American MineralogistVol. 57, pp. 85-102 (1972)

CHEMIOAL COMPOSITION AND PHYSICAL, OPTICAL,AND STRUCTURAL PROPERTIES OF BENITOITE,

NtrPTUNITE, AND JOAQUINITE1

Jo Lerno AND ARDEN L. Ar,nno, Diuision o! Geological and PlanetarySaences, Calilonuia Insti,tute of Techmology

Posadena. C alif ornia 91 109

Assrnect

The chemical composition, density, optical properties, and cell parameters ofbenitoite, neptunite, and joaquinite from San Benito County, California, arereported. Electron microprobe analyses show that benitoite is homogeneous,stoichiometric Ba.TiSLOe. Electron microprobe and emission spectrographic dataindicate that neptunite is fairly homogeneous and tha.t its formula is Lio uNa".tIG.a(Fe,Mg,Mn)".oTLoSi"oO-. Joaquinite is quite complicated chemically, containinghydroxyl and 1F-20 weight percent rare earth oxides. The density and'opticalproperties for each mineral and ceII parameters for benitoite and neptunite aresimilar to those given in the literature. Weissenberg and precession X-ray diffrac-tion data indicate that single crystals of joaquinite have varying proportions of amonoclinic and an orthorhombic component. The orthorhombic unit cell hasparameters a - 10.48 (2) A,b - 9.66 (2) A, and c - 22.26 (2) A; the monoclinicunit cell has parameters o - 10.51 (2) A, b - 9.66 (2) A, c = 11.82 (2) A, andp = 109.5 (2)'. Structural, chemical, and infrared data suggest the followingformula for joaquinite :

Bae.r(Srt "REu uTho.t) (Cao nNan

"Fe" tl,io "Mgo t)

Tir.tSt oO*.r(OII)*.r.

IwrnooucrroN

The three rare minerals benitoite, neptunite, and joaquinite occurin natrolite veins which cut a glaucophane schist inclusion in a ser-pentinite body near New Idria, San Benito County, California. Loud-erback and Blasdale (1909) described this occurrence in detail; theygave wet chemical analyses for benitoite and neptunite and physicaland optical properties for all three minerals. No further chemical datafor benitoite have been reported. Benitoite is of crystallographic in-terest because it is the only known member of the ditrigonal-dipyram-idal symmetry class; its structure was determined by Zachariasen(1930). Recently, Fischer (1969) refined the crystal structure ofbenitoite and confirmed Zachariasen's (1930) results.

Wet, chemical analyses of neptunite from this locality were givenby Bradley (1909) and more recently by Cannillo, Mazzi, and Rossi

' Contribution number 1979.

85

86 LAIRD AND ALBEE

(1966). A summary of the optical properties and geologic occurrencesof neptunite was given by Heinrich and Quon (1963). Berry (1963)and Cannillo et aI. (1966) reported crystal structure data for SanBenito neptunite.

Palache and Foshag (1932) determined the physical and crystal-lographic properties and chemical composition of California joaquin-ite. Other occurrences of joaquinite have been discussed by BelI (1963)and Semenov, Bukin, Balashov, and Sprensen (1967). Semenov ef ol.(1967) found 22.59 weight percent (RE)2OB (rare earth oxides) injoaquinite from southern Greenland and suggested that there weredifferent minerals of the joaquinite group;they also reported 11.5 and15.0 percent (RE)zOa in two samples of joaquinite from Californiabut did not give complete chemical analyses. All ihe published workon joaquinite has reported it as orthorhombic. However, J. E. Row-land (written communication from E. H. Nickel, 1970) has foundjoaquinite from San Benito County with a monoclinic unit cell.

The purpose of this work was to determine the complete chemicaleomposition and other properties of benitoite, neptunite, and joa-quinite from San Benito County, California. A secondary objectivewas to determine whether the electron microprobe techniques used inthis laboratory could be employed successfully to analyze such acomplex mineral as joaquinite.

ANer,vrrcer, Tncnxreups

Electron microprobe analyses were made with a three-channel Applied Re-search Laboratory model EMX mieroprobe using the techniques of Bence andAlbee (1968) and Albee and Ray (1970). The accelerating voltage was 15 kv atall times; and the beam current, pulse height selection, spot size, and countingtime were eonstant for each element analyzed but adjusted for each element andmineral to obtain maximum counting rates with minimum sample damage andcontamination. Simple silicates and oxides were used as standards for the majorelements; the rare ea.rth elements of joaquinite were analyzed relative to apatite,thalenite, and monazite. Ba and Ti in benitoite were analyzed relative to Ba-feldspar and synthetic rutile, respectively; whereas, Ba and Ti in neptunite andjoaquinite were analyzed relative to benitoite, assuming stoichiometry for thebenitoite. Previous studies in this laboratory indicate an accuracy for commonelements constituting more than about one weight percent of the sample of abouttwo percent of the amount present. Standards and correction parameters for lesscommon elements are not as good, and their abundanse in joaquinite somewhatdecreases the accuracy for the common elements.

