Anisotropic Thermal Expansion and a Second-Order Charge Order Transition in the Ferrimagnetic Dy₂CuZnMn₄O₁₂ Perovskite with Triple A-Site Cation Ordering

Abstract

Dy₂CuZnMn₄O₁₂ perovskite, belonging to the A-site columnar-ordered quadruple perovskite family with the general composition of A₂A′A″B₄O₁₂, was prepared by a high-pressure, high-temperature method at 6 GPa and 1500 K. Its crystal structure was studied by synchrotron powder X-ray diffraction between 100 and 800 K. The ideal cation distribution (without antisite disorder) was found to be realized within the sensitivity of the synchrotron X-ray diffraction method. Between 100 and 400 K, it crystallizes in space group Pmmn (no. 59) and has layered charge ordering of Mn³⁺ and Mn⁴⁺ at the B sites. Above 425 K, it crystallizes in space group P4₂/nmc (no. 137) with one crystallographic B site and an average Mn^3.5+ oxidation state. The charge ordering transition (at T_CO = 425 K) appears to be of the second order as no anomalies were found on differential scanning calorimetry curves and temperature dependence of the unit cell volume, and the orthorhombic a and b lattice parameters merge gradually. The compound demonstrates anisotropic thermal expansion with the c lattice parameter decreasing with increasing temperature above 280 K. A ferrimagnetic transition occurs at T_C = 116 K with an additional, gradual rise of magnetic susceptibilities below 45 K, probably due to increases of the ordered moments of the Dy sublattices.

Short abstract

A new member of the A-site columnar-ordered quadruple perovskites was prepared at high pressure, Dy₂CuZnMn₄O₁₂, with triple A-site ordering, ferrimagnetic properties, and a charge-order transition.

1. Introduction

A-site columnar-ordered quadruple perovskite oxides with the general composition of A₂A′A″B₄O₁₂ are a new playground in the perovskite science to play with compositional variations and their effects on physical properties.¹ In comparison with classical ABO₃ perovskites with one type of A sites and A-site-ordered quadruple perovskites, AA′₃B₄O₁₂, with intrinsically two types of A sites (A with a 12-fold coordination and A′ with a square-planar coordination), A₂A′A″B₄O₁₂ perovskites have an additional degree of freedom in the presence of the A″ site with a tetrahedral coordination (as the first coordination sphere). Therefore, a principally different, new set of cations, which are not usually located at A sites of perovskites, can be used for the A′′ site, such as Ga³⁺ and Zn²⁺.²⁻⁵

The parent structure of the A₂A′A″B₄O₁₂ perovskites has P4₂/nmc symmetry and is formed through large tilts of the a⁺a⁺c^– type.¹ This symmetry is retained in the majority of examples of such perovskites.^1,6 The next most commonly observed symmetry is P4₂/n, which is produced through full or partial rock-salt-type B-cation orderings, A₂A′A″B₂B′₂O₁₂.^1,6 The less common symmetries are P4₂mc (which is formed through polar distortions in CaMnTi₂O₆,⁷ CaMnTi_2–xV_xO₆,⁸ and NaYMnMnTi₄O₁₂)⁹ and Pmmn (which is formed through layered-type B-cation ordering in RMn₃O₆ (R = Gd to Tm)^10,11 and R₂CuMnMn₄O₁₂ (R = Dy and Y)¹² due to partial or full charge ordering of Mn³⁺ and Mn⁴⁺, respectively). The presence of charge-ordered structures based on the same element could suggest the existence of temperature-driven charge-(dis)ordered transitions as observed in simple perovskites (e.g., half-doped R³⁺_0.5A²⁺_0.5MnO₃)¹³ or quadruple perovskites (e.g., ACu₃Fe₄O₁₂).¹⁴ However, such charge-order transitions have not been observed so far in A₂A′A″B₄O₁₂ perovskites.^1,10

