Anion‐Π Meets H‐Bonding: Tuning Synergy Through the Flexibility‐to‐Preorganization Transition in Anion Receptor Design

Abstract

This study focuses on a series of receptors incorporating urea and pentafluoropyridine motifs to investigate the synergistic combination of hydrogen bonding and anion‐π interactions for anion recognition. Three receptors with various spacers are synthesized to evaluate the influence of molecular preorganization and rigidity on anion binding. Flexible receptor 1 and rigid receptors 2 and 3 are synthesized following a two‐step protocol, involving the construction of the urea fragment from isocyanate precursors and a nucleophilic aromatic substitution to install the tetrafluoropyridine motif as key steps. Computational analyses (density functional theory (DFT), noncovalent interaction (NCI) plots), nuclear magnetic resonance (NMR) spectroscopy, and mass spectrometry are employed to assess structural features and binding performance. DFT calculations reveal that all receptors allow complexation with chloride through dual urea and π‐anion sites. Structural rigidity in receptor 3 shows enhanced binding efficacy due to steric strain and additional C—H···Cl⁻ interaction from its naphthalene core. NMR titrations provide qualitative insights into binding events, with receptor 3 exhibiting the largest shieldings for all H‐bonds, in line with theoretical predictions. Mass spectrometry and collision‐induced dissociation experiments confirm receptor–anion complexation, with fragmentation patterns supporting the relative binding strengths. The overall ranking is 3 > 2 > 1, corroborating computational and experimental data.

Three anion receptors combining both hydrogen bonding and anion‐π interactions are synthetized, involving different spacers in order to evaluate the influence of molecular preorganization and rigidity on chloride binding. The naphthalene spacer in receptor 3 leads to more effective anion binding compared to flexible or less‐structured systems.

1. Introduction

Anions are ubiquitous and play a crucial role in various fields,^{[ 1} , ^{2 ]} including medicine, environmental science, and catalysis.^{[ 3 ]} As a result, anion recognition has become a fundamental area of supramolecular chemistry. In this context, several types of noncovalent interactions contribute to anion binding, such as hydrogen bonding (H‐bonding),^{[ 4} , ⁵ , ^{6 ]} van der Waals (VDW) forces, sigma–hole interactions,^{[ 7 ]} and anion–π interactions.^{[ 8 ]} Combining multiple interactions within a single receptor design remains challenging, as it requires an understanding of possible synergies and how they might impact anion binding.^{[ 9 ]}

The potential synergy between hydrogen bonding and anion–π interactions has been highlighted in both experimental and theoretical studies. Matile and coworkers have demonstrated that π‐acidic naphthalenediimide (NDI) systems bearing hydrogen bond (HB) donor arms can activate anionic substrates and stabilize transition states in catalytic contexts, revealing the powerful cooperative effects at play in anion–π catalysis.^{[ 10} , ¹¹ , ¹² , ^{13 ]} Complementary computational investigations by Alkorta et al. further emphasized the stabilizing role of such dual interactions, showing enhanced binding and charge‐transfer when anions and H‐bond donors act simultaneously on π‐acidic surfaces.^{[ 14} , ^{15 ]}

In this context, our study seeks to systematically examine this cooperative behavior within a series of structurally controlled receptors, enabling direct comparison between flexible and preorganized systems in order to deepen the understanding of how spatial arrangement and rigidity influence the interplay of hydrogen bonding and anion–π recognition. We disclose in this study, the exploration of the combination of hydrogen bonding—one of the most common noncovalent interactions—with anion–π interactions, which are comparatively less studied in anion recognition. For hydrogen bonding, we chose the urea functional group because of its well‐established anion binding properties. For anion–π interactions, we have incorporated the pentafluoropyridine (TFP) motif, which is known in the literature as anion–π interaction donor.^{[ 9 ]} However, the presence and effectiveness of the latter interaction within a designed receptor family remains uncertain or at least to clarify.^{[ 16 ]}

To investigate the synergy between these two interactions, we systematically varied the spacer unit connecting the urea and TFP moieties (Figure  1 ). The spacer plays a critical role in determining the flexibility/or preorganization of the receptor, which in turn affects the efficiency of anion recognition. To this end, we designed two types of spacer: 1) A flexible linear carbon chain (receptor 1), which allows conformational adaptability in the presence of the anion; and 2) rigid aromatic spacers, including a phenyl (receptor 2) or a naphthalene moiety (receptor 3), which impose a preorganized structure, creating a defined binding cavity where the relative positions of the urea and pyridine groups are fixed. The size of the cavity differs between the phenyl‐ and naphthalene‐based receptors, affecting the spatial arrangement of the interacting sites.

