American Mineralogist, Volume 74, pages870-876, 1989

Structural variations in natural Fo OH, and Cl apatites

JonN M. Hucnrs. Manvnr-r.nx CevrcnoN. KnvrN D. Cnowr-Bv

Department of Geology, Miami University, Oxford, Ohio 45056, U.S.A.

Ansrnl,cr
The crystal structuresof the hexagonal(P6r/m) end-member apatites with the formula
Car(POo)rX(X : F, OH, Cl) were refinedto R : 0.025,0.016,and 0.020 for fluorapatite,
hydroxylapatite, and chlorapatite, respectively.In accord with earlier studies,the F atoms
in fluorapatite lie in the mirror plane in the 2a (0,0,t/+)specialposition. In hydroxylapatite,
the OH speciesis disorderedin (0,0,2) positions 0.35 A above and below the mirror plane,
and in chlorapatite the Cl is also disordered, in positions 1.2 A above and below the
mirror planes. In chlorapatite, the Cl anion is so far displacedfrom the mirror plane that
an additional, weak bond developsbetween Ca(2) and a secondCl anion, thus increasing
the Ca(2) coordination.
In the three end-members, diferences in position of the column anions propagate
throughout the structure, but have minor, secondaryeffects on the Ca(l)O, polyhedron
and the POo tetrahedron in terms of cation-oxygen bond lengths and polyhedron orien-
tation; averageP{ and averageCa(lFO bond lengthsfor the three structuresare identical
within 0.005 A. The Ca(2)O,X(O) polyhedron, however, is greatly affectedby anion sub-
stitution. Individual Ca(2FX bond lengths are significantly different for the three struc-
tures, with the Ca(2FX bond length varying between2.311 and2.759 4,. The mean Ca(21
O distancerangesbetween2.461 and 2.4% A.
The atomic arrangementsof fluorapatite, hydroxylapatite, and chlorapatite suggestthat
these end-membersare immiscible in solid solution. The 1.2-A displacementof Ct from
the mirror plane in chlorapatite prohibits the existenceof F and OH as immediate neigh-
bors at certain sites in the anion columns becauseof prohibitively short interatomic dis-
tances.Structural adjustments that enable solid solution to occur in Cl-bearing binary or
ternary apatitesmay include (l) major shifts in column-anion positions compared to those
in end-member structures, (2) reduction of symmetry from hexagonalto monoclinic in
ternary apatites, and/or (3) ordering of anions within individual columns but disordering
of columns throughout the apatite structure.

of the end-members were published in the early 1970s,

INrnonucrroN i.e., P2,/b syntheticchlorapatiteby Mackie et al. (1972),
Natural apatites of composition Car(POo)r(F,OH,Cl) P2,/b synthetic hydroxylapatite by Elliott et aI. (1973),
exhibit large variations in F, Cl, and OH contents. Pure and P6r/m synthetic and natural fluorapatite by Sudar-
end-members are uncommon in nature, but binary and sanan et al. (1972).
ternary compositions are widely reported in igneous, In this paper,we report three new structurerefinements
metamorphic, and sedimentary rocks. Petrologistshave on natural P6./m nearend-member fluorapatite, chlor-
proposedthat the F-CI-OH variations in apatitesare po- apatite,and hydroxylapatites(binary and ternary apatites
tential geothermometers(e.g., Stormer and Carmichael, will be describedin a subsequentpaper).Thesestructures
l97l) and indicators of volatile fugacities in magmatic differ from those of their synthetic counterparts,perhaps
and hydrothermal processes(e.9., Candela, 1986; Bou- as a result of differencesin anion ordering causedby im-
dreau and McCallum, 1987). Other recent studies have purities and vacancies.The structural differencesbetween
shown that F, OH, and Cl concentrations correlate di- synthetic and natural crystals,and betweenend-member
rectly with properties such as etching rates and annealing and binary or ternary species,warrant careful consider-
characteristicsof U fission tracks in apatite (e.g., Green ation where apatites are used in thermodynamic, diffu-
et al., 1986;Crowley and Cameron, 1987). sion, fission-track,and partitioning studies.
Despite the ubiquitous nature of apatite minerals and
their potential usefulnessas petrogeneticindicators, rel- ExpnnrvrnNTAl- DETATLS
atively few structure studies of natural crystalshave been Near--end-memberspecimensfrom severalwell-docu-
undertaken. The P6r/m fluorapatite structure was de- mented localities were chosenfor study. The specimenof
scribed more than 50 years ago (N6ray-Szab6,1930; fluorapatiteisfromDurango,Mexico(Youngetal.,1969).
Mehmel, 1930),and the most recent modern refinements Electron-microprobe analysis and instrumental neutron
0003404x/89/07084870$02.00 870
Downloaded from https://pubs.geoscienceworld.org/msa/ammin/article-pdf/74/7-8/870/4221998/am74_870.pdf
by Yonsei University user
 HUGHESET AL.: STRUCTURALVARIATIONSIN APATITES 871

