American Mineralogist, Volume 75, pages 295-304, 1990

Crystal structuresof natural ternary apatites:Solid solutionin the

Ca.(POJTX(X : F, OH, Cl) system
JonN M. HucHrso Mlnvrlr-nN Carvrnnory KrvrN D. Cnowr-nv
Department of Geology, Miami University, Oxford, Ohio 45056, U.S.A.

Arsrn-l,cr
The crystal structures of P6r/m ternary apatite [Car(PO.)r(Fo.rrCL.rrOFtrr)]and a pre-
viously unrecognizedP2,/b tenary apatite [Car(PO.)r(Fo2ecl0oroHo2o); b : 2a, "( : l2ll
were refined to R values of 0.015 and 0.047, respectively,using three-dimensionalX-ray
data. Phosphatetetrahedraand Ca(l) polyhedra of both structuresare generallyvery sim-
ilar to analogouspolyhedra in the end-member fluor-, chlor-, and hydroxylapatite struc-
tures. The Ca(2) polyhedron, which includes among its ligandsthe column anions, exhibits
significant but regular variations in interatomic distances that can be directly correlated
to Cl content.
Solid solution in hexagonalternary apatite is achieved by a 0.4-A shift along c of the
Cl atom relative to its position in end-member chlorapatite. This adjustment affects a
Markovian sequen@of anions in the (0, 0, z) anion columns by providing a structural
environment that includes column OH speciesat a distance of 2.96 A from Cl. The shift
of the Cl atom is accompaniedby splitting of the Ca(2) atoms into two distinct positions
as a function of the kind of anion neighbor (F or OH vs. Cl). An additional nonequivalent
Cl site, similar to that in end-member chlorapatite, is also present. Those Cl atoms with
adjacent OH occupy a site different from Cl atoms adjacent to vacanciesin the anion
column.
The reduction of symmetry in monoclinic ternary apatite results from the ordering of
Cl and OH within the anion columns. The atomic positions of Cl and OH in the anion
column are equivalent to those in hexagonalternary apatite, but each is ordered into only
one of the two hexagonalsymmetry-equivalent sites.

INrnooucrroN mantle enrichment was addressedin studiesby Harrison

Like those of many common minerals, crystal structure and Watson (1984),Watson et al. (1985),and Exley and
studies of calcium phosphateapatites [Car(PO.)rX; (X : Smith (1982). Other recent studies have shown that F,
F, OH, Cl)l were long ago consideredcornplete.The hex- OH, and Cl concentrations correlate with etching rates
agonalP6r/m structure was determined 60 yearsago (N6- and annealing properties of U fission tracks in apatites
ray-Szab6,1930; Mehmel, 1930), and more recent re- (Green et al., 1986; Crowley and Cameron, 1987; Crow-
finements on end-members in the ternary system have ley et al., 1988; Crowley and Cameron, unpublished
illustrated monoclinic variants of the basic apatite struc-manuscript). The apatite structure has also received at-
ture (chlorapatite: e.g., Mackie et al., 1972; Hounslow tention as a host phasefor long-lived actinides in nuclear
and Chao, 1970;hydroxylapatite:e.g.,Elliot etal.,1973). waste formulations (e.g.,Weber et al., 1985; Weber and
Detailed structure descriptions and polyhedral arrange- Roberts, 1983).
ments are illustrated by Beeversand Mclntyre (1946) and Recent experiments in our laboratory demonstrate that
Hugheset al. (1989). many fundamental aspectsof the apatite atomic arrange-
Interest among geologistsin the apatite minerals has ment are not well understood, including mechanismsof
increasedin recentyears.Petrologistshave proposedthat ternary solution in natural crystals. In this study of ter-
the F, OH, and Cl variations in calcium phosphateapa- nary apatites we describe the crystal structures of hexag-
tites are potential geothermometers (e.9., Stormer and onal ternary apatile 'and a previously unrecognized
Carmichael, 197L) and indicators of P activity and vol- monoclinic variant of the ternary apatite structure.
atile fugacity in magmatic, metamorphic, and hydrother-
ExprnrunNTAL DETATLS
mal processes (e.g.,Yardley, 1985;Candela,1986;Boud-
reau and McCallum, 1987; Morrison and Valley, 1989). Hexagonal ternary apatite
Apatite is among the most important rare-earth element- Hexagonalternary F-OH-CI apatite wasrepresentedby
bearing phasesin igneous rocks (e.g., Henderson, 1984), a specimencollected from "ash F'in the Gunnison For-
and its role in crustal anatexis,granitoid production, and mation near Jackson Peak, southwesternUtah (Nielson,
0003{04Xl90/03044295$02.00 29s
296 HUGHESET AL.: CRYSTALSTRUCTURESOF TERNARY APATITES

TABLE1. Experimentaldata for hexagonal and monoclinicter- TABLE3. Anisotropic thermal parameters (x 104)for atoms in
nary apatites hexagonal ternary apatite
Hexagonal Monoclinic a
P23

