Metal‐Ion Regulated Light‐Free Z→E Isomerization of Azobenzene Glycomacrocycle

ABSTRACT

We reported a 26‐membered azobenzene‐embedded glycomacrocycle, that undergoes metal‐ion‐mediated, light‐free Z→E isomerization across an unprecedented kinetic spanning from seconds (Cu²⁺) to weeks (Co²⁺). The macrocycle was synthesized via an intramolecular glycosylation strategy incorporating an azobenzene‐triazole linker and an isopropylidene‐protected pentaerythritol unit. Upon addition of divalent metal perchlorates (Cu²⁺, Ni²⁺, Co²⁺, Pb²⁺) to the Z‐isomer, three distinct kinetic regimes were observed and characterized by absorption spectroscopy and electrospray ionization‐mass spectrometry. Beyond simple coordination‐driven isomerization, Cu²⁺, Ni²⁺, and Pb²⁺ were found to act as Lewis acids, promoting in situ hydrolysis of the isopropylidene acetal moiety. This metal‐induced chemical remodeling dramatically simplifies the macrocycle's conformational landscape, from a fragmented multi‐conformer ensemble to a singular global minimum, as validated by a conformational search method. A photokinetic experiment on the copper complex revealed near‐instantaneous re‐establishment of the E‐complex upon UV cessation, representing the fastest light‐free Z→E reversion observed in this system. Crucially, the macrocyclic framework is recoverable by metal scavenging with QuadraPure IDA, confirming the system's reversibility and recyclability. This work introduces a design principle wherein a glycomacrocyclic scaffold, photoresponsive azobenzene, metal‐ion‐mediated light‐free isomerizable, and Lewis‐acid‐labile protecting group cooperatively generate a programmable, temporally tunable molecular switch with broad potential in stimuli‐responsive materials and chemical logic applications.

Metal ion‐catalyzed Z→E isomerization of photoswitchable azobenzene glycomacrocycle: A new azobenzene‐based glycomacrocycle is capable of forming a 1:1 complex with divalent metal ions in both E‐ and Z‐isomers. The complexes undergo light‐free Z→E isomerization in a timescale spanning from seconds to weeks. Cu²⁺, Ni²⁺, and Pb²⁺ act as Lewis acid catalyst to promote in situ isopropylidene hydrolysis, which dramatically simplifies the conformational landscape of the deprotected macrocycle.

1. Introduction

Azobenzene‐metal complexes represent a unique class of dual‐functional molecular systems that bridge coordination chemistry with photoresponsiveness. The azobenzene moiety can either directly participate in metal coordination [1, 2] or function as a peripheral photoswitch to modulate the ligand part of the whole chemical structure. In the latter role, it can trigger light‐controlled processes at the metal center, such as redox transitions or photomagnetic effects [3, 4]. Alternatively, when embedded within a well‐designed molecular scaffold, the E→Z photoisomerization (or its reversion) induces a drastic conformational reorganization. Specifically, the contraction of the azobenzene unit width from 9 Å (E‐isomer) to 5.5 Å (Z‐isomer) activates or deactivates the reaction of the metal and its coordination environment [5, 6, 7, 8] This structural metamorphosis enables the development of reactivity‐contingent systems, with applications ranging from photocontrol of DNA cleavage [8] to the design of smart MRI contrast agents [9] that exhibit enhanced signal differentiation upon isomerization. Beyond these biomedical applications, azobenzene‐metal complexes offer versatile potential in the field of advanced materials science as well.

As a well‐established molecular scaffold, azobenzene undergoes E→Z isomerization under UV irradiation, with the reversible Z→E process triggered by visible light, thermal conditions, or a combination of both. In order to modulate this dynamic process for precise experimental control, researchers employ diverse strategies [9], including electrocatalysis, substituent engineering on the phenyl rings, designing pH‐responsive molecules, trapping in host‐guest systems, etc [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]. The use of ionic stimuli to modulate thermal Z→E isomerization represents a parallel and complementary strategy. Pioneering work by Shinkai and coworkers demonstrated that cation binding to azobenzene‐embedded crown ethers modulates both the rate and thermodynamics of thermal isomerization: large alkali metal cations (Rb⁺, Cs⁺) forming sandwich complexes with the cis‐isomers suppress back‐isomerization, while the trans/cis cavity size difference enables photoswitchable ion selectivity between Na⁺ and K⁺ [10, 11]. Subsequent studies by Jurczak and coworkers extended this concept to anion‐mediated acceleration: carboxylate binding to azobenzene‐urea receptors increases electron density on the CN═NC bond and accelerates thermal Z→E isomerization in a concentration and affinity‐dependent manner, a concept further elaborated with chiral urea aminoglucose receptors, where the bound enantiomer controls the isomerization rate [12, 13]. In a macrocyclic context, Cafeo et al. reported that bis‐carboxylate binding to a bis‐calix[4]pyrrole azobenzene receptor imparts guest‐dependent biexponential isomerization kinetics through coupled steric and electronic effects [14]. Beyond ionic control, Goulet–Hassens et al. established redox‐mediated pathways—both reductive (radical anion) and oxidative (hole catalysis)—enabling quantitative Z→E isomerization with quantum yields exceeding unity using red light [15]. These precedents collectively demonstrate that the thermal Z→E isomerization barrier is a tunable parameter sensitive to the electronic and coordination environment of the azobenzene moiety. Recently, Hossain et al. have reported a class of azobenzene macrocycles capable of undergoing metal‐ion‐mediated light‐free isomerization at room temperature [25]. In their system, the compact macrocyclic Z‐isomer isomerizes upon coordination with AgClO₄ or Cu(ClO₄)₂ in acetonitrile in the absence of light, giving E‐isomer at a slow or instant rate, depending on Ag⁺ or Cu²⁺, respectively. The C─N═N─C motif of azobenzene participated in the metal coordination, which was especially enforced in the Z‐isomer to involve the binding. Nevertheless, the study remains constrained by limited metal scope (Ag⁺, Cu^{2 +}) and relatively inflexible kinetic profiles. The use of transition‐metal divalent cations as programmable Lewis acid effectors capable of simultaneously driving isomerization and chemically remodeling a macrocyclic scaffold has remained unexplored.

