Subscriber access provided by READING UNIV
Letter
Molecular Dynamics of Photoinduced Reactions of Acrylic Acid: Products, Mechanisms and Comparison with Experiment Dorit Shemesh, and Robert Benny Gerber J. Phys. Chem. Lett., Just Accepted Manuscript • DOI: 10.1021/acs.jpclett.7b03015 • Publication Date (Web): 11 Jan 2018 Downloaded from http://pubs.acs.org on January 12, 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 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 Letters 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 20 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 Letters
Molecular Dynamics of Photoinduced Reactions of Acrylic Acid: Products, Mechanisms and Comparison with Experiment Dorit Shemesha and R. Benny Gerbera,b,* a
Institute of Chemistry and The Fritz Haber Research Center The Hebrew University, Jerusalem 91904, Israel b
Department of Chemistry, University of California Irvine, CA 92697, USA *Corresponding author:
[email protected]
ABSTRACT The photochemistry of acrylic acid is of considerable atmospheric importance. However, the mechanisms and the timescales of the reactions involved are unknown. In this work, the products, yields and reaction pathways of acrylic acid photochemistry are investigated theoretically by molecular dynamics simulations on the ππ* excited state. Two methods were used to describe the excited state: the semi-empirical OM2/MRCI and the ab initio ADC(2). Over one hundred trajectories were computed with each method. A rich variety of reaction channels including mechanisms, timescales and yields, are predicted for the single potential energy surface used. Main findings include: (1) Products predicted by the calculations are in good agreement with experiments. (2) ADC(2) seems to validate OM2/MRCI predictions on main aspects of mechanisms, but not on timescales. It is concluded that both semi-empirical and
1 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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 20
ab initio molecular dynamics simulations have useful advantages for the description of photochemical dynamics of carboxylic acids.
TOC GRAPHICS
Carboxylic acids are present in the atmosphere from industrial sources. Photochemistry initiated at the carbonyl chromophore is known to have important effects on the atmospheric chemical composition,1-7 causing in cases severe health problems in humans.8 Acrylic acid, the smallest α,β-unsaturated carboxylic acid, can decompose via several different pathways. One major pathway produces OH radicals, which are very reactive and likely contribute to the formation of ozone and photochemical smog. In order to assess the impact on the quality of the air, the kinetics and the mechanisms of acrylic acid degradation are of major interest. Four main channels have been proposed after excitation to the second singlet excited state (ππ* state): H2C=CHCOOH + hν
H2C=CHCO + OH
(1)
H2C=CH + HOCO
(2)
2 Environment ACS Paragon Plus
Page 3 of 20 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 Letters
H2C=CHOH + CO
(3)
H2C=CH2 + CO2
(4)
The first two reactions represent the cleavage of the C-O and C-C bonds, respectively. The third reaction represents decarbonylation and the fourth reaction decarboxylation. Each of the reactions has been observed in various experimental studies9-15, though product makeup is largely dependent on the experimental method used. Our study connects especially closely to the experimental work of Okumura et al.15 In his study a new experimental technique (time-resolved frequency comb spectroscopy) for identifying short-lived intermediates was described. Acrylic acid-d1 (H of OH group substituted by deuterium) was photolyzed. Intermediates detected include HOCO, but surprisingly also HOD, D2O and C2HD. In our simulations, we resolve the formation mechanisms of the HOCO, HOD and C2HD molecules. Previous theoretical studies assume that the reactions take place either in the S1 or in the T1 after population transfer to these states.16-17 Here we present a first dynamics study of these processes. In this paper, we assume that the excitation takes place to the S2 state, and that subsequent reactions occur on the same surface. Dynamics on the S2 surface has not been considered yet, so this study is the first to fill this gap. State switching is not considered in our approach. The justification of neglecting nonadiabatic transitions in our approach is obtained by comparing the experimental result to our simulations. A large number of simulated reactions are consistent with experiment, thus we believe that the assumption of a single state is validated, at least as a reasonable approximation. We certainly believe that a small percentage will internally convert to S1 (or even S0) and that reactions will take place on this surface. However, the short lifetimes seen here for the processes studied and the large energy gap between S2 and S1, imply, that S2 state reactions are of major importance and cannot be ignored. Additionally the strong correlation of the products with
