Iron(III) Ejection from a “Beheaded” TAML Activator: Catalytically

Aug 22, 2017 - Kinetic studies of the acid-induced ejection of iron(III) show that the ... its entire locus from the head to the tail diamido-N sectio...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/IC

Iron(III) Ejection from a “Beheaded” TAML Activator: Catalytically Relevant Mechanistic Insight into the Deceleration of Electrophilic Processes by Electron Donors Matthew R. Mills,† Longzhu Q. Shen, David Z. Zhang, Alexander D. Ryabov,* and Terrence J. Collins* Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213, United States S Supporting Information *

ABSTRACT: Kinetic studies of the acid-induced ejection of iron(III) show that the more electron-rich tetra-amido-N macrocyclic ligand (TAML) activator [FeIII{(Me2CNCOCMe2NCO)2CMe2}OH2]− (4), which does not have a benzene ring in its head component (“beheaded” TAML), is up to 1 × 104 times more resistant than much less electronrich [FeIII{1,2-C6H4(NCOCMe2NCO)2CMe2}OH2]− (1a) to the electrophilic attack. This counterintuitive increased resistance is seen in both the specific acid (kobs = k1[H+]/(K + [H+])) and phosphate general acid (kII = (kdiKa1 + ktri[H+])/ (Ka1+[H+])) demetalation pathways. Insight into this reactivity puzzle was obtained from coupling kinetic data with theoretical density functional theory modeling. First, although 1a and related complexes are six-coordinate in water, 4 has a strong tendency to repel the second aqua ligand favoring [LFe(OH2)]− and making appropriate the comparison of monoaqua-4 with diaqua-1a in the demetalation process. Second, dearomatization exerts a strong effect on the highest occupied molecular orbital (HOMO) energy of five-coordinate monoaqua-4, the presumed target in proton-induced demetalation, stabilizing it by ca. 51 kJ mol−1 compared with monoaqua-1a. Third, the monoaqua-4 HOMO is localized over the N−pπ system of all four N donors in contrast with monoaqua-1a, where N−pπ contributions from the head amides only mix with the aromatic ring π system. Fourth, addition of a second water ligand to monoaqua-1a giving [LFe(OH2)2]− reshapes the monoaqua-1a HOMO by shifting its entire locus from the head to the tail diamido-N sectionthis HOMO is by 54 kJ mol−1 less stable than the monoaqua-4 HOMO. These features provide the foundations for mechanistic conclusions concerning demetalation that (i) axial water ligands enable a favored path in the six-coordinate case of 1a, where a proton “slides” toward the Fe−N bond and (ii) early and late transition states are realized for 4 and 1a, respectively, with a larger free energy of activation for the beheaded TAML activator 4.



catalysis.11,12 As a result, TAML applications under acidic conditions were limited prior to the development of demetalation-resistant TAML activators, an advance that enabled studies of oxygen evolving catalysis for water oxidation at pH ≈ 1.13−15 Under the more common uses of TAML systems, catalysis of oxidations by H2O2 at neutral to moderately basic pH, ground-state instabilities play no detectable role. Instead, operational instability manifested in the cycling catalyst system is dominated by decay of the oxidized iron intermediate.5−7 For some time, we thought that the ground-state instability in TAML activators was wellunderstood and there was little else to be learned. But by examining the ground-state chemistry of the “beheaded” TAML catalyst 4 (Chart 1), a new dimension of acidic instability has been revealed that has importance to the mechanistic chemistry of catalysis and promises to direct iterative design to improve catalyst technical performances.

INTRODUCTION The efficacy of homogeneous oxidation catalysts can be influenced in principle by decomposition chemistry at any step in a catalytic cycle.1 FeIII-TAML activators (Chart 1; TAML = tetra-organic-N macrocyclic ligand complexes) are highly active catalysts that utilize hydrogen peroxide in water to achieve deep oxidative degradation of organic compounds including environmental pollutants.2−4 Comparative technical performances across the family are typically determined by the ratios of substrate oxidation to catalyst decomposition.5−7 However, TAML catalyst decomposition chemistry has been clearly characterized at both the resting FeIII state and the oxidized states. Decomposition insight gained from both these states has turned out to be valuable in catalyst and process design work, including identifying optimal reaction conditions, predicting substituent effects on catalytic efficacy and understanding how to measure catalyst lifetimes under various conditions.1,8,9 TAML ground-state instability has been found to originate exclusively from iron ejection from the macrocycle, which is subject to both specific acid10 and general acid © 2017 American Chemical Society

Received: April 14, 2017 Published: August 22, 2017 10226

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry

Kinetic Studies. Kinetic studies were conducted at 15, 25, 35, 40, 45, and 60 °C. Stock solutions of 4, 1a, perchloric acid, and phosphoric acid were prepared using high-performance liquid chromatography (HPLC) grade water. Concentrated NaOH and HClO4 were used for pH adjustments. Appropriate volumes of perchloric and phosphoric acids (if necessary) were added to quartz cuvettes, and the reactions were initiated by the addition of an aliquot of 4 or 1a stock solutions. The reaction progress was followed by measuring the decrease in absorbance at 368 nm (λmax for 4, ε = 7200 M−1 cm−1) or at 366 nm (λmax for 1a, ε = 4700 M−1 cm−1). The pseudo-first-order rate constants kobs were calculated from the absorbance versus time plots for exponential decay of 1a or 4 by fitting the data to the equation A = A∞ − (A∞ − A0) × e−kobst when the degree of demetalation was taken to greater than or equal to 90%. Each reported value of kobs is a mean value of three measurements. DFT Calculations. The DFT calculations were performed with Gaussian 09, rev. B.01,19 using Becke’s three-parameter hybrid functional (B3)20,21 along with the Lee−Yang−Parr correlation functional (LYP)22 and triple-ζ basis set 6-311G. The effect of water on the electronic states and optimized geometries was evaluated using the SMD continuum model.23 The default convergence criteria were adopted in geometrical optimizations. The water-attacking reaction energy profile was generated using relaxed potential surface scan.

