Temperature-Controlled Assembly/Reassembly of Two Dicarboxylate-Based Three-Dimensional Co(II) Coordination Polymers with an Antiferromagnetic Metallic Layer and a Ferromagnetic Metallic Chain

Two new dicarboxylate-based three-dimensional cobalt coordination polymers, [Co(Me2mal)(bpe)0.5(H2O)]n (1) and [Co(Me2mal)(bpe)0.5]n (2), were synthesized from dimethylmalonic acid (H2-Me2mal) in temperature-controlled solvothermal reactions. Lower temperatures (60–80 °C) favored the formation of 1, while higher temperatures (120 °C) favored the production of 2. Compound 1 is comprised of Co(II) corrugated layers linked by syn–anti carboxylate bridges from the Me2mal2− ligands and pillared through bis-monodentate bpe groups. Compound 2 is comprised of a three-dimensional network involving one-dimensional Co–carboxylate chains bonded by antisymmetric µ4-Me2mal2− ligands and aligned parallel to the [001] direction. The solvothermal retreatment of crystalline samples of 1 in a DMF/H2O solvent at 120 °C allowed the structural reassembly, with complete conversion within 2 over 48 h. Magnetic analyses revealed that compound 1 exhibits both spin-orbital coupling and antiferromagnetic interactions through a syn–anti carboxylate (Me2mal2−) bridge exchange pathway [Co–Co separation of 5.478 Å] and compound 2 showed a ferromagnetic interaction resulting from the short Co–Co separation (3.150 Å) and the small Co–O–Co bridging angles (98.5° and 95.3°) exchange pathway which was provided by µ4-Me2mal2− bridging ligand.


Introduction
Coordination polymers (CPs), hybrid crystalline materials comprised of organic and inorganic components whose structures are extended by coordination bonds, have attracted considerable interest in the field of condensed matter [1][2][3][4][5][6][7][8][9][10]. The beneficial features of these materials are attributed to the building blocks from which they are constructed, which have of both organic and inorganic parts, thus conferring hybrid properties. Consequently, they commonly possess novel and fascinating properties or functionalities, which originate from the hybridization of the inorganic-organic parts. Although CPs can have many other potentials and fascinating properties, including heterogeneous catalysis, gas storage, gas separation, and drug carriers [11][12][13][14][15][16][17][18], magnetism is also an important area of interest [19][20][21][22][23][24][25][26][27][28][29]. This is particularly true, when the paramagnetic metal centers are bridged by short ligands (such as azido anion, cyanide, and carboxylate groups) to produce extended structures, 1D chains, and 2D layers, which is the structural basis for transmitting significant magnetic interactions between spin carriers and the metal ions [30][31][32]. One main benefit of magnetic coordination polymers (MCPs) is they provide the possibility and opportunity for tuning the nature of magnetic interactions within the For compounds 1 and 2, the diffraction intensity data were collected at 150 K on a Bruker APEXII CCD diffractometer (Bruker, Karlsruhe, Germany) with graphite-monochromated Mo Kα radiation (λ = 0.7107 Å). The program SADABS (Bruker, 2016) was used for absorption corrections [45]. Direct methods were used to solve the structure, and the SHELX2014 program [46] was used to refine the structure with the full-matrix least-squares method against F 2 . All non-hydrogen atoms were refined anisotropic thermal parameters, whereas the hydrogen atoms on their respective carbon atoms were placed in ideal, calculated positions, using the riding model with isotropic thermal parameters. For compounds 1 and 2, the experimental details for X-ray crystallographic data, and the refinements are summarized in Table 1, and the selected bond distances and angles are listed in Tables 2 and 3.

