Polymorphism of Pb₅(PO₄)₃OH_δ within the LK-99 mixture

During the synthetic exploration targeting the polycrystalline compound LK-99, an unexpected phase, Pb₅(PO₄)₃OH_δ, was identified as a byproduct. We elucidated the composition of this compound through single-crystal X-ray diffraction analysis. Subsequent synthesis of the target compounds was achieved via high-temperature solid-state pellet reactions. The newly identified Pb₅(PO₄)₃OH_δ has an orthorhombic crystal structure with space group Pnma, representing a unique structure differing from the hexagonal apatite phases of Pb₁₀(PO₄)₆O and Pb₅(PO₄)₃OH. Comprehensive temperature- and magnetic-field-dependent magnetization studies unveiled a temperature-independent magnetic characteristic of Pb₅(PO₄)₃OH_δ. Solid-state nuclear magnetic resonance spectroscopy was employed to decipher the origins of the phase stability and confirm the presence of hydrogen atoms in Pb₅(PO₄)₃OH_δ. These investigations revealed the presence of protonated oxygen sites, in addition to the interstitial water molecules within the structure, which may play critical roles in stabilizing the orthorhombic phase.

1. Introduction

After the claim of the discovery of an ambient-pressure room-temperature superconductor (T_c > 400 K) LK-99, (Pb_10−xCu_x)(PO₄)₆O (Lee, Kim & Kwon, 2023), with a subsequent report of a levitation experiment at room temperature indicating strong diamagnetic signals (Lee, Kim, Kim et al., 2023), LK-99 has garnered unprecedented attention. Despite numerous attempts by various research groups to replicate and verify the superconductivity of LK-99 (Zhu et al., 2023; Timokhin et al., 2023; Kumar et al., 2023; Wu et al., 2023; Hou et al., 2023), increasing experimental evidence has begun to cast doubt on its superconducting nature (Garisto, 2023).

In the quest to unravel the mysteries of LK-99's superconductivity, our investigation has led to the discovery of several new phases. Among these, the novel structure of hy­droxy­lpyromorphite, Pb₅(PO₄)₃OH_δ, stands out and is different from the conventional hexagonal phase known for Pb₁₀(PO₄)₆O (Brückner et al., 1995; Kim et al., 1997; Barinova et al., 1998). The compound Pb₁₀(PO₄)₆O has been termed oxypyromorphite, a nomenclature that underscores its structural resemblance to pyromorphite, Pb₅(PO₄)₃Cl, and hy­droxy­lpyromorphite, Pb₅(PO₄)₃(OH), which adopts an apatite-like hexagonal structure, wherein O²⁻ anions substitute for the halide ions typically found in apatite structures. This substitution suggests the presence of vacancies at some halide sites, in contrast to the original proposition of a (Pb²⁺)₉(Pb⁴⁺)(PO₄)₆O₂ formula (Ito, 1968), which would negate the need for such vacancies. However, further studies confirmed the absence of Pb⁴⁺ cations in oxypyromorphite (Merker et al., 1970). Despite its intriguing properties, detailed crystal structure analysis of oxypyromorphite has yet to be carried out. In addition, the lead-based compounds Pb₄O(PO₄)₂, Pb₈O₅(PO₄)₂ and Pb₁₀(PO₄)₆O have been studied for several decades (Yang et al., 2001; Brixner & Foris, 1973; Krivovichev & Burns, 2003). The ferro-elastic Pb₈O₅(PO₄)₂ and its vanadium analog Pb₈O₅(VO₄)₂ (Dudnik & Kolesov, 1980; Kiosse et al., 1982) have also been discovered. Although X-ray and optical analyses on the single crystals of these compounds have been conducted, the crystal structures of both remain unresolved.

In this study, we present a synthetic strategy for the new Pb₅(PO₄)₃OH_δ compound using a high-temperature solid-state pellet reaction. Millimetre-sized single crystals were obtained from the reaction. Single-crystal and powder X-ray diffraction (XRD) experiments were conducted to determine the crystal structure and confirm the phase information. Accordingly, Pb₅(PO₄)₃OH_δ was found to crystallize in the orthorhombic crystal system with the space group Pnma. Different from the hexagonal apatite phase Pb₁₀(PO₄)₆O with the balanced charge of (Pb²⁺)₁₀(PO₄³⁻)₆O²⁻, the compound of Pb₅(PO₄)₃O cannot be charge-balanced with the sole existence of Pb²⁺ and full occupancies on all atomic sites. To confirm the existence of a proton (H⁺) to balance the charge in Pb₅(PO₄)₃OH_δ, high-magnetic-field (800 MHz or 18.8 T) solid-state nuclear magnetic resonance (NMR) spectroscopy was used to determine the chemical environments of proton sites and the presence of water molecules in the system.

