American Mineralogist, Volume 61, pages 956-962, 1976
The crystalstructureof eakerite,a calcium-tinsilicate
ANrHoNyA. Kossre.rorrl
Departmentof Chemistry,CaliforniaInstituteof Technology
Pasadena.
California 91109
nNoPrreRB. LnavrNs
Departmentof Geology,Uniuersityof Delaware
Newark,Delaware 197I I
Abstract
Eakerite, CazSnAlzSiuOr8(OH)r'2HrO,
contains crankshaft-likechains, similar to those in
feldspars,of composition AISLOT(OH), which are cross-linkedto form a kinked sheet.Al is
ordered, and the OH is bonded to it. Ca and Sn ions lie in sheetsbetween the kinked
aluminosilicatenetwork. The Ca ions are coordinated by 40, 2OH, and 2HrO in a square
antiprism; theseare edge-linkedinto chains which run acrossthe aluminosilicatechains and
which are cross-linkedby Sn octahedra.The OH and HrO are bonded to Ca and, by hydrogen
bonds, to other O; this strong bonding preventedtheir being distinguishedby thermogravimetric analysis.
Introduction
Eakerite (Leavenset al., 1970)is a rare tin silicate
found in hydrothermal fissuresin spodumene-bearing
pegmatite at King's Mountain, North Carolina. The
formula was given as CazAlrSnSi.Ol.(OH)6on the
basis of wet-chemical analysis of an I I mg sample
and from thermogravimetricanalysis,which showed
that all water is tightly bound.
The structural analysis described in this paper
shows that the chemical analysis is correct but that
the formula should be written CarSnAlrSi.O,r(OH),
.2H"O.
Experimental
Eakerite is monoclinic and crystallizes in space
grotp P2r/a; the cell parameters in Angstroms are
:
a:
t 5 . 8 9 2 ( 7 )b; : 7 . 7 2 1 ( 3 ) ; c : 7 . a 3 8 Q ) ; 0
1 0 1 . 3 4 ' ( 3 ) ;a n d t h e v o l u m e : 8 9 1 . 3 5A ' . T h e c a l c u lated density is 2.65 with Z : 2 molecules/unit cell
(Leavens et al., l97O).
Three-dimensionalintensitydata werecollectedusing a four-circle Picker card-automated diffractometer with filtered (0.002' Zr foll) MoKa radiation. The
crystalwas ground to a 0.25 mm diametersphereand
mounted with the b* axis alons the direction of the
I Presentaddress:Department ofBiology, Brookhaven National
Laboratory, Upton, New York
11973.
instrument. The linear absorption coefficient
(pMoKa) is 15.5 cm ', giving a transmissionfactor
of 0.69 for a 0.25 mm sphere.The moving crystal,
moving counter-measurementtechnique (0-N coupling) was used. Integratedintensitieswere measured
over a scan range taken 0.9o on both sides of the
Ka,-Ka, splitting at a rate of 2" /min. Individual
backgroundintensitieswere determinedby 30-second
stationary background counts taken on both sidesof
the peak. Three standard reflectionswere measured
every 60 reflectionsto monitor crystal alignment and
instrument stability.
ln all, l79l independentreflectionswere measured,
of which 1687 were consideredstatistically observable using the criterion F, 2 3o(F); o(F) was calculated from counting statistics and an instrumental
instability constant of 2 percent. The raw intensity
data of each reflection were corrected for background, Lorentz, and polarization effects. Absorption correctionswere not necessary,due to the spherical shape of the crystal and its low linear absorption
coefficient. Corrections for the effects of secondary
extinction and anomalous dispersionwere calculated
to be small and were ignored.
Solution of the structure
The crystal structure of eakerite was determined
using heavy-atom and vector superposition tech-
956
CRYSTAL STRUCTURE OF EAKERITE
niques. The two Sn atoms in the unit cell are necessarily located in a special position at the center of
symmet(y, chosen as 0, 0, 0 and Vz.,V2,0. The vector
peaks between Sn and the other atoms in the structure would, therefore, appear in two sets: (l) the
actual locations where the atoms would be found in
the real cell, and (2) a set displacedby Vz,V2,0.