Physical Measurements
The temperature dependence (dc) of the magnetic susceptibility measurements for compounds 1 and 2 were performed on microcrystalline samples, which were restrained in eicosane to prevent torquing, on a Quantum Design MPMS-7 SQUID (Quantum Design, San Diego, CA, USA) equipped with 7.0 T magnets and operated in the range of 2.0-300.0 K. The Pascal's constants [47] were used to estimate the diamagnetic corrections of both compounds from the experimental magnetic susceptibilities to achieve the molar paramagnetic susceptibilities. Elemental analysis (carbon, hydrogen, and nitrogen) of compounds 1 and 2 were made using an Elemental vario EL III analyzer (PerkinElmer, Taipei, Taiwan). Thermogravimetric (TG) analyses of both compounds were collected using a Seiko Instrumental, Inc., (Chiba shi, Japan) EXSTAR 6200 TG/DTA analyzer, operating under a 5 • C/min heating rate and a nitrogen atmosphere. Powder X-ray diffraction data were collected using a Rigaku MiniFlex-II X-Ray diffractometer (Tokyo, Japan), operating on 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). A Perkin-Elmer Spectrum RX1 FTIR spectrometer was used to collect the Fourier transform infrared (FTIR) spectra for both compounds (PerkinElmer, Taipei, Taiwan).

Synthesis and Characterization of Compounds 1 and 2
Compounds 1 and 2 were prepared via self-assembly processes by solvothermal reactions at different temperatures. The reaction of Co(OAc) 2 ·4H 2 O (0.05 mmol) with H 2 -Me 2 mal (0.05 mmol) and bpe (0.05 mmol) in a DMF/H 2 O (1 mL/5 mL) mixed solution at 60 or 80 • C for 48 h yielded needle-shaped purple crystals of 1. No crystals of 1 were formed when the reaction was carried out at room temperature. When the above reaction was performed at 120 • C for 48 h, however, red crystals of thermodynamically stable products of compound 2 were obtained. When crystals of 1 in the same solvent system were allowed to stand at 120 • C for about 48 h, the purple colored crystals of 1 were converted into red colored crystals of 2, thus confirming that crystal reassembly has occurred. The synthesis and structural conversion of 1 to 2 are depicted in Scheme 1. However, except for 1, no new product was formed, as determined by the PXRD patterns, but if the reaction was carried out in the absence of H 2 O/DMF as the solvent, this indicated that reassembly from 1 to 2 would be simultaneously assisted by both thermal and solvent factors. In viewing the crystal structures of 1 and 2, the removal of coordinated water from the Co(II) center and the rearrangement of the carboxylate of Me 2 mal 2− and bpe ligands would be expected to take place during the structural conversion process. We assume that the solvent molecules would provide intermolecular interactions, i.e., hydrogen bonding interactions that would be expected to stabilize the structure in the intermediate state and would benefit from the removal of the coordinated water molecules. Such solvent assisted structural transformations have been reported in the literature [48][49][50][51], and these findings can be compared to the reassembly of crystals from 1 to 2. In the reported literature, supramolecular interactions, such as layer-guest-layer hydrogen bonding interactions, stabilize the layer structure. As the temperature increased, the guest water molecules were removed and the structure was induced to transform into a 3D network. We, therefore, assume that the removal of coordination water molecules in 1 would be an important factor for the crystal reassembly to 2.   The phase purity for the bulk samples of 1, 2, and the products obtained by the crystal reassembly of 2 was confirmed by PXRD (Figures S1 and S2 in Supporting Information) and by elemental analysis. The thermal stabilities of compounds 1 and 2 were also examined by TG analysis ( Figure S3). The TGA curve for compound 1 showed consecutive weight loss steps at temperatures above 120 • C, which corresponded to the gradual removal of one coordinated water molecule (found: 6.28%; calcd: 6.04%). At temperatures above 250 • C, the host framework underwent a rapid weight loss, which was attributed to the elimination of the organic ligands and the gradual decomposition of the compound. Compound 2 showed one large weight loss at temperatures above 300 • C, corresponding to the elimination of the organic ligands, followed by the decomposition of the host framework.

