Lea A.
Brandner
a,
Mercedes
Linares-Moreau
a,
Guojun
Zhou
b,
Heinz
Amenitsch
c,
Simone
Dal Zilio
d,
Zhehao
Huang
b,
Christian
Doonan
*e and
Paolo
Falcaro
*a
aInstitute of Physical and Theoretical Chemistry, Graz University of Technology, 8010 Graz, Austria. E-mail: paolo.falcaro@tugraz.at
bDepartment of Materials and Environmental Chemistry, Stockholm University, Stockholm SE-106 91, Sweden
cInstitute of Inorganic Chemistry, Graz University of Technology, 8010 Graz, Austria
dCNR-IOM – Istituto Officina dei Materiali, SS 14, km 163.5, Basovizza, Trieste, 34149, Italy
eDepartment of Chemistry, The University of Adelaide, Adelaide, South Australia 5005, Australia. E-mail: christian.doonan@adelaide.edu.au
First published on 17th October 2023
3D-oriented metal–organic framework (MOF) films and patterns have recently emerged as promising platforms for sensing and photonic applications. These oriented polycrystalline materials are typically prepared by heteroepitaxial growth from aligned inorganic nanostructures and display anisotropic functional properties, such as guest molecule alignment and polarized fluorescence. However, to identify suitable conditions for the integration of these 3D-oriented MOF superstructures into functional devices, the effect of water (gaseous and liquid) on different frameworks should be determined. We note that the hydrolytic stability of these heteroepitaxially grown MOF films is currently unexplored. In this work, we present an in-depth analysis of the structural evolution of aligned 2D and 3D Cu-based MOFs grown from Cu(OH)2 coatings. Specifically, 3D-oriented Cu2L2 and Cu2L2DABCO films (L = 1,4-benzenedicarboxylate, BDC; biphenyl-4,4-dicarboxylate, BPDC; DABCO = 1,4-diazabicyclo[2.2.2]octane) were exposed to 50% relative humidity (RH), 80% RH and liquid water. The combined use of X-ray diffraction, infrared spectroscopy, and scanning electron microscopy shows that the sensitivity towards humid environments critically depends on the presence of the DABCO pillar ligand. While oriented films of 2D MOF layers stay intact upon exposure to all levels of humidity, hydrolysis of Cu2L2DABCO is observed. In addition, we report that in environments with high water content, 3D-oriented Cu2(BDC)2DABCO recrystallizes as 3D-oriented Cu2(BDC)2. The heteroepitaxial MOF-to-MOF transformation mechanism was studied with in situ synchrotron experiments, time-resolved AFM measurements, and electron diffraction. These findings provide valuable information on the stability of oriented MOF films for their application in functional devices and highlight the potential for the fabrication of 3D-oriented superstructures via MOF-to-MOF transformations.
An aspect of MOF devices that requires further study is how the material structure changes under typical working conditions, e.g. humidity. We hypothesized that the exposure of heteroepitaxially oriented MOF films to humidity could result in chemical and structural changes. In the literature, several Cu-based MOFs were reported to show sensitivity towards water, resulting in degradation of the crystal structure and loss of functional properties.27–33 For archetypical HKUST-1 (Cu3(BTC)2, BTC = 1,3,5-benzene tricarboxylate),34 it has been reported that a relative humidity (RH) above 50% RH results in the decomposition of the framework.35 Interestingly, for other MOFs (e.g. MOF-177), the interaction with water molecules triggers the re-organization of the unit cell into new crystalline phases.36 Recently, Fu and co-workers reported the epitaxial recrystallization from a single-crystal MOF to an oriented MOF superstructure (i.e. an ordered polycrystalline structure).37 Inspired by these works, we decided to examine the effect of water on heteroepitaxially grown Cu-MOF films. We examined two aspects: (i) the hydrolytic stability of different heteroepitaxially grown Cu-MOF films and (ii) the potential rearrangement of 3D-oriented MOF films into different ordered polycrystalline structures. In particular, in this study, we assess the influence of water on four different oriented Cu-MOF films (Cu2L2, Cu2L2DABCO, (L = 1,4-benzenedicarboxylate, BDC; biphenyl-4,4-dicarboxylate, BPDC) exposed to varying