Tetrastrontium dimanganese copper nonaoxide, Sr₄Mn₂CuO₉

Single crystals of Sr₄Mn_2.09Cu_0.91O₉ have been grown by flux synthesis and the structure, closely related to the hexagonal perovskite 2H, was solved from single-crystal X-ray data in space group P321. The structure of Sr₄Mn₂CuO₉ is composed of chains of face-sharing polyhedra with a sequence of two octahedra and one trigonal prism. The octahedra are filled by Mn atoms and the Cu atoms are randomly distributed at the centres of the square faces of the trigonal prism. A stacking fault is observed within one of the two chains, which can be attributed to a shifting of the chain along the c axis.

S1. Comment

Numerous compounds related to the perovskite 2H with general formula A_{1 + x}(A'_xB_{1 - x})O₃ (where 0 ≤ x ≤ 1/2, A is an alkaline earth, and A' and B are alkali, alkaline earth and d transition elements) have been synthesized over the last few years. Their structures consist of chains of B^IVO₆ octahedra and A'^IIO₆ trigonal prisms sharing faces along the c axis, separated by A cations. One of the reasons for the interest in studying such compounds is the possibility of using various combinations of d metals on the A' and B sites to modulate magnetic properties. For example, in the case of Sr₄Mn₂NiO₉ (x = 1/3; El Abed et al., 2001, 2002), it has been shown, by means of magnetic measurements and XANES, that the Mn^IV atoms occupy the B site and Ni^II atoms the A' site. The global magnetic susceptibility has been described with a simple model consisting of two independent spin sublattices, namely antiferromagnetically coupled (Mn⁺⁴)₂ dimers and paramagnetic Ni²⁺ cations at the centre of the trigonal prism. In order to verify if this description can be extended to other compounds of this family, the title compound, Sr₄Mn₂CuO₉, has been synthesized and its structure is presented here.

The structure of Sr₄Mn₂CuO₉ is composed of two different chains of face-sharing polyhedra, with a sequence of two octahedra and one trigonal prism. One of them is centred at (x = 0, y = 0), the other at (x = 1/3, y = 2/3) and (x = 2/3, y = 1/3). The octahedra are filled by Mn atoms, and the Cu atoms are randomly distributed at the centres of the square faces of the trigonal prism. The chains are held together by Sr atoms in three different irregular ten-, nine- and eight-coordinated sites (Fig. 1).

A stacking fault is observed in the chain located in (x = 2/3, y = 1/3), with the occurrence of an Mn octahedron instead of a trigonal prism (Fig. 2). This stacking fault may be attributed to a shifting of the chain along the c axis. This shifting would induce the existence of disordered Cu positions around the Mn2 position. It was not possible to refine properly a model with these positions, the density being delocalized. The stacking fault density is about 14% in this chain.

Even though it is difficult to distinguish clearly between Mn and Cu by X-ray diffraction because of their similar scattering factors, several indications allow us to confirm this cationic distribution. Firstly, from a chemical point of view, all the electron density found in the square face of the trigonal prism should correspond to Cu^II. Secondly, a partial substitution of Mn by Cu in the octahedra leads to a higher residual factor and non-definite positive atomic displacement parameters for some O atoms if the Cu ratio is too high. Finally, the calculations of bond-valence sums are in good agreement with the structure proposed (see below).

For the Mn1, Mn2 and Mn3 positions, the MnO₆ octahedra are quite regular, with Mn—O distances close to the values previously observed in SrMnO₃ (1.872 and 1.906 Å; Battle et al., 1988) and reported by Shannon (1976) (1.93 Å). This is also the case for the extra Mn position Mn2p, corresponding to the stacking fault, when considering the O4 and O3b positions (Fig. 2). It is observed that the Mn2p and O3b fractional occupancies of 0.139 (7) and 0.140 (9), respectively, are very similar, as expected.

The four Cu—O distances correspond to the values found in Sr₃CuPtO₆ (2.003 and 2.018 Å; Hodeau et al., 1992) and are slightly higher than the value calculated from Shannon's table (1.97 Å).

The short Mn—Mn distances are characteristic of Mn^IV cations in face-sharing octahedra (2.500 Å in SrMnO₃, for instance). However, the Mn2p—Mn3 distance is a little too small, and this is may be attributed to the constraints introduced in the refinement between Cu2 and Mn2p.

The calculated bond-valence (BV) sums are 4.0, 4.3 and 3.8 for Mn1, Mn2 and Mn3, respectively, and 3.8 for Mn2p with O4 and O3b as neighbours; the values for the bond-valence calculation are taken from Brown & Altermatt (1985). These values are in good agreement with the proposed charge balance of Sr₄⁺²Mn₂⁺⁴Cu⁺²O₉^-2.

