Dynamics in Flexible Pillar[n]arenes Probed by Solid-State NMR

Abstract

Pillar[n]arenes are supramolecular assemblies that can perform a range of technologically important molecular separations which are enabled by their molecular flexibility. Here, we probe dynamical behavior by performing a range of variable-temperature solid-state NMR experiments on microcrystalline perethylated pillar[n]arene (n = 5, 6) and the corresponding three pillar[6]arene xylene adducts in the 100–350 K range. This was achieved either by measuring site-selective motional averaged ¹³C ¹H heteronuclear dipolar couplings and subsequently accessing order parameters or by determining ¹H and ¹³C spin–lattice relaxation times and extracting correlation times based on dipolar and/or chemical shift anisotropy relaxation mechanisms. We demonstrate fast motional regimes at room temperature and highlight a significant difference in dynamics between the core of the pillar[n]arenes, the protruding flexible ethoxy groups, and the adsorbed xylene guest. Additionally, unexpected and sizable ¹³C ¹H heteronuclear dipolar couplings for a quaternary carbon were observed for p-xylene adsorbed in pillar[6]arene only, indicating a strong host–guest interaction and establishing the p-xylene location inside the host, confirming structural refinements.

1. Introduction

Host–guest chemistry is an important concept in the field of supramolecular chemistry that is driven by the interactions of molecular assemblies or ions via noncovalent interactions.¹ These interactions play a vital role in the design of advanced functional materials with exciting physical properties and applications in processes, such as adsorption, catalysis, energy storage, and molecular separations. Consequently, this area has become of increasing importance over the past few decades,²⁻⁵ and a wide range of supramolecular assemblies that adapt to guests⁶ has been discovered thanks to a large variety of tunable structural motifs and properties (e.g., solubility, functionality, and molecular flexibility). Among those, pillar[n]arenes (n = 5–15) have emerged as a novel class of easily functionalized supramolecular macrocycles⁷⁻¹⁰ whose structure consists of substituted phenolic moieties repeated n-times and connected in the para position by methylene linkages (Figure 1). For most values of n (except n = 7), the resulting architecture is a symmetrical cylindrical structure (side view, Figure 1) leading to a symmetrical polygon (top view) that yields a single pentagonal and hexagonal cavity for n = 5 and 6, respectively, and two pentagonal and/or hexagonal cavities for n > 7. The cavity plays an important role in hosting appropriately sized guest molecules for capture/molecular separation¹¹⁻¹⁶ and controlled delivery systems.^17,18

Figure 1.

Crystal structures of (a) perethylated pillar[5]arene EtP5-α (obtained at 240 K), (b) perethylated pillar[6]arene EtP6-β (298 K), (c) p-xylene in EtP6 pX@EtP6 (240 K), (d) m-xylene in EtP6 mX@EtP6 (298 K), and (e) o-xylene in EtP6 oX@EtP6 (100 K).²³ The side and top views are shown on the first and second rows. The pillar[n]arene host and xylene guests are denoted by “ball and stick” and “space filling” models, respectively, with carbons shown in gray, oxygens in red, and protons omitted for clarity in the ball and stick model while shown in white in the space-filling model. The two left panels of the third row provide the chemical structures of both EtP5-α and EtP6-β using color coding for different carbon environments (CH₃, orange; CH₂, light blue; OCH₂, green; CH, pink; CH₂C^IV, yellow; OC^IV, gray) consistent with those used throughout the NMR spectra assignments. The three right panels of the third row show a magnified view of the through space interaction between the p-xylene guest and EtP6 in pX@EtP6 (yellow circle) while no interaction is observed for mX@EtP6 and oX@EtP6 (see text for details).

Pillar[n]arenes (n = 5,6) have found the greatest interest, mostly due to their relatively small cavity sizes that enable them to host small molecules,¹⁰ combined with substituted alkyl and branched chains that strongly affect the host–guest properties.¹⁹⁻²² Perethylated pillar[n]arene (n = 5, EtP5; n = 6, EtP6) are examples of these substituted pillar[n]arenes that contain ethoxy groups (Figure 1a,b) with EtP6 existing as two polymorphs, a metastable EtP6-α phase and a crystalline EtP6-β phase.²³ Due to its large conformational flexibility, EtP6 has been found to adsorb a number of guest molecules,^14,23 and we have recently shown that EtP6-β adapts during adsorption of an o-xylene (oX)/m-xylene (mX)/p-xylene (pX) isomer mixture to efficiently capture pX with a high selectivity of 90% to form pX@EtP6 (Figure 1c),²³ while mX@EtP6 (Figure 1d) and oX@EtP6 (Figure 1e) are obtained by adsorption of the respective xylene isomer into EtP6-β. This is a step forward for the energy efficient separation of the xylene isomers, which are widely used as chemical feedstocks.^23,24

One important criterion for these supramolecular structures is their adaptivity and flexibility that dictates the adsorption of guest molecules in the cavity space and which are not yet understood. This adaptive behavior is not trivial to probe, especially in the solid state, given the requirements to access experimentally measurable observables that depend on dynamics and which need to be determined at the resolution of each local chemical environment. Solid-state nuclear magnetic resonance (NMR), often in conjunction with computational methods such as crystal structure prediction (CSP) and diffraction-based approaches, provides detailed, element-specific, and structural information on the atomic scale and plays an important role in supramolecular assemblies.²⁵⁻³¹ For example, we took advantage of the very high spectral resolution of the ¹³C NMR spectra of EtP5-α, EtP6-α and EtP6-β to support the conformational energy landscape exploration and identify the number of different carbons in the asymmetric unit cell.²³ NMR is also well suited to probe site selective molecular flexibility given its sensitivity to molecular motion over wide time scales from fast processes (subnanoseconds via relaxation measurements) to slower dynamics (milliseconds from line shape analysis), offering a unique access to the qualitative and quantitative description of motion.^28,32

Recent ²H NMR work has focused on the molecular dynamics on n-hexane-d₁₄ in pillar[5]arene that showed that the molecular diversity gave rise to different patterns of guest uptake and release.³³ Liquid state NMR has previously been used to investigate conformational properties³⁴⁻³⁷ and assess p-phenylene unit rotation in pillar[n]arenes, but little is known regarding the adaptive behavior of these materials in the solid state.

