Predicted Chemical Activation Rate Constants for HO2 + CH2NH: The

Subscriber access provided by United Arab Emirates University | Libraries Deanship

Article 2

2

Predicted Chemical Activation Rate Constants for HO + CHNH: The Dominant Role of a Hydrogen-Bonded Pre-Reactive Complex Mohamad Akbar Ali, Jason A. Sonk, and John R. Barker J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.6b06531 • Publication Date (Web): 16 Aug 2016 Downloaded from http://pubs.acs.org on August 18, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Predicted Chemical Activation Rate Constants for HO2 + CH2NH: The Dominant Role of a Hydrogen-Bonded Pre-Reactive Complex Mohamad Akbar Ali*, Jason A. Sonk, and John R. Barker* Department of Climate and Space Sciences and Engineering University of Michigan, Ann Arbor, MI 48109-2143 Email: [email protected], [email protected]

Abstract The reaction of methanimine (CH2NH) with the hydroperoxy (HO2) radical has been investigated by using a combination of ab initio and density functional theory (CCSD(T)/CBSB7//B3LYP+Dispersion/CBSB7) and master equation calculations based on transition state theory (TST). Variational TST was used to compute both canonical (CVTST) and microcanonical (µVTST) rate constants for barrierless reactions. The title reaction starts with the reversible formation of a cyclic pre-reactive complex (PRC) that is bound by ~11 kcal/mol and contains hydrogen bonds to both nitrogen and oxygen. The reaction path for the entrance channel was investigated by a series of constrained optimizations, which showed that the reaction is barrierless (i.e. no intrinsic energy barrier along the path). However, the variations in the potential energy, vibrational frequencies, and rotational constants reveal that the two hydrogen bonds are formed sequentially, producing two reaction flux bottlenecks (i.e. two transition states) along the reaction path, which were modeled using W. H. Miller’s unified TST approach. The rate constant computed for the formation of the PRC is pressure-dependent and increases at lower temperatures. Under atmospheric conditions, the PRC dissociates rapidly and its lifetime is too short for it to undergo significant bimolecular reaction with other species. A small fraction isomerizes via a cyclic transition state and subsequent reactions lead to products normally expected from hydrogen abstraction reactions. The kinetics of the HO2 + CH2NH reaction system differs substantially from the analogous isoelectronic reaction systems involving C2H4 and CH2O, which have been the subjects of previous experimental and theoretical studies.

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

1. Introduction Amines are emitted into the atmosphere from a variety of biogenic, oceanic and anthropogenic sources, including animal husbandry, marine emissions, biomass burning, and chemical manufacturing. In the future, amines may be released from carbon capture and storage technologies.1,

2

After emission into the atmosphere, amines undergo conversion reactions in

both the gas and aqueous phases. Amines undergoing aqueous phase reactions can participate in the growth of secondary organic aerosol (SOA) and brown carbon.1, 3-6 At ambient atmospheric concentrations amines are not very harmful, but they undergo reactions that produce known or suspected carcinogens, such as nitrosamines and nitramines.

1, 4, 6

7

The atmospheric chemistry of

amines has been discussed by Schade and Crutzen and reviewed by Nielson et al.,

1

who

summarized several detailed oxidation pathways initiated by the reactions amines with of atmospheric OH, Cl, O3, and NO3, in both gas and aqueous phases. The reaction of methylamine with hydroxyl radical (CH3NH2 + OH) is the first step in an important pathway.8, 9 OH radicals initially abstract the hydrogen from the C—H and N—H groups of the amine to generate carboncentered (alkyl) and nitrogen-centered (amino) free radicals. The subsequent reactions of the alkyl and amino free radicals lead to the production of several intermediates and products, including methanimine.10 Methanimine (CH2NH) is produced in the atmospheric photo-oxidation of methylamine and from the decomposition of methyl azides.11 Methanimine is isoelectronic with both formaldehyde and ethylene, and it can react in several ways. It has been suggested as a potential prebiotic precursor of glycine12,