In the case of Cu, the calculated BV sums, considering the four Cu—O distances for both square-planar sites, are 1.4. As the contributions of the two extra Cu—O distances are significant, around 0.1 for each bond, they should be included in the BV sum calculation, leading to a value of 1.6 for both Cu positions. The BV sums for Cu are less than 2 because the interatomic distances are constrained by the lattice structure. These internal strains can be correlated with the high atomic displacement parameters of the O atoms which belong to the trigonal prism. This oxygen motion or disorder tends to reduce the distances between the occupied Cu site and the O atoms that form the square around it (Fig. 2).

The synthesis of a pure powder sample is still in progress, in order to see if our magnetic model based on two independent sub-systems is also valid for this compound.

S2. Experimental

Single crystals of Sr₄Mn₂CuO₉ were grown by flux synthesis. The reagents, SrCO₃ (Aldrich 99.9+%), Mn₂O₃ (Aldrich 99.999%) and CuO (Aldrich 99.99+%) (total 1.5 g), were mixed thoroughly in stoichiometric proportions and placed in an alumina crucible. K₂CO₃ (about 20 g; Aldrich 99.99%) was added on top of the reagents. The filled crucible was covered and heated in air from room temperature to the reaction temperature of 1200 K at 60 K h^-1, held at this temperature for 48 h and subsequently cooled to 880 K at 6 K h^-1. The furnace was then turned off and the system allowed to cool to room temperature. The final product was recovered from the melt by washing with distilled water.

S3. Refinement

The common obverse/reverse twinning law was introduced. The twin matrice was [-1,0,0;0,-1,0;0,0,1] and the refined twin ratio was 0.1169 (14). Because of this twinning, the Flack parameter was not introduced. Both absolute configurations were tested and gave the same result. A model with the O2, O3 and O4 positions disordered over two sites was tested. The R and R_w factors were good (0.035 and 0.071, respectively), but a strong correlation was observed, with large s.u.s for the O occupancies. A refinement with Cu positions around the Mn2 position was also tested, but gave negative displacement parameters for some O-atom positions. Concerning the Cu disorder, no evidence of superstructure or lowering of the trigonal symmetry was observed, as in Sr₃CuPtO₆ (Hodeau et al., 1992). The total occupancies of O3a+O3b and Cu2+Mn2p were constrained to unity.

Computing details

Data collection: KappaCCD Software (Nonius, 1999); cell refinement: HKL DENZO (Otwinowski & Minor 1997); data reduction: HKL DENZO and SCALEPACK (Otwinowski & Minor 1997); program(s) used to solve structure: Please provide missing information; program(s) used to refine structure: JANA2000 (Petříček & Dušek, 2000); software used to prepare material for publication: JANA2000.

Figures

	Fig. 1. A view of the structure of Sr4Mn2CuO9 down the chain direction. The first chain is located at (x = 0, y = 0), and the second at (x = 1/3, y = 2/3) and (x = 2/3, y = 1/3).
	Fig. 2. A view of the two types of chains, at (x = 0, y = 0) (left) and (x = 1/3, y = 2/3) (right). Displacement ellipsoids are shown at the 90% probability level and atoms refined isotropically are represented by spheres of arbitrary radii.

Tetrastrontium dimanganese copper nonaoxide

Crystal data

Cu0.91Mn2.09O9Sr4	Dx = 5.337 Mg m^−3
Mr = 667.9	Mo Kα radiation, λ = 0.71073 Å
Trigonal, P321	Cell parameters from 13458 reflections
Hall symbol: P 3 2"	θ = 3.3–35.0°
a = 9.5817 (2) Å	µ = 30.90 mm^−1
c = 7.8290 (2) Å	T = 293 K
V = 622.48 (2) Å^3	Block, black
Z = 3	0.09 × 0.06 × 0.03 mm
F(000) = 908

Data collection

Nonius KappaCCD area-detector diffractometer	Rint = 0.067
Absorption correction: gaussian (Templeton & Templeton, 1978)	θmax = 35.0°, θmin = 3.6°
Tmin = 0.164, Tmax = 0.468	h = −15→15
11119 measured reflections	k = −14→15
1837 independent reflections	l = −12→12
1517 reflections with I > 2σ(I)

