Mechanism of Photoinduced Triplet Intermolecular Hydrogen Transfer

Triplet Intermolecular Hydrogen Transfer Between Cycloxydim and Chlorothalonil. Qi Yuan , Dimitrios Toroz , Nathan Kidley , and Ian R. Gould. J. P...
0 downloads 10 Views 1MB Size
Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

A: Spectroscopy, Photochemistry, and Excited States

Mechanism of Photoinduced Triplet Intermolecular Hydrogen Transfer Between Cycloxydim and Chlorothalonil Qi Yuan, Dimitrios Toroz, Nathan Kidley, and Ian R. Gould J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.7b12523 • Publication Date (Web): 16 Apr 2018 Downloaded from http://pubs.acs.org on April 16, 2018

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 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 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.

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 29 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

Mechanism Of Photoinduced Triplet Intermolecular Hydrogen Transfer Between Cycloxydim And Chlorothalonil Qi Yuan †, Dimitrios Toroz†, Nathan Kidley‡ and Ian R. Gould*† AUTHOR ADDRESS †

Institute of Chemical Biology, Department of Chemistry, Imperial College London, SW7 2AZ,

UK. E-mail: [email protected]

Syngenta, Jealott’s Hill, Bracknell, Berkshire, London, RG42 6EY, UK.

ACS Paragon Plus Environment

1

The Journal of Physical Chemistry 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 29

Abstract

The possible reaction mechanisms for the experimentally observed hydrogen transfer between herbicide Cycloxydim (CD) and triplet fungicide Chlorothalonil (CT) were identified with density functional theory (DFT) and time dependent density function theory (TDDFT) computations. Excited energy transfer (EET) calculations indicate that reactants for intermolecular hydrogen transfer were formed via energy transfer from triplet CT to ground state CD. Three possible reaction pathways after EET were identified, and hydrogen transfer from the hydroxyl group on the cyclohexane ring of CD to CT exhibited the lowest energy barrier. Natural Population Analysis (NPA) along the reaction pathways has confirmed that the pathways involved either electron transfer induced proton transfer or coupled electron-proton transfer, leading to different potential energy profiles. Electrostatic potential (ESP) study substantiated the reaction mechanisms in different pathways. This study suggests an explanation for the accelerated photodegradation of CD by CT and provides a pipeline for future studies of photoinduced intermolecular hydrogen transfer.

ACS Paragon Plus Environment

2

Page 3 of 29 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

Introduction Photodegradation is one of the major environmental fates of agrochemical compounds (e.g. herbicides, pesticides, and fungicides) after they have been applied to plants1. The photochemical processes that agrochemical compounds and residues undergo is of great importance to both the academic community and agrochemical industry since a better understanding of such processes can provide insight into the design of new pesticides. For example, it is important that photodegradation products have no adverse environmental impacts. To determine the photochemical fate of agrochemical compounds, both experimental2-9 and computational10-11 studies on the photodegradation of the active component of agrochemical compounds have been performed. However, application of adjuvants in combination with active ingredients12-13, as well as mixtures of different agrochemical compounds14 is becoming a dominant trait in modern agriculture because suitable adjuvants can improve performance and are time and cost efficient. Mixtures of agrochemical compounds, which are usually applied in herbicide tanks15-17, can also enhance efficacy and manage resistance. It should be noted, however, that the interaction between different agrochemical compounds might influence their photochemical fate, thus influencing the photostability and efficacy of the active compounds. Although several experimental studies18-19 have been performed on the photostability variations of mixed agrochemicals, the mechanism that alters the photostability upon mixture of agrochemicals are not fully understood. Thus, a theoretical insight can potentially provide better insight to the underlying mechanism. Cycloxydim (CD) is a cyclohexene oxime herbicide widely applied against a variety of weed species. Phototransformation is one of the major environmental fates for oxime agrochemicals

ACS Paragon Plus Environment

3

The Journal of Physical Chemistry 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 29

and has attracted great research interest, the photochemical fate of CD has been examined experimentally20. More recently, combination of experimental21 and theoretical22 investigations on phototransformation of tetrazoline oxime ethers has also been reported. Chlorothalonil (CT) is a broad spectrum foliar fungicide which is commonly used worldwide. The chemical structures of CD and CT are shown in Figure1. CD is used for weed control and CT is used for the control of fungi on tomato, asparagus and pea crops, thus the presence of both molecules on crop leaf surfaces is very likely.

