New structural insights into Fe₂P₂O₇ – unravelling an unresolved dispute and three reversible phase transitions

The debate in the literature whether the triclinic room-temperature crystal structure of iron(II) pyrophosphate (Fe₂P₂O₇) is centrosymmetric or not has been clearly resolved on the basis of new single-crystal X-ray intensity measurements. This study additionally revealed that Fe₂P₂O₇ undergoes three reversible phase transitions between −140 and 190°C, with the modifications denoted with decreasing temperature as β, α₃, α₂ and α₁. The room-temperature form, α₂-Fe₂P₂O₇, indeed crystallizes in a centrosymmetric but incommensurately modulated structure, a fact that has not been recognized for more than 40 years. For better comparison with the C-centred monoclinic thortveitite-type aristotype (space group type C2/m), the structure of the hettotype α₂-Fe₂P₂O₇ is described in the superspace group C1(αβγ)0 with a = 6.6393 (6), b = 8.4748 (6), c = 4.4839 (3) Å, α = 90.036 (5), β = 103.962 (7), γ = 92.929 (6)° and a modulation wavevector q = 0.4489 (3)a* + 0.2517 (3)b* + 0.3646 (3)c*. The α₂ modification undergoes two phase transitions towards periodic structures. On heating, a triclinic structure described in C1 with very similar lattice parameters is realized above 85°C for the corresponding α₃ modification. It can be considered as the non-modulated basic structure of the α₂ modification. At about 185°C, α₃-Fe₂P₂O₇ transforms to the thortveitite-type β modification, which remains stable up to at least 1000°C. On cooling the α₂ modification, a triclinic structure of the low-temperature α₁ modification forms below −140°C, which can be considered as a twofold superstructure of the α₃ modification with q = ½a* + ½b* + ½c*. The result of these phase transitions from the thortveitite-type β-modification via the triclinic α₃ phase and the incommensurately modulated triclinic α₂ modification to α₁-Fe₂P₂O₇ is the complete ordering of the pyrophosphate anion in the low-temperature phase with a P—O—P bridging angle of 151.91 (8)°. This ordering is accompanied by the lowering of the coordination number of one half of the Fe²⁺ ions from 6 to 5.

1. Introduction

Iron(II) pyrophosphate (Fe₂P₂O₇) is known to show multiple reversible phase transitions. Crystals of one modification were grown under high-temperature conditions (T > 1100°C) from partially molten samples. The corresponding room-temperature crystal structure has been reported on the basis of single-crystal X-ray data with triclinic symmetry, either with a non-centrosymmetric model in the space group P1 [a = 5.517 (2), b = 5.255 (2), c = 4.488 (1) Å, α = 98.73 (2), β = 98.22 (4), γ = 103.81 (2)°, V = 122.6 (2) ų, Z = 1 (Stefanidis & Nord, 1982)], or with a centrosymmetric model, using the non-standard setting C1 [a = 6.649 (2), b = 8.484 (2), c = 4.488(l) Å, α = 90.4 (3), β = 103.89 (3), γ = 92.82 (3)°, 245.4 (2) ų, Z = 2 (Hoggins et al., 1983)]. Since polymorphism had not been noticed at that time, polymorphic notation (α, β…) (Tolédano et al., 1998) was originally not used for this modification. Very recently, it was reported that this room-temperature phase, eventually designated α and likewise modelled in C1 on the basis of synchrotron powder data, shows two temperature-induced structural phase transitions on heating (Liang et al., 2024). One is reported to occur at about 73°C from the triclinic room-temperature α form to an intermediate α′ phase (assumed to be incommensurately modulated), and subsequently at about 183°C to a monoclinic β form, which is isotypic with β-Mg₂P₂O₇ and described in space group B2₁/c [non-standard setting of space-group type No. 14 with standard setting P2₁/c (Calvo, 1967; Lukaszewicz, 1967a)]. Room-temperature α-Fe₂P₂O₇ is a hettotype of the thortveitite aristotype structure [Sc₂Si₂O₇ (Zachariasen, 1930; Cruickshank et al., 1962)], which has monoclinic symmetry (C2/m, Z = 2, a ≃ 6.6, b ≃ 8.6, c ≃ 4.6 Å, β ≃ 103°). The thortveitite structure type or derivatives thereof are adopted by many representatives (Foord et al., 1993). A second polymorph of Fe₂P₂O₇ was prepared by solid-state reactions at 700°C (Parada et al., 2003). It crystallizes in the space group P2₁/c and was designated the γ polymorph due to its isotypism with the γ form of Co₂P₂O₇ (Kobashi et al., 1997). Compared with the reported α and β modifications of Fe₂P₂O₇, the γ polymorph has a different structural organization and thus shows no direct group–subgroup relationship (Müller & de la Flor, 2024) to the thortveitite structure or derivatives thereof.

The renewed interest in Fe₂P₂O₇ originates from its role as a precursor material for the preparation of the lithium iron phosphate cathode materials LiFePO₄ and Li₂FeP₂O₇ (Hu et al., 2008; Lee et al., 2012; Barpanda et al., 2012; Liu et al., 2015), as a possible impurity phase during charging/recharging of such cathodes (Ong et al., 2008), as a by-product during oxidative de­hydrogenation reactions employing the catalyst FePO₄ (Khan et al., 2010), or as the calcination product of Fe(II) hy­droxy­phosphono­acetate that – under incorporation of ammonia – shows an enhanced proton conductivity (Salcedo et al., 2020).

In view of these studies, it is even more surprising that the true nature of the stable triclinic room-temperature form of Fe₂P₂O₇ is still not fully clarified, including debates in the literature about the correct space group. The choice of the non-centrosymmetric model was initially based on two statistical methods with respect to the recorded X-ray diffraction data (Stefanidis & Nord, 1982), whereas it was argued that the X-ray diffraction data of the other study agreed statistically better with a centrosymmetric distribution (Hoggins et al., 1983). Some years later, the two models were compared, concluding that the centrosymmetric model is favoured over the non-centrosymmetric one, driven both by statistical and crystal-chemical arguments (Baur & Tillmanns, 1986). The centrosymmetric model was also supported by an infrared spectroscopic study (Baran et al., 1986). Nonetheless, a conclusive structure model for the triclinic room-temperature α-Fe₂P₂O₇ modification is missing to this day. Moreover, structural studies regarding the thermal behaviour of this modification with respect to possible phase transitions are restricted to the synchrotron powder study mentioned above. However, due to the limited information one can gain from powder diffraction data, the corresponding results of crystal structure modelling and refinement are frequently associated with certain inaccuracies, especially when it comes to incommensurately modulated structures, as was assumed for the reported α′ phase of Fe₂P₂O₇, for which no further structural details were given (Liang et al., 2024). Since M(II)₂P₂O₇ pyrophosphates with divalent cations of comparable size (M = Mg, Cr, Co, Ni, Cu, Zn, except M = Mn) show at least one phase transition on heating or cooling from the thortveitite aristotype structure to modifications with lower symmetry (Isupov, 2002; Palatinus et al., 2006), it appeared likely that phase transitions can also occur for iron(II) pyrophosphate in both directions.

In light of the background and problems described above, we have grown single crystals of Fe₂P₂O₇ and report here on the crystal structure refinement of the triclinic room-temperature phase (herein designated α₂), shining light on the unresolved problem regarding the structure models reported in the literature. Moreover, we could also determine the crystal structures of the corresponding low-temperature (designated α₁), intermediate high-temperature (α₃) and high-temperature (β) forms of Fe₂P₂O₇ from single-crystal X-ray data.

2. Experimental

2.1. Single-crystal growth

Thortveitite-related Fe₂P₂O₇ can be obtained by synproportionation reactions. Pale-greenish to light-brown single crystals were grown with edge lengths up to 4 mm (Fig. S1 in the supporting information) by chemical vapour transport reactions (Gruehn & Glaum, 2000; Binnewies et al., 2012) in a one-pot reaction in sealed and evacuated silica ampoules (volume ca 18 cm³), starting either from stoichiometric mixtures of FePO₄ and FeP [equation (1)] as described previously (Glaum et al., 1991), or, alternatively, from mixtures of FeP₃O₉, Fe₂O₃ and Fe [equation (2)]. For both reactions, I₂/P mixtures (≃100 mg I₂, ≃10 mg P per ampoule) were employed as transport agents in a temperature gradient 850 → 750°C [equation (3)]; the crystals grew on the wall of the silica ampoule placed in the colder sink region and were removed with diluted hydro­fluoric acid (5 wt%).

FePO₄ and FeP₃O₉ were prepared in the form of polycrystalline material by heating stoichiometric mixtures of Fe(NO₃)₃·9H₂O (Merck, p.A.) and (NH₄)₂HPO₄ (Fluka, 99%) in a porcelain crucible at 600°C for 24 h; FeP was synthesized in a sealed silica ampoule by reacting iron powder (Fluka, 99%) with red phospho­rus (Knapsack, electronic grade) at 750°C for 36 h with small amounts of iodine (Merck, p.A.) as a mineralizer and a slight surplus of 5% of phospho­rus relative to the stoichiometric amount. All synthesized educts were single phase according to X-ray powder diffraction.

2.2. Differential scanning calorimetry

The obtained triclinic α₂-Fe₂P₂O₇ crystals were finely ground, enclosed in aluminium crucibles with a pierced lid and subjected to a NETZSCH DSC-214 Polyma system in the temperature range −170 to 150°C (26.5 mg sample mass, flowing nitro­gen atmosphere with flow rate 40 ml min⁻¹ and a heating/cooling rate of 10°C min⁻¹) and to a NETZSCH DSC-200F3 Maia system in the temperature range 30 to 220°C (38.9 mg sample mass, flowing argon atmosphere with flow rate 20 ml min⁻¹ and a heating/cooling rate of 20°C min⁻¹).

2.3. Temperature-dependent powder X-ray diffraction

For temperature-dependent measurements, an HTK1200 Anton-Paar high-temperature furnace chamber was mounted on a PANalytical X'Pert PRO diffractometer (Bragg–Brentano geometry, Cu Kα radiation, X'Celerator multi-channel detector). The Fe₂P₂O₇ sample was finely ground and placed on a silicon zero-background sample holder. The zero point was previously calibrated with an LaB₆ standard and was automatically adjusted during the measurements with a PC-controllable alignment stage. The sample was heated under nitro­gen atmosphere at 5°C min⁻¹ to the respective temperature and kept for 5 min before measurement of each step to ensure temperature stability; temperature range: 25–1000°C with measurement each 25°C. Refinement of cell parameters was performed with the program TOPAS (Bruker, 2009).

2.4. Vibrational spectroscopy

2.4.1. Attenuated total reflectance Fourier-transform infrared spectroscopy

IR powder spectra of α₂-Fe₂Si₂O₇ were acquired at room temperature from 4000 to 370 cm⁻¹ on a Bruker Tensor 27 FTIR spectrometer equipped with a glo(w)bar MIR light source, a KBr beam splitter and a DLaTGS detector. The undiluted sample powder was pressed on the diamond window of a Harrick MVP 2 diamond attenuated total reflectance (ATR) accessory. Sample and background spectra were averaged from 32 scans at 4 cm⁻¹ spectral resolution. Background spectra were obtained from the empty ATR unit. Data handling was performed with the OPUS 5.5 software (Bruker, 2005).

2.4.2. Temperature-dependent micro-Raman spectroscopy

Micro-Raman measurements were performed at different temperatures on a confocal micro-Raman spectrometer Renishaw RM1000 equipped with a 20 mW Ar⁺ laser (488 nm) for excitation, an ultra-steep edge filter set facilitating measurements as close as >70 cm⁻¹ to the Rayleigh line, a Leica DLML microscope with an Olympus 20×/0.40 ultra-long working distance objective, a 1200 lines mm⁻¹ grating in a 300 mm monochromator and a thermo-electrically cooled CCD detector. The entrance slit and CCD readout were set to quasi-confocal mode. The spectral resolution of the system (apparatus function) was 5–6 cm⁻¹, and the absolute Raman shift was calibrated by the Rayleigh line and the 521 cm⁻¹ line of an Si standard. Spectra were acquired from −30 to 1600 cm⁻¹ to employ the Rayleigh line (0 cm⁻¹) as an internal standard. Instrument control and data handling was done with Galactic Grams32 software.

Samples were enclosed in a Linkam FTIR600 heating/cooling stage equipped with thin glass windows and electronic temperature control. A sample crystal was mounted close to the centre of the heated silver block of the stage and covered with a flat silver lid with a centre hole. Thus, the temperature gradient is considered minimal. The acquisition protocol was −150 to 225°C with steps of 25°C. The acquisition time was 120 s at every temperature step.

