Thermodynamic and kinetic study of cleavage of the N-O bond of N-oxides by a vanadium(III) complex: Enhanced oxygen atom transfer reaction rates for adducts of nitrous oxide and mesityl nitrile oxide

Article

pubs.acs.org/JACS

Thermodynamic and Kinetic Study of Cleavage of the N−O Bond

of N‑Oxides by a Vanadium(III) Complex: Enhanced Oxygen Atom

Transfer Reaction Rates for Adducts of Nitrous Oxide and Mesityl

Nitrile Oxide

†

,†

‡

Taryn D. Palluccio, Elena V. Rybak-Akimova,* Subhojit Majumdar, Xiaochen Cai, Megan Chui,

,

§

∥

Manuel Temprado,* Jared S. Silvia, Anthony F. Cozzolino, Daniel Tofan, Alexandra Velian,

Christopher C. Cummins,* Burjor Captain,* and Carl D. Hoﬀ*

,∥

,‡

†

Department of Chemistry, Tufts University, 62 Talbot Avenue, Medford, Massachusetts 02155, United States

Department of Chemistry, University of Miami, 1301 Memorial Drive, Coral Gables, Florida 33021, United States

‡

§

∥

́

Department of Physical Chemistry, Universidad de Alcala, Ctra. Madrid-Barcelona Km. 33,600, Madrid 28871, Spain

Department of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139,

United States

*

S Supporting Information

ABSTRACT: Thermodynamic, kinetic, and computational studies are reported for oxygen

atom transfer (OAT) to the complex V(N[t-Bu]Ar) (Ar = 3,5-C H Me , 1) from

3

6

3

2

compounds containing N−O bonds with a range of BDEs spanning nearly

−1

1

00 kcal mol : PhNO (108) > SIPr/MesCNO (75) > PyO (63) > IPr/N O (62) >

2

MesCNO (53) > N O (40) > dbabhNO (10) (Mes = mesityl; SIPr = 1,3-bis(diisopropyl)-

2

phenylimidazolin-2-ylidene; Py = pyridine; IPr = 1,3-bis(diisopropyl)phenylimidazol-2-

ylidene; dbabh = 2,3:5,6-dibenzo-7-azabicyclo[2.2.1]hepta-2,5-diene). Stopped ﬂow kinetic

studies of the OAT reactions show a range of kinetic behavior inﬂuenced by both the mode

and strength of coordination of the O donor and its ease of atom transfer. Four categories of

kinetic behavior are observed depending upon the magnitudes of the rate constants involved:

(

I) dinuclear OAT following an overall third order rate law (N O); (II) formation of stable oxidant-bound complexes followed

2

by OAT in a separate step (PyO and PhNO); (III) transient formation and decay of metastable oxidant-bound intermediates on

the same time scale as OAT (SIPr/MesCNO and IPr/N O); (IV) steady-state kinetics in which no detectable intermediates are

2

observed (dbabhNO and MesCNO). Thermochemical studies of OAT to 1 show that the V−O bond in OV(N[t-Bu]Ar)

3

−

1

is strong (BDE = 154 ± 3 kcal mol ) compared with all the N−O bonds cleaved. In contrast, measurement of the N−O bond

−1

in dbabhNO show it to be especially weak (BDE = 10 ± 3 kcal mol ) and that dissociation of dbabhNO to anthracene, N , and

2

3

a O atom is thermodynamically favorable at room temperature. Comparison of the OAT of adducts of N O and MesCNO to

2

the bulky complex 1 show a faster rate than in the case of free N O or MesCNO despite increased steric hindrance of the

2

adducts.

1

. INTRODUCTION

reactions of N O to middle transition metal complexes occur

faster by initial N coordination to form a nitrosoimide complex

2

1

Oxygen atom transfer (OAT) involving nitrogen oxides

XNO) reacting with metal complexes is of fundamental

L MNNO, which then undergoes an essentially barri-

n

(

erless OAT to a second equivalent of metal complex leading to

importance to developing catalytic oxidation chemistry for the

7

overall third-order reactivity. Caulton and co-workers had

most abundant pnictogen. In particular, N O is a signiﬁcant

greenhouse gas which is three hundred times more potent than

CO as a pollutant. It is a byproduct of several industrial pro-

cesses making its utilization of economic and environmental

signiﬁcance. A barrier to that is the relatively inert nature of

N O as an oxidant. Since the pioneering work of Taube and

Armor showing that oxidation of Cr by N O is accelerated

by a factor of 10 when it is complexed to Ru(NH ) , a

2

earlier performed computational studies in which novel modes

2

of coordination of N O were proposed, including that of a

2

6

3

nitrosoimido resembling that formulated by Lin et al. Despite

considerable interest in nitrous oxide complex chemistry, the

4

ﬁrst structure of bound N O was reported only recently and did

2

8

5

2+

not correspond to a nitrosoimide binding motif. There are

2

8

2+

relatively few experimental kinetic studies of OAT from N O to

2

3

5

9

metal complexes in solution, and the authors could ﬁnd no

number of studies have focused on whether oxidation by N O

2

occurs by initial coordination through the terminal O or

terminal N atom of the ligand. Recently Lin et al. have

performed computational studies and concluded that OAT

6

Received: May 29, 2013

Published: June 27, 2013

©

2013 American Chemical Society

11357

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

examples of third-order kinetics implying a ternary transition

state. Detailed kinetic, thermodynamic, and computational

studies of the reaction of Mo(N[t-Bu]Ar) with N O, which

(NHC) is bound to the C_n_i_t_r_i_l_eatom of MesCNO, was isolated

and structurally characterized. It was also found that formation

of NHC adducts with the oxidant MesCNO suppressed oxygen

atom transfer to main group elements. While reaction with

Lewis bases was stopped by blocking attack at the C_n_i_t_r_i_l_eatom,

qualitatively this was not observed to happen in reactions with

3

2

results in cleavage of the N−N rather than the N−O bond to

yield NMo(N[t-Bu]Ar) and ONMo(N[t-Bu]Ar) , have

3

10

been reported earlier.

While nitrous oxide represents an important substrate, other

nitrogen oxides are also of signiﬁcance and generally require

catalytic activation in OAT, the ﬁrst step of which is transfer to

an active metal complex. A general two step mechanism

involving initial oxidant binding followed by OAT is illustrated

for XNO in eq 1.

the Lewis acid metal complex V(N[t-Bu]Ar) (1), presumably

since attack in this case occurred directly at the terminal O of

MesCNO rather than at the interior C_n_i_t_r_i_l_eatom.

3

11

The initial goal of the current work was to quantitatively

compare the rates of OAT to 1 from MesCNO and its NHC

adduct SIPr/MesCNO simply to see which was faster. During

22

this time, Severin and co-workers reported characterization of

k₁

k

2

ML + XNO HooI XNOML → XN + OML

stable covalent adducts of NHCs with N O and also showed

n

2

23

k

(1)

−1

that these adducts readily react with V(Mes)₃. This prompted

us to extend our work to compare N O and IPr/N O (IPr =

The mechanism shown in eq 1 is deceptively simple. Portrayal

2

1

,3-bis(diisopropyl) phenylimidazol-2-ylidene) and to study a

of a general XNO−ML adduct hides the fact that the nature of

n

series of XNO compounds with N−O bond dissociation

enthalpies (BDEs) ranging from 10 to 108 kcal mol , a range

the intermediate formed can undergo signiﬁcant change as the

X group is varied and important aspects of spin state, bond

strength, charge transfer, and adduct geometry are not

displayed. These factors will all show up in the magnitudes of

the various rate constants involved as X is changed.

−1

of nearly 100 kcal mol . The N−O BDEs in the N-oxides

studied are all signiﬁcantly weaker than the V−O BDE in

−1

OV(N[t-Bu]Ar) (1O, BDE = 154 kcal mol ), which

3

−1

makes N−O bond cleavage exothermic by 46 to 144 kcal mol .

While all bond cleavage reactions studied involve 1 and a N−O

bond, a wide range of kinetic behavior was observed, allowing us

to characterize individual reaction steps in detail.

A number of benchmark studies have been made in metal-

mediated OAT, though results from one system can be

extrapolated to another only with considerable risk. Pioneering

12

work by Schultz and Holm for OAT cycling between

L Mo(IV)(O) and L Mo(VI)(O) showed that rates of OAT

2

. EXPERIMENTAL SECTION

spanned 4 orders of magnitude. In some cases, there was good

correlation between the rates of OAT and the BDE of the bond

being broken, but the substrate with the largest BDE studied,

Detailed description of synthetic, solution calorimetric, stopped ﬂow

kinetic, and computational methods are provided in Supporting

Information.

−

1

Ph AsO (103 kcal mol ), had an activation energy the same as

3

−1

that for the one with the smallest BDE, Ph SeO (43 kcal mol ).

2

13

3. RESULTS

Recent work by Abu-Omar has shown the importance of both

binding, the ﬁrst step in OAT, and the actual OAT reaction in

the second step. The importance of symmetry and geometry as

The main thrust of this paper is experimental mapping of the

range of behavior of the OAT reactivity of nitrogen oxides with

complex 1. All XNO reagents studied (Scheme 1) were found

well as electronic factors in OAT to M(silox) complexes has

3

14

been delineated by Wolczanski, Cundari, and co-workers

where the importance of achieving a nonlinear transition state

Scheme 1. (a) General Reaction Scheme of OAT to Complex

a

15

1

and (b) Structures of All the OAT Reagents Studied

was emphasized. Mayer and co-workers have examined the

eﬀects of spin-state change on gas-phase OAT to Cr(III)

porphyrin complexes. Interestingly, they found that above a

certain threshold bond strength, OAT in either direction did

not occur. “Ultrafast” and “ultraslow” rates of OAT to late

transition metal complexes have been studied by Brown and co-

1

6

workers.

Vasbinder and Espenson have shown that

coordination of a second molecule of PyO (Py = pyridine)

can serve to facilitate OAT of an already coordinated PyO

1

7

ligand. In addition to direct oxygen atom transfer, electron

transfer followed by oxygen atom transfer has been studied by

18

Fukuzumi and co-workers. Detailed theoretical studies of

OAT from metal oxo complexes to phosphines have been made

19

by Hall and co-workers which showed the importance of σ

donation from the incoming phosphine to the L MO π*

n

orbital when the O atom is abstracted from the metal. Veige

20

and co-workers have recently shown that OAT from Cr(V)O

to PR is ∼3000 times faster for PMe than for P(t-Bu) and is

3

consistent with a bimolecular side-on attack at the Cr(V)O

LUMO.

We have recently reported studies of OAT from mesityl

21

nitrile oxide (MesCNO) to main group PX_nacceptors.

a

During that work the blue nitrile oxide−carbene adduct SIPr/

MesCNO (SIPr = 1,3-bis(diisopropyl)phenylimidazolin-2-

ylidene), in which the C atom of the N-heterocyclic carbene

L = N(t-Bu)Ar; Ar = 2,6-diisopropylphenyl; Mes = mesityl. Note that

the L VNNO species is proposed, based on computational

3

results, to be the precursor complex for dinuclear OAT from N O.