13

and it has been detected spectroscopically in interstellar

clouds, suggesting that it is produced and reacts in interstellar space.14-18 High level theoretical studies of the atmospheric reactions of CH2NH initiated by HO radicals have been performed by Bukan et al.19, and Ali and Barker.20 Both studies utilized coupled cluster theory with the perturbative triples correction (CCSD(T)) and triple zeta basis sets. The calculated rate constants obtained in the two studies are in very good agreement with each other. Ali and Barker also computed reaction rate constants for hydroxyl radical reacting with formaldehyde and ethylene, which are isoelectronic with methanimine. The present work expands on that study by investigating the reaction of HO2 with CH2NH. To the best of our knowledge, there have been no experimental studies of the HO2 + CH2NH reaction. Direct investigation of this reaction probably poses experimental challenges, 2 ACS Paragon Plus Environment

Page 3 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


because methanimine is highly water-soluble and likely adsorbs strongly to surfaces. In the absence of experiments, theory can provide insight and estimates of rate constants and branching ratios, but the feasibility and ultimate accuracy of the calculations is limited by computational costs. As in our previous work on HO + CH2NH, we gain insight into the accuracy of the calculations by considering how well theory performs in describing the isolectronic reactions HO2 + CH2O and HO2 + C2H4, which have been studied previously by others.21-28 Because it has been suggested that these isoelectronic reactions proceed through concerted addition-elimination reactions,21,22,27,28 we have explored several mechanisms for the CH2NH system, including both (a) concerted addition/elimination and (b) addition of HO2 followed (after a delay) by elimination. We have also explored hydrogen abstraction by HO2. We have found that prereactive complexes (PRCs) are important in all of the reaction mechanisms, strongly affecting the rate constants and branching ratios.

2. Methods 2.1. Theory 2.1.1. Master equation By using a master equation formalism the time evolution of a chemical system can be modeled as a function of temperature and pressure. In general, the master equation describing chemical reactions and collisional energy transfer depends on both internal energy and angular momentum (a two-dimensional, or 2-D master equation): ∞

dN(E ', J ',t) = F(E ', J ',t)dE + ∑ ∫ R(E ', J '; E, J )N(E, J;t)dE dt J 0 ∞

channels

0

i=1

− ∑ ∫ R(E, J; E ', J ')N(E ', J ';t)dE − J

∑

(1)

ki (E ', J ')N(E ', J ';t)

where N(E ' J ';t) is the concentration of a chemical species with internal energy E' and angular momentum J' at time t. The first term on the right hand side is a source term that describes the production of the chemical species at energy E ' and angular momentum J'. This production term accounts for photo-activation, chemical activation, isomerization of other species, etc. The second and third terms describe the production and loss of N(E ' J ';t) via collisional energy transfer, where the transition rate coefficient R(E,J;E',J') is a pseudo-first order rate constant for 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 26

energy transfer from E',J' to a new E,J. The forward and reverse terms (i.e. R(E,J;E',J') and R(E',J';E,J)) are related by detailed balance, which depends on the density of states ρ(E,J). The last term on the right hand side describes the total rate of reaction via all available reaction channels; each channel has a unimolecular rate constant, k(E',J'), which depends on energy and angular momentum. In practice the 2-D master equation is reduced to a 1-D equation that depends only on energy. This 1-D master equation can then be solved numerically by standard methods. The method used in the MultiWell Program Suite29-31 is Gillespie’s stochastic simulation algorithm.32 The 1-D master equation requires knowledge of the 1-D density-of-states, ρ (E ') for each reactant and the 1-D microcanonical rate constant, k(E ') , for each reaction channel. In the present work, Eq. 1 was reduced to a 1-D master equation as first described by Smith and Gilbert33 and later extended slightly by Miller and Klippenstein.34 The 2-D microcanonical rate constant, k(E,J), is given by ‡

1 G ( E ′ − E0,J , J ′ ) k( E ′, J ′ ) = L h ρ ( E ′, J ′ ) ‡

(2)

where L‡ is the reaction path degeneracy; h is Planck’s constant; ρ (E ', J ') is the density of ‡

states of the reactant molecule; G ( E ′ − E0,J , J ′ ) is the sum of states of the transition state; and E0,J is the reaction critical energy, which depends on J. As explained by Gilbert and Smith,35 the reaction path degeneracy L‡ usually can be written as the product of a ratio of symmetry factors and a ratio of optical isomers,36 but sometimes direct enumeration is required. The 2-D sum and density of states can be summed over angular momentum to produce the 1-D terms needed for solving the 1-D master equation formalism:

(

)

J max ′

(

G E ′ − E0,0 = ∑ G ‡ E ′ − E0,J , J ′ ‡

ρi (E ′) =

J ′ =0

)

(3)

J ′ max

∑ ρ (E ′, J ′)

J ′ =0

i

(4)

′ is either the value of J ′ consistent with the highest internal energy in the numerical where J max calculation, or the value of J ′ for which a local minimum (i.e. the reactant) exists in the effective potential.


