inorganic compounds
Acta Crystallographica Section C
Crystal Structure
Communications
Fig. 1. The starting point of the fractional coordinates for all
atoms was taken from the work of Subramanian et al. (1988).
The reason for atoms O2 and O3 only being re®ned isotrop-
ically was the limited number of observed re¯ections with an
intensity/background ratio greater than 3. The ®rst model
obtained, with only (Tl,Hg) in the charge reservoir (CR) and
(Ca,Sr) at the smaller alkali-earth (AE1) metal site between
the Cu±O layers, resulted in unfavourable atomic displace-
ment parameters at both sites, i.e. a large U value for (Tl,Hg)
and a negative U value for (Ca,Sr). Both Tl and Hg were
earlier observed at the AE1 site in a high-pressure synthesized
2212-type phase (Wu et al., 1998), however, Wu et al. reported
this as a new stacking order. The CR on the other hand was
occupied by both Ca and Cu in another superconducting
material, also formed under high-pressure conditions (Chu et
al., 1997). Since our crystal was synthesized under pressure as
well, Ca was introduced at the CR site at the same time as a
minor amount of Tl was placed at the AE1 site. When re®ning
the occupancies, all sites were assumed fully occupied, except
for the O3 site, which ended up as 78 (5)% occupied after
re®nement. Ba, Hg, Cu and O were all placed at their
presumed sites, while Sr was placed at both AE sites with ®xed
values taken from EDS results. Tl and Ca were placed in the
CR and at the AE1 site. The total amounts of each of the two
metals were kept constant, according to EDS results, and only
their relative occupations at both sites were re®ned, using one
parameter. The metal distribution according to the re®nement
ISSN 0108-2701
The superconductor (Tl,Hg,Ca)2-
(Ba,Sr)2(Ca,Sr,Tl)Cu2O7.6
Martin Valldor,a* Ingrid Bryntsea and Andrzej Morawskib
aDivision of Inorganic Chemistry, Arrhenius Laboratory, Stockholm University,
SE-106 91 Stockholm, Sweden, and bHigh Pressure Research Center, UNIPRESS,
01-142 Warsaw, Sokolowska 29/37, Poland
Correspondence e-mail: beb@inorg.su.se
Received 18 June 2002
Accepted 22 July 2002
Online 21 August 2002
In the title 2212-type superconductor (thallium mercury
calcium barium strontium copper oxide), which contains both
Tl and Hg in the charge reservoir (CR), Sr is located at both
alkali-earth (AE) metal sites. Ca enters the CR at the same
time as Tl shares the smaller AE site, which increases the
apical CuÐCu distance signi®cantly. The structure causes the
superconducting Cu±O layers to become signi®cantly puck-
ered.
was (Tl0.582(2)Hg0.395Ca0.023(2))2(Ba0.435Sr0.565)2(Sr0.36Ca0.594(5)
-
Tl0.046(5))Cu2. This resulted in both atomic displacement
parameters ending up positive and reasonable.
Comment
In the ®nal model, only atom O3 had a large atomic
displacement. However, in many papers, this O atom is
displaced from the high-symmetry position (0, 0, z) to (x, x, z)
(see, for example, Subramanian et al., 1988), which could be a
reason for this relatively high value. Calculated bond lengths
and angles are reasonable (Table 1). The CR metal atoms
(Tl,Hg,Ca) are situated in distorted octahedra, with shorter
apical distances that suit the Hg2+ atom (40%), for which
linear coordination is common.
Ever since the 2212-type superconductors were discovered,
originating from Tl2Ba2CaCu2O8, which is as abbreviated 2212
(Subramanian et al., 1988), several closely related compounds
have been prepared, differing in their chemical substitutions.
Hg was tried at the Tl site (Bryntse, 1994) and Sr was intro-
duced at both the Ba and the Ca site in different amounts
(Maignan et al., 1995; Valldor et al., 2000). Single crystals of
(Tl,Hg)2Sr2CaCu2O8 were obtained using a high-pressure±
d
high-temperature technique (Valldor et al., 1999). The starting
composition Tl1.3Hg0.7BaSr1.5Ca0.5Cu2O8 d was also tried for a
single-crystal growth and this structural study will be
presented below. To ensure that the crystal analysed was
representative of the whole sample, unit-cell parameters were
From bond lengths between Cu and O, the bond-valence
sum of Cu is calculated as BVS(Cu) = 2.195 (9). Assuming
nominal oxidation states for all metal atoms, the calculated
oxygen content would be 7.62, which agrees well with the
re®ned value of 7.6 (2). The measured crystal was too small for
resistivity of susceptibility measurements, however, if the BVS
of Cu is the main property connected to Tc in this system, the
BVS±Tc comparison (Valldor et al., 2000) could be used to
conclude that Tc should be within the range 75±90 K. From the
same report, it is possible to conclude that both cell para-
meters and EDS analyses agree well with a Tc of about 80 K.
When comparing the interatomic distances in Tl2Ba2CaCu2O8
(Table 2; B, Subramanian et al., 1988) with those presented
here (Table 2, A), several things are made clear. The CR in A
is larger than that in B, which also could be an indication of a
minor Ca content at this site in A, since Ca2+ has a larger ionic
radius compared with both Tl3+ and Hg2+. The larger AE site
(AE2) has shorter metal-to-oxygen bonds in A, as expected,
also calculated from powder diffraction data using a Guinier±
Ê
È
Hagg camera on ground single crystals [a = 3.8353 (5) A and
c = 29.138 (6) A]. The unit-cell parameters obtained from the
Ê
Ê
single-crystal diffractometer data were a = 3.8380 (6) A and c =
Ê
29.145 (7) A. Ten elemental analyses (EDS, i.e. energy
dispersive spectroscopy) were performed on the single crystal
after diffraction data had been collected and the resulting
stoichiometry [Tl1.22(4)Hg0.79(4)Ba0.86(3)Sr1.48(4)Ca0.63(2)Cu2.03(6)
-
Ox] agreed with the starting composition (Tl1.3Hg0.7Ba-
Sr1.5Ca0.5Cu2Ox). The only signi®cant difference was the lower
Ba content in the product balanced by a higher Ca content.
The results from EDS were used to generate the occu-
pancies at the different sites in the structure, as illustrated in
i126 # 2002 International Union of Crystallography
DOI: 10.1107/S0108270102013148
Acta Cryst. (2002). C58, i126±i128
Valldor, Martin