2. Synthesis and experimental methods

2.1. Chemical synthesis

Pb₅(PO₄)₃OH_δ crystals were synthesized in two steps. The first step was synthesis of the precursors. The mixture of PbO (99.3%, BAKER ANALYZED) and PbSO₄ (99.1%, BAKER ANALYZED) was heated under a vacuum for 24 h with 1:1 mole ratio. After the reaction, we obtained the Pb₂(SO₄)O precursor with Pb₃(SO₄)O₂ (∼1.5 at.%]) impurity. Another precursor is Cu₃P, obtained by heating Cu powder (99.9%, Alfa Aesar) and P powder (99%, Beantown Chemical) for 48 h at 550°C under vacuum. The molar ratio of Cu and P was 3:1 and the powder XRD results of the two precursors are shown in the supporting information (Figs. S1a and S1b). The second step was mixing Pb₂(SO₄)O and Cu₃P in the ratio 5:3, remaining at 925°C for 20 h. Millimetre-size single crystals of Pb₅(PO₄)₃OH_δ were obtained. All the reaction products were powder pressed using a 0.25 inch (internal diameter) dry pellet pressing die made of carbon steel. A 2 ml alumina cylindrical crucible held the pressed pellet before it was sealed in a fused silica tube under vacuum with around 30 mTorr pressure. After the solid-state reaction, transparent single-crystalline samples can easily be identified on the bottom pellet, which can be separated from black-colored polycrystalline chunks and copper-colored solidified drops on the surface of the black chunks. The rest of the measurements were done using the single crystals separated mechanically from the mixture.

2.2. Structure determination and phase analysis

Pb₅(PO₄)₃OH_δ forms transparent, rod-like, brittle single crystals. A single crystal was selected, mounted on a nylon loop with Paratone oil and measured using an XtalLAB Synergy, Dualflex, Hypix single-crystal X-ray diffractometer. Data were collected using ω scans with Mo Kα radiation (λ = 0.71073 Å) and Ag Kα radiation (λ = 0.56087 Å), a micro-focus sealed X-ray tube, 65 kV, 0.67 mA. The total number of runs and images was based on the strategy calculation from the CrysAlisPro (Rigaku OD, 2017) program. Data reduction was performed with correction for Lorentz polarization. A numerical absorption correction was applied based on Gaussian integration over a multifaceted crystal model (Parkin et al., 1995). Empirical absorption correction used spherical harmonics, implemented in the SCALE3 ABSPACK scaling algorithm (Walker & Stuart, 1983). The structure was solved and refined using the SHELXTL software package (Sheldrick, 2015a, Sheldrick, 2015b). Tables 1 and 2 show the results of the single-crystal XRD. For the powder XRD measurements, single crystals were ground in an agate mortar and pestle, and the powder placed onto the 20 × 20 × 0.5 mm Rigaku Square groove. Powder XRD measurements were carried out using a Rigaku MiniFlex powder diffractometer in Bragg–Brentano geometry with Cu Kα radiation (λ = 1.5406 Å). Room-temperature measurements were performed with a step size of 0.01° at a scan speed of 0.5° per minute over a Bragg angle (2θ) range of 10–90°. GSAS II (Toby & Von Dreele, 2013) was used to perform the Rietveld refinement and analyze phase information.

Table 1
Experimental details

Values in parentheses are the estimated standard deviation from refinement.

Chemical formula	Pb5(PO4)3OH
Formula weight (g mol^−1)	1337.87
Space group	Pnma
Unit-cell dimensions (Å)	a = 13.5137 (4), b = 10.2904 (4), c = 9.3838 (3)
Volume (Å^3)	1304.93 (7)
Z, Z′	4, 0.5
Density (calculated) (g cm^−3)	6.810
Absorption coefficient μ (mm^−1)	64.725
Crystal size (mm)	0.076 × 0.056 × 0.049
F(000)	2240
Data collection
2θ range (°)	5.286–82.472
No. of reflections collected	48595
No. of independent reflections	4458
Rint	0.1257
Refinement method	Full-matrix least-squares on F^2
No. of data, restraints, parameters	4458, 0, 114
Final R indices	R1 [I > 2σ(I)] = 0.0402; wR2 [I > 2σ(I)] = 0.0681 R1 (all) = 0.0811; wR2 (all) = 0.0758
Δρmax, Δρmin (e Å^−3)	+3.79 and −5.99
RMS deviation from mean (e Å^−3)	0.511
Goodness-of-fit on F^2	1.047

Table 2
Atomic coordinates and equivalent isotropic atomic displacement parameters (Ų) of Pb₅(PO₄)₃OH

U_eq is one-third of the trace of the orthogonalized U_ij tensor. Values in parentheses are the estimated standard deviation from refinement.