In this work, we prepared a new member of the A-site columnar-ordered quadruple perovskite family with the composition of Dy₂CuZnMn₄O₁₂. In this compound, a new combination of the A/A′/A″ cations was used (R³⁺/Cu²⁺/Zn²⁺) to achieve the triple A-site cation ordering. This combination also gives an average oxidation state of Mn as +3.5—an ideal value for the realization of possible charge ordering. We indeed found that Dy₂CuZnMn₄O₁₂ crystallizes in space group Pmmn at room temperature with 1:1 charge order of Mn³⁺ and Mn⁴⁺ at the B sites. High-temperature structural studies showed the existence of a charge-disorder transition above 425 K, which is of the second order. We also observed anisotropic thermal expansion above 280 K and a ferrimagnetic transition below T_C = 116 K.

2. Experimental Section

Dy₂CuZnMn₄O₁₂ was prepared from a stoichiometric mixture of o-DyMnO₃, ZnO (99.9%), CuO (99.9%), and MnO₂ (Alfa Aesar, 99.9%) at about 6 GPa and about 1500 K for 2 h in Au capsules. After annealing at 1500 K, the samples were cooled down to room temperature by turning off the heating current, and the pressure was slowly released. We note that the oxygen content of a commercial well-crystallized single-phase MnO₂ chemical was confirmed by thermal analysis. Single-phase o-DyMnO₃ was prepared from a stoichiometric mixture of Dy₂O₃ (99.9%) and Mn₂O₃ (99.99%) by annealing in air at 1430 K for 60 h with several intermediate grindings. After the high-pressure synthesis, the sample was recovered as black powder (Figure S1). Therefore, some measurements (such as, specific heat, resistivity, and dielectric constant), which require dense pellets, could not be performed. We note that the melting point of Au at 6 GPa is above 1600 K;¹⁵ therefore, Au capsules could be safely used for the synthesis at 1500 K. We did not observe any reaction between Au and Dy₂CuZnMn₄O₁₂ or any damage of Au capsules.

X-ray powder diffraction (XRPD) data were collected at room temperature on a RIGAKU MiniFlex600 diffractometer using Cu Kα radiation (2θ range of 8–100°, a step width of 0.02°, and scan speed of 2°/min). Synchrotron XRPD data were collected from 100 to 800 K on beamline BL02B2 of SPring-8 (the intensity data were taken between 2.08° and 78.22° at 0.006° intervals in 2θ using a wavelength of λ = 0.619283 Å).¹⁶ High statistics data with a measurement time of 100 s were collected at 100, 295, and 600 K; at all other temperatures, the measurement time was 10 s. The sample was placed into an open Lindemann glass capillary tube (inner diameter: 0.2 mm), which was rotated during measurements. The Rietveld analysis of all XRPD data was performed using the RIETAN-2000 program.¹⁷

Magnetic measurements were performed on a SQUID magnetometer (Quantum Design, MPMS3) between 2 and 400 K in applied fields of 100 Oe and 10 kOe under both zero-field-cooled (ZFC) and field-cooled on cooling (FCC) conditions. High-temperature data were taken at 70 kOe from 300 to 470 K. Isothermal magnetization measurements were performed between −70 and 70 kOe at different temperatures. A small sample weight of 11.09 mg was used for the dc magnetic measurements because of a large magnetic moment (1.82 emu at 5 K and 70 kOe). Frequency dependent alternating current (ac) susceptibility measurements were performed using a Quantum Design MPMS3 instrument at a zero static dc field and at different frequencies (f) and different applied oscillating magnetic fields (H_ac) from 140 to 2 K using a 24.7 mg sample.

Differential scanning calorimetry (DSC) curves of a powder sample were recorded on a Mettler Toledo DSC1 STAR^e system under a N₂ flow between 125 and 670 K in an open Al capsule with a heating/cooling rate of 10 K/min. Two DSC runs were performed to check the reproducibility. No DSC anomalies were found between 125 and 670 K.