The emission spectrographic analyses are an average of duplicate burns(except for the joaquinite samples, which were only run once because of the smallamount of sample available) on a 3.4 m Wadsworth-mount spectrograph with a15,000 line per inch diffraction grating. A D.C. arc with a lg amp short circuit wasused. The samples were di luted in the proport ions: 1.0 sample,0. l Na,COe,0.5

BENITO'ITE, NEPTT]NITE, AND JOAQUINITE 87

graphite, and 0.9 quartz. For benitoite and neptunite 25 mg of this mixture were

weighed into the electrode and burned to completion; for joaquinite 10 mg were

used.X-ray diffraction powder work was done in a Guinier camera using qua'rtz-

monochromatized Cu Ka radiation. Single-crystal work was done in Supper

Weissenberg and precession camera"s with Fe Ka and Mo Ka radiation, respec-tively. The unit cell dimensions for benitoite and neptunite were calculated by

a least squares technique using previously indexed lines; 44 lines were fitted for

benitoite and 34 for neptunite. Indices were calculated for benitoite from the cellparameters of tr'ischer (1969) and for neptunite from the paramenters of Berry(1963).

The densities were determined on a Roller-Smith Berman balance usingtoluene. For benitoite and neptunite determinations were made on individualgrains, weighing from 4-40 mg, and averaged. Because of the limited amount ofmaterial available, the density of joaquinite was determined by weighing 2-4mg samples composed of several grains three times. Indices of refraction weredetermined with Na lieht a'nd index oils calibrated with a Zeiss Abb6 refrac-tometer. Those indices which were matches with the oils are cited as accurateto 0.001; indices bracketed by two oils are cited as accurate to 0.002.

The infrared spectrum of joaquinite was measured on a Perkin-Elmer 225infrared spectrophotometer using a KBr pellet, 15 mg sample to 400 mg KBr.The spectrum was normalized to that of a similar 400 mg KBr pellet, preparedsimultaneously, in the reference beam.

Benitoite, neptunite, and joaquinite were analyzed from five samples from theCalifornia Institute of Technology collection. Benitoite was chipped from sa"mpleD4, neptunite from samples D2636 and D6794, and joaquinite from samplesD2200, D7574, and D4, Joaquinite was taken from sample D6794 by dissolving theenclosing natrolite in concentrated HCl. In order to obtain chemical data con-sistent with other properties, densities, optical properties, and cell para,meters

were determined on grains of the same samples that were used for the chemicalanalyses.

Rpsur,rs AND DrscussloN

Benitoite

The chemical composition of benitoite is delineated in Table 1.Benitoite may be blue or white to colorless. Both colors occur withinsingle crystals and are separated by distinct contacts. Louderbackand Blasdale (1909) suggested a systematic relationship between colorchange and crystallographic axes, but no systematic and consistentrelationship was, apparent in the crystals available to us.

The SiOz, BaO, and TiOz contents of blue benitoite (analysis 1)are virtually iddntical to those of white benitoite (analysis 2) fromthe same crystal and closely approach the stoichiometric compositionof BaTiSisOn (SiOz, 43.59 percent; TiO2, 19.32 percent; BaO, 37.09percent). Our results are similar to the original analyses for blue andwhite benitoite by Louderback and Blasdale (1909) (Table 1, analy-

88 LAIRD AND ALBEE

ses 3 and 4) but do not support their reported differences in TiOz andBaO. Standard deviations for the averages of the individual analysesare less than half a percent, indicating both the blue and the whitebenitoite are homogeneous (see also Figure 1).

Electron beam scans and analytical profiles across color contactsindicate no detectable variation in compostion for Ba, Ti, and Si.Emission spectrographic analyses were made to see if differences ex-isted below the sensitivity of the electron microprobe (see Table 1).Although they show only small differences in the trace element con-tents of the blue and white benitoite, the difference in iron contentmay be significant.

Table 1. CheEl€l AMlyses of Benitoi te, San Benito Couty, Cal i fomla

1 2M roprobe Microproue

-F-r.s."

la l r t . % Wt. % I Fornula W t . % w t . %

l"o,lAlzo:

lrio2t -fFeoc

Mso

CaO

BaO

Na2O

Nb

S r

V

Z r

IotaI

43,L0

1 9 . 5 1

d

d

37.23

0, r.3

d

d

9 9 . 9 7

2 . 9 8

1 . 0 1

1 , 0 1

0 . 0 2

0 . 2

0 . 0 5

Trace

0 , I

Major

0 . 0 0 3 5

0 . 0 0 1 5

0 . 0 1 2

; l

0. 04

0 . 0 1

Trace

0 . 0 0 6

Maj or

o . t 4

0,0025

0 . 0 0 1

0 . 0 8

43.68

20.09

1 0 0 . 1 0

43.67

1 9 . 5 0

37.0i,

700. !2

a Emission spectroscopy analysis by E. Binghan. Looked for but noE found: Ag, As, Au, B,B e , B i , C d , C e , C o , C r , C u , c a , c d , I a , M o , M n , N d , N i , p b , ? r , S b , S c , S n , - i a , i t r , i f , 'W, Y, Yb, and Zn.

b Cat ior fomula proport ion aletemined by DorMliz ing af,hydrous oxygen to 9.

e Total Fe as FeO,

d Less than the sensit iv i ty of a @velength scan (- 0.10 Dt. percst) .

1. P1, blue color, Microprobe analysis is the average of 6 points.2. D4, wtr iLe color. From the see crystal as 1t; . l l icroprobu analyEis La the average of

6 p o i n t s .3, Blue color. Iouderback and Blasdale (1909).4 , W h i t e c o l o r , L o u d e r b a c k a n d B l a s d a l e ( 1 9 0 9 ) .

