the chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and...

11
American Mineralogist, Volume 61, pages 853-86j, 1976 The chemical compositions andoriginof thezeolites offretite, erionite, andlevyne WILlIenr S.WrsE Department of Geological Sciences, Uniuersity of Califurnia Santa Barbara. California93106 lNo R. W. TscHBRNrcn 532 Auenue A S nohomish, Washingt on 989 20 Abstract Since offretite, erionite, and levyne have similar structural elements, a chemical study of intergrowths from cavities in basalts was made in order to find chemical parameters that control the crystallization of eachmineral. Chemicalanalyses were made of parts of a complex intergrowth of all three zeolites from Milwaukie, Oregon; of offretite overgrowths on levyne from British Columbia and Arizona; of erionite overgrowthsfrom Rock lsland Dam, Wash- ington; and of intergrowthsof offretiteand erionite from Malpais Hill, Arizona, as well as of offretite and erionite from new localities in Oregon and Arizona. Erionite has Si/Al ratios between3.6 and 3.0 with K * Na ) Ca * Mg; offretite has Si/Al ratios between 2.8 and2.3 with Ca + Mg > K * Na; and levynehas Si/Al ratios between 2.2 and 1.8,and Ca )) Na ) K. Compositional breaksin Si contentsbetween eachmineral indicatethat their crystallization is controlled by the silica activity of the solution. The relatively common occurrence of erionite, rather than offretite, as a diagenetic mineral in siliceous tuffs is apparently related only to silica activity of the interstitialwater. Compositionsof the heulandite or clinoptilolite associatedwith offretite or erionite, respectively,have Si,/Al ratios higher than the range of erionite, suggesting those zeolites result lrom even higher silica activities. K seems an essential constituent for offretite and erionite but not for levyne. Overgrowths of offretite on levyne require increases in activitiesof K and silica. Higher activitiesof K commonly result in the concurrent crystallization of phillipsite with offretite and erionite. Introduction During the past few years,various intergrowths of offretite, erionite, and levyne have been described, including zonal growths and stacking faults in offre- tite and erionite crystals(Sheppard and Gude, 1969; and Kokotallo et al., 1972), offretite overgrowths on levyne (Sheppardet al.,1974) and erionite on levyne (Passaglia et al., 19741' Shimazu and Mizota, 1972). We have found and studied severalmore examples, including one intricate intergrowth of all three zeo- lites. By combining knowledge of growth sequences and results of detailed chemical studies with known crystal structure data, we can place some chemical controls on the origin of thesezeolites. The structures of the three zeolites all have in com- mon hexagonal layersbuilt of double 6-memberrings of tetrahedra(D6R) linked by single6-memberrings (S6R). Different stacking sequences (Fig. l) yield the different kinds of channelsand cagesin each of the frameworks. However, a change from one stacking sequence to another accounts for such features as stackingfaults in erionite (Kokotailo et al.,1972) and the oriented overgrowths on the basal planes of levyne. Chemical compositions of the three minerals show consistent differences, even though the basic struc- tural elements are the same. Typical cell contents can be represented by: offretite(Kr,ca,Mg)ruAlusi13o36. 15-l6Hro Z: I erionite (Kr,Nar,Ca,Mg)r ,Alo zSi,e aOs" ' l5HrO Z : 2 levyne(Ca,Nar,Kr)BAl"Si'ro,".18HrO Z=3 Erionite is the most common of these three zeo- 853

Upload: others

Post on 03-Jan-2021

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

American Mineralogist, Volume 61, pages 853-86j, 1976

The chemical compositions and origin of the zeolites offretite, erionite, and levyne

WILlIenr S. WrsE

Department of Geological Sciences, Uniuersity of CalifurniaSanta Barbara. California 93106

lNo R. W. TscHBRNrcn

532 Auenue AS nohomish, Washingt on 989 20

Abstract

Since offretite, erionite, and levyne have similar structural elements, a chemical study of

intergrowths from cavities in basalts was made in order to find chemical parameters that

control the crystal l izat ion of each mineral. Chemical analyses were made of parts of a complex

intergrowth of all three zeolites from Milwaukie, Oregon; of offretite overgrowths on levyne

from Brit ish Columbia and Arizona; of erionite overgrowths from Rock lsland Dam, Wash-

ington; and of intergrowths of offret i te and erionite from Malpais Hil l , Arizona, as well as of

offret i te and erionite from new local i t ies in Oregon and Arizona. Erionite has Si/Al rat ios

between 3.6 and 3.0 with K * Na ) Ca * Mg; offret i te has Si/Al rat ios between 2.8 and2.3

with Ca + Mg > K * Na; and levyne has Si/Al ratios between 2.2 and 1.8, and Ca )) Na ) K.