2

1

1358

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

to react rapidly and cleanly at room temperature in toluene

solution with paramagnetic complex 1 to produce the

diamagnetic complex 1O as conﬁrmed by NMR spectros-

copy. The experimental work is supported by theoretical

calculations to provide more in depth analysis; however, full

computational study of reaction mechanisms is beyond the

scope of the current work. Such work is in progress and will be

3.1.2. X-ray Structure of 1O. The X-ray structure of the

diamagnetic tetrahedral complex 1O is shown in Figure 2.

26

It resembles previously reported chalcogenide structures of 1

and that of a related oxo complex reported by Gambarotta and

25

co-workers. Space-ﬁlling representation of the structure

shown in Figure 2b illustrates the limited amount of free

space available for access to the O atom of 1O and by

inference that a complex 1−ONX leading to 1O would have

considerable steric constraint in the placement of the XN

group.

24

reported subsequently. Following structural and thermo-

chemical analysis, kinetic results are outlined.

3.1. Structural Analysis. 3.1.1. DFT-Calculated Structure

3.1.3. X-ray Structure of dbabhNO (dbabh = 2,3:5,6-

of 1. The structures of 1 and 1O give insight into parameters

inﬂuencing binding of oxidant and OAT reactivity in this

system. To date, crystals of 1 suitable for X-ray structure

determination have deﬁed isolation. DFT computational

studies, however, yield the minimum energy structure of 1

shown in Figure 1.

Dibenzo-7-azabicyclo[2.2.1]hepta-2,5-diene). The structures

of the majority of the OAT reagents studied here (Scheme 1)

are well established. However, this work also utilizes the novel

27

OAT reagent dbabhNO, which can be described as N O

2

bound to anthracene. The structure of this compound, which

has not been reported to date, is shown in Figure 3. It was

prepared following literature procedures and a sample was

recrystallized by slow evaporation of a methylene chloride/

heptane solution.

The N−N−O angle in dbabhNO of 111.3° compares with

2

that in N O of 180° and is in keeping with sp hybridization at

2

the central N of the N-nitrosoimide. In addition, for dbabhNO

the N−N (1.255 Å) and N−O (1.232 Å) bond lengths are

signiﬁcantly longer than for those N O: N−N (1.126 Å) and

2

8

N−O (1.186 Å). The computed structure of dbabhNO (see

Supporting Information) is in good agreement with the

crystallographic determination.

3

.1.4. Structures and Energies of Selected 1−ONX

Adducts. In order to clear up the nature of OAT in the

systems studied, the structures of selected 1−ONX adducts

were also computed. Since 1 has two unpaired electrons and

1

O is diamagnetic, it is clear that an intersystem crossing

between the triplet and singlet potential energy surfaces must

occur along the path of the reaction. This process can in

principle happen either before or after adduct formation.

Therefore, the structures of adducts for both the singlet and

triplet states were calculated. Computational results on

structures and energies of the adducts are summarized in

Table 1.

Figure 1. Optimized structure of 1 at the bp86/6-311G(d,p) (MDF10

for V with an additional set of f functions) level. Selected interatomic

distances (Å) and angles (deg): V−C1 = 2.45; V−C2 = 2.42; V−N1−

C1 = 92.1; V−N1−C3 = 138.3; C1−N1−C3 = 124.6; ∑(N−V−N) =

For the N O (O-bound), PyO, and MesCNO adducts,

2

minima were only found in the triplet potential energy surface.

For the rest, diﬀerent minima were obtained for both the

singlet and triplet states. The trigonal planar nature of 1 is not

greatly altered upon formation of the high-spin adducts in the

356.1.

2

The most stable conﬁguration is a high-spin d species

N

O (O-bound) and PyO adducts. The degree of pyramidal-

2

exhibiting one close aryl contact through the single unoccupied

d orbital (Figure 1). The same structural feature has been

ization of the VN moiety can be quantiﬁed by the sum of the

3

three N−V−N bond angles being about 353° in the adducts as

compared with ∑(N−V−N) = 356.1° in 1. While binding of

PyO results in little deviation from planarity and the complex

has a relatively long V−O bond length (2.007 Å), the situation

is quite diﬀerent for binding of PhNO in both triplet and singlet

previously observed in the X-ray structure of the V(N[Ad]Ar)

3

2

5

complex. Ligand binding to 1 will involve displacement of this

3

sort of η -allylic interaction, which will inﬂuence both the

thermodynamics (enthalpy of reaction) and kinetics (enthalpy

of activation). The process can in principle occur either through

associative displacement of this interaction by the incoming

base or by an intramolecular dissociative mechanism in which a

vacant site is generated. However, DFT calculations conﬁrm

that the corresponding conformer without this interaction is

not a minimum in the potential energy surface, indicating that

the ligand most likely associatively attacks 1 to essentially

dislodge the allylic interaction. Removal of this interaction in 1

by constrained optimizations accounts for about 4 kcal mol⁻in

the thermochemistry of ligand binding, and likely in the

enthalpy of activation as well.

states (and also for the IMe/N O and SIMe/PhCNO adducts).

2

For singlet 1−ONPh, the V−O bond of 1.751 Å is contracted

signiﬁcantly, being closer to that found in 1O (1.590 Å), and

the ∑(N−V−N) of 336.9° also compares to that for 1O of

336.7°. The structure of triplet 1−ONPh resembles that of the

singlet, and there is little change in ∑(N−V−N) for the two

complexes. It should be noted that the triplet form in the case

of PhNO, MesCNO, IMe/N O, and SIMe/PhCNO is diﬀerent

2

1

in character from that of the complexes of N O (O-bound) and

2

PyO. While DFT-calculated Mulliken spin densities for the

vanadium center are essentially 2 for the latter complexes,

1

1359

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

Figure 2. (a) Drawing of 1O with thermal ellipsoids at the 30% probability level. Selected interatomic distances (Å) and angles (deg): V1−O1 =

1

.590(3); V1−N1 = 1.883(2); O1−V1−N1 = 107.87(5); N1−V1−N1* = 111.03(5). (b) Space ﬁlling model showing access to O (red) bonded to

V (green); N (blue), C (gray), and H (white) are also shown.

When both singlet and triplet minima were found for a

certain adduct, the singlet state was generally lower in energy in

keeping with a stronger interaction between the metal fragment

and the ligand. However, in the case of the SIMe/PhCNO

adduct of 1, this pattern is reversed. This behavior can be easily

understood in terms of the alleviation of the steric strain

imposed by this bulky ligand in the triplet as compared with the

singlet state since the V−O bond length is longer in the high-

spin structure. Moreover, calculations were performed for a

truncated model where the 2,6-diisopropylphenyl substituents

on the nitrogen atoms of the NHC in the real system were

replaced by methyl groups and the mesityl group by a phenyl

ring. In the real system, SIPr/MesCNO, the energetic

diﬀerence favoring the high-spin over the low-spin structure

Figure 3. Drawing of dbabhNO with thermal ellipsoids at the 40%

probability level. Selected interatomic distances (Å) and angles (deg):

C1−N1 = 1.468(3); C8−N1 = 1.492(3); N1−N2 = 1.255(7); N2−

O1 = 1.232(8); N2−N1−C1 = 119.7(4); N2−N1−C8 =141.1(4);

C1−N1−C8 = 99.2(2); O1−N2−N1 = 111.3(8).

is expected to be even higher. For the IMe/N O truncated

2

model of IPr/N O, DFT calculations predict a very small

2

energetic diﬀerence between the singlet and triplet adducts.

Thus, it is conceivable that both species can coexist in

equilibrium once the adduct is formed.

Mulliken spin densities ranging from 1.20 to 1.50 were

obtained for the triplet adducts of PhNO, MesCNO, IMe/

2

A number of η complexes of PhNO are known, and the

2

N O, and SIMe/PhCNO, supporting substantial electron

structure of the minimum energy η singlet structure was

2

1

transfer from the metal to the ligand in these adducts.

computed. This species is 2.8 kcal mol⁻higher in energy than

a

Table 1. Computational Summary of Binding to 1

e

ΔH_b_i_n_d(kcal mol⁻¹

a

ligand

multiplicity

V−O (Å)

V−O−N (deg)

∑(N−V−N) (deg)

)

N O

triplet

singlet

triplet

singlet

triplet

singlet

triplet

singlet

2.164

140.6

352.7

343.3

335.9

353.7

339.3

342.6

340.1

338.9

336.3

338.7

12.1

3.1

2

b

c

d

N O

2

1.856

168.1

b

c

d

N O

2

1.710

170.5

−7.1

−2.3

4.1

PyO

2.007

1.849

1.869

1.756

1.874

1.796

1.888

1.751

134.0

122.6

131.4

156.9

126.8

137.8

123.9

142.3

MesCNO

f

IMe/N O

−18.6

−18.3

−11.8

−7.7

−24.1

2

IMe/N O

2

f

,

g

SIMe/PhCNO

PhNO

336.9

−31.3

a

b

c

d

Calculated at the bp86/6-311G(d,p) (MDF10 for V with an additional set of f functions) level. N-bound adduct. V−N bond length. V−N−N

e

f

angle. Sum of the three (N−V−N)

bond angles. calculations corresponding to the adduct with a truncated model where the 2,6-

g

anilides

diisopropylphenyl substituents of the nitrogens of the N-heterocyclic carbene in the real system have been replaced by methyls. Calculations

corresponding to the adduct with a truncated model where the mesityl group has been replaced by a phenyl ring.

1

1360

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

1

the corresponding η singlet adduct, presumably due to steric

Table 2. Experimental and Computational X−O BDE for

factors in the bulky trisanilide complex 1. Likewise, minimum

Selected Species

2

energy η -adducts were found computationally for the 1−ON /IMe

2

BDEcalc ₍_k_c_a_l_m_o_l−1

a

BDEexpt ₍_k_c_a_l_m_o_l−1

)

−1

bond

compound

dbabhNO

)

and 1−ONCPh/SIMe complexes being 7.0 and 15.8 kcal mol

b

N−O

V−O

10.4

9.7 ± 3.0

higher in enthalpy, respectively, than the corresponding most stable

1

IPr/N O

62.0

74.6

2

η -minima in the triplet state. Stable minima corresponding to an

η complex were not found for the rest of the XNO reagents

studied in the current work.

In binding of N O, no stable minimum energy conﬁguration

2

SIPr/MesCNO

PhNO

108.2

151.3

1O

154 ± 3

2

a

Calculated at the bp86/6-311G(d,p) level (MDF10 with an

was found for a 1,3-cyclic structure of the adduct of 1. A

b

1

additional set of f functions for V). This BDE value includes

rearrangement of the unstable dbabhN fragment to anthracene

and N₂.

minimum energy η (O-bound) triplet state was computed.

_T_h_e_c_o_m_p_u_t_e_d_e_n_t_h_a_l_p_y_o_f_b_i_n_d_i_n_g_o_f₊₁₂_.₁_k_c_a_l_m_o_l₋1 coupled

with an unfavorable entropy of binding means that this complex

is inherently unstable. Furthermore, N-bound adducts for both

the singlet and triplet states were found with an enthalpy of

binding of −7.1 and +3.1 kcal mol , respectively. These data

are in keeping with marginal thermodynamic stability since an

unfavorable entropy of binding will cancel the favorable

enthalpic term in the case of the singlet adduct. The bent

nature of the NNO ligand in 1−NNO as well as its singlet

ground state is in contrast to the ﬁrst reported structural

was measured. Utilizing the V−O BDE of 154 ± 3 kcal mol⁻

in 1O and data in eq 4 yields an N−O BDE of 11.4 ± 4.9

kcal mol⁻for dbabhNO. As a check on this value, the thermal

decomposition shown in eq 5 was investigated by diﬀerential

1

−1

1

8

determination of an M−NNO complex by Piro and Chang.

