Reversible structural switching of a metal–organic framework by photoirradiation

Varvara I. Nikolayenko, Simon A. Herbert and Leonard J. Barbour*
Department of Chemistry and Polymer Science, University of Stellenbosch, Matieland 7600, South Africa. E-mail: ljb@sun.ac.za

Received 3rd August 2017 , Accepted 23rd August 2017

First published on 23rd August 2017


A photoresponsive metal organic framework material undergoes switching of its pore volume and sorption capacity. UV irradiation of the crystals causes cyclisation within the bis-thienylcyclopentene bridging ligands, thereby altering the node positions relative to one another along the Zn–L–Zn linkages. Incorporation of conformational flexibility into the dicarboxylic acid co-ligands facilitates the change in the framework geometry enforced by photocyclisation.


Third generation1 metal–organic frameworks (MOFs) are dynamic materials capable of undergoing reversible structural changes, with applications including gas or vapour separation and storage,2 controlled drug delivery,3 chemical sensors4 and proton conductivity.5 In this context, there have been several reports of flexible porous materials for which the extent of guest uptake depends on structural transformations experienced by the host framework.6,7 Activation of an as-synthesised MOF typically requires the application of heat and/or vacuum in order to desorb the original guest molecules, and alternative activation techniques such as supercritical CO2 drying have also been reported.8 Subsequent targeted sorption of volatile guest molecules in appreciable quantities requires their presence at sufficient pressure (or partial pressure), and often also cryogenic conditions. Guest release is then achieved by reduction of pressure,9 inert gas flushing,10 heating,11 or a combination of these operations.12 For some applications (e.g. in situ reactivation of spent sensors) such procedures may be inconvenient and energy-intensive. However, this problem could be circumvented by incorporating stimulus responsive ligands into dynamic MOFs, thereby enabling remote control of the sorption and desorption capabilities of a material.

Photochromism is the reversible interconversion by photoirradiation between two isomers possessing distinct absorption spectra.13,14 Although photochromic compounds such as spirobenzopyrans,15 azobenzenes (ABs)16,17 and fulgides18,19 have been studied extensively, few exhibit photochromism in the solid-state.20 Moreover, because many photoinduced isomers lack thermal stability, their crystallographic analysis is quite rare. Upon irradiation with ultraviolet light, bis-3-thienylcyclopentenes (BTCPs, Fig. 1) can undergo photocyclisation in the solid-state to yield thermally stable ring-closed isomers.21 Subsequent exposure to visible light induces cycloreversion to the ring-open form and these compounds have shown impressive solid-state fatigue resistance (>100 cycles), with relatively high thermal stability of both isomers. Although BTCPs can exist in two conformations (parallel and antiparallel), photocyclisation only appears to occur for the antiparallel isomer.19 It has also been suggested that ring closure requires the non-bonded distance between reactive carbon atoms to be less than 4 Å,22,23 and that perfluorination of the cyclopentene ring enhances reversibility of photocyclisation and inhibits degradation.24,25


image file: c7cc06074b-f1.tif
Fig. 1 Reversible photocyclisation of bis-3-thienylcyclopentenes using ultraviolet and visible light.

Over the past two decades Irie's high-profile work on purely molecular BTCPs has done much to promote these compounds as stimulus-responsive materials.23,26 Considering that 1D coordination polymers based on related bis-3-thienylethene ditopic ligands were first reported in 1996 by Munakata et al.,27 it is quite surprising that similar ligands were only incorporated into MOFs quite recently.28,29 Indeed, according to the recent mini review by Castellanos et al.,30 the incorporation of photo switchable components into MOFs is not a novel concept, as illustrated by recent examples involving the inclusion of photoswitchable guests into framework cavities, bridging ligands with AB or BTCP sidearms, as well as nine reports of either AB31,32 or BTCP28,29,33–37 units incorporated into the metal–ligand–metal backbone. Only the recent report by Kitagawa et al. shows structural data to support the notion of bulk photoinduced ligand modification.37

