Thermodynamics and high-pressure kinetics of a fast carbon dioxide ﬁ xation reaction by a (2,6-pyridinedicarboxamidato-hydroxo)nickel( II ) complex †

The previously reported carbon dioxide ﬁ xation reaction by the planar terminal hydroxide complex [Ni(pyN 2Me2 )(OH)] 1 − in DMF has been further characterized by determination of the equilibrium constants K 298eq = 2.4 ± 0.2 × 10 5 M − 1 and K 223eq = 1.3 ± 0.1 × 10 7 M − 1 , as well as the volume of activation for the CO 2 binding ( Δ V ≠ 223on = − 21 ± 3 cm 3 mol − 1 ) and back decarboxylation ( Δ V ≠ 223o ﬀ = − 13 ± 1 cm 3 mol − 1 ) by high-pressure kinetics. The data are consistent with an earlier DFT computation, including the probable nature of the transition state, and support designating the reaction as one of the most completely investigated carbon dioxide ﬁ xation reactions of any type.

Fixation and utilization of CO 2 is an increasingly prominent problem [1][2][3][4] not only because CO 2 is a contributor to the greenhouse effect, but also because of its importance as a potentially cheap, nontoxic and abundant source of renewable synthetic fuels. 5There are many general approaches that can lead to fixation and utilization of CO 2 to form energetically more valuable molecules like carbon monoxide, methanol or methane, ranging from photochemical [6][7][8][9][10][11][12][13][14][15][16][17][18] to electrochemical 1,5,[19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34] to chemical reduction [35][36][37][38][39][40][41][42][43] pathways.[48][49] Besides redox methods, carbon dioxide fixation can also be achieved through non-redox processes. 41][52][53][54] Nature utilizes such a reaction within the catalytic cycle of carbonic anhydrases, where the terminal hydroxo group at the four-coordinate Zn II center plays a crucial role in the reversible hydration of CO 2 . 55Formation of a unidentate bicarbonate complex seems to be critical for the efficiency of that catalytic cycle. 56Some of us have recently observed that members of a set of planar terminal hydroxo complexes derived from N,N′-2,6-disubstituted phenyl-2,6dicarboxamide pincer ligands react rapidly and completely with CO 2 in DMF to form η 1 -bicarbonate products (Scheme 1). 3,57Reactions are second-order with rate constants (ca. 10 4 -10 6 M −1 s −1 ) that tend to decrease as the steric size of the phenyl R-substituents increases.The most notable feature is the second-order rate constant k on = 9.5 × 10 5 M −1 s −1 for reaction (1) at 298 K (Table 1), extrapolated from measurements at lower temperatures. 3[Ni( pyN Me2 )(OH)] 1− carries the smallest ortho-substituent in the set; the unsubstituted ligand forms octahedral [Ni( pyN 2 ) 2 ] 2− . 58 Comparison of k 298 on for reaction (1) with second-order rate constants for other systems including model reactions for the Zn(II) enzyme carbonic anhydrase reveals an increase of ≥10 3 M −1 s −1 .Further, under the approximation k on ≈ k cat /K M the rate constant overlaps the enzyme range of ca. 10 5 -10 8 M −1 s −1 . 55While precise comparisons cannot be made between synthetic or enzyme systems and reaction (1) owing to solvent differences (aqueous vs. DMF), it is nonetheless evident that this reaction is an unusually rapid CO 2 fixation process.Such a rapid CO 2 fixation appears to be related to several notable features of [Ni( pyN 2 Me2 )(X)] − : presence of the uncommon terminal Ni II -OH group in the reactant and unidentate η 1 -OCO 2 H bicarbonate in the product complex, despite the four-coordinate ligand environment around the Ni II center. 3lthough the reversible character of this process has been qualitatively demonstrated, 57 quantification of this equilibrium and the back decarboxylation reaction has not yet been reported.Since the efficiency of a CO 2 uptake is related to a CO 2 binding constant, the equilibrium constant K eq of reaction (1) was thermodynamically and kinetically determined in this study at 298 K and 223 K, respectively, contributing to understanding of the overall solution thermodynamics.Moreover, there are no literature data for binding constants of CO 2 to Ni-OH species.The only binding constant reported is for a Zn-hydroxo complex at 217 K. 53 Kinetic characterization of the reaction so far has disclosed a low enthalpic barrier and a large negative entropy of activation consistent with a secondorder reaction (Table 1).However, to fully comprehend the nature of this rather unique low activation pathway for CO 2 insertion, here we sought to determine the pressure effect on its kinetics, applying unique high-pressure cryo-temperature stopped-flow measurements at 223 K and 0.5-70 MPa.This is important not only with respect to the determination of the volume of activation (ΔV ≠ ) and thus visualization of a reaction transition state, but also considering an increase in solubility of CO 2 under low temperature and high pressure conditions, which can have an impact on the efficiency of CO 2 sequestering on a larger scale.
