Vibrational spectroscopy of the cryogenically cooled O- and N-protomers of 4-Aminobenzoic acid: Tag effects, isotopic labels, and identification of the E,Z isomer of the O-protomer

Abstract

4-Aminobenzoic acid (4ABA) is a biologically relevant, small organic molecule with two protonation sites: the amino group (N-protomer) and the carboxyl group (O-protomer). The O-protomer is energetically preferred in the gas-phase, while the higher energy N-protomer can be trapped using aprotic solvents such as acetonitrile during electrospray ionization. Here, we focus on the structure of the O-protomer, which can occur in three low-lying isomeric forms that result from different orientations of the OH groups relative to the benzene ring. We report the vibrational spectra of both N- and O-protomers of the cryogenically cooled ions in the gas phase over the spectral range 800–4000 cm⁻¹. The bands arising from the OH stretches are isolated from the nearby NH stretching fundamentals using isotopic labeling as well as by analysis of the shifts in these fundamentals upon attachment of D₂ and N₂ molecules to the OH groups of the O-protomer. The spectra of isomers derived from the different locations of the adducts were isolated using two-color, IR-IR photofragmentation spectroscopy. The docking motifs by which the O-protomer binds to another 4ABA molecule is also explored and found to feature a bifurcated arrangement involving attachment of both OH groups of the protonated head group to the carbonyl group of the neutral partner.

Graphical Abstract

1. Introduction

4-Aminobenzoic acid (para-aminobenzoic acid, 4ABA) is a versatile molecule [1–4] that is useful in the syntheses of a diverse array of compounds with practical applications [2, 5–9]. Here we are concerned with the protonated acid, H⁺−4ABA, a species that occurs in two isomeric forms according to whether protonation occurs at the amine or carboxyl groups, hereafter denoted the N- and O-protomers, with the calculated structures indicated in Fig. 1. Interestingly, the relative energies of the two protomers is strongly solvent dependent such that the O-protomer is 30–40 kJ/mol more stable in isolation (i.e., the gas phase), but the N-protomer is ~20 kJ/mol more stable in water. [10–12]. This scenario raises interesting possibilities when the cations are introduced into the gas phase using electrospray ionization [12–15]. In particular, spraying from acetonitrile (ACN) solutions yields the higher energy N-protomer, while spraying in water yields the O-protomer. Although both of these protomers have been characterized by vibrational spectroscopy and ion mobility [12, 13, 16], the spectral coverage of the IR studies is incomplete and the measurements were carried out on ions at relatively high temperature. Here and specifically address the structure of the protonated carboxylic head group of the O-protomer, which could occur in three rotameric forms (Fig.1A–C). Recently, Davies et al. [17] reported infrared spectra of protonated acetic acid in He nanodroplets, where they isolated the OH features of the E- and Z-OH and showed that the E,Z-rotamer is the only rotamer present in the gas phase. Here, the spectra are obtained using the tagging technique [18, 19], hence the various bands are found to shift according to the location of the tag, X = D₂ and N₂. We use this effect to our advantage by sorting out the isomers that differ according to the tag position using two-color IR-IR photobleaching and following the evolution of the band shifts displayed when the interaction is systematically modified in the series H⁺−4ABA•Xn=1,2, X = D₂ and N₂. This information, complemented by spectra of the various isotopologues arising from H/D substitution, yield an unambiguous assignment of the NH and OH stretches to one of the rotameric forms of H⁺−4ABA. Finally, we consider how the O-protomer binds to strong H-bond acceptors by comparing the spectrum of 4ABA binary complex with calculated patterns to establish the possible binding motifs.

Figure 1.

Geometry optimized structures of the possible rotamers of both the O- and N- protomers of H⁺−4ABA calculated at the B3LYP/aug-cc-pVTZ level of theory. ZPE corrected relative energies are displayed beneath the structures.