TABLE refinements
1. Crystaldataandresultsof crystal-structure andchlorapatite
hydroxylapatite,
for fluorapatite,

Crystalsize (mm) 0.08x0.13x0-16 0.25x0.20x0.10 0 35 x 0.25 x 0.20

Unit-celldimensions
Least squares
a (A) 9.398(3) I 418(2) 9.598(2)
b (A) 9.397(3) 9.416(2) 9.598(2)
c (A) 6.878(2) 6 875(2) 6.776(41
a (") 89.99(2) 90.01(3) 89.97(3)
6f) 90.02(2) 89.es(2) s0.02(3)
r (') 120.06(2) 11I 94(2) 119.96(2)
Structure refinement
9.3973 9.4166 9.5979
6.8782 6.8745 6.7762
Reflections 11 9 7 1201 1243
q7q 383 eo7
Unique reflections
R.",n" 0.014 0.012 0.010
Retlections/ > 30/ 328 344 354
H 0.025 0 016 0.020
B. 0 029 0.022 0.028
Ap residua (e A-3)
+ 0.388 0.333 0.326
o.745 0.332 0.374

Note. Numbers in parenthesesdenote one esd ot the least-significantdigit.

activation analysis(nlee) yielded a composition of (Cao.ro- reflection. One-third of the total counting time was spent
Feoo,Sroo,Ceoor)* ro(PrruSlo
orSoor)r,o,O,r(ForoCbor)r,or. Hy- determining background on both sides of the peak.
droxylapatite is representedby a specimen from Holly Data were reduced and corrected for Lorentz and po-
Springs, Georgia (National Museum of Natural History larization effects.Absorption effectswere correctedusing
no. R-9498); analysisof this specimenby electron micro- an empirical psi-scantechnique,using data obtained from
probe and rNAu{gave a composition of Ca4e8P2erO,r(Foou-360'psi scans.Equivalent reflections were then merged;
Cloo3oHoer)"r oo.The OH content was determined by dif- reflectionswith 1 < 3o, were consideredunobserved(Ta-
ference.The chlorapatite specimenis from Kragero, Nor- ble l).
way (American Museum of Natural History no. 23101), Crystal-structurecalculationswere undertakenin space
and yielded a compositionof (CaorrFeoo,NaoorCeoo,)*rr-group P6./m using the SDP-Plus set of programs (Frenz,
P3oror2(Fooeclo rr). r, by electron-microprobeand rNen 1985). Full-matrix least-squares refinementwas under-
analysis. taken by refining positional parameters,scalefactor, an-
Crystals were examined under plane-polarizedlight to isotropic temperaturefactors, and an isotropic extinction
insure that each specimenselectedwas an untwinned sin- factor. Weights proportionalto or2 were assignedto in-
gle crystal. Preliminary precession studies were under- dividual structure factors with a term to downweight the
taken, using a, b, and [1 10] axis photographs,to confirm contribution of intense reflections. Neutral-atom scatter-
that monoclinic superstructurereflectionswere not pres- ing factors,including terms for anomalousdispersion,were
ent and to verify the quality of X-ray diffraction maxima. used in the refinement.
X-ray data for all crystals were measuredusing an En- For fluorapatite, starting parameters were taken from
raf-Nonius cep+ single-crystaldiffractometer and graph- Sudarsanan(1972);occupancyofthe halogenswas mod-
ite-monochromated MoKa radiation. Unit-cell parame- eled with 95o/oF,located on the mirror plane at (0,0,1/+),
ters were refined (no symmetry constraints) using and 50/oCl, constrainedto lie at the Cl position obtained
diffraction anglesfrom 25 automatically centeredreflec- in the chlorapatite structure refinement. For hydroxyl-
tions. Refined parameters and the idealized hexagonal apatite, refinement was initiated using the parametersin
parameters(idealized as per Frenz, 1985) used for struc- the fluorapatite study, although some symmetry con-
ture refinement are given in Table l. straints were released.It was assumedthat the OH ions
t/+
Data were collectedfrom a quadrant ofreciprocal space were disordered along c about the mirror plane at z :
to 52 20. Intensities were measured using a 0/20 scan and that the hydroxyl O and H atoms were located in a
technique, with scan widths determined by the relation- half-occupied (0,0,2) position, necessaryin a P6'/m apa-
ship c.,: 1.00" + 0.34 tan d, to correct for dispersionof tite. The H positional parameter was refined, but the an-
the MoKa spectrum.Regularly interspersedintensity and isotropic temperature factors were fixed at those derived
orientation standards were monitored throughout data from the neutron-diffraction data ofKay et al. (1964). The
collection. A prescanofeach peak during data collection crystal-structure refinement of chlorapatite was under-
determined counting time, with a maximum of 90 s per taken using starting parametersfrom Mackie etal. (1972);