Here, we determine the dynamics of both guest-free EtP5-α and EtP6-β and the three xylene-adsorbed perethylated pillar[6]arenes over a range of time scale by probing site selective ¹³C ¹H heteronuclear dipolar couplings and accessing ¹H and ¹³C correlation times as a function of temperature (383–100 K). We find that the flexibility of the protruding OCH₂ groups in the guest-free pillar[n]arenes is reduced when there are fewer phenolic moieties, or at temperatures below 298 K, as well as by adsorption of xylene isomers; by contrast, other carbon groups have largely similar dynamics over the temperature range studied. We identify intermolecular ¹³C ¹H dipolar couplings at low temperatures in pX@EtP6 which are absent on both oX@EtP6 and mX@EtP6, which provides evidence for the location of xylenes in the EtP6 architecture and highlights the host–guest interactions. Finally, we exploit variable temperature spin–lattice relaxation measurements to access dynamics in the MHz regime, which confirm the flexibility of the extruding ethoxy groups of these pillar[n]arenes as opposed to the carbon atoms located in the ring core.

2. Experimental Section

2.1. Materials Synthesis

Guest-free EtP5-α(38) and EtP6-β(38) and the three xylene-adsorbed perethylated pillar[6]arenes²³ were synthesized using established literature procedures (Scheme S1).³⁸ Prior to adsorption, powder X-ray diffraction (PXRD, Figure S1) and NMR measurements, EtP5-α and EtP6-β were dried and heated under vacuum at a pressure of 10^–3 mbar to 433 K for 2 h to ensure no solvation and that the correct phases were obtained. pX@EtP6 and mX@EtP6 were synthesized using the xylene vapor adsorption method, whereas oX@EtP6 was prepared via solvent evaporation with adsorption time longer than 12 h to ensure the presence of one molecule of xylene per EtP6. Differential scanning calorimetry (DSC) data on EtP6-β identifies a phase change at 339 K (Figure S2). Thermogravimetric analysis (TGA) data on pX@EtP6,²³mX@EtP6 (Figure S3), and oX@EtP6 (Figure S4), combined with time-dependent sorption data from ¹H solution-state NMR spectroscopy spectra of dissolved crystals and single-crystal X-ray diffraction data on pX@EtP6, mX@EtP6, and oX@EtP6,²³ establish the stoichiometry of one xylene adsorbed per EtP6.

2.2. NMR Experiments

The ¹H and ¹³C solid-state NMR experiments at an external magnetic field B₀ = 9.4 T were performed on a Bruker Avance III HD NMR spectrometer equipped with a 4 mm HXY triple-resonance magic angle spinning (MAS) probe in double-resonance mode tuned to Larmor frequencies of ν₀(¹H) = 400.13 MHz and ν₀(¹³C) = 100.62 MHz. The B₀ = 14.1 T NMR experiments were performed on a 14.1 T Avance III DNP NMR spectrometer equipped with a low temperature 3.2 mm HXY triple-resonance MAS probe³⁹ in double-resonance mode tuned to ν₀(¹H) = 600.25 MHz and ν₀(¹³C) = 150.93 MHz. All experiments were obtained under MAS with the sample spinning at ν_r = 12.5 kHz, unless otherwise specified. ¹H pulses and SPINAL-64 heteronuclear decoupling⁴⁰ during ¹³C acquisition were performed at a radio frequency (rf) field amplitude of 83 kHz for all samples except the room temperature cross-polarization (CP) experiments on the guest-free samples where it was performed at 96 kHz. ¹³C pulses were performed at a rf field of 60 and 70 kHz at 9.4 and 14.1 T, respectively. For all data obtained at 14.1 T, a presaturation block consisting of 100 ¹H pulses separated by 1 ms was used (all pulse sequences are described in Figure S5 and Section S5 of the SI). For variable-temperature experiments, zirconia drive caps were used at 9.4 T and Vespel caps at 14.1 T. Additional ¹H one pulse quantitative spectra were obtained at B₀ = 20 T on a Bruker Avance III NMR spectrometer and under MAS at ν_r = 60 kHz using a 1.3 mm HXY triple-resonance MAS probe in double resonance mode tuned to a Larmor freuqency of ν₀(¹H) = 850.13 MHz; spectra were acquired with a rf field amplitude of 150 kHz.

In the variable-temperature CP experiments, the CP steps were performed with a ¹³C rf field of 41 kHz (at 9.4 T) and 70 kHz (at 14.1 T) while the ¹H rf field amplitude was ramped to obtain maximum signal at approximately 65 kHz (at 9.4 T) and between 70–96 kHz (at 14.1 T), dependent on samples and temperatures. An optimized contact time of 1.5–3.0 ms was used. Typically, ¹³C CP experiments were accumulated with 2048 scans (at 9.4 T) and 32–2048 scans (at 14.1 T), and used recycle delays of 1.3 × ¹H T₁(41) (with T₁ being the spin–lattice relaxation times measured as given below) that corresponds to the maximum signal-to-noise per unit time. Note that although ¹³C CP MAS experiments are not quantitative, only ¹³C integration within a chemically distinct carbon environment is given as its similar nature allows comparison of the number of carbons to be estimated.

Variable temperature ¹H and ¹³C spin–lattice relaxation times T₁’s were obtained with the saturation recovery and T₁ Torchia⁴² pulse programs, respectively. In the saturation recovery experiment, the magnetization is saturated by a presaturation block consisting of 100 ¹H pulses separated by 10 ms at 9.4 T or 1 ms at 14.1 T, followed by magnetization buildup during a variable τ delay and NMR detection. In the T₁ Torchia sequence,⁴² an initial ¹³C CP step creates ¹³C magnetization which then decays during a variable delay τ and ¹³C detection is achieved using a two-step phase cycle to account for the direct (unenhanced) ¹³C Boltzmann value rather than CP enhanced values. The data obtained via integrated intensities were fitted to stretch exponential functions of the form of 1 – exp[−(τ/T₁)^α] and exp[−((τ/T₁)^β] for the ¹H and ¹³C T₁ data, respectively, where α (between 0.75 and 0.96) and β (between 0.60 and 0.88) are the respective stretch exponential factors. Errors associated from the T₁ values are quoted to a 95% confidence level and are smaller than the symbol sizes in all figures.