3 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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
experiment supports the assumption that reactions occur on the S2 state. This does not exclude, that the same products might also be obtained on the S1 or S0 surface. Singlet to triplet state switching transitions in photochemistry for systems lacking heavy atoms are often neglected assuming inefficiency.7,
18-19
Whereas for certain types of systems, this is
compatible with available evidence, there are certainly examples where this assumption is invalid. One example is photolysis of HCl in Ar matrices, a system not involving heavy atoms.2021
Another example is in studies by Gonzalez et al on systems such as 2-thiouracil and
thymine.22-23 For internal conversion (S2 S1, non-adiabatic transition), the transition can certainly be very fast. The question here is whether the reactions on the S2 state happens on the same timescale as the non-adiabatic transition, i.e. whether the conical intersection is reached prior to the reaction on the S2 surface. For this system (acrylic acid) comparison of the products of the simulations vs. experiments supports our assumption, that indeed state-switching can be neglected. The objectives of this study are several. We explore the photochemistry of acrylic acid due to the importance of the process itself, but also to gain an understanding of how photochemistry in more complex systems can be approached. Most studies nowadays provide a very systematic and comprehensive picture by characterizing important structures such as transition states, conical intersections etc. on the relevant potential energy surfaces as has been done exceptionally in the work of Fang et. al on acrylic acid.16-17 From a methodological point of view, we will employ a different approach namely using Molecular Dynamics for providing mechanistic details, products and yields. Since there is a profound literature on this system, this enables us to check the Semi-Empirical Molecular Dynamics (SEMD) approach for its accuracy. We will compare SEMD both against experiment and Ab Initio Molecular Dynamics (AIMD). Due to the
4 Environment ACS Paragon Plus
Page 4 of 20
Page 5 of 20 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 Letters
computational expense, AIMD is feasible for acrylic acid, but not practical for much larger systems. Both approaches provide a complementary picture by contributing to different aspects of the photochemical reactions investigated. The advantages of using molecular dynamics simulations is that no prior knowledge of the characteristic points are necessary. Although the barrier heights are not known, this knowledge exists implicitly. Reaction pathways with lower barriers will result in higher yields. From a chemical point of view, we wish to obtain a better understanding of the products, mechanisms and timescales of these reactions. Molecular dynamics enable the calculation of yields of different reactions, provide an atomic level analysis of the reaction mechanisms and predict timescales of reactions. Two different potential energy surfaces have been employed: OM2/MRCI is a multiconfigurational semi-empirical method which is capable of describing excited states and other open shell systems.24 The method simplifies integrals obtained from ab initio methods and therefore results in a large speedup of the calculation. OM2/MRCI uses a multiconfigurational approach for treating these states, namely active orbitals and relevant configurations are chosen that describe best the system under review. The OM2/MRCI potential energy surface has been used to calculate vertical excitation energies and in conjunction with the molecular dynamics simulation. Here, the OM2/MRCI method and especially the OM2 method,25-28 have been previously used successfully by our group for a large variety of reactions.18,
29-33
In particular, the method was shown to be
qualitatively in good agreement with experiment in describing the photochemistry of carbonyls. Additionally, we have used the high-level ab initio method, ADC(2)34-35, for validation of excited state calculations in the Franck-Condon region, and in the dynamics simulations in time. Several works describing photodynamics of different systems using the ADC(2) potential have appeared in recent years.36-41 In some of the studies, ADC(2) based excited state lifetimes and yield of
5 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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
photoproducts were compared to results obtained using different other methodologies.36,
Page 6 of 20
38, 41
.