The (tpa )VNNO complex has a triplet ground state and a

Mes

linear NNO ligand binding architecture. This may be attributed

to electronic factors associated with the axial base ligand

Mes

scanning calorimetry between 30 and 150 °C. Utilizing the

binding in the (tpa )VNNO. It has the consequence that

−1

30

Mes

enthalpy of formation of anthracene (28.9 ± 2.4 kcal mol ) and

N O binds to the (tpa )V and the spin state and oxidation

30

2

−1

N O (19.6 kcal mol ) yields an absolute enthalpy of formation

2

number of V remain unchanged in the stable paramagnetic

−1

of dbabhNO (s) of 78.8 ± 3.3 kcal mol . The enthalpy of

dissociation of N O (eq 6) is known.

Mes

compound (tpa )VNNO. The computed structure for

30

2

1

−NNO can be viewed as having undergone spin pairing and

N O(g) → N (g) + O(g) ΔH = 39.5 kcal mol⁻

1

oxidative addition to form a V(V) complex 1N−NO. This

structure is in keeping with the nitrosoimido formulation

predicted by previous computational calculations in some

related species, as well as analogous structurally characterized

products.

(6)

2

6

,7

2

8

3

.2. Thermochemical Studies. 3.2.1. BDE in 1O. In

order to determine the enthalpic changes associated with

dissociation of the V−O bond in 1O, calorimetric studies

were made in toluene solution for OAT to 1 as shown in

eqs 2 and 3. Using values of the N−O BDE in MesCNO

Addition of eqs 5 and 6 yields eq 7. Correction of the solid

phase reaction in eq 7 to that in solution as shown in eq 8 was

made by measurement of the enthalpies of solution of

−

1

dbabhNO (s) (3.8 ± 0.4 kcal mol ) and anthracene (s)

−

1

(

5.0 ± 0.2 kcal mol ) in toluene and utilizing Hess’s law. The

value of the N−O BDE in toluene solution of dbabhNO is

1.4 ± 4.9 kcal mol⁻based on eq 4 and 8.0 ± 1.6 kcal mol

1

−1

1

based on eq 8. We adopt the average value 9.7 ±

3.0 kcal mol . This value is in perfect agreement with the

value obtained from DFT calculations of 10.4 kcal mol⁻

−

1

−

1

29

_a_n_d_P_y_O₍₆₃_.₃_k_c_a_l_m_o_l−1₎

29

(

53.1 kcal mol )

gives

1

estimates of ΔH

of 155.4 and 153.4 based on enthalpies

of reactions 2 and 3, respectively. An average value of 154 ±

V−O,disc

(

Table 2). This BDE value includes rearrangement of the

unstable dbabhN fragment formed on cleaving the N−O

bond to anthracene and N , which are the observed reaction

_k_c_a_l_m_o_l₋1 is adopted for this value in toluene solution (Table 2).

3

−1

2

The value obtained from DFT calculations of 151.3 kcal mol is in

good agreement with the experimental value.

products. Reaction 8 should have an entropy of reaction

positive enough to make it thermodynamically favorable at

room temperature. The authors are not aware of other

stable crystalline compounds that can provide a free O atom

3

.2.2. N−O BDE in dbabhNO. In order to determine

31

the N−O BDE of dbabhNO, the enthalpy of reaction 4

3

with a negative Gibbs energy of reaction. Additional synthetic

and thermochemical studies of the reactivity of dbabhNO are

in progress.

1

1361

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

Figure 4. (a) Time-resolved spectra obtained from the reaction between 1 (0.3 mM) and N O (66 mM) in toluene over 75 s at 25 °C. (b) Kinetic

2

trace at λ = 453 nm (red) with second-order ﬁt (black dashed line).

3

.2.3. Calorimetric Measurement of the Enthalpy of

Binding of PhNO to 1. The enthalpy of binding of PhNO to

was estimated using Calvet calorimetry at 30 °C in toluene

solution and resolving the thermogram into a rapid ﬁrst step

binding) and a slower secondary step (actual OAT). The

concentration as shown in Figure S8, Supporting Information.

Activation parameters for the high temperature third-order re-

⧧

−1

⧧

1

action were ΔH₃

= −4.1 ± 0.5 kcal mol and ΔS

=

rd‑order

−1

3rd‑order

−

1

−56 ± 2 cal mol K . The large negative activation entropy is

in keeping with a ternary transition state, while the negative

composite activation enthalpy suggests an exothermic binding

(

−1

derived value of approximately −45 kcal mol at 30 °C,

though approximate in nature, was higher than the DFT-

computed enthalpy of binding of −31.3 kcal mol (Table 1)

and prompted investigation by isoperibol calorimetry at 5 °C in

toluene. Ampoules of approximately 5 mL volume containing

the blue solution of fully dissociated PhNO were broken into

of N

−ONN−1 with a low barrier to oxidative addition.

At −14 °C, the kinetic traces could not be ﬁt well to either

ﬁrst- or second-order rate laws with respect to concentration of

. As discussed later, the behavior at intermediate temperature

O to 1 followed by formation of a transition state

2

−1

1

could be successfully modeled in terms of an overall mechanism

of mixed reaction order, but temperature resolution into two

limiting regimes, overall third-order (second-order in 1) at high

T and overall second-order (ﬁrst-order in 1) at low T, provided

the initial convincing evidence for the mechanism. At low

temperatures (−62, −53, −35 °C), the reaction was found to

obey a rate law ﬁrst order in concentration of 1 and in

3

50 mL of toluene solution containing a large excess of 1, and

the enthalpy of reaction generated in the ﬁrst ∼100 s of

reaction were measured. Under these conditions, NMR studies

show that OAT occurs to a negligible amount, and any small

heat generated would be absorbed into the ﬁnal baseline of

the thermogram. The value of ΔH = −54 ± 2 kcal mol is

the average of three separate determinations based on use of

sublimed PhNO as limiting reagent.

−1

concentration of N O. Spectra and analysis are provided in

2

Supporting Information, and the derived overall second-order

3.3. Kinetic Studies. The kinetics of the OAT reactions

−

1

−1

rate constants (k ) were 0.055 ± 0.005 M

s

(−62 °C),

1

were studied by the stopped ﬂow method, and in all cases time-

resolved UV−visible spectra showed conversion of 1 to 1O

−

1

−1

−1 −1

0

.119 ± 0.006 M

s

(−53 °C), and 0.386 ± 0.007 M

s

(

−35 °C). Activation parameters for the low temperature

(λ

= 453 nm) in toluene. For PhNO, variable-temperature

max

⧧

−1

second-order reaction were ΔH = 6.7 ± 0.3 kcal mol and

NMR experiments were necessary to study the kinetics of

1

observed and are discussed in the following order: (I) dinuclear

OAT following an overall third-order rate law (N O); (II)

formation of stable oxidant-bound complexes followed by OAT

in a separate step (PyO and PhNO); (III) transient formation

and decay of metastable oxidant-bound intermediates on the

⧧

−1 −1

ΔS = −32 ± 1 cal mol K .

O formation. Four categories of kinetic behavior were

The change in reaction order and in the sign of eﬀective

activation enthalpy as a function of temperature also suggests

2

stepwise bimetallic activation of N O in which the rate-

2

determining step changes with temperature. Dinuclear

activation of N O by the mechanism shown in eq 9 and

2

time scale of OAT (SIPr/MesCNO and IPr/N O); (IV)

2

steady-state kinetics in which no detectable intermediates are

observed (dbabhNO and MesCNO).

3.3.1. Category I: Dinuclear OAT Following a Third-Order

Rate Law (N O). OAT from N O to 1 occurs within minutes at

2

room temperature. Time-resolved spectra acquired over a

broad temperature range (−62 to 25 °C) showed clean con-

version with no detectable intermediates (Figure 4a). Kinetic

traces obtained at the three highest temperatures (7, 16, and

assuming a steady state in concentration of 1−NNO (in

keeping with the clean isosbestic points and failure to detect a

spectroscopically observable intermediate) yields the rate law

shown in eq 10.

2

5 °C) under pseudo-ﬁrst-order conditions (excess N O) did

2

not ﬁt to a rate law that is ﬁrst order in concentration of 1;

instead, as shown in Figure 4b, excellent ﬁts were obtained

from a rate law that is second order in concentration of 1 and

k k [1] [N O]

1

2

^r^a^t^e⁼(

k

+ k [1])

(10)

−

1

2

ﬁrst order in concentration of N O, yielding an overall third-

This will yield two limiting rate laws: if k₋₁≫ k [1], overall

2

order rate law.

third-order kinetics will be observed with k_o_b_s= (k k /k ). On

1

2

−1

−

2

−1

Derived third-order rate constants of 4670 ± 285 M

s

the other hand, if k [1] ≫ k , second-order kinetics will be

2

−1

−2

−1

−2 −1

(

25 °C), 5380 ± 512 M

s

(16 °C), and 6870 ± 136 M

followed with k = k .

obs

1

7 °C) were obtained from the slopes of the linear plots of the

The observed kinetic behavior over the entire temperature

range is consistent with this type of mechanism. This was

observed second-order rate constant (k ) versus N O

eff

2

1

1362

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

Figure 5. (a) Diode array spectral changes accompanying the reaction between 1 (0.3 mM) and PyO (2 mM) at −80 °C acquired over 580 ms,

showing growth of 1−OPy (longer wavelengths) with decay of 1 at shorter wavelengths and an isosbestic point near 570 nm. Inset shows kinetic

traces at λ = 453 and 650 nm. Longer reaction times show slow conversion to 1O. (b) Diode array spectral changes accompanying the reaction

between 1 (0.2 mM) and PyO (10 mM) at 0 °C acquired over 18 s, showing formation of 1O (λ = 453 nm) with concomitant decay of rapidly

formed 1−OPy occurring at longer wavelengths. Inset shows kinetic trace at λ = 453 nm with single exponential ﬁt.

⧧

−1

followed by full computer simulation of the data over all

temperatures to the rate law in eq 10, which was in excellent

agreement with the experimental data (see Supporting

Information). The activation parameters derived from the

ΔS = −7 ± 2 cal mol K⁻¹. Notable is the low entropy of

1

activation for ligand binding in this system. This may be

attributed to concomitant cleavage of the aryl contact in 1

(Figure 1) with a favorable entropic component due to more

degrees of freedom in the open form.