Page 5 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


‡

1 G ( E ′ − E) k( E ′ ) = L h ρ (E ′)

(5)

‡

where the critical energy E0,0 is for J = 0. For reaction channels that have no intrinsic energy barrier, the microcanonical rate constant must be obtained by using a variational treatment (see Ref.

37

and earlier references

cited therein). This can be accomplished by performing a series of constrained optimizations at fixed distances along the reaction path on the potential energy surface (PES). At each fixed distance, the potential energy is calculated, the optimized geometry is used to obtain the rotational constants, and a vibrational analysis is used to obtain the vibrational frequencies of the orthogonal degrees of freedom, after projecting out the reaction coordinate. Using these parameters, "trial" rate constants or reaction fluxes can be computed at each point along the reaction path. The point at which the minimum trial rate constant or reaction flux occurs is identified with the variational transition state. In a microcanonical calculation, Gd‡(E',J') , the sum of states, is computed at each distance (d) and expressed relative to a common zero of energy. For every E',J' of interest, the minimum value of Gd‡(E',J') along the reaction path is identified with the variational G‡(E',J') for the reaction channel. Thus the variational transition state usually occurs at different distances along the path, depending on the specific E',J'. To obtain the high-pressure limit rate constant, which is just a canonical rate constant, the microcanonical rate constant is thermally averaged: ∞

∞

∑ ∫ G (y, J ′)exp[−(y + V ‡

uni ∞, µ

k

(T ) =

J =0 E=0 ∞ ∞

∑∫

J =0 E=0

TS

+ BTS J ′( J ′ +1) / kBT ]dy (6)

ρ (x, J ′ ) exp[−(x + V0 + BR J ′( J ′ +1) / kBT ]dx

where x and y are the internal energies of the Reactant and the TS relative to their respective ZPEs; V0 and VTS, and BR and BTS are the respective potential energies and rotational constants (see Figure 1).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 26

Figure 1: Energy definitions for a unimolecular reaction. The total energy E is relative to the reactant zero point energy, V0 . For reactions with intrinsic energy barriers, the unimolecular rate constant at the highpressure limit, k∞uni (T ) , is computed with Eckart tunneling corrections for an unsymmetrical energy barrier38,39

k CTST (T ) = k∞uni (T ) = ΓL‡ ×

‡ ⎛ ΔE ‡ ⎞ kBT QTS exp ⎜ − 0 ⎟ h QR ⎝ kBT ⎠

(7)

where Γ is the quantum mechanical tunneling correction. The factor kB is Boltzmann’s constant,

ΔE0‡ is the zero-point energy corrected energy difference between the TS and the reactants, and ‡ and QR are the total partition functions for the transition state and the reactants, QTS

respectively, referenced to their respective zero point energies. Similar to microcanonical transition state theory, canonical transition state theory can be treated variationally for pathways ‡ with no intrinsic energy barrier. In this case Eq 7 is variationally minimized as a function of QTS

over the reaction path. The Eckart tunneling correction was a pragmatic choice for present purposes. Although not as accurate as some other methods, it is relatively inexpensive and sufficiently accurate for present purposes, since reactions with intrinsic barriers are relatively unimportant in this system. On occasion, when considering rate constants or fluxes at different points along a reaction path, several local minima may occur. Miller’s unified statistical model42 can be used to ‡ (E) . For two transition states, this takes the form find the effective sum of states Geff

⎡ 1 1 1 ⎤ G (E) = ⎢ ‡ + ‡ − ‡ ⎥ ⎣ G1 (E) G2 (E) Gmax (E) ⎦

−1

‡ eff

(8) 6

ACS Paragon Plus Environment

Page 7 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


‡ where Gmax (E) is the maximum sum of states in the region between the two transition states. ‡ Because Gmax (E) is usually much smaller than G1‡ (E) and G2‡ (E) , Eq. 8 reduces to

⎡ 1 1 ⎤ G (E) = ⎢ ‡ + ‡ ⎥ ⎣ G1 (E) G2 (E) ⎦

−1

‡ eff

(9)