Atoms	Wyckoff site	x	y	z	Occ.	Ueq
Pb1	8d	0.07377 (2)	0.55577 (2)	0.29630 (2)	1	0.019 (1)
Pb2	4c	0.02043 (2)	0.75	−0.02843 (3)	1	0.017 (1)
Pb3	4c	0.23350 (3)	0.75	−0.30356 (4)	1	0.025 (1)
Pb4	4c	0.2750 (2)	0.75	0.1556 (2)	0.71	0.0285 (2)
Pb5	4c	0.2703 (5)	0.75	0.1277 (6)	0.29	0.05 (1)
P1	8d	−0.1631 (1)	0.5324 (1)	0.07074 (1)	1	0.0123 (2)
P2	4c	0.0030 (2)	0.75	−0.4252 (2)	1	0.0153 (4)
O1	4c	−0.0913 (6)	0.75	−0.34610 (8)	1	0.0402 (2)
O2	4c	−0.071 (5)	0.75	0.4179 (7)	1	0.0260 (1)
O3	4c	0.1089 (4)	0.75	0.1753 (5)	1	0.013 (1)
O4	8d	−0.1751 (4)	0.3859 (4)	0.0955 (4)	1	0.0248 (9)
O5	8d	−0.1685 (4)	0.5588 (5)	−0.0902 (5)	1	0.030 (1)
O6	8d	−0.0618 (4)	0.5794 (5)	0.1249 (6)	1	0.031 (1)
O7	8d	−0.2451 (4)	0.6053 (5)	0.1488 (6)	1	0.033 (1)
O8	8d	0.0635 (4)	0.8677 (5)	−0.3915 (7)	1	0.041 (1)
H	8d	0.0993	0.85244	−0.322954	0.5	0.061

2.3. Solid-state NMR spectroscopy for detecting protons

¹H solid-state NMR experiments were conducted on a Bruker NEO spectrometer with a narrow-bore magnet with B₀ = 18.8 T [ν₀(¹H) = 800 MHz] at room temperature (298 K). Spectra were acquired using a Phoenix NMR 1.6 mm HXY magic-angle spinning (MAS) probe with samples packed into 1.6 mm (outer diameter) zirconia rotors. The MAS frequency was set to 8 kHz. ¹H chemical shifts were referenced to alanine (δ_iso = 1.38 p.p.m.) as a secondary reference with respect to tetra­methyl­silane (δ_iso = 0 p.p.m.). All ¹H spectra were acquired using a rotor-synchronized Hahn echo (90°–τ–180° acquisition) with 2.5 µs (100 kHz) π/2 pulses, an interpulse delay (τ) of 500 µs and a recycle delay of 2 s (see Table S1 for further details). Spectra were processed in TopSpin 4.1 (Bruker), and simulations of all spectra were prepared using ssNake v1.3 (van Meerten et al., 2019).

2.4. Magnetic measurements

Temperature- and magnetic-field-dependent DC and vibrating sample magnetometry (VSM) magnetization data were collected using a Quantum Design Magnetic Property Measurement System (MPMS3). Temperature- and magnetic-field-dependent DC magnetization measurements were taken on the powder sample loaded in the powder sample holder and put into a brass half-tube sample holder. In VSM measurements, a peak amplitude of 5 mm and an average of 2 s were used. An empty powder sample holder was measured under the same conditions to estimate the background to be subtracted from the measurements.

