Coexistence of Spin Canting and Metamagnetism in a One-Dimensional Mn(II) Compound Bridged by Alternating Double End-to-End and Double End-On Azido Ligands and the Analog Co(II) Compound ^†

Abstract

Two new compounds of general formula [M(N₃)₂(dmbpy)] in which dmbpy = 5,5′-dimethyl-2,2′-bipyridine, and M = Mn(II) or Co(II), have been solvothermally synthesized and characterized structurally and magnetically. The structures consist of zig-zag polymeric chains with alternating bis-µ(azide-N1)₂M and bis-µ(azide-N1,N3)₂M units in which the cis-octahedrally based coordination geometry is completed by the N,N’-chelating ligand dmbpy. The molecular structures are basically the same for each metal. The Mn(II) compound has a slightly different packing mode compared to the Co(II) compound, resulting from their different space groups. Interestingly, relatively weak interchain interactions are present in both compounds and this originates from π–π stacking between the dmbpy rings. The magnetic properties of both compounds have been investigated down to 2 K. The measurements indicate that the manganese compound shows spin-canted antiferromagnetic ordering with a Néel temperature of T_N = 3.4 K and further, a field-induced magnetic transition of metamagnetism at temperatures below the T_N. This finding affords the first example of an 1D Mn(II) compound with alternating double end-on (EO) and double end-to-end (EE) azido-bridged ligands, showing the coexistence of spin canting and metamagnetism. The cobalt compound shows a weak ferromagnetism resulting from a spin-canted antiferromagnetism and long-range magnetic ordering with a critical temperature, T_C = 16.2 K.

1. Introduction

Coordination compounds with azide ligand have been studied for decades, not only for their potential use as detonation agents or explosives [1,2,3,4,5], but also because of the intrinsic properties of N₃⁻ as a “pseudo halogen” [6,7] and as a subject of magnetism [8]. Many studies of stable azide coordination compounds have been reported, and the Cambridge Structural Database (2020 release) contains over 5000 items having at least one coordinated azide ligand [9].

Azide ligands can bind monodentately to metal ions in an end-on mode (N1), or they can bridge between two or more metal ions. Bridging azide between two metals can take place using one terminal nitrogen only (i.e., N1,N1, known as end-on mode, abbreviated in this manuscript as EO), or two terminal nitrogens (i.e., N1,N3 or end-to-end mode, abbreviated as EE). In the bridging mode, dinuclear compounds can be formed, like in M(azide)₂M species [10], as depicted in Figure 1, but also polynuclear compounds, including polymeric linear or zig-zag species of formula ∙∙∙(azide)₂M(azide)₂M(azide)₂ ∙∙∙. Mixed species with both terminal (non-bridging) and bridging azide are known in the literature [8,11,12,13,14], but are not so common.

In earlier reports from our laboratories, we have given attention to metal–azide compounds as bridging ligands for dinuclear compounds, or higher aggregates up to 2D networks [15,16,17,18] and we have found both ferromagnetic and antiferromagnetic cases. The ferromagnetic cases are only known for the EO azido-bridged compounds. It appears that the choice of the non-bridging co-ligands plays a major role on the formation of specific compounds and structures, and also whether EO, EE or a combination is found. As the azido ligands are rather small or narrow, to generate relative stable coordination spheres around the metal, the co-ligands should be rather bulky, as shown from the literature examples above.

In the present paper, we report on two new compounds that have alternating modes (EE and EO) of bridging azide, forming zig-zag chains in the solid state. The compounds have interesting magnetic properties and these are studied in detail using magnetic susceptibility and magnetization studies at low-temperature (down to 2 K). We feel that the study is relevant in the search for potentially cheap, stable and useful new magnets.

2. Results and Discussion

2.1. Synthetic Efforts

Both title compounds were easily synthesized by the hydrothermal method and the synthesis was found to be reproducible. Despite several attempts, no pure crystalline materials could be isolated for the similar Fe(II) compound. In most attempts, two crystalline forms could be obtained, one isomorphous with the Mn(II) compound and the other isomorphous with the Co(II) compound. Therefore, this compound was not considered suitable for the study of the magnetic properties. In contrast, pure bulk samples for the Mn(II) and Co(II) compounds were available, as evidenced by powder diffraction data, which are consistent with patterns simulated from single-crystal data (Figure S1).

