Structural diversity in the coordination compounds of cobalt, nickel and copper with N-alkylglycinates: crystallographic and ESR study in the solid state

Reactions of N-methylglycine (HMeGly), N-ethylglycine-hydrochloride (H2EtGlyCl) and N-propylglycine-hydrochloride (H2PrGlyCl) with cobalt(ii), nickel(ii) and copper(ii) ions in aqueous solutions resulted in ten new coordination compounds [Co(MeGly)2(H2O)2] (1), [{Co(MeGly)2}2(μ-OH)2]·2H2O (1d), [Cu(MeGly)2(H2O)2] (2α), [Co(EtGly)2(H2O)2] (3), [Ni(EtGly)2(H2O)2] (4), [Cu(μ-EtGly)2]n (5p), [Co(PrGly)2(H2O)2] (6), [Ni(PrGly)2(H2O)2] (7), and two polymorphs of [Cu(PrGly)2(H2O)2] (8α and 8β). Compounds were characterized by single-crystal and powder X-ray diffraction, infrared spectroscopy, thermal analysis and X-band electron spin resonance (ESR) spectroscopy. These studies revealed a wide range of structural types including monomeric, dimeric and polymeric architectures, as well as different polymorphs. In all monomeric compounds, except 2α, and in the coordination polymer 5p hydrogen bonds interconnect the molecules into 2D layers with the alkyl chain pointing outward of the layer. In 2α and in the dimeric compound 1d hydrogen bonds link the molecules into 3D structures. 1d with cobalt(iii), and 4 and 7 with nickel(ii) are ESR silent. The ESR spectra of 1, 3 and 6 are characteristic for paramagnetic high-spin cobalt(ii). The ESR spectra of all copper(ii) coordination compounds show that the unpaired copper electron is located in the dx2−y2 orbital, being in agreement with the elongated octahedral geometry.


Introduction
N-Alkylated-a-amino acids are present in nature and their biocatalytic properties, as well as chemical syntheses, are widely investigated. [1][2][3] They are useful building blocks in peptide science and have found application in structure-activity relationship studies. [4][5][6][7] The simplest modication of an amino acid is by N-methylation, so probably the most intensively investigated N-alkylated-a-amino acids are N-methyl-amino acids, especially N-methylglycine (sarcosine), N,N-dimethylglycine and N,N,N-trimethylglycine (betaine). 8 N-Methylation can be useful for conformational studies since introduction of N-methyl groups promotes conformational constraints and can also improve pharmacokinetic properties of some peptides. 9-11 N-Methylglycine is currently used as a dietary supplement, as a non-specic glycine transport inhibitor, and for treatment of schizophrenia and depression. 12,13 N-Ethylglycine acts as an inhibitor of glycine uptake and inhibits pain signaling and is a promising candidate for chronic pain treatment. 14 Structurally characterized coordination compounds with Nalkylated amino acids, especially those containing longer hydrocarbon chains at the amino nitrogen atom are quite rare. 15 To the best of our knowledge, no coordination compounds with N-ethylglycine and N-propylglycine have been structurally characterized up to now. There are a few structurally characterized coordination compounds with N-methylglycine. It was shown that the N-methylglycine moiety can occur in different forms: as an anion, a zwitterion or a cation. In the anionic form, the N-methylglycinato moiety acts as a bidentate ligand with O and N atoms involved in metal coordination. [16][17][18][19][20] In the case of the cationic and zwitterion forms, it acts as a monodentate, bidentate or bridging ligand. [21][22][23][24][25][26] Such coordination compounds may show different magnetic properties, depending on the metal oxidation state, local geometry around the metal center, metal-to-metal separation, bridging ligands, dimensionality of the complexes and non-covalent interacations.