3. Results and Discussion

Dy₂CuZnMn₄O₁₂ was found to be single-phase within the sensitivity of the laboratory XRPD data. All its reflections could be indexed in orthorhombic symmetry with a = 7.2599 Å, b = 7.2709 Å, and c = 7.7736 Å. Therefore, the Pmmn structural model and the structural parameters of DyMn₃O₆ (as the initial ones)¹⁰ were used in the Rietveld analysis of Dy₂CuZnMn₄O₁₂. Cu²⁺ (27 electrons) and Zn²⁺ (28 electrons) cations cannot be distinguished by (nonresonant) synchrotron XRPD. Therefore, we assumed the distribution of Cu²⁺ and Zn²⁺ based on their coordination preferences: Zn²⁺ cations have a strong preference for tetrahedral sites (among tetrahedral and square-planar ones) while Cu²⁺ cations have a strong preference for square-planar sites.¹⁸ We note that CaZnV₂O₆¹⁹ and CaCuFeReO₆²⁰ were recently prepared, where Zn²⁺ and Cu²⁺ cations are located in both square-planar (A′) and tetrahedral (A″) sites. However, these compounds were stabilized at much higher pressures of 12 and 15.5 GPa, respectively, and they were not stable at 6 GPa. Therefore, we can assume that usual tendencies in the site preferences for Cu²⁺ and Zn²⁺ cations are realized in Dy₂CuZnMn₄O₁₂ prepared at 6 GPa.

Small antisite disorders are often realized in A₂A′A″B₄O₁₂ perovskites with A = rare earths, A′ = Mn, and A″ = Mn. Such antisite disorders could be observed with synchrotron XRPD data as small (but detectable) deviations of occupation factors from unity (for ideal cation distribution models).^3−5,11 However, during the structural analysis of Dy₂CuZnMn₄O₁₂, all cation occupation factors converged to values close to unity. We report here the occupation factors at 600 K as the symmetry at 600 K is higher (see below) and the number of refined structural parameters is smaller. The following values were obtained (when all other structural and nonstructural parameters were refined simultaneously including atomic displacement parameters): g(Dy) = 0.985(2), g(Cu) = 0.482(3) (for a disordered model with the ideal g = 0.5), and g(Zn) = 0.995(6) with fixed g(Mn) = 1 and g(Cu) = 0.487(3), g(Zn) = 1.014(6), and g(Mn) = 1.018(2) with fixed g(Dy) = 1. We also note that when Mn was located at the tetrahedral A″ site (or the square-planar A′ site), its occupation factor was refined to be 1.265(7) (or 0.600(3) for the A′ site), that is, significantly larger than unity (or 0.5). Therefore, we concluded that the ideal cation distributions are realized in Dy₂CuZnMn₄O₁₂ within the sensitivity of synchrotron XRPD data.

Refined structural parameters, primary bond lengths, and bond valence sums (BVS)²¹ of Dy₂CuZnMn₄O₁₂ at 100 and 295 K are summarized in Tables 1 and 2 and Table S1. Experimental, calculated, and difference synchrotron XRPD patterns of Dy₂CuZnMn₄O₁₂ at 295 K are shown in Figure 1. The crystal structure of Dy₂CuZnMn₄O₁₂ at 295 K is presented in Figure 2a.

Table 1. Structure Parameters of Dy₂CuZnMn₄O₁₂ at 100 K (the First Line for Each Site) and 295 K (the Second Line for Each Site) from Synchrotron XRPD Data^a.

site	WP	x	y	z	Biso (Å2)
Dy1	2a	0.25	0.25	0.7785(2)	0.08(2)
		0.25	0.25	0.7784(2)	0.27(3)
Dy2	2a	0.25	0.25	0.2786(2)	0.15(3)
		0.25	0.25	0.2786(2)	0.40(3)
Cu	2b	0.75	0.25	0.7349(4)	0.41(5)
		0.75	0.25	0.7374(5)	0.61(6)
Zn	2b	0.75	0.25	0.2486(4)	0.03(3)
		0.75	0.25	0.2488(5)	0.30(5)
Mn1	4c	0.5	0	0	0.03(5)
		0.5	0	0	0.11(8)
Mn2	4d	0	0.5	0.5	0.10(5)
		0	0.5	0.5	0.26(8)
O1	8g	0.4402(7)	–0.0599(7)	0.2624(10)	0.61(8)
		0.4389(8)	–0.0589(8)	0.2610(11)	1.02(9)
O2	4f	0.0634(10)	0.25	0.0336(13)	0.19(12)
		0.0623(10)	0.25	0.0344(13)	0.08(11)
O3	4e	0.25	0.5396(11)	0.9105(11)	0.18(13)
		0.25	0.5380(11)	0.9083(12)	0.30(14)
O4	4f	0.5391(12)	0.25	0.4231(13)	0.61(16)
		0.5389(11)	0.25	0.4238(13)	0.71(16)
O5	4e	0.25	0.4330(10)	0.5386(13)	0.06(11)
		0.25	0.4353(10)	0.5405(13)	0.25(13)