2.2. Characterization of the Compounds by IR and Elemental Analysis

Both compounds display very similar infrared spectra (Figure S2), where bands resembling the free dmbpy ligand are easily recognized. Most characteristic is the broad strong band (doublet) near 2087 cm⁻¹, typical for the coordinated bridged azide ligand [19,20,21]. The elemental analyses of both compounds are in full agreement with the values calculated, and with the 3D structures presented below, indicating that no secondary phases were obtained with the used synthetic methodology.

2.3. Structure Description of the Compounds

Single crystal structures were determined for both synthesized compounds (1 and 2) [M(N₃)₂(dmbpy)], with M = Mn(II) and Co(II) and dmbpy = 5,5’-dimethyl-2,2’-bipyridine (C₁₂H₁₂N₂). The Mn(II) compound (1) crystallizes in space group P-1, while the Co(II) compound (2) crystallizes in space group P2₁/c. Even though the space groups are different, both compounds share the same molecular structure (see Figure 2 for the Mn(II) compound). The metal center is coordinated by the bidentate dmbpy ligand and two azido anions. Since both pseudohalides, azido N₃⁻ ligands, bridge between symmetry-related metal centers in the crystal, a one-dimensional polymeric structure is formed, in which each azido anion has a different function: two anions N3/N4/N5 form a centrosymmetric double-bridge, with the μ-1,1 mode of coordination (EO). Two other anions, N6/N7/N8, also form a centrosymmetric double-bridge, but this time with the μ-1,3 coordination mode (EE).

Both coordination modes alternate along the zig-zag polymeric chains, which run parallel to the a-axis regardless of the crystal system, triclinic (M = Mn) or monoclinic (M = Co). Relevant coordination bond lengths and angles are given in

Table 1. Selected bond lengths (in Å) and angles (deg.) for both compounds.

Compounds	[Mn(N3)2(dmbpy)]	[Co(N3)2(dmbpy)]
M–N1	2.272(3)	2.115(5)
M–N2	2.256(4)	2.133(5)
M–N3	2.215(4)	2.144(5)
M–N3	2.222(3) (a)	2.198(6) (c)
M–N6	2.208(3)	2.140(6)
M–N8	2.263(3) (b)	2.204(6) (d)
N1–M–N2	72.26(11)	76.7(2)
N3–M–N6	99.50(15)	95.2(2)
N1–M–N3	95.59(12) (a)	94.1(2)
N1–M–N6	162.97(15)	170.6(2)
N2–M–N6	91.83(14)	94.3(2)
N3–M–N3	79.08(12) (a)	78.6(2) (c)
N6–M–N8	90.56(12) (b)	88.8(2) (d)
N2–M–N3	168.67(12)	167.5(2)

Symmetry for the last N atom: ^(a) 1 − x, 1 − y, 1 − z; ^(b) 2 − x, 1 − y, 1 − z; ^(c) −x, −y, 1 − z; ^(d) 1 − x, −y, 1 − z.

. The trend of the well-known smaller ionic radii going from Mn(II) to Co(II) is clearly visible.

Along the chain, long and short M∙∙∙M separations alternate, 5.430(1) and 3.422(1) Å for M = Mn and 3.361(2) and 5.317(2) Å for M = Co. These distances are slightly shorter for the Co(II) compound as a result of the smaller ionic radii of the cations. The polynuclear zig-zag chains are packed efficiently in the crystal. With M = Mn, chains are parallel, and the dmbpy ligands of two neighboring chains are stacked in such a way that they interact with a rather short distance, 3.440 Å. For M = Co, the arrangement of chains in the crystal is slightly modified, because of the monoclinic cell symmetry. However, the relative position of two neighboring chains is essentially preserved, and the π–π interactions between the dmbpy ligands are even strengthened, with separations between ligand mean planes of 3.283 Å (Co). In Figure 3, both different packings are depicted.