Figure 1.

Target receptors 1–3.

Within this series, to assess the impact of receptor design on the synergistic effect of hydrogen bonding and anion–π interactions, we will study their complexation with chloride anions using a combination of theoretical (density functional theory (DFT) calculations), solution (nuclear magnetic resonance (NMR) spectroscopy), and gas‐phase (mass spectrometry) techniques. This systematic approach will provide insights into how structural preorganization and interaction synergy influence anion recognition, contributing to the rational design of more efficient anion receptors.

2. Computational Study: Geometries Optimization and Binding Analysis

The geometries of the three receptors were optimized using the APFD/Aug‐cc‐pVDZ level with the Gaussian 16 set of programs.^{[ 17 ]} For receptor 1, two initial conformations were considered (Figure  2 ). In the first conformation (1a), the spacer is extended, keeping the urea and TFP fragments far apart, to minimize steric hindrance. In the second conformation (1b), the spacer is folded, bringing the urea and TFP fragments closer together, which led to stabilization through weak intramolecular interactions. As a matter of fact, a comparison of the total energies of these two conformers, both of which converged to a local minimum with no imaginary frequencies, showed that 1b is more stable than 1a by 35.5 kJ mol⁻¹ in relative energies (See Table S3, Supporting Information), thereby suggesting that the receptor may naturally adopt a preorganized structure, in which the urea and TFP fragments are already positioned to interact with the anion without requiring additional energetic cost for folding. The calculation of the noncovalent interaction (NCI) plots using Multiwfn software^{[ 18 ]} on 1a and 1b was consistent with the energies of the receptor, as 1b has more intramolecular stabilizing interactions (Figure 2, interactions plotted in green) than 1a.

Figure 2.

Computed topology of the various receptors 1 and associated NCI plots. NCI analyses were performed using Multiwfn 3.7. RDG isosurfaces were generated with an isovalue of 0.4 a.u., using the medium quality grid. The scalar field sign(λ ₂)ρ was used to color the isosurface, with a default color range of −0.05–+0.05 a.u., corresponding respectively to attractive (blue), VDW (green), and repulsive (red) interactions. Density and gradient cutoffs were set to their default values in Multiwfn, namely ρ < 0.05 a.u. and |sign(λ ₂)ρ| < 0.05 a.u.

Consequently, we focused on a preorganized topology of the bonding cavity for receptors 2 and 3. Indeed, the presence of a phenyl and a naphthyl rigid spacers creates defined binding cavities. For example, the initial geometry of 3 was selected so that the urea protons face the cavity and the TFP fragment, maximizing potential interactions. The choice of this 3D topology is supported by 2D NMR results (see NMR section, vide infra). After optimization, both receptors reached a local minimum with no imaginary frequencies. In the particular case of receptor 2, two forms, 2a and 2b of close stability are found exhibiting a slightly different folding (Figure 2). 2a turns to be slightly more stable due to a stabilizing interaction between the urea motif and TFP ring. Remarkably, these two conformations are almost degenerate when considering relative free energies. As a consequence, both conformations will be considered for the chloride bonding study.

Upon addition of the chloride anion (Cl⁻) to the optimized geometries, three anion‐receptor complexes were obtained (Table  1 ). In all three cases, qualitative analysis using NCI plots, calculated with Multiwfn software,^{[ 18 ]} confirmed that the anion interacted simultaneously with both the urea and TFP fragments. The blue‐colored interaction plot in the NCI analysis indicates a strong favorable interaction between the anion and the urea protons, confirming the presence of hydrogen bonding. A green‐colored interaction plot, indicating a weaker but favorable interaction, was observed between the anion and the TFP fragment, providing evidence for anion–π interactions (Figure  3 ). Interestingly, the nature of the anion–π interaction differed between complexes. In [1 + Cl] ⁻ and [2a + Cl] ⁻ , it induces a η ² interaction, whereas a η ³ and η ⁶ interaction is observed in complexes [2b + Cl] ⁻ and [3 + Cl] ⁻ , respectively. Moreover, in the case of [3 + Cl] ⁻ , a third type of noncovalent interaction is evidenced, corresponding to a C—H^…Cl ⁻ interaction between the anion and one hydrogen of the naphthalene moiety located inside the binding cavity.