Variable-temperature 2D proton detected local field (PDLF) spectra correlating ¹³C NMR spectra in the direct frequency dimension ω₂ with ¹³C ¹H dipolar coupling spectra in the indirect ω₁ dimension were obtained using the windowed⁴³ sequence (wPDLF)⁴⁴ and R-type recoupling blocks.⁴⁵ The sequence starts with the reintroduction of the heteronuclear ¹³C ¹H dipolar coupling under MAS during the rotor synchronized evolution period t₁ using the symmetry-based R18₂⁵¹H recoupling block⁴⁶ which was optimized for maximum signal around the ¹H rf field amplitude of approximately 9 × ν_r (112.5 kHz). R18₂ also removes the homonuclear ¹H ¹H dipolar coupling⁴⁶ and the 180° phase shift in the recoupling block refocuses the (small) ¹H chemical shift anisotropy (CSA), while the synchronized 180° ¹³C pulse applied in the middle of t₁ prevents the same refocusing from occurring for the heteronuclear ¹³C ¹H dipolar coupling and refocuses the ¹³C chemical shift. The ¹³C CSA is averaged out over two rotor periods. The ¹³C magnetization is therefore only modulated by the ¹³C ¹H dipolar coupling in t₁ that yields a ¹³C ¹H dipolar coupling spectra in ω₁. Polarization transfer to ¹³C is subsequently achieved using the rotor synchronized PRinciples of Echo Shifting using a Train of Observations (PRESTO)⁴⁷ pulse sequence optimized for maximum signal for the protonated resonances to a length of × τ_r (142 μs), where τ_r is the rotor period (80 μs), and by varying the recoupling length of the R18₁⁷¹H recoupling block (which is also optimized to a similar ¹H rf field of approximately 9 x ν_r (112.5 kHz)). PRESTO is preferred to CP for polarization transfer as ¹H spin diffusion in the latter results in an increase of the signal intensity for the zero frequency signal.⁴⁴ Following Fourier transformation in the F1 dimension, an effective dipolar coupling constant κ_Rd_CH (with κ_R the scaling factor of the wPDLF sequence and d_CH the dipolar coupling constant, see SI Section S6 including Table S1 and Figure S6 for the experimental determination of κ_R) is obtained in the ω₁ frequency dimension.^44,48,49 The (scaled) ¹³C ¹H dipolar coupling spectra are then extracted at each ¹³C isotropic chemical resonances (δ_iso) and the dipolar coupling values are obtained from the distance between the outer singularities to yield site-specific motional averaged dipolar coupling ⟨d_CH⟩ values. Note that the small variation of these values obtained from each carbon resonance for a particular carbon subgroup (an example of which is given in Figure S7 for the CH₃ resonance of EtP6-β) has been used to provide estimated errors and we have chosen to give a single averaged ⟨d_CH⟩ value for each carbon subgroup.

Static dipolar coupling constants d_CH were calculated from eq 1 and carbon proton bond lengths. These were obtained from computed CSP²³ data for the EtP5-α and EtP6-β conformers or experimental low temperature high resolution powder neutron diffraction data from o-xylene⁵⁰ and m-, p-xylenes⁵¹ crystal structures for the xylenes.

Temperature calibrations were preformed prior to NMR data acquisition using either the ²⁰⁷Pb chemical shift thermometer of Pb(NO₃)₂^52,53 or the ⁷⁹Br T₁s⁵⁴ of KBr (extracted from polarization build-up curves using the saturation recovery pulse sequence) according to procedures outlined in the literature. All temperatures reported are actual sample temperatures and have an estimated accuracy of ±10 K. NMR data were processed with TopSpin and MATLAB R2019a.⁵⁵¹H and ¹³C spectra were referenced to H₂O at 4.8 ppm and the CH of adamantane at 29.45 ppm,⁵⁶ respectively, both relative to TMS primary reference at 0 ppm. Small deviations in the observed isotropic chemical shifts (±0.7 ppm in ¹³C CP MAS NMR spectra) is likely attributed to small changes in shim coil temperatures during variable temperature experiments.

3. Results and Discussion

3.1. NMR Structural Analysis

The ¹³C CP MAS NMR spectra of guest-free EtP5-α and EtP6-β (Figure 2a,b,²³Table 1) collected under MAS at 12.5 kHz and at a magnetic field of 9.4 T are extremely well resolved with full width at half-maximum lines typically around 30 Hz (or 0.3 ppm at 9.4 T), in agreement with the excellent crystallinity of these samples. Each different chemical subgroup can be readily assigned, and the remarkable resolution obtained enables the observation of all nonequivalent magnetically distinct carbon atoms in the asymmetric unit cells.²³ The ¹³C CP MAS NMR spectra of all xylene-adsorbed EtP6 adducts (Figure 2c–e) are all different from EtP6-β and from each other, as previously identified by CSP of the molecular conformational space. Therefore, the spectral identification of the xylenes resonances (red daggers in Figure 2c–e, Table 1) is not straightforward and is obtained based on comparisons with well-established isotropic chemical shift (δ_iso) values,^57−5913C-edited NMR experiments (Figures S8–S10) employing CP steps of various contact times, including their spectral deconvolution (Figures S11–S13), and the existence of CH dipolar couplings (Figures 3 and S20–S22). A detailed discussion on the spectral assignment of the xylenes-adsorbed EtP6 is available in Section S7 of the SI.

Figure 2.

¹³C CP MAS NMR spectra of (a) EtP5-α, (b) EtP6-β, (c) pX@EtP6, (d) mX@EtP6, and (e) oX@EtP6 obtained at a magnetic field of 9.4 T. The spectra for EtP5-α and EtP6-β are identical to those previously published.²³ Spectral assignments are given in the figure (see Figure 1) and are obtained from known δ_iso, ¹³C-edited CP experiments (Figures S8–S10), spectral deconvolution (Figures S11–S13), and 2D PDLF data (see below). The red daggers (†) denote signals arising from the xylene guests. The CH₃ originating from the o-xylene guest in (e) is unidentifiable due to spectral broadening and overlapping resonances with the CH₃ signals of the EtP6 host. Asterisks (*) and hashes (#) denote spinning sidebands and amorphous impurities, respectively.

Table 1. ¹³C NMR Assignments, ¹³C Isotropic Chemical Shifts δ_iso from Spectral Deconvolution, Calculated Static Dipolar Coupling Constants d_CH, Experimentally Found Motional Averaged Dipolar Coupling Constants ⟨d_CH⟩, and Order Parameters ⟨S_CH⟩ for Protonated Carbons in EtP5-α, EtP6-β, pX@EtP6, mX@EtP6, and oX@EtP6 at 298 and 243 K.