Figure 1 Chemical structures of CD(left) and CT(right) Monadjemi19 et al. reported accelerated photodegradation of CD in the presence of CT. Based on a series of photochemical degradation examinations and spectroscopic analysis of the products, they observed that under visible light CD is oxidized by triplet CT preferentially at the cyclohexane ring with the assumed hydrogen transfer process: 3

CT*+ CD→CTH· +CD-H

However, the proposed mechanism for the oxidation could not be identified with the Mass Spectrometry (MS). In addition, similar reaction pathways have been reported for agrochemical residues in surface water where herbicide molecules were photo-oxidized by triplet dissolvable organic matter (3DOM*). Canonica et al.23 examined the aqueous oxidation of phenylurea herbicides by triplet aromatic ketones and proposed electron-transfer followed by proton-transfer

ACS Paragon Plus Environment

4

Page 5 of 29 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

mechanism. Photo-oxidation of drug molecules by triplet ketones were studied and reported both experimentally and theoretically. Shah et al.24 and Li25 performed a theoretical study on the indirect photodegradation of sulphonamide antibiotics (SA) by 3DOM*, concluding that there was electron transfer from SA to DOM prior to proton migration. Jornet et al.26 studied the photoreactivity of benzimidazole moiety against selected aromatic ketones with combined laser flash photolysis and density functional theory methodologies and proposed electron-transfer followed by proton-transfer mechanism. Lucas et al.27 examined the phenolic H abstraction by triplet cyclic diketones with density functional theory calculation followed by Natural Population Analysis (NPA) and atom in molecules (AIM) analysis. They suggested that for mono-keto species the reaction pathway involves a coupled electron-proton transfer, while for di-keto species hydrogen migration was triggered by electron transfer. A major conclusion from this study was that electron transfer from the target molecule to the DOM molecule would facilitate the subsequent proton transfer process. Despite the experimental and theoretical study on the intermolecular photo reactions for pharmaceuticals, theoretical models on similar reaction pathways for agrochemicals has not been developed. Computational studies on the intermolecular oxidation of agrochemicals can provide valuable information on these processes and a better understanding of experimental findings. Herein, we performed computational studies on the oxidation of CD by CT- through the triplet hydrogen transfer process with density functional theory (DFT). Excited state properties of both CD and CT molecules as well as the CD/CT dimers were examined with time dependent density functional theory (TDDFT). Stationary points, including transition states and intermediate geometries along the possible reaction pathways were explored and compared in respect of energy barriers to determine the preferential reaction pathway. To understand the underlying

ACS Paragon Plus Environment

5

The Journal of Physical Chemistry 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 29

reason for the differences in the energy profiles for the possible pathways, natural population and surface charge analysis were performed, with the intermolecular interactions characterized. Computational studies on the effect of mixing herbicides and fungicides upon their photochemical fates have not been performed previously. The aim of this study is to provide computational examination of the CD/CT system to determine the photoreaction mechanism for intermolecular oxidation, which can be used to identify photochemical fate, predict novel photoproducts and assess the risks of efficiency loss before applying mixtures/adjuvants of herbicides.