levels of humidity (Fig. 1). Heteroepitaxial MOF films were grown from sacrificial Cu(OH)2 oriented nanostructures and exposed to 50% RH, 80% RH, and liquid water. Changes in crystallinity, chemical composition, and morphology were monitored over a period of 7 days, using X-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FTIR), and scanning electron microscopy (SEM). While the 2D-frameworks Cu2(BDC)2 and Cu2(BPDC)2 maintained their physical and chemical properties in the presence of water, we observed the hydrolysis and concomitant transformation from Cu2(BDC)2DABCO films to Cu2BDC2, as well as structural degradation of Cu2(BPDC)2DABCO starting from 50% RH. To elucidate the timescale and mechanism of the MOF transformation process, we performed time-resolved grazing incidence wide-angle X-ray scattering (GIWAXS) experiments using synchrotron radiation. These results were also correlated to time-resolved atomic force microscopy (AFM) measurements to study the corresponding changes in morphology and electron diffraction for structural changes. The collected data reveal that, at room-temperature, the MOF transformation process starts with the dissolution of the MOF crystals, then the initial 3D-oriented polycrystalline film undergoes heteroepitaxial recrystallization into a distinct 3D-oriented polycrystalline superstructure. We posit that this study (i) will help the progress of oriented MOF films towards micro- and opto-electronic devices and (ii) will advance the knowledge on ordered polycrystalline MOF superstructures.
Fig. 1 Schematic illustration of the structural evolution of heteroepitaxially grown, oriented MOF films under humid conditions. |
For the as-synthesized MOF films, diffraction plots (Fig. 2a) were consistent with those reported in the literature. Specifically, we observed diffraction peaks at different angles: 8.38°, assigned to the (001) plane of Cu2(BDC)2;19 5.93°, assigned to the (001) plane of Cu2(BPDC)2;19 8.16°, assigned to the (010) plane of Cu2(BDC)2DABCO;20,21 and 5.85°, assigned to the (010) plane of Cu2(BPDC)2DABCO.20 Azimuthal intensity in-plane XRD-scans confirmed the successful growth of the oriented MOF films (Fig. S2†). Reflections of residual Cu(OH)2 at 16.74° and 23.83°, assigned to the (020) and (021) planes, were also observed for Cu2(BDC)2, Cu2(BPDC)2, and Cu2(BDC)2DABCO in the out-of-plane XRD patterns. For Cu2(BPDC)2DABCO, these signals were absent in the diffraction pattern, indicating full consumption of the sacrificial Cu(OH)2 NBs. Next, we analysed the diffraction patterns of the films exposed to the different RH values. For the heteroepitaxially grown 2D frameworks Cu2(BDC)2 and Cu2(BDPC)2, the peak areas associated with the respective (001) reflections changed within 10–15% of the original signals under all tested conditions (Fig. 2b, c and S3†). It is noteworthy that the error bars overlap for all data points; this excludes significant changes in crystallinity with increasing RH and exposure time. In contrast, we found a significant decrease in signal intensity for Cu2(BDC)2DABCO and Cu2(BPDC)2DABCO films after exposure to the different humid environments (Fig. 2d and e). For Cu2(BDC)2DABCO, the peak areas assigned to the (010) reflection decreased by 90% after exposure to 50% RH for 1 d. In addition, the peak position shifts from 8.16° to 8.27° (Fig. S4†). Interestingly, at 80% RH and in water, the peak area stabilized at approximately 15–18% of the original value, but displayed an even larger shift to 8.37° (Fig. S5†), which is consistent with the (001) reflection of Cu2(BDC)2. It was also observed that for Cu2(BDC)2DABCO films, the diffraction signals associated with the (021) plane of the remaining Cu(OH)2, exhibited similar intensity at 50% RH, but decreased slightly at 80% RH and in water (Fig. S5†). For oriented Cu2(BPDC)2DABCO films, the peak areas attributed to the signal of the (010) reflection dropped by 75% after the first day under 50% RH, followed by another intensity decrease down to 20% of the original peak area after 7 d (Fig. S6†). At 80% RH and in water the integrated peak areas dropped by over 90% within 24 h, indicating almost complete loss of crystallinity (Fig. S6†).