			298 K, 9.4 T		243 K, 14.1 T
assignment	13C δiso (ppm)a	dCH (kHz)b	⟨dCH⟩ (kHz)c	⟨SCH⟩d	⟨dCH⟩ (kHz)c	⟨SCH⟩d
EtP5-α
CH3	14–17	–23.1	–7.2 ± 0.5	0.31 ± 0.02	–7.5 ± 0.5	0.32 ± 0.02
CH2	28–38	–23.0	–23.3 ± 0.8	1.01 ± 0.04	–22.8 ± 0.8	0.98 ± 0.04
OCH2	61–68	–22.8	–18.4 ± 0.7	0.81 ± 0.03	–19.8 ± 0.7	0.87 ± 0.03
CH	111–124	–23.8	–23.8 ± 0.8	1.00 ± 0.03	–23.4 ± 0.8	0.98 ± 0.04
EtP6-β
CH3	14–18	–23.1	–7.2 ± 0.5	0.31 ± 0.02	–7.2 ± 0.5	0.31 ± 0.02
CH2	27–35	–23.0	–22.4 ± 0.8	0.97 ± 0.04	–21.7 ± 0.8	0.94 ± 0.04
OCH2	62–67	–22.8	–18.1 ± 0.7	0.79 ± 0.03	–18.3 ± 0.7	0.80 ± 0.03
CH	111–118	–23.8	–23.9 ± 0.8	1.00 ± 0.04	–23.4 ± 0.8	0.98 ± 0.03
pX@EtP6
CH3	13–16	–23.1	–6.9 ± 0.5	0.30 ± 0.02	–7.3 ± 0.5	0.32 ± 0.03
CH3e	18–21	–23.7	–7.0 ± 0.5	0.30 ± 0.03	–7.1 ± 0.5	0.30 ± 0.02
CH2	27–38	–23.1	–21.6 ± 0.8	0.94 ± 0.04	–21.3 ± 0.8	0.92 ± 0.04
OCH2	61–65	–22.9	–18.4 ± 0.7	0.80 ± 0.03	–19.2 ± 0.7	0.84 ± 0.03
CH	110–118	–23.8	–22.4 ± 0.8	0.94 ± 0.03	–22.5 ± 0.8	0.95 ± 0.04
CHe	129–130	–23.7	–21.6 ± 0.8	0.91 ± 0.04	–22.6 ± 0.8	0.95 ± 0.04
mX@EtP6
CH3	14–17	–23.1	–6.9 ± 0.5	0.30 ± 0.02	–7.0 ± 0.5	0.30 ± 0.02
CH3e	21–22	–24.4	–6.2 ± 0.5	0.25 ± 0.02	–6.7 ± 0.5	0.27 ± 0.03
CH2	31–34	–23.1	–22.2 ± 0.8	0.96 ± 0.04	–21.2 ± 0.8	0.92 ± 0.04
OCH2	62–66	–22.8	–19.2 ± 0.7	0.84 ± 0.03	–20.0 ± 0.7	0.88 ± 0.03
CH	111–117	–23.7	–23.5 ± 0.8	0.99 ± 0.03	–23.2 ± 0.8	0.98 ± 0.03
CHe	126–131	–24.0	–22.5 ± 0.8	0.94 ± 0.04	–22.3 ± 0.8	0.93 ± 0.03
oX@EtP6
CH3	14–17	–23.1	–7.1 ± 0.5	0.31 ± 0.02	–7.0 ± 0.5	0.30 ± 0.02
CH3ef		–23.7
CH2	28–38	–23.2	–23.8 ± 0.8	1.03 ± 0.04	–22.2 ± 0.8	0.96 ± 0.04
OCH2	62–66	–22.7	–18.3 ± 0.7	0.81 ± 0.03	–18.8 ± 0.7	0.83 ± 0.03
CH	111–116	–23.9	–23.8 ± 0.8	1.00 ± 0.04	–22.8 ± 0.8	0.96 ± 0.04
CHef	128–129	–23.6	–17.0 ± 0.7	0.72 ± 0.03	–18.8 ± 0.7	0.80 ± 0.03

^aRange of ¹³C δ_iso obtained at room temperature are given for each carbon subgroup. Exact δ_iso for all individual carbons are provided in Tables S2–S6.

^bStatic dipolar coupling constants were calculated as described in the text in Section 3.2 and eq 1.

^cOnly the short-range ⟨d_CH⟩ constants are given (see text for details). Errors are estimated from the uncertainty in the determination of the position of the outer singularities of the ¹³C ¹H dipolar coupling spectra.

^dEstimated errors are calculated from the errors in ⟨d_CH⟩.

^eSignals from xylenes.

^fOverlapping resonances between the guest and the host in the ¹³C CP MAS NMR spectrum of oX@EtP6 prevents spectral assignment of the CH₃ and only allow tentative assignment of the xylene CH carbons.

Figure 3.

(a) ¹³C CP MAS spectrum, (b) PDLF spectrum and (c) selected site-specific ¹³C ¹H dipolar spectra for guest-free EtP6-β. Spectral assignments are given in the figure and correspond to those previously published.²³ The data presented above was obtained at 298 K and 9.4 T. ⟨d_CH⟩ is measured using the outer singularities of the dipolar coupling spectra as highlighted in the Experimental Section. Vertical light gray lines indicate the static limit dipolar coupling constants d_CH calculated from eq 1 and the computed CH distances obtained at the DFT level on the various conformers identified by CSP.²³ Asterisks (*) denote spinning sidebands.

Variable–temperature ¹³C CP MAS NMR spectra (Figures S14–S18) for all five materials were performed in the 383–100 K temperature range (down to only 243 K for EtP5-α). Upon cooling, significantly broader ¹³C NMR resonances are observed at low temperature (e.g., from 30 Hz at 298 K to 60 Hz at 100 K for the CH resonance of EtP6-β at 14.1 T) as anticipated from the macrocycles being trapped in a variety of conformations and leading to inhomogeneous broadening. There is minimal change in the intensity of the spinning sidebands, which likely indicates that the ¹³C CSA is largely unchanged in the temperature range studied here while also suggesting that accessing ¹³C CSAs is likely not a suitable method to obtain dynamics information in the kHz regime in these materials. There is also no evidence of signal coalescence due to chemical exchange.

Upon heating above 323 K, the ¹³C CP MAS NMR spectrum of EtP6-β remains very well resolved and the number of resonances halves (Figure S19) vs spectra at 298 K, indicating a crystal structure of higher symmetry. This change is in agreement with both the DSC data (Figure S2) that shows an endothermic peak at 339 K, and with refined XRD data at 433 K that indicated a transition from triclinic P1EtP6-β at room temperature to a metastable triclinic P1 state at 339 K with a half unit cell volume.²³ In contrast, the ¹³C CP MAS NMR spectrum of EtP5-α up to 383 K remains unchanged upon heating (Figure S14), and no polymorphic transition is observed. No change is also observed in the ¹³C CP MAS NMR spectra of pX@EtP6 or oX@EtP6 up to 323–330 K (Figures S16 and S18), which is consistent with TGA results that show that the adsorbed xylenes are only lost from the pores at temperatures exceeding about 348 K for pX@EtP6(23) and 340 K for oX@EtP6 (Figure S4).