Materials and physical measurements
All chemicals for the syntheses were purchased from commercial sources (Aldrich, Acros or Alfa Aesar) and used as received without further purication. N-Ethylglycine-hydrochloride (H 2 -EtGlyCl) and N-propylglycine-hydrochloride (H 2 PrGlyCl) were prepared by aminolysis of chloroacetic acid according to the method of E. Fischer (Scheme S1 †). 40 CHN analyses were performed on a Perkin-Elmer 2400 Series II CHNS analyzer in the Analytical Services Laboratory at the RuCer Bošković Institute, Zagreb, Croatia. The IR spectra were obtained in the range 4000-450 cm À1 on a Perkin-Elmer Spectrum Two™ FTIRspectrometer in ATR mode. The TGA measurements were performed at a heating rate of 10 C min À1 in the temperature range 25-600 C, under a nitrogen ow of 150 mL min À1 on a Mettler-Toledo TG/SDTA 851 e instrument. Approximately 10 mg of each sample was placed in a standard aluminum crucible (40 mL).
ESR measurements were conducted on a Bruker Elexsys 580 FT/CW spectrometer. The used microwave frequency was around 9.7 GHz; the magnetic eld modulation amplitude was 0.5 mT and the modulation frequency was 100 kHz. Samples were studied in the range from room down to liquid helium temperature.
Preparation of Co, Ni and Cu coordination compounds with the N-methylglycinato ligand Sodium hydroxide solution (0.08 g, 2 mmol in 10 mL water) was added to an aqueous solution containing N-methylglycine (0.53 g, 6 mmol) and corresponding metal acetate (2 mmol in 40 mL water). The mixture was stirred for a few minutes and le to stand at room temperature. Crystals of the coordination compounds [Co(MeGly) 2 2 ] n (2p), N-methylglycine (0.038 g, 0.5 mmol) and Cu(OH) 2 (0.24 g, 0.25 mmol) were mixed in 10 mL of water. Crystals of [Cu(m-MeGly) 2 ] n were obtained by slow evaporation of the solution. The coordination compounds were ltered off and washed with cold water (5 mL). A bulk sample of all compounds (except 1d) was taken for a powder X-ray diffraction experiment in order to conrm their purity. Powder patterns of the compounds were consistent with those calculated from the respective crystal structures (Fig. S1 †). Only a few crystals of [{Co(MeGly) 2 } 2 (m-OH) 2 ] (1d) were obtained from the solution that remained aer isolation of 1 aer several days.
[Co(MeGly) 2   [Cu(MeGly) 2   [Cu(m-MeGly) 2 ] n (2p). Synthesis as described above, with copper(II) hydroxide (0.24 g, 0.25 mmol). Dark blue crystals. The crystal structure of 2p is already published, however, the compound was synthesized by using a different copper(II) compound as the reactant. 20 Recrystallization of [Cu(MeGly) 2 2 ] (8a and 8b) suitable for X-ray structural analysis, were obtained by slow evaporation of the abovementioned reaction mixtures. A bulk sample of each compound was taken for powder X-ray diffraction experiment in order to conrm their purity. It was conrmed that the powder patterns of the synthesized compounds were consistent with powder patterns calculated from the respective crystal structures ( Fig. S1 †). The two polymorphs 8a and 8b crystallize from the same reaction mixture. In some cases only 8a crystallized from the solution while in others a mixture of the polymorphs crystallized. The two polymorphs could be crystallized separately from 8a in different solvents. The coordination compounds were ltered off and washed with cold water (5 mL). The crystals of 7 were not suitable for single-crystal X-ray structural analysis, however, other analyses (TGA, IR, PXRD) suggest the structural formula [Ni(PrGly) 2

X-ray diffraction analysis
Single-crystal X-ray diffraction data of the coordination compounds were collected by u-scans on an Oxford Diffraction Xcalibur3 CCD diffractometer with graphite-monochromated MoKa radiation. Room temperature single-crystal X-ray diffraction data for 2p, 3 and 4 were collected on an XtaLAB Synergy-S diffractometer with CuKa radiation. Data reduction was performed using the CrysAlis soware package. 42 Solution, renement and analysis of the structures were done using the programs integrated in the WinGX system. 43 All structures were solved by direct methods (SHELXS) and by dual-space methods (SHELXT), and the renement procedure was performed by the full-matrix least-squares method based on F 2 against all reections using SHELXL. [44][45][46] The non-hydrogen atoms were rened anisotropically. All hydrogen atoms were located in the difference Fourier maps. Because of poor geometry for some of them, they were placed in calculated positions and rened using the riding model. Hydrogen atoms of the coordinated and crystallization water molecules were found in difference Fourier maps and the O-H distances were xed to 0.85(1) A, and the H-H distances were xed to 1.39(2) A. Geometrical calculations were done using PLATON. 