2. Experimental and Computational Methods

A detailed description of our method has been published in references [19, 20]. Briefly, 10 mM solutions of 4ABA powder (obtained from Sigma Aldrich with no further purification) dissolved in a (pH ~2) solution of formic acid and H₂O, D₂O, or 1:1 ACN:H₂O underwent electrospray ionization (ESI), and the resulting ions were then transported to a 3D copper Paul trap using octupole radiofrequency ion guides. In the isotope substitution study, D₂O doped with H₂O was placed in the capsule for isotope exchange. The ions were then cooled to 15 K and “tagged” with He buffer gas doped with 25% X (X = D₂, N₂). The X_n=1,2-tagged species were then transferred to a double focusing tandem time-of-flight photofragmentation mass spectrometer described elsewhere [20]. Infrared radiation in the 600–4000 cm⁻¹ region was generated by a LaserVision OPO/OPA. Resonant excitation of the X_n=1,2-tagged ion results in the evaporation of weakly bound tag molecule(s), creating a photofragment ion, which is monitored continuously to generate the vibrational predissociation spectrum. For the isomer selective study, a second laser was used to perform double-resonance spectroscopy: a “probe” laser was fixed on a particular probe transition while another “bleach” IR laser intersects the same ion packet before it interacts with the probe laser. As the pump laser removes population from the species selected by the probe frequency, the photofragment yield from the probe laser appears as a series of dips to unveil the isomer-selective spectrum. Minimum energy structures were calculated using the Gaussian 09 program package [21] with DFT/B3LYP method. Calculations of involving the monomer of H⁺4ABA (with and without D₂ tag) were performed with the aug-cc-pVTZ basis set. The dimer, H⁺(4ABA)₂, were calculated with the 6–311++G(d,p) basis set.

3. Results and Discussion

3.1. IRPD spectra of the D₂-tagged N- and O-protomers of H⁺−4ABA

The spectra of the N- and O- protomers are easily isolated using the fact that the gas phase O-generated by spraying 4ABA dissolved in water, while the N-protomer is favored when it is sprayed in a mixed solvent consisting of acetonitrile and water (1:1 volume ratio) [11–16]. Figure 2 compares the D₂-tagged spectra of the ions generated from ACN:H₂O (Fig. 2A) and the formic acid solution (Fig. 2B). These spectra are indeed completely different and readily assigned to the expected protomers by cursory inspection of the band patterns. Specifically, the spectrum in Fig. 2A contains the carbonyl stretch at 1791 cm⁻¹ that is unique to the N-protomer [12], as well as a singular OH stretch in the vicinity expected for a nonbonded OH group of carboxylic acid [27, 28]. The −NH₃⁺ motif is also clear in the 3200–3400 cm⁻¹ region as observed in many protonated amino compounds [22–25], indicated by the blue arrow in Figure 2A. Thus, the spectrum in Fig. 2A is unambiguously assigned to the N-protomer. The fact that the spectrum in Fig. 2B does not contain the C=O stretch, on the other hand, indicates that none of the N-protomer is prepared with the ACN solution, and indeed the −NH₃⁺ signature is also missing while the OH stretching region is more complex as expected with two −COH groups on the protonated acid group. Consequently, the spectrum in Fig. 2B is that of the O-protomer. We note that, although these messenger-tag predissociation spectra are better resolved and more complete, parts of these spectra were reported earlier and are consistent with those obtained in this work. In particular, reported IRMPD spectra for the N-protomer in the NH/OH stretching region only show one peak in the NH₃ stretching region [16, 29] where three peaks were expected, possibly due to IRMPD transparency [30], which could suppress absorptions. For the O-protomer, peak assignments suffer from the closely spaced NH₂ and OH stretches as well as coupling-induced splitting on the NH₂ group. For example, the NH₂ antisymmetric stretch has been reported to be very close in energy with the OH stretches of the O-protomer [13, 16, 29], a typical scenario for systems with these functionalities [19, 23, 24, 31, 32]. Additionally, reported calculated spectra of the three O-rotamers showed that the OH stretches are close in energy [13]. Consequently, the isomeric composition of the O-protomer generated in the gas-phase is unresolved.

Figure 2.