BENITO,ITE, NEPTUN ITE, AN D JOAQU IN ITE 89

BoFormulo

propor t ion

TiFormulo

propor t ion

ooooSiFormulo

propor t ion *w,?,?e-Frc. 1. Scatter diagra,m for the cation formula proportions of benitoite samples

from San Benito County, California. Each point represents a complete microprobeanalysis. Formula proportions are determined by normalizing the analyses toanhydrous oxygen - 9.

Benitoite has very intense cathodoluminescence under the electronbeam and is commonly used by microprobe operators to check beamsize and shape. Many of these operators have found that benitoitecontains no detectable elements other than Ba, Ti, Si, and O and haveused it is a microprobe standard assuming stoichiometry.

The density of benitoite is 3.64(2) gm/cc; it is uniaxial positivewith <o : 1.756(1) and colorless and e : 1.800(1) and blue. Louder-back and Blasdale (1909) reported p:3.M-3.67 gm/cc, ', = 7.757,and e : 1.804.

From single-crystal work on benitoite from San Benito County,Fischer _(1969) reported that benitoite is hexagonal with the spacegroup P6c2; o : 6.6410 (7),E and^c :9.7597 (10) A.AGuinierX-raypowder pattern gives o : 6.63 (1) A and c : 9.73 (1) A. The d-spacingsare in good agreement with the ASTM index.

Neptunite

Chemical data for neptunite are summarized in Table 2. The mi-croprobe analyses are similar to those reported in the literature exceptthat Cannillo et al. (1966) reporbed 1.63 percent Li2O. It is impossibleto analyze for lithium with the microprobe because of its low atomic

90 LAIRD AND ALBEE

number; however, our analytical totals are one percent low, and quali-tative emission spectroscopic data indicate 1-2 percent LizO. Theformula LiNazK(Fe,Mg,Mn)2 (TiO)2 (SisOzz) was suggested for nep-tunite by Cannillo et aL. (196il. In the absence of Al, the formulaproportions can be calculated from the analyses by normalizing Si to8. Our analyses normalized in this manner give Li6.6Na2.1K6.e (Fe,Mg,Mn)2.s Ti2.eSisO2a. This formula is also obtained from the analy-ses by normalizing total cation charges (exclusive of Li) Lo 47 ortotal cations (exclusive of Li) to 15. The formula proportion of Li

lable 2, chemicalAnalyses of Neptunite, san Benito county, cal i fornia

2 3 4

Microprobe E . S . Microprobe

N t . 7 " i , I E . 7 , I , l E . 7 .wr. 7" Formula wr. % W E . 7 , Fomula !|lE. 7"

s i02

A 1 .0-

T i02

Ie0

MnO

Mgo

Ca0

Ba0

N a 2 O

Kro

r,i2 oAg

C r

C o

Nb

N i

l _l "I

Il ro ta l

', ',, iu . t + |t 1 . 7 1 c l

1 , .40

L , 1 3

d

o , 2 5

7 . 3 8

4 . 8 4

d

d

d

d

d

d

d

9 9 . 0 1

8 . 0 0

1 . 9 8

r . 4 5

0 . 1 8

0 . 3 8

0 . 0 1

2 , 1 2

0 . 9 2

0 . 5 9 -

0 . 0 4

8 c

1 . 5

r . 2

0 . o z

0 , 0 6

1-2

0 . 0 0 0 3

0.0023

o. o tg

0 . 0 1

0 . 0 0 2 8

0 . 0 0 5

0 . 0 0 6 5

54,09

d

1 7 . 9 9

to . 75"

1 . 6 5

2 . 1 6

0. 1,8

4 . 7 6

d

d.

d

iI

d

d

9 9 . O 9

8 . 0 0

2 . O O

o . 2 L

0.48

0 . 0 1

0 . 9 0

u , *

0 . 1

toc1 ?

r , 20 . 0 4

0 . 0 6

1-2

0.0002

0.0048

o. o t48

: 0 . 0 1

0.0038

0 . 0 0 6

0.00s4

1 7 . 1 8

TL.23

1 . 7 8

1 . 8 2

0 , 2 5

9 . t 4

5 . 3 9

00.23

52,87

1 7 . 8 3

1 1 . 6 9

0 . 8 5

7 , 4 4

1 . 5 6

9 . 5 6

5 . 0 8

I 0 0 . 8 8

52.29

L 7 . 3 5

I I .92

2 . 2 7

1 . 5 5

o , 6 2

6 . 8 1

1 . 6 3

100. 02

a Enission spectroscopy anal is is by E. Singhan. T,ooked for but not found: As, Au' B, Be, Bi,

C d , C u , G a , L a , M o , P b , P t , S b , S c , S n , S r , T a , I h , T 1 ' W ' Y , Y b , a n d Z n .o

c.aio. formula proporEion deEermine.al by norMliz ing Si to 8.

c Total I 'e as Feo.

d L"" th.n the senslt iv i ty of a {avelength scan (- 0. I0 wt. petcent).

e Cannot t leternine with Lhe microprobe.

f c. l "r l " t"d by assigning the def ic iency in the analyt ical total f rom 1.00 Eo Li2o.

g -- l o s s l D I e l n E e r t e r e n c e .