^aSource: Synchrotron powder X-ray diffraction (λ = 0.61928 Å); used d-space range: from 0.4916 to 7.1 Å. Crystal system: orthorhombic; space group Pmmn (no. 59, origin choice 2); Z = 2. Molecular weight: 865.6708 g/mol. The occupation factors (g) of all the sites are unity (g = 1). WP: Wyckoff position. No detectable impurities. 100 K: a = 7.25134(2) Å, b = 7.25958(2) Å, c = 7.76916(1) Å, and V = 408.9820(14) ų; R_wp = 7.02%, R_p = 5.15%, R_B = 2.86%, and R_F = 1.41%; ρ_cal = 7.030 g/cm³. 295 K: a = 7.25988(2) Å, b = 7.27087(2) Å, c = 7.77362(2) Å, and V = 410.3355(19) ų; R_wp = 7.26%, R_p = 5.33%, R_B = 3.22%, and R_F = 1.66%; ρ_cal = 7.006 g/cm³.

Table 2. Selected Bond Lengths (l (Å) < 2.8 Å), Bond Angles (deg), Bond Valence Sums, BVS, and Distortion Parameters of MnO₆, Δ, in Dy₂CuZnMn₄O₁₂ at 295 K^a.

Dy1–O5 × 2	2.288(9)	Dy2–O2 × 2	2.337(9)
Dy1–O3 × 2	2.325(8)	Dy2–O4 × 2	2.382(9)
Dy1–O2 × 2	2.412(9)	Dy2–O5 × 2	2.441(9)
Dy1–O1 × 4	2.669(7)	Dy2–O1 × 4	2.635(7)
BVS(Dy13+)	+3.38	BVS(Dy23+)	+3.13
Cu–O1 × 4	1.952(4)	Zn–O3 × 2	1.967(9)
BVS(Cu2+)	+1.91	Zn–O4 × 2	2.050(9)
		BVS(Zn2+)	+1.77
Mn1–O2 × 2	1.892(2)	Mn2–O5 × 2	1.901(2)
Mn1–O3 × 2	1.969(3)	Mn2–O4 × 2	1.932(3)
Mn1–O1 × 2	2.120(8)	Mn2–O1 × 2	1.958(8)
BVS(Mn13+)	+3.29	BVS(Mn24+)	+3.72
Δ(Mn1–O)	22.6 × 10–4	Δ(Mn2–O)	1.4 × 10–4
Mn1–O1–Mn2 × 2	144.73(9)	Mn2–O4–Mn2	140.32(9)
Mn1–O2–Mn1	147.75(9)	Mn2–O5–Mn2	145.35(9)
Mn1–O3–Mn1	134.33(9)

^aBVS = ∑^N_i=1ν_i, ν_i = exp[(R₀ – l_i)/B], N is the coordination number, B = 0.37, R₀(Dy³⁺) = 2.036, R₀(Cu²⁺) = 1.679, R₀(Zn²⁺) = 1.704, R₀(Mn⁴⁺) = 1.753, and R₀(Mn³⁺) = 1.76.

Figure 1.

Experimental (black crosses), calculated (red line), and difference (blue line at the bottom) synchrotron powder X-ray diffraction patterns of Dy₂CuZnMn₄O₁₂ at T = 295 K between 6 and 60°. The tick marks show possible Bragg reflection positions. The inset shows similar curves at T = 600 K between 6 and 26°.

