Layered lanthanum carbide halide superconductors La2C2(X,X′)2 (X,X′ = Cl, Br, I): Neutron powder diffraction characterization and electronic properties

Article abstract of DOI:10.1021/jp984646d

The superconducting properties of La₂C₂(X, X′)₂(X, X′ = Cl, Br, I) are investigated. Three different samples of La₂C₂Br₂, which are obtained under different synthesis conditions, show superconductivity at 7.03(5) K. La₂C₂I₂ (T_c≈1.7 K) exhibits slightly different transition temperatures for differently prepared samples. A total of five samples of La₂C₂(X, X′)₂ were investigated by high resolution neutron powder diffraction from room temperature to 1.5 K. Rietveld refinements yield the structural parameters, in particular of the C atoms. These allow us to follow in detail the variation of the interatomic distance in the C₂ group. On the basis of the low temperature structural parameters, extended Hueckel band structure calculations were performed and a band of low dispersion along the Γ-N direction in the Brillouin zone was identified. We discuss correlations of structural and electronic properties with the superconducting transition temperatures of the La₂C₂(X, X′)₂ phases.

Full text of DOI:10.1021/jp984646d

5446
J. Phys. Chem. B 1999, 103, 5446-5453
The Layered Lanthanum Carbide Halide Superconductors La2C2(X,X′)2 (X,X′ ) Cl, Br, I):
Neutron Powder Diffraction Characterization and Electronic Properties
Kyungsoo Ahn, Brendan J. Gibson, Reinhard K. Kremer, Hansju1rgen Mattausch,
Andres Stolovits,† and Arndt Simon*
Max-Planck-Institut fu¨r Festko¨rperforschung, Heisenbergstrasse 1, D-70569 Stuttgart
ReceiVed: December 7, 1998; In Final Form: April 20, 1999
The superconducting properties of La2C2(X, X′)2(X, X′ ) Cl, Br, I) are investigated. Three different samples
of La2C2Br2, which are obtained under different synthesis conditions, show superconductivity at 7.03(5) K.
La2C2I2 (Tc ≈ 1.7 K) exhibits slightly different transition temperatures for differently prepared samples. A
total of five samples of La2C2(X, X′)2 were investigated by high resolution neutron powder diffraction from
room temperature to 1.5 K. Rietveld refinements yield the structural parameters, in particular of the C atoms.
These allow us to follow in detail the variation of the interatomic distance in the C2 group. On the basis of
the low temperature structural parameters, extended Hu¨ckel band structure calculations were performed and
a band of low dispersion along the Γ-N direction in the Brillouin zone was identified. We discuss correlations
of structural and electronic properties with the superconducting transition temperatures of the La2C2(X, X′)2
phases.
I. Introduction
there is no indication of a decrease of Tc when 13C is substituted
into Y2C2Br2.9,14,15
In the course of our search for new superconductors, we found
superconductivity up to 11.6 K in the layered rare-earth metal
carbide halides RE2C2X2 (RE ) Y, La; X ) Cl, Br, I).1 They
crystallize in the Gd2C2Br2 structure types.2,3 Therein C-C units
are located in the octahedral voids in close packed metal atom
double layers. The latter are sandwiched between layers of
halogen atoms. Such slabs are connected via the van der Waals
interaction. Different stacking sequences (1s and 3s stacking
variants) are found (Figure 1). Superconductivity in layered
halide phases was reported for alkali metal intercalated Hf and
Zr nitride chlorides reaching transition temperatures up to 25.5
K.4-6
The chemical bonding of the C-C dumbbell located in a
(distorted) octahedral cage of metal atoms in the crystal structure
of RE2C2X2 phases was suggested to play an important role for
superconductivity. Metallic character of the rare-earth carbide
halides results from covalent overlap of the C2-π* and the metal
atom d-orbitals.16 Neutron powder diffraction measurements on
Y2C2X2 (X ) Br, I)9,17 revealed a slightly reduced C-C double
bond distance. Raman spectroscopy on crystallites of a series
of samples of RE2C2X2 (RE ) Y, Gd; X ) Br, I) proves that
the specific stretching and tilting vibrations of the C-C group
are well decoupled from the vibrations of the heavy metal atoms
forming the surrounding cage.17 Band structure calculations for
Gd2C2Cl216 and Y2C2Br27 point to a band of low dispersion along
a certain direction in the Brillouin zone. This “flat” band being
located close to the Fermi energy reflects the anisotropy of the
crystal structure. Nonlinearities in the temperature dependence
of the 13C-NMR Korringa relaxation rate in Y2C2Br2 and Y2C2I2
were explained within the scope of this band structure scheme.13
In a first study on the La2C2X2 (X) Cl, Br, I) system, we
experienced a somewhat different behavior to that found for
the Y2C2X2 compounds.18 In contrast to the behavior of the
yttrium system, samples of La2C2Br2-xIx show lower Tcs when
the iodine content x is raised. La2C2I2 does not show super-
conductivity above 2 K, while mixed I-Cl phases are super-
conductors with Tcs up to 3.7 K. La2C2Cl2 itself could not be
synthesized yet.