For example, non-adiabatic dynamics simulations of keto isocytosine shows that the lifetime predicted with ADC(2) is faster compared to SA-CASSCF.36 For this system, reaction channel yields strongly depend on the electronic structure methods used. In a recent study of the Surface Hopping Dynamics in 9H-Adenine38, it is shown that ADC(2) is very stable and that its performance is rated best among several other methods such as TD-DFT and CC2. Noteworthy is also the comparison between ADC(2) and OM2/MRCI. Compared to experiment, both methods have the closest agreement for the S0 population after 1 ps. However, the two methods differ in the yields from each deactivation pathway. Photodynamics of isocytosine was also calculated with the ADC(2) approach and tested against multireference quantum-chemical calculations (MR-CISD).41 Overall, the MR-CISD method confirmed the reliability of the ADC(2) method, though some inconsistencies in the prediction of the CI structure is found. The small number of studies so far using ADC(2) and OM2/MRCI prove the reliability of each method, but also indicate some of their limitations. It is therefore of major significance to get a clearer picture of the performances of ADC(2) and OM2/MRCI in describing photochemical reactions. This study provides a step towards this goal. The sampling of the initial conditions was carried out by running one trajectory on the OM2 ground state surface for 10 ps at a temperature of 300 K. From this trajectory 120 structures were sampled such that their excitation energy lies within ±0.5 eV of the vertical excited state energy. The velocity-Verlet algorithm with a time-step of 0.1 fs was employed throughout all the simulations. 120 trajectories were run also with an initial temperature of 300 K using OM2/MRCI method initiated on the second excited state with the same active space as described below. The maximum timescale was set to 100 ps for the OM2/MRCI and 34 ps for the ADC(2)
6 Environment ACS Paragon Plus
Page 7 of 20 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 Letters
simulations. ADC(2) simulation timescales were shorter than those of OM2/MRCI due to convergences problems. Increasing the iteration numbers would have been too costly. Figure 1 shows the structure of acrylic acid.
Figure 1: Structure of acrylic acid. Photoexcitation. Vertical excitation energies were calculated by OM2/MRCI. The ADC(2) method was used for calibrating the active space used in the OM2/MRCI method, see further details in the supporting information. OM2/MRCI vertical excitation energies are given in Table 1. State
Energy [eV]
Orbital excitation
Description nπ* ππ*
Oscillator strength 0.001 0.352
Dipole (Debye) 0.41 4.90
S1 S2
4.40 6.55
HOMO LUMO 77% HOMO -1 LUMO 90%
S3 S4
6.64 6.84
HOMO -3 LUMO 53% HOMO-2 LUMO 46 %
ππ*
0.0003 0.08
1.78 2.74
Table 1: Vertical excitation energies as calculated by OM2/MRCI with an active space of 10 electrons in 10 orbitals. For comparison, the vertical excitation energies calculated by ADC(2) with a cc-pVDZ basis set are given in Table 2.
7 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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
State S1 S2
Energy [eV] 4.89 7.21
S3
7.30
S4
8.29
Orbital excitation
Description
HOMO-1 LUMO 90 % HOMO LUMO 59 % HOMO-2 LUMO 32 % HOMO -2 LUMO 54 % HOMO LUMO 37 % HOMO-3 LUMO 52 % HOMO-1 LUMO+4 30 %
Page 8 of 20
nπ* ππ*
Oscillator strength 0.000011 0.45
Dipole (Debye) 4.03 1.22
OM2/MRCI state S1 S2
ππ*
0.13
3.59
S4
0.0012
0.89
S3
Table 2: Vertical excitation energy calculated by ADC(2)/cc-pVDZ on MP2/cc-pVDZ optimized geometry. Note that the ordering of the atomic orbitals by the OM2/MRCI method differs from the ADC(2) method. However, the state description can be easily compared by visualizing the orbitals. The comparison reveals, that the HOMO-3, HOMO-2 and LUMO orbitals are quite similar between both methods. However, the HOMO-1 and the HOMO orbitals are interchanged. The first excited state is predicted by both methods as an nπ* state. OM2/MRCI predicts this state at 4.40 eV with low oscillator strength making it a dark state. This is in excellent agreement with the experimental energy of 4.42 eV.42 The excitation energy predicted by ADC(2) is 4.89 eV, which is about 0.5 eV higher than the value calculated by OM2/MRCI. The second excited state for both methods is a ππ* state (of mixed π*(C=C)/π*(C=O) character, similar to experimental findings of Butler et al.43), which matches the excitation energy of 193 nm (6.4 eV) found in most experimental studies.15 The OM2/MRCI energy of 6.55 eV, therefore predicts this state very well, whereas the ADC(2) calculations predict it at 7.21 eV, much higher than the experimental value. Therefore, for calculating vertical excitation energy in this system, OM2/MRCI is superior to ADC(2). Additionally, higher lying states are predicted in a different orders using the two methods.
