)-sanidines.Several of these (K,Ca)-sanidines
I Secondpart of a serieson diffusive and were back-exchangedin molten NaCl (850'C, 2-
displacive transforma-
tion of felspars (Na,K,Ca)[(Al,Si)Si2O8]presentedIN MEMO-
l2h) to give topochemicallymonoclinic plagioclases.
RIAM FRITZ LAVES. The first part, "The high albite-monalbite Two sampleswere chosento prepare ternary series
and analbite-monalbil transitions," appeared n The American from the end-membersAbrrrAn,u.r-OrrrrAn,u.,and
Mineralogist,65 (l 192-l2l l, 1980). Abr22An2r.E-Orrt rALr, by homogentzingat 1000tC
"
0w3-004){'
/ 8r/0708-0763$02.00 763
764 KROLL AND BAMBAUER: TRANSFORMATION IN PLAGIOCLASE AND TERNARY FELDSPAR

weighed mixtures that had been pressedinto tablets cos'z(AlI l) vs.temperature.The resultantstraight line
(l kbar). In this way it is possibleto obtain topo- relationship can then be used to determine Tuoorby
chemically monoclinic ternary feldsparswhich can- extrapolationto l-cos'?(Alll) : 0 (Kroll et al., 1980,
not be directly synthesized. Fig. s).
Lattice parameters for the heat-treated, K-ex-
changed equivalents,their Na-back-exchangeprod- High temperatureX-ray powder camera
ucts, and the An , * ternary series (Tables I and 2) A Guinier-Jagodzinskitype camera that was de-
were determined by the Guinier-Jagodzinski tech- signed by Kroll (1971) and modffied by Schirmer
nique (CuKa, radiation). Further details on sample (1976)was usedto take the high temperaturephoto-
preparation and the procedureof determining lattice graphsin this study and that of Kroll et a/. (1980).Its
parametersare given by Kroll (1971) andKroll et al. main featuresare illustrated in Figure 1. The X-ray
(re80). beam (CuKc, radiation) penetratesthe sampleat an
angle of 45 degrees.The sampleholder is a Pt net of
Plagioclases 4 x 12 mm, sealedon a heat-resistantmetal frame
Only Na-rich plagioclasecan be prepared with a that is moved by * 5 mm on the focusingcircle. The
topochemically monoclinic Al,Si distribution. The furnace is a resistant-furnace,consisting of several
following sampleswere synthesized:Ano and An,, by bars of AlrO, ceramics around which a resistance
hydrothermal crystallization of gels at 1040"C, wire is wound. The bars are placed around the
P(HrO) : 5 bars for 15 days; An,o and An"o by dry sample holder and kept within a silica glass tube,
crystallizationof glassesat 1055"Cand l090oC, re- whose surfaceis coveredby Au foil to reflect the IR
spectively,for 15 days. The preparation of the gels radiation. The entire furnace is held in a water-
and glassesis describedby Kroll et al. (1980). cooled metal case.The maximum temperature that
Judging from the run duration and synthesistem- can be reachedis I100'C. This constructionprevents
peratures,the Al,Si distribution of the plagioclases any contamination of the sample during heating.
attainedequilibrium (alsocompareEberhard, 1967). . The temperatureis held constantto + 2.5oCbelow
about 500oCand + 0.5oCabove500"C by a thyris-
High temperatureX-ray meaturements tor-controller that is connectedto a ft-PtRh (DIN)
To follow the approximation to monoclinic cell di- thermocouple placed near the resistancewire. The
mensions with increasing temperature, the plagio- temperatureof the sample is measuredby a second
clasesand ternary feldsparswere X-rayed at elevated thermocouplein the immediate vicinity (0.5 mm) of
temperatures. The splitting of the (lll) line, Alll : that part of the Pt net which is penetrated by the X-
40(lll) - 40(l1l), was measuredand plotted as l- ray beam.The thermocouplewas calibrated by visu-

Table l. Lattice parameters of feldspars (a) that were prepared by K-exchange of low plagioclasesin molten KCI and subsequentdry
heat treatinent at ll00'C for l0 to 20 days, and (2) that were then back-exchangedin molten NaCl.