Size Elliptical,0.13x Approx.tetrahedron, ca(1) 41.7(6) 41.7 37(21 41.7 00

0 . 1 3 x 0 . 1 6m m 0.35mmon edge Ca(2)A 27.1(8) 13.2(8) 42(21 1 8 0 ) 00
Ca(2)B 32(21 2(2) 37(3) 6(2) 00
Cellparameters
Leastsquares P 2s.8(7) 27.1(7) 38(1) 32(11 00
a (A) s.45s(2) s.48q6) o(1) 43(2) s5(2) 57(4) 73(3) 00
b 9.464(2) 18.96q9) o(2) 52(2) 30(2) 110(s) 43(3) 00
o(3) 52(21 106(2) 68(3) 106(2) -s2(4) -81(5)
6.84s(2) 6.822(5) 29 29 220 29 00
F
a (') e0.03(2) 90.00(5)
89.98(2) 8s.9q6)
o(H) 30 30 120 30 00
Cl" 56 56 270 5o 00
7 119.99(1) 119.97(5) 270 56 00
Clo 56 56
ldealized
a (A) 9.4615 9.4877 A/ote.'Theform of the anisotropicdispla@mentparameteris exp[-(B,, .
b 18.9628 tf + p22.tc + Bs. P + p'". hk r B6- hl + p4- kl)1,
6.8491 6.8224
r (') 119.9739
Radiation MoKd,Zr filter Mor(a,graphite
monochromated Precessiona- and b-axis photogaphs confirmed the
Theta range 0-2e 0-26. absence of monoclinic superstructure reflections and
No. of reflections 1206 4172
Absorption yielded diffraction effectsconsistentwith space1;rovpP6/
correction psrscans psi scans,DTFABS m. Intensity data used for structure refinement were mea-
Unique reflections 386 2090
R."*" 0.014 0.030 sured with an Enraf-Nonius cao-+ diffractometer. Table
Reflectionswith I summarizes the details of data collection and refine-
I>3o, 307 ment.
n 0.015 o.o47
P 0.025 0.043 All crystal structure calculations were undertaken in
Ap residual(e/A3) spacegroup P6r/m using the sDp package of computer
(+) o.202 0.910 programs (Frenz, 1985),employing neutral-atom scatter-
(-) 0.362 0.594
ing factors with terms for anomalous dispersion.Starting
A/ote.'ForTable 1 and subsequent tables, numbers in parenthesesde- parameters were taken from fluorapatite (Hughes et al.,
note one esd of least significantfigure.
1989), and column-anion positions were initially located
using difference maps. Fluorine was fixed in the mirror
plane at (0, 0, %), and differencemaps supported the va-
1988); crystal abrasion during airborne transport of the lidity of this assumption. The thermal parametersof F,
ash yielded rounded, nearly spherical crystals. Compo- O(H), and Cl were fixed at the values of the hexagonal
sition of the crystals was determined by electron probe end-members(Hughes et al., 1989), and an isotropic ex-
microanalysis with the Southern Methodist University tinction factorwas refined.No attemptwas made to locate
JEoLsupERpRoBE using well-analyzed standards. (Full de- the H atom becauseofits low concentration.
tails of analysisare available from the authors.) Analysis Full-matrix, least-squaresrefinement was undertaken
of the Gunnison sample yielded (CaoruMn"o,FeoorSroo,- using data with 1 > 3o, and structure factors weighted
Nao orCeo o,)", 03(P2e4Si. 02SooJ", ,rO,r(Fo ,eclo 3jOHo 28)", oo, proportional to o;2, with a term to downweight intense
with OH determined by difference.For structure refine- reflections. Initial refinement was undertaken assuming
ment the composition was assumedto correspondto pure single sites for Ca(2) and Cl, and the refinement converged
calcium phosphate apatite, with column-anion compo- at R : 0.017. Refined thermal parametersof Ca(2) sup-
sition as determined by microprobe analysis. ported imposition of the disordered Ca(2) model of Su-
darsananandYoung(1978).In this split-sitemodel Ca(2),
occupancy.
TABLE 2. Positionalparametersand equivalentisotropictem- occupancywas constrained to be equal to Cl
peraturefactors,B (4")for hexagonalternaryapatite With the two Ca(2) positions the structure refined rou-
tinely, including anisotropic thermal parameters for both
z B(4") Ca(2) sites, to R : 0.015. Thermal parametersof the Cl
ca(1) ry3 1/s 0.0018(1) 0.98(1) ion were then releasedto assessany site disordering (val-
Ca(2)^ -0.0069s(9) 0.23861(9) th 0.64(2)
-0.0018(2) 1/t
ues were subsequently fixed as noted above). The Urr
C42). o.27o9(21 0.60(3)
P 0.36968(8) 0.40053(8) 1/t 0.68(1) value ofthe Cl atom refined to an unusually large value
o(1) o.4862(21 0.3315(2) 1/e 1.08(4) of 0.098 A2, suggestingdisordering of the Cl. The site was
o(2) 0.4651(2) 0.5884(2) 1/a 1.42(5)
o(3)
then split and multiplicities ofboth siteswererefined.The
0.2597(2) 0.34s0(2) 0.06ss(3) 1.68(3)
F 0 O 1/t 1.93 refinement convergedat R : 0.015, and the sum of the
o(H) 0 0 0.200(3) 1.31 multiplicities of the two chlorine sites convergedat 0.63
Cl" 0 0 0.368(2) 2.68
Cl^ 0 0 0.44(1) 2.68 atoms per unit cell, vs. 0.66 asdetermined by microprobe.
The significanceof the split Cl atoms is discussedin a
Nofe.'Subscripts A and B refer to positions for the split atoms.
later section. Examination of bond lengths showed one
 HUGHESET AL.: CRYSTALSTRUCTURESOF TERNARY APATITES 29'7

unusually short Ca(2)"-O(2)bond (2. I 6 A), suggestingthat

the Ca(2) disorder propagatesto disorder O(2). The ther-
mal parametersof O(2) support this positional disorder,
but a split O(2) model was not invoked.
Table 2 contains positional parametersand equivalent
isotropic B values for all atoms, and Table 3 presents
anisotropic thermal parameters.Table 4 lists selectedbond
lengths and tetrahedral bond anglesfor hexagonal ternary
apatite, and Table 5' gives observedand calculatedstruc-
ture factors.

Monoclinic ternary apatite

Monoclinic ternary apatite was found in a split of ma-
terial from which monoclinic chlorapatite was first de-
scribed (Hounslow and Chao, 1970).The crystal selected
was a pale pink crystal thought to representmonoclinic
chlorapatite. The crystal was mounted and oriented for
precessionstudies (MoKa radiation). Photographstaken
with c* (monoclinic first setting for comparison with the
hexagonalapatite cell) as the goniometer head axis dis-
played a superstructurealong one ofthe pseudohexagonal
a-axes (Fig. l); comparison of lattice parametersand in-
tensities of the superstructure reflections with those of
monoclinic chlorapatite (Hounslow and Chao, 1970)
demonstratedthat the material was not monoclinic chlor-
apatite. Microprobe analysis undertaken after data col-

' Copiesof Tables5, 8, and 9 may be orderedas Document Fig. 1. Precessionphotograph (a axis, OJevel)of monoclinic
AM-90-428from the Business Office,MineralogicalSocietyof ternary apatite. Superstructure reffections (example denoted by
America,1625I Street,N.W.,Suite414,Washington,D.C.20006, arrow) exist as (k : 2n + l, I + 0). MoKa radiation, Zr filter,
U.S.A.Pleaseremit $5.00in advancefor the microfiche. 4.0 h.