Considerable work has been devoted to an investigation of
the superconducting properties of the yttrium-based phases of
the rare-earth metal carbide halides.7-13 The experiments done
on the Y2C2X2 system have revealed a number of essential
experimental facts. For example, it has been demonstrated that
by a proper adjustment of the ratio of Br:I ≈ 1:3, the
superconducting transition temperature in samples of Y2C2(Br,
I)2 can be maximized to 11.6 K.7,12 An increase of the transition
temperature of Y2C2Br2 (Tc ) 5.05 K) by about 1.2 K can also
be achieved by an intercalation of Na atoms into the van der
Waals gap between the double layers of halogen atoms.11,12
A
comparison with the superconducting properties of the dicarbide
YC2 (CaC2 structure type) has shown that the pronounced
structural anisotropy in the yttrium carbide halides leads to a
significant increase of the electron-phonon coupling strength.9,10
The upper critical fields in the layered yttrium carbide halides
are enlarged by almost two orders of magnitude as compared
to YC2. The carbon isotope effect in YC2 and Y2C2Br2 is, as
well, surprisingly different: While a 12C/l3C isotope effect in
good agreement with BCS theory has been detected in YC2,
Here, we present a detailed investigation of the low temper-
ature structural and superconducting properties of La2C2Br2,
La2C2I2 (1s and 3s type), La2C2Br0.5I1.5, and La2C2Cl0.5I1.5. The
low temperature crystal structural properties were studied
employing high resolution neutron powder diffraction. Neutron
diffraction was particularly useful to gain precise positional
parameters of the carbon atoms. Subsequently, the refined
structural parameters were used as input for an extended Hu¨ckel
tight-binding calculation of the band structures.
* Prof. Dr. A. Simon: e-mail, ahn@simaix.mpi-stuttgart.mpg.de.
† Permanent address: Institute of Physics, Tartu University, Riia 142
EE-2400 Tartu, Estonia.
10.1021/jp984646d CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/15/1999

Layered Lanthanum Carbide Halide Superconductors
J. Phys. Chem. B, Vol. 103, No. 26, 1999 5447
Figure 1. Perspective view of the crystal structures of La2C2I2 (1s stacking variant left and 3s stacking variant right). I, La, and C atoms are drawn
with decreasing diameter, respectively. Stacking sequences are indicated and unit cells outlined.
II. Experimental Details
with an MPMS SQUID magnetometer between 2 and 10 K in
an external field of 1 mT. The samples were contained in dried
quartz glass ampoules under 1 bar He exchange gas to provide
sufficient thermal contact. The sample containers were designed
to give a negligible magnetic background signal.
A. Sample Preparation. Samples of La2C2X2 (X) Br, I)
were prepared from La metal (Johnson Matthey Inc., 99.99 %),
LaX3, and glassy carbon powder in ≈10 g batches. The trihalides
LaX3 (X ) Cl, Br) were prepared from La2O3 (Johnson Matthey
Inc., 99.9%) and NH4X according to procedures described
elsewhere,19,20 while LaI3 was prepared from La metal and I2.21
All trihalides were purified by repeated distillations in high
vacuum. The carbon powder (Aldrich, 99.99%) was heated and
degassed (1050 °C, 10-5 mbar) for 3 days and subsequently
stored in a dried argon atmosphere. In order to prepare 1s-La2
C2X2, stoichiometric mixtures with a 2% excess of carbon
powder were sealed in welded Ta tubes and heated to 980 °C
and 1050 °C in the case of X ) Br and X ) I, respectively.
The product for X ) I contained up to ≈10% of 3s-La2C2I2.
An increased carbon excess (≈10%) was found necessary to
obtain the 3s-La2C2I2 type. Halogen mixed phases were obtained
from appropriate mixtures of the metal, the trihalides, and 2%
carbon excess. The reaction period was about two weeks for
all samples. Two samples (1 and 2) of La2C2Br2 were prepared
from LaBr3 and La2C3. Details of the preparation of La2C3 are
given elsewhere.18 We used 3% less carbon for sample 1 and
3% excess carbon for sample 2. Sample 3 was obtained from a
reaction of La metal, LaBr3, and 15% excess carbon powder.
Since the trihalides and the reaction products are sensitive to
moisture, all handling was done under dried argon gas atmo-
sphere. The reactions yield polycrystalline material with brown-
ish luster for La2C2Br2 and with bronze color for the iodide
and the mixed phases. The 3s-La2C2I2 samples were found to
form into thin platelets up to 2 mm in diameter.
Below 2 K ac magnetic susceptibilities were determined in a
home-built top-loading single-shot 3He bath glass cryostat. The
ac coils being immersed in the pumped 4He outer bath and held
at constant temperature were operated at a frequency of 37 Hz
and generated ac fields of about 0.3 mT. The design of the
cryostat allows rapid mounting of the samples and is therefore
particularly suited for measurements on very air and moisture
sensitive rare-earth metal carbide halides. Samples of about 200
mg were pressed into cylindrical pellets to reduce the active
surface of the samples. To protect the pellets during mounting
they were enclosed in gelatin capsules (5 mm diameter, 12 mm
length) which had been desiccated in high vacuum for 24 h.
