Four 3D coordination polymers based on layers with single syn–anti carboxylate bridges: synthesis, structures, and magnetic properties

Four novel coordination polymers (CPs) based on a new 4-(3,5-dicarboxylphenyl) picolinic acid ligands (H3L), [M3(L)2(H2O)6]·4H2O (M3 = Mn3, 1; Co3, 2; Ni3, 3, Co1.01Ni1.99, 4), have been hydrothermally synthesized, and structurally and magnetically characterized. In these isomorphous CPs, octahedrally coordinated metal ions are linked by the single syn–anti carboxylate bridge (μ-COO) to give linear trinuclear motifs. The motifs are connected through the other single syn–anti carboxylate bridge (μ-COO) to give a 2D (4,4) layer, and the layers are interlinked by the L ligands into 3D frameworks. Magnetic measurement indicates that antiferromagnetic interactions between metal ions are mediated through the single syn–anti carboxylate bridges in 1 and 2, while the same carboxylate bridges in 3 transmit ferromagnetic couplings. The bimetallic CP 4 shows interesting complicated magnetic behaviors due to the competition effect of Co(ii) and Ni(ii) ions.


Introduction
Research efforts on metal-organic frameworks (MOFs) or coordination polymers (CPs) have disclosed great versatile multifunctional materials with fascinating structures and promising properties. 1-3 Such multifunctional materials have displayed excellent potential applications in the eld of gas storage, 4,5 catalysis, 6,7 uorescence, 8,9 and magnetism. 10,11 The general strategy to design CPs is to use organic ligands to connect metal ions or clusters into polynuclear clusters or polymeric networks. The choice of metal ions and organic linkers with specic coordination preferences is crucial to construct CPs through serendipitous synthesis or rational design. In this context, multi carboxylate ligands have been among the most extensively used ligands owing to their versatile and diverse in both coordination chemistry and magnetism. [12][13][14][15] For instance, the carboxylate groups can bind two or more metal ions in various bridging modes, such as syn-syn, syn-anti, and anti-anti. In addition, they can form strong hydrogen bonds in favor of the formation of extended supramolecular structures and reinforcement the coordination networks. More importantly, they can efficiently transmit ferromagnetic (FM) or antiferromagnetic (AFM) coupling, depending on the bridging mode and the metal ion. Among them, the pyridyl-polycarboxylate ligands, such as pyridyldi and tricarboxylates, have been largely explored, obtaining a variety of MOFs with fascinating structures and properties. 14,15b,c For instance, Zhao et al. 14c have synthesized a UiO type MOF based on 2,2 0 -bipyridine-5,5 0 -dicarboxylate (bpdc) ligand. This MOF exhibits high storage capacities for H 2 , CH 4 and CO 2 . In the previous paper, we described the series of Zn(II), Co(II), Ni(II), Mn(II) 15b and Ln(III)-CPs 15c based on 4-(3,5-dicarboxylphenyl)-2methylpyridine (H 2 L) ligands. These CPs exhibit the various topologies with interesting magnetic properties and luminescence properties. Encouraged by above results, we decided to explore a new rigid pyridyl-tricarboxylates ligand, 4-(3,5-dicarboxylphenyl) picolinic acid (Scheme 1), which still has been unexplored. The p-conjugated rigid backbone and the tetratopic connectivity would lead to some new features in the coordination Scheme 1 Structures of 4-(3,5-dicarboxylphenyl) picolinic acid (H 3 L) ligand.
chemistry. Here we present the synthesis, structures and magnetic properties of four isomorphic CPs: [M 3 (L) 2 (H 2 O) 6 ]$ 4H 2 O [M 3 ¼ Mn 3 for 1; Co 3 for 2; Ni 3 for 3; Co 1.01 Ni 1.99 for 4; H 3 L ¼ 4-(3,5-dicarboxylphenyl) picolinic acid]. These CPs consist of 3D frameworks, in which 2D layers with single syn-anti carboxylate bridges are interlinked by the L spacers. CPs 1 and 2 display AFM couplings, while the FM interaction was found in CP 3 and the competition effect of FM and AFM interactions exists in CP 4.