Herein, we address these challenges through a rational design of a mannose‐based scaffold featuring an azobenzene‐triazole linker. Our group has previously demonstrated the versatility of azobenzene‐embedded glycomacrocycles in applications such as chiral dopants for liquid crystals [26, 27]. In this work, we extend the scope of metal‐ion‐mediated light‐free Z→E isomerization to a broader range of divalent cations, including Ni²⁺, Co²⁺, and Pb²⁺, in addition to Cu²⁺. Upon coordination with these metal perchlorate salts, the macrocycles undergo structural transitions confirmed by UV‐vis absorption and electrospray ionization‐mass spectrometry (ESI–MS). By extending the range of the metal ions, we have achieved this specific isomerization spanning from seconds to week timescale level. Crucially, the system exhibits reversibility. The resulting metal complex solution can be reverted to E‐isomer noncomplexed form by using metal scavengers.

Our mannose‐based architecture (Scheme 1b) integrates several strategic features: (i) a conformationally labile 26‐membered ring accommodating triazole‐ester motifs for tunable metal affinity; (ii) a photoresponsive and coordination‐compatible azobenzene moiety capable of driving significant structural metamorphosis; (iii) a β‐glycoside unit with an isopropylidene‐protected pentaerythritol moiety, enhancing macrocyclic flexibility and future functional extensibility. This design offers four key advantages:

1. M^{2 +}‐triggered light‐free azobenzene Z→E isomerization,

2. the broadest reported metal scope (Ni^{2 +}, Co^{2 +}, Pb^{2 +}, Cu^{2 +}) for dark‐state switching,

3. a metal‐dependent kinetic range covering 7 orders of magnitude timescale (from seconds to weeks), and

4. a reversible process to noncomplexed E‐isomer by using commercially available metal scavengers.

SCHEME 1.

(a) Strategy of building the azobenzene embedded glycomacrocycle, composed of the glycomacrocycle moiety and a 1,2,3‐triazole linker. (b) Structure of compound 1 and its E↔Z isomerization pattern. (c) The total process of 1: photoresponsive E→Z isomerization, metal ion‐mediated light‐free Z→E isomerization, and the initial state reversibility in the presence of the metal scavenger.

2. Results and Discussion

2.1. Synthesis

Synthesis of the glycomacrocycle is achieved by using our previously developed intramolecular glycosylation strategy (Scheme 2) [28, 29, 30, 31]. The O‐propargyl‐functionalized azobenzene intermediate 4 is obtained through esterification reaction between isopropylidene acetal protected pentaerythritol 2 and azobenzene 3 in 41% yield, along with some di‐esterification product 5. CuAAC reaction between the azido‐functionalized thiomannoside 6 [30] and the azobenzene alkyne 4 led to the compound 7 which underwent the intramolecular glycosylation to the glycomacrocycle E‐1 (containing about 10% Z‐isomer) in 39% yield. The β‐anomeric configuration has been confirmed by the coupling constant of J _C1,H1 = 153.5 Hz (Figure S1) [32].

SCHEME 2.

Synthesis of glycomacrocycle E‐1.

2.2. Spectroscopic Analyses

The solvent compatibility of various metal perchlorate salts was prioritized, with acetonitrile selected for its optimal perchlorate solubility and photochemical inertness at λ > 300 nm. As shown in Figure 1, the electronic transitions of the azobenzene macrocycle were characterized via UV–Vis spectroscopy. The E‐1 isomer exhibited characteristic π→π* absorption bands around 350 nm and 300 nm, with the latter featuring a shoulder likely arising from the dissymmetric ortho‐azobenzene moiety. The n→π* transition appeared as a weak shoulder around 450 nm. Upon reaching the photostationary state (PSS₃₆₅) under UV irradiation, the Z‐isomer exhibited significantly diminished π →π* bands and a sustained, blue‐shifted (Δλ ≈ 20 nm vs. E‐1) n →π* transition. Notably, Z‐1 demonstrated good thermal stability, with negligible spectral evolution observed after 16 h in the dark at room temperature (Figure 1, inset).

FIGURE 1.

UV–Vis absorption spectra of E‐1, Z‐1 (PSS₃₆₅), and its thermal evolution (interval = 2 min) in darkness at room temperature for 16 h. All spectra were performed in acetonitrile at c = 50 µM. Inset: Z‐1 (PSS₃₆₅) time‐dependent absorbance change at 428 nm in the dark.

Following the precedent that azobenzene‐embedded rings can undergo metal‐ion‐mediated Z→E isomerization in the absence of light [24, 25], we introduced five equivalents of four different metal perchlorates to PSS₃₆₅ mixture of 1. The observed behaviors fell into three distinct kinetic categories in Figure 2:

1. Ni²⁺ and Co²⁺ (slow kinetics; Figure 2b,d): upon addition of Ni(ClO₄)₂ or Co(ClO₄)₂, the π–π* bands characteristic of the E‐isomer gradually recovered, accompanied by a bathochromic shift and subtle enhancement of the n–π* band. Significant intensity increases were observed at T+1 week compared to T+1 h, suggesting the initial formation of a Z‐complex followed by a slow, spontaneous reversion to the corresponding E‐complex. In addition, the T+1 week absorption intensity exceeded the E‐1·M²⁺ control reference.

2. Pb²⁺ (Moderate kinetics; Figure 2c): In the case of Pb(ClO₄)₂, the Z‐reverted and native E‐complexes control reached identical spectral intensities within one week. The resulted spectral profile is one broad absorption band, in contrast to the three other ions.

3. Cu²⁺ (Instantaneous kinetics; Figure 2a): The behavior with Cu(ClO₄)₂ was entirely distinct. The suppression of π→π* bands typical of the Z‐isomer was immediately observed at T+0, with spectra at T+0 and T+1 h overlapping together. This indicates an isomerization process occurring less than one minute, far exceeding the resolution of the steady‐state UV‐vis observing window. Notably, the n–π* band intensity initially exceeded the E‐1 reference at T+0, followed by a significant decline at T+1 w, compared to both T+0/T+1 h states and E‐1 complex reference.

FIGURE 2.

UV–Vis absorption spectra in acetonitrile of E‐1 and Z‐1 (PSS₃₆₅) and four different metal ions introduced as perchlorate salts (a) Cu(ClO₄)₂, (b) Ni(ClO₄)₂, (c) Pb(ClO₄)₂, and (d) Ni(ClO₄)₂. The E‐1 and Z‐1 (PSS₃₆₅) without additions of metal ions are shown in black line and grey dashed line, respectively. The time evolution T+0 (measured right after the addition), T+1 h (hour), T+1 d (day), T+1 w (week) has been picked for Z‐1 (PSS₃₆₅) with metal ions, and T+1 d (day) has been picked for E‐1 with metal ions.

The divergence in T+1 w behaviors across these metal systems can be rationalized by two coupled factors: the isopropylidene acetal deprotection (demonstrated by our subsequent ESI–MS analysis; see E‐1 metal complex MS study section) and the presence of Z‐isomer (∼10% Z in the isolated E‐1).