L. D2636. Microprobe analysis is the average of 12 Points on 7 grains.

2. D6794. Microplobe analysis is the average of 7 PJints on 4 grains.

3. Louderback and Blasdal-e (1909).4 . B r a d l e y ( 1 9 0 9 ) .5 . c a n n i l 1 o , e t a I . ( 1 9 6 5 ) .

BENITO,ITE, NEPTUNITE, AND JOAQUINITE 9I

in the analyses is calculated by assigning the defieiency of the ana-lytical total from 100 percent to LizO; if the totals are about halfa percent high, the formula proportion of Li would be 1.

In the reported analyses (Table 2), iron is given as FeO. Bancroft,Burns, and Maddock (1967) have determined from Miissbauer spec-troscopy on neptunite from California that at least 95 percent of theiron is present as Fe2*, corroborating the formula given above.

Figure 2 is a scatter diagram of the cation formula proportions ofneptunite for each point analyzed with the microprobe. Althoughthere is a scatter in the Fe, Mg, and Mn contents of single points,the total Fe*Mg*Mn proportion is constant. Electron beam scanphotographs also show that neptunite is quite homogeneous.

The density and optical properbies of neptunite agree well withthose reported in the literature for this and other occurrences. Thedensity is 3.20 (2) gm/cc compared with the values 3.19-3.23 gn/cclisted in Murdoch and Webb (1948). Optical properties of neptunitedetermined on sample D6794 are as follows: 27 (*) -40o ) optic plane

l l tOtOl ; U = b; Z A c = 17"; a = 1.692(1), pale yel low; F : 1.702(1), yellow orange; and 7 - 1.734(2), red orange to red brown.

NoFormulo

proport ionK

Formuloproport ion

TiFormulo

proport ion

Frc. 2. Scatter diagram for the cation formula proportions of neptunite samplesfrom San Benito County, California. Each point represents a iomplete micro-probe analysis, and analyses for each grain are grouped on the figure. Formulaproportions are determined by normalizing the analyses to Si = 8.

.1. Jr..t i

2.5 f , iz.o L. olo o olo 1D,,a

1.0;-o i7. . f .o.sL t t

a ala(,

,/

i t, t'.t7 '. t t

LAIRD AND ALBEE

Those reporbed by Larsen and Berman (1934) on California (?)neptunite are: 2V - 49o, Z t c = 16o, o = 1.690, F = 1.699, and

7 = 1.736.Cannillo et aI. (1966) proposed that neptunite is monoclinic, space

group Cc, but reported no new cell parameters. Guinier powder dataon neptunite from sample D6794 give cell dimensions ,and d-spacingssimilar to those reported by Berry (1963). Our parameters are d :

16.43(3)A, b = 12.49(2)A, c: 10.00(2)A, and B : 115.4(1)o; Berry(1963) reported a:16.7 A, b,: 12.4L, c: 10.04, and F = 115"44from single-crystal precession photographs.

Joaquinite

The density of the joaquinite from sample D6794 is 3.98(5)' gm/cc,compared with 3.89 gm/cc given by Palache and Foshag (1932). Ourdata indicate that joaquinite is biaxial positive'with a range in 2Vfrom -30-55o. The indices are: d = 1.753(1), F = 1.767 (1), andy : 7.822(2); a and B are colorless and y is pale yellow' Palache andFoshag (1932) reported 2V : 50" , a = 1.748, 13 : 1.767, and 7 r: 1.323.

The chemical composition of joaquinite is complex; electron micro-probe and emission spectroscopy data are presented in Table 3 withother chemical data on joaquinite taken from the literature. Thirty-eight points on 15 grains of 4 samples were analyzed with the micro-probe. Many grains are inhomogeneous, but joaquinite from sampleD6794 is quite homogeneous; all of our data on the density and opticaland structural properties of joaquinite were determined on grains ofthis sample.

The infrared spectrum of joaquinite is presented in Figure 3. Thesharp bands at -3500 and 3560 cm-l indicate that the joaquinite stmc-ture includes significant quantities of crystallographically-orderedhydroxyl groups. The band at -1610 cm-' (corresponding to an IIOHbending motion) and the broad absorption feature centered at -3400cm-1 (the OH stretch) are probably due to water adsorbed duringsample handling procedures. The C-H bands around 2900 cm-1 are dueto organic impurities introduced during sample handling. In this spec-trum the lower energy region is dominated by the strong Si-O absorp-tion near 1000 cm-1. However, infrared spectroscopic data on silicateminerals is insufficient to assign this Si-O absorption pattern to aparticular type of silicate.

We are unable to determine directly the amount of HzO that ispresent in joaquinite because of the small amount of material availableand the small size of the crystals. The reported HrO content is trasedon the difference between the analyzed oxide total and 100 percent and

BENITOITE, NEPTUNITE, AND JOAQUINITE 93

may not be very accurate. Qualitative emission spectrographic analysesindicate that Li is present but in an amount less than 1 percent Li2O.There may be some FezOg, although Semenov et aL. (1967) reportedonly 0.39 percent FezOa in joaquinite from southern Greenland. Thepresence of these components will affect the value reporbed as HzO.

The number of rare earth elements present in joaquinite makes itdifficult to do accurate microprobe analyses because their peaks areclose together. However, total rare earth oxide concentration deter-mined by electron microprobe and by emission spectroscopic analysesdiffers by only 0.5 weight percent for sample D6794 and by 2 weightpercent for sample D2200. Total rare earth oxide concentration in the38 microprobe analyses of joaquinite ranges from 15-20 weight per-cent; this is zero to 8.5 weight percent higher than that reported bySemenov et at. (1967) for two samples of joaquinite from California(see Table 3).

