Open Access Article
Jianmin
Jiao
a,
Heng
Li
a,
Wang
Xie
a,
Yue
Zhao
b,
Chen
Lin
*a,
Juli
Jiang
*a and
Leyong
Wang
*a
aState Key Laboratory of Analytical Chemistry for Life Science, Jiangsu Key Laboratory of Advanced Organic Materials, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing, 210023, China. E-mail: linchen@nju.edu.cn; jjl@nju.edu.cn; lywang@nju.edu.cn
bState Key Laboratory of Coordination Chemistry, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing, 210023, China
First published on 26th September 2023
X-ray crystallography is the most reliable method for structure elucidation and absolute configuration determination of organic molecules based on their single-crystal forms. However, many analytes are hard to crystallize because of their low melting points (an oily state at room temperature) or conformational flexibility. Here, we report the crystallization of a macrocycle, CTX[P(O)Ph] (host), which is a cyclotrixylohydroquinoylene (CTX) derivative, with 26 oily organic molecules (guests), which is applied for the structural determination of the guest with X-ray crystallography. With the aid of the host, CTX[P(O)Ph], the guest molecules were well-ordered with full occupancy in crystal structures. In most cases, at least one guest structure without any disorder could be observed; solvent masking was not necessary for the single crystal X-ray structural analysis, and thus the structures of the guests could be successfully determined, and the absolute configuration could be assigned reliably for chiral guests with this method. The crystallization mechanism was further discussed from theoretical and experimental perspectives, suggesting that the negative electrostatic potential surface of CTX[P(O)Ph] and noncovalent interactions between the host and guest were crucial for the ordered arrangements of the guest.
Recently, several strategies have been proposed to help study molecules that pose challenges in obtaining single crystals and address the limitations discussed above.3 The most famous one may be the “crystalline sponges (CSs)” method developed by the Fujita group. Small organic guest molecules are absorbed into the pores or channels of porous metal–organic frameworks (MOFs); the absorption is driven by strong π–π, CH–π, and charge–transfer interactions between guest molecules and the electron-deficient π-plane of MOFs.4 The orderliness is transferred from the crystalline networks (host molecules) to the pores, and finally, to the guest molecules so that the guest molecules become well-ordered and can be observed by SCXRD. The absolute configuration of the guests can also be well-defined because of the anomalous scattering of the heavy atoms in the networks.5 In addition, the CSs method can be performed with SCXRD analysis at nanogram to microgram scales of the guests, which allows this method to be combined with liquid chromatography4a or gas chromatography6 and ensures its application in studying complex samples and samples available in small amounts. The Yaghi group also used MOFs to determine the structure of a series of guest molecules.7 Coordination bonds were introduced between the MOFs and guest molecules to anchor the guests and further ensure their ordered arrangement. The chirality of the MOF backbone provides a reference for the assignment of the absolute configuration of the guests. Apart from MOFs, hydrogen-bonded frameworks, which are another kind of porous crystalline network, constructed by guanidinium and organosulfonate ions, are capable of encapsulating a range of guest molecules into the pores to elucidate their structure.8 In addition, non-network structures, such as supramolecular macrocycles9 and metal-macrocycles10 with inherent cavities and large organic molecules generating channels when packing9d,11 can also form co-crystals or clathrates with various guest molecules. For instance, tetraaryladamantanes developed by the Richert group have been used as crystallization chaperones to determine the absolute configuration of chiral molecules.11b In general, the suitable sizes and shapes of pores or cavities are indispensable in the design of host molecules suitable for the crystallization with guests because they can accommodate guest molecules, and more importantly, help orient the guests into an array in a periodic manner. However, there are some problems in such systems. For instance, only some fraction of the cavities are occupied by the guest molecules, while the other part remains vacant or is accommodated by residual solvent molecules. Some guest molecules have two or more different orientations in the pores. As a result, in some cases, the guest molecules were partially or even totally disordered,9b or their occupancy was considerably lower than 100%,10a,12 which greatly hindered molecular structure and the determination of its absolute configuration.