A three-dimensional sharpened Patterson, restricting sindltr to 0.35,was calculated(all crystallographic
programs used are from the XRAY '70 computing
package of James Stewart). The largest Patterson
peaks were situated along the y axis on sectionshaving l/6 unit cell separations.A survey of possible
Sn-X (Ca, Si, Al) vectors was made, and the vector
coordinates were used to calculate the Patterson positions of the X-X vectors and their transformations.
Unfortunately, the only set of prominent vectors
which this procedure could clearly identify were the
Ca-Ca vectors. Other vectors due to X-X [Si-Si (Al,
O)l interactions could not be identified unambiguously. The major problem faced in the early stages
of refinement was that using phasing information
from the Sn atoms located in the centered special
position introduced a falsemirror plane in all Fourier
maps perpendicularto they axis through the origin.
The Ca atom was placed in the phasing model
located in coordinate positions calculated from the
Patterson map. In theory, addition of atoms in the
phasing model located in general positions, and of
consistent orientation, will break the image ambiguity and lead to the correct structure. The phasing
contribution due to the Ca, however, was not large
enough to discriminateclearlybetweenimages.Three
of the four possible(Si, Al) atoms were chosenfrom
consistent Patterson and Fourier information and
were added to the phasing model. Even at this stage,
the image problem was not completely resolved and
led to doubt that a consistentset of atom positions
had been chosen.Further refinementon this particular model using the heavy-atom approach was therefore halted.
The position of the Sn atom at the origin and its
exaggerated influence on the initial phasing models
made the iterativeprocessof improving the phasesby
stepwiseaddition of correct atoms to the structure
less powerful than is normally observed in heavyatom problems of this type. On the other hand, this
specific symmetry of the Sn atom makes the structure
a prime candidate for solution by vector superposition. The vector superpositionwas accomplished
by overlaying two identical three-dimensional Patterson maps translated(V2,V2,0) from each other. Vec-
957
tor peak overlaps clearly identified the position of all
atoms in the structure, with the exception of two
oxygens. The orientations of all the atoms which had
previously been placed in the phasing model were
also shown to be correct. A cycle of least-squares
refinement was run varying positional parameters,
keeping temperature factors constant, giving a residualofR=0.32.
A difference map clearly showed the two remaining
oxygen positions. With all the non-hydrogen atoms
located, the R factor was still 0.30. Two cycles of
least-squaresrefinement varying the positional parameters and the isotropic temperature factors of all
the atoms led to a precipitousdrop in R to 0.061and
wR to 0.072. These cycles were run using the weighting scheme l/o(F)2, omitting reflections for which
both F, < 3o(F) and F" ( F'.
A cycle of least squareswas run, changing from
isotropic to anisotropictemperaturefactors,resulting
in a fit with residualsR = 0.044and wR : 0.059.This
refinement showed that the atoms had very little anisotropic motion, as is to be expected for this type of
compound.
It was originally thought that there might be Si-Al
disorder.This disorderwould be accompaniedby the
presence of average Si-O, Al-O bond distances
around the disordered sites.These were not found. Al
is tetrahedrallycoordinated to four oxygens(O3, 05,
07, O9), having mean Al-O bond distancesof 1.75
A. ntt the Si on the other hand have Si-O bond
lengthsof 1.60-1.64A, as expected.
A difference map was made, and the three hydrogen atom locations were found. One of the hydrogens
was bonded to oxygen 09, forming a hydroxyl, while
the other two were bonded to 08, giving a water
molecule.This configurationis consistentwith charge
considerations,sinceoxygen atoms bonded to silicon
are not expected to have attached hydrogen atoms,
and oxygens 8 and 9 are the only two oxygens in the
structure which could meet this criterion.
The theoretical hydrogen positions were calculated
from the bonding geometry and placed 1.0 A from
the oxygens 08 and 09. These hydrogen positions
were extremely close to those observed in the difference map. A final cycle of refinement,varying positional and anisotropic temperature factor parameters
but holding hydrogen parametersinvariant, gave residuals of R = 0.042 and wR : 0.055. A difference
map showed no identified peaks having over one
electron per A'.