Joaquinite is listed in the literature as orthorhombic. However,J. F. Rowland has found a monoclinic crystal from the Californialocality (E. H. Nickel, written communication, 1970). Single-crystal

f C€rcukLd !t rs3lBning the deflclency tn the lM1yEica1 br.l f lm 1.oo Eo tszo.

1. D6794, SaD 3dt!o CounLy, Californla. Mtcioplobe 6!€Lysis i6 the awerage of 6 Poinis on 5 sr.h8.

2, D22OO, Sad latto Counry, caltfodt., XicroProbe .na1yrl. 1s lhe gveraBe oI 2t Pointt on 7 gratnr

3 !7574, S.n Benlto CounEy, Californla Averdee of 6 !o1nt5 on 2 ar6tns4 n4, S.n lentlo coutrtyr catlfornia Average of 5 potnre otr I grain.

5 Sao lenlto counly, C.ltfornto ?olache ad loshc (1932). A@rysts rduced to 100 Petcent.6 san 3en1ro cou!!y, caltfoml. serend, d d (1967)

7 s.n Berlro eury, callfornia s@nov, 4 91 (1967)

3 s c r e e n l a d s a e . o v , 4 d ( 1 9 6 7 ) .

T.bLe 3 Ch4tcal Analyres of J@quhll€

*20

LL20

.'"r0,

\zo3

t RE2

F'Gu2

i :il' * l0 . 0 5

0 2 r

3 2 4

1 3 7

0 0 3

o 2 7

r 0 , 6 9

0 z L

o 2 6

3 2 7

L 2 5

D 1 6

,3 12

8l

1 1 3

4 0 5

13

g 2

o 2

1

3 5

1 5

0 5

L 7 5

0 2 r

2 4 2

1 1 7

0 5 3

0 3 3

93 23

3 3 7

2 4 6

0 1 2

0 3 l

2 0 3

3 0 9

0 0 5

0 1 7

3 3 0

o a 2

0 1 3

5 9 2

I 35 l7

1 1 4 1

2t 99

1 3 6

1 3 1

l t 32

0 1 9

0 0 5

2-42

1 1 3

t3 53

93 32

3 , 1 9

0 0 5

0 1 3

3 1 7

4 9 5

o , 1 7

O I L

0 t 3

6 1 9

34 93

1t 90

t2 5A

0 3 9

3 5 4

I a 2

t 1 5 1

o.o2

r 3 7

0 1 7

l7 73

93 t0

32 00

3 . 2 1

3 3 3

o . 1 7

1 3 3

3 . 3 6

0 3 5

Lt,azE

3 0 5

3 . 5

1 3 , 0

3 2

1 5 0

13 I

{ "

5 . 6

0 l

0 . 1

13 A2

9 , 2 0

0 3 9

4 , 7 3

{ o . *

o , 2 2

0 3 8

2 . 3 L

9.40

10 05

2 . 1 5

22.59

t 5 0

0 3 8

0 . 1 6

100.01

1 7 0

3 5 4

o 1 2

1 0 5

4 2 2

r by ! blngnam

. Ewerenqlh scan (- 0 10 Hl percent

s4 LAIRD AND ALBEE

@

N

I

@

oN

o o9 {

aruollsuql

BENITO.ITE, NEPTTJNITE, AND JOAQAINITE 95

X-ray diffraction work was undertaken in order to resolve this prob-lem; the results of this study and the structural data in the literaturefor joaquinite are delineated in Table 4.

Precession and Weissenberg photographs on a crystal of joaquinitefrom sample D6794 show it to be monoclinic with cell dimensionsa* = 10.51(2)L, b* = 9.66(2)A, c* : 11.82(2)A, and F = 109.5(2) ".Another crystal of joaquinite from sarnple D6794 gives an apparentlyorthorhombic diffraction pattern corresponding to cell dimensions o, =10.48(2)4, b" = 9.66(2)A, and co = 22.26(2)A. Because the o and baxes are closely similar for the monoclinic and orthorhombic cells andbecause co = 2c* sin B to a close approximation, the possibility mustbe considered that the apparently orthorhombic diffraction pattern isthe result of twinning of monoclinic individuals. As shown in Figure 4,the monoclinic reciprocal lattice points 4n, k,l correspond very closely(if not exactly) with points of the orthorhombic reciprocal lattice; infact, all of lhe 4n, /t, I reflections observed in the orihorhombic patterncan be explained by twinning of the monoclinic lattice on (001). How-ever, reflections with h + kn do not coincide in the two reciprocal lat-tices; hence, the complete orthorhombic pattern cannot be explainedby twinning of the known monoclinic cell. We conclude that joaquinite

occurs in two crystalline modifications that are very closely relatedbut are quite distinct.

As shown in Figure 4, the monoclinic cell is related to the orthorhombicone by the conditions oo ry a^, bo N b^, arrd c^* N2 co* N 8f 3 a.* cot B*.The last condition follows from the interesting coincidence of the 803reflection in the monoclinic reciprocal lattice and the 800 reflection inthe orthorhombic reciprocal lattice. The value of B calculated from thelast condition and the measured axial lengths is 109.4o, as comparedwith the directly measured 0 : 109.5o.