2.5. Single-crystal X-ray diffraction

Intensity data for Fe₂P₂O₇ crystals were collected at 197, 127, 27 and −173°C in a dry stream of nitro­gen on a Stoe STADIVARI diffractometer system equipped with a Dectris Eiger CdTe hybrid photon counting detector using Mo Kα radiation. The triclinic crystals, i.e. those investigated at 127, 27 and −173°C, were generally twinned by monoclinic pseudo-symmetry. By selecting tiny fragments, non-twinned crystals could be isolated and were attached to Kapton micro-mounts. To preclude artefacts in the F_obs maps due to low spatial resolution, intensity data were collected up to θ ≃ 40°. Data were processed with X-AREA (Stoe, 2024) and integrated with satellites up to the second order for the α₂ phase. A correction for absorption effects was applied using the multi-scan approach implemented in LANA (Koziskova et al., 2016). An initial model of the α₂ phase was obtained from charge flipping directly in 3+1-dimensional superspace using SUPERFLIP (Palatinus & Chapuis, 2007). The first refinement cycles of the α₃ phase were performed using coordinates of the basic structure of the α₂ phase. The structure of the α₁ phase was solved using the dual-space approach implemented in SHELXT (Sheldrick, 2015). The structures were refined using JANA2020 (Petříček et al., 2023). For the α₃ and β phases, two refinements were performed for each, one with a linear P₂O₇⁴⁻ group (bridging oxygen atom located on a centre of inversion) and one with the oxygen atom modelled as equally disordered about the centre of inversion.

Experimental details for data collections of the four Fe₂P₂O₇ structures are summarized in Table 1. Refinement details are compiled in Table 2 for α₂-Fe₂P₂O₇ and in Table 3 for β-, α₃- and α₁-Fe₂P₂O₇. Selected bond lengths and angles of the modifications are collated in Table 4. Further details of the crystal structure investigations may be obtained from the Cambridge Crystallographic Data Centre (CCDC) on quoting the deposition numbers listed at the end of Table 1. The data can be obtained free of charge via https://www.ccdc.cam.ac.uk/structures.

Table 1
Crystal data and details of data collections

Phase	β-Fe2P2O7	α3-Fe2P2O7	α2-Fe2P2O7	α1-Fe2P2O7
Formula weight	285.60	285.60	285.60	285.60
Diffractometer	Stoe STADIVARI	Stoe STADIVARI	Stoe STADIVARI	Stoe STADIVARI
Radiation, wavelength (Å)	Mo Kα, 0.71073	Mo Kα, 0.71073	Mo Kα, 0.71073	Mo Kα, 0.71073
Temperature (°C)	197	127	27	−173
Space group (No.)	C2/m (12)	C1 (2)	C1(αβγ)0	X1 (2)
Centering vectors		(0, 0, 0)^T, (1/2, 1/2, 0)^T	(0, 0, 0, 0)^T, (1/2, 1/2, 0, 0)^T	(0, 0, 0)^T, (−1/4, 1/4, 0)^T, (−1/2, 1/2, 0)^T, (−3/4, 3/4, 0)^T
q vector	–	–	0.4489 (3)a* + 0.2517 (3)b* + 0.3646 (3)c*	–
Crystal description	Light-brownish fragment	Light-brownish plate	Light-brownish plate	Light-brownish plate
Crystal dimensions (mm)	0.04 × 0.07 × 0.11	0.01 × 0.04 × 0.06	0.01 × 0.04 × 0.06	0.01 × 0.04 × 0.06
Formula units, Z	2	2	2	8
a (Å)	6.6154 (6)	6.6161 (5)	6.6393 (6)	13.3690 (5)
b (Å)	8.4624 (10)	8.4673 (4)	8.4748 (6)	16.9853 (4)
c (Å)	4.4920 (4)	4.4825 (3)	4.4839 (3)	6.3545 (3)
α (°)	90	90.135 (5)	90.036 (5)	50.200 (3)
β (°)	103.502 (7)	103.645 (7)	103.962 (7)	72.409 (4)
γ (°)	90	91.616 (7)	92.929 (6)	94.100 (4)
Volume (Å^3)	244.52 (4)	243.92 (3)	244.50 (3)	985.17 (11)
μ (mm^−1)	6.566	6.582	6.566	6.519
Density calc. (g cm^−3)	3.880	3.889	3.880	3.852
Range θmin–θmax (°)	3.98–40.76	3.92–41.04	2.84–37.04	2.80–41.06
Range h, k, l, m	−12 → 11,	−11 → 12,	−12 → 11,	−24 → 25,
	−14 → 15,	−9 → 15,	−14 → 13,	−10 → 10,
	−8 → 3	−8 → 8	−6 → 8,	−16 → 16
	–	–	−2 → 2	–
Measured reflections	3414	7257	21588	15018
Independent reflections	826	1563	5785	3198
Observed reflections [I > 3σ(I)]	815	1341	3402	2830
Ri	0.0119	0.0158	0.0360	0.0161
Absorption correction	Multi-scan	Multi-scan	Multi-scan	Multi-scan
Coef. of transm. Tmin, Tmax	0.47, 0.56	0.57, 0.69	0.36, 0.55	0.45, 0.54
Structure solution, refinement	–, JANA2020	–, JANA2020	SUPERFLIP, JANA2020	SHELXT, JANA2020
CSD Nos.	2455108 (single site), 2455107 (split site)	2455109 (single site), 2455106 (split site)	2455105	2455104

Table 2
Refinement details of the incommensurately modulated α₂-Fe₂P₂O₇

Phase	α2-Fe2P2O7
No. of parameters	251
Extinction coefficient (Becker & Coppens, 1974)	Isotropic type I, 690 (50)
Difference electron density min, max (e Å^−3)	−1.15, 1.13
R[F^2 > 3σ(F^2)], wR(F^2 all)
all reflections	0.0265, 0.0805
main reflections	0.0241, 0.0834
first-order satellites	0.0264, 0.0641
second-order satellites	0.0606, 0.1767
GooF	1.30

Table 3
Refinement details of the periodic β-, α₃- and α₁-Fe₂P₂O₇ structures

	β-Fe2P2O7 (split O3)	β-Fe2P2O7 (single O3)
No. of parameters	33	32
Extinction coefficient (Becker & Coppens, 1974)	Isotropic type I, 320 (90)	Isotropic type I, 330 (100)
Difference electron density min, max (e Å^−3)	−0.68, 0.58	−0.69, 0.62
R[F^2 > 3σ(F^2)], wR(F^2 all)	0.0188, 0.0800	0.0194, 0.0825
GooF	1.821	1.878

	α3-Fe2P2O7 (split O4)	α3-Fe2P2O7 (single O4)	α1-Fe2P2O7
No. of parameters	53	56	101
Extinction coefficient (Becker & Coppens, 1974)	Isotropic type I, 540 (60)	Isotropic type I, 540 (60)	Isotropic type I, 130 (30)
Difference electron density min, max (e Å^−3)	−0.18, 0.34	−0.15, 0.37	−1.02, 0.70
R[F^2 > 3σ(F^2)], wR(F^2 all)	0.0224, 0.0645	0.0218, 0.0627	0.0168, 0.0408
GooF	1.20	1.17	1.28

Table 4
Selected bond lengths (Å) and angles (°) in the Fe₂P₂O₇ modifications

(a) β-Fe₂P₂O₇.

Single-site model		Split-site model
Fe1—O1	2.1143 (6)	Fe1—O1	2.1142 (6)
Fe1—O1^i	2.1143 (6)	Fe1—O1^i	2.1142 (6)
Fe1—O2^ii	2.3141 (9)	Fe1—O2^ii	2.3141 (9)
Fe1—O2^iii	2.3141 (9)	Fe1—O2^iii	2.3141 (9)
Fe1—O2^iv	2.0780 (8)	Fe1—O2^iv	2.0782 (8)
Fe1—O2^v	2.0780 (8)	Fe1—O2^v	2.0782 (8)
P1—O1	1.5200 (9)	P1—O1	1.5201 (8)
P1—O2	1.5205 (8)	P1—O2	1.5203 (8)
P1—O2^vi	1.5205 (8)	P1—O2^vi	1.5203 (8)
P1—O3	1.5596 (4)	P1—O3	1.5759 (7)
P1—O3—P1^vii	180	P1—O3—P1^vii	163.5 (3)
		O3⋯O3^viii	0.453 (6)
Symmetry codes: (i) −x + 1, y, −z + 1; (ii) x, y, z + 1; (iii) −x + 1, y, −z; (iv) −x + 1/2, −y + 1/2, −z; (v) x + 1/2, −y + 1/2, z + 1; (vi) x, −y, z; (vii) −x, y, −z; (viii) −x, −y, −z.

(b) α₃-Fe₂P₂O₇.

Single-site model		Split-site model
Fe1—O1	2.0866 (8)	Fe1—O1	2.0868 (8)
Fe1—O1^i	2.1371 (9)	Fe1—O1^i	2.1370 (8)
Fe1—O2^ii	2.2347 (12)	Fe1—O2^ii	2.2347 (11)
Fe1—O2^iii	2.1128 (10)	Fe1—O2^iii	2.1127 (9)
Fe1—O3^iv	2.0377 (10)	Fe1—O3^iv	2.0376 (10)
Fe1—O3^v	2.4021 (13)	Fe1—O3^v	2.4022 (13)
P1—O1	1.5173 (8)	P1—O1	1.5173 (8)
P1—O2	1.5164 (10)	P1—O2	1.5164 (9)
P1—O3	1.5218 (10)	P1—O3	1.5219 (10)
P1—O4	1.5593 (4)	P1—O4	1.552 (8)
		P1—O4^vi	1.600 (8)
P1—O4—P1^v	180	P1—O4—P1^v	163.3 (3)
		O4⋯O4^vi	0.461 (7)
Symmetry codes: (i) −x + 1, −y, −z + 1; (ii) x, y, z + 1; (iii) −x + 1/2, −y + 1/2, −z; (iv) x + 1/2, y + 1/2, z + 1; (v) −x, −y, −z; (vi) −x + 1, −y, −z.

(c) α₂-Fe₂P₂O₇.

	Average	Minimum	Maximum
Fe1—O1	2.0724 (19)	2.039 (2)	2.112 (2)
Fe1—O1^i	2.151 (2)	2.119 (2)	2.174 (2)
Fe1—O2^ii	2.178 (2)	2.130 (3)	2.226 (3)
Fe1—O2^iii	2.134 (2)	2.104 (2)	2.168 (2)
Fe1—O3^iv	2.013 (2)	1.930 (2)	2.066 (2)
Fe1—O3^v	2.550 (3)	2.245 (3)	2.962 (3)
P1—O1	1.5201 (19)	1.504 (2)	1.526 (2)
P1—O2	1.520 (2)	1.481 (2)	1.525 (2)
P1—O3	1.512 (2)	1.493 (2)	1.529 (2)
P1—O4	1.5760 (19)	1.494 (3)	1.592 (3)
Symmetry codes: (i) −x1 + 1, −x2, −x3 + 1, −x4; (ii) x1, x2, x3 + 1, x4; (iii) −x1 + 1/2, −x2 + 1/2, −x3, −x4; (iv) x1 + 1/2, x2 + 1/2, x3 + 1, x4; (v) −x1 + 1, −x2, −x3, −x4.

(d) α₁-Fe₂P₂O₇.

Fe1a—O1a	2.0580 (11)	Fe1b—O1a	2.1742 (8)
Fe1a—O1b^i	2.1397 (8)	Fe1b—O1b^i	2.0860 (12)
Fe1a—O2a^ii	2.1005 (8)	Fe1b—O2a^v	2.1201 (7)
Fe1a—O2b^iii	2.1668 (8)	Fe1b—O2b^vi	2.1903 (8)
Fe1a—O3a^iv	1.9756 (12)	Fe1b—O3b^vii	2.0325 (11)
		Fe1b—O3b^iii	2.2463 (7)
P1a—O1a	1.5254 (9)
P1a—O2a	1.5238 (8)
P1a—O3a	1.5033 (12)
P1a—O4	1.5921 (8)
P1b—O1b	1.5194 (9)
P1b—O2b	1.5209 (9)
P1b—O3b	1.5228 (11)
P1b—O4	1.5795 (8)
P1a—O4—P1b	151.91 (8)
Symmetry codes: (i) x + 1/4, y − 1/4, z + 1; (ii) −x + 1/4, −y + 3/4, −z; (iii) −x − 1/4, −y + 1/4, −z + 1; (iv) x, y, z + 1; (v) −x + 1/2, −y + 1/2, −z; (vi) −x, −y, −z + 1; (vii) x + 1/4, y − 1/4, z.

3. Results

3.1. Thermal behaviour

As revealed by a combination of differential scanning calorimetry (DSC), temperature-dependent powder X-ray diffraction (PXRD) and Raman measurements, Fe₂P₂O₇ undergoes three reversible structural phase transitions in the range −120 to 190°C (Figs. 1–3).

Figure 1 (a) DSC curves of Fe2P2O7 in the range −160 to 140°C showing the reversible phase transitions α1 ⇌ α2 (left) and α2 ⇌ α3 (right). (b) DSC curves of Fe2P2O7 (heating red, cooling blue with curves of the first derivative in orange and lilac, respectively) in the range 30 to 220°C showing the reversible phase transitions α2 ⇌ α3 (left) and α3 ⇌ β (right).

