High-Pressure in Situ 129Xe NMR Spectroscopy: Insights into Switching Mechanisms of Flexible Metal-Organic Frameworks Isoreticular to DUT-49.

Flexible metal-organic frameworks (MOFs) are capable of changing their crystal structure as a function of external stimuli such as pressure, temperature, and type of adsorbed guest species. DUT-49 is the first MOF exhibiting structural transitions accompanied by the counterintuitive phenomenon of negative gas adsorption. Here, we present high-pressure in situ 129Xe NMR spectroscopic studies of a novel isoreticular MOF family based on DUT-49. These porous materials differ only in the length of their organic linkers causing changes in pore size and elasticity. The series encompasses both, purely microporous materials as well as materials with both micropores and small mesopores. The chemical shift of the adsorbed xenon depends on xenon-wall interactions and thus on the pore size of the material. The xenon adsorption behavior of different MOFs can be observed over the whole range of relative pressure. Chemical shift adsorption/desorption isotherms closely resembling the conventional, uptake-measurement-based isotherms were obtained at 237 K where all materials are rigid. The comparable chemical environment of the adsorbed xenon in these isoreticular MOFs allows to establish a correlation between the chemical shift at a relative pressure of p/p 0 = 1.0 and the mean pore diameter. Furthermore, the xenon adsorption behavior of MOFs is studied also at 200 K. Here, structural flexibility is found for DUT-50, a material with an even longer linker than that of the previously known DUT-49. Its structural transitions are monitored by 129Xe NMR spectroscopy. This compound is the second known MOF showing the phenomenon of negative gas adsorption. Further increase in the linker length results in DUT-151, a material with an interpenetrated network topology. In situ 129Xe NMR spectroscopy proves that this material exhibits another type of flexibility compared to DUT-49 and DUT-50. Further surprising observations are made for DUT-46. Volumetric xenon adsorption measurements show that this nonflexible microporous material does not exhibit any hysteresis. In contrast, the in situ 129Xe NMR spectroscopically detected xenon chemical shift isotherms exhibit a hysteresis even after longer equilibration times than in the volumetric experiments. This indicates kinetically hindered redistribution processes and long-lived metastable states of adsorbed xenon within the MOF persisting at the time scale of hours or longer.

Introduction

Metal–organic frameworks (MOFs) are highly porous hybrid materials based on inorganic secondary building units (SBUs) that are connected by organic linkers.[1,2] Their properties can be tuned by varying the length and functionalization of the organic linkers.[3,4] Thus, MOFs have become a research field of increasing interest.[1] Low densities (down to 0.124 g/cm3)[5] and high surface areas (up to 7839 m2/g)[6] render the materials suitable for various applications. They are explored with respect to gas and energy storage, gas separation, catalysis, sensing, and as drug delivery systems in biomedicine.[1,2,4,7−9] Before applying MOFs in gas storage and separation, it is especially important to characterize the adsorption processes inside the pore system. Flexible MOFs are a peculiar subgroup of MOFs with unique properties. They undergo structural transitions induced by external stimuli such as temperature, pressure, and presence of guest molecules. Famous examples are MIL-53, SNU-9, DUT-8, and many others.[10−12]

The recently discovered material DUT-49 (Dresden University of Technology No. 49) is the first flexible MOF showing the surprising phenomenon of a negative gas adsorption (NGA).[13] Adsorbed gases such as methane or xenon are partly released from the structure at increasing pressure during the structural transition from an open-pore (op) to a contracted-pore (cp) state. Inspired by this observation, an isoreticular series was synthesized, which differs from DUT-49 only in the length of the linker molecule (cf. Figure ). The xenon adsorption behavior of this isoreticular series is investigated within the present work using a high-pressure in situ 129Xe NMR spectroscopy.

Figure 1

Structure of the investigated metal–organic frameworks and their corresponding organic linkers (a, e) DUT-48, (b, f) DUT-46, (c, g) DUT-49, and (d, h) DUT-50. Color code: Cu—turquoise, O—red, C—gray, N—blue, H—white, cuboctahedral pore—green, tetrahedral pore—blue, octahedral pore—orange.

