view file

Canadian Mineralogist
(1989)
Yol.27, pp. 193-197
REFINEMENT OF THE CRYSTAL STRUCTURE OF LEUCOPHANITE
JOEL D. GRICE
NationalMuseumof Natural Sciences,
Mineral Sciences
Division, Ottowa,OntarioKIP 6P4
FRANK C. HAWTHORNE
Depaftmentof GeologicalSciences,Universityof Manitoba, Winnipeg,ManitobaR3T 2N2
ABSTRAC"T
INTRoDUC"IIoN
The crystal structure of leucophanitg [CaNaBeSi2O6F,
a 7.401Q), b 7.412Q), c 9.990Q) A, orthorhombic,
P21212yZ = 41,was refined using 926 observed(>2.5o1)
reflections collected with MoKa X rays and a crystal from
Mont Sainl-Hilaire, Quebec;a similar refinement of a crystal from Stokkoy, Norway gavenearly identical results.The
structure is similar to that previously determined, but
required a shift in origin (+ Yt, *t/n, +3A) and an interchange of tle x and / coordinates. The revised structure
refined to R : 3.0 and R* = 2.7V0 based on 926
observedreflections, as compared \irith R- 2070 for tlle
samedata and the previous model. Leucophanite is a shell
structure basedon the 2dimensional net (515f)that is also
tire basis of the melilite-group minerals. The lower symmetry and different cation-orderingpatternsin lzucophanite
as compared with the melilite-group minerals are mainly
dictatedby the local bond-valencerequirementsof F, which
is required by stoichiometry in the leucophanite structure.
The crystal strucfure of leucophanitewas reported
by Cannille et al. (1967).Theseauthors usedcrystals from Brevig, Norway to measureintensity data
using multiple Weissenbergfilms and a microdensitometer.With 586observedreflections,they were
ableto refine the structureto a reidual of R = 9.4Vo.
We obtained leucophanite crystals from the
Desourdy quarry, Mont Saint-Hilaire, Rouville
County, Quebec(specimennumber NMNS 53455)
and from Stokkoy, Vestfold County, Norway
(NMNS 46861).The Saint-Hilaire leucophanitewas
analyzed by Chao (1967) and determined to be
Cao.esNao.x([email protected],
almost
pure end-member leucophanite, CaNaBeSi2O6F.
The analytical data given by Brdgger (1890) for the
Norwegian material also correspond to nearly endmember leucophanite.The structuresdeterminedon
crystals from both localities confirm their chemical
purity.
Keywords: leucophanite, Mont Saint-Hilaire, structure,
cation ordering, nets.
E><pnnrNtsNIAL
Sotrtvnrnn
Nous avons affin6 la structure cristalline du leucopha4e
[CaNaBeSi2O6F,a 1.4OlQ), b 7,412Q), c 9.990Q) A,
orthorhombique, P21212pZ = 41,en utilisant un 6chantillon du mont St-Hilate et 926 r6flexions observ6es
(I>2,5o1) avecrayonnementMoKa. Un affinement semblable, r6alis6avecun cristal de Stokkoy (Norvbge),a donn6
desrdsultats presqueidentiques. La structure, qui ressemble i celle qui avait 6td ddterminde, requiert un d6placement de I'origine (+ 1/4,+ rA, + /q) et une permutationdes
axesr et /. L'affinernent a atteint un r€sidu R : 3.090 et
Rw = 2.7t/o pour les 926 rdflexions observdes,compar6 i
un r6sidu d'environ R- 20t/o aver,le m0mesdonn6eset
le moddle structural pr6c6dent, Qs min&al poss0deune
structrue en coqrrille calqueesur la maille bidimensionnelle
(5j5i), qui est aussi celle des mindraux du groupe de la
mdlilite. La sym6trie plus faible et les agencementsordonnes diffdrents parmi les cations du leucophane, en comparaison des membresdu groupe de la m6lilite, r6sultent des
exigeanceslocales en valences des liaisons impliquant le
fluor, dont la pr€senceest imposeepar la stoechiomdtrie.