The compounds were prepared as described. 46 To elucidate the reversible nature of reaction (1) and to determine its equilibrium constant K eq , a spectrophotometric titration was undertaken in which small aliquots of solutions of CO 2 in DMF were added to a 0.1 mM solution of [Ni( pyN 2 Me2 )(OH)] 1− in DMF at 298 K in a 1 cm path-length septum-capped and gas-tight UV/visible cuvette.Saturated solutions of CO 2 in DMF (0.2 M) 59,60 were prepared by bubbling CO 2 gas at room temperature through argon-saturated dry DMF in a 10 mL gas-tight Hamilton syringe.Because the concentration of CO 2 required for the titration was not in a high excess over the complex concentration, spectrophotometric data were analyzed with eqn (2) in which [ML] is the concentration of the product as a function of the concentration of CO 2 ([L 0 ]), [M 0 ] is the total complex concentration and K eq is the equilibrium constant.The change in the overall volume of a reaction mixture in the course of a titration is negligible because of the small amount of CO 2 titrant solution added.After the addition of CO 2 and prior to an absorbance measurement, the reaction mixture was allowed to equilibrate for 20 s.
The UV/visible spectra obtained in the course of the titration are depicted in Fig. 1(left).Observed absorbance changes and tight isosbestic points at 279 and 395 nm are consistent with previous results 3 describing the expected single-step reaction; the concentration of the product vs. [CO 2 ] is shown in Fig. 1(right).Analysis of the global spectra data with eqn (2) leads to K 298 eq = 2.4 ± 0.2 × 10 5 M −1 .Using K eq = k on /k off , we estimate the first order rate constant k off for the back decarboxylation reaction to be 4 s −1 at 298 K.We have previously demonstrated the reversibility of the binding reaction by its reversal upon passing dinitrogen through a solution of [Ni( pyN 2 Me2 )(OCO 2 H)] 1− . 3,57e only other quantified equilibrium for CO 2 fixation is bicarbonate formation by CO 2 insertion in a Zn-OH bond (K eq = 6 ± 2 × 10 3 M −1 in dichloromethane at 217 K, corresponding to ΔG = −3.8± 0.2 kcal mol −1 ). 53Under comparable low temperature conditions in DMF, reaction (1) is three to five orders of magnitude more efficient with estimated binding constants in the 4 × 10 7 to 8 × 10 8 M −1 range.The values were estimated from our K eq value and reaction enthalpies previously computed with BP86 and B3LYP functionals, respectively. 3 In terms of the mechanism of reaction (1), one can envision a transition state that is less or more compact than the product η 1 -OCO 2 H bicarbonate complex.The former will speak in favour of a four-coordinate transition state (hydroxo attack on CO 2 followed by a proton transfer), 61 whereas the latter supports a five-coordinate transition state (insertion of a carbon dioxide C-O bond into the Ni-OH bond). 3This ambiguity is possible to overcome by determining the volume of activation, ΔV ≠ , which will offer a clearer picture of the transition state in the solution.Given the information potentially forthcoming from the effects of applied pressure on reaction rates, [62][63][64] the high-pressure kinetics of reaction (1) was investigated.To determine the activation volume ΔV ≠ , pressure-dependent stopped-flow spectrophotometry at low temperature was employed.§ The decrease in absorbance of the Ni II complex at 450 nm and constant temperature, during the course of CO 2 binding, was monitored within a pressure range of 0.5-70 MPa (Fig. 2).