In [Mn(N₃)₂(dmbpy)], the N–N bond lengths are symmetrical in the EE azido ligands but are asymmetric in the EO ligands. The bite angle exhibited by the dmbpy ligand (72.26(11)°) is the largest distortion in the geometry of the cis-octahedral coordination sphere. The four-membered Mn₂(EO-N₃)₂ ring is planar as a result of the inversion center. The eight-membered Mn₂(EE-N₃)₂ ring adopts a chair configuration. The dihedral angle δ, defined by the N6/Mn1/N8 plane and the (EE-N₃)₂ plane, is 9.69(3)° and Mn1 sits 0.266(7) Å out of the (EE-N₃)₂ plane. The Mn–azido–Mn torsion angle τ, defined by the dihedral angle between the mean planes of Mn1-N6-N7-N8 and Mn1-N8-N7-N6, is 20.3(6)°. Similarly, in [Co(N₃)₂(dmbpy)], the N–N bond lengths are approximately symmetrical in the EE azido ligands but are asymmetric in the EO ligands. The bite angle exhibited by the dmbpy ligand (76.7(2)°) is the largest distortion in the geometry of the cis-octahedral coordination sphere. The four-membered Co₂(EO-N₃)₂ ring is planar as a result of the inversion center. The eight-membered Co₂(EE-N₃)₂ ring adopts a chair configuration. The dihedral angle δ, defined by the N6/Co1/N8 plane and the (EE-N₃)₂ plane, is 25.2(4)°. Co1 sits 0.66(1) Å out of the (EE-N₃)₂ plane. The Co–azido–Co torsion angle τ, defined by the dihedral angle between the mean planes of Co1-N6-N7-N8 and Co1-N8-N7-N6, is 47.7(6)°.

The azide ligand is well known for its versatility in coordination behavior, as explained in the introduction. When involved in metal-to-metal bridges, coordination modes EO and EE are frequent; however, the EO mode is roughly ten times more common than the EE mode, based on a survey of the Cambridge Structural Database [9]. Toggling EO and EE modes along a 1D polymeric structure is rare, but not unprecedented (see Table S1). Indeed, quite similar azido-bridged structures have been described using other ancillary ligands and a variety of transition metals: Schiff bases and Mn(II) [22,23] pyridine derivatives and Mn(II) [24,25], Co(II), Ni(II) [14] or Zn(II) [21] amine/pyridine derivatives and Ni(II) [11,26,27], among others. The nearest structurally related compound is certainly [Mn(N₃)₂(2,2′-bipyridine)], which crystallizes in space group P-1, with unit–cell parameters quite close to those of [Mn(N₃)₂(dmbpy)] [13,28] and in fact is even isostructural with the Fe(II) and Co(II) analogues [29].

Compounds based on azido-EE/EO double bridges are of interest in the field of magnetochemistry, because the type of interaction between magnetic centers is not unexpected. Therefore, it was decided to perform a detailed magnetic analysis down to very low temperatures. The results are described in the next section.

2.4. Magnetic Properties

For compounds with two different azide binding modes, one can expect interesting magnetic properties, as metal–metal exchange can occur via two different pathways. It is clear that the EE bridges, with their large metal∙∙∙metal separations, will always be involved in antiferromagnetic coupling (AF) [8]. The AF interaction is increased if the eight-membered metallacycle M(N₃)₂M’ is essentially planar. In contrast, EO bridges may promote ferromagnetism (F), provided the M–N–M angle is less than 108° [8]. Regarding the AF component, the triclinic compound (M = Mn) is expected to display a similar or even greater AF interaction than the monoclinic compounds (M = Co), since the former has an EE bridge almost flat, while the bridges in the latter have a butterfly conformation: the dihedral angle δ between the (N₃)₂ mean plane and the plane formed by M and the bonded N_azido atoms is δ = 9.69(3)° for M = Mn, and δ = 25.2(4)° for M = Co. Regarding the F component, both compounds fulfil the requirement for potentially having J > 0. The observed angles at EO bridges are 100.92(12) and 101.4(2)° for M = Mn, Co, respectively.

On the other hand, the six-coordinated metal ions used in this work are located in a slightly distorted octahedral ligand field, and their electronic configurations allow for either a low, or a high-spin state, depending on the crystal field splitting Δ_o for 3d orbitals. In the present case, the spectrochemical series indicates that N₃⁻ is a relatively weak ligand, while dmbpy is a relatively strong ligand. There is thus a competition between π-donation and π-backdonation in the coordination sphere, which makes the ground spin state not straightforward to anticipate. However, crystal structures are helpful in this regard, because Δ_o is also related to the strength of any Jahn–Teller (JT) effect in a crystal field with octahedral symmetry. Assuming a small tetragonal distortion, a symmetry measure accounting for the octahedral character of the field can be computed as $S\left(O_h\right)=5.39{\mathsf{\Delta }}^2-0.33\left|\mathsf{\Delta }\right|$, where Δ is the difference between the largest and the shortest coordination bond lengths. Structures with $S\left(O_h\right)<4.42$ are closer to the octahedron than to the trigonal prism geometry [30].