Figure 2.

Crystal structures of Dy₂CuZnMn₄O₁₂ at (a) T = 295 K in the charge-ordered state and (b) T = 600 K in the charge-disordered state.

The Pmmn structural model was used to obtain the lattice parameters between 100 and 450 K (Figure 3). At 425 and 450 K, the orthorhombic a_O and b_O lattice parameters already merged, suggesting a phase transition to a tetragonal structure. The orthorhombic splitting of reflections also disappeared (Figure 4). Therefore, at 450 K and above, we used the P4₂/nmc model³ to obtain temperature dependence of the lattice parameters. The refined structural parameters and bond lengths at 600 K are summarized in Tables 3 and 4, and the fitting patterns are shown in the inset of Figure 1. The crystal structure of Dy₂CuZnMn₄O₁₂ at 600 K is given in Figure 2b. We note that the Cu site could be split from the 2a site (with g = 1) to the 4c site (with g = 0.5) in the P4₂/nmc model.

Figure 3.

(a) Temperature dependence of the a_O, b_O, and a_T lattice parameters of Dy₂CuZnMn₄O₁₂ between 100 and 800 K. (b) Temperature dependence of the c_O and c_T lattice parameters (the left-hand axis) and the unit-cell volume (the right-hand axis) of Dy₂CuZnMn₄O₁₂. Errors are smaller than the symbol sizes. α_V is the volumetric coefficient of thermal expansion calculated between T = 450 K and T = 775 K. O, orthorhombic; T, tetragonal.

Figure 4.

Temperature dependence of the synchrotron X-ray powder diffraction patterns of Dy₂CuZnMn₄O₁₂ from 100 to 450 K (between 6 and 30°). The inset shows details near the 040_O, 400_O, and 400_T reflections to emphasize the disappearance of the orthorhombic (O) distortion. T, tetragonal.

Table 3. Structure Parameters of Dy₂CuZnMn₄O₁₂ at 600 K from Synchrotron X-ray Powder Diffraction Data^a.

atom	WP	g	x	y	z	Biso (Å2)
Dy	4d	1	0.25	0.25	0.22151(5)	0.788(8)
Cu	4c	0.5	0.75	0.25	0.7642(8)	1.04(7)
Zn	2b	1	0.75	0.25	0.25	0.83(5)
Mn	8e	1	0	0	0	0.500(10)
O1	8g	1	0.25	0.0629(5)	–0.0362(5)	0.68(7)
O2	8g	1	0.25	0.5393(5)	0.5849(5)	1.08(9)
O3	8f	1	0.4395(4)	–x	0.25	1.51(9)

^ag is the occupation factor. Source: Synchrotron powder X-ray diffraction (λ = 0.61928 Å); used d-space range: from 0.4916 to 7.1 Å. Crystal system: tetragonal; space group: P4₂/nmc (no. 137, origin choice 2); Z = 2. Molecular weight: 865.6708 g/mol. a = 7.30342(1) Å, c = 7.75007(2) Å, and V = 413.3886(12) ų; R_wp = 7.26%, R_p = 5.40%, R_B = 4.78%, and R_F = 3.97%; ρ_cal = 6.955 g/cm³.

Table 4. Bond Lengths (in Å; Below 2.8 Å), Bond Angles (in deg), Bond-Valence Sum (BVS), and Distortion Parameters of MnO₆ (Δ) in Dy₂CuZnMn₄O₁₂ at 600 K^a.