2.3. E‐1 Metal Complex MS Study

Accompanied by our spectroscopic investigations, we then performed ESI–MS experiments. The primary goal of these experiments was to determine the nature of the complexes generated when mixing E‐1 or Z‐1 with the different metal perchlorate salts. To this end, we used the same ligand/metal ratio (1/5) considered for the spectroscopic studies, but we reduced the concentrations to ensure compatibility with mass spectrometric conditions. Figure 3 presents typical electrospray spectra obtained in presence of Cu(ClO₄)₂ for E‐1 and Z‐1. Spectra obtained with the three other metals are given in the supporting information (Figures S4–S6).

FIGURE 3.

Positive ion ESI–MS spectra of a 1/5 mixture of (a) E‐1/Cu(ClO₄)₂ and (b) Z‐1/Cu(ClO₄)₂ (with Cu(ClO₄)₂ at 5 × 10⁻⁵ M) in acetonitrile, recorded on a 3D quadrupole ion trap. E‐1* and Z‐1* refer to glycoazomacrocycle without isopropylidene acetal function (see text for details).

The obtained electrospray spectra are rather simple and exhibit two families of complexes. The first group arises from the interaction of copper perchlorate with acetonitrile, leading to [Cu(CH₃CN)]⁺, [Cu(CH₃CN)₂]⁺, [Cu(ClO₄)(CH₃CN)]⁺ (weak), and [Cu(ClO₄)(CH₃CN)₂]⁺ at m/z 104.1, 145.1, 203.0, and 244.0, respectively. For the two formers, copper is in the (+1) oxidation state. This is not surprising as Cu²⁺ reduction under electrospray conditions is often observed. The second group of complexes is characterized by the interaction of the metal with the glycoazomacrocycle 1. Two different complexes, in which copper retains its +2 oxidation state, were observed, namely [Cu+1]²⁺ and [Cu(ClO₄)+1]⁺, at m/z 523.2 and 1145.4 (Figure 3b), respectively. The isotopic distribution easily confirms the double charge of the most intense complex, whereas the ion incorporating Cu(ClO₄) is of low intensity. Mass spectrometry data show that only the 1:1 stoichiometry was observed experimentally. Amidst the mass data recorded, there were instances of the isopropylidene acetal group removal (40 u mass shift). The pH values of the acetonitrile solutions with the same amount of M(ClO₄)₂ were examined. The loss of isopropylidene acetal observed for E‐1 in

Figure 3a could be ascribed to the acidic solution environment (here pH ≈ 1.8), which could trigger the deprotection process. In that case, the mass‐to‐charge ratios of the doubly‐charged and singly‐charged complexes are shifted downward by 20 and 40, respectively. However, based on all the experiments performed, we could not establish a clear and direct correlation between the signal intensity of unprotected species and the acidity of the solution. While the acidic conditions can be entirely discounted as a factor, it remains plausible that the deprotection effect is influenced by the presence of different metal sources.

We recorded the MS/MS spectrum of the ions detected at m/z 523.2 and 503.2 (Figure S2), and the product ions observed confirm that these ions do correspond to the [1+Cu]²⁺ and [1*+Cu]²⁺ complexes, respectively. Comparison of the mass to charge ratios observed with protected and unprotected complexes allows determining the product ions incorporating the isopropylidene protection (for example 452/492 and 672/712). We propose a dissociation scheme (Figure S3) to account for the most intense product ions detected. Finally, when examining the spectra obtained with Co²⁺, Ni²⁺, and Pb²⁺ (Figures S4, S5 and S6, respectively), an overall common behavior is observed with the four perchlorate salts: (i) similar spectra observed for E and Z macrocycles, (ii) a 1:1 stoichiometry with a systematic dominant doubly‐charged complex, (iii) a weak [M(ClO₄)+ligand]⁺ ion. Finally, in the particular case of Ni(ClO₄)₂ and Co(ClO₄)₂ (Figures S4 and S5), we could also observe several polynuclear species whose isotopic distribution is consistent with the [M₂(ClO₄)₃(CH₃CN)_n]⁺ formula (n = 2,3 for Co and 3,4 for Ni).

Therefore, the MS data reveals a clear hierarchy in Lewis acidity and catalysis efficiency: while Co²⁺ induces no acetal hydrolysis even after 24 h (Figure S4), Ni²⁺ and Pb²⁺ promote moderate to extensive deprotection (Figures S5 and S6), and Cu²⁺ triggers a near‐instantaneous conversion to diol from 10 min to full conversion after 24 h (Figures 3, S2, and S3). The purity‐driven intensity boost of Z→E complex over E‐complex reference in the case of Ni/Co‐reverted samples can be attributed to the presence of approximate 10% residual Z‐isomer in the synthesized E‐1 sample. Considering the E‐complex as the global energy minimum of the coordination system, the Z→E thermal reversion is inhibited by a high activation energy without metal ions. Upon coordination, the metal ion may stabilize the transition state, lowering this energy barrier. In the case of Ni²⁺ and Co²⁺, the ∼10% Z‐isomer in the E‐reference stayed at the metastable Z‐complex state too long due to the sluggish kinetics of these two metal ions. Conversely, the PSS samples offered a more concentrated initial Z‐isomer environment, facilitating a greater statistical drive to overcome the barrier within the observing window. Eventually, the metal‐ion‐mediated Z→E complex process populated the final E‐complex state more extensively. Nevertheless, at the T+1 h interval, both Ni²⁺ and Co²⁺ systems exhibited nearly identical π–π* recovery trends, suggesting comparable metal‐ion‐mediated Z→E isomerization rates. However, their plateaus at T+1 w diverged. ESI‐MS confirms that the Co²⁺ system remains entirely protected acetal state after 24 h. Consequently, its T+1 w absorbance remains relatively low, reflecting a “pure” contribution from the metal‐ion‐mediated Z→E isomerization. In contrast, MS data shows extensive hydrolysis in the Ni²⁺ system after 1 day. This transformation into a vicinal diol dissolves the fused ring structure on the 26‐membered big ring. This chemical transformation contributes to a subtle ring strain relaxation, allowing the azobenzene moiety to adopt a more planar conformation. This can explain the more pronounced π–π* bands of Z→E·Ni²⁺and E‐1·Ni²⁺ after 24 h.