The crystal structureof leucophanitewas refined
for crystals from both Mont Saint-Hilaire, Quebeg
and Stokkoy, Norway; only the experiment for the
Saint-Hilaire samplewill be describedin detail as thq
techniquesand resultsare very similar for both stnrcture determinations.
The Saint-Hilaire crystal chosenfor data sollection was ground to an ellipsoid that measures0.23
x 0.23 x 0.13 mm. X-ray-diffraction spots on
precessionphotographsare sharp and intense.These
photographs indicate orthorhombic symmetry and
systematic extinctions consistent with the unique
space-gxoupnQtzr (#19). Intensity data were collected on a Nicolet R3m automated four-circle
diffractometer using the method of Grice & Ercit
(1980. The data relevant to the structure refinement
are given in Table 1.
Srnucrunr RsrrNEtvrENr
(Iraduit par la R6daction)
Mots<lds: letcophane, mont St-Hilaire, structure, mise en
ordre des cations, mailles.
The initial starting parametersfor the atoms were
those of Cannillo et al. (1967). Theseparameters
193
194
THE CANADIAN MINERALOGIST
refined with isotropic temperature-factors to an R
of 20.2s/obut would not refine further., It is interesting to note that evenat this R index, the atomic coordinates had high, but not unreasonable, standaxd
deviationsof approximately t 0.001for the cations
and t 0.@2 for anions. These standard deviations
are in the order of those recorded by Cannillo et al.
(1967); of course, theseatomic parametersgaverise
to the samereasonablepolyhedra and bond lengths.
The magnitudesof the temperature factors were not
considered to be fully satisfactory, but most were
believable. However, the residual index indicated a
sdriouserror in the model.
To find the correctstructure,.E-mapswerecalculated from a set of normalized structure-factors.
With a set of starting coordinates for the heavier
atoms, a seriesof refinementswith difference-Fourier
maps was run, and the structural model completed.
TABLEl.
DATA
LEUC0PHNI'II: STRUCIURE-REFINEMENI
a ( E ): 7 . 4 0(' r2 )
b([): 7.a]2(2)
cU): 9.990(2)
The finat R valuesin Table I are for refinement with
anisotropic temperature-fastors, and the weighting
schemeincorporatedan isotropic, primary-extinction
correction. The final positional parameters and
equivalentisotropic temperature-faclorsare given in
Table2, and the anisotropic temperature-factorsare
given in Table 3. Bond lengths and anglesare glven
in Table 4, and an empirical bond-valencesrunmation is shown in Table 5. The observedand calculated structure-factors have been submitted to the
Depository of UnpublishedData, CISTI, National
ResearchCouncil of Canada, Ottawa, CanadaKIA
0s2.
Leucophanitefrom Stokkoy, Norway gavealmost
identical results, with a slightly improved R of 2.6t/o'
The very closesimilarity to the Saint-Hilaire material
establishesthe $tructure reported here asthe correct
one; the results for this second structure also have
been depositedin CISTI.
DISCUSSION
Comparison of the atomic coordinatesfor the leucophanite structure as determinedby Cawillo et al.
(L967) to those reported here may be made by tak958
No. of Fo:
Radlation,/lilono: Mo/graphite
No. of Fo>2.5d(t): 926
16.0m-'
r:
ing the coordinates in Table 2, interchanging x and
3.07,
Final i:
Mln. transmlsslon: 0.619
y and adding Vrto eachand adding % to e. The shift
2.7
lihx. tfansmlsslon: 0.685
Flnal aw:
in origin involves one of 64 possibleorigins in space
.