47 Drawings of the structures were prepared using PLATON and MERCURY programs. 48 The crystallographic data are summarized in Tables S1-S3. † Based on the crystal structures of polymorphs 8a and 8b as well as monoclinic (2b), (CSD refcode POBDIT) and triclinic (2a) polymorph of [Cu(MeGly) 2 (H 2 O) 2 ] the Hirshfeld surface was generated using program CrystalExplorer17. 49 Additionally, Hirshfeld surface ngerprint plots were generated representing 2D histograms of the d i and d e distances; d i corresponds to the distance from a point on the surface to the nearest nucleus inside the surface and d e corresponds to the distance from a point on the surface to the nearest nucleus outside the surface. 50 Powder X-ray diffraction (PXRD) data were collected on a Malvern Panalytical Aeris powder diffractometer in the Bragg-Brentano geometry with PIXcel 1D detector, using CuKa radiation (l ¼ 1.5406 A). Samples were contained on a Si sample holder. Powder patterns were collected at room temperature in the range from 5 to 30 (2q) with a step size of 0.043 and 7.14 s per step. The data were collected and visualized by using the Malvern Panalytical HighScore Soware Suite. 51 Crystal data for 1.
Crystal data for 5p.  (Table 1 and Fig. S2 †). Two polymorphs of the copper(II) coordination compounds with the N-propylglycinato ligand were obtained, 8a and 8b. The same synthetic procedures gave in some cases pure 8a, while in others simultaneous appearance of 8a and 8b. Recrystallization of 8a in some conditions gave pure 8a, in some cases pure 8b, and in one case a mixture of the two polymorphs (Table 1 and Thermal stability of all monomeric coordination compounds was evaluated by the initial loss of both coordinated water molecules. Nickel(II) coordination compounds, which dehydrate in the range 90-140 C, are the most stable, while copper(II) complexes lose coordinated water molecules at much lower temperatures (90-100 C). Cobalt(II) coordination compounds lose coordinated water molecules in the range 90-110 C. Further decomposition of the dehydrated coordination compounds proceeds with carbonization. The lowest decomposition temperature is observed in copper(II) coordination compounds (decomposition starts at ca. 200-210 C), while their cobalt(II) and nickel(II) analogues (aer dehydration) have similar thermal stabilities (decomposition starts at ca. 300-320 C). Full thermal analysis data are given in Table S4. † Infrared spectra of the coordination compounds were characterized by the presence of very strong and sharp bands of the antisymmetric and symmetric stretching of the carboxylate ion, n as (COO) occurring in the range of 1620-1580 cm À1 , and n s (COO) occurring in the range of 1400-1380 cm À1 . The difference between n as (COO) and n s (COO) is generally greater than 200 cm À1 indicating monodentate coordination mode of the carboxylate ion, as conrmed by the results of the X-ray analysis. [52][53][54] A sharp band of medium intensity, which was assigned as O-H stretching, n(OH, H 2 O), was observed in the range 3240-3461 cm À1 in the spectra of all monomeric coordination compounds. Comparing the spectra of the monomeric cobalt(II), nickel(II) and copper(II) compounds, the n(OH) bands occur at the highest wavenumbers in the spectra of the copper(II) compounds. This difference indicates a slightly larger decrease

Crystal structures of monomeric coordination compounds
All monomeric compounds (1, 2a, 3, 4, 6, 8a and 8b) are centrosymmetric with the metal atom lying on the inversion center (detailed crystallographic data are given in Tables S1-S3 †). The asymmetric units contain half of the coordination compound molecule, except in 2a where there are two independent halves of molecules. ORTEP plots of one representative molecular structure of a coordination compound with each Nalkylglycinato ligand: N-methylglycinato (1), N-ethylglycinato (4) and N-propylglycinato (8a), are presented in Fig. 1. Coordination compounds 2a, 3, 6, and 8b have analogous labeling schemes of the N-alkylglycinato ligands as the ones shown (Fig. S3 †). Cobalt(II) and nickel(II) coordination compounds with N-ethylglycinato ligands (3 and 4) are isostructural (Table S2 †). We were not able to obtain single-crystals of 7 of good quality to solve the crystal structure, however, thermal analysis, as well as infrared spectroscopy suggests that the molecular structure of 7 is equivalent to that of 6, 8a and 8b.