8 Environment ACS Paragon Plus
Page 9 of 20 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 Letters
However, for our simulations, these states are of less importance, as we mainly focus on the 2nd state. The ADC(2) method is used in this case to calibrate the active space. The active space was chosen such, that the OM2/MRCI method reproduces correctly the right ordering and energies for the first two excited states. Details about the calibration can be found in the supporting information. Without this calibration, the excited states with the OM2/MRCI method are predicted in the wrong order. It is therefore of great importance to use the ADC(2) method for assessing the correct excitations. Products and yields. Figure 2 shows a histogram of the products formed in the OM2/MRCI simulations.
Figure 2: Yields for different products obtained in the OM2/MRCI simulations. Note, that several molecules/fragments can be formed simultaneously. 9 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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
Only products above a yield of 6 % are shown. All these products, except for the H radical, have been either postulated by the above reaction schemes or detected experimentally in one of the studies mentioned above. In the following, we will describe the mechanisms leading to product formation. We will adhere to the proposed schemes of reactions, and will elaborate further when additional mechanisms have been found. Most importantly, we will provide mechanisms that account for the surprisingly C2HD and HOD products in the photolysis of deuterated acrylic acid (H of OH group substituted by deuterium)15.
Reaction pathways and mechanisms. In the following we will summarize the different pathways found in our simulations using OM2/MRCI. Pathway I and II are found in accordance what has been proposed in the introduction. Pathway III and IV were not found as proposed, but the products have been obtained by different mechanisms, as outlined below. Pathway 1: C-O fission (OH detachment). Direct OH detachment has been observed in 25 % of the trajectories. An additional 7 % of OH radicals are formed indirectly. The timescale for OH detachment is ultrafast (within 1000 fs). The potential energy surfaces for states S0, S1, S2 and S3 as calculated by OM2/MRCI are given in the supporting information. The S2 surface is mainly dissociative, with a small barrier of about 0.37 eV for the process. There is a substantial energy difference between S2 and S1. This supports our assumption that the S1 state is not reached by internal conversion in the timescale of the OH detachment (ultrafast, within 1000 fs). Pathway 2: C-C fission (HOCO detachment). OM2/MRCI simulations predict the C-C fission as a primary step in 5 % of the trajectories. In another 5 % of the trajectories the initial C-C fission is followed by a C-O fission resulting in the following fragments: CO + OH + C2H3.
10 Environment ACS Paragon Plus
Page 10 of 20
Page 11 of 20 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 Letters
Some of the trajectories also show H transfer between the two oxygens or H transfer between the carbon atoms prior to C-O fission. Note, that from the secondary reaction described here CO molecules are formed, which are a product in pathway III. However, the accompanying H2C=COH fragment from pathway III is found only as an end product in 1 % of the trajectories, meaning that the fragment is not stable and that it dissociates further. Pathway III is therefore not observed though it is postulated in the reaction schemes. CO2 formation.
Similarly, pathway IV has not been observed as postulated. CO2 is formed,
but from different mechanisms, without creating H2C=CH2 as a by-product. CO2 is created in six different minor pathways, summing up to a yield of 7 %. The main product within the CO2 formation is HCCH. Snapshots forming H2 are shown below in the pathway of HCCD formation. In the following, we will discuss major products formed in the simulation, that are not implied by the above postulated scheme, but are obtained directly in the experiment of Okumura.15 We have not found evidences for pathway III and pathway IV, although some of the products have been obtained by different mechanisms. H2O (HOD) formation: Water is created in about 19 % of the trajectories. In almost all trajectories the products are H2O, CO and HCCH. The dominant way to create water is depicted
in the following snapshots in Figure 3.