Chart 1. TAMLs of Generations One (1), Four (2), and Five (3)18 Plus Beheaded TAML 4, Which Shows Enhanced Resistance to Acid-Induced Iron Ejectiona

a Compound 5 was not synthesized; it and its Me4−nHnC2 (n = 1−3) derivatives were used for theoretical modeling only. Key for (X1, X2, R): a: (H, H, Me), b: (NO2, H, Me), c: (Cl, Cl, Me).



In the absence of general acids, ground-state demetalation of TAML catalysts begins to be apparent at pH values below 4.10 The speed of this process is significantly controlled by the nature of the “tail” part of TAML molecules (Chart 1). In particular, TAMLs of the first generation 1 can acquire 10 orders of magnitude in ground-state stability when tail methyl groups are replaced by electron-withdrawing fluorine atoms.10 Similarly, electron-withdrawing groups in TAML activators of the fourth generation (2, Chart 1) increase the ground-state stability.12 TAML 3, which contains a “tail” biuret moiety, also exhibits higher stability under acidic conditions.16 The examples summarized above pointed to the fact that, if electronwithdrawing substituents disfavor iron ejection, electrondonating substituents should facilitate it. Therefore, when recently synthesized7 electron-rich 4 was found to be 10 000fold more resistant to acid-induced demetalation than its prototype 1a,17 a counterintuitive stability-related mechanism became apparent, that is, substitution of the more electrondonating Me4C2 moiety for the less electron-donating aromatic ring in the “head” part of 1a disfavors an electrophilic process. To resolve this apparent incongruity, we looked at the demetalation behavior of 4 as a structural and mechanistic challenge and undertook detailed kinetic and computational studies of this novel mechanistic changeover for TAML science. Here we report the results of kinetic investigations of iron(III) ejection from TAML activator 4 induced by HClO4 (specific acid catalysis) and H3PO4 (general acid catalysis), both performed under variable temperature, together with results of structural and reactivity-probing computational studies. Both kinetic and density functional theory (DFT) results disclosed different behavior for 4 compared to the prototype TAML 1a that appear to be helpful in understanding properties of TAML activators relevant to catalysis.



RESULTS AND DISCUSSION HClO4−Induced Demetalation of 4. It is well-established that TAML activators of peroxides such as 1−3 undergo proton-induced demetalation10,16 at pH less than 4 and are subject of general acid demetalation at mildly acidic and neutral pH.11,12 Both types of iron(III) ejection proceed according to eq 1. [LFe(OH 2)2 ]− + 4H+ → H4L + Fe III

(1)

Complex 4 is demetalated similarly, and it has been previously reported that the pseudo-first-order rate constant kobs was a linear function of [HClO4] in the concentration range of 0.0025−0.25 M; that is, kobs = k1*[H+].17 Such kinetic behavior contrasted with that of TAMLs 1 and 3, for which an extra third-order term in [H+] was established (eq 2).10,16 kobs = k1*[H+] + k 3*[H+]3

(2)

It was very possible that for 4 the third-order pathway in H+ was trivially unobserved in the studied range of H + concentrations.17 Therefore, we extended the concentration of HClO4 to 2.5 M (10-fold) and discovered a new dependence of kobs on [HClO4], which is presented in Figure 1. Unexpectedly, “deceleration”, that is, the leveling of the rate of demetalation rather than “acceleration”, was observed with increasing [H+]. The inverse kobs appeared to be a linear function of inverse [H+] consistent with eq 3. kobs =

k1[H+] K + [H+]

(3)

Equation 3 is consistent with the initial protonation of 4 followed by the rate-limiting collapse of the protonated intermediate (Scheme 1). The leveling of kobs is observed at a very high [HClO4], and therefore it could result from the ionic strength effect. Thus, similar measurements were performed at [HClO4] + [NaClO4] = 2.5 M, which revealed that the presence of NaClO4 has an insignificant effect on kobs (see Figure S1 of Supporting Information). Fitting the data in Figure 1 to eq 3 resulted in the values of k1 and K of (3.9 ± 0.2) × 10−4 s−1 and 0.69 ± 0.07 M,

EXPERIMENTAL SECTION

Materials. Perchloric acid (70%) and phosphoric acid (85%) were purchased from Aldrich and used as received. Compound 1a is a GreenOx product. The tetramethylammonium salt of 4 was synthesized as previously described.7 Instrumentation. Spectrophotometric measurements were performed on a Hewlett-Packard Diode Array spectrophotometer (model 8452A) equipped with an automated and thermostated 8-cell positioner or with a Shimadzu UV-1800 instrument. Solution pH was determined using a Corning 220 pH meter. 10227

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry

Figure 1. Dependence of kobs on HClO4 concentration for demetalation of 4 (5 × 10−5 M) via eq 1 at 25 °C. The solid line was calculated using the best-fit values of k1 and K (eq 3).

Figure 2. Dependencies of kobs for the acid-induced demetalation of 4 as a function of HClO4 concentration at 25, 40, 50, and 60 °C, from the slopes of which the rate constants of (3.41 ± 0.05) × 10−4, (9.32 ± 0.07) × 10−4, (18.2 ± 0.3) × 10−4, and (32.5 ± 0.7) × 10−4 M−1 s−1, respectively, were calculated.