3. Results and discussion

The crystal structure of Pb₅(PO₄)₃OH_δ exhibits similarities to apatite-like lead compounds, including pyromorphite [Pb₅(PO₄)₃Cl], hy­droxy­lapatite [Pb₅(PO₄)₃(OH)] and the LK-99 precursor, Pb₁₀(PO₄)₆O. Previous determinations of hy­droxy­lapatite's structure, through neutron and X-ray powder diffraction analysis, identified the OH group positioned at the 4e Wyckoff site with a 0.5 site occupation factor (Kim et al., 1997). In contrast, Barinova et al.'s refinement using single-crystal diffraction data located the OH group at the 2b site, indicating full occupancy – a characteristic more aligned with the halide ion positions in pyromorphite-like compounds (Pb₅(PO₄)₃X, where X = F, Cl (Barinova et al., 1998). The structural framework of Pb₁₀(PO₄)₆O mirrors that of hy­droxy­lapatite, with the O4 atom situated at the 4e site, albeit with a reduced occupation factor of 0.25. Our investigation into a single crystal of Pb₅(PO₄)₃OH_δ revealed a deviation from the expected hexagonal structure to an orthorhombic Pnma space group, attributed to lattice parameter distortions, as shown in Fig. 1(a). The hydrogen atoms were refined with half occupancy at the 8d sites. However, considering the limitations of XRD in accurately detecting hydrogen positions, we employed high-field ¹H solid-state NMR (ssNMR) to further elucidate the hydrogen occupancies, providing a more comprehensive understanding of the structural intricacies of Pb₅(PO₄)₃OH_δ. To ascertain the phase purity and facilitate the investigation of its physical properties, powder XRD analysis was performed, with the results presented in Fig. 1(b). The data were refined using Rietveld refinement, and the green peak suggests a peak from a minor impurity phase.

Figure 1 The crystal structure and the powder XRD data of Pb5(PO4)3OHδ. (a) shows the crystal structure of Pb5(PO4)3(OH) from single-crystal XRD and a comparison of Pb atom distribution between the orthorhombic and hexagonal structures. (b) Powder XRD data of Pb5(PO4)3OHδ from in-laboratory diffraction measurements. The experimental data are plotted as red dots. The black line gives the Rietveld refinement. The blue line indicates the corresponding residual pattern (difference between observed and calculated patterns). Bars give Pb5(PO4)3OHδ peak positions from single-crystal XRD measurement. The green arrow indicates the peak from an impurity. On the upper right of the figure, a picture of the crystals is shown; the size of the chunk is about 5 mm.

The ¹H ssNMR spectrum features six underlying peaks with their isotropic chemical shifts (δ_iso) ranging from 0.8 to 5.2 p.p.m. in a 1:1.2:2.1:2.7:10.4:3.6 ratio from left to right (Fig. 2 and Table 3). This is surprising since the solved single-crystal XRD structure only includes one hydrogen position, while the NMR data indicate a far more complex ¹H environment. It is not unusual for ssNMR to detect additional structural features that are invisible to XRD techniques (Morris et al., 2017; Zhang et al., 2022; Inukai et al., 2016; Corlett et al., 2019; Li et al., 2013; Serrano-Sevillano et al., 2019). A few possible explanations for these hydrogen resonances are: (i) structural defects in the sample where the phosphate ion reacts with atmospheric water, (ii) water being incorporated into the structure, either occupying vacancies in the octahedral or tetrahedral holes in the crystal structure and/or (iii) atmospheric water being bound to the lead as a ligand.

Table 3
Isotropic chemical shifts (δ_iso) and percentages of ¹H resonances from ssNMR

	Peak
Parameter	a	b	c	d	e	f
δiso (p.p.m.)	5.2	2.2	2.0	1.5	1.3	0.8
Percentage (%)	5	6	10	13	50	17

Figure 2 ^1H MAS solid-state NMR analysis of Pb5(PO4)3OHδ. Experimental ^1H MAS NMR of Pb5(PO4)3OHδ is shown in blue, with corresponding analytical simulations in black, and deconvolution of the simulation is shown in color. Six ^1H resonances are color-coded and labeled as a–f, with their relative intensity percentages provided.

Since only a single spectrum for Pb₅(PO₄)₃OH_δ was acquired with no internal reference, only approximate estimation of the amount of hydrogen in the sample is possible. Precise quantification would require a low radiofrequency pulse, a different pulse sequence (i.e. Bloch decay), accurate site assignments of all hydrogen atoms and a series of standards to construct a calibration curve (Bharti & Roy, 2012; Pauli et al., 2012; Giraudeau, 2017; Holzgrabe, 2010; Pauli et al., 2015). Hence, we are limited to approximately quantifying the Pb₅(PO₄)₃OH_δ spectrum relative to another ¹H NMR spectrum of a known sample. First, a ¹H spectrum of alanine with the exact same experimental parameters as Pb₅(PO₄)₃OH_δ was acquired using a sufficiently long recycle delay (in this case, 2 s) to completely re-equilibrate the magnetization before the next scan. Second, a Hahn echo with a long interpulse delay (in this case, 500 µs) was used for these experiments in order to eliminate the broad ¹H background signal from the probe to allow for easier quantification of the spectra. It has been shown that the relative quantification of NMR spectra of two compounds is possible using the following equation (Malz & Jancke, 2005):