For Pb²⁺, the moderate coordination rate resulted in the same intensities of final E‐complex state. The broad single band can be attributed to the significantly larger ionic radius of Pb²⁺ (∼1.19 Å vs. Ni²⁺/Co²⁺/Cu²⁺ ∼0.7 Å). The bulky Pb²⁺ effectively fills the macrocyclic cavity, suppressing the structural fluctuation of the 26‐membered ring. The T+1 h has already begun to shift toward the single broad band shape, and this bulky metal ion offsets the increased flexibility typically associated with deprotection. This leads to a robust, synchronized equilibrium where Z‐reverted and E‐native complexes reach an identical endpoint.

The spectroscopic profile of the Cu²⁺ system is the most complex one, characterized by an instantaneous Z→E transition followed by a profound secondary decay after 1 day. This can be explained from the perspective of electronic orbital symmetry and Lewis acid‐catalyzed hydrolysis. The most striking kinetic disparity between the metal ions lies in the evolution of the π→π∗ transition region (∼350 nm) and n→π^∗ band (∼450 nm). Upon the addition of Cu²⁺ to the PSS₃₆₅ mixture, a near‐instantaneous surge in absorbance is observed within seconds, reaching a plateau that reflects the rapid Z→E thermal reversion and the unique Cu²⁺(d⁹ ) characteristics. From the π→π∗ transition region, we can see Cu²⁺ establishes an instantaneous electronic overlap with the azobenzene π‐system upon coordination. In addition, an immediate and massive enhancement across the n→π^∗ band at T+0 and T+1 h is nearly identical and significantly higher than that of the E/Z‐1 metal‐free reference in Figure 1, which could be ascribed to the d⁹ electronic configuration of Cu²⁺. Apart from the Ni²⁺(d⁸ ) and Co²⁺(d⁷ ) rigid coordination spheres, Cu²⁺(d⁹ ) has Jahn–Teller effect, which undergoes a spontaneous tetragonal distortion (elongation or compression of axial bonds) to lower the system energy. When Cu²⁺ enters the cavity, the effect may force the azobenzene N═N unit into a highly strained and asymmetrical coordination environment. This symmetry breaking would lift the forbidden nature of the n–π* transition, resulting in the significant enhancement of the intensity from 420 to 570 nm region. In this state, the whole molecular structure is highly distorted from the spectroscopic perspective. Because Cu²⁺ is potent, the distorted E‐complex is formed within the dead time of the observing window of steady‐state spectroscopies. At T+1 w, the spectrum of Z→E complex undergoes a secondary transformation: the n–π* band drops to the lowest compared to T+0/1 h and T+1 d of E‐native complex. The isopropylidene acetal is entirely converted into the corresponding vicinal diol after 24 h. With the evolution of time, the diol moiety could change the ambience to fit the Cu²⁺ coordinative satisfaction, relaxing the azobenzene from a Jahn‐Teller effect highly distorted conformation. This could explain the lower n–π* band of E‐1·Cu²⁺ with the time evolution.

The steady‐state spectroscopic evolution of the 26‐membered macrocycle is governed by a hierarchical interplay between coordination geometry and metal‐induced chemical remodeling. While the Ni²⁺ and Co²⁺ systems primarily reflect the slow kinetics pattern, the Pb²⁺ system demonstrates an ion dimension‐dependent effect. Most notably, the Cu²⁺ system reveals a dynamic structural transition from an initially strained Jahn‐Teller distorted state to a relatively relaxed diol‐coordinated environment. The chemical nature of the E‐complex and deprotected E‐1* and E‐1*·Cu²⁺ has been validated by ESI–MS analysis (Figure 3). Beyond confirming the formation of stable metal‐ligand complexes, the mass spectra provide proof of the acetal hydrolysis, tracking the conversion from the strained acetal scaffold to the relaxed diol form. For the unique Cu²⁺ case, we managed to observe the system in a time‐resolved window of E→Z→E isomerization by a photolysis experiment.

2.4. Conformational Search Analyses

While hybrid DFT functional offers high accuracy, the vast conformational space of this 26‐membered macrocycle requires an extensive sampling approach to capture the global landscape. Instead of optimizing some different starting points of the glycomacrocycle, a set of extensive conformational searches was performed using the Conformer‐Rotamer Ensemble Sampling Tool (CREST) algorithm at the GFN2‐xTB semi‐empirical level [33]. This ensemble‐based perspective proved more robust in correlating spectroscopic signatures than isolated high‐level cluster calculation, which often easily neglects the diversity of the flexible scaffold. Subsequent Boltzmann distribution and RMSD analysis reveal the conformational difference between the two forms, providing a detailed mapping of the potential energy surface (PES).

Set with sufficient metadynamics run times with this 26‐membered ring within an 8 kcal/mol energy window, we have carried out conformational sampling both in the gas phase and in acetonitrile to reduce the artefacts. The influence of the solvent environment is quantitatively captured by the global partition function (Q), which contracts from 2.977 in the gas phase to 1.581 in acetonitrile. This reduction in Q indicates that while the acetonitrile environment promotes a degree of pre‐organization, the isopropylidene acetal‐protected scaffold remains a multi‐conformer ensemble (320 states within 8 kcal/mol in acetonitrile, 111 states in the gas phase). In the gas phase, there are more low energy valleys from the RMSD calculations, however, the solvation consolidates these into a main one in acetonitrile. As summarized in Table S1, the CN═NC dihedrals of the top 10 conformers (C1–C10, 95.28% cumulative weight) in acetonitrile exhibit high local planarity (176.60° to 179.51°), seemingly mimicking the geometry of the diol‐form (179.69°). However, the RMSD matrix (Table S2) and visualization inspections reveal a profound topological distinction. While the azobenzene‐triazole moiety achieves local planarity, the macrocyclic backbone of C1–C10 is locked in a “compact‐folded” topology (Figure 4a). The structural variance within this ensemble is primarily localized to the stochastic fluctuations of the benzyl ether groups within a constrained, “clamped” cavity. Compared to the gas phase, where the top 5 conformers span three distinct families with RMSD gaps of 4–5.5 Å RMSD (Figure S8), the solvated topology ensemble is like a “grand basin” that still has a “forest” of micro‐states inside from the fluctuations from the main scaffold.

FIGURE 4.

Conformational portraits of macrocycle 1 from CREST/GFN2‐xTB analysis. (a) E _{isopropylidene}: overlay of the top 10 acetonitrile solvated conformers (C1–C10) in a compact‐folded shape. (b) Z _{isopropylidene}: singular constrained conformer with CN═NC dihedral of −26.9°, locating in a geometrically isolated area from the E ensemble. (c) E _diol: singular expanded‐stretched global minimum, resulting in an energy funnel endpoint.