For the Mn(II) compound, $S_{Mn}\left(O_h\right)~{10}^{-3}$, reflecting a very small departure from the ideal $O_h$ symmetry, which, in turn, confirms that the JT distortion indeed is not present, consistent with a high-spin configuration. For the Co(II) compound, the JT distortion is much more noticeable, with $S_{Co}\left(O_h\right)~{13\times 10}^{-3}$, in line with weak JT distortions, as expected for high-spin $d^7$ ions; (in particular, a low-spin configuration for Co²⁺, with an odd number of electrons in the $e_g$ orbitals would give rise to a strong JT effect, associated to a symmetry measure of the field $S\left(O_h\right)\gg {10}^{-2}$).

It is evident that the crystal structures for the newly synthesized compounds allow one to predict the ground spin state for each one, S = 5/2 and S = 3/2 for M = Mn(II) and Co(II), respectively. On the other hand, both antiferromagnetic and ferromagnetic interactions should alternate along the 1D chains. Structural features obtained from X-ray structures are, however, not enough to confidently assess the balance between F and AF interactions in these materials, and a comprehensive experimental study of magnetic susceptibility was thus warranted.

Mn Compound (1). The temperature dependences of χ_M and χ_MT are depicted in Figure 4. At 300 K, the χ_MT value per Mn(II) ion of compound 1 is 3.80 cm³ mol⁻¹ K, which is lower than the spin-only value of 4.38 cm³ mol⁻¹ K expected for a magnetically isolated octahedral high-spin Mn(II) ion with g = 2.00. Upon cooling, the χ_MT values decrease gradually and show the cusp around 3.5 K, with a χ_MT value of 0.317 cm³ mol⁻¹ K at 3.5 K, decreasing to a value of 0.194 cm³ mol⁻¹ K at 2.0 K. The monotonic χ_MT decrease at a high temperature is indicative of the existence of antiferromagnetic coupling. Upon cooling, the χ_M increases smoothly from 0.013 cm³ mol⁻¹ at 300 K to reach a plateau of 0.035 cm³ mol⁻¹ at about 14.6 K, and then increases rapidly, reaching a maximum of 0.098 cm³ mol⁻¹ at 2.5 K, before a slight decrease to a value of 0.096 cm³ mol⁻¹ at 2.0 K. The sharp increase in χ_M at a low temperature is likely due to a small amount of paramagnetic impurities, e.g., at crystal edges and vacancies along the plane, as is known for related cases [31,32], or perhaps to some features of uncompensated spin. The temperature dependence of 1/χ_M at temperatures above 110 K can be fitted by the Curie–Weiss law with C = 4.64(4) cm³ mol⁻¹ K and θ = −64.7(6) K (Figure S3). The negative Weiss constant suggests the presence of overall antiferromagnetic coupling between the adjacent Mn(II) ions.

To try to obtain intrachain magnetic couplings between Mn(II) ions through the double EO–N₃ and double EE–N₃ bridges, the magnetic susceptibility of compound 1 was fitted using the expression proposed by Cortés [13,28], for alternating chains of classical spins on the Hamiltonian H = −J₁ΣS_2iS_2i+1 − J₂ΣS_2i+1S_2i+2.

χ_M = [Ng²β²S(S + 1)/3kT][(1 + u₁ + u₂ + u₁u₂)/(1 − u₁u₂)]

(1)

where u_i = coth[J_iS(S + 1)/kT] – kT/[J_iS(S + 1)] (i = 1 and 2) with S = 5/2, J₁ and J₂ are the F and AF exchange constants through double EO–N₃ and double EE–N₃ superexchange pathways, respectively.

The data above 20 K were fitted and the best fit to experimental data led to J₁= 2.87(7) cm⁻¹, and J₂ = −11.3(9) cm⁻¹ with a fixed g = 2.0. The J₁ and J₂ parameters are consistent with the reported results for related compounds. As in the earlier theoretical and experimental studies, the magnetic couplings between Mn(II) ions via the double EO–N₃ bridge were usually ferromagnetic due to small Mn–N–Mn angles which would lead the orthogonality of the magnetic orbitals of the adjacent Mn(II) centers. In contrast, the antiferromagnetic couplings were usually dominated via the EE–N₃ bridge due to the well overlap of magnetic orbitals of the adjacent Mn(II) centers. The ferromagnetic interaction decreases with the increasing Mn–N_azide–Mn angle, while the antferromagnetic couplings via double EE–N₃ bridges were dependent on the δ angle, the dihedral angle between the N_azide–Mn–N_azide plane and the plane defined by the two azido bridges, the magnetic coupling decreases with the increasing δ angles [13]. The obtained magnetic coupling of two Mn(II) ions is ferromagnetic for a double EO–N₃ bridging with a small Mn–N_azide–Mn angle of 100.92(12)°, and is antiferromagnetic for a double EE–N₃ bridging with a small δ angle, 9.69(3)°, which are consistent with the reported results for related compounds [13,22,23,33,34,35,36,37,38]. In Table S1, a detailed overview of the literature is given with structural details and J values for a variety of Mn compounds, including the present two new compounds.

