Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2026 Apr 24;148(17):17662–17674. doi: 10.1021/jacs.5c21621

Ultrafast Formation of Aza-Quinone Methides from Suitably Substituted o‑Hydroxymethylanilines Involves Higher Excited States

Yifan Su †, Nađa Došlić ‡,*, Manuel Martinović ‡, Mladena Glavaš §, Josip Draženović §, Jiani Ma †,*, Nikola Basarić §,*
PMCID: PMC13154207  PMID: 42031365

Abstract

Aza-quinone methides (aza-QMs) are important intermediates for the synthesis of aza-heterocycles, yet mild methods for their generation remain scarce. Here, we combine experimental and theoretical studies to guide the rational design of aza-QM precursors in efficient photoreactions. Photochemical elimination of H2O from 2-aminobenzyl alcohol (1) serves as a promising starting point. We show that aza-QM formation from 1 is a relatively slow, multistep process involving heterolytic cleavage on the first singlet excited state (S1), formation of a contact ion pair, relaxation to the ground state (S0), and final deprotonation. In the newly designed N-Boc-O-Ac-aminobenzyl alcohol (5), the Boc group increases NH acidity and enforces a reactive conformation, while replacement of OH with an acetyl group introduces a low-energy nπ* state. Benzylic C–O bond elongation stabilizes this state, promoting OAc elimination and enabling a highly efficient pathway. Aza-QM generation from 5 is an ultrafast, intermediate-free process that proceeds via conical intersections from 5 (S1) to the aza-QM in S0. Notably, in 5, and possibly in 1, the benzylic C–O bond cleavage ultimately occurs along their respective dissociative nσ* states. We hope that these mechanistic insights will enable rational design of aza-QM precursors in multistep organic synthesis and in biological applications, where precise control of reactive intermediates is essential to avoid undesired interactions with biomolecular substrates.

graphic file with name ja5c21621_0014.jpg

graphic file with name ja5c21621_0012.jpg

Introduction

Quinone methides (QMs) are commonly encountered reactive intermediates in the chemistry and photochemistry of phenols, , that have been revitalized in the last ten years for the various applications in organic synthesis. − Still dormant, but very useful intermediates are their aza-analogues, aza-quinone methides (aza-QM), which are excellent synthons for the preparation of aza-heterocycles. − For example, they have been employed in the [4 + 2] cycloadditions in the construction of dihydroquinolines, , formal [4 + 3] cycloadditions with nitrones, [4 + 1] cycloadditions in the preparation of indoles, or triarylmethanes and spiroindolines. However, their chemistry has yet to flourish, which will be facilitated if new convenient methods for their preparation under mild conditions at room temperature are discovered. Namely, the generation of aza-QMs in the ground state typically involves the breaking of aromaticity, for which harsh conditions are needed. −

Photochemical synthesis, and particularly photocatalysis − has been in the focus of synthetic organic chemists in the past decade owing to numerous examples of complex molecular constructions, which were approached using photochemistry as the key step. , In that context, it is important to develop photochemical methodology to prepare aza-QMs, which would then be further used in organic synthesis, or biology. The photochemical methods are much milder compared to the thermal since they rely on the use of only photons (no reagents), at ambient temperature, which can be accomplished with spatial and temporal control. However, only a few photochemical strategies for generating aza-QMs have been reported to date. The most exploited is the approach by Kutetaladze et al., who developed an efficient protocol for the generation of aza-QMs based on the excited-state intramolecular proton transfer (ESIPT) from an amine to a carbonyl at the phenylene ortho-position. The reaction has been used in the synthesis of complex heterocyclic systems. − An older and far less explored strategy is the photodehydration of o-hydroxymethylaniline (1 in Chart ), originally proposed by Yang and Wan. The reactivity, stability, and aqueous acidity of aza-QM·1 (Scheme ) and its protonated form were studied a few years ago, but the topic has not been explored further. A likely reason is the reactivity of aza-QM·1 with the aniline precursor as a nucleophile, giving rise to oligomeric products. However, new reactions that can give rise to complex structures from simple precursors are needed, since these significantly increase the chemical space, particularly needed in the context of medicinal chemistry. , Here, we show that the generation of aza-QM from 1 is amenable to systematic improvement.

1. Investigated Aniline Derivatives 1–5 and 5a′ Undergoing Photoelimination to aza-QMs.

1

Open in a new tab

1. Photomethanolysis of 1–5, and Quantum Yields for the Formation of Methanolysis Products OCH 3 ·1–OCH 3 ·4 From 1–5 upon Irradiation in CH3OH Solutions (ΦR)­ ,

1

Open in a new tab

a The irradiations were performed in CH3OH solutions at 254 nm, and KI/KIO3 as an actinometer was used (Φ254 = 0.74). , The composition of the irradiated solution was analyzed by HPLC.

b Taken from ref .

Chemical substitution is a standard strategy for modulating photoreactivity. − Substituent effects can be broadly classified as inertial, steric, or electronic. Inertial effects arise from the mass of the substituent, which can slow nuclear motions such as torsional motion or peptide bond pyramidalization. Steric effects allow control over molecular conformation, with bulky groups restricting the system to more or less reactive geometries. Thus, the NEER effect (non equilibration of excited state rotamers) is a well-known effect in photochemistry affecting the reactivity. , Electronic effects, arising from the introduction of electron-donating or -withdrawing groups, can modify the topography of the relevant potential energy surfaces (PESs), as shown in adenine and 9-methyladenine, as well as in and small organic systems including malonaldehyde, acetylacetone or hexafluoroacetylacetone. − However, none of these strategies alone is sufficient to control the photochemistry of 1.

Here, we introduce chemical substituents in 1 that modify the intrinsic mechanism of the reaction by introducing new electronic states into play. This enables efficient aza-QM formation via elimination of small molecules (AcOH) from the aniline derivative 5 (Chart ). Specifically, the Ac group introduces a low-lying nπ* state localized on the –OAc moiety. Although this state lies above the initially excited ππ* state, small geometric distortions can bring the nπ* below the ππ* state, providing an additional pathway for ultrafast formation of aza-QM. In addition to studying the parent molecule 1 and derivative 5, we also examined how other nitrogen and oxygen substitutions affect the efficiency of the photoelimination reaction. The nitrogen atom was substituted with electron-withdrawing acetyl (Ac, in2, 3) or tert-butoxycarbonyl (Boc, in 4, 5) groups to increase the NH acidity and influence molecular conformation, while the alcohol OH was converted into a better leaving group by substituting it with an Ac group (in 3, 5).

The photochemical reactivity was investigated by preparative irradiations in CH3OH, where the formation of methanolysis products denote the photochemical pathways taking place via aza-QM intermediates. To get more insight into the processes taking place in the excited state and detect aza-QMs and other intermediates, steady-state and time-resolved fluorescence spectroscopy, and transient absorption spectroscopy (TAS) were conducted. Photochemical reaction mechanism for the generation of aza-QMs was proposed based on TAS measurements and computations, which corroborated experimental findings. We demonstrate that efficient ultrafast photoelimination takes place, albeit with different mechanisms. Aniline 1 delivers carbocation that subsequently deprotonates to aza-QM, while for 5, aza-QM intermediate is directly formed in the ground state via easily accessible conical intersections (CI). − Beyond this broad picture, however, understanding the mechanistic details and detecting all intermediates en route to the aza-QMs is important for their practical application, particularly in biological systems.

Results

Synthesis and Photochemistry

Preparation of 2–5 was based on a simple synthetic methodology for the introduction of Ac or Boc groups to the commercially available parent molecule 1 (see Schemes S1–S6 in the Supporting Information).

Photochemical reactivity of 1–5 was probed by irradiating CH3OH solutions, where the formation of methanolysis products (OCH 3 ·1–OCH 3 ·4) indirectly indicated elimination reaction via aza-QM intermediates (aza-QM·1–aza-QM·4, Scheme ). The photomethanolysis reaction of 1 was not clean, as reported by Yang and Wan. In addition to the anticipated product OCH 3 ·1, formation of numerous products took place (Figure S1 in the Supporting Information) due to nucleophilic attack of the amine in 1 to the aza-QM structure. On the contrary, photolysis of 2 afforded mostly OCH 3 ·2 (63%), but the irradiation over 46 h was required (Table S3 in the Supporting Information). Furthermore, photolyses of 3 and 4 gave the anticipated OCH 3 ·2 (43%), and OCH 3 ·4 (11%), and over long irradiations some side products were formed (Schemes S9 and S10 in the Supporting Information) in the competing cleavage reactions of the N-Ac, N-Boc or O-Ac groups. In contrast, the photomethanolysis of 5 was very clean and efficient. Irradiation of 5 in CH3OH–H2O (3:1 v/v, 4h, 300 nm) gave rise to the complete conversion and formation of two solvolysis photoproduct, OCH 3 ·4 and 4, which were isolated in 39% and 20% yield, respectively.

