You are on page 1of 9

1702 K.

KUGIMIYA
AND H. STEINFINK Inorganic Chemistry

COSTRIBUTIONFROM THE MATERIALS SCIENCE LABORATORIES,


DEPARTMENT UNIVERSITY O F TEXAS
O F CHEMICAL ESGINEERING, AT AUSTIS, AUSTIN,TEXAS78712

The Influence of Crystal Radii and Electronegativities on the


Crystallization of AB2X4Stoichiometriesl
BY K. KUGIMITU’AASD H. STEIKFINK

Received Murch 12, 1968

The quantitative relationships between compounds having AB204 stoichiometries and their crystal structures were investi-
gated. Diagrams relating the ratio of the radius of atom A to that of oxygen, the ratio of the radius of atom B to t h a t of
oxygen, and an artificial parameter described as a force constant between A and B atoms and derived from their electro-
negativities and the equilibrium distance between them show distinct areas for different structure types. In particular,
a plot of T J Y I , vs. the force constant provides good resolution among structures. A three-dimensional illustration in which
the axes are ratios of radii of -4 and B atoms to t h a t of oxygen and the force constant resolves the frequently occurring struc-
tures into separate volumes. A similar plot was constructed for compounds with AB4X4 stoichiometries where X is S , Se,
and Te. Although fewer known compounds are available for mapping, it was, nevertheless, possible to separate different
structures into different regions. A mathematical relationship between oxygen and cation occupancy of tetrahedral and
octahedral voids is derived which describes possible crystal structures,

Introduction The substitution of various cations into the tetrahedral


The ability to predict the type of crystallization in and octahedral positions in the close-packed oxygen
which a given stoichiometry of cations and anions will framework gives rise to hundreds of compounds in which
crystallize is a goal which has been pursued by crystal physical properties can be tailored to fit requirements
chemists for a considerable number of years. Many for a particular application. However, other struc-
attempts have been made to systematize a large amount tures are also found for compounds of this stoichi-
of information by plotting or correlating suitable phys- ometry. The olivine, CaFeeOl, KzSOa, BaAl2O4, and
ical and chemical parameters so that a scheme becomes phenacite structures constitute major types in which
evident which will permit such predictions. Evans2 many compounds crystallize. In addition to the
discussed the relationship between chemical compo- major structure types, a few of the compounds have
sitions and the crystal structures of metal alloys and structures which are related to two major types.
ionic compounds. Mooser and Pearson3 have reported The crystallization of spinels in the normal or inverse
that in compounds of the type A,Bn a correlation exists structure was successfully explained in terms of the
among the electronegativity differences, average stability energy of ions in tetrahedral or octahedral con-
principle quantum number, and the crystal structure. figuraton. Glasser and Glasser8 investigated the rela-
Later Pearson4 reported finding a better correlation by tion of ionic radii and resultant crystal structure for
taking account of the effect of the radius ratio for the some A4B203stoichiometries where A was Ba, Pb, Sr, Ca,
two atoms. Roth” studied a large number of perovskite- or Cd, and B was AI, Ga, Cr, Fe, or V. Because of the
and ABOa-type compounds and reported observing a interests in this type of structure, an attempt is made
relationship among crystal structure, radius ratio, and here to relate the type of crystallization to the intrinsic
polarizibility of the A and B atoms. Keith and Roy6 properties of the ions so that one can predict with a cer-
related the radius ratio and the tolerance factor of tain amount of assurance the crystal structure of an un-
perovskite-type compounds to their crystal structures. known compound with this stoichiometry.
Recently Kjekshus and Pearson’ surveyed the stoi- Crystal Parameters
chiometry AB and discussed the compounds which form A comparison of the normal spinel and olivine illus-
the nickel arsenide structure. trates the principles which need to be considered in
The mineral spinel, MgA12O4,has a crystal structure
the formation of such structures. Both types have a
which occurs in a large number of compounds having basic unit, A-B-0, consisting of four cations surrounding
the general formula AB204. Spinel-type compounds
one oxygen ion: three B cations in octahedral voids
have received much attention because of the discovery
and one X ion in a tetrahedral void. The structures are
that the magnetic properties which can be incorporated
formed by the close packing of the large oxygen ions,
into this structure permit their use in various electronic
cubic close packing in spinel and hexagonal close pack-
devices of considerable technological importance.
ing in olivine. I n the spinel structure the three B ions
(1) Research sponsored by the Air Force Office of Scientific Research, in octahedral voids and the A ion in a tetrahedral void
Office of Aerospace Research, United States Air Force. around oxygen are situated so as to maximize the A-B
( 2 ) R. C. Evans, “An Introduction t o Crystal Chemistry,” Cambridge
University Press, London, 1939. distances while a t the same time maximizing the B-B
(3) E. Mooser and W. R. Pearson. Acta C r y s t . . 12, 1015 (1959). distances which are less than A-B. The A ion sees
(4) W. B. Pearson, J . Phys. Chem. Solids, 23, 103 (1962).
(5) R . S. R o t h , J . Res. N a t l . BUY.Std., 58, 75 (195i). every B ion a t an equivalent distance. In the olivine
( 6 ) M. L. Keith and R. Roy, A m . ,Minevaiogisl, 39, 1 (1954). structure there are two basic A-B-0 units. One is the
(7) A . Kjekshus and W. B. Pearson, “Progress in Solid S t a t e Chemistry,”
H, Reiss, E d . , Pergamon Press Inc., New York, h-,Y . , 1965. (8) F P. Giasser and I,. S D. Glasser, J . A m . Cevam. .FOG.,46, 3 i i (1963).
Vol. 7, No. 9 , September 1968 OF AB2X4STOICHIOMETRIES
CRYSTALLIZATION 1763

unit described above for the spinel structure and the vestigation. It is quite possible, however, that the
other consists again of three B ions in octahedral voids choice of a correctly formulated difference function
and one A ion in a tetrahedral void, but the B-B dis- could yield results as good as or better than those shown
tances are first maximized and then A is placed into a here.
tetrahedral void so as to maximize the A-B distances. We have defined a tolerance factor for the spinel
The spinel cation arrangement will occur if the repul- structure as the ratio [ l l O ] / d ~ [ l O O ] . On the assump-
sion between A and B ions is greater than among the B tion that all ions are ideally spherical and rigid and that
ions, and, if the repulsion between A and B is smaller they contact each other along each direction as they
than that between the B ions, the olivine-type arrange- build up the ideal spinel structure, the length of the face
ment will occur. At the same time the repulsive forces diagonal is
among the cations must not be so large as to prevent
their contact with the oxygen ions or else the structure
becomes unstable.
The assumption is made in this development that the
ions are hard spheres although i t is recognized that the
and the lattice constant is 4(rb + ro). Thus the toler-
ance factor is
overlap of bonding electrons, bond angles, and the
polarizability influence the ion arrangement. How-
ever, the inclusion of these additional factors compli- T =
1 r,+
4 3 rb
+--------
1
Yo