Description of the Structure
Crystal Structures of Compound 1 [Co(Me 2 mal)(bpe) 0.5 (H 2 O)] n (1). An X-ray structural analysis showed that compound 1 crystallized in the monoclinic space group P2 1 /n and the asymmetric unit of 1 contains one crystallographically independent Co(II) center, one Me 2 mal 2− anion, one-half of a bpe ligand, and one coordinated water molecule. As depicted in Figure 1, the geometry of the Co(II) center is a distorted octahedron with a CoO 5 N coordination sphere. The equatorial positions on the octahedron are occupied by four oxygen atoms (O2, O3, and their symmetrical equivalents) derived from three Me 2 mal 2− ligands, while one nitrogen atom (N1) from one bpe ligand and one oxygen atom (O6) of a terminal water molecule occupied the axial positions. The Co-O bond lengths at the Co(II) vary from 2.0747(18)-2.124 (19) Å and a Co-N bond length is 2.151(2) Å, which falls in the range of values for typical octahedral Co(II) complexes [52,53]. The two carboxylate groups of the Me 2 mal 2− ligand adopt a syn-anti µ 2 :η 1 ,η 1 -bridging mode (Scheme 2a), in which the Co(II) ion is chelated by each of one oxygen atoms (O2 and O3), and connects to two crystallographically equivalent Co(II) ions through each of the other two oxygen atoms (O1 and O4). Thus, each Co(II) center is linked to four neighbors via three Me 2 mal 2− bridges and results in a corrugated Co-Me 2 mal layer that is located parallel to the ac crystal plane (Figure 2a). The two unique Co···Co distances in the layer spanned by the Me 2 mal 2− ligands are 5.478(1) and 5.258(1) Å. The adjacent Co-Me 2 mal layers are further pillared through bis-monodentate bpe ligands leading to a pillared-layer 3D framework with 1D channels along the crystallographic c axis ( Figure 2b). Similar structures have been reported in previous studies [54][55][56]. The bpe ligands are positioned alternately above and below the layers, in a trans array with the dimethyl groups of the Me 2 mal 2− ligand. The bridging bpe ligands separate the Co(II) ions by 13.668(2) Å, where the shortest interlayer Co···Co distance is 9.911(2) Å. The shortest distance of the centroid-centroid between the adjacent pyridyl rings of the bpe is 7.496(2) Å, which rules out any π-π interaction in compound 1. The shortest centroid-centroid distance between the adjacent pyridyl rings of the bpe ligand in compound 1 is 7.496(2) Å, which are considerably higher than the limit for π-π interactions between pyridyl rings, thus indicating there are no π-π interactions in compound 1. The intralayer hydrogen bonds between the oxygen atoms of the Me 2 mal 2ligand and the coordinated water molecule in 1 (2.704(3) and 2.768(3) A • for O6···O2 and O6···O3 and 150.4(2) and 163.9(4) o for O6-H6A . . . O2 and O6-H6B···O4) donate to the stabilization of the structure.   Scheme 1. The Schematic representation of the temperature-controlled assembly/reassembly of compounds 1 and 2.

Description of the Structure
[Co(Me2mal)(bpe)0.5(H2O)]n (1). An X-ray structural analysis showed that compound 1 crystallized in the monoclinic space group P21/n and the asymmetric unit of 1 contains one crystallographically independent Co(II) center, one Me2mal 2− anion, one-half of a bpe ligand, and one coordinated water molecule. As depicted in Figure 1, the geometry of the Co(II) center is a distorted octahedron with a CoO5N coordination sphere. The equatorial positions on the octahedron are occupied by four oxygen atoms (O2, O3, and their symmetrical equivalents) derived from three Me2mal 2− ligands, while one nitrogen atom (N1) from one bpe ligand and one oxygen atom (O6) of a terminal water molecule occupied the axial positions. The Co-O bond lengths at the Co(II) vary from 2.0747(18)-2.124 (19) Å and a Co-N bond length is 2.151(2) Å, which falls in the range of values for typical octahedral Co(II) complexes [52,53]. The two carboxylate groups of the Me2mal 2− ligand adopt a syn-anti µ2:η 1 ,η 1 -bridging mode (Scheme 2a), in which the Co(II) ion is chelated by each of one oxygen atoms (O2 and O3), and connects to two crystallographically equivalent Co(II) ions through each of the other two oxygen atoms (O1 and O4). Thus, each Co(II) center is linked to four neighbors via three Me2mal 2− bridges and results in a corrugated Co-Me2mal layer that is located parallel to the ac crystal plane (Figure 2a). The two unique Co···Co distances in the layer spanned by the Me2mal 2− ligands are 5.478(1) and 5.258(1) Å. The adjacent Co-Me2mal layers are further pillared through bis-monodentate bpe ligands leading to a pillared-layer 3D framework with 1D channels along the crystallographic c axis (Figure 2b). Similar structures have been reported in previous studies [54][55][56]. The bpe ligands are positioned alternately above and below the layers, in a trans array with the dimethyl groups of the Me2mal 2− ligand. The bridging bpe ligands separate the Co(II) ions by 13.668(2) Å, where the shortest interlayer Co···Co distance is 9.911(2) Å. The shortest distance of the centroid-centroid between the adjacent pyridyl rings of the bpe is 7.496(2) Å, which rules out any π-π interaction in compound 1. The shortest centroid-centroid distance between the adjacent pyridyl rings of the bpe ligand in compound 1 is 7.496(2) Å, which are considerably higher than the limit for π-π interactions between pyridyl rings, Scheme 2. The Schematic representations of (a) the symmetric µ 3 :η 1 ,η 1 ,η 1 ,η 1 -bridging Me 2 mal 2 ligand in compound 1, and (b) the antisymmetric µ 4 :η 1 ,η 2 ,η 2 -bridging Me 2 mal 2− ligand in compound 2.