Dy–O1 × 2	2.323(4)	Mn–O1 × 2	1.904(1)
Dy–O2 × 2	2.363(4)	Mn–O2 × 2	1.962(2)
Dy–O1 × 2	2.420(4)	Mn–O3 × 2	2.036(1)
Dy–O3 × 4	2.666(1)	Δ(MnO6)	7.6 × 10–4
BVS(Dy3+)	+3.19	BVS(Mn3+)	+3.47
Cu–O3 × 4	1.960(4)	Mn–O1–Mn × 2	147.15(6)
BVS(Cu2+)	+1.87	Mn–O2–Mn × 2	137.08(6)
Zn–O2 × 4	2.002(4)	Mn–O3–Mn × 2	144.25(6)
BVS(Zn2+)	+1.79

In the Pmmn model, there are two independent sites for the B cations, Mn1 (4c) and Mn2 (4d). The Mn1 and Mn2 sites have different BVS parameters of +3.29 and +3.72. More importantly, the Mn1 and Mn2 sites have very different octahedral distortion parameters of 22.6 × 10^–4 and 1.4 × 10^–4, respectively. These two facts give evidence that the Mn1 site should be occupied by Mn³⁺ cations resulting in a strong Jahn–Teller distortion of the Mn1O₆ octahedron. The Mn2 site should be occupied by Mn⁴⁺ cations without strong Jahn–Teller distortions of the Mn2O₆ octahedron. We note that BVS parameters for Mn³⁺ at B sites of perovskites (for example, in RMnO₃ with strong Jahn–Teller distortions where Mn definitely has an oxidation state of +3)^22,23 are often higher than expected (for example, +3.15 to +3.25). At 600 K in the P4₂/nmc model, there is one crystallographic site for the B cations. The BVS value of +3.47 was close to the expected average value of +3.5, and the octahedral distortion parameter had an intermediate value of 7.6 × 10^–4, reflecting the statistical presence of 50% of Mn³⁺ cations. Therefore, we conclude that a charge-order (CO) structure is realized below T_CO = 425 K. Figure 5 shows temperature dependence of the Mn–O, Cu–O, and Zn–O bond lengths. The Mn1–O1 bond shows a tendency for a gradual decrease with increasing temperature, and the Mn2–O1 and Mn–O1 bonds show a tendency for a gradual increase. Other bonds were nearly temperature independent within the sensitivity of the structural analysis based on synchrotron powder X-ray diffraction.

Figure 5.

Temperature dependence of (a) the Mn–O bond lengths and (b) the Cu–O and Zn–O bond lengths in Dy₂CuZnMn₄O₁₂.

No DSC anomalies were detected near T_CO = 425 K (Figure S2 of the Supporting Information). The temperature dependence of the unit cell volume (Figure 3b) shows no anomalies between 100 and 800 K. The orthorhombic a_O and b_O lattice parameters merge gradually when approaching T_CO = 425 K. All of these observations suggest that the structural phase transition at T_CO = 425 K is of the second order.

The orthorhombic a_O and b_O lattice parameters and tetragonal a_T lattice parameter monotonically increase with temperature from 100 to 800 K (Figure 3). On the other hand, the orthorhombic c_O lattice parameter slightly increases from 100 to 280 K and then decreases from 280 to 800 K (as c_T above 425 K). Therefore, Dy₂CuZnMn₄O₁₂ demonstrates anisotropic thermal expansion above 280 K. As shown in Figure 5, no detectable anomalies were observed on temperature dependence of the bond lengths. Therefore, the origin of the anisotropic thermal expansion is not clear at the moment. Neutron diffraction studies, which can locate oxygen atoms more accurately, will be needed to understand the structural evolution of Dy₂CuZnMn₄O₁₂ with temperature more precisely.

Temperature-driven structural phase transitions in A₂A′A″B₄O₁₂ perovskites have only been discovered so far in CaMnTi₂O₆⁷ and CaMnTi_2–xV_xO₆⁸ during a ferroelectric-paraelectric phase transition (P4₂mc ⇔ P4₂/nmc). The discovery of a charge-order transition in Dy₂CuZnMn₄O₁₂ (Pmmn ⇔ P4₂/nmc) could suggest the existence of such temperature-driven transitions in other A₂A′A″B₄O₁₂ perovskites that crystallize in space group Pmmn at room temperature (e.g., RMn₃O₆^10,11 and R₂CuMnMn₄O₁₂).¹² High-temperature DSC experiments were performed in the case of RMn₃O₆,¹⁰ but no anomalies were detected. Therefore, it was concluded that RMn₃O₆ does not show charge-order transitions up to 873 K. Our current results give evidence that such transitions could not be detected by DSC, and direct high-temperature structural studies are needed to discover such transitions.

