All gas-phase ab initio/density functional theory (DFT) calculations were performed using the Gaussian 16 suite of programs (Frisch et al., 2019). The stationary points of all molecular systems, including reactants, products, intermediates and transition states on the potential energy surface (PES) were optimized using the Minnesota 2006 exchange correlation functional such as M06-2X, in conjunction with the Pople’s split-valence 6-311++G (3df, 3pd) basis set (Frisch et al., 1984; Zhao and Truhlar, 2008). The long-range van der Waals interactions between the reactive species were accounted using Grimme’s empirical dispersion (GD3) corrections (Grimme et al., 2010). Previous studies have demonstrated that the current level of theory exhibits a reliable performance in addressing noncovalent interactions between gaseous molecules and in locating the transition states of atmospheric and combustion reactions (Ali et al., 2016; 2018; 2022; Ali, 2019; 2020; Dash and Ali, 2022; 2023; Ali and Balaganesh, 2023). Tight convergence criteria were applied during the wave function optimization of the reactive species, complexes, products and transition states. Unscaled vibrational frequencies at the same level of theory (M06-2X/6-311++G (3df, 3pd)) were utilized to compute zero-point energy (ZPE) corrections to the total energies of all molecular systems, to characterize the stationary points on the PES and for rate-constant calculations. Vibrational frequency analysis confirmed all positive frequencies for the reactants, complexes, intermediates and products, while a single imaginary frequency was observed for the transition states.

Additionally, single-point energy calculations were conducted at a higher-level of theory on the molecular structures optimized at a lower-level of theory to ensure an accurate description of the energetic parameters. Specifically, the CCSD(T)/6-311++G (3df, 3pd) level of theory was utilized to estimate the single-point energies of the gas-phase molecular geometries, which were initially optimized at the M06-2X/6-311++G (3df, 3pd) level of theory (Raghavachari et al., 1989). The basis set superposition error (BSSE) calculations were also performed using the counterpoise (CP) corrected method (Boys and Bernardi, 1970; Simon et al., 1996). The <Ŝ ²> eigenvalues were monitored to evaluate the spin contamination for the wavefunction of the open-shell radicals. The T1-diagnostic values obtained at the CCSD(T)/6-311++G (3df, 3pd) level of theory were analyzed to validate the single-reference method and were found to be within the acceptable range (i.e., ≤0.02) for all important species (Lee and Taylor, 1989). Overall, the combination of CCSD(T)//M06-2X functionals has been employed in numerous research studies, providing a reasonably accurate description of the thermochemistry and chemical kinetics of many atmospheric reactions (Ali et al., 2016; Ali et al., 2018; Ali, 2019; Ali, 2020; Ali et al., 2022; Dash and Ali, 2022; Ali and Balaganesh, 2023; Dash and Ali, 2023).

Comprehensive chemical kinetic calculations for the O ˙ H initiated oxidation reaction of aminomethanol (AM) using MultiWell suite of codes (Barker, 2001; Barker, 2009; Barker et al., 2023) were performed. This oxidation reaction involves a fast pre-equilibrium between the reactive species such as H₂NCH₂OH + O ˙ H and the pre-reactive complex H 2 NC H 2 OH ⋯ OH • stabilized by the van der Waals forces followed by a second step leading to the respective products as follows, H 2 NCH 2 OH + O ˙ H   k − 1 ⇆ k 1   H 2 NC H 2 OH ⋯ OH • (1) H 2 NC H 2 OH ⋯ OH •   k 2 →   H 2 N C ˙ HOH + H 2 O (2) where, k ₁ and k _-1 are the forward and reverse rate constant for the first bimolecular reaction and the k ₂ is the rate constant for the second unimolecular reaction. The kinetic rate constants for these bimolecular (k, in cm³ molecule⁻¹ s^-1) and unimolecular (k _uni, s^-1) reactions in the high-pressure limit defined by transition state theory are represented as follows, k 2 = Γ T × σ k B T h × Q T S Q I M × ⁡ exp   − ∆ E 0 R T (3)

Assuming that the pre-reactive complex was in equilibrium with the reactants and was at a steady state, then the overall rate constants is expressed as; k = k 1 k − 1 + k 2 k 2 (4)

If k_-1 >> k₂, the rate constant is rewritten as k = k 1 k − 1 k 2 (5) k = K e k 2 (6)

This kinetic model is reasonably correct at the high-pressure limit, where the pre-reactive complex can be stabilized by collisions with other atmospherics species. This approach was widely used in literature for the water-assisted reaction and the predicted rate coefficients are reasonably good agreement with the experimental values (Ali, 2019; Ali et al., 2022, Ali and Balaganesh 2023).

The different parameters of Eqs 3, 6 were breakdown and elaborately discussed the specifics of each component in Supplementary Material to prevent redundancy from earlier research. The k and k ₂ for the other plausible oxidation reactions of O ˙ H initiated oxidation reaction of AM such as for H₂NCH₂OH+ O ˙ H→H N ˙ CH₂OH + H₂O and H₂NCH₂OH+ O ˙ H→H₂NCH₂ O ˙ H + H₂O were also computed.