Materials and physical measurements
All the starting chemicals including 4-(3,5-dicarboxylphenyl) picolinic acid (H 3 L) were used as received. The Fourier transform infrared spectra were recorded in the range 500-4000 cm À1 on a NEXUS 670 FT-IR spectrometer using the KBr pellets. Elemental analysis (EA) was determined on an Elementar Vario E1 III analyzer. Inductively coupled plasma (ICP) analysis was performed on a Optima 7300 DV. Powder X-ray diffraction (PXRD) was recorded on a Bruker D8-ADVANCE diffractometer equipped with Cu-target tube at a scan speed of 1 min À1 . Magnetic measurements were carried out on a Quantum Design SQUID MPMS-5 magnetometer. Diamagnetic corrections were made with Pascal's constants. Thermogravimetric analyses were performed using a Mettler Toledo TGA/SDTA851 instrument at a heating rate of 5 C min À1 under the N 2 atmosphere.

Crystal structure analysis
Diffraction intensity data for 2, 3 and 4 was collected at 298 K on a Bruker Apex II CCD area detector equipped with graphitemonochromated Mo Ka radiation (l ¼ 0.71073 A). Empirical absorption corrections were applied using the SADABS program. 16 The structures were solved by the direct method and rened by the full-matrix least-squares method on F 2 using SHELXTL-2014 program, with all non-hydrogen atoms rened with anisotropic thermal parameters. 17 All of the hydrogen atoms attached to carbon atoms were placed in calculated positions and rened using the riding model, and the water hydrogen atoms were located from the difference maps. However, the hydrogen atoms of the guest water molecules (O10, O11) in CP 3 and 4 cannot be located because of the diffraction data of limited quality. These hydrogen atoms have been included in the formulas. All calculations were carried out with the SHELXTL crystallographic soware. A summary of the crystallographic data, data collection, and renement parameters for compound 2, 3 and 4 is provided in Table 1. CCDC reference numbers 1811142-1811144.

Synthesis and FT-IR spectra
CPs 1-4 were all synthesized by the hydrothermal reaction of MCl 2 $nH 2 O (n ¼ 4 or 6) and H 3 L. CPs 2-4 were obtained in a Teon-lined stainless steel vessel under 100 C for 3 days. But the same procedure for Mn(II) yielded microcrystalline products. We explore different ratios of Co/Ni source compared to the nal ratios of Co/Ni in the framework. Only one pure ratios of Co/Ni, ca. 1.01 : 1.99, have been observed. These CPs exhibit similar characteristic asymmetric (n as ) and symmetric (n s ) absorptions of the carboxylate groups. The n as (COO) and n s (COO) vibrations appear as strong bands at about 1627 and 1450 cm À1 , respectively.

PXRD and TGA
The phase purity of the bulk materials of CPs 1-4 was conrmed by PXRD experiments (Fig. S1 †). The PXRD patterns of samples in CPs 2-4 are in good agreement with those simulated from experimental data. And the PXRD patterns of CP 1 is also in good agreement with those of CPs 2-4, suggesting 1 is isomorphous with 2-4. This is also conrmed by FT-IR spectra and elemental analyses. The thermal stability of CPs 1-4 was measured under an air atmosphere (Fig. S2 †)