Scheme 1. Stoichiometric Mechanism of Demetalation of 4 Consistent with eq 3

1), were calculated using the Eyring plot shown in Figure S2 (Supporting Information). respectively. The calculated second-order rate constant k*1 = k1K−1 equals 5.6 × 10−4 M−1 s−1. This value is slightly higher than that obtained at [HClO4] below 0.25 M (3.4 × 10−4 M−1 s−1),17 which is absolutely normal in view of the different kinetic equations used, specifically, kobs = k*1 [H+] and eq 3. Different rate laws, specifically, eqs 2 and 3, are not the only noteworthy dissimilarity in the kinetic behavior between 1−3 versus 4. More strikingly, the k*1 observed for the compound with the highest donor capacity TAML ligand system, 4, is extremely low. The difference in k1* between 1a and 4 is ∼4 orders of magnitude, which defied prediction, since the beheaded TAML 4 is more electron-rich primarily due to replacement of sp2 carbon at two head N donors by sp3 carbon. The electronegativity of sp2 carbon is higher than that of sp3 carbon,24 and this effect is obviously stronger than the effect of the electron-donating methyl groups on the Me4C2 linker (see Theoretical Section). This electronic effect manifests itself in an increased value of pKa of the axially coordinated aqua ligand, which equals 10.3 and 11.4 for 1a18 and 4,17 respectively (eq 4). Therefore, complex 4 should be more susceptible toward electrophilic attack by the hydrated proton. The counterintuitive reactivities prompted us to collect additional kinetic data that could assist in accounting for the mechanistic dissimilarities. As a first step, we determined representative activation parameters for the demetalation of compounds 1a and 4 under acidic conditions. [LFe(OH 2)2 ]− ⇄ [LFe(OH)(OH 2)]2 − + H+

Table 1. Enthalpy (ΔH‡), Entropy (ΔS‡), and Free Energy (ΔG‡) of Activation (at 25 °C) for the k1* Pathways for 1a and 4 and the k3* Pathway for 1a TAML

rate constant

ΔH‡/kJ mol−1

ΔS‡/J mol−1 K−1

ΔG‡/kJ mol−1

1a 1a 4

k1* k3* k1*

21 ± 4 14.4 ± 0.9 50.8 ± 0.6

−170 ± 20 −95 ± 3 −141 ± 5

72 ± 14 43 ± 2 93 ± 5

The temperature dependencies of kobs for complex 1a at variable concentrations of HClO4 are shown in Figure 3. The rate constants are strongly curved upward, as it has been previously established.10,16 The kinetic data were fitted to eq 2. The rate constants k*1 and k*3 obtained are collected in Table 2. The corresponding Eyring plots are shown in Figures S3 and S4

(4)

Activation Parameters for the Acid-Induced Demetalation of 1a and 4. The kobs values for the beheaded activator 4 were collected in the temperature range of 25−60 °C. The H+ concentrations were varied from 0.1−0.5 M. Under such conditions eq 3 simplifies to the form kobs ≈ k1K−1[H+] = k*1 [H+] in agreement with the experimental data shown in Figure 2. The values of k*1 at variable temperatures were determined from the slopes of the straight lines. The corresponding activation parameters, ΔH‡ and ΔS‡ (Table

Figure 3. Dependencies of kobs for the acid-induced demetalation of 1a as a function of HClO4 concentration at 15, 25, 35, and 45 °C from bottom to top, respectively. The solid lines were calculated using the best-fit values of k1* and k3* (eq 2). 10228

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry (Supporting Information), and the calculated activation parameters are placed in Table 1. Table 2. Temperature Dependence of the Rate Constants for the Demetalation of 1a temperature/°C 15 25 35 45

k1*/M−1 s−1 1.2 1.6 2.6 2.8

± ± ± ±

k3*/M−3 s−1

0.3 0.3 0.5 0.5

(1.6 (2.1 (2.5 (3.1

± ± ± ±

0.1) 0.1) 0.2) 0.2)

× × × ×

105 105 105 105

The activation parameters for the k*1 pathway for complexes 1a and 4 collected in Table 1 reveal that the reactivity differential has an enthalpic origin; that is, the value of ΔH‡ for the demetalation of 4 is almost 30 kJ mol−1 higher than that for 1a. The values of ΔS‡ for 1a and 4 are much closer. General Acid-Induced Demetalation of 4. The differing kinetic behavior of TAML activators 1a and 4 described above prompted us to compare closely the speed of iron(III) ejection induced by general acids for both.11 Complexes 1a11 and 2a12 were found to exhibit measurable rates of demetalation in phosphate buffer at neutral pH with H2PO4− as the reactive demetalating species. A mechanistically similar process involving the less electron-rich compound 2b occurred at mildly acidic conditions, the reactive species this time being both H2PO4− and H3PO4.12 Thus, it was inspiring to probe 4 as a target for general acid-induced demetalation and to compare its kinetic features with those of 1a. Complex 4 is quite stable at pH 7 in the presence of 0.01 M phosphate. No iron ejection was detected at 25 °C for 16 h. The effect of phosphate becomes noticeable at pH 2 when kobs for reaction 1 is 30 times higher in the presence of 0.5 M phosphate as compared to kobs in its absence (7.5 × 10−5 vs 2.5 × 10−6 s−1). As is typical of a general acid mechanism,25 the values of kobs depend linearly on the total phosphate concentration in the range of 0.2−1.5 M without noticeable intercepts in this case (Figure 4A), that is kobs = kII[phosphate]

Figure 4. (A) Values of kobs for demetalation of 4 as a function of total phosphate concentration at different pH. (B) Rate constants kII (slopes of straight lines such as in (A)) against pH at 25 °C. See text for details. The solid line was calculated using the best-fit values of ktri and Ka1 (eq 6).