Sample 0r Ab An Heat treatment ax* bc d y V No.

-l
N o .* | morFl r[ oc] t[ oays] t ffl r8l tBl tol L t"l tX'l fines
'r28.5(1) 40
i 4 s ? ( 1 1 0 )9 s . 9 0 4.1 1100 10 8 . s 7 0 4 ( 6 ) 1 J . o 3 o s (6 ) ? , 1 7 6 8 (s ) e o 1 1 s . e ? s (4 ) s o

14s5(B4) 83.s 0 16.s 1100 10 8 . 4 8 2 6 (7 ) 1 3 . o i 3 e ( B ) 7 . 1 6 4 e (6 ) e E 1 1 s . 9 6 8 (4 ) e o ?11.1(1) 37

4173( E4) 0 8 3 . 5 1 6 . 5 b a c k - e x c h a n q e dB ' 1 6 0 8 ( 8 ) 1 2 . 8 6 s 7 (B ) 7 . 1 o e e ( 6 ) e 3 . s 5 1 ( 6 ) 1 1 6 . 3 2 s ( s ) e o . 2 6 4 ( 5 ) 6 6 7 . 6 ( 1 ) 4 7

1 4 5 4 ( 1 6 6 )? 8 . 5 0 s1,s 1i00 10 8 . 4 5 8 3 ( 7 ) r 3 . o o e B (9 ) ? . 1 6 i 1 ( s ) 9 o 1is.e62( s) eo ?De.?('t) 27

1 4 8 4 ( ' t 7 o )? 4 . 7 E 2s.3 11oo 10 8 . 4 3 0 0 ( ' 8 ) 1 3 . o o 3 s (B ) z . i s s 4 ( 5 ) e o 1i5.s62( s) 90 785,2(1) 21

4'l7o(17a) a d.1639( 6) p.a742(

74.7 25.3 back-exchangE8 6 ) ? . ' t o | 7 ( 6 ) e 3 . 4 ? ? ( 5 ) 1 1 6 . 3 0 6 (4 ) e o . 3 4 s ( 4 ) 6 6 8 . 1 ( 1 ) 5 6
1 4 s 4 ( 1 s 3 )? 4 . 4 0 2s.6'1100 10 8 . 4 2 8 6 ( 8 ) 1 3 , o o 1 o (z ) ? . 1 5 6 s ( 4 ) e o 1 1 s . e s 6 (4 ) 9 0 ?o4.e(1) 22

1480(193) o 7 4 . 4 2 5 . 6 b a c k - e x c h a n q e d8 ' 1 5 5 0 ( 1 2 ) 1 z . B ? 4 3 ( 1 3 )? . 1 0 6 1 ( e ) 9 3 . 4 7 ? ( 1 o ) 1 1 6 , 2 7 5 ( 8 ) 9 0 . 4 0 4 ( ? ) 6 6 7 . 3 ( 2 ) 34

1455( 31) ?2.2 E 27.8 11OO 10 8.4141( 5) 12.eelo( s) ?.1s35( 3) eo 1 1 s . e s 4 (3 ) e o ?03.3(1) 41

' l ) ' t 2 . 8 7 6 6 ( s ) 7 . 1 0 7 7 ( s ) e 3 . 5 0 8 ( 6 ) 1 1 6 . 2 5 2 (s ) e o . 4 0 9 ( s ) 6 6 e . 3 ( 1 ) 36
1479( 3'l) O ? 2 . 2 2 7 . a b a c k - e x c h a n g e ds , 1 6 2 7 (
r+ss(ror) zz.o 0 28.0 1i00 10 8 . 4 1 7 7 ( 1 o )1 2 . e e 7 3 ( 9 ) 7 . 1 s 3 3 ( 8 ) e o 1 1 5 . e 4 1 (? ) e o ?03.8(2) 22

