Bioinspired manganese(II) complexes with a clickable ligand for immobilisation on a solid support

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains. Accepted Manuscript Dalton Transactions


Introduction
][3] Fixation on a support allows not only achieving site isolation but also controlling site nuclearity.Indeed, for monomeric sites, oligomerisation and degradation are often observed in solution with molecular bioinspired complexes.Therefore, it is important to develop new ligands with side functions allowing reactions that are orthogonal to complexation for fixation on a solid support.One possibility is to use the Huisgen azidealkyne [2 + 3] cycloaddition reaction referred to as "click chemistry" by Sharpless and co-workers. 4Indeed, this efficient and versatile strategy, developed mainly for the synthesis of elaborate organic molecules (such as inhibitors for pharmaceutical applications), removes constraining considerations of protection/deprotection and cross-reactivity. 5,68][9] In the case of silica-based supports, either azide or alkyne functions are grafted onto the solid support using the suitable organosiloxane molecules before carrying out the Huisgen cycloaddition reaction.In the case of immobilisation of a metal complex, it is preferred to have the alkyne function in the ligand and the azide in the tether to avoid complexation of azide to the metal ion.This modular approach has been used for example in the grafting of Co and V complexes with Schiff base ligands, or Pt and Pd organometallic complexes to develop heterogeneous catalysts. 10,11lectronic supplementary information (ESI) available: Crystallographic details of the complexes, crystal packing figures of 1-4, EPR simulation of 1 and N 2 sorption isotherms of the porous materials.CCDC 905065, 911104-911106 for 1-4.3][14][15][16][17][18][19] In particular, catechol dioxygenases are ring-cleaving enzymes that oxidise catechol derivatives with a concomitant ring opening, which is a critical step in the aerobic degradation of aromatic compounds by bacteria. 19Therefore, these systems can be used as models to develop new materials for the degradation of aromatic chemicals.2][23] This motif, called the 2-His-1-carboxylate facial triad, is observed in catechol dioxygenases as well as in other metalloproteins. 15,24n this work a new series of manganese complexes with two different biomimetic clickable ligands N,N′-bis(( pyridin-2-yl)methyl)prop-2-yn-1-amine (L 1 ) and N-((1-methyl-1H-imidazol-2yl)methyl)-N-(pyridin-2-ylmethyl)prop-2-yn-1-amine (L 2 ) is presented (Scheme 1).Both ligands can mimic the facial triad present in the active site of the MndD and possess an alkyne side function that allows for an efficient grafting onto a solid support using click chemistry.

Synthesis of the ligands
The ligands L 1 and L 2 were synthesised according to experimental procedures reported in the literature.In the case of L 1 , two different synthetic methods were explored.The first one is a two-step protocol starting with the reduction with NaBH 4 of the Schiff base resulting from the condensation of 2-picolylamine with 2-pyridinecarboxaldehyde which is followed by an electrophilic substitution with propargyl bromide to obtain the desired L 1 ligand. 25,26The yield of the first step was 96% with a simple work-up whereas the purification by column chromatography on silica in the second step only yielded 11% of the product likely due to the less affinity of amino-pyridinic sub-strates to the acidic nature of the silica.Yet this latter should not be taken as a standard reaction yield as results found in the literature show average yields around 90%. 27 The second method is a straightforward one-step protocol consisting of a double electrophilic substitution involving propargylamine and 2-picolyl chloride. 28The reaction time of 12 h was prolonged up to 5 days allowing us to retrieve L 1 in a 90 to 96% yield, compared to a 32% yield for 12 h.Therefore this one-step protocol was the appropriate synthetic route to the L 1 ligand (Scheme 2a).
L 2 is an original ligand that has been synthesized using an adapted version of the first method, i.e., condensation of an amine and an aldehyde, namely 2-picolylamine and 1-methyl-2-imidazolecarboxaldehyde, respectively, followed by a reduction 29 and an electrophilic substitution by propargyl bromide (Scheme 2b).The first step provided the secondary amine in 96% yield.It is interesting to note the absence of undesired by-products in the second step, in contrast to the synthesis of L 1 that required purification by column chromatography.In fact, re-dissolution in CH 2 Cl 2 and filtration through KBr provided the pure ligand L 2 in 80% yield.