Next, we analysed the chemical changes in the different MOF films by monitoring characteristic vibrational modes over time at the different RH values. We focused on the asymmetric and symmetric carboxylate stretching modes and compared to the pristine samples and previously reported spectra23,24,42 to determine the coordination and protonation state of the carboxylic ligands. For the as-synthesized Cu2(BDC)2 films, the asymmetric and symmetric carboxylate stretching modes were observed at 1579 cm−1 and 1398 cm−1, respectively. For pristine Cu2(BPDC)2, the corresponding vibrational bands appeared at 1583 cm−1 and 1423 cm−1. In the case of Cu2(BDC)2DABCO, we recorded absorption bands at 1630 cm−1, assigned to νasymm(COO−), and overlapping signals at 1433 cm−1 and 1394 cm−1, associated with the νsymm(COO−) vibration and a benzene ring mode.27 As-synthesized Cu2(BPDC)2DABCO films exhibited an asymmetric carboxylate stretching mode at 1623 cm−1 and a broad band at 1406 cm−1, assigned to νsymm(COO−).
During the exposure period of 7 days, the vibrational modes of Cu2(BDC)2 and Cu2(BPDC)2 preserved both intensity and position; no additional bands appeared in the spectra for all tested conditions (Fig. S7†). These data indicate the negligible influence of water on the chemical composition of the samples. For the Cu2(BDC)2DABCO films, the IR spectra revealed several changes induced by humid conditions. After exposure to 50% RH for 1 d, the carboxylate modes lost approximately half their intensity, indicating a substantial degradation of the sample. Additionally, a new vibrational band appeared at 1579 cm−1; this mode could be ascribed to asymmetric carboxylate stretching of Cu2(BDC)2 (Fig. S8†). Vibrational modes attributed to the DABCO pillar ligand in the region at 2960 cm−1 (in-phase νasymm(CH2)), at 1318 cm−1 (in-phase γtwist(CH2), in-phase νasymm(NC3)) and at 1060 cm−1 (in-phase νasymm(NC3), νasymm(C–C), γtwist(CH2)) were also observed43 and showed only a minimal change in intensity when exposed to 50% RH over time (Fig. S9†). Similar changes in the spectra were observed for Cu2(BDC)2DABCO films exposed to 80% RH and water; however, in these samples, the initial asymmetric COO− band disappeared completely, resulting in carboxylate modes consistent with pure Cu2(BDC)2 MOF (Fig. S8†). Additionally, the vibrational bands attributed to DABCO disappeared in water, indicating the removal of the pillar ligand from the framework (Fig. S9†). At 80% RH, the intensity of the DABCO-related band at 1060 cm−1 decreased over time and Raman maps suggest that DABCO molecules rearrange and aggregate in different regions of the film surface (Fig. S10†). For Cu2(BPDC)2DABCO films, a similar drop in intensity of the asymmetric carboxylate stretching vibration was observed within the first 24 h at 50% RH. At the same time, new bands at 1586 cm−1 and 1540 cm−1 emerged, matching the vibrational modes of Cu2(BPDC)2 (Fig. S11†). At 80% RH and in water, the asymmetric COO− mode of the initial framework disappeared completely and only the vibrational signals similar to Cu2(BPDC)2 remained (Fig. S11†). While the in-phase νasymm(CH2) mode of the DABCO shows little intensity for as-synthesized and exposed films, the vibrational bands at 1318 cm−1 and 1060 cm−1 are stable at 50% RH, but disappear at high water loading, similar to the observations for Cu2(BDC)2DABCO (Fig. S12†).