The temperature- and pressure-dependent microscopic rate constants k(E) have also been computed for the O₂ addition reaction to the aminomethanol radicals generated in Eq. 2. This was accomplished using the Rice−Ramsperger−Kassel−Marcus (RRKM)/master equation (ME) theory, implemented in the MultiWell suite of programs. The MultiWell code facilitates the computation of non-steady-state effects including unimolecular decomposition processes, isomerization, collision energy transfer and chemical activation for the complex rate-constant calculations. To perform these calculations, molecular and energetic parameters such as vibrational frequencies, moments of inertia and reaction barriers are required as input data. Using this data, the MultiWell suite computes sum and density-of-states, followed by the evaluation of microscopic rate-constant k(E). The RRKM/ME microscopic rate-constant k(E) is defined as follows, k E = m ≠ m σ e x t σ e x t ≠ g e ≠ g e 1 h G ≠ E − E 0 ρ E . (7)

The details of each term can be found in the Supplementary Material. Temperature and pressure-dependent rate constants and branching ratios of the products were evaluated by incorporating N₂ gas as the bath gas. The collisional energy transfer process was addressed using the conventional temperature-dependent exponential-down model with a <ΔE > _down parameter (which represents the average energy loss per the collision of the active compound with the bath gas molecule), with an approximate value of ∼200*(T/300)^0.85 cm⁻¹ (Goldsmith et al., 2012). Lennard-Jones (L-J) parameters were employed to account for the frequency of collisions between the active compound and the bath gas (N₂) collider. The L-J parameters for N₂ gas, specifically σ (N₂) = 3.74 Å and ε/k_B (N₂) = 82 K, were sourced from the literature, while the same parameters for all wells were adopted from our previous study (Dash and Ali, 2022).

For the barrierless reactions i.e., AM ˙ +O₂→AM-O O ˙ and product complexes to individual product molecules Inverse Laplace Transform (ILT) method was incorporated to determine the rate-coefficients (Robertson et al., 1995). Consistent with studies on numerous analogous reactions, this method has proven effective, with the Arrhenius’s activation energy equating to the critical energy of the reaction (E ₀) (Firaha et al., 2018). Additionally, statistical rate theories that neglect non-statistical effects were applied, including slow intramolecular vibrational energy redistribution (IVR) as discussed in previous study (Mazarei and Barker, 2022).

The pressure-dependent total rate constants k ^bimol (T, M) for aminomethanol radical ( AM ˙ ) + O₂ have been computed using, k b i m o l T , M = Γ   K e q × k ∞ u n i 1 − f AM ˙ + O 2 (8) where, Γ represents the quantum mechanical tunneling corrections, f AM ˙ + O 2 is the branching fraction ( f ) of the chemical reaction returning to the respective reactive species and k ∞ u n i is the rate constant at the high-pressure limit. The tunneling was incorporated for the chemical activation distribution in all our chemical kinetic calculations using the keywords “CHEMACT” and “TUN”.

Finally, the calculated rate constants were fitted at the high-pressure limit ( k ∞ ) in the temperature range of 200–400 K to the modified Arrhenius expression, which is as follows, k ∞ T = A × T n × ⁡ exp − E a R T (9) where A is the pre-exponential factor, T is the temperature, n is the temperature exponent and E _a is the activation energy. The coordinates of equilibrium geometries, vibrational harmonic frequencies and rational constants of all important species involved in the O ˙ H initiated oxidation reaction of AM are listed in the Supplementary Material.

The oxidation reaction between the aminomethanol (AM) and O ˙ H radical occur through the abstraction of H-atom from three different H-bearing functional groups (i.e., –CH₂, –NH₂ and –OH) of AM. Generally, the H atom of –CH₂, –NH₂ and –OH groups along with the presence of lone pair of electrons on the N and O atoms facilitate the hydrogen bonding (H-bonding) between the AM and the O ˙ H radical. Interestingly, the –NH₂ and –OH groups of AM can freely rotate around the single bonds to form the inter molecular H-bonding with the O ˙ H radical based on its attacking direction. It leads to different rotational conformations for the AM- O ˙ H radical complex. In addition, formation of such intermolecular H-bonding in the pre-reactive and the transitions state (TS) structures has a great effect on the energetics of the reaction. Hence, we have comprehensively explored the different reaction pathways and the corresponding energy barriers using different rotational conformations of AM. It is evident from previous studies that the AM exhibits four different rotational conformations, which we named as AM0, AM1, AM2 and AM3 as presented in Figure 2 (Hays et al., 2013). It shows that the AM0 is energetically more stable and can be consider as ground state conformation of AM. The geometrical features of ground state conformation of AM are good in agreement with that of previous study (Hays et al., 2013). Other rotational conformations such as AM1, AM2 and AM3 exhibit ∼0.2, 0.8 and 4 kcal/mol higher energy compared to the AM0, respectively. The trend in the conformational stabilities of AM is in accordance with the study of Weaver and co-workers (Hays et al., 2013). Hence, the AM0 is used as a reference to provide the further intricate details about the mechanistic pathways of the AM oxidation reaction with the O ˙ H radical.