Refinement

Refinement on F^2	Weighting scheme based on measured s.u.'s w = 1/[σ^2(I) + 0.0009I^2]
R[F^2 > 2σ(F^2)] = 0.034	(Δ/σ)max = 0.001
wR(F^2) = 0.076	Δρmax = 1.28 e Å^−3
S = 1.04	Δρmin = −1.55 e Å^−3
1837 reflections	Extinction correction: Becker & Coppens (1974), type I
82 parameters	Extinction coefficient: 0.09 (2)

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

	x	y	z	Uiso*/Ueq	Occ. (<1)
Sr1	0.32498 (17)	0.0	0.5	0.0158 (4)
Sr2	0.36024 (12)	0.0	0.0	0.0083 (4)
Sr3	0.02373 (13)	0.68803 (11)	0.75067 (7)	0.0102 (3)
Mn1	0.0	0.0	0.3366 (2)	0.0072 (4)
Mn2	0.666667	0.333333	0.4326 (2)	0.0146 (5)
Mn3	0.666667	0.333333	0.1028 (3)	0.0103 (4)
Cu1	0.0544 (4)	0.0544 (4)	0.0	0.0154 (11)	0.33333
Cu2	0.7176 (4)	0.3365 (3)	0.7644 (3)	0.0046 (5)*	0.287 (2)
Mn2p	0.666667	0.333333	0.8096 (17)	0.0046 (5)*	0.139 (7)
O1	0.8470 (10)	0.0	0.5	0.010 (2)
O2	0.0060 (8)	0.1630 (8)	0.1997 (6)	0.020 (2)
O3a	0.5118 (10)	0.3304 (9)	0.5605 (8)	0.0306 (17)*	0.860 (9)
O3b	0.549 (5)	0.180 (4)	0.636 (4)	0.0306 (17)*	0.140 (9)
O4	0.6837 (7)	0.1816 (7)	0.9630 (5)	0.011 (2)
O5	0.5171 (8)	0.1694 (9)	0.2621 (5)	0.010 (2)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Sr1	0.0143 (3)	0.0249 (6)	0.0119 (4)	0.0124 (3)	0.0041 (2)	0.0083 (4)
Sr2	0.0072 (4)	0.0092 (6)	0.0092 (3)	0.0046 (3)	−0.0015 (2)	−0.0030 (4)
Sr3	0.0120 (4)	0.0109 (4)	0.0097 (3)	0.0072 (4)	0.0000 (2)	0.0019 (3)
Mn1	0.0091 (6)	0.0091 (6)	0.0034 (6)	0.0045 (3)	0.0	0.0
Mn2	0.0172 (7)	0.0172 (7)	0.0093 (7)	0.0086 (4)	0.0	0.0
Mn3	0.0070 (4)	0.0070 (4)	0.0171 (8)	0.0035 (2)	0.0	0.0
Cu1	0.0229 (14)	0.0229 (14)	0.0094 (12)	0.0181 (14)	−0.0003 (6)	0.0003 (6)
O1	0.009 (3)	0.012 (3)	0.010 (3)	0.0061 (16)	0.0007 (13)	0.001 (3)
O2	0.015 (2)	0.024 (3)	0.027 (2)	0.014 (2)	0.008 (2)	0.022 (2)
O4	0.010 (3)	0.011 (2)	0.018 (2)	0.009 (2)	−0.0001 (19)	−0.002 (2)
O5	0.007 (3)	0.014 (3)	0.008 (2)	0.004 (2)	0.0007 (17)	0.0019 (18)

Geometric parameters (Å, º)