Downloaded from https://pubs.geoscienceworld.org/msa/ammin/article-pdf/74/7-8/870/4221998/am74_870.pdf

by Yonsei University user
 872 HUGHESET AL.: STRUCTURALVARIATIONSIN APATITES

TABLE2, Positional parameters and equivalent isotropic tem- halogen occupancy was modeled with 0.86 atoms of Cl
perature factors (Ar) for end-memberapatites disorderedin a (0,0,2) position and 0. I I F atoms fixed at
y z B(A,) (0,0,Yr),in accord with the microprobe data.
Data-collection procedures and the results of the re-
Ca(1)
F % h 0.0010(1) 0.91(1)
finements are summarized in Table l. Table 2 contains
OH 43 h 0.00144(8) 0.929(7) positional parametersand equivalent isotropic B values
cl 43 % 0.0027(1) 0.99(1) for the end-member apatites,and Table 3 records aniso-
Ca(2) tropic thermal parameters for atoms in the three struc-
T -o.0o712(7) o.24227(7) Va 0.77(1)
OH -0.00657(5) o.24706(5) V4 0.8ss(e) tures. Table 4 contains bond lengths and angles for the
cl 0.00112(6) 0.2s763(6) V4 1.14(1) three polyhedra of the end-member structures. In this
P case,the term "bond angle" is defined as the angle that
F 0.36895(8) 0.39850(8) V4 0.57(1) each cation-anion bond makes with each of the unit cell
OH 0.36860(6) 0.39866(6) '/4
0.62(1)
cl 0.37359(7) (7)
0.40581 th 0 77(1) axes.This usageis meant to facilitate comparison among
o(1) the structuresby providing a common referencesystem.
F 0.4849(2) 0.3273(3) 0.ee(4) Table 5t contains observed and calculated structure fac-
OH 0.48s0(2) 0.3289(2) V4 1.00(3) tors for the three structures.
cl 0.4902(2) 0.3403(2) V4 1.34(4)
o(2) Srnucrun-c,L vARTATToNSIN FLUoRApATITE.
F 0.4667(2) 0.587s(3) V4 1.19(5)
OH ITYDROXYLAPATITE, AND CHLORAPATITE
0.4649(2) 0.5871(2) V4 1.25(3)
cl 0.4654(2) 0.s908(2) Y4 1.47(4) A detailed description of the P6r/m apatites was pro-
o(3) vided by Beeversand Mclntyre (1946). For the purposes
F o.2575(2) 0.3421(2) 0.070s(2) 1.32(3)
OH 0.2580(1 0.3435(1) 0.0703(2) 1.s7(2) of this paper, we regard the structure as being made up
)
cl 0.2655(2) 0.3522(21 0.0684(3) 1.88(3) of Ca(l) and Ca(2) coordination polyhedra hexagonally
disposedabout a central [001] hexad (Fie. l). The F, Cl,
FO 0 V4 1.93(6) and OH ions lie in columns on these hexads at (0,0,2)
o(H) 0 0 0.1979(6) 1.31(8)
HO 0 0.04(2) 3.3
cr 0 0 0.4323(4) 2.68(5) ' A copy of Table 5 may be ordered as Document AM-89-412
Note-'Numbers in parenthesesdenote one esd of the least-significant
from the BusinessOfrce, Mineralogical Societyof America,1625
digit. I Street,N.W., Washington,D.C. 20006, U.S.A. Pleaseremit
$5.00 in advancefor the microfiche.