A fundamental topological isolation emerges from the Z‐acetal conformational search, which converged to a singular state in both gas phase and acetonitrile, producing the exact same conformation. The Z‐scaffold, despite occupying a bigger space than the E‐acetal forms, is locked into a singular, highly constrained topology by the isopropylidene acetal anchor, which enforces a CN═NC dihedral angle of −26.9°. The shape is transformed vastly from the E‐acetal ensembles, indicating the E/Z‐acetal forms are vastly far on the PES. The great conformational change suppresses all thermally accessible conformational pathways connecting the Z‐region to the E‐acetal forest. This is entirely consistent with the observed thermal stability of PSS₃₆₅ Z‐1: in the absence of any chemical or light stimulus, no spectroscopic evolution is detected over 16 h, confirming that the Z‐acetal singularity is a kinetically trapped state as shown in Figure 4b.

Regardless of the computational environment, the conformational search for the diol converged to a singular, unique conformer (100% Boltzmann weight). It indicates on the PES, that the diol is like a funnel that everything falls into one destination in the end. The deprotection of the acetal anchors triggers a thermodynamic singularity, where the entire population funnels into a singular, unique state in both gas phase and acetonitrile (1 vs 111/320 states). This state is characterized by an “expanded‐stretched” geometry (Figure 4c) with a fully relaxed macrocyclic ring—a topology that is energetically and physically inaccessible to the fused‐ring precursor. Topological hindrance signifies the inability of the semi‐rigid 26‐membered ring to undergo the requisite global reorganization to bridge divergent basins on the potential energy surface without chemical deprotection.

This three‐way portrait—a topologically isolated Z‐singularity, a frustrated E‐acetal forest, and a thermodynamically relaxed E‐diol funnel—provides a unified structural foundation for interpreting the divergent spectroscopic behaviors (the evolution of the π→π∗ absorption band) of the four metal systems. The Co²⁺ system represents a stochastic recovery. ESI‐MS and UV–Vis monitoring confirm that Co²⁺ lacks the Lewis acidity required to catalyze the hydrolysis of the isopropylidene acetals. The macrocycle of PSS₃₆₅ drops randomly into the available E‐form conformational forest. This explains why the Co²⁺ spectrum perfectly replicates the initial E‐1 profile characteristic bands around 300–350 nm, where the higher‐energy sub‐band remains dominant. Ni²⁺ has induced major hydrolysis over 1 day as determined by ESI‐MS. Given the similar ionic radius to Co²⁺, it alters the PES by the Lewis acidity, resulting in a “conformational purification” signature in the UV–Vis. As Ni²⁺ catalyzes the removal of the acetal, the PES shifts from the energy basin into the funnel with a singular endpoint. In UV–Vis, we observe a slow but steady increase in the lower‐energy sub‐band of the π→π∗ absorption region, which eventually surpasses the higher‐energy sub‐band. This can be ascribed to the PES shifting from the complicated fused‐ring ensemble into the planar diol funnel, reaching the new global minimum. The Pb²⁺ system is similar but influenced by the big ionic radius refitting, giving a broad merged band with a higher side of the lower‐energy sub‐band. Cu²⁺ has the quickest kinetics, and the Cu²⁺‐induced full hydrolysis further purifies the PES to the diol form. The final spectrum exhibits the most defined lower‐energy sub‐band feature.

2.5. Photokinetic of Cu(II) Complex

In the previous section, we discerned that Cu²⁺ induced the Z→E isomerization in the absence of light, albeit at a distinguished pattern compared to Co²⁺, Ni²⁺, and Pb²⁺. To further explore the kinetics, we subjected the E‐1^⚫ Cu(II) complex to a photokinetic experiment (Figure 5). The freshly prepared PSS₃₆₅ Z‐1 treated with 5 eq. Cu(ClO₄)₂ within 1 h, i.e., the E‐1^⚫ Cu(II) complex, was irradiated at 365 nm, while the absorbance at 480 nm (n–π* band) was monitored in real‐time to track the isomerization process.

FIGURE 5.

Photokinetic experiment of 1^⚫ Cu(ClO₄)₂ complex. The violet signal is the indicator of the intervention of UV irradiation at 365 nm, and the blue signal represents the instant UV–Vis absorption at 480 nm. The irradiation window time control was kept at 1 min each.

Remarkably, we observed the suppression of the n–π* band while the UV irradiation was on, and the system exhibited an unprecedentedly fast reversion to the initial state upon the cessation of the irradiation. As depicted in Figure 5, in the window of UV irradiation, the observed spectral suppression could stem from either the formation of a highly strained Z‐complex or a transient photo‐induced dissociation, where the reduced cavity of the Z‐isomer or the significantly bent azobenzene pushed out the Cu²⁺ ion. However, the moment the light source was removed, an immediate and robust recovery of the 480 nm absorbance ensued. This near‐instantaneous reversion, manifesting within mere seconds, suggests that even if the photodissociation occurs, the Cu²⁺ ion remains in close proximity to the macrocyclic coordination site, waiting for the rapid re‐establishment of the E‐1^⚫ Cu(II) complex. So far, this is the fastest revert process of light‐free Z→E isomerization we have observed by the photolysis.

In contrast, the other three metal ions displayed significantly disparate temporal behaviors in 72 h evolution (Figure S7). Following the addition of Pb²⁺ to the PSS₃₆₅ solution, an initial spectroscopic surge was observed, yet the subsequent evolution toward the E‐complex spectrum remained protracted, spanning several hours. The Ni²⁺ and Co²⁺ systems were even more sluggish. These divergent timescales, ranging from seconds (Cu²⁺) to hours (Pb²⁺) and days (Ni²⁺/Co²⁺), highlight the unique role of Cu²⁺. This kinetic hierarchy serves as a valuable benchmark for the development of temporally programmable molecular switches, where the switching rate can be precisely tuned by the choice of the metal effector.

2.6. E‐1 Regenerated by Using Metal Scavenger

To evaluate the structural robustness and reusability of the glycomacrocycle scaffold, we employed QuadraPure IDA, a universal metal scavenger featuring iminodiacetic acid functional groups, to sequester the coordinated metal ions from the complex solutions. As illustrated in Figure 6, the addition of the scavenger led to a significant attenuation of the complex‐specific bands, confirming the successful release of the metal‐free ligand. This in situ regeneration provides critical evidence that the macrocyclic framework remains chemically intact after coordination and potential structural transformations.

FIGURE 6.

UV–Vis spectra of the E‐1, E‐1·M(II) complex, and E‐1· M(II) complex treated with QuadraPure IDA after 1 week. (a) with Cu(ClO₄)₂ and (b) with Co(ClO₄)₂.