Crystal structures
Single crystal X-ray structure analyses revealed that CPs 2, 3 and 4 are isomorphous, crystallizing in the P2 1 /c space group and exhibiting 3D frameworks in which M(II)-carboxylate layers are interlinked by the L ligands. The relevant parameters are summarized in Table S1. † The structure of 2 is described here in details. The asymmetric unit consists of one and a half of Co(II) ion, one L ligand, three coordinated H 2 O molecules and two guest H 2 O molecules. As shown in Fig. 1a, there are two crystallographically independent Co(II) ions in 2 (Co1 and Co2). Co1 assumes a distorted octahedral [NO 5 ] geometry ligated by one pyridyl nitrogen atom (N1A) and four carboxylate oxygen atoms (O1, O4E, O5A, and O6C), and a H 2 O molecule (O7). Co2 resides at an inversion center and is coordinated in the trans-octahedral geometry by four H 2 O molecules (O8, O8D, O9 and O9D) at the equatorial plane and two equivalent carboxylate oxygen atoms (O2 and O2D) at axial positions. The Co-N/O bond distances for both Co1 and Co2 fall in the range of 2.158(3)-2.013(3) A. The Co-O/N distances are comparable to those previous Co(II) CPs based on the similar pyridine-tricarboxylate ligands. 15d,e Adjacent Co1 and Co2 atoms are bridged by single m-syn, anticarboxylate bridges. Through such bridges, each Co2 is linked to two equivalent Co1 atoms to form a centrosymmetric linear trinuclear motif (Fig. 1b). In the trinuclear motif, the Co1/Co2 distance separated by the single m-syn, anti-carboxylate bridge are 5.372(9) A. It is notable that the trinuclear motif is reinforced by the hydrogen bonds between the coordinated H 2 O molecules (O7 and O8) with O8-H8C/O7 ¼ 168.5 (1) (1) A, giving a 2D layer parallel to the ab plane (Fig. 1b). Topologically, each trinuclear motif serves as a 4-connecting node, which reduces to the (4,4) topology. The layers are pillared into a 3D framework by the L ligands (Fig. 1c).
The L ligand serves as a hexadentate bridge ligand and each ligand binds ve Co(II) ions through its three carboxylate groups and one pyridyl nitrogen atom in different modes: the carboxylate attached to the pyridyl ring and one pyridyl nitrogen atom (N1) connects with two Co(II) ions, which induces a syn-anti coordination fashion for the carboxylate group, with the Co1F-O6-C14-O5 and Co1-O5-C14-O6 torsion angles being 53.7(5) and 170.6 (3) , respectively, and another carboxylate group also binds two Co(II) ions in a syn-anti fashion, while third carboxylate group connects with one Co(II) ion in monodentate fashion. The structures in 2 have two lattice water (O10, O11) per asymmetric unit that forms rich hydrogen bonds with carboxylate oxygen atoms and coordinated H 2 O molecules. The relevant hydrogen bonding parameters are given in Table S2. † Each lattice H 2 O molecule (O10, O11) donates its two hydrogen atoms to form two hydrogen bonds with carboxylate oxygen atoms (O1, O5, O3) and coordinated H 2 O molecule (O9), and each O10 and O11 atom also acts as a bifurcate hydrogen acceptor to interact with four different coordinated H 2 O molecules (O7, O8, O9, O9[Àx + 1, y + 1/2, Àz + 1/2]). Two tetrahedral geometries around O10 and O11 are formed by the four hydrogen bonds. Moreover, the coordinated H 2 O molecule (O8) uses its two hydrogen atoms to form one hydrogen bond with coordinated H 2 O molecule (O9). Five H 2 O molecules (O7, O8, O9, O10, O11) constitute a hydrogen-bonding ten-membered ring [graph set R 5 5 (10) 18 (Fig. S3 †)].