Spectrum of H⁺−4ABA•D₂ sprayed in H₂O:ACN (A) vs. in H₂O (B). Peaks assignments for trace A and trace B (fingerprint region) are in good agreement with electronic structure calculations (Fig. S1) and is colored accordingly. The presence of the carbonyl stretch at 1791 cm⁻¹ in trace A unambiguously signifies the presence of the N-protomer. The NH and OH stretching region of trace B remains unassigned. Blue arrow indicates the −NH₃⁺ stretching region for various protonated amino compounds in Ref. [22–25]. Calculated frequency are scaled by a factor of 0.968 according to the CCCBDB [26].

3.2. Assignment of the OH and NH stretches in the O-protomer through isotopic substitution and tag perturbations

To separate contributions from NH versus OH stretches in the O-protomer spectrum, we performed isotopic substitution, specifically involving incorporation of a single H atom into the otherwise perdeuterated H⁺−4ABA scaffold to quench coupling between the two NH groups. Note that in the absence of isotope fractionation, the single H atom can occupy two or three distinct sites depending on which rotamer(s) are present. The NH₂ symmetric and antisymmetric stretches are expected to collapse into a single peak, whereas the distant OH stretches, which are only weakly coupled, are expected to remain largely intact. The isotopically labeled spectrum is shown in Figure 3A. Comparison with the spectrum of the all H isotopologue Figure 3B reveals the emergence of a new peak at 3495 cm⁻¹ (blue) while the two peaks at 3441 cm⁻¹ and 3542 cm⁻¹ in the single H isotopologue (blue in Fig. 3B) are missing. Three peaks (red and orange) are persistent and therefore assigned to the OH stretches. The 3495 cm⁻¹ feature falls almost exactly at the midpoint of the two missing bands and is therefore assigned to the uncoupled NH stretch of the NHD moiety. This procedure then establishes the v^sym and v^asym contributions of the NH₂ group to the spectrum of the all H isotopologue (Fig. 3B) as blue features in Fig. 3B (3441 cm⁻¹ and 3542 cm⁻¹, respectively).

Figure 3.

A) Isotopically labeled vibrational spectrum of H⁺−4ABA•D₂ (which occurs as the 4-amino-d₂-benzoic acid-d₁ and 4-amino-d₁-benzoic acid-d₂ isotopomers). Predissociation spectra of B) H⁺−4ABA•D₂, C) H⁺−4ABA•(D₂)₂, D) H⁺−4ABA•N₂, and E) H⁺−4ABA•(N₂)₂. Blue peaks represent NH stretches, red for bound OH, and orange for free OH.

Having identified the −NH₂ contribution to the D₂ tagged spectrum of the all H H⁺−4ABA isotopologue, we next consider the assignments of the three remaining peaks (3478, 3554, and 3595 cm⁻¹) that did not shift in the single H spectrum in the context of the OH stretching contribution from the remaining two OH groups. The fact that there are more than two peaks raises the possibility that more than one rotamer is in play as well as the perturbation of the intrinsic spectrum by the D₂ tag. To address the tag effect, we extended the study to include addition of a second D₂ as well as complexation with a more strongly binding adduct, in this case N₂. The relative strengths reflect the respective gas-phase basicities of 495 and 424 kJ/mol [33] for N₂ and D₂, respectively, and hence the expected redshift of the bound OH frequency should be larger for N₂. Since the charge is localized on the carboxyl group for the O-protomer, the tag species is expected to interact only with the OH groups, thus only the peaks associated with the OH groups are expected to shift while the NH groups remain unchanged. Figure 3B–D summarize the result, where the 3441 and 3542 cm⁻¹ peaks are recovered in all spectra, confirming the assignment that they are the v^sym and v^asym stretches of the NH₂ group, respectively. Since the isotope study verified that the OH groups are very weakly coupled, we expect that attachment of a single tag will leave one OH at its intrinsic location, ${\nu }_{OH}^{\mathit{\text{free}}}$, while the other bound OH, $v_{OH}^{\mathit{\text{bound}}}$, shifts in ways that depends on the binding strength of the adduct. If the dominant rotamer is isomer B or C in Figure 1, where both OH groups point in the same direction, then we would expect two OH stretching peaks since the tag species would break the degeneracy of the two ${\nu }_{OH}^{\mathit{\text{free}}}$. The case for isomer A is more complex, since the tag species could either bind to the E- or Z- positions, which could have different binding strengths [34], hence different redshifting magnitudes. That should yield a pattern of four bands, whereas the observed spectrum of the single tag only has three, which does not correspond to either of the scenarios outlined for the behavior of a single species.

