Theoretical Study on the Mechanism of Aqueous Synthesis of Formic

Feb 3, 2015 - To determine the feasibility of using the [Ru3+]-EDTA catalyst to produce CO-free hydrogen gas in WGSR, two probable reaction paths have...
0 downloads 12 Views 5MB Size
Download PDF
Article pubs.acs.org/IC

Theoretical Study on the Mechanism of Aqueous Synthesis of Formic Acid Catalyzed by [Ru3+]‑EDTA Complex Zhe-Ning Chen,† Kwong-Yu Chan,† Jayasree K. Pulleri,† Jing Kong,‡ and Hao Hu*,† †

Department of Chemistry, The University of Hong Kong, Pokfulam Road, Hong Kong, China Department of Chemistry and Center for Computational Sciences, Middle Tennessee State University, 1301 Main Street, Murfreesboro, Tennessee 37132, United States

‡

S Supporting Information *

ABSTRACT: Because formic acid can be effectively decomposed by catalysis into very pure hydrogen gas, the synthesis of formic acid, especially using CO and H2O as an intermediate of the water gas shift reaction (WGSR), bears important application significance in industrial hydrogen gas production. Here we report a theoretical study on the mechanism of efficient preparation of formic acid using CO and H2O catalyzed by a water-soluble [Ru3+]-EDTA complex. To determine the feasibility of using the [Ru3+]-EDTA catalyst to produce CO-free hydrogen gas in WGSR, two probable reaction paths have been examined: one synthesizes formic acid, while the other converts the reactants directly into CO2 and H2, the final products of WGSR. Our calculation results provide a detailed mechanistic rationalization for the experimentally observed selective synthesis of HCOOH by the [Ru3+]EDTA catalyst. The results support the applicability of using the [Ru3+]-EDTA catalyst to efficiently synthesize formic acid for hydrogen production. Careful analyses of the electronic structure and interactions of different reaction complexes suggest that the selectivity of the reaction processes is achieved through the proper charge/valence state of the metal center of the [Ru3+]-EDTA complex. With the catalytic roles of the ruthenium center and the EDTA ligand being carefully understood, the detailed mechanistic information obtained in this study will help to design more efficient catalysts for the preparation of formic acid and further to produce CO-free H2 at ambient temperature.

■

INTRODUCTION Hydrogen is considered to be the most ideal fuel for clean energy.1−4 Production and transportation of hydrogen becomes an important field of study in energy sciences. Direct production of hydrogen gas is often achieved with the water gas shift reaction (WGSR) (eq 1).5−8 CO + H 2O → CO2 + H 2

Indirect production of hydrogen could be made with synthesis and decomposition of hydrogen-rich compounds. With a 4.4 wt % of hydrogen, formic acid has received considerable attention as a potential hydrogen-storage material.11−13 High-purity hydrogen gas, without CO, could be obtained from HCOOH decomposition using soluble organometallic complexes13−19 or insoluble heterogeneous catalysts20−26 in aqueous solution at ambient temperatures. Research along this direction suggested that the formic acid generation from CO and H 2 O and the formic acid decomposition could be combined together to constitute essentially another type of WGSR.27,28 The synthesis and decomposition of HCOOH could even be coupled together to become a promising alternative to produce high-purity hydrogen (eq 2).27,28

(1)

On industrial scale, the reaction is usually carried out in gas phase and is facilitated by heterogeneous catalysts such as commercially available inorganic complexes containing Fe/Cr/ Cu and/or Cu/Zn/Al.5,8 Because the reaction is exothermic with a standard enthalpy change of −9.84 kcal/mol,7 H2 production and CO consumption is thermodynamically more favorable at lower temperatures. The catalytic efficiency at low temperature thus becomes the key requirement for the WGSR catalysts. Transition-metal carbonyls such as Fe(CO)5 and Ru(CO)5 have attracted great attention because of their good catalytic performance in mild reaction conditions.6,7,9,10 However, the industrial production of the hydrogen gas with WGSR was often complicated by the unreacted CO left in the reaction mixture, which creates additional technical and economic difficulties in separating the products. © XXXX American Chemical Society

CO(g ) + H 2O(l) → HCOOH(l) → CO2 (g) + H 2(g) (2)

For this two-step WGSR, the mechanism of the decomposition process has been extensively scrutinized.29−35 In Received: August 29, 2014

A

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

frequencies were calculated to confirm the correctness of the structure of either a local minimum or a transition state (TS). The solvation effects were considered using the SMD64 model with one explicit solvent water molecule incorporated at important positions in the model structure. Charge analyses were performed using the Mulliken schemes.65 The optimized structures were used to calculate the free energies at the level of unrestricted ωB97X-D functional66 with TZP basis sets. The ωB97X-D functional was reported to provide a good description of nonbonded interactions67 and showed reliable performance in recent benchmark tests of transition metal systems.68,69 Here TZP stands for a basis set that employs a 6-311+G(2d,p) all-electron basis set70 for the main group elements and a def2-QZVP basis set with the Stuttgart/Dresden effective core potentials (SDD def2-QZVP)61−63,71 for the ruthenium atom. The free energy at ωB97X-D/TZP level in aqueous phase was calculated according to eq 3:72−74