TABLE4. Selected bond lengths (A) and tetrahedral bond angles (") for hexagonalternary apatite

ca(1f-o11;r'a'c 2.403(21
-O(2;o" 2.450(21
-O(31o" 2.806(2)
Mean 2.553
Ca(2X-O(1)E 2.761(21
-o(2f 2.433(2)
-o13y'' 2.522(21
-o13Y' 2.3s3(2)
_F 2.2s13(e)
-o(Hr 2.317(3)
-cti 2.63(4)
Mean.,y 2.49',1
Ca(28-O(1)B 2.741(2)
-O(2)' 2.164(21
-o(sy,r 2.530(21
-O(3yt 2.314(21
-cta 2.6ss(4)
Mean*y 2.432
PA-O(1)o 1.537(3)
-o(2r 1.540(2)
-o(0;^i 1.528(21
Mean 1.533
O(1)^-Pa-O(2f 111 . 0 6 ( 1 2 ) o(2r-PA-o(3r 107.62(9)
o(1r-P^-o(3r 111.34(9) o(3r-PA-o(3I 107.69(9)
o: r,y,z,Table2;B: -y, x- y,zic: y - x, -x,zio: -x,-1, -zl,: y,y - x,-zlF: x- y,x, -zio = - x,-y,V2 + 2,": y,y - x,
y 2+ z i ' : x - -
Y ,x , l z + z t J : x , y , V z z l K : - y , x - y , 1 / z- 2 i L : . y - x , - x , l z - z .
298 HUGHESET AL.: CRYSTALSTRUCTURESOF TERNARYAPATITES

Tmu 6. PositionalparametersandequivalentisotropicB (Ar) metry constraints) using diffraction angles from 25 au-
for monoclinic ternaryapatite tomatically centered reflections (Table l). The cell was
defined with c as the unique monoclinic axis and D as the
superstructureaxis; thus b : 2a. Resultsof the cell refine-
Ca(1), 0.3313(1) 0.58243(5) 0.0020(1) 1.14(21
V3 0.58333 0.0018 ment and details of three-dimensionaldata collection are
Ca(1),, 0.3356(1) 0.58434(5) 0.4980(2) 1.51(2) presentedin Table l.
Y3 0.58383 0.4982 The data setand precessionphotographswereexamined
Ca(2), 0.2511(1) 0.24809(s) 0.2498(2) 1.48(21
0.24887 0.24739 1/e for systematicextinctions; refinement was initiated in space
Ca(2),, -0.0038(1) 0.622s8(6) 0.7492(2) 1.48(2) P2lb becauseofthe absenceof hkl, k:2n + |
-0.00523 $oup
0.62295 Vc reflections and by analogy to other monoclinic apatites.
Ca(2),, 0.2550(1) 0.37ss4(5) 0.750q2) 1.49(21
0.25410 0.37444 Vt The zero-moment test yielded an intensity distribution
Pl 0.4018(1) 0.43534(6) 0.2505(2) 1.01(2) not inconsistent with the centrosymmetric spacegroup.
0.40053 0.43484 Vt
Pu 0.6292(1) 0.26557(6) 0.2508(2) 1.02(2) Subsequentsuccessfulrefinement in P2r/b (and unsuc-
0.63032 0.26543 V4 cessfulrefinement in P2, and PD) supported the choice of
Pu' 0.03120) 0.45093(6) 0.7502(21 1.02(2) spacegroup.
0.03085 O.45O27 Ta
o(1), 0.3346(3) 0.4939(2) 0.2510(5) 1.41(6) Full-matrix least-squaresrefinementwasinitiated using
0.3315 0.4931 V4 the atomic positions for Ca, P, and O in P2,/b chlorapatile
O(1)" 0.4880(4) 0.3268{2) 0.7s11(5) 1.41(7) (Mackie et al., 1972). During early stagesof refinement,
o.4862 0.3274 Tt
O(1)'u 0.1534(3) 0.5829(2) 0.2492(51 1.41(7) no column anions were included and isotropic B values
0.1547 0.5843 V4 were fixed. After initial refinement, equivalent data were
o(2), 0.5894(4) 0.4825(21 0.246s(6) 1.63(7)
0.5884 0.4826 l/c averagedand isotropic thermal parameterswere refined.
o(2)', 0.s346(4) 0.3121(2) 0.2490(6) 1.62(7) The R value convergedat 0.146 without addition of col-
0.5349 0.3117 V4 umn anions. A differencemap then revealedthat the four
O(2),u 0.1242(4) o.s447(21 0.7476(6) 1.60(7)
0.1233 0.5442 lq largest peaks were near (0, t/t, z), which are equivalent to
o(3), 0.3441(4) 0.3800(2) 0.0709(6) 2.00(7) (0, 0, z) positions in space$oup P63/m. Based on the
0.3450 0.3799 0.0699
O(3),' 0.7399(4) 0.2926(21 0.7010(6) 1.8s(8)
bond dislance to Ca(2), one site was assignedto Cl and
0.7403 0.2927 0.0699 the other three were modeled with F scatteringfactors to
O(3)'" 0.0846(4) 0.4217(2) 0.5710(6) 1.eq7) representthe iso-electronicF and O ions. Occupancywas
0.0853 0.4225 0.5699
O(3)'" 0.3508(4) 0.3817(2) 0.4331(6) 2.00(7) refined, and fixed isotropic ,B values of F : 1.0 and Cl :
0.3450 0.3799 0.4301 1.5 A' were assigned.The R value convergedat 0.081,
o(3), 0.736s(4) 0.2932(21 0.4329(6) 2.10(8) and a differencemap revealed no missing atoms.
0.7403 0.2927 0.4301
o(3)", 0.086s(4) 0.4258(2) 0.9338(6) 2.05(8) The psi-scan<orrecteddata were further adjustedusing
0.0853 0.4225 0.9301 the program DTFABS of Walker and Stuart (1983). After
Cl" - 0.000(1) 0.2507(6) 0.130(2) 1.39(20) re-averagingdata and refining all atoms except column
n V4 0.132
Cl^ -0.0000(s) 0.2494(21 0.4345(7) 1.22(6) anions in the anisotropic mode, R : 0.049.
0 V4 0.44 Comparison of positions of the column anion siteswith
F 0.0000(8) 0.2502(41 0.237(11 1.40(10)
0 V4 1/t those in hexagonalternary apatite suggestedthat one site
o(H) 0.0006(8) 0.2498(41 0.340(1) 0.14(10) assignedto O andlor F in the early stagesofrefinement
0 V4 0.30 was equivalent to a chlorine position in hexagonalternary
Note: For all atoms, the atomic coordinates in the second row are cal- apatite. After assigning Cl to that position and further
culated from the parameters of equivalent atoms in hexagonal ternary refining the structure by alternatively refining column-
apatite [without disordered Ca(2)] with appropriatetransformationto the
cellfrcltntlb. anion occupancy and isotropic thermal parameters, R
'Roman numeral subscripts
designate monoclinic, symmetricallyde- convergedat0.047. The data did not support anisotropic
generate equivalentsof hexagonalatoms.
thermal-parameter refinement for column anions. In the
final refinement with isotropic temperaturefactors for the
column anions, the -Bvalue of the column hydroxyl oxy-
lection confirmed that the crystal was a F-, OH-, Cl-bear- gen refined to an atypically low value (0.14 A'z). Anion
ing ternary apatite. Microprobe analysis yielded the occupancy convergedat Cl : 0.37, F : 0.38, and O(H)
formula Ca,or(P,ooSio or)"r.orO,r(ForeoHo24CL.47)rr 00,with : 0.29 atoms per formula unit, comparedto Cl : 0.47,
OH obtained by difference. F : 0.29 and O(H) : 0.24 atoms per formula unit by
Prior to structure analysis we realized that the crystal, microprobe. The differencesare attributed to the inferior
the only one isolated to date, was not of ideal quality for quality of the crystal; diference maps showedno unusual
a high-precision X-ray structure study. Abundant inclu- peaksin the anion columns.
sions (-3-4 volo/o?)and the large crystal size precluded a A zero-sigmacutoffand unit weights for all reflections
superior refinement, but the unique nature of the prelim- were used throughout all refinement cycles.The data set
inary data prompted further study. essentiallyconsistsoftwo groups ofreflections, the very
The crystal was mormted on an Enraf-Nonius ca>+ dif- weak superstructurereflectionsand the strong"hexagonal
fractometer. Lattice parameters were refined (no sym- apatite" reflections.Exhaustiveexperimentationwith var-
 HUGHES ET AL,: CRYSTAL STRUCTURES OF TERNARY APATITES 299