Theoretical methods Geometry optimization and excited state property computation: Ground state minimum energy geometries of CD and CT were obtained by optimization using density functional theory28-29 (DFT) with the M06-2X hybrid functional of Truhlar and Zhao30 with the 6-311G(d, p)31-32 basis set. Energies and oscillation strengths of the first 20 singlet and triplet excited states for CD and CT were calculated using time dependent density functional theory (TDDFT)33 at M06-2X/6-311+G(d, p)34 level. The molecules were then modelled together as the initial geometry for the CD/CT dimer and optimized using the M06-2X functional and 631G(d) basis set35, two local minimum dimers for the ground state (S0) and first triplet state (T1) of CD/CT complex dimers were obtained. Vibrational frequencies for the dimers were calculated at the same level of theory, and the dimers were confirmed to be minima by the absence of any imaginary frequencies. For the excited energy transfer (EET) from CT to CD, the fragmentexcitation difference (FED) calculation36 was performed at M06-2X/6-311+G(d,p) level.

ACS Paragon Plus Environment

6

Page 7 of 29 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

Examination of reaction mechanisms: Potential Energy Surfaces (PES) of the hydrogen transfer process for both S0 and T1 states of the dimers corresponding to different reaction pathways were plotted using constrained geometry optimizations. The corresponding N-H bond lengths were fixed with a step size of 0.1 Å. The PESs were obtained at the M06-2X/6-31G(d) level. Transition states along the reaction coordinates for hydrogen migration were optimized at the same level. To obtain the transition state structure for reaction pathways where PES was ‘flat’ on the peak, force constants were computed at every point over the optimization with the ultrafine integration grid applied. The transition states were confirmed to be connected with the desired reactants and products with subsequent Intrinsic Reaction Coordinate37-38 (IRC) calculations. Energies of critical points on the potential energy surfaces were refined by M06-2X-D3/6-311+G(d, p) single point calculations, using the D3 version of Grimme’s dispersion with the original D3 damping function39 to further correct for possible intermolecular BSSE interactions. Natural Population Analysis (NPA)40 was also performed on the stationary points at the M06-2X/6-31G(d) level to examine the charge and electron population along the reaction pathways. Finally, the Atom in Molecules (AIM)41 method was applied to study the type of intermolecular interaction at selected stationary points. Results from NPA and AIM analysis were used to analyze whether the reaction mechanism was electron transfer followed by proton transfer or coupled electron-proton transfer. Solvation effects were not considered as the corresponding experiments were carried out without solvent19.To further understand the reaction mechanism for the different pathways, the surface area distribution of electrostatic potential (ESP) was calculated. In addition, the PESs for the first and second singlet excited states (S1 and S2) were examined by calculating the excitation

ACS Paragon Plus Environment

7

The Journal of Physical Chemistry 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 29

energies for the dimers along the reaction pathways using TDDFT with the M06-2X/6-311+G(d, p) level of theory. All geometry optimization and energy profile calculations were performed with the Gaussian 09 program42, the AIM calculation and surface area distribution of ESP were calculated using Multiwfn43, FED calculation for excitation energy transfer were performed with Q-Chem 4.344.

Results and discussion Geometries optimization and excitation energy transfer The geometries of two CD/CT dimers (denoted as dimer I and dimer II, respectively) were obtained by varying the initial orientations of CT with respect to CD followed by optimization with no constraints. Structures of the obtained dimers are illustrated in Figure 2a (dimer I) and 2b (dimer II). The single point energy of dimer II was 0.30 kcal/mol higher than that of dimer I, which indicates that co-existence of both dimers is possible. The Cartesian coordinates and energies of dimers I and II are given in the electronic supporting information (ESI). For both dimers, the nitrogen atom labelled N63 in CT was near the cyclohexane ring of CD, with nearby hydrogen atoms H27, H29 and H41 as is shown in Figure 2. The corresponding N-H bond lengths were 2.49 Å for N63-H27, 2.74 Å for N63-H29 and 3.08 Å for N63-H41, respectively. Intermolecular N-H hydrogen bonds were reported to be strengthened at excited states and subsequent electron transfer were facilitated45-47. Thus the intermolecular hydrogen bonds between N63 and H27, H29 and H41 could be strengthened at the excited states and intermolecular hydrogen transfer for H27, H29 and H41 on CD molecule to N63 on CT molecule are possible. It should be noted that the dimers identified here are not the only ones that could undergo intermolecular hydrogen transfer. Dimers of CT approaching CD from the opposite side

ACS Paragon Plus Environment

8

Page 9 of 29 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

compared to the ones shown in Figure 2 have also been examined, with corresponding potential energy profiles for intermolecular hydrogen transfer from CD to CT calculated. However, the steric effect from the thiane ring of CD significantly increased the energy barrier for the reactions, indicating that such reaction mechanisms are not preferred in this case, thus the corresponding structures and energies are not discussed in this study.