One way to resolve this paradox is to add a second tag, which should yield a single band for rotamers B and C (Fig. 1) since the quasi-equivalent OH groups are both tagged, and two bands for the E,Z isomer (A in Fig. 1) if the intrinsic frequencies of the distinct OH groups are different and both are tagged. Harmonic calculations, for example, predict that the Z-OH group falls 48 cm⁻¹ below that of E. The O-protomer spectrum of the 4-ABA•(D₂)₂ cluster is presented in Fig. 3C and indeed displays two bands (red) in addition to the blue bands from the NH₂ group. Note that the highest energy feature disappears from the spectrum, indicating that this transition is due to the free OH of one of the distinct OH groups in the E,Z isomer. Based on the calculated behavior, this is assigned to the E-OH group pointing toward the ring, while the lower energy feature is retained and therefore is due to the $v_{Z-OH}^{\mathit{\text{bound}}}$ fundamental. This exercise does not, however, reveal the location of the Z-OH group as only the E-OH feature was removed with the second tag. This remaining ambiguity can be resolved, however, by introduction of the more strongly bound N₂ molecule, which yields the spectrum displayed in Fig. 3D. Indeed, a new band appears 89 cm⁻¹ below the $v_{Z-OH}^{\mathit{\text{bound}}}$ (D₂) feature, which is expected for the shift when D₂ is replaced by N₂ on the Z-OH group. Now, there are four bands in addition to the two from the NH₂ group, which is the pattern expected for the E,Z isomer described above. In that analysis (3D), the two highest frequency OH bands are the non-bonded, and hence intrinsic features of the E,Z O-protomer, as indicated in Fig. 3B. Curiously, however, that the N₂ tagged spectrum contains a peak that is very close to the assigned D₂-tagged Z-OH group. And the new assignment for the free Z-OH band appears very close to that assigned to the bound OH of the E-OH in the (D₂)₂ spectrum (Fig. 3C). This suggests that there is an accidental near degeneracy in play such that the locations of the bound OH groups are overlapping free positions, and the D₂-tagged E-OH falls very close to the frequency of the free Z-OH! This scenario is consistent with the pattern in the H⁺−4ABA•(N₂)₂ spectrum (Fig. 3E), which establishes the frequencies of the bound E and Z-OH groups. Note that fingerprint region is not complicated by tag effects (Fig. S2).

3.3. Isolation of the tag site isomers in the H⁺−4ABA•D₂ spectrum with IR²MS³ isomer-selective spectroscopy

Although self-consistent, this analysis of the tag shifts relies on two accidental near coincidences such that bands that fall in similar locations in the patterns are assigned to different species. This unfortunate circumstance could be resolved by introduction of yet another tag with intermediate binding energy, but a more systematic way to unambiguously establish the assignment is to leverage the fact that two isomers are invoked to explain the H⁺−4ABA•D₂ spectrum (Fig. 3B), which differ according to which of the distinct OH groups is bound to the D₂ molecule. As such, the spectra of each isomer can be isolated using two-color, IR-IR photobleaching, a method in which the populations of the isomers are removed upon excitation of the transitions associated each. By setting a probe laser on a specific transition monitoring the photofragmentation, the population of the isomer responsible for this transition is continuously monitored. When a second, powerful laser (the bleach laser) is scanned through the entire spectrum prior to the interrogation by the probe laser, all of the features associated with the probed isomer are revealed as population depletion or dips as a function of the bleach laser frequency. The method requires three stages of mass selection to separate the actions by the bleach and probe lasers, and is therefore classified as an IR²MS³ method of secondary ion analysis [20].