TABLE7. Bond lengths (A) and tetrahedral bond angles (") for atoms in monoclinicternary apatite

Ca(1),-O(1)l 2.400(41 Ca(1)il-O(1), 2.402(41

-o(1)" 2.4O7(3) -o(1), 2.390(3)
-O(1)'' 2.390(4) -O(1)'' 2.41344)
-o(2), 2.425(41 -o(21' 2.461(41
-o(2)', 2.452(3) -o(2)" 2.437(41
-o(2)'" 2.445(4) -o(2)'' 2.446{4)
-o(3)' 2.837(4) -O(3),u 2.845(3)
-O(3)" 2.814(5) -O(3)'" 2.752(4)
-o(3)", 2.746t3) -O(3)u 2.769(s)
Mean 2.546 Mean 2.546
Ca(2),-O(1),,, 2.788(4) Ca(2),,-O(1)l 2.7s3(3)
-o(2),, 2.334(3) -O(2)'' 2.333(4)
-o(3), 2.514(41 -O(3)', 2.518(3)
-o(3),' 2.339(4) -O(3)'' 2.328(41
-o(3), 2.546(41 -o(3L 2.s45(3)
-o(3)" 2.322(41 -o(3L, 2.327(4)
-Cl" 2.542(12) 2.s3(1)
-Clo 2.706(5) -Clo 2.717(s)
-F 2.404(9) -F 2.3s7(e)
-o(H) 2.471(9) -o(H) 2.482(sl
Mean.,, 2.474 Mean.,y 2.474
Ca(2),,,-O(1),1 2.791(41
-o(2)' 2.33s(3)
-o(3)' 2.332(4)
-O(3)'' 2.5O7(4)
-O(3)'" 2.328(41
-o(3)'' 2.559(5)
_ule 2.s44(8)
2.70s141-
-F 2.404(61
-o(H) 2.479(61
Mean",y 2.475
P,-O(1), 1.532(4) P,r-O(1)il 1.535(3)
-o(2), 1.542(3) -o(2)', 1.540(4)
-o(3), 1.526(4) -o(3)" 1.527(4)
-o(3)," 1.526(4) -O(3)u 1.s24(4)
Mean 1.532 Mean 1.532
o(1),-P,-O(2), 110.93(18) o(1),,-P,,-O(2)rl 110.84(18)
o(1),-P,-O(3), 111.35(221 o(1),,-P,,-O(3)rl 111.62(20)
o(2),-P,-O(3),u 111.90(22) o(1),,-P,,-O(3), 111.72(21)
o(2),-P,-O(3), 1O7.37(21) o(2),,-Pr,-O(3),, 107.54(231
o(2),-P,-O(s),v 106.93(21) o(2),r-P,,-O(3)u 106.83(23)
o(3),-P,-O(3),v 108.1s(19) o(3)r-Pr-O(3)u 108.06(20)
Pil,-O(1)il, 1.s3s(3)
-o(2). 1.s41(3)
-o(3)'' 1.530(4)
-O(3)u' 1.524(4)
Mean 1.533
o(1),,r-P,,r-O(2)r,l 110.97(21)
o(1)r,r-Pr,r-O(3)il, 111.48(19)
o(1),,,-P,,,-O(3)v, 111.54(19)
o(2),,,-P,,,-O(3)r, 107.8809)
o(2)il,-Pil,-O(3)vl 106.42(20)
o(3),r,-Pr,-O(3)vl 108.34(24)

ious weighting schemesand cutoff values demonstrated atoms [except O(H)], and Table 9 (seefootnote l) gives
that their use biased the data set against the weaker su- observed and calculated structure factors.
perstructure reflections, and thus unit weights for all re-
flections were used. This choice of refinement parameters
CoNrrcuru.TroN oF ANroN coLUMNs AND
may affect the final R value and thermal parameters,but
MECHANISMS OF.SOLID SOLUTION
it does not bias the data set nor ultimately the positional
parametersagainst the superstructurereflections. Hexagonal variants
Table 6 presents positional parameters and isotropic In end-member hexagonalapatite structuresthe posi-
thermal parameters for all atoms in the structure, and tions ofcolumn anions arenot constrainedbyunlike anion
Table 7 gives selectedbond lengthsand tetrahedral bond speciesin the [0, 0, z] anion columns. Thus the column
anglesfor the structure. Table 8 (seefootnote l) presents anions occur at sitesthat result in the ideal bond distance
anisotropic thermal parametersfor all cations and oxygen to the trigonally disposedCa(2)atoms in the mirror planes
300 HUGHESET AL.: CRYSTALSTRUCTURESOF TERNARY APATITES

