The Influence of β-Ammonium Substitution on the Reaction Kinetics of Aminooxy Condensations with Aldehydes and Ketones

Abstract

The click-chemistry capture of volatile aldehydes and ketones by ammonium aminooxy compounds has proven to be an efficient means of analyzing the carbonyl subset in complex mixtures, such as exhaled breath or environmental air. In this work, we examine the carbonyl condensation reaction kinetics of three aminooxy compounds with varying β-ammonium ion substitution using Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). We determined the activation energies for the reactions of the aminooxy compounds ATM, ADMH and AMAH with a panel of ketones and aldehydes that included acrolein and crotonaldehyde. The measurements indicate that the activation energies for the oximation reactions are quite low, less than 75 kJ/mol. ADMH is observed to react the fastest with the carbonyls studied. We postulate this result may be attributed to the ADMH ammonium proton effecting a Brønsted-Lowry acid-catalyzed elimination of water during the rate-determining step of oxime ether formation. A theoretical study of oxime ether formation is presented to explain the enhanced reactivity of ADMH relative to the tetraalkylammonium analog ATM.

Graphical Abstract

INTRODUCTION

Carbonyl compounds are of significant interest as biological markers of various disease states. For example, some aldehydes, such as malondialdehyde, acrolein, n-hexanal, n-heptanal, n-nonanal, 4-hydroxy-2-hexenal and 4-hydroxy-2-nonenal, are by-products of lipid peroxidation and are being used to assess and monitor oxidative stress.^[1] Levels of certain aldehydes and ketones have been found to be elevated in the brains of preclinical Alzheimer’s Disease (AD) patients and in brain areas susceptible to late stage AD and thus have been reported as AD biomarkers.^[2] The presence of carbonyl compounds in exhaled breath, such as hydroxy-acetaldehdye and 2-butanone, has been linked to lung cancer.^[3] In terms of ex vivo interest, toxic aldehydes and ketones are prevalent in cigarette smoke and electronic cigarette aerosols.^[4]

Monitoring atmospheric carbonyls is also of great interest. These are derived from a variety of sources including direct emissions from motor vehicles or other combustion sources^[5] and from photochemical reactions of precursor molecules with atmospheric ozone.^[6] Carbonyl compounds play a central role in atmospheric chemistry close to the tropopause, and this is directly relevant to the issue of ozone depletion.^[7] Thus, it is crucial to understand the reactions of carbonyl compounds, particularly with respect to carbonyl derivative formation for purposes of monitoring or analysis.

The condensation reaction between an aminooxy moiety (RONH₂) and the carbonyl group of an aldehyde or ketone — known as an oximation reaction — is a versatile ‘click chemistry’ reaction^[8] that generates a robust oxime ether linkage. Oxime ether formation is central in many ligations of proteins with polysaccharides^[9] as well as for the analytical derivatization of oxidized cellular metabolites.^[10] Indeed, oximation has shown much promise as a tool for in situ ligation and application in living organisms.^[11]

In spite of the importance of oximation reactions in the analysis and quantification of trace carbonyl compounds in environmental air, exhaled breath and other biological samples, few studies have examined the influence of structural features in both aminooxy compounds and carbonyls on oximation kinetics.^[12] The effect of carbonyl compound structural features on the reaction rate of hydrazone formation, a process analogous to oximation, has recently been studied.^[13],[14] We report herein our findings on the reaction kinetics of three quaternary ammonium aminooxy compounds reacting with a panel of ketones and aldehydes. Specifically, we studied the kinetics of 2-(aminooxy)-N, N, N-trimethylethan-1-ammonium iodide (ATM), 2-(aminooxy)-N, N-dimethylethan-1-ammonium chloride (ADMH), and 4-(2-(aminooxy)ethyl)-morpholin-4-ium chloride (AMAH) in reactions with three ketones (acetone, methyl isobutyl ketone, and 2-heptanone) and three aldehydes (propanal, acrolein, and crotonaldehyde) to understand the effects of structural features on rates of oximation (Scheme 1). When reacted with carbonyls, these positively charged aminooxy compounds enable soft ionization, such as electrospray ionization, of the resultant carbonyl adducts, which is a requirement for many different types of mass spectra analysis.^[15] The kinetics of oximation studied here were carried out by direct mass spectrometric measurements using Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). This mass spectrometry procedure offers the advantage of measuring fast reaction rates that are otherwise difficult or impossible to monitor using other spectrometric techniques. We also disclose the results of a theoretical study of oxime ether formation by ADMH relative to ATM using unrestricted density functional theory (UDFT) with B3LYP hybrid exchange-correlation functional^[16] and 6-311++G(d,p) basis set^[17] to probe the role of the β-ammonium ion as a means to rationalize the observed difference in oximation rates.

