LiLa6I12Os, a substitutional derivative of rhombohedral La(La6I12Os)

Article abstract of DOI:10.1016/j.jssc.2004.08.031

The series of quaternary rare-earth-metal halide cluster compounds ALa ₆I₁₂Z with transition metal interstitials Z and alkali or alkaline-earth metal cations A has been expanded to include A=Li. The compounds synthesized by high-temperature solid-state techniques for Z=Os, Ir, Pt, Ru are isotypic with rhombohedral R₇X₁₂Z (R3, Z=3). The refined single X-ray crystal structure of (Li_0.967La _0.033)La₆I₁₂Os is reported, along with supportive results from a Rietveld analysis of neutron powder diffraction from a different sample, ⁷Li MAS-NMR, and electronic resistivity and magnetic susceptibility measurements. The samples show continuous Li _1-xLa_x cation compositions and are generally semiconductors, but their complex paramagnetic properties are not those of simple spin-only systems.

Full text of DOI:10.1016/j.jssc.2004.08.031

ARTICLE IN PRESS
Journal of Solid State Chemistry 177 (2004) 4465–4474
www.elsevier.com/locate/jssc
LiLa I Os,a substitutional derivative of rhombohedral La(La I Os)
6
12
6 12
Ã
Elizabeth A. Jensen,John D. Corbett
Department of Chemistry, Iowa State University, Room 353, Spedding Hall, Ames, IA 50011-3020, USA
Received 16 April 2004; received in revised form 2 August 2004; accepted 9 August 2004
Abstract
6
The series of quaternary rare-earth-metal halide cluster compounds ALa I12Z with transition metal interstitials Z and alkali or
alkaline-earth metal cations A has been expanded to include A=Li. The compounds synthesized by high-temperature solid-state
¯
techniques for Z=Os,Ir,Pt,Ru are isotypic with rhombohedral R X Z (R3; Z=3). The refined single X-ray crystal structure of
7
12
(
6
Li0.967La0.033)La I12Os is reported,along with supportive results from a Rietveld analysis of neutron powder diffraction from a
7
different sample, Li MAS-NMR,and electronic resistivity and magnetic susceptibility measurements. The samples show
continuous Li1 La cation compositions and are generally semiconductors,but their complex paramagnetic properties are not
those of simple spin-only systems.
ꢀx
x
r 2004 Elsevier Inc. All rights reserved.
6
Keywords: Rhombohedral; LiLa I12Os; Semiconductors
1
. Introduction
as well as main group elements (B,C,N,Si,etc.). The
structure, Fig. 1,consists of R trigonal antiprismatic (3)
¯
6
Reduced rare-earth-metal (R) halides form only three
structures that contain isolated R X (Z)-type clusters in
clusters that are only slightly distorted from ideal
octahedra,and have Z bound in their centers. All twelve
edges of each metal cluster are bridged by one of two
crystallographically distinct X. Six inner X1 atoms bridge
the R–R edges of the basal faces (as X ). The other six
edges around the cluster waist are bridged by bifunc-
tional X2 atoms that also form exo bonds to metal
i–a
vertices of adjacent clusters (X ). Reciprocally,the
6
12
which X=Cl,Br,I and Z is a centered transition metal
or other interstitial atom,namely with R X Z, R X Z,
and R X Z stoichiometries. Other cluster halides
7
12
6 10
i
1
2 17
contain condensed bioctahedral units,oligomers of four
octahedra,or infinite chains or layers of octahedra,all
of which are centered. The variety of bridging halide
modes possible leads to a vast array of compositions and
structures [1–3],and these can be further extended
through the inclusion of cations of the more active
metals.
metal atom vertices of each cluster are terminally bonded
to X2 atoms which bridge R–R waist edges of neighbor-
a–i
ing clusters as well (X ). The seventh R atom (crystal-
¯
lographically R2) is located on a 3 site halfway between
¯
The rhombohedral (R3) R X Z structure is the most
R Z clusters along c so that it is surrounded by six
6
7
12
versatile example,known for R=Sc,Y,and most of
La–Lu with the three heavier halides,although the
bromides have been considerably less well studied. These
compounds accept a great variety of interstitial elements
Z,including most of the transition metals in groups 7–11
trigonal antiprismatic X1 atoms from the clusters above
and below. A more detailed formulation of R X Z is
7
12
+
thus R [R (Z)(X1 ) (X2 )6/2(X2 )6/2
3
i
i–a
a–i
ꢀ3
]
[4,5].
6
6
Experience with other structures and compositions
has shown that introduction of alkali or alkaline-earth
metal cations and the corresponding electrons or halide
anions may produce new compounds with more reduced
or more open structural frameworks of the metal–halide
Ã
Corresponding author. Fax: +1-515-294-5718.
E-mail address: jcorbett@iastate.edu (J.D. Corbett).
0022-4596/$ - see front matter r 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2004.08.031

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4466
triiodide LaI was prepared by direct reaction of the
3
elements,purified by vacuum sublimation in contact
with only Ta [15],and stored in tightly capped vials in
the He-filled glovebox. Iodine (Fisher,99.8%) for this
purpose was used as received. Lanthanum (ꢁ99.99%)
was obtained from the Ames Laboratory in the form of
thin cold-rolled sheets,which were stored within the
glovebox,scraped with a scalpel to remove any surface
films (oxides or ‘‘tarnish’’),and cut into small pieces
immediately before use. Typically, ꢁ0.25 g of reactants
were loaded into Ta tubes that had previously been
cleaned,crimped,and welded on one end. Their open
ends were then tightly crimped,arc-welded under Ar,
and sealed into well-evacuated and baked fused silica
jackets before being heated in a furnace. Reactions
were generally allowed to take place around 900 1C for
Fig. 1. The structure of L0iLa
chains of La 12Os clusters. Small black,large black,large gray,and
white spheres represent Os,La,Li and I,respectively. In the parent
compounds,La 12Os,the Li position is occupied by La2.