3
The temperature of the He bath was determined with a
calibrated Ge resistor (Lake Shore) being in close contact with
the sample. The coil setup was calibrated against a standard Al
sample.
D. X-Ray Powder Diffraction. The product phases were
characterized using powder X-ray diffraction measurements
either with the modified Guinier technique22 or a STOE powder
diffractometer. Calibration of the lattice parameters (see Table
1) was done with Si as an external standard.
E. Neutron Powder Diffraction. Neutron powder diffraction
measurements were performed with the D1A powder diffrac-
tometer (λ ) 191.1 pm) at the ILL (Grenoble, France). High
resolution patterns (6° < 2Θ < 160°, steps of 0.05°) were
collected on microcrystalline powder samples of about 10 g
which were either encapsulated in thin-walled vanadium con-
tainers (i.d. 8 mm) or in quartz glass tubes (o.d. 8 mm, wall
thickness < 0.5 mm). Dried helium at 1 bar was used in the
ampoules as heat exchange gas.
Structure refinements from the powder patterns were carried
out with the FAT-Rietan 97 program.23 From the 2659 data
points containing 216 contributing reflections, 27 parameters
were refined, of which 13 were global (1 instrumental, 6
background parameters, 5 profile parameters, 1 asymmetry
parameter) and 14 local (1 overall scale factor, 4 lattice
B. Resistivity Measurements. Electrical resistivity was
determined on sintered pellets of 5 mm diameter and a thickness
of about 1 mm by the van der Pauw method in fields up to 7 T
using the cryostat of an MPMS magnetometer (Quantum
Design). Currents of 10 mA (dc) or less were supplied by a
Keithley 2400 sourcemeter. The voltage drop across the contacts
was determined by a HP 34420A nanovoltmeter. The pellets
were enclosed into a vacuum tight copper can and pressed onto
four gold plated spring pins.
C. Magnetic Susceptibility Measurements. The magnetic
susceptibilities of powder samples (≈100 mg) were measured

5448 J. Phys. Chem. B, Vol. 103, No. 26, 1999
Ahn et al.
TABLE 1: Lattice Parameters at Room Temperature According to X-Ray Powder Diffraction Measurements of All
Investigated Samples
no.
compd
type
a (pm)
b (pm)
c (pm)
â (deg)
1
2
3
4
5
6
7
8
La2C2Br2
La2C2Br2
La2C2Br2
La2C2I2
1s
1s
1s
1s
3s
1s
1s
1s
745.1(3)
749.6(3)
747.4(1)
764.4(3)
762.4(2)
760.4(1)
758.9(1)
760.5(5)
405.6(1)
405.9(1)
406.08(8)
413.8(1)
414.1(1)
410.9(1)
408.61(5)
408.4(1)
1002.6(2)
1007.0(2)
1005.1(1)
1077.6(3)
1090.3(3)
1071.4(1)
1075.2(2)
1061.8(3)
94.31(3)
94.12(4)
94.27(2)
93.10(4)
100.74(4)
93.40(1)
93.42(1)
93.14(3)
La2C2I2
La2C2Br0.5I1.5
La2C2Cl0.5I1.5
La2C2BrI
TABLE 2: Parameters Used for the Extended Hu1ckel Band
Structure Calculations
element orbital
Hii
ú1a
ú2a
c1b
c2b
La
6s
6p
5d
5s
5p
2s
2p
-7.67 2.140
-5.01 2.080
-8.21 3.780 1.381 0.7765 0.4586
I
-18.0
-12.7
-21.4
-11.4
2.679
2.322
1.625
1.625
C
a Coefficients used in double ú expansion. b Slater-type orbital
exponents.
TABLE 3: Superconducting Transition Temperatures of the
Investigated La2C2X2 Samples Determined from dc and ac
Susceptibility Measurements (Tocn: onset temperature; Tmc id
midpoint temperature)
:
no.
compd
type method
Tocn (K)
6.97(5)
7.01(5)
7.03(5)
1.60(5)
1.72(5)
1.65(3)
1.82(3)
Tmc id (K)
1
2
3
4
5
6
7
8
La2C2Br2
La2C2Br2
La2C2Br2
La2C2I2
1s
1s
1s
1s
3s
1s
1s
1s
dc
dc
dc
ac
ac
ac
ac
ac
6.0(1)
6.3(1)
6.83(5)
1.55
1.65
1.09
La2C2I2
La2C2Br0.5I1.5
La2C2Cl0.5I1.5
La2C2BrI
1.70
1.0
1.45(8) (2.3(1))
constants, 6 positional, and 3 isotropic displacement parameters).
Profile matching was achieved using a pseudo-Voigt function.
III. Band Structure Calculation
Figure 2. Electrical resistivity (a) and dc magnetic susceptibility (b)
of La2C2Br2 (sample 2). The inset in (a) displays the high temperature
resistivity. In (b) fc and zfc refer to field cooled and zero-field cooled
susceptibility, respectively.