TABLE3, Anisotropic thermal parameters (A,) for atoms in end-memberapatites

P33 0.2
ca(l)
F 0.00394(6). 0.00394 0.0033(1
) 0.00394 0 0
OH 0.00408(5) 0 00408 0.0033(1
) 0.00408 0 0
UI 0.00462(6) 0.00462 0.0023(2) 0.00462 0 0
Ca(2)
F 0.002s6(6) (6)
0.00361 0.0034(1
) 0.00299(9) 0 0
OH 0.00308(5) 0.00353(5) 7(8)
0.0041 0.0031
2(7) 0 0
cl 0.00449(6) 0.00526(6) 0.0037(1
) 0.004s6(8) 0 0
P
T 0.00208(7) 0.00231(7) 0.0032(1
) 0.002s(1
) 0 0
OH 0.00223(5) 0.00250(6) 0.0034(1
) 0.00254(8) 0 0
cl 0.0031
0(6) 0.00330(6) 0.0034(1
) 0.00368(9) 0 0
o(1)
T 0.0037(2) o.oo44(2) 0.0055(4) 0.0053(3) 0 0
OH 0.0036(2) o.0044(2) 0.0059(3) 0.0054(2) 0 0
cl 0.00s9(2) 0.0075(2) 0.0054(4) 0.0102(3) 0 0
o(2)
F 0.0038(2) 0.0030(2) 0.0088(5) 0 0028(3) 0 0
OH 0 0035(2) (2)
0.0031 0.0105(3) 0.0032(2) 0 0
UI 0 0049(2) 0.0036(2) 0.0109(5) 0.0040(3) 0 0
o(3)
F 0.0043(2) 0 0077(21 0.0055(3) 0.0076(2) -0.0032(4) -0.0052(5)
OH 0.0050(1
) 0.0096(1
) 0 0070(2) 0.0101(2) -0.0050(3) -0.0082(3)
cl 0.0062(1
) 0.0113(2) 0.0076(3) 0.0119(2) -0.0070(4) -0.0109(5)
x
F 0.0029(3) 0.0029 0.022(1
) 0.0029 0 n
o(H) 0.0030(3) 0.0030 0.01
2(1) 0.0030 0 n
cl 0.0056(1
) 0.0056 0.027(1 0.0056 0 n
)
Note.'The form of the anisotropic temperaturetactor is exp - llrrh" I pek2 + PsP + Pphk + P$hl + Pakn. Numbers in parenthesesdenote one
esd of the least-significantdigit.

Downloaded from https://pubs.geoscienceworld.org/msa/ammin/article-pdf/74/7-8/870/4221998/am74_870.pdf