Notably, the regeneration process highlights a fundamental distinction between the metal systems. For the Co²⁺ system, the absorption profile reverts to its initial E‐1 state, confirming that the metal ion was simply “nested” within the rigid, acetal‐protected cavity without altering the covalent scaffold (Figure 6b). In contrast, for the Cu²⁺ system, the scavenging process reveals the emergence of a chemically remodeled diol‐macrocycle. Although the spectrum reverts to an E‐isomer profile, this regenerated species represents a transition from the acetal‐protected precursor to a diol environment. The restoration of the azobenzene signature across the two systems proves that the observed spectroscopic shifts, including the rapid reversion unique to Cu²⁺, do not stem from the macrocyclic degradation. This selection of metal effector allows a recyclable and programmable logic application.

3. Conclusion

In this work, we have synthesized an azobenzene‐embedded glycomacrocycle that exhibits a remarkable light‐free metal‐ion‐regulated Z→E isomerization, spanning from mere seconds to weeks. Unlike previously reported systems, the expansive 26‐membered ring and its inherent structural flexibility accommodate a diverse range of metal salts (Cu²⁺, Ni²⁺, Co²⁺, and Pb²⁺), each dictating a unique isomerization kinetics and spectroscopic profile. Our UV–Vis and ESI–MS investigations reveal that this kinetic diversity is not only a function of ionic radius, but is fundamentally regulated by a metal‐induced chemical remodeling of the macrocyclic scaffold. Specifically, Cu²⁺ and Ni²⁺ act as Lewis acid catalysts to promote the in situ hydrolysis of the isopropylidene acetals, a process that dramatically simplifies the conformational landscape to one singular pure conformation. CREST‐based conformational search analyses corroborate this interpretation: the acetal‐protected scaffold occupies a fragmented energy ensemble PES, whereas deprotection of the acetal anchors triggers a thermodynamic singularity, funneling the entire conformational population into a single expanded‐stretched global minimum inaccessible to the protected precursor—a ‘conformational purification’ driven by Lewis acid catalysis rather than light. This topological transformation of the PES provides a unified mechanistic rationale for the spectroscopic divergence among the four metal ions at the E‐form terminus.

The unprecedented second‐timescale isomerization kinetics observed in the Cu²⁺ system are further investigated by a photokinetic experiment, which demonstrates near‐instantaneous re‐establishment of the E‐1^⚫ Cu(II) complex upon cessation of UV irradiation. Moreover, the system's reversibility to the metal‐free E‐isomer, achieved by employing a commercially available metal scavenger, establishes a recyclable, programmable molecular switch platform. Collectively, the metal‐dependent light‐free isomerization kinetic hierarchy, the Lewis‐acid‐driven conformational remodeling, and the demonstrated reversibility position this glycomacrocycle as a versatile design platform. The scaffold has the potential to be the probe candidate for biological environments with metal‐ion gradients. Beyond biomedical contexts, the seconds to weeks’ broad time magnitude kinetic tunability by simple metal ion selection is directly translatable to programmable molecular logic gates, where each metal ion encodes a distinct time‐domain, reversible or non‐reversible outputs. The glycomacrocycle's molecular architecture, separating the photoresponsive, coordination, and protecting‐group functions into discrete structural elements, further invites the integration of this building block into metal‐organic frameworks or stimuli‐responsive polymers. These prospects, rooted in the molecular design principles established here, illustrate new chemically programmable molecular switches operating entirely in the dark.

4. Experimental Section

4.1. General Information

The synthetic protocol of azobenzene glycomacrocycle E‐1 is given below. All commercially available compounds and solvents were purchased from Sigma–Aldrich affiliated to Merck KGaA, Carlo Erba Reagents, and TCI chemicals without further purification. QuadraPure IDA metal scavenger was purchased from Sigma–Aldrich affiliated to Merck KGaA, and used without pre‐treatments. ¹H, ¹³C, COSY, HSQC NMR data were acquired by JEOL ECS‐400 spectrometer with a broadband code. High‐resolution mass spectra (HRMS) were performed on a Bruker maXis mass spectrometer by the SALSA platform from the ICOA laboratory. Optical rotation [α]_D ²⁵ was measured using a JASCOP‐2000 polarimeter at 589 nm where no absorption occurred for all compounds. Spectroscopic measurements were performed with spectroscopic grade solvents from Carlo Erba Reagents and Sigma‐–Aldrich affiliated to Merck KGaA. Absorption spectra at room temperature were recorded using Agilent Cary‐4000 or Cary‐5000 UV spectrometers. The samples were kept in a dark environment to avoid possible isomerization.