t0,0,zl parent how solid solution is achievedin hexagonalternary

apatite. The structure refinement of hexagonal ternary
apatite described below demonstrates that one mecha-
z:314 nism that enablessolid solutionin hexagonal temary apa-
Ca(2) tites involves a 0.4 A shift of the Cl atom toward its
associatedmirror plane, relative to the Cl position in end-
member chlorapatite. This structural adjustment allows
an OH neighbor at the OH position near the adjacent
mirror plane.
Anions in the [0, 0, z] columns are located at different
positions in end-member and ternary apatite structures.
In hexagonal end-member chlorapatite, the Cl atom is
unfettered by adjacent non-Cl anions in the [0, 0, z)col-
z:I/ 4 umns, and the atom half-occupies a position at (0, 0,
0.5677) about the z : 3/qmirror plane. In hexagonalter-
nary apatire,however, the Cl atom position shifts in re-
spons€to the presenceofnon-Cl anion neighbors.The Cl
I responseinvolves a shift of 0.4 A closer to the z : 3/q
mirror plane, to a half-occupied (0, 0, 0.632) position.
Fig.2. Depictionof possibleanionpositionsin thehexagonal This shift in Cl position resultsin an atomic arrangement
ternaryapatitestnrcture.Stippledplanesrepresentmirror planes that accommodatesan OH neighbor at the OH position
at z : t/t,%,eachcontaininga triangleof Ca(2)atoms(connected disordered below the adjacent plane at z : t/t lO(H)' in
by "bonds").Atom Cl represents a Cl atom disorderedbelow Fig. 21.The Cl-O distance is a reasonable2.96 A, com-
the mirror plane at z : 3/q,and other column anionsrepresent
pared with 2.52 Lcalculated with anion positions in the
thefrvepossibleanionneighbors associatedwith themirror plane
at z : lc (Cl' : Cl disorderedabovethat plane;O(H) : OH hexagonalend-member chlor- and hydroxylapatite struc-
oxygendisorderedabovethat plane;F: fluorineatornat 0, 0, tures.
Yr;and O(H)' and Cl" respresent OH oxygenand Cl atomsdis- Associatedwith the shift of the Cl position is splitting
orderedbelowthe mirror planeat z : Vc). of the single Ca(2) site into two disordered sites. In a
structure refinement of another ternary apatite, Sudarsa-
nan and Young (1978) demonstratedthat the shift of the
at z : t/r,3h(Fig. 2). In hexagonalfluorapatite, the F atoms Cl atoms closer to the Ca(2) triangles causesa portion of
lie in the plane of the mirrors at the center of the Ca(2) the Ca(2)atoms (a portion equivalent to Cl concentration)
triangles. In hexagonalhydroxylapatite, the OH ions are to disorder to a new position to maintain the ideal Ca-Cl
displaced from the Ca(2) triangle and occupy disordered, bond distance (ca. 2.70 A;. fwo Ca(2) positions thus re-
mirror-related sites near (0, 0, 0.20) and (0, 0, 0.30) [and sult Ca(2)A,in which Ca(2) atoms bond to F or OH species,
(0, 0, 0.70)and (0, 0, 0.80)1,eachhalf-occupied.Like OH, and Ca(2)",in which Ca bonds to the Cl" atoms. An added
the larger Cl atoms in chlorapatite are also displacedfrom complication not recognizedin the Sudarsananand Young
the centersof the Ca(2) triangles and are disordered be- sample is evident, however, in two sets of split sites ob-
tween mirror-related sitesnear (0, 0, 0.43) and (0, 0, 0.07) served for the Cl ion in our sample.
[and (0, 0, 0.93) and (0, 0, 0.57)] that are 1.2 A from the As noted previously, thermal parametersof the Cl ion
plane. in our refinement suggestedsplit Cl sites. Upon splitting
The three hexagonal end-member anion columns are of the Cl into two sites, the positions of the split atoms
incompatible in solid solution. Figure 2 depicts an ex- refined to those of Cl" [ca. (0, 0, 0.37)], the "normal"
ample of a Cl atom (the atom labeledCl in Fig. 2) located ternary apatite Cl position constrained by a OH and by
at the Cl position (z : 0.57) disorderedbelow the mirror bonding to Ca(2)"and Cl^ [ca. (0, 0, 0.44)], a Cl position
plane at z : 3/q.For that particular Cl atom, possible adjacent to a vacancy. [As shown by Sudarsananet al.
column-anion neighborsassociatedwith the adjacentmir- (1977) and Wilson et al. (1977), vacanciescommonly oc-
ror plane at z : t/r are (l) a Cl located above the z : Yq cur in apatite anion columns.l The Cl^ position bonds to
plane(Cl-Cl'distance:0.92A;, 1Z;an OH ion disordered Ca(2)", and is identical to the Cl position in hexagonal
above the mirror plane [CI-O(H) distance : l.8l A], (:) end-member chlorapatite, as in both casesthe Cl is un-
an F ion in the plane of the mirror at (0, 0, y4) (Cl-F fettered by unlike anion neighbors.Refinement of Cl oc-
distance : 2.16 A), (+) a OH ion disordered below the cupanciesindicated that approximately 3oloof the column-
mirror plane [CI-O(H)' distance: 2.51A], and (5) a Cl anion sites are occupied by Cl", representinga Cl atom
ion disordered below the z: Yrmirror plane (Cl-Cl" dis- or a sequenceof Cl atoms bounded by a vacancy in the
tance : c/2 : 3.4 A;. On the basis of anion positions in anion column. In hexagonalternary apatite, Cl thus oc-
the hexagonalend-member structures,all of these anion cupies two sets of mirror-related sites about the mirror
neighbors for the Cl ion except Cl" are prohibited based p l a n ea t z : t / q I C l u : ( 0 , 0 , 0 . 3 7 ) (, 0 , 0 , 0 . 1 3 ) ; c l o : ( 0 ,
on anion-anion distances.It is thus not immediately ap- 0,0.44), (0, 0, 0.06)l; as Cl atoms are distributedequally
 HUGHESET AL.: CRYSTALSTRUCTURESOF TERNARY APATITES 301