between the oxygen atoms of the Me2mal 2-ligand and the coordinated water molecule in 1 (2.704(3) and 2.768(3) A° for O6···O2 and O6···O3 and 150.4(2) and 163.9(4) o for O6-H6A…O2 and O6-H6B···O4) donate to the stabilization of the structure.  Topological analysis of the 3D structure of compound 1 indicated that each Co(II) center with a CoNO5 coordination environment and µ3-Me2mal can be viewed as a six-connected node and a three-connected node, respectively, while each bpe ligand that is bonded to two Co(II) centers can be treated as a two-connector. Such connectivity repeats in infinity, thus producing the Co-Me2mal layer and the 3D framework of 1 as schematically represented in Figure 3. Analysis by the TOPOS software package showed that the framework of 1 could be explained as a binodal (3,4)-connected ins topology with the Schläfli symbol (6 3 )(6 5 .8) [57]. Topological analysis of the 3D structure of compound 1 indicated that each Co(II) center with a CoNO 5 coordination environment and µ 3 -Me 2 mal can be viewed as a six-connected node and a three-connected node, respectively, while each bpe ligand that is bonded to two Co(II) centers can be treated as a two-connector. Such connectivity repeats in infinity, thus producing the Co-Me 2 mal layer and the 3D framework of 1 as schematically represented in Figure 3. Analysis by the TOPOS software package showed that the framework of 1 could be explained as a binodal (3,4)-connected ins topology with the Schläfli symbol (6 3 )(6 5 .8) [57]. [Co(Me2mal)(bpe)0.5]n (2). A single-crystal X-ray diffraction analysis reveals that 2 is a three-dimensional framework consisting of well-isolated one-dimensional cobalt/oxygen chains bridged by a bpe ligand along with a Me2mal ligand. The asymmetric unit of compound 2 contains two Co(II) atoms at special positions, two Me2mal 2− groups, and one bpe ligand (Figure 4). The Co1 [Co(Me 2 mal)(bpe) 0.5 ] n (2). A single-crystal X-ray diffraction analysis reveals that 2 is a three-dimensional framework consisting of well-isolated one-dimensional cobalt/oxygen chains bridged by a bpe ligand along with a Me 2 mal ligand. The asymmetric unit of compound 2 contains two Co(II) atoms at special positions, two Me 2 mal 2− groups, and one bpe ligand (Figure 4). The Co1 shows a distorted octahedral with a CoO 4 N 2 coordination sphere, where two oxygen atoms (O1, O1A) of the carboxylate from the two Me 2 mal 2− ligands and two nitrogen atoms (N1, N1A) from the two bpe ligands make up the equatorial plane, and the two carboxylate oxygen atoms (O3, O3A) from two Me 2 mal 2− ligands occupy the axial positions. The Co2 center adopts a CoO 6 distorted octahedral geometry bonded to six carboxylate oxygen atoms from four different Me 2 mal 2− ligands, where the four oxygen atoms (O1, O2, and their symmetric equivalents) occupy the equatorial plane and the two oxygen atoms, O3 and its symmetric equivalent, are coordinated to the axial positions. The Co−O bond distances (ranging from 1.992(2)-2.163(2) Å) and Co−N (2.067(3) Å) are all within the normal ranges for octahedral Co(II) complexes [42,43]. The Me 2 mal 2− ligand acts as an