Theoretical Study on the Mechanisms, Kinetics, and Toxicity Evaluation of OH-Initiated Atmospheric Oxidation Reactions of Coniferyl Alcohol

Abstract

In this paper, we investigated the mechanisms, kinetics, and toxicity evaluation of the OH-initiated reaction of coniferyl alcohol (4-(3-hydroxy-1-propenyl)-2-methoxyphenol) in the atmosphere using theoretical calculations. The initial reaction of coniferyl alcohol with OH radicals had two pathways, H-abstraction and OH-addition reactions. The total reaction rate constants were 2.32 × 10⁻⁹ cm³ molecule⁻¹ s⁻¹ (in gas-phase) and 9.44 × 10⁹ s⁻¹ M⁻¹ (in liquid-phase) for the preliminary reactions of coniferyl alcohol with OH radicals at 298 K, respectively, and the half-lives of the total reaction (including all initial H-abstraction and OH-addition reactions) of coniferyl alcohol with OH radical in the atmosphere, urban and remote clouds were 8.3 × 10⁻² h, 5.83 × 10³ h and 9.27 × 10² h, respectively. The temperature had a strong and positive influence on the initial reaction rate constant. The branching ratios of H-abstraction and OH-addition reactions were 3.68% and 97.69%, respectively, making the OH-addition reactions become dominant reactions. The ecotoxicity evaluation revealed that the toxicity levels of coniferyl alcohol and its products were similar and non-toxic. However, all these products have developmental toxicity, with most of them having no mutagenicity. Therefore, further attention should be paid to the oxidation process and product toxicity evaluation of coniferyl alcohol in the atmosphere.

1. Introduction

In recent years, the share of fossil fuels in total energy consumption is gradually decreasing due to electricity shortage, the greenhouse effect, and the continuous deterioration of the atmosphere and water environment [1], and the global progress toward an energy transition is rapid [2]. Biomass has become the fourth largest energy source in the world [3], accounting for a significant portion of total energy consumption [4]. Biomass combustion is the most common and widely used thermochemical method of biomass conversion [3] and will remain one of the main ways of biomass energy utilization [5] for quite some time to come. Additionally, the pollutants produced by combustion are of global importance as they affect regional and global air quality [6,7], climate [8], and human health [9]. Exposure to wood-burning smoke particles can lead to elevated morbidity and mortality from lung and cardiovascular disease [9]. Based on these effects, we recognize the necessity of increasing research on the sources of wood smoke particles.

Biomass consists mainly of three complex organic polymers, including lignin, hemicellulose, and cellulose [5], of which lignin has been detected at 18–35% [10]. Lignin is a complex compound composed of structural units of phenylpropyl, hydroxyl, and methoxy substituents, so its pyrolysis produces a large number of phenolic compounds—methoxyphenols (MPs) [11]. Methoxyphenols are widespread pollutants in the atmosphere, while the discharge of pulp mill (paper mill) effluent leads to the contamination of aquatic ecosystems by MPs [12]. MPs are well-known to be potential tracers of wood smoke emissions [13] due to the fact that they are the main component of wood smoke fine particulate matter, accounting for about 45% of the total fine particulate matter mass.

MPs can react with some important oxidants in the atmosphere, such as OH radicals, NO₃ radicals, Cl radicals, and ozone. These oxidation processes and oxidation products may result in either direct or indirect impacts on the atmosphere, aquatic ecosystems, and human health. Hence, for the assessment of the toxicity of MPs and their unfavorable impacts on the atmospheric environment, it is also necessary to investigate the photochemical behavior of MPs in the atmosphere. More recently, academia has paid much attention to the reactions of MPs in the atmosphere, and homogeneous and non-homogeneous reactions between several MPs and various gas-phase oxidants have been studied, focusing on degradation mechanisms and kinetics [14,15,16,17,18,19].