by Yonsei University user
 HUGHES ET AL.: STRUCTURAL VARIATIONS IN APATITES 873

positions. The Ca(1)polyhedron is nine-coordinated,u'ith TnaLe4. Bond lengths (A) and angles (") with unit-cellaxes for
six shorterbondsthat definean approximatetrigonal prism cation polyhedrain end-memberapatites
and three longer Ca(lfO(l) bonds that emergethrough
the prism faces.The Ca(l)Onpolyhedrashare(001) pin- A
1.534(2) 1.533(3)
A
Po-o(1)o A
1.537(3)
acoid facesto form chains parallel to c. The Ca(2) poly- a-P"-O(1)A 22.14 2178 20.80'
hedron can be consideredto be a CaOrX octahedron,but b-Po-o(1)o 142.14 14't.78 140.80.
90' orp 9tr
a seventhweak bond to O(l) (0.15 valenceunits) exists o-PA-O(1)a
PA-O(2r A
1.s3e(2) A
1.537(2) A
1.538(2)
in fluorapatite and hydroxylapatite, redefining the poly- 90.25'
a-Po-O(2)o 88.86' 89.26'
hedronasCaO,X(O).The Ca(2)atomswithin mirror planes b-P^-O(2)o 31.14. 30.74 29.75
at z :0.25 and 0.75 form trianglesthat are centered o-PA-O(2)a 90. 90. 9e
around the [001] zone axescontainingthe X anions.Ad- Po-O(3)o' 1.532(2)
A A
1.529(1) 1.524e)A
120.71" 't20.79 120.81'
jacent Ca(l) and Ca(2) polyhedra are linked through oxy- a-Po-O(3)o
D-Po-O(3)o 90.24 89.97' 89.84'
gen atoms sharedwith POl tetrahedra. o-Pn-O(3)o 143.70 143.87" 143.86'
Mean 1.535A 1.532A 1.530A
Cation polyhedra
Ca(1 )o-O(1 )ABc 2.39s(2)A 2.404{0lA 2A07e)A
Changesinduced in the natural apatites as a result of 134.44 134.68' 135.86.
a-Ca(1f-o(1f
differencesin the column anion propagatethroughout the b-Ca(1)A-O(1)A 70.58' 70.21" 67.68.
crystal structure. The effects on the Ca(l)O, polyhedra c-Ca(1)A-O(1)a 44.46 44.69 45.89'
and the PO* tetrahedra are minor, whereasthose on the ca(1)a-o(2)o EF 2.457e)A 2.452e)A 2.448e1A
a'Ca(1)^-O(2)" 131.42' 131.16' 131.81'
Ca(2)OrX(O) polyhedron are significant. b-Ca(1)a-O(2)D 56.09" 56.06' 56.19'
Phosphatetetrahedra for the three structuresare com- o-Ca(1)^-O(2)o 13463. 134.81' 134.38.
pared in the onrep (Johnson,1965)plot ofFigure 2. As ''
Ca111^-9,r,o 2807e)A A
2.802(1) A
2.7e3(3)
seenin that figureand Table 4, differencesin anion content a-Ca(1)A-O(3)o 106.83' 106.86' 10790.
b-Ca(1)^-O(3)" 16.32" 16.32. 15.39'
effect only minor changesin both interatomic distances 100.09. 100.13' 99.93'
o-Ca(1)^-O(3)"
and anglesofthe tetrahedra.Setsofanalogousbond lengths Mean 2.s54A 2.5s3A 2.549A
for the three structuresdiffer by a maximum of 0.008 A.
Analogous O-P-O anglesexhibit variations on the order Ca(2)A-O(1)B 2]01e) A 2.711{[)A A
2.e01(2)
a-Ca(2)^-O(1)" 165.21' 1 6 4 .11' 162.03"
of 0. 1-1.1". Among the three structures,averagebond b-Ca(2)A-O(1)B 74.79 75.89. 77.97"
lengths vary within a ran3eof only 0.005 A. Differences c-Ca(2)^-O(1)B 90' 90 90.
of approximately l'in the angulardispositionof the P- Ca(2)A-O(2)c 2.374e)A 2.353(1) A A
2.306(2)
O(l) and P-O(2) bonds with respectto the 4 and b axes a-Ca(2)^-O(2)c 94.00' 93.49" 93.39'
b-Ca(2)^-O(2)c 26.01. 26.51" 26.61"
(Table 4) may reflect the slight displacementof Ca(2) in- a-Ca(2)^-O(2)c 9tr 90" 90'
duced by changesin position of the column anion. A
2.349(2) 2.343(1)A A
2.331(2)
Ca(2)^-O(3)E'
The Ca(l)On polyhedra in the end-member apatite a-Ca(2)A-O(3)F 109.91' 109.82" 112.19.
structuresalso show little responseto anion substitution, b-Ca(2)^-O(3)' 7753 78.32. 77.69.
o-Ca(2)a-O(3)F 159.84' 160.05' 157.74'
as documentedin Figure 2 and Table 4. The largestrange
Ca121o-9131or A
2.501(2) 2.509(1)A 2.s44e)A
in individual Ca-O bonds involves Ca(lFO(3), which 36.23 35.70' 35 02'
a-Ca(2)^-O(3)^
exhibitsa variation of 0.014A amongthe threestructures. b-Ca(2)^-O(3)^ 97.01. 97.74 98.15'
Averagebond lengthsfor the Ca(l) polyhedra ofthe three o-Ca(2)^-O(3)^ 11958' 119.49' 118.92.
apatitesdiffer by only 0.005 A. Small (- l-3) deviations Ca(2)a-X 2.31089)A 2.3851(4A A
2.759(1)
a-Ca(2)A-XA 58.57' EO noo 63.63.
in orientations of bonds with respectto the crystallograph- 178.57. 17127" 153.40'
b-Ca(2)A-X^
ic axesresult from small differencesin the z parameterof o-Ca(2)A-XA 90" 81.37' 63.40.
Ca(l) in the three structures. Mean 2.463A 2.461A 2.493A
The major structural responseto anion substitution in -x, ziD:
: : x , y , z ( T a b l e 2 l ; B : - y , x - Y ,z i C : y - x ,
N o t eA
the natural end-member apatites studied occurs in the - x , - y , - z t E : y , y - x , - z ; F : x - y , x , - z ; G = - x , - y , V 2+
-!'
Ca(2)O,X(O) polyhedron (Fig. 2 and Table 4). Large dif- z ; H : y , y - x , y 2 + z i l : x - y , x , 1 h+ z i J : x , y , y 2 - z ; K :
x - y , V 2 - z ; L : y - x , - x , V 2- z .
ferencesin position ofthe column anion (F, Cl, OH species)
are especially evident in Figure 2. Individual Ca(2lX
bond lengthsare significantly different for the three struc-
tures,with the Ca(2FX bond lengthvarying between2.31I hexadsand mirror planesat z : 0.25 and 0.75. Each F is
and 2.759 A gaUte 4). The Ca(2FO(1) bond length is bonded to three Ca(2) atoms that form a triangle within
similar in fluorapatite and hydroxylapatite (2.70 and 2.7 | the mirror plane. The OH anions are disordered 0.35 A
A, respectively),but is 0.20 A longer in chlorapatite. above or below the mirror planes, and Cl ions are dis-
ordered 1.2 A above or below the planes.The disordering
of Cl and OH from the 2a special position reflects the
Geometry of the column-anionpositions structural accommodation required between the column
The column anions in the three natural apatites ex- anion and the triangle of Ca(2) atoms. The Ca(2YCa(2)
amined in this study lie on [001] hexads.The F anions distances(and areas of the correspondingtriangle) vary
lie in the 2a special position at the intersection of the from4.002A rc.gqA,)to +.oa+A Q.zz\"11s+.213A

Downloaded from https://pubs.geoscienceworld.org/msa/ammin/article-pdf/74/7-8/870/4221998/am74_870.pdf

by Yonsei University user
 874 HUGHESET AL.: STRUCTURALVARIATIONSIN APATITES

C a ( 2 )l z

.li'3,,,9j.1{
'.1 o!.t]..'.Q(ot2) //-o
(3l
,..::j';i{\{-\\
."'o(3 \,/ o(1.)
)(3) i
o(1);,

Fig. l. Schematicdepiction ofa portion offour unit cells ofthe apatite structure projected onto the (001) plane (after Braggand
Claringbull, 1965). Solid, dashed,and dotted lines indicate bonds in the different polyhedra.