asymmetrical µ 4 -bridging ligand (Scheme 2b). One of two carboxylates in a Me 2 mal 2− adopt a anti, syn-syn, µ 3 :η 2 ,η 1 -bridging mode with terminal bonding to the Co2A center through an O2 atom with the Co1 and Co2 centers bridged by the O2 atom, while the other one of two carboxylate groups show an anti,syn µ 2 :η 2 -bridging mode that connects the Co2 to Co1A centers through its one carboxylate oxygen atom (O3) with an uncoordinated oxygen atom (O4). Such an asymmetrical µ 4 -bridging structure was first obtained in the complex with malonate-related ligands. Therefore, the Co1 and Co2 centers are connected by one syn-syn carboxylate and two µ 2 -oxygen atoms derived from two carboxylate groups and lead to the formation of an edged-shared zigzag Co−Me 2 mal chain along the c axis (Figure 5a). The intrachain Co1−Co2 bond distance is 3.150 Å, and the Co1−O1−Co2 and Co1−O3−Co2 bond angles are 95.35 • and 98.51 • , respectively. The adjacent Co−Me 2 mal chains are further crosslinked through bis-monodentate bpe ligands leading to a 3D network structure with 3D channels (Figure 5b). In the crystal, the adjacent porous 3D framework is interpenetrated leading to a two-fold interpenetrated network, in which only the small 1D channels along the a axis can be observed ( Figure S4). In the 3D interpenetrated framework, the bridging bpe ligands separate the Co(II) ions by 13.287(4) Å, where the shortest interchain Co···Co distance is 9.352(2) Å. The shortest centroid-centroid distance between adjacent pyridyl rings of bpe is 7.496(2) Å ruling out any π-π interactions in compound 2.

Magnetic Properties
Solid-state, temperature-dependence of magnetic susceptibility data of compounds 1 and 2 were measured in the 2.0-300 K range under a 1.0 kOe applied field.
The temperature dependence of χMT values of compound 1 is shown in Figure 6. The spin-only value at 300 K is 3.34 cm 3 K mol -1 , which is in agreement with that of the measured, calculated by a high-spin Co(II) ion with S = 3/2 with a strong spin-orbit coupling [58]. The χMT value of 1 continues

Magnetic Properties
Solid-state, temperature-dependence of magnetic susceptibility data of compounds 1 and 2 were measured in the 2.0-300 K range under a 1.0 kOe applied field.
The temperature dependence of χ M T values of compound 1 is shown in Figure 6. The spin-only value at 300 K is 3.34 cm 3 K mol -1 , which is in agreement with that of the measured, calculated by a high-spin Co(II) ion with S = 3/2 with a strong spin-orbit coupling [58]. The χ M T value of 1 continues to decrease with cooling of the temperature from 300 to 10 K, below 10 K the value of χ M T decreases more rapidly to 1.40 cm 3 K mol -1 at 2.0 K. The monotonic decrease in the χ M T value with decreasing temperatures are characteristics of overall antiferromagnetic interactions and/or spin-orbital couplings in compound 1. Above 50 K, the magnetic susceptibility data obeyed the Curie-Weiss law with a Curie constant C = 3.58 cm 3 mol -1 K and Weiss θ = −23.15 K ( Figure S5). The negative θ value indicates that antiferromagnetic coupling and the spin-orbital coupling existed in compound 1.