Synthesis of the Mn(II) complexes
Complexes 1-4 were synthesized in MeOH using a stoichiometric amount of the ligand and metal salt (Scheme 3).For compounds 1 and 2 involving chloride and bromide counteranions, respectively, the corresponding manganese salt was available and directly used for the reaction.In the case of 3 and 4, involving azide anions, manganese nitrate was reacted with the ligand and 4 equivalents of sodium azide were added to exchange the nitrate anion by the azide one.Except for 2, each compound precipitated as light brown powder in MeOH within 30 min.Those powders were then redissolved in a minimum amount of MeOH, while the solution containing 2 was used as it is.The crystals were grown using the same technique in all the cases, i.e., slow diffusion of diethyl ether in a methanolic solution of the compound to afford crystals within a couple of days.Apart from 3, all the crystals were stable in air at room temperature.Molecules of MeOH found in the structure of 3 were lost when left at room temperature, damaging the crystallinity of the solid.

Description of the structures
Complexes 1-4 are double halide or pseudohalide-bridged neutral dimers of Mn(II) and exhibit the same structural layout (Fig. 1).They differ in the nature of the donor groups from the organic ligand counterpart, i.e., amine and pyridines from L 1 (1-3), or amine, pyridine and imidazole from L 2 (4), as well as the coordinating anion composing the double bridge between the metal ions and completing the metal environment, i.e., chloride (1), bromide (2), or azide (3 and 4).Selected bond distances and angles for 1-4 are compiled in Table 1.
Within all complexes, the Mn(II) ions are located in a distorted N 3 X 3 [X = Cl (1), Br (2) and N (3 and 4)] octahedral environment.In all the complexes, the Mn-N bond lengths from the tridentate ligands are similar and in good agreement with similar complexes involving a double bridge with double chloride, [30][31][32] bromide 33 and end-on azide [34][35][36][37] bridges.It is important to note that the Mn-N bond length involving the central amine nitrogen atom of the ligand is usually ∼0.1 Å longer than the two external ones involving the pyridine and/ or imidazole nitrogen atoms.In the case of 4, involving L 2 , one of the Mn-N bond lengths (N2) or the external nitrogen atoms of the ligand is shorter (∼0.1) and the Mn-N bond length for the central (N3) one is longer (∼0.1 Å) than those observed for complexes 1-3 involving L 1 .The environment of the metal ions is completed by three X atoms with two of them involved in the double bridge between the two Mn(II) ions.As expected, Mn-Cl bond lengths are shorter than the Mn-Br ones.Moreover, Mn-X bond lengths within the double bridge (X = Cl1, Br1) are longer (∼0.1 Å) than the terminal ones (X = Cl3, Br3).
The distortion of the octahedral environment of the metal ion is also exemplified by the distribution, far from the ideal values of 90 and 180°, of the X-Mn-X angles (X = N, Br and Cl) ranging from 69.72 (5) In the crystal lattice of 1 and 2, structural packing is built by weak hydrogen bonding H⋯X (X = Cl and Br, respectively) interactions leading to a dense 3D network.In the case of 3, the co-crystallized methanol molecule, through the hydrogen atom of the alcoholic group, forms hydrogen bonds with the terminal nitrogen atom of the double end-on azido-bridge.Finally, for 3 and 4, the crystal packing, in both cases, is assumed to be formed by weak interactions as van der Waals interactions for example.