NMR spectroscopy offers two general approaches to study porous materials like MOFs. On the one hand, it is possible to investigate the structure and dynamics of the material itself by observing the NMR-active nuclei in the corresponding compound, such as 13C or 1H.[14−16] On the other hand, it is possible to load the materials with NMR-active probe molecules or atoms, thus exploiting host–guest interactions. Gases like CO2 or Xe are often applied as probe.[17−25]

Due to its large electron cloud, the NMR chemical shift of xenon exhibits a high sensitivity. Therefore, 129Xe (I = 1/2) with a natural abundance of 26.4% is a well-suited probe to study surfaces.[26−29] The chemical shift of the adsorbed xenon in porous materials like zeolites[28−30] or MOFs[31] can be written asThe first term, δs, reflects all contributions due to interactions between xenon atoms and different surface sites of the pore system. The second term, δXe–Xe, describes the interactions between the adsorbed xenon atoms and depends on the density of xenon ρXe and the temperature.[26−28]

Since the adsorbed xenon is mobile and can undergo rapid exchange and diffusion processes inside the pore system, two fundamentally different cases must be distinguished: (i) Compounds with pores large enough to adsorb xenon but rather small interconnecting windows/channels between these pores prevent fast exchange processes between neighboring pores. Note that the diameter of xenon is ca. 4.4 Å. (ii) Compounds with pores interconnected by sufficiently large windows/channels allow a rapid xenon diffusion between different pores. A striking example for case (i) is the zeolites Na-A and Ag-A.[32−35] Xenon clusters of different sizes are then formed inside different pores—without a fast exchange between neighboring pores. This results in an ensemble of lines of different chemical shifts caused by the influence of Xe–Xe interactions, which depend on the cluster size (cf. eq ). Recently, this behavior was also observed for a special MOF denoted as MFU-4 with a remarkably small pore aperture of only 2.4 Å diameter.[36] Quite often, however, case (ii) is observed. A rapid diffusion of xenon atoms through the pore system then results in a single line with a chemical shift representing the weighted average along the trajectory of xenon. Note that the shape of this line is not necessarily symmetric. For materials with anisotropic pore systems, the line shape of the adsorbed xenon can be anisotropic as demonstrated previously.[37−42] Note that the line shape not only depends on intracrystalline exchange and diffusion but also on intercrystalline exchange processes.[41] Thus, 129Xe NMR spectroscopy provides information about adsorption processes as well as xenon dynamics in various porous materials.[26−51]

Here, high-pressure in situ 129Xe NMR spectroscopy is used to investigate host–guest interactions under controlled thermodynamic conditions. The home-built high-pressure apparatus[51] allows in situ experiments at variable pressures up to saturation and at temperatures down to 195 K. In addition to the previously studied DUT-49, a novel series of MOFs including the materials DUT-48, DUT-46, and DUT-50 was studied. These MOFs are built from copper paddle-wheels and carbazole-3,6-dicarboxylate forming metal–organic polyhedrons (MOPs). These MOPs constitute small cuboctahedral pores. Connecting the MOPs with tetratopic, organic linkers result in a framework consisting of three different types of pores. In addition to the cuboctahedral pores, tetrahedral and octahedral cavities are formed (cf. Figure a–d).[52] The length of the organic linkers is varied within the MOF family (Figure e–h), resulting in MOFs of different pore sizes and inner surface areas (Table ). The functionalization of the organic linkers as well as the MOP structure remains unchanged. DUT-49 shows an extremely high methane storage capacity.[52] Comparative investigation of the new isoreticular series thus allows to study differences in the adsorption behavior that are purely introduced by different linkers. Furthermore, it is speculated that the pore size may correlate with the chemical shift of the adsorbed xenon due to similarity in the surface chemistry provided by this special MOF series.

Table 1

Calculated Pore Sizes of DUT-48, DUT-46, DUT-49, and DUT-50

	octahedralpore (Å)	tetrahedral pore (Å)	cuboctahedral pore (Å)
DUT-48	18.7	14.8	10.7
DUT-46	21.8	15.0	10.7
DUT-49	24.8	17.6	10.5
DUT-50	30.7	21.0	10.7

Experimental Section

The dried MOFs were stored under an inert, dry atmosphere. For the experiments, the samples were transferred into a single-crystal sapphire tube and activated overnight (100 °C, 10–5 mbar). The high-pressure in situ experiments were carried out using a home-built apparatus.[51] A defined xenon pressure can be adjusted. All experiments were carried out on an Avance 300 NMR spectrometer (Bruker, Germany) equipped with an HR probe. The NMR spectra were recorded at a resonance frequency of 83.04 MHz for 129Xe using a pulse length of 11 μs and relaxation delays of 15 s. A BIOSPIN SA BCU-Xtreme unit (Bruker, Germany) was used for sample cooling. Measurements were carried out at two different temperatures of 237 and 200 K. To obtain adsorption/desorption isotherms, the chemical shift of the adsorbed xenon was observed over the whole range of pressure up to a relative pressure of p/p0 = 1. Equilibration times of 15 min after pressure changes and 30 min after temperature changes were used. The temperature calibration was carried out using the 1H signal of methanol.[53,54] As a second method of temperature calibration, the condensation pressure of xenon was used to determine the corresponding temperature from the phase diagram. The chemical shift is referenced relative to xenon gas extrapolated to zero pressure.