6
I12Os,showing a [110] section of several
6
I
1
week.
Phases were identified and yields were estimated by
7
I
visual examination of Guinier X-ray powder diffraction
patterns (Enraf-Nonius FR-552) of the products and
comparisons with calculated patterns. Positions of the
indexed diffraction lines of known phases and those of
Si (NIST,SRM-640b,added as an internal standard)
were used in the calculation of lattice parameters and
their standard deviations by least-squares refinements.
Reported parameters were refined from X-ray powder
patterns composed of at least 90% of an ALa I Z-type
clusters. The cations mainly fulfill charge requirements,
as in the compounds AR I Z (A=K,Cs; R=La,Pr;
6
10
Z=Mn,Fe,Os) [6],Cs R I Z (R=Ce,Pr; Z=Co,Os)
4
6 13
[
the R2 position in R X Z structures by Ca resulted in
7],and Cs La I Co [8]. Earlier attempts to substitute
2 10 17 2
7
12
only partial substitution in (Ca0.65Pr0.35)Pr I Co,
6 12
(
ducts [5]. Another investigation yielded CsEr I C with
Ca1 Gd )Gd I Fe,and (Ca
Gd )Gd I Co pro-
1ꢀz z 6 12
ꢀy
y
6 12
6
12
6
12
a comparable stoichiometry,but with a different cation
position and differently interconnected clusters [9]. A
complete substitution of the isolated R2 atom in the
R X Z structure type was first discovered in the
product (as visually estimated) and with at least 10
indexed diffraction lines that matched. It should be
realized that only the overall composition of each
sample was known unless a (Li,La)La In Z structure
7
12
6
12
isotypic NaLa I Os [10] during attempts to synthesize
6
12
was refined as no fractionation of the products was
possible. It also became evident that an excess of the
lithium reagent was necessary in order to gain samples
close to LiLa I Z. The excess LiI (or other) was not
analogues of K La I Os [11]. We have recently
6 14
4
reported a corresponding series of new ALa I Z phases
6
12
with A=Na,Mg,Ca,Sr and some of their electrical and
magnetic properties [12,13]. The following details the
particular problems with and properties of the remain-
ing LiLa I Z phases.
6
12
identifiable.
The most common impurity observed was LaOI. This
phase appeared in almost every pattern,usually in
p10% yield,and presumably arose from adventitious
water evolved from the fused silica,oxygen in the Ta
tubing,and other impurity sources.
6
12
2
. Experimental section
Initially,reaction mixtures with stoichiometry LiLa 6
I Os were loaded. A number of samples were also
The general techniques,including the methods of
1
2
loading the reactions and running these within welded
Ta containers and Guinier powder diffraction with an
internal standard on the products,have been described
before [10,14]. All manipulations were carried out within
loaded for mixed cations on the isolated site,namely
Li La )La I Os. The values calculated for the A2
(
x
1ꢀx
6 12
contents (below) may of course have been affected by
errors in measuring alkali metal reactants,evaporative
losses during the welding process (rare),and or
stoichiometric deviations caused by formation of LaOI
during reaction (which sequesters a portion of the
loaded La and I).
N - or He-filled gloveboxes.
2
2
.1. Syntheses
Elemental Li (Alfa Aesar,99.9%) and Os or other Z
(
Alfa Aesar,99.95%) were stored in a He-filled glovebox
and used as received. The binary iodide LiI (Alfa Aesar,
9%) was dried and sublimed under dynamic vacuum
before storage in this glovebox. The rare-earth-metal
2.2. Crystallography
9
Irregular black crystals were mounted in thin-walled
glass capillaries with a small amount of degassed

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4467
Apiezon L grease,and Laue photos were taken to
determine their suitability. Good crystals were found in
the products of reactions with loaded composition
Li La I Os. Diffraction data were collected up to
powdered samples were sieved to obtain an average
particle diameter of 200 mm (40–60 mg),diluted with dry
3
chromatographic Al O (total volume ꢁ1 cm ) to
2
3
minimize contact between particles,and this was sealed
within an evacuated Pyrex ampoule about 5 cm long.
This method generally gives values correct within a
factor of two and good temperature dependencies
[21,22].
2
6 12
2
y=601 at room temperature with graphite-monochro-
mated MoKa radiation on a Rigaku AFC6R diffract-
ometer. Routine indexing and cell reduction readily gave
the R-centered hexagonal unit cell. The reflection data
were empirically corrected for absorption with the aid of
1
space groups: R3 (No. 148) or R3 (No. 146). The
0 c-scans. Systematic absences indicated two possible
2.6. Void space calculation
¯
structure was solved by direct methods [16] and refined
¯
The AVOID program (part of the XTEL Program
Library from Indiana University Molecular Structure
Center) was used to calculate void spaces in the
La I Os-type structure. Given the appropriate input
in the expected centrosymmetric group R3:
2
.3. Neutron powder diffraction
7
12
file of atomic positions and element identities,the
program calculates the larger voids in the crystal lattice
as well as distances to neighboring atoms. In each
iteration,the user fills the previous void with an atom
and then the program calculates the next largest void.
3
A 3 cm sample of a loaded composition Li La I Os
3
6 12
was used for neutron diffraction,which was carried out
on the General Purpose Powder Diffractometer (GPPD)
of the Intense Pulsed Neutron Source (IPNS) at
Argonne National Laboratory. Diffraction patterns
were collected at an angle of 145.91 at room temperature
with the sample enclosed in a capped cylindrical
vanadium container. The resulting data were refined
by the Rietveld method [17,18] using GSAS [19].