contrast, the mechanism of the generation of HCOOH from CO and H2O has been relatively less studied, even though an abundance of reports suggest that HCOOH can be prepared by the reduction of CO2 via electrochemical, homogeneous, or heterogeneous catalytic approaches.36−43 Recently, Shukla et al.44 reported the direct synthesis of HCOOH from CO and H2O can be accomplished efficiently over water-soluble [Ru3+]EDTA catalyst45 in moderate conditions of pressure (5−20 atm) and temperature (10−40 °C). An oxidation-addition and reduction-elimination mechanism was proposed for the catalyzed HCOOH synthesis but without further experimental validation. Furthermore, even though experiments hinted a high selectivity of HCOOH synthesis by the [Ru3+]-EDTA complex, how the catalytic selectivity is achieved is unknown. Obviously, if one can determine the catalytic mechanism of the [Ru3+]EDTA complex and subsequently confirm the selected synthesis of formic acid instead of direct WGSR, it might be feasible to combine the synthesis and decomposition of HCOOH in the same reacting apparatus to produce CO-free H2 as reported in recent experimental work.27,28 The catalyst in this reaction, that is, the water-soluble [Ru3+]EDTA complex,45 has a range of accessible oxidation states and has been explored in a wide variety of areas including catalysis and bioinorganic applications.46−49 It has shown notable catalytic properties mimicking enzymatic hydrocarbon oxidation by cytochrome P450 under homogeneous conditions.46,50 The catalytic versatility of the compound also makes the mechanistic study important in understanding the origin of catalytic power and in helping design new efficient catalysts. Our investigation in this paper focuses on the mechanism and kinetics of the production of HCOOH from CO and H2O using the [Ru3+]-EDTA catalyst. The primary interest is to determine and understand the origin of the selectivity of HCOOH production with the [Ru3+]-EDTA catalyst. For this reason, two plausible reactions catalyzed by the [Ru3+]-EDTA complex are investigated in the current work. One is the synthesis of HCOOH, and the other is the direct production of H2 from CO and H2O. The former corresponds to the first half of the WGSR via formic acid, and the latter corresponds to the direct WGSR without formation of formic acid. The nature of the electronic state of the effective catalyst is investigated. Generally, d5 complex [Ru3+]-EDTA should be a radical with one unpaired electron. But the proper charge/valence state of the metal center, instead of the radical character, is suggested to be the key contributor to the selective catalysis. To understand the effects of the buffer and solution pH, we also study the reaction processes over a protonated catalyst, the results of which demonstrate the importance of maintaining the solution pH. Our calculations show good agreement with experimental results and also in many aspects provide sound explanations for experimental observations. The results thus will provide a good foundation for rational design of a more effective catalyst, supplementing many existing theoretical studies on the mechanism of HCOOH dehydrogenation29−35 or direct reaction between CO plus H2O.51−57

■

⎛ RT ⎞ ωB97X‐D/TZP ωB97X‐D/TZP B3LYP/DZP ⎟ Gaq = Egas + ΔGthermo(aq) + ΔGsolv + RT ln⎜ ⎝ P ⎠

(3) The first term in the right-hand side is the electronic energy computed at ωB97X-D/TZP level in gas phase. The second term is the thermal correction to the free energy of the solute in the aqueous phase at B3LYP/DZP level. The third term is the solvation free energy. The last term denotes the free energy correction from the gas-phase standard state (1 atm) to the solution-phase standard state of 1 M, which is 1.89 kcal/mol at 298.15 K. The concentration effect of water molecules was considered if the solvent water molecule directly participates in the reactions. Note that the solvation free energy ΔGsolv was obtained with the corresponding SMD calculations at the level of B3LYP/6-31G(d) to make it consistent with the specific methods used in the development of the SMD solvation model.64 All calculations were carried out using the Gaussian 09 program.75

■

RESULTS Reaction Catalyzed by [Ru(EDTA)CO]− Catalyst. The metal complex ethylenediaminetetraacetato ruthenate [Ru(EDTA)(H2O)]−, prepared by reducing [Ru(HEDTA)(Cl)]K·H2O in water, was considered as the initial catalyst in aqueous solution.44 The electronic structure of the complex deserves a careful analysis here. The EDTA4− ion can donate at most 12 electrons from two nitrogen and four oxygen atoms of the ligand to the ruthenium center. Water molecule is a twoelectron donor ligand. The Ru3+ cation has five d electrons available in the coordination complex. Since the total number of available electrons should be no more than 18 for a transition metal complex, two different coordination patterns are allowed for the complex [Ru(EDTA)(H2O)]−, namely, complex I and II, depending on whether an oxygen (I) or nitrogen (II) atom is away from the ruthenium center to avoid oversaturated electrons in the metal center (Figure 1). In either case, there are total of 17 coordination electrons in the ruthenium center. The d5 complex [Ru(EDTA)(H2O)]− should be a radical. The significance of the radical character of the catalyst will be discussed in the later section. The approaching of CO to the ruthenium center to form a coordination complex was considered to be the initial step of the catalyzed hydration of CO.27,28,44 The remaining reactions can thus be considered to occur between H2O and this Ru-CO complex. As shown in Figure 1, our calculations demonstrate that the substitution of H2O by CO is thermodynamically favored. The stabilization likely comes from the effect of a synergistic π* back-bonding when carbon monoxide bonds to a transition metal. For the two isomers III and IV of the [Ru3+]EDTA carbonyl complex produced, respectively, from isomers I and II, our calculations show that isomer III, with an oxygen

COMPUTATIONAL DETAILS

All geometry optimizations were performed with the hybrid density functional theory (DFT) at the level of unrestricted B3LYP58,59 using DZP basis sets. Here, DZP stands for a basis set that employs 631G(d) all-electron basis set60 for the main group elements (H, C, N, and O) and the corresponding basis sets with the Stuttgart/Dresden effective-core potential (SDD)61−63 for the ruthenium atom. Analytical B