11 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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 3: Snapshot of simultaneous HCCH, CO and H2O formation Here it can be seen, that the molecule is rearranged such that all the products are produced simultaneously. Within 40 fs the products already reach a sufficient mutual separation. In all the trajectories in which water is created the OH is abstracted as one entity and therefore the water created will correspond to experimentally detected HOD. This means, that all the isomers of HCCH (H2CC, HCCH and the cleaved species H + CCH) are formed from hydrogens, and from this pathway no C2HD (as found in Okumura’s experiment15) is formed. CO formation: CO is created from two main pathways. The first is the formation of CO within the same pathway of water creation, as described above. The second pathway is via C-C fission (pathway III), followed by C-O fission. Minor pathways involve different steps with C2-C3 cleavage to yield the CO molecule. C2HD formation: Let us first note that C2H2 is created as a product in the water formation pathway as discussed above. However, in the above pathway, the deuterium is always found in the water molecule and not in the isomer of HCCH. Here we will focus on pathways which form C2HD, that is, where the hydrogen that is bonded to the oxygen is transferred to either C1 or C2 during the reaction. Fig. 4 shows snapshots from a trajectory in which C2HD is formed, but in this case the D atom is loosely bound, and towards the end of the simulation becomes detached from C2HD. This formation of H atoms is a prediction, which current experiments did not explore. Figure 4 shows snapshots of the mechanism leading to HCCD formation.
12 Environment ACS Paragon Plus
Page 12 of 20
Page 13 of 20 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 Letters
Figure 4: Snapshots of a trajectory showing the HCCD formation. SEMD vs AIMD dynamics: Figure 5a shows a histogram of the timescales of OH formation as computed by the OM2/MRCI method. One event at 12760 fs was excluded.
13 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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 5: (a) Histogram of timescales for the OH detachment pathway as calculated by OM2/MRCI. (b) Histogram of timescales for the OH detachment pathway as calculated by ADC(2). Figure 5b shows the histogram for the same reaction pathway as calculated by the ADC(2) method, which agrees qualitatively with the OM2/MRCI method. The bulk of OH formation events in OM2/MRCI is by a factor of ~2 faster than in the ADC(2) simulations. A rare OH formation event, both in OM2/MRCI and in ADC(2) is delayed, with a timescale of t > 2.25 ps. Simulations performed by the ADC(2) methods predict 42 % of OH detachment. In contrast with the OM2/MRCI method simulations, all these events are happening directly.
14 Environment ACS Paragon Plus
Page 14 of 20
Page 15 of 20 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 Letters
The common trajectory between both methods is the OH detachment discussed at length above. Two other minor trajectories have been seen within both methods. H transfer from the OH group to the carbonyl group occurs within both methods. The percentage for this pathway is 6 % within the ADC(2) simulations, and about 2 % for the OM2/MRCI simulations. About 8 % of the trajectories calculated by ADC(2) shows H detachment, compared to 28 % within OM2/MRCI. OM2/MRCI predicts many different reactions, which are not observed using ADC(2). One possible reason is that the simulation time with ADC(2) may not be enough in order to reveal the large multitude of reactions. Other reasons for differences between the two different methods may be as follows. The barriers are probably higher for ADC(2) than for OM2/MRCI. Therefore, those reactions are not seen within ADC(2). Probably OM2/MRCI is less accurate than ADC(2). However, one has to keep in mind, that OM2/MRCI predicts pathways which are observed experimentally, so this validates the method from the experimental side. On the other hand, there is no experimental validation for the timescales, so the OM2/MRCI results must be trusted less than ADC(2) for this property. Finally, validation of both methods is best achieved by comparison to the experiment. From the comparison, it is very obvious, that the semi-empirical method succeeds very well to predict the products observed experimentally. Likewise, the ADC(2) method is able to predict most of the pathways. In the few cases where the ADC(2) simulations do not predict an observed product, the likely reason is that the relevant trajectories were not propagated for sufficiently long time. Both potentials therefore describe correctly the main photodissociation channels. In summary, the dynamics simulations, both the semi-empirical OM2/MRCI and the ab initio ADC(2), proved successful in predicting the main experimentally obtained products. To an extent, though the experimental data on this is limited, theory accounts also for the relative yields
15 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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
of products formed. This lends support for the simple model used, in which the processes are assumed to take place on the singlet S2 potential surface, ignoring effects of possible nonadiabatic transitions at conical intersections, as well as of singlet to triplet non-adiabatic transitions. Our results do not, of course, exclude the possible occurrence of such non-adiabatic transitions. We merely establish that the single-state model can account for most if not all experimental data for this system. Future studies will explore computationally the role of nonadiabatic processes. However, it is highly desirable to aim also at new experiments that may throw light on the possible participation of non-adiabatic channels. The comparison with experiments seems to validate the semi-empirical MD simulations with regard to prediction of products and yields. On the other hand, comparison of OM2/MRCI-MD with the more rigorous ADC(2)-MD shows significantly differences, especially on the dynamics in time. Qualitative mechanisms predicted by SEMD are validated by AIMD, but the quantitative timescales are not, and firm conclusions on this will benefit from future ultrafast time-domain experiments. There seem excellent prospects of progress in photodynamics by combining increasingly powerful simulation methods with experiment. To conclude, the great success in predicting all the experimentally obtained products, supports the use of the semiempirical method for the study of photochemical processes. In particular, the semiempirical method opens the way for predicting products and yields of photochemistry of much larger systems which is a major breakthrough.