3. Results and discussion
3.1. Synthesis and structure
7
.4. Li MAS-NMR
2
The first problems encountered had to do with
demonstrating that attempts to substitute Li for La in
the seventh position in La (Z)I to give some version of
A small amount of a LiLa I Os sample was ground
12
6
7
12
and diluted with dry chromatographic Al O to total
2
Li La La ZI were successful and that there was no
x 6 12
3
1ꢀx
substitution of more Li in other interstitial sites in the
3
volume of ꢁ1 cm in an N -filled glovebox,and this was
2
then loaded into the NMR spinner capsule and tightly
capped. The capsule was spun at 4 kHz on a Bruker
MSL-300 spectrometer at room temperature while the
spectrum was measured. Powdered dry LiCl was used as
a reference.
lattice.
Previous refinements of the lattice constants of
various ALa I Os phases from Guinier X-ray powder
6
12
diffraction patterns showed that the sizes of the unit
cells reasonably increase as the crystal radius of A is
increased. The cell volume increase in these originates
almost entirely from a lengthening of c relative to a,as
the La I Z clusters move apart in that direction to
2
.5. Physical property measurements
6
12
Magnetic susceptibility measurements were per-
accommodate the larger cations,the separations of the
basal I2 remaining fairly constant. For compounds with
A=Na,Ca,Sr and Z=Os,the unit cell volumes were
formed on a Quantum Design MPMS SQUID magnet-
ometer. Unsieved powdered samples (40–70 mg) were
loaded into improved fused silica containers [20] in a
He-filled glove box. Magnetizations of the samples at 2
or 5 K were checked as a function of the applied field
3
˚
found to increase by about 25–70 A over that for
La I Os. [12] However,for A=Li, Z=Os,the unit cell
7
12
volume is indistinguishable from that of the parent
La I Z,making it impossible to establish the substitu-
(
0–6 T) to screen for possible magnetic impurities,but
7
12
+
tion of Li for La
3+
none was observed. Magnetic susceptibilities were
measured at 1,3,or 5.5 T between 5 and 400 K. The
data were corrected for the susceptibility of the contain-
er as well as for the standard diamagnetic core values.
The susceptibility of each sample was fitted to a
generalized Curie–Weiss equation with an additional
temperature-independent (van Vleck) term w0:
Bulk electrical resistivities were measured over the
temperature range of 100–300 K with an Hewlett-
Packard 4342A Q-meter operating at ꢁ34 MHz. The
cations by dimensional shifts in
their Guinier patterns alone (Table 1). Furthermore,it
can be seen that the unit cell volumes change much less
as the relatively large transition metal interstitials Z are
varied. For the lithium compounds with Z=Ru,Os,Ir,
3
˚
Pt,the volume increases only about 25 A ,2470 to
3
˚
2495 A ,with most of the increase taking place in
(Table 2).
a
Changes in dimensions with the amount of Li loaded
might be expected to be small too,but this analysis

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4468
Table 1
Refined unit cell constants for some Li
a,b
x
La
6
I
12Os compositions
˚
Volume
˚
Loaded
composition
No. of
lines
a(s) (A)
c(s) (A)
3
˚
(s) (A )
c
c
c
c
Li0.36La
Li0.71La
Li1.43La
Li2.85La
6
6
6
6
I
I
I
I
12Os
12Os
12Os
28
33
14
17
19
13
12
16.156(4)
16.146(5)
16.147(5)
16.182(5)
16.04(1)
10.992(5)
10.968(4)
10.972(4)
10.996(4)
11.132(8)
11.03(4)
2485(2)
2476(2)
2477(2)
2494(2)
2480(4)
2491(13)
2479(2)
12Os
d
Li
Li4.28La
Li La 12Os
3 6
La I12Os
c
6
I
12Os
16.15(3)
d
7
6
I
16.153(4)
10.971(4)
a
¯
Space group R3 (No.148), Z ¼ 3:
b
Dimensions from least-squares refinements of Guinier X-ray
powder diffraction patterns with Si as an internal standard; l ¼
(
1
:540598 A; 23 1C.
c
Loaded stoichiometry corrected for LiI ꢂ yH
2
O content.
d
Loaded with metallic Li.
Fig. 2. Unit cell volumes of loaded compositions Li
x
La
6
I
12Os (circles)
3
12Os (broken line; 2476.9(9) A ) (data in Table 1). Bars
˚
and of La
7
I
indicate 3s error levels.
Table 2
a
6
Refined unit cell constants for some (Li,La)La I12Z compounds
3
Volume (A )
metallic Li because of contact with traces of moisture.
The only other phase observed in these reactions was
LaOI (5–10%,if any).
˚
c (A)
˚
Compound
La 12Os
No. of lines
a (A)
7
I
33
20
33
16
22
16.146(2)
16.131(2)
16.146(5)
16.172(7)
16.187(4)
10.971(3)
10.960(1)
10.968(4)
10.957(9)
10.995(4)
2476.9(9)
2469.8(6)
2476(2)
2482(3)
2495(2)
b
LiLa
LiLa
LiLa
LiLa
6
6
6
6
I
I
I
I
12Ru
12Os
12Ir
The fact that the lattice constants of LiLa I Z are
6
12
indistinguishable from those of La I Z (for the same Z)
7 12
12Pt
made it very difficult to infer the insertion of Li in the
La2 site with certainty (other than the fact that LiI was
not an immediate product). However,additional evi-
dence for the incorporation of cations resulted from
comparison of the powder diffraction intensities of the
1 0 1 and 1 1 0 reflections,which are the lowest-angle
lines in the pattern (near 2y=101). For La I Os the
a
As in Table 1.
b
1 6
Loaded compositions were Li La I12Z.
(
i.e.,of Li La I Os samples,0 oxo1) was greatly
x
6 12
complicated by the discovery that the reactant LiI used
initially was,despite previous vacuum sublimation and
7
12
1 1 0 line is calculated to be more intense than the 1 0 1
line,whereas for LiLa I Os the opposite is true. The
glovebox storage,contaminated by LiI ꢂ yH O (yp3).