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

1). As shown, the T1 → T2 path leads to HCOOH formation, and the T1 → T3 → T4 path leads to the direct formation of Scheme 1

the WGSR products (CO2 and H2). In the transition state T1, the H2O molecule approaches CO in the [Ru(EDTA)(CO)]− complex. The hydroxyl group of H2O attacks the coordinated CO group. In a concerted fashion, the EDTA ligand is protonated to HEDTA by the attacking of the other proton of H2O. The product of the concerted reaction is a metallocarboxylic acid (MCA) intermediate [Ru]−COOH. Note that this MCA intermediate can undergo different reaction paths leading to different products. The T1 → T2 path via the MCA intermediate leads to the formation of formic acid and is thereafter referred as the formic acid formation path. In the transition state T2, the carboxyl carbon in the MCA intermediate is attacked by the proton in HEDTA ligand, forming a formic acid molecule in aqueous solution. (Note that we use the combined font of boldface and italic to highlight the atom under consideration.) Since there is no valence change of any atom in this reaction, the T1 → T2 path is an acid−base process. The transition states T3 and T4 correspond to the direct formation of CO2 and H2 as in the ordinary WGSR. The former is formed in a typical decarboxylation process, in which the hydrogen atom of the MCA intermediate ([Ru]−COOH) approaches [Ru] to form a CO2 and a ruthenium hydride complex [Ru]−H. This path correlates with a valence change of the hydrogen atom from a proton to a hydride, with the carboxyl carbon atom of the MCA complex [Ru]−COOH being formally oxidized from +2 to +4 via T3 to produce CO2. Thus, this path belongs to a redox process. The formation of the T4 transition state is a proton-hydride coupling process between [Ru]−HEDTA and [Ru]−H to form hydrogen gas. Therefore, the T1 → T3 → T4 path is a direct path to produce CO2 and H2 without going through the HCOOH intermediate. Figures 2−5 and the first part of Table 1 summarize the optimized geometries and the energies of all the transition states in different paths. In general, there are three or four possible reaction sites depending on whether complex III or IV is the catalyst, respectively. Reactions at all these sites were examined in the current work. The energies in the first part of Table 1 show that both complexes III and IV are catalytically competitive for addition of H2O molecule (T1), with an activation free energy of 23.2−27.8 kcal/mol for complex III and 19.5−23.8 kcal/mol for complex IV. In contrast, formation

Figure 1. Selected structures of initial complexes [Ru(EDTA)(H2O)]− and [Ru(EDTA)(CO)]−. Numbers in the parentheses are the free energies relative to complex I in kcal/mol. Atoms of Ru, N, O, C, and H are labeled in red, blue, orange, gray, and white, respectively.

atom of EDTA away from the Ru center, is 7.1 kcal/mol more stable than isomer IV, with a nitrogen atom away from the Ru center. Both isomers were considered as possible reactive species in the present work since the overall reaction activity is not always determined by the stability of an individual species in the path. When CO bonds to a transition metal to form an oversaturated free radical metal carbonyl complex, some previous calculations suggested that the M−CO angle can bend to allow delocalization of the unpaired electron on a carbonyl group and thus to achieve 18-electron low-energy configurations.76,77 Note that the bent Ru−CO conformation was not presented for the [Ru(EDTA)(CO)]− complex in the current work. Starting from the [Ru3+]-EDTA carbonyl complex, elementary steps of the two reaction paths were considered (T1−T4) to construct a clear picture of the catalytic processes (Scheme C

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 2. Optimized structures of the transition state T1 for the addition of H2O molecule to [Ru(EDTA)CO]− (bond length in Å). Numbers in parentheses are the imaginary frequencies in cm−1. Letters a, b, c, and d represent different oxygen active sites; III and IV indicate the initial structure with either O or N atom away from Ru, as shown in Figure 1.

this approach is supported by the agreement between our computed apparent activation free energy (T1 → T2) of 23.2 kcal/mol and the experimental result of 24.6 kcal/mol.44 The direct WGSR path to produce CO2 and H2 (T1 → T3 → T4) without forming the HCOOH intermediate was also examined in the present work. As shown in the first part of Table 1, even though the energy of the complex IV is more favorable than that of III for transition state T3, the activation barrier of 55.2−60.4 kcal/mol is too high for complex IV to be an effective catalyst for the decarboxylation process. As discussed previously, T3 belongs to a redox process in which the carboxyl carbon of the MCA intermediate ([Ru]−COOH) was oxidized from +2 to +4, while the proton of [Ru]−COOH was reduced to a ruthenium hydride [Ru]−H. This process should be generally disfavored because of the relatively low electron density in the Ru3+ center. In other words, formation of T3 is likely to be difficult if the metal center is in a high oxidation state. This proposition is in agreement with the observation that a process similar to the formation of T3 could always be observed when the metal center is in a zero valence state, such as the transition-metal carbonyls with a Ru0 or Fe0 metal center.51−57 The next transition state T4 corresponds to a proton-hydride coupling between [Ru]−COOH and [Ru]−H leading to the formation of hydrogen gas. As shown in Table 1,

of T2 showed clear difference between complexes III and IV. The activation barrier of T2 originated from complex III is as high as 43.6−63.1 kcal/mol, while it is only 22.5−27.0 kcal/mol if started from complex IV. This suggests that formation of T2 from complex III is very unlikely. Moreover, Ob (Figures 2 and 3) seems to be the most probable site for the formation of HCOOH in the T1 → T2 path with an apparent activation free energy of 23.2 kcal/mol. Previous experiments of hydration of CO suggested that the activation parameters are ΔH⧧ = 47.3 ± 2.5 kJ/mol and ΔS⧧ = −187 ± 7 J/K/mol,44 corresponding to an activation free energy of 24.6 kcal/mol at ambient temperature. The computed activation free energy thus is in a good agreement with experimental results. Formic acid, the final product of the T1 → T2 path, is known to be a strong organic acid. Its ionization process will make direct contribution to the thermodynamics of the reaction. To better describe the interaction between the formic acid and the water solvent, a simple model, with one explicit H2O molecule introduced around the hydroxyl group, was constructed for the HCOOH formation path (T1 → T2) in the current work. It was expected that the explicit interaction between the added H2O molecule and the hydroxyl group of the intermediate, transition, and product states would provide a better description for the solvation effect. The effectiveness of D

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 3. Optimized structures of the transition state T2 from MCA intermediate to prepare HCOOH, in which the proton transfers from ligand HEDTA to [Ru]−COOH (bond length in Å). Numbers in parentheses are the imaginary frequencies in cm−1.