16 Environment ACS Paragon Plus
Page 16 of 20
Page 17 of 20 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 Letters
Supporting information Supporting information is available and includes details on the methodology and the potential energy surface along the OH detachment coordinate as calculated by OM2/MRCI. This material is available free of charge via the Internet at http://pubs.acs.org. Acknowledgements We would like to acknowledge Prof. Mitchio Okumura for helpful discussions on this project. We are grateful to Dr. Laura McCaslin and Dr. Natalia Karimova for constructive comments on the manuscript. RBG gratefully acknowledges support from the National Science Foundation through the Center for Aerosol Impacts on Chemistry of the Environment under grant no. CHE1305427.
References (1) Griffith, E. C.; Carpenter, B. K.; Shoemaker, R. K.; Vaida, V. Photochemistry of Aqueous Pyruvic Acid. Proc. Natl. Acad. Sci. 2013, 110, 11714-11719. (2) Shepler, B. C.; Epifanovsky, E.; Zhang, P.; Bowman, J. M.; Krylov, A. I.; Morokuma, K. Photodissociation Dynamics of Formaldehyde Initiated at the T1/S0 Minimum Energy Crossing Configurations. J. Phys. Chem. A 2008, 112, 13267-13270. (3) Nadasdi, R.; Zugner, G. L.; Farkas, M.; Dobe, S.; Maeda, S.; Morokuma, K. Photochemistry of Methyl Ethyl Ketone: Quantum Yields and S1/S0-Diradical Mechanism of Photodissociation. ChemPhysChem 2010, 11, 3883-3895. (4) Houk, K. N. Photochemistry and Spectroscopy of Beta, Gamma-Unsaturated CarbonylCompounds. Chem. Rev. 1976, 76, 1-74. (5) Naik, P. D.; Upadhyaya, H. P.; Kumar, A.; Sapre, A. V.; Mittal, J. P. Photodissociation of Carboxylic Acids: Dynamics of OH Formation. J. Photochem. Photobiol., C 2003, 3, 165-182. (6) George, C.; Ammann, M.; D'Anna, B.; Donaldson, D. J.; Nizkorodov, S. A. Heterogeneous Photochemistry in the Atmosphere. Chem. Rev. 2015, 115, 4218-4258. (7) Turro, N. J.; Ramamurthy, V.; Scaiano, J. C. Modern Molecular Photochemistry of Organic Molecules. University Science Books: Sausalito, CA, 2010.
17 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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
(8) Rao, Z.; Chen, Z.; Liang, H.; Huang, L.; Huang, D. Carbonyl Compounds over Urban Beijing: Concentrations on Haze and Non-Haze Days and Effects on Radical Chemistry. Atmos. Environ. 