Based on the shapes of the DSC curves, the low-temperature (α₁) ⇌ room-temperature (α₂) phase transition is endothermic on heating and exothermic on cooling and proceeds slowly with a somewhat large hysteresis (onset at −121.8°C on heating and at −131.1°C on cooling). The room-temperature (α₂) ⇌ intermediate high-temperature (α₃) transition likewise is endothermic on heating and exothermic on cooling but proceeds with a greater transition enthalpy [Fig. 1(a)]. The hysteresis is moderate, with onset temperatures of 84.3°C on heating and at 89.6°C on cooling. The determined temperatures for the latter transition are in a similar range as reported previously (∼73°C; Liang et al., 2024) and are consistent with the temperature-dependent PXRD data (Fig. 2), which show a weak but clearly visible step-change in some reflections at around 88°C. The subsequent α₃ ⇌ β (high-temperature) phase transition is indicated by barely perceptible effects in the DSC curves, seen as very weak shoulders [Fig. 1(b)] with onsets of about 186°C on heating and about 199°C on cooling. This behaviour indicates only very small energetic differences between the α₃ and β phases. The determined temperatures for the latter phase transition are in rough agreement with the previously reported value of about 183°C (Liang et al., 2024). On the other hand, the α₃ ⇌ β phase transition is clearly visible in the temperature-dependent PXRD measurements and is characterized by a convergence of reflections from the triclinic α₃ phase to the monoclinic high-temperature β form, which is completed at about 190°C (Fig. 2). The absence of a well resolved peak in the DSC curve and the slow convergence of the lattice parameters indicate a higher order for the α₃ ⇌ β phase transition, while the other phase transitions are most likely of first order. The evolution of lattice parameters of the α₃ and β phases with temperature is given in Fig. S2.

Figure 2 Temperature-dependent PXRD measurements of Fe2P2O7 in the range 25–1000°C with the stability fields of the modifications indicated.

Raman spectra of Fe₂P₂O₇ have been recorded between −150 and 225°C in 25°C intervals. The observed bands in the spectra vary with temperature (Fig. 3) and provide additional evidence for the three reversible structural phase transitions already identified by thermal analysis. The continuously increasing number of bands with decreasing temperature also indicates a decrease in the symmetry of the corresponding crystal structures.

Figure 3 Fe2P2O7. Temperature-dependent Raman spectra for the modifications α1 (T ≤ −140°C), α2 (−140 ≤ T ≤ 70°C), α3 (70°C ≤ T ≤ 190°C) and β (T ≥ 190°C).

The ambient-temperature Raman and IR spectra (Fig. 4) match those reported in the literature (Baran et al., 1986). These spectra show the typical signals for metal(II) pyrophosphates (Rulmont et al., 1991; Popović et al., 2005). The main spectral features are readily assigned to ν_as(PO₃) and ν_s(PO₃) (1000 to 1200 cm⁻¹), ν_as(POP) (900 to 1000 cm⁻¹), ν_s(POP) (IR: 714, Raman: 719 cm⁻¹), δ_as(PO₃) (500 to 600 cm⁻¹), and δ_s(PO₃) (426 cm⁻¹) according to the literature (Baran et al., 1986). With respect to the focus of this study, the energy, shape and intensity (Raman) of the band related to the symmetric stretching vibration at 719 cm^–1 deserves particular attention. Apart from a slightly lower relative intensity in the incommensurately modulated α₂ modification and the commensurately modulated α₃ modification at −125 and −150°C, respectively (Fig. 3), this emission shows no variation over the entire temperature range. This suggests that the bond angle and P—O distances in the (P—O—P) core of all pyrophosphate groups in the modifications α₁, α₂, α₃ and β are virtually identical, despite the structural complexity of these modifications described in subsequent sections. In line with observations for iron(II) silicates [orthopyroxene (Mg,Fe)­SiO₃ (Goldman & Rossman, 1977); Fe₂SiO₄ (Runciman et al., 1973)], the weak broad hump arising around 1400 cm^–1 in the Raman spectra of the α₂ and the α₃ modifications at the lowest temperatures (Fig. 3, −125 and −150°C) can be attributed to electronic transitions within the low-symmetry split levels of the ⁵T_2g (ideal O_h symmetry) ground state.

Figure 4 α2-Fe2P2O7. IR and Raman spectra at ambient temperature.

3.2. Crystal structures

3.2.1. The thortveitite-type aristotype structure (β-Fe₂P₂O₇)

Above 190°C, the β-Fe₂P₂O₇ modification is stable and exists at least up to 1000°C, which was the highest temperature accessible for our measurements. Since monoclinic thortveitite-type β-Fe₂P₂O₇ is the aristotype of the triclinic α₃-, α₂- and α₁-Fe₂P₂O₇ modifications (hettotypes), its crystal structure is discussed first. The basic structural features of β-Fe₂P₂O₇ can be described as made up of alternating layers (metal cations; pyrophosphate anions) parallel to (001) [Fig. 5(a)].

Figure 5 β-Fe2P2O7 (thortveitite-type, C2/m, Z = 2). Crystal structure with schematic polyhedra [Fe(II)O6] (blue) and P2O7^4− groups (yellow) (a) viewed along [010] and (b) as a projection onto (001); atoms given as spheres of arbitrary radius; the unit cell is indicated in red. ORTEP style representations of the polyhedra from refinement of the bridging oxygen atom O3 (c) on a single site and (d) with split sites; ellipsoids are given at the 50% probability level; symmetry codes refer to Table 4. [Additional symmetry codes (ix) x, y, −1 + z; (x) x + ½, −y + ½, z; (xi) −x + 1, y, −z − 1; (xii) −x + ½, −y + ½, −z − 1.]

The P₂O₇⁴⁻ anion (composed of two corner-sharing PO₄ groups) is located on a 2/m position [Fig. 5(b)]. Owing to the imposed symmetry, the conformation of the pyrophosphate anion is staggered. The bridging oxygen atom shows enlarged anisotropic displacement parameters (ADPs) and can either be modelled on a single site [Fig. 5(c)] or on split sites with half occupancy [Fig. 5(d)]. Although the crystal structure of the eponymous mineral thortveitite, Sc₂Si₂O₇, is usually described with the bridging oxygen atom located at the 2/m position (Cruickshank et al., 1962), it is still debated whether the bridging oxygen atom in the aristotypic phases should be considered as lying at the 2/m position or as dynamically disordered around the mirror plane. Combined neutron/X-ray diffraction studies of thortveitite-type Mn₂P₂O₇ (Stefanidis & Nord, 1984) corroborate the disorder model. The split-position refinement features slightly better residuals (Table 3), which is not surprising as the electron density around the origin is modelled with more parameters. A detailed discussion of the differences between a single-site refinement and a split-position refinement is given in the next section for the crystal structure of α₃-Fe₂P₂O₇. For simplicity, we use the one-site (2/m) model to describe the crystal structure of β-Fe₂P₂O₇ in the following. The bond lengths in the P₂O₇⁴⁻ group follow the general trend in pyrophosphate anions (Clark & Morley, 1976; Durif, 1995), where the P—O bond lengths to the terminal atoms are shorter [O1: 1.5200 (9) Å, O2: 2 × 1.5205 (8) Å] than to the bridging atom [O3: 1.5596 (4) Å].

Fe²⁺ ions in between the layers of pyrophosphate anions are located on twofold rotation axes. The coordination number of the unique Fe²⁺ ion is 6, and the distorted octahedral coordination polyhedron is defined by the non-bridging atoms of the P₂O₇⁴⁻ anion with four short [2 × 2.0780 (8); 2 × 2.1143 (6) Å] and two longer [2 × 2.3141 (9) Å] Fe—O bonds to the oxygen atoms located at opposite sides. By edge sharing, the [FeO₆] polyhedra form a honeycomb pattern where one third of the octahedral voids in the resulting layer remain unoccupied [Fig. 5(b)].

3.2.2. The crystal structure of α₃-Fe₂P₂O₇

On cooling below 190°C, monoclinic β-Fe₂P₂O₇ transforms into a triclinic structure, α₃-Fe₂P₂O₇, obtained by a translationengleiche symmetry descent (Müller & de la Flor, 2024) of index 2. To simplify the comparison with the known thortveitite phases, the triclinic primitive setting was transformed into the C-centred setting of the parent structure, and the relationship of the setting in space group C1 to the reduced setting is a = −b_red − c_red, b = −b_red + c_red, c = −a_red. The metrical deviation from monoclinic symmetry is most pronounced for the γ angle of 91.616 (7)°.

The crystal structure of α₃-Fe₂P₂O₇ (Fig. 6) is closely related to that of the aristotype β phase. Owing to the symmetry descent from monoclinic to triclinic, the P₂O₇⁴⁻ unit no longer has 2/m symmetry but is located with its bridging atom O4 on an inversion centre; bond lengths and angles in the anion are very similar to the monoclinic modification (Table 4). The unique Fe²⁺ ion of α₃-Fe₂P₂O₇ sits on a general position and retains its octahedral coordination polyhedron, however with a larger distortion [2.0377 (10) ≤ d(Fe−O) ≤ 2.4021 (13) Å].

Figure 6 a3-Fe2P2O7 (C1, Z = 2). Crystal structure with schematic polyhedra [Fe(II)O6] (blue) and P2O7^4− groups (yellow) (a) viewed along [010] and (b) as a projection onto (001); atoms given as spheres of arbitrary radius; the unit cell is indicated in red. ORTEP style representations of the polyhedra from refinement of the bridging oxygen atom O4 (c) on a single site and (d) with split sites; ellipsoids are given at the 50% probability level; symmetry codes refer to Table 4. [Additional symmetry codes; (vii) x, y, z − 1; (viii) −x + 1, −y, −z; (ix) −x + 1, −y, −z − 1; (x) x + ½, y + ½, z; (xi) −x + ½, −y + ½, −z − 1]. Note the monoclinic pseudo-symmetry, notably a pseudo-reflection plane parallel to (010) passing through the P2O7^4− unit.

Again, the split-position refinement of O4 [Fig. 6(d)] featured slightly better residuals (Table 3). The highest electron density of the bridging O4 atom is observed at the origin, even in the split-position refinement (Fig. 7).

Figure 7 2 × 2 Å^2 (y, z) Fobs contour plot of the electron density resulting from the split-position refinement of α3-Fe2P2O7 showing an electron density maximum at the origin, corresponding to a linear P2O7^4− unit. Contours are drawn at the 1 e Å^−3 levels. Red discs represent the refined position of the O4 atom and its image by inversion at the origin.

For a simplified discussion with regards to the linearity of the pyrophosphate group, we will only consider displacement in the direction of the longest eigenvector of the ADP tensor of bridging atom O4, which is virtually parallel to [001]. As expected, the mean square displacement in that direction is more pronounced for the single-position model with U = 0.110 Ų, compared with U = 0.042 Ų for the split-position model. However, the latter can still be considered as a strong displacement indicative of additional disorder.

To show the issues in interpreting such models, Fig. 8 compares the one-dimensional probability distribution of the two models: for the single-position model a normal distribution with the corresponding variance is shown; for the split-position model, two normal distributions with weight ½ spaced by the distance of the two disordered positions are summed. Remarkably, the split-position model features a basically flat distribution at the origin. The small dip at the origin is due to the impossibility of modelling a flat distribution as the sum of two normal distributions and could be remedied by adding an additional position at the origin. It should be stressed that the probability distributions in Fig. 8 represent the densities of finding the barycentre of the atom at the given position, not an electron density distribution. Thus, even though the split-position refinement results in a nominal P—O—P angle of 163.3 (2)°, a linear angle is just as likely according to the refined split-position model. We conclude that in the case of a significant overlap of probability densities, geometric parameters must not be derived from atomic positions, as they are meaningless.

Figure 8 Comparison of one-dimensional normal distributions of the positions of O4 in α3-Fe2P2O7 using the largest eigenvalues of the ADP tensor in the single position and in the split-position models. The distributions of the individual atoms in the split-atom model are given in black.

In summary, based on the present data, the P₂O₇⁴⁻ group indeed adopts a linear configuration within experimental precision. In fact, configurations with angles ca 160–180° appear to possess similar likelihoods. To determine whether linear P₂O₇⁴⁻ units are actually realized, one could resort to high-resolution neutron diffraction experiments, since there the atomic positions are not convolved with diffuse electron clouds.

3.2.3. The crystal structure of α₂-Fe₂P₂O₇

The room-temperature phase α₂-Fe₂P₂O₇ exhibits an incommensurately modulated structure, as shown by the appearance of satellite reflections (Figs. S3 and S4). The basic structure of α₂-Fe₂P₂O₇, i.e. the structure with the reference positions for the modulated structure (Fig. 9), is the triclinic C1 structure of α₃-Fe₂P₂O₇ (Fig. 6), and the two structures appear to be virtually identical. Compared with the α₃ modification, the γ angle [92.929 (6)°] shows an even larger deviation from monoclinic metrics.