Table 1.

Comparison of selected bond distances in receptors 1–3 and their associated chloride complexes, computed at the APFD/aug‐ccpvdz level.

	Fragment contribution
Entry	Urea	H‐bonda)	1		[1 + Cl] −	2b	[2b + Cl] −	3	[3 + Cl] −
1		Na—Ha	1.011		1.044	1.010	1.046	1.009	1.042
2		Nb—Hb	1.011		1.034	1.010	1.036	1.009	1.028
3		Ha ···Cl	–		2.045	–	2.015	–	2.030
4		Hb ···Cl	–		2.136	–	2.165	–	2.220
	Ar C—H	H‐bond
5		C—Hf	–		–	–	–	1.088	1.095
6		Hf ···Cl	–		–	–	–	–	2.417
7	Anion–π	Cl···centroïd	–		3.409	–	3.631	–	3.291
8		Ca···Cl	–		3.956	–	3.537	–	3.565
9		Cb···Cl	–		3.825	–	4.109	–	3.518
10		Nc···Cl	–		3.602	–	4.429	–	3.559
11		Cc···Cl	–		3.377	–	4.180	–	3.548
12		Cd···Cl	–		3.465	–	3.631	–	3.581
13		Ce···Cl	–		3.778	–	3.258	–	3.548

^a)distances in Å.

Figure 3.

Computed forms of the various [receptor + Cl]⁻ complexes and associated NCI plots.

To assess the complexation quantitatively, the interaction energies of the three anion‐receptor complexes were calculated, yielding values ranging from −192.4 kJ mol⁻¹ for [1 + Cl]⁻ to −243.1 kJ mol⁻¹ for [3 + Cl]⁻ (Table S1, Supporting Information). Thus receptor 3 presents the highest affinity for chloride and 1 the lowest. [2a + Cl]⁻ and [2b + Cl]⁻ exhibit intermediate values (Table S1, Supporting Information) allowing an overall ranking 1 < 2 < 3. Taking into account that [2a + Cl]⁻ exhibits a significative lower interaction energy than [2b + Cl]⁻, only the latter will be further considered. The associated basis set superposition error (BSSE) have been estimated, and are rather low (2.7–3.1 kJ mol⁻¹, see Table S1, Supporting Information). This finding is consistent with BSSE estimated for complexes involving trifluoroterephthalonitrile‐containing receptors.^{[ 19 ]} Finally, note that for the sake of comparison, we have also carried out the computational study at the B3LYP/6‐311++G(d,p) level, by including the D3 version of Grimme's dispersion with Becke–Johnson damping (GD3‐BJ).^{[ 20 ]} Results are gathered in the Section 4.6 of the supporting information. We could deduce from this comparison that the optimized geometries of both neutral and complexes are close, and that the computed interaction energies are very similar. Consequently, the fact that B3LYP also performed well is an important point with a view to studying systems of increasing size, because APFD calculations are slightly longer than those involving B3LYP.

Modification of interatomic distances before and after anion binding are indicative of the effective complexation with chloride. In addition, the interstitial distances between the anion and fragments capable of generating weak interactions are also worth investigating in order to establish the receptor–host relationship. To this end, selected calculated binding distances are summarized in Table 1 and deserve comments. Table 1 is organized according to the nature of the spacer and the weak interactions involved in the binding of the chloride anion.

The urea fragment and the H–bond interactions were first examined (Table 1, entries 1–4). For a HB to be considered significant, the sum of the VDW radii^{[ 21 ]} of the anion and the interacting atom (2.95 Å) should be greater than the H···Cl⁻ distance.^{[ 22 ]} This condition was fulfilled in all observed N—H···Cl⁻ interactions, confirming the presence of hydrogen bonding as evidenced in NCI plots.