for all the reactant, intermediate, transition, and product states investigated in the current work (Figure S1 and Table S1 in Supporting Information). Only a negligible mount of unpaired spins spread over other atoms. The radical character therefore does not make any special contribution to the catalysis in the calculations reported up to this point. To confirm that the Ru3+ radical complex is indeed responsible for the catalytic synthesis of formic acid, reactions catalyzed by the nonradical catalyst [Ru(EDTA)(CO)]0 and [Ru(EDTA)(CO)]2−, in which ruthenium has the formal valences of +4 and +2 respectively, are investigated here too. Unlike [Ru(EDTA)CO]−, only one structure is allowed for the nonradical species [Ru(EDTA)(CO)]0. As shown in Figure 7, all possible coordination sites of the EDTA ligand, two nitrogen and four oxygen atoms, coordinate to Ru to form a saturated electronic configuration with 18 electrons in the ruthenium center. For the other nonradical species [Ru(EDTA)(CO)]2−, there are two probable isomers, complex [Ru2+]-III and [Ru2+]-IV, in which either an oxygen or a nitrogen atom of EDTA ligand points away from the ruthenium center, which prevents more than 18 electrons from being in the ruthenium center.

this process is kinetically feasible with a relatively low activation free energy (10.5−16.7 kcal/mol) for complex IV of [Ru3+]− EDTA carbonyl species. Our calculations indicate that Od (Figures 4 and 5) is the most active oxygen site for the path T1 → T3 → T4 with a high apparent activation free energy of 55.2 kcal/mol. Thus, the direct WGSR path to produce CO2 and H2 (T1 → T3 → T4) without HCOOH intermediate is unlikely to be kinetically important for the catalyst we studied here. Comparison of the energetics of the two reaction paths, as shown in Figure 6, suggests that the dominant product for the reaction between H2O molecule and CO coordination complex [Ru(EDTA)(CO)]− should be formic acid, rather than the final WGSR products of CO2 plus H2. This conclusion is in line with the previous experimental observation.27,28,44 Consequently, the current computation result supports the plan to combine the synthesis and decomposition of HCOOH in the same reacting apparatus to produce CO-free H2.27,28 Reaction Catalyzed by [Ru(EDTA)CO]0 and [Ru(EDTA)CO]2−. The d5 catalytic species [Ru(EDTA)(CO)]− should be a radical with one unpaired electron. Spin analyses show that no less than 90% unpaired spins localize at the ruthenium center E

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 4. Optimized structures of the decarboxylation transition state T3 from MCA intermediate to generate CO2, in which the hydrogen transfers from [Ru]−COOH to [Ru] center (bond length in Å). Numbers in parentheses are the imaginary frequencies in cm−1.

the Table, although [Ru2+]-IV is more competitive than [Ru2+]III isomer in all possible reaction paths, the apparent reaction barrier for [Ru2+]-IV is always too high to be kinetically significant. The apparent activation free energies for the preparation of HCOOH (path T1 → T2) over the nonradical catalyst [Ru2+]-IV are as high as 46.0−48.1 kcal/mol. Moreover, the calculated apparent activation free energies for producing CO2 and H2 directly (path T1 → T3 → T4) are even higher at 68.1−73.8 kcal/mol. The results suggest that both the nonradical catalysts [Ru(EDTA)(CO)]0 and [Ru(EDTA)(CO)]2− are catalytically incompetent for both the preparation of formic acid and the direct WGSR in current experimental conditions.27,28,44 Reaction Catalyzed by [Ru(HEDTA)CO]0 in Acidic Condition. The catalytic paths explored in previous sections did not indicate any consumption or generation of proton in the reaction. Thus, the apparent reaction rate should be independent of pH. This is in line with previous experimental observation.44 But the same experiment also reported that disodium hydrogen phosphate (Na2HPO4) and potassium dihydrogen phosphate (KH2PO4) in appropriate portions were required to achieve a desired pH buffer to preserve the catalytic species in the reaction mixture. Some recent experiments also indicated that this reaction could not occur without an appropriate buffer solution.27 The dependence on the pH of

The second part of Table 1 summarizes the energies calculated for the nonradical catalyst [Ru(EDTA)(CO)]0. Surprisingly, our calculations show that the activation free energies for the addition of H2O molecule over [Ru4+]-IV is very low, at only 0.9−4.3 kcal/mol. However, the proton transfer process in the next step of the formation of HCOOH (T2) is kinetically infeasible, with an activation free energy of 30.3−32.6 kcal/mol. Compared with the parallel path catalyzed by the radical species, the height of this barrier is higher by ∼7 kcal/mol. Moreover, our calculations show that the hydrogen transfer from [Ru]−COOH to [Ru] center to produce CO2 (T3) over [Ru4+]-IV is kinetically infeasible with an activation free energy of 56.0−61.6 kcal/mol. Clearly, both paths for the formation of aqueous HCOOH (T1 → T2) and for the direct production of CO2 and H2 (T1 → T3 → T4) are kinetically prohibitive via the nonradical catalyst [Ru(EDTA)(CO)]0. Recent experiment also suggested that [Ru(EDTA)CO]− is sensitive in the air, which can be easily oxidized to a ruthenium species in +4 valence and becomes deactivated.27 Our calculations therefore provided a sound explanation for the experimental observations. The third part of Table 1 summarizes the calculated energies for the nonradical catalyst [Ru(EDTA)(CO)]2−. The [Ru2+]III isomer is much more favored than [Ru2+]-IV isomer with a free energy difference of 16.5 kcal/mol. Hence it would be difficult for [Ru2+]-IV to be an efficient catalyst. As shown in F

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 5. Optimized structures of the transition state T4 for the proton-hydride coupling between HEDTA and [Ru]−H to obtain H2 gas (bond length in Å). Numbers in parentheses are the imaginary frequencies in cm−1.