betweenthe sitesin eachset,the locally monoclinic struc-

ture averagesover the entire crystal to hexagonal sym-
PQtm P2ltb
metry.
Miscibility of F, Cl, and OH in hexagonalternary apa-
tite is thus a result of a Markovian anion sequencein the
10,0, zl anion columns. The Cl" atom disordered below
the mirror plane at z :3/q (atom Cl in Fig. 2) cannot have
cla cle
an F atom [ocated at (0, 0, Yr)]or OH disordered above 0.44 0.43
the mirror plane at z : Vqllocated at (0, 0, 0.30); atom
O(H) in Fig. 2l as a nearestcolumn neighbor along -c. Cle
To changethe chemical speciesin the sequence,the Cl
atom must be adjacent to a OH ion disorderedbelow the
0.37 o(H)
0.34
mirror plane at z : t/t lat (0, 0, 0.20)1,2.96 Adistant. The
F atoms can then be accommodatedas neighbors of that
o(H)
OH ion, thus effectingternary solution.
0.30
Monoclinic variants
m b
F
Monoclinic apatite variants include end-memberchlor-
and hydroxylapatite that contain few or no "reversal sites"
o(H) 0.24
in the anion columns to bring about disorder. In these
0.20
structuresthe OH and Cl ions are ordered into one of the
two equivalent sites oftheir hexagonalcounterparts,and Cls Cle
thus monoclinic(P2,/b) symmetry results(Hounslow and 0.13 0.13
Chao, 1970;Mackie etal.,1972; Elliot et al., 1973).The
monoclinic ,ernary apatite describedhere differs from these
end-members in that all three column-anion speciesare
cln
accommodatedin a previously unrecognizedP2r/b vai- 0.06
ant. We presentthe generalfeaturesof the structure; def-
inition of finer details, such as disordering of the equiv-
alents ofhexagonal Ca(2), undoubtedly presentas indicated [0,0,2] 10,1'14,zl
by Ca thermal parameters and Ca(2)-Cl" bond lengths,
Fig. 3. Correlationof anion sitesin P6t/ m andP2'/ b ternary
await more suitable crystals.
apatites.Atomsrepresentpossibleanionsitesassociated with the
The positions of noncolumn atoms in P2r/b ternary mirror plane(P6r/m)or theb-glideplane(P2,/b)at z : Yr.Atom
apatite are similar to tho sein P6,/ m ternary apatite (Table size is arbitrary, and the z positional parameteris given. The
6). The reduction of symmetry from hexagonalto mon- shift of the O(H) position betweenstructuresis largelya result
oclinic in ternary apatite results from ordering ofthe col- of the differencein Cl content.
umn anions into one of the two symmetry-equivalentcol-
umn-anion sites present in hexagonal ternary apatite.
Figure 3 depicts the correlation between column-anion positions at the cenier of the Ca(2) triangles in the 2a
sites in the two ternary structures. special positions at (0, 0, th) and (0,0,3/+),related by the
In the monoclinic ternary apatite structure one position 6, screw axis. In monoclinic ternary apatite the F atoms
for Cl [Cl"] existsnear (0, 0.25, 0.43), which is equivalent occupy similar positions [hexagonalequivalents : (0, 0,
to the hexagonalposition of (0, 0, 0.43). The monoclinic 0.24) and (0, 0, 0.74) relatedby the 2' axis.
ternary structure thus accommodatesClo in a position In hexagonalternary apatite OH is displacedfrom the
similar to one of the symmetry-equivalent Clo sites in mirror plane and half-occupiesmirror-related sites at (0,
hexagonal ternary apatite (0, 0, 0.44)1.However, as in 0, 0.20)and (0, 0,0.30). The zparameterof OH, however,
monoclinic chlorapatite,equivalent positions are no long- is sensitive to Cl content. Sudarsananand Young (1978)
er in the same column and related by mirrors but occur demonstrated the covariance of zot and Cl content in
in neighboring columns and are related by b-glide planes. apatites.Using their relationship, OH in apatite with 0.47
A secondposition for chlorine (Clu)existsin monoclinic Cl atoms per formula unit (as in monoclinic ternary apa-
ternary apatite at (0.00,0.25,0.13), which is equivalent tite) would be locatednear (0, 0, 0.18) and (0, 0, 0.32).
to the hexagonalposition of (0, 0, 0.13). As shown pre- In the monoclinic structure OH, like the Cl atoms, is
viously, in hexagonal ternary apatite the majority of the ordered into one of these sitesat (0, 0.25, 0.34), which is
chlorine atoms are located in the mirror-related Cl" sites equivalent to (0, 0, 0.34) in hexagonalapatite.
at (0, 0, 0.37) and (0, 0, 0.13). In monoclinic ternary The reduction of symmetry in monoclinic ternary apa-
apatite only one ofthese two sitesis occupiedin any given tite results from ordering of OH and Cl into one of the
column. two equivalent hexagonal sites. It is not clear from this
In hexagonalternary fluorapatite the F atoms occupy initial investigation what proportion of natural ternary
302 HUGHESET AL.: CRYSTALSTRUCTURESOF TERNARY APATITES