Quantum yield of the photomethanolysis reaction (ΦR) was measured by the use of KI/KIO3 as an actinometer (Φ254 = 0.74). , The ΦR values for 1 (0.40) and 5 (0.49) are similar, whereas for 2–4 are significantly lower (Scheme ). Note, however, that the methanolysis reaction of 5 is clean, contrary to the reaction of 1.

Irradiation of 5 was also conducted in CH3CN to which ethyl vinyl ether (EVE) was added, which is a known quencher of QMs, reacting with them in the Diels–Alder cycloadditions. ,− The irradiation gave rise to the tetrahydroquinoline cycloadduct 6, isolated in 51% yield (Scheme ). The isolation of 6 is a strong indication that aza-QM intermediate is formed in the photoelimination of acetic acid from 5.

2. Irradiation of 5 in the Presence of Ethyl Vinyl Ether (EVE).

2

Open in a new tab

Photophysical Properties

The photophysical properties of 1–5 were investigated in CH3CN solutions (for all data see Figures S6–S15 and Tables S4 and S5). The absorption spectrum of 1 is characterized by two low-energy absorption bands at ≈290 and ≈240 nm (Figure S6 and Table S4 in the Supporting Information), that based on the computations correspond to ππ* transitions (vide infra). Compared to 1, in the absorption spectra of 2–5 (Figures S7–S10), the lowest energy absorption bands are hypsochromically shifted (≈up to 10 nm), have lower oscillator strength and appear as shoulders, whereas the main absorption bands at ≈240 nm are not significantly affected. Thus, the energies of the S1 states of 2–5 are positioned higher in energy by the introduction of the electron withdrawing Ac or Boc groups, and based on computation can be assigned to ππ* transitions (vide infra). Note that Ac and Boc groups influence the nature and energy ordering of the low-lying excited states (S2–S4), as they introduce nπ* and ππ* states localized on the carbonyl groups of 2–5.

Fluorescence spectra were measured in CH3CN, and quantum yields of fluorescence were measured by use of N-acetyltryptophanamide in water as a reference (Φf = 0.12). Fluorescence spectra of 1 and 4 in CH3CN have maxima at 332 nm, and 303 nm, respectively. On the contrary, the spectra of 2, 3 and 5 show the typical dual fluorescence with the main bands at 305–325 nm, and additional weaker bathochromically shifted bands at ≈450–490 nm. The later emission likely originates from minima of initially higher-lying excited states that, during the relaxation dynamics, descend below the initially excited ππ * state (vide infra). Moreover, the Φf for 1 is significantly higher compared to the values for 2–5 (particularly for 2, 3 and 5 which also show dual fluorescence) where the fluorescence is at the limits of the detection (Φf < 0.01, Table S4 in the Supporting Information). Thus, the introduction of the N- and/or O-substituents affected the emissive properties of the molecules by increasing the accessibility of nonradiative deactivation channels, and for 2, 3, and 5 by population of another weakly emissive species.

Fluorescence decays for 1–5 in CH3CN were measured by time-correlated single photon counting (Figures S11–S15 and Table S5 in the Supporting Information). The decays could not be fit to single exponential functions; a sum of two or three exponents was used. The decay for 1 was fit to a sum of two exponents, with a small contribution of a very fast decay component (≈150 ps), which could not be estimated precisely on the setup used, and the major decay component of 2.6 ns. Such a dual fluorescence may originate from the two H-bonded conformers of 1 (vide infra). Similarly, in derivative 4, two exponential decay of fluorescence may be attributed to different conformers that are emissive. On the other hand, for molecules 2, 3, and 5, the fluorescence detected at 320 nm was fit to a sum of three exponents that have very short (≈100–150 ps) major decay component. The fast decay is related to the efficient nonradiative deactivation (and hence low Φf), and probably, to the population of species emitting at higher wavelengths.

Conformational Equilibria and Computed Spectra

Geometries and relative energies of the most stable conformers of 1–5 optimized with MP2/cc-pVDZ, together with their ADC(2)/cc-pVZD vertical excitation energies and dominant natural transition orbitals (NTOs) pairs contributing to the two lowest excited electronic states are given in Figures S16–S28 and Tables S6–S18 in the Supporting Information. Cartesian coordinates of all relevant structures are given in the Supporting Information Section 7.

We focus here on compounds 1 and 5, since they undergo significantly more efficient H2O or AcOH elimination reaction in the investigated series (Scheme ). A comparison with compounds 2–4 is presented in the Discussion Section. In the electronic ground state, the lowest energy conformer of 1, denoted 1a features a O···H–NH hydrogen bond (see Figure S16 for the geometry and numeration of atoms), while the O–H···NH2 conformer, denoted 1b, is found 0.45 (0.42) kcal/mol higher in energy in the gas phase (CH3CN) (Table S6). We note that a synchronous elimination of H2O to generate aza-QM is possible only from conformer 1a in which the amino group donates the proton to the hydroxyl group (d(N–H···O) ≈ 2.11 Å). However, the H2O elimination does not necessarily proceed via a synchronous mechanism (vide infra), as previously reported in the literature.

In the most stable conformer of 5, denoted 5a, the H-bond is formed between the amide group as a donor and the carbonyl oxygen atom of the Ac group (Figure S26 and Table S16). The distance between the proton of the amide group and the mentioned oxygen atom is 1.97 Å, which is 0.14 Å shorter that in 1a, implying possibility of a more efficient synchronous elimination of the AcOH. In conformer 5b, the corresponding N–H···O distance increases to 2.09 Å, reducing the likelihood of synchronous AcOH elimination.

Vertical excitation energies of 1a are listed in Table S7. The two lowest excited singlet states appear at 4.78 eV (S1) and 5.83 eV (S2). In CH3CN, these states shift to 4.79 eV (S1) and 5.78 eV (S2), in good agreement with the experimental values. The two transitions arise from two linear combinations of ππ* excitations delocalized on the phenyl ring. Figure S17 shows the dominant NTOs corresponding to these transitions. Apart from excitation on the phenyl ring, excitation to either S1 or S2 induces charge redistribution on the nitrogen and oxygen atoms. These states can therefore be regarded as precursors for the OH detachment (vide infra), since nuclear motion in these states is likely to elongate the C–O bond and potentially lead to the crossing with dissociative nσ*states, which lie very high in energy at the Franck Condon (FC) geometry. ,

The two lowest excited states of 5a are found at 4.85 (4.86) eV and 5.60 (5.59) eV in the gas phase (CH3CN) (Table S16). At the FC geometry, the third electronic state is only 0.28 eV above the second singlet. From the analysis of the dominant NTOs (Figure S27) one can see that S1 and S2 can be described as ππ* states, while the S3 state has nπ* character with the excitation localized on the Ac group. Extension of the C–O bond (vide infra) stabilizes the S3(nπ*) state, which then becomes S1(nπ*). In 5b, the nπ* state at 5.79 eV is 0.89 and 0.20 eV above the two ππ*states at 4.90 and 5.59 eV. This explains the dual fluorescence in the steady-state spectrum, arising from the presence of both a ππ* and an nπ* minimum on the S1 PES.

Nonadiabatic Dynamics for 1 and 5

To elucidate the mechanism of photoinduced C–O bond cleavage and assist in TAS assignment (vide infra), gas-phase surface-hopping nonadiabatic dynamics simulations (NAD) of compounds 1 and 5 were performed using the fewest-switches algorithm (see Computational Methods and Section 5 in the Supporting Information). These gas-phase simulations aim to reveal plausible photochemical pathways and are not intended for direct quantitative comparison with experiments performed in pure or mixed solvents. The proposed mechanisms were validated and refined through molecular dynamics simulations in explicit water and reaction-path calculations in implicit solvent. NAD simulations were performed for the most stable conformer of 1, 1a, and model compound 5a′ in which the tert-butyl group of 5 was substituted with a methyl group (Figure ). Since the tert-butyl group is not photochemically active in the excitation range of 270–282 nm, its substitution with a methyl group has a negligible effect on the photodynamics (compare the vertical excitation energies in Table S18 and the character of the lowest excited states in Figures S27 and S28), while it accelerates the calculations. SH trajectories were initiated from the S1 state of 1a and 5a′ and propagated for 1 ps in the manifold of three excited states (S1–S3) and the ground state (S0). At the end of the simulation, 10% of 1a trajectories (8/80) and 28% of 5a′ trajectories (20/78) reached the ground state, indicating faster relaxation and photochemistry for 5a′ (see Figure S30). For 1a, the number of trajectories reaching the ground state is too low to draw conclusions about the dominant deactivation mechanism. Two illustrative SH trajectories showing photoelimination of H2O via a S1/S0 CI and the cleavage of the C–O bond in the S1 state leading to the formation of a contact ion pair (CIP) are analyzed in Figures S31 and S32. As noted, the majority of trajectories remained trapped in the S1 state, consistent with its high fluorescence quantum yield of 1.