⧧

−1

⧧

kinetic model are ΔH = 6.3 ± 0.3 kcal mol and ΔS =

1

−1

⧧

−1

−

34 ± 1 cal mol

K

for k ; ΔH = 13 ± 1 kcal mol and

At higher temperatures, binding of PyO is too rapid to be

observed by the standard stopped ﬂow method, and conversion

to the bound 1−OPy complex occurs rapidly on mixing. This

allows for convenient following of the second step in the OAT

reaction sequence as the intermediate complex 1−OPy

converts to 1O and Py. Time-resolved spectral data are

shown in Figure 5b. The broad, lower energy absorbance band

assigned to 1−OPy (600−800 nm) decays steadily as 1O

1

K

−1

⧧

−1 −1

⧧

ΔS = −6 ± 1 cal mol

for k ; and ΔH = 3.0 ± 0.1

−1

2

−1

⧧

−1

kcal mol and ΔS = −27 ± 1 cal mol

K

for k . The

2

⧧

predicted third-order activation parameters of ΔH

=

3

rd‑order

⧧

= (6.3 + 3.0 − 13) = −3.7 kcal mol⁻¹and ΔS

⧧

ΔH₍

ΔS₍

k

1 2

+k

−k−1

)

3rd‑order

⧧

= (−34 + (−27) − (−6)) = −55 cal mol K⁻¹are

−1

k

1 2

+k

−k−1

)

in excellent agreement with the direct experimental estimates from

the high-temperature data discussed above. The crossover in reaction

mechanism is due to the high activation energy for the dissociation

(

λ

= 453 nm) grows in. The observed rate constants

max

measured at high temperatures showed essentially no dependence

on PyO concentration (Figure S14, Supporting Information).

The reaction was thus treated as zero order in PyO and the

average value of k_o_b_swas taken as the true ﬁrst-order rate constant

step (k ) causing it to overtake k at higher T. The enthalpy of

−

1

2

0

⧧

binding of N O, ΔH = (ΔH − ΔH ) = (6.3 − 13) = −6.7

2

k

1

k

−1

−

1

kcal mol , is in agreement with the computed enthalpy of binding

−1

of N O (N-bound, singlet) of −7.1 kcal mol (Table 1). The high-

2

(

k ) for product formation from 1−OPy. Activation parameters

2

temperature third-order reaction has an apparent negative enthalpy

of activation and appears to speed up with decreasing temperature.

If this reaction were an elementary third-order step rather than a

composite one, the change to second-order kinetics and

concomitant slowing with temperature would not be observed.

⧧

associated with k were derived from an Eyring plot: ΔH = 15 ±

2

−1

⧧

−1

1

kcal mol and ΔS = −4 ± 2 cal mol K . The OAT

2

⧧

reaction shows a high value of ΔH and a negative ΔS that is

small in absolute value.

The kinetics of PhNO binding was also studied by the

stopped ﬂow method at low temperatures (−80 to −53 °C).

The corresponding time-resolved spectral changes revealed

rapid formation of intense absorbance bands at λ_m_a_x= 477 and

Provided k₋₁has a higher enthalpy of activation than does k , the

2

rate will decrease more rapidly with decreasing temperature and

cause the observed change in the rate-limiting step. This

observed behavior is predicted from the analysis of time-resolved

spectral data, supporting the proposed stepwise bimetallic

mechanism (eq 9).

7

40 nm that are assigned to the 1−ONPh adduct (Figure 6).

The rates of formation of the 1−ONPh adduct were quantiﬁed

by following the growth in absorbance at λ = 740 nm under

pseudo-ﬁrst-order conditions; observed rate constants depended

linearly on PhNO concentration, and y-intercepts were very close

to zero (Figure S15, Supporting Information). Activation pa-

rameters for PhNO binding were derived from an Eyring plot:

3.3.2. Category II: Formation of a Stable Adduct Followed

by OAT (PyO and PhNO). The reaction between excess PyO

and 1 could be readily separated into ligand binding at low

temperatures (−80 to −53 °C) and OAT from 1−OPy at high

temperatures (0−20 °C). Low-temperature time-resolved

spectral changes are shown in Figure 5a. A rapid increase in

the absorption bands at longer wavelengths was observed, with

concomitant decay occurring under ∼550 nm. The observed

rate constants obtained from ﬁtting kinetic traces to a single-

exponential function depended linearly on PyO concentration,

⧧

−1

⧧

−1 −1

ΔH = 5.7 ± 0.5 kcal mol and ΔS = −14 ± 3 cal mol K .

1

The rather dramatic changes in spectra upon binding of

PhNO to 1 are in contrast to those described earlier for the

binding of PyO (Figure 5a). In order to provide additional

information on the nature of these adducts observed at low

1

temperature by stopped ﬂow spectrophotometry, H NMR

and the second-order rate constants (k ) representing for-

1

studies of ligand binding to 1 at −70 °C were performed for

both PhNO and PyO. As shown in Supporting Information,

PhNO yielded a diamagnetic complex with narrow peaks in its

mation of a 1−OPy adduct were obtained from the slopes of

these plots as shown in Figure S13, Supporting Information.

The near zero y-intercepts suggest that this process can be

regarded as essentially irreversible over the temperature range

studied. An Eyring plot was used to obtain activation param-

1

H NMR spectrum. In contrast, PyO was conﬁrmed to form a

paramagnetic complex upon mixing at −70 °C, and very slow

OAT to form Py and 1O was observed at this temperature.

⧧

−1

eters for this binding process: ΔH = 6.8 ± 0.4 kcal mol and

1

1363

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

Figure 6. (left) Diode array spectral changes upon reaction of 1 (0.3 mM) with excess PhNO (1 mM) at −80 °C over 7.5 s. (right) Kinetic trace at

λ = 740 nm (red) with single exponential ﬁt (black dashed line).

Therefore, the diﬀerences in UV−vis spectra observed between

−OPy and 1−ONPh can be explained by the diﬀerent spin

states of vanadium (high-spin versus low-spin) in these adducts

as determined by NMR studies and in keeping with DFT

calculations described earlier. Furthermore, TD DFT calculated

UV−vis spectra for both species are in keeping with the

experimental spectral changes.

readily calculated. As shown in eq 13, the value of +26

−1

1

kcal mo₋l exceeds the observed activation energy of +19

1

kcal mol , and so expulsion of NPh can be ruled out not only

33

based on production of PhNNPh but also in terms of

energetics. Additional evidence in keeping with expulsion of a

free triplet nitrene is generated by the observation that in the

presence of excess free PhNO, a well-known trapping species

34

for triplet nitrenes, azoxybenzene is formed, and in the

presence of excess 1, 1NPh is also formed. However, more

complex mechanisms in which a free nitrene is not generated

Over a 1 h time period, 1O formation from 1−ONPh was

not observed to occur to a measurable extent at −70 °C.

1

However, the room temperature H NMR spectrum of PhNO

35

can also be imagined, and only direct spectroscopic detection

of the triplet nitrene can prove its generation. That is beyond

the scope of this investigation.

and 1 show clear formation of 1−ONPh, which slowly con-

verts to 1O and other products. As shown in Figure S3,

Supporting Information, the NMR signals due to 1−ONPh and

3.3.3. Category III: Transient Formation and Decay of a

1

O were suitable for kinetic analysis. Rate constants for the

Metastable Intermediate on the Time Scale of OAT (IPr/N O

−1

2

formation of 1O from 1−ONPh were 0.00031 ± 0.00006 s

and SIPr/MesCNO). Stopped ﬂow kinetic measurements of the

−1

(

22 °C), 0.0015 ± 0.0003 s (35 °C), and 0.0034 ± 0.0007 s

45 °C). Derived activation parameters (Figure S5, Supporting

reaction between excess IPr/N O (2−12 equiv) and 1

2

(

0.25 mM) were performed over a broad temperature range

⧧

−1

⧧

Information) are ΔH = 19 ± 2 kcal mol and ΔS = −9 ± 5

2

(−62 to −35 °C). As shown in Figure 7, time-resolved spectral

changes showed a rapid, intense growth in absorbance across

the entire spectral window with a band centered near λ = 475 nm.

A slower decay process follows where the intense band at 475 nm

is found to blue shift to the expected maximum for 1O

cal mol⁻K .

1

−1

In the absence of added reagents, the reaction occurs

according to eq 11. Decomposition of in situ 1−ONPh to 1O

(λ = 453 nm).

For technical reasons, the kinetic data were analyzed at

λ = 725 nm, and traces were ﬁt to the general two term rate

3

6

law, rate = A exp(−k t) + A exp(−k₂_,_o_b_st). Plots of k₁_,_o_b_s

1

1,obs

2

versus IPr/N O concentration were linear with nonzero

2

is nearly quantitative in terms of V containing products, and the

unknown organic products likely result from decomposition of

phenyl nitrene, which may be generated either as a free or

loosely bound species during the reaction. The yield of

PhNNPh was ∼30% of the anticipated PhNNPh based on

reaction stoichiometry (eq 11). The unknown products include

a dark precipitate and an unknown paramagnetic compound

with broad peaks in the region of 6−8 ppm. The most likely

mechanism is direct expulsion of a nitrene as shown in eqs 12

and 13.

y-intercepts, while plots of k2,obs versus IPr/N O concentration

2

appeared to saturate. Full analysis of the experimental data is

provided in Supporting Information. The results obtained from

analysis of the experimental data were veriﬁed by global

modeling as described in Supporting Information. The

proposed mechanism of OAT from IPr/N

through the initial reversible formation of an intensely

absorbing intermediate adduct (k , k−1), which subsequently

decays in the OAT step (k

O to 1 occurs

2

1

), forming 1O. Spectra computed

2

by the kinetics program for the species involved were consistent

over the temperature and concentration ranges studied and are

shown in Figure 8. The modeled spectrum of 1−ON /IPr is in

2

good agreement with that obtained computationally by TD

DFT calculations for the singlet adduct of the truncated model

IMe/N O (see Supporting Information). Derived activation

2

parameters from the kinetic model were in good agreement

⧧

with experimentally determined values and are ΔH = 5.6 ± 0.4

1

kcal mol⁻and ΔS = −18 ± 2 cal mol

1

⧧

−1

K

−1

(for ligand

2

1

⧧

−1

⧧

Based on the measured enthalpy of binding of PhNO (−54

binding, k ) and ΔH = 13 ± 1 kcal mol and ΔS = 1 ± 2

1

2

−

1

−1

−1 −1

kcal mol ), the calculated N−O BDE in PhNO (108.2 kcal mol ,

cal mol

K

(for OAT, k ).

2

−1

see Table 2) and a gap of 18 kcal mol between the triplet

and singlet nitrene, the energetics of nitrene expulsion is

Analogous to the reactions with IPr/N O, the time-resolved

2

32

spectral changes accompanying the reaction between excess

1

1364

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

Figure 7. Time-resolved spectra obtained from the reaction of 1 (0.25 mM) with IPr/N O (1 mM) at −62 °C acquired over 26 s. (a) Rapid

2

formation of the intermediate is observed initially across the entire spectral window, which maximizes within ca. 2 s. (b) Decay of the intensely

absorbing intermediate (purple traces) superimposed on its prior accumulation, ultimately ending with the expected spectrum of 1O (black trace).

Inset shows the kinetic trace at λ = 725 nm with double-exponential ﬁt.

Figure 8. Spectra computed of colored components by the kinetics

the adduct formed between 1 and MesCNO/SIPr as shown in

program, obtained from modeling at −62 °C with [1] = 0.25 mM and

o

eq 14a. An alternative mechanism considered is shown in

eq 14b, in which the colored intermediate is due to a metastable

organic intermediate, the deoxygenated adduct SIPr/MesCN.