Miller's unified model applies to microcanonical sums of states or rate constants, but an ‡ analogous equation can be written for the partition function QTS or thermal rate constant in a

canonical system. The canonical version of Eq. 9 is not as accurate as the microcanonical version, but it is a useful pragmatic approximation. In previous work,10 the semi-empirical Inverse Laplace Transform (ILT) approach of Forst40 was used to compute microcanonical rate constants for the O2 + CH2NH2 reaction, which has no intrinsic energy barrier. This reaction is part of the same PES as HO2 + CH2NH. It was included in the present calculations by using the same ILT approach with the parameters found previously.10 The ILT semiempirical method requires the density of states of the reacting intermediate and the empirical Arrhenius parameters of the high-pressure limit rate constant, A∞ and E∞ . As in our previous work, E∞ for the dissociation reaction was replaced with E0, the critical energy for the reaction, in order to obtain more accurate values of k(E) near the reaction threshold:

k ( E ) = A∞

ρ ( E − E0 )

(10)

ρ ( E)

where ρ(E) is the density of states of the dissociating intermediate. The critical energy E0 is the enthalpy of reaction at 0 K for the endothermic dissociation reaction: E0 = ∆fHr(0). The Preexponential factor was computed from E0 and the value of kuni at 300 K:

(

A∞ = kuni ( 300 K ) exp Δ f H r ( 0 ) / RT

)

(11)

2.2 Electronic Structure and Kinetics Calculations All electronic structure calculations were carried out using the GAUSSIAN 09 software suite.41 Geometries of stationary points on the PES were optimized using the B3LYP+D4244

density functional with the CBSB745 basis set. The B3LYP+D functional includes the Grimme 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 26

empirical dispersion correction,44 which is useful in searching for van der Waal complexes. Optimized geometries of the stationary points are tabulated in Supporting Information. Vibrational frequencies and rotational constants where calculated using the same level of theory. Single point energy calculations were performed with the optimized geometries to obtain better estimates of the energies; these single point calculations were carried out using coupled cluster theory with the perturbative triples correction, CCSD(T)46, and the same basis set. For open- and closed-shell species, unrestricted and restricted Hartree-Fock (UHF/RHF) reference wave functions were used, respectively. Intrinsic reaction coordinates (IRC) calculations47,

48

were

carried out to confirm the identities of the reactants and products for each transition states. The T1 diagnostic was used for all relevant species to measure the amount of multi-reference character of the wave function; all species were found to have T1≤0.04, the maximum that gives acceptable results using a single reference wave function.49 The entrance channels for the addition reactions have no intrinsic energy barriers. For these cases a better description of the wave function is obtained by allowing the singlet and triplet states to mix (the “GUESS=MIX” keyword was used). The “GUESS=MIX” option mixes the HOMO and LUMO orbitals to break spatial symmetry.

2.3 Kinetics Calculations Kinetics calculations were carried out using programs in the MultiWell Program 29-31

Suite

and a new program "ktools", which will be added to the MultiWell package in a future

release. Ktools computes rate constants using both canonical and microcanonical transition state theory. In the absence of an intrinsic barrier, ktools will calculate the variational reaction rate constant. If multiple reaction bottlenecks are found along the reaction path, ktools automatically utilizes W. H. Miller’s unified transition state model40 to calculate an overall rate constant. Ktools reports the J-resolved 2-D sum and density of state arrays (including the 2J+1 rotational degeneracy factor); it also reports 1-D sum and density of states arrays obtained by summing the 2-D arrays over J (see equations (3) and (4)). Finally, ktools reports canonical (thermal) rate constants obtained in two ways: by averaging the microcanonical rate constant via equation (5) and by using thermal partition functions in the conventional manner: ( µ )CTST and CTST, respectively.