Furthermore, the temperature dependences of χ_MT for compound 1 under 100 Oe were also collected (see Figure S4), showing similar behavior to that under 1000 Oe except for that in the low-temperature range, in which the χ_MT value slightly increases with decreasing temperature below 6.0 K to a maximum of 0.381 cm³ mol⁻¹ K at 4.5 K and then sharply decreases to 0.076 cm³ mol⁻¹ K at 2.0 K. This suggests that a possible mechanism involving weak ferromagnetic correlations is operative within compound 1 below 6.0 K and the final decrease may be due to antiferromagnetic interactions between the chains and/or saturation effects. These weak ferromagnetic correlations can be attributed to spin canting, i.e., the antiferromagnetically coupled local spins within the –Mn–(EE-N₃)₂–Mn–EO-N₃)₂–chains are not perfectly antiparallel, but are canted with respect to each other, resulting in uncompensated residual spins [13,22,23,33,34,35,36,37].

To further characterize the low-temperature magnetic behavior of compound 1, ZFC/FC magnetization measurements under a field of 10 Oe were carried out. As shown in Figure S5, the ZFC/FC magnetizations were found to be non-bifurcated and show a sharp maximum at 3.4 K, suggesting the occurrence of antiferromagnetic ordering. The temperature dependence of the AC susceptibility of compound 1 was also measured at H_dc = 0 Oe and H_ac = 3.5 Oe at different frequencies (Figure S6), which shows the sharp frequency-independent value of the χ_M′ signals with the peak maximum at 3.4 K. The absence of χ_M″ signals confirms the onset of antiferromagnetic ordering with a conventional T_N = 3.4 K and implies the existence of a magnetic phase transition.

The isothermal field dependence data of the magnetization of compound 1 was collected at 1.8 K (Figure 5), in which the magnetization shows a sigmoidal shape with an abrupt increase at a field above ~0.3 kOe to reach a value of 0.36 Nβ at 70 kOe. This sigmoidal magnetization clearly indicates a field-induced magnetic transition of metamagnetic nature [39,40,41]. In this metamagnetic transition, the net moments of spin-canting Mn–N₃ chains aligned antiparallel under a weak applied field by weak interchain antiferromagnetic interactions are overcome by a stronger external field and result in the state transition from antiferromagnetic (AF) to paramagnetic (P). The critical field of magnetic transition, H_C, at 1.8 K, was estimated to be about 0.48 kOe as determined by dM/dH (Figure 5, inset). The M value of 0.36 Nβ at 70 kOe is far below the expected saturation value of 5.0 Nβ for an isotropic high-spin Mn(II) system, confirming the antiferromagnetic nature of 1. Moreover, at 1.8 K, a small butterfly-shaped magnetic hysteresis loop was obtained, indicating a soft magnetic behavior (see Figure S7). The spin canting angle was estimated to be about α = 0.30°, based on the equation sin(α) = M_R/M_S (M_R = 0.026 Nβ; obtained by extrapolating the high-field linear part of the magnetization curve at 1.8 K to zero field, and M_S = 5.0 Nβ) [42,43,44,45].

Generally speaking, spin canting can arise from two contributions: (i) the presence of an antisymmetric exchange Dzyaloshinsky−Moriya interaction [46] and (ii) the existence of single-ion magnetic anisotropy [47,48,49,50,51,52]. The presence of an inversion center between adjacent spin centers can result in the disappearance of the antisymmetric exchange. Hence, the lack of antisymmetric exchange in compound 1 would be expected, due to the existence of a crystallographic inversion center between the Mn(II) ions (see symmetry codes in caption of Figure 2). However, the Mn1 center in 1 displays a distorted octahedral geometry due to the small bite angle of 72.26(11)° of the dmbpy chelating ligand. Such distorted octahedral geometries were reported as originating from the weak single-ion anisotropy of the high-spin Mn(II) site [53,54,55,56,57]. Therefore, it is assumed that the spin-canted antiferromagnetism in 1 can be attributed to weak single-ion magnetic anisotropy by a distorted metal coordination environment at low temperature. Similar spin canting behavior has been observed in a few other Mn(II) compounds containing chains of alternating double EO and double EE bridging modes of azides [34,58].