Luo et al.29 published a study of [Zn(L1)bpdc]·2DMF·H2O (DMOF, Fig. 2) reporting the as-synthesised crystal structure, thermal analysis and CO2 sorption isotherms. When colourless crystals of DMOF were exposed to UV irradiation they became blue and the transformation could be reversed using visible light. Since similar observations had already been reported for BTCP molecules23,26,27 (both free and metal-coordinated), the authors ascribed the colour change to bulk transformation of L1 from the ring-open to the ring-closed conformation. The study was concluded with low pressure CO2 sorption experiments for an activated sample under different irradiation conditions. At 1 bar a marked difference in uptake capacity was observed, depending on the irradiation wavelength used (5.0 cm3 g−1 under visible irradiation and 20.1 cm3 g−1 under UV irradiation).


image file: c7cc06074b-f2.tif
Fig. 2 Preparation of four MOFs containing BTCP. DMOF-DMOF3 were obtained from various combinations of Zn(NO3)2 with the ligands L1, L2, 4,4′ biphenyl dicarboxylic acid (bpdc) and 4,4′-oxybisbenzoic acid (oba).

Independently of the abovementioned report29 we also prepared crystals of DMOF (see Fig. S8, ESI) and extended the study by determining the crystal structure of the activated form, DMOF′. A bulk powdered sample of activated DMOF′ was subjected to CO2 sorption analysis at 25 °C in the range 1 to 20 bar, yielding a type I isotherm (Fig. S13, ESI). At 20 bar, the pressure limit of the instrument, DMOF′ absorbs approximately 1.5 molecules of CO2 per formula unit. Note that the guest occupancy is only 0.15 CO2 molecules per host formula unit at 1 bar. Irradiation of the powdered sample using a high-power 365 nm photodiode resulted in a distinct and immediate colour change from clear to deep blue (iDMOF′). When we repeated the CO2 sorption experiment for iDMOF′ after irradiation we obtained an isotherm similar to that recorded for DMOF′ (see Fig. S13, ESI). The powder diffractograms of DMOF′ and iDMOF′ are remarkably similar, indicating that UV irradiation does not result in an appreciable (or lasting) change in the framework periodicity of the bulk material. Irradiation of single crystals of DMOF′ produces a dramatic colour change with no concomitant loss of crystal singularity (Fig. 3). A series of single-crystal diffraction (SCD) analyses were carried out under a range of different conditions with no bulk ring-closure being observed (i.e. DMOF′ and iDMOF′ appear to be the same structure). Detailed examination of iDMOF′ crystals (Fig. S9, ESI) shows that conversion of the material might occur near the surface of a crystal, but in insufficient yield to influence the overall crystal structure determination.


image file: c7cc06074b-f3.tif
Fig. 3 Photomicrographs of a single crystal of DMOF′ before (left) and iDMOF′ after (right) irradiation with UV light.

We note that the apparent lack of bulk conversion may be due to experimental parameters such as irradiation wavelength, intensity and exposure time, as well as the particle size distribution of the material. Alternatively, the apparent failure of DMOF′ to undergo bulk photocyclisation could be a consequence of non-optimal structural features such as poor framework flexibility, a high degree of interpenetration, and the distance between the two photoactive carbon atoms. Although the reactive C⋯C distances in DMOF, DMOF′ and iDMOF′ are within the 4 Å limit19–23 (i.e. 3.652(7), 3.675(7) and 3.668(9) Å, respectively), it is possible that the overall rigidity of the host framework does not permit the conformational flexibility required by L1 for bulk ring closure to occur.

Following our unsuccessful transformation of DMOF′, we systematically altered each of the two bridging ligands in our search for a MOF with more suitable structural characteristics for bulk photoconversion (Fig. 2). In particular, we investigated the effects of substituting L2 for L1, and of utilising the more flexible oba in place of bpdc. Having established that a colour change alone is not a reliable indication of bulk photocyclisation, we focused primarily on using the powerful structural technique of SCD analysis to unequivocally verify bulk conversion. Therefore, it was imperative to find a system that undergoes all of the necessary conversions as single-crystal to single-crystal (SC–SC) transformations; of the four different materials explored, only one yielded the desired result.

