Breathing Metal–Organic Polyhedra Controlled by Light for Carbon Dioxide Capture and Liberation

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Jun 2021Breathing Metal–Organic Polyhedra Controlled by Light for Carbon Dioxide Capture and Liberation Yao Jiang†, Peng Tan†, Shi-Chao Qi, Chen Gu, Song-Song Peng, Fan Wu, Xiao-Qin Liu and Lin-Bing Sun Yao Jiang† State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 , Peng Tan† State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 , Shi-Chao Qi State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 , Chen Gu State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 , Song-Song Peng State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 , Fan Wu State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 , Xiao-Qin Liu State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 and Lin-Bing Sun *Corresponding author: E-mail Address: [email protected] State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 211816 https://doi.org/10.31635/ccschem.020.202000314 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail Metal–organic polyhedra (MOPs) have emerged as versatile platforms for artificial models of biological systems due to their discrete structure and modular nature. However, the design and fabrication of MOPs with special functionality for mimicking biological processes are challenging. Inspired by the breathing mechanism of lungs, we developed a new type of MOP (a breathing MOP, denoted as NUT-101) by directly using azobenzene units as the pillars of the polyhedra to coordinate with Zr-based metal clusters. In addition to considerable thermal and chemical stability, the obtained MOP exhibits photocontrollable breathing behavior. Upon irradiation with visible or UV light, the configuration of azobenzene units transforms, leading to reversible expansion or contraction of the cages and, correspondingly, capture or liberation of CO2 molecules. Such a breathing behavior of NUT-101 is further confirmed by density functional theory (DFT) calculation. This system might establish an avenue for the construction of new materials with particular functionality that mimic biological processes. Download figure Download PowerPoint Introduction Metal–organic polyhedra (MOPs), or supramolecular coordination cages, are an emerging class of discrete nanostructure constructed by organic linkers and metal ions/clusters.1–5 By virtue of designable structure, tunable porosity, and diverse chemical/physical properties,6–8 MOPs show potential in various applications like separation,9–13 catalysis,14–18 drug delivery,19–21 and sensing.22–24 More interestingly, unlike metal–organic frameworks (MOFs) with infinite networks, MOPs are discrete molecules with particular geometries and sizes, which can be adopted as artificial models to simulate certain biological architectures and processes.25 Breathing is an essential biological process for human life.26 Lungs are the primary organs responsible for breathing, which undergo reversible expansion and contraction during inhalation and exhalation (Figure 1a).27,28 Inspired by lungs, researchers have engaged in developing artificial systems to mimic the breathing behavior,29–33 and the breathing mechanism is based on the configuration change of frameworks upon external stimuli.34–40 A representative example is MIL-53, whose framework is responsive to water; the pore dimension experiences reversible increase/decrease upon hydration/dehydration, respectively, leading to an unusual adsorption characteristic.41 Among various stimuli, light receives much attention because it is a remote stimulus with rapid responsiveness, high precision, and few undesired byproducts.42–45 Currently, only a couple of breathing MOFs that respond to light have been successfully achieved,46–48 and a breathing, photoresponsive MOP has never been reported. Due to the relevance of MOPs to biological self-assembly, the development of breathing MOPs is extremely desirable yet challenging. Figure 1 | Scheme depicting the breathing behavior of lungs and MOPs. (a) The inhalation and exhalation process along with the corresponding expansion and contraction of human lungs. (b) Guest molecule capture and liberation during expansion and contraction of breathing MOPs upon photoirradiation. Download figure Download PowerPoint Here, we report for the first time a breathing MOP, NUT-101 (NUT stands for Nanjing Tech University) by directly utilizing azobenzene units as the pillars of polyhedra to coordinate with Zr-based metal clusters. Due to the strong Zr−O bonds, the MOP shows considerable thermal stability (up to 300 °C) and chemical stability (in aqueous solutions with a pH range from 1.0 to 11.0). In contrast to pendant responsive motifs, azobenzene unit pillars lead to expansion and contraction of cages during configuration transformation, thus we observe a breathing behavior response to light of NUT-101 (Figure 1b). Upon