Magnetic properties
Fig. 2 shows the magnetic properties of 1-4 in the form of χ M T versus T plots and M versus H plots, χ M and M being the magnetic susceptibility and the magnetisation per two Mn(II) ions respectively, while T and H are the absolute temperature and the applied magnetic field, respectively.
At room temperature, χ M T for 1 and 2 is ca.8.75 cm 3 mol −1 K. Whereas this value is expected for two magnetically isolated high-spin Mn II ions (S Mn = 5/2), the corresponding values for 3 and 4 are somewhat greater (ca.9.5 cm 3 mol −1 K).Upon cooling, χ M T for 1 remains constant until 25.0 K and it further decreases to 7.70 cm 3 mol −1 K at 2.0 K, suggesting the occurrence of a very weak antiferromagnetic interaction.In contrast, χ M T for 2-4 continuously increases with the decrease of the temperature to reach maxima at 5.5 K (ca.11.3 cm 3 mol −1 K) for 2 and at 10 K (ca.14.8 cm 3 mol −1 K) for 3 and 4.After these maxima, χ M T decreases to 7.5 (2), 13.7 (3) and 13.9 cm 3 mol −1 K (4) at 2.0 K.The increase of χ M T in the high temperature region for 2-4 is indicative of the occurrence of a ferromagnetic interaction between the paramagnetic Mn(II) ions, whereas the decrease at low temperatures can be attributed to weak intermolecular antiferromagnetic interactions and/or zero-field splitting of the S = 5 ground spin state.However, because of the large isotropic character of the six-coordinate high-spin Mn(II) ion, the zero-field splitting effects are expected to be negligible.2).The inset shows the magnetization curve at 2.0 K and the solid line is the theoretical Brillouin curve for two magnetically independent spin sextuplets, S = 5/2.
Table 1 Selected bond distances (Å) and angles (°) The expression of the magnetic susceptibility for a dimanganese(II) unit derived from the Van Vleck's equation (the spin Hamiltonian being defined as H = −JS Mn1 •S Mn2 ) is given by eqn (1), 38 where x = J/kT and θ is the Weiss constant defined as θ = zJ′S-(S + 1)/3k with zJ′ accounting for the magnetic interaction between the z nearest dinuclear units.The other parameters have their usual meaning.The best-fit parameters obtained are listed in Table 2.
As indicated above, the decrease of χ M T at low temperatures for 2-4 could be alternatively interpreted by the presence of a zero-field splitting (D).Having this possibility in mind, the magnetic data for 2-4 were also analysed through the Hamiltonian of eqn (2) using VPMAG. 39 The least-squares fit of the experimental data to eqn (2) gave values for the g and J parameters similar to those obtained through eqn (1) (Table 2), but the values for the D parameter [4.9 (2), 1.4 (3) and 1.1 cm −1 (4)] are too high to be considered real and they are meaningless.In fact, the reported D values for the high-spin d 5 Mn II ion in an octahedral environment are lower than 1 cm −1 . 40So, the decrease of the χ M T values has to be mainly attributed to intermolecular antiferromagnetic interactions.
Field dependence magnetisation plots of 1-4 at 2.0 K are shown in the inset of the corresponding plot (Fig. 2).The isothermal magnetisation data for 1 are well below the Brillouin curve for two magnetically isolated S = 5/2 Mn(II) ions (with no ZFS), while they are well above for 2-4, in agreement with the occurrence of antiferro-(1) or ferromagnetic (2-4) interactions between the Mn(II) ions.
The main magneto-structural parameters for a series of [Mn II (μ-X) 2 Mn II ] (X = Cl or Br) dinuclear complexes which were the subject of previous studies are listed in Table 3.As far as we know, no theoretical magneto-structural correlation has been reported for this family of compounds.As one can see therein, no correlation seems to exist between the structural parameters (angles or distances) and the magnetic coupling parameter ( J), this circumstance being most likely due to the low number of compounds, the small variation of the structural parameters and the experimental errors.However, there must exist a correlation between the nature and the magnitude of the magnetic exchange and the structural parameters, the value of the Mn-X-Mn angle being the most determinant one, as is known for other metal ions and bridging ligands. 41In fact, the experimental data from Table 3 predict a change from ferro-to antiferromagnetic behaviour at an angle of ca.97°for the bridgehead chloro atom.The value of this crossover Mn-Cl-Mn angle is very close to that observed for the bis(μ-hydroxo)dicopper(II) complexes (ca.97.5°). 42Compound 2 is the first example of a bis(μ-bromo)dimanganese(II) complex studied for its magnetic properties and so, we can only compare it with the chloro derivatives listed in Table 3.
Concerning the bis(μ 1,1 -azido)dimanganese(II) complexes (3 and 4), an experimental relationship between the magnetic coupling constant ( J) and the Mn-N-Mn bridging angle (θ), J (cm −1 ) = 0.552θ (deg) − 53.8, was proposed. 36,43Table 4 shows the magneto-structural parameters for double end-on azido-bridged Mn(II) complexes.3 and 4 follow roughly the above relationship and they are in good agreement with the fact that the value of the exchange coupling constant increases with the Mn-N-Mn bridging angle, which is consistent with theoretical predictions. 41A careful inspection of the data listed in Table 4 shows that the values of the coupling parameter, J, generally increase with an increase in the difference between the two Mn-N(bridging azide) bond distances (Δd in Table 4), 36,43 which is an increase in the asymmetry of the azido bridging.In this sense, a better magneto-structural correlation is observed when we use both structural parameters (θ and Δd ).Fig. 3 shows the values of J versus the corresponding values of (θ + 10Δd ) from Table 4.The best linear fit gives the equation J (cm −1 ) = 0.553[θ (deg) + 10Δd (Å)] − 54.111, which is slightly better than J (cm −1 ) = 0.576θ (deg) − 54.040, where the influence of Δd has been neglected.