1 E s 2 ( 1 3 6 )6 4 . ? 0 3s.s 1tu0 20 8 , 3 9 9 9 ( 1 7 ) 1 2 . s ? 8 o ( 1 3 ) 7 . ' t4 s 6 ( 1 2 ) 9 0 . 7 1 5 ( 1 2 ) i 1 6 . i 0 0 ( 1 s ) e 0 . 0 ? 5 ( 1 3 )6 e e . e ( 3 ) 32

4431( 9i) 52.2 o 47,8 11oo 20 8 . 3 3 ? s ( 2 0 ) ' t 2 , s s 16 ( ' t ? ) 7 . 1 2 s 2 ( 1 8 ) 9 2 . 0 4 2 ( 14 ) 1 1 6 . 0 6 s ( 1 6 )e o . 3 e ? ( 1 6 )6 9 0 . 7 ( 4 ) 3 1

* Bracketed numbe!s co!!espond to the sample numbe!s of cof,lett and Eberha!d (196?).
** StandaId eDDors ars given in parentheses and refe! to the Last decimal place(s7.
 KROLL AND BAMBAUER: TRANSFORMATION IN PLAGIOCI,IISE AND TERNARY FELDSPAR 765

Table 2. Lattice parameters of a series of ternary feldspars with a topochemically monoclinic Al,Si distribution. They were prepared
from samples(31) listed in Table I by dry homogenization at l000oC of weighed mixtures.

Sample 0r Ab An Homogehization a*i o c q

PY V No.
ol
No.* LmorF] r[oc] tInours] t8l t fil tFl t tll tol LA J IINES

1455(31) 72.2 o 27.4 a.atzz('to) 12.ss73( e) ?.is33( B) e0 11s.941(7) eo ?03.8(2) 22

1529a 65 7 . 2 2 7, 8 1 0 0 0 26 B . i B B z ( e ) 1 2 . s s 3 3 (7 ) ? . 1 s o 2 ( s ) e o 1 1 6 . 0 0 ? (s ) e 0 780.4(1) 27
1529b 55 17.2 27.4 10EO 26 ) 0 . 1 s e ( 1 6 )6 e 6 . 5 ( 5 ) 1 e
e . s o a r ( r : ) 1 2 . 9 7 7 4 ( ' t E?) . i 4 6 0 ( 1 2 ) e 0 . z 1 B ( 1 3 )1 1 6 . 1 0 2 ( 1 2 e
1523 45 2 ' , 1 . 22 7 . 8 1 0 0 D 26 a.srse(ro) i2.s63o( a) ?.142s(6) e1.40s(B) 1i6.13?( 6) e0.23s(7) 6eE.e(2) 31
1529c 35 37.2 27.4 1EEO 26 8 . 2 7 7 4 ( e ) t z . s + o + ( e ) ? . 1 3 b 0 ( 7 ) e ' t . e e 1 (7 ) 1 1 6 . 1 7 2 (6 ) e 0 , 3 0 e ( 6 ) 6 8 5 , 3 ( 1 ) 2 9
l s33 25 47,2 2?,A 1000 16e a.246't(1s) 12.s2s\(iz) ?.i2Bo(11) e2.4s6( 9) i16.212( e) e0.345( 8) 680.?(2) 26
1326 15 5'7.2 2?,A 1000 14 8.22x( e ) 1 2 . s 1 3 6 (E ) 7 . ' t 2 1 7 ( B ) e 2 , ? a 6 ( 6 ) 1 1 6 . 2 4 2 (? ) e 0 . 3 e s ( E ) 6 7 7 . 1 ( 2 ) 3 1
't47e(3't) o 7 2 . 2 2 ?, A 8.1627( 7) 12.8766(e) 7.1077( s) e3.508( 6) 116.2s2( s) e0.40e( s) 668,3(1) 36