The extended Hu¨ckel method24,25 was used for the band
structure calculations on La2C2I2 using the crystal structure
parameters as refined from the powder neutron diffraction
measurements (Table 2). The calculations were performed on
a 588 k-point mesh for the monoclinic cell. The electronic
density of states obtained from the band structure was smoothed
with Gaussians of a half-width of 0.02 eV. Band dispersion was
calculated for selected directions in the Brillouin zone for 40
discrete k-points in each direction.
temperatures (see the inset Figure 2a) which may hint at an
27
additional scattering mechanism, e.g., a narrow band near EF
which is also evidenced from the band structure calculations
(see below).
Figure 3 displays the electrical resistivity in magnetic fields
of a sample of La2C2Br2 (sample 3). Magnetic fields shift the
superconducting transition to lower temperatures. The absolute
resistivity is by a factor of 400 larger than in sample 2, and
zero resistivity is reached at a somewhat lower temperature
(6.60(5) K). This noticeable difference between the two samples
is attributed to the considerable amount of excess carbon in
sample 3.
Using a 10% criterion as a measure for the upper critical field,
we estimate Bc2(0 K) by an extrapolation with an even poly-
nomial of fourth degree to be 5.6(2) T and 4.3(2) T for samples
3 and 2, respectively (see Figure 4). According to these results
the upper critical field of La2C2Br2 is somewhat higher than of
isoelectronic Y2C2Br2 (3.0(2) T, Tc ) 5.05 K).9
IV. Results and Discussion
A. Superconducting Properties. As verified on three dif-
ferent samples, 1s-La2C2Br2 becomes superconducting below
7.03(5) K (onset temperature Tocn from dc susceptibility, see
Table 3). Figure 2 displays the electrical resistivity and the dc
susceptibilities of a 1s-La2C2Br2 (sample 2) which is monophasic
according to a neutron powder diffraction study.26 Zero resistiv-
ity is reached at 7.03(5) K. Diamagnetic shielding is complete
while the Meissner fraction is only about 10% of the expected
value. This difference is frequently observed and may be
attributed to strong pinning effects at grain boundaries or crystal
imperfections in the individual sample. The electrical resistivity
shows a remarkable convex temperature dependence at high
We note the positive curvature of the phase boundary line of
both samples close to Tc(0). This behavior is similar to results

Layered Lanthanum Carbide Halide Superconductors
J. Phys. Chem. B, Vol. 103, No. 26, 1999 5449
Figure 6. ac susceptibility of a sample of (a) La2C2Br0.5I1.5, (b)
La2C2Cl0.5I1.5, and (c) La2C2BrI (sample 6, 7, and 8, respectively).
Figure 3. Resistivity of sample 3 measured in external magnetic fields
as indicated.
Figure 7. Transition temperatures versus the mean ionic radius jr of
the halogen constituents in the systems La2C2(X, X′)2(X ) Cl, Br, I).
4 give the results of this work; data represented by b were taken from
ref 18.
Figure 4. Upper critical field of samples 2 (O) and 3 (b) determined
from a 10% drop criterion from the normal resistivity. Full lines are
fits to an even polynomial in Tc (see text).
slight difference in the onset temperature ∆Tocn ≈ 0.12(5) K.
Flux exclusion is complete at lowest temperatures for both
samples as can be concluded from the comparison with an Al
standard sample.
Figure 6 displays the ac susceptibilities of three samples with
mixed halogen constituents (samples 6, 7, and 8 in Table 3),
which are all of the 1s stacking type. The onset transition
temperatures for samples 7 and 8 are higher than for 1s-La2
C2I2. This finding confirms the trend established before that
replacement of I by Br or Cl raises Tc.18
The superconducting transition temperature of a sample of
La2C2BrI (sample 8) which had been investigated before and
had been found at 2.7 K, according to the present study is
somewhat lower and appears at 2.32(5) K.18 However, a second,
even sharper transition with onset at about 1.5 K and midpoint
at ≈1 K is seen which could indicate that this sample is
inhomogeneous even though impurity phases were not visible
in the X-ray powder diffraction pattern. There is also indication
of an inhomogeneity of sample 7 not obvious from the structure
investigation. However, the bulk exhibits a sharp transition.
Figure 5. ac susceptibility of a sample of 1s-La2C2I2 (O, not sintered
pellet; ], sintered pellet) and 3s-La2C2I2 (+) (sample 4 and 5,
respectively).
found for other highly anisotropic systems such as intercalation
28
compounds of TaS2 or the organic superconductor (BEDT-
TTF)2Cu(NCS)2 with BEDT-TTF being bis(ethylenedithio)-
tetrathiafulvalene.29 It has also been seen in high Tc cuprate
superconductors.30
Figure 7 displays the transition temperatures of the systems
La2C2(X,X′)2 as a function of the averaged halide ion radius.