Reactions were performed under pseudo-first-order conditions, with CO 2 concentrations ranging between 0.5 and 1.5 mM and a Ni II concentration of 0.1 mM.The temperature of the reaction cell and the solutions was maintained at 223 ± 0.1 K. Table S1 † contains the resulting second order rate constants (k on ) derived from the slopes of plots of the observed rate constants as a function of CO 2 concentration at variable pressures (Fig. 3(left)).At least five kinetic runs were performed per data point up to 70 MPa at 223 K.Under these conditions, the volume of activation is independent of pressure and can be evaluated from the slope of the linear plot (ln(k on ) vs. pressure) 62 in Fig. 3(right). 35alysis of the data leads to ΔV ≠ on = -21 ± 3 cm 3 mol −1 for reaction (1).This value is entirely consistent with the negative value of ΔS ≠ for this reaction (Table 1) and supports an associative mechanism where bond breaking does not have a significant role in the rate-determining step.Unlike many classical substitution reactions, there is neither a leaving ligand nor a change in charge.Thus, the observed ΔV ≠ corresponds to the intrinsic volume change and directly reflects the transition state at the molecular level.Activation volumes for the same type of reaction are reported only for CO 2 uptake by [M(NH 3 ) 5 (OH)] 2+ complexes in aqueous solution (M = Co III , -10.1 ± 0.6; Rh III , -4.7 ± 0.8; Ir III , -4 ± 1 cm 3 M −1 ). 61These values are significantly less negative than that of reaction (1), suggesting a more compact transition state in that reaction.By way of comparison, the partial molar volume of CO 2 in DMF is ca.43 cm 3 M −1 . 65By a dilution experiment (stopped-flow mixing of [Ni( pyN 2 Me2 )(OCO 2 H)] 1− with degassed DMF at 223 K) at variable pressures we have directly determined the activation volume for the back decarboxylation reaction ΔV ≠ off = -13 ± 1 cm 3 mol −1 (Fig. 4, Table S1 †).This enabled us to construct a volume profile (Fig. 5), which clearly indicates that the {[Ni( pyN 2 Me2 )(OH)] 1− ⋯CO 2 } § transition state is more compact than the product bicarbonate complex.In addition, based on the kinetic parameters at 223 K (Table S1 †) the corresponding binding constant could be obtained as K 223 eq = k on / k off = 1.3 ± 0.1 × 10 7 M −1 , which is in accord with the estimations based on the thermodynamic parameters (vide supra).
Such a volume profile supports a five-coordinate transition state with bidentate bicarbonate coordination on the way to the four-coordinate product.The four-coordinate structure of the starting complex and the binding of the dinegative pincer ligand enhance the nucleophilicity of a bound hydroxide.This favors an initial nucleophilic attack of a hydroxide and a cycloaddition pathway that should promote fast kinetics.The analogous reactions for the [M(NH 3 ) 5 (OH)] 2+ systems (M = Co III , Rh III or Ir III ) 61 have three orders of magnitude smaller rate constants than that in this study.This parallels a significantly lower nucleophilicity of OH − coordinated to the +3 metal centers within the positively charged [M(NH 3 ) 5 (  57 Results for both systems indicate a mechanism proceeding by a nucleophilic attack at the carbon atom of CO 2 by a hydroxide, formation of a cyclic transition state as depicted in Fig. 5, rupture of one Ni-O bond, and reorientation of the resulting monodentate bicarbonate by a Ni-O-C pivot motion.The predicted final perpendicular orientation of the bicarbonate relative to the NiN 3 plane is observed in crystalline salt of the reaction products. 