2
6 12
The stoichiometric parameter x was back-calculated
from the amount of contaminated ‘‘LiI’’ loaded in each
reaction,assuming that LaOI was the product and the
maximum y was 3,although y may not have been
constant throughout. Consequently,some subsequent
reactions were loaded with metallic Li (3:0pxp7:0),
therefore avoiding the impurity problem,and their
same pertains when heavier cations (i.e., A=Na,Ca,
Sr),substitute for La2 [12]. This of course assumes that
other factors,such as preferred orientation within the
samples,did not skew the observed intensities. However,
the single crystals examined were irregularly shaped,
which did not suggest any tendency toward preferred
orientation. Table 3 summarizes the results of an
assortment of such comparisons that do demonstrate
successful substitution of the light cation for La2 in the
isolated position. Naturally the calculated pattern of
results compared to those loaded with LiI ꢂ yH O (with
2
corrected x). The reaction product from loaded compo-
sition Li La I Os using Li metal did not show any
3
6 12
1
evidence of any other phase in its X-ray powder pattern.
The numerical results are presented in Table 2 and
shown in Fig. 2. No clear trend of unit cell dimensions
with x is indicated. Only three examples gave unit cell
volumes more than 3s different from that of La I Os,
La I Os (with no atom at 0,0, ),which is unknown for
6 12
2
this rhombohedral R X Z structure,would have
7
12
intensity ratios that were very similar to those of
LiLa I Os (e.g.,1.61 for the Li compound and 1.69
for the hypothetical empty one).
6
12
7
12
all three slightly larger than the parent compound,and
none had a volume less than that of the parent.
Reactions loaded with X4 equivalents of Li per formula
The occupancy problem was quantified by X-ray
refinement of a single crystal selected from a high
quality product of a sample loaded as Li La I Os. A
2
6 12
unit showed 15–25% LiI ꢂ yH O in the powder patterns
summary of some crystallographic parameters is given
in Table 4,and the resulting atom positions and
anisotropic displacement parameters are listed in
Table 5. Important interatomic distances and angles
2
of the products,while reactions with less did not.
Unexpectedly,a small amount of LiI ꢂ yH O was
2
observed in the products of some reactions loaded with

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4469
Table 3
Guinier X-ray powder pattern relative line intensities
are given in Table 6. Isotropic refinement of the isolated
1
a
position at 0,0,
as only La gave R1,w R2 values of
2
b
22.1%,63.8%,respectively,and a Ueq value for La2 that
was considerably greater than the other atoms’ Ueq.
When this position was refined as 100% Li,its Ueq
vanished but R1 and wR2 improved to 9.0%,27.0%,
respectively. The refinement of the Li occupancy (with
Ueq fixed at the average value of the other atoms) failed
to converge,but the constrained anisotropic refinement
of the position shared by Li and La2 resulted in R1,w R2
of 4.3%,9.3%,with occupancies of 96.7(1)% Li and
3.3(1)% La2. The Ueq for Li/La2 was about twice as
large as that of the other atoms,but the cavity may be a
bit oversized (see below). Since the Li–La2 site in the
end members of the series Li La La I Os are
Compound
(101)
(110)
(101)/(110)
La
La
La
7
I12Os
7
I12Os
6
I12Os
Calculated
Observed
Calculated
Calculated
Observed
Observed
10.20
10.01
26.90
26.40
15.97
33.29
28.60
17.30
15.90
16.40
9.98
0.36
0.58
1.69
1.61
1.60
1.66
c
LiLa
LiLa
6
I
12Os
12Os
6
I
Li
3
La
6
I
12Os
20.00
a
Error levels were not included in the intensity output; the data are
probably not better than 70.01.
Loaded compositions.
b
c
7
Same structure as La I12Os but with the La2 position unoccupied.
Table 4
X-ray crystallographic data for (Li0.967La0.033)La
x
1ꢀx
6 12
6
I12Os
evidently fully occupied,we presume all intermediate
systems are as well.
Space group, Z
a
R 3¯ (No. 148),3
˚
a, c (A)
16.147(5),10.972(4)
2477(2)
5.122
3
V (A )
˚
Table 6
3
˚
d
calc (g/cm )
Important interatomic distances and angles (A,deg) in (Li 0.967La0.033
)
ꢀ1
˚
La I12Os (A)
6
m (MoKa,mm )
Crystal dimensions (mm)
22.561
0.11 Â 0.10 Â 0.10
Rigaku AFC6
MoKa,0.71069
2y2o
7h, +k, 7l,60
2573
Atoms
d
Atoms
d
Diffractometer
˚
Radiation (A)
Scan mode
Os–La1
Li/La2–I1
2.8743(9)
3.108(1)
3.246(1)
3.250(1)
3.414(1)
La1–I1
3.224(1)
3.228(1)
4.050(2)
4.079(2)
4.316(1)
Li/La2–I1
La1–La1
La1–La1
Octants measured,2 y (max) (deg)
No. measured reflections
No. independent reflections, Rint (I40)
No. independent observed reflections
ðI ¼ 2sIÞ
a
La1–I2
a
La1–I2
1628,0.0697
766
b
c
La1–I2
Atoms
I1–I1
Atoms
Angle
Angle
No. of variables
34
Relative transmission factor range
Sec. extinction coeff. (10
Goodness of fit
0.7656–1.0000
4.5(5)
0.950
La1–Os–La1
La1–Os–La1
I1–Li/La2–I1
I1–Li/La2–I1
I1–Li/La2–I1
90.41(3)
89.59(3)
180.00(4)
87.94(3)
92.06(3)
Li/La2–I1–La1
La1–I1–La1
La1–I2–La1
La1–I2–La1
La1–I2–La1
89.61(3)
77.77(4)
77.79(4)
173.56(4)
95.91(4)
ꢀ5
)
R1,w R2 ðIX2sIÞ
R1,w R2 (all data)
Largest residual peaks (e/A )
0.0429,0.0710
0.1570,0.0928
3
˚
˚
1.66 (0.8 A from I1), ꢀ1.91
a
˚
0 A from Os)
Waist bridging I2.
Exo bonded I2.
Shortest I–I contact.