* Bracketed numbsrs correspond to t,he sample numbers of Corlett and Ebernard (1967).
** Standard errors are given in parentheses and refer to the tast decimal place(s).

ally observing the melting of several different sub- quired monoclinic topochemistryat ll00"C, but dis-
stancesranging between212"C and 1063'C (AgNOr, placively inverted to triclinic symmetry 61 seeling.
KrCrrOr, Ba(NOr)r, NaCl, Ag, Au). They were Their 7 anglesare definitely larger than those of syn-
placed on the sampleholder itself or usedin place of thetic high-temperatureplagioclases,which were K-
it (Ag,Au). We estimatethe accuracyof the temper- ixchanged by Kroll and Miiller, 1980 (Figure 2a).
ature measurementto t 5oC. This indicates that the two samples in question are
considerablymore disorderedthan high-temperature
Results plagioclaseswith the same An content, but without
Topochemistry
of plagioclasesand their K-exchange further experiments it is not possible to decide
products whether they are truly topochemically monoclinic.
For monoclinic topochemistrythe room temperature
The low plagioclaseswhich were K-exchangedand monoclinic/triclinic inversion in this K-exchanged
then heated at llOOoC acquired monoclinic topo- serieswould correspondto that in the analbite-sani-
chemistry41d mel6slinic cell dinensions when their dine series.
An content was less than about 33 mol%o.Samples The a and y curves of the K-exchanged synthetic
with more than 33 mol%oAn, specificallyOr*rAnr' high-temperatureplagioclasesseparateat Ao,r-ro.
and OrrrrAnorr, retained triclinic cell dimensions, The An o lattice parameterscould not be evaluated
even after prolonged heating, as is seenfrom Figure with certainty becausethe powder diagram shows
2a. It is possible, however, that these samples ac- considerablepeak broadening, which indicates tri-
clinic geometry. We conclude that high-temperature
plagioclasesare topochemically monoclinic in the
\ - m o v e m eonfst o m p lheo t d e( r 5 m m ) compositional range Aq to Anr5-2o, but are triclinic
p-\- when the An content is greater than 15-20 moI%o.
m o t ofro r s o m p tm
e ovement
This is confrmed by a structure refinement of an
a w o t e r c- o o l efdu r n o c ce o s e An,., high-temperature plagioclase that was pre-
pared by dry heating from a low plagioclase.It was
s i t i c og l o s s t u b ew i t hA u f o i l
found to be topochemically triclinic (Kroll, 1978),
r e s i s t o n cf u
ernoce and accordingly, its K-exchange equivalent is geo-
s o m p theo t d e( rP t n et ) metrically triclinic (marked by crossesin Fig. 2a). By
contrast, Arr2r', which was K-exchangedin the low
monochromotic structural stateand subsequently heat-treated,is geo-
- r o yb e o m
metrically monoclinic, thus demonstrating its mono-
clinic topochemistry.
The structural differences between high-temper-
X - r o y t i t mo n f o c u s s i ncgi r c t e ature plagioclasesand Na-back-exchanged(K,Ca)-
Fig. l. Schematic drawing of the high+emperature X-ray sanidines are further illustrated by comparison of
camera used in this work and that of Kroll et al. (1980\. their lattice angle y (Fig. 2b). The y valuesof the Na-
766 KROLL AND BAMBAUEK TRANSFORMATION IN PI,/IGIOCLASE
AND TERNARY
FELDSPAR