Table 1. Energetics of Important Stationary States Catalyzed by [Ru]−EDTA Complexa (1) [Ru3+]−EDTA T1-III T1-IV T2-III T2-IV T3-III T3-IV T4-III T4-IV

Oa

Ob

Oc

23.2 20.2 43.6 26.5 83.4 56.4 37.1 16.7

24.3 23.2 63.1 23.1 76.9 55.8 51.7 11.8

27.8 19.5 58.8 27.0 73.6 60.4 33.6 16.5

(2) [Ru4+]−EDTA

(3) [Ru2+]−EDTA

Od

Oa

Ob

Oc

Od

23.8

4.3

1.4

3.9

0.9

22.5

30.3

31.6

32.6

31.2

55.2

56.0

61.6

10.5

17.5

17.6

Oa

Ob

Oc

42.7 48.1 52.4 36.6 90.1 70.7 56.4 41.0

45.3 46.5 59.4 32.3 78.5 68.1 70.3 35.6

43.4 46.0 57.3 36.8 77.2 73.8 47.8 41.2

(4) [Ru3+]−HEDTA Od

Oa

Ob

Oc

24.2

26.9

29.0

43.6

62.1

60.3

76.9

77.4

74.4

37.9

52.7

33.5

46.1 32.3 67.6 36.0

a Relative activation free energies are computed with respect to the H2O and most stable form of corresponding [Ru]−EDTA carbonyls. The first, second, third, and fourth part correspond to reactions catalyzed by [Ru3+]−EDTA, [Ru4+]−EDTA, [Ru2+]−EDTA, and [Ru3+]−HEDTA complexes, respectively. [Ru3+]−EDTA, [Ru4+]−EDTA, and [Ru2+]−EDTA stand for the catalyst in buffer solution with ruthenium center in +3, +4, and +2 valence state, respectively. [Ru3+]−HEDTA stands for protonated catalyst in acidic condition with ruthenium center in +3 valence state. Oa, Ob, Oc, and Od denote different oxygen active sites. All values are in kcal/mol.

the buffer solution seems to be contradictory to the pHindependence of the reaction path. The inconsistency can be satisfactorily explained by the HCOOH formation path reported in previous sections. The key factor is the protonation of EDTA in low pH. As shown in Figure 7, the presumed catalytically active form in acidic condition is now [Ru3+]−HEDTA-III. The protonated oxygen atom loses the coordination capability and moves away from ruthenium center. Consequently, the activation free energy for

the H2O addition process (T1) is 24.2 kcal/mol when Oa is the active site (Table 1, fourth part). However, the following proton transfer process (T2) for HCOOH formation is not kinetically favorable, with an activation free energy of 43.6 kcal/ mol at the Oa site. Thus, the overall T1 → T2 path for the HCOOH generation is kinetically infeasible with [Ru3+]− HEDTA-III catalyst. The direct path T1 → T3 → T4 for producing CO2 and H2 is disfavored too, with the activation G

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 6. Free energy profiles of formic acid formation and direct WGSR catalyzed by isomer IV of [Ru3+]−EDTA carbonyl complex. Path T1 → T2 leads to HCOOH formation, and path T1 → T3 → T4 leads to the direct formation of the WGSR products, i.e., CO2 and H2. The MCA intermediate [Ru(HEDTA)COOH]− is shared by both paths.

■

DISCUSSION The structure and energetics of four possible catalytic paths (T1, T2, T3, T4) for the preparation of formic acid or direct WGSR over the Ru−EDTA catalysts are determined in the current work. It is shown that the IV form of catalyst is always more active than the III form for both the preparation of aqueous HCOOH (T1 → T2) and the direct production of CO2 and H2 (T1 → T3 → T4). The charge analyses are performed to understand the origin of the difference of catalytic activity of different forms of catalyst. The electron density of Ru center for the IV form is higher than that for the III form (0.609 vs 0.704 au). The density difference is probably caused by the different coordination patterns between EDTA and Ru, in which some negatively charged oxygen atoms of EDTA move away from the Ru atom in the III form. The activation of CO in the formation of T1 in the current work is in fact achieved subtly and elegantly by the [Ru]− EDTA complex. Carbon monoxide is a typical π acid ligand. There is significant synergistic π* back-bonding effect when carbon monoxide bonds to transition metals. Normally, this back-bonding effect correlates to the activation of the C−O bond, which is often favored when carbon monoxide interacts with a metal atom with a relatively high electron density. On the other hand, T1 is formed as a concerted water addition process with the electrophilic carbon of [Ru]−CO being attacked by the lone-pair electrons of water oxygen; simultaneously, the nucleophilic oxygen atom of EDTA ligand is attacked by the water proton (Figure 8). To have an efficient, concerted formation of the C−O (COOH group) and O−H (HEDTA ligand) bonds to generate an MCA intermediate, the electron density in the ruthenium center must not be too high to reduce the electropositive character of the carbon atom of [Ru]−CO. Considering the two issues, the electron density in ruthenium center must reach a balance for kinetic performance. This seems to provide a conceivable explanation for the results that both complexes III and IV show a good performance for the T1 step. The contribution of the ruthenium center to the formation of T2 is even more important. To form T2 from the MCA

Figure 7. Selected structures for the nonradical catalysts [Ru(EDTA)(H2O)] ([Ru4+]-IV) and [Ru(EDTA)(CO)]2− ([Ru2+]-III and [Ru2+]-IV) and for the radical catalyst in acidic condition [Ru(HEDTA)(H2O)] ([Ru3+]−HEDTA-III)).

free energy for transition state T3 being as high as 74.4−76.9 kcal/mol. The calculation results of the [Ru3+]−HEDTA-III catalyst, the proposed dominant catalytic form in acidic condition, show low catalytic efficiency for both HCOOH generation and direct WGSR. Therefore, a buffer solution is required to avoid the acidification of the catalyst. This is very important since the product HCOOH is a strong acid and can release proton to lower the pH of the solution with the progress of the reaction. H