1.

1

Open in a new tab

Top: time evolution of the potential energies of the three lowest singlet excited states (orange, purple, blue) and the ground state (black) of 5a′. The red dots indicate the populated current state at a given time. The insets show the dominant NTO pairs describing the active electronic state at 10, 70, and 74 fs. Middle: time evolution of the C11–O14 bond. Bottom: the oscillator strengths of the S1 state. The CI between the initially populated S1(ππ*) state and the S1(nσ*) states is encountered at ≈75 fs.

The relaxation of 5a′ is faster (Figure S30). The excited-state lifetime of 5a′ was determined by fitting its population decay to a monoexponential function, yielding the excited-state lifetime of τ ≈ 3100 fs. The dominant relaxation mechanism is the cleavage of the benzylic C–O bond (Figure ). However, analysis shows that 5a′ undergoes fragmentation through two distinct mechanisms. The first mechanism, similar to the one found in 1, is illustrated in Figure . The top panel displays the potential energy evolution of the three lowest singlet excited states and the ground state. The active state, that is the state populated during the dynamics is indicated by red dots. The trajectory evolves in the adiabatic S1 state and at 95 fs reaches the S1/S0 CI through the C11–O14 bond cleavage (middle panel). This is accompanied by the decrease of the N15–H16···O14 distance from 2.29 Å at the start of the dynamics to 1.88 Å at the S1/S0 CI. In our simulation, the leaving fragment is •OAc not AcOH. The H-transfer likely occurs either in the CI region or after reaching the ground state; however, since ADC(2) is a single-reference method, the system’s evolution near or beyond the S1/S0 CI cannot be reliably tracked. In the insets we show the dominant NTO pairs of the current state. Up to 75 fs, the trajectory stays in the ππ*state, after which it encounters a CI with an nσ* state and switches to that state. Once in the nσ* state, the S1/S0 CI is reached in less than 20 fs. The state’s character change is also evident from the sharp change in the S1 state oscillator strength, shown in the bottom panel of Figure .

Figure shows a slower trajectory of 5a′ that reaches the S1/S0 CI at ≈750 fs (top), also by C11–O14 bond cleavage (middle). The main difference from the previous case is the rapid loss of the oscillator strength (bottom) around 20 fs, caused by a nonadiabatic transition from the initially excited ππ* state to the nπ* state (S3 at the FC geometry, Tables S17 ad S18) with the excitation localized on the OAc group (middle inset). This change in the S1 state character is accompanied by an increase in the ground state potential energy by ≈3 eV (black curve). However, the system does not reach the CI with S0 in the S1 (nπ*) state because the C11–O14 bond is not stretched enough. Around 740 fs, a CI with the nσ* state is encountered (see Figure S33). This leads to a sudden extension of the C11–O14 bond and the CI with the ground state.

2.

2

Open in a new tab

Top: time evolution of the potential energies of the three lowest singlet excited states (orange, purple, blue) and the ground state (black) of 5a′. The red dots indicate the populated current state at a given time. Middle: time evolution of the C11–O14 bond. Bottom: the oscillator strengths of the S1 state. The CI between the initially populated ππ* state and the dark nπ* state is encountered at ≈20 fs. At 740 fs a CI with the nσ* state is reached. The insets and Figure S33 show the dominant NTO pairs of the S1 state at t = 0, 150, and 740 fs.

Reaction Paths for the Formation of aza-QM from 1 and 5

The reaction mechanisms suggested by NAD simulations can be better understood in a static framework. In the gas phase, NAD simulations indicate that both 1a and 5a′ undergo homolytic cleavage of the benzylic C–O bond, although with different efficiencies. This process involves significant geometric rearrangement and, according to the simulations, may proceed through more than one electronic state. These results, however, require careful consideration, as solvent effects can substantially alter the reaction by modifying the relevant PESs. Furthermore, experiments were carried out in CH3CN, but trace amounts of water cannot be excluded. Since water can form intermolecular hydrogen bonds with 1a and 5a and directly influence their photochemical reactions, molecular dynamics simulations in explicit water were performed. The results (Figure S29 in the Supporting Information) show that the two compounds experience different solvation environments. For 1a, hydrogen bonds form between the amino nitrogen and water (N16···O ≈ 2.8 Å) and between the hydroxyl oxygen and water (O14···O ≈ 2.7 Å), motivating the inclusion of an explicitly hydrogen-bonded water molecule in the reaction-path calculations, which bridges the amino and hydroxyl groups. In contrast, 5a is stabilized by an intramolecular hydrogen bond, which largely prevents the formation of a persistent intermolecular hydrogen bond with water.

Figures and S34 show the reaction paths for the C–O bond cleavage in 1a·H 2 O and 1a, respectively, optimized in CH3CN using the COSMO model. For comparison with reaction paths optimized in the gas phase see Figures S35–S37 in the Supporting Information. At first glance, the reaction paths for 1a·H 2 O and 1astarting from the respective FC geometries (dark blue line), passing through the minimum of the ππ* state, and involving a barrier of ∼0.4 eV relative to the FC geometries to reach the S2(nσ*)/S1(ππ*) CI, followed by the S1(nσ*)/S0 CI at 4.43 eV (1a) and 3.96 eV (1a·H 2 O)appear qualitatively similar. They indicate that only a negligible fraction of 1a or 1a·H 2 O can overcome this barrier, even when the ground-state zero-point energy contribution (≈0.2 eV) is taken into account. However, the main difference between the reaction profiles occurs after the S2(nσ*)/S1(ππ*) CI, where they diverge significantly. In both systems, the nonadiabatic population transfer between the ππ* and nσ* states may be inefficient but in 1a·H 2 O the extended plateau of the ππ* state indicates that a radical pairlikely a CIP under experimental conditionscan form. This implies that the signature of CIP-1 and carbocation intermediates (cat-1) may be detected by TAS. Overall, based on the present calculations, rapid C–O bond dissociation in 1 appears unlikely.

3.

3

Open in a new tab

Potential energy profiles for C–O bond cleavage 1a·H 2 O in CH3CN. The reaction paths are divided in three segments: from the FC geometry (the computed point 1 from left to right) to the S1(ππ*) minimum (point 5 from left to right), then to the ππ*/nσ* minimum energy CI (point 11), and then to the S1(nσ*)/S0 CI (point 22). Three locally excited ππ*states (blue) and two dissociative nσ* states (red) are shown, together with the ground state (orange). Vertical excitation energies were computed at each geometry using ADC(2)/cc-pVDZ. Insets show the dominant NTO pairs of the lowest ππ* state and the dissociative nσ* state, along with the molecular geometry at the S1/S0 CI. For comparison with 1a see Figure S34 and for gas-phase reaction paths see Figures S35–S37 in the Supporting Information.

Based on the NAD simulations, two reaction profiles were constructed for C11–O14 bond dissociation in 5a (Figure ). The pathway shown in Figure a shares similarities with those of 1a and 1a·H 2 O, as it involves a barrier (0.37 eV in 5a) leading to the S2(nσ*)/S1(ππ*) CI. However, the key difference between 1a and 5a lies in the shape of the ππ* state after this intersection. In 5a, the ππ* surface rises steeply after the S2(nσ*)/S1(ππ*) CI. This increases the probability of nonadiabatic population transfer to the dissociative nσ* state and it is expected to promote faster and potentially more efficient fragmentation of 5a.

4.

4

Open in a new tab

Potential energy profiles for C–O bond cleavage in 5a in CH3CN. Reaction paths: (a) from the FC geometry (point 1) to the ππ*/nσ* CI (point 9) and then to the S1/S0 CI, and (b) from the FC geometry (point 1) to the nπ* minimum (point 7) and then to the S1/S0 CI (point18) via the S2(nσ*)/S1(nπ*) CI. Two ππ* states (blue), two nπ* states (green), and two dissociative nσ* states (red) are shown together with the ground state (orange). Vertical excitation energies were computed at each geometry using ADC(2)/cc-pVDZ in implicit solvent. Insets show the dominant NTOs pairs of the key electronic transitions. Geometries at the FC, ππ*/nπ* CI, and S1/S0 CI are depicted, and the chromophores on which the excitations are localized are highlighted.

The reaction profile shown in Figure b features first a low-barrier transition from the initially populated ππ* state (denoted π1π1*, dark blue) to a nπ* state, denoted n1π3*, (S3 at the FC geometry, green). Upon reaching the minimum of the n1π3* state, a slight elongation of the C11–O14 bond leads to a peaked CI with the dissociative n1σ* state. Relative to the FC geometry, the barrier to the n1π3*/n1σ* CI is 0.43 eV, slightly higher than for the first pathway. However, the narrow barrier and peaked topography of this CI suggests that nonadiabatic transitions from the n1π3*/to the n1σ* state are more probable, leading to more efficient fragmentation.

