The Impact of Structural Defects on Iodine Adsorption in UiO-66

: Radioactive I 2 (iodine) produced as a by-product of nuclear ﬁssion poses a risk to public health if released into the environment, and it is thus vital to develop materials that can capture I 2 vapour. Materials designed for the capture and storage of I 2 must have a high uptake capacity and be stable for long-term storage due the long half-life of 129 I. UiO-66 is a highly stable and readily tuneable metal-organic framework (MOF) into which defect sites can be introduced. Here, a defective form of UiO-66 (UiO-66-FA) was synthesised and the presence of missing cluster moieties conﬁrmed using confocal ﬂuorescence microscopy and gas sorption measurements. The uptake of I 2 vapour in UiO-66-FA was measured using thermal gravimetric analysis coupled mass spectrometry (TGA-MS) to be 2.25 g g − 1 , almost twice that (1.17 g g − 1 ) of the pristine UiO-66. This study will inspire the design of new efﬁcient I 2 stores based upon MOFs incorporating structural defects. acid in DMF and heating the solution to 120 ◦ C for 24 h. UiO-66-FA was synthesised using the same general procedure but using a mixture of DMF and formic acid as solvent. Powder X-ray diffraction (PXRD) analysis of as-synthesised UiO-66 and UiO-66-FA (Figure 1) conﬁrmed the phase purity of both samples [30,31]. The PXRD pattern of UiO-66-FA shows two additional Bragg peaks between 5 and 7 ◦ attributed to the presence of reo regions within the fcu structure caused by missing cluster defects [29,32]. These Bragg peaks are also broad, highlighting the disorder of these defect sites throughout the parent structure of UiO-66. with unsaturated zirconium clusters and increased access to the framework of UiO-66-FA results in a high I 2 uptake.


Introduction
Nuclear power is responsible for approximately 21% of the UK's energy production as of 2020 [1]. The products produced as a result of the fission of uranium pose a danger to the environment and public health, and so materials that can capture and store such fission products are of significant interest. 131 I and 129 I are volatile fission products with half-lives of 8 days and 1.57 × 10 7 years, respectively [2]. 131 I poses a serious risk to human health as it has been linked to the occurrence of thyroid cancer [3]. 129 I is less hazardous due to its low energy beta and gamma emissions, but poses a long-term environmental risk due to bioaccumulation [4]. It is vital also to prevent I 2 escaping into the environment as it can spread over a wide area due to its high solubility in water [5]. The capture of I 2 has been investigated using a wide variety of porous materials such as aerogels [6], zeolites [7], porous organic polymers [8] and covalent-organic framework [9] materials. The main drawbacks to using these materials are that they can be non-specific for I 2 , have low uptakes, or have amorphous structures that prevent determination of preferred binding sites, thus restricting an understanding of the mechanism of action of the material.
Metal-organic framework (MOF) materials are often highly porous and crystalline and are well-known for their tuneable structures, potential high chemical stability and high surface areas [10]. The tunability of MOFs has allowed them to be used in a wide range of applications including catalysis [11], molecular separations [12] and the capture of gases such as CO 2 [13] and SO 2 [14]. The storage of I 2 by MOFs has been reported previously with uptakes reaching as high as 7.35 g g −1 in the case of the ionic liquid-doped material PCN-333 [15]. Proof-of-concept studies have also shown that MOFs can be used for the long-term storage of I 2 using glass sintering [16] and pressure-induced amorphization [17]. Functionalisation of the MOF structure is a common route to increasing I 2 adsorption and relies on the introduction of electron donating [18,19] or reactive groups [20] onto the ligands. Doping MOFs with Ag [21] and Cu [22] ions can also improve the I 2 uptake of the host MOF. Detailed studies into the adsorption mechanism of I 2 have also been carried out and highlight the importance of the structure of the MOF [23,24] and of metal cluster nodes for the efficient capture of I 2 [25]. The introduction of structural defects has been shown to increase the catalytic activity of MOFs [26,27] and increase the uptake of CO 2 by UiO-66 [28]. However, the impact of such defects on the adsorption of I 2 has not been investigated previously and we were interested to determine whether this was an appropriate methodology for improved I 2 adsorption ( Figure S1, Supplementary Materials).