The field-induced magnetic phase transition for compound 1 was further investigated by the measurements of various fields of the FC magnetic susceptibilities, χ_M(T), and the field dependence of the magnetizations, M(H), at different temperatures. As shown in Figure S8, the maximum of χ_M(T) shifts to lower temperatures with increasing applied field, until the χ_M(T) reaches a plateau at a field larger than 600 Oe, confirming that the weak interchain antiferromagnetic interaction is overcome by a stronger external field. As shown in Figure S9, at 2.0 K, the stepwise M(H) curve clearly indicates a field-induced magnetic transition from AF to P. This stepwise magnetization becomes less pronounced with increasing temperature, and the differentials of these curves show peaks that shift to lower fields with increasing temperature (see Figure S10), indicating the phase transitions of metamagnetism.

Combining M(H), FCM and the frequency-independent χ_M′ data, the magnetic phase (T, H) diagram has been plotted in Figure 6. The value of H_C decreases with increasing temperature and finally disappears at about 3.4 K. The solid line of H_C (T) in Figure 6, on an analysis of the M–H curves, signifies a typical magnetic transition from AF to P corresponding to metamagnetic materials.

Co compound (2). The temperature dependences of χ_M and χ_MT of compound 2 are shown in Figure S11 and Figure 7, respectively. As the temperature decreases from 300 K, the χ_M value increases smoothly, reaching a rounded maximum of 0.021 cm³ mol⁻¹ at about 70 K, and then decreases slightly reaching a value of 0.015 cm³ mol⁻¹ at 20 K. Upon further cooling, the χ_M value increases rapidly to a sharp maximum of 0.017 cm³ mol⁻¹ at 12.8 K, after slightly decreasing, χ_M value increases again to 0.017 cm³ mol⁻¹ at 2.0 K. The temperature dependence of 1/χ_M at temperatures above 100 K has been fitted by the Curie–Weiss law with a Curie constant C = 4.85 cm³ mol⁻¹ K and a Weiss constant θ = −130.5 K (see Figure S12).

The large negative Weiss constant suggests the presence of strong spin-orbital coupling and/or overall antiferromagnetic interactions between the adjacent Co(II) ions. As shown in Figure 7, at 300 K, the χ_MT value per Co(II) of compound 2 is 3.38 cm³ mol⁻¹ K, which is larger than the spin-only value of 1.87 cm³ mol⁻¹ K for a magnetically isolated octahedral Co(II) ion (S = ³/₂), with g = 2.00. Upon cooling, the value of χ_MT decreases monotonically to attain a local minimum value of 0.299 cm³ mol⁻¹ K at 18.2 K, which is indicative of the existence of antiferromagnetic coupling. After a very small increase to a maximum value of 0.277 cm³ mol⁻¹ K at 14.6 K, the χ_MT value decreases again with further cooling to 2.0 K. The increase in χ_MT below 18.2 K is field-dependent, as shown in the inset of Figure 7; this suggests that a mechanism of weak ferromagnetic correlations due to spin canting antiferromagnetism is operative within compound 2. The final decrease in χ_MT value may be attributed to antiferromagnetic interactions between the chain and/or saturation effects. Similar to the observations in compound 1, the lack of antisymmetric magnetic interactions in compound 2 would be expected because of the presence of inversion centers in the P2₁/c crystal structure. Thus, the spin canting of compound 2 originates from the single-ion anisotropy of the Co(II) ion, which is in agreement with the reported Co(II) spin canting compounds containing the same bridging mode of azide [59,60,61].

In order to substantiate the low-temperature magnetic properties of compound 2, ZFC/FC magnetization studies were carried out at 50 Oe. As shown in Figure 8, upon cooling, both ZFC and FC magnetizations increase abruptly at temperatures below 18 K and a divergence between ZFC/FC below 16.2 K is observed, suggesting the occurrence of magnetic ordering for the formation of an ordered state and the existence of an uncompensated moment below the critical temperature of Tc = 16.2 K. Upon cooling, both ZFC and FC magnetizations increase again below 5.0 K, which may be due to the spin-reorientations of the domain wall. The existence of magnetic ordering was also confirmed by the AC magnetic susceptibility measurements of compound 2 performed at H_dc = 0 Oe and H_ac = 3.5 Oe at different frequencies (Figure S13). As can be seen from Figure S13, both χ_M′ and χ_M″ signals are frequency-independent, where the χ_M′ signals show two peaks at ca. 14.8 K and 5.2 K with two corresponding non-zero χ_M″ signals formed at temperatures below 16.2 and 6.2 K. The presence of χ_M′ and χ_M″ peaks at about 15 K are the result of the formation of an ordered state with an uncompensated moment and the peaks of χ_M′ and χ_M″ at about 5 K may be caused by spin-reforestation [62,63]. These data confirm the occurrence of magnetic ordering by weak ferromagnetism due to spin canting, which is consistent with the obtained results from ZFC/FC magnetizations data. Due to the presence of weak non-zero χ_M” signals, a coercive magnetic behavior would be expected below T_c.