+ 42 +
Yo rb
Yo

yo
cates the problem to an extent that it becomes intract-
able by the procedure adopted here. The bond and becomes unity if the ideal spinel structure is as-
stretching force constant, K , is considered to represent sumed. In order to compare the calculated lattice con-
this interplay of forces and is given by stant of the spinel structure as obtained from the ionic
radii with the experimentally measured value, Lexptlis
K = UN(Y)”~ b + taken as the average of 1/1/2[110] and [loo] and the
calculated lattice constant is
where a and b are constants, N is the bond order, re is
the equilibrium distance, and X, and Xb are the electro-
negativities of the A and B ions, r e s p e c t i ~ e l y . ~If~ ’ ~
the assumption is made that a and N are constants and No attempt was made to calculate the lattice constants
that b is small and constant so that it can be neglected, of other structures.
then K can be written as
Computations of Parameters
Tables I and I1 list the electronegativities and ionic
radii of the cations and anions used in the calculations
where re2 = (ra + + +
ro)2 (rb + +
r o ) 2 1.155(ra r o ) . of Kab,K a b / K b b , the tolerance factor, the various radius
ratios, and the products of electronegativities. The
(rb -/- T o ) , ra, rb, and ro are the radii of the A, B, and 0
ions. A further simplification can be introduced by radii shown are for octahedral coordination and were
considering re to be approximately constant and then multiplied by 0.94 to obtain the value for tetrahedral
Kab = XaXb. The ratio K a b / K b b can be considered as a coordination. If the cationic radii in the interstices did
measure of the respective repulsive forces between A-B not contact the anions, then they were adjusted as fol-
and B-B and from the known arrangements i t can be lows. (1) If the cationic radii were too small, they
postulated that the spinel structure will be preferred were assigned values of 0.414r0 and 0.225r0 for the octa-
when this ratio is small, and the olivine structure will be hedral and tetrahedral sites, respectively. (2) If the
preferred when this ratio becomes larger. I n addition tetrahedral ion was too large to fit the interstice, then
to these electrostatic interactions, the radius ratios the length of the tetrahedral oxygen edge was cal-
of the various cations and anions greatly influence the culated. If the octahedral ion radius was now too
crystallization. small, it was expanded to touch the six oxygen ions.
The correlation between electronegativity differ- The conditions expressing this are : if
ences and average quantum number has successfully
grouped crystal structures in binary compounds. The
0.94ra
~~
r, + Yb + ro
2- 1.414
1.225
choice of the type of difference function, whether simply
xa - Xb or a more complicated expression, will strongly then r b = 1.154r, -t 0 . 1 5 4 r 0 and ra has the value shown
influence the resultant map when this method is applied in Table I. (3) If the octahedral ion was too large for
to compounds containing three different atoms. We the interstice, then the length of the octahedral oxygen
found that maps based on xa - Xb, or other simple dif- edge was calculated and the tetrahedral cationic radius
ference expressions, were not as effective in separating was enlarged to fit the void. The conditions expressing
crystal structures as the expression adopted in this in- this are: if
(9) W. Gordy, J . Chem. Phys., 14, 305 (1946).
(10) A. F. Wells, “Structural Inorganic Chemistry,” 3rd ed, Oxford Uni-
0.94, +
ro rb
<- ro +
versity Press, London, 1962. 1.225 1.414
i\ND H STEINFIKK
17634 K. KUGIMIYA Tnorgnnic C'hrmistry