a
:(lFol-lFcl)/tlFol
groupP2Qr21,but only.16of thesepossibilitieslead
a y . t ' l { ( l F o l - l F c l ) ' z / r w l F o l 2 l t ,w < - 2 ( F o )
to changesin the phases.The interchange of the x
and y coordinatesreflects the closeproximity of this
structure to tetragonal symmetry. The structure is
in fact closely related to the tetragonal structure of
TABLE2. LEUCoPHNI1!: PoSITIoNALANDIlERl.lALPARAI4EIERS
melilite, Ca2MgSipT. The lowering of symmetry
from tetragonal to orthorhombic for the leuu(eq)xIotR2
Slte
cophanite structure results from the ordering of Ca
ee(2)
0.6462(l) 0.75es(r)
ca
0.0788(1)
and Na at two distinct sites, as opposed to the sin0.6782(2) 0.2465(2) 2 o o ( 5 )
l,la
0.1021(2)
0 . 3 5 9 2 ( 6 ) 0 . 5 3 0 7 ( 4 ) 28(r0)
Be
0.1200(6)
gle Ca site in melilite (Smith 1953). The occupancy
(
3
)
76
sil
o.rooo(r) 0 . 3 5 8 0 ( r ) 0 . 0 2 4 r ( r )
for thesetwo sites refined to 1.043(4)for Ca and
si2 0.2371(2) 0 . 9 9 r 7 ( 2 ) 0 . 9 9 9 4)( 1 '|7812)
05(7)
0l
0 . 0 9 1 2 ( 4 ) 0.3548(4) 0.8645(2)
0.996(9) for Na, indicating cc rlete order.
02
0 . 1 4 8 3 ( 4 ) 0 r 1 5 8 0 ( 4 ) 0 . 0 8 9 5 ( 3 ) r 0 0 (7 )
1 0 r( 7 )
03
0.0986(4) 0 . 9 1 3 6 ( 4 ) o . 4 1 l (l 3 )
The most prominent feature of the leucophanite
e8(7)
04
0.0800(4) 0.9129(4) o.9o3o(3)
structure (Fig. 1) is the tetrahedral [BeSip6F] sheet
0 . r 7 6 9 ( 4 ) 0 . r 6 2 6 ( 4 ) 0 . 5 9 3 5 ( 3 ) 're 7 ( 7 )
05
or(6)
06
0.2553(4) 0.5076(4) 0.5892(3)
perpendicular to [001]. This strongly bonded struc0..1216(3) 0.3526(4) o.3729(2) 1 5 9 ( 7 )
F
ture module is linked into a 3-dimensional structure
by more weakly bonding alkali and alkaline earth
cations, and thus is a sheet $tructure in the termi
nology of Hawthorne (1985, 1986).If we consider
(XIOgEI)
IE.TPERAIURE-FACIORS
TABTf3. LEU@PHAI{ITE
: A}TISOTROPTC
the sheet as a 2-dimensional net (e.9., Smith 1977'
otz
0rr
orr
Hawthorne & Smith 1986), it is representedby the
general symbol (51514);
this signifies that there are
2r (3)
25(3)
6(3)
116(3)
8e(3)
Ca
and one 4-connectedververtices
two
3-connected
-64(8)
45(8)
52(7)
2 5 0 ( r 0 ) r7r(e)
Na
180(9)
0(0)
4(r5)
4e(r8)
0(0)
in the net, and that all the shortest circuits from
Be
30(18)
tex
d
t
a
l
-2(4)
0(4)
56(4)
73(4)
9 l( 5 )
stl
-J LCJ
-r r (4)
ai,i
69(4)
e8(4)
stz
68(4)
i each vertex are Pentagonal.