The isomer-selective scans revealed two distinct patterns that are overlapping in the non-selective spectrum (Fig. 3B), which are displayed as the depletion or “dip” spectra in Figure 4C and 4E. These were obtained by setting “probe” laser at the two positions, 3595 cm⁻¹ (Fig. 4C) and 3554 cm⁻¹ (Fig. 4E) indicated by the black arrows in Fig. 4, while scanning the spectrum with another “bleach” laser from 2900–3800 cm⁻¹. The 3595 cm⁻¹ band corresponds to the highest energy band in Fig. 3B, assigned to the $v_{E-OH}^{\mathit{\text{free}}}$ fundamental. As such, we expect the Z-OH to be bound to the D₂ tag, and indeed one of the three remaining dips in Fig. 4C appears at the $v_{Z-OH}^{\mathit{\text{bound}}}$ band (red at 3478 cm⁻¹). The more interesting case is that obtained with the probe laser set at 3554 cm⁻¹, which we provisionally assigned to the $v_{Z-OH}^{\mathit{\text{free}}}$ band, should reveal the location of the bound E-OH group as a second dip. Interestingly, the bleach scan (Fig. 4D) only displays a single dip beyond the two expected for the NH₂ bands. This confirms that the transition from the bound E-OH group, $v_{E-OH}^{\mathit{\text{bound}}}$, indeed accidentally falls under the $v_{Z-OH}^{\mathit{\text{free}}}$ band, confirming the analysis based on tag shifts.

Figure 4.

A) IRMPD spectrum of H⁺−4ABA reproduced with permission from Ref. [13] B) Non-selective predissociation spectra of H⁺−4ABA•D₂ in H₂O. C) and E) Double resonance spectra with probe positions at 3595 cm⁻¹ and 3554 cm⁻¹, respectively, denoted by vertical black arrows. Dashed red and cyan vertical arrows on top of trace 4A at 3482 and 3550 cm⁻¹, respectively, denote previously reported Z- and E-OH frequencies for protonated acetic acid [17]. Calculated spectra of the E,Z-isomer of H⁺−4ABA•D₂ with D₂ at the Z- (D) and E-position (F) at the B3LYP/aug-cc-pVTZ level of theory. Calculated frequencies are scaled by 0.96. Optimized structures of the isomers are shown to the right with ZPE corrected relative energies displayed above them. Blue peaks represent NH stretches, red for Z-OH, and cyan for E-OH. Horizontal arrows represent the D₂ tag-induced shift of the Z-OH (red) and E-OH (cyan) groups.

The assignment of the two free OH transitions for E,Z isomer yields a close approximation to the expected spectrum for the bare ion. These positions are indicated by the dashed lines in Fig. 4A, which displays the IRMPD spectrum reported by Polfer for the bare ion at room temperature [13, 29]. Given the broadening expected for the non-linear method applied to ions with significant internal energy, the agreement of the observed features with those obtained from the tagged species is excellent, in turn establishing that the E,Z isomer was also in play in those studies. Curiously, the reported free Z- and E- OH frequencies measured in this study (3554 and 3595 cm⁻¹, respectively) are higher than those observed in protonated acetic acid (H⁺-AA, 3482 and 3550 cm⁻¹) [17]. This observation indicates the range of frequencies exhibited by the protonated carboxylic acid motif (R-C(OH) ⁺₂) with different R-groups. Although in both cases, the $v_{Z-OH}^{\mathit{\text{free}}}$ occurs at a lower frequency than the $v_{E-OH}^{\mathit{\text{free}}}$, the splitting between the two is significantly different (41 and 68 cm⁻¹ for H⁺−4ABA and H⁺-AA, respectively).