Figure 2 Local minimum energy geometries of the CD/CT dimers I (a) and II (b), obtained at M06-2X/6-31G(d) level. Major atoms involved in the hydrogen transfer process are labelled. Figures were generated by vmd48 Minimum geometries of the T1 state corresponding to dimer I and dimer II were also optimized at the M06-2X/6-31G(d) level of theory and confirmed with frequency calculation at the same level. The M06-2X-D3/6-311+G (d, p) refined single point energy of dimer II is 0.85 kcal/mol higher than that of dimer I. Figure 3a and 3b show the geometries for the T1 minimum of dimer I and dimer II, the NH bond lengths are 2.42 Å for N63-H27, 2.69 Å for N63-H29 and 3.00 Å for N63-H41, respectively. The spin density distribution among CD and CT molecules of the two

ACS Paragon Plus Environment

9

The Journal of Physical Chemistry 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 29

dimers are also shown in Figure 3, which can be regarded as the indication for localization of unpaired electrons. For both dimers, more than 90% of the Mulliken spin density is located on CD, which implies that the dimers were both in fact combinations of triplet CD and singlet CT. Full information of the Mulliken spin density results is available in the ESI. Such results disagree with the experimental conclusion that the photo-oxidation reactants were singlet CD and triplet CT. The reason for such disagreement could be that both dimers were formed via intermolecular excited energy transfer (EET) from triplet CT to singlet CD.

Figure 3 Minimum geometries of T1 state for dimer I (a) and dimer II (b), with the spin density iso-surface plotted in orange. Both dimers are confirmed to be CT+3CD To further examine the EET process, excited state properties for both dimers were calculated. Table 1 shows the T1 and T2 excitation energies and excitation assignments. For dimer I, excitation energy for T1 and T2 states were 3.23 eV and 3.36 eV, for dimer II the excitation energies were 3.19 eV and 3.45 eV, respectively. The character of the triplet excitation states can be analyzed using the excitation assignment by molecular orbital (MO) pairs. For both dimers,

ACS Paragon Plus Environment

10

Page 11 of 29 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

the T1 states were mainly associated with promotion of one electron from the HOMO-1 to the LUMO+2, as can be seen in Table 1, both orbitals were mainly localized on CD molecule, indicating that for both dimers the T1 state corresponded to local excitation of CD molecule and the spin density would be located on CD molecule. The T2 state of dimer I, on the other hand, was associated with HOMO-4 to LUMO and HOMO-3 to LUMO+1 promotions. MOs for the T2 excitation in dimer I were localized on CT molecule, suggesting a localized excitation in the CT molecule, thus the spin density would be located on CT. For T2 state in dimer II, the excitation contribute was almost evenly shared by 4 MO pairs, with 3 pairs of them localized on CT. For both dimers, the T1 state corresponded to the pi-pi* excited state for CD, and the T2 state corresponds to the pi-pi* excited state for CT. Thus, for T1 states, unpaired electrons would be located preferentially on CD molecule while for T2 states unpaired states would be located mainly on CT molecule. Analysis on the excitation assignments for the first two triplet excited states of the dimers explained the results in Figure 1 that the spin density for both dimers at T1 minimum structures mainly located on CD molecule. In addition, excited state energy and oscillator strengths for isolated CD and CT molecules were also calculated. The T1 excitation energies of CT and CD were 3.35 eV and 3.14 eV, respectively. The T1 and T2 excitation energies for both dimers were in qualitative agreement with that for the isolated CD and CT molecules, which is further indication for that for that the T1 and T2 excited states for both dimers can be mainly be attributed to the local excitation of CD and CT molecule, respectively.