Our results extend the data range of our previous publication
and confirm the monotonic decrease with increasing halide ion
radius. The good correlation of the experimental data into a
narrow band supports our finding for the Y2C2(X,X′)2 system
that, to first approximation, the averaged halide ion radius may
be taken as a suitable parameter to describe the superconducting
transition temperature in the RE2C2X2 compounds. In contrast
to the Y2C2(X,X′)2 system, however, we observe no local
maximum in the superconducting transition temperatures of the
La2C2(X,X′)2 compounds.7,12 If we assume such a maximum
Figure 5 displays the ac susceptibilities of samples of 1s-
La2C2I2 and 3s-La2C2I2 (samples 4 and 5 in Table 3). The
transition to superconductivity occurs at temperatures below 2
K. This is why these transitions had been missed in our previous
investigations.18 For both samples the transitions to supercon-
ductivity are remarkably sharp. The transition width was
considerably reduced by additionally sintering the pressed
pellets. In contrast to La2C2Br2, the two differently prepared
samples, exhibiting 1s and 3s stacking, respectively, show a

5450 J. Phys. Chem. B, Vol. 103, No. 26, 1999
Ahn et al.
Figure 8. Observed (b), calculated (full line), and difference pattern for La2C2I2 (3s type) at 1.5 K. Vertical bars indicate the positions of the
reflections used to simulate the pattern. The broad feature in the background at low angles is due to diffuse scattering from the quartz glass
container.
Figure 9. Observed (b), calculated (full line), and difference pattern for La2C2Cl0.5I1.5 (1s type) at 1.5 K. Vertical bars indicate the positions of the
reflections used to simulate the pattern. The broad feature in the background at low angles is due to diffuse scattering from the quartz glass
container.
TABLE 4: Positional Parameters for La2C2X2 (X ) Br, I) as
Refined from the Neutron Powder Diffraction Patterns
Collected at 1.5 Ka
3s- Gd2C2Br2 structure type. These first calculations proved all
compounds to adopt the 1s type, except La2C2I2, which
crystallizes with both structure variants. Refinements of the
patterns of two different samples of La2C2I2 revealed exclusively
the 3s type for one sample but a phase mixture of 0.92:0.08 for
the 1s:3s type for the second sample. Single phase refinements
of all patterns were carried out where appropriate (see Figures
8 and 9).
no.
compd
atom
x
z
Beq (104 pm2)
La
Br
C
0.1447(4) 0.1567(4)
0.1717(5) 0.6554(5)
0.4280(4) 0.0364(5)
0.5(1)
0.5(1)
1.5(1)
2
1s-La2C2Br2
La
I
C
0.1425(3) 0.1396(4)
0.1752(5) 0.6573(4)
0.4258(4) 0.0292(3)
0.2(1)
0.2(1)
0.2(1)
4
5
1s-La2C2I2
3s-La2C2I2
All samples showed no evidence for impurity phases, e.g.,
starting materials or decomposition products above the level of
sensitivity of powder diffraction measurements. For mixed
halide samples the occupation factors were also refined, and
the resulting composition was found to be in good agreement
with the values expected from the initial ratio of the trihalides
in the starting mixtures (see Table 5). Most likely, because of
the significant ion size mismatch and the concomitant local
lattice stress, the isotropic thermal displacement parameters for
La
I
C
0.4069(2) 0.1408(4)
0.7952(4) 0.3428(3)
0.0819(3) 0.0284(2)
0.1(1)
0.2(1)
0.4(1)
a The Rwp factors (Rp factors) amounted to 5.9% (4.6%), 10.8%
(8.5%), and 5.6% (4.3%) for ls-La2C2Br2 (no. 2), 1s-La2C2I2 (no. 4),
and 3s-La2C2I2 (no. 5), respectively.
to exist, the transition temperatures of the La2C2(X,X′)2
compounds already lie on the descending branch above the local
maximum.
B. Neutron Diffraction. The neutron powder diffraction
patterns were refined on the basis of the space group C2/m (no.
12) with all atoms in position 4i in Wyckoff notation (cf. Tables
4 and 5) of the structure for Gd2C2Br2, either in its 1s or 3s
stacking variant.2,3 Initially, the Rietveld refinements had been
started assuming the simultaneous presence of the 1s- and the
the halogen atoms in the sample of La2C2Cl0.51(3) 1.49(3) were
I
found to be unusually large. On cooling from room temperature
to 1.5 K, apart from a monotonic decrease of the lattice
parameters (cf. Table 6 and 7), no indications for structural phase
transitions were observed in the diffraction patterns of all
samples. The C-C distance within error bars is independent of
temperature. The experience with the preparation of 1s- and
3s-La2C2I2 suggests the possibility that the C2 group might, to

Layered Lanthanum Carbide Halide Superconductors
J. Phys. Chem. B, Vol. 103, No. 26, 1999 5451
TABLE 5: Positional Parameters and Occupation Factors for La2C2X0.5I1.5 (X ) Cl, Br) as Refined from the Neutron Powder
Diffraction Patterns Collected at 1.5 Ka
no.