where n_x/n_y is the molar ratio, I_x/I_y is the total integrated intensity ratio and N is the number of nuclei corresponding to the resonance line. Using the total integrated intensities from a spectrum of alanine and Pb₅(PO₄)₃OH_δ, we are able to write equation (2) for approximately quantifying the spectrum as

where I is the total integrated intensity of the spectrum, M is the molecular weight, m is the mass of the sample packed in the NMR rotor and S is the number of moles of hydrogen in the sample. Using equation (2), we were able to estimate the amount of hydrogen in Pb₅(PO₄)₃OH_δ to be approximately 82.5 µg or 1.15 wt%.

While Fig. 1(b) depicted the primary phase of the transparent single crystal, Pb₅(PO₄)₃OH_δ, the data also suggested the presence of minor impurity phases within the sample. Single-crystal and powder XRD analyses have identified Cu₂S as the predominant impurity. Prior to evaluating the magnetization data, it is essential to consider the potential magnetic contributions from Cu₂S impurities. If Cu₂S were to significantly influence the magnetic behavior, one might expect to observe a β-to-γ phase transition around 370 K in resistivity, heat capacity and magnetic susceptibility measurements (Zhu et al., 2023). VSM measurements, conducted without background subtraction across a temperature range of 1.8 to 400 K and presented in Fig. 3, do not reveal any magnetic transition near 370 K. The large positive magnetization is also different from previous work (Zhu et al., 2023). The presence of undetected impurity phases, a common occurrence in the solid-state synthesis of LK-99, cannot be overlooked. Fig. 3 presents the findings from temperature-dependent magnetization investigations, conducted under identical experimental conditions to those applied to an empty sample holder to guarantee precision. The magnetization profiles depicted in Fig. 3, obtained through zero-field-cooled-warming (ZFCW) and field-cooled (FC) methods, exhibit distinctive Curie–Weiss-like behavior fitting is shown in Figs. S2a and S2b). A significant deviation between the ZFCW and FC data is evident around 50 K, where a kink-like anomaly, potentially arising from oxygen trapped in the measured sample, is identified in the ZFCW magnetization trajectory. This anomaly is accentuated in the derivative d(MT/H)/dT plot, prominently featured in the inset at the upper-left corner of the figure, further emphasizing its significance. The field-dependent magnetization is also shown in Fig. S3b.

Figure 3 Temperature-dependent magnetization of Pb5(PO4)3OHδ. ZFCW and FC magnetization as a function of temperature is given by VSM in the range of 1.8–400 K with a field of 1 kOe. The left inset shows d(MT/H)/dT as a function of temperature in the low-temperature range. The right inset shows the derivative of d(MT/H)/dT in the high-temperature range.

In conclusion, the orthorhombic phase of Pb₅(PO₄)₃OH_δ was successfully synthesized via solid-state reaction techniques with the objective of creating LK-99. The determination of the phase composition and crystal structure of the product was achieved through an integrated approach, utilizing both single-crystal and powder X-ray diffraction analyses, complemented by solid-state NMR spectroscopy.

Supporting information

Computing details

(I)

Crystal data

HO13P3Pb5	Dx = 6.810 Mg m^−3
Mr = 1337.87	Mo Kα radiation, λ = 0.71073 Å
Orthorhombic, Pnma	Cell parameters from 15385 reflections
a = 13.5137 (4) Å	θ = 2.6–41.0°
b = 10.2904 (4) Å	µ = 64.73 mm^−1
c = 9.3838 (3) Å	T = 293 K
V = 1304.93 (7) Å^3	Rod-like
Z = 4	0.08 × 0.06 × 0.05 mm
F(000) = 2240

Data collection

XtalLAB Synergy, Dualflex, Hypix diffractometer	Rint = 0.126
Absorption correction: analytical	θmax = 41.2°, θmin = 2.6°
Tmin = 0.057, Tmax = 0.177	h = −24→24
48595 measured reflections	k = −18→18
4458 independent reflections	l = −17→17
2940 reflections with I > 2σ(I)