To correlate the changes observed in the XRD patterns and FTIR spectra with changes of the films' morphologies, we examined the samples as-synthesized and after exposure to the different environments, using scanning electron microscopy (SEM). Pristine films of the 2D frameworks Cu2(BDC)2 and Cu2(BPDC)2 showed small, plate-like crystals with diameters of approximately 150 nm (Fig. 3a and b). In contrast to this, micrographs of Cu2(BDC)2DABCO films revealed partially intergrown, cuboid crystals, with an average length of 600 nm and width in the 150–400 nm range (Fig. 3c). Pristine Cu2(BPDC)2DABCO films also exhibited a cuboid morphology with crystal lengths up to 2.2 μm and an average width of 500 nm (Fig. 3d). These morphologies are comparable with previous reports of heteroepitaxially grown Cu-MOF films from Cu(OH)2 substrates.19–21
When compared to the micrographs of as-synthesized samples, the morphologies of Cu2(BDC)2 and Cu2(BPDC)2 exposed to different RH values remain unchanged (Fig. 3e, f, S13 and S14†) indicating minimal impact of humidity and water on the 2D-layered MOF crystal shapes. However, significant changes were observed in the case of Cu2(BDC)2DABCO. After 7 days at 50% RH, the original large cuboid crystals are replaced by small plate-like crystals (Fig. 3g and S15†). We note that the new crystalline morphology is similar to the one of Cu2(BDC)2 MOFs (Fig. 3a and e). At higher water loadings, the transformation from a cuboid to a plate-like morphology occurs at a faster rate (e.g. < 1 d at 80% RH and in H2O). These observations suggest the transformation from Cu2(BDC)2DABCO to Cu2(BDC)2.
For Cu2(BPDC)2DABCO, the overall cuboid crystal shapes remained prominent after the 7 d exposure to 50% RH, but less defined crystal edges and increased surface roughness were observed (Fig. 3h). Under high humidity conditions, individual crystals became gradually more difficult to distinguish and micrographs show holes covering the surface of the remaining material (Fig. S16†).
Our experimental results show that humidity has negligible effects on the crystallinity, chemical composition and morphology of heteroepitaxially grown Cu2(BDC)2 and Cu2(BPDC)2 films. Thus, these oriented 2D Cu-MOF frameworks are structurally stable even in highly humid environments. The similar behaviour observed for both linkers (BDC and BPDC) suggests that for these MOFs, the small hydrophobic pore windows prevent an effective mass transfer of water molecules into the larger in-plane pores. Thus, the MOF crystal orientation could have a stabilizing contribution. We note that the effect of hydrophobic pore windows was also previously reported for hydrophobic frameworks, which showed a significantly higher affinity towards non-polar molecules over polar guests.47
In contrast, in the exposed 3D-oriented Cu2(BDC)2DABCO film, reduced crystallinity and decreased intensity of the carboxylate vibrations in the IR spectra are ascribed to the hydrolysis of the framework. In addition, the appearance of new vibrational bands and a plate-like morphology indicate the formation of Cu2(BDC)2. Although partial decomposition of powdery Cu2(BDC)2DABCO at 50% RH has been previously observed,27 the transformation from 3D-oriented Cu2(BDC)2DABCO films to Cu2(BDC)2 coatings has not been reported to date. In the PXRD patterns recorded of Cu2(BDC)2DABCO exposed to 50% RH, the shift of the (010) reflection to higher 2θ values indicates the partial hydrolysis of the framework as previously described.27 However, the co-presence of vibrational bands that can be ascribed to Cu2(BDC)2DABCO and Cu2(BDC)2, indicates that at 50% RH the MOF-to-MOF transformation is incomplete. In contrast, when Cu2(BDC)2DABCO is exposed to 80% RH or liquid water, the final diffraction patterns and IR spectra match those of pure Cu2(BDC)2, suggesting a complete transformation from the pillar-layered MOF to the 2D framework. We note that there is an emerging research in the synthesis of MOF superstructures;16,17,37 for example, it has been demonstrated that a Zn2(BDC)2DABCO MOF single crystal can evolve in an iso-oriented MOF superstructure.48 While in single-crystal transformations, the crystal shape is typically preserved,49 our SEM data indicates that the initial Cu2(BDC)2DABCO crystals are replaced by Cu2(BDC)2 crystals with different morphology (see RH = 80% Fig. S15†). This morphological change could indicate that the formation of the 2D MOF occurs via recrystallization of building units that were initially part of the Cu2(BDC)2DABCO framework (Fig. 4c). Considering that DABCO has a water solubility of 61 g/100 g H2O,50 whereas H2BDC is rather insoluble,51 it is expected to have a high concentration of BDC, originated from the dissolution of the original framework, available for recrystallization. On the other hand, after hydrolysis of the coordination bonds, DABCO is expected to be easily removed from the framework. This hypothesis is confirmed by the absence of the DABCO vibrational bands from Cu2(BDC)2DABCO films after exposure to liquid water. We further note that Fischer and co-workers reported that Cu2(BDC)2DABCO does not form in the presence of water (40 °C), but the synthesis yields Cu2(BDC)2 crystals instead.52 This suggests that, under the investigated humid conditions, the recrystallization of Cu2(BDC)2DABCO at room temperature is unlikely while the formation of Cu2(BDC)2 is favoured.