The zero-point energy (ZPE) corrected potential energy surface (PES) for the H-abstraction reaction of the most stable conformation of aminomethanol (AM0) by the O ˙ H radical is depicted in Figure 3. In this reaction, the O ˙ H radical abstracts the H-atom from three different H-bearing moieties ( i . e ., –CH₂, –NH₂ and –OH) of AM0, leading to the formation of three distinct AM0 radicals, namely, the carbon-centered NH₂ C ˙ HOH radical, nitrogen-centered N ˙ HCH₂OH radical and oxygen-centered NH₂CH₂ O ˙ radical. These radicals are obtained through an energetically favorable pre-reactive complex (PRC), followed by the transition states (TSs) for the transfer of the H-atom. The optimized geometries of AM0, O ˙ H radical, PRC, TSs, post-reactive complexes (PORCs) and the final radical species of O ˙ H initiated AM oxidation reaction are presented in Figure 4. In detail, the O ˙ H radical attacks the AM0 and forms a PRC, which is stabilized by the formation of H-bonding between the H atom of O ˙ H radical and the O atom of AM0 as shown in Figure 4. The PRC is stabilized with respect to the reactants by −5.26 kcal/mol. Subsequently, the H-bonding stabilized O ˙ H radical rearranges in the reactive space of AM0 in a way to form a bond angle between the H- atom of –CH₂, –NH₂ and –OH groups and the O ˙ H radical close to that of H₂O molecule, to viably eliminate as a H₂O molecule through a transition state.

For example, the bond angle between the H atom of the –CH₂ group and O ˙ H is found to be 94.6° ( i . e ., ∡HC-H··· O ˙ H = 94.6°) in the TS-CH (see Figure 4). The presence of two inequivalent H-atoms on –NH₂ group form two different TSs. The TSs associated with the back and front H-atoms are designated as TS-NH_a and TS-NH_b, respectively. The bond angles are found to be ∡H_bN-H_a··· O ˙ H = 96.5°, ∡H_aN-H_b··· O ˙ H = 103.4° and ∡O-H··· O ˙ H = 101.2° in the TS-NH_a, TS-NH_b and TS-OH, respectively. Computed results reveal that the abstraction of H-atom from the –NH₂ group through the transition state, TS-NH_b (TS-NH_a) is more favorable with an energy barrier of 4.7 (5.7) kcal/mol followed by the abstraction from the –CH₂ (6.5 kcal/mol) and –OH (6.93 kcal/mol) groups. The H transfer through the transition states namely, TS-CH, TS-NH_a, TS-NH_b and TS-OH forms the corresponding post-reactive complexes of water molecule and AM0 radicals, which are stabilized by −26, −21, −21 and −17 kcal/mol with respect to the reactants, respectively. Finally, these post-reactive complexes separate into the AM0 radical and water molecules in each reaction pathway.

To account for the effect of other rotational conformers on the oxidation reaction, the energies of PRCs, TSs and PORCs have been computed for the abstraction of H atom from –CH₂, –NH₂ and –OH groups of other rotational conformations of AM (i.e., AM1 and AM2). The complete reaction profile for the H-abstraction reaction of AM1 along with the energies is presented Supplementary Figure S1. The H abstraction from the –CH₂ group of AM1 proceeds via three TSs such as TS-CH_a TS-CH_b and TS-CH_c. Among these TSs, TS-CH_a is linked to the PRC1 while TS-CH_b and TS-CH_c are linked to the PRC2 (see Supplementary Figure S2). However, all these–CH TSs are linked to the same product complex and lead to the formation of a single NH₂ C ˙ HOH radical conformation. On the other hand, the abstraction of H from –NH and –OH groups of AM1 occurs via the TSs namely, TS-NH and TS-OH, which are linked to the pre-reactive complex, PRC1. These TSs of AM1 subsequently form the N- and O- centered radicals similar to that of AM0+ O ˙ H radical reaction. Similarly, we evaluated the energetics of all important species of AM2+ O ˙ H oxidation reaction and the reaction profile is shown in Supplementary Figure S3. The optimized geometries of the reactive species, intermediates, TSs, post-reactive complexes and C-, N-, O-centered radicals of AM2+ O ˙ H reactions are presented in Supplementary Figure S4.