Scheme 1.

Aminooxy panel and oximation of aldehydes or ketones (R = ammonium moiety; ¹R, ²R = alkyl group).

EXPERIMENTAL SECTION

Materials

All reagents and solvents, including deuterated acetone (acetone-d₆), acetone, methyl isobutyl ketone, 2-heptanone, propanal, acrolein, crotonaldehyde and methanol, were purchased from Sigma-Aldrich. ATM (1, Scheme 1),^[18] and AMAH (2)^[19] were prepared according to literature procedures. ADMH was synthesized in three steps as follows (Scheme 2): (1) reaction of ethanolamine with N-hydroxyphthalimide (NHP) under standard conditions^[20] to obtain phthaloyloxy amine 6; (2) hydrazinolyis by treatment of 6 with methylhydrazine at 0 °C in dry dichloromethane followed by Kugelrohr distillation of the deprotected aminooxy product; and (3) acidification using aqueous hydrochloric acid. Recrystallization of the crude salt from isopropyl alcohol afforded ADMH as a white solid.

Scheme 2.

Synthesis of ADMH. Reagents and conditions: a. N-hydroxy-phthalimide, PPh₃, DIAD, THF, 0 °C to rt, 12h; b. CH₃NHNH₂, CH₂Cl₂, 0 °C, 4.5h; c. 6M HCl, reflux, 22h.

Measurement of reaction kinetics

The respective quaternary ammonium aminooxy reagent ATM, ADMH or AMAH (5.0 ×10⁻⁷ mol) was added to 200 μL spectroscopic grade methanol to form a 2.5×10⁻³ mol/L solution. Acetone (5.0 ×10⁻⁷ mol), or other carbonyl compound, then was added to the solution, either at 21 °C, 0 °C, or −21 °C. The corresponding acetone-d₆ adduct of ATM, ADMH or AMAH (5.0 ×10⁻⁷ mol) was added to each mixture to serve as an isotopically labeled internal standard for quantification of the formed aminooxy-carbonyl adduct by FT-ICR-MS. The temperature range of −21 °C to 21 °C was used for kinetics measurement because aminooxy compounds are used for derivatization of carbonyl compounds in biofluids at temperatures in the range of room temperature to −78 °C.^[21] A mixture of 90% ethylene glycol and 10% ethanol with dry ice in a thermally insulated container was used to achieve −21 °C. The 2 mL vials containing the reaction solutions were placed in this dry ice bath. Aliquots of the reaction mixture (15 μL) were analyzed by FT-ICR-MS at different time intervals. The unreacted carbonyl compound concentration C(t) was calculated by subtraction of the reacted carbonyl compound concentration from the initial concentration C_o. All kinetic studies were done following this procedure.

FT-ICR-MS analyses

The methanol solutions of the aminooxy reagent–carbonyl mixtures were analyzed on a hybrid linear ion trap FT-ICR-MS instrument (Finnigan LTQ-FT, Thermo Electron, Bremen, Germany) equipped with a TriVersa NanoMate ion source (Advion BioSciences, Ithaca, NY) with an electrospray chip (nozzle inner diameter 5.5 μm). The TriVersa NanoMate was operated in positive ion mode by applying 2.0 kV with no head pressure. Initially, low-resolution MS scans were acquired for 1 min to ensure the stability of ionization, after which high mass accuracy data were collected using the FT-ICR analyzer where MS scans were acquired for 8.5 min and at the target mass resolution of 100,000 at m/z of 800. The quaternary aminooxy compound and its adducts were assigned on the basis of their accurate mass by first applying a small (typically <0.0005) linear correction based on the observed mass of the internal standard.