visible light irradiation, the cages of NUT-101 expand and the capture of CO2 molecules takes place; upon UV light irradiation, the cages contract, leading to the liberation of CO2. Moreover, such a capture–liberation process is fully reversible upon dynamic photoirradiation. The relevance of unusual adsorption behavior to the structural change of NUT-101 is addressed by density functional theory (DFT) calculation. Experimental Methods The sample of NUT-101 was synthesized by the solvothermal method. Bis(cyclopentadienyl)zirconium dichloride (0.1 mmol) and azobenzene-4,4′-dicarboxylic acid (0.03 mmol) were dissolved in a mixed solution of N,N-dimethylformamide (DMF; 4 mL) and tetrahydrofuran (THF; 2 mL) followed by the addition of deionized water (0.65 mL). The mixture was treated ultrasonically for 15 min and heated to 65 °C for 12 h. After cooling to room temperature, the resultant block crystals were filtered and washed with DMF and acetone. The obtained samples were characterized by proton nuclear magnetic resonance (1H NMR) spectroscopy, infrared (IR) spectroscopy, X-ray diffraction (XRD), atomic force microscopy (AFM), dynamic light scattering (DLS), X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis (TGA), transmission electron microscopy (TEM), scanning electron microscopy (SEM), and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). All adsorption isotherms were measured using an ASAP 2020 analyzer (Micromeritics, Norcross, GA). A xenon lamp equipped with a filter was used as the UV/Vis light source. All DFT calculations were performed by employing the functional of Perdew–Burke–Ernzerhof, which had been implemented in the Gaussian 16 package. Other experimental details, characterization methods, and calculation details are available in Supporting Information. Results and Discussion Synthesis and structural analysis NUT-101 was synthesized from Cp2ZrCl2 (Cp:η5-C5H5) and azobenzene-4,4′-dicarboxylic acid in a mixture of DMF, THF, and water. Single-crystal XRD (SC-XRD) indicated that NUT-101 (Figure 2b) crystallizes in the monoclinic system with C2/c space group ( Supporting Information Table S1). Structurally, the isolated NUT-101 has a discrete candy-like geometry with V2E3 (V: vertex, E: edge) topology where two trinuclear Zr clusters, Cp3Zr3(μ3-O)(μ2-OH)3, of secondary building units (SBUs) are the vertexes and three azobenzene-containing ligands are edges; Cl- balances the positive charges of NUT-101 ( Supporting Information Figures S5 and S6). The free spaces in the cages of fresh NUT-101 are occupied by structurally disordered guest molecules, including unreacted ligands, DMF, THF, and H2O. As calculated by PLATON, the total potential guest-accessible volume of NUT-101 is 4947.9 Å3, which is 43.9% of the total unit volume (11,268.0 Å3).49 Powder XRD (PXRD) patterns were collected to assess the phase purity of NUT-101. The PXRD pattern ( Supporting Information Figure S7) of as-synthesized NUT-101 is in line with the simulated pattern from SC-XRD data, suggesting bulk purity of as-synthesized sample. Because NUT-101 lacks three-dimensional connectivity, structural rearrangement takes place after activation.12 Thus, the PXRD pattern of the activated sample does not match that of the as-synthesized sample.50 When the activated sample was resoaked in fresh DMF, the PXRD pattern was restored and comparable with the initial one ( Supporting Information Figure S8). NUT-101 can be characterized by 1H NMR spectroscopy because of its good solubility in methanol. In the 1H NMR spectrum ( Supporting Information Figure S2), the chemical shift at 6.67 ppm is the characteristic signal for the Cp protons,51 while the chemical shifts at 7.55–7.86 ppm are ascribed to the proton signals for the azobenzene-containing ligand, which verify the formation of discrete cages of NUT-101 with a highly symmetric scaffold. Mass spectrometry (MS) is considered an effective approach to determine discrete molecules as they can be well preserved during the ionization process.52 MALDI-TOF-MS was employed to give further evidence for the formation of NUT-101. As shown in Figure 2b, a dominant peak is clearly observed at m/z of 1876.7745, which is consistent with NUT-101, [M-2Cl-+H+]3+ (where M is the intact NUT-101, with the addition of one H+ and with the loss of two Cl-), indicating the presence of intact molecular cages in solution (Figures 2c and 2d). The 1H NMR and MALDI-TOF-MS results thus confirm the successful fabrication of NUT-101 and the preservation of its structure in solution. Figure 2 | Structural illustration and MS spectra of NUT-101. (a) Scheme for the synthesis of NUT-101. Zr clusters [Cp3Zr3O(OH)3(RCO2)3] and organic linkers (azobenzene-4,4′-dicarboxylic acid) form NUT-101. Zr, cyan; C, black; N, blue; and O, red. H atoms, counteranions, and solvent molecules have been omitted for clarity. The large yellow sphere in a NUT-101 molecule represents the free space inside