Trapping an Ester Hydrate Intermediate in a π-Stacked Macrocycle with Multiple Hydrogen Bonds

Ester hydrates, as the intermediates of the esterification between acid and alcohol, are very short-lived and challenging to be trapped. Therefore, the crystal structures of ester hydrates have rarely been characterized. Herein, we present that the mono-deprotonated ester hydrates [CH3OSO2(OH)2]−, serving as the template for the self-assembly of a π-stacked boat-shaped macrocycle (CH3OSO2(OH)2)0.67(CH3OSO3)1.33@{[ClLCoII]6}·Cl4·13CH3OH·9H2O (1) (L = tris(2-benzimidazolylmethyl) amine), can be trapped in the host by multiple NH···O hydrogen bonds. In the solution of CoCl2, L, and H2SO4 in MeOH, HSO4− reacts with MeOH, producing [CH3OSO3]− via the ester hydrate intermediate of [CH3OSO3(OH)2]−. Both the product and the intermediate serve as the template directing the self-assembly of the π-stacked macrocycle, in which the short-lived ester hydrate is firmly trapped and stabilized, as revealed by single-crystal analysis.


Introduction
Esters are important chemicals because of their wide applications in a variety of products ranging from medicine to biodiesel [1][2][3]. Esters can be produced via esterification reactions between the corresponding acids and alcohols via an intermediate state involving two transition states [4,5]. To trap the short-lived intermediate species, some special strategies have to be adopted. Previous reports have demonstrated that reactive species can be stabilized using the cavities of porous materials [6][7][8][9][10][11][12][13][14][15][16]. For instance, air and moisturereactive white phosphorus can be safely stored in some tetrahedral cages, as demonstrated by Nitschke et al. and Wu et al. [17][18][19]. Furthermore, various reactive intermediates, including a sulfenic acid [20][21][22], a selenenic acid [23,24], an S-nitrosothiol [25][26][27], and a Senitrososelenol [28], have been isolated for the "peripheral steric protection" of pre-designed molecular cavities [29]. Inspired by the exceptional functionalities of cavities on stabilizing reactive intermediates, Fujita et al. performed a simple and ubiquitous reaction between an amine and an aldehyde in a porous network and successfully observed a trapped transient hemiaminal via single-crystal X-ray diffraction analysis (SCXRD), which usually is a very short-lived intermediate [30]. Different from the above-mentioned strategies, reactive intermediates may also act as the templates for directing the assembly of molecular cages or macrocycles. As a result, the reactive intermediates may be stabilized and trapped by the resulting assemblies, thus allowing structural characterizations. Recently, our group started to using the tripodal ligands tris (2-benzimidazolylmethyl) amine and tris(2-naphthimidazolylmethyl) amine to construct hierarchical assemblies based on π-stacked cages [31][32][33][34][35]. Here, we report the in situ-generated ester hydrate intermediate [CH 3 OSO 2 (OH) 2 ] − templates and the assembly of a boat-shaped π-stacked macrocycle, in which the short-lived ester hydrate is firmly trapped and stabilized, thus enabling the structural determination of the intermediate via single-crystal X-ray analysis.   Keeping the reactant solution undisturbed at 10 • C for seven days, purple cubic crystals of (CH 3 OSO 2 (OH) 2 ) 0.67 (CH 3 OSO 3 ) 1.33 @{[ClLCo II ] 6 }·Cl 4 ·13CH 3 OH·9H 2 O (1) could be harvested. The successful trapping of the ester hydrate [CH 3 OSO 2 (OH) 2 ] − was further indicated via single-crystal X-ray analysis of compound 1. The formula of compound 1 was determined using a combination of single-crystal X-ray crystallography (Table S1) and TG analysis ( Figure S2). Single-crystal X-ray analysis of compound 1 revealed a πstacked boat-shaped macrocycle composed of six [ClLCo II ] + ions, in which the ester hydrate intermediate [CH 3 OSO 2 (OH) 2 ] − and the esterification product [CH 3 OSO 3 ] − were trapped in a ratio of 1:2.

Physical-Property Characterization of Compound 1
The phase purity of compound 1 was confirmed via the powder X-ray diffusion pattern (PXRD) measurement ( Figure S5). In the infra-red (IR) spectrum of compound 1, the absorption band at 1395 cm −1 can be assigned to υ (S=O) and the band at 1022 cm −1 can be assigned to υ (C-O) [37,38] (Figure 3a). The existence of S VI was confirmed using the X-ray photoelectron spectroscopy (XPS) study. The observed peaks at 168.15 and 169.25 eV corresponded to S VI 2p 3/2 and S VI 2p 1/2 , respectively [39,40] (Figure 3b). A temperature-dependent magnetization study of compound 1 was also performed under 1 kOe field in the 2-300 K range. The χ m T value of 14.95 cm 3 K mol −1 at 300 K was higher than the spin-only value of six isolated high-spin Co II ions (11.25 cm 3 mol -1 K) (Figure 3c) [41][42][43]. This result implies an obvious unquenched orbital contribution. Upon cooling, the χ m T value kept almost constant until 10 K, and then began to decrease, reaching a value of 12.14 cm 3 mol -1 K at 2 K, implying very weak antiferromagnetic couplings between Co II ions. In the range of 300-2 K, the magnetic susceptibility data followed Curie-Weiss law, giving θ = -1.13 K and C = 14.79 cm 3 K mol -1 , confirming the dominant weak antiferromagnetic interactions. To give further insights into the magnetism of compound 1, the field-dependent magnetizations were measured. The magnetization increased slowly with the increasing field and reached a value of 19.10 Nβ at 80 kOe without obvious hysteresis (Figure 3d), which is consistent with the weak antiferromagnetic couplings between Co II ions [44,45].