Finally, the energetics and the barrier height values of H abstraction reactions from AM0, AM1 and AM2 by O ˙ H radical have been compared and presented in Supplementary Table S1. It shows that the energies of PRCs of AM0, AM1 and AM2 are in the range of −4.6 to −5.8 kcal/mol. These values are in good agreement with the PRC energies of similar electronic systems. For example, Franco et al., have investigated the abstraction of H atom from the different conformers of methanediol (Franco et al., 2021). They showed that, the PRC between OHCH₂OH and O ˙ H is stabilized by −5.22 kcal/mol with respect to the reactants in the most stable conformation of methanediol. Ali et al., have studied the O ˙ H + CH₃OH reaction profile under tropospheric conditions (Ali et al., 2021). They showed that the H abstraction occurred through a hydrogen bond stabilized PRC, which exhibits a relative energy of −4.97 kcal/mol. Du and co-workers have explored the gas-phase reaction of ethanol with O ˙ H radical (Xu et al., 2019). Results of this study reveal that the PRC between O ˙ H radical and CH₃CH₂OH is stabilized by the H-bonding. It exhibits a relative energy of −5.2 kcal/mol. On the other hand, González et al. have explored the rate constants for the O ˙ H radical reaction with CH₃NH₂ using experimental and theoretical methods (González et al., 2022). They have found that the PRC between O ˙ H radical and CH₃NH₂ is stabilized through formation of H-bond between the N(CH₃NH₂) and H(OH). It exhibits a relative energy of −6.69 kcal/mol similar to that of PRC1 (−5.8 kcal/mol) formed with the conformer AM1. Clearly, the energies of PRCs formed between O ˙ H radical and different rotational conformers of AM are in close agreement with those of previous studies. The computed barrier height values for the abstraction of H-atom from the different moieties of AM by O ˙ H radical are closely corroborating with that of abstraction of H from the respective moieties of similar electronic systems. For example, Franco et al., have reported the barrier height values in the range of 6.03–6.95 and 8.25–10.22 kcal/mol for the abstraction hydrogen from the –CH₂ and –OH groups of methanediol, respectively (Franco et al., 2021). Similarly, Ali et al., and Xu et al., have reported the barrier height values around 5.8 (7.5) and 5.8 (6.8) kcal/mol for H abstraction from –CH (–OH) groups of methanol and ethanol, respectively (Xu et al., 2019; Ali et al., 2021). Onel et al., and Tian et al., have reported the barrier height values for –NH abstraction of alkylamines around 4.4-9.5 kcal/mol (Tian et al., 2009; Onel et al., 2013). Our computed results show that the barrier height values for the abstraction of H from –CH₂, –NH₂ and –OH groups are in the range of 4.12–6.50, 3.76–5.50 and 6.5–9.25 kcal/mol, respectively. These results also corroborate that the barrier heights are good in agreement with the previous studies. This analysis not only provides more confidence but also substantiate our choice of ab initio and DFT methods.

It is worth to mention here that, the TS energies for the major H-abstraction pathways i.e., from the –CH₂ and –NH₂ channels of NH₂CH₂OH + O ˙ H radical reaction are found to be around 1-2 kcal/mol. Previous studies show that the O ˙ H radical initiated H-abstraction reactions from different reactants are favorable even with the slight positive TS energies. For example, Nguyen et al., have studied the H-abstraction from the CH₃OH using ab initio/RRKM methods (Nguyen et al., 2019). They reported the TS energies for H-abstraction around 2.3 and 0.5 kcal/mol. Further, they satisfactorily reproduced the experimental rate constants using these TS energies. Baidya et al., have explored the H-abstraction reaction of CHF₂CH₂OH with O ˙ H radical to unveil the atmospheric implications of chlorofluorocarbons (Baidya et al., 2018). All the TSs reported in this study exhibit positive energies in the range of 0.8–3.1 kcal/mol. However, they reported that their calculated k _OH value is in good agreement with the experimental value. Rahbar et al., have investigated the kinetics and mechanism of the O ˙ H radical initiated atmospheric oxidation of catechol over the temperature range 200-400 K (Rahbar et al., 2021). However, catechol shows high positive TS values (1.27–8.8 kcal/mol) for the H-abstraction reaction. The authors pointed out that the computed rate constants are close to that of experimental value. In light of these analogous studies, we believe that the slight positive energies (∼1-2 kcal/mol) for the TSs are acceptable within the troposphere, which is the focal point of interest in our present investigation.

The rate constant ( k OH ) and the branching fraction values have been computed for the hydrogen abstraction reactions at 300 K using all possible conformations. The k OH values for the abstraction of H from different channels (–CH₂, –NH₂ and –OH) of AM0 are plotted in Figure 5A and listed in Supplementary Table S2. The sum of the rate constant values from different channels of AM0 (i.e., k _OH-AM0) are presented in Table 1. This data indicates that the rate constants for H abstraction from different channels gradually decrease within the temperature range 200–400 K. The k _OH-AM0 value for O ˙ H radical reaction with AM0 rotational conformation is found to be 1.40 × 10⁻¹² cm³ molecule⁻¹ s⁻¹ at 300 K. Similarly, the k OH for each H abstraction channel of AM1 and AM2 has been computed and plotted in Figures 5B, C and the numerical values are presented in Supplementary Tables S3, S4. The sum of the rate constant values from different channels of AM1 and AM2 (i.e., k _OH-AM1 and k _OH-AM1) has been found to be 4.40 × 10⁻¹² and 1.39 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹, respectively at 300 K (see Table 1). This results in an overall rate constant for the hydrogen abstraction from the all AM conformations to be around 1.97 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ (see Figure 5D; Table 1). The computed total k _OH value of AM is closely aligns with that of CH₃NH₂ (1.97 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹) (Onel et al., 2013), CH₃NHCH₃ (6.27 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹) (Onel et al., 2013), CH₃CH₂NH₂ (2.50 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹) (Onel et al., 2013), NH₂CH₂CH₂OH (7.27 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹) (Xie et al., 2014), CH₃CH₂OH (9.06 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹) at 300 K. The k _OH values are also computed using basis set super position error (BSSE) corrections. It is evident from Supplementary Table S5 that the BSSE corrections does not altered the rate constants appreciably. Overall, a negative temperature dependence in the k _OH of AM has been observed over the temperature range 200–400 K as can be seen from Figure 5D, similar to that of amines and alcohol reactions. The energetic values coupled with the computed rate constant results clearly align with those of previous studies, which further corroborating the reliability of the computational methodology and the scheme adopted in this study. Additionally, the branching fraction values for the abstraction of H from –CH₂, –NH₂ and –OH groups to successively form product radicals NH₂ C ˙ HOH, N ˙ HCH₂OH and NH₂CH₂ O ˙ are estimated to be around 77%, 20% and 3%, respectively. Xie et al., have shown that the branching ratio is around 82%, 17% and 1% for the abstraction of H from –CH₂, –NH₂ and –OH groups of monoethanolamine (MEA) (Xie et al., 2014). The computed branching fraction values of AM are in good agreement with that of MEA (Xie et al., 2014). The branching fraction values clearly suggest a strong preference for the C-centered H-abstraction over N- and O- centered H-abstraction. Interestingly, the computed branching fraction values also strongly adhere to the Evans−Polanyi relationship bond enthalpy of O−H > N−H > CH relationship (Evans and Polanyi, 1938).