Results and Discussion

The flexibility of DUT-49, i.e., its phase transformations, does not only depend on the type of adsorbed gas but also on the temperature. In the case of xenon adsorption, it behaves like a rigid MOF at 237 K but is flexible at 200 K. Therefore, isothermal in situ 129Xe NMR experiments were performed at these two temperatures for the isoreticular series studied here as well.

Experiments at 237 K

129Xe NMR spectra were recorded at various pressures covering the entire relative pressure range at constant temperature. Such measurements result in chemical shift adsorption/desorption isotherms. Figure compares the chemical shift isotherms for the purely microporous material DUT-48, as well as DUT-46 and DUT-50, which exhibit micropores and small mesopores (cf. Table ). This comparison reveals striking differences for the three chemical shift isotherms. First, the shape of the chemical shift isotherm of DUT-48 resembles a reversible type I(b) isotherm.[55] This observation can be explained by the fact that the 129Xe NMR chemical shift is correlated with the amount of xenon adsorbed in the pores as predicted by the second term of eq . This correlation is the reason for the similarity between conventional adsorption/desorption isotherms obtained by uptake measurements (in mass or volume units) and the chemical shift isotherms presented here. Similar correlations were already observed by our group[56] for the microporous MOFs UiO-66 and UiO-67 (Universitetet i Oslo).[57] After a steep, almost linear increase at a low relative pressure, the isotherm starts to level off for p/p0 > 0.4. The initial regime with its linear behavior is well-known for other microporous materials such as zeolites.[28,55] In contrast to DUT-48, the isotherm of DUT-50 exhibits the characteristic S-shape (type IV) expected for mesoporous materials.[55] At lower relative pressures, the present micropores are preferentially filled and only a monolayer of xenon atoms is adsorbed in the mesopores. With increasing relative pressure, multilayer adsorption occurs inside the mesopores, which finally results in capillary condensation. Therefore, a plateau is finally observed, indicating that the maximum possible density is reached, as also observed for the microporous DUT-48. However, xenon desorption during pressure release shows a hysteresis compared to the adsorption branch for the mesoporous DUT-50, a characteristic behavior for capillary condensation inside mesopores. The isotherm of DUT-46 is located between those of the two cases described above. The chemical shift isotherm shows a type I(b) shape. However, the chemical shift increase is slower than that for DUT-48 and the plateau is finally reached at a higher relative pressure. This is due to the larger pores of the material. Nevertheless, the obtained isotherm is reversible.

Figure 2

129Xe NMR adsorption and desorption isotherms of DUT-48, DUT-46, and DUT-50 at 237 K.

It is, furthermore, interesting to note that the chemical shift of the adsorbed xenon exceeds the chemical shift of liquid xenon at a given temperature and high relative pressure. The chemical shift of 129Xe at p/p0 = 1 can be determined very precisely by the extrapolation of the chemical shift since the slope of the isotherm decreases and is almost zero at relative pressures close to one. It can be concluded from this observation in combination with eq that the xenon density inside the pores only slightly changes at sufficiently high relative pressures, i.e., that the pore system is filled at p/p0 close to 1. A chemical shift difference Δδ is observed between the chemical shift δ̅ resulting from the extrapolation of δ to p/p0 = 1 and the chemical shift δB of bulk liquid xenon of ca. 197 ppm at the given measurement temperature. This difference arises due to the fact that a certain fraction of xenon atoms is in direct contact with the MOF surface, whereas others in the middle of a sufficiently large pore only interact with other xenon atoms as in the bulk liquid. Due to the presence of exchange processes, an averaged chemical shift results. The inset of Figure illustrates this idea. Under the simplifying assumption that the surface is chemically homogeneous and the chemical shift of xenon in the surface contact is given by δS, one can then writeHere, pS and pB are the probabilities of a xenon atom to be in surface contact or in the bulk liquid, respectively. Thus, Δδ depends on the probability pS that xenon interacts with the surface. It can be written asNS and NB are the number of xenon atoms in contact with the surface and in the bulk liquid, respectively.