FCC magnetic susceptibilities showed sharp rises below T_C = 116 K in a small applied magnetic field of 100 Oe, suggesting the development of a strong ferromagnetic (FM) component (Figure 6) with additional, gradual rises below about 45 K. The additional anomalies could be more clearly seen on the ZFC curve and on the differential dχT/dT versus T curves (Figure S3). At H = 100 Oe, a difference between the ZFC and FCC χ versus T curves was observed below T_C. At H = 10 kOe, almost no difference was detected between the ZFC and FCC χ versus T curves. The χ^–1 versus T curves followed the Curie–Weiss law, and a fit by the law between 300 and 400 K gave a positive Curie–Weiss temperature, θ = +125 K, indicating predominantly FM interactions between magnetic ions. The calculated effective magnetic moment (μ_calc) is 17.485μ_B (for the calculation we used 10.6μ_B for Dy³⁺, 4.899μ_B for Mn³⁺, 3.873μ_B for Mn⁴⁺, and 1.732μ_B for Cu²⁺ and an equation μ_calc² = 2 μ_Dy² + 2 μ_Mn(III)² + 2 μ_Mn(IV)² + μ_Cu²). The experimental effective magnetic moment was about 9% smaller (μ_eff = 15.99μ_B). While the origin of this reduction is not clear at the moment, the same tendency was observed for all other members of the R₂CuZnMn₄O₁₂ (for example, μ_calc = 9.00μ_B versus μ_eff = 8.29μ_B for R = Lu) and R₂CuGaMn₄O₁₂ series.³

Figure 6.

Left-hand axis presents ZFC (filled symbols) and FCC (empty symbols) dc magnetic susceptibility (χ = M/H) curves of Dy₂CuZnMn₄O₁₂ measured at 100 Oe and 10 kOe (multiplied by 10). Right-hand axis shows the FCC χ^–1 versus T curves at 10 kOe with the Curie–Weiss fit (black line). Parameters of the fit are shown on the figure.

Temperature dependence of magnetic susceptibilities of Dy₂CuZnMn₄O₁₂ was qualitatively similar to that of the parent Dy₂MnMnMn₄O₁₂ (= DyMn₃O₆).¹⁰ But the absolute values of magnetic susceptibilities were about 10 times larger in Dy₂CuZnMn₄O₁₂ at both H = 100 Oe and 10 kOe. This fact suggests that the ordered moments in Dy₂CuZnMn₄O₁₂ were much larger than those of Dy₂MnMnMn₄O₁₂. Recent neutron diffraction studies found that the ordered moments were significantly reduced (in comparison with the expected full moments) in an isostructural compound Y₂MnMnMn₄O₁₂ due to competing exchange interactions, and there was a competition between antiferromagnetic and ferrimagnetic ground states.²⁴ The magnetic frustration could be reduced through the introduction of nonmagnetic Zn²⁺ cations into the A″ site in Dy₂CuZnMn₄O₁₂ and through the full charge ordering of Mn³⁺ and Mn⁴⁺, resulting in larger ordered magnetic moments. The gradual rise of magnetic susceptibilities below 45 K could be caused by the gradual increase of the ordered moments of the Dy³⁺ sublattices, but it was difficult (from the available data) to determine at what temperature the Dy³⁺ sublattices start to order.