The ¹³C CP MAS NMR spectrum of mX@EtP6 at 323 K (Figure S17), however, shows the disappearance of the adsorbed xylene peaks and accounts for m-xylene desorption from the pores (Figure S3). Upon cooling this sample back to room temperature, the ¹³C CP MAS NMR spectrum (data not shown) indicates that the material has not returned back to EtP6-β as this polymorph is only formed above 433 K.²³ We ascribe this difference of behaviors between mX@EtP6 and pX@EtP6/oX@EtP6 to the smaller cavity of the former preventing the m-xylene guest (Figure 1) to be fully accommodated in the pores and facilitating this removal upon heating.

3.2. Temperature-Dependent Motional Averaged Site-Selectivity in Guest-Free Pillar[n]arenes and Xylene-Adsorbed Pillar[6]arenes

Heteronuclear dipolar couplings are dependent on distance and motion,⁶⁰ and the magnitude of this dipole–dipole coupling is given by the following expression for ¹³C ¹H

1

where d_CH is the dipolar coupling constant between the ¹³C and ¹H nuclei in Hz, μ₀ is the vacuum permittivity, ℏ is the reduced Planck constant, γ_C and γ_H are the respective gyromagnetic ratios for the ¹³C and ¹H nuclei, and r_CH is the distance between the carbon and hydrogen atoms. Motional averaged dipolar couplings ⟨d_CH⟩ can be obtained by two-dimensional (2D) proton detected local field (PDLF)^43,44,61 experiments that correlate the ¹³C isotropic chemical shifts with their corresponding ¹³C ¹H dipolar spectra, providing site-selective heteronuclear dipolar coupling constants (see the Experimental Section for further details).

The room-temperature ¹³C CP MAS NMR spectrum of EtP6-β is given in Figure 3a with the corresponding 2D PDLF spectrum (Figure 3b) showing dipolar coupling for all protonated carbons as expected (Figure 3c) and allowing the corresponding ¹³C ¹H dipolar spectra to be extracted at each ¹³C shifts which revealed significant ⟨d_CH⟩ differences between each carbon subgroup. For example, smaller ⟨d_CH⟩ of −7.2 ± 0.5 and −18.1 ± 0.7 kHz are obtained for the CH₃ and OCH₂ carbons of the ethoxy group, respectively, while larger values of −22.4 ± 0.8 and −23.9 ± 0.8 kHz are extracted for the CH₂ and CH carbons of the pillar[6]arene backbone ring (Tables 1 and S3). While no dipolar coupling splitting is apparent for the quaternary OC^IV carbons, partially resolved small couplings of −2.9 ± 0.3 kHz are obtained for the CH₂C^IV carbons and is likely due long-range through space coupling to the nearby methylene CH₂ ring group. Similar long-range dipolar couplings (−4.0 ± 0.3 and −6.1 ± 0.5 kHz) can also be observed for the CH₃ and CH₂ environments, respectively, which arise from spatial proximity with protons on the nearby carbons.

Motion can be quantified by a site-specific order parameter ⟨S_CH⟩ (eq 2, Table 1 and Figure 4) that compares ⟨d_CH⟩ with the static limit dipolar coupling constants d_CH in the absence of motion and ranges from 0 for isotropic motion to 1 for a rigid system

2

⟨S_CH⟩ obtained for each carbon subgroup in EtP6-β are found to be 0.31 ± 0.02 for CH₃, 0.97 ± 0.04 for CH₂, 0.79 ± 0.03 for OCH₂, and 1.00 ± 0.04 for CH (Table 1) at room temperature. There is therefore no (or limited) motion for the CH₂ and CH carbons situated in the arene core of the pillar[n]arene ring. However, both CH₃ and OCH₂ carbons in the ethoxy group show motional averaging caused by dynamics which is ascribed to rotational and librational motions of these carbons. While this effect is fairly small for the OCH₂ carbon (⟨S_CH⟩ = 0.79), motion is particularly pronounced for the CH₃ group which ⟨d_CH⟩ is approximately one-third of the d_CH yielding ⟨S_CH⟩ = 0.31 and indicates an increase of motion further away from the arene core.

Figure 4.

(Left) Temperature dependency of the motional averaged CH dipolar coupling order parameters ⟨S_CH⟩ and (right) selected ¹³C ¹H dipolar coupling spectra of the OCH₂ signals at various temperatures for (a) guest-free EtP5-α (stars), (b) guest-free EtP6-β (circles), (c) pX@EtP6 (squares), (d) mX@EtP6 (diamonds), and (e) oX@EtP6 (triangles). The different carbon subgroups can be identified with the following color coding for CH₃ (orange), CH₂ (light blue), OCH₂ (green), and CH (pink) (Figure 1). Data recorded at room temperature have been collected at both 9.4 and 14.1 T. Error bars in ⟨S_CH⟩ (ΔS_CH) are consistently smaller than 0.04 and are obtained from estimated errors in the determination of ⟨d_CH⟩ and small variations in the dipolar coupling values across one carbon subgroup (see Figure S7); these errors are less than the symbol size. Data below 243 K were not recorded for EtP5-α. The dotted line in (b) indicates a polymorphic transition in EtP6-β from triclinic P1 to a metastable triclinic P1 state with higher symmetry at 339 K (see Figure S2). Dashed lines in (c) and (d) represent the onset temperatures at which xylenes are lost as identified by the TGA data for pX@EtP6(23) and oX@EtP6 (Figure S4) and both TGA data (Figure S3) and changing NMR spectrum (Figure S17) for mX@Et6. Vertical light gray lines in the dipolar coupling spectra indicate the static limit dipolar coupling constants d_CH.