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

For both the HCOOH production and ordinary WGSR, the IV form is a more effective catalyst due to the appropriate electron density in the ruthenium center. However, acidification will change the structure of the catalyst, because the protonated oxygen atom of EDTA will move away from ruthenium center. This is an important factor to be considered since the reaction mixture will be constantly acidified with the increased production of HCOOH as the proton source. This renders the importance of pH buffer in the current reaction process. Theoretical and computational analysis reported in this work determines the active form of the catalyst for the preparation of HCOOH should be the d5 radical species of [Ru(EDTA)CO]− ([Ru3+]−EDTA). Our calculations also show that the nonradical species, [Ru(EDTA)CO]0 ([Ru4+]−EDTA) and [Ru(EDTA)CO]2− ([Ru2+]−EDTA), are ineffective to catalyze this process. Because T1 and T2 belong to an acid−base process, while T3 and T4 are the transition state for two-electron redox process, the radical state does not play significant mechanistic roles in those elementary processes. The key interaction for the effective catalysis lies in the proper amount of electron density in the metal center. As discussed, the electron density in ruthenium center correlated to the activity of catalyst. For nonradical species [Ru(EDTA)CO]0, T2 would not be kinetically favorable for the insufficient electron density in ruthenium center with a formal valence of +4 (0.788 au in ruthenium center). On the other hand, for [Ru(EDTA)CO]2−, the T1 step is not kinetically feasible because of the excessive electron density in ruthenium center with a formal valence of +2. This goes against the nucleophilic attacking of H2O to [Ru]−CO (0.292 au for III form and 0.282 au for IV form in ruthenium center). Furthermore, it has been shown that the [Ru 2+ ]-IV isomer cannot be an efficient catalyst for thermodynamic disadvantages. A catalytic mechanism has been proposed in a previous experimental report44 in which a water molecule is activated by oxidation-addition to the ruthenium center. HCOOH could then be formed in a succeeding reduction-elimination process.44 However, the mechanism reported in the present calculations does not show any involvement of the oxidationaddition and/or reduction-elimination in the metal center. The HCOOH formation path is simply an acid−base process. None of the elementary processes (T1 − T4) show a valence change of the ruthenium center. In addition to the ruthenium center, the EDTA ligand is also important for the catalyzed proton transfer process. In step T1, EDTA ligand acts as a proton acceptor, while, in T2 and T4 processes, EDTA ligand acts as a proton donor. Last but not least, EDTA ligand might also be important for protecting the radical center from being attacked by some other active species in solution, since the d5 complex [Ru(EDTA)CO]− ([Ru3+]− EDTA) is a radical species with a unpaired electron to be attacked easily.

Figure 8. Optimized structure, Mulliken charges, and Mulliken spin densities (in parentheses) of the most effective transition state (T1) for the addition of H2O molecule to form the MCA intermediate.

intermediate, both the carboxyl carbon of [Ru]−COOH and the hydrogen of HEDTA ligand are electropositive (Figure 9).

Figure 9. Optimized structures, Mulliken charges, and Mulliken spin densities (in parentheses) of MCA intermediate and the most effective transition state (T2) for the subsequent formation of formic acid.

Thus, it seems the formation of T2 could be disfavored due to the charge repulsion. The calculation results suggest that the participation of the ruthenium center is the key to solve this dilemma. The O−H bond of HEDTA ligand can be activated via the antibonding orbital accepting electrons from the ruthenium center. Furthermore, the electropositivity of both the hydrogen of the HEDTA ligand and the carbon of the [Ru]−COOH both can be reduced by the participation of ruthenium center (Figure 9). Thus, a high electron density in the ruthenium center should favor the formation of T2. This analysis expects that the IV form should be kinetically more efficient to form the transition state T2 as compared to the III form, because the IV form has a higher electron density in the metal center than the III form does. This is consistent with the calculation results reported in previous section. For the direct CO2 and H2 production path, decarboxylation of the MCA intermediate leads to the formation of T3 as the rate-limiting step. This process is a redox process, in which the positively charged proton of [Ru]−COOH changes to a negatively charged hydride as [Ru]−H. Higher electron density in the ruthenium center apparently will facilitate the redox process and promote the subsequent proton-hydride coupling process (T4). Generally, the IV form of catalyst with a higher electron density in metal center is more effective than the III form for the direct path to CO2 and H2, as shown by the energetic data reported in Table 1.

■

CONCLUSIONS After an extensive theoretical survey for the energetics and mechanism of the preparation of formic acid (T1 → T2) and direct WGSR (T1 → T3 → T4) catalyzed with the [Ru3+]− EDTA catalyst, we show that it is feasible to synthesize aqueous HCOOH via the T1 → T2 path as an intermediate of WSGR. This step might be further combined with HCOOH decomposition, which can be efficiently carried out with many currently available catalysts. The direct WGSR is kinetically very much disfavored for the [Ru3+]−EDTA catalyst; I