= 450oCof K-feldspar with ?itn = 980oCof Na-feld-

spar (Kroll et al., 1980,Table 3).
. s y n t hh r g h - pl
l . untreoied
9r.0oi not lowplog : heot- treoted
not lowptog . |(- exchonged Displacivetransformationof ternaryfeldspars
heot- treoted
rt"l No-bock-exch Figure 3a, b demonstratesthe linear dependence
906 of l-cos'z(AlI l) on temperaturewithin the two topo-
chemically monoclinic ternary seriesOr,o-roAbrr r-r-
An,"., and Or,r-rrAbrrr-r72Arr2r", The points of inter-
902 section on the abscissacorrespondto the displacive
transformation temperatures(compare Kroll et al.,
. synthhigh-T pt : l(-exchonged 1980, Fig. 5). In Figure 4 these temperatures are
heot- treoted
9 3 + not lowplog K-exchonged
plotted vs. the Or content within the two series
o n o t l o wp l 0 g K-exchonged (curves2, 3). Curve I showsthe variation of ?:i"prin
d,d['] heot- treoted alkali feldspars (An ) that equilibrated at l000oc
(Kroll et al.,1980,Eq. l, Fig. 8). By analogyto curve
92
l, a parabolic fit was chosenfor curves2 and 3. The
ordinate intercept of these curves indicates the tem-
peratures at which topochemically monoclinic
9l plagioclasesAn,u, and Anr^ would become metri-
cally monoclinic if permitted by the solidus temper-
atures(compareFig. 6).
90

An[mot%] Ab172-572
0155-
15An27.B
Fig. 2. (a) Variation with composition of a and 7 anglesof Kroll
and Miiller's (1980)K-exchangedequivalentsof high-temperature :>. +
c,
plagioclases(dots) plus K-exchanged and subsequentlyheated low
plagioclases(diamonds). (b) Variation with composition of the 7
L

><
I b
anglesof Kroll and Miiller's (1980)highaemperatureplagioclases L

(dots) plus Na-back-exchanged(K,Ca)-sanidines(diamonds). o

+ a\
(:f,

=0 568

back-exchanged(K,Ca)-sanidinesAn25.3,Anru."and
Aar, ate larger than those of the high-temperature U). Ab:rs-z:s0156
-16An16
5
"
plagioclases.
30
This con-firmstheir larger degreeof dis- I

order, which is anticipated from Frgure 2a where

their K-equivalents are found to be monoclinic. In 0
contrast,the K-exchangedhigh-temperatureplagio-
clasesare ldslinis when their An-content exceeds15-
20 moI%aIn Figure 2b, the An,u.,sampleplots near
the high curve becausean original high-temperature
plagioclaseof Anr6.sis closeto monoclinic topochem-
istry and thus cannot be distinguished from a Na- ro 288 716
back-exchanged(K, Ca)-sanidine. 0 sootu' looo
The (K,Ca)-microclinesdisorder much faster than T r - , . o[ yo C]
low plagioclases of the same composition when Fig. 3. Variation with temperature of the param€ter l-
heated at the sametemperature(e.9., ll00'C). This cos2lAlll; within the topochemically monoclinic series (a)
Or,o-soAb4r-r,,An,", and (b) Or15-55Ab57 2-112An27 g. The in-
is probably due to the fact that the diffusive transfor-
dicated temperatur€s of the displacive transformation, Tot"or,
mation of a (K,Ca)-feldsparoccursat a much lower fitting. Al I I : a0(l I l)
were determinedfrom linear least-squares
temperature than that of a plagioclase.Compare 7},' - 40(lIl), CuKay
 KROLL AND BAMBAUER: TRANSFORMATION IN PLAGIOCLASE AND TERNARY FELDSPAR 767

1 18 " C
1400

1200
o

1000 n^, o
5U4 =6
<
---_
800 an
Ot
a (J4

- E I

600
2
400
0
200 400 600 8m 1000 1200
I r - , . o[y" C ]