Based on the reaction profiles shown in Figures and , which connect selected minima and CI points, it is natural to question the robustness of the proposed mechanisms. However, the optimized reaction paths presented in Figures S39–S41 in the Supporting Information demonstrate that the reactions do not require passage through minimum-energy CI points. Instead, they proceed through extended regions of avoided crossing, where two electronic states, separated by up to ∼1 eV, interact strongly and gradually exchange electronic character. The presence of this extended ππ*/nσ* avoided-crossing region makes the mechanism significantly more robust than suggested by Figures and alone.

Femtosecond and Nanosecond TA Measurements

The TAS was used to probe for the formation of aza-QMs and other reactive intermediates. Laser flash photolysis (LFP) measurements were conducted in the Ar- or O2-purged solutions, where the difference was anticipated due to the known quenching of triplets and radicals by O2 (but not aza-QMs). Moreover, the measurements were conducted in CH3CN and CH3CN–PBS (4:1 v/v), where the addition of protic solvent was expected to affect the proton transfer pathways, which may be involved in the elimination of H2O or AcOH. In the aqueous solvent the spectra and decay kinetics were measured also at different pH (in the presence of 50 mM phosphate or PBS buffer). It was reported that aza-QM from 1 was formed in a two-step process, that involved elimination of OH–, followed by the deprotonation of aza-QM from the carbocation (the reported pK a of the aza-QM is 10.6). To assign the transients, the experimental data was compared to simulated spectra. For 1 and 5, the ADC(2)/cc-pVDZ electronic absorption spectra (Figures S44 and S45) show good agreement with the experimental UV–vis spectra, with deviations from experiment of 25 and 3 nm, respectively. Based on the comparison between experimental and computational results for 1 and 5 (Tables S7, S17, S18 and Figures S44 and S45 in the Supporting Information), the most of femtosecond transient absorption (fs-TA) experiments used a 266 nm pump source, which excites the molecules to the S1 state. In addition, some measurements were conducted with the excitation at 245 and 300 nm.

By use of the laser irradiation at 266 nm, 1 in CH3CN is excited to the FC geometry of the S1 state (Figure S46) showing computed absorption at ≈370 nm and 550–700 nm (Figure S47). After that, it transfers very fast to IM1 with the typical signal at 440 and 550 nm. The fs-TA in CH3CN–PBS solution (Figure a–d) is almost the same with that seen in pure CH3CN, while the transformation from the FC S1 to IM1 is faster in protic solvent. In aprotic solvent, the characteristic signal of IM1 is observed within 77 ps, while in a protic solvent, it is within 677 fs. The peak positions of IM1 in CH3CN–PBS solution is the same with that in CH3CN, while the intensity is different (Figures a–d and S46). Based on the NAD simulations (Figure S32) computed reaction energy profile (Figure ) and computed TA spectra (Figures S48 and S53 in the Supporting Information), the IM1 was assigned to the singlet excited state CIP (denoted as CIP-1 (S1)) formed by heterolytic cleavage of the OH– in the excited state. Subsequently, the signal of CIP-1 (S1) transformed to a new species IM2 absorbing at around 407 nm, as well as to IM3 and other species absorbing at 360 and 505 nm. It is suggested that the CIP-1 (S1) deactivates to CIP-1 (S0) through a slight torsion of the NH2 group. Our computations (Figure S49) predict that CIP-1 (S0) absorbs between 350 and 400 nm, depending on the distance between the two ions. We therefore assign IM2 detected at 407 nm to CIP-1 (S0). In the ground state CIP-1 (S0) dissociates to give cat-1 (S0), which shows a computed absorption around 440 nm (Figure S50). This species subsequently deprotonates on a longer time scale to form aza-QM·1 (S0) absorbing at 360 nm (Figure S51) corresponding to the species detected on the millisecond time scale. Recall also that 1 is fluorescent with the lifetime in nanoseconds (Table S5). The TA spectrum of 1a computed at the S1(ππ*) minimum displays two intense transitions at 360 and 505 nm (Figure S52). Therefore, contributions from multiple species to the TA spectra on the nanosecond time scale should be taken into account. Consequently, the most persistent transient species detected by fs-TA on nanosecond time scale were assigned to CIP-1 (S0) absorbing at 407 nm, aza-QM·1 (S0) with the absorption at 360 nm, and relaxed 1a (S1) at 360 and 505 nm (for the comparison of all computed spectra of intermediates with the experimental spectra see Figures S47–S54 in the Supporting Information). The fs-TA data for 1 in CH3CN–PBS solution was also fit globally. However, due to overlapping of the spectral features of too many species, a reliable model that could fit data could not be obtained (Figures S55 and S56 in the Supporting Information).

5.

5

Open in a new tab

(a–d) fs-TA spectra of 1 in CH3CN–PBS (4:1, v/v, pH = 7.4) (λex = 266 nm). (e) TAS of 1 (λex = 266 nm, A 266 = 0.30, c = 5.48 × 10–4 M) in O2-purged CH3CN–H2O (4:1, pH = 7.0, 50 mM sodium phosphate buffer). (f) Decay of TA at 390 nm for the solution of 1 without, and in the presence of NaSCN (0.11 M).

The computed PES for the photoelimination from 1 (Figure ) indicates an energy barrier (vide supra). Consequently, it is plausible that excitation at different wavelengths may give rise to different observed rate constants for the formation of CIP-1 (S1) and subsequent transients. Therefore, we conducted fs-TA measurements for 1 in CH3CN–PBS solution at different excitation wavelengths, 245 and 300 nm (Figures S57 and S58 in the Supporting Information). The spectra look qualitatively similar. However, no signal corresponding to the FC geometry of 1 was observed under 245 nm excitation, indicating that it rapidly formed CIP-1 (S1). On the contrary, a slower conversion of the FC of 1 to the CIP-1 (S1) under 300 nm excitation (within 500–600 fs) was observed. However, no significant difference was observed in the kinetics for the formation of CIP-1 (S0) and other transients under different excitation wavelengths.

In the ns-TA spectra of 1 in Ar-purged CH3CN (Figure S60 in the Supporting Information), two transients were detected that absorbed at 300–450 nm and decayed with lifetimes τ ≈ 1.4 ± 0.2 μs and τ ≈ 19 ± 1 ms, and were not quenched by O2, precluding their assignment to the triplet excited state of the molecule or radicals. The short-lived transient with the absorption at λ > 400 nm was tentatively assigned to the solvated cat-1 (S0), whereas the long-lived transient with the absorption band at 360 nm was tentatively assigned to aza-QM·1. Attempts to quench the transient by EVE failed, and other quenching studies that would prove the assignment were not possible due to the very low intensity of the transient absorption. Note that we did not observe the transformation of one species to another with the decay of the fist matching the rise kinetics of the second, as reported by Sung et al. The decay of TA was fit to a sum of two exponents not indicating that one transient was formed from the other.

In the aqueous solution of 1 at pH 4 (Figure S62 in the Supporting Information), two transients were detected in both Ar- and O2-purged solution, that decayed with the lifetime of τ ≈ 4.8 ± 0.3 μs and τ > 100 μs, and their intensity was weak to allow further studies. On the contrary, at pH 7, the TA signal was stronger. Two transients were detected (Figures e and S64 in the Supporting Information); the short-lived one absorbed with the maximum at ≈ 400 nm and decayed with the lifetime of τ ≈ 71 ± 1 μs. The longer-lived transient absorbed with the maximum at 360 nm, decayed with τ ≈ 300 ± 30 μs. The purging by O2 did not affect the intensity or the decay kinetics of the transients, suggesting that triplet excited states or radicals were not involved. To assign the transients detected at pH 7, quenching experiments were performed with nucleophiles NaN3, and NaSCN (Figures S65–S69 in the Supporting Information), the ubiquitous quenchers of species such as cations, aza-QMs or QMs. ,,, Both nucleophiles quenched the faster decay component (for the rate constants see Table ), which was assigned to cat-1 (S0). The longer-lived decay component was also quenched by NaN3, NaSCN and EVE (Figures S65–S69 in the Supporting Information), but we could not estimate precisely and accurately reliable rate constants. Consequently, we assign the transient with the lifetime of τ ≈ 71 ± 1 μs to the carbocation cat-1 (S0), or protonated aza-QM, and the long-lived transient (τ ≈ 300 ± 30 μs) is aza-QM·1 (S0).