The bimolecular reactions between the aminomethanol (NH₂CH₂OH) + O ˙ H radical and the NH₂ C ˙ HOH + O₂ would also be possible in their excited electronic states. It is evident from previous studies that, the computationally characterized potential energy surface (PES) and the corresponding rate constants of (for example, ethyl alcohol (H₃C-CH₂-OH) (Xu et al., 2019), monoethanolamine (NH₂CH₂CH₂OH) (Xie et al., 2014), methanediol (OH-CH₂-OH) (Franco et al., 2021)) similar electronic systems agrees well with that of experimental studies. However, these studies not included any excited state photochemical interactions to obtain the rate constants. Further, Al-Hashimi and co-workers reported that the interactions between the O-Anisidine and O ˙ H radical occurs in the ground state rather than in the excited state (Abdel-Rahman et al., 2021). Priya et al., have explored the abstraction reaction mechanism of O ˙ H radical with 2-methoxyphenol (Priya and Lakshmipathi, 2017). They showed that the H-abstraction reaction can occur in ground state than the in the excited state. Similar to these studies, we believe that the interactions between the aminomethanol and the gaseous species could be corresponding to ground state rather than to excited state.

In recent decades, significant advances in atmospheric chemistry have spurred the development of new theoretical approaches for exploring intricate details of ground-state chemical reactions and their underlying mechanisms. Nevertheless, an equivalent synergy between theory and experimentation remains absent in the realm of atmospheric photochemistry involving electronically excited states. The modeling of molecular photochemistry necessitates a meticulous consideration of non-adiabatic effects, specifically, the coupling between electronic states and molecular motion. This presents formidable challenges, as it contradicts several conventional approximations in theoretical chemistry. Notably, non-adiabatic effects challenge the venerable Born-Oppenheimer approximation, while classical treatments of nuclear dynamics may prove inadequate and non-equilibrium phenomena can challenge established reaction rate theories.

A plethora of methodologies have emerged to address these challenges, including MCTDH (Multi Configuration Time Dependent Hartree) (Manthe et al., 1992; Beck, 2000; Meyer, 2012), trajectory surface hopping (TSH) (Tully and Preston, 1971; Hammes-Schiffer and Tully, 1994; Fabiano et al., 2008) and ab initio multiple spawning (AIMS) (Ben-Nun et al., 2000; Ben-Nun and Martínez, 2002; Yang and Martínez, 2011). However, the application of these techniques to investigate atmospheric photochemistry encounters various challenges. These encompass complexities in simulating spectroscopically relevant choices and atmospheric modeling interests, the intricate electronic architecture of multichromophoric volatile organic compounds (VOCs), the diverse excited-state dynamics triggered by solar irradiation, the protracted dynamics of excited VOCs, the prominence of intersystem crossings or collisional processes and the modulating influence of aqueous environments such as those found in atmospheric aerosols and clouds. Given that non-adiabatic dynamic approaches demand considerable computational resources beyond our current capacity and each method carries its own limitations, we envision these studies as prospects for future exploration.

The aminomethanol radicals produced in the initial reactions of aminomethanol + O ˙ H are highly reactive and undergoes subsequent reactions with atmospheric O_2. The NH₂ C ˙ HOH has been considered for further reactions with atmospheric oxygen due to its large branching fraction. The zero-point energy (ZPE) corrected potential energy surface of NH₂ C ˙ HOH + O₂ reaction has been computed and shown in Figure 6, while the optimized geometries of intermediates, transition state structures and the product complexes (PCs) are depicted in Figure 7. The triplet oxygen molecule (³O₂) reacts with NH₂ C ˙ HOH radical and barrierlessly added to the C-site to form a peroxy radical (IM-0A) intermediate, as shown in Figure 6. However, the attacking direction of O₂ molecule on NH₂ C ˙ HOH and the feasibility for the rotation of –NH₂ and –OH groups around the CO and CN single bonds lead to the formation of different peroxy radical intermediate rotational conformers. Similar to aminomethanol, the IM-0A also exhibits other rotational conformations, for example, IM-0B and IM-0C as presented in Supplementary Figure S5. Previous studies have shown that the rotational conformations of the peroxy radical have a negligible effect on the formation of the end products (Xie et al., 2014). Hence, we have chosen the highly stable NH₂C(O O ˙ )HOH radical (IM-0A) for further studies and the reactions between other rotational conformations and O₂ have been excluded. The IM-0A strongly stabilized and located well below the reactants with a relative energy of −39 kcal/mol with respect to the reactants. The transfer of H occurred from −CH, –NH₂ and –OH groups to the O-site within the peroxy radical intermediate (IM-0A). We consider the H transfer followed by the breaking of CO or OO bonds occurs through two consecutive steps as an indirect mechanism while the same occurs in a single step as a direct mechanism. Overall, five indirect (via TS-1A, TS-3A, TS-5A, TS-7A and TS-8A) and three direct reaction paths (via TS-10A, TS-11A and TS-12A) have been studied for the NH₂ C ˙ HOH + O₂ reaction as shown in Figure 7.