(7.91 A) for fluorapatite, hydroxylapatite, and chlor-
apatite, respectively.
Becauseof its larger size and the longer Ca-Cl bond
Iength as compared to Ca-OH and Ca-F, the Cl anions
in chlorapatite are displacedfrom the (0,0,%)specialpo-
sition on the mirror plane to two equivalent half-occupied
positions at (0,0,0.4323)and (0,0,0.0677).It has been
shown (Mackie et al., 1972)that in pure chlorapatite, the
Cl anions are all ordered above the plane in a given col-
umn and below the plane in adjacent [001] columns along
D.This ordering yields monoclinic P2r/b chlorapatitewith
b^ono: 2 x bn"; the monoclinic Cl end-memberhasbeen
foundin nature,asreportedby Hounslowand Chao(1970).
In the near-end-memberhexagonalapatite refined in this
study, the column anions are not ordered becauseno su-
perstructurereflectionswere observedon precessionpho-
tos. It is possible,however,that clustersofCl anionsare
ordered above or below the plane in a given column, but
the senseofordering can be reversedby F ion "impurities"
or anion vacancies,as discussedby Elliott et al. (1973) and
Hounslow (1968). Thus the structure may contain regions
where the column anions are locally ordered, but overall
the averageapatite structure has P6r/m symmetry (Fig.
3). In chlorapatite, the Cl anion is displaced so far from
the mirror plane (1.2 A; ttrat a weak bond (0.09 valence
units) forms between Ca(2) and a second Cl anion, Cl',
located one-half unit cell away along c [Ca(2)-Cl' bond
distance : 3.27 A1. fne shght overbonding of Ca(2) be-
causeof this weakCa(2)-Cl' interactionis balancedthrough
reduced bonding between Ca(2) and,O(l) in chlorapatite.
Kay et al. (1964) refined the crystal structureof hydrox-
ylapatite using three-dimensionalX-ray (R : 0.075) and
neutron (R : 0.055) data. They determined that the OH
anions are disordered, half-occupying equivalent (0,0,2)
positions 0.35 A above and below the mirror planes.It
has been suggested(Elliott et al., 1973) that if insufficient
F anionsor vacancies(<0.15 per site?)exist in the anion
columns, OH anions will order above the plane in a given
column and below the plane in the adjacentcolumn along
b, again giving rise to monoclinic P2,/b symmetry with
b*""" : 2 x bn,. In our diffraction experiments, long-
exposureprecessionphotographsand diffractometerscans
showed no evidence of the monoclinic superstructure.
The arrangementof hydroxyls in the [001] columns of
the hydroxylapatite structure is shown schematically in
Figure 3. Clustersofhydroxyls exist that are orderedabove
and below the mirror plane. In natural hydroxylapatites,
however, enoughF "impurities" or anion vacanciesexist
in a single column to act as "reversal" sites, thus giving
rise to a locally ordered structure that averagesto the
disorderedP6r/m model.
Innpr-rclrroNs FoR TERNARyApATrrEs
Apatite minerals with the formula Car(POo)rX(X : F,
OH, CD have long been consideredas miscible isostruc-
tures even though few crystal-structurerefinementshave
been published on binary or ternary members of the sys-
tem. Observations on the hexagonalnear<nd-members
apatites of this study, however, clearly demonstrate that
the structural configurationsof the column anions (F, Cl,

Downloaded from https://pubs.geoscienceworld.org/msa/ammin/article-pdf/74/7-8/870/4221998/am74_870.pdf

by Yonsei University user
 HUGHESET AL.: STRUCTURALVARIATIONSIN APATITES 875

3lt4
z J/+

2 1.t4
oQ) o (1 ) 13t4
T II4
314
z=Il4

o(2) F OH CI
1A
Fig. 3. Depiction(to scale)of anion columnsin hexagonal
o(3) hydroxylapatite,
fluorapatite, Column"F' shows
andchlorapatite.
o(3) F anionsin fluorapatitelocatedon mirror planesat z : lq,3/+in
Ca(1) successive unit cells.Column"OH" depictsthreesuccessive hy-
droxylsin hydroxylapatite disordered0.35A abovethe mirror
planesand threesuccessive hydroxylsdisorderedbelowthemir-
ror planes,with the senseof disorderingreversedby an F anion
(stippled)"impurity" at z : l3/+.Column"Cl" showsthreesuc-
cessiveCl anionsin chlorapatitedisorderedabovethe mirror
planesandthreesuccessive Cl anionsdisorderedDelowthemirror
planes.The vacancy(Z) at z : l3/cin chlorapatitemustexistin
orderto reversethe senseofordering,as F and OH species are
prohibited. Radii:F: OH : 1.4A,Cl : 1.7A. Scaleof H atom
(solidcircles)is arbitrary.Scalebar : I A.