the molecular cage. (b) MALDI-TOF-MS spectrum of NUT-101. The (c) observed and (d) calculated isotope patterns of [M-2Cl-+H+]3+. M represents the intact assembly of NUT-101. Download figure Download PowerPoint The optical microscope image reveals that NUT-101 crystallizes in quadrilateral crystals (Figure 3a), and the quadrilateral morphology can be confirmed by SEM (Figures 3b and 3c). In the high-resolution TEM (HRTEM) image (Figure 3d), discrete NUT-101 cages with an average size of approximately 3 nm × 2 nm are directly observed. AFM analysis shows that the molecular size of MOPs is approximately 3 nm ( Supporting Information Figure S9). The DLS technique was also employed to analyze the dimension of NUT-101. As provided in Supporting Information Figure S10, the average size of NUT-101 is approximately 3.5 nm, which agrees with the value determined by HRTEM and AFM. To obtain the elemental composition of NUT-101, energy-dispersive spectroscopy (EDS) was employed. The elements of C, N, O, Cl, and Zr are detected in NUT-101 ( Supporting Information Figure S11) and are homogeneously distributed as displayed in elemental mapping images (Figure 3e). In addition, XPS was conducted to further analyze the composition of NUT-101 ( Supporting Information Figures S12 and S13). The results indicate that NUT-101 consists of C, N, O, Cl, and Zr, which agrees with the EDS data. Figure 3 | Optical and electron microscopy analyses of NUT-101. (a) Optical microscope, (b and c) SEM, (d) HRTEM, and (e) elemental-mapping images of NUT-101. Download figure Download PowerPoint Stability examination Thermal and chemical stability of NUT-101 were examined in detail, considering their importance in practical applications such as adsorption.53 To investigate the thermal stability, in situ variable-temperature PXRD patterns in the range of room temperature to 330 °C were recorded (Figure 4a). The PXRD pattern changes at 90 °C, indicating the structural transformation of NUT-101. Such an XRD pattern is identical to activated NUT-101 as discussed before, suggesting the removal of free solvents at elevated temperatures.54 Upon further increasing the temperature to 300 °C, the PXRD pattern remains constant. This means that the crystalline structure can be retained up to 300 °C. When the temperature reaches 330 °C, the characteristic PXRD peaks disappear, reflecting the collapse of the MOP structure. The thermal stability of NUT-101 is also verified by the TGA displayed in Supporting Information Figure S14. Figure 4 | Stability of NUT-101. (a) The in situ variable-temperature PXRD patterns of NUT-101. (b) MALDI-TOF-MS spectra of NUT-101 with various pH values. Download figure Download PowerPoint To evaluate the chemical stability, NUT-101 was soaked in aqueous solutions with a pH value range from 1.0 to 13.0 overnight, and after soaking, the sample was analyzed by MALDI-TOF-MS. As shown in Figure 4b, there is no obvious change in MS spectra for NUT-101 after treatment with solutions in the pH range between 1.0 and 7.0, demonstrating the stability under acidic conditions. When the pH value of the solution is 13.0, however, the m/z of 1876.7745 ascribed to the cage structure becomes invisible, suggesting that NUT-101 decomposes under strongly basic conditions. Therefore, NUT-101 exhibits good chemical stability in aqueous solutions with a wide pH range (1.0–11.0). The high chemical stability of NUT-101 is attributed to the strong Zr−O bonds, which is similar to Zr-based MOFs like UiO-66.55 The thermal and chemical stability of NUT-101 allows facile application to gas adsorption. Photoresponsive behavior It is known that azobenzene and its derivatives are able to undergo reversible isomerization upon light irradiation, and this process can be monitored by UV/Vis absorption spectroscopy.56 In a typical UV/Vis absorption spectrum of azobenzene, the absorption band at low wavelength can be ascribed to the π–π* transition of the trans-azobenzene isomer, while the absorption band at high wavelength can be ascribed to the n–π* transition of the cis-azobenzene isomer.57 Owing to the presence of azobenzene units in NUT-101, the photoswitchable behavior of NUT-101 was monitored by UV/Vis absorption spectroscopy. Apparently, the absorption spectrum of NUT-101 in Figure 5a presents two signature absorbance bands at approximately 375 and 455 nm, corresponding to the trans-azobenzene and cis-azobenzene isomers, respectively. Upon UV light irradiation, the absorbance band at 375 nm is weakened, while the absorbance band at 455 nm simultaneously becomes intense. This observation is consistent with the trans to cis photoisomerization of azobenzene units. As shown in Figure 5a, the cis to trans isomerization can be reversed upon visible light irradiation. Moreover, the trans/cis isomerization of azobenzene units within NUT-101 is reversible upon alternating irradiation between UV and visible light (Figure 5b). Fourier