4.2. Synthesis

(5‐(Hydroxymethyl)‐2,2‐dimethyl‐1,3‐dioxan‐5‐yl)methyl‐(E)‐2‐(2‐((2‐(prop‐2‐yn‐1‐yloxy)phenyl)diazenyl)phenoxy)acetate (4) and (2,2‐dimethyl‐1,3‐dioxane‐5,5‐diyl)bis(methylene)bis(2‐(2‐((E)‐(2‐(prop‐2‐yn‐1‐yloxy)phenyl)diazenyl)phenoxy)acetate (5). To a solution of compounds 3 (630 mg, 2.03 mmol) and 2 (349.4 mg, 2.03 mmol) in CH₂Cl₂ (4 mL) was added a solution of the EDC (389.4 mg, 2.03 mmol) and DMAP (26.4 mg, 0.20 mmol) in dry CH₂Cl₂ at 0°C. The solution was stirred (0°C to rt) until the reaction was finished as indicated by TLC. The reaction mixture was diluted with CH₂Cl₂ and washed with water (3 times). The organic layer was separated, dried over MgSO₄, and concentrated. The residue was purified by flash chromatography (CH₂Cl₂/EtOAc = 8/2) to afford 4 (203 mg, 41%) and 5 (106 mg, 24%) as red solids. Compound 4: R_f = 0.42 (Petroleum ether/EtOAc = 5/5); [α]_D: −0.7 (c = 0.5, CH₃CN); mp: 77°C; ¹H NMR (400 MHz, CDCl₃): δ 7.66 (dd, J = 8.0, 2.0 Hz, 1H, H_Ph), 7.63 (dd, J = 8.0, 1.6 Hz, 1H, H_Ph), 7.47–7.39 (m, 2H, H_Ph), 7.26–7.24 (m, 1H, H_Ph), 7.10–7.07 (m, 2H, H_Ph), 7.03 (d, J = 8.0 Hz, 1H, H_Ph), 4.95–4.94 (m, 4H), 3.65 (d, J = 12.0 Hz, 1H, OCH), 3.61 (d, J = 12.0 Hz, 1H, OCH), 3.40 (s, 2H, CH₂), 2.56 (t, J = 2.4 Hz, 1H, CH≡), 1.38 (s, 3H, CH₃), 1.36 (s, 3H, CH₃); ¹³C NMR (100 MHz, CDCl₃): δ 169.6 (C═O); 155.3, 155.2, 144.4, 143.4 (C_q); 132.4, 132.3, 122.6, 122.3, 118.1, 117.7, 115.8, 115.6 (CH); 98.6 (C_q), 78.6 (C≡), 76.3 (CH≡), 67.5, 64.5, 64.4, 62.0, 57.8 (CH₂); 39.0 (C_q), 24.8, 22.7 (CH₃); HRMS (ESI) m/z: Calcd for C₂₅H₂₈N₄O₇ [M+H]⁺: 469.1975, Found: 469.1970. Compound 5: R_f = 0.29 (Petroleum ether/EtOAc = 5/5); [α]_D: −6.5 (c = 0.5, CH₃CN); mp: 84°C; ¹H NMR (400 MHz, CDCl₃): δ 7.65 (dd, J = 7.6, 2.0 Hz, 2H, H_Ph), 7.63 (dd, J = 8.0, 1.6 Hz, 1H, H_Ph), 7.45–7.35 (m, 4H, H_Ph), 7.23 (d, J = 8.0 Hz, 2H, H_Ph), 7.08–7.00 (m, 4H, H_Ph), 6.98 (d, J = 7.2 Hz, 2H, H_Ph), 4.94–4.89 (m, 8H), 4.12 (s, 4H, 2xCH₂), 3.57 (s, 4H, 2xCH₂), 2.55 (t, J = 2.4 Hz, 2H, 2xCH≡), 1.26 (s, 6H, 2xCH₃); ¹³C NMR (100 MHz, CDCl₃): δ 168.8 (C═O); 155.4, 155.2, 143.5, 143.2 (C_q); 132.2, 122.6, 122.3, 118.0, 117.7, 115.9, 115.5 (CH); 98.8 (C_q), 78.7 (C≡), 76.3 (CH≡), 67.3, 63.9, 62.0, 57.8 (CH₂); 37.4 (C_q), 23.6 (CH₃); HRMS (ESI) m/z: Calcd for C₄₂H₄₁N₄O₁₀ [M+H]⁺: 761.2823, Found: 761.2819.

4.2.1. Compound 7

To a solution of 6 (200 mg, 0.427 mmol) and 4 (267.2 mg, 0.427 mmol) in CH₂Cl₂/H₂O (4/1, 12.5 mL), were added CuSO₄ ^.5H₂O (53.3 mg, 0.213 mmol) and ascorbic acid (75.2 mg, 0.427 mmol). The mixture was stirred at rt for 16 h until the reaction was finished as indicated by TLC. The reaction mixture was diluted with CH₂Cl₂ and washed with water (3 times). The organic layer was separated, dried over MgSO₄, and concentrated. The residue was purified by flash chromatography (Petroleum ether/EtOAc = 5/5) to get the product 7 (417 mg, 87%) as a red solid. R_f = 0.58 (Petroleum ether/EtOAc = 3/7); [α]_D: + 16.3 (c = 0.5, CH₃CN); mp: 92°C; ¹H NMR (400 MHz, CDCl₃): δ7.62 (s, 1H, CH‐triazole), 7.54‐7.50 (m, 2H, H_Ph), 7.34–7.19 (m, 22H, H_Ph), 7.13 (dd, J = 8.8, 0.8 Hz, 1H, H_Ph), 7.00–6.92 (m, 3H, H_Ph), 5.48 (d, J = 1.6 Hz, 1H, H₁), 5.37 (s, 2H, CH₂), 4.96–4.83 (m, 3H), 4.84 (s, 2H, CH₂), 4.64 (d, J = 12.4 Hz, 1H), 4.56 (d, J = 12.4 Hz, 1H), 4.55‐4.53 (m, 3H), 4.48 (d, J = 11.2 Hz, 1H), 4.36 (dd, J = 11.2, 2.4 Hz, 1H, H₆), 4.36 (dd, J = 11.6, 5.6 Hz, 1H, H_6’), 4.22 (s, 2H, CH₂), 4.20–4.18 (m, 1H, H₅), 3.92 (t, J = 2.0 Hz, 1H, H₂), 3.81–3.80 (m, 2H, H_3,4), 3.56 (d, J = 12.4 Hz, 2H, 2xCH), 3.50 (d, J = 12.4 Hz, 2H, 2xCH), 3.28 (s, 2H, CH₂), 2.16 (s large, 1H, OH), 1.29 (s, 3H, CH₃), 1.28 (s, 3H, CH₃); ¹³C NMR (100 MHz, CDCl₃): δ169.6, 166.0 (CO); 155.6, 155.2, 144.7, 143.5, 143.4, 138.0, 137.9, 137.7, 133.9 (C_q); 132.6, 132.2, 131.4, 129.3, 128.6, 128.3, 128.1, 128.0, 127.8 (CH); 124.7 (CH‐triazole), 122.6, 122.0, 118.2, 118.1, 115.8, 115.7 (CH); 98.6 (C_q), 85.4 (C₁), 80.2 (CH), 75.9 (C₂), 75.2 (CH₂), 74.1 (CH), 72.2, 72.1 (CH₂); 70.6 (C₅); 67.5 (CH₂), 65.0 (C₆), 64.5, 64.0, 62.4, 61.9, 50.6 (CH₂), 38.9 (C_q), 25.0, 22.4 (CH₃); HRMS (ESI) m/z: Calcd for C₆₀H₆₄N₅O₁₃S [M+H]⁺: 1094.4221, Found: 1094.4229.