.A
2.60
Ca(l) polyhedron are minimal; however, variations with-
.--r-------------o-----__._!1e_11__c_{_11:_91:__"_"_:11 jl*. in the Ca(2) polyhedron, which includes among its ligands
2.50
B.
the column anions, are regular and can be related to the
Cl content.
2.50 Ao
a- - > --' -' -' -- --'+------{----'
-------.-
Phosphate tetrahedra in both ternary structures have
Deatr Ca(2)-O: 5 oxygens
2.40
Bo
statistically identical mean P-O bond distances(1.532-
.c. 1.533 A). These values are identical, within +3 esd, to
those observed for end-member fluor-, hydroxyl-, and
A -----a"' chlorapatite (Hughes et al., 1989). Individual P-O bond
) 1<
;----""""""1s distancesand O-P-O anglesare essentiallyconstant with
ca(2)-o.)
2.65
the exceptionof thoseinvolving O(3) atoms,which exhibit
o very small but systematic variations correlative with Cl
o A.D
2.40 F tBn- content.
o-----------_____l
.t) oft
MUN
-_____-___- In both the ternary apatitesdiscussedherein and in end-
2.30 TEn_----r-----
HEX cl member phosphate apatites discussedby Hughes et al.
(1989),Ca(l) bonds to three O(l) and three O(2) at short
a 2.20 _ Cz(D-O(2)
distances(-2.3-2.4 A) and three O(3) at longer distances
E.
2.45
a--a-------------t------G'-------_-----------l
(-2.8 A). Differencesin interatomic distancesamong the
()
AB Ca(2)-O(3) five ternary and end-member structures are minor, al-
2.35
F. though there is a slight tendency for the mean Ca(l)-O
2.'.10 distancesto decrease( - 0.0 I A; with increasingCl content
a Gie.a).
2.60 Ca(2)-CI The Ca(2) polyhedron in phosphate apatites is an ir-
G. regularly shapedpolyhedron that is at least 6-coordinate
4 . 5 0 ._- - ' - - - - - - - - ' -
Trian8les with significant [one O(2), four O(3), and one X column anion]. The co-
bonding
. ----)---.ca-cl ordination number may be larger if Ca(2) bonds to an
4.30 ' O(l) atom or to an additional Cl atom. In hexagonalend-
, Ttl"nglo witb oo or
minor Ca-Cl bonding o
member apatites, Ihe Ca(2) in the F and OH speciesis
4 1 0 ' l a{ * - - - - -
,,-
7-coordinate if a long bond to O(1) (ca. 2.7 A) is included
Ca(2)_Ca<2,
3.90
_ _.__ _ _ _ - - _. - _ _._ in the coordination sphere,whereasin chlorapatite it is
S-coordinateif both O(l) (ca. 2.9 A) and a weak bond to
a secondCl atom are included. In the ternary hexagonal
VoCl apatite reported in this study, Ca(2)is disorderedinto two
Fig. 4. Selected interatomic distances in natural F, Cl, OH sites, as discussedpreviously. The disordered Ca(2)opo-
ternary and end-member apatites. Ternary data (TER-HEX and sition is similar to that recognized in the end-member
TER-MOI{) are reported in this paper; end-member data (Cl, F, apatite structures,with Ca(2)" bonding to F, OH, and Cl^
OH) are reported by Hughes et al. (1989). The data for ternary at distancesof -2.29,2.32,and2.63 A,respectively.The
hexagonal apatite represented by triangles are interatomic dis-
atom in the Ca(2)" position bonds to Clr, a unique Cl
tancesfor the averageCa(2) atom prior to refinement with a split-
position recognizedin ternary and Cl-bearingbinary apa-
atorn model; "A" and "B" in B, C, D, and E refer to split Ca(2)
atoms. Dotted line at 2-7 A in F included for reference; dashed tites. A similar disordering of Ca(2) is expected in the
line in G taken from Sudarsanan41d fsrrng (1978). Seetext for ternary monoclinic apatite based on the magnitude of the
further discussion. Size of symbols is generally larger than + I thermal parameters(Table 6), but the split atom was not
esd for each data point. modeled becauseof the lower quality of the data collected
from the inferior monoclinic crystal.
Interatomic distancesbetween Ca(2) and its ligands vary
apatites possessesmonoclinic symmetry, nor which phys- regularly with Cl content in the five apatites we have
ical or chemical conditions may result in monoclinicity studied (this study and Hughes et al., 1989).The average
in ternary (Cl, OH, F) species. The structures presented of the four Ca(2)-O(3)bonds remains approximately con-
here, however, illustrate the unaddressed complexities in stant at 2.42-2.43 A Gig. 4) in all of the structures.With
apatite crystal chemistry in that two crystals of similar increasingCl content, the Ca(2)-O(2)bond decreasessig-
composition differ in symmetry as a result of column- nificantly, whereasthe Ca(2)-O(l) distance increasessig-
anion ordering. nificantly.
In a previous study of end-member hexagonalapatites,
ClrroN PoLYHEDRA Hugheset al. (1989) interpreted the increasein the Ca(2)-
The topological variations exhibited by cation poly- O(l) distancein chlorapatite (vs. fluor- and hydroxylapa-
hedra in both ternary apatites result from perturbations tite) as a mechanism to reduceoverbonding of Ca(2) that
introduced by Cl atoms in the column-anion sites. Effects results from its coordination to a secondCl atom in the
on interatomic distances in the P tetrahedron and the [001] column. In the ternary structures,it appearsthat the
 HUGHESET AL.: CRYSTALSTRUCTURESOF TERNARY APATITES 303