We hypothesize a second contribution to the growth of Cu2(BDC)2: by SEM we noted that in the Cu2(BDC)2DABCO films, we could observe regions exposing uncoated Cu(OH)2 NBs (Fig. S15†). The presence of these residual NBs in pristine Cu2(BDC)2DABCO films was also confirmed by XRD, showing the (021) reflection of Cu(OH)2 in the diffraction pattern. When immersing Cu2(BDC)2DABCO films in liquid water, we then observed the formation of plate-like Cu2(BDC)2 crystallites in these regions of exposed NBs. X-ray diffraction analyses of these films showed that the appearance of Cu2(BDC)2 was concomitant with a slight decrease of the (021) Cu(OH)2 intensity. This suggests that after degradation of the original framework, Cu2(BDC)2 could be partially formed from unreacted Cu(OH)2 directly on the NBs.
Overall the data suggest that immersion of Cu2(BDC)2DABCO films in water triggers both, a heteroepitaxial transformation into Cu2(BDC)2 and a fresh heteroepitaxial growth from the exposed Cu(OH)2.
In order to assess how the transformation process influences the overall crystalline orientation of the MOF film, we investigated the azimuthal angle dependence of the intensity of the respective (100) reflections of Cu2(BDC)2DABCO and Cu2(BDC)2 before and after exposure to 80% RH and liquid water. The intensity profile of the as prepared 3D-oriented Cu2(BDC)2DABCO film showed two intensity maxima of the (100) reflection at approx. 60° and 240° (Fig. 5a). The (001) planes in the same sample have their maxima shifted by 90°, complying with the tetragonal MOF unit cell53 and confirming the in-plane order in the as-synthesized film. After exposure to 80% RH, the intensity profile of the (100) reflection ascribed to Cu2(BDC)2 shows also two intensity maxima shifted by 180° (Fig. 5b); the profile of the (001) reflection confirms the decomposition of Cu2(BDC)2DABCO crystals. Similar intensity profiles were observed when Cu2(BDC)2DABCO films were immersed in water; however, in this case, a broader distribution of crystalline orientations was recorded (Fig. S17†). Notably, the position of the maxima of the new Cu2(BDC)2 phase coincides with the maxima of the pristine oriented Cu2(BDC)2DABCO. This indicates that the (100) planes of Cu2(BDC)2DABCO might take an active role in the recrystallization of the 2D MOF layers by directing the growth of Cu2(BDC)2 during the MOF-to-MOF transformation process. The mechanism could be explained by the lattice match between the two frameworks, which was also recently reported for the epitaxial recrystallization of oriented MOF nanostructures from a labile Cu-MOF single crystal.37 This epitaxial recrystallization from 3D-oriented Cu2(BDC)2DABCO films to Cu2(BDC)2 coatings, here supported by chemical, structural and morphological investigations, is the first evidence that an oriented MOF superstructure could be transformed into an iso-oriented, chemically and structurally distinct MOF superstructure. To further examine the kinetics and morphological changes of this process, we performed additional in situ experiments (vide infra).