Composit ional breaks in Si contents between each mineral indicate that their crystal l izat ionis control led by the si l ica activi ty of the solut ion. The relat ively common occurrence of

erionite, rather than offretite, as a diagenetic mineral in siliceous tuffs is apparently related

only to si l ica activi ty of the interst i t ial water. Composit ions of the heulandite or cl inopti lol i te

associated with offretite or erionite, respectively, have Si,/Al ratios higher than the range of

erionite, suggesting those zeolites result lrom even higher silica activities. K seems an essential

constituent for offretite and erionite but not for levyne. Overgrowths of offretite on levyne

require increases in act ivi t ies of K and si l ica. Higher act ivi t ies of K commonly result in the

concurrent crystallization of phillipsite with offretite and erionite.

Introduction

During the past few years, various intergrowths ofoffretite, erionite, and levyne have been described,including zonal growths and stacking faults in offre-tite and erionite crystals (Sheppard and Gude, 1969;and Kokotallo et al., 1972), offretite overgrowths onlevyne (Sheppard et al.,1974) and erionite on levyne(Passaglia et al., 19741' Shimazu and Mizota, 1972).We have found and studied several more examples,including one intricate intergrowth of all three zeo-lites. By combining knowledge of growth sequencesand results of detailed chemical studies with knowncrystal structure data, we can place some chemicalcontrols on the origin of these zeolites.

The structures of the three zeolites all have in com-mon hexagonal layers built of double 6-member ringsof tetrahedra (D6R) l inked by single 6-member rings

(S6R). Different stacking sequences (Fig. l) yield thedifferent kinds of channels and cages in each of theframeworks. However, a change from one stackingsequence to another accounts for such features asstacking faults in erionite (Kokotailo et al.,1972) andthe oriented overgrowths on the basal planes oflevyne.

Chemical compositions of the three minerals showconsistent differences, even though the basic struc-tural elements are the same. Typical cell contents canbe represented by:

of f ret i te(Kr ,ca,Mg)ruAlus i13o36. 15- l6Hro Z: I

erionite (Kr,Nar,Ca,Mg)r ,Alo zSi,e aOs" ' l5HrO Z : 2

l evyne (Ca ,Nar ,K r )BA l "S i ' r o , " . 18HrO Z=3

Erionite is the most common of these three zeo-

853

Page 2: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

854 W. S. WISE AND R. W. TSCHERNICH

l4- Hedron

6R

R

e : C A G

lites, because it occurs as a diagenetic replacement ofsilicic tuffs as well as an amygdule mineral. Levyneforms in the cavities of basalts, such as those in theColumbia Plateau, Oregon; the Garron Plateau, Ire-land: and the Deccan basalt of India. From our ob-servations most levyne crystals have at least a thin,fibrous overgrowth of offretite. Offretite also occursas single crystals or sprays of crystals in amygdules ofalkalic, mafic lavas. Although the interpretation ofzeolite sequences in cavities is tenuous, the relativelylarge crystals allow detailed chemical study of inter-growths, and thereby suggest some sequences ofchemical changes during the formation of the differ-ent phases.

Experimental methods

All the chemical analyses in this study were per-formed by an electron probe microanalyzer on sam-ples that ranged from relatively thick plates oflevyneto fibrous growths of offretite. Whole clusters andintergrowths were embedded in epoxy and ground toexpose the centers of crystals or intergrowths.Mounts were then polished and coated with approxi-mately 250 A of carbon. An accelerating voltage of l5kV was used for all elements with a sample current ofl0 nA. Standards used were andesine (Ca and Al), K-feldspar (K and Si), olivine (Mg), hematite (Fe),sanbornite (Ba), and celestite (Sr). Emission data

dxo

qo|Nt

(J

- HEDRON

I

O F F R E T I T E E R I O N I T E LEVYNEFtc. l .Summaryof thecrystalst ructuresofof f ret i te,er ioni te,andlevyne(af terGardandTai t , l9T2;Breck, l974;and BarrerandKerr ,

1959). All three structures are built ofsingle 6-member rings oftetrahedra (S6R) arranged in various stacking sequences, but all retaininghexagonal symmetry.

Ofretite. The S6R stacking sequence is AABAAB . . . , developing three kinds of cages: double 6-member rings (D6R) betweensuccessive,4 rings; e cages (also called cancrinite cages) between successive D6R's; and l4-hedra (also called gmelinite-type cavities)between successive B rings. Cation locations: K+ in €-cages; a few Ca'+ in D6R's; hydrated Mg'?+ in the l4-hedra; and hydrated Ca'+ inthe wide channels. The space group is P6m2.

Erionile. The S6R stacking sequence is AACAAB . . . , giving a c dimension double that of offretite and developing three kinds of cages:D6R; e cages; and 23-hedra between successive B rings and successive C rings. Cation locations: K+ in 6 cages; a few Ca2+ in D6R's; allothers surrounded by water molecules in the very large 23-hedron. Space group symmetry is P6"/mmc.

Leuyne. The S6R stacking sequence is BBCAABCCA . . . , giving a c dimension three times that of offretite, and developing two kindsof cages: D6R, and l7-hedra, which are stacked in a one D6R and two l7-hedra sequence parallel to the c axis. Cation positions have notbeen determined. Space group symmetry is R3m.