From the receptor to the corresponding complex, the N_a—H_a and N_b—H_b distances increase. Notably, all N_a—H_a distances vary in a similar manner (1.011–1.044, 1.010–1.046, and 1.009–1.042), indicating an almost identical contribution to chloride complexation. The N_b—H_b distances also increase in all complexes, but to a lesser extent. It clearly highlights the influence of the 3D topology imposed by the rigid phenyl and naphthyl spacers in compounds 2b and 3. A more detailed study about the effect of the different functional groups onto the N—H‐bond lengthening is provided in Section 4.3 of the supporting information.

The H_a···Cl⁻ distances decrease from receptor 1 to receptors 2b and 3, suggesting a strengthening of the hydrogen bonding interaction from 1 to 2b and 3 as suggested by computed H‐bonding energies (Table  2 , vide infra). Moreover, all H_b···Cl⁻ distances are significantly longer than their H_a analogs, likely indicating a greater contribution of H_a to Cl⁻ complexation. This evolution of bond distances thus seems to indicate a positive influence of the presence of rigid spacers in 2b and 3 on the efficiency of hydrogen bonding in anion complexation.

Table 2.

HB binding energies (BE) in kJ mol⁻¹.

Complex	(Na—Ha···Cl−)	(Nb—Hb···Cl−)	(Hf···Cl−)
[1 + Cl]−	−57.2	−47.9	–
[2b + Cl]−	−60.8	−45.5	–
[3 + Cl]−	−58.4	−41.4	−30.2

From the topological analysis, we may estimate the binding energy of each HB, by using the electron density at their associated bond critical point.^{[ 23 ]} Assuming a charged HB toward Cl⁻, the different HB binding energies are summarized in Table 2.

One may observe that there is a very good correlation between the computed BE and the H _x ···Cl distances (Figure S26, Supporting Information). HB distances can therefore be used confidently to discuss HB strength for the various complexes. These estimates notably confirm the strengthening of the N_a—H_a···Cl HB on going from receptor 1 to receptor 3. One can also notice that the HBs involving H_a are systematically stronger than those involving H_b, resulting in the more pronounced lengthening of the N_a—H_a bond. This is in agreement with the stronger variation of the chemical shift of H_a upon addition of chloride during NMR experiments (vide infra).

The contribution of the C_aromatic—H fragment to the complexation event through hydrogen bonding was also investigated. For [3 + Cl]⁻, the presence of C—H···Cl⁻ bonding was confirmed by the lenghtening of the C—H_f distance upon anion interaction. In addition, the C—H_f···Cl⁻ distance being 2.417 Å, is shorter than 2.95 Å, further confirming the presence of a HB. The computed BEs (Table 2) show that this HB remains weaker than those involving the urea motif but partly explains the weakening of the N_b—H_b···Cl HB within the [3 + Cl]⁻ complex. These observations support a clear positive contribution from the rigid naphthalene platform.

For anion–π interactions, the distance between the anion and the centroid of the aromatic ring should be ≤4 Å, same as each atom of the aromatic ring and the anion for a η ₆ anion‐π interaction.^{[ 24 ]} These conditions were met for 1 and 3 (Table 1). For 2b, only C_c···Cl, C_e···Cl, and C_d···Cl were shorter than 4 Å indicating a weaker interaction with η ³ type.

First, the centroid···Cl⁻ distance appears significantly shorter in [3 + Cl]⁻ compared to [1 + Cl]⁻ and [2b + Cl]⁻, suggesting a more beneficial effect of increased steric strain in the binding cavity. It is also worth noting that all C···Cl and N···Cl bond distances in [3 + Cl]⁻ are ≈3.5 Å, suggesting an η ₆ coordination mode. In contrast, the C_c···Cl and C_d···Cl bond distances are shorter compared to other C···Cl bonds in 1, indicating an η ² coordination mode. Similarly, C_c···Cl, C_d···Cl, and C_e···Cl distances are sensibly shorter in 2b indicating an η ³ coordination mode. Overall, the bond distances are more homogeneous in 1 than in 2b. This difference may also indicate a better adaptability of the flexible arm.