The temperature dependency of ⟨S_CH⟩ was obtained by measuring site selective ⟨d_CH⟩ for EtP5-α from 383 K down to 243 K (Figure 4a, Table 1) and for EtP6-β over an extended temperature range from 383 K down to 100 K (Figure 4b). In EtP6-β (and for all samples), the ⟨S_CH⟩ values for the CH₃ groups remain largely constant at 0.31 ± 0.02, indicating that this group still possesses significant motion even at 100 K. This is consistent with temperatures lower than 100 K required to “freeze” the rapid 3-site hopping motion of CH₃ in various biomolecules.⁶²⁻⁶⁴ In contrast, the ⟨S_CH⟩ values of the OCH₂ increase significantly upon cooling from 0.79 ± 0.03 at 298 K to 0.95 ± 0.03 at 100 K, supporting reduction in motion and lower flexibility by the pillar[n]arene at lower temperatures. In EtP5-α, while the room temperature ⟨S_CH⟩ values for the CH₃, CH₂, and CH carbon subgroups are virtually identical to those determined for EtP6-β (Table 1), a difference was observed for the OCH₂ group upon cooling. An increase in ⟨d_CH⟩ from −18.3 ± 0.7 kHz in EtP6-β to −19.8 ± 0.7 kHz in EtP5-α is observed as evidenced by larger splitting of the outer singularities in the ¹³C ¹H dipolar spectra at 243 K (Figure 4b) and results in slightly larger ⟨S_CH⟩ values in EtP5-α (0.87 ± 0.03) than in EtP6-β (0.80 ± 0.03). Similarly, at higher temperature (383 K), the ¹³C ¹H dipolar coupling spectra of the OCH₂ group yield larger ⟨d_CH⟩ values (−16.6 ± 0.7 and −14.5 ± 0.6 kHz) and smaller ⟨S_CH⟩ values (0.73 ± 0.03 vs 0.64 ± 0.03) in EtP5-α than in EtP6-β, respectively. This indicates more restricted motion and increased hindrance which is likely due to the reduced void space of the smaller EtP5-α cavity versus EtP6-β.

Variable-temperature 2D PDLF NMR experiments were also recorded on the three guest-adsorbed xylene adducts in EtP6 to access ⟨d_CH⟩ and ⟨S_CH⟩ (Table 1, Figure 4c,d,e for pX@EtP6, mX@EtP6, and oX@EtP6, respectively). There, the trends are largely similar to EtP6-β with temperature independent ⟨S_CH⟩ around 1 for the CH₂ and CH carbons in the pillar[n]arene core, around 0.3 for the CH₃, and increasing toward 1 for the OCH₂ group as temperatures are lowered into the static regime. Although the room temperature ⟨S_CH⟩ values for the CH₃, CH₂, and CH carbons are within error of each other for EtP6-β and the xylene-adsorbed adducts, there is a slight increase in the room temperature ⟨S_CH⟩ values obtained for the OCH₂ group in EtP6-β/oX@EtP6/pX@EtP6 vs mX@EtP6 (Table 1). This small difference is enhanced further upon cooling to 243 K, and the data therefore seems to suggest marginally slower dynamics of the OCH₂ group in mX@EtP6 than in EtP6-β, oX@EtP6/pX@EtP6. In contrast to the latter two phases, the xylene in mX@EtP6 lies on top of the EtP6 host rather than within the void space, as illustrated in Figure 1; therefore, the interaction of the m-xylene with the protruding ethoxy groups is likely to cause slower dynamics, at least for the OCH₂ subgroup. These experiments therefore highlight small change in structure flexibility between guest-free and guest-adsorbed EtP6 assemblies.

The room-temperature PDLF data on the three xylene adducts (Figures S20–S22) also partially resolved the dipolar coupling observed in the xylenes themselves. While the corresponding ⟨d_CH⟩ for the xylene CHs in pX@EtP6 and mX@EtP6 indicate limited motion with ⟨S_CH⟩ values found in the 0.91–0.94 ± 0.04 range (Table 1), the xylene CHs in oX@EtP6 show considerably more motion with smaller ⟨S_CH⟩ values of 0.72 ± 0.04 at 298 K. This indicates that the o-xylene has a significant amount of spatial freedom to allow for mobility and that the CH and CH₃ motion of the xylene is not completely limited upon loading into the EtP6 cavity at room temperature.

3.3. Host–guest Interaction Probed by Dipolar Coupling in Xylene-Adsorbed Pillar[6]arenes

No large dipolar coupling is observed at room or low temperatures for the quaternary carbons of either xylenes or pillar[6]arene host as expected (Figures 5a and S20–S23); however, surprisingly, upon cooling pX@EtP6 to 100 K, strong dipolar couplings of −23.4 ± 0.8 kHz were observed for the CH₂C^IV carbon (Figures 5b and S24c). These couplings in pX@EtP6 do not originate from either of the CHs in the xylene (at 129.4 and 129.8 ppm) or a long-range interaction in the EtP6 architecture (no coupling is observed in the CH₂C^IV of EtP6-β as revealed in Figure 3 at room temperature and Figure S23 at 100 K) but rather from the CH₂C^IV carbons (125–131 ppm). Therefore, this coupling was ascribed to intermolecular heteronuclear dipolar coupling between the quaternary CH₂C^IV carbon of the EtP6 host and protons of p-xylene identifying EtP6 p-xylene spatial interaction and strong host–guest interaction. These results are in sharp contrast to the 100 K PDLF data for mX@EtP6 and oX@EtP6 adducts (Figures S25 and S26, respectively) for which no coupling is observed for CH₂C^IVs suggesting an absence of host–guest interaction or that the coupling is still averaged out at 100 K.

Figure 5.

Comparison of selected CH₂C^IV 13C ¹H dipolar spectra for (a) EtP6-β and (b) pX@EtP6 obtained at 298 K and 9.4 T and at 100 K and 14.1 T. The polarization transfer to ¹³C during the PRESTO block of 2D PDLF sequence was optimized for maximum signal on the protonated resonances (see Section 2.2), which accounts for the signal-to-noise of these quaternary carbon resonances.

These results strongly support the structures illustrated in Figure 1. In particular, in pX@EtP6, p-xylene is located in the center of the EtP6 cavity which is stabilized by strong π–π stacking with two aromatic rings from EtP6 (Figure 1c), yielding strong ⟨d_CH⟩ between the CH₂C^IV carbons of the pillar[6]arene backbone with p-xylene protons. It is also likely that this coupling arises preferentially from the aromatic protons of p-xylene rather than the methyl protons due to methyl group rotation as discussed above. In oX@EtP6, similar rotational dynamics prevent coupling of the methyl protons of o-xylene located inside the cavity to the EtP6 backbone while the aromatic protons are positioned outside the cavity (Figure 1e) from which a small static dipolar coupling would only be expected (0.44 kHz based on the smallest 4.1 Å distance with the EtP6 CH₂C^IV carbon). The small EtP6 cavity in mX@EtP6 is too small to host m-xylene (Figure 1d), resulting in this xylene to be excluded and the absence of dipolar coupling interaction with EtP6.