\ cxltiltEu

Page 3: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

OFFRETITE. ERIONITE, AND LEVYNE 855

were reduced with a modified version of the computerprogram Enpenn 7 (Rucklidge and Gasparrini, 1969,Department of Geology, University of Toronto). Inorder to overcome some difficulties in analyzing zeo-lites by Etvtx methods, large beam diameters (10 to 20micrometers) were used, the beam was moved afterevery lO-second counting period, and only the largestcrystals or crystal clusters were analyzed. Where pos-sible multiple spots were analyzed; the averaged re-sults are listed in Table l Unit-cell contents calcu-lated from these analyses are listed in Table 2.

Most of the samples were photographed with ascann ing e lec t ron mic roscope; charac ter is t i cmorphologies are illustrated in Figures 2 through 7.The samples were coated with 200 to 400 A of AuPd.The thinner coats are less satisfactory because ofsome charging effects on thin needles.

If sufficient material was available, each samplewas X-rayed to verify identification. However, offre-tite and erionite intergrowths can only be identifiedby the optical methods described by Sheppard andGude (1969). Our chemical and X-ray data support

Tnrln l. Averaged microprobe analyses of offretite, erionite, levyne, and some associated zeolites

I

s io2 57.07

NZ03 15.1+3

F"203 2.32

Mco 0.56

CaO 5 ,97

Naro 0 .0

KzA 2.39

Total : 8\ .71+

2

5)+. oo

0 . 93

0 . 39

8 . oo

0 . 032 q R

8 1 . 5 9

3

) 1 . t I

r 7 .70

2 . O 7

o . I+2

7 ' l +o

0 . 0

2 . 3 \

8l+. 90

l+

20,62

o . 2 6

0 . 28

10. 00

o . 5 2

0. 8 l+

88.25

5

55.97

r9. 9 l+

0 . 39

o . t or c A

I l , 0 5

RR q i

6

55.99

20.62

0 . 0

0 . 0

7 .6r

3 . 81

88 . 25

7

\7 .9L

0 . 0

0 . 0

9 . 9 8t i ?

o . \ 8R "

" l

B

5r.78

20.79

0 . 0

0 . 0A ) "

U , J O

90.59

o

50.29

22.L8

0 . 0

0 . 0

10 . b7

L . z 5

I . l+2

85.62

t l

56 .U+ , 7 .2 I

20.01+ 1? .10

0 . r 0 0 . 0 9

o .25 1. o l+

8 . o )+ \ . 95

0 . 7 3 0 . 1 5

3 . 2 2 \ , 5 6

88 .83 85 .2L

11p

si02 5l+.77

Aro3 23.53

F .203 0 .03

MCo 0,10

cao )+.67

Sr0 0 .0

Bao 1 .09

NarO 0 .51

Kzo 6 .73

Total 91.1\

l t h

o J . ) l

f ) , u I

0 . 0 8

1. )+8

4 , > 3

u . ) o

0 . 1 5

0 . 0

L . 6 2

87. 00

T2

57.80

o. oBo . 9 7

> . z o

0 . 0

0 . 0

o . 5 9

5 . 2 6

85 . 1 r

rzp63.7\18 . 90

0 . 0 8

0 . 0? 7 2

0 . 0

u . ) o

3 . 0 8

) , 5 v

L 1 C

59.581 ) a i

0 . 0 2

0 . 5 6

0 . 0

0 . 0

0 . 7 9

r. 8)+

78 .79

57.30f o . z J

0 . 02I t ?

\ . 6 2

0 . 0

0 . 0

0 . r 7l+ . fB

83.69

14

) t . c )

0 . 0

o ,3 l+

L ' 9 r

0 . 0

0 . 0

\ .7t+

o b . b 4

60 . f i

l o . ! y

0 . r 0

0 . 88

5 . 3 !

0 . 0

0 . 0

0 . 2 0

3 . 7 3

85 . 50

59.85

L5.97

0 . 0

0 . 0

o . 0

r . o l

) 7 1

6 . 2 7

89 . o3

t)c ro

65 .90 62 .52

13 .10 16 . I+8

0 . 0 4 0 . 0 1

o . 05 t . r 2

\ . l+t 2. \9

0 . 0 0 . 0

o . \ 2 o . o

o . 5 8 r . 7 2

2 .39 l + . 8L

85 .90 89 .18

1. Erionite in conplex intergrorth (see figue 8c), quary at corner of Lava and River Rds., Milvaukie' ore.2. Offretite in complex intergrowlh (see figue 8c), quury at corner of Lava and River Rds., Milmukie' Ore.3. Erionite with offretite in complex intergrorth (see fi-gue 8c), quarry at corner of Iava and River Rds., Milvaukie' ore.l+. Levyne ln conplex intergrorth (see figue 8c), quarry at corner of Lava ancl River Rds., MiJ.mukie,Ore.5. Offretite in complex intergrorth (see figue 8c), qurry at comer of Lava and Rlver Rds., Mllwaukie' Ore.5, Offretite on levyne, along Douglas Leke Rd,, southvest of Westwold., s. Brlt lsh Colwbia.7. Levya6' along Douglas I€ke Rd,, aouthwest of Westvold, s, Brit ish Colubia8. Offretite on levyne, along Queen Creek, 5 uiles vest of Superior, Pinal Co., Arizona9. Levyne, along Queen Creek, 6 Eiles vest of Superior, Pinal Co., Arizona