Overall, the theoretical analysis of the weak interactions involved in chloride binding highlights the significant role of hydrogen bonding and anion–π interactions in complex stabilization. Compared to receptor 1, the presence of rigid spacers in 2 and 3 impacts the overall topology of the corresponding complexes and results in a slight enhancement of the efficiency of hydrogen bonding Ha···Cl⁻. However, a decrease of the H_b···Cl⁻ is counterbalanced in 3 by a stronger anion‐π interaction. In addition, the naphthalene platform in [3 + Cl]⁻ contributes positively through C—H···Cl⁻ interactions, further enhancing the binding capacity. Noteworthy, in [1 + Cl]⁻, the anion‐π interaction occurs through a nearly η ² coordination mode. In contrast, [2 b + Cl]⁻ displays a η ³ coordination mode while [3 + Cl]⁻ exhibits η ⁶ coordination mode. These observations highlight the interplay between structural rigidity, steric effects, and three weak interactions in optimizing chloride anion binding efficiency, making 3 the best candidate.

For gaining quick information, we have attempted to compare the above analysis with a more qualitative overview. The qualitative analysis of weak interactions can be carried out using the bond distances from Table 1 and computed H‐bond binding energies in Table 2. For each interaction, several contributing fragments can be considered. Each of these fragments can be assigned four levels of contribution depending on the increasing importance of their contribution to chloride ion complexation phenomenon (Table  3 ): − (no contribution—pink), + (weak contribution—orange), ++ (moderate contribution—yellow), or +++ (enhanced contribution—green).

Table 3.

Qualitative ranking of receptor abilities.

Weak interaction	Fragment	[1 + Cl]−	[2b + Cl]−	[3 + Cl]−
Urea—H Bond	Na—Ha	+++	+++	+++
	Nb—Hb	++	++	+
	Ha …Cl	+++	+++	+++
	Hb …Cl	++	++	+
Car—H—Cl	C—Hf	–	–	+
	Hf …Cl	–	–	+
π‐Cl	Centroid…Cl	++	–	++
Overall ranking		1 < 2b < 3

Thus, the N_a—H_a distance remains very similar regardless of the complex ([1 + Cl]⁻, [2b + Cl]⁻, or [3 + Cl]⁻), indicating a comparable (strong) contribution of this fragment. On the other hand, the N_b—H_b bond is shorter for the [3 + Cl]⁻ complex than for its analogs 1 and 2b, indicating a weaker contribution in 3.

The length of the H_a···Cl or H_b···Cl bond also indicates the contribution of the urea fragment to the complexation phenomenon. In this context, the H_b···Cl bond is longer for [3 + Cl]⁻, whereas H_a···Cl bonds are significantly shorter (below 2.1 Å) for all the complexes, especially for [2b + Cl]⁻ and [3 + Cl]⁻.

Thus, it appears that the rigid phenyl and naphthyl motifs have a pronounced effect on the parameters derived from the urea fragment. The C—H_f and H_f···Cl bonds are specific to the rigid naphthalene motif. They are only present in [3 + Cl]⁻ and therefore contribute exclusively to this complex.

The tetrafluoropyridine motif, which is likely to contribute through the weakest interaction, is the last fragment analyzed. The Cl···centroïd distances are significantly shorter for [1 + Cl]⁻ (3.409 Å) and [3 + Cl]⁻ (3.291 Å) compared to [2 + Cl]⁻ (3.631 Å). This could indicate a detrimental effect due to a sterically more constrained binding cavity. Furthermore, the C···Cl and N···Cl distances are homogeneous in complex [3 + Cl]⁻, suggesting a η ⁶ coordination mode. In contrast, the complexes [1 + Cl]⁻ and [2b + Cl]⁻ exhibit differentiated C—Cl bond distances. The C_c···Cl and C_d···Cl distances appear shorter in [1 + Cl]⁻ , likely confirming an η ² coordination mode. In [2b + Cl]⁻ the C_c···Cl, C_d···Cl and C_e···Cl distances appear shorter, confirming an η ³ coordination mode.

From this qualitative analysis, we confirmed a significant contribution of the rigid and templating naphthalene fragment to the complexation phenomenon through the addition of H···Cl, C_ar—H—Cl, and π—Cl bonds. This ranking 3 > 2 > 1 is in agreement with the calculated interaction energies.