(5)

Scheme 2. Stoichiometric Mechanism of General Acid Demetalation of 4

In turn, the second-order rate constants kII increase with decreasing pH in the range of 0.5−3.0 as shown by representative examples in Figure 4B confirming a generalacid mechanism. A two-plateau sigmoidal dependence of kII on pH in Figure 4B indicates two reactive species, namely, H2PO4− and H3PO4, that demetalate 4 with the rate constants kdi and ktri, respectively, the subscripts “di” (H2PO4−) and “tri” (H3PO4) referring to the di- and triprotonated phosphate, respectively (Scheme 2). Thus, the expression for kII is given by eq 6. kII =

kdiK a1 + k tri[H+] K a1 + [H+]

is, when the reaction rate does not practically depend on pH; that is, kII ≈ ktri, and ΔH° for H3PO4 is low.26 The corresponding Eyring plot is shown in Figure S6, from which the value of ΔH‡ (64.4 ± 0.5 kJ mol−1) was calculated and used together with the value of ktri = 9.1 × 10−4 M−1 s−1 for obtaining the entropy of activation ΔS‡ (−87 ± 8 J mol−1 K−1). The kinetic data obtained for the phosphate-induced demetalation of 4 are summarized in Table 3 where the relevant data for other TAML activators (1a and 2a,b) are also included for comparison. The trend, which was revealed previously for the H+-induced iron(III) dissociation, holds here as well4 is significantly more resistant to general acids. The benchmark comparison should now be made with compound 2b, for which the resistance to the H3PO4 species is additionally

(6) +

The dependence of kII on [H ] shown in Figure S5 (Supporting Information) suggests that the kdiKa1 term in eq 6 is negligible. Fitting the data in Figure S5 to eq 6 resulted in the rate constants kdi and ktri of ∼0 and (9.1 ± 0.8) × 10−4 M−1 s−1, respectively, and the equilibrium constant Ka1 of (3.9 ± 0.9) × 10−2 M (pKa1 = 1.4 ± 0.2). The latter agrees with the previously reported pKa1 values of 1.726 and 1.827 for pKa1 of H3PO4. Since kdi ≈ 0, the activation parameters for ktri were calculated from the temperature dependence of kII at pH 1, that 10229

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry Table 3. Rate Constants (both in M−1 s−1), Enthalpies (ΔH‡, in kJ mol−1), Entropies (ΔS‡, in J mol−1 K−1), and Free Energies (ΔG‡, in kJ mol−1) of Activation for the General Acid Pathway of Demetalation of Complexes 1a, 2a,b, and 4 TAML 4 4 1a 2a 2b 2b a

pathway −

H2PO4 H3PO4 H2PO4− H2PO4− H2PO4− H3PO4

ka,b a, c 9.1 × 10−4 b 1.3 × 10−3 a 0.12a 3.0 × 10−3 a 0.31b

ΔH‡

ΔS‡

ΔG‡

reference this work

64.4 47

−87 −143

90.3 89.6

42

−150

86.7 75.8

11 12 12 12

kdi. bktri. cToo small to be measured.

enhanced by two NO2 groups. Nevertheless, comparatively very electron-rich 4 is more stable than 2b by a factor of 340! Computed Properties of 4 Inspired by the Kinetic DataComparative HOMO-Based Analyses. The deviation from predicted susceptibility toward acid-induced demetalation revealed by 4 shows that insight into electron donating/withdrawing effects alone, that have previously always been reliable in guiding the design of TAML activators, do not provide a clear clue for the enhanced resistance of 4 compared to that of 1 and 2. Another line of interpretation must be found. First, the four potentially sterically consequential aliphatic methyl groups in 4, as opposed to 1 with its flat aromatic ring, could affect 4 to override electronic effects. However, given the sterically undemanding nature of the proton, such an explanation for the differing specific acid behavior is hard to envision. Second, proton attack could depend on the highest occupied molecular orbitals (HOMOs) of 1a and 4, where significant variation in energies and distributions could result from the structural differences. Third, impacts on axial coordination need also to be considered. As an opening line of investigation, the HOMOs of 1a and 4 were calculated, and the relative energies were compared to immediately highlight the significance of the latter two lines of thought. The HOMO and sHOMO of 1a have been reported previously.11 Here, extra attention was paid to the HOMO of both the mono- and diaqua forms of 1. The calculations (see below) reveal that the diaqua form is problematic for 4 but accessible for 1a, where the latter result is in agreement with experimental observation.10 Therefore, the diaqua form of 1a was compared theoretically with the monoaqua forms of both 1a and 4 in Figure 5. As can be seen, the two HOMOs of 1a vary significantly, though the energy gap between them is just 3.36 kJ mol−1 with the second H2O ligand causing a slight destabilization. Interestingly, if the monoaqua form of 1 were the dominant species in water, proton attack would be expected at head amide nitrogens (Figure 5A). Since the diaqua species dominates, attack at the tail nitrogens is the major expected pathway (Figure 5B).10 The HOMO of 4 (Figure 5C) is different both geometrically and energetically. The electron density is distributed evenly across both the “head” and “tail” amide nitrogens, implying that proton attack at either end is equally realistic. The relative energies of [LFe(OH2)]− HOMOs of 1a and 4 are strikingly different, the latter being stabilized by 50−55 kJ mol−1 (Scheme 3). We interpret this large energy difference as resulting in part from (i) dearomatization of the ligand in 4 and in part from (ii) the increased electron-donating effect of the Me4C2 linkage. It is important to note that the dominant species in water for 1a is B in which the HOMO is localized exclusively on the tail

Figure 5. HOMOs of [LFe(OH2)]− (A) and [LFe(OH2)2]− (B) forms of 1a and [LFe(OH2)]− form of 4 (C). Hydrogens other than of H2O are omitted for clarity. The calculated energy differences between [LFe(OH2)]− of 4 and [LFe(OH2)]− and [LFe(OH2)2]− of 1a equal −50.8 and −54.2 kJ mol−1 (Scheme 3).