Computational study

To compare with experiment, four ATM and ADMH adducts were studied that differ in the R₁ and R₂ side chains (Figure 1). The adduct structures correspond to the carbonyls used in the oximation reactions: acetone (R₁=CH₃, R₂=CH₃), propanal (R₁=C₂H₅, R₂=H), 2-heptanone (R₁= C₅H₁₁, R₂= CH₃), and acrolein (R₁=H, R₂=C₂H₃). Low energy structures of ATM and ADMH with R₁ = H and R₂ = H were determined first from conformational searches with the MMFF94 force field^[22] using GMMX.^[23] The set of initial candidate structures then were optimized to a minimum with B3LYP/6-31G in order to identify relevant low energy conformations which subsequently were explored with the 6-311+G(d,p) basis set. Transition states were calculated for conformational changes between relevant intermediates and for oxime ether formation. These transition states were confirmed with frequency and IRC calculations. Structures were obtained using both the isolated molecule in gas phase, and in the presence of an explicit methanol molecule.

Figure 1.

ADMH and ATM molecules indicating backbone atom labels referenced throughout the text. Side chains R₁ and R₂ are adduct dependent.

RESULTS AND DISCUSSION

FT-ICR-MS provides the highest accuracy and mass-resolving power.^[24] For example, FT-ICR-MS has been applied for direct determination of kinetic constants of phosphoglucomutase and its phosphorylated products,^[25] various enzymatic kinetics,^[26] and to determine the kinetics and exchange thermodynamics of dimethylamine for ammonia in ammonium methanesulfonate clusters.^[27]

We initially assumed the reaction kinetics of the aminooxy reagents reacting with carbonyl compounds to be an elementary first order reaction for both reactants as well as an irreversible reaction. For equimolar aminooxy reagent and carbonyl compound, the reaction rate is given by the following equation

$$r=\frac{d{C}_{carbonyl}}{dt}=-k{C}_{Aminooxy}{C}_{carbonyl}=-k{C}_{carbonyl}^2$$ (1)

where k is the specific reaction rate or rate constant, C is the concentration of the reactants, t is the reaction time. The following equation can be obtained by integration of equation (1) from time zero to t for the concentration of carbonyl compounds from initial C_o to C(t)

$$\frac{1}{C(t)}-\frac{1}{C_0}=kt$$ (2)

To verify that the reaction is an irreversible elementary first order reaction for both reactants, plots of 1/C vs time were made for all reactions. Figure 2 shows representative experimental data points of 1/C vs t for AMAH reacted with acetone at 21 °C, 0 °C, and −21 °C, respectively. For representative FT-ICR-MS spectra of this process, see the spectra overlay of the AMAH reaction with acetone at different times at 21 °C (Figure S1, Supporting Information). Linear regression with high R² (>98%) were used to obtain the reaction rate constant k in equation (2) for all three aminooxy compounds reacted with the carbonyl compounds. Therefore, the oximation reactions can be assigned as overall second order reactions. The corresponding values of all rate constants k at 21 °C, 0 °C, and −21 °C were obtained from the slope of the linear plot of 1/C vs t.

Figure 2.

The relationship between 1/C and oximation reaction time between AMAH and acetone at the given temperature.

After the rate constant k was determined at three different temperatures, the activation energy and frequency factor were determined using the Arrhenius equation,

$$\mathrm{k}={\mathrm{k}}_{\circ }\text{exp}(-{\mathrm{E}}_{\mathrm{a}}∕\text{RT})$$ (3)

$$lnk=-\left(\frac{E_a}{R}\right)\cdot \left(\frac{1}{T}\right)+ln{k}_o$$ (4)