3.4. Differences in the tag binding energies to the E- and Z-OH groups and anticooperativity upon N₂ binding to both

Although the tag perturbations are rather large in this system, their behavior is also useful in that they encode how the orientation of the OH group affects the binding energy to that group. In both cases, the Z-OH group exhibits a much larger shift than the E-OH (see Table 1). Note that both molecules are predicted to bind with their axes roughly perpendicular to the OH group, which is typical for H-bond to cations [35]. In the case of the Z-OH, D₂ binds with the center of the bond along the OH axis, while in the more weakly bound E-OH configuration, it is off this axis by about 30.8°. This misalignment appears to reduce the strength of this interaction, likely a result of the steric interference by the nearby CH group on the ring. Another interesting aspect of the tag shifts is the that for the more strongly interacting N₂ complexes, attachment of a second molecule causes the OH groups originally bound to N₂ to undergo blue shifts with magnitudes (14 and 10 cm⁻¹) such that the Z-OH is more affected than the more weakly interacting E-OH group. This represents an anti-cooperative effect [36–39] between the two OH groups such that binding to one of them weakens the bond already in play with the other. This effect is not observed as strongly in D₂ because D₂ is a weaker H-bond acceptor than N₂.

Table I.

Summarizing the experimental and calculated assigned normal modes of the O- and N- protomer. Calculated frequencies were computed at the B3LYP/aug-cc-pVTZ level of theory. The N-protomer calculated frequencies were scaled by 0.968 according to the CCCBDB. The calculated frequencies of H⁺−4ABA•X (X = D₂, N₂) O-protomer were scaled by 0.96.

Band label	Observed freq. (cm−1)	Calculated scaled freq. (cm−1)	Description
O-protomer
$v_{N{H}_2}^{sym}$	3441	3439	NH2 symmetric strethch
$v_{N{H}_2}^{\mathit{\text{asym}}}$	3542	3549	NH2 antisymmetric strethch
$v_{E-OH}^{\mathit{\text{free}}}$	3595	3614	Free E-OH stretch of the E,Z-rotamer
$v_{Z-OH}^{\mathit{\text{free}}}$	3554	3569	Free Z-OH stretch of the E,Z-rotamer
$v_{E-OH}^{b-D_2}$	3554	3547	D2 bound E-OH stretch of the E,Z-rotamer
$v_{Z-OH}^{b-D_2}$	3478	3456	D2 bound Z-OH stretch of the E,Z-rotamer
$v_{E-OH}^{b-N_2}$	3492	3502	N2 bound E-OH stretch of the E,Z-rotamer
$v_{Z-OH}^{b-N_2}$	3389	3388	N2 bound Z-OH stretch of the E,Z-rotamer
$v_{E-OH}^{b-2N_2}$	3502	-	2N2 bound E-OH stretch of the E,Z-rotamer
$v_{Z-OH}^{b-2N_2}$	3403	-	2N2 bound Z-OH stretch of the E,Z-rotamer
N-protomer
$v_{N{H}_3}^{\mathit{\text{umbrella}}}$	1470	1467	NH3 umbrella mode
νC=0	1791	1749	Carbonyl stretch
$v_{N{H}_3}^{sym}$	3235	3212	Symmetric NH3 stretch
$v_{N{H}_3}^{\mathit{\text{asym}}}$	3273	3316	Asymmetric NH3 stretch
	3316	3358	Asymmetric NH3 stretch
νOH	3574	3610	Free OH stretch