6
compd
atom
occup
x
z
Beq (104 pm2)
La
I
Br
C
1
0.1434(2)
0.1750(3)
0.1750(3)
0.4264(3)
0.1419(1)
0.6615(3)
0.6615(3)
0.0297(2)
0.4(1)
0.8(1)
0.8(1)
0.6(1)
1s-La2C2Br0.5I1.5
0.74(3)
0.26(3)
1
La
Cl
I
1
0.1449(3)
0.1730(3)
0.1730(4)
0.4280(3)
0.1409(2)
0.6674(3)
0.6674(3)
0.0298(2)
0.5(1)
1.9(1)
1.9(1)
0.8(1)
7
1s-La2C2Cl0.5I1.5
0.74(2)
0.26(2)
1
C
a The Rwp factors (Rp factors) amounted to 8.8% (6.8%) and 4.8% (3.8%) for 1s-La2C2Br0.5I1.5 and for 1s-La2C2Cl0.5I1.5, respectively.
TABLE 6: Temperature Dependence of the Lattice
Parameters of La2C2Cl0.5I1.5 (no. 7) According to Neutron
Powder Diffraction
T (K)
a (pm)
b (pm)
c (pm)
â (deg)
dC-C (pm)
1.5 758.57(3) 409.81(2) 1067.82(5) 93.272(2) 131.5(4)
10
293
758.55(3) 409.81(2) 1067.76(5) 93.273(2) 131.9(4)
761.09(4) 411.01(2) 1072.34(6) 93.268(3) 131.6(5)
some extend, be replaced by single C atoms which would
occupy the centers of the metal atom octahedra. Rietveld
refinements assuming this structure model, however, failed to
reveal scattering density in the center of the octahedra beyond
the level of sensitivity of the method. According to the
calculations, the replacement of C2 by C amounts to 0.3(6)%.
It has been pointed out in previous investigations7,17 that the
C-C bond length in the RE2C2X2 phases is considerably
reduced as compared to that commonly found for a C-C double
bond observed, e.g., in ethene (dC-C ) 133 pm).31 However,
the C-C bond length in the La carbide halides spans a
somewhat wider range than found in the Y carbide halides
Y2C2X2 (X ) Br,I) phases (see Table 7).
Figure 10. Mean La to halogen atom distances dLa-Xh versus the
distances dLa-C of the axial La atom to the carbon atoms in the phases
La2C2(X, X′)2. (a: 1s-La2C2I2 (Tc ) 1.60(5) K); b: 1s-La2C2Br0.5I1.5
(Tc ) 1.65(3) K); c: 1s-La2C2Cl0.5I1.5 (Tc ) 1.82(3) K); d: 1s-La2C2Br2
(Tc ) 7.03(5) K); e: 3s-La2C2I2 (Tc ) 1.72(5) K)).
A central question we pursued with our investigations was
to reveal correlations between the structural and the supercon-
ducting (viz. electronic) properties. According to our results
there is, however, no direct correlation of the transition
temperatures with the C-C distances. As a general trend we
rather observe an increase of the superconducting transition
temperatures with decreasing distance of the axial (end-on to
C2) La atoms to closest halogen atoms in the crystal structure
and also with the distance of the axial La atoms to the carbon
atoms (Figure 10). All 1s type samples come to lie very close
to a straight line in a dLa-Xh versus dLa-C plot, while the change
of the stacking variant from 1s to 3s in the case of La2C2I2
gives rise to a much increased dLa-C distance. A close correlation
of the superconducting transition temperatures with the mean
ionic radius of the halogen constituents has already been reported
for the phases Y2C2(X,X′)2 and La2C2(X, X′)2 (X, X′ ) Cl, Br,
I),7,18 whereas a correlation of the transition temperatures with
the La-C interatomic distances has not been described before.
Figure 10, in addition, indicates to first order a linear dependence
of the interatomic distances of the axial La to the halogen atoms
and of the distances of the La to the C atoms.
Figure 11. (a) Band dispersion for 3s-La2C2I2 in selected directions
of the Brillouin zone and (b) corresponding density of states DOS
(Fermi level at -9.2 eV).
approximated by the formula (RE3+)2C24-(X-)2, suggesting
semiconducting behavior. From the molecular point of view,
the metallic behavior has been attributed to backbonding from
occupied C2-π* to empty d-states of the metal atoms.7,16
The metallic character of La2C2X2 (X ) Br, I) is verified by
the results of our extended Hu¨ckel calculations. Figure 11
displays the total density of states (DOS) calculated for the
observed crystal structure of the 3s type of La2C2I2. From the
C. Band Structure Calculations. Within the Zintl-Klemm
concept, the chemical bonding for the phases RE2C2X2 may be
TABLE 7: Lattice Parameters and C-C Distances at 1.5 K According to Neutron Powder Diffraction of All Investigated
Samples
no.
compd
type
a (pm)
b (pm)
c (pm)
â (deg)
dC-C (pm)
2
4
5
6
7
La2C2Br2
La2C2I2
La2C2I2
La2C2Br0.5I1.5
La2C2Cl0.5I1.5
1s
1s
3s
1s
1s
746.09(5)
761.90(3)
761.32(2)
758.57(3)
757.64(3)
404.93(3)
412.52(2)
413.24(1)
409.81(2)
407.58(2)
1001.16(9)
1075.13(7)
1085.90(4)
1067.82(5)
1072.66(5)
94.163(5)
93.143(4)
100.835(2)
93.272(2)
93.384(3)
134.1(7)
132.2(5)
128.4(4)
131.5(4)
129.7(5)

5452 J. Phys. Chem. B, Vol. 103, No. 26, 1999
Ahn et al.