The activation energy E_a and frequency factor k_o were determined from linear regression of the plots of the natural logarithm of ln k vs. the reciprocal of temperature 1/T for AMAH, ADMH, and ATM reactions with three ketones (acetone, methyl isobutyl ketone, and 2 heptanone) and three aldehydes (propanal, acrolein and crotonaldehyde) (Figure 3).^[24] Table 1 lists calculated activation energy E_a and frequency factor k_o for the reactions of AMAH, ADMH and ATM with these ketones and aldehydes. In general, ADMH has lower activation energies followed by AMAH, then ATM for reactions with these carbonyl compounds. It is intriguing to speculate that the lower activation energies of ADMH and AMAH relative to ATM in all cases (except for 2-heptanone where ATM activation energy is 0.1 kJ mol⁻¹ lower than AMAH) might be attributed to the proximal, acidic ammonium proton, which may function as a catalyst for oxime ether formation. Indeed, elegant studies by Kool et al.^[13],[28] demonstrated that rate enhancements ensue when intramolecular protonation of analogous hydrazine-carbonyl adducts can proceed to speed breakdown of corresponding tetrahedral intermediates leading to hydrazone products. Since ATM has no such acidic proton, intramolecular protonation is not possible. The activation energies of AMAH are slightly higher than that of ADMH likely because the morpholine-based compound is sterically more demanding than ADMH. Analogously, the activation energies of the We first examined the relevant initial low-energy structures increasing carbon number. Of all the carbonyl compounds under study, propanal has the lowest activation energy with all three aminooxy compounds. Saturated aldehydes are generally more reactive than ketones. The propanal reactions with the three aminooxy compounds also have a lower frequency factor k_o in comparison with the corresponding acetone reactions, which lowers the rate constants of propanal relative to acetone for reactions with the same aminooxy compound at 21 °C (Table 2).

Figure 3.

Plots of the natural logarithms of the oximation reaction rate constants vs. the reciprocal of reaction temperatures for aminooxy reagents reacting with carbonyls: A, acetone; B, propanal; C, methyl isobutyl ketone; D, 2-heptanone; E, acrolein; F, crotonaldehyde.

Table 1.

Activation energies and frequency factors for AMAH, ADMH and ATM reactions with acetone, propanal, 2-heptanone, methyl isobutyl ketone (MIBK), acrolein, and crotonaldehyde.

ADMH
Ea (kJmol−1)	32.2	29.2	33.2	33.1	45.7	49.8
Ko (M−1·s−1)	3.84 ×105	6.54 ×104	8.56 ×104	1.64 ×105	2.58 ×108	6.20 ×108
AMAH
Ea (kJmol−1)	35.4	30.7	37.3	34.4	48.2	54.2
Ko (M−1·s−1)	1.78 ×106	2.28 ×105	5.04 ×106	3.64 ×106	4.66.×108	4.86 ×109
ATM
Ea (kJmol−1)	36.8	34.5	37.2	37.0	51.9	74.0
Ko (M−1·s−1)	1.95 ×107	4.92 ×105	5.62 ×106	5.72 ×105	2.92 ×109	7.02 ×1012

Table 2.

Rate constants k (M⁻¹·s⁻¹) for reactions of aminooxy panel with carbonyls at 21 °C.

	acetone	propanal	2-heptanone	MIBK	acrolein	crotonaldehyde
ADMH	0.73	0.42	0.11	0.22	2.00	0.88
AMAH	0.91	0.79	1.20	2.84	1.27	1.14
ATM	5.64	0.36	1.38	0.15	1.72	0.50

The rate constant k of the acetone reactions with ATM and ADMH is higher than those of the MIBK and 2-heptanone reactions for the same aminooxy compound because of lower activation energies and higher frequency factor k_o (Table 1). The frequency factor k_o of ketone reactions with ATM and ADMH decreases with increasing number of ketone carbons because of the slower molecular motion of larger-sized ketones with larger molecular weight as well as steric effects, especially with the iso-butyl group of MIBK. It is interesting to note that although the activation energy of ATM reactions with the carbonyl compounds is higher than those of ADMH, the frequency factor k_o of ATM reactions are also consistently higher than those of ADMH. The contribution of the frequency factor k_o makes the rate constant k of ATM reactions with acetone and 2-heptanone even higher than those of ADMH and AMAH, but not for the ATM reaction with MIBK.