Figure 9 a2-Fe2P2O7. Section of the crystal structure (–0.03 ≤ z ≤ 0.13) in a commensurate approximation (P1, Z = 792) with schematic polyhedra [Fe(II)O6] (blue), [Fe(II)O5] (teal) and P2O7^4− groups (yellow) viewed along [001].

The 3+1-dimensional superspace group of α₂-Fe₂P₂O₇ is C1(αβγ)0. In the known monoclinic incommensurately modulated thortveitite-type structures of α₂-Cr₂P₂O₇ (Palatinus et al., 2006), α₂-Zn₂P₂O₇ (Stöger et al., 2014) or β-Zn₂As₂O₇ (Weil & Stöger, 2010), the modulation wavevector q lies in the (a*, c*) plane for symmetry reasons, which means that translation symmetry is fully retained in the [010] direction. In contrast, α₂-Fe₂P₂O₇ has a distinct b* component of about one quarter: q ≃ 0.449a* + 0.252b* + 0.366c* at 27°C. The vector q itself changes with temperature, e.g. for a −118°C measurement (i.e. shortly before the transition to the α₃ form) q ≃ 0.452a* + 0.291b* + 0.381c*.

To simplify the modelling and discussion of the modulated P₂O₇⁴⁻ group in the α₂ structure, we placed the origin of the basic structure onto the inversion centre where this group is located. Electron density maps in superspace suggested that the positional modulation of the bridging O4 atom in the incommensurately modulated structure is best described by a function with a single point of discontinuity per period. To both sides of the point of discontinuity, the O4 atom is located at distinct positions without ever adopting any intermediate position (for details, see below).

The modulation of the bridging oxygen atom (O4) is fundamentally different from the monoclinic incommensurate thortveitites. There, the bridging oxygen atom was modelled using a segment valid for half of the internal space moved away from the 2/m position. The other half of the internal space was generated by a symmetry operation acting on 3+1-dimensional superspace. Thus, any X₂O₇ group in the actual structure is distinctly bent and located to either side of the 2/m position, with two points of discontinuity per internal space period. In α₂-Fe₂P₂O₇, the electron density of O4 in superspace adopts a distinctly different shape and the positional modulation is better described as a sawtooth-like function with a single point of discontinuity per period in the internal space (Fig. 10). The best results in the sense of giving the most reasonable interatomic distances were obtained by modelling the O4 atom with x-harmonics (Petříček et al., 2016) up to order 2 (4 refined coefficients). To obtain reasonable inter­atomic distances and angles within the PO₄ groups (Fig. 11) for all t values, the other phosphorus and oxygen atoms needed to be modelled with x-harmonics, even though the point of discontinuity was not obvious from superspace electron densities maps (Fig. 12). The points of discontinuities were fixed at half integer t, in accordance with the point of discontinuity of O4. All phosphorus and oxygen atoms were modelled with x-harmonics up to second order for positional and displacement parameters.

Figure 10 2 Å-wide (x2, x4) superspace section at the origin (position of the bridging O4 atom) of α2-Fe2P2O7 showing a single point of discontinuity per internal space period at half-integer t values. Contours are drawn at the 2 e Å^−3 levels. The refined centre of gravity of the O4 atom is represented by a blue line. Positional modulation is distinctly less pronounced in the x1 and x3 directions and is not shown.

Figure 11 t-plot of P—O distances in α2-Fe2P2O7 with a point of discontinuity at t = ½.

Figure 12 2 Å-wide (x1, x4) superspace section of α2-Fe2P2O7 centred at the O3 atom showing a single point of discontinuity per internal space period at half-integer t values, which is not obvious from the electron density. Contours are drawn at the 2 e Å^−3 levels. Positional modulation is even more subtle in the x1 and x3 directions and is not shown.

In the given model, the modulation function of the O4 atom passes through the origin at integral t, which means that the P—O—P angle becomes arbitrarily close to 180° in parts of the structure (Fig. 13). In fact, for integral t, the highest electron density is observed at the origin (Fig. 14). Even when introducing a further point of discontinuity at t = 0, the two resulting sections of the modulation function converged to nearly meet at the origin.

Figure 13 t-plot of the P—O—P angle in α2-Fe2P2O7, which becomes linear at integral t.

Figure 14 2 × 2 Å^2 (x2, x3) Fobs contour plot of the electron density of α2-Fe2P2O7 at t = 0 centred at the origin (position of O4) showing an electron density maximum at the origin, corresponding to a linear P2O7^4− unit. Contours are drawn at the 1 e Å^−3 levels.

We assume that a situation similar to the non-modulated α₃ phase arises at integral t, with the U²² ADP tensor element of O4 being significantly increased at t = 0 (Fig. 15). U²² describes the displacement of O4 nearly perpendicular to the P—P segment in the [010] direction. Observe that at half-integer t, the U²² element is likewise increased, indicating additional disorder at the point of discontinuity.

Figure 15 t-plot of the U^22 and U^33 ADP tensor elements of the bridging O4 atom in α2-Fe2P2O7. The remaining tensor elements are significantly smaller and are not shown.

Another special feature in the known incommensurate thortveitite structures concerns the metal oxide layers and the modulation of the M—O distances. In α₂-Fe₂P₂O₇, the Fe²⁺ ion is five- and sixfold coordinated in different positions of internal space (Fig. 16), owing to sometimes close and sometimes more distant O3 atoms. The ratio of five- to six-coordination is approximately 1:1. Note that close to the point of discontinuity, the Fe1—O3 bond length of the close O3 atom is unreasonably short. This is most likely a refinement artefact and can be remedied by using x-harmonics for Fe1 as well. However, then other Fe1—O distances become too large. In fact, the Fe1 atom connects to six distinct P₂O₇⁴⁻ groups with distinct points of discontinuities in the internal space. Thus, the Fe1 atom would have to be described using six distinct points of discontinuity per internal space period. This appears excessive, and we ultimately applied harmonic modulation functions up to the second order for positional and displacement parameters of the Fe1 atom. Fig. 9 shows an excerpt of a slab parallel to (001), exemplifying the distribution of fivefold and sixfold coordination around the Fe²⁺ ion. Note that, in contrast to the monoclinic incommensurate thortveitites, translation symmetry is lost in the [010] direction (although it is close to a fourfold superstructure).

Figure 16 t-plot of the Fe−O distances in α2-Fe2P2O7 showing distinct regions of fivefold and sixfold coordination.

3.2.4. The crystal structure of α₁-Fe₂P₂O₇

The α₁-Fe₂P₂O₇ modification forms below −140°C and is likewise triclinic. It is a twofold superstructure of α₃-Fe₂P₂O₇ with the modulation wavevector q = ½a* + ½b* + ½c* and obtained by a klassengleiche symmetry descent (Müller & de la Flor, 2024) of index 2. The `second-order satellite' of an h + k even reflection has again h + k even and therefore does not violate the reflection condition for the C1 space group of α₃-Fe₂P₂O₇. The modulation wavevector of α₁-Fe₂P₂O₇ does not relate to that of α₂-Fe₂P₂O₇, notably with respect to the b and c components, which are closer to one quarter and one half in α₂-Fe₂P₂O₇, respectively. The crystal structures of α₁- and α₂-Fe₂P₂O₇ are both modulated variants of the α₃ phase. However, due to the significant change in the modulation vector, it is difficult to classify the α₁ → α₂ transition as being of the lock-in type, unlike the other known incommensurate thortveitite-type phases.

To better relate the unit cells of α₁- and α₃-Fe₂P₂O₇, the structure of the former is described in the unusual setting (a_α1, b_α1, c_α1) = (2a_α3, 2b_α3, ½a_α3 + ½b_α3 + c_α3). In this setting, the additional centring vectors are −¼a + ¼b, −½a + ½b and −¾a + ¾b, and the resulting non-standard centring is represented by the Bravais symbol `X'. The crystallographic information file (CIF) for α<sub>3</sub>-Fe<sub>2</sub>P<sub>2</sub>O<sub>7</sub> according to the standard setting in P1 is available in the supporting information. The γ angle in the chosen X1 setting is directly comparable to that of α<sub>3</sub> and has moved even further away from 90° than that of the α<sub>2</sub> modification (Table 1). The symmetry relationship between α<sub>1</sub>-Fe<sub>2</sub>P<sub>2</sub>O<sub>7</sub> in X1 and α<sub>3</sub>-Fe<sub>2</sub>P<sub>2</sub>O<sub>7</sub> in C1 is of the isomorphe type with index 2 (Müller & de la Flor, 2024). Every second centre of inversion is lost, in particular the point-group symmetry of the P<sub>2</sub>O<sub>7</sub><sup>4−</sup> unit is reduced from 1 in the α<sub>3</sub> modification to 1 in the α<sub>1</sub> modification. The Fe1 atom of α<sub>3</sub>-Fe<sub>2</sub>P<sub>2</sub>O<sub>7</sub> is split into two fully occupied positions, Fe1a and Fe1b, in α<sub>1</sub>-Fe<sub>2</sub>P<sub>2</sub>O<sub>7</sub>. In general, split positions are named according to the α<sub>3</sub>-Fe<sub>2</sub>P<sub>2</sub>O<sub>7</sub> phase with `a' or `b' appended.

The P₂O₇⁴⁻ group adopts the angle ∠(P—O—P) = 151.91 (8)° and appears in two orientations related by the inversion operation (Fig. 17). This situation is energetically more favourable than a linear P—O—P angle and is realized in all M(II)₂P₂O₇ low-temperature structures: M = Mg 144° (Calvo, 1967); Cr 144.9, 140.1° (Palatinus et al., 2006); Co 143° (Krishnamachari & Calvo, 1972); Ni 137° (Lukaszewicz, 1967b); Cu 156.8 (Effenberger, 1990); and Zn 140.6, 148.5° (Stöger et al., 2014). Again, the averaged values of the P—O_bridging bond lengths (average 1.586 Å) are greater than the corresponding values of the P—O_terminal bond lengths (average 1.519 Å).

Figure 17 An Fe^2+ layer and adjacent P2O7^4− layers in α1-Fe2P2O7 (a) viewed along [010] and (b) projected on (001), showing fivefold (teal) and sixfold (blue) coordination of Fe. The unit cell is indicated by red lines. The non-standard −¼a + ¼b centering translation is indicated by a red arrow. (c) ORTEP style representations of the polyhedra; ellipsoids are given at the 50% probability level; symmetry codes refer to Table 4. [Additional symmetry codes: (viii) −x, −y + 1, −z − 1; (ix) x + ¾, y + ¼, z − 1; (x) x + ½, y + ½, z − 1; (xi) −x + ¼, −y + ¾, −z − 1.]

One of the two unique Fe²⁺ ions (Fe1a) has changed its coordination number from six to five, accompanied by an overall shortening of the remaining five Fe—O bond lengths in comparison with six-coordinate Fe²⁺ ions of about 0.05 Å for averaged values. The corresponding coordination polyhedron around Fe1a is a distorted square pyramid [the geometry index τ₅ = 0.34 (Addison et al., 1984), where the ideal values for τ₅ are 0 and 1 for a square pyramid and a trigonal bipyramid, respectively], with the apical atom O2b exhibiting the longest bond of 2.1668 (3) Å in the polyhedron. The `next-nearest' oxygen atom to augment the coordination number of Fe1a to 6 is atom O3a at a distance of 3.1254 (4) Å. This distance is by far too long to be considered as a primary bond because the contribution of this long bond to the bond-valence sum (Brown, 2002) is only 0.023 valence units using the bond-valence parameters provided by Gagné & Hawthorne (2015); the calculated bond-valence sum when considering the five close oxygen atoms is 1.932 (2) valence units. The second Fe²⁺ ion (Fe1b) retains octahedral coordination, which appears much more regular than in the high-temperature α₃ modification; the mean bond length for equatorial oxygen atoms is 2.10 Å, while for axial oxygen atoms it is 2.22 Å. The calculated bond-valence sum for Fe1b is 2.059 (2) Å.

In the α₁-Fe₂P₂O₇ modification, the sequence of coordination polyhedra along a is 5–6–6–5 5–6–6–5 according to the notation introduced by Palatinus et al. (2006), whereby the number represents the coordination number of the central atom, the dash connects two edge-sharing polyhedra and a space separates two neighbouring polyhedra that do not share any atom (Fig. 17).