transform IR (FTIR) spectroscopy was employed to further study the photoresponsive behavior of NUT-101. The IR band at the region of 550–600 cm−1 is ascribed to the C–N bending vibration of azobenzene moieties ( Supporting Information Figure S15).46 Upon UV light irradiation, the band intensity decreases due to the transformation of trans azobenzene to cis azobenzene. The IR band is recovered upon visible light irradiation, indicating the recovery of trans isomers. Noteworthily, the PXRD results demonstrate that the crystalline structure of NUT-101 is not affected upon photoirradiation ( Supporting Information Figure S16), indicating that the configuration transformation of azobenzene units in NUT-101 is local. A similar result has also been reported in which no change in the XRD patterns occurred in a photoresponsive MOF with azobenzene-based dicarboxylic acid as the pillar ligand because the photoinduced structural change is local.48 Analogously, it is hard to detect any obvious change for NUT-101 by 1H NMR upon photoirradiation due to the local and instantaneous structural change of the azobenzene group in NUT-101 upon photoirradiation ( Supporting Information Figure S17). Also, the color of NUT-101 does not change upon photoirradiation due to its intrinsic dark color ( Supporting Information Figure S16). In short, the results demonstrate that azobenzene units as the pillars of cages can undergo reversible isomerization upon light irradiation, leading to expansion and contraction of cages, and subsequent tunable adsorption capacity of gas molecules as shown below. Figure 5 | Photoswitchable behavior of NUT-101. (a) Alteration in the UV/Vis spectra of NUT-101 upon UV/Vis light irradiation. (b) Reversible changes in absorbance as a function of cycles upon alternating UV/Vis light irradiation. Download figure Download PowerPoint Adsorption performance To investigate the photoresponsive adsorption performance of NUT-101, the adsorption isotherms of CO2, CH4, and N2 were collected upon UV/Vis light irradiation, respectively. Generally, the open metal cation sites in MOFs or MOPs serve as charge-dense binding sites for CO2, which is adsorbed more strongly at these sites owing to its greater quadrupole moment and polarizability compared with CH4 and N2. Also, it is reported that the μ2-OH group in MOPs or MOFs is a strong binding site to interact with CO2.58 For NUT-101, as mentioned above, it is confirmed that the μ2-OH group was included in the metal cluster of Cp3Zr3(μ3-O)(μ2-OH)3, which is the adsorption active site for CO2. Unfortunately, there are no suitable active sites to bind to CH4 and N2 in NUT-101. Thus, irrespective of the light used for irradiation, NUT-101 shows higher adsorption capacity of CO2 than CH4 and N2 (Figure 6a), indicating the selective adsorption of CO2 in the cages. Moreover, the change in adsorption capacity upon light irradiation is quite different for different gases. Upon visible light irradiation, the cages of NUT-101 expand and can capture 1.20 mol CO2 per mol MOP. UV light irradiation leads to the contraction of cages, and the adsorption capacity decreases to 0.85 mol CO2/mol MOP. This corresponds to 29.2% difference in CO2 adsorption capacity (Figure 6b). Because there were no suitable active sites for CH4 and N2 in NUT-101, a lower adsorption capacity of CH4 and N2 is achieved regardless of photoirradiation. In comparison with CO2, accordingly, the differences in the adsorption capacity of CH4 (8.6%) and N2 (4.8%) upon photoirradiation are relatively lower as well. Figure 6 | Experimental and DFT calculations of CO2 capture and liberation by NUT-101. (a) Adsorption isotherms of CO2, CH4, and N2 on NUT-101 upon photoirradiation. (b) Change amount of CO2, CH4, and N2 on NUT-101 upon photoirradiation. (c) In situ dynamic capture and liberation of CO2 on NUT-101. (d) Adsorption cycles of NUT-101 for CO2 upon photoirradiation. DFT-optimized possible structures of NUT-101 upon (e) visible and (f) UV light irradiation. As per calculations, the CO2 molecule can be captured in the cavity of NUT-101 upon visible light irradiation (e) but not captured upon UV light irradiation (f). Download figure Download PowerPoint Furthermore, in situ capture and liberation of CO2 upon dynamic photoirradiation of NUT-101 were recorded (Figure 6c). Initially, CO2 adsorption on NUT-101 was performed by visible light irradiation. Interestingly, once the visible light is switched to UV light, the adsorption capacity of CO2 decreases. At this stage, UV light irradiation induces structural change, leading to the contraction of cages and subsequently the liberation of CO2. After the switch of UV to visible light, the adsorption capacity of CO2 can be recovered to the initial adsorption trend. The maximum liberation amount of the captured CO2 triggered by photoirradiation is calculated to be 30%. Obviously, the controllable capture and liberation of CO2 over NUT-101 can be