EPR characterization of the complex in solution
The molecular complex 1 is conveniently dissolved in MeOH-EtOH solution.Such a solution exhibits at room temperature a 6 line EPR signal centered at g = 2.001 ± 0.001, typical of Mn(II).It is assigned to the hyperfine coupling between the electronic spin (S = 5/2) and the nuclear manganese spin (I = 5/2) for the allowed transitions (Δm S = ±1, Δm I = 0).Upon simulation, the hyperfine coupling constant A is 97 ± 1 G (Fig. S5 †).At low temperatures, a pair of forbidden lines due to cross terms in the spin Hamiltonian lie between each of the main hyperfine lines (Fig. 4). 49A quantitative study was performed with complex 1 at room temperature in order to determine the proportion of monomeric and dimeric species present in the solution.The spin concentration was measured using manganese(II) perchlorate in methanolic solution at various concentrations for calibration.Accordingly, only 1-1.5% of this dimeric complex yields EPR active monomeric species in solution at room temperature.

Grafting onto silica
The above solution was used to graft the manganese complex in the nanopores of a mesoporous silica.2D hexagonal LUS silica (MCM-41 type) prepared using cetyltrimethylammonium tosylate as the surfactant was chosen as a support. 50,51It possesses a narrow pore size distribution of 3.8 ± 0.1 nm and a large mesopore volume of 0.78 ± 0.01 cm 3 g −1 .3][54] The manganese complex 1 was covalently grafted by using click chemistry.As an illustrative example, the azidefunctionalized silica (N 3 @SiO 2 ) was reacted with 1 in a MeOH-MeCN mixture in the presence of the tris-(triphenylphosphine) copper(I) bromide catalyst 55 to obtain the hybrid material MnL 1 @SiO 2 , with 2.5 wt% manganese.This complex loading corresponds to ∼80% yield for the click reaction.Note that the use of such a Cu(I) complex avoids the displacement of Mn(II) from the L 1 ligand by Cu(I) during the grafting of the complex.Indeed, there are no traces of Cu(II) that would form under aerobic conditions, according to the EPR spectrum (Fig. 4).On the other hand, the signal reveals the presence of Mn(II) as in solution.Indeed, the proportion of EPR active species remains similar to that in solution, consistent with most of the species being still dimeric.In addition, the signal is broader than in solution both at room temperature and at 116 K. Therefore, this broadening is not due to dynamic effects but due to distribution of species on the surface.This gives evidence of the heterogeneity of the environment around the grafted species as usually observed in such hybrid materials. 2he hexagonal array of the internal pores of the material was unaltered all through the different steps of the synthesis, as shown from the X-ray diffraction patterns (Fig. 5).The decrease of pore volume from 0.78 cm 3 g −1 down to 0.48 cm 3 g −1 after grafting confirmed the presence of the metal complex inside the pores (ESI, Fig. S6 †).Concomitantly, the specific surface area underwent a diminution from 990 m 2 g −1 to 340 m 2 g −1 .This is consistent with partial pore filling due to internal molecular functionalization of the silica.