The difference in the transition temperatures of 1s type and
3s type La2C2I2 is unprecedented in the Y2C2X2 phases. There,
the Tcs of mixed halide samples, Y2C2(X,X′)2, pass over two
changes of the stacking sequence without showing any jumps.
This finding was considered as strong evidence for the assump-
tion that the X-Y-C2-Y-X slabs themselves are in the first
place responsible for superconductivity in RE carbide halides.
The difference in Tc for the 1s- and 3s-La2C2I2 samples could
point to an additional sensitivity of the electronic conditions in
the I-La-C2-La-I slabs to packing effects in the halogen
atom double layer. However, the observation that pure 3s type
samples of La2C2I2 can only be prepared using a considerable
excess of carbon while mixed 1s/3s type samples result when
the carbon excess is reduced close to stoichiometric mixtures
strongly indicates that the 1s type samples exhibit a carbon
deficiency, La2C2-xI2. Such a substoichiometric carbon content
which is below detection level may easily be achieved by a
substitution of C42- groups by isoelectronic C4- anions without
noticeably altering the crystal structure. As experienced in the
case of Y2C2I2, partial replacement of C2 by C1 leads to a
marginal decrease of the lattice parameters only, while it may
induce a marked reduction of Tc.7
As a central result we can identify, from the band structure
calculations, a band of low dispersion in the direction (Γ-N)
which corresponds to the direction of the crystallographic a axis.
We emphasize that the flat band originates from orbital
interactions within the layer and not from interactions perpen-
dicular to the layers. Hence the flat/steep band scenario in the
band structure proposed to be essential for the RE2C2X2
superconductors is not a trivial consequence of the layered
character of the crystal structure of these phases.7
Figure 12. Density of states (DOS) and band dispersion for (a) 3s-
La2C2I2 and (b) 1s-La2C2I2 plotted for the Γ-N (-1/2, 1/2, 0) direction
in the Brillouin zone.
atomic orbital projections we assign the character of the bands
as follows: (1) bands between -12.9 and -14.2 eV are mostly
of I-p character and some contributions from C2-π* levels;
(2) a strong mixing of C2-π* levels (40%) and La-d orbitals
(60%) occurs between -10.8 eV and -5.3 eV; and finally (3)
the broad metal d bands extend from -5.3 eV upwards (not
plotted in the figure). Figure 11 shows the band dispersion
curves plotted over some selected directions in the Brillouin
zone. Along the direction Γ-A, the bands exhibit negligible
dispersion which may be expected from the two-dimensional
characteristics of the crystal structure. The local maximum in
the DOS around EF, however, rather originates from bands of
low dispersion along Γ-N () -1/2, 1/2, 0), (Figure 12). The
direction Γ-N corresponds to the direction of the crystal-
lographic a axis or the projection of the C-C group onto the
crystallographic a-b plane. This in-plane band of low dispersion
should be essential for superconductivity in these phases.7
From the Rietveld refinement results it is found that especially
the 1s and 3s stacking variant of La2C2I2 show a difference in
the C-C bond length. C-C distances of 128.4(4) pm and 132.2-
(5) pm are found for 3s- and ls-La2C2I2, respectively (Table 7).
The variation of the La-La distances of the 3s and 1s form
does not exceed 0.3%. The expansion of the C2 unit may be
expected to lower the energy of the C2-π* states and to affect
the overlap of antibonding C2-π* with La-d states. A calculation
of the electronic band structure of 1s-La2C2I2 seems to indicate
that these structural changes indeed influence the position
relative to EF of the low dispersive bands along Γ-N, while
we observe no significant alteration of the bands along the Γ-A
direction. As a consequence of the different C and La-C
interaction the DOSs near EF calculated for identical composition
(see section V) exhibit differences for 1s-La2C2I2 and 3s-La2C2I2.
The Fermi level for 1s-La2C2I2 comes to lie in a local minimum,
while EF for 3s-La2C2I2 is shifted towards a local maximum.
The different DOS for 3s-La2C2I2 as compared to 1s-La2C2I2
could be the origin of the increased superconducting transition
temperature of 3s-La2C2I2 as observed in the experiments.
Volume dependent band structure calculations reveal a subtle
flattening of the low dispersive band along Γ-N when the cell
volume is decreased, e.g., by introducing larger halogen
constituents (Figure 7).
Acknowledgment. We thank E. Suard and P. Cross at the
ILL for their kind assistance in the neutron experiment, E.
Bru¨cher and G. Siegle for the resistivity and susceptibility
measurements, W. Ro¨thenbach for the X-ray powder diffraction
measurements.
References and Notes
(1) Simon, A.; Mattausch, Hj.; Eger R.; Kremer, R. K. Angew. Chem.,
Int. Ed. Engl. 1991, 30, 1188.