Herein, a kind of nonporous phosphorylated macrocyclic host molecule named CTX[P(O)Ph] is reported that can help a range of oily organic compounds crystallize, and thus, aid the elucidation of their structure and absolute configuration (Fig. 1). It was found that most kinds of common organic guests could crystallize with CTX[P(O)Ph] as the host, suggesting its potential application for diverse analytes. Such multi-component crystals can be classified into clathrates and co-crystals. Clathrates are formed by the inclusion of guest molecules in the host molecule assembly with cavities or channels. In co-crystals, both the host and guest molecules are involved in the construction of hydrogen-bonded frameworks.13 In the packing mode of the crystal structures of the guest molecules@CTX[P(O)Ph], the guest molecules are present in the layers or channels formed by the assembly of the host molecules, and no hydrogen-bonded framework is found. Therefore, the crystal structures in this work are clathrates. As a macrocycle, CTX[P(O)Ph] could trap the guest molecules above its cavity through host–guest interactions, which lowered the guest molecules' degrees of motion freedom and induced them to form an array in a uniform fashion. On the other hand, different from traditional macrocycles that have high symmetry and well-defined cavities, such as pillararene14 and cucurbituril,15 the small size of the cavity of CTX[P(O)Ph] was a censer-like shape that could not fully accommodate guest molecules and the guest molecule avoid existing in the cavity in low occupancy or disordered form. As a consequence, there was at least one guest molecule without any disorder in most cases, and the occupancy of all guest molecules was 100%. Such a regular alignment was in favor of determining the chemical structure, especially the absolute configuration of oily compounds.
![]() | ||
| Fig. 1 The crystallization of CTX[P(O)Ph] with oily compounds for structural elucidation. One host–guest complex is shown as an example. | ||
![]() | ||
| Fig. 2 The crystal structure of guest molecules obtained by crystallization with CTX[P(O)Ph] (OPTEP plots with 50% probability). Flack values (x) are given for the chiral guests. | ||
NMR spectroscopy is a powerful approach for structure determination of organic compounds. However, there are still some organic molecules whose structures are difficult to define with NMR spectroscopy. For instance, 4-chloro-3-fluoroanisole (GM 12) and 3-bromo-4-chloroanisole (GM 13) are hard to distinguish from their corresponding isomers (3-chloro-4-fluoroanisole and 4-bromo-3-chloroanisole) with NMR spectroscopy because of the similar electron-withdrawing effect of halogen atoms. They are both challenging to crystalize because of their oily state at room temperature. In contrast, their structures could be well-defined after crystallization with CTX[P(O)Ph] (Fig. 4). The halogen atoms could be identified easily and accurately based on the intensity of electron peaks and the bond length. The simple examples provided here demonstrate that this method can accurately define some molecules whose structures are hard to determine by NMR spectroscopy.
![]() | ||
| Fig. 4 Crystallization with CTX[P(O)Ph] can define some guest structures that are challenging to determine by NMR spectroscopy. | ||
As shown in Fig. 2 and S19–S26,† eight chiral guests were studied, including three pairs of enantiomers, 2-phenylethanol (GM 19–20), 2-pentanol (GM 21–22), and epichlorohydrin (GM 23–24), and two chiral molecules, methyl-5-oxo-L-prolinate (GM 25) and R-styrene oxide (GM 26). All crystals existed in a triclinic or monoclinic crystal system. In each crystal structure, two or four independent guest molecules were in the minimal asymmetric unit and all these guest molecules exhibited the same chirality with limited restraints. Flack parameters close to 0 indicate an unambiguous assignment of absolute configuration, and a threshold of 0.1 is commonly used.11b For GM 19–22 and GM 25–26, the flack values calculated using Parsons' method were within the acceptable threshold for credible absolute configuration assignment.18 Inversion tests were carried out for GM 23–24 because their flack values were slightly larger than 0.1.11b,19 Much higher flack values of the opposite chirality after inversion confirmed the correct chirality assignment (Table S30†).