Figure 3

Correlation of the mean pore diameter, d, and the chemical shift difference, Δδ, of the adsorbed xenon at 237 K for DUT-48, DUT-46, DUT-49, and DUT-50. The inset shows the model of a xenon-filled pore.

The probability for a xenon atom to be in surface contact, however, will depend on the average pore size of the material. To test the validity of this consideration, the correlation between Δδ and the average pore diameter of the studied samples (cf. Table ) is shown in Figure . For the materials of the considered isoreticular series, Δδ obviously follows the expected trend: An increasing average pore size leads to a decreasing Δδ since the probability of a surface contact decreases in larger pores. Assuming spherical pores, one would expect a 1/d dependence for pS. The data do not exactly follow this dependence but show a similar trend. The deviation is not unexpected because the MOFs exhibit not only mesopores but also micropores and the above-mentioned assumptions are certainly an oversimplification. In any case, our observations show that Δδ can be used as an easily measurable estimate for the average pore size in MOFs of comparable surface chemistry.

Table 2

Average Pore Diameter d (for Detailed Information about the Calculation, see Table S2) and Chemical Shift Difference Δδ of the Adsorbed Xenon Relative to the Bulk Liquid Xenon at 237 K for DUT-48, DUT-46, DUT-49, and DUT-50

	d (Å)	Δδ (ppm)
DUT-48	15.66	32.0
DUT-46	17.59	24.2
DUT-49	20.80	21.3
DUT-50	26.98	19.2

Experiments at 200 K

As mentioned above, DUT-49 is rigid at 237 K but becomes flexible at 200 K.[13] The other members of the described isoreticular series were found to be rigid at 237 K as well. To characterize the influence of decreasing temperature, their behavior at 200 K was also studied here. Since the 129Xe chemical shift is very sensitive to structural changes, differences between chemical shift isotherms and the conventional xenon uptake measurements must now be taken into account.

For DUT-48, an almost reversible type I(b) isotherm is also observed at 200 K (Figure ). Only a minor hysteresis occurs within the experimental error. This means that the microporous DUT-48 is rigid at both measurement temperatures applied here.

Figure 4

129Xe NMR adsorption and desorption isotherms of DUT-48 (top) and DUT-46 (bottom) at 200 K.

DUT-46 contains both micropores and small mesopores (see Table ) and exhibits a type IV adsorption/desorption isotherm. However, the volumetric uptake-derived isotherm does not exhibit a detectable hysteresis. This can be explained by the rather small mesopores (21.8 Å diameter) where capillary condensation not yet plays a major role. However, the NMR-derived 129Xe chemical shift adsorption/desorption isotherm exhibits a considerable hysteresis. It should be noted that the uptake-derived isotherm was measured at an even shorter equilibration time of about 400 s for each data point than the NMR-derived chemical shift isotherm measured with 15 min equilibration for each data point. The data in the hysteresis range at relative pressures between 0.1 and 0.3 were even measured after 75 min equilibration because the equilibrium is not reached after 15 min. This follows from the observation that line width and position still change after 15 min (cf. Figure S2). The characteristic difference between the two isotherms, i.e., the absence of a hysteresis in the volumetric uptake-derived data, can thus not be explained by insufficient equilibration time for the NMR experiment. Our explanation for this behavior is a diffusion-limited redistribution of the adsorbed xenon in the pore system. Typically, the small pores fill first at low pressures. However, the windows of the smallest pores are only about 4 Å in size and the van der Waals diameter of xenon amounts to 4.4 Å. At low pressures, xenon is thus probably not (or only to a limited extent) adsorbed in these small cuboctahedral pores. However, a redistribution of the adsorbed xenon in the pore system can occur at a higher pressure, i.e., the small pores are increasingly filled. During desorption, the cuboctahedral pores are still filled and the desorption takes place at a lower relative pressure. Thus, the hysteresis in the NMR-derived isotherm could be explained. It is already known for another MOF that xenon is able to pass even smaller pore apertures and is trapped inside afterward.[36] However, this behavior can be observed at lower temperatures and pressures for DUT-46. In agreement with this idea, the integral intensity of the signal for adsorbed xenon remains almost constant during the whole time of the equilibration experiment of 75 min (cf. Figure S3). Only minor changes occur after the first few minutes of equilibration. Thus, the significant change in chemical shift and line shape occurs most likely due to the rearrangement of the already adsorbed xenon atoms. Further work is in progress to clarify this observation. In any case, it can be stated that DUT-46 and DUT-48 do not show the counterintuitive low-temperature NGA behavior of DUT-49. The reason for this is most likely the linker structure. NGA is a spontaneous desorption of gas during pressure increase.[13] Adsorbed xenon is released because the mesopores are contracted. This is accompanied by a pronounced deformation (bending) of the linkers. The energy needed for this bending is overcompensated by the higher adsorption enthalpies for guests like methane or xenon.[58] Obviously, the organic linkers used in DUT-48 and DUT-46 are too short and rigid to allow adsorption-driven bending in contrast to DUT-49. However, DUT-50 exhibits an even longer linker than DUT-49 in contrast to DUT-48 and DUT-46. The 129Xe NMR spectra recorded at 200 K are shown in Figure for the adsorption experiment, i.e., steadily increasing pressure along with the volumetric uptake-derived adsorption isotherm. The chemical shift of the adsorbed xenon steadily increases from 120 to 150 ppm with a relative pressure of up to 0.28. At a relative pressure of about 0.3, a sudden shift jump of nearly 100 ppm occurs. This jump is caused by the structural transition from an open-pore (op) state to a closed-pore (cp) state. The resulting higher probability pS for xenon to be in contact with the surface in the contracted pores causes a sudden increase in chemical shift similar to that in DUT-49.[31]