1. Rate Constants (k q) for the Quenching of Transients Determined by Ns-TA.

compound conditions quencher k q/M–1 s–1
1 CH3CN–H2O (4:1), pH 7 NaN3 (1.3 ± 0.1) × 108
1 CH3CN–H2O (4:1), pH 7 NaSCN (1.0 ± 0.2) × 105
5 CH3CN EVE 33 ± 1
5 CH3CN ethanolamine (5.8 ± 0.2) × 104
5 CH3CN–H2O (4:1), pH 7 NaN3 (3.48 ± 0.06) × 105
5 CH3CN–H2O (4:1), pH 7 NaSCN (1.00 ± 0.03) × 105
Open in a new tab

The fs-TA spectra for 5 in CH3CN–PBS solution (Figure ) are identical to those seen in pure CH3CN (Figure S70). Upon excitation at 266 nm, a signal appears at 346 and 460 nm (Figure a), which is assigned to 5 (S1). This signal stops rising after 727 fs, consistent with the nonadiabatic simulations for the model compound 5a′. Subsequently, a signal emerges at 400 nm (Figure b), that is assigned to aza-QM·4 (S0) based on computed spectra (Figures S72 and S74 in the Supporting Information). The formation of aza-QM·4 (S0) likely occurs before 1.57 ps, but the signal is initially obscured by the stronger absorption of 5 (S1). The global fitting of the fs-TA data for 5 indicated only two species (for the decay associated spectra see Figure S75 in the Supporting Information). The shorter time constant was assigned to the lifetime of 5 (S1), and the long-lived species to aza-QM·4 (S0). Thus, upon excitation, the S1 state of 5 was generated, followed by synchronous elimination of acetic acid to generate aza-QM·4 (S0), most probably via a CI, since other intermediates en route were not observed.

6.

6

Open in a new tab

(a,b) fs-TA spectra of 5 in CH3CN–PBS (4:1, v/v) (λex = 266 nm). (c) TA spectra of 5 (λex = 266 nm, A 266 = 0.30, c = 4.00 × 10–4 M) in O2-purged CH3CN–H2O (4:1, pH = 7.0, 50 mM sodium phosphate buffer). Inset: Decays at 390 nm for the O2-purged solutions of 5 (c = 4.00 × 10–4 M) in CH3CN and CH3CN–H2O (4:1) at pH 7. (d) Decay of the TA at 400 nm for the solution of 5 in CH3CN, without and in the presence of EVE (c = 4.00 × 10–4 M).

fs-TA spectra for 5 in CH3CN–PBS (4:1, v/v) were also measured upon excitation at 245 and 300 nm (Figures S76 and S77 in the Supporting Information). Qualitatively, they look similar as those upon excitation at 266 nm. However, the dynamics is different. Upon excitation at 245 and 266 nm the signal at 400 nm assigned to aza-QM·4 (S0) appears early (≈ 1 ps). However, the signals were different under 300 nm excitation. During the early stage of fs-TA, signals at 346, 425, and 460 nm appeared simultaneously. Based on the computation (Figure S73 in the Supporting Information), the signal at 425 nm was assigned to the relaxed 5 (S1) n,π* state. Subsequently, the signal at 346 nm decayed faster than the signal at 425 nm and the generation of aza-QM·4 (S0) was first observed at 18.8 ps. Thus, the fs-TA measurements are in line with the computations. Excitation at longer wavelength gives aza-QM·4 (S0) less efficiently in the slower process.

In contrast to the ns-TA spectra of 1, only one transient species was detected by ns-TA from 5, formed within the laser pulse, absorbing with the maximum at ≈ 400 nm (Figures c and S79–S88 in the Supporting Information). The transient decayed with the lifetime of τ ≈ 0.5 ± 0.2 s in CH3CN solution, and was not quenched by O2 (Figures S80 and S83). The transient had somewhat slower decay kinetics in the aqueous solution, where it decayed with the lifetime τ ≈ 0.21 ± 0.02 s (Figure S84 in the Supporting Information), and was also not quenched by O2. Based on the insensitivity to O2, position of the maximum in the spectra, and comparison to the computed spectra (Figures S72 and S74 in the Supporting Information) the transient was assigned to aza-QM·4 (S0). To verify the assignments, the quenching studies were performed with NaN3, NaSCN and ethanolamine, as well as with EVE. − The values of all rate constants (Table ) are in accord with the assignment of the transient to aza-QM·4. In particular, the quenching by EVE (Figure d), undoubtedly proved the assignment. Namely, other plausible electrophilic species such as carbocations may also be formed, and react with nucleophiles. However, the reactions of carbocations with nucleophiles proceed faster and the Diels–Alder reactions are not possible. Consequently, fs-TA and ns-TA experiments for 5 clearly gave rise to aza-QM·4, which was formed efficiently from 5 (S1).

Discussion

Based on the experimental and computational results, a mechanism for the H2O elimination from the parent molecule 1 and AcOH elimination from derivative 5 can be proposed. The overall mechanism for the H2O elimination from 1 to aza-QM·1 (S0) is shown in Scheme . After the initial population of 1 (S1), it undergoes the cleavage of the C–O bond at the benzylic position. For this to occur, a crossing to an nσ* statewhere the σ* orbital is localized on the bond to be cleavedmay be required, as indicated by the computed PES connecting the FC region to the S1/S0 CI. This result is noteworthy in the context of reactions that generate QMs or aza-QMs. Furthermore, NAD simulations and reaction-path scans in solution indicate that both OH· radical and H2O elimination are inefficient due to a broad energy barrier along the pathway from the 1 (S1) FC to the nσ* CI. Thus, 1 is characterized by high value of Φf, which is also the consequence of the presence of two conformers (1a and 1b), where only 1a is reactive, and 1b is emissive.

3. H2O Photoelimination Mechanism from 1 .

3

Open in a new tab

The cleavage of the C–O bond in solution is most likely heterolytic, as indicated by the fs-TA experiments. The discrepancy between experiment and theory likely reflects current methodological limitations, as current computational approaches cannot reliably describe heterolytic bond dissociation either in the gas phase or within implicit solvent models. Fragment-based quantum chemical methods capable of distinguishing between homolytic and heterolytic pathways are only now emerging and have so far been primarily applied to ground-state reactions in the gas phase.

Notwithstanding these limitations, our computations for 1a·H 2 O, conducted in implicit CH3CN and including an explicitly hydrogen-bonded water molecule bridging the NH2 and OH groups (Figure ), are generally in accord with the fs-TA experiments. Compared to gas-phase studies, we found that solvation strongly alters the topography of the PESs near the S2(nσ*)/S1(ππ*) CI, making excited-state OH– elimination and formation of CIP-1 (S1) more favorable, in agreement with the fs-TA observations, which detect the formation of the first intermediate (IM1). Thereafter, the CIP-1 (S0) is generated via an easily accessible CI. Subsequently, CIP-1 (S0) undergoes dissociation to form free solvated carbocation cat-1 (S0), which was observed by TAS on the nano- and microsecond time scale. The N–H bond cleavage to form aza-QM·1 (S0) takes place over relatively a long time of 71 μs due to its strong basicity. We also note a possibility for the direct formation of aza-QM·1 (S0) from 1 (S1) based on computed PES and NAD, which is however inefficient minor pathway. Due to the direct formation of aza-QM·1 (S0) from 1 (S1) the rise kinetics of aza-QM·1 (S0) during the decay of cat-1 (S0) was not detected.

The computations for 5a conducted for the gas phase and in implicit CH3CN solution give rise to the same conclusions, which are in accordance with experiments. In the photochemistry of 5, the substitution of nitrogen and oxygen by Boc and As group is beneficial. The steric bulk of the Boc forces the molecule to adopt the reactive conformation for the synchronous elimination of AcOH. Moreover, the Ac group brings in the dark nπ* state, locally excited on the Ac. Although higher in energy then the lowest ππ* state in the FC region, upon elongation of the C–O bond the energy of the nπ* state is lowered and becomes S1. Therefore, two ultrafast relaxation pathways are available to 5. The first mechanism, similar to that operating in 1, involves a direct crossing between the initially excited ππ* state and the dissociative nσ* state (Figure a). The second mechanism proceeds via internal conversion from the ππ* to the localized nπ*state, which was detected by fluorescence and fs-TA, followed by crossing with the nσ* state (Figure b). In both cases, the elimination of AcOH requires the involvement of an nσ* state. Therefore, 5 exhibits dual fluorescence with the emission from the ππ* and nπ* states. Once the nσ* is reached, the cleavage of the C–O bond takes place delivering the aza-QM·4 via the S1/S0 CI. In any case, both pathways directly populate aza-QM·4 (S0), which was shown experimentally by fs-TA (Scheme ). We emphasize the long lifetime of this aza-QM, which allows it to react with various nucleophiles and in relatively slow Diels–Alder reactions forming cycloadducts. Consequently, photochemistry of 5 or similar molecules, may be used in organic synthesis for the construction of more complex aza-heterocycles. Moreover, it may be of interest in biological systems in the development of photocages since it directly involves transformation to aza-QMs with the elimination of a potentially biologically active cargo in the form of carboxylic acid.