Coniferyl alcohol (CA), as a methoxyphenol, is one of the four alcohol monomers that form lignin and is widely used in the study of the synthesis mechanism and chemical structure of some natural plant products with high R&D and market value [20]. The emission rate of CA is 3.6–18 μg g⁻¹ for burning wood, and its concentration in the atmosphere is 0.3–0.5 ng m⁻³ [21]. In a previous study, a kinetic study of the degradation reaction of coniferyl alcohol was performed over a range of OH radical concentrations, temperatures, and relative humidities to acquire kinetic data to perform deeper calculations (atmospheric lifetime) [22]. Several other studies have discussed the homogeneous and heterogeneous reactions of CA that adsorbed on silica particles with NO₃ radicals [23] and ozone [24], respectively. The OH radical is produced primarily by the photolysis of ozone. Some of its subsequent reactions are cyclic processes involving CO, CH₄, and NO_x, where OH radicals are regenerated [25]. Additionally, the global average atmospheric concentration of OH radicals in the troposphere is approximately 10⁶ molecules cm⁻³ [26], which are very reactive oxidants and play a predominant role in tropospheric photochemistry. Until now, though, there are insufficient data on the reaction mechanism and product properties of CA initiated with OH, and further studies are still urgently needed to assess the photochemical degradation mechanisms as well as the toxicity evolution of these chemicals in the atmospheric environment. Furthermore, the photochemical degradation of these chemicals and their toxic effects on the aquatic environment have not been investigated. Moreover, the liquid phase environments, such as clouds and fog in the atmosphere (the atmospheric liquid phase), need to be considered. Previous studies have reported that some methoxyphenols and their degradation products (guaiacol, ribavirin, etc.) can be toxic to aquatic organisms when deposited into the aqueous environment [14,27]. Thus, it is necessary to assess the ecotoxicity of CA and its products to the aquatic environment. In conclusion, the investigation of atmospheric oxidation reactions for OH-initiated CA is vital to refine the atmospheric behavior as well as the toxic evolution of CA.

2. Computational Methods

2.1. Electronic Structures Calculations

In this work, all electronic structure calculations were conducted using the Gaussian 16 software program. The hybrid M06-2X function [28] of DFT and the 6-31+G(d,p) basis set were chosen to optimize the geometrical structures of reactants (Rs), pre-reacted complexes (RCs), transition states (TSs), intermediates (IMs), and products (Ps). According to the study of benchmarking at different levels (CCSD, CCSD(T), M06-2X, etc.), the M06-2X function is reliable with little error from other levels [29]. In addition, the M06-2X function has good prediction results for thermochemistry, reaction mechanisms, and kinetics of some chemical systems [30,31] and has been successfully applied in simulations for similar studies with excellent performance [32,33,34]. At the same level, we calculated the harmonic vibration frequencies, where the results were analyzed, and stationary points and minimum were determined. In order to gain precise energy parameters, the same M06-2X/6-311++G (3df,2p) level was used to calculate the single point energies (SPEs) for all optimized configurations. The reactants, intermediates, and products have positive vibrational frequencies, while the transition state has only one virtual frequency. Moreover, the performed intrinsic reaction coordinate (IRC) [35] calculations verify that all transition states are correctly connected with the corresponding reactants and products along the minimum energy path (MEP). We also have considered the factors of recrossing effect, torsional anharmonicity, and anharmonicity [36,37]. However, these factors were neglected due to their minor influence on the main study of this article, as well as for reasons based on computational costs.

There are three potential reaction mechanisms for the reaction of OH radicals with CA: H-atom abstraction (HAA) and radical addition (RAF). In the calculations, we considered the water environment for the simulation process and used a universal solvation model (SMD) [38] to assess the solvent effects in the water environment.

2.2. Rate Constant Calculations

All rate constants for the reaction of CA with OH radicals were calculated using transition state theory (TST) [39] with Wigner tunneling correction using the Kinetic and Statistical Thermodynamical Package (KiSThelP) [40], based on thermodynamic results for the reactants, intermediates, and transition states:

$$k^{TST}\left(T\right)=\sigma \frac{k_{b}T}{h}{\left(\frac{RT}{P^0}\right)}^{\Delta n}{e}^{-\frac{{\Delta G}^{0,‡}\left(T\right)}{-k_{b}T}}$$

where σ refers to the reaction path degeneracy, k_b represents the Boltzmann constant, h represents Planck’s constant, and T is the temperature. ΔG^0,‡ (T) is the standard Gibbs free energy of activation of the reaction at T temperature. Δn is 0 or 1 for unimolecular or bimolecular reactions, respectively. RT/P⁰ has the unit of the inverse of a concentration.

In this study, we adopted the Collins–Kimball theory [41] to take into account the effect of the diffusion-limited based on the TST calculations.

$$k=\frac{k_a{k}_d}{k_a+k_d}$$

where k_a is the thermal rate constant, and k_d is the steady-state rate constant of the non-reversible bimolecular diffusion-controlled reaction. Additional details on the calculation method and results are provided in the Supporting Information.