Page 9 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For practical calculations, non-linear reactants and transition states were approximated as symmetric tops. Each symmetric top possesses a degenerate 2-dimensional rotation (the 2D adiabatic "j-rotor") and a 1-dimensional rotor (the "k-rotor"). For non-symmetric molecules, the rotational constant for the j-rotor is approximated as the geometric mean of the two rotational constants that are most similar, in the conventional manner.38 The k-rotor is assumed to exchange energy freely with the vibrational modes, but it also obeys the requirements that the quantum number K, is limited to the range from –J to +J and that all rotational energies are ≥0. In the present work, sums and densities of states were calculated assuming the various degrees of freedom are separable and of the following types: harmonic oscillators, free rotations, and hindered internal rotations. Energy levels for the separable harmonic oscillators and free rotations are from harmonic frequencies and rotational constants obtained from the quantum chemistry calculations. Eigenvalues for the 1-D separable hindered internal rotations were computed internally from the 1-D Schrödinger equation, based on effective masses and potential energies computed as functions of dihedral angle (see the MultiWell User Manual for details and references). The total density of states for a chemical species was obtained by convolving the separate densities of states for each degree of freedom, according to the Stein-Rabinovitch50 extension of the Beyer-Swinehart algorithm.51 An energy grain size of 10 cm-1 was used for all density of states and master equation calculations.

2.4 Master Equation Calculations Microcanonical k(E) values were computed by the MultiWell master equation code from the 1-D sums and densities of states and Eq. 5, as described above. The microcanonical rate constants and densities of states were used in master equation simulations to obtain rate constants and branching ratios at pressures ranging from 0.001 to 1000 bar, and at temperatures ranging from 200 to 350 K. As in our earlier work,10 collisional energy transfer between the excited species and Nitrogen gas (N2) was modeled using a conventional exponential-down model with the average energy of down-steps assumed to be down = 200 cm−1. The bimolecular rate constant describing collisions between N2 and the intermediate (“Well”) was based on LennardJones collisions with net parameters obtained using the usual combining rules from parameters for the two collision partners: N2 gas (σ = 3.74 Å and ε/kB = 82 K)52 and the Well ( σ = 5.20 Å 9 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 26

and ε/kB = 212 K, used for all Wells).10 The latter parameters were estimated by Rissanen et al.,10 who used n-butane as a model. The double arrays used in MultiWell for the master equation simulations consisted of 1000 array elements with 10 cm−1 energy grains, with the quasi-continuum regime evaluated up to 50000 cm−1. At each temperature and pressure, master equation simulations were initiated using the chemical activation energy distribution, which is appropriate for recombination reactions.29,

53

The master equation simulations consisted of 107 stochastic trials, each with a

simulated time duration corresponding to an average of 100 collisions. Tests showed that the chemical activation was essentially completely quenched by collisional thermalization during this time period. Each simulation resulted in the computed fractional yields of intermediates and bimolecular products. The temperature- and pressure-dependent chemical activation total bimolecular constant is given by: 29, 53 rec ktot (T , M ) = Keq k∞uni (T ) ⎡⎣1− f HO2 +CH2 NH ⎤⎦

(12)

where k∞uni (T ) is the high-pressure limit rate constant for PRC → HO2 + CH2NH, Keq is the equilibrium constant, and fHO2 +CH 2 NH is the calculated fractional yield of reactants regenerated by the back-reaction of PRC during the 100 collision time period.

3. Results and Discussion 3.1. Potential Energy Surface The stationary points on the PES for HO2 + CH2NH are shown in Figure 2. The reaction is found to occur via weakly bound PRCs with Cs symmetry. Optimized structures of the reactants, PRC1, and TS1 are shown in Figure 3. The optimized geometries of other complexes and transition states are supplied in the Supporting Information. When chemical species have more than one conformation, the energy of the most stable conformer was used.


Page 11 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2: Potential Energy Surface for HO2 + CH2NH (addition pathways).

Figure 3: Optimized structure of reactants, pre-reactive complex and transition state obtained using B3LYP+D/CBSQB7. In PRC1 the two reactant moieties adopt geometries that are only slightly perturbed versions of the separated reactants, suggesting that valence electrons are not significantly 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 26