then r z = O.866rb - 0 . 1 3 4 and


~ ~ r b has the value shown Several ambiguities exist about structures for the
in Table I. The above modifications in the ionic radii same materials. The compound CuFegO4 is reported to
were used only in the calculation of Kab and Kab/'Kt,b be a spinel and also to have the Mn304structure. It is
TABLEI therefore listed under both structures in Table 111.
ELECTRONEGATIVITIES, X,
The inverse spinel structure and the Mns04structure are
AND IONICRADII,' Y , OF CATIONS reported for CdIn204, and CuRh204 is said to have the
----Valence I---- ---T'alence 2- ----Valence 3--- spinel structure.ll Both the spinel and olivine struc-
Ele- Ele- Ble- tures are reported for GeMgzOJ,lland SrCr20dis said to
ment )I, A x ment 11, A x ment v, b x
Li 0.68 0 97 Be 0.35 L 4 i AI 0.51 1.47 have the stuffed silica structure although i t is not defi-
Sa 0.97 1.01 M g 0 66 1 23 Sc 0 81 1.20 nite.8 The compound CaAl204 has a distorted stuffed
K 1.33 0.91 Ca 0 99 1.04 Ti 0 76 1.32
Cu 0.96 1 75 Y 0 88 1 45 V 0 74 1.45 silica structure jx SrFe20jhas an CaFepOi structure* al-
Rb 1.47 0.89 RIn 0 80 1 60 Cr 0 63 1.56 though it was previously reported to have hexagonal
Ag 1.26 1.42 Fe 0.74 1 64 Mn 0.66 1 60
Cs 1.67 0.86 Co 0.72 1 70 Fe 0.64 1 64
symetry." C r ~ i c k s h a n k ~reported
' that Calnn04,
Au 1.37 1.12 Ni 0.69 1 75 Co 0.63 1 70 CdIn204,SrInn04,and BaIneOl have the distorted spinel
TI 1.47 1.44 Cu 0 72 1 7,5 Ga 0.62 1.82
Zn 0.74 1 66 Y 0 92 1.11 structure and that their structures are related to that of
Ge 0.73 2.02 Kh 0.68 1.45 CaFe204. For AB204compounds, in which A is Sr or
---Valence4--- Si 1.12 0.99 In 0 81 1 49
Ele- Pd 0.80 1 35 Sb 0.76 1.82
Ba and B is Er, Tm, or Lu, the CaFe204structure is re-
ment jl) a x Ag 0.89 1.42 La 1.14 1.08 ported by Bhargava, et u Z . , ' ~ though S ~ h w a r zreported
'~
Si 0.42 1.74 Cd 0.97 1.46 Ce 1.07 1.06
S 0.37 2,jo Sn 0 93 1 72 Pr 1.06 1 07 that those compounds have other structures. Quey-
Ti 0.68 1.32 Ba 1.34 0.97 Sd 1.04 1.07 roux'j studied the compounds CaYb204 and CaDysOl
v 0.63 1.45 Pt 0 80 1 44 Sm 1 00 1.07
and reported that they are similar to CaFe204. For
&In 0.60 1 60 Hg 1 10 1 44 Eu 0.98 1 01
Ge 0.53 2 02 Pb 1 20 1 55 Gd 0.97 1,ll comparison with compounds having the formula AB204,
0 50 1 43 0 97 Tb 0.93 1.10
Se 2.48 Ra
Dg 0.92 1.10
several compounds with compositions XB2X4,where X
Zr O,i9 1.22
IVZO 0.70 1.30 --T Ialence 5---
Ho 0 91 1 10 is sulfur, selenium, or tellurium, are also listed in Table
Ru 0.67 Er 0.89 1.11
Ele-
Tm 0.87 1.11
111.
Pd 0.65
0.71
l.33
1.72
ment 7, A x Yb 0 86 1.06 The equation for L e a l e d , derived with the assumption
Sn
0.70 2 01 v 0 59 1.45 Lu 0 85 1.14
Te
0.69 1.23 Au 0.85 1.42
stated previously, provided agreement between ob-
Ce 0.94 1.06 lib
Pr 0 92 1,o7 Mo 0.66 1.30 T1 0.95 1.44 served and calculated lattice constants with an average
1.10 Sb 0.62 1 82 Bi 0.96 1.67
Tb 0.81
0.62 2 21 deviation of 1.5% except for PbFe20d. The same equa-
Hf 0.78 1.23 I
W 0 70 1.40 Ta 0.68 1 33 -----Palence 6---. tion, but using modified ionic radii according to the
Re 0.72 1.46 'V 0 66 1.40 Ele-
0.66 146 ment r ) .$ x models previously discussed, gave calculated values
Ir 0.68 1.55 Re
Pt 0.65 1.44 Bi 0.74 1.67 s 0 30 2.41 which were 10-15yo larger than observed values. It
Pb 5j I-'* 0.89 1.14 Cr 0 52 1 56
0.84
0 88 1.22 Se 0.42 2.48
would have been reasonable to expect that these mod-
Th 1.02 1 11 C'
Pa 0.98 1.14 .\Io 0.62 1 30 ified radii should give closer agreement and the failure
u Te 0 56 2.01
0.97 1,22
w 0.62 1.40
to do so can be ascribed to distortion which the cations
Re 0 61 1.46 apparently undergo, giving rise to deviations from
Os 0.69 1 52
IJ n 80 ~~ I 22 sphericity which permit them to fill the available space
L. H. Ahrens, Genchem. Cosn7ochinl. ilcta, 2, 155 (1952) in the close-packed oxygen layers. The two calculated
lattice constants of PbFes04, for the different valences
for the cations, are 8.8 Ak,which is significantly larger
T A B L EI1
than the experimental value, 7.51 A. It must be as-
X , ASD IONIC
ELECTRONEGATIVITIES, RADII,T , OF XNIOKS
sumed that either the reported value or the reported
Element v,A X
structure is wrong.
0 1.40 3.50
S 1.84 2.44 Figure 1 is a plot of KabDS. y a / ' r b for known structures.
Se 1.98 2.48 It is seen that the different structures cluster into cer-
Te 2.21 2.01 tain areas of the diagram, although some compounds
actually lie on the boundaries as shown. The phenacite
Results and Discussion structure, Be2SiOd, provides an exception and is found
scattered throughout the diagram. Compounds with
Information regarding cornpounds listed in Table I11
the Mn304structure and those with a closely related
without asterisks was obtained from Wyckoff ;I1 struc-
distorted spinel structure are almost all contained in
tures of compounds with a single asterisk were obtained
the normal and inverse spinel areas. It i s interesting to
from the ASTM card file.12 The degree of inversion of
observe that structures formed by slight deviations from
the 'pine' structures marked with the asterisk
typestructures are found on the boundary lines bet.lveen
was taken from Wellslo and the numbers next to the
the two major structure areas. The large intermediate
compounds indicate the particular reference from which
the information was ~ b t k e d . ~ ~ - ~ ~ (13) H. D . Bhargava, L. IVI. Kovba, L. I. hlartynenko, and V. I. Spitsyn,
Dokl. A k a d . Nauk S S S R , 161,594 (1965).
(14) H. Schu-arz, Z. Nulz~IJojlsch..19b,955 (1964).
(11) R . W. G. Wyckoff, "Crystal Structures," Vol. 3, 2nd ed, Intersci- (15) F. Queyroux, Coingt. Rend., 261, 4430 (1965).
ence Publishers, Inc., New York, S . Y . , 1965. (16) J. Flahaut, hZ.Guittard, & I.
Patrie, &I, P. Pardo, S. M. Golabi, a u d
(12) Powder Diffraction File, American Society for Testing Xaterials, L. llomange, A d a C ~ y s t . 19,
, 14 (1965).
Philadelphia, Pa., 1963. (17) F. H . Cruickshank, J . I x o i g , Nucl. Chrm., 26, 937 (1964).
Vol. 7 , No. 9, SePtember 1968 OF ABZX4 STOICHIOMETRIES
CRYSTALLIZATION 1765