2016, 124, 207-216. (9) Rosenfeld, R. N.; Weiner, B. R. Photofragmentation of Acrylic-Acid and MethacrylicAcid in the Gas-Phase. J. Am. Chem. Soc. 1983, 105, 6233-6236. (10) Miyoshi, A.; Matsui, H.; Washida, N. Detection and Reactions of the HOCO Radical in Gas-Phase. J. Chem. Phys. 1994, 100, 3532-3539. (11) Kitchen, D. C.; Forde, N. R.; Butler, L. J. Photodissociation of Acrylic Acid at 193 nm. J. Phys. Chem. A 1997, 101, 6603-6610. (12) Osborne, M. C.; Li, Q.; Smith, I. W. M. Products of the Ultraviolet Photodissociation of Trifluoroacetic Acid and Acrylic Acid. Phys. Chem. Chem. Phys. 1999, 1, 1447-1454. (13) Upadhyaya, H. P.; Kumar, A.; Naik, P. D.; Sapre, A. V.; Mittal, J. P. Dynamics of OH Formation in the Dissociation of Acrylic Acid in its npi* and pipi* Transitions Excited at 248 and 193 nm. J. Chem. Phys. 2002, 117, 10097-10103. (14) Zhang, R. R.; Qin, C. C.; Long, J. Y.; Yang, M. H.; Zhang, B. Ultrafast Predissociation Dynamics of Excited State of Acrylic Acid. Acta Phys-Chim Sin 2012, 28, 522-527. (15) Fleisher, A. J.; Bjork, B. J.; Bui, T. Q.; Cossel, K. C.; Okumura, M.; Ye, J. Mid-Infrared Time-Resolved Frequency Comb Spectroscopy of Transient Free Radicals. J. Phys. Chem. Lett. 2014, 5, 2241-2246. (16) Fang, W. H.; Liu, R. Z. Photodissociation of Acrylic Acid in the Gas Phase: An Ab Initio Study. J. Am. Chem. Soc. 2000, 122, 10886-10894. (17) Fang, W. H. Ab Initio Studies toward Understanding Photoisomerization of Acrylic Acid in the Gas Phase. Chem. Phys. Lett. 2000, 325, 683-692. (18) Shemesh, D.; Lan, Z. G.; Gerber, R. B. Dynamics of Triplet-State Photochemistry of Pentanal: Mechanisms of Norrish I, Norrish II, and H Abstraction Reactions. J. Phys. Chem. A 2013, 117, 11711-11724. (19) McQuarrie, D. A.; Simon, J. D. Physical Chemistry. University Science Books: Sausalito, CA, 1997. (20) Krylov, A. I.; Gerber, R. B. Photodissociation Dynamics of HCl in Solid Ar: Cage Exit, Nonadiabatic Transitions, and Recombination. J. Chem. Phys. 1997, 106, 6574-6587. (21) Niv, M. Y.; Krylov, A. I.; Gerber, R. B.; Buck, U. Photodissociation of HCl Adsorbed on the Surface of an Ar12 Cluster: Nonadiabatic Molecular Dynamics Simulations. J. Chem. Phys. 1999, 110, 11047-11053. (22) Sanchez-Rodriguez, J. A.; Mohamadzade, A.; Mai, S.; Ashwood, B.; Pollum, M.; Marquetand, P.; Gonzalez, L.; Crespo-Hernandez, C. E.; Ullrich, S. 2-Thiouracil Intersystem Crossing Photodynamics Studied by Wavelength-Dependent Photoelectron and Transient Absorption Spectroscopies. Phys. Chem. Chem. Phys. 2017, 19, 19756-19766. (23) Mai, S.; Richter, M.; Marquetand, P.; Gonzalez, L. The DNA Nucleobase Thymine in Motion - Intersystem Crossing Simulated with Surface Hopping. Chem. Phys. 2017, 482, 9-15. (24) Koslowski, A.; Beck, M. E.; Thiel, W. Implementation of a General Multireference Configuration Interaction Procedure with Analytic Gradients in a Semiempirical Context Using the Graphical Unitary Group Approach. J. Comput. Chem. 2003, 24, 714-726. (25) Weber, W.; Thiel, W. Orthogonalization Corrections for Semiempirical Methods. Theor. Chem. Acc. 2000, 103, 495-506.