UiO-66 was chosen in this work due its high stability and the established synthesis and routes to the preparation of defective derivatives. Coordination modulation by undertaking the synthesis of UiO-66 in the presence of formic acid can introduce defects into UiO-66. The competitor ligand formate binds to metal clusters in place of the bridging terephthalate linker to produce missing linker defects. If there is sufficient missing linker, defects will occur within the overall stable geometry with certain metal clusters absent ( Figure S2) [29]. We have analysed the presence of missing cluster defects within a sample of defect UiO-66, designated UiO-66-FA, using Brunauer-Emmett-Teller (BET) surface area analysis, thermal gravimetric analysis and confocal fluorescence microscopy. The uptake of I 2 has been measured and confirmed using Raman spectroscopy and thermal gravimetric analysis coupled with mass spectrometry. The presence of structural defects results in a nearly 100% enhancement in the adsorption capacity of I 2 within UiO-66-FA compared to pristine UiO-66.

Results and Discussion
UiO-66 was synthesised by dissolving ZrCl 4 and terephthalic acid in DMF and heating the solution to 120 • C for 24 h. UiO-66-FA was synthesised using the same general procedure but using a mixture of DMF and formic acid as solvent. Powder X-ray diffraction (PXRD) analysis of as-synthesised UiO-66 and UiO-66-FA (Figure 1) confirmed the phase purity of both samples [30,31]. The PXRD pattern of UiO-66-FA shows two additional Bragg peaks between 5 and 7 • attributed to the presence of reo regions within the fcu structure caused by missing cluster defects [29,32]. These Bragg peaks are also broad, highlighting the disorder of these defect sites throughout the parent structure of UiO-66.
material PCN-333 [15]. Proof-of-concept studies have also shown that MOFs can be used for the long-term storage of I2 using glass sintering [16] and pressure-induced amorphization [17]. Functionalisation of the MOF structure is a common route to increasing I2 adsorption and relies on the introduction of electron donating [18,19] or reactive groups [20] onto the ligands. Doping MOFs with Ag [21] and Cu [22] ions can also improve the I2 uptake of the host MOF. Detailed studies into the adsorption mechanism of I2 have also been carried out and highlight the importance of the structure of the MOF [23,24] and of metal cluster nodes for the efficient capture of I2 [25]. The introduction of structural defects has been shown to increase the catalytic activity of MOFs [26,27] and increase the uptake of CO2 by UiO-66 [28]. However, the impact of such defects on the adsorption of I2 has not been investigated previously and we were interested to determine whether this was an appropriate methodology for improved I2 adsorption ( Figure S1, Supplementary Materials).
UiO-66 was chosen in this work due its high stability and the established synthesis and routes to the preparation of defective derivatives. Coordination modulation by undertaking the synthesis of UiO-66 in the presence of formic acid can introduce defects into UiO-66. The competitor ligand formate binds to metal clusters in place of the bridging terephthalate linker to produce missing linker defects. If there is sufficient missing linker, defects will occur within the overall stable geometry with certain metal clusters absent ( Figure S2) [29]. We have analysed the presence of missing cluster defects within a sample of defect UiO-66, designated UiO-66-FA, using Brunauer-Emmett-Teller (BET) surface area analysis, thermal gravimetric analysis and confocal fluorescence microscopy. The uptake of I2 has been measured and confirmed using Raman spectroscopy and thermal gravimetric analysis coupled with mass spectrometry. The presence of structural defects results in a nearly 100% enhancement in the adsorption capacity of I2 within UiO-66-FA compared to pristine UiO-66.

Results and Discussion
UiO-66 was synthesised by dissolving ZrCl4 and terephthalic acid in DMF and heating the solution to 120 °C for 24 h. UiO-66-FA was synthesised using the same general procedure but using a mixture of DMF and formic acid as solvent. Powder X-ray diffraction (PXRD) analysis of as-synthesised UiO-66 and UiO-66-FA ( Figure 1) confirmed the phase purity of both samples [30,31]. The PXRD pattern of UiO-66-FA shows two additional Bragg peaks between 5 and 7° attributed to the presence of reo regions within the fcu structure caused by missing cluster defects [29,32]. These Bragg peaks are also broad, highlighting the disorder of these defect sites throughout the parent structure of UiO-66.  Confocal fluorescence microscopy (CFM) can be used to visualise mesoporous defects within electro-synthesised MOFs [27]. In these systems, Lewis acid sites found in defect sites and boundaries catalyse the formation of a fluorescent oligomer from furfuryl alcohol, the monomer of which is not fluorescent ( Figure S3). On exposure of UiO-66 and UiO-66-FA to furfuryl alcohol it was noted that the UiO-66-FA sample had a darker colour than the sample exposed to UiO-66 ( Figure S4). This suggested that a more fluorescent oligomer was being produced by UiO-66-FA as the oligomer has a dark brown colour compared to the colourless furfuryl alcohol. This conclusion was supported by CFM which showed a uniform spread of high intensity fluorescence throughout the sample of UiO-66-FA exposed to furfuryl alcohol. The sample of UiO-66 exposed to furfuryl alcohol shows weaker fluorescence which is not evenly distributed across the sample (Figure 2). The relatively small amount of fluorescence and its location for the UiO-66 sample can be explained by the presence of Lewis acid sites at the edges of the crystals, with the high intensity fluorescence across the sample of UiO-66-FA confirming the presence of an increased number of defects within the structure of UiO-66-FA compared to UiO-66.