ACS Paragon Plus Environment

11

The Journal of Physical Chemistry 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 29

Table 1 Calculated triplet excitation energy (∆E) and excitation assignment for T1 and T2 states of dimers I and II Dimer I

Excited State

∆E (eV)

Excitation Assignment a

T1

3.23

H-1 → L+2 (73%)

T2

3.36

H-4 → L (57%)

Occupied orbital

Unoccupied orbital

H-3 → L+1 (21%)

II

T1

3.19

H-1→ L+2 (69%)

T2

3.45

H-4 →L (29%)

H-3 → L (18%)

H-3 → L+1 (15%)

H-5 → L (14%)

a

H-1, H-4, H-3, L, L+1, etc. represent HOMO-1, HOMO-4, HOMO-3 LUMO, LUMO+1 orbitals, respectively. The percentage contribution of MO pairs to the excited states are given in parentheses.

ACS Paragon Plus Environment

12

Page 13 of 29 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

A complementary study on the intermolecular EET were also performed via Fragment-Excitation Difference (FED) calculations. The CT molecule was selected as the donor, while CD was selected as the acceptor. The fragment excitation for T1 and T2 for both dimers are shown in Table 2. It can be observed that for both dimers, the majority of the spin density for the T2 state is located on the donor (CT), while for the T1 state, most of the spin density is located on the acceptor (CD), which clearly indicates triplet excitation energy transfer (TEET) for both dimers. The coupling term for EET of dimers I and II were calculated to be 2.3 meV and -11.0 meV, respectively. It can be suggested based on the excited state assignment to molecular orbitals and the FED calculations that the minimum structures of the T1 states for both dimer I and dimer II were formed via excitation energy transfer from triplet CT to singlet (ground state) CD.

Table 2 Fragment-Excitation Differences for T1 and T2

Dimer I

II a

State

X(D)a

X(A)

T1

-0.021

-1.979

T2

-1.948

-0.052

T1

-0.082

-1.918

T2

-1.937

-0.063

X(D) and X(A) correspond to spin population on donor (CT) and acceptor (CD), respectively.

Intermolecular hydrogen migration pathways: The T1 minimum structures of dimers I and II were found to be generated via triplet excitation energy transfer from CT molecule to CD molecule and such minimum structures will be used as reactants for the triplet hydrogen transfer. From the S0 and T1 minimum structures of dimers I and II, we have identified three possible reaction sites in the cyclohexane ring of CD for

ACS Paragon Plus Environment

13

The Journal of Physical Chemistry 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 29

subsequent intermolecular hydrogen transfer. H27, H29 and H41 of CD molecule as shown in Figure2 are likely to transfer to the CT molecule. Intermolecular transfer of H27, H29 and H41 will be denoted as pathways 1, 2 and 3.

The NH bond lengths of the reactant dimers were 2.42 Å for N63-H27, 2.69 Å for N63H29 and 3.00 Å for N63-H41, respectively. Potential energy curves along the three identified reaction pathways for the S0, S1, S2 and T1 states were calculated and are provided in the ESI. As shown in Figure S1, the PES at S0, S1 and S2 states for all three pathways exhibited ascending trend and the intermolecular hydrogen transfer were not favourable at such states. Transition States (TSs) corresponding to hydrogen migration for pathways 1 and 2 were obtained at the M06-2X/6-31G(d) level. The TS for pathway 3 was obtained using additional evaluation of force constants at every step with the ultrafine integration grid due to the flatness of the potential energy surface (Figure S1).

Figure 4 Geometries for the transition state structures for pathway 1(4a), 2(4b) and 3(4c)

ACS Paragon Plus Environment

14

Page 15 of 29 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

The transition states were confirmed with subsequent IRC calculations which indicated that the TSs were connecting the desired reactants and products. Structures for the TSs of pathways 1,2 and 3 are shown in Figure 4a, 4b and 4c, respectively. Geometries for the three TSs and the corresponding imaginary frequencies as well as the IRC plots are provided in the ESI.