In detail, the transfer of H atom from –CH and –OH groups to the O-site of peroxy radical (IM-0A) is favored through the transition states TS-1A and TS-8A. On the other hand, the presence of two different H atoms in –NH₂ group and different attacking directions of ³O₂ lead to three different reaction pathways, which proceed through the TSs, TS-3A, TS-5A and TS-7A for the H abstraction from –NH₂ group. Among all these TSs, the TS-1A and TS-3A are situated above the reactants with an energy of 1.5 and 0.5 kcal/mol followed by TS-5A, TS-7A and TS-8A (−1.3, −3.7 and −10.0 kcal/mol). It indicates that the H-transfer from –CH group to O-site of IM-0A requires a high energy of ∼40 kcal/mol to form the intermediate IM-1A. The transfer of two different H atoms from –NH₂ group respectively requires 40 (via TS-3A) and 35 (via TS-7A) kcal/mol to form the intermediates, IM-2A and IM-3A. Further, different spatial arrangement of O₂ also facilitates a distinct transition state, TS-5A for the H transfer from –NH₂ group and it leads to the previous intermediate, IM-2A. The reaction proceed through TS-5A shows a barrier height more than 35 kcal/mol. The H transfer process from that of –OH group requires a relatively low energy of 29 kcal/mol via the TS-8A to form IM-4A. The barrier heights for these H-transfer reactions strongly comply with that of H transfer from the similar functional groups. The intermediates IM-1A to IM-4A are stabilized by ∼ -29, −20.2, −20.6 and −16 kcal/mol with respect to the reactants, respectively. The breaking of O-O and C-O bond in IM-1A and IM-4A through the transition states TS-2A and TS-9A form the product complexes, PC-1A and PC-4A, respectively. While, PC-2A and PC-3A can be formed by the breaking of C-O bond of IM-2A via the TSs, TS-4A1 and TS-4A2. The IM-3A also leads to the same product complexes, PC-2A and PC-4A via the TSs, TS-6A1 and TS-6A2. Figure 6 shows that, the IM-1A is almost barrierlessly (0.3 kcal/mol) dissociated into the product complex PC-1A, which is composed of hydrogen bonded amino formic acid (H₂NCOOH) and O ˙ H radical (see Figure 7). The PC-1A is a highly stable product complex with an energy of ∼ -101 kcal/mol, which eventually separates into amino formic acid (NH₂COOH) and O ˙ H radical. While, the IM-2A, IM-3A and IM-4A exhibit a barrier height of ∼13, 16 and 13 kcal/mol for the conversion into PC-2A, PC-3A and PC-4A, which are the hydrogen bond stabilized product complexes of formimidic acid (HN = C(H)-OH), formamide (H₂N-CHO) and O ˙ ₂H radical. The formimidic acid and formamide product complexes (PC-2A, PC-3A and PC-4A) located above than that of amino formic acid (PC-1A) with an energy of −32, −36 and −49 kcal/mol, respectively (see Figure 6). Finally, these product complexes dissociate into the final products of the ³O₂ + NH₂ C ˙ HOH reaction, P-2A (HN = C(H)-OH + O ˙ ₂H), P-3A (HN = C(H)-OH + O ˙ ₂H) and P-4A (H₂N-CHO + O ˙ ₂H) with the energies of −22, 23.5 and −38 kcal/mol, respectively.

On the other hand, the reactions proceed via the transition states, TS-10A, TS-11A and TS-12A directly form the product complexes. In detail, these TSs initiates the simultaneous transfer of H atom from –NH₂ and –OH groups to the O-site of peroxy radical and the breaking of C-O bond to form the post-reactive product complexes, PC-2A, PC-3A and PC-4A, respectively. Similar to TS-3A and TS-7A, the TS-10A and TS-11A involve the transfer of inequivalent H atoms of –NH₂ to the O-site. The energy barriers have been found to be 18, 17 and 4.3 kcal/mol for the transfer of H atom from –NH₂ (via TS-10A, TS-11A) and –OH (via TS-12A) groups to the O-site of peroxy radical. Overall, it is evident from Figure 6 that, the reaction pathway proceeding through the transition state, TS-12A to form P-4A (NH₂CHO) is energetically more favorable followed by that proceeding through TS-10A and TS-11A to form P-2A and P-3A (NHCHOH) by the simultaneous H transfer and CO bond breaking process rather than reaction pathways that proceeding through TS-1A, TS-3A, TS-5A, TS-7A and TS-8A.