A stable azirine formed by reaction of an NHC with a nitrile

[

IPr/N O] = 2 mM. A known spectrum of IPr/N O was used while

2

o

2

all other species were calculated. The computed spectra of 1 and 1O

are in good agreement with their actual spectra.

37

SIPr/MesCNO (2−12.5 equiv) and 1 (0.3 mM) at low

temperatures (−80 to −62 °C) could be divided into two parts

as shown in Figure 9. The ﬁrst stage of the reaction is also

characterized by a rapid, intense growth in absorbance over

the entire spectral region (Figure 9a). In the second stage, there

is a leveling oﬀ in the absorbance near that seen for 1O at

has been reported recently though it was found to be pale

yellow. TD DFT calculations predict a low intensity band at

λ = 473 nm for the SIPr/MesCN fragment (See Supporting

Information), but the absorption observed at longer wave-

lengths due to the intermediate formed in stopped ﬂow

experiments cannot be attributed to this fragment. However,

the TD DFT computed spectrum of the triplet 1−ONCPh/

SIMe truncated model for the adduct (See Supporting

Information) is in keeping with kinetic modeling studies in

which the spectral changes in Figure 9 are in fact due to

complex formation and decay in a manner similar to that

4

53 nm, but at wavelengths above 500 nm, there is steady

decrease in absorption (Figure 9b, inset). At higher temper-

atures, however, the intensely absorbing species was not

detected, and only clean, rapid conversion of 1 to 1O was

observed with isosbestic points near λ = 415 and 525 nm.

Two possible sources of the colored intermediate were

considered as shown in eqs 14a and 14b. The accepted mech-

anism that best ﬁts the kinetic data and calculated spectro-

scopic data is one in which the colored intermediate is due to

described for IPr/N O above. Full analysis of the experimental

2

data and details of kinetic modeling are provided in the

Supporting Information. Derived activation parameters are

⧧

−1

⧧

−1

ΔH = 4.9 ± 0.4 kcal mol , ΔS = −19 ± 2 cal mol

K

1

Figure 9. Time-resolved spectra obtained from the reaction of 1 (0.3 mM) with SIPr/MesCNO (3 mM) at −80 °C acquired over 7.5 s. (a) Initial

stage of the reaction (0−0.6 s) where rapid adduct formation is observed. (b) Spectral changes during the second phase of the reaction (0.6−7.5 s)

showing decay of the broad lower energy bands (purple traces) superimposed on its prior accumulation. Inset shows kinetic traces at λ = 453 and

577 nm.

1

1365

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

⧧

−1

⧧

(

−

for ligand binding, k ) and ΔH = 9.3 ± 0.3 kcal mol , ΔS =

1

2

9 ± 1 cal mol⁻

1

K

−1

(for OAT, k ).

.3.4. Category IV: Steady-State Kinetics in Which No

2

3

Detectable Intermediates Are Observed (dbabhNO and

MesCNO). The kinetics of the OAT reaction between 1 and

dbabhNO were measured at low temperatures (−71 to

−44 °C) using excess dbabhNO (3.3−33.3-fold excess). The

intense absorbance band at λ = 453 nm corresponding to 1O

grew in within seconds as shown in Figure S32, Supporting

Information. Absorbance bands assignable to an intermediate

dbabhNO−1 adduct were not detected in the time-resolved

spectra; only clean formation of 1O from 1 with isosbestic

points near λ = 417 and 526 nm was observed. Kinetic traces ﬁt

well to single-exponential rate law, and linear plots of the

Figure 10. Rate of production of 1O as a function of time at

−

62 °C under similar conditions. [XNO] = 3 mM (SIPr/MesCNO,

o

MesCNO); 2 mM (PyO, PhNO, dbabhNO, IPr/N O); 66 mM

2

(

N O) and [1] = 0.3 mM (for reactions with PhNO, MesCNO,

2 o

observed rate constant (k ) as a function of dbabhNO

dbabhNO, SIPr/MesCNO, N O), 0.25 mM (IPr/N O), or 0.2 mM (PyO).

obs

2

concentration yielded second-order rate constants (k_a_p_p),

^w^h^e^r^e^,^a^s^s^u^mⁱⁿ^g^a^s^t^e^a^d^y^s^t^a^t^eⁱⁿⁱⁿ^t^e^r^m^e^dⁱ^a^t^e^a^d^d^u^c^t^,^kapp

=

following the same pattern if it is considered not as reaction of 1

k k /(k + k ) = k /(1 + k /k ) for product formation. Since

1

2

−1

2

1

−1

2

with N O but as reaction between 1 and the 1−NNO complex as

2

intermediates were not detected in these reactions, the

composite rate constant cannot be broken down further.

Activation parameters for the composite reaction were derived

shown in Scheme 1. From the kinetic point of view, an ideal OAT

reagent would have a large value for binding (k ), a low value for

1

dissociation (k ), and a huge value for actual OAT (k ). The

−1

2

⧧

−1

⧧

from an Eyring plot: ΔH = 6.3 ± 0.3 kcal mol and ΔS =

ratio k /k is the K for substrate binding, and for a good OAT

1

−1

eq

−1

−

13 ± 1 cal mol K .

reagent this will necessarily be high, as will the k /k ratio.

2

−1

Reactions of excess MesCNO (12−93-fold excess) with

(0.3 mM) showed similar spectroscopic behavior to that of

dbabhNO as shown in Figure S30, Supporting Information.

^Uⁿ^lⁱ^k^e^d^b^a^b^h^N^O^,^t^h^e^lⁱⁿ^e^a^r^p^l^o^t^s^o^f^kobs versus MesCNO

concentration appeared to give nonzero intercepts, which

^s^u^g^g^e^s^t^s^t^h^a^t^k⁻1 plays a role in the kinetic scheme. Derived

The substrates in Scheme 1 can be divided into linear

1

(

MesCNO and N O) and bent (PhNO, dbabhNO, IPr/N O,

2

SIPr/MesCNO, and 1NNO) molecules. Only for

MesCNO and N O can the N bonded to O be viewed as an

2

sp hybrid N atom with true delocalized π-bonding. For PyO,

the bonding is primarily dative donation of the basic lone pair

of electrons on the Py N atom to O. These and other factors all

contribute to inﬂuence the rate of OAT in these systems.

It should also be recognized that the act of binding may be

quite diﬀerent for diﬀerent substrates and this blurs the

distinction between simple binding and oxidative addition/

OAT. Examples of both high-spin (PyO) and low-spin (PhNO)

adduct formation were detected by low T NMR experiments.

Thus, the spin crossover event could be considered part of the

act of binding or part of the actual OAT itself depending upon

the actual potential energy diagram. It is typical to discuss this

in terms of the minimum energy crossing point (MECP) as

⧧

activation parameters from an Eyring plot were ΔH = 5.3 ±

0

.8 kcal mol⁻and ΔS = −24 ± 3 cal mol K . Despite a

1

⧧

−1 −1

⧧

−1

slightly lower value of ΔH (∼1 kcal mol ), the rate constant

−1

for reaction of MesCNO at −62 °C (81 M s ) is over 20

times smaller than that for dbabhNO at the same temperature

(

1820 M⁻s ) due to the more unfavorable entropy of

1

−1

activation for OAT.

4. DISCUSSION

This paper describes exploration of factors involved in OAT to

complex 1 from the nitrogen oxide substrates shown in Scheme

. Kinetic, thermodynamic, and computational studies are

reported and a relatively complete, albeit complex, picture

39

1

delineated by Harvey. However, it is clear from this work that

the crossing point can occur before (PhNO) or after adduct

formation (PyO). The computed minimum energy structures

of the paramagnetic PyO and diamagnetic PhNO adducts of 1

are shown in Figure 11.

emerges. Thermochemical studies showed that the V−O bond

−1

in 1O (BDE = 154 ± 3 kcal mol ) is one of the strongest

38

metal oxo bonds to be experimentally determined. Computa-

tional studies indicate that in the absence of a stabilizing allylic

interaction in 1 (see Figure 1), this would be close to 158

In addition to diﬀering in spin state (S = 1 for 1−OPy and

S = 0 for 1−ONPh), there is a signiﬁcant diﬀerence in the

V−O bond lengths (2.007 vs 1.751 Å) and the degree of

pyrimadalization of the V center. The 1−OPy complex

resembles the starting complex 1 in the sense that the three

N atoms of the amides are nearly planar, whereas for 1−ONPh,

the adduct is more product like in that it resembles 1O.

Furthermore, there is a diﬀerent arrangement of the N[t-Bu]Ar

groups; for 1−ONPh, the t-butyl groups are all pointed up

toward bound PhNO, and the aryls groups are pointed down

and away from it. For 1−OPy however the computed minimum

energy structure has one t-butyl group pointing down and two

pointing up, probably due to favorable π-stacking interactions

−1

kcal mol . Reaction of 1 with XNO is exothermic by 46

PhNO) to 143 (dbabhNO) kcal mol for the compounds

shown in Scheme 1, which range from 10−108 kcal mol in

−1

(

−1

terms of the BDE of the N−O bond broken.

Because individual rate constants and mechanism vary in

these systems, it is useful to view the overall eﬃcacy of the

XNO reagents under a common set of conditions as shown in

Figure 10.

The net rate of OAT, without regard to detailed mechanism,

is in the order PhNO ≪ N O < PyO < IPr/N O < MesCNO <

dbabhNO ≈ SIPr/MesCNO. Derived rate constants and their

assignments, as well as activation parameters and N−O BDE

data are collected in Table 3.

2

40

between the Py and arene rings of the anilide ligands. These

same considerations will play a role in the transition state for

binding as well. In the present work, we have not computed the

The range of OAT reagents studied, except for N O, which

undergoes dinuclear OAT, all follow the general kinetic scheme

in eq 1. The reaction of N O and 1 can also be viewed as

2

39

MECPs since that is beyond the scope of our primary

experimental goals in the current paper. However, it is clear

2

1

1366

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

i

Table 3. Data for Reactions of XNO and 1

a

k refers to bimolecular rate constants for XNO binding, k refers to rate constants for OAT, and k refers to composite bimolecular rate constants

1

2

app

b

c

for reactions where individual rate constants could not be resolved. Calculated from gas phase data in ref 30. Values taken from kinetic model.

d

e

f

Value taken from ref 29. Value extrapolated from Eyring plot for OAT at high T. Calculated at the BP86/6-311G(d,p) level; this value includes

g

dissociation of the deoxygenated adduct to its components in solution. Derived from thermochemical data in this paper and corresponds to the

formal reaction 8. As discussed in Supporting Information, a value of k of 3560 M

however in reality it is the k /k ratio that is known accurately rather than individual values for k and k . It can be concluded only that k is very

fast even at low T. BDE in kcal mol , k

h

−1 −1

s

at −62 °C can be determined from modeling studies;

2

−1

2

i

−1

⧧

−1

⧧

−1 −1

, k at −62 °C, ΔH in kcal mol , and ΔS in cal mol K .

assigned

Figure 11. Optimized structure at the bp86/6-311G(d,p) (MDF10 for V with an additional set of f functions) level for (a) 1−OPy (triplet) and (b)

1

η -1−ONPh (singlet). For selected interatomic distances and angles, see Table 1.

that the ﬁnal analysis must include the entire potential energy

surface, from binding to spin crossover to expulsion of the

reduced OAT reagent, and that spin crossover as well as formal

oxidation of V from (III) to (V) may in some cases occur after

initial binding and in some cases be part of the binding process

has low-lying π* orbitals available to accept electron density

from the high-spin complex 1, the 1−ONX adduct can be

viewed as {XNO⁻}{1 } (α = 0−1) in which the V(III) has

already been oxidized to V(III+2α) where α represents the

degree of charge transfer.