The orihorhombic cell found here agrees with the parameters re-ported by Palache and Foshag (1932), Bell (1963), and Semenov et aI.(1967). Because of its close relation with the monoclinic cell, we havechosen the orthorhombic axes to correspond with those of the mono-clinic cell. Therefore, in Table 4 Lhe a axis of earlier workers is rede-fined as bo and the b axis as oo.

Systematic extinctions for the monoclinic crystal indicate possiblespace groups C2, Cm, or C2/m; the orthorhombic crystal may belongto space group Cc?m, Ccm?, ot Ccmm.

Diffraction patterns from the monoclinic crystal show additionalweak, somewhat diffuse reflections in positions corresponding to re-flections from an orthorhombic crystal; the orientation relationship is

I'AIRD AND ALBEE96

roi

d

ow

z

6! o

ic ' do q

. iu b 06 Ho o.i -o

r 5 0C r ^

' e oo . i

N O . Oo o Bo t r Ei o - d

u . d G! q ( d

b0 .H( $ l d c

- d 3 0 0o v . a . Ho a qt ! € o o

. c o o€ ^ d o od o d { J od a . B ! !

6 ^ O A F l lo i N c 4

s v ao 6 . r r( d d i f q d d

H i . i . id o . . t r t r

t u F q ^ ' _ b ! !o d l o oc r d l H q

! t 6 . . d . H( ! ( d d l l d i d

. i . J v oJ l .J ( ! d

! ! H ^ HO O n > O . ^q H O O H > | X. i l . d c q t r . r u !

d . t E , o 5 5( J ( J . O O O O! @ Q ( J

. - OX > r ! X o oU U ( l . P P !t r t r ! , O t r . d . d

o o d ( ! o o c )( ) O F - l i O c q F Q

O O . O o t r t ru u o o u r d $. i . J , ! ! . d @ o

O O F 1 O - ^F q F q . d € : f t +

i p o 6C d 6 J d N rd ( ! a J o s o a@ o @ o @ a Q

st

o .

E sd B

. U qo . H

. d . d p

. J a t g

cr bo(H

o d o

U XH t r ( so q )

H P H

( ! . d

Q O . dl u !

i F ' H( ! t rtr q, .ilJ > r s! d c )q E ' datl .ol c,

. U-J ( ! l .d

Bo o

i o n ), o x q t$ ( ! Ht s b D

C J OO U

u !. r o

oo

H . d

i N d ) < - h \ o N

N

I

"€ lstst:i

N N NN

d a N6 0 @ n

o 6 i 6d J o

i

o

I

H

dutl

EIst?tj lo l o l t {u t ( ) t o

e<

N N N

@ @ @

{ 9 " :o 6 Ni N

r

g

o

o

\ lN I

N O O ON N € T

o o o

o

g

H

o

E I ! JE l o oo t o c

,<o oi n H

6 0 n( n o \ t

1 : 1O 6 Ni N

o

!

p!r

AJ

E l u ;E l @ oOl O tr(JI E A

o6 0

9 16 N

N

n

od

N

o.t

!

U|r

o

dE l u !E l o oo l o H(JI E O

h a ' <9 i @. : f Q d

N

N

o

o

|{opk

6 i ' 4< f , a $

NN

oO

d Eu oq D

> o ( d 0

@ c )

o l, a lG I

97BEN ITOITE, NEPTUNITE, AND JOAQUINITE

o 6o t r o6o++ co*, co

+of;, oo

' Monoc l i n i c rec ip roco l l o t t i ce po in t s o t k = O+ Monocl in ic rec iprocol lo t t ice points ot f = 1o Or lhorhombic rec iprocol lo t t ice poin ls ot , ( = Oo Orthorhombic rec iprocol lo t t ice points ot f = 1

Frc. 4. Relationship between the orthorhombic (o) and monoclinic (m) re-ciprocal lattices of joaquinite. Recip.rocal axes are shown by the symbol *,

the same as that shown in Figure 4. Similarly, the orthorhombic crystalshows a few very faint, but sharp reflections that correspond to amonoclinic crystal and its twin across (001). Therefore, althoughcrystals of joaquinite appear to be simple on a macroscopic scale, theyare composites of orthorhombic and monoclinic components at theX-ray diffraction level.

Electron beam scans were done with the microprobe on sampleD6794 of joaquinite in order to determine if a difference in composi-tion corresponding to the monoclinic-orthorhombic component bound-aries could be seen. There is no compositional variation in Ba, Sr, Ce,Na, Fe, Mg, or Ti at a magnification of 5,000 and a spot size of -1

p,m. The two phases either must have identical compositions, and hencebe polymorphs, or else the structural domains must be too fine to beresolved with the electron beam scans.

Using a density of 3.98 gm,/cc, we computed the cation formulaproportions for our monoclinic and orthorhombic cells. With the mono-clinic cell, the number of Si cations for all analyzed points is 16.0(4) ;

o o o

T

o o E

98 LAIRD AND ALBEE

the orthorhombic parameters give 32.0(7) for the number of Si cationsfor all analyzed points. Figure 5 shows the cation formula proportionsfor each analyzed point of joaquinite calculated by normalizing theformula proportions to Si : 32. Patt, of the scatter in the joaquiniteanalyses is due to measurement error, but much of this scatter repre-sents inhomogeneity. Although joaquinite is more complicated thanbenitoite or neptunite, all three minerals were analyzed in the samemicroprobe runs; and the precision for joaquinite should not be muchless than that for neptunite and benitoite (compare Figures 1 and 2with 5).