2.3. Toxicity Assessment

Both the computer-based Ecological Structure Activity Relationships (ECOSAR) modeling software [42] and the Toxicity Estimation Software Tool [43] have been employed to estimate the ecotoxicity and human health impacts of CA and its products, respectively, where ECOSAR has been found and adopted as a practical prediction tool. ECOSAR was used to assess the acute toxicity and chronic toxicity of these compounds for aquatic organisms at three different trophic levels in green algae, fish, and daphnia. More specifically, acute toxicity was illustrated by lethal concentration (LC₅₀) for fish and daphnia and maximum median effect concentration (EC₅₀) for green algae, while chronic toxicity was illustrated by the chronic toxicity value (ChV) for fish, daphnia, and green algae. The T.E.S.T. software was employed to predict its developmental toxicity and mutagenicity in health toxicology for human health effects. If the concentration of complex compounds has multiple effects, the lowest toxicity value should be elected as the most conservative estimate based on the precautionary principle of the Chemical White Paper [44].

3. Results and Discussion

3.1. Preliminary Reactions of OH Radicals with CA

In this study, the OH-initiated reactions in the atmospheric liquid phase are investigated by selecting the lowest energy conformation from the cis and trans conformations of CA as the optimal structure for the subsequent reactions (Figure S1). CA can react with hydroxyl radicals through two possible pathways: H-abstraction (OH radicals abstract H atoms from benzene rings, hydroxyl, methoxy, and carbon–carbon double bonds on branched chains) and OH-addition reactions (OH radicals are added to unsaturated bonds of substituents and to six species of C atoms on benzene rings). Therefore, nine H-abstraction pathways and eight OH radical addition pathways are considered (Figure 1). To make comparisons of thermodynamic data in the atmospheric liquid phase, the Gibbs free energy change ΔG and the Gibbs free energy barrier (energy barrier) changes ΔG^‡ are consistently used to reflect the difficulty of all elementary reactions. We have calculated the energy barriers between all reactants and the respective transition states. In addition, the calculated rate constants include reactants, transition states, and pre-complexes.

3.1.1. H-Abstraction Mechanisms

As indicated in Figure 1, there are nine different pathways for the preliminary reaction of H-abstraction (R1–R9), including OH radicals from hydroxyl, methoxy, benzene ring, and carbon–carbon double bonds on branched chains. H atoms in different positions can be attacked by OH radicals to form the respective intermediates and water. More specifically, when the OH radical abstracted the H atom from the hydroxyl group of CA, IM1 was generated via transition state TS1 with the energy barrier of 7.54 kcal mol⁻¹ and the reaction energy of 14.77 kcal mol⁻¹, and IM7 was generated through transition state TS7 with the energy barrier of 4.59 kcal mol⁻¹ and high exothermic exergy of 38.10 kcal mol⁻¹. For the abstraction of the H atom from the hydroxyl group of the CA by the methoxy, the product IM8 was generated through transition state TS8 with the energy barrier of 7.41 kcal mol⁻¹ and the reaction energy of −19.78 kcal mol⁻¹. In addition, OH radical abstracted the H atom from the benzene ring to produce the intermediates IM5, IM6, and IM9 via the transition states TS5, TS6, and TS9 with corresponding energy barriers of 11.31, 9.37, and 10.82 kcal mol⁻¹ and exothermic energy of 5.79, 5.12, and 3.55 kcal mol⁻¹, respectively. In the abstraction of the H atom from the olefinic group on the branched chain by the OH radical, three different intermediates IM2, IM3, and IM4, were generated via the transition states TS2, TS3, and TS4 with corresponding Gibbs free energy barriers of 6.11, 8.35, and 7.24 kcal mol⁻¹ and exothermic energy of 41.14, 8.93, and 14.00 kcal mol⁻¹, respectively.

For the initial reaction of hydrogen abstraction, the thermodynamically most favorable pathway is the one that forms IM7 with a high exothermic energy of 38.10 kcal mol⁻¹ and the lowest energy barrier of 4.59 kcal mol⁻¹. Therefore, the R7 pathway may be the main reaction pathway for hydrogen abstraction. Secondly, the second thermodynamically favorable pathway is the formation of IM2, with a Gibbs free energy change of 41.14 kcal mol⁻¹ and an energy barrier of 6.11 kcal mol⁻¹. Furthermore, the reaction energy barriers for the R1, R4, and R8 pathways of hydrogen abstraction are 7.54, 7.24, and 7.41 kcal mol⁻¹, respectively, which are only higher by 1.13–1.43 kcal mol⁻¹ than the R2 pathway. In contrast, the hydrogen abstraction pathways of R3, R5, R6, and R9 are relatively infeasible, with energy barriers that are 3.26–5.2 kcal mol⁻¹ higher than the R2 pathway.