Table 1 shows that the activation energies of the acrolein and crotonaldehyde reactions with the three aminooxy compounds are generally higher than those of the saturated aldehydes and ketones. Acrolein and crotonaldehyde, as α,β-unsaturated aldehydes, have decreased electrophilicity at the carbonyl carbon relative to saturated aldehydes^[29] as well as offer another mode for nucleophilic addition to occur, namely conjugate addition, which may account in part for the large difference in frequency factor between acrolein/crotonaldehyde and the other carbonyl compounds. The activation energies of ADMH and AMAH with the α,β-unsaturated aldehydes are closer and considerably lower than those of ATM. In fact, the E_a of ATM with crotonaldehyde is 1.5 times higher than that of ADMH. ADMH and AMAH are therefore more reactive towards acrolein and crotonaldehyde than ATM. In comparison with propanal, acrolein has a much higher frequency factor k_o for reaction with the aminooxy compounds. Therefore, even though the activation energies of the acrolein reactions are higher than those of propanal, the much higher frequency factor k_o increases the rate constants of acrolein relative to propanal for a given aminooxy compound. Table 2 lists the rate constants k for reactions of the aminooxy compounds with the carbonyl panel at 21 °C. The rate constant k of crotonaldehyde is lower than that of acrolein for reaction with the same aminooxy compound because of the associated higher activation energy, given in Table 1. Nevertheless, it is surprising that the frequency factor k_o of crotonaldehyde is also higher than that of propanal and acrolein for reactions with the three aminooxy compounds.

Molecular Modeling

Oxime condensations proceed through a tetrahedral hemiaminal intermediate followed by a rate-limiting dehydration, as originally detailed by Jencks.^[30] To probe whether the proximal ammonium hydrogen of ADMH (or AMAH) can influence the rate of oxime ether formation relative to ATM, we conducted a computational study comparing ADMH and ATM. Specifically, we postulated that intramolecular proton transfer at the tetrahedral intermediate stage, such as postulated with hemiaminal intermediates 7a-c (Scheme 3), might accelerate the loss of water relative to an intermediate in which the β-ammonium ion has no hydrogen available, such as when using ATM. We first examined the relevant initial low-energy structures of ATM and ADMH and the consequences for the oxime ether formation mechanism. There are several hydrogen bond donor sites (Figure 1: N₅, O₇) and acceptor sites (N₅, O_4, O₇) common to both molecules. The main difference between ADMH and ATM is the presence of the hydrogen bond acceptor site on N₁, which is only present in ADMH adducts. Initial optimization of ADMH and ATM molecules revealed the N₁ site was a key difference, as all low-energy conformers of ATM displayed an elongated chain, while all low-energy conformers of ADMH involved intramolecular hydrogen bonding between one of the three hydrogen bond donor sites and the N₁ hydrogen bond acceptor.

Scheme 3.

Postulated intramolecular H-bonding of ADMH-carbonyl adducts.

Table 3 shows the relative energies of the lowest energy minima of ADMH hemiaminal where R₁=H and R₂=H, while the corresponding structures are shown in Figure 4, A-D. The ADMH global minimum (Fig. 4, structure A) was found to contain a hydrogen bond involving O₇ and N₁. A second low-energy structure involving the same O₇-N₁ hydrogen bond was found to lie 15.9 kJ mol⁻¹ above structure A (Fig. 4, structure B), related to structure A through torsion of the C₂-C₃-O₄-N₅ backbone dihedral. While structure A (postulated as 7c, Scheme 3) is energetically the most favorable, the folding of the backbone from the extended structure involves formation of an eight-member ring and thus is likely to be entropically unfavorable. Conversely, the lowest energy structure involving a hydrogen bond between N₁ and O₄ involves a 5-member ring (Fig. 4, structure D) and is 47.9 kJ mol⁻¹ above the global minimum. Thus, structure D is unlikely to play a role in defining ADMH reactivity. Structure C is the lowest energy conformation involving a hydrogen bond between N₅ and N₁ and is just 11.5 kJ mol⁻¹ above structure A. In fact, structure C (postulated as 7a, Scheme 3) is the third lowest energy structure overall and, as hydrogen bonding involves formation of a six-member ring, is likely to be entropically favored. Thus, even though structure A is the global minimum, we conclude that structure C is likely to play a key role in activating ADMH for water dissociation.

Table 3.