involved and the two species are bound together mostly by electrostatic forces. The PRCs present on the HO2 + CH2NH PES are analogous to those on the HO + CH2NH PES,20 but the binding energy for the former (~11 kcal/mol) is more than twice that of the latter (~5 kcal/mol). The larger binding energy can greatly affect the kinetics. The structures of the PRCs present in the HO2 + CH2O and C2H4 systems are also similar to those in the HO2 + CH2NH system, but the energies are very different. The binding energy of PRC1 is around 11 kcal/mol for the CH2NH system compared to 6 kcal/mol and 4 kcal/mol for the CH2O and C2H4 systems respectively. The five membered ring structure of PRC1 and its large binding energy of ~11 kcal/mol are due the combined effects of the N—H and O—H hydrogen bonds. Because of its strong binding energy, PRC1 is a key intermediate in the reaction between HO2 and CH2NH. The reaction proceeds by first forming PRC1, which can re-dissociate back to reactants, or can react further by two distinct pathways: via radical addition to the C-atom in CH2NH to form Int−2, which is a substituted amino radical (•NHCH2OOH), or via a concerted addition reaction to form Int−1, which is a peroxy radical (H2NCH2OO•). The radical addition reaction proceeds through transition state TS2 in which the C=N double bond is partially broken and a partial bond is formed between the O-O group and one side of the double bond. The concerted addition pathway can be visualized as initial chemical binding of the HO2 moiety to the C-atom in CH2NH (as in the radical addition pathway), immediately followed by a “tailbiting” isomerization, in which the N-atom (which has free-radical character) abstracts the Hatom from the HO2 group via a five-member cyclic transition state. The radical and concerted addition reactions have barrier heights of 31.6 and 13.5 kcal/mol, respectively. The intermediates Int-1 and Int-2 can undergo reverse isomerization to regenerate PRC1, or they can dissociate to produce two sets of bimolecular products: HO + HC(=O)NH2 and O2 + CH2NH2. (Note that if these products were observed in an experiment, they might mistakenly be interpreted as the results of O-atom and H-atom metathesis reaction, respectively, but the theoretical calculations show that they are the result of more complicated process.) Because of their large energy barriers, both thermal dissociation reactions are expected to be slow, but the second reaction is the faster of the two. The present results show that the addition of HO2 radical to the nitrogen atom in CH2NH is not favored, in agreement with a conclusion of Rissanen et al.10 The subsequent unimolecular dissociation channel producing OH + c-OCH2NH is very 12 ACS Paragon Plus Environment

Page 13 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


slow, since the barrier is very high (~30 kcal/mol). In the HO2 + CH2NH system, transition states TS1 and TS3 are higher in energy than the reactants. This is in contrast to HO2 + CH2O, where the lowest energy transition state is lower than the energy of the reactants. Compared to the isoelectronic analogues, the transition states in the HO2 + CH2NH reaction system have larger imaginary frequencies, suggesting faster quantum mechanical tunneling. In CH2NH, the three H-atoms reside in three unique structural environments, suggesting that three unique hydrogen abstraction reactions are possible. Abstraction from the C—H and N—H groups in CH2NH to produce •CHNH + H2O2 and CH2N• + H2O2, respectively. The present calculations show that all three of the abstraction reactions proceed by first forming a one of three PRCs, followed by transfer of a hydrogen atom and final separation of products (see Figure 4). Intrinsic Reaction Coordinate calculations show that after a PRC is formed, an internal rotation takes place until the terminal O-atom of the HO2 moiety approaches an H-atom in the CH2NH moiety. The barrier heights for hydrogen abstraction from the cis-CH, the transCH, and the NH are 18.9 kcal mol-1, 20.9 kcal mol-1, and 10.6 kcal/mol, respectively. Thus abstraction from the N-H bond is strongly favored. This is also true for abstraction by HO free radicals, but the barrier heights for H-abstraction by HO2 radicals are much higher than those for abstraction by HO radicals.

Figure 4: Hydrogen abstraction pathways on the HO2 + CH2NH potential energy surface.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 26

3.2. Transition states for barrierless reactions The PRCs that play such important roles in the HO2 + CH2NH system are formed via entrance channels that have no local maxima on the PES (i.e. they have no intrinsic energy barriers). The potential energy profile for the entrance channel forming PRC1 is shown in Figure 5 and optimized geometries at some points along the reaction pathway are shown in Figure 6. In addition, the lengths of the two hydrogen bonds (N---HOO and HOO---H) are shown as a function of the N--O bond distance (RN-O) in Figure S4 (Supporting Information). The Gibbs free energy profile for the entrance channel forming PRC1 is shown in Figure S5 for several temperatures. Two maxima in the Gibbs free energy are clearly present on each curve. At RNO

>7 Å, the interactions between HO2 radical and CH2NH are very weak. At RN-O = 7 Å, the N---

HOO hydrogen bond (HB) is very long (7.0 Å), while the HOO--H HB is only (3.8 Å). At this point, the binding energy is

Predicted Chemical Activation Rate Constants for HO2 + CH2NH: The

Recommend Documents