0 Fig. 5. Variation with temperature of the parameter l-

?0 40 60 cos2(AlI l) for four plagioclases.The intcrsectionwith the abscissa
corresponds to Tdi"elonly in caseof Aq.
%]
0rImot
Fig. 4. Variation of T6i"o1 idus curve at ! Anrs. Thus, there is only a small field
with Or contentin the ternary seriesof
Figure 3. Curve 3: Or15-55Ab57 2-17zAn r r, curve 2: Or,o-roAb735_
where (Na, Ca)-monalbiteexists.This result is in ac-
335An165. Curve I is a plot of the variation of Tai"o1in the alkali
cordance with the results obtained from the K-ex-
feldspars calculated from Eq. I of Kroll el a/. (1980) for an
equilibration temperatureof 1000'C. change experiments(Fig. 2a). The pronounced in-
crease of Tooo,with An content is as expected:
topochemically monoclinic anorthite is /ar from
Displacive transformation of plagioclases reaching monoclinic geometry even near its melting
The variation of l-cos'z(Alll) with temperatureis temperature(Czank, 1973).
plotted in Figure 5 for the four plagioclasesdescribed Three other data points (Ano,s,ro)
are plotted in Fig-
in the experimental section. The topochemistry of ure 6. They were obtained from the line splittings re-
thesesamplesis not known a priori. A comparisonof ported by MacKenzie (1952, Table l, "Crystallized
their temperatures of equilibration, [ooo, and the dry" and Table 3). They were linearized as l-
temperaturesat which the straight lines intersectthe cos'?(Alll) and extrapolatedto 939'C (Ano),967"C
abscissa,possibly representing Zooo,,gives an in- (An,), and 1023"C (tu,o). Although no crystalliza-
dication. tion temperatureswere given for the o'dry" samples,
As we discussbelow, a temperatureof intersection the extrapolation temperatures presumably corre-
that is higher than [0",, doesnot correspondto Zooo,. spond to the displacivetransition. This would not be
Otherwise,a topochemicallymonoclinlcAl, Si distri- true for MacKenzie'sAn o samplenor for any of his
bution would have acquired equilibrium in a triclinic hydrothermally crystallized plagioclases.Possible
unit cell. This is, however, a contradiction in itself reasonsfor the discrepancybetween Z*' in Mac-
because symmetrically non-equivalent tetrahedral Kenzie's and our samplesare discussedby Kroll (in
sites are energeticallynon-equivalent and thus can- preparation).
not contain equal amounts of Al (Laves, 1960).
Therefore,only the Ann samplewhere f*"1 (l040oc) Dffi sive transformation of plagioclases
) Io*p,(l005oC) is topochemicallymonoclinic. The temperature of the difusive transformation
In Figure 6, the curve of the displacivetransforma- Iu,* of Na-feldspar is = 980oC. We expect it to in-
tion in plagioclasesis drawn through the four data creasewith increasing An content based on the fol-
points available: An"(984'C, Kroll et a/., 1980, Eq. lowing reasoning.If we assumethat [,o is independ-
l), An6(1005oC, Fig. 5), An,65(ll95oc,Fig. 4) and ent of An content, a sample of Anro prepared at
An rs(l4l8oC, Fig. 4). This curve intersectsthe sol- l000oc would be topochemicallymonoclinic and its
768 KROLL AND BAMBAUEK TRANSFORMATION IN PLAGIOCLASE AND TERNARY FELDSPAR