4. Discussion

To date, Fe₂P₂O₇ is the fourth example of a thortveitite-like structure where incommensurability of one of the modifications occurs, here realized for the room-temperature α₂ phase intermediate between the commensurate α₃ and α₁ phases at higher and lower temperatures, respectively. The other examples are Cr₂P₂O₇ (Palatinus et al., 2006), Zn₂P₂O₇ (Stöger et al., 2014) and Zn₂As₂O₇ (Weil & Stöger, 2010). High-temperature β-Cr₂P₂O₇ crystallizes in the thortveitite aristotype and on cooling subsequently transforms into a structurally unknown modification at 91°C, then into incommensurate α₂-Cr₂P₂O₇ at 72°C, followed by commensurate α₁-Cr₂P₂O₇ at 12°C. For Zn₂P₂O₇, a similar temperature-dependent phase transition sequence is observed. High-temperature β-Zn₂P₂O₇ again adopts the thortveitite aristotype structure and transforms in a sluggish phase transition of second order over a wide temperature range into incommensurate α₂-Zn₂P₂O₇, followed by commensurate α₁-Zn₂P₂O₇ (isotypic with α₁-Cr₂P₂O₇) at 135°C. In comparison with the three pyrophosphate phases, Zn₂As₂O₇ is unique as it does not exhibit a high-temperature β-modification with a commensurate structure. Instead, β-Zn₂As₂O₇ has an incommensurately modulated structure and transforms below −6°C into commensurate α-Zn₂As₂O₇ (isotypic with α₁-Cr₂P₂O₇).

Irrespective of the type of incommensurability or whether the incommensurate phase is intermediate between commensurate high- and low-temperature phases or occurs as a high-temperature form only, a common feature is clearly visible for all four examples. Based on the high-temperature β phases with a coordination number of 6 for the crystallographically unique M²⁺ ion, superstructures are realized for the commensurate low-temperature modifications where the crystallographic sites of M²⁺ ions are multiplied and now exhibit coordination numbers of 6 or 5. In the (intermediate) incommensurate structures, the coordination of the M²⁺ ions varies from 5 to 6, and hence the formation of [MO₅], [MO₆] or [MO₅₊₁] polyhedra is observed. A change of the coordination number seems not to be the sole reason for the incommensurability in the structures. In fact, the competition between the dynamics of the X₂O₇ groups due to avoiding of a linear O—X—O bridging angle and the coordination of the M²⁺ ions are decisive for this behaviour. The actual reasons for the change of the coordination number from 6 to 5 still need to be elucidated for the different M²⁺ cations. For M²⁺ = Cr and Fe with their d orbitals not completely filled, electronic effects seem to be the driving forces. Interestingly, Fe₂P₂O₇, which is associated with a weak Jahn–Teller effect of second order for the Fe²⁺ ion (electronic configuration high-spin d⁶), has a much lower ordering temperature (≃ −140°C) into the commensurate low-temperature structure than Cr₂P₂O₇, which is associated with a stronger Jahn–Teller effect of first order for Cr²⁺ (electronic configuration d⁴). Also, the local coordination environment and distortions of the corresponding [MO₅] and [MO₆] polyhedra in the two commensurate low-temperature structures are different, which leads to a different linkage pattern of 5–6–6–5 5–6–6–5 in α₁-Fe₂P₂O₇ versus 5–6–5 5–6–5 in α₁-Cr₂P₂O₇. On the other hand, for M = Zn with its completely filled d orbitals such electronic effects are missing, and geometric factors or packing features might primarily be responsible for the incommensurability.

The unit-cell volume (Table 1) of Fe₂P₂O₇ decreases with temperature from α₃ (123.15 ų per formula unit at −127°C) to α₂ (122.25 ų at 27°C) and α₁ (121.96 ų at 127°C) before it increases for β (122.26 ų at 107°C) and constantly with further temperature (Fig. S2). The associated negative thermal expansion (NTE) observed for the α3, α2 and α1 forms has been studied previously and is attributed to a cooperative Jahn–Teller effect of the Fe²⁺ ions (Liang et al., 2024). Some other pyrophosphate compounds A₂P₂O₇ with A cations smaller than 0.97 Å also exhibit NTE in certain temperature ranges (Zeng et al., 2023), including Cu₂P₂O₇ and Zn₂P₂O₇. In the copper and zinc compounds, however, various phonon modes are described as being responsible for the NTE property (Mochizuki et al., 2024; Wang et al., 2024).

5. Conclusions

The crystal structures of four different phases of Fe₂P₂O₇ could be unambiguously determined on the basis of new single-crystal X-ray intensity measurements, which also led to a correction of the existence ranges and space-group symmetries of these phases as reported by Liang et al. (2024). It was shown that Fe₂P₂O₇ is already incommensurately modulated under room-temperature conditions [full polymorphic notation (Tolédano et al., 1998): α₂|70 to −140°C|C1(αβγ)0|Z = 2|–|Type = hettotype of thortveitite, incommensurately modulated], and both crystal structure models for this triclinic phase reported more than 40 years ago (Stefanidis & Nord, 1982; Hoggins et al., 1983) are incorrect. Reversible phase transitions to periodic crystal structures, β|>190°C|C2/m (15)|Z = 2|–|Type = thortveitite, α₃|190 − 70°C|C1 (2)|Z = 2|–|Type = hettotype of thortveitite; basic structure of α₂, and α₁|<−140°C|X1 (2)|Z = 8|–|Type = hettotype of thortveitite; twofold superstructure of α₃, reveal a complex thermal behaviour. Fe₂P₂O₇ is the fourth representative of the thortveitite-type M₂X₂O₇ family for an incommensurately modulated crystal structure, existing either as an intermediate phase between high- and low-temperature forms (M = Cr, X = P; M = Zn, X = P) or as a high-temperature form (M = Zn, X = As).

We suggest that the incommensurability of the observed α₂-Fe₂P₂O₇ structure is a result of the competition between the structural distortions caused by the dynamical behaviour of the X₂O₇ group (avoidance of a linear bridging angle) and the coordination of the metal cations associated with a change in the coordination number from 6 to 5 caused by Jahn–Teller effects.

Supporting information

Computing details

(alpha1_standardsetting)

Crystal data

Fe2O7P2	Z = 2
Mr = 285.6	F(000) = 276
Triclinic, P1	Dx = 3.852 Mg m^−3
Hall symbol: -P 1	Mo Kα radiation, λ = 0.71073 Å
a = 5.58850 (16) Å	Cell parameters from 25284 reflections
b = 6.3545 (3) Å	θ = 2.8–41.2°
c = 7.3692 (16) Å	µ = 6.52 mm^−1
α = 81.40 (1)°	T = 100 K
β = 86.007 (9)°	Plate, light brown
γ = 72.204 (3)°	0.06 × 0.04 × 0.01 mm
V = 246.29 (6) Å^3

Data collection

STOE STADIVARI diffractometer	3198 independent reflections
Radiation source: Axo_Mo	2830 reflections with I > 2σ(I)
Graded multilayer mirror monochromator	Rint = 0.016
Detector resolution: 13.33 pixels mm^-1	θmax = 41.1°, θmin = 2.8°
rotation method, ω scans	h = −9→10
Absorption correction: multi-scan STOE LANA, absorption correction by scaling of reflection intensities. J. Koziskova, F. Hahn, J. Richter, J. Kozisek, "Comparison of different absorption corrections on the model structure of tetrakis(µ2-acetato)- diaqua-di-copper(II)", Acta Chimica Slovaca, vol. 9, no. 2, 2016, pp. 136 - 140. Afterwards a spherical absorption correction was performed within STOE LANA.	k = −9→11
Tmin = 0.454, Tmax = 0.542	l = −8→13
15018 measured reflections

Refinement

Refinement on F^2	0 restraints
R[F^2 > 2σ(F^2)] = 0.017	0 constraints
wR(F^2) = 0.041	Weighting scheme based on measured s.u.'s w = 1/[σ^2(Fo^2) + (0.02P)^2] where P = (Fo^2 + 2Fc^2)/3
S = 1.28	(Δ/σ)max = 0.002
3198 reflections	Extinction correction: B-C type 1 Gaussian isotropic (Becker & Coppens, 1974)
101 parameters	Extinction coefficient: 130 (30)

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

	x	y	z	Uiso*/Ueq
Fe1a	0.23189 (2)	0.84797 (2)	0.306934 (17)	0.00402 (3)
Fe1b	−0.17832 (2)	0.65341 (2)	0.158440 (17)	0.00363 (3)
P1a	0.30718 (4)	0.31844 (4)	0.40545 (3)	0.00323 (5)
P1b	0.72307 (4)	0.19797 (4)	0.10898 (3)	0.00310 (5)
O1a	0.14313 (12)	0.55312 (10)	0.33618 (9)	0.00479 (15)
O1b	0.87765 (12)	−0.04010 (10)	0.17448 (9)	0.00465 (15)
O2a	0.40161 (12)	0.30369 (11)	0.59740 (8)	0.00484 (15)
O2b	0.55895 (12)	0.21262 (10)	−0.05143 (9)	0.00520 (15)
O3a	0.18749 (13)	0.14056 (11)	0.38785 (9)	0.00687 (16)
O3b	0.88492 (12)	0.35431 (10)	0.06697 (9)	0.00496 (15)
O4	0.55648 (13)	0.28457 (11)	0.28032 (9)	0.00705 (16)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Fe1a	0.00420 (5)	0.00321 (5)	0.00499 (5)	−0.00137 (4)	−0.00098 (4)	−0.00062 (3)
Fe1b	0.00371 (5)	0.00343 (5)	0.00411 (5)	−0.00135 (4)	−0.00003 (3)	−0.00100 (3)
P1a	0.00351 (8)	0.00311 (8)	0.00302 (8)	−0.00092 (6)	−0.00028 (6)	−0.00027 (6)
P1b	0.00315 (8)	0.00287 (8)	0.00326 (8)	−0.00102 (6)	0.00005 (6)	−0.00016 (6)
O1a	0.0047 (2)	0.0034 (2)	0.0059 (2)	−0.00088 (18)	−0.00104 (18)	0.00041 (18)
O1b	0.0047 (2)	0.0031 (2)	0.0058 (2)	−0.00081 (18)	−0.00105 (18)	0.00028 (18)
O2a	0.0046 (2)	0.0065 (2)	0.0035 (2)	−0.00166 (18)	−0.00069 (18)	−0.00062 (18)
O2b	0.0047 (2)	0.0063 (2)	0.0047 (2)	−0.00178 (19)	−0.00115 (18)	−0.00053 (18)
O3a	0.0082 (3)	0.0046 (2)	0.0088 (3)	−0.0029 (2)	−0.0020 (2)	−0.0010 (2)
O3b	0.0056 (2)	0.0045 (2)	0.0058 (2)	−0.00307 (18)	0.00054 (18)	−0.00078 (18)
O4	0.0058 (2)	0.0090 (3)	0.0059 (2)	−0.0015 (2)	0.00244 (19)	−0.0024 (2)

Geometric parameters (Å, º)

Fe1a—O1a	2.0580 (7)	Fe1b—O3b^iii	2.2463 (9)
Fe1a—O1b^i	2.1397 (7)	P1a—O1a	1.5254 (6)
Fe1a—O2a^ii	2.1005 (7)	P1a—O2a	1.5238 (9)
Fe1a—O2b^iii	2.1668 (10)	P1a—O3a	1.5033 (8)
Fe1a—O3a^iv	1.9756 (8)	P1a—O4	1.5921 (8)
Fe1b—O1a	2.1742 (8)	P1b—O1b	1.5194 (6)
Fe1b—O1b^i	2.0860 (7)	P1b—O2b	1.5209 (8)
Fe1b—O2a^v	2.1201 (9)	P1b—O3b	1.5229 (8)
Fe1b—O2b^vi	2.1903 (7)	P1b—O4	1.5795 (8)
Fe1b—O3b^vii	2.0325 (7)
O1a—Fe1a—O1b^i	79.99 (3)	O1a—P1a—O4	104.60 (4)
O1a—Fe1a—O2a^ii	94.08 (3)	O2a—P1a—O3a	113.17 (4)
O1a—Fe1a—O2b^iii	93.57 (3)	O2a—P1a—O4	103.17 (4)
O1a—Fe1a—O3a^iv	149.50 (3)	O3a—P1a—O4	109.93 (4)
O1b^i—Fe1a—O2a^ii	169.60 (3)	O1b—P1b—O2b	112.46 (4)
O1b^i—Fe1a—O2b^iii	92.81 (3)	O1b—P1b—O3b	112.18 (4)
O1b^i—Fe1a—O3a^iv	92.99 (3)	O1b—P1b—O4	105.78 (4)
O2a^ii—Fe1a—O2b^iii	78.95 (3)	O2b—P1b—O3b	111.76 (4)
O2a^ii—Fe1a—O3a^iv	96.36 (3)	O2b—P1b—O4	110.77 (4)
O2b^iii—Fe1a—O3a^iv	116.53 (3)	O3b—P1b—O4	103.34 (4)
O1a—Fe1b—O1b^i	78.59 (3)	Fe1a—O1a—Fe1b	99.68 (2)
O1a—Fe1b—O2a^v	85.93 (3)	Fe1a—O1a—P1a	128.53 (4)
O1a—Fe1b—O2b^vi	163.49 (3)	Fe1b—O1a—P1a	128.74 (5)
O1a—Fe1b—O3b^vii	96.62 (3)	Fe1a^viii—O1b—Fe1b^viii	99.92 (2)
O1a—Fe1b—O3b^iii	83.59 (3)	Fe1a^viii—O1b—P1b	128.13 (5)
O1b^i—Fe1b—O2a^v	87.48 (3)	Fe1b^viii—O1b—P1b	131.93 (4)
O1b^i—Fe1b—O2b^vi	96.72 (3)	Fe1a^ii—O2a—Fe1b^v	103.60 (3)
O1b^i—Fe1b—O3b^vii	156.35 (3)	Fe1a^ii—O2a—P1a	127.94 (4)
O1b^i—Fe1b—O3b^iii	81.59 (3)	Fe1b^v—O2a—P1a	126.38 (4)
O2a^v—Fe1b—O2b^vi	78.01 (3)	Fe1a^iii—O2b—Fe1b^vi	99.15 (3)
O2a^v—Fe1b—O3b^vii	115.48 (3)	Fe1a^iii—O2b—P1b	113.70 (4)
O2a^v—Fe1b—O3b^iii	166.15 (3)	Fe1b^vi—O2b—P1b	144.96 (4)
O2b^vi—Fe1b—O3b^vii	93.76 (3)	Fe1a^ix—O3a—P1a	146.26 (4)
O2b^vi—Fe1b—O3b^iii	111.59 (3)	Fe1b^x—O3b—Fe1b^iii	105.16 (3)
O3b^vii—Fe1b—O3b^iii	74.84 (3)	Fe1b^x—O3b—P1b	126.48 (4)
O1a—P1a—O2a	111.85 (4)	Fe1b^iii—O3b—P1b	126.78 (4)
O1a—P1a—O3a	113.24 (4)	P1a—O4—P1b	151.91 (6)