Lowest energy structures of ADMH adduct where R₁=H and R₂=H relative to the global minimum structure A in gas phase (kJ mol⁻¹).

Structure	Hydrogen Bond	ΔE (kJ mol−1)
A	N1─O7	0.00
B	N1─O7	15.9
C	N1─N5	11.5
D	N1─O4	47.9

Figure 4.

Lowest energy ADMH structures where R₁=H and R₂=H. A: global minimum; B: intermediate in path between A and C; C: lowest energy N₁-N₅ hydrogen-bonded structure; D: lowest-energy N₁-O₄ hydrogen-bonded structure; E: lowest energy ATM structure. Dashed lines indicate hydrogen bonds.

To further analyze the role of structure C in activating ADMH, the energy barriers for conversion to the O₇-N₁ hydrogen-bonded structure A were investigated. We found that although structure C cannot access structure A directly, it could do so through two low energy barriers via intermediate structure B. The transition structure connecting C to B involves a shift of the hydrogen bond donor from N₅ to O₇ with a 3.5 kJ mol⁻¹ barrier. The transition structure is 0.04 kJ mol⁻¹ higher in energy than structure B, and so the reverse reaction from structure B to C is effectively barrierless (especially when accounting for zero point energy). The transition from structure B to A is accessed through dihedral torsion of C₂-C₃-O₄-N₅, with activation barrier energy of 3.4 kJ mol^-1. Thus, structure A can be accessed from structure C, and the resulting orientation of the molecule is such that water dissociation to form the oxime ether only requires bond order changes without significant nuclear motion.

In the case of ATM, conformational studies revealed that all low-lying structures had an extended backbone chain with different conformations related to each other through backbone dihedral torsions. There are no intramolecular interactions that serve to increase the rate of oxime ether formation like those observed in the case of ADMH. Studies of the mechanism used the lowest energy conformation (Fig. 4, structure E). A key feature of the lowest energy conformation was that the lone pair on N₅ was orientated towards the hydroxide group containing O₇. Thus, any unimolecular reaction in which the hydroxide leaves with the N₅ proton would require inversion of the N₅ center. Even though a unimolecular reaction mechanism is not feasible, ATM oxime ether formation requires inversion of N₅, as discussed below.

Table 4 shows the computed activation barriers of ATM and ADMH. Initial studies determined the unimolecular pathway for oxime ether formation in gas phase (Table 4, columns 2 and 3). The ADMH barriers in gas phase are all indicative of a relatively facile reaction, with barriers of 92.3-97.8 kJ mol⁻¹ depending on the R₁ and R₂ groups (Table 4, column 2). The transition structures display the same nuclear configuration regardless of the specific R₁ and R₂ groups. Figure 5 shows the transition structures where R₁ = CH₃ and R₂ = CH₃, along with the normal mode vectors associated with the imaginary frequency. Figure 5, structure F demonstrates that the gas-phase oxime-ether transition vector involves the C₆-O₇ bond stretching and N₁-H-O₇ proton transfer to result in water dissociation, corroborating the role of the N₁ hydrogen in ADMH.

Table 4.

Computed dissociation barriers (kJ mol⁻¹) for four different ATM (left) and ADMH (right) adducts in the gas phase.

	ΔE (kJ mol−1) in gas phase		ΔE (kJ mol−1) with explicit methanol
	ADMH	ATM	ADMH	ATM
2-heptanone	97.8	187.3	70.2	120.3
acetone	94.2	192.4	69.2	119.0
propanal	92.4	215.0	83.0	137.9
acrolein	92.3	196.6	69.8	122.9

Figure 5.

Oxime ether formation transition structures where R₁ = R₂ = CH₃ of ADMH, F: in gas phase, G: with explicit methanol; and ATM, H: in gas phase, I: with explicit methanol.