achieved upon dynamic photoirradiation. It is reported that the breathing adsorption behavior of MIL-53 can be directly monitored by in situ electron microscopy due to its significant structural changes during hydration/dehydration at the lattice level.59 In contrast to water adsorption, however, the gas adsorption in breathing materials cannot be monitored by in situ electron microscopy technology because of technical problems at present. The studies of gas adsorption in breathing materials remain dominated by molecular simulation at present.60 Hence, DFT calculations were performed to support the unusual adsorption behavior (Figures 6e and 6f). Upon visible light irradiation, the binding energy of NUT-101 adsorbing CO2 is negative, suggesting that the CO2 molecule can be captured within the cage ( Supporting Information Table S4). However, the binding energy of NUT-101 adsorbing CO2 is positive upon UV light irradiation, indicating that NUT-101 is not inclined to capture CO2 molecules in this state. Thus, accompanied with the structural changes of NUT-101 upon UV light irradiation, the captured CO2 molecules can be liberated from the cages. According to the results from simulation and calculation, the decreased CO2 adsorption capacity in NUT-101 upon UV irradiation can be further attributed to both the changed intrinsic porosity and extrinsic porosity of NUT-101. The intrinsic porosity of NUT-101 was decreased upon UV light irradiation because of the contracted framework, which results in a lower adsorption capacity for CO2. In contrast, the extrinsic porosity was changed upon UV light irradiation because of the changed arrangement of NUT-101, which further results in a changed adsorption capacity for CO2. Moreover, the DFT calculations for CH4 and N2 in NUT-101 upon photoirradiation were performed as well ( Supporting Information Figure S18 and Table S4). The binding energy of NUT-101 adsorbing CH4 and N2 was much higher than that of the binding energy of NUT-101 adsorbing CO2 regardless of photoirradiation, which further demonstrates that the NUT-101 has no adsorptivity for CH4 and N2. Overall, the results of the DFT calculation accord with that of the adsorption experiments, proving the breathing behavior of NUT-101 responds to light. The ideal adsorption solution theory (IAST) model is adopted to estimate the selectivity of CO2/CH4 and CO2/N2 upon photoirradiation.61 As shown in Supporting Information Figure S19, the CO2/CH4 selectivity (273 K, 1 bar) is 7.3 upon visible light irradiation, which decreases to 5.2 upon UV light irradiation. Similarly, the CO2/N2 selectivity (273 K, 1 bar) is 49.5 upon visible light irradiation, which decreases to 23.9 upon UV light irradiation. The results indicate a higher priority adsorption of NUT-101 for CO2 than CH4 and N2. For practical applications, recyclability is a crucial process for adsorbents. Figure 6d shows three cycles of NUT-101 for CO2 adsorption upon reversible photoirradiation. The adsorption capacity and change amount of NUT-101 for CO2 upon UV or visible light irradiation remains at the same level, suggesting the breathing-like process is reversible. Conclusion We have developed a breathing MOP (NUT-101), which features a discrete nanocavity, considerable thermal and chemical stability, and pillared photocontrollable motifs. Unlike pendant photoresponsive motifs, azobenzene units as the pillars lead to expansion and contraction of cages in the process of configuration transformation, and hence the obtained NUT-101 can breathe. It was experimentally demonstrated that the capture and liberation of CO2 in NUT-101 can be modulated by photoirradiation. Visible light irradiation causes the expansion of cages, which results in the selective capture of CO2. Inversely, the cages of NUT-101 contract once exposed to UV light, leading to the liberation of CO2. Moreover, the breathing behavior of NUT-101 is confirmed by the DFT calculation. The present investigation might provide an avenue for the design and development of more elaborate bionic materials with special properties and applications. Supporting Information Supporting Information is available. Conflict of Interest There are no conflicts to declare. Acknowledgments This study was supported by the National Science Fund for Excellent Young Scholars (no. 21722606), the National Natural Science Foundation of China (nos. 21676138, 21878149, 21808110, and 21576137), and the China Postdoctoral Science Foundation (no. 2018M632295). The authors are grateful to the High-Performance Computing Center of Nanjing Tech University for supporting the computational resources. References 1. Yaghi O. M.; Michael O. K.; Ockwig N. W.; K.; Synthesis and the of M.; K.; and of Liu of Liu a of on Liu Sun Photoresponsive by Polyhedra Liu Liu Sun of Polyhedra in with and Liu W.; Chen Polyhedra with and Wu K.; K.; M.; Chen from for of M.; M.; for of M.; in of a of from Sun M.; Sun with for for in A