Positional and anisotropic thermal parametersof
the atoms, with their standarddeviations,are given in
A. A KOSSIAKOFFAND P. B. LEAVENS
T l s I - r l . A t o m i c c o o r d i n a t e s ( a s f r a c i i o ncseol lf e d g e s ) a n d a n i s o t r o p i c t h e r m a l p a r a m e t el 0
r s- (, )Uoi fj e a k e r i t e
utr
u22
u3r
o.r2(2)
0.23(4)
0 . 1 4( 6 )
0 . 2 7( 2 )
-0 .0s(4)
0 . 0 2( 5 )
0.60(e)
0.63(9)
0 . 7 1( e )
0.62(8)
0.ss(8)
o.s1(8)
-0.01(s)
0.04(5)
0.02(5)
0 . 1 8( s )
0 . 0 5( 5 )
0.11(5)
- 0 . 0 7( 5 )
0 . 0 1( 6 )
0.07(s)
o.96(22)
1.s3(23)
L.LL(22)
L.56(23)
L.40(24)
0.39<22)
r.35<24)
0.96<23)
r.34(22)
2 . 3 7( 2 4 ' )
1.o1(21)
0.99(22)
0.21(15)
0 . 2 s( 1 5 )
-0.11(16)
0.14(15)
0 . 2 5( l s )
o . 3 s( 1 5 )
0.32(15)
o.%(1s)
0.35(16)
- 0 . 0 7 ( r . 6)
-0 . 32(15)
-0.27(r.s)
o.82(22)
0.86(22)
o.92(22)
1.54(2s)
L.43<23)
1.04(23)
r.31(24)
t.79(26>
L.O9<22'
L.29<2L'
1.33(23)
L.66(23)
-0.15(15)
o.o2(1s)
o.ls(15)
0.r.0(17)
0.28(15)
-0.14(14)
0.14(15)
-o.1r(16)
-0 .24(16)
0 . 0 5( 1 6 )
0.01(15)
-0.12(17)
0.26(16)
0.01(16)
0.36(16)
0 . 2 1( 1 5 )
0 . 2 8( l s )
0 . 2 9( 1 6 )
0.11(16)
0 . 0 1( 1 5 )
-0 .08(15)
0.0
(1)
0 .3r.11
o .0220(2)
0.0
0 . 7 8 (3 )
0.0021(2) o.e2(5)
-0.3L72(2) o.s5(9)
sl (1)
sl (2)
sl(3)
0 . 4 s 5 4( 1 )
0.1661(1)
0.4932(L>
0.L697(2)
o .1445(2)
0 . 2 1 8 (12 )
0.7L32(2> 0.43(8)
(2) 0.s4(8)
o.0391
0.334e(2) 0.37(e)
0( 1 )
0(2)
0( 3 )
0(4)
0.0197(2)
0.s051(3)
o.2L43(2)
o .L225(3)
0.1993(5)
-0 .0164(5)
0.3083(s)
0 .0234(5)
-0 .1614(5)
-0.2694(6)
0.2372(s)
0.1411(s)
0(s)
0( 6 )
0( 7 )
o (8) (H2o)
0.3625(2)
0 . 4 4 7 0( 2 )
0,2362(3)
0.3r93(3)
0 .1448(s)
0 . 3 3 4 (05 )
0 .0349(5)
(s)
0.0030
-0 .2316(s)
0.1545(s)
0 . 4 4 8 (56 )
(6)
0.1531
o (9) (oH)
0( 1 0 )
0 (11)
0.2017(3)
0.43e7(3)
0.0922(3)
- 0 . 1 8 1 1 ( s ) 0 . 9 7( 2 2 ' t . 5 7 ( 2 4 > L . 2 2 ( 2 2 )
0 . r031(s)
0 . 4 9 7 (75 '
0.228r(6)
r . 0 2 < 2 4 ) 1 . 9 1 ( 2 s ) o.9L(22)
0 . 2 2 r 6 ( s , o . 4 L 2 4 ( s ) 1 .r 0 ( 2 3 ) L . 5 3 ( 2 4 ) r . 2 4 ( 2 2 )
ln
the
leaet
uz3
o.04(2)
-0.05(9)
o.01(6)
0,0
0 . 3 0 0 s( 1 )
0.2703(1)
devlatlon
ut3
0 . 8 2( 3 )
0.89(3)
1.09(7) 1.00(6)
0.59(10) 0.65(9)
Sn
Ca
41
The standard
Erz
slgniflcant
flgure(s)
appear
ln
parentheses.