10. Of f re t i te , road cu t 2 . I n i les s . o f Owl rhee Dm, Malheur Co. , OregonLI . o f f re t i te (+er ion i te ) , ra i l road cu t , n ,e , s id .e Malpa is H i l I , 5 n i les s . w ink l@an, P ina l Co, , Ar izonaI lp Ph i l l i ps i te , ra i l road cu t , n .e . s i t le Ma lpa is H iU, 6 n iLes s . Wink le@n, P ina l Co. , Ar izona11h Heu lmdi te , ra i l road cu t , n .e . s ide Mai lpa is H i11 , 5 mi les s . Wink lenan, P ina l Co. , Ar izona

12. offretite, Rock Island D@I2p PhiUipsite, Rock Island DmL2c ClinoptiLolite, Rock fsland De

13. ETionite,along nev roacl cuts 0.1 i l i le north of Clifton, Greenlee Couty, Arizona

L\. Erionite, blocks of basalt 2 miles north of Thub Butte, eastern Grahm County, Arizona

15. Erionite, quarry at Cape Lookout, Tallmook Co., oregonI5p Phluipsite, querry at Cape Lookout, Til luook Co., oregonIlc Clinoptilolite, quarry at Cape lookout, Til lmook Co., Oregon

15. Erionite, Road cuts at Yaquina Head, IJincoln Co., Oregon

Page 4: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

W. S, WISE AND R, W. TSCHERNICH

Tnst-r 2. Cell contents* and cell dimensions of analyzed offretites, erionites, and levynes

SmpleNwber

I

2

3

l+

5

5

T

B

9

l0

11

I2

13

1l+

15

L O

Mineral ( s )

' ' i teraonate

+ +offretite

er ion i te( +offretite )

rr

fewne

offretite

offretite

le{fne

offretite

levyne

offretite

offretite( +erionite )

offretitef + c r i a n i + a )

er ion i te

erionite

erionite

er ion i te

Locality K

Mi lwauk ie , Ore . O.75

M i l m u k i e , O r e . O , 7 9

MilvauJ<ie, Ore. O.?L

Milvaulie, Ore. 0.21t

Mi lvauk ie , Ore . 1 .17

near Westwo ld , B .C. 1 ,10

near Westvo ld , B .C. 0 .15

near Super io r , Az . 1 .58

near Super io r , Az . 0 .113

OFyhee Dm, Ore . O.92

Malpa is H i1 t , Az . 1 .38

R o c k I s . D m , I . 5 7Washington

Cl i f ton , Az . L .26

Thmb But te , Az , 1 .38

uape rcoKour , u re . a .uo

raqu lna nead, u re . l . Jo

A1 Si

L .5S ) - 3 .5 r

5 .23 l 2 . 77

I+. 9)+ 13 . 05

j . \ 5 12 . r -

5 . 3 3 \ 2 . 6 8

5 . \ 9 12 .66

6 . 5 9 1 1 . 5 0

5 . \ 8 12 . \ ' f

6 . L j 1 1 . 8 2

5 . 3 1 1 2 . 8 3

\ . 6 9 1 3 . 3 0

\ . \ 5 1 3 . 5 3

! . 5 0 1 3 . 5 0

\ . 4 8 1 3 . 5 8

4 , 5 3 r 5 . O )

t + . 28 13 . 78

1 3 . 3 3 5 ( 8 )

1 3 . 3 8 5 ( 5 ) 2 3 . 0 ( r )

1 3 . 3 1 5 ( 1 ) 7 . 5 5 6 ( 5 )

1 3 . 2 7 8 ( B )

1 3 . 2 8 0 ( 3 ) 1 5 . 1 1 ( r )

1 3 . 2 8 3 ( 3 ) 1 5 . 1 r ( r )

8 . 2 6 7 ( 3 ) 1 5 . 1 1 ( 3 )

B . z t+5Q) 15 .13 (2 )

Na Ca Mg Fe

o . 0 1 . 5 8 0 . 1 9 0 . 0

o . o 2 . 0 6 o . r L 0 . 0

o . o r - 9 6 o . r 5 o . o

0 . 2 3 2 . 1 r 0 0 . 1 0 0 . 0 1

0.69 I . 5 l+ 0. 13 0. o l+

0 . 1 0 r . 8 \ 0 . 0 0 . 0

0 . 9 ) + 2 . 5 7 o . o o . o

0 . 1 5 r . 9 7 0 . 0 0 . 0

0 . 5 8 2 . 6 3 o . o o . o

o . 3 2 1 . 9 1 + 0 . 0 9 o . O 2

o . o 7 r . 2 3 0 . 3 5 o . o 2

o . o 2U . J 40 . 98o . 2 7

0 .08 1 .17 0 . l +1 0 .01

o . 2 5 ) - . 2 O 0 . 1 1 0 . 0

o . o 9 1 . 3 0 0 . 2 9 O . O 2

0 . 7 4 0 . 5 9 0 . 3 7 0 . 0

* Based on 35 orygens for+ Except for smples 6 and

dev ia t ion app l ied to thettThese ce11 contents have

all three ninerals11, these are leas tfigure imediatelybeen correctetl for

squiles refinenents, the figure in parenthesisp r e c e d . i n g i t . _ T h u s , 1 5 . 1 1 ( 1 ) i s 1 5 . 1 1 i 0 . 0 1 .inclucled clay (iron-rich snectite) vhich ranged