Based on the crystal data of GM 19–26, the success of absolute configuration determination could be attributed to the following points. The existence of phosphorus atoms in CTX[P(O)Ph] enhanced the anomalous scattering effect when chiral guests were absorbed, which is highly important for absolute configuration assignment.2,8c,20 This also avoided the introduction of extra heavy atoms to guest molecules because the anomalous scattering was already adequate.21 In addition, crystallizing with chiral guest molecules was often concomitant with the lowering of the symmetry of the crystals. Low guest molecule occupancy possibly leads to poor data quality and global pseudo-symmetry problems, weakening the reliability of absolute configuration assignment.12 Therefore, 100% occupancy of GM 19–26 is desired to determine the absolute configuration. Additionally, the well-ordered arrangement of the guest in GM 19–26 is vital for the determination of absolute configuration, which is reflected by the application of only a few restraints and the good accordance between the Fo map and refined structure. Partial disorder of the guest molecule possibly leads to an increase in the flack value. The solvent-masking procedure has to be run for severe disorders.22 However, this is not recommended for use in the refinement of chiral molecules.12 The electron density in voids should have been assigned to guest molecules, but it might be masked. The absence of its contribution renders flack parameters unsuitable for evaluation of the reliability of absolute configuration determination. In addition, due to the regular arrangement of the guests, no unassignable electron density is observed in refinement without solvent masking.
To sum up in this section, characteristics, such as the anomalous scattering of phosphorus atoms, regularly ordered guest molecules, full occupancy, and no requirement for the masking of solvents, work together to achieve satisfactory flack values and contribute to credible absolute configuration determination.
:
1 to 1
:
1. However, nerol, a kind of alcohol with a longer carbon chain cannot crystallize with the host; in fact, although its hydroxyl group is supposed to form a hydrogen bond with CTX[P(O)Ph]. With the lengthening of the carbon chain, both molecular polarity and volume change. Comparison of the two physical properties of the guests that crystallize with CTX[P(O)Ph] and the guests did not reveal molecular volume23 and hydrophilicity as possible key factors (Fig. 5 and Table S31†). As shown in Fig. 3, most guest molecules were above the cavity of CTX[P(O)Ph]. It is obvious that the space was not adequate for very large molecules. The largest test molecules were 2,6-diisopropylaniline (GM 3 with the largest molecular volume of 195.23 Å3) and 3-bromo-4-chloroanisole (GM 13 with the largest molecular weight of 221.48 g mol−1). In addition, some smaller molecules, such as menthone and perillyl alcohol, could not crystallize with CTX[P(O)Ph] (Table S31†), which was suspected to be related to the polarity or hydrophilicity of the guest molecules. The hydrophilicity and hydrophobicity were expressed as the logarithm of the partitioning coefficient between 1-octanol and water (log
P).24 Guest molecules with moderate hydrophilicity or polar functional groups tended to crystallize with the host. Once the hydrophobic part of the guest molecule was quite large, it preferentially remained in solution instead of being captured by the host. The most hydrophobic test molecule was 2,6-diisopropylaniline (GM 3) with a log
P value of 3.18. This phenomenon was also consistent with the fact that CTX[P(O)Ph] could not solubilize in a non-polar oil sample under heating, not to mention crystallization. It should be noted that the success of clathrate crystallization is quite complicated and depends on many properties of the host and guest. Therefore, we could only draw a qualitative inference that very large and hydrophobic molecules probably could not be crystallized with CTX[P(O)Ph].