A plot of the temperature dependence of the χ M T value for compound 2 is shown in Figure 7. The χ M T value is 3.13 cm 3 K mol −1 at 300 K which is significantly greater than the spin-only value for a the high-spin Co(II) center, while it is in agreement with the values observed of the magnetic moment for high-spin Co(II) complexes in an octahedral environment with strong spin-orbital coupling. The χ M T value decreases slowly to a minimum of 3.03 cm 3 K mol −1 at 55 K. Below 55 K, the χ M T value sharply increases to reach a maximum of 9.12 cm 3 K mol −1 at 2.0 K, indicating the existence of the ferromagnetic interaction in compound 2. The data of the temperature dependence of magnetic susceptibilities above 50 K followed the Curie−Weiss law giving a θ = −2.90 K and a C = 3.21 cm 3 K mol −1 ( Figure S6). The Curie constant of 2 was larger than the theoretical spin-only value of the Co(II) ion, indicating that an orbital contribution existed in compound 2. Thereby, the weak and negative values of θ are unable to indicate the antiferromagnetic interaction between Co(II) centers in 2 because of the significant strong spin-orbital coupling of the octahedral Co(II) ions. to decrease with cooling of the temperature from 300 to 10 K, below 10 K the value of χMT decreases more rapidly to 1.40 cm 3 K mol -1 at 2.0 K. The monotonic decrease in the χMT value with decreasing temperatures are characteristics of overall antiferromagnetic interactions and/or spin-orbital couplings in compound 1. Above 50 K, the magnetic susceptibility data obeyed the Curie-Weiss law with a Curie constant C = 3.58 cm 3 mol -1 K and Weiss θ = −23.15 K ( Figure S5). The negative θ value indicates that antiferromagnetic coupling and the spin-orbital coupling existed in compound 1. A plot of the temperature dependence of the χMT value for compound 2 is shown in Figure 7. The χMT value is 3.13 cm 3 K mol −1 at 300 K which is significantly greater than the spin-only value for a the high-spin Co(II) center, while it is in agreement with the values observed of the magnetic moment for high-spin Co(II) complexes in an octahedral environment with strong spin-orbital coupling. The χMT value decreases slowly to a minimum of 3.03 cm 3 K mol −1 at 55 K. Below 55 K, the χMT value sharply increases to reach a maximum of 9.12 cm 3 K mol −1 at 2.0 K, indicating the existence of the ferromagnetic interaction in compound 2. The data of the temperature dependence of magnetic susceptibilities above 50 K followed the Curie−Weiss law giving a θ = −2.90 K and a C = 3.21 cm 3 K mol −1 ( Figure S6). The Curie constant of 2 was larger than the theoretical spin-only value of the Co(II) ion, indicating that an orbital contribution existed in compound 2. Thereby, the weak and negative values of θ are unable to indicate the antiferromagnetic interaction between Co(II) centers in 2 because of the significant strong spin-orbital coupling of the octahedral Co(II) ions. Because of the contribution of the spin-orbit coupling for Co(II) ions, it is not possible to find a suitable analytical expression that describes the temperature-dependent magnetic susceptibility for Co(II) centers of the layered and chained polymeric structures in compounds 1 and 2, respectively. However, Rueff et al. successfully proposed a phenomenological approach for a low-dimensional polymeric Co(II) compound that permits the magnitude of the magnetic coupling and the spin-orbit coupling effects. They assumed the phenomenological equation [59]: where A + B are the Curie constants and the E1 and E2 are the activation energies, which correspond to the parameters of the spin-orbit coupling and the magnetic coupling interaction, respectively. The E2 is related to the constant of the magnetic coupling (J) according to the Ising chain approximation, χMT ∞ exp(+J/2kT). This equation sufficiently pronounces the spin-orbit coupling, which affects the splitting of the low-temperature divergence of the susceptibility between discrete levels and the exponential. Some reasonable values for magnetic interactions and interactions of spin-orbit coupling have been described in several studies on Co(II) coordination polymers with 1D and 2D structures [60][61][62]. Because of the contribution of the spin-orbit coupling for Co(II) ions, it is not possible to find a suitable analytical expression that describes the temperature-dependent magnetic susceptibility for Co(II) centers of the layered and chained polymeric structures in compounds 1 and 2, respectively. However, Rueff et al. successfully proposed a phenomenological approach for a low-dimensional polymeric Co(II) compound that permits the magnitude of the magnetic coupling and the spin-orbit coupling effects. They assumed the phenomenological equation [59]: where A + B are the Curie constants and the E 1 and E 2 are the activation energies, which correspond to the parameters of the spin-orbit coupling and the magnetic coupling interaction, respectively. The E 2 is related to the constant of the magnetic coupling (J) according to the Ising chain approximation, χ M T ∞ exp(+J/2kT). This equation sufficiently pronounces the spin-orbit coupling, which affects the splitting of the low-temperature divergence of the susceptibility between discrete levels and the exponential. Some reasonable values for magnetic interactions and interactions of spin-orbit coupling have been described in several studies on Co(II) coordination polymers with 1D and 2D structures [60][61][62].
The results obtained by the Rueff s procedure are quite consistent with the experimental data. For compound 1, the data above 10 K was fitted and the parameters of the fitting are A + B = 3.72 cm 3 K mol -1 , practically close to the Curie constant found from the Curie-Weiss law, 3.58 cm 3 K mol -1 . The E 1 /k was +49.73 K which is in the same magnitude (the order of +100 K) to those reported for Co(II) compounds.