The complete formula of joaquinite cannot be directly determinedwithout knowing the Fe3*, Li, and (OH) contents. However, a numberof lines of evidence suggest that joaquinite is basically a sorosilicate,.a2si2o7.

Joaquinite from sample D6794 is homogeneous and representativeof all the analyses. Table 5 gives the formula proportions with Sinormalized to 32 for this sample. Several features are present in theseanalyses:

a) The total number of cations other than Si (31.18) is nearly equalto that of Si. Figure 5 suggests that the variation in the totalFe f Mg * Mn may be partially matched by the range in Ticontent but appears unrelated to changes in other analyzedelements. Li has an ionic radius similar to Fe, Mg, and Mn andmay account for some of the Fe * Mg * Mn variation; a rangein LirO content from zero to 0.55 weight percent (0 to 2 formulaproportions of Li) will match the Fe * MS f Mn variation.With the presence of Li, then, the total number of cations, ex-cluding Si, equals the total number of Si cations.

b) The total positive charge ranges from 209 to 214 for all theanalyzed points, with an average value of 210. In a simple anhy-drous silicate this charge requirement is most closely met by asorosilicate, basically 16(Cr3*Siroz); /r*SirOr, Bz'*SizOo, andD2**SirO8 do not fit.

c) The small variation in the formula proportion of Ba about thevalue of 8 in Figure 5 suggests that Ba occupies a distinct positionin the joaquinite structure. The formula proportion of Ti variesmore than that of Ba, but it too approaches 8, suggesting adistinct position for Ti. The cations in Table 5 are ordered by theiratomic radii. Ba is probably in eight-fold coordination and Ti insix-fold coordination. The rare earth elements plus Sr total abouteight and may occupy an eight-fold coordinated position. Theremaining cations, except Si and K, total about eight and may

BENITO'ITE, N EPTANITE, AND JOAQUINITE 99

a ' a

BoFormulo

proport ion

5 rFormulo

proport ion

:RE, ThFormulo

proportion

N oFormulo

proporl ion

5.O

4 5

IFe,Mg,Mn 4'o

Formulo 3.5proporlion a.o

2 .5

2 0

T iFormulo

proporlion

9.O

8.5

8.o

7.5<----1+/r+

D6794

t'tIror7 .5L

'l

. .l'tJry' t . t '

/ , t o

t,

. l/

/,'/. /

. J t .

aa. l o

. /,

| . 'h .

o a/ at;L

t.7

a

o

a/a a

4.orI

"- l" " 12.5L

a

a L7

4r

'tbt,

,

, 'ra,

a ,

a

- / a- a /

a

o o a

a

/, /, /, /.

DZZOO -2*- .-l---t-

l t lvtiFrc. 5. Scatter diaeram for the cation

D7574 D4

formula proportions of joaquinite

samples from San Benito County, California. Each point represents a completemicroprobe analysis, and analyses for each grain are grouped on the figure.Formula proportions are determined by normalizing the analyses to Si = 32.

occur in six-fold coordination. Hence, a possible generalized

distribution is

Ba-rottt (Sr, E-E)-rvr" (Cu, Na, Fe, Li, Mg)4vr Ti-rtr Sirrtv

with possible coordination indicated by superscripts.d) Since the actual mineral contains univalent, divalent, trivalent,

100 LAIRD AND ALBEE

Table 5. Joaqulnlte formula proportlons for the averaged analysls of D6794and orobable end nembers.

Ionic TormulaRadii ProDortioB

B^*2 r.34 s.6sK+ 1.33 o.o4s.*2 L.Lz r..70

Total Formula

sr '14 32.00

E catlons = 64.00

Ecatlon charges = 210,7

(@) = 11 .5

O = 99.6

+?L a -

c"*3l?

Nd+3

Th ' -

s**3

cd+3

oy+3

Y+3

t . r 4 0 . 7 2

1 . 0 7 3 . 5 8

1 . 0 6 0 . 4 2

1 .04 1 .05

L . 0 2 0 . 0 5

1 . 0 0 0 . 2 2

0 . 9 7 0 . 0 8

o .92 0 "06

0 .92 0;34

6 . 53

c"*2+

N e '_ + 2-be

+!t

+lLT i - '

uJ2+fL

s i

0 . 9 9 0 . 2 0

0 . 9 7 3 . 3 2

0 . 7 4 3 . 1 3

0 .58 (0 .82 )a

0 . 6 8 8 . 1 4

0 . 6 6 0 . 0 7

0 .42 32 .00

a calculated by asslgning the difference in total cationg (except Li) fron64 .00 t o L i .

Proportlons of possible formula end nembers

^+

4.18 [Ac(sr2or) <orlr] +.?s2.29 lBc(sL2o6)(oH)l8.20 IBD(st2o7)]1.33 [82(si2os) (orr)z]

d2 c+3 D+4 si{4 o (oH)4 .L8 8 .36 20 .90 8 .35

2 .29 2 .29 4 .58 L3 .74 2 .29

8.20 8.20 16.40 57. ,40

2 . 6 6 2 . 6 6 6 . 6 5 2 . 6 6

4 . 1 8 1 3 . 1 5 6 . 4 7 8 . 2 0 3 2 . 0 0 9 8 . 6 9 1 3 . 3 1

quadrivalent, and (OH) ions, there must be a number of coupledsubstitutions which can be expressed as theoretical end membets.A generalized hydrous sorosilicate may be written as

(A*, B'*, C'*, Dn*1ooo'tse-0-' SieOr-,(OH)".