Table l, and the observed (F,) and calculated (F")
structure factors are given in Table 2.2
parallel to a. Thesechains, which can be seenparticularfy well in Figure 2, arevery much like thosein the
feldspars,for examplesanidine(Taylor, 1933;Deeret
Description of the structure
al., 1963). As in the feldspars, the chains in eakerite
The eakerite structure is illustrated as a stereopair are cross-linkedto form a seriesof roughly square
in Figure 1; Figures 2 and 3 show the structurepro- rings. Here the resemblanceends. In the feldspars,
jected perpendicularto the b and to the c axis. Bond pairs of chains form a continuous, discrete,kinked
lengthsand angles are given in Tables 3 and 4.
band complexly bonded to four other bands around
Eakerite is composedof irregular, kinked sheetsof it. In eakerite each chain is bonded to the chains on
composition AlSi3Or(OH) which are parallel to the either side, with four successivebonds to alternate
a-b plane. The sheets are bonded together by inter- sides. This alternate linking results in a pattern belayer Sn in 6-fold coordination and Ca in 8-fold tween any two adjacent chains of three four-memcoordination. There are four HrO moleculesper unit bered tetrahedral rings alternating with a ring of
cell, each bonded to 2 Ca. The correct structural twelve tetrahedra.
formula is CarSnAlrSi6O,r(OH)r.2H2O,with Z : 2.
In eakerite each chain has an eight-tetrahedron
The Al is fully ordered and is bonded to three O and repeat. The chains zig-zag back and forth, with segone OH. Hydroxyl ions rarely are bonded into the ments of four tetrahedra alternately parallel [l l0]
tetrahedral network in silicates.Besideseakerite,ju- and I l0] (Fig. 3). The Al tetrahedra are at the
goldite and other members of the pumpellyite group ends of these segments.Each segmentof four tetraare examplesof such bonding (Allmann and Donnay, hedra is about 8.5 A long, about the same length as
1973).ln all of these,OH is bonded to Al rather than the correspondingunit in the feldspars,but because
Si, and Al is in an ordered position in the network.
of the zig-zag, the two-segment, eight-tetrahedra, rep e a t d i s t a n c e( a n d a a x i s ) i s o n l y 1 5 . 8 9A .
The aluminosilicatesheet
It is convenient to think of the aluminosilicate The cation sheet
sheet as composedof crankshaft-likechains roughly
The Sn and Ca atoms lie in sheetsalmost exactlyin
(001) plane, between the aluminosilicatesheets.
the
'For a copy of the structurefactor data,Table2, order DocuIn
Figure
3 they seem to be in large holes formed by
ment AM-76-024from the BusinessOffice,MineralogicalSociety
of America,1909K Street.N.W. 20006.Pleaseremit $1.00in the l2-membered rings, but the kinking of the sheets
advancefor the microfiche.
makes these holes less coherent than they appear in
CRYSTAL STRUCTURE OF EAKERITE
959
FIc. l. Stereoscopic
pair viewsof the crystalstructureof eakerite,vieweddown 6
that projection. Each of the cations does lie between
two chains and is bonded to oxygens in those two
chainsonly (and to HrO moleculesin the caseof Ca).
Sn is bonded to six unlinked oxygens in a nearly
regular octahedron. Ca is in an irregular square antiprism composed of two OH (O9), two HrO (O8),
two unlinked O (O4, 06), and two O linking Al and
Si (O3, 05). The Ca-O bond lengths vary by about
0.2 A, and on the averagethe bond lengths to OH
and HrO are slightly longer than those to the oxygens.
Figure 4 shows the polyhedral Ca-Sn sheet of
eakerite;it is composedof chains of edge-sharingCa
antiprisms parallel to b, cross-linkedby the Sn octahedra. The atoms comprising the sharededgesof the
antiprisms are OH or HrO. The sheetcontains large
holes,surrounded by 6 Ca polyhedra and 2 ofSn; the
axesofthese holesare parallel to (210) and (2T0).The
chains of Ca polyhedra are almost identical to those
in herderite, CaBePO,(OH,F) (Lager and Gibbs,
1974,Fig. lb). In herderitethe chains also are parallel to b, and the D dimension of the two minerals is
similar: eakerite7.72A, herderite7.66A.In herderite
the chains are alternatelycross-linkedto each other
stQ *
l9
83
c
O
s ; t D a t O a rc o O
"O
oO o"@
"ro@
Ftc 2. The structure of eakerite, viewed down b. Numbers give the atomic coordinates, in percent, along b. Bonds within the
tetrahedral network are indicated by solid lines; other bonds, including hydrogen bonds, by dotted lines. Hydrogen bonds to hydroxyl are
not included becauseof drafting difficulties. Broken lines indicate bonds between atoms in adjacent unit cells.