Concerning the values obtained for antiferromagnetic exchange interactions, it is weak but significant (E 2 /k = 0.46 K, corresponding to J = −0.92 K), which is in agreement with the antiferromagnetic property of compound 1 and consistent with some other reported 2D Co(II) compounds [61,62]. As described in the crystallographic part, compound 1 is comprised of Co(II) ions connected by carboxylate groups in a syn-anti bridging mode thus giving a Co-Me 2 mal layer, which is further linked in a three-dimensional network by bpe ligands. Thus, the overall antiferromagnetic exchange interaction can be attributed to magnetic interaction within the Co-carboxylate layer. Magnetic exchanges through the syn-anti carboxylate bridges for Co(II) and Mn(II) ions are usually reported as antiferromagnetic due to the good overlap of magnetic orbitals [54][55][56]. For compound 2, the data above 30 K were fitted and the best fit parameters of the A + B value was 3.18 cm 3 K mol -1 , which is in good agreement with those reported in the literature for the Curie constant, E 1 /k, the effect of the distortion of coordination site and spin-orbit coupling, was +99.41 K and −E 2 /k was 4.07 K, corresponding to magnetic interactions of J = 8.14 K within the Ising chain approximation, which is consistent with values reported for several 1D Co(II) complexes [59][60][61]. These fitting results indicate that the distinct ferromagnetic exchange is dominated between Co(II) through one O−C−O and two µ 2 -O bridges. The intrachain Co1−O−Co2 angles and Co···Co distance are 98.5 • , 95.3 • and 3.150(1) Å, respectively. Nevertheless, the shortest Co···Co distance in the chain is 9.352 Å. The magnetic interactions of compounds 1 and 2 can be compared to those of Co(II) compounds containing similar structures in the literature [60,63], in which a weak antiferromagnetic interaction (J = −1.30 K) was dominated in a syn-anti carboxylate-bridged Co(II)-malonate layer and a weak ferromagnetic coupling (J = 3.72 K) was transmitted in a Co(II)-based chain with one O−C−O and two µ 2 -O bridges. This is in agreement with magneto-structural analyses: a small Co−O−Co angle resulting in a Co−Co ferromagnetic coupling and a large Co−O−Co angle resulting in a Co−Co antiferromagnetic interaction [63].
The ferromagnetic coupling of compound 2 was further estimated by isothermal magnetization data. As shown in Figure S7, the magnetization increased sharply and then reached a saturation plateau (2.53 Nβ at 7 T) with a fast saturation of the magnetization. The fast saturated magnetization confirms the existence of ferromagnetic interactions within 2 and the saturation value was consistent with the theoretical values (2−3 Nβ) expected for Co(II) compounds. Indeed, no magnetic hysteresis loop was detected indicating the absence of magnetic ordering in 2 above 2.0 K.

Conclusions
In summary, we report on the temperature-controlled synthesis and characterization of two new dicarboxylate-based 3D Co(II) coordination polymers. The formation of compound 1 was favored at lower temperatures of 60-80 • C, but a higher temperature of 120 • C was favored to yield compound 2. Indeed, a crystal reassembly from compound 1 to compound 2 was also observed by solvothermal treatment of 1 at 120 • C in DMF/H 2 O. The structure of compound 1 contains corrugated layers of Co(II) connected through syn-anti carboxylate bridges of Me 2 mal − ligands and bis-monodentate bpe pillars. Compound 2 shows a 3D porous framework involving one-dimensional Co-carboxylate chains connected by antisymmetric µ 4 -Me 2 mal 2− and bis-monodentate bpe ligands. Magnetic measurements indicate that antiferromagnetic interactions through the syn-anti carboxylate bridges of the Me 2 mal 2− ligands were dominated in compound 1, while compound 2 revealed ferromagnetic interactions resulting from the short Co-Co separation (3.150 Å) and small Co-O-Co bridging angles (98.5 • and 95.3 • ) exchange pathway of the µ 4 -Me 2 mal 2− bridges. The studies demonstrate that dimethylmalonic acid has great substantial for use in the preparation of coordination polymers with multipurpose structural topologies and unusual magnetic properties.