3.4. Temperature-Dependent Relaxation Studies of Guest-Free and -Absorbed Pillar[n]arenes

T₁ relaxation is a measure of the time for the spin population to recover to equilibrium after a perturbation and is mediated by fluctuations of the local magnetic fields, as quantified by the correlation times of the motion τ_c with corresponding frequencies τ_c^–1 on the order of the Larmor frequency, i.e., MHz. Site-specific ¹³C spin–lattice relaxation rates T₁^–1 for all carbons have been obtained versus temperatures in the 383–243 K and 298–100 K temperature range at 9.4 T (Figures S27) and 14.1 T (Figures S28), respectively, for EtP6-β, all xylene-adsorbed EtP6 adducts and EtP5-α (data only available at 9.4 T for this phase). ¹H T₁^–1 were also obtained (Figures S27–S29) and suggest the same similar motional process likely due to the lack of resolution (for a discussion of the ¹H data, Section S10 of the SI). Illustration of the ¹³C T₁s obtained at room temperature and 9.4 T are given in Tables S2–S6 for EtP5-α, EtP6-β, pX@EtP6, mX@EtP6, and oX@EtP6, respectively, and we have chosen to give a single T₁ value (with associated errors) for each carbon subgroup as these are within errors of each other. The following general trend is observed in all of the guest-free and xylene-adsorbed pillar[n]arenes: the CH₃ group has the shortest T₁ (approximately 2 s at room temperature) of all the carbon environments, as it is well-known that methyl groups are relaxation sinks due to their facile three-site hopping motions and efficient ¹³C ¹H heteronuclear dipole–dipole coupling relaxation; the ¹³C T₁ of the OCH₂ moieties are also relatively short (approximately 20–40 s) and likely due to rotation around the O–C bond; these T₁ are in contrast with the ones of the CH₂/CH groups that are in the 10² s range and suggest limited motional freedom and rigidity of these pillar[6]arene core groups; OC^IV and CH₂C^IV carbons yield the longest T₁ as the dominant relaxation mechanism of CSA (see below) is less efficient than dipolar coupling to ¹H for these nonprotonated carbons. Note that upon loading of p- and m-xylene, the OCH₂ group shows an increase in T₁ at room temperature, suggesting that guest addition lowers the flexibility of the pillar[n]arenes. oX@EtP6 shows a reduction in nearly all T₁ in comparison to EtP6-β; however, this is likely attributed to the more amorphous nature of this material.

The ¹³C T₁^–1 rates for each carbon subgroup in EtP5-α, EtP6-β, and the xylene-adsorbed adducts typically increase with increasing temperatures (Figure S28b), pass through maxima at 165–168 K (at 14.1 T) for the majority of resonances (excluding the CH₂ in pX@EtP6, CH in oX@EtP6 and the CH₃, CH₂, CH, and OC^IV in mX@EtP6), and then decrease. At these T₁^–1 maxima, the motion is near the ¹³C Larmor frequency ω_0,C (in rad·s^–1) with the following expression eq 3 being satisfied⁶⁵

3

leading to a τ_c value of 6.5 × 10^–10 s for these materials.

Assuming negligible contribution from spin-rotation and scalar coupling relaxation mechanisms, ¹³C T₁^–1 rates can generally be expressed (eq 4)

4

as the sum of both ¹³C ¹H heteronuclear dipolar coupling (eq 5)

5

and ¹³C CSA relaxation (eq 6)^65,66

6

mechanisms with n the number of protons attached to ¹³C, Δδ_C the (reduced) anisotropy (sensitivity of the chemical shift interaction to the orientation), and η_C asymmetry parameter (deviation from axial symmetry) of the second rank ¹³C chemical shift tensor with principal components δ₁₁, δ₂₂ and δ₃₃ as defined in Section S12 of the SI (all other terms are defined above). The local magnetic fields fluctuation term of the CSA expression is magnetic field dependent and proportional to the square of the Larmor frequency and anisotropy.

¹³C relaxation generally arises from ¹³C–¹H heteronuclear dipole–dipole coupling for protonated carbons with small CSA, i.e., CH₃, OCH₂, and CH₂, and from ¹³C CSA for quaternary aromatic carbons, i.e., OC^IV and CH₂C^IV, as confirmed by comparing the magnitude of the local dipolar and CSA magnetic fields term in eqs 5 and 6. For example, in mX@EtP6 (similar observations were made on the other materials), the calculated local dipolar magnetic fields term for CH₃ (6 × 10⁹ s^–2) is 2 orders of magnitude larger than the calculated CSA term (8 × 10⁷ s^–2 at 14.1 T assuming a typical ¹³C Δδ_C for this carbon of 25 ppm),⁶⁷ while for OC^IV, the CSA term largely dominates even at the lower magnetic field (1 × 10⁹ s^–2 at 9.4 T with a ¹³C Δδ_C of −142 ppm vs 4 × 10⁷ s^–2 for dipolar). However, for the remaining aromatic CH sites, ¹³C relaxation derives from cross terms between dipolar and CSA interactions⁶⁸ as both local magnetic field contributions are comparable (2 × 10⁹ s^–2 for dipolar vs 1–3 × 10⁹ s^–2 at 9.4–14.1 T for CSA using an aromatic CH with a ¹³C Δδ_C of −147 ppm)⁶⁹ and is further suggested by the slight magnetic field dependency of the T₁.

¹³C T₁^–1 maxima and dominant relaxation mechanism(s) allow experimental access to the local magnetic fields term by combining eq 3 and either eq 5 (for heteronuclear dipolar coupling relaxation), eq 6 (for CSA relaxation), or eq 4 (for both mechanisms). For example, in keeping with mX@EtP6, the experimentally determined local dipolar magnetic fields term for CH₃ (3 × 10⁹ s^–2) compares well with the calculated value (6 × 10⁹ s^–2). These equations were then used to obtain τ_c for all materials (CH₃, OCH₂, and OC^IV in Figure 6a–e; CH₂, CH, and CH₂C^IV in Figure S30a–e) and the room-temperature τ_c are the shortest for the CH₃ and OCH₂ groups supporting motion. The temperature dependence of the correlation frequencies τ_c^–1 was subsequently modeled with an Arrhenius equation of the form

7

with τ_c,0^–1, E_a, and R the attempt frequency, activation energy of the thermally activated motional process, and universal gas constant, respectively, and are given in Table 2 for ¹³C and Table S7 for ¹H. The E_a for CH₃ in EtP6-β (6 kJ mol^–1) is significantly smaller than in EtP5-α (11 kJ mol^–1) and is likely due to the smaller ring size of the latter hindering molecular rotation. Upon addition of any guest of EtP6-β, the E_a for CH₃ increases to 8–10 kJ mol^–1 which suggests restricted motion caused by their spatial proximities. No significant difference is however observed between the different guest-adsorbed materials or for OCH₂ groups. There, much smaller changes in T₁ and τ_c times are measured and would therefore indicate that, within the temperature range probed, all materials experience the same motional processes.