18 Environment ACS Paragon Plus
Page 18 of 20
Page 19 of 20 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 Letters
(26) Dral, P. O.; Wu, X.; Sporkel, L.; Koslowski, A.; Thiel, W. Semiempirical QuantumChemical Orthogonalization-Corrected Methods: Benchmarks for Ground-State Properties. J. Chem. Theory Comput. 2016, 12, 1097-1120. (27) Tuna, D.; Lu, Y.; Koslowski, A.; Thiel, W. Semiempirical Quantum-Chemical Orthogonalization-Corrected Methods: Benchmarks of Electronically Excited States. J. Chem. Theory Comput. 2016, 12, 4400-4422. (28) Dral, P. O.; Wu, X.; Sporkel, L.; Koslowski, A.; Weber, W.; Steiger, R.; Scholten, M.; Thiel, W. Semiempirical Quantum-Chemical Orthogonalization-Corrected Methods: Theory, Implementation, and Parameters. J. Chem. Theory Comput. 2016, 12, 1082-1096. (29) Epstein, S. A.; Shemesh, D.; Tran, V. T.; Nizkorodov, S. A.; Gerber, R. B. Absorption Spectra and Photolysis of Methyl Peroxide in Liquid and Frozen Water. J. Phys. Chem. A 2012, 116, 6068-6077. (30) Gerber, R. B.; Shemesh, D.; Varner, M. E.; Kalinowski, J.; Hirshberg, B. Ab Initio and Semi-Empirical Molecular Dynamics Simulations of Chemical Reactions in Isolated Molecules and in Clusters. Phys. Chem. Chem. Phys. 2014, 16, 9760-9775. (31) Lignell, H.; Epstein, S. A.; Marvin, M. R.; Shemesh, D.; Gerber, R. B.; Nizkorodov, S. Experimental and Theoretical Study of Aqueous Cis-Pinonic Acid Photolysis. J. Phys. Chem. A 2013, 117, 12930-12945. (32) Shemesh, D.; Gerber, R. B. Femtosecond Timescale Deactivation of Electronically Excited Peroxides at Ice Surfaces. Mol. Phys. 2012, 110, 605-617. (33) Shemesh, D.; Blair, S. L.; Nizkorodov, S. A.; Gerber, R. B. Photochemistry of Aldehyde Clusters: Cross-Molecular Versus Unimolecular Reaction Dynamics. Phys. Chem. Chem. Phys. 2014, 16, 23861-23868. (34) Schirmer, J. Beyond the Random-Phase Approximation - a New Approximation Scheme for the Polarization Propagator. Phys. Rev. A 1982, 26, 2395-2416. (35) Dreuw, A.; Wormit, M. The Algebraic Diagrammatic Construction Scheme for the Polarization Propagator for the Calculation of Excited States. Wires Comput. Mol. Sci. 2015, 5, 82-95. (36) Hu, D. P.; Liu, Y. F.; Sobolewski, A. L.; Lan, Z. G. Nonadiabatic Dynamics Simulation of Keto Isocytosine: A Comparison of Dynamical Performance of Different Electronic-Structure Methods. Phys. Chem. Chem. Phys. 2017, 19, 19168-19177. (37) Kungwan, N.; Kerdpol, K.; Daengngern, R.; Hannongbua, S.; Barbatti, M. Effects of the Second Hydration Shell on Excited-State Multiple Proton Transfer: Dynamics Simulations of 7Azaindole:(H2O)(1-5) Clusters in the Gas Phase. Theor. Chem. Acc. 2014, 133, 1480. (38) Plasser, F.; Crespo-Otero, R.; Pederzoli, M.; Pittner, J.; Lischka, H.; Barbatti, M. Surface Hopping Dynamics with Correlated Single-Reference Methods: 9H-Adenine as a Case Study. J. Chem. Theory Comput. 2014, 10, 1395-1405. (39) Stojanovic, L.; Bai, S.; Nagesh, J.; Izmaylov, A. F.; Crespo-Otero, R.; Lischka, H.; Barbatti, M. New Insights into the State Trapping of UV-Excited Thymine. Molecules 2016, 21, 1603. (40) Szabla, R.; Gora, R. W.; Janicki, M.; Sponer, J. Photorelaxation of Imidazole and Adenine Via Electron-Driven Proton Transfer Along H2O Wires. Faraday Discuss. 2016, 195, 237-251. (41) Szabla, R.; Gora, R. W.; Sponer, J. Ultrafast Excited-State Dynamics of Isocytosine. Phys. Chem. Chem. Phys. 2016, 18, 20208-20218.
19 Environment ACS Paragon Plus
The Journal of Physical Chemistry Letters 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
(42) Morita, H.; Fuke, K.; Nagakura, S. Electronic-Structure and Spectra of Acrylic-Acid in Vapor and Condensed Phases. B. Chem. Soc. Jpn 1976, 49, 922-928. (43) Arendt, M. F.; Browning, P. W.; Butler, L. J. Emission-Spectroscopy of the Predissociative Excited-State Dynamics of Acrolein, Acrylic-Acid, and Acryloyl Chloride at 199 nm. J. Chem. Phys. 1995, 103, 5877-5885.
20 Environment ACS Paragon Plus
Page 20 of 20