Refinement

Refinement on F^2	Hydrogen site location: inferred from neighbouring sites
Least-squares matrix: full	H-atom parameters constrained
R[F^2 > 2σ(F^2)] = 0.040	w = 1/[σ^2(Fo^2) + (0.0195P)^2] where P = (Fo^2 + 2Fc^2)/3
wR(F^2) = 0.076	(Δ/σ)max = 0.001
S = 1.05	Δρmax = 3.79 e Å^−3
4458 reflections	Δρmin = −5.99 e Å^−3
114 parameters	Extinction correction: SHELXL-2019/1 (Sheldrick 2019), Fc^*=kFc[1+0.001xFc^2λ^3/sin(2θ)]^-1/4
0 restraints	Extinction coefficient: 0.00025 (3)

Special details

Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq	Occ. (<1)
Pb1	0.07377 (2)	0.55577 (2)	0.29630 (2)	0.01897 (6)
Pb2	0.02043 (2)	0.750000	−0.02842 (3)	0.01674 (6)
Pb3	0.23350 (3)	0.750000	−0.30356 (4)	0.02470 (8)
Pb4	0.27499 (19)	0.750000	0.1556 (2)	0.0285 (2)	0.71
Pb5	0.2703 (5)	0.750000	0.1277 (6)	0.0496 (14)	0.29
P1	−0.16310 (10)	0.53243 (13)	0.07074 (14)	0.0123 (2)
P2	0.00299 (15)	0.750000	−0.4252 (2)	0.0153 (4)
O1	−0.0913 (6)	0.750000	−0.3461 (8)	0.0402 (19)
O2	−0.0171 (5)	0.750000	0.4179 (6)	0.0257 (13)
O3	0.1089 (4)	0.750000	0.1753 (5)	0.0134 (10)
O4	−0.1751 (4)	0.3859 (4)	0.0955 (4)	0.0248 (9)
O5	−0.1685 (4)	0.5588 (5)	−0.0902 (5)	0.0297 (11)
O6	−0.0618 (3)	0.5794 (5)	0.1249 (6)	0.0311 (11)
O7	−0.2451 (4)	0.6053 (5)	0.1488 (6)	0.0333 (12)
O8	0.0635 (4)	0.8677 (5)	−0.3915 (7)	0.0406 (14)
H8	0.099291	0.852435	−0.322954	0.061*	0.5

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Pb1	0.02213 (10)	0.01654 (10)	0.01823 (9)	−0.00122 (8)	−0.00192 (8)	0.00298 (7)
Pb2	0.01586 (12)	0.02079 (14)	0.01357 (11)	0.000	−0.00303 (10)	0.000
Pb3	0.02997 (17)	0.02127 (15)	0.02286 (15)	0.000	0.00819 (13)	0.000
Pb4	0.0111 (3)	0.0524 (5)	0.0221 (5)	0.000	0.0032 (3)	0.000
Pb5	0.0111 (9)	0.098 (2)	0.040 (3)	0.000	0.0058 (15)	0.000
P1	0.0110 (5)	0.0105 (5)	0.0154 (6)	0.0008 (5)	0.0014 (5)	0.0000 (4)
P2	0.0142 (8)	0.0203 (9)	0.0115 (8)	0.000	0.0010 (7)	0.000
O1	0.029 (4)	0.066 (6)	0.026 (4)	0.000	0.011 (3)	0.000
O2	0.022 (3)	0.041 (4)	0.015 (3)	0.000	0.002 (2)	0.000
O3	0.010 (2)	0.018 (3)	0.012 (2)	0.000	0.0001 (17)	0.000
O4	0.039 (3)	0.0122 (18)	0.024 (2)	−0.0076 (19)	0.007 (2)	−0.0007 (15)
O5	0.040 (3)	0.033 (2)	0.0160 (19)	0.012 (2)	−0.0050 (19)	0.0031 (18)
O6	0.018 (2)	0.041 (3)	0.035 (3)	−0.011 (2)	−0.0083 (19)	0.008 (2)
O7	0.031 (3)	0.025 (2)	0.044 (3)	0.005 (2)	0.020 (2)	−0.012 (2)
O8	0.028 (3)	0.033 (3)	0.061 (4)	−0.001 (2)	−0.012 (3)	−0.019 (3)

Geometric parameters (Å, º)