represents the calcul-atetl

up to 2/,,

their assertion that offretites are optically negativeand erionites are optically positive.

The unit-cell dimensions l isted in Table 2 wereobtained by least-squares refinement of powder X-ray diffraction data, using the U. S. Geological Sur-vey's FoRTRAN IV computer program W7214. Sil icon(Natl Bur. Stand. Snna 640, a : 5.43088 A) was usedas an internal standard for oscil lations over the peakslisted by Sheppard and Gude (1969) and Sheppard etal. (1974). Since very small quantit ies of two samples(6 and ll) were available, the X-ray powder datacould not be refined, and a was calculated from dnoo.

Descriptions of analyzed samples

Milwaukie, Oregon

Intricate intergrowths of erionite, offretite, andlevyne occur in some vesicles in basalt, exposed in asmall quarry (now a vacant lot at the corner of Lavaand River Roads, Milwaukie, Oregon). The basalt, aflow in the Yakima Basalt section west and south ofPortland (Trimble, 1963), has a coarse, open texturewithout phenocrysts. Small crystals of sti lbite andchabazite are common in the vesicles, but the erion-

ite-offretite-levyne intergrowths are sparsely distrib-uted.

There are five stages in the development of theintergrowths; most of the intermediate stages can befound in some vugs (Figs. 2-4). The cross section(Fig. 8c) was drawn from the section cut for micro-probe analysis.

Stage l. Erionite crystallized as needles clustered inthe form of a tapering hexagonal prism (Point l, Fig.8c, and Analysis 1). Inclusions of an iron-rich smec-tite color these clusters pinkish orange. Identif icationof the erionite was made by both X-ray and opticalmethods.

Stage 2. Irregular crystals of offretite with someerionite interlayered with clay formed hollow, hex-agonal cylinders, twice as high and enveloping thefirst stage prism (Fig. 2). The entire growth is inoptical continuity, but the layers shown in Figure 8cas erionite have positive elongation, while the re-mainder is negative. Only the negative part (offretite)was analyzed as Point 2 (Analysis 2).

Stage 3. Compact subparallel needles oyerlie theopen, porous growth of Stage 2. This zone (Figs. 3and 4) includes less clay and is composed entirely of

Page 5: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

OFFRETITE, ERIONITE, AND LEVYNE E57

Ftc. 2, Scanning clcctron micrograph of an offre-tite-crionitc-clay (iron-rich smectite) intergrowth, Milwaukie,Oregon. The open intergrowth has all crystals in optical continuityand the c axes are nearly vertical.

erionite (identified by the positive elongation). Be-cause this zone is so narrow, the analysis of a portionof this zone (Analysis 3) probably includes someoffretite.

Stage 4. Thin plates of levyne (about 0.05 mmthick, Analysis 4) grew from the base of the corecylinder outward at least 0.5 mm. There are severalspecimens of multiple, parallel plates; however, thoseshown in Figure 3 are typical. The c axes of the levyneand earlier-formed erionite and offretite are parallel.

Stage 5. An overgrowth of very thin offretite fibers(Analysis 5) added as much as 0.1 mm to all theexposed parts of the intergrowth. Figure 4 illustratesthe'long, wispy needles in an early part of this stage.Continued growth would cover this exposed surface,giving it an appearance like that in Figure 5. Identi-fication of this offretite was based on its negativeelongation.

The balls of plates in Figure 2 and 3 are iron-richsmectite that recrystallized after stage 5 of zeolitegrowth.

Offretite ouergrowths on leuyne

New localities with particularly striking offretiteovergrowths on levyne have been found southwest ofWestwold, along the road to Douglas Lake, southernBritish Columbia, and along Queen Creek, 6 mileswest of Superior, Pinal County, Arizona. The West-

Ftc. 3. Scanning electron micrograph of offretite-levync inter'growth, Milwaukie, Oregon. The clay-filled cylinder is composedlargely of offretite with some intergrown erionite, and the levyneplates have a thin overgrowth of offretite (white fibers)'

wold offretite overgrowths are extraordinarily thick'1.0 mm on levyne plates 0.2 mm thick (Analyses 6and 7). Although scarce, the Queen Creek over'growths are exceptional in that they are thin on rela-

Frc, 4. Scanning clectron micrograph of offretitc-lcvync intcr-growth, Milwaukie, Oregon. (Dctailed view of part of Fig, 3.) Thcview shows thejoin between the levyne plate and the offretite corecylinder, but with long fibers of offretite overgrowing both earlierformed structures.