TABLE
I11
STRUCTURES
CRYSTALDATAFOR KNOWN THE ABtX4 COMPOSITION
HAVING

,Ht*LCllf

C.f.?O(

The first three columns correspond to A, B, and X. T h e single and double asterisks and numbers in the columns immediately fol-
lowing the anions are references. T h e succeeding columns are: valence states of A and B cations; three radius ratios Y ~ / I ' ~Yb/?x,
, and
ye/rb; product of electronegativities xILxh;force constant Kat,; ratio of force constants K,b/Kbb; tolerance factor; calculated unit cell
constants; experimentally observed unit cell constant; degree of spinel inversion.

area between the spinel region and the CaFezO4 region and 2-3. One assumption of valence distribution
contains CaDy204and CaYbz04 which are monoclinic places it into an incorrect area while the other assump-
with a cation arrangement similar to CaFezO4; UCa~04 tion assigns i t to the correct structure region. Thus the
which has a tetragonal symmetry; CaIn204, SrIn204, compound MoV'AgT204lies almost on the boundary
and BaInz04 which are distorted spinels and have a between the KzAgI4 and the olivine structure areas,
structure related to that of CaFez04;BaV~04and SrV204 while M0'~"''~04 falls properly into the inverse spinel
which are defect perovskite structures; CaTizO4 which is region. Similarly Ge'1Mn'"z04 is contained in the
body-centered orthorhombic ; and CaV204 and Ca- spinel region while Ge1VMn"204 lies properly in the
Ga204which are reported to have structures depending olivine region. The compounds NiA1204 and CdGazO4
on the conditions of the formation of the crystals. The which fall into the normal spinel area are reported to be
compound GeMg204 which can have both the spinel partially inverted spinels with X = 0.38 and 0.25, re-
and olivine structures lies on the boundary between the spectively, and the reported inverse spinel structure for
spinel and olivine regions. Some monoclinic crystal CdInz04, which lies in the normal spinel region, is
structures which deviate from major structure types stated to be doubtful.'l The compound Fey04 which
are also observed on boundaries. lies in the normal spinel region near the boundary line
There are a few known compounds which do not fall with the inverse spinel region is known to have an in-
into the proper regions in Figure 1. Those are MoAgz- verted structure. Consequently the notable excep-
04,GeMnz04, MnV204, GeFeeO4, GeCoz04, MnFez04, tions are phenacites, MgTi,O4, BaYbzO4, and Ba-
SnCoz04,NiA1204,CdInz04,MgTiz04, CdGa204, B aCr2- Crz04. Phenacites are scattered throughout the
04,BaYbz04, and phenacites. All of these except for diagram of Figure 1, MgTiz04lies in the inverse spinel
NiAl2O4, CdInzO,, MgTi204, CdGaz04, BaInz04, and region although its structure is reported to be normal,
phenacite have two possible valence states, mainly 4-2 BaYbz04 lies in the CaFez04structure area although i t
AND H. STEINPINK
1766 K. KUGIMIYA Inorganic Chemistry

24
'

0.0 __3 0.5 1.0 1.5

0 s p i n e l ; A unknown Q K2SOb s t r u c t u r e
0 normal s p i n e l * K2 MgP4 structure
e i n v e r s e s p i n e l ; A known 0 CaFe204 s t r u c t u r e
0 Mn O4 s t r u c t u r e BaAl 0 s t r u c t u r e
B distorted spinel structure 0 phengctte s t r u c t u r e
a olivine structure o t h e r s t r u c t u r e s a n d unknowns
Flgure 1 -Plot of K,i, VA rddlus ratlo of cations of k i i o ~ i icrystdl structures for compounds hdving the forinu~ciAB204

the structures to be predicted for unknown compounds,


generated by the elements listed in Table I, which fall
onto the boundaries can be the distorted structure or
some other structures related to two adjacent major
structure types. Uncertainty also exists for unknown
structures which fall into the part of the inverse spinel
area which is mapped on the basis of only two known
compounds, MoNa204and WNa204.
There are several regions which contain no or only a
very few known compounds. Those regions are left
blank in Figure 1 and none of the boundary lines in
Figure 2 is extended past K,h = (1.25. The known
compounds lie nithin relatively small values of K a b a t
the fixed Ya/Yh values. The generated compounds
Figure 2.-Plot of K,,, v s . f f : t h i , for computer-generated com- which have large K,b values are presumably difficult to
pounds. Each dot represents a compound AB204. The bound- prepare and may form multiphase material containing
aries are transferred onto this diagram from Figure 1 compounds such as AB03 +
BO or A 0 +
BeOawhich
satisfy the stoichiometry ABs04. Perhaps an addi-
has hexagonal symmetry, and BaCr204 falls in the tional thermodynamic variable, such as pressure, is
stuffed silica structure region although it has a mono- required to prepare these compounds.
clinic crystal structure. The correlations among Kat,, r,/rl,, and Y I , / Y ~ are
Among the areas of the various structures, the spinel schematically illustrated in three dimensions in
region covers the largest area in the diagram and con- Figure 3. The regions of the spinel structure and of the
tains almost half of the compounds whose structures olivine structure are shown by solid lines, and the re-
were known. This may imply that the spinel structure gions of the K2S04structure, the K2;tlgFastructure, and
is one of low potential energy and is therefore a preferred others are omitted except for their base lines on the
structure type. zero level in order t o keep the diagram simple. In the
The ions listed in Table I were used t o generate 1236 volume of high K a b or low K & values only a few com-
compounds of the type ABzO4and they are plotted in pounds exist, The surfaces which define the spinel and
Figure 2 . The boundaries shown in this figure are the olivine volumes are simply continued t o separate those
same as those of Figure 1. The positions of some bound- spaces.
aries are subject to error since there are very few known Again fairly good separations of compounds are ob-
structures close to the boundaries themselves and thus served in Figure 3. The exceptions are MoMgzO~,
Vol. 7 , No. 9,September 1968 CRYSTALLIZATION O F AB2Xd STOICHIOMETRIES 1767

Crystal Chemistry of ABzX4 (X = S, Se, Te)