The combination of hydrogen bonding and anion–π interactions is highly dependent on the structural nature of the spacer unit. Our study suggests that preorganization by rigid spacers enhances anion binding by reinforcing both interactions, with larger cavities further enhancing this effect. These findings provide valuable guidance for the rational design of more efficient anion receptors.

We next concentrated our efforts in the preparation of receptors 1, 2, and 3, and in the analysis of complexation to chloride by mean of NMR and mass spectrometry.

2.1. Synthesis

The synthesis of the receptors follows a straightforward and mild two‐step procedure (Scheme  1 ). For 1 and 3, a nucleophilic addition reaction is carried out at room temperature to form the urea intermediate. This is followed by a nucleophilic aromatic substitution (SN_Ar) to introduce the pyridine moiety.^{[ 25 ]} Receptors 1 and 3 were obtained with respective yields of 89% and 35%. For 2, the synthetic sequence required further optimization and was inverted. The SN_Ar occurred first followed by the nucleophilic addition, leading to 2 in a yield of 75%.

Scheme 1.

Synthesis routes of 1, 2, and 3.

For receptor 3, a 2D NMR study confirmed the structure optimized by DFT calculations. The NOESY spectrum revealed interactions between H_a and H_b, indicating that both urea protons are in the same plane (Figure  4 ). In addition, the interaction between H_b and H_f suggests that the urea moiety is located within the preorganized cavity. Furthermore, the HOESY spectrum is in complete agreement with the presence of the pyridine moiety in a plane parallel to the urea through a correlation between F_a and H_f.

Figure 4.

Overlap of ¹H NMR spectra (CD₃CN) of 3 (red) and 3 + 25 eq of Cl⁻ (green).

2.2. ¹H NMR Qualitative Insights

Unfortunately, the comparison of the association constant obtained from a ¹H NMR titration, between 1‐Cl (755 L mol⁻¹) and 2‐ or 3‐Cl, could not be carried out. Attempts to obtain quantitative results from a formal titration of receptors 2 and 3 failed partly because of partial and rapid degradation of receptor 2 in solution (CD₃CN or CD₃COCD₃), or strong overlap of the characteristic urea and naphthalene proton signals in the naphthalene analog 3.

In this context, a qualitative NMR approach is presented in this study. It will be followed by a full mass spectrometry analysis.

¹H NMR allows us to gain information by analyzing which protons are affected by the addition of the anion. Observing the chemical shift shielding from the receptor to the corresponding complexes is a simple experimental way of qualitatively assessing the strength of H‐bonds, as illustrated by Figure 4.

For all three receptors, 25 equivalents of TBACl were added, and the Δδ values of H_a, H_b, and H_f were calculated. The urea protons are strongly impacted by the anion addition; in each case, the H_a proton is sensibly more affected than H_b (Table  4 ). Receptor 3 exhibits the largest ΔδH_a and ΔδH_b values, indicating stronger interactions compared to receptors 2 and 1. Furthermore, the presence of an H_f···Cl⁻ interaction is also detected with a ΔδHf of 0.84 ppm, confirming an additional weak interaction in [3 + Cl]⁻. This qualitative analysis allows the establishment of the 3 > 2 > 1 capability ranking, which is in good agreement with both the above qualitative data and DFT results.

Table 4.

Qualitative ranking of receptor abilities by ¹H NMR.

Proton	Δδ [ppm]
	1‐Cl	2‐Cl	3‐Cl
Ha	3.38	3.75	3.80
	++	++	++
Hb	1.80	2.07	2.46
	+	+	++
Hf			0.84
	–	–	+
Overall	+	+	++
Ka a )	744	nd	nd

^a)unit in L mol⁻¹.

From this NMR qualitative analysis, it turns out that receptor 3 should be the most effective. This finding is in good agreement with ranking 3 > 2 > 1 determined from the calculated interaction energies and DFT calculations.

The next step will be to compare the receptor binding abilities using mass spectrometry. This will provide further insights into their binding properties.