CI
are, on the average,disorderedabout the mirror plane. In
any column there are thus four possible anion sites as-
o(H) sociatedwith each mirror plane: an OH site located 0.35
F A below the plane, an OH site 0.35 A above the plane, a
o(2) Cl site 1.2 A below the plane,and a Cl site 1.2 A above
the plane. On average,each of these four possible sites
would be one-fourth occupiedby its respectiveoccupant,
OH or Cl. Anion-anion distances(Fig. a) calculated for
column anions using end-member data preclude such a
disordered P6 r/ m anangement,however.
Fig. 2. onrrr (Johnson,1965)drawingsof POotetrahedra Examination of possible anion arrangementsand Cl-
and Ca(l) and Ca(2)polyhedrafor the threeapatitestructures.
OH interionic distancesdemonstrateswhy complete dis-
Eachoverlayis a superposition ofthe analogous polyhedrafrom
the threeend-memberstructuresdrawnto the samescaleand ordering is not possible.For example,associatedwith one-
with coincidentcentralcalions. fourth of the mirror planes at z : t/qwe would find a Cl
anion disordered1.2 A above the plane. In this model,
possible neighbors along +c include an OH anion below
OH) are not miscible. In this section, we develop further the adjacentplane at z:3h lCl-OH distance: l.8l Al,
structural constraintsthat must govem binary and ternary an OH arrior' abovethe adjacent plane [CI-OH distance
solution ofapatites. :2.52 A, whereasideal non-hydrogen-bonded Cl-O dis-
Consider, for example, a hypothetical binary apatite tance : 3.2 Al, a Cl anion below the adjacent plane (CI-
with the compositionCar(POo)r(OHorClor). If we assume Cl distance : 0.92 A), and a Cl anion abovethe adjacent
that the anion positions in this binary apatite are the same plane,at a distanceof c/2 : 3.41 A. Only the last config-
as in the end-members, then its structure would have uration (Cl anions above the plane, ad infinitum) is rea-
anion columns that contain both Cl and OH anions that sonable based on interionic distances.Vacanciescan, of

Downloaded from https://pubs.geoscienceworld.org/msa/ammin/article-pdf/74/7-8/870/4221998/am74_870.pdf