In contrast to ADMH barriers, unimolecular ATM barriers are significantly larger (Table 4, column 3). Exploration of possible unimolecular oxime ether formation pathways in ATM revealed that a proton is removed concertedly with hydroxide dissociation, forming water as the leaving group. While transition states involving several proton sources were located, the final deprotonation site is N₅. The transition state in the direct N₅ deprotonation pathway involves inversion of the N₅ center to form a four membered ring consisting of N₅-H-O₇-C₆, from which the oxime ether product directly results. Figure 5, structure H shows the transition structure and associated normal mode of the gas-phase dissociation in ATM, demonstrating the late transition barrier where the C₆-O₇ bond has already broken, but proton transfer has not yet occurred. The large barriers (187.3-215.0 kJ mol⁻¹) suggest that ATM oxime ether formation does not occur as an intramolecular process and must involve at least a second molecule in an intermolecular process.

Possible candidates for inclusion in intermolecular reaction mechanisms involved in the experimental rate measurements included water, ethanol, and methanol, all of which are likely to result in similar activation barriers, as the key moiety is the hydroxyl group. To examine the intermolecular pathways, transition structures were recomputed in the presence of an explicit methanol molecule (Table 4, columns 4 and 5). Several positions for the methanol molecule interacting with ATM were explored with the lowest identified arrangement forming a hydrogen-bond donor with N₅ and a hydrogen-bond acceptor with O₇ (Fig. 5, structure I). In ATM, the transition barriers were reduced by 67.0-77.1 kJ mol⁻¹ when explicit methanol was included (table 4, column 5). The ATM mechanism in the presence of methanol is similar to the unimolecular process with the required inversion of the N₅ center. However, in the presence of methanol, two protons are transferred concertedly with hydroxide dissociation, through a six-membered ring formed of the hydrogen-bonding network between the secondary amine (N₅) and the methanol hydroxide, and between the methanol hydroxide and ATM hydroxide (O₇).

Similarly to ATM, ADMH transition states were re-optimized in the presence of methanol, maintaining the same hydrogen-bonding pattern as in ATM. ADMH barriers were reduced by 9.4-27.6 kJ mol⁻¹ (table 2, column 4). The lowest energy ADMH structures in the presence of explicit methanol do not significantly change the ADMH backbone structure (Fig. 5, structure G). As can be seen by the difference between F and G, the presence of methanol has a small effect on the transition vector, indicating inclusion of explicit methanol is not required for describing the transition structure of ADMH.

Comparing the calculated activation barriers with those from experiment, both approaches are in agreement that ADMH barriers are lower than those for ATM molecules. Additionally, the order of the energy differences between the different R₁ and R₂ groups was the same in both experiment and theory, with molecules with ketone derived side chains having a smaller ΔE than those from aldehydes. However, given that the experimental ΔE values all lie within 2.2 kJ mol⁻¹ of each other, the ordering is within the expected error range of the theoretical method. In fact, both computational and experimental results ultimately indicate that ΔE values are similar regardless of the particular side chain, in agreement with the computed transition structures and reaction pathways which indicate the particular R₁ and R₂ side chains do not play a significant role in in defining reactivity. Despite the aforementioned agreement between experiment and theory, the difference in computed activation energy between ADMH and ATM were an order of magnitude greater in computational analysis than ascertained experimentally, corresponding to around 45 kJ mol^-1. While such discrepancy is at the limit of typical errors in the computational methodology, the error is substantially larger than the experimental difference. The effect of solvent polarization of transition structures was ruled out through IEFPCM implicit methanol solvent model calculations,^[31] which showed very similar response in both ATM and ADMH.

Conclusion

The kinetic study between three quaternary ammonium aminooxy compounds and select ketones and aldehydes indicate the effects of structural features of aminooxy compounds and carbonyl compounds on the reaction rate. The activation energies for these oximation reactions are quite low, less than 75 kJ/mol. Of the panel of aminooxy compounds, ADMH has a lower activation energy for reactions with the carbonyls studied. We postulate the presence of a β-ammonium proton may enhance the key dehydration step from resultant hemiaminal intermediates to afford the oxime ether adducts. Molecular modeling supports this possibility in that a structure featuring intramolecular hydrogen bonding in an eight-member configuration was found to be the global minimum energy conformation, likely arising from a kinetically favored six-member hydrogen bonding arrangement. These results suggest that incorporation of structural features capable of enhancing the dehydration of hemiaminal intermediates will increase oximation efficiency and may serve as a useful guide for development of more effective aminooxy reagents for analytical derivatization or conjugation purposes.