Sr1—O1^i	2.698 (2)	Mn2—O5^v	2.013 (8)
Sr1—O1^ii	2.698 (12)	Mn2—O5^ii	2.013 (8)
Sr1—O2^iii	2.834 (6)	Mn3—Mn2	2.582 (3)
Sr1—O2^iv	2.834 (8)	Mn3—Cu2^ix	2.691 (3)
Sr1—O3a	2.790 (7)	Mn3—Cu2^xix	2.691 (3)
Sr1—O3a^v	2.900 (8)	Mn3—Cu2^xi	2.691 (3)
Sr1—O3a^vi	2.900 (14)	Mn3—Mn2p^ix	2.292 (16)
Sr1—O3a^vii	2.790 (8)	Mn3—O4^ix	1.891 (7)
Sr1—O3b	2.24 (3)	Mn3—O4^xix	1.891 (6)
Sr1—O3b^vii	2.24 (5)	Mn3—O4^xi	1.891 (9)
Sr1—O5	2.550 (5)	Mn3—O5	1.956 (5)
Sr1—O5^vii	2.550 (8)	Mn3—O5^v	1.956 (8)
Sr2—O2^viii	2.475 (7)	Mn3—O5^ii	1.956 (8)
Sr2—O2^iv	2.475 (9)	Cu1—Mn1	2.6863 (16)
Sr2—O4^ix	2.707 (6)	Cu1—Mn1^viii	2.6863 (16)
Sr2—O4^x	2.648 (6)	Cu1—Cu1^xiii	0.903 (5)
Sr2—O4^xi	2.648 (6)	Cu1—Cu1^iv	0.903 (5)
Sr2—O4^vii	2.707 (10)	Cu1—O2	2.055 (8)
Sr2—O5	2.582 (5)	Cu1—O2^xiii	2.582 (6)
Sr2—O5^xii	2.582 (7)	Cu1—O2^viii	2.055 (8)
Sr3—O1^xiii	2.563 (6)	Cu1—O2^iv	2.077 (9)
Sr3—O2^xiv	2.674 (7)	Cu1—O2^xx	2.077 (9)
Sr3—O2^xv	2.767 (12)	Cu1—O2^xii	2.582 (6)
Sr3—O3a^xvi	2.447 (8)	Cu2—Mn2	2.641 (3)
Sr3—O3a^xv	2.919 (9)	Cu2—Mn3^xxi	2.691 (3)
Sr3—O3b^xiii	2.83 (5)	Cu2—Cu2^v	0.820 (6)
Sr3—O4^xiii	2.510 (5)	Cu2—Cu2^ii	0.820 (5)
Sr3—O4^xvii	2.716 (5)	Cu2—Mn2p	0.593 (10)
Sr3—O5^iii	2.632 (10)	Cu2—O3a	2.513 (10)
Sr3—O5^xv	2.697 (10)	Cu2—O3a^v	2.069 (10)
Mn1—Mn1^iii	2.559 (2)	Cu2—O3a^ii	2.046 (7)
Mn1—Cu1	2.6863 (16)	Cu2—O3b	1.86 (3)
Mn1—Cu1^xiii	2.6863 (18)	Cu2—O3b^v	1.33 (3)
Mn1—Cu1^iv	2.6863 (18)	Cu2—O3b^ii	1.95 (6)
Mn1—O1^xviii	1.946 (7)	Cu2—O4	2.060 (6)
Mn1—O1^i	1.946 (7)	Cu2—O4^v	2.082 (6)
Mn1—O1^ii	1.946 (5)	Cu2—O4^ii	2.545 (9)
Mn1—O2	1.871 (8)	Mn2p—Mn2	2.955 (16)
Mn1—O2^xiii	1.871 (6)	Mn2p—Mn3^xxi	2.292 (16)
Mn1—O2^iv	1.871 (10)	Mn2p—Cu2	0.593 (10)
Mn2—Mn3	2.582 (3)	Mn2p—Cu2^v	0.593 (10)
Mn2—Cu2	2.641 (3)	Mn2p—Cu2^ii	0.593 (10)
Mn2—Cu2^v	2.641 (3)	Mn2p—O3a	2.444 (15)
Mn2—Cu2^ii	2.641 (3)	Mn2p—O3a^v	2.444 (16)
Mn2—Mn2p	2.955 (16)	Mn2p—O3a^ii	2.444 (14)
Mn2—O3a	1.777 (9)	Mn2p—O3b	1.90 (3)
Mn2—O3a^v	1.777 (12)	Mn2p—O3b^v	1.90 (4)
Mn2—O3a^ii	1.777 (8)	Mn2p—O3b^ii	1.90 (5)
Mn2—O3b	2.08 (3)	Mn2p—O4	1.952 (12)
Mn2—O3b^v	2.08 (4)	Mn2p—O4^v	1.953 (11)
Mn2—O3b^ii	2.08 (5)	Mn2p—O4^ii	1.953 (13)
Mn2—O5	2.013 (5)
O2—Cu1—O2^xiii	68.8 (3)	O3a^v—Cu2—O3a	66.6 (4)
O2—Cu1—O2^viii	163.9 (3)	O3a^v—Cu2—O3a^ii	76.3 (3)