Magnetic properties
CPs 1-4. The magnetic susceptibility of 1-4 was performed on polycrystalline samples under 1 kOe in the range of 2-300 K (Fig. 2-5). For Mn(II) CP 1, the cT values per Mn 3 at 300 K is about 13.15 emu K mol À1 , comparable with the value expected for three magnetically isolated Mn(II) ions (13.13 emu K mol À1 ). Upon cooling, the cT value decreases continuously, which indicates the occurrence of antiferromagnetic interaction in this compound. However, the c value increases monotonically (Fig. 2a). The data follow the Curie-Weiss behavior above 70 K with C ¼ 13.93 emu K mol À1 and q ¼ À9.45 K. The negative q value further conrms the overall antiferromagnetic behaviors between Mn(II) ions. Due to the lack of an appropriate formula for such system, the J values (J is the interaction between Mn(II) ions) can not be quantitatively estimated. However, magnetostructural comparisons with previous compounds indicate that the m-syn, anti-carboxylate bridges induce antiferromagnetic interactions between Mn(II) ions. 19,20 In addition, the antiferromagnetic couplings is also conrmed by the isothermal magnetization measurement at 2 K (Fig. 2b). The magnetization curve is increases quasilinearly with the eld lied and the value (11.13 Nb) at 70 kOe is much lower than the saturation value. Co(II) CP 2. As is shown in Fig. 3a, the cT values per Co 3 at room temperature for this compound is about 9.72 emu K mol À1 , which is much larger than the spin-only value for three isolated uncorrelated Co(II) ions with S ¼ 3/2. However, this is typical of pseudo-octahedral Co(II) systems with a signicant contribution from unquenched orbital momentum. Upon cooling, the cT value decreases continuously, while the c value increases gradually, indicating the antiferromagnetic interaction between Co(II) ions. The c À1 vs. T curve follow the Curie-Weiss law above 140 K with C ¼ 10.20 emu K mol À1 , q ¼ À13.60 K. It is noted that the negative q value is not necessary for the antiferromagnetic interactions because of the effect of rstorder spin-orbital coupling in Co(II) system. In fact, the cT value of 2 is only 0.85 Â 3 emu K mol À1 per Co(II) at 2 K, which is very lower than that expected for the effective ground Kramer's doublet with a typical g av value of 4.3, indicative of antiferromagnetic couplings. 21 Moreover, isothermal magnetizations Fig. 1 The structure of CP 2: (a) the coordination environment of Co(II) ions; (b) the 2D (4,4) Co(II)-carboxylate layer and the intralayer hydrogen bonds; (c) the 3D structure. (Symmetry codes: (A) x, Ày + 1/2, z À 1/2; (B) Àx + 3, Ày + 1, Àz + 1; (C) Àx + 3, y + 1/2, Àz + 1/2; (D) Àx + 4, Ày + 1, Àz + 1; (E) x, Ày + 1/2, z + 1/2; (F) x, Ày + 1/2, z À 1/2). were also measured in the range of 0-70 kOe at 2 K. As is shown in Fig. 3b, the magnetization rises much more rapidly as the led was lied from zero and the magnetization value of 6.97 Nb at 70 kOe is far from saturation, which supports the dominant antiferromagnetic coupling. In spite of no appropriate molde to quantitatively analyze the magnetic coupling in this system, However, magneto-structural comparisons with the related reports about Co(II) compounds indicate that the m-synanti-carboxylate bridges induce antiferromagnetic interactions between Co(II) ions. This magnetic data results are in accordance with that in literature. 19a,22 Ni(II) CP 3. As is shown in Fig. 4a, the cT value per Ni 3 at room temperature for this compound is about 4.00 emu K mol À1 , which is much larger than the spin-only value (3.00 emu K mol À1 ) for three isolated Ni(II) ions with S ¼ 1 spins. Upon cooling, the cT value increases slowly up to 4.59 emu K mol À1 from 300 K to 20 K, and then the value rapidly increases to a maximum value of 10.06 emu K mol À1 at 2.0 K, which indicates that the ferromagnetic interaction is induced by single syn-anti-carboxylate bridges between Ni(II) ions. The c À1 vs. T curve follow the Curie-Weiss law above 10 K with C ¼ 3.94 emu K mol À1 , q ¼ 2.85 K. The Curie constant (C) is in accordance with the theoretical value of an S ¼ 1 free ion (C ¼ 1 emu K mol À1 for three isolated Ni(II)). The positive q value suggests ferromagnetic interactions between Ni(II) ions. The ferromagnetic coupling is also supported by the isothermal magnetization measurements at 2 K (Fig. 4b). The magnetization rst sharply increases and then slowly increases to 6.10 Nb at 70 kOe. The value (J) of 3 can not be given because of the lack of an appropriate model for such system. However, the magnetic couplings are consistent with the previous compounds with the syn-anti carboxylate bridge mediating ferromagnetic interactions. 22f,23 Nevertheless, magneto-structural studies on Ni(II) compounds with the syn-anti carboxylate bridges are scarce.
Bimetallic CP 4. The cT value at 300 K is about 5.9 emu K mol À1 , falling within the values of 2 (9.72 emu K mol À1 ) to 3 (4.00 emu K mol À1 ) (Fig. 5). 24 The value is higher than the spinonly values for three isolated high-spin Co(II) (5.64 emu mol À1 K) and Ni(II) (3.00 emu mol À1 K). However, it is consistent with the rst-order orbital moment contribution of Co(II) ions in the octahedral eld. Upon cooling, the c value increases continuously, while the cT value rst increases slightly to a round maximum at about 80 K and then slowly decreases to a minimum at 5.87 K, aer that it increases sharply to a maximum of 6.99 emu K mol À1 at 4.27 K, and nally drops to 5.85 emu K mol À1 at 2.0 K. The data follow the Curie-Weiss law above 10 K with C ¼ 5.90 emu K mol À1 and q ¼ 0.59 K. The increase of cT value in the range of 80-300 K suggests a ferromagnetic interaction in this structure. The low-temperature behavior of cT vs. T may be attributed to the cooperation or competition of several effects such as spin-orbital coupling and magnetic interactions between the different metal(II) ions. Isothermal measurements (Fig. 5) revealed that the molar magnetization rises continuously to 3.58 Nb at 50 kOe with increased eld, which is far from saturation, supporting the competitive effects of the magnetic coupling and the anisotropy. No hysteresis loops were observed at 2.0 K in CP 4 (Fig. 5  inset). The eld-cooled (FC) and zero-eld-cooled (ZFC) curves are measured at 20 Oe from 2 to 20 K (Fig. S4 †). The two curves are identical, indicating no occurrence of the phase transition or ordering. This is also conrmed by the thermal ac susceptibility curves (Fig. S5 †). The ac curves show the frequency independent behaviors. The real component (c 0 ) shows no maximum, meanwhile, no imaginary signal (c 00 ) is observed.