amides. That the electronic effects of the four methyl substituents on the Me4C2 linker are minor was supported by calculations of the HOMO energies of structural resemblants of 4, in which methyl groups were one by one replaced by hydrogens to generate the limiting-case hypothetical species 5 (Chart 1) with four hydrogen atoms at the H4C2 linker. The HOMO energies of complexes with one, two (at different carbons), three, and four methyl groups have values of 0.26, 0.27, 1.8, and −4.9 kJ mol−1, respectively, relative to 5. Thus, the electron-donating effect of the methyl groups in 4 compared to 1 is unlikely to be a significant factor in the demetalation reactivity but rather the dearomatization/ rehybridization in 4, that is, replacement of the phenylene ring by the aliphatic unit. As a result, two less electronegative sp3 carbons instead of two more electronegative sp2 carbons affect the character of the HOMO. Computed Properties of 4 Inspired by the Kinetic Datathe Coordination Number of the Iron(III) of 4 in Water. Compound 1a is a square pyramid in the solid state and an octahedral species with axial aqua/hydroxo ligands in water.10,28 Our preliminary attempts to optimize the geometry of 4 in a form [LFe(OH2)2]− reveal that the Fe−O bond distance of the second H2O ligand is significantly longer compared to the first one17 suggestive of the equilibrium shown 10230

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry Scheme 3. Experimental and Calculated Energy Differencesa of [LFe(OH2)]− and [LFe(OH2)2]− Species of 1a, 2b, and 4

structure for the hypothetical species 5 failed. The structural design of the 4 family of TAML activators does not accommodate two axial aqua ligands as much as the 1 family does. The consequences will be discussed below. Mechanistic Considerations. First Round. “A case of steric hindrance toward the proton” is not very newa quoted fragment from the title of the paper by Brown and Kanner from 1966.29 The authors have pointed to the pKa values of pyridine and its 2-tBu and 2,6-tBu2 derivatives (4.38, 4.68, and 3.58, respectively). Though the first two values are expectedly normal, the pKa of the disubstituted by the bulky groups molecule is counterintuitively smaller, which was “attributed to steric effects accompanying addition of the proton” to the free base.29 This rationale is incompatible with transfer to the present case if for no other reason than the second-order rate constants k1* for both 1a and 4 are well below the diffusion control limit,30 which is typical of protonation of bases.25 Additionally, different rate laws for 1a and 4 (cf. eqs 2 and 3, respectively) implicate mechanistic differences for the protoninduced demetalation of 1a and 4. Consider first the behavior of 4 as depicted in Scheme 1. The question that must be asked is “what is the site of protonation in {4,H+}.” The pK value found experimentally (−0.16) is higher than the pKa of the hydronium ion (−1.7),25 which practically rules out the option that pK = −0.16 corresponds to the protonation of axially coordinated waterLewis acids decrease pKa’s of coordinated ligands including H2O.31 Also, the pK = −0.16 could reflect the protonation of the amide oxygens of 4, because the amide functionalities in TAML activators are planar (objective for oxygen over nitrogen protonation),32,33 and pKa’s of free amides are around zero (acetamide) or −2 (benzamide).25 The HOMO of 4 shown in Figure 5C does not rule out oxygen protonation. It should however be mentioned that the pKa values of free amides should differ from those in TAML activators 1−4. Here, while protonic hydrogen is replaced by the iron cation of charge 3+, perhaps suggesting that the pKa’s could be more negative. The spectrum of 4 at [H+] up to 2.5 M remains practically unchanged. Thus, there is insufficient experimental evidence to permit identification of the site of protonation for {4,H+}. Therefore, we investigated this problem theoretically. The results and conclusions are presented in the next section. Theoretical Investigation of Protonation of 1 and 4 in Water. The protonation of the [LFe(OH2)2]− and [LFe(OH2)]− forms of 1a and of the [LFe(OH2)]− form of 4 was investigated. No energy minimum was detected for [LFe(OH2)2]− of 1a; the proton goes immediately to the tail nitrogen forming the N−H bond (Figure S8). In contrast, the [LFe(OH2)]− forms of 1a and 4 reveal energy minima in both cases. Diverse behavior of mono- and diaqua forms of 1a is consistent with the kinetic data for 1a and 4; just the mono diaqua transition increases the reactivity toward proton. At the same time the monoaqua form of 1a could be considered as a frozen, that is, less reactive, model of the diaqua form, which could suggest possible intrinsic features of this rate-limiting step10 for its more reactive [LFe(OH2)2]− form. The computed structures of protonated [LFe(OH2)]− forms of 1a and 4 are presented in Figure 6A,B, respectively. The proton locations in {1a,H+} and {4,H+} vary. In both cases, protonation leads to a pronounced elongation of the Fe−O bond (data are in Figure 6). In the case of 1a the incoming proton finds itself between the oxygen atom of the axial water and the tail amide nitrogen (Figure 6A) with N···H+ and O···

a (A) Calculated HOMO energies of 1a and 4; (B) ΔG≠ for protoninduced demetalation of 1a and 4; (C) triphosphate-induced demetalation of 2b and 4.

in eq 7 lying to the right. This fact prompted us to look more deeply at the decrease in coordination number of iron(III) from six to five in 4 in light of the dearomatization/rehybridization effect. [LFe(OH 2)2 ]− ⇄ [LFe(OH 2)]− + H 2O

(7)

Thus, the abundance of six versus five coordinate iron(III) TAML species has been explored theoretically for 1a and 4. In both cases, the [LFe(OH2)]− species (with “resident” water) were set static with the Fe−O bond distance to 2.2 Å. The second H2O molecule (“incoming” water) was allowed to move toward iron(III) along the opposite axial direction. The resident Fe−OH2 bond distance and the energy change were both monitored as a function of the separation distance between iron(III) and the incoming water oxygen. The distance−energy plots for 1a and 4 are shown in Figure S7 (Supporting Information). For 1a, a local energy minimum is observed at a distance of ca. 2.35 Å when the incoming Fe···O separation was varied in the range from 3.0 to 2.15 Å (Figure S7A). Thus, a comparable value for the Fe−OH2 bond distance in [LFe(OH2)2]− is observed for 1a compared to 2.097 and 2.11 Å found by X-ray crystallography in square pyramid [LFe(OH2)]− species 1a10 and 2c,12 respectively. The distance−energy plot for 4 is different (Figure S7B). As the Fe···O separation decreases, the energy increases much faster than for 1a followed by a sharp energy decline at the Fe···O separation around 2.4 Å. This is accompanied by the simultaneous rapid growth of the resident Fe−OH2 bond distance, which becomes as large as 3.9 Å. In other words the incoming H2O has a strong tendency to replace the resident H2O in the case of 4, and the probability of producing the [LFe(OH2)2]− species is therefore minimal. Quite surprisingly, the five versus six coordination of 4 is not of steric origin but results primarily from the dearomatization and associated hybridization changes at the amide-N connecting carbons. Attempts to optimize the six-coordinate 10231