(
b
a
c
Refined cell constants from 14 Guinier lines, l=1.540562,23 1C.
Table 5
6
X-ray positional and anisotropic displacement parameters for (Li0.967La0.033)La I12Os
a
Atom
x
y
z
Occ.
B
eq
La1
I1
0.83818(6)
0.96155(8)
0.76503(8)
0
1
2
1
2
0.15261(8)
0.3245(1)
0.0230(2)
0.0308(2)
0.0375(3)
0.0129(3)
0.05(2)
1
0.0228(2)
0.0292(2)
0.0206(3)
0.0130(3)
0.05(2)
1
I2
ꢀ0.0019(1)
1
Os
Li
0
0
1
0.967(1)
La2
0
0.05
0.033(1)
0.05
Atom
U
11
U
22
U
33
U
23
U
13
U
12
La1
La1
I2
0.0228(5)
0.0292(5)
0.0206(4)
0.0130(4)
0.0228(4)
0.0329(5)
0.0372(6)
0.0242(4)
0.0270(5)
0.0402(7)
0.0129(7)
–0.0050(4)
0.0090(5)
0.0130(6)
0
–0.0048(4)
–0.0036(5)
–0.0078(5)
0
0.0117(4)
0.0130(4)
0.0036(4)
1
2
Os
U
U
U
11
11
11
U11
1
2
Li
0.05(2)
0.05
0.05(3)
0.05
0
0
0
0
U11
1
2
La2
U11
a
Hexagonal setting.

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4470
The foregoing efforts do not establish whether Li may
also be occupying interstitial sites elsewhere in the
lattice. Therefore,the void spaces in the structure of
LiLa I Os were calculated using atomic coordinates
6
12
from the single crystal X-ray refinement,initially with
1
the 0,0, position empty. The first iteration of the
program showed that this position is the most regular
2
and largest available. It has six I1 neighbors,all at
˚
3
.11 A,a reasonable Li–I distance (the sum of crystal
+
–
˚
radii of six-coordinate Li and I is 2.96 A [23]). A
second void is unsymmetrical,with two I1 and two I2
˚
nearest neighbors at 3.12 A as well as one I1 and one I2
˚
next-nearest neighbors at 3.22 A. All other voids have
four or fewer I neighbors and are smaller,with
separations from neighboring iodines less than the sum
+
ꢀ
˚
of crystal radii of four-coordinate Li and I (2.79 A).
The six largest voids are summarized in Table 7. If a
second Li site exists in this structure,it would most
likely be in the second void,but no suitable residual was
observed at that position according to X-ray or neutron
diffraction. These calculations did not take into account
possible effects of the Li cation on positions of the
Fig. 3. Neutron powder diffraction pattern at 2y=145.91 for the
sample refined as Li0.57La0.43(La6I12Os). In the top portion the black
crosses are: observed data,the solid gray line: calculated pattern.
Vertical ticks beneath mark the calculated positions of diffraction
6 2 3
peaks for: (Li0.57La0.43)La I12Os,LaOI,LiI ?H O,and LaI ,respec-
tively,and the difference curve is plotted at the bottom.
neighboring atoms in the (Li0.967La0.033)La I Os struc-
6 12
ture. There is no lanthanide example of rhombohedral
1
Some diffraction peaks were not accounted for by the
calculated (Li0.57La0.43)La I Os pattern,and contam-
‘
‘R X Z’’ (without an atom at 0,0, ) available for
6
12
6 12
2
comparison,although there are zirconium analogues,
e.g.,Zr Cl H [24].
ination by impurity phases was suspected. The Guinier
X-ray powder pattern taken before neutron diffraction
6
12
Fig. 3 shows the neutron powder diffraction pattern
at room temperature collected from a composition of
Li La I Os. The refinement began with the symmetry
and atomic parameters previously established by single
crystal X-ray diffraction in the space group R,with the
occupancy of the isolated position set to 100% Li.
showed ꢁ10% of LaOI (P4/nmm) and 5–10% LaI
3
(P6 /m). These two phases were added to the refinement
3
and accounted for some of the less intense diffraction
peaks. Although not seen in the X-ray powder pattern,
3
6 12
LiI ꢂ yH O (Pm3m) was also observed in the neutron
2
refinement (p5%),presumably because of the more
efficient scattering of neutrons by these relatively light
atoms. About five peaks,the most intense of which was
ꢁ100 counts/ms (located at ꢁ14 ms) remained unex-
plained. The final isotropic R and R were 8.2% and
Isotropic refinement gave unacceptably large U for Li
iso
2
0.40 A ),about 20 times larger than the average U for
˚
(
iso
the other atoms. A fractional La atom was allowed to
share the Li position,and the occupancy refined to
wp
p
2
7(3)% Li,43(3)% La2 with U =0.062(2) A . Sum-
2
˚
5
5.3%,respectively,while the large reduced w value was
probably a result of the failure to fully account for all of
the peaks in the spectrum.
iso
maries of data collection and isotropic refinement
parameters and structural results are given in Tables 8
and 9. An attempted anisotropic refinement was not
well-conditioned.
Most of the results of the neutron refinement were
consistent with results of X-ray diffraction. Table 10
compares the distances. The isotropic displacement
parameters from neutron vs. X-ray refinements are also
comparable. Some of the distance differences in Table
Table 7
Largest calculated void spaces in LiLa
1
0 must arise from the ꢁ40% difference in the lithium
6
I12Os
vs. lanthanum occupancies. The distances within the
structures are consistent with those found in structural
Void
x
y
z
Neighbors
a
1
2
b
analyses of other A(Li X Z) isotypes [11–14]. Subtle
6 12
˚
6 I at 3.11 A
1
0
0
effects of the alteration of the A atom are also observed.