SCD analysis of [Zn2(L2)bpdc]·2DMF (DMOF1) reveals that it is structurally similar and isometric to DMOF, with the exception that L2 is disordered equally over two similar positions. The reactive C⋯C distances in the two disordered components are 3.850(2) Å and 3.480(2) Å, and the solvent-accessible space accounts for ca. 32% of the total volume. Several crystals were activated by heating at 200 °C under reduced pressure for 24 hours and did not exhibit any discernible deterioration in single-crystal (SC) quality. The reactive C⋯C distance in the apohost structure DMOF1′ increases slightly to 3.725(1) Å and the guest-accessible volume is approximately the same as that of the solvate. Upon UV irradiation DMOF1 rapidly changes from colourless to blue, reverting back to colourless upon green laser irradiation (532 nm, ∼800 mW). SCD analysis of the UV-irradiated form reveals that bulk ring closure does not occur under the conditions investigated; the reactive C⋯C distance of iDMOF1′ measures 3.762(8) Å, and DMOF1′ and iDMOF1′ appear to be the same structure.

[Zn2(L1)(oba)2]·3DMF (DMOF2) consists of a three-dimensional, twofold-interpenetrated metal–organic framework. The metal cluster node comprises two symmetry-independent Zn ions, each in a distorted square pyramidal coordination environment. The structure contains an estimated 27% guest-accessible volume, which is occupied by three crystallographically unique DMF solvent molecules.

Activation of the crystals by heating at 150 °C under dynamic vacuum for 12 hours resulted in loss of crystal singularity, as did activation by immersion8 in supercritical CO2 (sc-CO2). Although UV irradiation caused the activated polycrystalline material to change from colourless to blue, loss of crystal singularity prevented us from studying this material further with a view to obtaining reliable structural data. We note that the reactive C⋯C distance in DMOF2 is 3.883(8) Å.

[Zn2(L2)(oba)2]·4DMF (DMOF3) shares several structural features with DMOF2, viz space group, secondary building unit, degree of interpenetration and framework connectivity. Moreover, DMOF3 contains an estimated 27% guest-accessible volume (Fig. 4a and Table S5, ESI), which is occupied by disordered DMF solvent molecules. The reactive C⋯C distance of L2 is 3.735(7) Å. Activation of DMOF3 using a conventional approach (heating under dynamic vacuum) once again resulted in a polycrystalline material. Although crystals of DMOF3 developed cracks when immersed in sc-CO2 for 24 hours, it was possible to isolate a fragment suitable for SCD analysis (DMOF3′). Desolvation causes the crystal to shrink by 14%, as inferred from the difference in the unit cell volumes of DMOF3 and DMOF3′. The guest-accessible volume of the compressed framework DMOF3′ comprises 9 to 15% (see Table S5, ESI) of the total crystal volume in the form of isolated pockets (Fig. 4b). The largest pocket contains a small amount of residual electron density, presumably due to the presence of diffuse entrapped CO2 molecules (we believe that DMOF3′ is a transiently porous38 phase). Notably, contraction of the framework is accompanied by a significant reduction in the reactive C⋯C distance to 3.513(1) Å.


image file: c7cc06074b-f4.tif
Fig. 4 Projections along the crystallographic a axis showing the guest-accessible space39 in (a) DMOF3, (b) DMOF3′ and (c) iDMOF3′. Solvent molecules and host hydrogen atoms have been omitted for clarity.