To further study the magnetic ordering of compound 2, the isothermal field-dependent magnetization was collected at 2.0 K. As depicted in Figure S14, the initial increase in magnetization shows a positive curvature to 0.175 Nβ at 70 kOe, which is far below the theoretical value of saturation for an isotropic high-spin Co(II) system, and the absence of saturation of magnetization; this is an indication of an overall antiferromagnetic interaction between the Co(II) ions in compound 2. In addition, when the field is less than 10 kOe, a hysteresis loop is clearly observed at 2.0 K, suggesting the soft magnet property of compound 2 (Figure S14, inset). The hysteresis loop shows a remanent magnetization (M_r) of ≈0.007 Nβ and a coercive field of ≈350 Oe. Based on the value of M_r at 2.0 K, the canting angle of compound 2 is estimated to be approximately 0.20°, where M_S is 2.15 Nβ for an octahedral Co(II) at 2 K with the effective spin of S’ = ¹/₂ and a common value of g’ = 4.3 [60].

Finally, to rule out any contributions of Mn(II)/Mn(III) or Co(II) oxides in the low-temperature magnetic behaviors in compounds 1 and 2, the AC magnetic data and 2.0 K field-dependent magnetization were collected using the thermal decomposed samples of 1 and 2 that had been heated at 350 °C for two hours (Figures S15–S18), in which the disappearance of peaks in χ_M′ and/or χ_M″ and the absence of magnetic hysteresis loops in field-dependent magnetization are obtained, excluding the contribution of the behavior of Mn(II)/Mn(III) or Co(II) oxides.

3. Concluding Remarks

The results presented and discussed above have shown that the new compounds of general formula [M(N₃)₂(dmbpy)] for M = Mn (1) and Co (2) show very similar and almost identical linear chain structures, with alternating azide double bridging anions (EE and EO). Using magnetic studies down to a very low T (2K), the existence of intrachain ferro- and antiferromagnetic interactions was established, and these interactions are dominated by the double EE and double EO azido ligand bridges. Overall, the compounds are found to behave as antiferromagnets, and the study is relevant in the search for potentially cheap, stable and useful new magnets.

Compounds 1·(Mn) and 2·(Co) exhibit spin-canted antiferromagnetism at very low-temperatures, which is ascribed to the presence of single-ion anisotropy. Furthermore, below the Néel temperature, T_N, field-induced magnetic transitions have also been observed, and these are indicative for metamagnetism in the case of 1·(Mn). Such coexistence of spin-canted antiferromagnetism and metamagnetism in 1D Mn(II) compounds with alternating double EE and double EO azido ligands is unprecedented. Although the spin-canted antiferromagnetism might seem incompatible with the crystal structure of 2, because of the presence of an inversion center between the bridged Mn(II) centers, the observation of the spin canting in 2 should be attributed to a structural phase transition, or a distortion in the crystal at low temperature, thereby removing the inversion center. Such types of distortions have been reported before [34,58]. The weak interchain interactions present in both compounds are ascribed to the π–π stacking interactions between the dmbpy rings.

4. Material and Methods

4.1. General Remarks

The starting materials (metal salts, sodium azide and the ligand 5,5’-dimethyl-2,2’-bipyridine, C₁₂H₁₂N₂), and used solvents were purchased from commercial sources (analytical reagent grade) and used without further purification. All the compounds were synthesized solvothermally under autogenous pressure. Azido derivatives are potentially explosive and should be handled with great care and prepared only in small quantities by trained persons.