Pb1—Pb4	3.624 (2)	Pb3—O5^vi	2.573 (4)
Pb1—Pb5	3.681 (6)	Pb3—O8^iv	2.725 (5)
Pb1—O2	2.608 (4)	Pb3—O8	2.725 (5)
Pb1—O3	2.347 (3)	Pb4—O3	2.252 (6)
Pb1—O5^i	2.601 (5)	Pb4—O7^ii	2.379 (5)
Pb1—O6	2.450 (5)	Pb4—O7^vii	2.379 (5)
Pb1—O7^ii	2.552 (5)	Pb5—O3	2.226 (9)
Pb2—Pb5	3.681 (6)	Pb5—O7^vii	2.581 (7)
Pb2—O3	2.255 (5)	Pb5—O7^ii	2.581 (7)
Pb2—O4^i	2.592 (5)	P1—O4	1.534 (4)
Pb2—O4^iii	2.592 (5)	P1—O5	1.536 (4)
Pb2—O6	2.527 (5)	P1—O6	1.538 (5)
Pb2—O6^iv	2.527 (5)	P1—O7	1.526 (5)
Pb3—O1^v	2.753 (7)	P2—O1	1.474 (7)
Pb3—O4^i	2.528 (4)	P2—O2^viii	1.498 (6)
Pb3—O4^iii	2.528 (4)	P2—O8	1.495 (5)
Pb3—O5^v	2.573 (4)	P2—O8^iv	1.495 (5)
Pb4—Pb1—Pb5	4.13 (11)	O7^vii—Pb4—Pb1	88.76 (15)
O2—Pb1—Pb4	95.16 (11)	O7^ii—Pb4—Pb1^iv	88.76 (15)
O2—Pb1—Pb5	96.41 (12)	O7^ii—Pb4—Pb1	44.60 (13)
O3—Pb1—Pb4	37.09 (13)	O7^vii—Pb4—Pb1^iv	44.60 (13)
O3—Pb1—Pb5	35.30 (14)	O7^vii—Pb4—O7^ii	77.5 (3)
O3—Pb1—O2	69.77 (15)	Pb1^iv—Pb5—Pb1	65.77 (11)
O3—Pb1—O5^i	85.80 (15)	Pb2—Pb5—Pb1	60.61 (10)
O3—Pb1—O6	75.47 (17)	Pb2—Pb5—Pb1^iv	60.61 (10)
O3—Pb1—O7^ii	74.55 (17)	O3—Pb5—Pb1	37.53 (9)
O5^i—Pb1—Pb4	67.03 (12)	O3—Pb5—Pb1^iv	37.53 (9)
O5^i—Pb1—Pb5	64.62 (13)	O3—Pb5—Pb2	35.03 (16)
O5^i—Pb1—O2	154.68 (15)	O3—Pb5—O7^vii	76.0 (2)
O6—Pb1—Pb4	105.49 (12)	O3—Pb5—O7^ii	76.0 (2)
O6—Pb1—Pb5	101.73 (14)	O7^vii—Pb5—Pb1	84.6 (2)
O6—Pb1—O2	81.89 (19)	O7^ii—Pb5—Pb1	43.88 (14)
O6—Pb1—O5^i	85.71 (17)	O7^vii—Pb5—Pb1^iv	43.88 (14)
O6—Pb1—O7^ii	146.06 (16)	O7^ii—Pb5—Pb1^iv	84.6 (2)
O7^ii—Pb1—Pb4	40.88 (12)	O7^ii—Pb5—Pb2	104.4 (2)
O7^ii—Pb1—Pb5	44.49 (14)	O7^vii—Pb5—Pb2	104.4 (2)
O7^ii—Pb1—O2	102.13 (18)	O7^vii—Pb5—O7^ii	70.5 (3)
O7^ii—Pb1—O5^i	76.61 (17)	O4—P1—O5	108.5 (2)
O3—Pb2—Pb5	34.53 (16)	O4—P1—O6	110.6 (3)
O3—Pb2—O4^i	77.19 (15)	O5—P1—O6	108.2 (3)
O3—Pb2—O4^iii	77.19 (15)	O7—P1—O4	109.5 (3)
O3—Pb2—O6^iv	75.55 (14)	O7—P1—O5	110.5 (3)
O3—Pb2—O6	75.55 (14)	O7—P1—O6	109.5 (3)
O4^iii—Pb2—Pb5	49.98 (12)	O1—P2—O2^viii	109.8 (4)
O4^i—Pb2—Pb5	49.98 (12)	O1—P2—O8	111.5 (3)
