3.1. Floating Zone Crystal Growth of Pb-Doped Bi-2201 at IFW Dresden
The large-size lead-doped Bi-2201 crystals have been prepared by crucible-free FZ method in 1997 for the first time [
18]. Prior to the growth as-sintered feed rod was pre-melted at a fast rate of 25-27 mm/h in infrared (IR) image furnace in order to obtain a high dense rod with density above 90% of the crystal density. For the feed rod preparation, Bi
2O
3, PbO, SrCO
3 and CuO powders of, at least, 3N purity were used as starting materials. A number of feed rod compositions were investigated. Initial powder mixture was mixed in ethanol and fired at 720-840 C for 24 h in air. The rods were pressed at 1 kbars by cold isostatic press and finally fired at 840 ℃ for 24 h in the air. The best feed rod composition was detected to be Bi
1.74Pb
0.38Sr
1.88CuO
y. The crystal growth was initiated by forming a molten zone between the feed and seed rods using IR heating, and the zone was passed through the feed rod at a rate of 5.0 to 0.3 mm/h [
18].
Large size and bulk Bi
1.6Pb
0.4Sr
2.05CuO
y single crystals were grown at IFW-Dresden in a similar way. The 4-mirror type image furnace produced by CSI (Japan) equipped with 4x300 W halogen lamps was used for the growth,
Figure 1. We have found that the using of pre-melted feed rod does play an important role for a stable growth run because liquid penetration onto a feed rod was strongly suppressed in this case [
21]. Therefore, for high dense pre-melted feed rod a stable molten zone (MZ) could be maintained over a long lasting experiment (over 1-2 days). In contrast, using of low dense as-sintered rods with density less than 75% always terminated a FZ growth due to the dramatic liquid penetration. Thus, a straight, high dense and equal-diameter pre-melted rod is crucial to stabilize a MZ for long lasting growth. Also, similar to undoped Bi-2201, a high temperature gradient (~150-300 ℃/cm) in a hot zone generated by correctly focused halogen lamps is needed to maintain a stable molten zone. Long size (about 10 mm in length) pre-melted rod was cut into two pieces, and the shorter piece (~ 2 cm in length) was utilized as a seed and the longer one as a feed rod. Both feed and seed rods were coaxially mounted on upper and lower shafts inside a transparent quartz tube. The used O-ring sealing is gas tight and protects a growth atmosphere from air humidity. Slow growth rates of 0.5-1.0 mm/ were applied in order to get a faster grain selection. Both feed and seed rods were counter rotated (15-20 rpm) to provide a homogeneous temperature distribution in radial direction. The bulk Pb-substituted Bi-2201 crystals with the actual composition Bi
1.8Pb
0.38Sr
2.01CuO
y were successfully fabricated by using the CSI image furnace with infrared heating. The growth was performed in Ar/O
2 = 90/10 gas mixture, the actual pulling rate was 1.0-0.8 mm/h. As-grown samples were non-superconducting above 4 K from the magnetization measurements, and demonstrated a superconducting transition at
Tc~22-23 K after a post-growth heat treatment under vacuum conditions [
18].
The Pb-doped Bi-2201 single crystals cleaved by sharp scalpel from as-grown ingot are depicted on
Figure 2 (left). The biggest samples had dimensions up to
15×4×3 mm
3 with a clear (010) cleavage plane. The weak PbO and Bi
2O
3 evaporation from MZ was detected during a growth. The boiling points of PbO and Bi
2O
3 oxides are 1480 ℃ and 1890 ℃, respectively. The evaporated material deposits on inner wall of quartz tube forming a slightly yellowish thin layer. Therefore, the IR heating was not affected by it. Therefore, there was not need to increase a lamps power and, in turn, the MZ remained stable over the growth.
Typically, Bi
1.6Pb
0.4Sr
2.05CuO
y samples with thickness of 2-3 mm along c-axis show broadening of neutron rocking curves due to weak shoulders around a few Bragg peaks reflecting the co-existence of slightly misoriented grains. The mosaicity of these crystals was found to be about 2-3
o. In contrast, crystals with the thickness of 0.3-0.5 mm along c-axis demonstrate better mosaicity of about 0.5-0.7
o [
17].