46The herewith visualized mechanism differs from that reported for [M(NH 3 ) 5 (OH)] 2+ (M = Co III , Rh III or Ir III ), 61 where the rate-determining step in decarboxylation proceeds according to a dissociative mechanism that involves breakage of the oxygen-carbon bond on the carbonate ligand.In contrast, our back reaction has a clearly associative character, involving a Ni-O bond formation and closure of a chelate bicarbonate ring.We should point out that the reactions between CO 2 and [M(NH 3 ) 5 (OH)] 2+ operate in aqueous solutions, where a hydrogen-bonding network assists in proton transfer from the bound hydroxide to one of the CO 2 oxygen atoms.On the other hand, our reaction operates in DMF, within a hydrophobic environment of the sterically bulky ligand, and there is no possibility that the CO 2 oxygen atoms can be involved in some prominent hydrogen-bonding interactions.This makes the CO 2 oxygen atoms even more nucleophilic and prone to bind to the Ni II center, resulting in a fivecoordinate transition state.This may also explain the fact that our interpretation of the overall volume profile is consistent with, and may be said to support, the computed reaction pathway, although these computational models do not address hydrogen bonding. 3

Conclusions
Here, we have revealed the thermodynamics of CO 2 binding to [Ni( pyN 2 Me2 )(OH)] − , which was previously reported as the fastest CO 2 fixation reaction by a metal-bound hydroxide. 3sed on the determined equilibrium constant of 2.4 ± 0.2 × 10 5 M −1 at 298 K and 1.3 ± 0.1 × 10 7 M −1 at 223 K, it seems that the thermodynamics of this process is also highly efficient, though there is limited literature information for comparisons. 53ur high-pressure kinetic studies revealed a large negative activation volume for both the CO 2 uptake and the back decarboxylation reaction.These results support a very compact transition state, which can be visualized as a five-coordinate species with a bicarbonate bound in a bidentate fashion.The observed character of the transition state, in accord with computational results, 3 explains a very low activation enthalpy and a negative activation entropy consistent with rapid CO 2 fixation.Reaction (1) is currently the most thoroughly characterized carbon dioxide fixation reaction by a metal hydroxide complex, and among the more thoroughly examined carbon dioxide fixation reactions of any type.
Thermodynamics and high-pressure kinetics of a fast carbon dioxide fixation reaction by a (2,6-pyridinedicarboxamidato-hydroxo)nickel(II) complex † O. Troeppner, a D. Huang, b R. H. Holm b and I. Ivanović-Burmazović* a The previously reported carbon dioxide fixation reaction by the planar terminal hydroxide complex [Ni( pyN 2 Me2 )(OH)] 1− in DMF has been further characterized by determination of the equilibrium constants K 298 eq = 2.4 ± 0.2 × 10 5 M −1 and K 223 eq = 1.3 ± 0.1 × 10 7 M −1 , as well as the volume of activation for the CO 2 binding (ΔV ≠223 on = −21 ± 3 cm 3 mol −1 ) and back decarboxylation (ΔV ≠223 off = −13 ± 1 cm 3 mol −1 ) by high-pressure kinetics.The data are consistent with an earlier DFT computation, including the probable nature of the transition state, and support designating the reaction as one of the most completely investigated carbon dioxide fixation reactions of any type.

Fig. 1
Fig. 1 UV/visible spectral changes upon addition of small aliquots of CO 2 in DMF to a 0.1 mM solution of [Ni( pyN 2 Me2 )(OH)] 1− in DMF at room temperature (left).Plot of the concentration changes of the bicarbonate reaction product ([ML], based on the absorbance change at 381 nm) as dependent on [CO 2 ]; the red line is a fit to eqn (2) (right).

Fig. 2
Fig. 2 Spectral changes upon the reaction of 1 mM CO 2 and 0.1 mM [Ni( pyN 2 Me2 )(OH)] 1− in DMF at room temperature (left).The kinetics trace (black) is overlaid with the one-exponential fit to the data (red) for the same reaction system at 223 K and 0.5 MPa (right).

Fig. 4
Fig. 4 Plot of ln(k off ) vs. pressure for the reaction of 0.08 mM [Ni( pyN 2 Me2 )(OH)] 1− with CO 2 in DMF at 223 K (left).Corresponding time resolved UV/visible spectra in the course of the back decarboxylation reaction; 100 s observation time (right).