r1n4)
301)
1 7 ( r)1
0l
12e(12't
e 6 0 r)
9 0 ( r)1
It is of interest to consider the substitution of
e(lr)
r3(10)
6 7 ( r)
n0nl)
I (12)
02
124(r2')
8 ( 1 2 ) -9(r'r)
r09
0(1r)
n(12\
n 6 (r r)
06 2)
on the basic melilite+ype (515f) net. The
cations
'
r
1
(
1
2
)
-8(r2)
-290r)
n 5 (r ? )
65(r2)
il4(n)
04
31( 10)
t 0 ( ' l )r
may be written as X2YZ2O7,where
minerals
11502)
66n2) 1 r 0 ( r 2 ) -14(r2)
melilite
20112)
2'I8 ( 1)1 - 6 ( 1 0 )
r 0 4 ( 1)r
r o l( 1 2 )
99(il )
- 3 ( r 2 ) X: Ca, Na(+Ba, St . . .), I= Mg, N, Zn, Be
- 1 8 (1r)
(9)
162(12)
1 9 1 0 2 ) 1 2 3 ( r)
(+Co, Fe, Ni . . .), Z = Si,Al; Iications occupythe
Ideal Fomula: CaNaBeSl:0oF
P2r2t2t
space Group:
zi4
r95
THE CRYSTAL STRUCTI.JRE OF LEUCOPHANITE
rArLE4. LEUCoPHAI{ITE:
(R)ND ATOLES
(o)
SELECIEO
nftnAloillc oISTNCES
Ca squaE antl-DrlsB
Ca-01
Ca-04
Ca-06
Ca-old
Ct-02e
Ca-04d
&-05b
Ca-F c
@!n
2.401(4)
2.440(4\
2.381(4)
2.349G\
3.014(4)
2.650(4)
2.396(4)
2.48€G1
z:5I5*
0l -Cn-05b
05b-Ca-04
04 -Ca-F c
F c-Ca-01
06 -Ca-04d
o4d-C!-Otd
0ld-Ca-02e
02€-ca-06
Be tetrahedrcn
ii,iiii fr#
gi,xili *-H"on
1.642(6)
1.598(6)
r.577(5)
1.670(5)
I.&Z-
7 5 .(rr )
5 5 . 6 ( l)
s'l l EtrahedDn
ila square anil_prl9
Ia-03
Na-F
tla-010
Na-02d
!ln-03e
Ia-05e
l{a-05d
Na-F a
@an
2.398(4)
2.729(41
2.569(4)
2.479(4)
2,921141
2.530(4)
2.3/.ll4)
2.461(4)
2]553-
05 -Be-06
F -6e-05
F -Be-06
04d-Be-05
04d-Be-06
04d-Be-F
@an
F -tta-ozd
02d-Ia-03
03 -l,la-Ole
ole-tia-F
06d-Na-F
03e-tra-02d
F a-Na-03
05a-Na-0te
i
I
{
i*n
sfl-02
Sll-ola
Sll-03€
sil-06d
1.659(3)
r.596(3)
r .6se(4)
'r
.60r(3)
T3Z6-
st2-04
st2-02b
Sl2-03s
st2-05d
man
1.618(3)
1.662(3)
1.658(3)
t.612(3)
TldB-
t
)
l
5a vertex, and Z cations occupy the 53 vertex; this
type of vertex occupancy seemsto be the norm for
incorporation of divalent cations into polymerized
tetrahedral modules. Notably, gugiaite contains Be
at the 5a vertex. This contrasts strongly with leucophanite, in which Be occupies a 53vertex and Si
occupiesa 5avertex. Thus we have the questionof
why leucophanite is orthorhombic (and not
tetragonal), and also why the differeut ordering
schemein the tetrahedral sheet is different.
One notable aspectof leucophanite is that it has
both Na and Ca in equal amoutrts as intramodule
cations, and that theseare ordered in the structure.
By itself, this cannot be the determinative factor for
the orthorhombic symmetry, as synthetic .,sodamelilite", NaCaAlSi2O, (Louisnathan 1970), is
tetragonalwith the melilite strucfure, and hasNa and
Ca disordered at a single site. The occurrenceof Be
in the sheetcannot be the primary reason gugiaite,
Ca2BeSi2OT(Kimata & Ohashi l9g2), has the
tetragonal melilite structure. Zachariasen(1930)suggestedthat the presenceof F in leucophanite causes
the orthorhombic symmetry. This is probably the
case,but doesnot addresstle reaons(s)for the different ordering schemesof the leucophanite and melilite sheets.Furthermore, it is uot clear why F could
not disorder within a tetragonally symmetricalsheer.