Fig. 5 (a) Azimuthal intensity profile of the (100) and (001) reflection of Cu2(BDC)2DABCO before exposure to 80% RH, (b) azimuthal intensity profile of the (100) reflection of newly formed Cu2(BDC)2 after exposure to 80% RH and of the (001) reflection of Cu2(BDC)2DABCO, confirming the absence of the DABCO pillar ligand. The average intensity of the (001) reflection of Cu2(BDC)2DABCO in (b) is similar to the lowest intensity measured in the azimuthal scan of Cu2(BDC)2 (see Fig. S17† for plots with intensities). |
For the heteroepitaxially grown Cu2(BPDC)2DABCO MOF film exposed to 50% RH, the crystallinity also decreased significantly within the initial 24 h, suggesting the hydrolysis of the network. However, this drop is not as drastic as for Cu2(BDC)2DABCO under the same environmental condition (Cu2(BDC)2DABCO: 90% drop of the signal after 24 h at 50% RH), consistent with the hypothesis that higher relative humidity is needed to effectively displace the linkers in MOFs with larger pore sizes. Interestingly, the IR spectra of exposed Cu2(BPDC)2DABCO films show the appearance of vibrational bands that could be ascribed to the presence of Cu2(BPDC)2. This suggests that during the storage in humid environments, the crystalline Cu2(BPDC)2DABCO 3D framework tends to evolve into a Cu2(BPDC)2 2D network. A transformation mechanism similar to that of Cu2(BDC)2DABCO seems plausible due to the low water-solubility of the BPDC linker. However, unlike the previously described MOF-to-MOF transformation with BDC, the XRD diffraction patterns show minimal crystallinity with increasing water loading (Fig. S6†). This is consistent with the SEM micrographs, which show how Cu2(BPDC)2DABCO crystals progressively lose their sharp crystal edges and holes appear on the crystal surface (Fig. S16†); this morphological change is typically indicative of degradation by-products with low crystallinity.33,54
High-resolution GIWAXS patterns were recorded using a humidity chamber set at 50% RH, 80% RH, and increasing humidity from 5–90% RH. For these in situ studies, a time resolution of 60 s between the acquisition of the diffractograms was selected. Additionally, we monitored the transformation of Cu2(BDC)2DABCO films immersed in water, using a customized chemical cell at an increased time resolution of 1.1 s. For data analysis, we considered the out-of-plane component of the collected 2D images to monitor the (010) reflection of Cu2(BDC)2DABCO, and the (001) reflection of Cu2(BDC)2, and correlated the structural changes to the water loading during the in situ measurements. Since the d-spacing of the investigated reflections of the 3D and 2D framework is similar (1.08 nm for Cu2(BDC)2DABCO, and 1.05 nm for Cu2(BDC)2), the collected reflections partially overlap. To selectively follow the degradation of oriented Cu2(BDC)2DABCO crystals, we also monitored the (110) reflection (Fig. S4†). In addition, to examine the role of Cu(OH)2 in the transformation process, we monitored the (021) reflection ((021) does not overlap with second order reflections from Cu2(BDC)2DABCO or Cu2(BDC)2 in the region at 11.5 nm−1).