Page 6: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

858 W. S. WISE AND R. W. TSCHERNICH

FIc. 5. Scanning electron micrograph of offretite overgrowths onlevyne from Queen Creek, near Superior, Arizona. Note that theoffretite fibers have grown only on the basal faces of the levyne.

tively thick levyne (Fig. 5), and clearly show growthonly on the basal planes (Analyses 8 and 9).

Owyhee Dam, Oregon

Fibers of offretite forming hexagonal prisms occursparsely at Owyhee Dam, Malheur County, Oregon

(see Staples et al., 1973, p. 406, for a description ofthe locality and host rock). The prisms are about 0.3X 1.0 mm and are composed entirely of offretite(Analysis l0).

Offre t it e-e rioni t e int e rg row ths

Particularly interesting occurrences of offretitewith erionite intergrowths are found at Malpais Hill,6 miles south of Winkleman, Pinal County, Arizona,and two miles south of Horseshoe Dam, MaricopaCounty, Arizona. Malpais Hill is a Plio-Pleistoceneolivine basalt volcano that erupted subaqueously in ashallow lake. Vesicular blocks, colored orange by thehydrated and oxidized glass, contain vugs that arelined with celadonite and montmorillonite, which arein turn covered by small crystals of offretite, heuland-ite, and phillipsite. The Horseshoe Dam locality is abed of river gravels containing orange vesicular cob-bles of olivine basalt having the same mineralogy asthe Malpais Hill basalt.

The Malpais Hill offretite crystals are sharp, hex-agonal prisms about 0.1 mm long (Fig. 6). However,the crystals are actually intergrowths of offretite anderionite; the zones shown in Figure 8a were identifiedby the sign of elongation. The analyzed spots (Analy-sis ll) undoubtedly included several of the narrowerionite bands, and therefore represent an average forthe crystal. The cell parameters from X-ray diffrac-tion data likewise represent averages.

Ftc. 6. Scanning electron micrograph of offretite from MalpaisHill, Pinal County, Arizona. The hexagonal prisms actually areintergrown offretite and erionite (see Fig. 8a), growing on celado-nite.

Frc. 7. Scanning electron micrograph of erionite from CapeLookout, Tillamook County, Oregon. The clusters of striatedprisms are typical of erionite crystals having grown in basalt cav-ities.

Page 7: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

OFFRETITE. ERIONITE, AND LEVYNE

Frc. 8. Sketches of offretite-erionite-levyne intergrowths.(a) Alternating erionite (shaded) and offretite in a stacking fault arrangement (associated with phillipsite),

Malpais Hi l l , Ar izona.(b) Overgrowth of erionite (shaded) on offretite from Rock Island Dam, Washington.

(c) The complex intergrowths of offretite-erionite (shaded)-levyne, Milwaukie, Oregon. The numbers

refer to stages of growth and to analyses in Tables I and 2. (1 ) erionite with clay, (2) offretite wjth clay and

erionite, (3) erionite, (4) levyne, and (5) offretite overgrowth.

Zonal growths of offretite and erionite within asingle crystal occur at Rock Island Dam, Washing-ton, which is the type locality of paulingite, describedby Kamb and Oke (1960). This growth (Fig. 8b) issimilar to that observed in some offretite crystalsfrom Mt. Simiouse near Montbrison, Loire, France(Sheppard and Gude, 1969). Again the analyses ofthese crystals (Analysis 12) probably represent anaverage of the two minerals. Long, very thin fibers oferionite also occur in the same rocks. Comblete anal-ysis on these fibers, about 5 pm in diameter, were notobtained, but K:Na:Ca is approximately 50:30:20,typical of many erionites but far more soda-rich thanthe offretite crystals.

Erionite localities

Two localities in southeastern Arizona (near Clif-ton, Greenlee County, and 2 miles north of ThumbButte, Graham County) have yielded erionite in ves-icles of mid-Tertiary olivine basalts. The erioniteforms bundles and sprays up to 2 mm in length

(Analyses 13 and l4). Common associated mineralsare opal, quartz, mordenite, clinoptilolite, and phil-lipsite.

Erionite.also occuis in amygdules of Miocene tho-leiitic basalts in two localities along the coast of Ore-gon at Cape Lookout, Tillamook County (Analysis15 and Fig. 7), and at Yaquina Head, LincolnCounty (Analysis l6). The zeolite assemblage at CapeLookout is complex but includes mordenite, clinop-tilolite, phill ipsite, and dachiardite. The YaquinaHead erionite occurs with clinoptilolite.