The energy barriers of the second and seventh pathways are 6.26 and 9.29 kcal mol⁻¹, respectively, in the liquid phase, while in the atmospheric liquid phase, we reached a different conclusion that the most favorable pathway is the one that produces IM2. Comparing the obtained Gibbs free energy changes and energy barriers with the reactions in the gas phase, it can be found that the other pathways (R1–R7, R9) have higher energy barriers (0.15–25.68 kcal mol⁻¹) in the atmospheric liquid phase, except for the R8 pathway (0.38 kcal mol⁻¹ lower energy barrier); the exothermic energy of the R1 and R7 paths is lowered by 1.91 and 0.57 kcal mol⁻¹, respectively, and the other pathways (R2–R6, R8–R9) have higher exothermic energy (1.14–2.95 kcal mol⁻¹) in the atmospheric liquid phase.

3.1.2. OH-Addition Mechanisms

We also identify eight OH-addition reaction pathways given the fact that hydroxyl groups can undergo addition reactions with the benzene ring and carbon–carbon double bonds on the branched chains in CA (Figure 1). Specifically, the OH radical can attack the carbon–carbon double bond (C=C) to form the products IM10 and IM11 with reaction energy barriers of 2.92 and 5.92 kcal mol⁻¹. In addition, different products (IM12–IM17) are produced by adding OH to the six carbon sites (C4–C8 and C10) of the benzene ring. The energy barriers associated with all the aforementioned reactions are quite low, ranging from 2.92 to 10.87 kcal mol⁻¹. Additionally, each reaction is exothermic and has a heat of reaction within the range of 6.00–31.95 kcal mol⁻¹. Taken together, these findings strongly suggest that these reactions can readily occur under atmospheric conditions. For OH-addition reactions, the addition of OH radicals to produce IM10 at the C1 site has the lowest energy barrier, which is the most favorable pathway in the addition process.

A similar conclusion can be reached under atmospheric liquid phase as gas phase conditions, i.e., the thermodynamically most favorable pathway for OH-addition reactions remains the tenth pathway. Notably, the reaction energy barriers in the atmospheric liquid phase are generally lower than those observed in the gas-phase reactions, with reductions ranging from 0.26 to 4.95 kcal mol⁻¹ for all pathways except for the R14 and R15 pathways, which have higher energy barriers of 3.37 and 1.76 kcal mol⁻¹, respectively; Additionally, with the exception of R12 and R16 pathways (0.53 and 2.38 kcal mol⁻¹ higher exothermic energy, respectively), reaction heats for all other pathways are lower (0.65–1.89 kcal mol⁻¹). The atmospheric liquid phase enhanced the feasibility of most preliminary addition reactions while also modifying the reaction rate and the order of reactivity of the addition pathways. It is worth noting that the energy barrier for OH-addition reactions is decreased to a greater extent in the atmospheric liquid phase compared to that of H-abstraction reactions. To investigate subsequent transformations, all addition products were chosen as they can function as highly activated radicals that can undergo a series of reactions to produce more stable compounds.