![]() | ||
| Fig. 5 The volume and hydrophilicity of guest molecule possibly affects the formation of clathrate crystal. | ||
Although CTX[P(O)Ph] is a macrocycle, its cavity is too small and shallow to fully accommodate the whole organic molecules (Fig. 6A). The whole guest molecules stayed above the cavity, instead of inside the cavity (Fig. S28†). The nitrogen adsorption–desorption isotherm reported in our previous work also suggested CTX[P(O)Ph] to be a nonporous structure with low Brunauer–Emmett–Teller (BET) surface area.25 It is important to understand why CTX[P(O)Ph] can act as a kind of host molecule to crystallize with various guests and arrange them in order. We suspected this phenomenon was associated with its unique structure. Apart from the extremely rigid skeleton, three converging P
O bonds were another feature of CTX[P(O)Ph]. The oxygen atoms on it provided hydrogen bond acceptors to recognize guest molecules. More importantly, they exhibited an electron-rich trend toward the center. The electrostatic potential (ESP) was calculated to understand the clathrate crystallization mechanism of CTX[P(O)Ph] (Fig. 6A and S31†).26 The three oxygen atoms on P
O bonds exhibited obvious negative potential regions with ESP surface minima of −2.50, −2.49, and −2.40 eV, respectively. The electron-rich cavity of CTX[P(O)Ph] also showed a negative potential of approximately −1 eV (Fig. S31†). They worked collectively and a negative electrostatic potential surface was formed above the cavity, which meant that the host molecule needed to be solvated urgently to release the excessively concentrated negative electrostatic potential. As a result, the guest molecules were located above the cavity to solvate the host through hydrogen bonds and some weaker C–H⋯O noncovalent interactions. This hypothesis was also supported by the phenomenon observed in non-polar guests. The guest would stay on the side of the host (GM 9) or would not be captured by the host (menthone etc.) when its polarity was too low to solvate the host. At this time, CTX[P(O)Ph] would be solvated by benzene rings of another host instead through C–H⋯O interactions in the crystal structure (Fig. S29 and S30†).
The negative electrostatic potential surface acted as a site to capture the guests firmly. The periodically arranged hosts allowed the guests to form an array by following them with 100% occupancy and in the same order. Consider the crystal structures of GM 19@CTX[P(O)Ph], GM 17@CTX[P(O)Ph] and GM 9@CTX[P(O)Ph] as examples. An independent gradient model based on the Hirshfeld partition (IGMH) was established to concretize the noncovalent interactions.27 In GM 19, the hydrogen bonds (1.870 and 1.980 Å) and C–H⋯O interactions (2.367 and 2.520 Å) worked together to anchor the guests above the cavity (Fig. 6B). Their complexation energy was calculated to be −34.57 kcal mol−1 by DFT calculation (Table S41†). The IGMH domains (Fig. 6B) clearly illustrated that the interaction region was distributed above the oxygen atoms and inside the cavity or around the negative ESP surface. According to the IGMH analysis (Tables S38–S40†), three converging P
O bonds and the electron-rich cavity of CTX[P(O)Ph] account for 34.0 and 44.1% of the noncovalent interactions between GM 19 and CTX[P(O)Ph] respectively. In GM 17, only C–H⋯O interactions (2.508 and 2.666 Å) were observed in the crystal structure (Fig. 6C). Despite the moderate complexation energy (−15.25 kcal mol−1) and the absence of hydrogen bonds, the host-guest association was similar to the one above with the contributions of noncovalent interaction being 26.5 and 45.3% for P
O bonds and the cavity of CTX[P(O)Ph] respectively. Therefore, the guest was forced to form an array in order with the host and its structure could be determined well (Fig. 6C and Tables S35–S37†).
In contrast, for GM 9 (toluene molecule), the space above the cavity of CTX[P(O)Ph] was occupied by the benzene ring of another CTX[P(O)Ph] molecule. Toluene molecules were on the side of CTX[P(O)Ph] and nearly no noncovalent interactions were found between them in the crystal structure (Fig. 6D), which could also be demonstrated by the quite low complexation energy (−2.05 kcal mol−1). In addition, the small IGMH domains were mainly distributed between the toluene molecule and the side chain of CTX[P(O)Ph] (Fig. 6D). The contribution of three P
O bonds and the cavity was close to zero (Table S32†). Due to the absence of interactions between the guest and the negative ESP surface of the host, two disordered components were modeled, and more restraints had to be applied to refine the guest in the crystal structure of GM 9@CTX[P(O)Ph]. Based on a comparison of these three examples (Table S42†), it is obvious that noncovalent interactions between the guest and the negative ESP surface of the host play a vital role in the ordered arrangement of the guest.
Footnote |
| † Electronic supplementary information (ESI) available: General information, synthetic procedures, theoretical calculation and all X-ray crystallographic data. CCDC 2251917–2251943. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3sc02995f |
| This journal is © The Royal Society of Chemistry 2023 |