As-grown lead-substituted Bi-2201 crystals are usually non-superconducting above 3 K but post-growth heat treatment in vacuum dramatically enhances
Tc up to 22-23 K (Figure2, right). Annealing conditions and effects of heat treatment on superconductivity and material stability of Bi
1.6Pb
0.4Sr
2.05CuO
y crystals are summarized in
Table 1 [
17]. Since the substitution of Bi
3+ with Pb
2+ is believed to generate holes in a sample, lead doping makes the Bi-2201 compound to be heavily overdoped.
The low-temperature superconductivity does vanish in as-grown samples after annealing in the oxygen flow (at least, T
c is below the low temperature limit 2 K of our experimental facility). In contrast, annealing of as-grown lead-doped Bi-2201 samples in vacuum (~ 10
−5 mbar) at 550 ℃ for 7 days increases
Tc up to 22-23 K. This is a direct hint for extra oxygen expel from as-grown Pb-doped Bi-2201 samples reflecting a reduction of sample doping from heavily overdoped to slightly overdoped. By adjusting the annealing atmosphere and temperature, the superconductivity transition temperature could be varied between 0 and 23 K. Nevertheless, the lead-substituted Bi-2201 samples became air sensitive after post-growth heat treatment under vacuum conditions above 550 ℃. These crystals were completely powdered after keeping on a desk at room temperature for a week. Only crystals annealed either in Ar flow (5N purity) or in vacuum below 450 ℃ remain stable in air, and have no visible sign of crystal decomposition and retain superconductivity. Thus, the superconductivity in a lead-doped Bi-2201 material is highly sensitive to annealing process and may result in samples with various
Tc if the heat treatment was carried out under different conditions [
43]. It has been found in the literature that the whole region from slightly overdoped to heavily overdoped samples can be covered by combining the Pb
2+ partial doping with annealing at reduced atmosphere.
Concerning the effect of lead doping on the structural modulation, it is found that the modulation preserves by the La doping, while the substitution of Pb for Bi acts effectively for a modulation suppression [
44]. It was shown by Ikeda et al. [
45] that structural modulation disappears in a narrow range near the Pb-solubility region of
x= 0.4 at
y = 0.125 in Bi
2-xPb
xSr
2-yCuO
z.
3.2. SCXRD Investigations on Pb-Doped Bi-2201
The majority of previous structural characterization works for the Bi-2201 family of cuprates were performed on crystals grown from off-stoichiometric oxides mixtures, therefore the “best” crystals chosen for SCXRD might not always be representative of the bulk. Floating zone crystal growth, in opposite, guarantees a homogenous composition distribution along a crystal ingot [
17,
18], which is beneficial for the reproducibility of SCXRD analysis, too. Unfortunately there are only few reports about the crystal growth by this technique and, consequently, a few structural works on such crystals available in the literature since the 80’s [
34]. In this section, we thoroughly analyzed centrosymmetry as well as modulation suppression on two heavily Pb-doped Bi-2201 crystals, before and after post-growth annealing.
Single crystal X-ray diffraction data acquisition was accomplished on a Bruker D8 Venture (MoKα λ = 0.71073 Å) equipped with a PHOTON 100 CMOS detector. The measurements were performed at room temperature. Indexing was performed using APEX3 software [
46]. Data integration and absorption corrections were performed using the SAINT and SADABS [
46,
47] software, respectively. Crystal structure was solved by dual-space methods implemented in the SHELXT [
48] program and refined by full-matrix least-squares method on F
2 with SHELXL [
49]. The composition of the crystal was taken from EDX analysis.
Generally, our SCXRD data confirm the average crystal structure of Pb-doped Bi-2201 described in the literature. SCXRD reflections can be indeed indexed in a
C-centered orthorhombic cell with very close values of lattice parameters
a = 5.388(2) Å,
b = 24.608(1) Å,
c = 5.279(2) Å (a non-annealed crystal with
Tc = 3 K, further denoted as OD3K) or
a = 5.3947(6) Å,
b = 24.605(3) Å,
c = 5.2786(6) Å (an annealed crystal with
Tc = 23 K, further denoted as OD23K). A notable feature of the OD23K crystal is a very weak incommensurate modulation with
q 0.21
c*, clearly seen in 0
kl reciprocal layers (see
Figure 3). For a comparison, the OD3K crystal demonstrate rather thin diffuse streaks along
a* instead of well-defined spots of satellite reflections,
Figure 4. This modulation in OD23K cannot be entirely associated with the post-annealing, it is present also in the sample before annealing. The mismatch between the BiO-slab and the copper oxide perovskite block cannot be eliminated by Pb-doping alone, and the observed modulation is evidently a residual feature. The
q-vector value is close to those reported by Gao et al. [
36] for a Pb-doped single layer Bi
2Sr
2−xCa
xCuO
6 cuprate as well as to those previously reported for undoped Bi-2212 [
50] (
q 0.21
a*, sp.gr.