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

(15) Fellay, C.; Dyson, P. J.; Laurenczy, G. Angew. Chem., Int. Ed. 2008, 47, 3966−3968. (16) Fukuzumi, S.; Kobayashi, T.; Suenobu, T. ChemSusChem 2008, 1, 827−834. (17) Boddien, A.; Loges, B.; Junge, H.; Beller, M. ChemSusChem 2008, 1, 751−758. (18) Rieckborn, T. P.; Huber, E.; Karakoc, E.; Prosenc, M. H. Eur. J. Inorg. Chem. 2010, 4757−4761. (19) Boddien, A.; Mellmann, D.; Gaertner, F.; Jackstell, R.; Junge, H.; Dyson, P. J.; Laurenczy, G.; Ludwig, R.; Beller, M. Science 2011, 333, 1733−1736. (20) Ting, S.-W.; Cheng, S.; Tsang, K.-Y.; van der Laak, N.; Chan, K.Y. Chem. Commun. 2009, 7333−7335. (21) Ting, S.-W.; Hu, C.; Pulleri, J. K.; Chan, K.-Y. Ind. Eng. Chem. Res. 2012, 51, 4861−4867. (22) Hu, C.; Ting, S.-W.; Tsui, J.; Chan, K.-Y. Int. J. Hydrogen Energy 2012, 37, 6372−6380. (23) Gu, X.; Lu, Z.-H.; Jiang, H.-L.; Akita, T.; Xu, Q. J. Am. Chem. Soc. 2011, 133, 11822−11825. (24) Huang, Y.; Zhou, X.; Yin, M.; Liu, C.; Xing, W. Chem. Mater. 2010, 22, 5122−5128. (25) Zhou, X.; Huang, Y.; Liu, C.; Liao, J.; Lu, T.; Xing, W. ChemSusChem 2010, 3, 1379−1382. (26) Tedsree, K.; Li, T.; Jones, S.; Chan, C. W. A.; Yu, K. M. K.; Bagot, P. A. J.; Marquis, E. A.; Smith, G. D. W.; Tsang, S. C. E. Nat. Nanotechnol. 2011, 6, 302−307. (27) Baby, J. P. Low Temperature Catalytic Liquid Phase Water Gas Shift Reaction for Carbon Monoxide Free Hydrogen Generation. Ph.D. Thesis, The University of Hong Kong, 2014. In this thesis, the authors combined the CO hydration with HCOOH dehydrogenation in a customized reactor. The [Ru3+]-EDTA and Pt/Ru/BixOy (PRB) catalyzed the generation and decomposition of HCOOH, respectively. In the customized reactor, the product gas leaves through a chamber that is isolated from the reactant chamber, which is fed with CO and H2O, but the bottom parts of both chambers are connected with a common liquid phase where HCOOH is formed. This arrangement prevents mixing of reactant and product gases, and hydrogen generated from the aqueous phase can leave without CO contamination. (28) Pulleri, J. K.; Ting, S. W.; Lam, F. L. Y.; Hu, C.; Chan, K.-Y. In Generation of CO-Free Hydrogen by Liquid-Phase Bicatalytic WGS Reaction at 40°C via HCOOH Intermediate; AIChE 2012 Annual Meeting, Conference Proceedings; American Institute of Chemical Engineers: New York, 2012; p 510e. (29) Gao, W.; Keith, J. A.; Anton, J.; Jacob, T. Dalton Trans. 2010, 39, 8450−8456. (30) Zhou, S.; Qian, C.; Chen, X. Catal. Lett. 2011, 141, 726−734. (31) Luo, Q.; Feng, G.; Beller, M.; Jiao, H. J. Phys. Chem. C 2012, 116, 4149−4156. (32) Wang, H.-F.; Liu, Z.-P. J. Phys. Chem. C 2009, 113, 17502− 17508. (33) Gao, W.; Keith, J. A.; Anton, J.; Jacob, T. J. Am. Chem. Soc. 2010, 132, 18377−18385. (34) Hu, C. Q.; Ting, S. W.; Chan, K. Y.; Huang, W. Int. J. Hydrogen Energy 2012, 37, 15956−15965. (35) Hu, C. Q.; Ting, S. W.; Chan, K. Y.; Huang, W. Int. J. Hydrogen Energy 2013, 38, 8720−8731. (36) Lu, X.; Leung, D. Y. C.; Wang, H. Z.; Leung, M. K. H.; Xuan, J. ChemElectroChem 2014, 1, 836−849. (37) Leitner, W. Angew. Chem., Int. Ed. 1995, 34, 2207−2221. (38) Jessop, P. G.; Ikariya, T.; Noyori, R. Chem. Rev. 1995, 95, 259− 272. (39) Jessop, P. G.; Joo, F.; Tai, C. C. Coord. Chem. Rev. 2004, 248, 2425−2442. (40) Himeda, Y. Eur. J. Inorg. Chem. 2007, 3927−3941. (41) Federsel, C.; Jackstell, R.; Beller, M. Angew. Chem., Int. Ed. 2010, 49, 6254−6257. (42) Cokoja, M.; Bruckmeier, C.; Rieger, B.; Herrmann, W. A.; Kuhn, F. E. Angew. Chem., Int. Ed. 2011, 50, 8510−8537.

thus, the catalyzed synthesis of HCOOH is highly selective. The active catalyst [Ru(EDTA)CO]− is a radical, but the HCOOH synthesis reaction shows no correlation to the radical character of the catalyst. Instead, the proper charge/valence state of the metal center plays the major catalytic role. Calculation results suggest that both ruthenium center and EDTA ligand take part in the catalytic processes. The ruthenium ion in a +3 valence state has a suitable electron density to facilitate the T1 and T2 processes to produce HCOOH. The EDTA ligand acts as both the proton acceptor and donor to catalyze the reaction. Moreover, the ruthenium radical center might be protected by chelate EDTA ligand from being attacked by some other active species in solution. This detailed mechanistic information provides a clear picture to facilitate experimental design of more efficient catalysts for production of formic acid and further to prepare CO-free H2 at ambient temperature.

■

ASSOCIATED CONTENT

* Supporting Information S

Mulliken spin densities for all the transition states and important intermediates; optimized Cartesian coordinates of important structures in standard xyz file format. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors are thankful for the financial support from the Research Grants Council of Hong Kong, University of Hong Kong strategy research themes on the topics of “Clean Energy” and “Computation and Information” and from the National Science Foundation of China. J.K. is thankful for the technical assistance from Mr. D. John at the National Institute for Computational Sciences.