3.1.3. Pre-Reactive Complexes and Kinetics

To quantitatively assess the contribution of each preliminary reaction path to the overall reaction, we calculated the reaction rate constants and branching ratios in the atmospheric liquid phase within the temperature range of 193 to 298 K at 1 atm. The reaction rate constants, k_abs (sum of rate constants for R1–R9), k_add (sum of rate constants for R10–R17), and k_total (the sum of k_abs and k_add) for each reaction path are shown in Table S2. The branching ratio (Γ) is a proportional relationship between the odds ratios of the different reaction pathways, which is calculated as the ratio of the individual pathway of the H-abstraction and OH-addition reactions to the total rate constant (Γ_add = k_add/k_total and Γ_abs = k_abs/k_total). After calculation, k_abs = 3.47 × 10⁸ s⁻¹M⁻¹, k_add = 9.09 × 10⁹ s⁻¹M⁻¹, k_total = 9.44 × 10⁹ s⁻¹M⁻¹ in the atmospheric liquid phase at 298 K and 1 atm. The rate constant of the H-abstraction reactions accounted for only 3.68% of the overall reaction rate constants at 298K, making a negligible contribution to the overall reaction. On the other hand, the rate constant of the addition reactions accounted for 97.69% of the total reactions and had a significant contribution. In the gas phase, k_abs = 3.02 × 10⁻¹⁰ cm³ molecule⁻¹ s⁻¹, k_add = 2.02 × 10⁻⁹ cm³ molecule⁻¹ s⁻¹, k_total = 2.32 × 10⁻⁹ cm³ molecule⁻¹ s⁻¹ at 298 K and 1 atm, which is different from the experimental results [22] due to the different conditions for the calculation of the rate constants. Therefore, the rate constant of the H-abstraction reactions accounts for 12.99% of the overall reaction rate constant and made a negligible contribution to the overall reaction. On the other hand, the rate constant of the addition reactions accounted for 87.01% of the overall reactions, which is significant. In the thermodynamic calculations, the energy barrier of the OH-addition reaction path is lower than that of the H-abstraction path. In particular, the R10 and R11 paths are addition reactions occurring on carbon–carbon double bonds with low energy barriers that are prone to occur, which is consistent with the calculations of the rate constants. Furthermore, the overall rate constant increases from 6.12 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ to 2.32 × 10⁻⁹ cm³ molecule⁻¹ s⁻¹ as the temperature increases from 193 K to 298 K. This indicates that the rate constant is proportional to temperature, which is consistent with a previous study [22]. Comparative analysis showed that the IM10 generation pathway is predominant, followed by R7, R2, and R11 paths, in good agreement with the thermodynamic data.

As shown in Figure 2, k_abs, k_add, Γ_abs, and Γ_add are temperature dependent. k_abs, k_add, and Γ_abs all increase with temperature, while Γ_add decreases. When the temperature rose from 193 K to 298 K, k_abs and k_add increased from 3.25 × 10⁻¹² cm³ molecule⁻¹ s⁻¹ and 5.79 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ to 3.02 × 10⁻¹⁰ cm³ molecule⁻¹ s⁻¹ and 2.02 × 10⁻⁹ cm³ molecule⁻¹ s⁻¹, respectively; Γ_abs increased slightly from 3 × 10⁻⁴ to 3.68 × 10⁻². Therefore, it was shown that the H-abstraction reactions are more significant at higher temperatures. Conversely, when the temperature increased from 193 to 298 K, Γ_add decreased from 99.97 to 96.32, processes that can readily highlight the importance of addition reactions at lower temperatures. The branching ratios of the H-abstraction pathway and OH-addition pathway at 298 K in the atmospheric liquid phase are 3.68% and 97.69%, respectively, indicating that H-abstraction reactions are dominant.

The atmospheric lifetime and half-life (τ_1/2) of the OH-initiated transformation of CA can be calculated from the total rate constant and the concentration of OH radicals [45,46]:

$$\tau =\frac{\ln 2}{k\left[\mathrm{O}\mathrm{H}\right]}$$

where k is the overall rate constant for the reactions of CA with OH radicals, and [OH] is the concentration of OH radicals.

The average global atmospheric concentration of OH radicals in the troposphere is approximately 10⁶ molecules cm⁻³ [26]. The atmospheric half-life of CA with OH radicals is calculated to be 8.3 × 10⁻² h at 298 K and 1 atm. The concentrations of OH radicals in urban and remote clouds are 3.5 × 10⁻¹⁵ and 2.2 × 10⁻¹⁴ mol L⁻¹, respectively [47]. At 298 K and in the atmospheric liquid phase, the half-lives of CA with OH radicals in the urban cloud and remote cloud are approximately 5.83 × 10³ h and 9.27 × 10² h, respectively. These findings suggest that the reaction rate of CA in the atmosphere via the OH-initiated photochemical conversion process is faster than that in the atmospheric liquid phase.

3.2. Subsequent Reactions of IM2, IM7, and IM8

3.2.1. Reaction with NO₂

IM7, an intermediate from the H-abstraction reactions, is selected for the subsequent reaction with NO₂. Since a substituent is already present at the C4 position, IM7 can add NO₂ at the C6 position, accompanied by the release of 5.14 kcal mol⁻¹ energy to generate IM19. We add H₂O as a bridge for the H-transfer to form the final product P2. The “water bridge” leads to an increase in the energy barrier and a decrease in the exothermic energy. The high exotherm of the process is 27.93 kcal mol⁻¹, and the energy barrier is 20.89 kcal mol⁻¹.