Figure 2 of Kroll and Bambauer (1971),which was

1600 redrawn by Smith (19'14,1,Fig. 7-60). Contours are
I ['C] drawn for 25"C,500oC,and l000oC. If the samples
Or* rAnr, , and Orrr rAn T8are in fact topochemically
1400 ./'
aonsslinis, the contours continue along the dashed
curvestb the Or-An side.
1200 Transition temperaturesas given by Laves (1952)
6tidus and MacKenzie (1952) for some natural anortho-
1000 MA$l)-6
t ct.o
claseswere added. Lacking suitable data, we were
not able to redetermine these temperatures, and
therefore they are not directly comparable to our
o MocKenzie (1952)
800 o t h i swo rk
data. However, for one synthetic sample,
OrroAbroAn,o(MacKenzie, 1952, Table l), which
most probably is topochemically monoclinic, ?:i.er
600 was redetermined and is in excellent agreementwith
the contours.
Ab 10 20 30 As noted by Smith (1974,I, p. 353),the temper-
aturesfor the natural specimensare low comparedto
An[mol%] the synthetic samples(0'C to 250'C). This nlay be
Fig. 6. Variation of 766o1and Ta;6 in sodic plagioclases.As
discussedin the text, Zu,"ntand ?din are probably identical in the
due to some replacement of (Na,K,Ca) by larger
plagioclase series. For rapid cooling, therefore, the monoclinic/ mono- and divalent cations (Rb,Sr,Ba).
triclinic transition curve corresponds to the displacive In addition, Kroll e/ a/. (1980) observefrom their
transformation. For slow (equilibrium) cooling it corresponds to synthetic alkali feldsparsthat a decreaseof the equil-
the diffusive transformation. The two data pohts that plot above
the solidus curve were found by extrapolation of curves 2 and 3 in
Figure 4.
o MocKenzie(1952)
:synthetic An
' MocKenzie(1952)
:noturot
"o Loves(1952)noturol
displacive inversion would occur at about 1450"C et ol (1980)
Kroll ond
(Fig. 6). This would mean that a topochemically thisworksynthetic80
monoclinic Al,Si distribution had equilibrated in a
triclinic unit cell. We reject this possibility for rea-
sonsstatedabove.Consequently,the increaseof Iai"or
with increasingAn content causesan increaseof Zo* ab
to at least the sameextent.
i
| -l

Furthermore, in alkali feldspars that are crystal- lt // t

lized just above T*, (Kroll et al., 1980,Fig. 8) fo*o,is

lower than 4* . This temperature difference de-
creaseswith decreasingOr content and equals zeto at ,*10/r4,';'X,';).'
Ab,oo.It is highly unlikely that a further decreaseof iiFts/'^ monoclinic
the relative size of the large cation (Na,Ca) with re-
spect to the framework size (<Si,Al - 0>) would
10 60 80 0r
cause Ii,"o, again to fall below Zainas in the alkali %l
lmot
feldspars.We have thus excludedthe possibility Zu," -['C] in the ternary system Or-Ab--
Fig. 7. Variation of Ta;"o1
( Zooo,as well as Tom) Iai"pr. Therefore, we con-
An. fdi"pr is indicated beside the data points. Contours drawn for
clude that, as for Na-feldspar, Zai.preeu&ls?u- in the 25'C, 500'C, l000oC are based on the synthetic samples.The
plagioclasesseries(Fig. 6). transition plane corrcspondsto T.qu : 1000- I 100'C. Increasing
Al,Si order within topochemically monoclinic samples due to a
Summary decreaseof 4q,ir lowers the plane (Kroll et al., 1980, Eq. 3).
Sourcesof data: Laves (1952), MacKenzie (1952), Ktoll et al.
The temperatures of the displacive transformation (1980,Eq. I for Z*,1: 1000'C), tlis paper:Figure 3 and 6 where
ofalkali feldspars,plagioclasesand ternary feldspars ?ai"pr : 1140'C is taken at ttre intersection of the curve of the
are summarized in Figure 7. This diagram replaces displacive transformation with the (dry) solidus curve.
 KROLL AND BAMBAUER TRANSFORMATION IN PLAGIOCI,/ISE AND TERNARY FELDSPAR 769