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

(alpha1)

Crystal data

Fe2O7P2	Z = 8
Mr = 285.6	F(000) = 1104
Triclinic, X1	Dx = 3.852 Mg m^−3
Hall symbol: -X 1	Mo Kα radiation, λ = 0.71073 Å
a = 13.3690 (5) Å	Cell parameters from 25284 reflections
b = 16.9853 (4) Å	θ = 2.8–41.2°
c = 6.3545 (3) Å	µ = 6.52 mm^−1
α = 50.200 (3)°	T = 100 K
β = 72.409 (4)°	Plate, light brown
γ = 94.100 (4)°	0.06 × 0.04 × 0.01 mm
V = 985.17 (11) Å^3

Data collection

STOE STADIVARI diffractometer	3198 independent reflections
Radiation source: Axo_Mo	2830 reflections with I > 2σ(I)
Graded multilayer mirror monochromator	Rint = 0.016
Detector resolution: 13.33 pixels mm^-1	θmax = 41.1°, θmin = 2.8°
rotation method, ω scans	h = −23→24
Absorption correction: multi-scan STOE LANA, absorption correction by scaling of reflection intensities. J. Koziskova, F. Hahn, J. Richter, J. Kozisek, "Comparison of different absorption corrections on the model structure of tetrakis(µ2-acetato)- diaqua-di-copper(II)", Acta Chimica Slovaca, vol. 9, no. 2, 2016, pp. 136 - 140. Afterwards a spherical absorption correction was performed within STOE LANA.	k = −18→31
Tmin = 0.454, Tmax = 0.542	l = −9→11
15018 measured reflections

Refinement

Refinement on F^2	0 constraints
R[F^2 > 2σ(F^2)] = 0.017	Weighting scheme based on measured s.u.'s w = 1/[σ^2(Fo^2) + (0.02P)^2] where P = (Fo^2 + 2Fc^2)/3
wR(F^2) = 0.041	(Δ/σ)max = 0.044
S = 1.28	Δρmax = 0.70 e Å^−3
3198 reflections	Δρmin = −1.02 e Å^−3
101 parameters	Extinction correction: B-C type 1 Gaussian isotropic (Becker & Coppens, 1974)
0 restraints	Extinction coefficient: 130 (30)

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

	x	y	z	Uiso*/Ueq
Fe1a	0.095495 (11)	0.211439 (10)	0.54104 (3)	0.00403 (6)
Fe1b	0.123799 (10)	0.034640 (10)	0.49497 (3)	0.00363 (6)
P1a	0.125929 (19)	0.279521 (18)	−0.08701 (5)	0.00323 (10)
P1b	−0.126281 (19)	0.235257 (18)	0.08900 (5)	0.00310 (10)
O1a	0.13231 (6)	0.20387 (5)	0.21694 (14)	0.0048 (3)
O1b	−0.13217 (6)	0.30665 (5)	−0.21457 (14)	0.0046 (3)
O2a	0.19830 (5)	0.39910 (5)	−0.29371 (15)	0.0048 (3)
O2b	−0.16545 (6)	0.11402 (5)	0.26405 (15)	0.0052 (3)
O3a	0.14705 (6)	0.24080 (5)	−0.24729 (15)	0.0069 (3)
O3b	−0.18775 (5)	0.25471 (5)	0.28734 (15)	0.0050 (3)
O4	0.00104 (5)	0.27928 (5)	0.00426 (15)	0.0070 (3)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Fe1a	0.00552 (5)	0.00392 (5)	0.00401 (5)	0.00206 (4)	−0.00257 (4)	−0.00310 (4)
Fe1b	0.00391 (5)	0.00405 (5)	0.00414 (5)	0.00161 (4)	−0.00206 (4)	−0.00320 (4)
P1a	0.00335 (8)	0.00298 (8)	0.00295 (8)	0.00093 (6)	−0.00135 (6)	−0.00196 (7)
P1b	0.00310 (8)	0.00336 (8)	0.00281 (8)	0.00121 (6)	−0.00125 (6)	−0.00217 (7)
O1a	0.0063 (2)	0.0046 (2)	0.0035 (2)	0.00252 (18)	−0.00261 (19)	−0.0027 (2)
O1b	0.0062 (2)	0.0046 (2)	0.0035 (2)	0.00244 (18)	−0.00266 (19)	−0.0027 (2)
O2a	0.0044 (2)	0.0033 (2)	0.0052 (2)	0.00089 (18)	−0.00161 (19)	−0.0026 (2)
O2b	0.0057 (2)	0.0037 (2)	0.0055 (2)	0.00172 (18)	−0.0023 (2)	−0.0030 (2)
O3a	0.0102 (3)	0.0069 (2)	0.0064 (3)	0.0034 (2)	−0.0044 (2)	−0.0055 (2)
O3b	0.0051 (2)	0.0064 (2)	0.0051 (2)	0.00215 (18)	−0.00185 (19)	−0.0048 (2)
O4	0.0035 (2)	0.0084 (3)	0.0090 (3)	0.00219 (19)	−0.0029 (2)	−0.0059 (2)

Geometric parameters (Å, º)

Fe1a—O1a	2.0580 (11)	Fe1b—O3b^iii	2.2463 (7)
Fe1a—O1b^i	2.1397 (8)	P1a—O1a	1.5254 (9)
Fe1a—O2a^ii	2.1005 (8)	P1a—O2a	1.5238 (8)
Fe1a—O2b^iii	2.1668 (8)	P1a—O3a	1.5033 (12)
Fe1a—O3a^iv	1.9756 (12)	P1a—O4	1.5921 (8)
Fe1b—O1a	2.1742 (8)	P1b—O1b	1.5194 (9)
Fe1b—O1b^i	2.0860 (12)	P1b—O2b	1.5209 (9)
Fe1b—O2a^v	2.1201 (7)	P1b—O3b	1.5228 (11)
Fe1b—O2b^vi	2.1903 (8)	P1b—O4	1.5795 (8)
Fe1b—O3b^vii	2.0325 (11)
O1a—Fe1a—O1b^i	79.99 (3)	O1a—P1a—O4	104.60 (4)
O1a—Fe1a—O2a^ii	94.08 (4)	O2a—P1a—O3a	113.17 (4)
O1a—Fe1a—O2b^iii	93.57 (4)	O2a—P1a—O4	103.17 (4)
O1a—Fe1a—O3a^iv	149.50 (3)	O3a—P1a—O4	109.93 (6)
O1b^i—Fe1a—O2a^ii	169.60 (4)	O1b—P1b—O2b	112.46 (6)
O1b^i—Fe1a—O2b^iii	92.81 (3)	O1b—P1b—O3b	112.18 (5)
O1b^i—Fe1a—O3a^iv	92.99 (4)	O1b—P1b—O4	105.78 (4)
O2a^ii—Fe1a—O2b^iii	78.95 (3)	O2b—P1b—O3b	111.76 (4)
O2a^ii—Fe1a—O3a^iv	96.36 (4)	O2b—P1b—O4	110.77 (5)
O2b^iii—Fe1a—O3a^iv	116.53 (4)	O3b—P1b—O4	103.34 (6)
O1a—Fe1b—O1b^i	78.59 (3)	Fe1a—O1a—Fe1b	99.68 (3)
O1a—Fe1b—O2a^v	85.93 (3)	Fe1a—O1a—P1a	128.53 (6)
O1a—Fe1b—O2b^vi	163.49 (3)	Fe1b—O1a—P1a	128.74 (7)
O1a—Fe1b—O3b^vii	96.62 (4)	Fe1a^viii—O1b—Fe1b^viii	99.92 (3)
O1a—Fe1b—O3b^iii	83.59 (3)	Fe1a^viii—O1b—P1b	128.13 (7)
O1b^i—Fe1b—O2a^v	87.48 (4)	Fe1b^viii—O1b—P1b	131.93 (6)
O1b^i—Fe1b—O2b^vi	96.72 (4)	Fe1a^ii—O2a—Fe1b^v	103.60 (2)
O1b^i—Fe1b—O3b^vii	156.35 (3)	Fe1a^ii—O2a—P1a	127.94 (5)
O1b^i—Fe1b—O3b^iii	81.59 (4)	Fe1b^v—O2a—P1a	126.38 (5)
O2a^v—Fe1b—O2b^vi	78.01 (3)	Fe1a^iii—O2b—Fe1b^vi	99.15 (3)
O2a^v—Fe1b—O3b^vii	115.48 (4)	Fe1a^iii—O2b—P1b	113.70 (5)
O2a^v—Fe1b—O3b^iii	166.15 (4)	Fe1b^vi—O2b—P1b	144.96 (5)
O2b^vi—Fe1b—O3b^vii	93.76 (4)	Fe1a^ix—O3a—P1a	146.26 (5)
O2b^vi—Fe1b—O3b^iii	111.59 (2)	Fe1b^x—O3b—Fe1b^iii	105.16 (4)
O3b^vii—Fe1b—O3b^iii	74.84 (4)	Fe1b^x—O3b—P1b	126.48 (5)
O1a—P1a—O2a	111.85 (5)	Fe1b^iii—O3b—P1b	126.78 (6)
O1a—P1a—O3a	113.24 (5)	P1a—O4—P1b	151.91 (8)

Symmetry codes: (i) x+1/4, y−1/4, z+1; (ii) −x+1/4, −y+3/4, −z; (iii) −x−1/4, −y+1/4, −z+1; (iv) x, y, z+1; (v) −x+1/2, −y+1/2, −z; (vi) −x, −y, −z+1; (vii) x+1/4, y−1/4, z; (viii) x−1/4, y+1/4, z−1; (ix) x, y, z−1; (x) x−1/4, y+1/4, z.

(alpha2)

Crystal data

Fe2O7P2	Z = 2
Mr = 285.6	F(000) = 276
Triclinic, C1(αβγ)0†	Dx = 3.880 Mg m^−3
q = 0.4489(3)a* + 0.2517(3)b* + 0.3646(3)c*	Mo Kα radiation, λ = 0.71073 Å
a = 6.6393 (6) Å	Cell parameters from 10653 reflections
b = 8.4748 (6) Å	θ = 3.9–37.0°
c = 4.4839 (3) Å	µ = 6.57 mm^−1
α = 90.036 (5)°	T = 300 K
β = 103.962 (7)°	Plate, light brown
γ = 92.929 (6)°	0.06 × 0.04 × 0.01 mm
V = 244.50 (3) Å^3

† Symmetry operations: (1) x₁, x₂, x₃, x₄; (2) −x₁, −x₂, −x₃, −x₄; (3) x₁+1/2, x₂+1/2, x₃, x₄; (4) −x₁+1/2, −x₂+1/2, −x₃, −x₄.