(2) Schwanitz-Schu¨ller, U.; Simon, A. Z. Naturforsch. B 1985, 40, 710.
(3) Mattausch, Hj.; Kremer, R. K.; Eger, R.; Simon, A. Z. Anorg. Allg.
Chem. 1992, 609, 7.
(4) Yamanaka, S.; Hotehama, K.; Kawaji, H. Nature 1998, 392, 580.
(5) Kawaji, H.; Hotehama, K.; Yamanaka, S. Chem. Mater. 1997, 9,
2127.
(6) Yamanaka, S.; Kawaji, H.; Hotehama, K.; Ohashi, M. AdV. Mater.
1996, 8, 771.
(7) Simon, A.; Yoshiasa, A.; Ba¨cker, M.; Henn, R. W.; Felser, C.;
Kremer, R. K.; Mattausch, Hj. Z. Anorg. Allg. Chem. 1996, 622, 123.
(8) Simon, A. Angew. Chem. 1997, 109, 1879; Angew. Chem., Int. Ed.
Engl. 1997, 36, 1788.
(9) Henn, R. W. Dissertation Thesis, University Karlsruhe (TH), 1996.
(10) Henn, R. W.; Schnelle, W.; Kremer, R. K.; Simon, A. Phys. ReV.
Lett. 1996, 77, 374.
(11) Ba¨cker, M.; Simon, A.; Kremer, R. K.; Mattausch, Hj.; Dron-
skowski, R.; Rouxel, J. Angew. Chem., Int. Ed. Engl. 1996, 35, 752.
(12) Ba¨cker, M. Dissertation Thesis, University Stuttgart, 1997.
V. Conclusions
The neutron powder diffraction investigations prove very
helpful in establishing phase purity of these new superconduc-
tors. Refinement of the diffraction patterns of phase-pure
samples also allows to improve the accuracy of the positional
parameters of C atoms, as compared to X-ray data. The onset
of superconductivity in La2C2Br2 is found close to 7 K in a set
of three samples, irrespective of details in the sample prepara-
tion.
(13) Herrling, E.; Dormann, E.; Henn, R. W.; Kremer, R. K.; Simon,
A. Synth. Met. 1998, 92, 13.
(14) Gulden, Th. Dissertation Thesis, University Stuttgart, 1997.
(15) Gulden, Th.; Henn, R. W.; Jepsen, O.; Kremer, R. K.; Schnelle,
W.; Simon, A.; Felser, C. Phys. ReV. 1997, B56, 9021.
(16) Miller, G. J.; Burdett, J. K.; Schwarz, C.; Simon, A. Inorg. Chem.
1986, 25, 4437.
(17) Henn, R. W.; Strach, T.; Kremer, R. K.; Simon, A. Phys. ReV. 1998,
B58, 14346.

Layered Lanthanum Carbide Halide Superconductors
J. Phys. Chem. B, Vol. 103, No. 26, 1999 5453
(18) Ahn, K.; Mattausch, Hj.; Simon, A. Z. Anorg. Allg. Chem. 1997,
623, 619.
(19) Taylor, M. D.; Carter, C. P. J. Inorg. Nucl. Chem. 1962, 24, 387.
(20) Meyer, G.; Ax, P. Mat. Res. Bull. 1982, 17, 1447.
(21) Krause, B. H.; Hook, A. B.; Warner, F. Acta Cryst. 1963, 16, 848.
(22) Simon, A. J. Appl. Crystallogr. 1970, 3, 11.
(23) Izumi, F. The RietVeld Method; Young, R. A., Ed.; University
Press: Oxford, 1993; Chapter 13.
(24) Hoffmann, R. J. Chem. Phys. 1963, 39, 1397.
(25) Whangbo, M.; Hoffmann, R. J. Am. Chem. Soc. 1978, 100, 6093.
(26) Kremer, R. K.; Ahn, K.; Henn, R.; Mattausch, Hj.; Schnelle, W.;
Stolovits, A.; Simon, A. Conference Proceedings of the 1st Conference on
Anomalous and Complex Superconductors, to appear in Physica C.
(27) Gratz, E.; Zuckerman, M. J. Handbook of the Physics and Chemistry
of Rare Earth; Gschneider, K. A., Jr., Eyring, L., Eds.; North Holland, 1982;
Chapter 42.
(28) Coleman, R. V.; Eiserman, G. K.; Hillenius, S. J.; Mitchell, A. T.;
Vincent, J. L. Phys. ReV. 1983, B27, 125.
(29) Oshima, K.; Urayama, H.; Yamochi, H.; Saito, G. Physica 1988,
153-155C, 1148.
(30) Bauhofer, W.; Biberacher, B.; Gegenheimer, B.; Joss, W.; Kremer,
R. K.; Mattausch, Hj.; Mu¨ller, A.; Simon, A. Phys. ReV. Lett. 1989, 63,
2520.
(31) Pauling, L. The Nature of the Chemical Bonding; Cornell University
Press: Ithaca, New York, 1960.