2α

+2α

itself. A prime example of that is N O itself, for which the only

Despite these complexities, the enthalpy of activation for

binding in all systems where ligand coordination rates were

measured (Table 3) is on the order of 6 ± 1 kcal mol⁻

independently of whether the adduct formed is in low or high

2

computed stable binding mode is through formal oxidation of 1

to produce a computed low-spin V(V) complex, which can be

described as being 1N−NO. In the case where the ligand

1

1367

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

spin and is not much higher than that estimated for breaking

as the fact that it may approach and bind from either end of the

molecule giving it a larger open steric window for binding. Of

3

−1

the η -arene interaction present in 1 of 4 kcal mol . This

implies that ligand binding to 1 lacking this interaction should

special note is the very slow rate of binding of N O, which,

2

−1

be only on the order of 2 kcal mol , in keeping with the

generally low activation enthalpy for ligand addition to a metal

center with a vacant orbital available for bonding. Since the

enthalpies of activation for ligand binding are relatively

constant, the entropy of activation plays a strong role in

determining the rate constant values. Detailed computational

study of oxidant binding in this system is planned, in particular

the entropy of activation may be inﬂuenced by the nature of the

complex formed.

according to computational studies, involves formation of a

low-spin adduct. The reasons for the slow binding are

attributed to rearrangements that occur during what is in eﬀect

an oxidative addition reaction. These are discussed further

under the individual substrate sections.

4.2. PhNO. This reagent has the slowest rate of OAT in part

because the N−O BDE is highest for it. The rate of binding

occurs reasonably fast at low temperatures, but OAT does not

occur until near 0 °C. PhNO is isoelectronic with O but unlike

2

Following brief discussions of rate constant data, each OAT

reagent is discussed separately. As mentioned earlier, full

computational study of this system is planned and may yield

additional insight, and the focus of the current paper is the

range of experimental results.

O its ground state is a singlet. NMR studies show that the

2

1

2

1−ONPh adduct is low spin, as are both the η and η binding

modes of O . However, unlike O , which is considerably more

stable in its η -binding form, computational studies show that

for PhNO, the η¹is more stable than the η -binding,

presumably for steric reasons. The enthalpy of binding of

4

1

2

4

.1. Rate Constants for Substrate Binding (k ), OAT

1

−1

(

k ), and Composite Reactions (k k /(k + k )). For PyO

2

PhNO to form η -1−ONPh (−54 kcal mol ) compares with

1

2

−1

2

−1

and PhNO, it was possible to measure both the rate of initial

binding and the rate of the actual OAT reaction itself and to

characterize the intermediate adducts by low T NMR

spectroscopy. The intermediate adducts formed between IPr/

that of η -1−O (−75 kcal mol ).

2

The most probable mode of OAT is decomposition of the

3

low-spin complex 1−ONPh to generate free NPh and 1O in

a unimolecular reaction. Due to the high heat of binding of

N O or SIPr/MesCNO and 1 were observed to rapidly form

PhNO and concomitant stability of 1−ONPh, even generation

2

−1

and decay in stopped ﬂow experiments. Nevertheless, it proved

possible to obtain individual rate constants by kinetic modeling.

For MesCNO and dbabhNO, however, no detectable

intermediate complexes were observed, and only the composite

steady-state rate constants could be determined. This provides

some upper and lower limits for estimation of rate constants

of the lower energy (by 18 kcal mol ) triplet nitrene is

−1

endothermic as an elementary step by about 8 kcal mol , and

the reaction is driven forward only by trapping of the nitrene.

There are therefore two MECP’s in this reaction, a cross over

from the triplet 1 and singlet PhNO to form the singlet

complex in ligand binding, followed by OAT, which involves

reverting back to a triplet state of bound 1−ONPh, which then

provided the steady-state assumption is valid. For N O, the rate

2

3

of binding could be resolved by kinetic analysis, but this

involves binding through N rather than O. For systems where

individual rate constants have been resolved, either directly or

through kinetic modeling, a summary of individual rate

eliminates NPh and singlet 1O. The observed enthalpy of

−1

activation for actual OAT (19 kcal mol ), as well as some

qualitative trapping experiments are all consistent with this

mechanism; however, it will not be proven until spectroscopic

observation or other incontrovertible evidence is found.

Nonetheless, the slowest nature of PhNO for OAT among all

the XNO reagents studied here seems to be the consequence of

the high stability of the adduct as well as the extrusion of the

unstable triplet nitrene.

constants is presented in Table 4 for both binding (k ) and

1

actual OAT (k ) at the common temperature of −62 °C.

2

Table 4. Measured or Modeled Values of Rate Constants at

−62 °C

4

.3. PyO. PyO has the fastest rate of binding of all the

k (M⁻s⁻¹

1

)

substrate

1−NNO

SIPr/MesCNO 7.9 ± 0.3

k₂(s⁻¹)

substrate

PyO

PhNO

1

substrates studied here. Furthermore, ligand binding was

essentially irreversible at low temperature for the high-spin

complex formed. Due to the basicity of pyridine, the electron

density on the donor O atom of PyO would be expected to be

the most basic, and the stability of high-spin O bonded donors

(13 ± 1) × 10³

(5.6 ± 0.1) × 10

very fast

b

3

2

−1

SIPr/MesCNO (3.3 ± 0.1) × 10

IPr/N O

(3.1 ± 0.1) × 10

2

−5

a

IPr/N O

(7.8 ± 0.5) × 10

(3.8 ± 1.2) × 10⁻PhNO

PyO

8.5 × 10

4.7 × 10

2

−10

a

N O

2

would be expected to be in the order PyO > MesCNO > N O.

2

a

Value extrapolated from Eyring plot (See Supporting Information).

See footnote h in Table 3.

The NHC adducts of MesCNO and N O would be expected to

2

b

also have increased electron density on the donor O atom

compared with the free ligands. Despite rapid and irreversible

adduct formation and a lower or similar N−O BDE than the

The rate of binding is in the order PyO > PhNO > SIPr/

MesCNO > IPr/N O ≫ N O and spans almost 6 orders of

NHC adducts of MesCNO and N O, respectively, PyO

2

magnitude. The rates of actual OAT, where they could be

determined, show an even wider range of values (10 orders of

magnitude) with the descending order: SIPr/MesCNO > IPr/

N O ≫ PyO ≫ PhNO. In reactions with N O, the proposed

underwent relatively slow OAT, and the activation barrier for

−1

k (15 kcal mol ) is high enough that it yielded a readily

2

separable two-step process. A plausible explanation is that there

are no π* orbitals available to accept electron density and that

this forces a higher energy MECP involving a larger geometric

rearrangement to involve instead the σ* orbitals. While

intuitively reasonable, this premise will be tested by more

detailed computational studies in progress.

2

intermediate 1NNO cannot be detected, but kinetic

analysis suggests it to be a rapid OAT reagent to free 1.

The rate constants for binding of the substrates to 1 at

−

62 °C in Table 4 can be compared with the recently

41

4

−1 −1

published value of 2.6 × 10 M

s

for O binding to form a

4.4. N O. Due to its importance as a ligand and the fact that

2

1

low-spin η -1−O adduct. The very fast rate for O , faster even

than PyO (Table 4), may be attributed to its small size, as well

it undergoes dinuclear activation, N O is discussed in more

2

5

detail. The pioneering work of Taube’s group had shown

1

1368

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

2

+

much earlier that the preformed complex (NH ) RuN O

underwent rapid OAT with Cr but that reaction between Cr

3

5

2

+

2+

and free N O required months under similar conditions. This

2

sets the precedent for dinuclear OAT involving nitrous oxide₈

between diﬀerent metallic fragments. Chang and co-workers

have reported the ﬁrst crystal structure of a nitrous oxide

Mes

complex (tpa )VNNO, which might be expected to be similar

in nature to any intermediate complex formed between N O

2

Mes

and 1. The presence of the coordinating base in the (tpa )V

system results in formation of a linear high-spin complex. In

contrast, DFT studies done here suggest that for 1−NNO the

binding should be low spin with a doubly bent N O group.

2

Computational studies by Caulton and co-workers have sug-

gested a new binding form for N O as a dianion, and that is also

2

in keeping with our calculated structure of 1NNO.

Binding of N O to 1 through O is thermochemically

2

−1

unfavorable (ΔH = 12.1 kcal mol , Table 1), and also the

high-spin N-bound adduct is unstable (ΔH = 3.1 kcal mol ,

bind

−1

bind

Table 1). Likewise, the diamagnetic V(V) complex is computed

to be of marginal stability. The relatively slow nature of binding

at high temperatures is in keeping with the fact that binding is a

reversible process to yield the low-spin 1NNO adduct.

Once formed, the key factor determining the fate of the

nitrosoimido species is the k /k ratio of its dissociation back

−1

2

to 1 and N O versus its rate of trapping by free 1 to yield a

2

bridging N O intermediate complex 1NNO−1, which

2

−1

Figure 12. Potential energy (kcal mol ) diagram for reaction 21 +

N O → 1O + 1 + N by dinuclear OAT. At low T, the rate-

subsequently fragments with a low barrier to 1O and 1−NN,

2

which rapidly loses N to yield the observed products. The

2

determining step occurs at ligand binding (TS1) and the reaction

obeys the rate law dP/dt = k [1][N O]. At high temperatures, the

rate-determining step occurs at formation of the dinuclear complex

proposed termolecular reaction for N O is in keeping with

2

1

2

6

recent calculations of Lin et al. who have concluded that OAT

studied for Cp V by Bergman and co-workers actually occurs

1NNO−1 (TS2), and the reaction obeys the rate law dP/dt =

2

by initial binding through N of N O to one V center, followed

(k

1

k

2

/k−1)[1] [N

2

O]. See text for additional discussion.

2

by essentially barrierless OAT to a second V center. The

derived reaction proﬁle for OAT between N O and 2 mol of 1

is shown in Figure 12.

the N−N bond of N O by Mo(N[t-Bu]Ar) was initial binding

2

3

to form the highly potent nitrosylating agent ONNMo(N[t-

The reaction proﬁle is computed based on the experimental

Bu]Ar) . Activation parameters for the rate-determining step

3

−1

⧧

−1

⧧

−1

enthalpies of activation for k (6.3 kcal mol ), as well as k

were ΔH = 9.7 kcal mol and ΔS = −24 kcal mol .

1

−1

43

(

13 kcal mol ) and k (3.0 kcal mol ). The equilibrium

Generation of free NO, though energetically allowed, was

2

binding enthalpy (determined from k /k at various temper-

atures) is computed to be −6.7 kcal mol . This analysis shows

that there is a much greater enthalpy of activation for the k₋₁

excluded based on competition and trapping experiments.