For example,comparison of the present Li vs. Na [11]
results,the closest relative in the group,shows consistent
˚
4 I at 3.12; 2 I at 3.22 A
2
3
4
5
6
0.77
0.55
0.10
0.15
0
0.09
0.59
0.74
0.62
0
0.50
0.42
0.43
0.27
0.74
˚
3 I at 2.69 A
˚
2 I at 2.65; 2 I at 2.82 A
˚
2 I at 2.61; I at 2.69 A
˚
3 La1 at 2.60; I at 2.74 A
˚
differences. From Li to Na, d(A–I1) increases 0.21 A in
˚
comparison with 0.12 A in Shannon’s CN6 crystal radii
a
[23] for the cations. The difference presumably reflects
the presence of greater covalent shortening for lithium.
The La2 position.
b
˚
Uncertainty in the distance values are probably 70.01 A or less.

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4471
7
previous static Li NMR study of cubic
˚
A parallel effect shows in 0.03–0.04 A increases in
A
d(La1–I1) on the basal faces. All other changes are of
˚
˚
distance La1–I2 of 0.04 A,which parallels the larger
Li Zr Cl Mn [25],a similar compound made up of
2 6 15
the order of 70.01 A except for an increase in the exo-
Mn-centered Zr clusters surrounded by edge-bridging
6
i
a–a
Cl atoms and interconnected by Cl bridges showed a
triplet spectrum centered at –0.4 ppm (vs. solid LiCl).
The Li in the Zr compound has six Cl neighbors
arranged nearly octahedrally (D2d symmetry),but the
site is only one-third occupied. The greater chemical
shift in the Zr compound,as compared to the
˚
increase in a (0.20 A) vs. c (0.07 A) for this change.
˚
3
.2. Physical properties
7
Room temperature Li MAS-NMR data from a
sample of powdered LiLa I Os showed a single sharp
LiLa I12Os sample,is probably the result of decreased
6
6
12
peak centered at ꢀ10.22 ppm (vs. solid LiCl) accom-
panied by a pair of spinning sidebands (Fig. 4). The
single resonance confirms a single Li position of high
symmetry in the compound,consistent with the single
crystal picture in which Li is at the center of a (nearly
octahedral) trigonal antiprismatic arrangement of iodine
atoms. The absence of any other peaks in the spectrum
is further evidence that there are no other competitive
sites in the structure significantly occupied by Li.
shielding because of the more electronegative Cl
neighbors.
Bulk resistivities of three powdered compositions
were measured by the ‘‘Q’’ method,La
I12Os and two
7
Li La 12Os compositions. All of the samples showed
I
6
x
decreasing resistivity with increasing temperature,in-
dicative of semiconducting behavior,as the example in
Fig. 5 shows,and room temperature resistivities ranged
between 133 and 204 mO cm (Table 11). The electrical
7
Fig. 4. Li NMR of LiLa
6
I12Os sample between 7150 ppm with LiCl(s) as reference. The single peak is centered at ꢀ10.22 ppm.

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4472
Table 8
Neutron crystallographic data for the powdered sample loaded as
Li La 12Os
3
6
I
Space group, Z
˚
a, c (A)
R 3¯ ; 3
16.1709(5),10.9981(7)
2490.7(2)
5.220
3
V (A )
˚
3
d
calc (g/cm )
Absorption coefficient
Diffractometer
0.032752(5)
IPNS-GPPD
145.9
Detector bank,2 y (deg)
Histogram scale factor
3.62(2)
R
wp, R
p
0.0817,0.0533
4596
3.053
Number of data points
2
Reduced w
Diffractometer zero point
Refined background and (pseudo-Voigt) profile coefficients.
ꢀ9.09(5)
Background coefficients
Profile coefficients
6
Fig. 5. The electrical resistivity of a sample of Li0.5La I12Os by the Q
1
2
3
4
5
=0.199(4)
=0.0052(5)
a ¼ 3ð3Þ
method. The extrapolated resistivity at 298 K is 172 mO cm and the
b ¼ 0:0403ð9Þ
0
mean fractional temperature dependence (dr=dT ꢂ r) is ꢀ0.013 Kꢀ
1
.
=ꢀ0.00026(5)
=0.000012(2)
=ꢀ0.00000027(5)
b ¼ 0:006ð1Þ
1
g ¼ 3:1ð6Þ
1
s1 ¼ 109ð8Þ
s2 ¼ ꢀ0:2ð10Þ
Table 11
Electrical resistivities of ALa
6
I12Z compounds by the ‘‘Q’’ method
a
Loaded composition
r
298 K (mO cm)
Table 9
La
Li0.5La
Li La
7
I
12Os
12Os
12Os
167,169,184
172,173
133,204
Positional parameters for Li0.57La0.43La
6
I12Os refined from neutron
a
6
I
powder diffraction data
2
6
I
2
˚
Atom Fraction
x
y
z
U
iso (A )
a
Multiple values correspond to repeat experiments on the same
sample.
La1
Os
I1
1
0.8821(5)
0
0.8376(4)
0
0.6155(10) 0.4644(8)
0.1519(7)
0
0.0230(9)
0.0304(3)
1
1
ꢀ0.008(1) 0.0241(2)
I2
Li
1
0.57(3)
0.6863(8)
0
0.7630(10) ꢀ0.005(1) 0.0355(2)
1
0
0
0.062(2)
0.062(2)
insulators or poor semiconductors. One example is
La I Fe ,which contains pairs of iodine-sheathed
La Fe octahedra in which some of the iodine atoms
2
1
2
La2
0.43(3)
0
12 17
2
a
6
¯
Space group R3 (No.148).