3.2.2. Reactions with O₂

As shown in Figure 3, the intermediates IM2, IM7, and IM8 produced by the dominant paths of the H-abstraction reactions are selected for subsequent reactions with O₂. IM2, one of the most favorable intermediates in the H-abstraction process, can be added to O₂ at the C1 position with a reaction energy barrier of 10.06 kcal mol⁻¹ to form the peroxide adduct IM18, with an exothermic reaction energy of 9.17 kcal mol⁻¹. Subsequently, IM18 undergoes NO addition and NO₂ elimination reactions without energy barriers, followed by the addition of O₂ at the C1 position and removal of HO₂ to achieve H-abstraction with the energy barrier of 11.52 kcal mol⁻¹ and high exothermic exergy of 51.08 kcal mol⁻¹, leading to the formation of the product P1. Similarly, IM7 and IM8 can also react with O₂ by adding O₂ at the C6 and C9 positions, respectively, leading to NO addition and NO₂ elimination reactions without energy barriers. The OH radicals then achieve H-abstraction, resulting in the formation of the products P3 and P4 with exothermic reaction energies of 32.11 kcal mol⁻¹ and 39.13 kcal mol⁻¹, respectively. In the presence of the atmospheric liquid phase, the exothermic energy rises in the range of 1.75–2.74 kcal mol⁻¹, and the energy barrier decreases in the range of 2.73–4.52 kcal mol⁻¹.

3.3. Subsequent Reactions of IM10–IM16

3.3.1. Reactions with NO₂

As shown in Figure 4, we select the intermediates IM13–IM16 resulting from the OH-addition reaction for the subsequent reaction with NO₂. These OH adducts (IM13–IM17) are activated radicals, allowing NO₂ to add to the neighboring and para-position of the incoming OH groups. Due to the steric hindrance effect, NO₂ can only add to one neighboring site of the incoming OH group on IM13 and IM15, while it can add to one neighboring and opposite site of the incoming OH group on IM14 and IM16. All these addition reactions are strongly exothermic, with exothermic reaction energies ranging from 21.91 to 29.21 kcal mol⁻¹ and no energy barrier, indicating that in the atmosphere, these reactions are simple, spontaneous processes that can readily form nitrogen dioxide-hydroxy-CA adducts (IM25–IM30). Consequently, these adducts can directly eliminate H₂O to generate nitro-CA (P2, P5, and P6) via high exothermic energies of 19.92–31.75 kcal mol⁻¹. Considering the induction effect of nitro and acetyl groups, the path NO₂ addition to IM10 was designed to produce IM10-2 with a high exothermic energy of 40.61 kcal mol⁻¹, and then IM10-2 elimination of H₂O to produce IM10-3 with a high energy barrier of 51.76 kcal mol⁻¹ and a reaction energy of 6.57 kcal mol⁻¹ (Figure S3), which is 10.68 kcal mol⁻¹ higher than the energy barrier of the similar path IM24 elimination of water. Thus, it was concluded that the elimination of water from allyl alcohol is more challenging than from IM25.

(H₂O)_n is added as a “bridge” for the elimination of water, and the energy barrier of water eliminated by the bimolecular reaction (H₂O)₂ is 6.66–16.07 kcal mol⁻¹ lower, and the exothermic energy of the reaction is elevated in the range of 0.92–7.16 kcal mol⁻¹ compared to the direct elimination of H₂O.

Moreover, the products P2, P5, and P6 can all be formed by the same OH-adducts through two different pathways. For example, P2 can be generated from IM13 and IM15 through a set of chemical reactions, while P5 and P6 can be produced by IM14 and IM16 through a series of reactions.

3.3.2. Reactions with O₂

As shown in Figure 5 and Figure S4, all of the OH adducts (IM10–IM17) were selected for reaction with O₂. For IM10 and IM11, O₂ enters the branched radical C2 and C3 sites to generate IM31 and IM33, with exothermic energies of 13.11 and 18.04 kcal mol⁻¹, and energy barriers of 8.97 and 6.90 kcal mol⁻¹, respectively. Subsequently, the addition of NO continued without an energy barrier and was followed by the elimination of NO₂ and the H-abstraction of O₂ to form products P7 and P8. In the presence of the atmospheric water phase, the energy barrier for the formation of IM31, IM33, and P8 decreased by 8.03–15.96 kcal mol⁻¹, while the energy barrier for the formation of P7 increased by 7.78 kcal mol⁻¹.