Discussion
Generally speaking, the anti-anti carboxylate bridges induces antiferromagnetic interactions between paramagnetic metal ions owing to the geometry of the magnetic orbital. Moreover, the double syn-syn carboxylate bridges also usually induce antiferromagnetic couplings, while the syn-anti carboxylate bridges may transmit weak antiferro-or ferromagnetic interactions. 25 Aer the literature research, we found that the single carboxylate bridges in the Mn(II) compounds always induce antiferromagnetic interactions. The magnetic coupling of CP 1 is antiferromagnetic, which is in accord with the previous compounds in literature. 19,20 However, only a few Mn(II) compounds shows the spin-canted ferromagnetic ordering above 2 K. In the context, many Co(II)-carboxylate compounds have been reported. However, the most of them are double carboxylate bridged. Thus, the Co(II) species with single syn-anti carboxylate bridges are still lacking. Moreover, the magnetic analysis for octahedral Co(II) is complicated due to its single-ion effects. Therefore, it is difficult to extract magnetostructural correlations. Some literatures survey suggest that single synanti carboxylate bridges between Co(II) indued antiferromagnetic couplings. In 2, the adjacent Co(II) ions bridged through the single syn-anti carboxylate bridges is a new examples in this series. Ni(II) CP 3 with the single syn-anti carboxylate bridges are much less than Co(II) and Mn(II) species with the similar bridges in the literature. Only few exceptions of this series are known in which single syn-anti carboxylate bridges all transmit ferromagnetic interactions between Ni(II) ions. The magnetic behaviors in bimetallic CP 4 depend mainly upon the random spin distribution from their parent CPs 2 and 3. Co(II) CP 2 show a typical antiferromagnetic coupling with antiparallel arrangement of spins, while Ni(II) CP 3 with parallel arrangement of spins show a typical ferromagnetic coupling. Thus, when the second spins, Ni(II) ions are incorporated into Co(II) system, the Co(II) ions would be substituted by Ni(II) ions. Therefore, in CP 4, there are two competing effects, including the decreasing anisotropy. Such effect is similar to that in previous bimetallic Co(II)-Ni(II) system. 24 Although CPs 1-4 are isostructural, magnetically, they exhibit the interesting different magnetic behaviors through the (syn-anti-COO) single bridge: 1 and 2 are antiferromagnetic, 3 is ferromagnetic and 4 is complicated magnetic behaviors. It is noted that so far, no CPs based on such M(II)-carboxylate layers have been reported. However, our results can be supported by previously reported M(II)carboxylate compounds (M ¼ Co(II), Mn(II), Ni(II)) with squaregrid layers.

Conclusions
In summary, we present four isomorphous CPs derived from a new pyridine-tricarboxylate ligands, in which adjacent M(II) ions are connected through single syn-anti carboxylate bridges into M(II)-carboxylate layers and then the layers are interlinked into 3D frameworks by the L spacers. The magnetic studies demonstrated that the single syn-anti carboxylate bridges transmit antiferromagnetic (AFM) interactions in the Mn(II) and Co(II) CPs while ferromagnetic (FM) in the Ni(II) species. The bimetallic Co(II)-Ni(II) exhibit the competition effect of FM and AFM interactions. Although the theory of composition-effects is unclear, the experimental ndings provide a good approach to improve magnetic behaviors by mixing metals. Thus, this work not only represents new structures based on M(II)-carboxylate layers, but also represents the interesting magnetic behaviors are related to the nature of the metal(II) ions.

Conflicts of interest
There are no conicts to declare.