i–I
bridge one edge in each of two adjacent clusters (I )
i–a–a
and others have unusual I
functionality,bridging an
edge of one cluster and forming terminal exo-bonds to
two other clusters [26]. Its resistivity showed a negative
temperature coefficient and an ꢁ350 mO cm value at
room temperature,though this may have been an
underestimate owing to a LaI2 impurity. In the
anomalous metallic a-K R I Os (R=La,Pr),the
Table 10
Selected interatomic distances in (Li
a
˚
12Os structures (A)
x
La1ꢀx)La
Atoms
2.884(2) 2.8743(9) La1–I1
6
I
Atoms
Neutron X-ray
Neutron X-ray
Os–La1 ( Â 6)
3.245(5) 3.224(1)
3.269(5) 3.228(1)
3.262(3) 3.246(1) La1–La1 4.071(2) 4.050(2)
Li/La2–I1 ( Â 6) 3.158(5) 3.108(1) La1–I1
4
6 14
b
i–a
La1–I2
La1–I2
La1–I2
octahedral clusters are interconnected by planar I
b
c
3.266(3) 3.250(1) La1–La1 4.085(2) 4.079(2)
d
3.374(3) 3.414(1) I1–I2
bridges to form two-dimensional networks. These net-
works give rise to low and substantially temperature-
independent bulk resistivities of ꢁ120–240 mO cm,in the
range of a poor metal [11]. Given the structure of the
present compounds,there seems to be no reason to
expect other than semiconducting properties. The metal
clusters are linked to adjacent clusters three dimension-
ally by noncoplanar iodine bridges,and the metal
cations are isolated from the clusters as well as from
each other by iodine atoms.
4.370(2) 4.316(1)
a
Room temperature data. Different refined occupancies of Li,0.967
in X-ray sample,0.57 in neutron diffraction sample,are presumably
reflected in the distance differences as well.
b
Waist bridging I2.
Exo bonded I2.
c
d
Shortest I–I contact.
resistivities of most noncondensed ternary and quatern-
ary lanthanide halide cluster compounds have not been
reported,but those known have nearly always been
Relatively few magnetic susceptibility studies of these
cluster halides have been reported,although the simple

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4473
MO scheme put forth in early studies [27] suggested
relatively simple spin-only results might pertain,the
electronic configurations being readily altered via
changes in the electron counts of different interstitials.
Magnetic susceptibilities were measured for four differ-
ent Li La I Os compositions at a field strength of
either 1 or 3 T and between ꢁ5 and 300 K. Magnetiza-
tion data as a function of field showed no evidence of
ferromagnetic impurities in any of the samples. The
susceptibility data for each sample were fitted to a
presumed Curie–Weiss equation modified by an addi-
tional temperature-independent paramagnetic (van
more complicated susceptibilities presumably arise from
correspondingly more complex magnetic structures.
In one explanation,the ground state of the cluster can
be described as Curie–Weiss paramagnetic at lower
temperatures,but at higher temperatures,or under the
influence of the external magnetic field,one or more
excited states become electronically populated,resulting
in mixtures of ground and excited state(s) paramagnetic
characteristics. Spin–orbit coupling between La cluster
atoms and the heavy transition metal Z is also a real
possibility. To explore this further,the susceptibilities of
a sample of the same Li La I Os loaded composition
x
6 12
2
6 12
Vleck) term w . The results are collected in Table 12.
Most samples showed small y values but significant
deviations from a Curie–Weiss description,especially at
was measured at both 1 and 5.5 T. Both measurements
showed similar Curie–Weiss behavior at low tempera-
tures with similar magnetic moments (Table 12) and
nearly temperature-independent susceptibility at higher
temperature, Fig. 7,though neither displayed the abrupt
transition that had been observed in others at 3 T. The
behavior of the susceptibilities at 5.5 T was less clear
above about 100 K,giving an array of widely scattered
points (possibly indicating that the sample shifted inside
its container),whereas the data for the 1 T measurement
continued to fall along a well-defined although non-
linear curve. In addition,none of the effective magnetic
moments calculated at lower temperature (Table 12) is
0
1
00–200 K and above,as typified by that shown in
Fig. 6. These often quite sharp deviations appear in 1=w
vs. o plots as regions of different slope and do not seem
to depend either on the particular elements or the formal
cluster electron counts. Since all of the samples were
obtained with high yield (95–100%,estimated by
powder X-ray diffraction patterns) of the ALa I Z
6
12
product,and none showed evidence of ferromagnetic
impurities,the possibility of deviations arising from
sample contamination seems low. Instead,the observed
Table 12
Curie–Weiss parameters and calculated effective magnetic moments for Li
x
La
6
I12Os products
ꢀ
4
ꢀ4
Loaded composition
Field (T)
Temperature (K)
w0; 10 (emu/mol)
C,(emu K/mol)
y (K)
meff (m
B
)
w300; 10 (emu/mol)
Li0.5La
6
I
12Os
3
6–300
2–350
6–300
2–350
2–350
2–320
5.95(6)
8.77(4)
6.27(2)
9.16(5)
5.91(2)
9.8(1)
0.0302(6)
0.0270(1)
0.0146(2)
0.0128(3)
0.0146(2)
0.0208(3)
ꢀ5.1(3)
ꢀ2.26(9)
–4.8(2)
0.492
0.465
0.342
0.320
0.341
0.408
7.38(6)
9.79(4)
6.93(2)
9.54(5)
6.34(2)
9.7(1)
LiLa
6
I12Os
1
Li
2
Li
2
Li
2
Li
3
La
La
La
La
6
I
6
I
6
I
6
I
12Os
12Os
12Os
12Os
3
1
ꢀ2.4(1)
ꢀ5.1(1)
ꢀ1.92(7)
5.5
1
Fig. 6. Magnetic susceptibilities measured at 3 T for a sample loaded
as Li La 12Os. Circles and triangles represent w and 1=w; respectively,
2
6
I
Fig. 7. Magnetic susceptibility of an Li
2
La
6
I12Os sample at 1 T
each corrected by a van Vleck term (w ),and the line marks the best
Curie–Weiss fits to all data.
(circles) and 5.5 T (triangles). Fits to the modified Curie–Weiss
equation (right scale) are shown by the straight lines.