Figure 5

129Xe NMR adsorption spectra (top) and 129Xe NMR and volumetric uptake-derived xenon adsorption isotherms (bottom) of DUT-50 at 200 K.

Further pressure increase results in a moderate increase for the chemical shift of xenon in the cp state. However, a second signal at a lower chemical shift occurs at relative pressures beyond 0.45. This can be assigned to the xenon adsorbed in DUT-50op. The transformation of DUT-50 from the cp state back to the op state takes place over a relatively wide relative pressure range. At p/p0 = 1, only one signal for xenon remains, indicating the complete transformation of DUT-50 into the op state.

The coexistence of the cp and op states over a relatively wide relative pressure range between 0.46 and 0.92 could be explained by two models: (i) The structural transformation is a collective phenomenon of entire crystallites and the sample contains crystallites that exhibit different transition pressures. (ii) The pore opening starts locally at a certain pressure and does then steadily proceed into other regions of the crystallite at increasing pressure (for example, from the surface into the bulk). It is anticipated that the exchange rate of xenon between the op and cp states would be slower for the intercrystalline exchange described in (i) than for the intracrystalline exchange in (ii). Therefore, we measured the exchange rate by 2D EXSY spectroscopy (Figure ). The exchange rate can be calculated from the dependence of the intensity ratio between cross peaks and diagonal peaks on the mixing time τm. For a two-site exchange, the ratio between the intensities of the cross peaks, Icross, and diagonal peaks, Idiagonal, is described by the function[59]Figure demonstrates that the data indeed follow this function. An exchange rate kex of 95 Hz could be measured corresponding to a characteristic exchange time constant of 11 ms. This is a typical value for intercrystalline exchange processes in microcrystalline materials.[51,60] It is thus anticipated that the op–cp structural transition in DUT-50 is a collective switching of entire crystallites in an ensemble of crystallites with a slightly varying switching pressure and not a local phenomenon continuously spreading out within one crystallite. During the desorption experiment (see Figure ), the structural change from DUT-50op to DUT-50cp occurs at p/p0 = 0.3. The material stays in this state even at a very low pressure down to 0.01. This means that the described structural transitions are not reversible.

Figure 6

EXSY spectra of xenon adsorbed on DUT-50 at p = 2.0 bar and 200 K (top) and correlation between the mixing time and the intensity ratio of cross peaks and diagonal peaks measured as a function of the mixing time. The red line was determined by fitting the data to eq .

Figure 7

129Xe NMR desorption spectra (top) and 129Xe NMR and volumetric uptake-derived xenon desorption isotherms (bottom) of DUT-50 at 200 K.

The adsorption/desorption isotherm obtained from the 129Xe NMR spectra can be compared with the volumetric xenon isotherms (cf. Figures and 7, bottom). During adsorption, the jump in the chemical shift takes place, coinciding with the NGA transition. The appearance of a second signal for the adsorbed xenon shows a gradual transformation of the cp state back to the op state. During desorption, the structural change of DUT-50 is only observable within the 129Xe NMR isotherm, rendering 129Xe NMR spectroscopy highly useful to better understand the structural transformation. It can thus be stated that DUT-50 is the second known material showing a negative gas adsorption phenomenon.