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry

much weaker than in the 1a case. The molecules in Figure 6 remind one of high-energy intermediates similar to the corresponding transition states for proton-induced demetalation of 1a and 4. The N−H bond is nearly formed in the case of 1a (Figure 6A) suggesting a late transition state, while there is still a long way to go to make a N−H bond for 4 (Figure 6B). Hence, an early transition state is more likely. The idea that late and early transition states prevail for 1a and 4 correlates with lower and higher free energies of activation, respectively. The computational results indicate that in both cases the axial water serves as a directing group for the proton guiding it toward the targeted Fe−N bond. It is important to note that the pronounced elongations of the Fe−O bonds upon protonation are expected to correspond with the coordination of second water for 4 at iron and a strengthening of the existing bond for 1a. It is also noteworthy that all three O−H bonds in 4 are similar, but there are just two similar bonds in 1a (Figure 6B,A, respectively). The third proton has a tendency to stay closer to, or attack, the amide nitrogen en ruote to cleaving the Fe−N bond. Taking into account the finding that the second H2O ligand is weakly bound to iron(III) in the case of 4, axial protonation should favor hydronium ion dissociation with concomitant establishment of an Fe−OH2 ligation to maintain the favored five coordination. A plausible sequence of events is presented in Scheme 4. The reactive species A collapses to the products or loses hydronium transforming into unreactive five-coordinate complex. Scheme 4. Tentative Mechanism of Demetalation of 4 Which Agrees with the Kinetic eq 3

Figure 6. Theoretically optimized protonated forms of [LFe(OH2)]− derived from 1a (A) and 4 (B). Numbers refer to the estimated N−H and Fe−O bond distances in Å.

H+ separations of 1.05 and 1.75 Å, respectively. A comparison of the calculated N−H bond distance with those in the free ligand 6 (average of four: 0.79 ± 0.05 Å)34,35 indicates that the N−H bond is significantly formed. The proximity of the proton to the nitrogen causes an elongation of the Fe−N bond by ca. 0.2 Å compared to the three remaining Fe−N bonds. The structure in Figure 6A suggests that the proton tends to attack tail amide nitrogen from the equatorial ligand plane via or with concomitant formation of an aqueous bridge.

The mechanism in Scheme 4 leads to the rate expression 8, which agrees with the experimental rate law 3. As it was mentioned above, the value of Ka may be too low to be observed experimentally for 4. Therefore, it is very likely that Ka ≪ K. If so, eq 8 simplifies to eq 3 accounting for the leveling of kobs at high [H+] due to dissociation of weakly bound hydronium ion, which is described by the equilibrium constant K. kobs =

In the case of 4, the Fe−O bond distance is less elongated (3.55 Å), and the hydronium ion H3O+ is rather intact. The N···H+ and O···H+ separations are 2.74 and 0.97 Å, respectively. The two other O−H bonds in 4 are of similar length, and the corresponding Fe−N bond experiences a reduced although still considerable elongation. More interesting is that the proton in question sits right above the Fe−N bond and is actually closer to iron (2.66 Å) than to nitrogen. The Fe−H−N angle is acute (41°). The significantly lengthened Fe−O bond is consistent with expected repulsion between iron(III) and H3O+. The only slightly distorted hydronium ion H3O+ suggests the proton attraction to the amide nitrogen is

k[H+] K a + K + [H+]

(8)

The lifetime of intermediate A in Scheme 4 in the case of 1a is significantly shorter, and presumably the dissociation of hydronium ion as in 4 takes a different course, because the proton “slides” toward the Fe−N bond resulting in its cleavage. Thus, the tendency to dissociate of the 4−hydronium species caused by dearomatization is another proposed reason for its stabilization. Mechanistic Considerations. General Acid Pathway. In the proposed mechanism of the proton-induced iron(III) ejection, the key feature of the sliding proton from the axially coordinated hydronium ion reflects the general acid-catalyzed 10232

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry mechanism of reaction 1 suggested previously.11 It involves axial coordination of H2PO4−11 or H3PO412 in the case of 1 and 2, respectively, followed by proton delivery at the Fe−N bond. The kinetic behavior of 4 revealed in this work is virtually identical to that of 2b, but 4 is demetalated much slower. In other words, similar retardation effects are typical of “beheaded” 4 in both the specific and general acid-catalyzed transformations. The retardation mechanism in the general acid case is obviously similar to the specific acid case and originates from the low HOMO energy of 4, which extends to lowering the ability of the central iron to accommodate two axial ligands. The rate constants ktri and kdi for 2b differ by a factor of 100 (Table 3). If a similar gap exists for 4, the value of kdi for 4, which could not be reliably estimated, may be ∼9 × 10−6 M−1 s−1. This allows us to compare the reactivity of 1a and 4 toward H2PO4− using the value of kdi for 1a from Table 3. The ratio obtained suggests that 4 is 140 times more stable than 1a, which follows the same trend as in the proton-induced demetalation.

observed resting state properties influence other states in TAML catalytic cycles.