4. Photochemical Generation of aza-QM·4 by Elimination of AcOH from 5 .

4

Open in a new tab

The inefficiency of aza-QM generation from compounds 2–4 can be explained by the ordering and character of the excited states. In 2 and 3, the nitrogen is substituted with an electron-withdrawing Ac group. While this increases NH acidity, it also introduces a low-lying nπ* state on the amide. In 2a, this state is S2 at the FC geometry and as shown in Figure S42 it is readily populated from the initially excited ππ* state, as the barrier to the ππ*/nπ* CI is below 0.1 eV. From the nπ* minimum, the system reaches the S1/S0 CI via amide pyramidalization and CO elongation, leaving the –OH group intact. In 3a, the amide-localized nπ* state (S2) lies 0.33 eV below the –OAc-centered nπ* state (S4), similarly redirecting excitation to a nonreactive site. In 4a, due to the Boc substitution, the two lowest excited states are of ππ* character. However, the charge redistribution in the S1 state leaves the dissociative C–O–H moiety largely unaffected by electronic excitation (see electron density differences in Figure S43), indicating that no significant driving force acts along the C–O coordinate. Only in 5 is excitation efficiently localized at the dissociative site, enabling efficient aza-QM generation.

Conclusion

Our combined experimental and computational investigation has elucidated the mechanistic details of the photochemical elimination of H2O or AcOH from suitably substituted anilines, leading to the formation of aza-QMs.

The formation of aza-QM from 1 is a stepwise process involving OH–elimination in the excited-state, which delivers the corresponding CIP in the S1 state. The CIP deactivates to S0 and dissociates to the free cation detected by ns-TA. Subsequent deprotonation yields aza-QM from 1. In contrast, the formation of aza-QM from 5 is an ultrafast process occurring without detectable intermediates, proceeding from 5 (S1) to the corresponding aza-QM (S0) via internal conversion at conical intersections. For both molecules, calculations indicate the involvement of dissociative, higher-lying nσ* states, which decrease in energy as the benzylic C–O bond elongates and cross with the initially excited S1(ππ*) state (ππ* → nσ*). For the model compound 5, the calculations further suggest a possible indirect ππ* → nπ* → nσ* pathway involving an intermediate nπ* state, which was detected by fluorescence and fs-TA.

The mechanistic details unveiled herein are necessary in the rational design of aza-QM precursors for their efficient generation. In particular, it was demonstrated that N-Boc and O-Ac substitutions have beneficial effects by increasing the population of the reactive conformer for elimination and introducing a localized, easily accessible nπ* state, which provides an additional pathway for the elimination of the OAc fragment from the molecule. A thorough understanding of these mechanistic features is also crucial for the application of aza-QMs in biological contexts, where reactive intermediates may interact with biological substrates and lead to undesired effects.

Computational Methods

The conformational space of 1–5 in the electronic ground state was examined using the Møller–Plesset second-order (MP2) method and a range of correlation-consistent basis sets in the gas phase and in solution, where the COSMO implicit solvation model was used to approximate CH3CN. , The photophysical properties of the most stable conformers were subsequently analyzed with the adiabatic diagrammatic construction second-order [ADC(2)] method. − Nonadiabatic dynamics simulations were performed with Tully’s fewest switches surface-hopping algorithm using the Zagreb surface hopping program available in the Quantics suits of programs. , For detail on the computational procedure see Supporting Information Sections 4 and 5.

Supplementary Material

ja5c21621_si_001.pdf (6.1MB, pdf)

Acknowledgments

This research was funded by the National Natural Science Foundation of China (22322301) and Croatian Science Foundation (HRZZ grants no. HRZZ IP-2022-010-4658 to ND and HRZZ-IP-2019-04-8008 and HRZZ-IP-2024-05-8565 to NB). ND and NB are grateful to the ADRIS foundation for supporting this work. Dedicated to Professor Yu Fang on the occasion of his 70th birthday.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c21621.

  • Contains detailed experimental procedures for the synthesis of molecules and photochemical experiments, photophysical data, UV–vis and fluorescence spectra, computational data on conformational equilibria, vertical excitations, nonadiabatic dynamics, and reaction path analyses, transient absorption spectroscopy data, coordinates of the optimized structures and copies of 1H and 13C NMR spectra (PDF)

∥.

NMR Center, Rud̵er Bošković Institute, Bijenička cesta 54, 10000 Zagreb, Croatia

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.