2.3. Mass Spectrometry

Our previous studies have shown that chloride/receptors complexes described as stable species in the gas phase according to computational studies, can be easily observed experimentally when using electrospray ionization.^{[ 19} , ²⁶ , ²⁷ , ²⁸ , ^{29 ]} This is again the case with the present receptors. Employing electrospray ionization on equimolar solution (10⁻⁵  m) of receptor/NBu₄Cl prepared in a 90/10 ACN/water mixture, results in the formation of abundant peaks that may be ascribed to [(receptor)_n + Cl]⁻ ions (n = 1,2), as illustrated by the Figure S21 of the Supporting Information, obtained for the 1/NBu₄Cl mixtures. Note however that in the particular case of receptor 2, these complexes could be observed only when solutions were prepared just before the mass spectrometry experiments. It turned out that this compound rapidly degrades as soon as it is solubilized, leading to a weak complex that is impossible to study over the long term and thus to compare with 1 and 3 analogs. In spite of this peculiar behavior, we were able to record the collision‐induced dissociation (CID) spectrum of the 1:1 complex for the three receptors, in order to get some insights about their structure. These spectra are gathered in Figure  5 .

Figure 5.

MS/MS spectrum of a) [1 + Cl]⁻, b) [2 + Cl]⁻, and c) [3 + Cl]⁻ recorded on a 3D quadrupole ion trap.

First, these CID spectra are characterized by 2 distinct processes: 1) deprotonation of the receptor by loss of hydrogen chloride (−36u); and 2) fragment ions arising from the cleavage of covalent bonds within the receptor. As these spectra were recorded with a 3D quadrupole ion trap (Bruker Amazon Speed ETD), the low mass range by default is limited by a low mass cutoff of 27% of the precursor ion. In addition, modifying the cutoff to try to observe the m/z 35/37 ion resulted in the complete loss of signal. Consequently, this instrument is unable to detect the possible formation of Cl⁻ by elimination of the intact receptor. Therefore, we also recorded these spectra with a Q‐TOF instrument which does not suffer from this limitation, and we did not observe any Cl⁻ ion, thereby suggesting a rather strong interaction between the anion and the receptor. Starting with [1 + Cl]⁻, two main product ions are observed (Figure 5). One can see that elimination of HCl leading to deprotonated receptor [1‐H]⁻ is a minor process. But this process is indicative of a strong interaction of chloride with the urea motif, as formation of the [receptor‐H]⁻ ions probably implies removal of one of the urea proton. Implicitly, removal of the Ha proton should be more favorable as the resulting anion would be stabilized by a mesomeric effect involving the bis‐CF₃ phenyl ring. This is confirmed by DFT calculations which indicate that the Ha deprotonated receptor is more ≈22 kJ mol⁻¹ more stable than the Hb deprotonated form (Figure S26, Supporting Information). The dominant fragmentation process gives rise to a m/z 166 which certainly imply the fluorinated pyridine moiety. This ion may be obtained by a SN₂‐like mechanism involving the attack of Cl⁻ onto the C(H₂)—O carbon and departure of the [OC₅F₄N]⁻ ion (m/z 166, Figure S22, Supporting Information). Alternatively, one may envisage deprotonation of H_a as a first step, followed by an intramolecular attack of the pyridine by an SN_Ar process, and ultimately formation of m/z 166. This assumption is supported experimentally by the MS3 spectrum of [1‐H]⁻ (Figure  6a). The fact that this product ion is totally shifted by 16 mass units (m/z 182) when the carbonyl group (C=O) of receptor 1 is replaced by a thiocarbonyl group (C=S), also supports this assumption.^{[ 30 ]} This process may be favored by the folding of the receptor around Cl⁻ through the combination of hydrogen bonds and anion‐π interactions, and may suggest the interaction of Cl⁻ with the pyridine motif. The behavior upon collision of [2 + Cl]⁻ and [3 + Cl]⁻ ions is similar, and different from that of [1 + Cl]⁻. First, the peak associated with the deprotonation of the receptor is much more intense, and remarkably corresponds to the base peak for [3 + Cl]⁻. This may be due to the fact that deprotonation of both nitrogens of urea results in very stable anions strongly stabilized by mesomeric effect, either by the bis‐CF₃ phenyl ring (H_a) or the aromatic spacer (H_b). DFT calculations show that with the introduction of an aromatic spacer, removal of the Hb proton becomes slightly favored (Figure S26, Supporting Information) over removal of Ha. We also recorded the CID spectra of both deprotonated receptors (Figure 6b,c). One can see that all the product ions observed in Figure 5b,c are also present for the deprotonated receptors. In addition, the relative intensities of m/z 377, 257, and 237 ions on the MS/MS spectra of [2 + Cl]⁻, and [2 − H]⁻, are strikingly similar. These data suggest a two‐step process for receptors 2 and 3: deprotonation by loss of HCl followed by dissociation of the deprotonated receptors. Possible mechanisms accounting for the formation of the various fragments are given in supporting information (Figure S23 and S24, Supporting Information), and assume that a mixture of deprotonated forms resulting from the removal of both H_a and H_b could be at the origin of the different product ions. Finally, we also recorded a series of MS/MS/MS spectra according to the 512 → 257 → sequence, which confirm that m/z 237 ions probably arise from m/z 257 by loss of hydrogen fluoride (HF). A similar dissociation is observed for m/z 562 (Figure 6c), leading to m/z 542.