O2—Cu1—O2^iv	80.0 (3)	O3a^v—Cu2—O3b	37.9 (15)
O2—Cu1—O2^xx	98.4 (4)	O3a^v—Cu2—O3b^v	57.1 (18)
O2—Cu1—O2^xii	125.8 (3)	O3a^v—Cu2—O3b^ii	90.1 (12)
O2^xiii—Cu1—O2	68.8 (3)	O3a^v—Cu2—O4	99.6 (4)
O2^xiii—Cu1—O2^viii	125.8 (3)	O3a^v—Cu2—O4^v	167.1 (4)
O2^xiii—Cu1—O2^iv	68.5 (3)	O3a^v—Cu2—O4^ii	122.6 (3)
O2^xiii—Cu1—O2^xx	121.4 (3)	O3a^ii—Cu2—O3a	66.9 (4)
O2^xiii—Cu1—O2^xii	74.57 (19)	O3a^ii—Cu2—O3a^v	76.3 (3)
O2^viii—Cu1—O2	163.9 (3)	O3a^ii—Cu2—O3b	93.6 (10)
O2^viii—Cu1—O2^xiii	125.8 (3)	O3a^ii—Cu2—O3b^v	38 (3)
O2^viii—Cu1—O2^iv	98.4 (4)	O3a^ii—Cu2—O3b^ii	51.8 (14)
O2^viii—Cu1—O2^xx	80.0 (3)	O3a^ii—Cu2—O4	161.5 (5)
O2^viii—Cu1—O2^xii	68.8 (3)	O3a^ii—Cu2—O4^v	99.6 (3)
O2^iv—Cu1—O2	80.0 (3)	O3a^ii—Cu2—O4^ii	127.6 (5)
O2^iv—Cu1—O2^xiii	68.5 (3)	O3b—Cu2—O3b^v	91.3 (19)
O2^iv—Cu1—O2^viii	98.4 (4)	O3b—Cu2—O3b^ii	74.5 (18)
O2^iv—Cu1—O2^xx	168.8 (3)	O3b—Cu2—O4	93.8 (11)
O2^iv—Cu1—O2^xii	121.4 (3)	O3b—Cu2—O4^v	154.8 (14)
O2^xx—Cu1—O2	98.4 (4)	O3b—Cu2—O4^ii	85.2 (14)
O2^xx—Cu1—O2^xiii	121.4 (3)	O3b^v—Cu2—O3b	91.3 (19)
O2^xx—Cu1—O2^viii	80.0 (3)	O3b^v—Cu2—O3b^ii	87 (3)
O2^xx—Cu1—O2^iv	168.8 (3)	O3b^v—Cu2—O4	125 (3)
O2^xx—Cu1—O2^xii	68.5 (3)	O3b^v—Cu2—O4^v	112.3 (16)
O2^xii—Cu1—O2	125.8 (3)	O3b^v—Cu2—O4^ii	165 (3)
O2^xii—Cu1—O2^xiii	74.57 (19)	O3b^ii—Cu2—O3b	74.5 (18)
O2^xii—Cu1—O2^viii	68.8 (3)	O3b^ii—Cu2—O3b^v	87 (3)
O2^xii—Cu1—O2^iv	121.4 (3)	O3b^ii—Cu2—O4	146.6 (14)
O2^xii—Cu1—O2^xx	68.5 (3)	O3b^ii—Cu2—O4^v	97.0 (11)
O3a—Cu2—O3a^v	66.6 (4)	O3b^ii—Cu2—O4^ii	77.9 (14)
O3a—Cu2—O3a^ii	66.9 (4)	O4—Cu2—O4^v	80.3 (3)
O3a—Cu2—O3b	44.1 (15)	O4—Cu2—O4^ii	70.0 (3)
O3a—Cu2—O3b^v	90.2 (18)	O4^v—Cu2—O4	80.3 (3)
O3a—Cu2—O3b^ii	30.4 (11)	O4^v—Cu2—O4^ii	69.7 (2)
O3a—Cu2—O4	128.6 (2)	O4^ii—Cu2—O4	70.0 (3)
O3a—Cu2—O4^v	123.5 (4)	O4^ii—Cu2—O4^v	69.7 (2)
O3a—Cu2—O4^ii	77.1 (3)

Symmetry codes: (i) −y, x−y−1, z; (ii) −x+y+1, −x+1, z; (iii) y, x, −z+1; (iv) −x+y, −x, z; (v) −y+1, x−y, z; (vi) −x+1, −x+y, −z+1; (vii) x−y, −y, −z+1; (viii) y, x, −z; (ix) x, y, z−1; (x) y, x−1, −z+1; (xi) −x+y+1, −x+1, z−1; (xii) x−y, −y, −z; (xiii) −y, x−y, z; (xiv) y, x+1, −z+1; (xv) x−y, −y+1, −z+1; (xvi) −x+y, −x+1, z; (xvii) y, x, −z+2; (xviii) x−1, y, z; (xix) −y+1, x−y, z−1; (xx) −x, −x+y, −z; (xxi) x, y, z+1.