4.5. MesCNO. The rate of reaction of MesCNO is

intermediate between that of the two NHC adducts studied

1

−1

−

1

step than for the k step. However, the entropy change for the

(IPr/N

2

O and MesCNO/SIPr), which are discussed together

O in

2

−1

k₋₁step (−6 cal mol K ) is less negative than that associated

in the next section. MesCNO most closely resembles N

2

−

1

−1

with the k step (−27 cal mol K ). Thus, as the temperature

terms of structure and bonding, and yet it reacts much faster.

This is attributed to the fact that while both ligands are

computed to bind in an endothermic reaction to 1,

2

is raised, k₋₁will increase much more rapidly than k . This has

2

the consequence of moving the transition state in the reaction

from TS1 to TS2 in Figure 12 with the counterintuitive result

that third-order kinetics is followed at high T and second-order

at low T. The slowness of the reaction, compared with other

binding reactions, is reﬂected primarily in the unfavorable

entropy of activation.

−1

coordination of MesCNO (4.5 kcal mol ) is less endothermic

−

1

than the O-binding of N

with MesCNO, it is postulated that k−1 plays a possible role in

slowing down the reaction. Provided k−1 ≫ k , the net rate for

OAT via a steady-state mechanism becomes kobs = k

K k . Experimental plots of kobs versus MesCNO concen-

eq 2

O (12.1 kcal mol ). In reaction of 1

2

k /k−1 =

1 2

It is of interest to compare reaction of 2 mol of 1 with N O

2

to the quite diﬀerent results when Mo rather than V is the

tration, particularly at higher temperatures, appear to give

ⁿ^oⁿ^z^e^r^oⁱⁿ^t^e^r^c^e^p^t^s^s^u^g^g^e^s^tⁱⁿ^g^t^h^a^t^k−1 may play a role in the

kinetic scheme.

metal center as shown in eq 15 where N O undergoes dinuclear

2

4

.6. IPr/N O and SIPr/MesCNO. As shown in Table 1,

2

DFT results show enhanced enthalpies of binding of the NHC

adducts of N O and MesCNO compared with the parent

2

substrates. This is presumably due to increased electron density

on the terminal O atom due to inductive eﬀects from the highly

basic NHCs. Computations predict that the adducts between

4

2

activation with cleavage of the N−N rather rather than the

N−O bond. The rate-determining step in dinuclear cleavage of

IPr/N O and 1 should be energetically similar for both low-

2

and high-spin states using a truncated ligand system (IMe

1

1369

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

rather than IPr) suggesting that both species may in principle

dbabhNO to O, anthracene, and N and it is gaining the

2

coexist in equilibrium. The TD DFT computed UV−vis

resonance energy of anthracene and stability of N , which give

2

spectrum for the singlet 1−ON /IMe adduct (See Supporting

this stable compound such a low formal BDE. The initial stages

of OAT, however, need not involve complete fragmentation of

the molecule. As discussed above for SIPr/MesCNO, following

spin pairing dissociation could conceivably proceed by a rapid

two-step process as shown in eq 18:

2

Information) is in good agreement with the actual spectral data

of the intermediate detected in stopped ﬂow experiments. For

SIPr/MesCNO, DFT calculations predict that a high-spin

triplet state is most stable.

Global modeling of the IPr/N O kinetic data yielded

2

⧧

activation parameters for ligand binding (k ) of ΔH = 5.6 ±

1

−

1

⧧

−1 −1

0

(

.4 kcal mol and ΔS = −18 ± 2 cal mol

K

and for OAT

1

⧧

−1

⧧

−1 −1

k ) of ΔH = 13 ± 1 kcal mol and ΔS = 1 ± 2 cal mol K .

2

⧧

These data can be compared with SIPr/MesCNO: ΔH =

1

−

1

⧧

−1

4

.9 ± 0.4 kcal mol and ΔS = −19 ± 2 cal mol

K

for

2

1

⧧

−1

⧧

Such a two-step process would mean that not all of the

energy of the reaction is released in step 1 because generation

of anthracene and dinitrogen in step 2 will be exothermic. This

could conceivably be rapid even on the stopped ﬂow time scale.

Based on thermochemistry alone, dbabhNO would be expected

to be the most rapid OAT reagent. It is possible that the fact

that it is not the fastest is due simply to a low rate of binding

and that in fact OAT is “instantaneous” once bound. More

detailed experimental and computational work on dbabhNO is

in progress.

ligand binding and ΔH = 9.3 ± 0.3 kcal mol and ΔS =

2

−

1

−1

−

9 ± 1 cal mol

K

for OAT. It is worth noting that

MesCNO reacts faster than N O and that order is retained in

2

the NHC adducts as well.

An unanswered question in these complex systems is at what

point do they fragment to smaller pieces. This is illustrated for

SIPr/MesCNO with the two possible OAT pathways shown in

eqs 16 and 17.

5

. CONCLUSION

The original goal of this work was to answer the simple

question of whether adducts of MesCNO and N O reacted

2

faster, slower, or at the same rate as the unbound substrates.

Fortuitous kinetic results allowed us to take a more complete

view of OAT in the XNO series. While the facts lie in the rate

and activation parameters themselves, the mechanistic picture

that evolves is not as simple as the authors anticipated. It seems

clear, as pointed out by other workers in this area, that detailed

mechanistic understanding must necessarily involve a complete

picture involving binding, spin pairing, electron transfer, bond

cleavage, and fragment elimination. These need not occur in

that order and may appear in diﬀerent steps for diﬀerent

substrates. For example, PhNO undergoes binding and electron

transfer and spin pairing in one step; in contrast PyO simply

binds in a high-spin state and later undergoes electron transfer

and spin pairing before going onward to complete OAT. It is

tempting to state that initial binding is the key factor in the

OAT process in most cases, but that hides the fact that binding

may be diﬀerent for diﬀerent substrates. The main diﬀerence in

As shown in eq 16, actual OAT could occur with dissociation

of an intact SIPr/MesCN fragment, which would subsequently

44

dissociate rapidly to MesCN and SIPr. A second possibility as

shown in eq 17 is that the NHC would initially dissociate from

the bound adduct, followed by OAT and dissociation of

MesCN. In this second scenario, the NHC could be viewed as a

catalyst that delivers MesCNO in a conﬁguration more suitable

for OAT and then dissociates. The accelerating eﬀect of NHC

coordination suggests their possible utility as catalysts for

45

substrate activation. That acceleration is attributed to both

increased favorability of ligand binding (k /k ) and also the

1

−1

accessibility of π* orbitals to participate in the OAT (k ) step.

2

the rate of binding of PyO and N O is the entropy of activation

4.7. dbabhNO. The net rate of reaction of dbabhNO is

2

to form the complex. Surprisingly, the enthalpies of activation

overlap within experimental error. This may relate to the

transition state for binding being relatively distant from the V

center for the high-spin PyO adduct, which remains para-

intermediate between SIPr/MesCNO and IPr/N O with the

2

important diﬀerence that no intermediate adduct is detected.

This implies that the steady-state approximation is valid, that

intermediates decay at about the same rate as they are

produced. The fact that the overall reaction is relatively fast and

near the range expected for rates of ligand binding (unlike

MesCNO) implies a more eﬃcient OAT than that for

magnetic, compared with N O, which is computed to yield a

2

low-spin complex in which oxidation of 1 has already occurred

and which is formulated as a bent diamagnetic complex

1N−NO. This may require a much more constricted reac-

tion pathway despite the fact that both are “ligand binding”

events. More detailed computational and additional experimental

work is in progress to further deﬁne this critical process. The

current work presents the experimental backdrop for those

studies, but has also generated several surprising results:

3

−1 −1

MesCNO. It can be postulated that k > 1.8 × 10 M

s

1

for dbabhNO at −62 °C. For IPr/N O, which reacts more

2

−

1 −1

slowly than dbabhNO, k = 785 ± 54 M

s

and k = 0.308 ±

2

1

_.₀₁₄_s₋1 at −62 °C. It is clear that k is larger for dbabhNO

0

1

and also that the net rate of reaction is faster. Since

intermediate does not build up, it can be concluded that k₂

−1

must also be greater for dbabhNO than for IPr/N O.

a. The V−O bond in 1O (BDE = 154 kcal mol ) is

strong enough for 1 to cleave all NO bonds except for

NO itself in a direct reaction.

2

−

1

The extremely weak N−O BDE of 10 kcal mol based on

thermochemical measurements would suggest extremely facile

OAT to 1 based on thermodynamics alone. However, the

thermodynamic BDE is based on full fragmentation of

b. Rates of ligand binding play a role in the overall OAT

process, but it is complex since both low-spin and

1

1370

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

high-spin adducts occur depending upon XNO. Crossing

may occur before or after adduct formation. In some

cases, ligand binding is the rate-determining step; in

some cases, OAT from the bound complex is the rate-

determining step.

(E.V.R.-A.), and CRIF-0639138 (E.V.R.-A.), (b) the Spanish

Ministry of Economy and Competitiveness (MINECO) under

Grant CTQ2012-36966 (M.T.), (c) the Natural Sciences and

Engineering Research Council of Canada postdoctoral fellow-

ship (A.F.C.), and (d) a GAANN Fellowship (T.D.P.).

c. The relatively slow rate of OAT of PyO is attributed to

slow conversion of the high-spin 1−OPy adduct to a

low-spin form. The lack of an available NO π* orbital

may force a more bent transition state than for other

adducts.

REFERENCES

■

(

1) Abu-Omar, M. M. in Physical Inorganic Chemistry. Reactions,

Processes, and Applications; Bakac, A., Ed.; Wiley: Hoboken, NJ, 2010;

pp 75−108.

2) Ravishankara, A. R.; Daniel, J. S.; Portmann, R. W. Science 2009,

d. The slow rate of OAT for PhNO, despite detection of a

stable low-spin adduct, is due to the stability of the

adduct and a relatively large enthalpy of activation for

(

3

(

26, 123−125.

3) Zhang, F.; Chen, X.; Zhuang, J.; Xiao, Q.; Zhong, Y.; Zhu, W.

3

expulsion of the unstable NPh fragment.

Catal. Sci. Technol. 2011, 1, 1250−1255.

(4) Tolman, W. B. Angew. Chem., Int. Ed. 2010, 49, 1018−1024.

(5) Armor, J. N.; Taube, H. J. Am. Chem. Soc. 1971, 93, 6476−6480.

e. No simple correlation was found between BDE and rates

of OAT. Despite a formally stronger N−O BDE and

increased steric hindrance, the adducts SIPr/MesCNO

(

6) (a) Yu, H.; Jia, G.; Lin, Z. Organometallics 2009, 28, 1158−1164.

(

b) Yu, H.; Jia, G.; Lin, Z. Organometallics 2008, 27, 3825−3833.

and IPr/N O undergo more rapid OAT to the bulky

2

c) Yu, H.; Jia, G.; Lin, Z. Organometallics 2007, 26, 6769−6777.

complex 1 than MesCNO and N O themselves. This is

2

(

7) Fan, H.; Caulton, K. G. Polyhedron 2007, 26, 4731−4736.

8) Piro, N. A.; Lichterman, M. F.; Harman, W. H.; Chang, C. J. J.

attributed primarily to electronic factors making available

an NO π* orbital.

Am. Chem. Soc. 2011, 133, 2108−2111.