Other intermediates IM12–IM17 generated by OH-addition reactions can be added to the benzene ring with O₂, where O₂ only adds to the neighboring positions of the incoming OH groups on IM13 and IM15, generating the peroxide adducts IM37 and IM40. For IM12, IM14, IM16, and IM17, O₂ can add to both neighboring and para-positions of the incoming OH groups, leading to the formation of the peroxide adducts IM35, IM36, IM38, IM39, and IM41–IM44. Specifically, the pathways from IM12 to IM35, IM14 to IM38, and IM16 to IM42 are exothermic in the range of 0.69–6.02 kcal mol⁻¹, and the remaining pathways are endothermic in the range of 0.08–7.63 kcal mol⁻¹; and these reactions only need to overcome the higher energy barriers, ranging from 10.84 to 21.83 kcal mol⁻¹. Under the atmospheric aqueous phase conditions, the exothermic energy of the reactions increased by 0.96–5.61 kcal mol⁻¹ compared to the atmospheric environment, while the reaction energy barrier decreased by 2.25–5.39 kcal mol⁻¹. Therefore, these reactions are more likely to occur in the atmospheric liquid phase.

As shown in Figure 6, all of the peroxides (IM35–IM43) were selected for reaction with NO. Since these peroxyl radicals, IM35–IM43, are not yet final stable products, all can add NO and eliminate NO₂ without an energy barrier, releasing thermal energy of 7.95–30.78 kcal mol⁻¹ to produce nitrogen dioxide and eight different intermediates (IM45, IM47, IM49, IM51, IM53–IM56, and IM58). Then the possible subsequent pathways of IM45, IM47, IM49, and IM51 are ring cleavage reactions and release high thermal energy of 17.72–31.06 kcal mol⁻¹. C4–C5 bond cleavage via TS46 generates IM46 and an energy barrier of 2.02 kcal mol⁻¹, and C4–C10 bond cleavage via TS48 generates IM48 and an energy barrier of 0.77 kcal mol⁻¹. In addition, cleavage of the C5–C6 bond by two different pathways, TS50 and TS52, yielded IM50 and IM52 for energy barriers of 15.49 kcal mol⁻¹ and 153.56 kcal mol⁻¹, respectively. Afterward, the H-abstraction of O₂ or H-abstraction of OH radicals is carried out separately with intense exothermic energy of 65.57–74.40 kcal mol⁻¹, yielding dialdehyde (P9–P11) and HO₂ or H₂O and the energy barrier of 7.04–19.54 kcal mol⁻¹. Among them, the product P11 can be generated by two different pathways.

Since the structures cannot undergo ring cleavage, IM53 and IM56 directly undergo H-abstraction of O₂ and H-abstraction of OH radicals, respectively, with high exothermic exergies of 38.61 and 106.18 kcal mol⁻¹, respectively, and high energy barriers of 15.26 kcal mol⁻¹. After that, IM57 removes a CH₃OH to generate the final product, dialdehyde (P15). In addition, the structures of IM55, IM56, and IM58 readily undergo both H-abstraction and ring cleavage reactions, so they are combined into one step in the reaction mechanism. These pathways are highly exothermic, with high exothermic energies of 60.36, 33.62, and 14.57 kcal mol⁻¹, respectively, and are barrier-free reactions. In the atmospheric liquid phase, the exothermic energy of the majority of the reactions increases by 2.38–24.10 kcal mol⁻¹, and the energy barriers are lower than those in the atmosphere by 0.55–15.91 kcal mol⁻¹, with only the energy barrier of IM47 ring cleavage pathway elevated by 2.93 kcal mol⁻¹. Thus, most subsequent reactions are favored in the atmospheric liquid phase than in the gas phase, but there are also low energy barriers (3.70 kcal mol⁻¹) in which the formation of IM45 is still favorable.

3.4. Toxicity Assessment

In this study, ECOSAR and T.E.S.T. are chosen for the toxicity prediction of CA and its end products(P1–P15). For the ecotoxicity assessment, the acute and chronic toxicities are predicted for aquatic organisms at three different trophic levels (fish, daphnia, and green algae); for human health effects, developmental toxicity and mutagenicity are evaluated.