CONCLUSION The iterative design protocol,8 by which TAML oxidation catalysts have been produced and perfected, has engaged the tools of imaginative coordination chemistry aimed at accessing higher oxidation states of iron.36 Catalyst decomposition studies to identify vulnerable ligand degradation sites for replacement and kinetically derived structure−activity relationships have been used to probe how ligand-substituent electronic and steric effects manifest in the technical performance parameters.6,7 The current study adds a novel dimension to the design landscape expanding understanding of groundstate reactivity. TAML activator 4 shows exceptional stability under acidic conditions relative to previous TAML activators, for example, 1 and 2. This is especially interesting, as all previous work had suggested that substitution of the “tail” CR2 unit with electron withdrawing groups such as CF2 or NR enhanced the stability of these complexes, while substitution on the aromatic “head” unit had little effect. The resistance of 4 to acid demetalation is exaggerated by the hyperbolic dependence on [H+] as opposed to previous TAML activators, which exhibit a positive curvature associated with a third-order term in [H+]. Acceleration of the demetalation by a general acid was revealed by the example of triphosphate, H3PO4 being the reactive species. Here again, 4 is more stable than 1 or 2. The stabilization of 4 first seemed counterintuitive, because its main structural difference was the substitution of the electron-rich Me4C2 linker in place of the aromatic unit in the head part of traditional TAML catalysts. Thus, normal electron donation/ withdrawal thinking would have 4 as more electron-rich with an expected increased reactivity toward electrophiles. The reactivity-controlling structural and electronic properties of 4 have been disclosed through theoretical DFT modeling. The key factors stemming from dearomatization/rehybridization of the head unit are (i) the lowered HOMO energy of 4 compared to 1a and (ii) the associated tendency to reduce the coordination number from six to five in water. The DFT studies also suggest that in acid-induced demetalation, an early transition state for 4 is achieved, as opposed to a late transition state for 1a in which the proton can be viewed as sliding from a protonated axial water to the targeted nitrogen atom. As a result, the free energy of activation for 4 is higher. The pronounced difference found for 4 over all other catalysts suggests an avenue of research into understanding how the

Corresponding Authors



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b00921. Plots showing the effect of added NaClO4 on kobs for HClO4-induced demetalation; Eyring plots for the k1* for both 1a and 4 and k3* for 1a used to derive the enthalpy and entropy of activation for HClO4-induced demetalation; dependence of the rate constant kII for the H3PO4-induced demetalation of 4 as a function of [H+]; Eyring temperature dependence for the H3PO4-induced demetalation of 4 at pH 1; and computed distance− energy plots (PDF)





AUTHOR INFORMATION

*E-mail: [email protected]. (A.D.R.) *E-mail: [email protected]. (T.J.C.) ORCID

Matthew R. Mills: 0000-0001-8975-2855 Alexander D. Ryabov: 0000-0002-5255-1395 Terrence J. Collins: 0000-0003-2611-9184 Present Address †

Department of Chemistry, University of Missouri, Columbia, Missouri 65211, United States. Author Contributions

The manuscript was written through contributions of all authors. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS T.J.C. thanks the Heinz Endowments for support. M.R.M. thanks the R. K. Mellon Foundation for support through a Presidential Fellowship.



REFERENCES

(1) Crabtree, R. H. Deactivation in homogeneous transition metal catalysis: causes, avoidance, and cure. Chem. Rev. 2015, 115, 127−150. (2) Collins, T. J.; Khetan, S. K.; Ryabov, A. D. Chemistry and applications of iron-TAML catalysts in green oxidation processes based on hydrogen peroxide. In Handbook of Green Chemistry; Anastas, P. T., Crabtree, R. H., Eds.; WILEY-VCH Verlag GmbH & KgaA: Weinheim, Germany, 2009; pp 39−77. (3) Ryabov, A. D. Green challenges of catalysis via iron(IV)oxo and iron(V)oxo species. Adv. Inorg. Chem. 2013, 65, 118−163. (4) Mills, M. R.; Arias-Salazar, K.; Baynes, A.; Shen, L. Q.; Churchley, J.; Beresford, N.; Gayathri, C.; Gil, R. R.; Kanda, R.; Jobling, S.; Collins, T. J. Removal of ecotoxicity of 17α-ethinylestradiol using TAML/peroxide water treatment. Sci. Rep. 2015, 5, 10511. (5) Chanda, A.; Ryabov, A. D.; Mondal, S.; Alexandrova, L.; Ghosh, A.; Hangun-Balkir, Y.; Horwitz, C. P.; Collins, T. J. The activitystability parameterization of homogeneous green oxidation catalysts. Chem. - Eur. J. 2006, 12, 9336−9345. (6) Emelianenko, M.; Torrejon, D.; Denardo, M. A.; Ryabov, A. D.; Collins, T. J.; et al. Estimation of rate constants in nonlinear reactions involving chemical inactivation of oxidation catalysts. J. Math. Chem. 2014, 52, 1460−1476.