Exposure of the crystal fragment to 365 nm UV irradiation for 1 h produced a colour change from orange (DMOF3′) to dark blue (iDMOF3′) (Fig. 5a) and a further visual reduction in SC quality. The SCD structure of iDMOF3′ confirms unequivocally that irradiation of DMOF3′ results in bulk ring closure; the reactive C⋯C distance in iDMOF3′ is 1.486(2) Å, which is typical for a C–C single bond (Fig. 5a).


image file: c7cc06074b-f5.tif
Fig. 5 (a) The molecular structures of L2 and photomicrographs of a single crystal before (left) and after (right) UV irradiation. (b) A structural overlay of the asymmetric units of DMOF3′ (orange) and iDMOF3′ (blue).

In order to understand how photoinduced ring closure in L2 affects the overall framework structure of DMOF3, it is essential to compare the crystal structures of DMOF3′ and iDMOF3′ in detail. Interestingly, cyclisation of L2 has little effect on its shape, molecular volume and pyridyl N⋯N separation distance (see Fig. 5b and Fig. S28 and Video 2, ESI). Indeed, the most noticeable change experienced by L2 upon cyclisation is evident in the relative orientations of the two N⋯C(4) vectors passing through the two pyridyl moieties. For the ligated pyridyl groups these vectors are usually almost collinear with the N–metal coordination bonds. Therefore a significant change in the relative orientations of these vectors in a bis(pyridyl) ditopic ligand would necessarily influence the relative positions of coordinated metal ions in a grid-like coordination framework, even if the intra-ligand N⋯N′ distance remains relatively constant. In DMOF3′ steric repulsion between the ethyl groups attached to the reactive carbon atoms of L2 offsets the two N⋯C(4) vectors such that their closest points of contact are approximately 3 Å apart. After cyclisation, the N⋯C(4) vectors of L2 in iDMOF3′ are almost in the same plane. This conformational change in L2 (see Fig. S28 and Video 3, ESI) requires a concomitant change in the geometry of the 3D grid defined by the metal cluster nodes. The strain associated with this change is alleviated by conformational changes in the flexible oba bridging ligands (a more detailed account of the structural changes is given in the ESI).

We note that the changes due to photocyclisation affect the pore volume of the system dramatically (Fig. 4). Before UV irradiation each unit cell contains at least 262 Å3 of guest-accessible space (see Table S5, ESI). Photoconversion results in deformation of the framework, which in turn reduces the unit cell volume by approximately 253 Å3, leaving behind negligible guest-accessible pockets of 20 Å3 each.

To verify porosity switching of this material a bulk powdered sample of DMOF3′ was subjected to CO2 sorption analysis at 25 °C in the range 1–20 bar, yielding a type I isotherm (Fig. S34, ESI). At 20 bar, DMOF3′ absorbs approximately 4 molecules of CO2 per formula unit. Ex situ irradiation of the powdered sample, using a high power 365 nm photodiode resulted in a distinct and immediate orange to blue colour change (iDMOF3′). When the CO2 sorption experiment was repeated for iDMOF3′ after irradiation for 3 hours, an isotherm was obtained wherein CO2 uptake had decreased by half, with the material absorbing 2 molecules of CO2 per formula unit. Since SCD established that UV photoirradiation of DMOF3′ results in complete loss of porosity, it is postulated that the bulk phase material had only undergone partial ring-closure during ex situ irradiation.

Of the materials studied, only DMOF3 underwent activation and bulk cyclisation as SC–SC transformations, allowing unambiguous elucidation of all the relevant structures. The availability of structural data for the material, both before and after photomodification, serves as a reliable verification that the process occurs as a bulk phenomenon and provides insight into the transformation. We believe that the isostructural pair of materials DMOF and DMOF1 do not easily undergo bulk conversion. Since DMOF2 did not survive solvent removal as a SC–SC process, we turned our attention to DMOF3.