Ca(2)-O(2)bonds may also contribute to the Ca(2) over-

bonding, as a significant decreasein Ca(2)-O(2) bond Rnrnnrxcns crrED
length occurs with increasingCl content. Beevers,C.A., and Mclntyre, D.B. (1946) The atomic structure of fluor-
Refined Ca(2)-Cl distances in both ternary and end- apatite and its relation to that oftooth and bone material. Mineralogical
member structuresare closeto the ideal value of -2.7 A. Magazine,27, 25+257.
The short Ca(2)-Cl, distance(-2.54 A) reported in Table Boudreau,A.E., and McCallum, I.S. (1987) Halogensin apatitesfrom the
Stillwater Complex: Preliminary results (abs.).EOS, 68, 1518.
7 for the monoclinic apatite is an artifact ofthe refinement
Candela,P.A. (1986) Towards a thermodynamic model for the halogens
procedure,as we did not model the Ca(2) split atoms. We in magmatic systems: An application to the meh-vapor-apatite equi-
expectthat if Ca(2)were allowed to disorder in split-atom libria. Chemical Geology, 57, 289-3Ol.
positions, readjustmentsin both Ca-Cl distances[Ca(2)o- Crowley, K.D., and Cameron, M. (1987) Annealing of etchable fission-
Cl"; Ca(2)"-Clul would occur, and both would approach track damage in apatite: Effects of anion chemistry. Geological Society
of Arnerica Abstracts with Programs, 19, 6t l-632.
the ideal bond distanceof -2.7 A. Crowley, K.D., C-ameron,M., and McPherson, B.J. (1988) Annealing of
The Ca(2) atoms are arrangedin triangles that are cen- etchablefission-track damagein F-, OH-, Cl-, and Sr-apatite: l. Sys-
tered around [001] containing the column anions. The tematics and preliminary interpretations. Proceedings of 6th Intema-
size of these triangles can be related to the amount of Cl tional Fission Track Dating Conference, Besangon,France.
Elliot, J.C., Mackie, P.E., and Young, R.A. (1973) Monoclinic hydroxy-
in an apatite structure and to the presenceor absenceof
apatite.Science,180, 1055-1057.
Ca-Cl bonds in a given Ca(2)triangle.The Ca(2)ntriangles Exley,R.A., and Smith, J.v. (1982)Therole ofapatite in mantleenrichment
bonded to the column anions OH, F, and Clo exist in an processesand in the petrogenesisofsome alkali basaltsuites.Geochim-
essentiallyCl-free environment and exhibit similar Ca(2)- ica et CosmochimicaActa, 46, 1375-13E4.
Frenz, B.A. (1985) Enraf-Nonius structure determination package.SDP
Ca(2) distances(-3.97-4.07 A; Fig. 4). In contrast,the
Users Guide (Version 4), Enraf-Nonius, Delft, The Netherlands.
Ca(2)utriangles in ternary apatite that involve significant Green, P.F., Duddy, I.R., Gleadow, A.J.W., Tingate, P.R., and Laslett,
Ca-Cl bonding [e.g., Ca(2)"-Cl"] exhibit a very regular G.M. (1986) Thermal annealingof fission tracks in apatite: l. A qual-
inverse relationship between Ca(2)-Ca(2)distanceand Cl itative description. Chemical Geology, 59, 237-253.
content (Fig. a). The latter relationship was first recog- Harrison, T.M., and Watson, E.B. (1984) The behavior of apatite dwing
crustal anatexis: Equilibrium and kinetic considerations.Geochimica
nizedby Sudarsananand Young (1978)and wasattributed
et Cosmochimica Acta. 48. 1467-1477.
by those authors to Ca-Cl bonding requirements. Henderson, P. (1984) Rare earth element geochemistry.Elsevier, Am-
sterdam.
Iivrpr,rc^lrroNs FoR BTNARyCr,-F lpLrrrrs Hounslow, A.W., and Chao, G.Y. (1970) Monoclinic chlorapatite from
Solid solution in ternary apatite is made possible by a Ontario. Canadian Mineralogist, lO, 252-259.
Hughes,J.M., Cameron, M., and Crowley, ICD. (1989) Structural varia-
0.4 A shift of the Cl atom relative to its position in end- tions in nalural F, OH, and Cl apatites. American Mineralogist, 74,
member chlorapatite.The movement of the Cl 0.4 A clos- 870-876.
er to the plane of Ca(2) atoms enablesa OH neighbor to Mackie, P.E., Elliot, J.C., and Young, R.A. (1972) Monoclinic structure
occupy a position at the next column anion site, at a of synthetic Ca,@On),CI,clrlorapatite. Acta Crystallographica,B28, 1840-
I 848.
distanceof 2.96 A. fnis configuration,however, doesnot Mackie, P.E., and Young, R.A. (1974) Fluorine-chlorine interaction in
permit a F ion as a neighbor to Cl in the anion column. fluor-chlorapatite.Joumal of Solid State Chemistry, ll, 319-329.
It is thus not clear from the ternary structure how solid Mehmel, M. (1930) Uber die Struktur des Apatits: I. Zeitschrift fiir lkis-
solution results in fluor-chlorapatite. tallographie,75, 321-331.
Morrison, J., and Valley, J.W. (1989) Partitioning of F-CI-OH between
Using synthetic crystals,Mackie and Young (1974) ad,-
amphibole, apatite, and biotite in granulites from the Adirondack
dressedthe incompatibility of F and Cl in apatite anion Mountains, New York (abs.).EOS, 70, 493.
columns. They concluded that solid solution in pure Naray-Szab6,S. (1930)The structureofapatite (CaF)Ca,(POo),. Zeitschrift
F-Cl apatites is made possible by creation of new F sites Iiir Kristallographie, 75, 387-398.
in the anion column, which yields reasonableCl-F dis- Nielson, D.R. (1988) Depositional environments and petrology of the
Middle Jurassic Carmel Formation near Gunlock, Washington County,
tances.In preliminary refinementsof natural, near-binary
Utah, 108 p. M.S. thesis,Brigham Young University.
chlor-fluorapatite, we have found that small amounts of Stormer, J.C., and Carmichael,I.S.E. (1971) Fluorine-hy&oxyl exchange
OH stabilize F in the 2a specialposition, and solid so- in apatite and biorite: A potential igneous geothermometer. Contribu-
Iution is made possible by splitting of the chlorine sites. tions to Mineralogy and Petrology, 3l, l2l-131.
The concentration of OH ne@ssaryto stabilize this Cl-F Sudarsanan,tr(., and Young, R.A. (1978) Structural interactionsofF, Cl,
and OH in apatites.Acta Crystallographica,B34, 140l-1407 .
structure is not known; however, we are continuing ex- Sudarsanan,K., Young, R.A., and Wilson, A.J.C. (1977) The structures
periments to elucidatethe nature of solid solution in nat- of some cadmium "apatites" Cdr(MOo)rX: L Determinations of the
ural fl uor-chlorapatites. structuresof Cdr(VOo)rI,Cdr(POn)rBr,Cdr(AsOo).Brand CddVO),Br.
Acta Crysrallographica, B33, 3 136-31 42.
AcxNowr,BocMENTs Walker, N., and Stuart, D. (1983) An empirical method for correcting
diffractometer data for absorption effects. Acta Crystallographica, A39,
We acknowledgeBart Kowallis of Brigham Young University and Georye l 58-166.
Y. Chao of Carleton University for providing specimens for this study. Watson, E.B., Harrison, T.M., and Ryerson,F.J. (1985) Diftsion of Sm,
Paul Robinson of Southern Illinois University kindly duplicated our ex- Sr, and Pb in fluorapatite. Geochimica et Cosmochimica Actz, 49,
perimental results for hexagonal apatite using a Rigaku ercss diftactom- l8 l3-1823.
eter. The National Science Foundation provided support for this project Weber,W.J., Wald, J.W., andMatske, Hj. (19E5)Self-radiationinactinide
through grantsEAR-8717655and CHE-84 18897(J.M.H.), EAR-85 I 762 I host phases of nuclear waste forms. Materials Research Society Sym-
(K.D.C. and M.C.), and EAR-8916305(J.M.H. and M.C.). posium Proceedings,44, 679-686.
304 HUGHES ET AL.: CRYSTAL STRUCTURESOF TERNARY APATITES

Weber, W.J., and Roberts, F.P. (1983) A review of radiation effectsin Yardley, B.W.D. (1985)Apatite composition and the fugacitiesof HF and
solid nuclear waste forms. Nuclear Technology,60, 178-19E. HCI in metamorphic fluids. Mineralogical Magazine, 49, 77-79.
Wilson, A.J.Q. Sudarsanan,trL, and Young, R.A. (1977) The structures
of some cadmium "apa.tites" Cdr(MO.)rX: II. The distributions of the
halogen atoms in Cd'(VO4)!I, C4(PO"),8r, C<t,(AsOn),8r, Cd,(VOo),Br MeNuscnrpt REcETvEDAucusr 2l', 1989
and Cd,(PO),CI. Acta Crystallographica,B33, 3142-3154. Mem;scnrpt AccEprEDNoverr.rneR28, 1989