Conclusions
A series of dinuclear manganese(II) complexes with two μ-X bridges (X = Cl, Br or N 3 ) has been synthesised using two different clickable ligands.These ligands allow both coordination to a metal ion and also grafting onto an azide-functionalized support using the Huisgen alkyne-azide cycloaddition reaction.Within this series we present the first μ-Br bridged Mn(II) dinuclear complex for which a complete magnetic property study has been performed.It turned to be weakly ferromagnetic compared to the μ-Cl analogue, which is weakly antiferromagnetic.It is also noteworthy that a better magnetostructural correlation is observed when both structural parameters (θ and Δd ) are considered for the bis(μ 1,1 -azido)dimanganese(II) complexes.Successful stabilisation of the complex mainly in its dimeric form has been achieved by grafting onto mesoporous silica.This kind of hybrid nanocomposite material, where confinement is at stake, could help by comparison to better understand the catalytic mechanism of MndD. 16,23,56The study of the catalytic properties of these silica-supported manganese(II) complexes is under progress.Alternatively, the alkyne function present in the ligand opens up other opportunities such as using these complexes as building blocks for the synthesis of metal-organic frameworks (MOFs). 57perimental Synthesis of the ligands N,N′-Bis(( pyridin-2-yl)methyl)prop-2-yn-1-amine ligand (L 1 ).The synthesis was based on a published protocol. 28A mixture of propargylamine (1.50 g, 27.3 mmol) and potassium carbonate (22.8 g, 163.8 mmol) was stirred in 140 mL of acetonitrile for 5 min.Then 2-picolyl chloride hydrochloride (9.84 g, 60.0 mmol) dissolved in 140 mL of acetonitrile was added.The resulting solution was stirred under reflux for 5 days.The solution was filtered and the filtrate was evaporated under reduced pressure.The resulting brown oil was dissolved in 100 mL of distilled water and extracted thrice with dichloromethane.The organic phases were gathered and dried over sodium sulphate.The solvent was finally evaporated to afford L 1 in 94% yield.hydride (0.86 g, 22.7 mmol) was added portionwise and the solution was further stirred for two more hours at room temperature.The solvent was evaporated under reduced pressure and the orange oil thus obtained was redissolved in a mixture of water-methanol (25 mL/50 mL), before being extracted thrice with 50 mL of dichloromethane.The combined organic phases were washed with 50 mL of water and dried over sodium sulphate.The solvent was finally removed to obtain the product as an orange oil in 96% yield. 1

Synthesis of the manganese(II) complexes
[Mn 2 (L 1 ) 2 (Cl) 2 (μ-Cl) 2 ] (1).A solution of L 1 (47 mg, 0.2 mmol) in 2 mL of methanol was slowly added to a solution of MnCl 2 (40 mg, 0.2 mmol) in 2 mL of methanol.The solution was stirred for 1 h, and then the solid was filtered and washed with methanol to obtain 38 mg of the complex in 52% yield.This powder was redissolved in a minimum amount of methanol, and pale yellow crystals of 1 were obtained by slow diffusion of diethylether into this solution.
[Mn 2 (L 1 ) 2 (N 3 ) 2 (μ-N 3 ) 2 ]•2CH 3 OH (3).A solution of L 1 (24 mg, 0.1 mmol) in 2 mL of methanol was slowly added to a solution of Mn(NO 3 ) 2 •4H 2 O (25 mg, 0.1 mmol) in 2 mL of methanol.The solution was stirred for 30 min, and then a solution of NaN 3 (26 mg, 0.4 mmol) in 3 mL of methanol was added.The mixture was further stirred for 1 h, and the solid thus formed was filtered and washed with methanol.The solid was redis-solved in methanol, and slow diffusion of diethylether into that solution provided crystals suitable for X-ray diffraction.
Synthesis of MnL 1 @SiO 2 The manganese complex was grafted onto LUS mesoporous silica where azide functions were previously incorporated using the so-called molecular stencil pattering technique described elsewhere. 53esoporous silica LUS.A solution of cetyltrimethylammonium tosylate (CTATos) (1.96 g, 4.3 mmol) in distilled water (71 mL) was stirred at 60 °C for 1 h.In the meantime, a sodium silicate solution (49 mL) was also stirred at 60 °C for 1 h.The silicate solution was added to the surfactant one by pouring it slowly on the edge of the recipient.After a vigorous shaking by hand, the resulting white mixture was placed in two autoclaves and heated in a microwave oven for 10 min at 180 °C.The autoclaves were cooled in an ice bath for 10 min before filtration and washing with distilled water (about 100 mL).The white solid obtained was dried at 80 °C to obtain 3.1 to 3.4 g of LUS.Elemental analyses: Si (23.63%),C (30.91%), H (6.31%), N (1.72%), S (0.20%).
N 3 @SiO 2 .LUS silica was functionalized with trimethylsilyl functions (TMS) using a reported protocol to afford TMS@SiO 2 . 53Then, TMS@LUS (500 mg) was pre-treated at 130 °C under argon for 2 h, and at 130 °C under vacuum for 2 more hours.After cooling under argon down to room temperature, a solution of 3-azidopropyl triethoxysilane (371 mg, 1.5 mmol) dissolved in 5 mL of dry toluene was added.The suspension was stirred for 1 h, 10 mL of toluene was added, and the mixture was stirred at 80 °C for 18 h.After filtration and washing with 20 mL of toluene, 20 mL of EtOH 70% and 20 mL of acetone, the solid was finally dried overnight at 80 °C.Elemental analysis (wt%): C: 5.18%, H: 1.80%, N: 2.06%.