References

  1. Rokita, S. E. Quinone Methides; Wiley: Hoboken, 2009. [Google Scholar]
  2. Ge L., Guo Y., Zhang S.-L., Phillips D. L., Basarić N., Ma J.. Mechanistic Investigations of Photochemical Generation of Quinone Methides. Phys. Chem. Chem. Phys. 2025;27:15272–15292. doi: 10.1039/d5cp01322d. [DOI] [PubMed] [Google Scholar]
  3. Singh M. S., Nagaraju A., Anand N., Chowdhury S.. ortho-Quinone Methide (o-QM): a Highly Reactive, Ephemeral and Versatile Intermediate in Organic Synthesis. RSC Adv. 2014;4:55924–55959. doi: 10.1039/C4RA11444B. [DOI] [Google Scholar]
  4. Bai W.-J., David J. G., Feng Z.-G., Weaver M. G., Wu K.-L., Pettus T. R. R.. The Domestication of ortho-Quinone Methides. Acc. Chem. Res. 2014;47:3655–3664. doi: 10.1021/ar500330x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Uyanik M., Nishioka K., Kondo R., Ishihara K.. Chemoselective Oxidative Generation of ortho-Quinone Methides and Tandem Transformations. Nat. Chem. 2020;12:353–362. doi: 10.1038/s41557-020-0433-4. [DOI] [PubMed] [Google Scholar]
  6. Wojciechowski K.. Aza-ortho-xylylenes in Organic Synthesis. Eur. J. Org. Chem. 2001;2001:3587–3605. doi: 10.1002/1099-0690(200110)2001:19<3587::AID-EJOC3587>3.0.CO;2-5. [DOI] [Google Scholar]
  7. Yang B., Gao S.. Recent Advances in the Application of Diels-Alder Reactions Involving o-Quinodimethanes, Aza-o-Quinone Methides and o-Quinone Methides in Natural Product Total Synthesis. Chem. Soc. Rev. 2018;47:7926–7953. doi: 10.1039/C8CS00274F. [DOI] [PubMed] [Google Scholar]
  8. Liao H.-H., Miñoza S., Lee S.-C., Rueping M.. Aza-Ortho-Quinone Methides as Reactive Intermediates: Generation and Utility in Contemporary Asymmetric Synthesis. Chem.Eur. J. 2022;28:e202201112. doi: 10.1002/chem.202201112. [DOI] [PubMed] [Google Scholar]
  9. Steinhagen H., Corey E. J.. A Convenient and Versatile Route to Hydroquinolines by Inter- and Intramolecular Aza-Diels-Alder Pathways. Angew. Chem., Int. Ed. 1999;38:1928–1931. doi: 10.1002/(SICI)1521-3773(19990712)38:13/14<1928::AID-ANIE1928>3.0.CO;2-1. [DOI] [PubMed] [Google Scholar]
  10. Lee A., Younai A., Price C. K., Izquierdo J., Mishra R. K., Scheidt K. A.. Enantioselective Annulations for Dihydroquinolones by in Situ Generation of Azolium Enolates. J. Am. Chem. Soc. 2014;136:10589–10592. doi: 10.1021/ja505880r. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Zheng Y.-S., Tu L., Gao L.-M., Huang R., Feng T., Sun H., Wang W.-X., Li Z.-H., Liu J.-K.. Accessing Benzooxadiazepines via Formal [4 + 3] Cycloadditions of Aza-o-Quinone Methides with Nitrones. Org. Biomol. Chem. 2018;16:2639–2642. doi: 10.1039/C8OB00201K. [DOI] [PubMed] [Google Scholar]
  12. Liu Y.-Y., Yu X.-Y., Chen J.-R., Qiao M.-M., Qi X., Shi D.-Q., Xiao W.-J.. Visible-Light-Driven Aza-ortho-quinone Methide Generation for the Synthesis of Indoles in a Multicomponent Reaction. Angew. Chem., Int. Ed. 2017;56:9527. doi: 10.1002/anie.201704690. [DOI] [PubMed] [Google Scholar]
  13. Liao H.-H., Chatupheeraphat A., Hsiao C.-C., Atodiresei I., Rueping M.. Asymmetric Brønsted Acid Catalyzed Synthesis of Triarylmethanes - Construction of Communesin and Spiroindoline Scaffolds. Angew. Chem., Int. Ed. 2015;54:15540–15544. doi: 10.1002/anie.201505981. [DOI] [PubMed] [Google Scholar]
  14. Walden D. M., Jaworski A. A., Johnston R. C., Hovey M. T., Baker H. V., Meyer M. P., Scheidt K. A., Cheong P. H.-Y.. Formation of Aza-ortho-quinone Methides under Room Temperature Conditions: Cs2CO3 Effect. J. Org. Chem. 2017;82:7183–7189. doi: 10.1021/acs.joc.7b00697. [DOI] [PubMed] [Google Scholar]
  15. Michelin C., Hoffmann N.. Photosensitization and Photocatalysis Perspectives in Organic Synthesis. ACS Catal. 2018;8:12046–12055. doi: 10.1021/acscatal.8b03050. [DOI] [Google Scholar]
  16. Prier C. K., Rankic D. A., MacMillan D. W. C.. Visible Light Photoredox Catalysis with Transition Metal Complexes: Applications in Organic Synthesis. Chem. Rev. 2013;113:5322–5363. doi: 10.1021/cr300503r. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Romero N. A., Nicewicz D. A.. Organic Photoredox Catalysis. Chem. Rev. 2016;116:10075–10166. doi: 10.1021/acs.chemrev.6b00057. [DOI] [PubMed] [Google Scholar]
  18. Skubi K. L., Blum T. R., Yoon T. P.. Dual Catalysis Strategies in Photochemical Synthesis. Chem. Rev. 2016;116:10035–10074. doi: 10.1021/acs.chemrev.6b00018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kärkäs M. D., Porco J. A., Stephenson C. R. J.. Photochemical Approaches to Complex Chemotypes: Applications in Natural Product Synthesis. Chem. Rev. 2016;116:9683–9747. doi: 10.1021/acs.chemrev.5b00760. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Pitre S. P., Overman L. E.. Strategic Use of Visible-Light Photoredox Catalysis in Natural Product Synthesis. Chem. Rev. 2022;122:1717–1751. doi: 10.1021/acs.chemrev.1c00247. [DOI] [PubMed] [Google Scholar]
  21. Mukhina O. A., Cronk W. C., Kumar N. N. B., Sekhar M. C., Samanta A., Kutateladze A. G.. Intramolecular Cycloadditions of Photogenerated Azaxylylenes: An Experimental and Theoretical Study. J. Phys. Chem. A. 2014;118:10487–10496. doi: 10.1021/jp504281y. [DOI] [PubMed] [Google Scholar]
  22. Mukhina O. A., Bhuvan Kumar N. N., Arisco T. M., Valiulin R. A., Metzel G. A., Kutateladze A. G.. Rapid Photoassisted Access to N,O,S-Polyheterocycles with Benzoazocine and Hydroquinoline Cores: Intramolecular Cycloadditions of Photogenerated Azaxylylenes. Angew. Chem., Int. Ed. 2011;50:9423–9428. doi: 10.1002/anie.201103597. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kumar N. N. B., Mukhina O. A., Kutateladze A. G.. Photoassisted Synthesis of Enantiopure Alkaloid Mimics Possessing Unprecedented Polyheterocyclic Cores. J. Am. Chem. Soc. 2013;135:9608–9611. doi: 10.1021/ja4042109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mukhina O. A., Kuznetsov D. M., Cowger T. M., Kutateladze A. G.. Amino Azaxylylenes Photogenerated from o-Amido Imines: Photoassisted Access to Complex Spiro-Poly-Heterocycles. Angew. Chem., Int. Ed. 2015;54:11516–11520. doi: 10.1002/anie.201504455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kuznetsov D. M., Mukhina O. A., Kutateladze A. G.. Photoassisted Synthesis of Complex Molecular Architectures: Dearomatization of Benzenoid Arenes with Aza-o-xylylenes via an Unprecedented [2 + 4] Reaction Topology. Angew. Chem., Int. Ed. 2016;55:6988–6991. doi: 10.1002/anie.201602288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kuznetsov D. M., Kutateladze A. G.. Step-Economical Photoassisted Diversity-Oriented Synthesis: Sustaining Cascade Photoreactions in Oxalyl Anilides to Access Complex Polyheterocyclic Molecular Architectures. J. Am. Chem. Soc. 2017;139:16584–16590. doi: 10.1021/jacs.7b07598. [DOI] [PubMed] [Google Scholar]
  27. Yang C., Wan P.. Photosolvolysis of 2-Aminobenzyl Alcohol in Aqueous Solution. J. Photochem. Photobiol., A. 1994;80:227–232. doi: 10.1016/1010-6030(94)85005-4. [DOI] [Google Scholar]
  28. Wu T.-H., Su Z.-S., Sung R., Sung K.. Aza-ortho-Quinone Methide and Its Conjugated Acid: Reactivity, Stability and Acidity. ChemPhysChem. 2020;21:307–312. doi: 10.1002/cphc.201901133. [DOI] [PubMed] [Google Scholar]
  29. Wang M., Li S., Wang J., Zhang O., Du H., Jiang D., Wu Z., Deng Y., Kang Y., Pan P., Li D., Wang X., Yao X., Hou T., Hsieh C.-Y.. ClickGen: Directed Exploration of Synthesizable Chemical Space via Modular Reactions and Reinforcement Learning. Nat. Commun. 2024;15:10127. doi: 10.1038/s41467-024-54456-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Chow S. Y., Nelson A.. Embarking on a Chemical Space Odyssey. J. Med. Chem. 2017;60:3591–3593. doi: 10.1021/acs.jmedchem.7b00423. [DOI] [PubMed] [Google Scholar]
  31. Rydholm E., Bastys T., Svensson E., Kannas C., Engkvist O., Kogej T.. Expanding the Chemical Space Using a Chemical Reaction Knowledge Graph. Digital Discovery. 2024;3:1378–1388. doi: 10.1039/D3DD00230F. [DOI] [Google Scholar]
  32. MacDonell R. J., Schuurman M. S.. Site-Selective Isomerization of Cyano-Substituted Butadienes: Chemical Control of Nonadiabatic Dynamics. J. Phys. Chem. A. 2019;123:4693–4701. doi: 10.1021/acs.jpca.9b02446. [DOI] [PubMed] [Google Scholar]
  33. MacDonell R. J., Corrales M. E., Boguslavskiy A. E., Bañares L., Stolow A., Schuurman M. S.. Substituent Effects on Nonadiabatic Excited State Dynamics: Inertial, Steric, and Electronic Effects in Methylated Butadienes. J. Chem. Phys. 2020;152:084308. doi: 10.1063/1.5139446. [DOI] [PubMed] [Google Scholar]
  34. Zimmerman H. E.. The Meta Effect in Organic Photochemistry: Mechanistic and Exploratory Organic Photochemistry. J. Am. Chem. Soc. 1995;117:8988–8991. doi: 10.1021/ja00140a014. [DOI] [Google Scholar]