The metal atom in the structures of all monomeric compounds (1, 2a, 3, 4, 6, 8a and 8b) is octahedrally coordinated by two N,O-bidentate N-alkylglycinato ligands in the equatorial positions and two water molecules occupying the axial coordination sites ( Fig. 1 and S3 †). The amino nitrogen atoms are arranged in the trans-position. The copper(II) ion in compounds 2a, 8a and 8b exhibits the typical Jahn-Teller distorted [4 + 2] coordination geometry. The longer axial bonds are toward the coordinated water molecules (Table S5 †).
In all monomeric compounds, except 2a, hydrogen bonds interconnect the molecules into 2D layers (Fig. 2a). All metal atoms within the hydrogen-bonded 2D layer are coplanar. Although the alkyl chains are of different lengths (methyl in 1, ethyl in 3 and 4, and propyl in 6, 8a and 8b) the hydrogen bond motif within the layer is the same in all compounds except 8b which has two additional hydrogen bonds. In all monomeric compounds the hydrogen atom from the amino nitrogen atom N1 serves as a hydrogen bond donor to the carboxylate oxygen atom O2 which is not coordinated to the metal atom. The shortest N/O hydrogen bond, d(N1/O2) ¼ 2.970(3) A, is in compound 1 (Table S8 †). Additionally, the coordinated water    (Table S8 †). Fig. 2 (upper row) shows hydrogen bonds forming 2D layers in compounds 1, 3 and 6 as a representative of the monomeric compounds. The alkyl chains in 1, 3 and 6 and in all monomeric compounds, except 2a, point outward of the 2D layers forming only weak van der Waals contacts (Fig. 2, lower row). Geometries of the intermolecular hydrogen bonds are given in Table S8. † In the triclinic polymorph of [Cu(MeGly) 2 (H 2 O) 2 ] (2a) (two independent halves of the molecules in the asymmetric unit) hydrogen bonds link the molecules into a 3D structure (Fig. 3). Coordinated water molecules, as well as amino nitrogen atoms, serve as hydrogen bond donors to the carboxylate oxygen atoms, both between symmetrically dependent and independent molecules. The shortest hydrogen bond length is that between the symmetrically dependent molecules and involves the oxygen atom from the coordinated water molecule (O1W) and the carboxylate oxygen atom O21 which is not coordinated to the copper atom (Table S8 †). Two 2D layers of hydrogen bonds forming a 3D supramolecular structure are shown in Fig. S4. † The monoclinic polymorph of [Cu(MeGly) 2 (H 2 O) 2 ] (2b) also forms a 3D supramolecular structure. The main structural difference between the two polymorphs is the orientation of the water molecule in the complex molecule (Fig. S5 †). As a consequence, the two polymorphs have slightly different intermolecular contacts. Hirshfeld surfaces and ngerprint plots showing intermolecular contacts are given in Fig. S6. † Both polymorphs [Cu(PrGly) 2 (H 2 O) 2 ] (8a and 8b) crystallize in the monoclinic crystal system but with different unit cell parameters and space groups (8a in I2/a and 8b in P2 1 /c, see Table S3 †). There is only a small difference in the molecular conformation of 8a and 8b, mostly in the orientation of the coordinated water molecules (Fig. S7 †). However, this small difference has a signicant impact on the crystal packing. In the crystal structure of 8b there are two additional bifurcated hydrogen bonds. Amino nitrogen atom connects two molecules through the N-H/O carboxylate hydrogen bond and the hydrogen bond involving the coordinated water molecule toward carboxylate oxygen atoms of two neighbouring molecules (Fig. S8 †). The difference in the hydrogen bonding between two neighbouring complex molecules in polymorphs can be  described by graph-set notation of hydrogen bond motifs. 55 In 8a two rings are formed -R 2 2 (8) and R 2 2 (10), and in 8b there are ve rings formed by six hydrogen bonds -2R 1 2 (6), 2R 2 1 (4) and R 2 2 (8) (Fig. S8 †). Non-covalent interactions in the crystal structures of the polymorphs were further investigated by Hirshfeld surface analysis. The 2D ngerprint plots with the decomposition of the dominant types of intermolecular contacts in 8a and 8b are presented in Fig. 4.