Physical-Property Characterization of Compound 1
The phase purity of compound 1 was confirmed via the powder X-ray diffusion pattern (PXRD) measurement ( Figure S5). In the infra-red (IR) spectrum of compound 1, the absorption band at 1395 cm −1 can be assigned to υ(S=O) and the band at 1022 cm −1 can be assigned to υ(C-O) [37,38] (Figure 3a). The existence of S VI was confirmed using the X-ray photoelectron spectroscopy (XPS) study. The observed peaks at 168.15 and 169.25 eV corresponded to S VI 2p3/2 and S VI 2p1/2, respectively [39,40] (Figure 3b). A temperature-dependent magnetization study of compound 1 was also performed under 1 kOe field in the 2-300 K range. The χmT value of 14.95 cm 3 K mol −1 at 300 K was higher than the spin-only value of six isolated high-spin Co II ions (11.25 cm 3 mol -1 K) (Figure 3c) [41][42][43]. This result implies an obvious unquenched orbital contribution. Upon cooling, the χmT value kept almost constant until 10 K, and then began to decrease, reaching a value of 12.14 cm 3 mol -1 K at 2 K, implying very weak antiferromagnetic couplings between Co II ions. In the range of 300-2 K, the magnetic susceptibility data followed Curie-Weiss law, giving θ = -1.13 K and C = 14.79 cm 3 K mol -1 , confirming the dominant weak antiferromagnetic interactions. To give further insights into the magnetism of compound 1, the field-dependent magnetizations were measured. The magnetization increased slowly with the increasing field and reached a value of 19.10 Nβ at 80 kOe without obvious hysteresis (Figure 3d), which is consistent with the weak antiferromagnetic couplings between Co II ions [44,45].

Materials and Physical Measurements
The ligand tris(2-benzimidazolylmethyl) amine (L) was synthesized according to the procedure reported in the literature [46], and all the other reagents were commercially

Materials and Physical Measurements
The ligand tris(2-benzimidazolylmethyl) amine (L) was synthesized according to the procedure reported in the literature [46], and all the other reagents were commercially obtained and used without further purification. Powder X-ray diffraction (PXRD, Miniflex 600, Akishima, Rigaku, Tokyo, Japan) patterns were performed on a Rigaku Miniflex 600 diffractometer with Cu-Kα radiation using flat plate geometry. High Resolution Mass Spectrometry (HR-MS, Impact II UHR-TOF, Bruker, Billerica, MA, USA) measurements were performed on a DECAX-30000 LCQ Deca XP system. X-ray photoelectron spectroscopy (XPS, Axis Supra, Shimadzu, Manchester, United Kingdom) studies were performed on an AXIS SUPRA Kratos system and the C1s line at 284.8 eV was used as the binding energy reference. Thermogravimetric analysis (TGA/DSC 1, Mettler Telodo, Zurich, Switzerland) was performed on a Mettler-Toledo TGA/DSC 1 system with a heating rate of 10 K/min under an argon atmosphere. Fourier-transform infrared (FTIR, Nicolet iS 50, Thermo Fisher, Waltham, MA, USA) spectra were recorded in the range of 500-4000 cm −1 on a Thermo Nicolet is50 FT-IR spectrometer at room temperature. Magnetic measurements (MPMS-5S SQUID, Quantum Designwere, San Diego, CA, USA) were performed on an MPMS-5S SQUID magnetometer.

Crystallography
Single-crystal X-ray data were harvested on a Bruker D8 Venture diffractometer with Mo-K α radiation at 200 K. Structures were solved using intrinsic phasing method with SHELXT and refined via the full-matrix least-squares technique on F 2 with SHELXTL 2014 program [47]. The graphical user interface for the solving and refining process adopted Olex2 software [48]. In the process of solving crystals, some reasonable restriction commands such as DFIX, SIMU, and OMIT were used. All the H atoms were geometrically generated and refined using a riding model. The X-ray crystallographic coordinates for structures reported in this article have been deposited at the Cambridge Crystallographic Data Centre (CCDC) under deposition numbers 2271453. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/ data_request/cif (accessed on 22 June 2023). Detailed crystallographic data are listed in Table S1.

Computational Methods
We calculated the complete reaction pathway of methanol and vitriol (Figure 1). The geometry structures and the frequency calculations of reactants, products, and intermediate structures, as well as transition-structures (TS) have been optimized using the M062X method and the 6-31G* basis set [49]. Using the same method and basis set, the vibrational frequency calculations were carried out to confirm the local minima intermediate structure and the TS (one negative eigenvalue) on the potential energy surface (PES). The M062X method and the 6-31G* basis set were employed to calculate the intrinsic reaction coordinates (IRCs) [50], verifying that the TS were indeed the lowest saddle points for the expected structure connecting the reaction path. All calculations were performed using Gaussian 16 program [51].

Conclusions
In conclusion, the reactive ester hydrate intermediate [CH 3 OSO 2 (OH) 2 ] − has been successfully trapped and stabilized as the template for directing the assembly of π-stacked boat-shaped macrocycles. The intermediate is firmly trapped in the macrocycle by multiple NH···O hydrogen bonds. This achievement allowed for the structural determination of the intermediate using single-crystal X-ray analysis. This strategy for stabilizing reactive species may be planted to other systems, and, thus, provides a new means of giving insights into the structural transformations that occur during chemical reactions.

Data Availability Statement:
All data related to this study are presented in this publication.