Figure 6.

Low‐energy a) MS3 spectrum of [1 − H]⁻ (m/z 464) and MS2 spectra of b) [2 − H]⁻ (m/z 512) and c) [3 − H]⁻ (m/z 562) recorded on a 3D quadrupole ion trap. Excitation amplitude: 0.30, 0.20, and 0.25 V, respectively.

Like with many receptors, the CID spectrum of [(receptor)₂ + Cl]⁻ ions (not shown) is characterized by the elimination of one intact molecule of receptor, leading to the 1:1 complex. So, we also tried to determine the relative chloride affinities of the three receptors by means of the kinetic method.^{[ 19} , ³¹ , ^{32 ]} This method implies the formation of heterodimers of the type [receptor 1 + receptor 2 + Cl]⁻ which is then subject to a CID experiment. If weak interactions are involved within the dimer, these ions should dissociate to generate two monomers [receptor 1 + Cl]⁻ and [receptor 2 + Cl]⁻, the most intense product ion corresponding to the molecule having the highest affinity for chloride. Due to its instability in solution, it was not possible to generate heterodimers with receptor 2. On the other hand, the [1 + 3 + Cl]⁻ ion could be easily generated by electrospray (m/z 1063) by infusing an equimolar 1:1:1 mixture of 1/3/NBu₄Cl, and its low‐energy CID spectrum is given in the Supporting Information (Figure S25, Supporting Information). The dominant fragmentation process channel corresponds to the formation of the [3 + Cl]⁻ monomer, clearly indicating that receptor 3 has a higher affinity for Cl⁻ than receptor 1 in the gas phase, therefore confirming the relative order deduced from DFT calculations and NMR experiments. The fact that the ion [1 + Cl]⁻ is only present in trace amounts is consistent with an important difference in computed interaction energies in favor of 3 (≈50 kJ mol⁻¹).

3. Conclusion

Our study demonstrates that synthetic receptors designed to combine hydrogen bonding and anion–π interactions significantly enhance chloride ion recognition. Incorporating rigid, preorganized frameworks, such as the naphthalene spacer in receptor 3, leads to more effective anion binding compared to flexible or less structured systems. Both theoretical (DFT calculations) and experimental (NMR spectroscopy, mass spectrometry) evidence confirm that structural preorganization and the synergy of multiple weak interactions are key factors in optimizing anion recognition. Receptor 3, in particular, exhibits superior chloride affinity due to a unique combination of N—H···Cl⁻, C—H···Cl⁻, and Cl···π interactions. From our data, we set the following ranking: 3 > 2 > 1, which highlights the crucial role of molecular design, especially the use of rigid and pre‐organized architectures, in achieving new‐generation receptors for supramolecular anion recognition.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supplementary Material

Dedicated to Professor François Terrier, a pioneer in physical organic chemistry whose work and legacy continue to inspire

Zayene Olfa, Hu Jun, Gaucher Anne, Plais Romain, Barday Manuel, Moreau Xavier, Salpin Jean‐Yves, Prim Damien. ChemPhysChem. 2025; 26:e202500378. 10.1002/cphc.202500378

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.