10233

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234

Article

Inorganic Chemistry (7) DeNardo, M. A.; Mills, M. R.; Ryabov, A. D.; Collins, T. J. Unifying evaluation of the technical performances of iron-tetra-amido macrocyclic ligand oxidation catalysts. J. Am. Chem. Soc. 2016, 138, 2933−2936. (8) Collins, T. J. Designing ligands for oxidizing complexes. Acc. Chem. Res. 1994, 27, 279−285. (9) Collins, T. J. TAML oxidant activators: A new approach to the activation of hydrogen peroxide for environmentally significant problems. Acc. Chem. Res. 2002, 35, 782−790. (10) Ghosh, A.; Ryabov, A. D.; Mayer, S. M.; Horner, D. C.; Prasuhn, D. E., Jr.; Sen Gupta, S.; Vuocolo, L.; Culver, C.; Hendrich, M. P.; Rickard, C. E. F.; Norman, R. E.; Horwitz, C. P.; Collins, T. J. Understanding the mechanism of H+-induced demetalation as a design strategy for robust iron(III) peroxide-activating catalysts. J. Am. Chem. Soc. 2003, 125, 12378−12379. (11) Polshin, V.; Popescu, D.-L.; Fischer, A.; Chanda, A.; Horner, D. C.; Beach, E. S.; Henry, J.; Qian, Y.-L.; Horwitz, C. P.; Lente, G.; Fabian, I.; Münck, E.; Bominaar, E. L.; Ryabov, A. D.; Collins, T. J. Attaining control by design over the hydrolytic stability of Fe-TAML oxidation catalysts. J. Am. Chem. Soc. 2008, 130, 4497−4506. (12) Ellis, W. C.; Tran, C. T.; Roy, R.; Rusten, M.; Fischer, A.; Ryabov, A. D.; Blumberg, B.; Collins, T. J. Designing green oxidation catalysts for purifying environmental waters. J. Am. Chem. Soc. 2010, 132, 9774−9781. (13) Ellis, W. C.; McDaniel, N. D.; Bernhard, S.; Collins, T. J. Fast water oxidation using iron. J. Am. Chem. Soc. 2010, 132, 10990−10991. (14) Demeter, E. L.; Hilburg, S. L.; Washburn, N. R.; Collins, T. J.; Kitchin, J. R. Electrocatalytic oxygen evolution with an immobilized TAML activator. J. Am. Chem. Soc. 2014, 136, 5603−5606. (15) Panda, C.; Debgupta, J.; Diaz Diaz, D.; Singh, K. K.; Sen Gupta, S.; Dhar, B. B. Homogeneous photochemical water oxidation by biuret-modified Fe-TAML: evidence of FeV(O) intermediate. J. Am. Chem. Soc. 2014, 136, 12273−12282. (16) Panda, C.; Ghosh, M.; Panda, T.; Banerjee, R.; Sen Gupta, S. Fe(III) complex of biuret-amide based macrocyclic ligand as peroxidase enzyme mimic. Chem. Commun. 2011, 47, 8016−8018. (17) Mills, M. R.; Weitz, A. C.; Zhang, D. Z.; Hendrich, M. P.; Ryabov, A. D.; Collins, T. J. A “beheaded” TAML activator: a compromised catalyst that emphasizes the linearity between catalytic activity and pKa. Inorg. Chem. 2016, 55, 12263−12269. (18) Warner, G.; Mills, M.; Enslin, C.; Pattanayak, S.; Panda, C.; Sen Gupta, S.; Ryabov, A. D.; Collins, T. J.; Panda, T. K. Reactivity of Ntailed (‘biuret’) TAMLs in water: Kinetics of the catalyzed oxidation of Orange II by H2O2. Synthesis and X-ray characterization of an Nphenyl biuret TAML. Chem. - Eur. J. 2015, 21, 6226−6233. (19) 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.; 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 B.01; Gaussian, Inc., 2009. (20) Becke, A. D. Density-functional exchange-energy approximation with correct asymptotic behavior. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098−3100. (21) Becke, A. D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648−5652. (22) Lee, C.; Yang, W.; Parr, R. G. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789.

(23) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions. J. Phys. Chem. B 2009, 113, 6378−6396. (24) Hinze, J.; Jaffe, H. H. Electronegativity. I. Orbital electronegativity of neutral atoms. J. Am. Chem. Soc. 1962, 84, 540−546. (25) Bender, M. L.; Bergeron, R. J.; Komiyama, M. The bioorganic chemistry of enzymatic catalysis; John Wiley & Sons: New York, 1984. (26) Smith, R. M.; Martell, A. E. Critical stability constants; Plenum Press: New York, 1976; Vol. 4. (27) Migneault, D. R.; Force, R. K. Dissociation constants of phosphoric acid at 25 °C and the ion pairing of sodium with orthophosphate ligands at 25 °C. J. Solution Chem. 1988, 17, 987−997. (28) Ryabov, A. D.; Collins, T. J. Mechanistic considerations on the reactivity of green FeIII-TAML activators of peroxides. Adv. Inorg. Chem. 2009, 61, 471−521. (29) Brown, H. C.; Kanner, B. Preparation and reactions of 2,6-ditert-butylpyridine and related hindered bases. A case of steric hindrance toward the proton. J. Am. Chem. Soc. 1966, 88, 986−992. (30) Espenson, J. H. Chemical kinetics and reaction mechanisms, 2nd ed.; McGraw-Hill, Inc.: New York, 1995. (31) Bertini, I.; Gray, H. B.; Lippard, S. J.; Valentine, J. S. Bioinorganic chemistry; Univ. Sci. Books: Mill Valley, CA, 1994. (32) Fraenkel, G.; Franconi, C. Protonation of amides. J. Am. Chem. Soc. 1960, 82, 4478−4483. (33) Szostak, R.; Aube, J.; Szostak, M. An efficient computational model to predict protonation at the amide nitrogen and reactivity along the C-N rotational pathway. Chem. Commun. 2015, 51, 6395− 6398. (34) Collins, T. J.; Kostka, K. L.; Uffelman, E. S.; Weinberger, T. L. Design, synthesis, and structure of a macrocyclic tetraamide that stabilizes high-valent middle and later transition metals. Inorg. Chem. 1991, 30, 4204−4210. (35) Uffelman, E. S. Macrocyclic tetraamido-N Ligands that stabilize high valent complexes chromium, manganese, iron, cobalt, nickel, and copper. Ph. D. Thesis, California Institute of Technology, 1992. (36) (a) Chanda, A.; Popescu, D.-L.; Tiago de Oliveira, F.; Bominaar, E. L.; Ryabov, A. D.; Münck, E.; Collins, T. J. High-valent iron complexes with tetraamido macrocyclic ligands: Structures, Moessbauer spectroscopy, and DFT calculations. J. Inorg. Biochem. 2006, 100, 606−619. (b) Collins, T. J.; Ryabov, A. D. Targeting of high-valent iron-TAML activators at hydrocarbons and beyond. Chem. Rev. 2017, 117, 9140−9162.

10234

DOI: 10.1021/acs.inorgchem.7b00921 Inorg. Chem. 2017, 56, 10226−10234