_----T----------
I ;:r // :,
\: The same procedure which was carried out for the
compounds of the ABz04type was applied to the com-
pounds having the formula AB2X+where X is S, Se,
and Te. The results are listed in Tables 111 and I V
and a plot is shown in Figure 4. Three major crystal
structure areas exist: the spinel and AgzHgI4, the
CaFe204,and the Th3P4 structure areas. Two inter-
mediate regions can be recognized, namely, a rhombo-
hedral and an orthorhombic one. The broken lines in
Figure 4 are the boundary lines corresponding to the
areas shown in Figure 1. The square of the ratio of
the radius of selenium to that of oxygen, ( r ~ ~ / r , ,is) ~2,
and in order to adjust the scale the force constant was
multiplied by I/z. These transferred boundary lines
produce regions which do not conflict with the areas
occupied by the major crystal structure types. In-
verse spinels lie in the inverse spinel region and the
horizontal line separates the map into two areas of
major structures. The similarity between the dia-
grams in Figure l and 4 is nevertheless somewhat ten-
uous because few compounds of the AB2X4 type having
\ structures lying on the left side of Figure 4 are reported,
while three major structure areas of compounds having
Figure a.-Three-dimensional surface resulting from a plot of the formula AB204are located there.
K,b vs. rt,/ro os. rn/ra for known cry:tal structures of compounds
An important feature in Figure 4 is that the spinel
having the AB204 composition.
structure area and the AgzHgI4 structure area are not
separated. It is also interesting that no compound of
GeMg204, PbFe204, SiNiz04, BaIn20d, SrInzOa, and the AgzHgI4 structure is found among the ABzOl-type
phenacites. The compound CaIn204lies on a sep- compounds. It can be postulated that in addition to
arating surface. Among these compounds the struc- the larger radius of the anions S, Se, and Te, which
ture of GeMg20dis not certain;ll the compound PbFe20d permit the cations to fit easily into the tetrahedral sites,
is a spinel but a valence distribution 2-3 places it out of some other force due to these anions becomes impor-
the spinel structure region, while a 4-2 valence distribu- tant, for instance, a site preference energy. This
tion properly places i t in the spinel region. It is reason- postulate is advanced on the basis that even though
able to assume therefore, that in PbFe204 the valences compounds with the formula AB204 have similar values
are 4-2 or perhaps mixed 4-2 and 2-3 and in MoAg204 a for the force constants and radius ratios as compounds
similar valence distribution occurs as previously stated. ABTX4 which have the AgqHgI4 structure, they still do
The spinel structure region in Figure 3 is almost a not crystallize with this structure. The Th3P4struc-
rectangular parallelepiped, except for the areas of large ture to which no compound with the formula ABz04
and small values of the force constant. This feature is belongs is also present as a major structure type in
also observed for regions of other structures and implies Figure 4. No ABzOa-type compound has both the
that the radius ratio is a dominant factor in the deter- small force constant and appropriate radius ratio value
mination of the structure. The radius ratios of all to permit this crystallization. Thus it is not unreason-
spinels except SiNi204are within the limits able that the Th3P4structure is not observed among the
ra/ro = 0.36-0.65 compounds of the AB204 type if the congruence be-
tween Figures l and 4 holds.
rb/ro = 0.35-0.69 The density distribution of the compounds having the
where ra is the tetrahedral radius. Phenacites and a formula AB2X4is quite similar to that of the compounds
few compounds of other structures are also found within with the formula ABz04 shown in Figure 2. Super-
these limits. position can be achieved by shifting every dot in Figure
I n addition to SiNi204,BaIn204,and SrIn204,as pre- 2 by half the magnitude of the force constant. This fig-
viously discussed, MoNa20d and WNa204, which have ure immediately shows that many computer-generated
low values of Kab,are also exceptions from this relation. compounds lie in the area where no AB2X4 compounds
CaTizO4, CaFezO4, CaV204, and CaCr2Od fall nearly on are found in Figure 4. The major structures similar
the face of the spinel region. The radius ratios for them to the AB204 types are expected to occupy this area.
are ra/ro = 0.66. It is also observed that the known
compounds having the olivine structure have larger Possible Structures for AB2X4Stoichiometries
Kab values than the compounds with the spinel struc- An expression relating various cation configurations
ture. around anions in close packing can be derived. Let
1765 K. Kuc1MIu.4 AND H. STEINFINK Inorganic Chemistry