Both polymorphs exhibit a pair of long sharp spikes with short d i and d e values (bottom le of the plot. The upper associated with the donor atom, the lower one with the acceptor) representing the O water -H/O carboxylate hydrogen bonds. There is also a close C-O carboxylate /C-O carboxylate contact (Fig. S9 †) in 8b (d(O/C) ¼ 2.970(2) A), which is characteristic for trans-(aminocarboxylato)copper(II) polymeric coordination compounds. 37 Crystal structure of the dimeric compound 1d In 1d each cobalt(III) atom is octahedrally coordinated by two Nmethylglycinato ligands and two hydroxyl groups forming a distorted octahedron (Fig. 5a). This structure is a dihydrate, the only one among the investigated compounds. The structure is dimeric with two hydroxyl groups linking two cobalt(III) atoms. Such coordination is typical for cobalt(III) coordination compounds with amino carboxylates, that is, glycinate, 56,57 alaninate, 58 Table S5 †) in the monomeric compounds. The crystal structure is stabilized by an extensive hydrogen-bonding network. Both water molecules of crystallization are involved in hydrogen bonding forming 2D layers (Fig. 5b) but only one water molecule (O1W) is involved in the linkage between the layers thus forming a 3D network. Hydrogen bonds in 1d are given in Table S9. † Crystal structure of the copper 2D coordination polymer 5p The copper atom in 5p is coordinated by two N-ethylglycinato ligands in the equatorial plane and the axial coordination sites are occupied by carboxylate oxygen atoms from the neighbouring complex units (Fig. 6). The copper(II) ion exhibits the Jahn-Teller distorted coordination geometry with four shorter equatorial bonds to the nitrogen and carboxylate oxygen atoms of two N-alkylglycinato anions, and the longer axial bonds to the carboxylate oxygen atoms of neighbouring complexes (Table  S7 †).  The packing is similar as that in the monomeric structures. 2D layers are formed with the alkyl chain pointing outward of the layer, however here the molecules are interlinked by covalent bonds (Fig. 6). Only one type of hydrogen bond is present in the structure, the intermolecular N1-H/O1 hydrogen bond ( Fig. S10 and Table S9 †). Each complex is involved in four hydrogen bond chains (two as hydrogen bond donors and two as acceptors) propagating in two dimensions (Fig. S10 †).

Inuence of the alkyl chain on crystal packing
Packing index (PI) was calculated for all structurally characterized complexes and for compounds published in CSD ( obtained with all three N-alkylglycinates, these complexes were studied in more detail. The distance between hydrogen bonded layers is shortest for the N-ethylglycinate complex 3 (8.03 A), being in accordance with the efficient packing, and longest for N-propylglycinate complex 6 (9.93 A), while for the N-methylglycinate coordination compound 1 it is slightly greater than in 3 (8.35 A) (Fig. S11 †). This result may be surprising, however, Nethylglycinate ligand has larger conformational freedom than N-methylglycinate, which allows it to fold in a more efficient way. On the other hand, N-propylglycinate with an extra CH 2 group is large enough to form interpenetrated alkyl chains between the hydrogen bonded layers, thus signicantly increasing the interlayer distance (Fig. 2, lower row).