Amaa was used in the original works). In opposite, some literature testifies a full disappearance of the satellite patterns in heavily Pb-doped Bi-2201 [
34,
45]. In in the following discussion we show that these weak reflections can be convincingly described in a non-centrosymmetric space group
Ccc2.
The Pb,Bi-2201 structure can be solved and refined either in the centrosymmetric
Cccm or in non-centrosymmetric
Ccc2 space group. It is worth noting that E-statistics indicate that the structure is rather non-centrosymmetric, Sheldrick’s |E
2 – 1| criterion is 0.769, see
Figure 5, data for the OD23K crystal.
We tested both Cccm and Ccc2 solutions and ascertained that the positions of the metal atoms Bi, Sr, and Cu found by ab-initio structure solution by SHELXT are basically equivalent in both space groups. However, refined anisotropic temperature factors (ADP) for the metal atoms as well as crystallographic coordinations of Bi/Pb by O atoms are clearly different. The structure solution in space group Cccm showed structural arrangement very similar to one described previously for the parent 2201 phase, with ribbons of strong Bi-O bonds, and splitting of the oxygen site in the Bi plane. Moreover, the Bi atoms in this model had very large anisotropic factors along the c-axis.
Lowering symmetry to
Ccc2 and the refinement of this model resulted in R
1 = 4.78% and reasonable ADP along the
c-axis. The atomic coordinates, temperature factors, and site occupancies for the Bi-2201 crystal together with the details of the refinement are listed in
Tables S1–S3. As can be seen from these tables, the stereochemistry of the various ions is as expected for this structure type (
Figure 6). In distinction to Y. Ito et al. [
51], where the Pb dopant was placed in both Bi and Sr sites, we assume that the Sr sites are occupied predominantly by Sr, whereas Pb/Bi sites are mixed occupied. Since scattering powers of Pb and Bi are almost equal for wavelengths which are not close to an absorption edge, these atoms cannot be distinguished by conventional X-rays. The experimental composition measured by EDX and averaged over all measurements found to be Pb:Bi:Sr:Cu = 0.35(3):1.69(8):2.01(9):1, Pb/(Bi+Pb) ratio is 0.17, thus, Cu oxidation state can be estimated as +2.21, providing ideal oxygen composition. SCXRD refinement reveals that the Pb/Bi site may be slightly underoccupied, as compared to the EDX data. There is no extra-oxygen position found within the Bi-O layers, in contrast to the undoped Bi-2201, e.g., Mironov et al. [
30].
The crystal structure of a non-annealed OD3K crystal was also refined in the space group
Ccc2 using 12508 reflections measured (3.318° ≤ 2Θ ≤ 61.82°), 1070 unique (
Rint = 0.1170, R
sigma = 0.0462) in all calculations. The final
R1 was 0.0852 (I > 2σ(I)) and
wR2 was 0.1868 (all data). For comparison, the refinement in the centrosymmetric space group
Cccm has converged to
R1 = 0.1040 (I > 2σ(I)) and
wR2 = 0.2298 (all data). The atomic coordinates, temperature factors, and site occupancies for the OD0K crystal together with the details of the refinement are listed in
Tables S4–S6.
In the
Table 2, we put together different structural models proposed in the literature for the doped Bi-2201 and Bi-2212 compounds.
Thus, the substitution of Bi by Pb has yielded the observation of superconductivity in Bi-2201 materials that do not contain extra oxygen within the Bi-O layers, also this suggests a lowering of the oxygen content to 6 per formula unit. The subtle difference between structures with space groups Ccc2 and Pnan originates in the position of the oxygen atoms with respect to Bi in the Bi-O layers.