Data collection

STOE STADIVARI diffractometer	5785 independent reflections
Radiation source: Axo_Mo	3402 reflections with I > 3σ(I)
Graded multilayer mirror monochromator	Rint = 0.036
Detector resolution: 13.33 pixels mm^-1	θmax = 37.0°, θmin = 2.8°
rotation method, ω scans	h = −12→11
Absorption correction: multi-scan STOE LANA, absorption correction by scaling of reflection intensities. J. Koziskova, F. Hahn, J. Richter, J. Kozisek, "Comparison of different absorption corrections on the model structure of tetrakis(µ2-acetato)- diaqua-di-copper(II)", Acta Chimica Slovaca, vol. 9, no. 2, 2016, pp. 136 - 140. Afterwards a spherical absorption correction was performed within STOE LANA.	k = −14→13
Tmin = 0.360, Tmax = 0.554	l = −6→8
21588 measured reflections

Refinement

Refinement on F^2	4 constraints
R[F^2 > 2σ(F^2)] = 0.027	Weighting scheme based on measured s.u.'s w = 1/[σ^2(Fo^2) + (0.04P)^2] where P = (Fo^2 + 2Fc^2)/3
wR(F^2) = 0.081	(Δ/σ)max = 0.038
S = 1.10	Δρmax = 1.13 e Å^−3
5785 reflections	Δρmin = −1.15 e Å^−3
251 parameters	Extinction correction: B-C type 1 Gaussian isotropic (Becker & Coppens, 1974)
0 restraints	Extinction coefficient: 690 (50)

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

	x	y	z	Uiso*/Ueq
Fe1	0.489014 (18)	0.189884 (13)	0.51774 (3)	0.01065 (4)
P1	0.21133 (3)	0.00069 (3)	−0.08996 (5)	0.00791 (5)
O1	0.37351 (9)	0.00434 (8)	0.21420 (14)	0.00976 (14)
O2	0.22535 (10)	0.14714 (7)	−0.27860 (15)	0.01152 (15)
O3	0.20953 (11)	−0.14842 (8)	−0.27222 (16)	0.01532 (17)
O4	0	0	0	0.0260 (3)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Fe1	0.01073 (6)	0.00775 (6)	0.01171 (7)	−0.00107 (4)	−0.00038 (4)	−0.00073 (4)
P1	0.00647 (9)	0.00886 (9)	0.00787 (9)	−0.00084 (6)	0.00095 (7)	0.00025 (6)
O1	0.0097 (2)	0.0107 (2)	0.0078 (2)	0.00078 (18)	0.00006 (18)	−0.00088 (18)
O2	0.0129 (2)	0.0095 (2)	0.0125 (3)	0.00168 (19)	0.0035 (2)	0.0020 (2)
O3	0.0201 (3)	0.0104 (3)	0.0127 (3)	−0.0036 (2)	−0.0006 (2)	−0.0029 (2)
O4	0.0076 (4)	0.0447 (7)	0.0279 (5)	0.0007 (4)	0.0086 (4)	0.0069 (5)

Bond lengths (Å)

	Average	Minimum	Maximum
Fe1—O1	2.0724 (19)	2.039 (2)	2.112 (2)
Fe1—O1^i	2.151 (2)	2.119 (2)	2.174 (2)
Fe1—O2^ii	2.178 (2)	2.130 (3)	2.226 (3)
Fe1—O2^iii	2.134 (2)	2.104 (2)	2.168 (2)
Fe1—O3^iv	2.013 (2)	1.930 (2)	2.066 (2)
Fe1—O3^v	2.550 (3)	2.245 (3)	2.962 (3)
P1—O1	1.5201 (19)	1.504 (2)	1.526 (2)
P1—O2	1.520 (2)	1.481 (2)	1.525 (2)
P1—O3	1.512 (2)	1.493 (2)	1.529 (2)
P1—O4	1.5760 (19)	1.494 (3)	1.592 (3)

Symmetry codes: (i) −x₁+1, −x₂, −x₃+1, −x₄; (ii) x₁, x₂, x₃+1, x₄; (iii) −x₁+1/2, −x₂+1/2, −x₃, −x₄; (iv) x₁+1/2, x₂+1/2, x₃+1, x₄; (v) −x₁+1, −x₂, −x₃, −x₄.

(alpha3_single)

Crystal data

Fe2O7P2	Z = 2
Mr = 285.6	F(000) = 276
Triclinic, C1	Dx = 3.889 Mg m^−3
Hall symbol: -C 1	Mo Kα radiation, λ = 0.71073 Å
a = 6.6161 (5) Å	Cell parameters from 11950 reflections
b = 8.4673 (4) Å	θ = 3.9–41.0°
c = 4.4825 (3) Å	µ = 6.58 mm^−1
α = 90.135 (5)°	T = 400 K
β = 103.645 (7)°	Plate, light brown
γ = 91.616 (7)°	0.06 × 0.04 × 0.01 mm
V = 243.92 (3) Å^3

Data collection

STOE STADIVARI diffractometer	1563 independent reflections
Radiation source: Axo_Mo	1341 reflections with I > 3σ(I)
Graded multilayer mirror monochromator	Rint = 0.016
Detector resolution: 13.33 pixels mm^-1	θmax = 41.0°, θmin = 3.9°
rotation method, ω scans	h = −11→12
Absorption correction: multi-scan STOE LANA, absorption correction by scaling of reflection intensities. J. Koziskova, F. Hahn, J. Richter, J. Kozisek, "Comparison of different absorption corrections on the model structure of tetrakis(µ2-acetato)- diaqua-di-copper(II)", Acta Chimica Slovaca, vol. 9, no. 2, 2016, pp. 136 - 140. Afterwards a spherical absorption correction was performed within STOE LANA.	k = −9→15
Tmin = 0.568, Tmax = 0.689	l = −8→8
7257 measured reflections

Refinement

Refinement on F^2	0 constraints
R[F^2 > 2σ(F^2)] = 0.022	Weighting scheme based on measured s.u.'s w = 1/(σ^2(I) + 0.0016I^2)
wR(F^2) = 0.065	(Δ/σ)max = 0.009
S = 1.20	Δρmax = 0.34 e Å^−3
1563 reflections	Δρmin = −0.18 e Å^−3
53 parameters	Extinction correction: B-C type 1 Gaussian isotropic (Becker & Coppens, 1974)
0 restraints	Extinction coefficient: 540 (60)

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

	x	y	z	Uiso*/Ueq
Fe1	0.49493 (3)	0.190859 (18)	0.50929 (4)	0.01288 (5)
P1	0.21345 (4)	0.00054 (3)	−0.09088 (6)	0.00889 (7)
O1	0.37354 (13)	0.00275 (10)	0.21365 (18)	0.01153 (18)
O2	0.22526 (16)	0.14811 (10)	−0.2778 (2)	0.0151 (2)
O3	0.21855 (18)	−0.14925 (11)	−0.2769 (2)	0.0193 (2)
O4	0	0	0	0.0513 (9)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Fe1	0.01281 (8)	0.00944 (8)	0.01400 (9)	−0.00045 (5)	−0.00155 (5)	−0.00023 (5)
P1	0.00697 (11)	0.01237 (12)	0.00713 (11)	−0.00035 (8)	0.00129 (8)	0.00088 (8)
O1	0.0116 (3)	0.0134 (3)	0.0082 (3)	0.0010 (2)	−0.0004 (2)	−0.0004 (2)
O2	0.0196 (4)	0.0125 (3)	0.0131 (3)	0.0039 (3)	0.0034 (3)	0.0035 (3)
O3	0.0276 (5)	0.0125 (3)	0.0144 (4)	−0.0052 (3)	−0.0011 (3)	−0.0028 (3)
O4	0.0086 (6)	0.114 (2)	0.0357 (10)	0.0056 (9)	0.0132 (7)	0.0186 (12)

Geometric parameters (Å, º)

Fe1—O1	2.0866 (8)	Fe1—O3^v	2.4021 (13)
Fe1—O1^i	2.1371 (9)	P1—O1	1.5173 (8)
Fe1—O2^ii	2.2347 (12)	P1—O2	1.5164 (10)
Fe1—O2^iii	2.1128 (10)	P1—O3	1.5218 (10)
Fe1—O3^iv	2.0377 (10)	P1—O4	1.5593 (4)
O1—Fe1—O1^i	79.99 (3)	O1—P1—O3	112.59 (5)
O1—Fe1—O2^ii	86.54 (4)	O1—P1—O4	104.34 (4)
O1—Fe1—O2^iii	94.07 (3)	O2—P1—O3	111.88 (6)
O1—Fe1—O3^iv	155.39 (4)	O2—P1—O4	107.10 (4)
O1—Fe1—O3^v	79.81 (3)	O3—P1—O4	107.46 (5)
O1^i—Fe1—O2^ii	84.18 (3)	Fe1—O1—Fe1^i	100.01 (3)
O1^i—Fe1—O2^iii	160.67 (4)	Fe1—O1—P1	130.27 (5)
O1^i—Fe1—O3^iv	94.44 (3)	Fe1^i—O1—P1	129.20 (5)
O1^i—Fe1—O3^v	82.33 (3)	Fe1^vi—O2—Fe1^iii	102.92 (4)
O2^ii—Fe1—O2^iii	77.08 (4)	Fe1^vi—O2—P1	120.14 (6)
O2^ii—Fe1—O3^iv	116.95 (4)	Fe1^iii—O2—P1	134.47 (7)
O2^ii—Fe1—O3^v	162.25 (3)	Fe1^vii—O3—Fe1^v	104.32 (4)
O2^iii—Fe1—O3^iv	98.19 (4)	Fe1^vii—O3—P1	133.27 (7)
O2^iii—Fe1—O3^v	114.94 (4)	Fe1^v—O3—P1	119.95 (6)
O3^iv—Fe1—O3^v	75.68 (4)	P1—O4—P1^viii	180
O1—P1—O2	112.88 (5)

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

(alpha3_split)

Crystal data

Fe2O7P2	Z = 2
Mr = 285.6	F(000) = 276
Triclinic, C1	Dx = 3.889 Mg m^−3
Hall symbol: -C 1	Mo Kα radiation, λ = 0.71073 Å
a = 6.6161 (5) Å	Cell parameters from 11950 reflections
b = 8.4673 (4) Å	θ = 3.9–41.0°
c = 4.4825 (3) Å	µ = 6.58 mm^−1
α = 90.135 (5)°	T = 400 K
β = 103.645 (7)°	Plate, light brown
γ = 91.616 (7)°	0.06 × 0.04 × 0.01 mm
V = 243.92 (3) Å^3

Data collection

STOE STADIVARI diffractometer	1563 independent reflections
Radiation source: X-ray tube	1341 reflections with I > 3σ(I)
Graded multilayer mirror monochromator	Rint = 0.016
Detector resolution: 13.33 pixels mm^-1	θmax = 41.0°, θmin = 3.9°
rotation method, ω scans	h = −11→12
Absorption correction: multi-scan STOE LANA, absorption correction by scaling of reflection intensities. J. Koziskova, F. Hahn, J. Richter, J. Kozisek, "Comparison of different absorption corrections on the model structure of tetrakis(µ2-acetato)- diaqua-di-copper(II)", Acta Chimica Slovaca, vol. 9, no. 2, 2016, pp. 136 - 140. Afterwards a spherical absorption correction was performed within STOE LANA.	k = −9→15
Tmin = 0.568, Tmax = 0.689	l = −8→8
7257 measured reflections

Refinement

Refinement on F^2	0 constraints
R[F^2 > 2σ(F^2)] = 0.022	Weighting scheme based on measured s.u.'s w = 1/(σ^2(I) + 0.0016I^2)
wR(F^2) = 0.063	(Δ/σ)max = 0.014
S = 1.17	Δρmax = 0.37 e Å^−3
1563 reflections	Δρmin = −0.15 e Å^−3
56 parameters	Extinction correction: B-C type 1 Gaussian isotropic (Becker & Coppens, 1974)
0 restraints	Extinction coefficient: 540 (60)

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

	x	y	z	Uiso*/Ueq	Occ. (<1)
Fe1	0.49492 (3)	0.190858 (17)	0.50928 (4)	0.01288 (5)
P1	0.21344 (4)	0.00054 (3)	−0.09088 (6)	0.00888 (6)
O1	0.37352 (13)	0.00272 (10)	0.21365 (18)	0.01151 (18)
O2	0.22526 (16)	0.14812 (10)	−0.2778 (2)	0.0150 (2)
O3	0.21854 (18)	−0.14926 (10)	−0.2769 (2)	0.0193 (2)
O4	−0.0016 (10)	−0.0263 (5)	−0.0139 (18)	0.0287 (12)	0.5

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Fe1	0.01280 (8)	0.00943 (8)	0.01399 (8)	−0.00045 (5)	−0.00156 (5)	−0.00023 (5)
P1	0.00697 (10)	0.01234 (11)	0.00712 (10)	−0.00035 (8)	0.00128 (8)	0.00087 (8)
O1	0.0115 (3)	0.0135 (3)	0.0082 (3)	0.0010 (2)	−0.0004 (2)	−0.0004 (2)
O2	0.0195 (4)	0.0124 (3)	0.0131 (3)	0.0039 (3)	0.0034 (3)	0.0035 (2)
O3	0.0275 (5)	0.0125 (3)	0.0144 (4)	−0.0052 (3)	−0.0012 (3)	−0.0027 (3)
O4	0.0090 (6)	0.050 (3)	0.0317 (16)	0.002 (2)	0.0127 (9)	0.001 (2)

Geometric parameters (Å, º)