On the other hand, temperature dependence of magnetic susceptibilities of Dy₂CuZnMn₄O₁₂ was different from that of Dy₂CuMnMn₄O₁₂ at small magnetic fields and at temperatures below about 40 K.¹² In Dy₂CuMnMn₄O₁₂, magnetic susceptibilities decreased at low temperatures. This fact suggests different magnetic structures for these two compounds. The magnetic susceptibility behavior of Dy₂CuMnMn₄O₁₂ was consistent with the determined magnetic structure, where the Dy1 and Dy2 sublattices order antiferromagnetically relative to each other but with different magnitudes, and the resulting uncompensated moment on the Dy sublattices aligns antiferromagnetically relative to the FM Mn sublattices.¹²

Isothermal magnetization measurements (Figure 7) showed a behavior typical for soft ferrimagnets with a coercive field, H_C, of about 200 Oe at T = 5 K and a large saturation magnetization, M_S, of about 25.6μ_B (at T = 5 K and H = 70 kOe). The near saturation value was already observed from about H = 20 kOe. This saturation value is smaller than the full magnetization of 35μ_B expected for a full FM alignment (using the maximum value of 10 μ_B for Dy³⁺). On the other hand, Dy³⁺ cations have a noticeable single-ion anisotropy; therefore, the full moment of Dy³⁺ cannot be reached in powder samples. Magnetic structures of related compounds, Dy₂CuMnMn₄O₁₂ and Y₂CuGaMn₄O₁₂, were investigated by neutron diffraction. It was found that the B-site Mn sublattices are ordered ferromagnetically but with reduced magnetic moments (2.2μ_B per Mn in Y₂CuGaMn₄O₁₂ and 2.5μ_B per Mn⁴⁺ and 3.3μ_B per Mn³⁺ in Dy₂CuMnMn₄O₁₂). Considering uncertainties in ordered moments on the Dy and Mn sublattices, the maximum contribution of 1μ_B from the Cu sublattice of Dy₂CuZnMn₄O₁₂ can be neglected in the discussion below.

Figure 7.

M versus H curves of Dy₂CuZnMn₄O₁₂ at 5, 25, 60, and 100 K (f.u.: formula unit). The left-hand inset compares M versus H curves of Dy₂CuZnMn₄O₁₂ and Lu₂CuZnMn₄O₁₂ at 5 K. The right-hand inset compares M versus H curves of Dy₂CuZnMn₄O₁₂ and Dy₂MnMnMn₄O₁₂¹⁰ at 5 K. M_S is the magnetization value at H = 70 kOe and T = 5 K.

M_S was about 9.5μ_B (at T = 5 K) in a related compound without magnetic rare-earth cations, Lu₂CuZnMn₄O₁₂ (the inset of Figure 7), given an average moment of 2.4μ_B per Mn, which is close to the experimentally determined values in the related compounds. Assuming a similar magnetic structure with similar ordered moments at the B sublattices in Dy₂CuZnMn₄O₁₂ and Lu₂CuZnMn₄O₁₂, the difference in the M_S values could be attributed to the Dy³⁺ sublattices. The difference was about 16μ_B, which was significantly larger than the maximum Dy³⁺ ordered moment. This fact suggests that both Dy³⁺ sublattices (Dy1 and Dy2) give FM contributions to the magnetic structure with an average value of 8μ_B. Therefore, M versus H curves of Dy₂CuZnMn₄O₁₂ (and comparison with the related compounds) suggest that the Mn1 and Mn2 sublattices are ordered ferromagnetically and the Dy1 and Dy2 sublattices are ordered ferromagnetically relative to each other and to the Mn sublattices.

4. Conclusions

In conclusion, we prepared a new member of the A-site columnar-ordered quadruple perovskites, Dy₂CuZnMn₄O₁₂, which crystallizes in the charge-order structure at room temperature with Pmmn symmetry. Using direct high-temperature structural studies, we could detect a charge-disorder transition above 425 K, while other methods could not detect such a transition. Triple A-site cation ordering was realized through a new combination of the A/A′/A″ cations (Dy³⁺/Cu²⁺/Zn²⁺). Dy₂CuZnMn₄O₁₂ exhibits a ferrimagnetic transition below 116 K with a large saturation magnetization and involvement of the Dy³⁺ sublattices at low temperatures.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.3c02835.

• Bond distances of Dy₂CuZnMn₄O₁₂ at 100 K, DSC curves, details of M vs H curves, M vs T curves, and differential magnetic susceptibility curves, ac magnetic susceptibility curves, and a photograph of the as-synthesized powder (PDF)

The author declares no competing financial interest.