9) (a) Jana, A.; Roesky, H. W.; Schulzke, C. Dalton Trans. 2010, 39,

32−138. (b) Demir, S.; Montalvo, E.; Ziller, J. W.; Meyer, G.; Evans,

f. OAT with N O occurs in a termolecular process

2

(

1

involving reversible formation of an N-bonded adduct

1−NNO, which then forms 1−NNO−1 prior to OAT.

W. J. Organometallics 2010, 29, 6608−6611. (c) Harrold, N. D.;

Waterman, R.; Hillhouse, G. L.; Cundari, T. R. J. Am. Chem. Soc. 2009,

131, 12872−12873. (d) Harman, W. H.; Chang, C. J. J. Am. Chem. Soc.

2007, 129, 15128−15129. (e) Kaplan, A. W.; Bergman, R. G.

Organometallics 1997, 16, 1106−1108.

The rate-determining step changes as a function of

temperature due to the activation parameters for the

individual binding and OAT step involved as discussed in

the text.

(10) (a) Laplaza, C. E.; Odom, A. L.; Davis, W. M.; Cummins, C. C.;

g. The unique OAT reagent dbabhNO undergoes rapid

−1

Protasiewicz, J. D. J. Am. Chem. Soc. 1995, 117, 4999−5000.

(b) Cherry, J. P. F.; Johnson, A. R.; Baraldo, L. M.; Tsai, Y. C.;

Cummins, C. C.; Kryatov, S. V.; Rybak-Akimova, E. V.; Capps, K. B.;

Hoff, C. D.; Haar, C. M.; Nolan, S. P. J. Am. Chem. Soc. 2001, 123,

OAT with 1 and its N−O BDE of 10 kcal mol implies

3

that thermochemically generation of a O atom from this

stable reagent is thermodynamically favored at room

temperature.

7

(

271−7286.

h. It appears that bent structures such as SIPr/MesCNO,

11) Parmon, V. N.; Panov, G. I.; Uriarte, A.; Noskov, A. S. Catal.

IPr/N O, and 1−N O undergo more rapid OAT than

2

Today 2005, 100, 115−131.

linear precursors such as N O and MesCNO. The

2

(12) Schultz, B. E.; Holm, R. H. Inorg. Chem. 1993, 32, 4244−4248.

(13) McPherson, L. D.; Drees, M.; Khan, S. I.; Strassner, T.; Abu-

Omar, M. M. Inorg. Chem. 2004, 43, 4036−4050.

enhanced rate of OAT for the NHC bound complexes

may lead to their utilization as accelerants in some

reactions in designed complexes.

(14) (a) Veige, A. S.; Slaughter, L. M.; Wolczanski, P. T.; Matsunaga,

N.; Decker, S. A.; Cundari, T. R. J. Am. Chem. Soc. 2001, 123, 6419−

Additional work is in progress to further map the kinetics,

thermodynamics, and complete reaction proﬁle for OAT

involving the N−O bond and transition metals of potential

catalytic utility.

6

420. (b) Veige, A. S.; Slaughter, L. M.; Lobkovsky, E. B.; Wolczanski,

P. T.; Matsunaga, N.; Decker, S. A.; Cundari, T. R. Inorg. Chem. 2003,

2, 6204−6224.

15) Crestoni, M. E.; Fornarini, S.; Lanucara, F.; Warren, J. J.; Mayer,

4

(

J. M. J. Am. Chem. Soc. 2010, 132, 4336−4343.

ASSOCIATED CONTENT

Supporting Information

Experimental details and supporting data as mentioned in the

text. This material is available free of charge via the Internet at

http://pubs.acs.org.

■

(16) Fortner, K. C.; Laitar, D. S.; Muldoon, J.; Pu, L.; Braun-Sand, S.

B.; Wiest, O.; Brown, S. N. J. Am. Chem. Soc. 2007, 129, 588−600.

*

S

(17) Vasbinder, M. J.; Espenson, J. H. Organometallics 2004, 23,

3

(

355−3358.

18) Park, J.; Morimoto, Y.; Lee, Y. M.; Nam, W.; Fukuzumi, S. J. Am.

Chem. Soc. 2011, 133, 5236−5239.

19) Keith, J. M.; Tomic, Z. D.; Zaric,

A: Chem. 2010, 324, 15−23.

20) (a) O’Reilly, M. E.; Del Castillo, T. J.; Abboud, K. A.; Veige, A.

(

́

S. D.; Hall, M. B. J. Mol. Catal.

AUTHOR INFORMATION

Corresponding Author

elena.rybak-akimova@tufts.edu; manuel.temprado@uah.es;

ccummins@mit.edu; captain@miami.edu; c.ho ﬀ @miami.edu.

Notes

■

(

S. Dalton Trans. 2012, 41, 2237−2246. (b) O’Reilly, M. E.; Ghiviriga,

I.; Abboud, K. A.; Veige, A. S. J. Am. Chem. Soc. 2012, 134, 11185−

1

1195.

(21) Cai, X.; Majumdar, S.; Fortman, G. C.; Frutos, L. M.;

The authors declare no competing ﬁnancial interest.

Temprado, M.; Clough, C. R.; Cummins, C. C.; Germain, M. E.;

Palluccio, T.; Rybak-Akimova, E. V.; Captain, B.; Hoff, C. D. Inorg.

Chem. 2011, 50, 9620−9630.

ACKNOWLEDGMENTS

The authors dedicate this paper to Professor Robert E. Gawley

1948−2013). This material is based upon work supported by

a) the National Science Foundation under Grants CHE-

■

(

1

(22) Tskhovrebov, A. G.; Solari, E.; Wodrich, M. D.; Scopelliti, R.;

Severin, K. Angew. Chem., Int. Ed. 2012, 51, 232−234.

(23) Tskhovrebov, A. G.; Solari, E.; Wodrich, M. D.; Scopelliti, R.;

Severin, K. J. Am. Chem. Soc. 2012, 134, 1471−1473.

111357 (C.C.C.), CHE-0615743 (C.D.H.), CHE-0750140

1

1371

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Journal of the American Chemical Society

Article

(

24) More detailed computational, as well as additional experimental,

studies of OAT between XNO and diﬀerent metal complexes are in

progress.

(

25) Ruppa, K. B. P.; Desmangles, N.; Gambarotta, S.; Yap, G.;

Rheingold, A. L. Inorg. Chem. 1997, 36, 1194−1197.

26) The structure of the analogous 1S has been reported: Brask, J.

K.; Dura-Vila, V.; Diaconescu, P. L.; Cummins, C. C. Chem. Commun.

002, 902−903.

27) Carpino, L. A.; Padykula, R. E.; Barr, D. E.; Hall, F. H.; Krause,

J. G.; Dufresne, R. F.; Thoman, C. J. J. Org. Chem. 1988, 53, 2565−

572.

28) (a) Fickes, M. G.; Davis, W. M.; Cummins, C. C. J. Am. Chem.

(

2

(

2

(

Soc. 1995, 117, 6384−6385. (b) Proulx, G.; Bergman, R. G. J. Am.

Chem. Soc. 1995, 117, 6382−6383.

(

29) Luo, Y. R. Handbook of Bond Dissociation Energies in Organic

Compounds ; CRC Press: Boca Raton, FL, 2003.

(

30) http://webbook.nist.gov.

31) It is standard to use TΔS ≈ 10 kcal mol at rt for each gain in

−1

total number of particles of products minus reactants, see: Page, M. I.

Angew. Chem., Int. Ed. Engl. 1977, 16, 449−459. In reaction 8, since

there is a two particle advantage for the products, the favorable TΔS

term of approximately −20 kcal/mol will overcome the ΔH of

approximately 10 kcal/mol to yield an estimated ΔG on the order of

−

10 ± 5 kcal/mol.

(

32) Gritsan, N. P.; Platz, M. S. Chem. Rev. 2006, 106, 3844−3867.

(

33) (a) Platz, M. S. Acc. Chem. Res. 1995, 28, 487−492. (b) Borden,

W. T.; Gritsan, N. P.; Hadad, C. M.; Karney, W. L.; Kemnitz, C. R.;

Platz, M. S. Acc. Chem. Res. 2000, 33, 765−777.

(

34) (a) Liang, T. -Y.; Schuster, G. B. J. Am. Chem. Soc. 1987, 109,

7

803−7810. (b) Takaoka, A.; Moret, M.-E.; Peters, J. C. J. Am. Chem.

Soc. 2012, 134, 6695−6706.

35) For example, a dinuclear reaction with a rate-limiting

mononuclear ﬁrst step between an activated form of 1−ONPh and

−ONPh to yield PhNNPh and two 1O or an activated form of 1−

ONPh and 1 to yield 1O and 1NPh.

36) In most systems, the change in absorbance at λ = 453 nm (λ

(

1

(

max

for 1O) was followed. For reactions with IPr/N O, however, the

2

absorbance was too intense in that region of the spectrum. Analysis of

data at 453 nm, when possible, was in agreement with analysis at λ =

7

(

6

(

25 nm.

37) Moerdyk, J. P.; Bielawski, C. W. J. Am. Chem. Soc. 2012, 134,

116−6119.

38) The most commonly cited OM bond strength measured for a

complex in solution is that of Watt, G. D.; McDonald, J. W.; Newton,

W. E. J. Less-Common Met. 1977, 54, 415.

(

1

(

39) Carreon-Macedo, J. L.; Harvey, J. N. J. Am. Chem. Soc. 2004,

26, 5789−5797.

40) Temprado, M.; McDonough, J. E.; Mendiratta, A.; Tsai, Y. C.;

Fortman, G. C.; Cummins, C. C.; Rybak-Akimova, E. V.; Hoff, C. D.

Inorg. Chem. 2008, 47, 9380−9389.

(

41) Cozzolino, A. F.; Tofan, D.; Cummins, C. C.; Temprado, M.;

Pallucio, T. D.; Rybak Akimova, E. V.; Majumdar, S.; Cai, X.; Captain,

B.; Hoff, C. D. J. Am. Chem. Soc. 2012, 134, 18249−18252.

(

42) Cherry, J. P. F.; Johnson, A. R.; Baraldo, L. M.; Tsai, Y. C.;

Cummins, C. C.; Kryatov, S. V.; Rybak-Akimova, E. V.; Capps, K. B.;

Hoff, C. D.; Haar, C. M.; Nolan, S. P. J. Am. Chem. Soc. 2001, 123,

7

(

271−7286.

43) Khoroshun, D. V.; Musaev, D. G.; Morokuma, K. Organo-

metallics 1999, 18, 5653−5660.

44) Some NHC/nitrile adducts have been reported to be stable; no

adduct exists between MesCN and SIPr to the authors’ knowledge.

45) Enders, D.; Niemeier, O.; Henseler, A. Chem. Rev. 2007, 107,

606−5655.

(

5

1

1372

dx.doi.org/10.1021/ja405395z | J. Am. Chem. Soc. 2013, 135, 11357−11372

Article Doi

Thermodynamic and kinetic study of cleavage of the N-O bond of N-oxides by a vanadium(III) complex: Enhanced oxygen atom transfer reaction rates for adducts of nitrous oxide and mesityl nitrile oxide

DOI: 10.1021/ja405395z

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja405395z

Full text of DOI:10.1021/ja405395z

Products guided by the article

R&D Labs maybe for 120-12-7

Relevant to this article

Hot Product