In a gugiaite-compositionmelilite-Iike sheet,F cannot oscupy any of the bridging anion positions in
the sheetbecauseoflocal bond-valenceconstraints.
Only a single nonbridging aniou position is left to
accommodatethe F. This nonbridging anion (O2 in
the nomenclatureof Smith 1953)is bonded to one
Si and three Ca atoms; hence it also cannot be
replaced by F, again becauseof local bond-valence
restrictions. However, if the B€ occupies a 53 vertex, tlen the local environment of the O, anion (one
0la-sll-02
o3e-sll-02
03e-Sll-01.
o6d-sir-02
06d-Sl
l-0la
06d-sl1-03e
man
112.9(21
105.0(2)
110.9(2)
r04.6(2)
.l16.2(2)
106.2(2,
IM:t-
512 tetnhedmn
02b-s't2-04 107.8(2)
03s-sl2-04 lll.3(2)
0 3 q - s l 2 - 0 2 b1 0 5 . 3 ( 2 )
osa-srz-04 il2.0(2)
osd-sl2-ozb 111.5(2)
05d-Sl2-03q 108.7(21
rean
nlI?-
Be and three Ca atoms) is ideal for the incorporation of F. Thus we seetJratthe determinative factor
for the difference in the ordering scheme among
tetrahedra in leucophaniteis the needto incorporate
F into the sheetand to satisfy its local bond-valence
requirements.
The next problem is more difficult: why is leucophaniteorderedand orthorhombic rather tlan disordered and tetragonal? The 53 vertices are
occupiedby th Be * /z Si, suggestingthe possibility of ordering and a lowering of symmetry. GehIenite, CarAl(AlSi)O7, has lzN + /zSi occupying
the 53 vertices and yet is tetragonal (Louisnathan
l97l). However, Louisnathan(1971)also remarked
on the occurrenceof optically biaxial gehlenite, and
the presenceof very fine-scaletwinning in somegehlenite, suggestingthat incipient ordering occurs on
a very fine scale.In beryllosilicates,Be rarely if ever
disorderswith Si. This being the case,the occurrence
of Be at the 53 vertex in the melilite-like sheetwill
require these two vertices to be symmetrically distinct, lowering the symmetry of the structure. Given
the opportunity, the Ca and Na also order to produce
TASLE5.
BONIFYATEilCE
SIJIiIiIATIOII'
FORI"EUCOPMXITE
[a
0t
M
03
04
05
06
0.2n
0.326
0.085
3:i93
0.162
0.187
0.080
0.293
0.303
0.149
0.207
I
'1.t39
B€
r Brom (1981).
st(2)
1.072
0.908
0.910
0.90t
0.910
0.45s
0.490
0.54{r
1.057
1.950
3,947
o.re8 3:i3i o.4ar
Otl
St(l)
I.0lt
1 ,027
3.849
1.827
2.056
2.087
1.905
1.959
2.115
0,867
r96
THE CANADIAN MINERALOGIST
@Na
OGa
["{1e"
nsit
ffisiz
Frc. 1. A Z-axis projection of tle leucophanitestructure showing only the upper sheet
of tetrahedra within the unit cell. The lower tetrahedra, consistingof the same
topologrcal 2-dimensional net (SiSr4)is offset slightly in x and y coordinates.
the final structure. However, the determinative factor for the existenceof leucophaniteis the local bondvalence requirements around the F that must be
incorporated into the structure for overall electroneutrality.