Crystal structure determination and refinement
Single crystal X-ray diffraction data for 2-3 were collected at 293(2) K on a Bruker APEX II diffractometer using graphitemonochromated Mo-K α radiation while data for 1 and 4 were collected and treated on a Gemini Oxford Diffractometer at 150 K and the related analysis software. 58Data reduction for 2-3 were performed with the SMART and SAINT software. 59he structures were solved by direct methods using the SHELXS-97 program, 60 and refined on F 2 s by full-matrix leastsquares with the SHELXL-97 program included in the WINGX software package. 60,61Absorption correction was applied using SADABS. 62For complexes 1 and 4, an absorption correction based on the crystal faces was applied to the datasets (analytical). 63Structures of 1 and 4 were solved by direct methods combined with Fourier difference syntheses and refined against F using reflections with [I/σ(I) > 3] using the CRYSTALS program. 64All non-hydrogen atoms were refined using anisotropic terms.The hydrogen atoms were placed at the calculated positions and refined with isotropic parameters as riding atoms.The hydrogen atoms for the co-crystallized methanol molecules were located from difference maps for complex 3.The final geometrical calculations and the graphical manipulations were carried out with PARST95, 65 PLATON, 66 (for complexes 2-4) and DIAMOND 67 programs.
Crystal data, refinement results, atomic coordinates, selected bond lengths and bond angles for 1-4 are summarized in Tables S1-S12.†

Physical characterization
Infrared spectra were recorded on KBr pellets using a Mattson 3000 IRTF spectrometer.Nitrogen sorption isotherms at 77 K were determined with a volume device Belsorp Max on solids that were dried at 80 °C under vacuum overnight.Low angle X-ray powder diffraction (XRD) experiments were carried out using a Brucker (Siemens) D5005 diffractometer with Cu Kα monochromatic radiation.Variable-temperature magnetic susceptibility measurements were made using a Quantum Design SQUID susceptometer, using an applied field of 1000 G. Diamagnetic corrections of the constituent atoms were estimated from Pascal constants.EPR spectra were recorded using a Brucker Elexsys e500 X-band (9.4 GHz) spectrometer with a standard cavity.The EPR calibration was performed using manganese(II) perchlorate as the spin reference in the same range of concentrations as for the samples.The spin quantification was performed by double integration of the signal.The simulated spectra were calculated using the EasySpin toolbox from Matlab.

Scheme 3
Scheme 3 Synthesis of the manganese(II) complexes.

Fig. 2
Fig. 2 Temperature dependence of χ M T for 1 (top left), 2 (top right), 3 (bottom left) and 4 (bottom right).The solid line represents the best fit based on the parameters discussed in the text (Table2).The inset shows the magnetization curve at 2.0 K and the solid line is the theoretical Brillouin curve for two magnetically independent spin sextuplets, S = 5/2.

Table 2
Best-fit parameters for 1-4 (see the text)