From our results, we infer that the secondary ligands need to have sufficient flexibility to accommodate the change in coordination directionality of the photoactive bridge. We further believe that transformation of the framework geometry of DMOF3′ is facilitated by the lower degree of interpenetration (i.e. lower steric hindrance to changes in the shape of the framework). Although a practical application of photoresponsive MOFs (i.e. reversible pore modulation upon irradiation) might involve the use of polycrystalline particles, it is important to show that high-yield bulk conversion of such materials is feasible. In this regard, we have shown that even relatively large particles such as single crystals are able to undergo bulk conversion. This is consistent with some previous studies that have employed SC–SC transformations to obtain unequivocal structural evidence for photoresponsive phenomena. In particular, we note the reports of Kitagawa et al. and Baroncini et al. which have both presented structural evidence for SC–SC photo-switching of porosity.37,40 The work presented here serves as crystallographic evidence that MOF materials such as DMOF3, with photochromic components, can also undergo light modulated structural switching.

We thank the National Research Foundation (NRF) of South Africa for financial support.

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. S. Horike, S. Shimomura and S. Kitagawa, Nat. Chem., 2009, 1, 695 CrossRef CAS PubMed.
  2. J. R. Li, J. Sculley and H. C. Zhou, Chem. Rev., 2012, 112, 869 CrossRef CAS PubMed.
  3. P. Horcajada, C. Serre, M. Vallet-Regí, M. Sebban, F. Taulelle and G. Frey, Angew. Chem., Int. Ed., 2006, 45, 5974 CrossRef CAS PubMed.
  4. Z. Hu, B. J. Deibert and J. Li, Chem. Soc. Rev., 2014, 43, 5815 RSC.
  5. J. A. Hurd, R. Vaidhyanathan, V. Thangadurai, C. I. Ratcliffe, I. L. Moudrakovski and G. K. H. Shimizu, Nat. Chem., 2009, 1, 705 CrossRef CAS PubMed.
  6. J. Seo, R. Matsuda, H. Sakamoto, C. Bonneau and S. Kitagawa, J. Am. Chem. Soc., 2009, 131, 12792 CrossRef CAS PubMed.
  7. I. Beurroies, M. Boulhout, P. L. Llewellyn, B. Kuchta, G. Ferey, C. Serre and R. Denoyel, Angew. Chem., Int. Ed., 2010, 49, 7526 CrossRef CAS PubMed.
  8. A. P. Nelson, O. K. Farha, K. L. Mulfort and J. T. Hupp, J. Am. Chem. Soc., 2009, 131, 458 CrossRef CAS PubMed.
  9. V. Bon, N. Kavoosi, I. Senkovska and S. Kaskel, ACS Appl. Mater. Interfaces, 2015, 7, 22292 CAS.
  10. U. Eiden and E. U. Schliinder, Chem. Eng. Process., 1992, 31, 63 CrossRef CAS.
  11. M. F. De Lange, K. J. F. M. Verouden, T. J. H. Vlugt, J. Gascon and F. Kapteijn, Chem. Rev., 2015, 115, 12205 CrossRef CAS PubMed.
  12. M. Alhamami, H. Doan and C.-H. Cheng, Materials, 2014, 7, 3198 CrossRef PubMed.
  13. Photochromism, ed. G. H. Brown, Wiley-Interscience, 1971 Search PubMed.
  14. Photochromism, Molecules and Systems, ed. H. Dürr and H. Bouas-Laurent, Elsevier, 1990 Search PubMed.
  15. M. Kamenjicki, I. K. Lednev and S. A. Asher, Adv. Funct. Mater., 2005, 15, 1401 CrossRef.