3.4.1. Acute and Chronic Toxicity of CA to Aquatic Organisms

The toxicity values of CA and its end products and the toxicity evaluation criteria are listed in Figure 7 and Table S5, respectively. In accordance with the toxicity classification criteria developed by the European Union, the predicted values for acute toxicity, such as the LC₅₀ for fish and daphnia and the EC₅₀ for green algae of CA, are 7.95 × 10², 4.19 × 10², and 2.29 × 10² mg L⁻¹, respectively, which are all categorized as not harmful (LC₅₀ > 100 and EC₅₀ > 100). As for the chronic toxicity of CA, the ChV values on fish, daphnia, and green algae are 71.1, 33.1, and 50.7 mg L⁻¹, respectively, which are also categorized as not harmful (ChV > 10). Since CA is not toxic to fish, daphnia, and green algae, it can be classified as a non-toxic substance that is probably not harmful to aquatic organisms. Even though CA is not toxic to fish, daphnia, and green algae, it may hasten the emergence and development of drug resistance. Thus, it is necessary to degrade CA, and its degradation products should be tested.

3.4.2. The Toxicity of the Products in the NO₂ and O₂ Addition Pathways

The NO₂ addition pathway products (P2, P5, and P6) have the same and slightly lower toxicity values than CA but at the same level of toxicity. In terms of acute toxicity, there is no significant toxicity to fish, daphnia, and green algae, as the LC₅₀ values for fish and daphnia and the EC₅₀ values for green algae are 438 and 240 mg L⁻¹ above 100 mg L⁻¹. In terms of chronic toxicity, none of the products are toxic to aquatic organisms, as the ChV values for green algae are above 10 mg L⁻¹.

In addition, the toxicity values of the O₂ addition pathway products are mostly much greater than those of the CA and NO₂ addition pathway products with the same toxicity levels, except for P4 and P14.

3.4.3. Developmental Toxicity and Mutagenicity

The predicted results for developmental toxicity and mutagenicity of CA and final products (P1–P15) are obtained using the T.E.S.T. software, as shown in the Figure S6 and Table S7. According to the consensus method for prediction of developmental toxicity and mutagenicity, the value of developmental toxicity for CA is 0.56, so the developmental toxicity result is developmental toxicant; the mutagenicity value is 0.14, so the mutagenicity result is negative (mutagenicity ≤ 0.50), i.e., no mutagenicity. All final products have developmental toxicity, with P2 exhibiting the most pronounced developmental toxicity (developmental toxicity value of 0.83). Most of the products remain mutagenic, except for P9 and P12. These findings highlight the potential impact of CA and its transformation products on human health, emphasizing the need for closer monitoring of their degradation products and detection of further degradation pathways.

4. Conclusions

This study investigated the mechanism of the OH-initiated reactions of CA in the atmosphere using theoretical calculations and evaluated its toxicity and final products. Thermodynamic and kinetic studies were conducted on the initial reactions of CA with hydroxyl radicals and the subsequent reactions of the primary products with O₂, nitrogen dioxide, and nitric oxide. The main conclusions are as follows:

(1)

All of the preliminary OH-addition reactions and H-abstraction reactions of CA have been studied, and the role of the atmospheric liquid phase was examined. Except for the R8 pathway, the energy barrier of initial H-abstraction reactions decreased due to the atmospheric liquid phase, whereas that for most of the initial OH-addition reactions decreased, except for R14 and R15. Furthermore, the reaction order of different pathways changed, with the long-branched chain being more susceptible to initial reactions than the benzene ring. The subsequent reactions of CA with O₂/NOx were studied using favorable pathways IM2, IM7, IM8, and IM10–IM16.

(2)

In the atmospheric liquid phase, the total reaction rate constants for the preliminary reactions of CA with OH radicals at 298 K were 2.32 × 10⁻⁹ cm³ molecule⁻¹ s⁻¹ and 9.44 × 10⁹ s⁻¹M⁻¹, with half-lives of 8.3 × 10⁻² h, 5.83 × 10³ h, and 9.27 × 10² h, respectively. The initial reaction rate constant was positively correlated with temperature. OH-addition reactions played a significant role in the initial reaction with OH radicals in the atmospheric liquid phase, with branching ratios of H-abstraction reactions and OH-addition reactions being 3.68% and 97.69%, respectively.

(3)

Ecotoxicity evaluation revealed that CA and its products had similar levels of toxicity, which were non-toxic. However, all these products had developmental toxicity, and most of them were also mutagenic. Therefore, the ecotoxicity testing of the CA and degradation products and the effects on human health should be closely monitored.