We have compared the energetics of important reactive species of NH₂ C ˙ H₂OH + O₂ reaction with that of similar species of previous studies. For example, the relative energy of IM-0A (−39 kcal/mol) is comparable to that of OH C ˙ HOH + O₂ (−39.4 kcal/mol) (Franco et al., 2021), CH₃ C ˙ HNH₂+O₂ (−36 kcal/mol) (Rissanen et al., 2014). Similarly, the barrier height values for the H transfer from–CH (40.5 kcal/mol) via indirect mechanism closely match with that of CH₃ C ˙ HNH₂+O₂ (38 kcal/mol) (Rissanen et al., 2014), CH₂NH₂+O₂ (39.5 kcal/mol) (Rissanen et al., 2014) and CH₂OH + O₂ (40.5 kcal/mol) reactions (Dash and Ali, 2022). On the other hand, the barrier height values for transfer of–NH hydrogen via an indirect mechanism (35-39 kcal/mol) are close to that of CH₂NH₂+O₂ (35.8 kcal/mol) (Rissanen et al., 2014) reaction whereas via direct mechanism (16-18 kcal/mol) are well in agreement with the NH₂ C ˙ HCH₂OH + O₂ (16.3-18.6 kcal/mol) (Silva, 2012; Xie et al., 2014) mechanistic studies. On the other hand, the direct (4.5 kcal/mol) and indirect (29 kcal/mol) –OH hydrogen transfer energy barrier values are also in similar range of OH C ˙ HOH + O₂ (6.7 kcal/mol) (Franco et al., 2021) and CH₂OH + O₂ (25 kcal/mol) (Dash and Ali, 2022) reactions, respectively. These results strongly authenticate the close resemblance of energy parameters of the current study with that of previous studies.

Overall, the computed results suggest that the reaction pathways, which proceed through the TS-1A, TS-3A, TS-5A, TS-7A and TS-8A from the intermediate IM0-A exhibit high energy barriers. These energy barriers are in the range of (30-40 kcal/mol) as shown in Figure 7. Hence, these reaction pathways are excluded from the further studies of reaction kinetics and branching ratio analysis due to the high energy barriers. Relatively low energy barrier pathways which occurred via the transition states TS-10A, TS-11A and TS-12A were adopted for further kinetics and branching ratio calculations.

The rate constants and the branching ratio values of ³O₂ + NH₂ C ˙ HOH reaction have been computed for the most favorable reaction pathways (via TS-10A, TS-11A and TS-12A), which form the product compounds P-2A, P-3A and P-4A. We have treated all these reaction pathways in a master equation to evaluate the temperature and pressure-dependent rate constants. The computed rate constants at the atmospheric conditions relevant to the troposphere, i.e., at 1 atm pressure and 298 K for the overall reaction is 5.5 × 10⁻¹² cm³ molecule⁻¹ s⁻¹. The rate constant of O₂ + NH₂ C ˙ HOH was compared with those of isoelectronic systems, namely, CH₃ C ˙ HOH + O₂ (Silva et al., 2009) and OH C ˙ HOH + O₂ (Ali and Balaganesh, 2023) reactions. We have found that the overall rate constant of O₂ + NH₂ C ˙ HOH decreases with the temperature similar to that of OH C ˙ HOH + O₂ as shown in Figure 8. The total rate constant of NH₂ C ˙ HOH + O₂ reaction is mainly contributed from the rate constant of P-4A reaction rather than that of P-2A and P-3A (see Supplementary Figure S6). Further, branching ratio analysis suggests that ³O₂ + NH₂ C ˙ HOH reaction predominantly leads to the formation of P-4A (NH₂CHO) with the branching fraction of ∼99% over the temperature range 200–400 K, which is akin to the previous analysis on similar reaction system (Ali et al., 2021). The branching fraction values for the other products are negligibly small and independent of temperature and pressure.

We have also computed the ZPE corrected PES of nitrogen centered radical, namely, N ˙ HCH₂OH + O₂ reaction and presented in Supplementary Figure S8. The CH-direct, CH-indirect and OH-indirect reaction pathways shows the high positive energies for the TSs around 8.9 (TS5), 27.6 (TS1) and 19.5 (TS2) kcal/mol and hindered by the strong positive energy barriers, which are around 10.5, 29.2 and 22.9 kcal/mol, respectively. The large positive energies for the TSs are not encouraging and the corresponding reaction pathways will likely not be traversed in the upper atmosphere.