4.2.2. Glycomacrocycle E‐1

A mixture of glycosyl donor‐acceptor pair 7 (87 mg, 0.08 mmol) in CH₂Cl₂ (2.8 mL) was treated with 4Å molecular sieve (400 mg) under Argon for 6 h at rt. To this mixture, cooled to −78°C, was added NIS (35 mg, 0.16 mmol) and TfOH (7 µL, 0.08 mmol). The mixture was kept stirring at ‐78°C for 4 h until the reaction was finished as indicated by TLC. The solution was diluted with CH₂Cl₂, then filtered. The organic layer was washed with 20% aq. Na₂S₂O₃ and water (3 times), dried over MgSO₄, concentrated. The residue was purified by CombiFlash Rf+ (Petroleum ether/EtOAc 7/3 – 5/5) to give the desired macrocycle E‐1 (29 mg, 39%) as a red solid. The E/Z mixture in CDCl₃ was heated at 60°C for 48 h to E‐1 for the NMR characterization and optical rotation measurement. R_f = 0.52 (Petroleum ether/EtOAc = 5/5); [α]_D: −326.4 (c = 0.25, CH₃CN); mp: 77°C; ¹H NMR (400 MHz, CDCl₃): 7.78 (s, 1H, CH‐triazole), 7.48‐7.46 (m, 2H, H_Ph), 7.38‐7.16 (m, 17H, H_Ph), 7.01–6.93 (m, 4H, H_Ph), 5.46‐5.39 (m, 2H, OCH₂), 5.08‐4.98 (m, 2H, NCH₂), 4.95‐4.86 (m, 3H), 4.78 (d, J = 12.4 Hz, 1H), 4.66 (d, J = 12.4 Hz, 1H), 4.57‐4.54 (m, 2H), 4.47 (d, J = 12.0 Hz, 1H), 4.41 (dd, J = 12.0, 2.4 Hz, 1H), 4.19‐4.10 (m, 3H), 3.82 (s, 1H, H₁), 3.73‐3.68 (m, 2H), 3.62 (s, 2H, CH₂), 3.51–3.42 (m, 3H), 3.38 (dd, J = 9.6, 2.8 Hz, 1H), 3.24‐3.19 (m, 1H, H₅), 3.10 (d, J = 9.6 Hz, 1H), 1.35 (s, 3H, CH₃), 1.29 (s, 3H, CH₃); ¹³C NMR (100 MHz, CDCl₃): δ 169.1, 165.9 (CO); 154.7, 154.0, 144.6, 144.1, 143.8, 138.4, 137.9, 137.7 (C_q); 132.2, 131.9, 128.7, 128.4, 128.3, 128.2, 128.1, 128.0, 127.8, 127.7, 127.6 (CH); 124.7 (CH‐triazole), 122.3, 122.0, 120.0, 118.7, 115.6, 114.6 (CH); 101.1 (C₁), 98.8 (C_q), 81.8 (CH); 79.5, 74.5 (CH), 74.2 (CH₂); 73.8, 73.0 (CH); 71.5, 68.6, 66.9, 65.3, 64.6, 64.1, 63.1, 61.5 (CH₂); 50.9 (NCH₂), 37.9 (C_q), 23.9, 23.6 (CH₃). HRMS (ESI) m/z: Calcd for C₅₄H₅₈N₅O₁₃ [M+H]⁺: 984.4031, Found: 984.4027.

4.3. Mass Spectrometry Studies

Studies of the interactions of the glycoazomacrocycle with the four metals were performed on a 3D ion trap instrument (Bruker Amazon Speed ETD), allowing MSⁿ experiments to be performed when necessary. Complexes were generated in the gas phase by electrospray ionization. To this end, 1/5 mixtures of glycoazomacrocycle/M(ClO₄)₂ (M(ClO₄)₂ at 5 × 10⁻⁵ M) were prepared. Starting from 10⁻³ M stock solutions of E‐1, irradiated E‐1, and M(ClO₄)₂ solubilized in acetonitrile and purified water, respectively, the final mixtures of E‐1/M(ClO₄)₂ and Z‐1/M(ClO₄)₂ prepared in pure acetonitrile were introduced in the electrospray source by a syringe pump (4 to 5 µL/min). Typical experimental conditions were as follows:

ESI conditions: capillary voltage:—4500 V; End plate offset: −500 V; Dry gas: 7 L/min / Dry gas temperature: 200°C, Nebulizer gas: 11 PSI; Cap exit 140 V and Trap drive 52.4. All spectra were recorded in the “Maximum Resolution mode”.

MS analysis: ICC mode: “on” (50000 ions) and acquisition time: auto. For MSⁿ analysis: ICC mode “off” / accumulation time 1 ms to reach c.a. 40000 ions for the selected precursor/Isolation window 3.5 Da / Fragmentation delay 40 ms/ amplitude of fragmentation: 0–0.3 / Low mass cutoff: auto

4.4. QuadraPure IDA Experiment Procedures

QuadraPure IDA metal scavenger, a chelating agent, is capable of complexing with Fe, Al, Ga, In, Cu, V, Pb, Ni, Zn, Cd, Be, Mn, Co, Sr, Ba according to the specification sheet by Sigma–Aldrich, which perfectly covers the metal ions used in this work. We applied it to two complex solutions to check the chelating caption ability and recover the ligand. The macroporous beads span the particle size from 350–700 µm, which cannot be counted stoichiometrically. Cu(II) and Co(II) perchlorate salted samples were selected as the fastest and the slowest dynamic representatives. A certain number of beads below 1/3 liquid level were added. The resulting mixture was stirred for 1 week at room temperature to allow the scavenger to fully bind to the metal ions. The scavengers were then removed by filtration with microfilter syringe just once, avoiding evaporation.

4.5. Conformational Searches

Conformational searches were performed using the CREST (Conformer‐Rotamer Ensemble Sampling Tool) program (version 3.0) developed by Grimme and coworkers, employing the GFN2‐xTB semi‐empirical tight‐binding method throughout. All calculations were conducted for the E _{isopropylidene}, E _diol and Z _{isopropylidene} forms of glycomacrocycle 1. For each species, conformational sampling was carried out both in the gas phase and in acetonitrile using the Analytical Linearized Poisson‐Boltzmann (ALPB) solvation model [35] to assess the influence of the solvent environment and suppress hydrogen bonding artefacts inherent to gas phase sampling. The iMTD‐GC (iterative metadynamics with geometry optimization and genetic Z‐matrix crossing) workflow was employed with an energy window of 8 kcal/mol above the lowest conformer. Metadynamics run times were set sufficiently long to ensure convergence of the ensemble for the macrocyclic scaffold. Sampling of nonconvalent complexes and aggregates (NCI mode) of CREST was opened for all entries. The resulting conformational ensembles were analyzed by Boltzmann‐weighted population distribution at 298.15K, and pairwise root mean square deviation (RMSD) matrices were computed on the heavy‐atom framework to assess topological diversity among low‐energy conformers. The global partition function Q was extracted directly from the CREST output. Final geometries of representative conformers were visualized using VMD (Visual Molecular Dynamics, version 2.0.0a6) [36]. Cartesian coordinates of all reported conformers are provided in the Supporting Information.

Conflicts of Interest

The authors have no conflicts of interest to declare.

Supporting Information

Experimental details (including NMR spectra, photochemical and photophysical characterizations, mass spectrometry and conformational search calculations) are given in the supporting information.

Supporting information

Supporting File: chem71333‐sup‐0001‐SuppMat.pdf

Data Availability Statement

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