0

ARTICLE IN PRESS
E.A. Jensen, J.D. Corbett / Journal of Solid State Chemistry 177 (2004) 4465–4474
4474
consistent with the ideal spin-only values, ꢁ1.7 m and
and Robert Henning for the neutron diffraction data
collection and refinement,and Shu Xu for the NMR
measurements. This work was supported by the Na-
tional Science Foundation—Solid State Chemistry—
via Grants DMR-9510278,-9809850,and -0129785,
and was carried out in the facilities of the Ames
Laboratory,US Department of Energy.
B
ꢁ
2.9 m for one and two unpaired electrons,respec-
B
tively,rather all of the meff values fall in the range
.342–0.492 m . These numbers indicate that a spin-only
0
B
description is not appropriate for these phases,although
some past work involving similar clusters has resulted in
seemingly good agreements. Compounds with isolated
(
closer to the expected spin-only values include Zr I Mn
M,R) X -type clusters that gave m
values much
6
12
eff
6
12
References
[
27] (1.84(7) m ),Sc I C [28] (1.8(1) m ),Sc I C [4]
B
6 11
2
B
7 12
(
isostructural with the ALa I Z series,1.7 m ),
6
12
B
[
1] A. Simon,H.j. Mattausch,G.J. Miller,W. Bauhofer,R.K.
Kremer,in: K.A. Gschneidner,L. Eyring (Eds.),Handbook on
the Physics and Chemistry of Rare Earths,vol. 15,Elsevier
Science Publishers,Amsterdam,1991,p. 191.
CsZr I C [29] (1.48 m ),La I Fe [26] (1.11(2) mB),
and a-K La I Os [11] (0.77 m ). All of these examples
were predicted to have a single unpaired electron per
cluster. In addition,CsLa I Z [6] with isolated R X
6 12
clusters,would have one unpaired electron for Z=Fe
and two for Z=Mn,consistent with the observed meff
6
14
B
12 17
2
4
6 14
B
[2] J.D. Corbett,J. Alloys Compd. 229 (1995) 10.
[3] M. Lulei,J.D. Martin,L.M. Hoistad,J.D. Corbett,J. Am. Chem.
Soc. 119 (1997) 513.
6
10
[
[
[
4] D.S. Dudis,J.D. Corbett,S.J. Hwu,Inorg. Chem. 25 (1986) 3434.
5] T. Hughbanks,J.D. Corbett,Inorg. Chem. 27 (1988) 2022.
6] M. Lulei,J.D. Corbett,Z. Anorg. Allg. Chem. 622 (1996) 1677.
values of 1.4 and 2.9 m ,respectively. None of the above
B
was reported to have deviations in magnetic suscept-
ibility at higher temperatures comparable to those found
here except for isostructural Sc I C [4] for which a
[7] M. Lulei,J.D. Corbett,Inorg. Chem. 35 (1996) 4084.
[
[
8] M. Lulei,P.A. Maggard,J.D. Corbett,Angew. Chem. Int. Ed.
Engl. 35 (1996) 1704.
9] H.M. Artelt,T. Schleid,G. Meyer,Z. Anorg. Allg. Chem. 618
7
12
flattening at ꢁ300 K was postulated to result from
‘thermal population of neighboring levels’’.
‘
(
1992) 18.
[
[
[
[
[
[
10] S. Uma,J.D. Corbett,J. Solid State Chem. 161 (2001) 161.
11] S. Uma,J.D. Martin,J.D. Corbett,Inorg. Chem. 38 (1999) 3825.
12] E.A. Jensen,J.D. Corbett,Inorg. Chem. 41 (2002) 6199.
13] E.A. Jensen,J.D. Corbett,J. Solid State Chem. 172 (2003) 132.
14] E.A. Jensen,J.D. Corbett,J. Solid State Chem. 144 (1999) 175.
15] J.D. Corbett,Inorg. Synth. 22 (1983) 31.
4
. Conclusions
The quaternary lanthanide halide cluster compounds
Li1ꢀx
temperature solid-state techniques in the rhombohedral
La I Os (0oxo1) have been synthesized by high-
6 12
[16] SHELXTL,Version 5.1; Bruker AXS: Madison,WI,1998.
[17] H.M. Rietveld,J. Appl. Crystallogr. 2 (1969) 65.
[
18] R.A. Young,The Rietveld Method,Oxford University,New
York,1993.
structure type known for many R(R X Z) phases.
6
Single crystals of the compounds were structurally
characterized to show that Li occupies the isolated R2
12
[
19] A.C. Larson,R.B. Von Dreele,GSAS: General Structure
Analysis System,Low Alamos National Laboratory,Los Alamos,
NM,1986–1994.
position halfway between R I Z clusters along c. The
6
12
compounds are semiconductors according to isotropic
electrical resistivity measurements. However,the mag-
netic susceptibility data are complex,yielding low meff
values relative to their ideal formulations and deviations
from linearity above ꢁ100 K.
[20] A.M. Guloy,J.D. Corbett,Inorg. Chem. 35 (1996) 4669.
[21] J. Shinar,B. Dehner,B.J. Beaudry,D.T. Peterson,Phys. Rev. B
3
7 (1988) 2066.
[
[
[
[
[
22] J.-T. Zhao,J.D. Corbett,Inorg. Chem. 34 (1995) 378.
23] R.D. Shannon,Acta Crystallogr. A 32 (1976) 751.
24] R.P. Ziebarth,J.D. Corbett,J. Solid State Chem. 80 (1989) 56.
25] J. Zhang,J.D. Corbett,Inorg. Chem. 30 (1991) 431.
26] M. Lulei,J.D. Martin,J.D. Corbett,J. Solid State Chem. 125
(
1996) 249.
27] T. Hughbanks,G. Rosenthal,J.D. Corbett,J. Am. Chem. Soc.
10 (1988) 1511.
28] D.S. Dudis,J.D. Corbett,Inorg. Chem. 26 (1987) 1933.
Acknowledgments
[
1
The authors thank Jerome Ostenson and Serguei
Bud’ko for the magnetic measurements,Chris Murphy
[
[29] J.D. Smith,J.D. Corbett,J. Am. Chem. Soc. 107 (1985) 5704.