  16. E. Merino, Chem. Soc. Rev., 2011, 40, 3835 RSC.
  17. O. S. Bushuyev, A. Tomberg, T. Friščić and C. J. Barrett, J. Am. Chem. Soc., 2013, 135, 12556 CrossRef CAS PubMed.
  18. A. Santiago and R. S. Becker, J. Am. Chem. Soc., 1968, 90, 3654 CrossRef CAS.
  19. Y. Yokoyama, Chem. Soc. Rev., 2000, 100, 1717 CrossRef CAS.
  20. K. Horie, K. Hirao, N. Kenmochi and I. Mita, Makromol. Chem., Rapid Commun., 1988, 9, 267 CrossRef CAS.
  21. H. Tian and S. Yang, Chem. Soc. Rev., 2004, 33, 85 RSC.
  22. V. Ramamurthy and K. Venkatesan, Chem. Rev., 1987, 87, 433 CrossRef CAS.
  23. S. Kobatake, K. Uchida, E. Tsuchida and M. Irie, Chem. Commun., 2002, 2804 RSC.
  24. M. Frigoli, C. Welch and G. H. Mehl, J. Am. Chem. Soc., 2004, 126, 15382 CrossRef CAS PubMed.
  25. W. Li, Y. Cai, X. Li, H. Ågren, H. Tiana and W.-H. Zhu, J. Mater. Chem., 2015, 3, 8665 CAS.
  26. S. Kobatake, S. Takami, H. Muto, T. Ishikawa and M. Irie, Nature, 2007, 446, 778 CrossRef CAS PubMed.
  27. M. Munakata, L. P. Wu, T. Kuroda-Sowa, M. Maekawa, Y. Suenaga and K. Furuichi, J. Am. Chem. Soc., 1996, 118, 3305 CrossRef CAS.
  28. D. E. Williams, J. A. Rietman, J. M. Maier, R. Tan, A. B. Greytak, M. D. Smith, J. A. Krause and N. B. Shustova, J. Am. Chem. Soc., 2014, 136, 11886 CrossRef CAS PubMed.
  29. F. Luo, C. B. Fan, M. B. Luo, X. L. Wu, Y. Zhu, S. Z. Pu, W.-Y. Xu and G. C. Guo, Angew. Chem., Int. Ed., 2014, 53, 9298 CrossRef CAS PubMed.
  30. S. Castellanos, F. Kapteijn and J. Gascon, CrystEngComm, 2016, 18, 4006 RSC.
  31. R. Lyndon, K. Konstas, B. P. Ladewig, P. D. Southon, C. J. Kepert and M. R. Hill, Angew. Chem., Int. Ed., 2013, 52, 3695 CrossRef CAS PubMed.
  32. A. Schaate, S. Dühnen, G. Platz, S. Lilienthal, A. M. Schneider and P. Behrens, Eur. J. Inorg. Chem., 2012, 790 CrossRef CAS.
  33. Y.-K. Li, J.-J. Zhang, Z.-J. Bian, Y.-X. Fu, F. Liu, C.-H. Wanga, X. Mab, J. Hua and H.-L. Liu, Chin. Chem. Lett., 2016, 27, 518 CrossRef CAS.
  34. M. Han, Y. Luo, B. Damaschke, L. Gýmez, X. Ribas, A. Jose, P. Peretzki, M. Seibt and G. H. Clever, Angew. Chem., Int. Ed., 2016, 55, 445 CrossRef CAS PubMed.
  35. C. B. Fan, Z. Q. Liu, L. L. Gong, A. M. Zheng, L. Zhang, C. S. Yan, H. Q. Wu, X. F. Feng and F. Luo, Chem. Commun., 2017, 53, 763 RSC.
  36. B. J. Burnett and W. Choe, CrystEngComm, 2012, 14, 6129 RSC.
  37. Y. Zheng, H. Sato, P. Wu, H. J. Jeon, R. Matsuda and S. Kitagawa, Nat. Commun. DOI:10.1038/s41467-017-00122-5.
  38. L. J. Barbour, Chem. Commun., 2006, 1163 RSC.
  39. J. L. Atwood, L. J. Barbour and A. Jerga, Science, 2002, 296, 2367 CrossRef CAS PubMed.
  40. M. Baroncini, S. D’Agostino, G. Bergamini, P. Ceroni, A. Comotti, P. Sozzani, I. Bassanetti, F. Grepioni, T. M. Hernandez, S. Silvi, M. Venturi and A. Credi, Nat. Chem., 2015, 7, 634 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available: Synthesis, experimental methods. CCDC 1566319–1566328. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c7cc06074b

This journal is © The Royal Society of Chemistry 2017