The time-resolved data reveals that the transformation of Cu2(BDC)2DABCO to Cu2(BDC)2 is highly dependent on the applied water loadings. GIWAXS data show that up to 50% RH, diffraction patterns remain identical to the as-synthesized sample (Fig. S18†), indicating water-stability up to this humidity value. However, once 50% RH is reached, the intensities of the (010) and (110) Cu2(BDC)2DABCO reflections rapidly drop and shift to higher q values (Fig. S18 and S19†). The shift was previously observed by Chabal and co-workers with a laboratory X-ray diffractometer and attributed to the partial hydrolysis of Cu2(BDC)2DABCO powder exposed to 50% RH.27 In our GIWAXS investigation, from 50% to 70% RH, the intensity and the d-spacing of the diffraction signal further decreases, indicating enhanced sensitivity towards higher water contents after this intermediate stage (Fig. S19†). Between 70% and 80% RH, the (110) reflection of Cu2(BDC)2DABCO eventually disappears and only the diffraction pattern of Cu2(BDC)2 remains (Fig. 6a, S19, and Movie S1†), suggesting the end of the heteroepitaxial conversion from Cu2(BDC)2DABCO to Cu2(BDC)2. The integrated intensity of the (010)-(001) region and the (110) reflection of Cu2(BDC)2DABCO over the whole experiment time span reveals that the transformation is not a linear process, and it occurs in two steps (Fig. 6b). At 50% RH, an initial sharp decline can be ascribed to the partial hydrolysis of the MOF.27 However, the (110) plateaus until 70–80% RH, where a second abrupt decline of intensity evidences a significant degradation of the remaining 3D framework. Due to the overlapping signals of Cu2(BDC)2DABCO and Cu2(BDC)2 in the region between 5.6 and 6.4 nm−1, this second step could not be observed for the (010) reflection and it is notably difficult to assess when the formation of the 2D framework takes place. However, by plotting the peak position maxima of the (010)-(001) reflection range versus the applied humidity, the two-step mechanism is also evident (Fig. 6c). In accordance with the integrated intensity profile of the (110) reflection of Cu2(BDC)2DABCO, the first step takes place at 50% RH, showing a shift from 5.81 nm−1 to 5.89 nm−1, which is consistent with the previously reported increase of the (010) interlayer distance caused by the partial hydrolysis of the 3D-framework.27 The second shift in the q-value at 70–80% RH can then be attributed to the recrystallization of the Cu2(BDC)2 film upon dissolution of the DABCO ligand. When RH is increased by 10% min−1 and held at a target value of 80%, the transformation to Cu2(BDC)2 is complete within 20 min (Fig. S20†). Under immersion in water, the kinetics of the process is faster and it takes only a few seconds (Fig. S21†). This suggests that, under these conditions, the coordination bonds of Cu2(BDC)2DABCO are almost instantly broken. When examining the Cu(OH)2 patterns, we note that the signal assigned to the (021) reflection of Cu(OH)2 remains unchanged up to 50% RH, and shows only a slight decrease in intensity for higher RH values including liquid water (Fig. S18–20†). This slight consumption of the residual metal hydroxide concomitant to the formation of Cu2(BDC)2 suggests a partial heteroepitaxial growth process on aligned Cu(OH)2 nanobelts.
High magnification SEM micrographs recorded after the in situ experiment at 50% RH show crystals of Cu2(BDC)2DABCO with rounded corners and soft edges, which is in accordance with the reduced crystallinity observed for this intermediate stage of the transformation process (Fig. S22†). In contrast, after exposure to 80% RH for 30 min, the previously described regions of concentrated polycrystalline Cu2(BDC)2 were observed while the surrounding Cu(OH)2 nanostructures seemed mostly unaffected (Fig. S23†). In case of exposure of Cu2(BDC)2DABCO to liquid water, the images show densely populated regions surrounded by single Cu2(BDC)2 particles on the previously unreacted Cu(OH)2 NBs (Fig. S24†). These results are consistent with the previously recorded SEM micrographs (vide supra). SEM-EDX analysis, comparing pristine Cu2(BDC)2DABCO and Cu2(BDC)2 after transformation at 80% RH, shows that nitrogen is effectively removed in the transformed film. This indicates that DABCO is indeed released from the framework and its sublimation62 is enhanced under high-vacuum conditions during the SEM-EDX measurement (Fig. S25†).