We have conducted Born-Oppenheimer molecular dynamics (BOMD) simulations to investigate the formation of formamide at 300 K through the reactions involving the reactive species ( O ˙ H + AM and NH₂ C ˙ HOH radical + O₂) using the CP2K code (VandeVondele et al., 2005). The computational methodology details are provided in the Supplementary Material. The snapshots captured at different time intervals from the trajectory of the 3ps BOMD simulation for the H abstraction from each channel of AM are shown in Figure 9. These snapshots clearly demonstrates that the abstraction of H from –CH₂, –NH₂ and –OH groups during the O ˙ H + AM reaction follows the similar paths as discussed in Section 3.2. For example, the initial geometry for the H abstraction by the O ˙ H radical from –NH₂ group is depicted in Figure 9 at 0 femtoseconds (fs). During the BOMD simulation, a pre-reactive complex like geometry formed between the AM and O ˙ H radical at 62 fs. Subsequently, we have observed the transfer of H from –NH₂ to O ˙ H radical occurring through a transition state-like geometry at 230 fs. Further, this forms a N ˙ HCH₂OH + H₂O complex at 250 fs. Similar mechanisms are also observed in the cases of H abstraction from –CH₂ and –OH groups of AM as illustrated in Figure 9. Additionally, BOMD simulations have been employed to simulate the reaction between NH₂ C ˙ HOH radical and O₂. Snapshots captured at different time intervals during the BOMD simulations between NH₂ C ˙ HOH and O₂ are presented in Supplementary Figure S7, which are consistent with the earlier formamide formation reaction through TS-12A. These BOMD simulations strongly substantiate the predominant formation of formamide during the NH₂CH₂OH + O ˙ H/O₂ reactions.

Our previous investigation (Ali, 2019) has demonstrated the occurrence of a high-temperature reaction between ammonia (NH₃) and formaldehyde (H₂CO) during biomass burning, resulting in the formation of aminomethanol (NH₂CH₂OH). This prompts a keen interest in elucidating the oxidation reaction mechanism and kinetics of aminomethanol, leading to the formation of various compounds, including potentially hazardous or carcinogenic byproducts. The possible atmospheric decomposition pathway resulting from O ˙ H radical initiated oxidation of NH₂CH₂OH is portrayed in Figure 10. We have computed the atmospheric lifetime (τ) of NH₂CH₂OH due to its interactions with O ˙ H radicals using the following the equation, τ = 1 k × O ˙ H (10) where, k denotes the rate constant of NH₂CH₂OH + O ˙ H radical reaction (5.33 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹) at 230 K, relevant to an altitude of 12 km. The averaged concentration of O ˙ H radicals in the upper troposphere is denoted by O ˙ H and has been taken as 1.0 × 10⁶ molecule cm^-3 akin to that of earlier studies (Ali and Balaganesh, 2023). Computed results reveal that, the aminomethanol has a lifetime of 5 h in the presence of O ˙ H radicals and produce the NH₂ C ˙ HOH radicals as the major product when compared to N ˙ HCH₂OH and NH₂CH₂ O ˙ radicals. Under tropospheric conditions, the major radical product, i.e., NH₂ C ˙ HOH undergoes further reaction with molecular oxygen (³O₂) to yield formamide (NH₂CHO). We have employed k of 7.80 × 10⁻¹² cm³ molecule⁻¹ s⁻¹ at 225 K to calculate the atmospheric lifetime of the NH₂ C ˙ HOH radicals. Our results show that, the lifetime of NH₂ C ˙ HOH is approximately 13 microseconds (μs), indicating a rapid formation of formamide under atmospheric conditions. The current study also suggests that the formations of amino formic acid (NH₂COOH) and formimidic acid (NHCHOH) are unfavorable. It is important to mention here that the formation of formamide involves the reaction of aminomethanol with O ˙ H and O₂. Analysis of NH₂CH₂OH+ O ˙ H reactions indicates that the transition states (TSs) tend to exhibit a slight positive value for the H-abstraction reactions. Consequently, these reactions pose challenges under astrochemical conditions, where temperatures typically remain below 100 K. Nevertheless, the reactions involving the carbon-centred NH₂ C ˙ HOH radical with O₂ may be viable in such conditions, as the corresponding aminomethanol-based peroxy radical intermediate, NH₂CH(O O ˙ )OH required less than 5 kcal/mol for the formamide formation. Therefore, investigating the comprehensive reaction pathways for the formamide formation from aminomethanol under astrochemical conditions holds significant promise for future studies in this field.

Barnes et al., have reviewed the mechanistic details and atmospheric chemistry of amides (Barnes et al., 2010). They showed that the formamide could further react with O ˙ H radicals and produce C- and N-centered formamide radicals as shown in Figure 10. However, Zhu and co-workers have revealed that the C-centered formamide radicals have exclusively been formed and rule out the possibility for the formation of N-centered radicals using the combination of experimental and theoretical studies (Bunkan et al., 2016). They have also found that the isocyanic acid (HNCO), which is known to be a potentially hazardous compound for the human health, is only the product formed during the NH₂CHO + OH and O₂ reaction.

On the other hand, the branching fraction (20%) of N-centered radicals indicates the formation of N ˙ HCH₂OH in significant quantities. It suggests that the N-centered radicals are also prone to further reactions with atmospheric gaseous compounds. However, N-centered radicals usually react slowly with the atmospheric oxygen and potentially leads to the formation of carcinogenic nitrosamines or nitramines through the bi-molecular reactions with the other trace compounds of the atmosphere (NO and NO₂).

Overall, investigating the branching fraction and ensuing rate constants associated with the hydrogen abstraction from –CH₂, –NH₂, and –OH groups of aminomethanol by hydroxyl radicals offers initial insights into the predominant formation of formamide. Additionally, this study provides additional insights into the formation of N-centered radicals in significant quantities. Nevertheless, conducting further investigations on the reactions of O₂, NO and NO₂ with N-centered radicals would undoubtedly contribute to significantly understand their impact on the atmosphere.