O4^iii—Pb2—O4^i	65.31 (18)	O1—P2—O8^iv	111.5 (3)
O6—Pb2—Pb5	100.19 (12)	O8—P2—O2^viii	107.9 (3)
O6^iv—Pb2—Pb5	100.19 (12)	O8^iv—P2—O2^viii	107.9 (3)
O6—Pb2—O4^iii	150.17 (14)	O8^iv—P2—O8	108.2 (4)
O6^iv—Pb2—O4^i	150.17 (14)	P2—O1—Pb3^ix	179.6 (5)
O6^iv—Pb2—O4^iii	96.77 (15)	Pb1—O2—Pb1^iv	100.0 (2)
O6—Pb2—O4^i	96.77 (15)	P2^x—O2—Pb1^iv	110.2 (2)
O6^iv—Pb2—O6	88.0 (3)	P2^x—O2—Pb1	110.2 (2)
O4^iii—Pb3—O1^v	82.81 (18)	Pb1^iv—O3—Pb1	116.8 (2)
O4^i—Pb3—O1^v	82.81 (18)	Pb2—O3—Pb1^iv	107.63 (14)
O4^iii—Pb3—O4^i	67.17 (19)	Pb2—O3—Pb1	107.63 (14)
O4^i—Pb3—O5^v	92.11 (15)	Pb4—O3—Pb1	103.97 (15)
O4^iii—Pb3—O5^vi	92.11 (14)	Pb4—O3—Pb1^iv	103.96 (15)
O4^i—Pb3—O5^vi	152.01 (14)	Pb4—O3—Pb2	117.3 (2)
O4^iii—Pb3—O5^v	152.01 (14)	Pb5—O3—Pb1^iv	107.17 (16)
O4^iii—Pb3—O8	74.03 (16)	Pb5—O3—Pb1	107.17 (16)
O4^i—Pb3—O8^iv	74.03 (16)	Pb5—O3—Pb2	110.4 (3)
O4^iii—Pb3—O8^iv	102.50 (18)	Pb5—O3—Pb4	6.87 (18)
O4^i—Pb3—O8	102.50 (18)	Pb3^i—O4—Pb2^i	98.09 (15)
O5^v—Pb3—O1^v	75.81 (15)	P1—O4—Pb2^i	114.1 (2)
O5^vi—Pb3—O1^v	75.81 (15)	P1—O4—Pb3^i	133.9 (2)
O5^vi—Pb3—O5^v	99.8 (2)	Pb3^ix—O5—Pb1^i	108.18 (15)
O5^v—Pb3—O8	131.03 (16)	P1—O5—Pb1^i	128.9 (3)
O5^v—Pb3—O8^iv	88.65 (17)	P1—O5—Pb3^ix	122.7 (3)
O5^vi—Pb3—O8^iv	131.03 (16)	Pb1—O6—Pb2	96.54 (15)
O5^vi—Pb3—O8	88.65 (17)	P1—O6—Pb1	148.3 (3)
O8^iv—Pb3—O1^v	151.60 (12)	P1—O6—Pb2	114.9 (3)
O8—Pb3—O1^v	151.60 (12)	Pb1^xi—O7—Pb5^xi	91.6 (2)
O8—Pb3—O8^iv	52.8 (2)	Pb4^xi—O7—Pb1^xi	94.52 (17)
Pb1^iv—Pb4—Pb1	66.95 (4)	Pb4^xi—O7—Pb5^xi	4.14 (13)
O3—Pb4—Pb1	38.94 (7)	P1—O7—Pb1^xi	134.1 (3)
O3—Pb4—Pb1^iv	38.94 (7)	P1—O7—Pb4^xi	126.5 (3)
O3—Pb4—O7^vii	79.79 (17)	P1—O7—Pb5^xi	128.0 (3)
O3—Pb4—O7^ii	79.79 (17)	P2—O8—Pb3	99.5 (3)

Symmetry codes: (i) −x, −y+1, −z; (ii) x+1/2, y, −z+1/2; (iii) −x, y+1/2, −z; (iv) x, −y+3/2, z; (v) x+1/2, y, −z−1/2; (vi) x+1/2, −y+3/2, −z−1/2; (vii) x+1/2, −y+3/2, −z+1/2; (viii) x, y, z−1; (ix) x−1/2, y, −z−1/2; (x) x, y, z+1; (xi) x−1/2, y, −z+1/2.