Products for hydrogen transfer were marked as radicals and geometries for the radicals formed along reaction pathways 1 2 and 3 were obtained by full optimization on the final points of the potential energy curves. Gibbs free energies of the stationary points along all three pathways were obtained by summing the M06-2X-D3/6-311+G(d,p) single point energies with the Gibbs free energy correction terms obtained at M06-2X/6-31G(d) level of theory. Reaction rate constants for hydrogen transfer in all pathways were estimated using transition state theory, and the quantum tunnelling effect was treated with Wigner’s correction6:  ||   =  +

    where υ is the imaginary frequency for the transition state, kB is the Boltzmann constant and T is the absolute temperature, where 298 K was used. The quantum tunnelling transmission coefficient for hydrogen transfer in pathways 1, 2, and 3 were 3.08, 4.53 and 1 respectively. The quantum tunnelling transmission coefficient for pathway 3 was set to be 1 since the absolute imaginary frequency for the TS was small and did not contribute much to the quantum tunnelling effect. The calculated reaction rate constants at 298 K for reaction pathways 1, 2 and 3 were 3.08e-3 s-1, 2.23e-10 s-1 and 3.51 s-1, respectively. The

ACS Paragon Plus Environment

15

The Journal of Physical Chemistry 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 16 of 29

relative free energies and corresponding N-H bond lengths are shown in Table 3. Pathway 3 has the lowest energy requirement for hydrogen transfer, among the products formed via hydrogen transfer, radicals upon hydrogen migration in pathway 3 have the lowest energy compared to the reactant dimer, which indicates that pathway 3 was the most probable reaction mechanism for triplet intermolecular hydrogen transfer. Pathway 2 is the least favourable due to the high energy requirement and low rate constant. Intermolecular hydrogen transfer through pathway 1 was also feasible because of the modest energy barrier and the quantum tunnelling effect. For pathways 1 and 2, the corresponding N-H bond lengths at the transition states were at 1.37 Å and 1.36 Å respectively, whereas for pathway 3, the N-H distance at the transition state was 1.60 Å, with subsequent hydrogen migration almost barriourless (see Figure S1). Such geometry and energy differences indicate that reaction mechanisms might be different for pathway 3 as compared to pathways 1 and 2. Table 3 Relative free energies and N-H distances for stationary points along different reaction pathways Pathways Dimer TS/IM E(kcal/mol) Radical Dimer N-H TS/IM bond(Å) Radical

1 0 22.9 6.3 2.42

2 0 31.5 11.2 2.69

3 -1.7 15.1 5.6 3.00

1.37 1.03

1.36 1.03

1.60 1.04

Proton transfer after electron transfer and coupled proton-electron transfer:

ACS Paragon Plus Environment

16

Page 17 of 29 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

To further identify the mechanism for the hydrogen transfer along the above three pathways, Natural Population Analysis(NPA) charges of all atoms for the stationary points in all three reaction pathways have been calculated and compared. Molecular charges for CD and CT were computed by summing the natural charges for each atom in CD and CT respectively. The electron distribution for both molecules are represented by the total atomic NPA population. Charge and electron population for CD molecule (or CD radical as the product for hydrogen transfer) at the stationary points along the three reaction pathways are given in Table 4. The molecular charge for CD molecule in the reactant dimers was close to zero for all three pathways, which indicates that for the reactants both CD and CT were in the neutral state. For the transition states complex the CD molecule carried a positive charge for all three pathways. For the transition state complex in pathway 3, CD molecule had a positive charge of 0.836, which suggested intermolecular charge transfer before hydrogen migration. For pathways 1 and 2, on the other hand, intermolecular charge transfer before hydrogen migration is less feasible since CD molecule only had positive charges of 0.587 and 0.286, respectively. The different extent of intermolecular charge transfer might influence the intermolecular attraction, thus influencing the energy requirement for hydrogen transfer in the different pathways. According to the electron population analysis, the product radicals for CD molecule has 175 e electron population (e representing unity charge density) compared to 176 e in the reactant dimers for all three pathways. Such a result suggests that an electron had transferred from CD to CT along with proton transfer prior to or coupled with the transferred proton. For pathways 1 and 2, at the transition state structures CD has 0.6 e and 0.3 e transferred to CT, respectively in agreement with the NPA charge results, which indicates coupled electron and proton transfer. The energy barrier in pathway 1 is lower than that in pathway 2, as a result of the larger extent of