■

REFERENCES

(1) Vanvorst, W. D.; Kelley, J. H.; Veziroglu, T. N. Int. J. Hydrogen Energy 1983, 8, 857−866. (2) Schlapbach, L.; Zuttel, A. Nature 2001, 414, 353−358. (3) Lee, T. B.; McKee, M. L. Inorg. Chem. 2009, 48, 7564−7575. (4) Momen, G.; Hermosilla, G.; Michau, A.; Pons, M.; Firdaouss, M.; Hassouni, K. Int. J. Hydrogen Energy 2009, 34, 3799−3809. (5) Jacobs, G.; Davis, B. H., Low Temperature Water-Gas Shift Catalysts. In Catalysis; The Royal Society of Chemistry: Cambridge, U.K., 2007; Vol. 20, pp 122−285. (6) Laine, R. M.; Crawford, E. J. J. Mol. Catal. 1988, 44, 357−387. (7) Ford, P. C. Acc. Chem. Res. 1981, 14, 31−37. (8) Newsome, D. S. Catal. Rev.: Sci. Eng. 1980, 21, 275−318. (9) Ford, P. C.; Rinker, R. G.; Ungermann, C.; Laine, R. M.; Landis, V.; Moya, S. A. J. Am. Chem. Soc. 1978, 100, 4595−4597. (10) Sunderlin, L. S.; Squires, R. R. J. Am. Chem. Soc. 1993, 115, 337−343. (11) Enthaler, S. ChemSusChem 2008, 1, 801−804. (12) Joo, F. ChemSusChem 2008, 1, 805−808. (13) Johnson, T. C.; Morris, D. J.; Wills, M. Chem. Soc. Rev. 2010, 39, 81−88. (14) Loges, B.; Boddien, A.; Junge, H.; Beller, M. Angew. Chem., Int. Ed. 2008, 47, 3962−3965. J

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry (43) Fernandez-Alvarez, F. J.; Iglesias, M.; Oro, L. A.; Polo, V. ChemCatChem 2013, 5, 3481−3494. (44) Shukla, R. S.; Bhatt, S. D.; Thorat, R. B.; Jasra, R. V. Appl. Catal., A 2005, 294, 111−118. (45) Khan, M. M. T.; Chatterjee, D.; Merchant, R. R.; Paul, P.; Abdi, S. H. R.; Srinivas, D.; Siddiqui, M. R. H.; Moiz, M. A.; Bhadbhade, M. M.; Venkatasubramanian, K. Inorg. Chem. 1992, 31, 2711−2718. (46) Chatterjee, D. Coord. Chem. Rev. 1998, 168, 273−293. (47) Chatterjee, D.; Mitra, A.; De, G. S. Platinum Met. Rev. 2006, 50, 2−11. (48) Chatterjee, D.; Ember, E.; Pal, U.; Ghosh, S.; van Eldik, R. Dalton Trans. 2011, 40, 10473−10480. (49) Chatterjee, D.; van Eldik, R. J. Mol. Catal. A: Chem. 2012, 355, 61−68. (50) Shepherd, R. E. Inorg. Chim. Acta 1993, 209, 201−206. (51) Torrent, M.; Sola, M.; Frenking, G. Organometallics 1999, 18, 2801−2812. (52) Barrows, S. E. Inorg. Chem. 2004, 43, 8236−8238. (53) Rozanska, X.; Vuilleumier, R. Inorg. Chem. 2008, 47, 8635− 8640. (54) Zhang, F. L.; Zhao, L.; Xu, C. M.; Chen, Y. Inorg. Chem. 2010, 49, 3278−3281. (55) Chen, Y.; Zhang, F.; Xu, C.; Gao, J.; Zhai, D.; Zhao, Z. J. Phys. Chem. A 2012, 116, 2529−2535. (56) Kuriakose, N.; Kadam, S.; Vanka, K. Inorg. Chem. 2012, 51, 377−385. (57) Schulz, H.; Gorling, A.; Hieringer, W. Inorg. Chem. 2013, 52, 4786−4794. (58) Becke, A. D. Phys. Rev. A 1988, 38, 3098−3100. (59) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785− 789. (60) Petersson, G. A.; Bennett, A.; Tensfeldt, T. G.; Allaham, M. A.; Shirley, W. A.; Mantzaris, J. J. Chem. Phys. 1988, 89, 2193−2218. (61) Häu ssermann, U.; Dolg, M.; Stoll, H.; Preuss, H.; Schwerdtfeger, P.; Pitzer, R. M. Mol. Phys. 1993, 78, 1211−1224. (62) Kuchle, W.; Dolg, M.; Stoll, H.; Preuss, H. J. Chem. Phys. 1994, 100, 7535−7542. (63) Leininger, T.; Nicklass, A.; Stoll, H.; Dolg, M.; Schwerdtfeger, P. J. Chem. Phys. 1996, 105, 1052−1059. (64) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2009, 113, 6378−6396. (65) Mulliken, R. S. J. Chem. Phys. 1955, 23, 1833−1840. (66) Chai, J. D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10, 6615−6620. (67) Brookes, N. J.; Ariafard, A.; Stranger, R.; Yates, B. F. J. Am. Chem. Soc. 2009, 131, 5800−5808. (68) Laury, M. L.; Wilson, A. K. J. Chem. Theory Comput. 2013, 9, 3939−3946. (69) Zhang, W. J.; Truhlar, D. G.; Tang, M. S. J. Chem. Theory Comput. 2013, 9, 3965−3977. (70) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650−654. (71) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297−3305. (72) Ho, J. M.; Klamt, A.; Coote, M. L. J. Phys. Chem. A 2010, 114, 13442−13444. (73) Liu, G.; Wu, J. M.; Zhang, I. Y.; Chen, Z. N.; Li, Y. W.; Xu, X. J. Phys. Chem. A 2011, 115, 13628−13641. (74) Ribeiro, R. F.; Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2011, 115, 14556−14562. (75) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,

N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision D. 01, Gaussian, Inc.: Wallingford, CT, 2009. (76) Hasanayn, F.; Nsouli, N. H.; Al-Ayoubi, A.; Goldman, A. S. J. Am. Chem. Soc. 2008, 130, 511−521. (77) Nsouli, N. H.; Mouawad, I.; Hasanayn, F. Organometallics 2008, 27, 2004−2012.

K

DOI: 10.1021/ic5021127 Inorg. Chem. XXXX, XXX, XXX−XXX