'TABLE I\'
ADDITIONAL COMPOUNDS
WITH THE AB& SIOICHIOMETRP~
2 3 .34 . c9 e58 1 30 I u4 ,94 uti i4 SE 2 3 .31 . E8 .54 1.33 .05 .V3
2
2
3
3
.34
.3L
.h2
a5-3
.54
.59
1-33
1.32
1u3
.b4
-91
.99 tructilre
vu C E SE
* G PQ SE
2
2
3
3
.)I .54
~ 3 1 .F4
* 58
I59
1.30
1.32 .
* 05
n5
.95
*95
2
2 3
2
3
3
.34
.3r
,34
.5T
954
*53
s
.b2
.b*
bo 1.32
1.32
1.31
.
* 04
b9
.Ul
.95
.9b
-91
nknown MG NN> SE
V G Su SE
-4ti 0'1 SE
2
2
2
3
3
3
r.i!
-31
-31
153
.SI
*49
ab0
,b%
.64
1.32
1 .32
1.31
.'15
.05
a96
.91
.98
:ructLre
nknown
2 3 I34 -51 67
I 1.35 -81 .9I UG l a SE 2 1 .31 .a7 -67 1.35 1.00
2 3 ,3* -50 .bl 1.35 .dl -99 *b U Y st L 3 .31 .e5 .D7 1.3Y 1.00
2 3 I34 * 50 ab1 1.32 -81 .99 rvli Y SE ? 3 .31 .ih .bl 1.37 1.00
2 3 * 34 * *9 -60 1.35 .MI *9L) .IG ti? SE 1 3 -31 I * 6 .68 1.3s 1.00
2 3 .34 .LU -70 1.37 .L(l 1.00 40 EU SF. 1 3 ~ 3 1 $65 a70 1.37 1.01 ublc spine:
2 3 .34 .*I .lI 1.37 .SI 1.00 Yti 1rr sc 2 3 .31 .44 .ll 1.37 I.OZ
2 3 I34 .*I a72 1.30 .u* 1.01 iublc s p i n e l ulj Y Y SE 2 3 .31 .r3 -12 1.30 1.02
2 3 .34 .4b .73 1.40 * 7R 1.01 * G LL SE 2 3 .31 .43 173 1.60 1.03
2 3 a34 .44 .77 1.48 .7Y 1.03 C 4 L A SE 2 3 .41 .5H ,a2 1.12 .99
2 3 -51 .67 -82 1.12 -70 ,91 C h CE SE 2 3 .*I .=A 111 1.LQ \.Ill
2 3 .51 .58 81 1.10 .ne
.
# .I1 1.00 C A P9 SE 2 7 .bl .5* 1.11 1.01
2 3 -51 *58 .nR 1.11 .lI
c4 NO
S"
s
s
2 3 .51 .57 nq 1.11 .lI
1.00
1.01
C A NO SE
C A s u SE
2
2
3
3
.e7
.il
.53
*51
.u9
.D3
1.11
!.I1
1.02
1.03
CA
CA
CA
6')
T*
5
5
2
2
2
3
3
3
:1.51 -54
.53
.5l
e93
r9b
1.00
1.11
1.15
1.14
*7I
bP
.b9
1.02
1.03
C A G 1 SE
C A ra SE
2
2
3
3
.41
.*7
.IY
nu7
-96
l,OO
1.15
1.14
1.04
1.06
1.05 C A D V SE 2 3 I41 -16 1-31 1.1s 1.06
CA OY 5 2 3 151 .51 1.01 1.14 ,b 9 1.05 CA Y SE 2 3 .*7 146 1.01 1.15 .Ob 1.0h
CA 7 s 2 3 .51 .40 l.Ul 1.15 .bR 1.05 c4 neSE 2 3 a41 I 46 1.0; 1.1r .04 1.07
CA un s 2 3 -51 -49 1102 1.14 ab9 1.05 C 4 ER S C 2 3 .*l 145 1.05 1.15 .*4 1.07
C A ti4 S 2 3 .51 .4R 1.05 1.15 ,b8 1.06 rthorhombic c 4 T.1 SE 2 3 .41 .44 1.01 1.15 1.011
C I 19 5 2 3 .51 .&I 1.07 1.15 rbd L.01 type A C4 Y* 5E 2 3 * 41 si3 i .U* 1.10 1 .OH
CA I H 5 2 3 .51 .41 I.OM 1.10 .71 1.07 C A L l l SE 2 3 .L7 443 1.09 1.19 1.09
CL LU s 2 3 .51 .ab 1.09 1.19 b6 1.08 tructure 5" L4 BE 2 3 .57 .ER * 92 e01 1.01
CA sc s 2 3 -51 .I4 1.15 1-25 ,b3 1.09 nknown S Y CE SE 2 3 *53 .5* .Yd .U5 .n3 1.03
'14 LA 5 2 3 .Sl e h2 .92 1.07 -61 1.00 5 U PO SE 2 ? $53 -54 .99 1.0b .07 1.04
5H CF 5 2 3 .51 .5R -90 1.05 .b8 1.02 5R NO Sh 2 7 -53 .53 1.01 I .OD .03 1.06
5R PQ 5 2 3 -51 .5R .99 1.06 .67 1.02 SH S r SE 2 3 .53 1'1 1.05 1.06 .03 I-Ob
I R hrl S 2 3 .57 .51 1.01 1.06 a61 1.03 SR GO SE 2 3 .53 .49 1.09 1.10 .03 1.07
SR S U S 2 3 .57 -54 1.05 1.06 e61 1.05 i n SE 2 3 -53 947 1.13 1.OV 1.08
SI on s 2 3 -57 .53 1.09 i.10 -65 1.0b 5R uv SE 2 3 .53 .46 1.1) 1-09 1.09
S R 7R S 2 3 -57 041 1.13 1.09 .b5 1.01 TY Y SE 2 3 .53 .Lb 1.14 1.10 1.09 8rtnarhamt i c
SR OY 5 2 3 -51 .50 I.lb 1.09 b5 1.08 FU NO SE 2 3 *53 .46 I.lh 1.09 1.09 type B
SR V S 2 3 -57 ~ 5 0 1.11 1.10 .b5 1.00 5R E U SE 2 3 .53 .r5 1,IR 1.10 1.10
si? UO s 2 3 -57 .49 1.16 1-09 .b5 1.00 FR T " SE 2 3 .53 .&I 1.21 1.10 1.11
SP EU S 2 3 .57 .4R 1.18 1.10 .65 1.0'4 5 H V R SE 2 3 *53 .a3 1.22 1.05 1.11
SR
S I
TM
VH
5
5
2
2
3
3
-57
.57
.*7
.47
1.21
1.22
1.10
1.0s
.65
bO
1.10
1.10
SR LIJ SE
U A LA SL
2
2
3
3 ..s3
o*
.*3
.5B
1.24
1.10
1.13
1.05
,03
03
1.11
1.05
59
SR
LIJ
SC
5
S
2
2
3
3
e51
.
.51
.46
.*4
l.2b
1.30
1.13
1.19
a63
.b0
1.10
1.12
t r u c ture
nknown
HA CF: SL
A b PQ SE
2
2
3
3 .b4
b4
.54
.54
1.18
1.19
1rU3
I .u*
03
03
1.07
1,0n
RA
EA
EA
LA
CE
P9
5
S
5
2
2
2
3
3
3 .
.bY
bR