3.5. The proton bound dimer: complexation with 4ABA

Having clarified the assignments in the protomers, it is of interest to explore the docking motif by which the O-protomer attaches to neutral 4ABA in light of the previous work that identified the key linkage in the binary complex starting with the N-protomer [12]. In that case, the dimer is asymmetrical such that the NH⁺ attaches to the C=O group leading to a “head-to-tail” arrangement (representative structure shown in Fig. 5E). The spectrum of the D₂ tagged H⁺-(4ABA)₂ binary complex is displayed in Fig. 5A, along with the calculated (scaled) harmonic spectra for several minimum energy candidate structures identified in our search of the potential energy landscape. We can immediately establish that this is different from the species generated from the N-protomer by absence of the free C=O stretch (red in Fig. 5E) in the experimental spectrum [12]. On the other hand, two of the sharp bands in the region of the free NH region are in the same locations as found in the O-protomer, while the third feature highest in energy (red, $v_{OH}^{\mathit{\text{free}}}$) falls at the same location as the free acid OH stretch in the N-protomer. All of these telltale bands are consistent with the spectrum predicted for the lowest energy isomer (Fig. 5B), which features a linkage between the two carboxylic acid groups. Interestingly, however, this is different from the motif found for the proton bound dimer of formic acid, which occurs with a C=O--H⁺--O=C type arrangement where the extra proton binds the two carbonyl groups together [40]. The structure in Fig. 5B, on the other hand, exhibits an unusual bifurcated motif in which the Z,Z form of the protonated O-protomer binds quasi-symmetrically to the C=O group on the neutral partner. That configuration also rationalizes the two very diffuse bands that appear in the 2500–3000 cm⁻¹ range (labeled θ and χ in Fig. 5A) as due to the symmetric and asymmetric stretching modes of the two OH groups on the O-protomer that are attached to the oxygen atom. Such broadening is often observed in such cases where displacements in the soft, intermolecular modes strongly modulate the frequency of the bound OH groups [41]. In this case, the hinging mode corresponding to the flex of the two moieties at the linkage is a likely candidate for the coupling that leads to the breadths of the bands. We also explored the generality of this bifurcated motif in the brief survey of the H⁺−4ABA•H₂O complex, which is also calculated dock in the bifurcated motif as evidenced by the appearance of two broad features that occur at 3390 and 3265 cm⁻¹. The IRMPD spectrum of the bare complex is included in Fig. S3.

Figure 5.

A) Vibrational spectrum of H⁺-(4ABA)₂ sprayed in H₂O. B-E) calculated spectra of the possible binding motifs of H⁺-(4ABA)₂ at the B3LYP/6–311++G(d,p) level of theory. Calculated frequencies were scaled by 0.95. Representative structures and their corresponding ZPE corrected relative energies are displayed above the calculated spectra. Blue represents NH stretches, red for OH stretches, cyan for free E-OH, and orange for transitions that involve the shared proton.

4. Summary

We report the vibrational spectra of the D₂- and N₂-tagged O- and N-protomers of protonated 4-aminobenzoic acid (H⁺−4ABA), which are obtained by electrospray ionization of 4-aminobenzoic acid at pH=~2 in water and a 1:1 solution of water and acetonitrile. The spectrum of the isotopologue with one H and three D atoms in all exchangeable sites confirms the assignments of the NH stretches in the spectrum of the all-H isotopologue and establishes that the OH stretches on the protonated carboxylic acid head group are essentially uncoupled. The tag molecules introduce a substantial perturbation of the OH stretching vibrations of the O-protomer, and isomer-specific, two-color IR-IR photobleaching spectroscopy confirms that the single tag species occurs in two isomeric forms. These correspond to the two distinct OH groups in the E,Z rotamer, which is the only structural isomer of the O-protomer generated under these conditions. The tag shift is observed to be much larger for the bound Z-OH group than that for the E-OH. Addition of a second N₂ tag results in blue shifts of both the bound Z- and E-OH groups, thus establishes a weak anticooperativity effect where a second H-bond weakens the first. The spectrum of the proton bound dimer, prepared by attachment of 4ABA to the O-protomer of H⁺−4ABA, indicates that this system binds through contacts between the two carboxylic acid functionalities. This linkage is distinct from that reported earlier that started from the N-protomer where the dimer adopted a “head-to-tail” arrangement. The linkage identified here for the O-protomer features a bifurcated interaction where both OH groups of the Z,Z-rotamer bond to the carbonyl group of the neutral acid.

Highlights.

• 4-Aminobenzoic acid (4ABA) has two protomers: N-protomer or carboxylic O-protomer

• The O-protomer has three low energy rotamers: E,Z; Z,Z; and E,E

• Isomer-selective spectroscopy confirms only the E,Z isomer is present

• Site-specific binding shows that the E- and Z-OH have different H-bond strengths

• The O-protomer binds to a neutral 4ABA through a bifurcated motif to the Z,Z isomer