Discussion

As with most zeolites, the chemical compositionsof these three minerals are nonstoichiometric and canbe characterized by the Si/Al ratio in the frameworkand the content of the charge-balancing cations. Fig-ure 9, a plot of the relationships between the Si con-tent and the relative proportions of monovalent anddivalent cations, shows that the three phases differmainly with respect to their Si/Al ratios. The fields

Page 8: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

860 W. S, WISE AND R. W. TSCHERNICH

. _ a a ^ 1 6

st 2 ' A -

f l l p

a a

L"o - - -0rsp

for offretite and erionite may overlap, but if Points 3,ll, and 12 represent analyses of mixtures (Figs. 8aand 8b), a miscibility gap may exist. The sharpnesswith which the phases can be distinguished opticallysuggests that gradations between the phases do notoccur. We conclude, therefore, that the differentstacking sequences are accompanied and probablycontrolled by differences in the Si/Al ratio. The anal-yses of the offretite and levyne intergrowths demon-strate a significant compositional gap.

The Si variation in the erionite field is small fi3.5silicons per 36 oxygens at Clifton, Arizona, to 14.lper 36 in a sedimentary erionite from MalheurCounty, Oregon, giving Si/Al values between 3.0 and3.6), but there is a wide variation in the relativeproportions of divalent and monovalent cations

Erionites

(R++/R+ ratio). Offretite compositions are somewhatmore restricted; Si varies from 13.3 in those crystalsintergrown with erionite to 12.0 in overgrowths onlevyne, giving Si/Al ratios of 2.0 and 2.8; R++/R+varies only between 2/ | and | / l. The levynes plottedhere vary from I1.5 and 12.4 Si per cell, giving Si/Alratios of |.7 to 2.2, and have a great predominance ofdivalent cations.

Table 2 and the analyses in Sheppard and Gude(1969) show that the sum of all non-frameworkcations varies between 2.5 and 4.0. with values near 3most common. Moreover, the crystal structure workof Gard and Tait (1972), illustrated in Figure l,shows thal a single layer within any of these struc-tures has room for only four cations. An averageerionite with 4 Al and therefore a framework charse

75

Of f ret i tes

3/t ul t / |

( R** / R* )

FIc. 9. Plot of Si-(Ca + Mg)-(Na, + Kr) of the analyzed minerals in Table l. Numbers correspond toanalyses. Circles are erionites (solid circles are from Sheppard and Gude, 1969), squares are offretite (half-filled squares are intergrowths of offretite and erionite). Triangles are levynes. Arrows connect the growthsequence of the Milwaukie intergrowths, and lines connect the offretite overgrowth with the host levyne.Sample B is Beech Creek, Oregon, offretite-levyne (Sheppard et al., 1974); Sample F is original offretite,Montbrison, France (Sheppard and Gude, 1969). Diamonds labeled p, h, or c indicate associatedphillipsite, heulandite, or clinoptilolite, respectively.

t / ?

5 t LevYne

Page 9: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

OFFRETITE. ERIONITE, AND LEVYNE

of -4 can be balanced with any combination of non-framework cations, including 4 monovalent ones. Asthe framework charge increases with decreasing Sicontent, more divalent cations are required to bal-ance the charge, without exceeding the limit of fourcations.

Since Al is mostly in the D6R cages (Gard andTait,1972), the a dimension of the unit cell is respon-sive to the Al substitution (Fig. l0). The trend foreach mineral is generally the same, although the best-fit l ines are not coincident.

Figure I l, a plot of Ca * Mg, Na, and K, showsthat erionite has a wide range of compositions butgenerally with K * Na ) Ca * Mg. The offretitefield has Ca * Mg > K + Na, while the levynes haveCa ) Na ) K. The new data support the general-izations made by Sheppard and Gude (1969) thatmost offretites contain very little Na, and that erion-ites and offretites contain at least 0.75 K per layer.

It is also clear from Figure I I that erionites withvery little Na also exist, and therefore the formationof the erionite or offretite framework does not de-pend on Na content. The occurrence of erioniterather than offretite as a diagenetic mineral in si-liceous tuffs is dependent only on silica activity,which is controlled to a large extent by the tuff com-position and the pH of the water.

The experimental work of Aiello and Barrer (1970)

and the structural studies of Gard and Tait (1972)indicate that "each cancrinite cage (e-cage) 'collects'

around the K ion as a precursor to crystallization,and the frame is subsequently built up by con-densation of these cancrinite cages" (Gard and Tait,1972, p.832). Table 2 and Figure I I show that levynecontains very little K, which can be explained by thelack of €-cages in the levyne structure.

Two other zeolite minerals intimately associatedwith erionite and offretite in the assemblages studiedhere are heulandite or clinoptilolite and phillipsite(Fig. 9). The heulandite and clinoptilolite with higherSi/Al ratios are clearly the result of increases in silicaactivity beyond values that erionite is able to accom-modate. The associated phillipsites, having nearly thesame Si/Al ratio, are probably responding to someother variable, such as high as+ ot low 4s,o.