b8
.58
-58
h2 1.10
1.10
1.19
1.0s
1.03
1.04
a65
.bl
a66
].Ob
1-06
1.07 h3P4
I)A NO SL
Y A S r SE
H A GO SE
2
2
2
3
3
3
. .
* 84
04
ebb
a53
.51
9
.
1.21
1-26
1.30
l.Ut
1.04
l.UU
03
n3
03
1.08
1.10
1.11
0L NQ S 2 3 ,ha -57 1.21 1.04 .b6 1.07 HA T U SE 2 3 .64 .47 1.35 1.01 .n3 1,12
HA SU S 2 3 * bll 1.26 l.O+ ) r t ht oy rpheo m
B bic
8 A GO S 2 3 m b8
.54
-53 1.30 1.08
.6b
64
1.09
1.10
U A D I SE
3 1 Y LE $3,
s 06
a b4
.4b
,4b 1:: 1-01
1.08 03
103 I1.13
:!
8 1 14 S 2 3 .bH .*I 1.35 1.01 .64 1.12 BA MPI SE 2 3 h4 ~ 4 0 1-30 1.07
RA OV
R I Y
5
s
2 3
2 3
.ha
. 6 ~
.50
.90
1.31
1.31
1.07
1.00
*64
.
.b4
1.12
1.12
i(A E 9 SE
114 7 " SE
2
2
3
3 .
-64
b*
-45
.b4
1.42
1.65
1.08
1.01 103
1.14
1.15
wr P8T r
81
2 tl
S 2 3
ea
.be
60
h0 .
rn

.41
49
40 1.30
1.42
1.45
1
1-06
.01
1-08
b4
.64
b4
1.12
1.13
1.14
H A Y R SE
dA LU SE
2
2
3
3
.h4
.64
-43
.43
1.46
1-48
1.03
1.11
,r3
-03
1.15
1,1b
EA Y B S 2 3 a60 a41 1.46 1.03 -61 1.Ir
EA LU S 2 3 bR 146 1.48 1.11 e62 1.1s
S L sc s 2 3 1 bO *4+ 1.5h 1.16 .59 1.17
t r v c tu re
?known

The columns represent the ions A, B, X, valence state of A, valence of B, radius ratio
Q y8/'rx, radius ratio r,,!lrxs radius ratio Y,/YI,
electronegativity product xaxb,force constant Kat,,ratio K a b / K b ~and
, tolerance factor.

I
8 I
0.5
I
1.0
I
1.5
O-O.-> 'apb

0 s p i n e l ; h unknoxn o CaFe20q
e i:.i;erse s p i n e l e orthorhombic
0 Ag2Hg14 0 rhombohedral
@ Th p 0 s t r u c t u r e unknowri
3 4
Figure .f.-l'lot of Kciljt i s . Y ~ / ' Y I ~for chalcogenide compositions A B p & (X = Se, Se, T c ) for knonv and sornr unkno\rn structures.
Vol. 7, No. 5 , September 1968 O F AI32X4 STOICHIOMETRIES
CRYSTALLIZATION 1769

NtQ and N,Q be the number of atoms of type Q at- TABLE V


tached to n oxygen atoms in tetrahedral and in octa- POSSIBLE OF ION DISTRIBUTIONS
COM~INATIONS AROUND OXYGEN

hedral voids, respectively. Thus one-fourth of Q in a


tetrahedral void and one-sixth of Q in an octahedral
7
-

Tetrahedral
I11 FOR THE CASEn = 2
FOR SOLUTON
Oxygen 1-----
Octahedral
-------Oxygen
Tetrahedral
2----
Octahedral
-
void belong to each oxygen, and on the average one- elements elements elements elements
A B A B A B A B
fourth of A and one-half of B in compounds with the
0 0 0 0 0 2 3 3
formula AB204 are attached to an oxygen ion. The 0 1 3 -0
following equations can be derived 0 2 3 1
0 3 3 0
1 0 0
i 3
1 1 .0. -3
1 2 2 1
Thus 1 3 -
9 0
2 0 1 3
2 1 1 2
2 2 1 1
-0 3 1 o
3 0 0 3
where all N's and n's are positive integers. 3 1 0 2
The degree of inversion X of a spinel is related to 3 -3 0 1
no^ by 3 3 0 0
0 1 0 0 0 1 3 3
N~A 0 1 3 2
x=
+
N ~ A no^ 0
0
2
3
3
3
1
0
and usually No* 5 no^. Since for the inverse spinels 1 0 2 3
1 1 2 2
1 2 2 la
1 3 2 0
a This row represents arrangement for inverse spinel.
X can be also expressed as
TABLE VI
= )*(1 THECATION ARRANGEMENTSAROUND A N ANION FOR THE CASE
2 n n = 4 CORRESPONDING
TO THE DISTRIBUTION IN MgGanSOc
No. of anions in unit cell
For thecasen = 1 with same cation nearest No. of Mg ions in No. of Ga ions in

3Nt.4 + ~ N O=A 3 neighbors


16
octahedral voids
3
tetrahedral voids
1
3NtB -!r 2No~
= 6
8 1 2
8 2 2
and two solutions exist, (I) and (11). 16 0 3

NtA = 1 no^ = 0 and two A atoms and one B atom are associated in the
(1) octahedral voids. The other solutions are
NtB = 2 no^ = 0
The iigzHgI4 structure has this type of cation arrange- N ~=
A 0 no^ = 3
(IV)
ment. NtB = 0 no^ = 6
A 1 No* = 0
N ~=
(11)
NtB = 0 NoB = 3
The spinel structure and the olivine structure are ex-
amples of this type of cation arrangement.
For thecasen = 2
3Nta +2No~ 6 =
Generally the number of solutions for the case n = m
is given by
3NtB + 2 N o B 12 =
(Im/21 + + 1)
Six solutions exist among which two solutions give the
where [m/2] is m / 2 if m is even and [m/2]is (m - 1)/2
same cation distribution as (I) and (11) above, while a
if m is odd, and all solutions for n = m are contained in
third solution
the solutions for n = I , if and only if m is a divisor of 1.
There are several possible combinations of ion dis-
tributions around the two oxygens for each solution.
I n Table V the possible combinations for the third
describes the inverse spinel structure. One oxygen is solution are shown. Mathematically all configurations
associated with one B atom in the tetrahedral void, are consistent with respect to charge distributions and
A N D LIPPAKU
1770 BENNETT,COTTON,LEGZUINS, inorgimic Chemistry