A. A. KOSSIAKOFFAND P B. LEAVENS
/A
'8\
85
'-""
@--"'"'
i $?
--
i
i
F^o
1..'
...''\
84
l6
tl8
AI
32
f@
Frc. 3. The structure of eakerite, viewed down c. Numbers give atomic coordinates, in percent, along c. Bonds within the tetrahedral
network are indicated by solid lines; other bonds, including hydrogen bonds, by dotted lines.
to form a sheet;this sheetcan be produced by remov- linear, hydrogen bonds as a function of the separaing the Sn octahedra from the eakerite sheetand by tion of the oxygen ions (cited in Donnay and Allman,
linking the Ca polyhedra directly to each other. Be- 1970). The assumption that the bond is linear procauseofthe interveningSn octahedra,the holesin the
eakerite sheet are larger than those in the herderite
Tnnle 3. Bond leneths of eakerite
sheet. Both the folding of the tetrahedral sheet in
eakeriteand the open linking of the polyhedral sheet
Dtstance(R)
Atom
ots tance (8)
Alon
ofthe high charge
can be thought of as consequences
of the Sn ion.
L.723(4)
Hydrogen bonds
Both the HrO (O8) and the OH (O9) are clearly
oversaturated.The method of Donnay and Allman
(1970) gives H2O (O8) an excessof 0.46 charge and
OH (O9) an excessof 0.21 charge (Table 4). This
oversaturation indicates that hydrogen bonds are
present. The distance between HrO (O8) and 07,
which links Si and Al. is 2.79 A. and that between
H,O, which links 2Si, is 2.76 A (f igs. 2, 3), much
closer than the normal minimum O-O distance in
inorganic structuresof 3.3 A, and in the typical range
for hydrogen bonds. Lippincott and Schroeder(1955)
calculatedthe fractional bond valenceof asvmmetric.
sn-0(1)
s n - O( 4 )
s n - o( 6 )
z . 0 L t ,( 4 )
2.023(4)
2.061(4)
c a - 0( 3 )
c a - 0( 4 )
c a - 0( 5 )
c a - O( 6 )
C a - O( 8 )
c a - 0 ( 9)
ca-0(8)'
ca-0(9)'
2.42L(s)
2.4L2(4)
2.5r1(4)
2.40L(4)
2.623(4)
2.4s8(4)
2.5o2(4)
2.622(4)
A 1 - 0( 3 )
A 1 - 0( s )
] - 7 5 5( 4 )
r.746(4)
The
standard
cleviatlon
l s given in parentheses'
in
A1-0(7)
A 1 - o( 9)
L . 74 7( 5 )
sr(1)-o(1)
si(1)-o(2)
si(1)-o(5)
si(1)-o(r0)
1.599(4)
1.63r(4)
1.614(s)
1.536(4)
si(2)-o(3)
s1(2)-o(4)
s1(2)-o(7)
si(2)-o(11)
r.620(5)
r.60s(4)
1.603(4)
r.634(6)
sl(3)-o(2)
si(3)-o(6)
si(3) -0(10)
s i ( 3 )- o ( 1 1 )
L,634(6)
1.606(4)
1.6r0(5)
| . 6 2 7( 5 )
the least
signiflcant
flgure(s)
CRYSTAL STRUCTURE OF EAKERITE
961
o