4.2. Synthesis

Synthesis of catena-poly-[Mn(N₃)₂(dmbpy)] (1) A mixture of Mn(NO₃)₂·4H₂O (50 mg, 0.2 mmol), dmbpy (37 mg, 0.2 mmol) and NaN₃ (26 mg, 0.4 mmol) in H₂O/EtOH (3:1 v/v, 20 mL) was sealed in a Teflon-lined autoclave and heated at 130 °C for 2 days. After cooling to room temperature at a rate of 10 °C h⁻¹, yellow-colored crystals of 1 were obtained (yield 32%). Anal. Calcd. (%) for C₁₂H₁₂MnN₈: C, 44.59; H, 3.74; N, 34.67%. Found: C, 44.45; H, 3.92; N, 34.42%. Main IR band (KBr pellet, cm⁻¹): doublet at 2089s [ν(N₃⁻)].

Synthesis of catena-poly-[Co(N₃)₂(dmbpy)] (2). This compound was prepared following a procedure similar to that of compound 1, except that Co(NO₃)₂·6H₂O (58 mg, 0.2 mmol), was used instead of Mn(NO₃)₂·4H₂O. Brown-colored crystals were obtained, with a yield of 40%, containing 2. Anal. Calcd. (%) for C₁₂H₁₂CoN₈: C, 44.05; H, 3.70; N, 34.24%. Found: C, 43.85; H, 3.82; N, 34.63%. Main IR band (KBr pellet, cm⁻¹): doublet at 2085s [ν(N₃⁻)].

Physical Measurements. Elemental analyses of the obtained compounds (C, H and N) were performed using a Perkin–Elmer 2400 series II CHN analyzer. Infrared spectra were recorded as KBr pellets in the range 4000–400 cm⁻¹ (4 cm⁻¹ resolution) on a Perkin–Elmer 100 FT-IR spectrometer, which was calibrated using polystyrene and CO₂ bands. The temperature dependence DC and AC magnetic susceptibility measurements were performed on powdered samples, restrained in eicosane to prevent torquing, on a Quantum Design MPMS-7 SQUID (Superconducting Quantum Interference Device) and a PPMS (Physical Property Measurement System) magnetometer, equipped with 7.0 T and 9.0 T magnets (Quantum Design, San Diego, CA, USA), respectively, operated in the range of 2.0–300 K. Diamagnetic corrections were estimated from Pascal’s constants [64] and subtracted from the experimental susceptibility data to obtain the molar paramagnetic susceptibility of the compounds. Powder X-ray diffraction (PXRD) measurements of 1 and 2 were carried out on a Siemens D-5000 diffractometer (Siemens, Karlsruhe, Germany) running in a step mode with a step size of 0.02° in θ and a fixed time of 10 s at 40 kV, 30 mA for Cu-Kα (λ = 1.5406 Å).

4.3. X-ray Crystallography

Diffraction data (

Table 2. Crystal data and refinement parameters.

Compound/Deposition in CCDC	1/1954571	2/1954572
Formula	C12H12MnN8	C12H12CoN8
Fw	323.24	327.23
Crystal size (mm3)	0.5 × 0.3 × 0.2	0.30 × 0.30 × 0.18
Space group	P-1	P21/c
a (Å)	7.8058(18)	7.397(2)
b (Å)	9.538(3)	18.373(4)
c (Å)	10.601(3)	10.661(3)
α (°)	113.630(8)	-
β (°)	102.610(8)	110.571(9)
γ (°)	93.223(8)	-
V (Å3)	696.4(3)	1356.5(6)
Z, Z’	2, 1	4, 1
Diffractometer	Bruker X2S	Bruker X2S
Radiation	Mo-Kα	Mo-Kα
T (K)	200	200
Abs. coef. (mm−1)	0.954	1.272
Transmission fact.	0.56–0.83	0.51–0.80
Refl. collected	6450	8922
Sinθ/λ (Å−1)	0.62	0.61
Rint (%)	4.26	8.68
Completeness (%)	97.4	99.2
Data/parameters	2658/192	2575/194
Restraints	0	0
R1, wR2 [I > 2σ(I)]	5.87, 16.64	6.53, 16.48
R1, wR2 [all data]	7.07, 17.60	8.26, 17.88
GOF on F2	1.035	1.003

) were collected at 200 K on a Bruker SMART X2S (Bruker AXS Inc., Madison, WI, USA) benchtop diffractometer [65], using the Mo Kα radiation (λ = 0.71073 Å) and the structures were refined with SHELXL [66,67]. Crystals for M = Co(II) were found to be twinned by a twofold rotation about the c* reciprocal axis. One batch scale factor was refined, which converged to 0.24. All H atoms were placed in calculated positions and refined as riding to their carrier atoms.