Examination of the bond-valencevalues for leucophanite (Table 5) showsthat the deviations from
regularity in the various coordination polyhedra can
be ascribed to local bond-valence requirements. In
particular, the Si-O bond lengths to the Si-O-Si
bridging anions are lengthenedto allow for the additional bonding to the intramodule Ca and Na. Note
also how the relative ordering of Ca and Na has
occurred such that Ca can form more and stronger
bonds to those anions that are most underbonded
in the tetrahedral sheet [the nonbridging O(l) and
the Si-O-Be bridging O(4), O(5) and O(Q anionsl,
whereas Na forms more than one bond to those
anions that are most strongly bonded in the tetrahedral sheet[the Si-O-Si bridging O(3) anion]. The
control of this local connectivity and geometryis the
local bond-valence requirements, around both Ca
and Na. Thus leucophanite can be thought of as a
cation-ordered derivative of the melilite-type structure, the determining factor for the existence of
which is dependent on the stoichiometric need to
incorporate F into the structure.
ACICNIOWLEDGEMENTS
The authors thank Robert I. Gait of the Royal
Ontario Museum, Toronto, and H. Gary Ansell of
the Geological Survey of Canada, Ottawa, for
providing crystalsof leucophanite.Financial support
and
*as p.ovided to F.C.H. by the Natural Sciences
Engineering ResearchCouncil of Canada.
RSFERENcES
Bnoccrn, W.C. (1390):Die Mineraliender Syenitpegmatitgange der subnorwegischenAugit-,,und
Nephelinsyenite. Leukophan. Z. Kristallogr,
Mineral. 16, ?'46'n8.
Browll, I.D. (l9Sl): The bond-valencemethod: an
empiricalapproachto chemicalstructureand bond\ng. In Structure and Bonding in Crystals II (M.
OtKeeffe& A. Nawotsky, eds.).AcademicPress,
New York.
Cam.u,r,o,E., Grusnrrerrt,G. & Ttzzott, V. (1967):
The crystalstructureof leucophanite.Actq Crystallogr. 23,255-259.
TI{E CRYSTAL STRUCTURE OF LEUCOPHANITE
197
Cueo, G.Y. (1967):Leucophanite,elpidite and narsarsukite from the Desourdyquarry, Mont St-Hilaire,
Quebec.Can. Mineral, 9, 286-287,
LoursNArsar,S.J. (1970):The crystalstructureof synthetic soda melilite, CaNaAlSi2O?.Z. Kristallogr.
1'3t,3t+321.
Gr.rcr,J.D. & Encn, T.S. (1980:The srystalstructure
of moydite. Can. Mineral. M,675-678.
(1971):Refinementof the crystalstructureof
a natural gehlenite,Ca2Al(Al,Si)p7. Can.Mineral.
10, 822-837.
Hewruonre, F.C. (1985):Towardsa structuralclassification of minerals: tbe vIdvTz(Do minerals.
Am. Mineral. 10, 455473.
(1986): Structural hierarchy in vrM*rrrT"O,
minerals. Can. Mineral. A, 625-642.
Sr"rrnr,J.V. (1953):Reexaminationof the crystalstructure of melilite.Am. Mineral.3E, 643-661.
(1977):Enumerationof 4-connected
3-dimensionalnetsand classificationof frameworksilicates.
I. Perpendicularlinkagefrom simplehexagonalnet.
Am. Mineral, 62, 703-709.
& Srvrnn, J.V. (1986): Enumeration of
4-connected3-dimensionalnetsand classificationof
framework silicates. 3-dimensionalnets basedon Zacuaruaser,W.H. (1930):On meliphaniteand leuinsertion of 2-sonnectedverticesinto 3-connected
cophanite.Z. Kristallogr, 74, 226-229.
planenets.Z, Kristallogr,175, 15-30.
KrMArA,M. & Onesu, H. (1982):The crystalstructure
Neues Jahrb.
of synthetic gugiaite, Ca2BeSi2O7.
Mineral., Abh. 143,210-222.
ReceivedMay 3, 1988,revisedmanuscriptacceptedSe?
tember30, 1988.