Implications for natural zeolite growth

In an earlier paper (Wise and Tschernich, 1976),we argued that the activities of SiO, and HrO are theprincipal variables that control the crystallization ofhigh-silica zeolites. The compositional gap with re-spect to Si that probably exists between erionite andoffretite suggests that again 4s1e, plals a primary role .Since cation composition fields overlap, activities ofthese cations are apparently not important. Any twoof the structures do not seem able to grow contempo-

ct)a5()

oosf

(DN

5d

30.8

3 0 9

3 r . o

3r I

31.2

5 t . 3

13.35 \ Bo

\ 6\o

\

Offrelite

t otr

\ ro \

i l \ 1 3 1 4\ s 6 0'

\ ^ 1 5\ .

n 1 6o - \

Erioni te ' \O

Si otoms per 36 oxygen oioms

Frc. 10. Variation of the a dimension with Si content of the unit cell, based on 36 oxygens. Symbols are

the same as in Fig.9. Error bars represent plus and minus one standard deviation. The 400 peak appears

only in offretite and erionite patterns.

Page 10: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

862 W. S. WISE AND R, W, TSCHERNICH

C o + M g

C o + M g > N o + K

o l 4s 1 2a

No+K>Co + Mg

N o KFtc. I L (Ca + Mg)-K-Na plot of offretite, erionite, levyne. The symbols and lines are the same as Fig. 9.

raneously from the same solution. The reason whythe Si/Al ratio affects the stacking sequence is notknown, but the problem probably could be solved bydetermining Al locations in all three zeolites. Forexample, if the 6 Al's in levyne only occur in the D6Rgroups, the negative framework charge around thesecages may be too strong to be shielded by cationsforming € cages. Therefore, lower structural energiesare obtained by rotation of D6R cages (Fig. l).

From the data presented here and the experimentalwork of Aiello and Barrer (1970), it is reasonablyclear that K+ is necessary to form the e -cages prior toframework construction of offretite on levyne, andthe crystallization of offretite may be a response toincreases in either 4s1e, or ax+ or both. The crystalli-zation of levyne (low Si/Al and low K) can causeincreases in these constituents.

AcknowledgmentsWe are greatly indebted to Mr. Robert Mudra of Mesa, Ari-

zona, for generously providing samples and showing us the Ari-

zonazeolite localities he has found. Mr. and Mrs. Robert Hillsdonprovided samples of the British Columbia zeolites. We appreciatethe discussions and suggestions of Dr. James Boles concernlngzeolite crystallization. Thanks also go to the Research Committee,University of California, Santa Barbara, for computer time usedfor reduction of microprobe and X-ray diffraction data.

References

AIELLo, R. exo R. M. BARRER (1970) Hydrothermal chemistry ofsilicates. Part XIV. Zeolite crystallization in presence of mixedbases.,/. Chem Soc. (A), 1470-14'15.

BARRER, R. M. lNo I. S. Ksnn (1959) Intracrystal l ine channels inIevynite and some related zeolites. Trans. Faraday Soc 55,t9l5-t923.

Bnncx, D. W (1974) Zeolite Molecular Sieues, John Wiley andSons, New York, 771 p.

Glno, J. A. aNo J. M. Tlrr (1972) The crystal structure of thezeolite offretite, K, ,Ca, ,Mgo, [Sir, "Alu rOru] 15.2 H2O. ActaCrystallogr. B 28, 825-834.

Krnrs, W. B. eNo W. C. OKE (1960) Paulingite, a new zeolite, inassociation with erionite and filaform pyrite. Am. Mineral. 45,79-91.

KororlrLo, G. T., S. Slwnur lNn S. L. LewroN (1972) Direct

Page 11: The chemical compositions and origin of the zeolites ...ington; and of intergrowths of offretite and erionite from Malpais Hill, Arizona, as well as of offretite and erionite from

OFFRETITE, ERIONITE, AND LEVYNE 863

observation of stacking faults in the zeolite erionite. Am. Min-eral. 57, 439-M4.

Pnssnclte, E., E. Glt-t AND R. RTNALDT (1974) Levynes and erion-ites from Sardinia, ltaly. Contrib. Mineral. Petrol.43,253-259.

Sunrlno, R. A. rNp A. J. Guon, 3no (1969) Chemical composi-tion and physical properties of the related zeolites, offretite anderionite. Am. Mineral. 54, 875-886.

G. A. DrssoRoucH AND J. S. Wsrrr, Jn. (1974)Levyne-offretite intergrowths from basalt near Beech Creek,Grant County, Oregon. Am. Mineral. 59,837-842.

Suturrzu, M. aNo T. MrzorA (1972) Levyne and erionite fromChojabaru, Iki Island, Nagasaki Prefecture, Japan. J. Jap. As-soc. Mineral. Petrol. Econ. Geol.67.418-424.

Srrelrs, L. W., H. T. EveNs, Jn. ero J. R. LtNosev (1973)Cavansite and pentagonite, new dimorphous calcium vanadiumsilicate minerals from Oregon. Am. Mineral.5t' 405-411.

Tnrvne, D. E. (1963) Geology of Portland, Oregon and adjacentareas. U. S Geol. Suro. Bull. lll9, l19 p.

Wrsr, W. S. rNo R. W. TscuenNtcu (1976) Chemical composit ionof ferrierite. Am. Mineral 61. 60-66.

Manuscript receiued June 6, 1975;accepted for publication October 14, 1975.