To further examine the morphological changes during the heteroepitaxial transformation from Cu2(BDC)2DABCO to Cu2(BDC)2, time-resolved AFM measurements were performed in a small environmental chamber under humid Ar flow (80% RH at exit of the chamber). The in situ morphological analysis revealed that within a few minutes in humid conditions, terrace-like steps and rounded crystal corners appear, indicating a degradation of the original cuboid morphology for the Cu2(BDC)2DABCO film. Over time (Fig. 6d, e, S26, and Movie S2†), we observed the formation of the plate-like morphology of Cu2(BDC)2, which preferentially grows in an ordered fashion, spreading from Cu2(BDC)2DABCO crystal edges towards the centre of the crystal faces. The higher reactivity of crystal edges was previously noticed and ascribed to a higher concentration of defects.63 During the transformation, the crystallite size changed from approx. 530 × 290 nm (initial Cu2(BDC)2DABCO) to 110 × 80 nm (transformed Cu2(BDC)2) (Table S1†). In addition to this, 3D AFM topography images show that the maximum height of 648 nm of the Cu2(BDC)2DABCO film is reduced to 560 nm after exposure to 80% RH for 100 min (Fig. 6d, e). This decrease in the film thickness can be attributed to partial material loss or vertical contraction as the initial framework progressively transforms from the surface towards underlying layers. Similar changes in film thickness were estimated from SEM micrographs recorded at a 45° tilting angle (Fig. S27†).
To validate the phase transition from Cu2(BDC)2DABCO to Cu2(BDC)2, selected area electron diffraction (SAED) was used to analyse the nanocrystals scratched from the Si substrates. For a single pristine Cu2(BDC)2DABCO crystallite, the obtained electron diffraction pattern revealed sharp spots (Fig. 6f) with a calculated d-value of d010 = 1.08 nm, which agrees well with our PXRD data (d = 1.08 nm) and the reported crystal structure of the MOF.20,53 Then, to access the intermediate stage of the transformation, a Cu2(BDC)2DABCO film was exposed to 80% RH for 10 min, solvent-exchanged to dichloromethane, evacuated, and stored under inert conditions (Ar) until measurement. The electron diffraction of this sample shows the previously described diffraction spots of the Cu2(BDC)2DABCO phase (d010 = 1.08 nm) as well as broad diffraction rings attributed to Cu2(BDC)2, indicating the co-existence of both MOF phases (Fig. 6g). The diffraction rings consist of overlapping diffraction spots, which can be attributed to the electron diffraction from multiple Cu2(BDC)2 crystallites being included in the selected area. The calculated d-value of d010 = 0.59 nm corresponds to that of reported Cu2(BDC)2.19 For the film exposed to 80% RH for 3 h, the electron diffraction pattern shows only diffraction rings of Cu2(BDC)2 (d010 = 0.59 nm), confirming the complete MOF-to-MOF transformation in the crystal phases (Fig. 6h).
In order to determine how the MOF-to-MOF transformation influences the porosity of the different frameworks, powdery Cu2(BDC)2DABCO was exposed to H2O and the formation of the 2D framework confirmed by PXRD (Fig. S28†). N2 adsorption isotherms then revealed that the transformed Cu2(BDC)2 shows similar properties as pristine Cu2(BDC)2, with BET surface areas of 160 m2 g−1 and 145 m2 g−1, respectively (Fig. S28†).
Overall, these changes observed in morphology, in situ X-ray scattering, and electron diffraction demonstrate that at 80% RH the transformation from Cu2(BDC)2DABCO films to Cu2(BDC)2 occurs preferentially via a dissolution and recrystallization pathway. The time-resolved experiments further demonstrate that a 3D-ordered MOF superstructure can undergo a heteroepitaxial recrystallization mechanism.
While the degradation product of Cu2(BPDC)2DABCO was found to be amorphous, we observed conditions (i.e. RH > 70%) in which an ordered film of Cu2(BDC)2DABCO transforms into an ordered film of Cu2(BDC)2. By combining time-resolved synchrotron experiments and in situ AFM measurements, we identified salient aspects of the transformation mechanism, including the hydrolysis of the original framework and the recrystallization of the building blocks as an oriented 2D MOF coating. The recrystallization was demonstrated to follow a heteroepitaxial pathway. These results highlight that in conditions that allow a MOF-to-MOF transformation, an original organization of crystals in a 3D-oriented superstructure could lead to oriented, but distinct polycrystalline MOF systems. Our results suggest that the conversion among 3D-oriented superstructures could be a valuable alternative synthetic approach for the fabrication of oriented MOF films.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3sc04135b |
This journal is © The Royal Society of Chemistry 2023 |