References
Corlett, M. and Eberhard, E. (1967)Das Material ftir chemische
und physikalische Untersuchungen an Plagioklasen. (Tei-l I der
Laboratoriumsuntersuchungen an Plagnioklasen). Schweize-
rische Mineralogische und Petrographische Mitteilungen, 47,
303-316.
Czank, M. (1973)Strukturuntersuchungen von Anorthit im Tem-
peraturbereichvon 20oC bis l430oC. Ph. D. Thesis,ETH, Zii-
rich.
Eberhard,E. (1967)Zur Syntheseder Plagioklase.Schweizerische
Mineralogischeund Petrographische Mitteilungen, 47, 385-398.
Kroll, H. (1971) FeldspAteim System KAlSi3O6-CaAl2Si2O8-
NaAlSi3O6: Al,Si-Verteilung und Gitrerparameter, Phasen-
Transformationen und Chemismus. Inaugural-Dissertation der
Westf. Wilhelms-Universitiit. Mtnster.
Kroll, H. (1978)The structure of heat-treatedplagioclasesAn28,
An52, An69 and the estimation of Al,Si order from lattice pa-
rameters(abstr.).Physicsand Chemistry of Minerals, 3,'16-77.
Kroll, H. and Bambauer,H. U. (1971)The displacivetransforma-
tion of (K,Na,Ca)-feldspars. Neues Jahrbuch fiir Mineralogie,
Ab 20 /,0 60 g0 0r Monatshefte,1971,413416.
Kroll, H. and Miiller, W.F. (1980)X-ray and electron-opticalin-
lmot%l vestigation of synthetic high-temperature plagioclases. Physics
Fig. 8. Approximate representationof the phase relations of and Chemistryof Minerals, 5,255-277.
rapidly cooled, high-temperature ternary feldspars. The plane of Kroll, H., Bambauer,H.U. and Schirmer,U. (1980)The high al-
the displacive transformation is taken from Figure 7, the solvus bite-monalbite and analbite-monalbite transitions. The Ameri-
(900"C, 0,5 kbar) and three tie-linesfrom Seck(1971).The phases can Mineralogist, 65, | 192-l2l l.
that occur above the transition plane are Or-containing (Na,Ca)- Laves, F. (1952) Phase relations of the alkali feldspars. I. In-
monalbites and An-containing monalbite-sanidine mixed crystals; troductory remarks. II. The stable and pseudo-stablephase rela-
(Na,Ca,K)-analbitesoccur below the plane. tions in the alkali feldspar system.The Journal of Geology, 60,
436450 and 549-574.
Laves, F. (1960) AllSi-Verteilungen, Phasen-Transformationen
ibration temperature of a (Na,K)-monalbite by und Namen der Alkalifeldspiite. Zeitschrift fiir Kristallo-
graphie, 113,265-296.
100"C causesZooo, to decreaseby 25"C to 30oC.The
MacKenzie,W.S. (1952)The eflectof temperatureon the symme-
ternary feldspars were prepared from K-exchanged try of high temperature soda-rich feldspars. American Journal
plagioclasesheated at I l00oc, whereasnatural feld- of Science,Bowen volume, 250A, 319-342.
sparsgrew at much lower temperatures.Thus, as ex- Schirmer, U. (1976) Die Ordnungen der diffusiven und der dis-
pected, all natural specimensinvert at lower temper- plaziven Umwandlungen bei Na-reichen Alkalifeldspiiten. Di-
plomarbeit, Miinster.
atures than do synthetic ones. The plane of the
Seck,H.A. (a971)KoexistierendeAlkalifeldspite und Plagioklase
displacive transformation in Figures 7 and 8 in- im SystemNaAlSi3Os-KAlSi3O3-Ca2Al2Si2O8-H2O bei Tem-
dicates approximately the highest temperatures at peraturenvon 650oCbis 900"C. NeuesJahrbuchftir Mineralo-
which the transition occurs on quenching. The lowest gie, Abhandlungen, 115, 315-345.
temperatures could only be found if we knew [,, in Smith, J.V. (1974) Feldspar Minerals. I. Crystal structure and
physical properties.Springer-Verlag:Heidelberg.
the ternary field.
Viswanathan,K. (t971) A new X-ray method to determine the
anorthite content and structural state of plagioclases.Contribu-
Acknowledgment
tions to Mineralogy and Petrology,30, 332-335.
We would like to thank Ms. G. v. Cdlln and Mrs. I. Schmie-
mann for their careful assistancein collecting the data and draw- Manuscript received, May 20, 1980;
ing tle figures. acceptedfor publication, January 6, 1981.