BENITOITE, NEPTUNITE, AND IOAQUINITE

Possible substitutions are:

c,3*(si,or) .,4.*Dn*(si,or)(oH) /*C'*(Sirou)(OH),

B'*Dn*(Si,Or) B'*Cu*(SirO.)(OH) /*B'*(si,oJ(oH).

Br'*(Sirou)(OH), /,r*(siror)(oH)o

The analysis may then be expressed (see Table 5)

4.1S [.4.C(Si,OJ(OIr),] + 2.2s [BC(Si,O6)(OID] + 8.20 [BD(Si,O')]

+ 1.33 [,B,(Si,oE)(oH)']

or

[(Ko.oeNaa. grl,io. rr) r-.. r, (Bar.ouSrt.roCas. 2qFe3. 1sM8o.oz) >-rs. rE

(.8,E)u. n?(Tho.ouTir. rn) r-r.ro]"=r,. ooSi, 'Oe8't(OH) 13's'

This solution is not unique; other combinations of end members

are possible.Suppori for the suggestion that joaquinite is a sorosilicate is pro-

vided by the following observation: The O and OH formula propor-

tions determined by reducing the chemical analysis to a number of

possible coupled substitutions are close to those determined by assum-

ing that the total number of cations is 64.00 and by assigning the

deficiency in the analytic total from 100 percent to HzO (98'7 us' 99'6

and 13.3 as. 11.5, respectivelY).We suggest, then, that a general formula for joaquinite is

16 {[(.,4.*, B'* , C"* , D4+)rvrrr(a+ , B'* , Dn*)rorlcharce-6-e sirrvo"-,(oH)']

and that the basic unit contains two eight-fold coordinated and two

six-fold coordinated positions. The formula for the california occur-

rence of joaquinite is then

[Bar.rott t (Srr.rEEo.uTho.r)r-r. .ott t (Cao.rNar.aFee.r l , io.aMgo' ') ,- ' 'utt

' Tir. r t t ]r=rr.o Siur.otvO,, ' t(OH) tt ' t '

(1e70)Further single-crystal x-ray diffraction work is necessary, however,

to determine an exact structural formula for joaquinite'

t02 LAIND AND ALBEE

AcKNoWT,EDGMENTS

sxpported by NSF grant (GA-12862). The microprobe laboratory has beendeveloped with the support of the National science Foundation, tlie Jet propul-sion Laboratory, and the Union pacific Foundation.

RnrpnnNcnsAr,non, A. L., arrn L. R"ry (1g70) Correction factors for electron probe micro-

analysis of silicates, oxides, carbonates, phosphates, and sulfates. AnaI. chem.42,740V1414.

Ber'rcnorr, G. M., R. G. BunNs, euo A. G. Maooocr (lg62) oxidation state of ironin neptunite from Miissbauer spectroscopy. acta crystallogr. 22, gB4-gBE.

Bnr-r,, K. w. (1963) X-ray diffraction examination of a canadian occurrance ofjoaquinite. Dept. Mines Tech,. suru., Geol. suru, can. Mineral. gcd,. Internat.8ep. MS-63-82.

BnNco, A. E., ewo A. L. Ar,sen (1g68) Empiricar correction factors for the electronmicroanalysis of silicates and oxides. J. Geol. 76,8g2408.

Bnnnv, L. G. (1963) Neptunite: Unit cell and X-ray powder data. Con. Mineral.7,679481.

Bn.or,rr, W. M. (1909) On the analysis of the mineral neptunite from San BenitoCounty, California. Amer, J. Scz'. 28. llFl6.

CewNrr,r,o, E.,F.Mmzr, awn G. Rossr (1g66) The crystal structure of neptunite.A ct a'C r y s t allo g r. 21, 2N-208.

Frscron, K. (1969) verfeinerung der Kristarstruktur von Benitoit, BaTi(sLo,).Z. Rr;^taUo gr. t2g, n2-248.

Iletwnrcu, E. W., ervn S. IL QuoN (1963) Neptunite from Seal Lake, Labrador.Can. Mineral. j,6E0-{.E4.

Per,ecuo, C., er.m W. F. Fossec (1932) The chemical nature of joaqrinite. Amer.Mineral.lz, 808_812.

Snunnov, E. I., Y. I. Burrr*, y. A. Beusnov, AND If. SlnrNsnx (lg62) Rareearths in minerals of the joaquinite group. Amer. Mineral s2, lz62-lz69.

Suor,rN, Y. f., exo Y. F, Snrpnr,pv (lg70) The crystal structures of the rare earthpyrosilicates. A cta C r y s tallo g r. B.26, 4g44gZ.

ZecrenrasiN, W. E. (1930) The crystal structure of benitoite, BaTiSiaO". Z.K ris t allo g r. 7 4, 139-146.

Manuscri,pt recei,ued, Apri,I I, 19TI; accepted, Jor publi,cation, August pS, 1921,