by Yonsei University user
 876 HUGHESET AL.: STRUCTURALVARIATIONSIN APATITES

Anion Associated With Mirror Plane agonal symmetry elements, and thus a variety of anion
sites suitable for the three anions in ternary apatites are
tt z = 0.75 created.
(l)
oH" oHu ct, ctu AcxNowr-nncMENTS
We thank I. D. Brown for kindly providing constantsfor calculation of
:
.-A
oH- 3 . 4 1 t:':'2.70 4.30 1.81...i. Ca-Cl bond strengths.This work was supported by grants from the Na-
tional ScienceFoundation(CHE-8418897and EAR-8717655to Hughes
and EAR-8 5 17621to Crowley and Cameron).Specimensusedin this study
were provided by GeorgeE. Harlow of the American Museum of Natural
elO
(Jtlr. 4.12 3.41, 5.01 ) \'). History and P. J. Dunn of the National Museum of Natural History,
= N
.W Smithsonian Institution. The manuscript was improved by thorough re-
views by G. Y. Chao and R C. Rouse.
tl
lt::::-l::::::
q.)N
;ecl_ ttt';'.'s:;)ttttl
::r::r:ll
.j....1.81::.l3 . 4 1 ;:,::,'.'.0.92,,1, RsF'nRENcnscITED
I :::: ii:i:i:::1:
Beevers,C.A., and Mclntyre, D.B. (1946) The atomic structure of fluor-
U) apatite and its relation to that oftooth and bone material. Mineralogical
0
4- Magazine,27, 254:257.
Clx 5.01 4.30 5.90 3.41
Bragg, L.B., and Claringbull, G.F. (1965) Crystal structuresof minerals,
Eu
409 p. G. Bell and Sons,London.
Boudreau,A.E., and McCallum, I.S. (1987) Halogensin apatitesfrom the
Stillwater Complex: Preliminary results (abs.).EOS, 68, 1518.
Fig. 4. Anion-aniondistances(A) for columnanionsusing Candela,P.A. (1986) Towards a thermodynamic model for the halogens
in magmatic systems:An application to the melt-vapor-apatite equi-
end-member dalapresented in thispaper.Subscript"a" indicates libna. Chemical Geology, 57, 289-301.
anion disorderedabovethe associated mirror plane,subscript Crowley, K.D., and Cameron, M. (1987) Annealing of etchable fission-
"b" indicatesaniondisordered belowtheassociated mirror ptane. track damagein apatite: Effectsof anion chemistry. GeologicalSociety
Shadedcellsindicateconfigurations prohibitedbecauseof un- of America Abstracts with Programs, 19,631-632.
reasonably shortinterionicdistances orjuxtapositionofH ions. Elliott, J.C., Mackie, P.E., and Young, R.A. (1973) Monoclinic hydrox-
All valueswerecalculated assumingc: 6.83A, the averageof ylapatite.Science,180, 1055-1057.
co, and c.,. Frenz, B.A. (1985) Enraf-Nonius structure determination package.SDP
usersguide, Version 4. Enraf-Nonius, Delft, The Netherlands.
Green, P.F., Duddy, I R., Gleadow, A.J.W., Tingate, P.R., and Laslett,
course,initiate a new anion sequencein the column, al- G.M. (1986) Thermal annealingof fission tracks in apatite 1. A qual-
itative description. Chemical Geology, 59, 237-253.
though it has not been establishedthat stoichiometricva- Hounslow,A.W. (l968) Crystal structur€softwo naturally occurringchlor-
canciesexist in ternary apatites, nor how chargebalance apatites Ph.D dissertation, Carleton University, Northfield, Minne-
would be maintained. This simple binary example can be sota.
extended to ternary members including F anions in the Hounslow, A.W., and Chao, G.Y. (1970) Monoclinic chlorapatite from
Ontario. Canadian Mineralogist, 10, 252-259.
mirror plane, as well, becausethe Cl-F distance (2.17 L)
Johnson, C.K. (1965) A FORTRAN thermal ellipsoid plot program for
would lie betweenthe two CI-OH valuesdescribedabove. crystalstructureillustrations. U.S. National TechnicalInformation Ser-
It is thus clear that the ternary members of the apatite vice, ORNL-3794.
systemare not ideal solutions of the hexagonalend-mem- Kay, M.I., Young, R.A., and Posner, A.S (1964) Crystal structure of
hydroxylapatite.Nature, 204, I 050-1052.
bers. There are several possibleways, however, that ter-
Mackie, P.E., Elliott, J.C., and Young, R.A. (1972) Monoclinic structure
nary solid solution can be accommodated.Extensive re- of synthetic Ca,(PO).CI, chlorapatite. Acta Crystallographia, 828, I 840-
arrangementof the end-member anion configuration, as I 948.
yet undefined,may occur in the ternary systemto accom- Mehmel, M. (1930) Uber die Struktur des Apatits. I. Zeitschrift fiir Kris-
modate the Cl ion; anion vacancies may also play an tallographie, 75, 323-33 1.
N6ray-Szab6,S. (l 930) The structureofapatite (CaF)Cao(POo),. Zeitschrift
essential role in stabilizing these apatite structures. Fi- fiir Kristallographie, 75, 387-398.
nally, a completely disordered hexagonal structure may Stormer, J.C, and Carmichael,I.S.E. (1971) Fluorine-hydroxyl exchange
be realizedwherethere is anion ordering withinindividual in apatite and biotite: A potential igneousgeothermometer.Contribu-
columns, but disordering among the columns themselves. tions to Mineralogy and Petrology,31, l2l-131.
In a newly initiated study, we are analyzing the crystal Sudarsanan,K., Mackie, P.E., and Young, R.A. (1972) Comparison of
synthetic and mineral fluorapatite, Ca,(POo),F,in crystallographicde-
structures of binary and ternary apatites in the OH-F-CI tail. MaterialsResearchBulletin, 7, l33l-1338.
system.Limited data obtained to date indicate that binary Young, E.J , Myers, A.T., Munson, E.L., and Conklin, N M (1969) Min-
and ternary apatites respond to solid solution in at least eralogy and geochemistryoffluorapatite from Cerro de Mercado, Du-
two ways: (l) by large(-0.4 A) shifts of anion positions rango,Mexico. U.S. GeologicalSurveyProfessionalPaper650-D, D84-
D93.
relative to the end-membersand (2) by reduction of ter-
nary apatite symmetry to monoclinic, where Ca(2) and Maxuscrum RECETvEDOcronpn 14, 1988
the column anions are no longer constrained by the hex- Ma.NuscnrvrACcEFTED
MmcH 7, 1989

Downloaded from https://pubs.geoscienceworld.org/msa/ammin/article-pdf/74/7-8/870/4221998/am74_870.pdf

by Yonsei University user