F I c . 4 T h e p o l y h e d r a l C a - S n s h e e ti n e a k e r i t e .S p o t s i n d i c a t e S n a t u n i t c e l l c o r n e r s
T e s r - E4 . B o n d a n s l e so f e a k e r i t e
Atom3
0 ( 1 )- s n - o( 4 )
0(1) sn-0(6)
0(4)-sn-0(6)
i\nglc
89.3(1)
39.7(2)
s 4 . 4( 2 )
0(3)'ca-a(5)
0(l) ca 0(6)
,)('l)_ca_0(t)
0(3)-ca 0(4)
rl(l)-ca-0(8)'
rl(l)_ca_0(9)'
r0s.0(2)
1053(r)
73 . 4 ( 7 )
1 3 6. ) ( 2 )
81.7(2)
64 9(1)
0 (5) -ca-0 (6)
0(s)-ca-o(8)
0(5)-ca-0(e)
0 ( s ) - c a - 0 ( 4)
0(5)-ca-0(3)'
0(5)-ca-0(e)'
35 2(3)
78.8(2)
64.9(1)
75.0(1)
110.3(1)
LA5.8(2)
0((r)-ca-0(3)
r)(r))-ca-0(q)
0(6)-ca-0(4)
c(6)-ca-0(8)'
r l ( 6 )- c a - O ( 9 ) '
3 0. 5 ( 1 )
r 4 o. 7 ( 2 )
69.5(2)
r 3 9. 2 ( 3 )
77.8(2)
0(3)-ca-0(9)
0 ( 3 )- c a - 0( 4 )
0(3)-ca-0(8)'
0(8)-ca-c(9)r
69.6(2)
r.41.4(3)
r . 4 33 ( 2 )
125.1(7)
0 ( 9 )- c a - 0( 4 )
0(9)'-ca-0(B)'
0 ( 9 )- c a - o( 9 ) '
0 ( 9 )- c a - 0( 4 )
r 2 1 . 4( 1 )
17.8(2)
r40 6(r)
7 2. 4 ( 7 )
0(5)-A1-0(9)
o ( 5 ) - A 1 - o( 7 )
0 ( 5 ) - A 1 - o( 3 )
0 (9) -Ar -0 ( 7)
0 (9) -At-0 (3)
c(7)-A1-0(3)
99.4(3)
7r4.4(2)
1 D 9. 9 ( 2 )
1170(2)
101.4(l)
1r3.1(3)
o( r o ) - s i ( 1 )- o ( 2 )
0 ( 1 0 )- s l ( 1 ) - 0 ( 5 )
0 ( r . 0 )- s i ( r ) - 0 ( 1 )
0(2)-si (r) -o (1)
0(5)-s1(1)-0(1)
107.3(3)
108.1(3)
r10.7(3)
r o s . 6( 2 )
1 1 5 . 7( 2 )
0(3)-si(2)-0(4)
0 ( 3 )- s r ( 2 ) - 0 ( 7 )
0 ( 3 ) - s l ( 2 )- 0 ( 1 1 )
0( 4 )- s i ( 2 ) - 0 ( 7 )
0(4)-si(2)-0(11)
110.8(2)
rc8.3(3)
1Ci.3(3)
109.9(3)
ro9 .9 (2)
0 ( 7 )- s i ( 2 )- 0 ( 1 r )
c ( 3 )- s i ( 2 ) - s i ( 3 )
0 ( 4 ) - s i ( 2 )- s r ( 3 )
0 ( 7 )- s i ( 2 ) - s i ( 3 )
0 ( 1 1 )- s i ( 2 ) - s i ( 3 )
1 1 0. 9 ( 3 )
r 0 3 . 9( 2 )
3 9. 9 ( 2 )
r32 .4 (2)
L 2 4. 0 ( 2 )
0 ( 6 ) - s i ( 3 )- 0 ( 1 0 )
0(6)-si(l) -0 (2)
0 ( . ) )- s i ( 3 ) - 0 ( 1 1 )
0 ( 1 0 )- s i ( 3 ) - 0 ( 2 )
0(10)-si(3)-0(11)
109.9(2)
109.1(1)
112.2(3)
1 0 8 . 2( 3 )
109,4(2)
0(2)-s1(3)-0(11)
0(6)-si(3)-si(2)
o ( 1 0 ) - s i ( 3 )- s i ( 2 )
0 ( 2 ) - s i ( 3 )- s i ( 2 )
1 0 1. 9 ( 2 )
91.8(2)
1 3 r. 1( 3 )
ro4.5(2)
vides the maximum charge transferral to the secondary oxygen, but the bond usually is bent, and transferral is below the maximum. Using their values,
h o w e v e r , 0 . l 8 3c h a r g ei s t r a n s f e r r e dl o 0 2 , a n d 0 . 1 7 0
chargeto 07. This reducesthe overchargeon HrO to
0 . 1 , b u t o v e r c h a r g e sO z 6 y 0 . 1 2 6 .