stoichiometry, but a modifying principle can be intro- will be given by the expression I[lcm(4,n)/4], where 1
duced which will make some distributions more likely is an integer and the numerator is the least common
than others by assuming that the oxygen ions will pre- multiplier of 4 and n.
fer identical or nearly identical cation environments. It is quite possible that the local energy around oxy-
The ion distribution around oxygens characterized by gen due to the ionic charges becomes higher as n in-
0112 and 0121 is that of an inverse spinel. The cation creases. If n is small, e . g . , 1, then each oxygen ion has
arrangement of MgGazS418is an example of the case the same cation environment. For n = 2 the cation
n = 4, which is a partially inverted structure of solu- environment becomes different for the inverted spinel
tion V,and the cation arrangement is listed in Table VI. structure and these differences among oxygen ions
The number of integral formula weights per unit cell would become more pronounced as n increases. The
fact that spinel, olivine, and AgzHgI3 structures occur
(18) C . Romers, B. A. Blaisse, and D. J. W. Ijdo, Acta Cryst., 23, 634
so frequently indicates that structures in which the
(1967). local surroundings of anions differ greatly are unstable.

COSTRIBUTION FROM THE DEPARTMENT O F CHEMISTRY,


MASSACHUSETTS
1SSTITUTE OF TECHNOLOGY, CAMBRIDGE, MASSACHUSETTS 02139

The Crystal Structure of Cesium Tetrakis(hexafluoroacety1acetona to)yttrate(II1).


A Novel Stereoisomer Having Dodecahedral Eight-Coordination1
BY M. J. B E S K E T T , F. A. C O T T O S , P.LEGZDIr';S,2 ASD S. J. L I P P A R D

Received March 1 4 . 1968

T h e crystal and molecular structures of cesium tetrakis(hexafluoroacetylacetonato)yttrate(III), Cs [Y(CF3COCHCOCFs)a],


have been determined from 1650 independent, nonzero reflections collected with a counter diffractometer. T h e compouad
crystallizes in the orthorhombic space group Pbcn with four molecules in a unit cell of dimensions: a = 8.679 9z 0.005 A ,
b = 21.518 i 0.008 d, and c = 17.553 f 0.008 A ( p o b s d = 2.10 g cm-3; pcalcd = 2.12 g cmV3). T h e structure, excluding
hydrogen atoms, was solved from Patterson and electron density maps and refined by least-squares methods to a conventional
R factor of 6.57,. T h e molecules are monomeric y i t h the yttrium being dodecahedrally coordinated to the eight essentially
equivalent oxygen atoms (average Y-0, 2.323 (4) A ) . Both coplanar ligands show dihedral folding of 7.6" along the 0-0
line in the chelate rings. In contrast with other dodecahedral ML4 ( L = bidentate ligand) structures, each ligand in this
structure spans adjacent vertices betdeen the two bisphenoids constituting the dodecahedron, resulting in over-all D, sym-
metry. T h e monomeric anions are linked together in infinite columns parallel to the crystallographic a axis by a close Cs-F
association (3.2-3.7 A), but other intermolecular distances are not unusual. T h e structure is in accord with the mass spec-
tral evidence for the existence of a strong ion pair.

Introduction structural possibility, i t was decided to carry out a


During the course of our investigation of the solid- single-crystal X-ray analysis of Cs [Y(HFA)*], where
state and solution properties of several anionic eight- HFA represents the hexafluoroacetylacetonate ion,
coordinate P-diketonate complexes of yttrium(III),3 , 4 CFaCOCHCOCF3-.
i t was discovered that certain large but flexible organic Experimental Procedures
cations, such as (C6HB)4Asf,gave rise t o a polymor- Crystalline Cs[Y(HFA)h] was prepared by the reaction of
phism that was not observed for the simple inorganic Cs(HFA) with YC13 in aqueous ethanol as described previouslys
cations (Na+, K+, C S + ) . ~ It
> ~was considered possible and was recrystallized from the same solvent. Anal. Calcd
that this polymorphism might correspond to geometric for C S Y C ~ ~ H ~ OC, ~ ~ : H, 0.38; F, 43.43. Found:
~ F22.85;
isomerism for the eight-coordinate anions. This type C, 22.71; H, 0.46; F, 43.27. T h e analyses were performed by
Galbraith Microanalytical Laboratories, Knoxville, Tenn.
of isomerism, although considered to be a possibility,'j T h e well-formed, transparent, needle-like crystals showed
has never been observed in tetrachelate metal atom mmm morphology under optical examination, and preliminary
complexes.' As part of an investigation t o test this Weissenberg (Okl and lkl levels) and precession (h01, hll, hkO,
and h k l levels) photographs, taken with Cu K a radiation,
(1) Work supported by t h e National Science Foundation under Grant KO. confirmed this as the Laue symmetry. Moreover, the observed
7034X.
(2) Arthur D. Little Predoctoral Fellow, 1966-1967. systematic absences-0kZ for k # 2n, h01 for l # 2n, and hk0
(3) F. A. Cotton, P. Legzdins, and S. J. Lippard, J . Chem. Phys., 46, 3461 for h+ k # 2%-suggested the unique choice of the orthorhombic
(1966). space group Pbcn (no. 60). T h e following unit cell dimensions
(4) S. J. Lipyard, Proceedings of the 9 t h International Conference on Co-
ordination Chemistry, S t . Moritz, 1966, p 47.
were obtained, a t 22O, by a method described earlier,g with 28
( 5 ) H. Bauer, J. Blanc, and n. L. Ross, J . A m . Chem. Soc., 86, 5125 values measured on a General Electric XRD-5 manually oper-
(1964). ated single-crystal diffractometer using Cu K a radiation (X(Ka1)
(6) J. 1,. Hoard and J. V. Silverton, I m r g . Chem., 2, 235 (1963).
(7) For general discussions of the stereochemistry of eight-coordinate
complexes see: (a) S. J , Lippai-d, Progr. Ixorg. Chrm., 8, 109 (1967): (b) (81 S . J . Lippard, J . A m . Cheriz. Soc., 88, 4300 (1'366).
E, L. Muetterties and C. M. Wright, Quart. Rev. (London), 21, 109 (1967). ('3) M. J. Bennett, P. A. Cotton, and J. Takats, ibid.,90, U03 (1068).

You might also like