ACS Paragon Plus Environment

17

The Journal of Physical Chemistry 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 18 of 29

electron transfer which contributed to the intermolecular attraction. For pathway 3, however, CD has almost one electron transferred to CT at the transition state complex, therefore, a barrierless migration of proton afterwards can be expected. Our conclusion on intermolecular electron transfer and proton migration is in agreement with the suggestion by Lucas et al.49 that for intermolecular hydrogen transfer from phenols to diketone molecules, the extent of NPA electron abstraction by diketones before proton migration had a positive relationship with their reactivity. Since the CD and CT molecules carried a certain amount of charge, accurate separation of the dimers into different fragments and calculation of interaction and complexation energy was not achievable. However, the interaction between the transferring hydrogen from CD and hydrogen acceptor of CT can be qualitatively characterized by the corresponding atomic charges. For the TS in pathway 1, the NPA atomic charge transferring hydrogen H27 was 0.37, with the atomic charge for N63 on CT molecule -0.48; for the TS in pathway 2, the atomic charge for H29 and N63 was 0.33 and -0.41; for the TS structure in pathway 3 the atomic charge for H41 was 0.53 and for N63 was -0.52. From the atomic charges, we can infer that the N-H intermolecular interaction was strongest in pathway 3, leading to the lowered energy requirement for hydrogen migration. Table 4 NPA charge and population for CD molecule/radical at the stationary points Pathways Dimer NPA charge TS/IM Radical Dimer Electron TS/IM population(e) Radical

1 -0.003 0.587 -0.009 176.0 175.4

2 -0.003 0.286 -0.004 176.0 175.7

3 0.000 0.836 0.008 176.0 175.2

175.1

175.1

175.0

ACS Paragon Plus Environment

18

Page 19 of 29 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

To further justify the intermolecular interaction and clarify electron transfer versus proton transfer an Atom in Molecules (AIM)41 analysis was performed on the stationary points. Properties of the bond critical points (BCPs) between the transferring hydrogen atom (H27, H29 and H41 for pathways 1, 2 and 3, respectively) and nitrogen atom on CT (N63) were calculated using Multiwfn39. The BCPs in all pathways are shown in the ESI (Figure S 2). Crucial properties for interaction of the corresponding N-H atoms including total electron density (ρ), Laplacian of electron density (∇2 ρ) and ratio of potential energy density |Vc| and kinetic-energy density Gc (|Vc|/Gc) for the BCPs are listed in Table 5. The N-H interaction type can be determined based on the nature of the BCPs41. For typical covalent (or “shared”) interactions, ρ is usually > 0.1 a.u., and ∇2 ρ < 0, with positive values for very polar bonds; for ionic (or “closed shell”) interactions, ρ is usually small (~10-2 a.u. for hydrogen bond interactions) with ∇2 ρ > 0; with intermediate values of ρ and ∇2 ρ ~ 0, the interaction is a mixture of the above two effects. It has been suggested more recently by Varadwaj and Marques50 that the ratio of potential energy density |Vc| and kinetic-energy density Gc (|Vc|/Gc) can characterize interaction types as well. Interactions with |Vc|/Gc > 2 are considered as covalent interaction, those with |Vc|/Gc < 1 are thought to be closed shell interactions (N-H hydrogen bond in our case), while 1