Finally, another MOF denoted as DUT-151 (cf. Figure S4) with an even longer linker than that of DUT-50 (four 1,4-substituted phenylene units in the ligand backbone) was also investigated with respect to its xenon adsorption behavior. Due to its interpenetrated network structure, the MOF was not included into the pore size estimation experiments.

Nevertheless, the adsorption experiments at 200 K were performed for DUT-151 as well. The material shows structural transformations during adsorption at this temperature, which, however, differ from the behavior of DUT-50. After the activation (DUT-151act) and start of the adsorption experiment at low pressure, the material is found in the closed-pore state (DUT-151cp). At low pressure, the signal of the adsorbed xenon is observed in the range between 118 and 158 ppm (Figure ). This signal can be assigned to xenon adsorbed on the closed pores at low pressures. Beyond a relative pressure of 0.2, a second signal for the adsorbed xenon occurs at a higher chemical shift of 202 ppm. Thus, two signals can be observed simultaneously in this pressure range. However, the signal intensity of the first signal at a lower chemical shift decreases with increasing pressure and the signal finally disappears.

Figure 8

129Xe NMR adsorption spectra (top) and 129Xe NMR adsorption and desorption isotherms (bottom) of DUT-151 at 200 K.

We assume that the interpenetrated networks shift relative to each other beginning from a relative pressure of 0.2, i.e., the network seems to transform from a closed-pore structure to an open-pore structure (DUT-151op). The chemical shift jump and the appearance of the second signal is clearly visible within the adsorption isotherm of DUT-151 at 200 K (Figure , bottom). This pore opening allows more xenon adsorption as seen from the increasing total signal intensity. The tendency of the lattice to close, which must be overcome by adsorbed xenon, as well as Xe–Xe interactions may cause a higher chemical shift of the second signal. During desorption, only one signal of the adsorbed xenon is recorded and a hysteresis is observed. The chemical shift jump is observable during desorption at a lower relative pressure of 0.15. The starting value of the chemical shift of the adsorbed xenon is reached again after desorption of gas, thus the adsorption-induced structural transition seems to be reversible. The reason for the difference in adsorption and switching behavior compared with those of DUT-49 and DUT-50 (see above) is the lattice interpenetration present in DUT-151. This was found to prevent large-scale structural contraction and NGA due to a reduced void. However, sufficient void space in the pores is required for such a structural transition.

Conclusions

High-pressure in situ 129Xe NMR spectroscopy is well-suited to investigate host–guest interactions and adsorption processes in porous materials such as MOFs. The following conclusions are drawn from our studies:

Chemical shift measurement of the adsorbed xenon as a function of pressure yields chemical shift adsorption/desorption isotherms, which compare well with the conventional, uptake-measurement-based isotherms for rigid MOFs. This is because the chemical shift and the density of the adsorbed xenon are correlated.

The difference Δδ between the chemical shift of the adsorbed xenon extrapolated to a relative pressure of one and the chemical shift of bulk liquid xenon correlates with the average pore diameter for MOFs of comparable chemical composition, i.e., surface chemistry. This correlation between Δδ and the mean pore diameter is demonstrated for the series of isoreticular MOFs studied here (DUT-48, DUT-46, DUT-49, and DUT-50). The chemical shift difference of the adsorbed xenon decreases with increasing mean pore diameter. The quantity Δδ is easily and accurately measurable by high-pressure in situ 129Xe NMR spectroscopy and allows an estimation of the mean pore size.

Characterization of the adsorption processes at lower temperature (200 K) shows interesting adsorption phenomena. Similar to DUT-49, the novel compound DUT-50 also shows structural changes of the pore system during adsorption, causing a negative gas adsorption (NGA) as can be visualized by 129Xe NMR spectroscopy. 2D EXSY experiments allow to measure the exchange between the coexisting open and contracted pores. The measured exchange rate suggests that structural transitions in DUT-50 are a rather collective phenomenon of entire crystallites.

The interpenetrated compound DUT-151 shows a different switching behavior compared with DUT-49 and DUT-50. The transition is not a structural contraction but a xenon-adsorption-induced mutual shift of the two interpenetrating networks from a closed-pore to an open-pore state of the MOF.

The nonflexible compound DUT-46 exhibits a hysteresis in the 129Xe NMR isotherm due to diffusion-limited adsorption/redistribution processes at low temperature.