  35. Chakraborty P., Couto R. C., List N. H.. Deciphering Methylation Effects on S2(ππ*) Internal Conversion in the Simplest Linear α,β-Unsaturated Carbonyl. J. Phys. Chem. A. 2023;127:5360–5373. doi: 10.1021/acs.jpca.3c02582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mirón G. D., Semelak J. A., Grisanti L., Rodriguez A., Conti I., Stella M., Velusamy J., Seriani N., Došlić N., Rivalta I., Garavelli M., Estrin D. A., Kaminski Schierle G. S., González Lebrero M. C., Hassanali A., Morzan U. N.. The Carbonyl-Lock Mechanism Underlying Non-Aromatic Fluorescence in Biological Matter. Nat. Commun. 2023;14:7325. doi: 10.1038/s41467-023-42874-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Whitesell J. K., Minton M. A., Tran V. D.. The NEER Principle. Ground-State Conformational Bias in Triene Photocyclizations. J. Am. Chem. Soc. 1989;111:1473–1476. doi: 10.1021/ja00186a049. [DOI] [Google Scholar]
  38. Mazzucato U., Momicchioli F.. Rotational Isomerism in trans-1,2-Diarylethylenes. Chem. Rev. 1991;91:1679–1719. doi: 10.1021/cr00008a002. [DOI] [Google Scholar]
  39. Satzger H., Townsend D., Zgierski M. Z., Patchkovskii S., Ullrich S., Stolow A.. Primary Processes Underlying the Photostability of Isolated DNA Bases: Adenine. Proc. Natl. Acad. Sci. U.S.A. 2006;103:10196–10201. doi: 10.1073/pnas.0602663103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Bhattacherjee A., Pemmaraju C. D., Schnorr K., Attar A. R., Leone S. R.. Ultrafast Intersystem Crossing in Acetylacetone via Femtosecond X-ray Transient Absorption at the Carbon K-Edge. J. Am. Chem. Soc. 2017;139:16576–16583. doi: 10.1021/jacs.7b07532. [DOI] [PubMed] [Google Scholar]
  41. Sapunar M., Ayari T., Došlić N.. Comparative Study of the Photodynamics of Malonaldehyde and Acetylacetone. Chem. Phys. 2018;515:622–627. doi: 10.1016/j.chemphys.2018.07.042. [DOI] [Google Scholar]
  42. Haugen E. A., Hait D., Scutelnic V., Xue T., Head-Gordon M., Leone S. R.. Ultrafast X-ray Spectroscopy of Intersystem Crossing in Hexafluoroacetylacetone: Chromophore Photophysics and Spectral Changes in the Face of Electron-Withdrawing Group. J. Phys. Chem. A. 2023;127:634–644. doi: 10.1021/acs.jpca.2c06044. [DOI] [PubMed] [Google Scholar]
  43. Chakraborty P., Couto R. C., List N. H.. Electronic and Inertial Effects of Methylation on Excited-State Hydrogen Transfer. J. Phys. Chem. A. 2026;130:1090–1103. doi: 10.1021/acs.jpca.5c07439. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Sobolewski A. L., Domcke W.. Conical Intersections Induced by Repulsive 1πσ* States in Planar Organic Molecules: Malonaldehyde, Pyrrole and Chlorobenzene as Photochemical Model Systems. Chem. Phys. 2000;259:181–191. doi: 10.1016/S0301-0104(00)00203-2. [DOI] [Google Scholar]
  45. Conical Intersections, Theory, Computation and Experiment. Advanced Series in Physical Chemistry; Domcke, W. , Yarkony, D. R. , Köppel, H. , Eds.; World Scientific Publishing co., 2011; Vol. 17 [Google Scholar]
  46. Matsika S., Krause P.. Nonadiabatic Events and Conical Intersections. Annu. Rev. Phys. Chem. 2011;62:621–643. doi: 10.1146/annurev-physchem-032210-103450. [DOI] [PubMed] [Google Scholar]
  47. Rahn R. O.. Potassium Iodide as a Chemical Actinometer for 254 nm Radiation: Use of Iodate as an Electron Scavenger. Photochem. Photobiol. 1997;66:450–455. doi: 10.1111/j.1751-1097.1997.tb03172.x. [DOI] [Google Scholar]
  48. Goldstein S., Rabani J.. The Ferrioxalate and Iodide-Iodate Actinometers in the UV Region. J. Photochem. Photobiol. 2008;193:50–55. doi: 10.1016/j.jphotochem.2007.06.006. [DOI] [Google Scholar]
  49. Veljković J., Uzelac L., Molčanov K., Mlinarić-Majerski K., Kralj M., Wan P., Basarić N.. Sterically Congested Adamantylnaphthalene Quinone Methides. J. Org. Chem. 2012;77:4596–4610. doi: 10.1021/jo3002479. [DOI] [PubMed] [Google Scholar]
  50. Škalamera Đ., Bohne C., Landgraf S., Basarić N.. Photodeamination Reaction Mechanism in Aminomethyl p-Cresol Derivatives: Different Reactivity of Amines and Ammonium Salts. J. Org. Chem. 2015;80:10817–10828. doi: 10.1021/acs.joc.5b01991. [DOI] [PubMed] [Google Scholar]
  51. Arumugam S., Popik V. V.. Photochemical Generation and the Reactivity of o-Naphthoquinone Methides in Aqueous Solutions. J. Am. Chem. Soc. 2009;131:11892–11899. doi: 10.1021/ja9031924. [DOI] [PubMed] [Google Scholar]
  52. Boens N., Qin W., Basarić N., Hofkens J., Ameloot M., Pouget J., Lefèvre J. P., Valeur B., Gratton E., VandeVen M., Silva N. D., Engelborghs Y., Willaert K., Sillen A., Rumbles G., Phillips D., Visser A. J. W. G., Van Hoek A., Lakowicz J. R., Malak H., Gryczynski I., Szabo A. G., Krajcarski D. T., Tamai N., Miura A.. Fluorescence Lifetime Standards for Time and Frequency Domain Fluorescence Spectroscopy. Anal. Chem. 2007;79:2137–2149. doi: 10.1021/ac062160k. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Sobolewski A. L., Domcke W., Dedonder-Lardeux C., Jouvet C.. Excited-State Hydrogen Detachment and Hydrogen Transfer Driven by Repulsive πσ* States: A New Paradigm for Nonradiative Decay in Aromatic Biomolecules. Phys. Chem. Chem. Phys. 2002;4:1093–1100. doi: 10.1039/b110941n. [DOI] [Google Scholar]
  54. Ashfold M. N. R., Cronin B., Devine A. L., Dixon R. N., Nix M. G. D.. The Role of πσ* Excited States in the Photodissociation of Heteroaromatic Molecules. Science. 2006;312(5780):1637–1640. doi: 10.1126/science.1125436. [DOI] [PubMed] [Google Scholar]
  55. Tully J. C.. Molecular Dynamics with Electronic Transitions. J. Chem. Phys. 1990;93:1061–1071. doi: 10.1063/1.459170. [DOI] [Google Scholar]
  56. McClelland R. A., Kanagasabapathy V. M., Banait N. S., Steenken S.. Reactivities of Triarylmethyl and Diarylmethyl Cations with Azide Ion Investigated by Laser Flash Photolysis. Diffusion-Controlled Reactions. J. Am. Chem. Soc. 1991;113:1009–1014. doi: 10.1021/ja00003a040. [DOI] [Google Scholar]
  57. Osuna S., Swart M., Baerends E. J., Bickelhaupt F. M., Sola M.. Homolytic versus Heterolytic Dissociation of Alkalimetal Halides: The Effect of Microsolvation. ChemPhysChem. 2009;10:2955–2965. doi: 10.1002/cphc.200900480. [DOI] [PubMed] [Google Scholar]
  58. Wang Q., Agarawal V., Hermes M. R., Motta M., Rice J. E., Jones G. O., Gagliardi L.. Distinguishing homolytic vs heterolytic bond dissociation of phenylsulfonium cations with localized active space methods. J. Chem. Phys. 2024;161:014106. doi: 10.1063/5.0215697. [DOI] [PubMed] [Google Scholar]
  59. Ellis-Davies G. C. R.. Reverse Engineering Caged Compounds: Design Principles for their Application in Biology. Angew. Chem. Int. Ed. 2023;62:e202206083. doi: 10.1002/anie.202206083. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Klamt A.. Conductor-like Screening Model for Real Solvents: A New Approach to the Quantitative Calculation of Solvation Phenomena. J. Phys. Chem. 1995;99:2224–2235. doi: 10.1021/j100007a062. [DOI] [Google Scholar]
  61. Klamt A., Jonas V., Bürger T., Lohrenz J. C. W.. Refinement and Parametrization of COSMO-RS. J. Phys. Chem. A. 1998;102:5074–5085. doi: 10.1021/jp980017s. [DOI] [Google Scholar]
  62. Schirmer J.. Beyond the Random-Phase Approximation: A New Approximation Scheme for the Polarization Propagator. Phys. Rev. A:At., Mol., Opt. Phys. 1982;26:2395–2416. doi: 10.1103/PhysRevA.26.2395. [DOI] [Google Scholar]
  63. Trofimov A. B., Schirmer J.. An Efficient Polarization Propagator Approach to Valence Electron Excitation Spectra. J. Phys. B:At., Mol. Opt. Phys. 1995;28:2299–2324. doi: 10.1088/0953-4075/28/12/003. [DOI] [Google Scholar]
  64. Dreuw A., Wormit M.. The Algebraic Diagrammatic Construction Scheme for the Polarization Propagator for the Calculation of Excited States. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2015;5:82–95. doi: 10.1002/wcms.1206. [DOI] [Google Scholar]
  65. Sapunar M., Piteša T., Davidović D., Došlić N.. Highly Efficient Algorithms for CIS Type Excited State Wave Function Overlaps. J. Chem. Theory Comput. 2019;15:3461–3469. doi: 10.1021/acs.jctc.9b00235. [DOI] [PubMed] [Google Scholar]
  66. Coonjobeeharry J., Spinlove K. E., Sanz Sanz C., Sapunar M., Došlić N., Worth G. A.. Mixed-Quantum-Classical or Fully-Quantized Dynamics? A Unified Code to Compare Methods. Philos. Trans. R. Soc., A. 2022;380:20200386. doi: 10.1098/rsta.2020.0386. [DOI] [PubMed] [Google Scholar]
  67. Worth, G. Quantics [Computer software], GitLab, 2025. https://gitlab.com/quantics/quantics (accessed April 07, 2026).

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ja5c21621_si_001.pdf (6.1MB, pdf)

Articles from Journal of the American Chemical Society are provided here courtesy of American Chemical Society

RESOURCES