Fe1—O1	2.0868 (8)	P1—O1	1.5173 (8)
Fe1—O1^i	2.1370 (8)	P1—O2	1.5164 (9)
Fe1—O2^ii	2.2347 (11)	P1—O3	1.5219 (10)
Fe1—O2^iii	2.1127 (9)	P1—O4	1.552 (8)
Fe1—O3^iv	2.0376 (10)	P1—O4^vi	1.600 (8)
Fe1—O3^v	2.4022 (13)	O4—O4^vi	0.461 (7)
O1—Fe1—O1^i	79.99 (3)	O2—P1—O3	111.89 (5)
O1—Fe1—O2^ii	86.54 (3)	O2—P1—O4	113.1 (2)
O1—Fe1—O2^iii	94.07 (3)	O2—P1—O4^vi	101.1 (2)
O1—Fe1—O3^iv	155.39 (4)	O3—P1—O4	99.3 (2)
O1—Fe1—O3^v	79.82 (3)	O3—P1—O4^vi	115.38 (19)
O1^i—Fe1—O2^ii	84.19 (3)	O4—P1—O4^vi	16.7 (2)
O1^i—Fe1—O2^iii	160.68 (4)	Fe1—O1—Fe1^i	100.01 (3)
O1^i—Fe1—O3^iv	94.43 (3)	Fe1—O1—P1	130.26 (5)
O1^i—Fe1—O3^v	82.33 (3)	Fe1^i—O1—P1	129.22 (5)
O2^ii—Fe1—O2^iii	77.08 (4)	Fe1^vii—O2—Fe1^iii	102.92 (4)
O2^ii—Fe1—O3^iv	116.95 (4)	Fe1^vii—O2—P1	120.14 (6)
O2^ii—Fe1—O3^v	162.26 (3)	Fe1^iii—O2—P1	134.47 (6)
O2^iii—Fe1—O3^iv	98.19 (4)	Fe1^viii—O3—Fe1^v	104.32 (4)
O2^iii—Fe1—O3^v	114.93 (4)	Fe1^viii—O3—P1	133.28 (7)
O3^iv—Fe1—O3^v	75.68 (4)	Fe1^v—O3—P1	119.94 (6)
O1—P1—O2	112.89 (5)	P1—O4—P1^vi	163.3 (3)
O1—P1—O3	112.58 (5)	P1—O4—O4^vi	87.6 (13)
O1—P1—O4	106.3 (3)	P1^vi—O4—O4^vi	75.7 (13)
O1—P1—O4^vi	102.2 (3)

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

(beta_single)

Crystal data

Fe2O7P2	F(000) = 276
Mr = 285.6	Dx = 3.880 Mg m^−3
Monoclinic, C2/m	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: -C 2y	Cell parameters from 6240 reflections
a = 6.6154 (6) Å	θ = 4.0–40.8°
b = 8.4624 (10) Å	µ = 6.57 mm^−1
c = 4.4920 (4) Å	T = 470 K
β = 103.502 (7)°	Fragment, light-brownish
V = 244.52 (4) Å^3	0.11 × 0.07 × 0.04 × 0.03 (radius) mm
Z = 2

Data collection

STOE STADIVARI diffractometer	826 independent reflections
Radiation source: Axo_Mo	815 reflections with I > 3σ(I)
Graded multilayer mirror monochromator	Rint = 0.012
Detector resolution: 13.33 pixels mm^-1	θmax = 40.8°, θmin = 4.0°
rotation method, ω scans	h = −12→11
Absorption correction: multi-scan STOE LANA, absorption correction by scaling of reflection intensities. J. Koziskova, F. Hahn, J. Richter, J. Kozisek, "Comparison of different absorption corrections on the model structure of tetrakis(µ2-acetato)- diaqua-di-copper(II)", Acta Chimica Slovaca, vol. 9, no. 2, 2016, pp. 136 - 140. Afterwards a spherical absorption correction was performed within STOE LANA.	k = −14→15
Tmin = 0.472, Tmax = 0.558	l = −8→3
3414 measured reflections

Refinement

Refinement on F^2	0 constraints
R[F^2 > 2σ(F^2)] = 0.019	Weighting scheme based on measured s.u.'s w = 1/[σ^2(Fo^2) + (0.04P)^2] where P = (Fo^2 + 2Fc^2)/3
wR(F^2) = 0.083	(Δ/σ)max = 0.010
S = 1.88	Δρmax = 0.62 e Å^−3
826 reflections	Δρmin = −0.69 e Å^−3
32 parameters	Extinction correction: B-C type 1 Gaussian isotropic (Becker & Coppens, 1974)
0 restraints	Extinction coefficient: 330 (100)

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

	x	y	z	Uiso*/Ueq
Fe1	0.5	0.190935 (18)	0.5	0.01503 (8)
P1	0.21350 (4)	0	−0.09113 (6)	0.00973 (9)
O1	0.37349 (14)	0	0.21291 (19)	0.01322 (18)
O2	0.22244 (12)	0.14875 (8)	−0.27750 (16)	0.01845 (17)
O3	0	0	0	0.0555 (10)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Fe1	0.01568 (13)	0.00997 (12)	0.01659 (13)	0	−0.00198 (8)	0
P1	0.00872 (14)	0.01254 (15)	0.00779 (14)	0	0.00167 (10)	0
O1	0.0140 (3)	0.0152 (3)	0.0088 (3)	0	−0.0006 (2)	0
O2	0.0261 (3)	0.0134 (3)	0.0147 (3)	0.0052 (2)	0.0024 (2)	0.00335 (19)
O3	0.0107 (6)	0.119 (3)	0.0407 (11)	0	0.0137 (7)	0

Geometric parameters (Å, º)

Fe1—O1	2.1143 (6)	Fe1—O2^v	2.0780 (8)
Fe1—O1^i	2.1143 (6)	P1—O1	1.5200 (9)
Fe1—O2^ii	2.3141 (9)	P1—O2	1.5205 (8)
Fe1—O2^iii	2.3141 (9)	P1—O2^vi	1.5205 (8)
Fe1—O2^iv	2.0780 (8)	P1—O3	1.5596 (4)
O1—Fe1—O1^i	80.33 (2)	O2^iv—Fe1—O2^v	98.48 (3)
O1—Fe1—O2^ii	84.52 (3)	O1—P1—O2	112.72 (3)
O1—Fe1—O2^iii	81.93 (3)	O1—P1—O2^vi	112.72 (3)
O1—Fe1—O2^iv	94.05 (3)	O1—P1—O3	104.33 (4)
O1—Fe1—O2^v	158.02 (3)	O2—P1—O2^vi	111.76 (4)
O1^i—Fe1—O2^ii	81.93 (3)	O2—P1—O3	107.37 (3)
O1^i—Fe1—O2^iii	84.52 (3)	O2^vi—P1—O3	107.37 (3)
O1^i—Fe1—O2^iv	158.02 (3)	Fe1—O1—Fe1^vii	99.67 (3)
O1^i—Fe1—O2^v	94.05 (3)	Fe1—O1—P1	129.900 (17)
O2^ii—Fe1—O2^iii	162.25 (3)	Fe1^vii—O1—P1	129.900 (17)
O2^ii—Fe1—O2^iv	76.37 (3)	Fe1^viii—O2—Fe1^iv	103.63 (3)
O2^ii—Fe1—O2^v	115.92 (3)	Fe1^viii—O2—P1	120.15 (4)
O2^iii—Fe1—O2^iv	115.92 (3)	Fe1^iv—O2—P1	133.72 (5)
O2^iii—Fe1—O2^v	76.37 (3)	P1—O3—P1^ix	180

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

(beta_split)

Crystal data

Fe2O7P2	F(000) = 276
Mr = 285.6	Dx = 3.880 Mg m^−3
Monoclinic, C2/m	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: -C 2y	Cell parameters from 6240 reflections
a = 6.6154 (6) Å	θ = 4.0–40.8°
b = 8.4624 (10) Å	µ = 6.57 mm^−1
c = 4.4920 (4) Å	T = 470 K
β = 103.502 (7)°	Fragment, light-brownish
V = 244.52 (4) Å^3	0.11 × 0.07 × 0.04 × 0.03 (radius) mm
Z = 2

Data collection

STOE STADIVARI diffractometer	826 independent reflections
Radiation source: Axo_Mo	815 reflections with I > 3σ(I)
Graded multilayer mirror monochromator	Rint = 0.012
Detector resolution: 13.33 pixels mm^-1	θmax = 40.8°, θmin = 4.0°
rotation method, ω scans	h = −12→11
Absorption correction: multi-scan STOE LANA, absorption correction by scaling of reflection intensities. J. Koziskova, F. Hahn, J. Richter, J. Kozisek, "Comparison of different absorption corrections on the model structure of tetrakis(µ2-acetato)- diaqua-di-copper(II)", Acta Chimica Slovaca, vol. 9, no. 2, 2016, pp. 136 - 140. Afterwards a spherical absorption correction was performed within STOE LANA.	k = −14→15
Tmin = 0.472, Tmax = 0.558	l = −8→3
3414 measured reflections

Refinement

Refinement on F^2	0 constraints
R[F^2 > 2σ(F^2)] = 0.019	Weighting scheme based on measured s.u.'s w = 1/[σ^2(Fo^2) + (0.04P)^2] where P = (Fo^2 + 2Fc^2)/3
wR(F^2) = 0.080	(Δ/σ)max = 0.010
S = 1.82	Δρmax = 0.58 e Å^−3
826 reflections	Δρmin = −0.68 e Å^−3
33 parameters	Extinction correction: B-C type 1 Gaussian isotropic (Becker & Coppens, 1974)
0 restraints	Extinction coefficient: 320 (90)

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

	x	y	z	Uiso*/Ueq	Occ. (<1)
Fe1	0.5	0.190938 (17)	0.5	0.01500 (7)
P1	0.21350 (4)	0	−0.09111 (6)	0.00970 (9)
O1	0.37350 (13)	0	0.21293 (18)	0.01317 (18)
O2	0.22246 (12)	0.14872 (8)	−0.27746 (16)	0.01839 (17)
O3	0	0.0268 (5)	0	0.0338 (13)	0.5

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Fe1	0.01565 (12)	0.00994 (12)	0.01656 (13)	0	−0.00198 (8)	0
P1	0.00869 (14)	0.01253 (15)	0.00775 (14)	0	0.00166 (9)	0
O1	0.0140 (3)	0.0151 (3)	0.0088 (3)	0	−0.0006 (2)	0
O2	0.0261 (3)	0.0132 (3)	0.0146 (3)	0.0051 (2)	0.0023 (2)	0.00333 (18)
O3	0.0112 (5)	0.053 (4)	0.0413 (11)	0	0.0139 (7)	0

Geometric parameters (Å, º)

Fe1—O1	2.1142 (6)	P1—O1	1.5201 (8)
Fe1—O1^i	2.1142 (6)	P1—O2	1.5203 (8)
Fe1—O2^ii	2.3141 (9)	P1—O2^vi	1.5203 (8)
Fe1—O2^iii	2.3141 (9)	P1—O3	1.5759 (7)
Fe1—O2^iv	2.0782 (8)	P1—O3^vii	1.5759 (7)
Fe1—O2^v	2.0782 (8)	O3—O3^vii	0.453 (6)
O1—Fe1—O1^i	80.32 (2)	O1—P1—O3	104.18 (4)
O1—Fe1—O2^ii	84.52 (3)	O1—P1—O3^vii	104.18 (4)
O1—Fe1—O2^iii	81.92 (3)	O2—P1—O2^vi	111.75 (4)
O1—Fe1—O2^iv	94.05 (3)	O2—P1—O3	100.16 (15)
O1—Fe1—O2^v	158.02 (3)	O2—P1—O3^vii	114.49 (14)
O1^i—Fe1—O2^ii	81.92 (3)	O2^vi—P1—O3	114.49 (14)
O1^i—Fe1—O2^iii	84.52 (3)	O2^vi—P1—O3^vii	100.16 (15)
O1^i—Fe1—O2^iv	158.02 (3)	O3—P1—O3^vii	16.5 (2)
O1^i—Fe1—O2^v	94.05 (3)	Fe1—O1—Fe1^viii	99.68 (3)
O2^ii—Fe1—O2^iii	162.24 (3)	Fe1—O1—P1	129.898 (17)
O2^ii—Fe1—O2^iv	76.37 (3)	Fe1^viii—O1—P1	129.898 (17)
O2^ii—Fe1—O2^v	115.92 (3)	Fe1^ix—O2—Fe1^iv	103.63 (3)
O2^iii—Fe1—O2^iv	115.92 (3)	Fe1^ix—O2—P1	120.17 (4)
O2^iii—Fe1—O2^v	76.37 (3)	Fe1^iv—O2—P1	133.71 (5)
O2^iv—Fe1—O2^v	98.48 (3)	P1—O3—P1^x	163.5 (3)
O1—P1—O2	112.72 (3)	P1—O3—O3^vii	81.73 (16)
O1—P1—O2^vi	112.72 (3)	P1^x—O3—O3^vii	81.73 (16)

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