L i k e w i s e ,O H ( O 9 ) i s 3 . 0 2A f r o m O l , b o n d e dt o S i
and Sn, and markedly undersaturated (Table 5).
However, the large O9-Ol separation permits a
charge transferof only 0.096charge,leavingOH (O9)
o v e r s a t u r a t e db y 0 . 1 1 8 c h a r g e , a n d O 1 , u n d e r saturatedby 0. I 80 charge.The averagecompensated
valence on oxygen ions in the eakerite structure as
calculated by the method of Donnay and Allman
( 1 9 7 0 )i s 2 . 0 0 1 ,i n g o o d a g r e e m e nw
t ith the required
imbalances
local
residual
2, and suggestingthat the
on Ol, 08 (H,O), and 09 (OH) are real.
The strong bonding of the water molecules,both
by bonds to Ca and hydrogen bonds to other oxygens,explainswhy water is held to such high temperatures when eakerite is heated(Leavenset al., l97O)
and why the water and hydroxyl were not distinguished on the thermogravimetriccurve of eakerite.
Leavenset al., (1970) noted that the relative abundance of external forms on crystals of eakerite does
A. A, KOSSIAKOFFAND P. B. LEAVENS
962
T,c.sI-r5. Valence bond strengths of the oxygen atoms of eakerite
ot!
AI
tt2
1.04
.97
1
2
.28
.28
5
6
,it
.75
.91
1. 0 0
1.02
,78
.25
7
8(H2o)
1.03
\ri
1 .7 3
r.94
2.Or
1.99
1.00
I .03
.73
.24
.28
I
sheets.These two features, along with the large holes
oriented along [210] in the polyhedral sheets,seem
adequateto account for the morphologicalanomalies
of eakerite crystals,which require a pseudo-halving
of the c axis.
Acknowledgments
1.99
L,94
We would like to thank Robert H. Wood and Daniel Appleman
for their encouragement and advice.
1.81
.46
References
ALLMAN,Ruooln eNp GesntrLLsDoNNev (1973)The crystal
r.01
97
structureof julgoldite.Mineral. Mag 39,271-281.
1. 9 5
.98
DEER,W. A , R. A. Howre nNo J. ZussneN(1963)Rock-Forming
L1
Minerals.YoI. 4, FrameworkSilicates.Longman,London.435
p.
figure(s)
appear
The standard devlatlon
in the least slSniflcant
in pd eocheses.
DoNNAy,G.tnnrsLls nNn Rupou ALLMAN(1970)How to recognizeo2 , OH-, and H'O in crystalstructuresdeterminedby Xrays.Am Mineral.55' 1003-1015.
DoNNAy,J. D. H. luo Dnvto Hrnxen (1937)A newlaw of crystal
morphologyextendingthe Law of Bravais.Am. Mineral 22,
not conform to the law of Bravais as extended by
446-467.
D o n n a y a n d H a r k e r ( 1 9 3 7 ) ,s i n c et h e f o r m s { 2 1 0 }a n d Lrcen, Geoncs A. nNo G. V. Gtsss(1974)A refinement
of the
crystalstructureof herderite.Am. Mineral 59,919-925.
{410i were more common and prominent than expected, and the forms {100} and {110} less so. Al- Lre.veNs,Prrrn B., JonNS. WHrrE,JR.,rxo Max H. Hnv (1970)
Eakerite,a new tin silicate.Mineral.Rec. l' 92-96.
though the structure of eakeritecontains two silicate
LrrerNcorr, E R. eNo R. ScHnososn(1955)One-dimensional
chains per cell in the b axis direction, it has two such
modelofthe hydrogenbond."/. Chem.Phys 23' 1099-1106.
9(oil)
t0
.2r
.26
L.2l
1.98
doubled elements in the d axis direction: the 4 tetrahedron sub-repeatin the silicate chains, and the
two chains of Ca polyhedra per cell in the polyhedral
OctoLer6, 1974;accepled
Manuscriplreceiued,
for publication,April 30, 1976.