The nickel(II) coordination compounds were ESR silent within the measured temperature range, as it is usually the case for nickel(II) (non-Kramer's system with S ¼ 1). 62 The dimer coordination compound 1d was also ESR silent in the whole temperature range as expected for coupled integer spins of cobalt(III) ions. 62 The cobalt(II) coordination compounds had no signal at room temperature but aer lowering the temperature below 100 K, the signals appeared. The recorded spectra of these coordination compounds at two selected temperatures are shown in Fig. 7.  The recorded spectra are characteristic for paramagnetic high-spin cobalt(II) ions (S ¼ 3/2, d 7 ). Octahedral cobalt(II) ion usually has large zero-eld splitting that results with only the lowest states (m s ¼ AE1/2) thermally occupied, thus only one ESR line is observed with highly anisotropic g-values. 34,62,63 No hyperne interaction between electron spin S ¼ 1/2 and nuclear spin I ¼ 7/2 for cobalt(II) ions was detected. 64 Therefore, the following reduced form of spin-Hamiltonian was assumed: In eqn (1), the constant m B is Bohr magneton, B is external magnetic eld, g is g-tensor, S is electron spin operator for the effective cobalt spin of S ¼ 1/2. The spectra were simulated by using EasySpin soware. 65 The obtained g-values and parameters used for the simulation of cobalt(II) coordination compounds are given in Table 2 while the simulated spectra are shown in Fig. 7. The same parameters were used for the simulations at different temperatures while only line-width of the used Lorentzian lines were changed with temperature. g-Strain parameters were used as factors for line-broadening to obtain better agreement with the experimental spectra.
The representative ESR spectra of the investigated copper(II) coordination compounds, obtained at the selected temperatures, are shown in Fig. 8. Hyperne interaction between electron spin S ¼ 1/2 and nuclear spins I ¼ 3/2 was not detected and therefore the form of spin-Hamiltonian (1) was used for the simulation. 65 The simulated spectra are shown in Fig. 8, while the parameters used for the simulations are given in Table 3. As was mentioned before for the cobalt coordination compounds, the spectra were simulated taking into consideration only the temperature change of line-width of assumed Lorentzian lines.
Although 2p and 5p are coordination polymers, simulations show that their magnetic structures are monomer-like and similar to those of 2a, 8a and 8b. This is due to the fact that the closest Cu/Cu distance in the polymeric chain is 5.4 A and 5.5 A in 2p and 5p, respectively. From the obtained g-values, given in Table 3, one can see that g x z g y < g z for all copper complexes so the unpaired copper electron is located in the d x 2 Ày 2 orbital. This is in agreement with the elongated octahedral copper geometry where g z is in the direction of the axial distortion. 66

Conclusions
Structural diversity was found to depend both on the metal ion and chain length. Cobalt(II) and nickel(II) coordination compounds are monomers of the general formula Table 2 The values of spin-Hamiltonian parameters obtained from the spectral simulations of Co(II) coordination compounds   2 ] when R ¼ methyl and propyl, and also polymeric coordination compounds of the type [Cu(RGly) 2 ] n , R ¼ methyl and ethyl. Two polymorphs of the copper(II) coordination compound with the N-propylglycinato ligand were obtained, 8a and 8b, with signicant differences in non-covalent interactions due to the orientation of the coordinated water molecule. Conditions for obtaining pure forms were found by varying solvents. In all monomeric compounds, except copper(II) with N-methylglycinate, hydrogen bonds interconnect the molecules into 2D layers. Although the alkyl chain in the monomers is of different length the hydrogen bond motif within the layers is the same in all compounds except 8b which has two additional hydrogen bonds. In copper(II) with N-methylglycinate the hydrogen bonds link the molecules into a 3D structure. Oxidation of cobalt(II) to cobalt(III) occurred upon standing of the solution of monomeric 1 in air, resulting in the formation of 1d with dimeric molecules linked into a 3D structure. 5p is a coordination polymer with 2D layers similar to those in the monomeric compounds. The effect of the alkyl chain length in the cobalt(II) and nickel(II) compounds is seen in the efficiency of crystal packing: monomeric N-ethylglycinato complexes pack most efficiently. ESR spectroscopy shows that cobalt(III) and nickel(II) coordination compounds are ESR silent. Cobalt(II) coordination compounds have ESR spectra characteristic for paramagnetic high-spin cobalt(II) ions (S ¼ 3/2, d 7 ). ESR spectra of copper(II) coordination compounds show that the unpaired copper electron is located in the d x 2 Ày 2 orbital, being in agreement with the elongated octahedral coordination in all copper(II) coordination compounds. Spectra of the polymeric coordination compounds 2p and 5p are similar to those of the monomeric copper(II) coordination compounds due to large Cu(II)/Cu(II) distances in these polymers and therefore weak spin-spin interactions between them.

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