Figure 6.

(Left) ¹³C spin–lattice relaxation rates T₁^–1 against correlation times τ_c and (right) corresponding ¹³C correlation frequencies τ_c^–1 Arrhenius plots. Data shown in black and blue outlines were obtained at 9.4 and 14.1 T, respectively, for (a) guest-free EtP5-α (stars), (b) guest-free EtP6-β (circles), (c) pX@EtP6 (squares), (d) mX@EtP6 (diamonds), and (e) oX@EtP6 (triangles). Selected carbon subgroups have been plotted here with the following color coding for CH₃ (orange), OCH₂ (green), and OC^IV (gray) (Figure 1) while plots giving the three other carbons are given in Figure S30. The associated errors are smaller than the symbol sizes. In the left panels, the solid (−) lines are those obtained from a dipolar coupling relaxation mechanism (eq 5) for CH₃ (orange) and OCH₂(green) and the dotted (··) lines from a CSA relaxation mechanism (eq 6) for OC^IV(gray) at both fields, using the experimentally determined local magnetic fields terms (values from EtP6-β were used for EtP5-α as no T₁ minimum was found in the temperature range studied). A T₁ minima was found for OC^IV in oX@EtP6 in the temperature range studied at 9.4 T; therefore, this data was used to extract correlation times and is plotted for this series. In the right panels, the lines are fit to the experimental data using the Arrhenius equation.

Table 2. Comparison of the Attempted Frequencies τ_c,0^–1 and Activation Energy Barriers E_a of Guest-Free EtP5-α, Guest-Free EtP6-β, pX@EtP6, mX@EtP6, and oX@EtP6 Obtained from the Arrhenius Plots of the ¹³C Correlation Frequencies τ_c^–1.

carbon subgroup	τc,0–1 (s–1)	Ea (kJ mol–1)a
EtP5-αb
CH3	5 × 1012	11
CH2	6 × 1010	5
OCH2	6 × 1010	2
CH	4 × 1010	5
CH2CIV	5 × 1012	4
OCIV	4 × 1010	5
EtP6-β
CH3	4 x1011	6
CH2	6 × 1010	5
OCH2	2 × 1011	7
CH	3 × 1010	4
CH2CIV	8 × 109	3
OCIV	1 × 1010	4
pX@EtP6
CH3	5 × 1012	10
CH3c	5 × 1012	3
CH2	2 × 1011	7
OCH2	7 × 1011	8
CH	7 × 1010	5
CH2CIV	1 × 1010	4
CHcde
OCIV	2 × 1010	4
mX@EtP6
CH3	2 × 1012	8
CH3c	4 × 1012	1
CH2	5 × 1010	5
OCH2	1 × 1012	9
CH	8 × 109	3
CH2CIV	7 × 109	3
CHcde
OCIV	5 × 109	2
oX@EtP6
CH3	2 × 1012	8
CH3cd
CH2	5 × 109	2
OCH2	2 × 1011	7
CH	3 × 109	2
CH2CIV	5 × 109	3
CHcde
OCIV	7 × 109	3

^aErrors are in the order of 1 kJ mol^–1.

^bData in the 383–243 K temperature range only available.

^cDenotes signals from xylenes.

^dOverlapping resonances between guest and host in the ¹³C CP MAS NMR spectra prevent measurement of T₁ times.

^eSignal to noise of some of the signals are also too weak for accurate determination of T₁ times.

τ_c were also extracted for the methyl groups of the xylene guests in both pX@EtP6 and mX@EtP6, and the temperature dependency of their frequencies was used to extract E_a values (Figure S31a,b). Activation energies are small (1–3 kJ mol^–1, Table 2) and significantly less than the ones determined for the CH₃ groups of the host (10 kJ mol^–1 in pX@EtP6; 8 kJ mol^–1 in mX@EtP6), indicating that the CH₃ groups have significantly higher degree of motion in the xylenes than pillar[6]arene. Additionally, further comparison between the xylene CH₃’s in pX@EtP6 vs mX@EtP6 reveals higher E_a in the former and supports the xylene location inside the arene core.

4. Conclusions

We employed variable-temperature multinuclear NMR experiments to provide detailed understanding of the dynamics in guest-free perethylated pillar[n]arene (n = 5,6) and xylenes-adsorbed pillar[6]arenes. Site-selective ¹³C ¹H dipolar spectra, enabled by the highly resolved ¹³C CP MAS NMR spectra, permit the quantification of order parameters that reveal differential dynamics properties. Protruding carbons were found to have faster dynamics than those in the core, while the larger void size of EtP6-β than EtP5-α results in a less restricted OCH₂ motion. ¹³C ¹H dipolar spectra also identified spatial proximity in pX@EtP6, not detected in oX@EtP6 and mX@EtP5, demonstrating significantly strong π–π stacking of p-xylene located in the center of the void validating structural models. Temperature-dependent correlation frequencies from relaxation times measurements tentatively suggest oX@EtP6 to have the largest size conformation and show extensive motional dynamics of the perethylated and xylene methyl carbons.

This work demonstrates the capture of structural transformations resulting from host–guest interactions and motional effects in adaptative pillar[n]arene materials, which could have implications for processes such as competitive loading, molecular separation, and drug release. This adds to our understanding of motion in flexible molecular solid state systems and opens up new perspectives in the rational design of materials with enhanced physical properties.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jpcc.1c02046.

• NMR data, additional method details, synthesis details and their supporting data (PDF)

Author Contributions

F.B. designed the project. A.R.H. performed all NMR experiments and analyses and interpreted the data with F.B. S.P. assisted with NMR data acquisition at 14.1 T. A.R.H. loaded the pillar[n]arenes with M.L. A.R.H. and F.B. wrote the manuscript with input from all authors.

The authors declare no competing financial interest.