Confocal fluorescence microscopy (CFM) can be used to visualise mesoporous defects within electro-synthesised MOFs [27]. In these systems, Lewis acid sites found in defect sites and boundaries catalyse the formation of a fluorescent oligomer from furfuryl alcohol, the monomer of which is not fluorescent ( Figure S3). On exposure of UiO-66 and UiO-66-FA to furfuryl alcohol it was noted that the UiO-66-FA sample had a darker colour than the sample exposed to UiO-66 ( Figure S4). This suggested that a more fluorescent oligomer was being produced by UiO-66-FA as the oligomer has a dark brown colour compared to the colourless furfuryl alcohol. This conclusion was supported by CFM which showed a uniform spread of high intensity fluorescence throughout the sample of UiO-66-FA exposed to furfuryl alcohol. The sample of UiO-66 exposed to furfuryl alcohol shows weaker fluorescence which is not evenly distributed across the sample (Figure 2). The relatively small amount of fluorescence and its location for the UiO-66 sample can be explained by the presence of Lewis acid sites at the edges of the crystals, with the high intensity fluorescence across the sample of UiO-66-FA confirming the presence of an increased number of defects within the structure of UiO-66-FA compared to UiO-66.  UiO-66-FA shows a higher BET surface area than UiO-66, 1705 and 1170 m 2 g −1 , respectively, as determined from the N 2 adsorption isotherm, consistent with the removal of linkers and clusters to form defect sites in UiO-66-FA. The pore size distribution data ( Figure S5) confirm that larger pores are present in UiO-66-FA with significant peaks above 8 Å radius attributed to the large pores created due to missing cluster moieties. In contrast, UiO-66 only shows pores of less than 9 Å radius [30]. The calculated micropore volume also supports the formation of larger pores with 0.30 cm 3 g −1 in UiO-66 and 0.73 cm 3 g −1 in UiO-66-FA. PXRD, CFM and BET results confirm the presence of defects in UiO-66-FA. I 2 adsorption was carried out in a sealed flask by heating for 3 days the activated, desolvated solid MOF sample with solid I 2 , each inside an open glass vessel. The MOF sample was then removed from the vessel for further analysis. TGA analysis of the I 2loaded samples showed a drop in mass between 100 and 200 • C attributed to loss of I 2 as monitored by mass spectrometry (Figure 3). The I 2 uptake over three repeat cycles gave an average uptake for UiO-66 and UiO-66-FA of 1.17 and 2.25 g g −1 , respectively, consistent with the increased porosity due to defects in the structure of UiO-66-FA. The weight drop observed at around 500 • C is linked to decomposition of the ligand and the percentage drop for UiO-66-FA is less than that of UiO-66, reflecting fewer ligands present in the UiO-66-FA. The Raman spectrum of solid I 2 shows a peak at 180 cm −1 assigned to the υI-I stretching vibration. This peak is shifted in both I 2 -loaded UiO-66 samples (Figure 4), and the presence of this single peak rules out the presence of triiodide that would produce a peak between 110 and 140 cm −1 [33]. Other observable features in the complete Raman spectra ( Figure S6) include peaks at 1600 cm −1 , 1150 cm −1 and 850 cm −1 assigned to the C-C bonds in the aromatic ring of the ligand [34]. The overlapping peaks seen at 1400 cm −1 are due to the COO stretching vibration of the linker overlapping with another aromatic ring Raman peak. There is no change to these peaks upon I 2 adsorption in both UiO-66 and UiO-66-FA, confirming that the linker remains intact after I 2 adsorption. I 2 has strong interaction with unsaturated zirconium clusters and increased access to the framework of UiO-66-FA results in a high I 2 uptake. spectively, as determined from the N2 adsorption isotherm, consistent with the removal of linkers and clusters to form defect sites in UiO-66-FA. The pore size distribution data ( Figure S5) confirm that larger pores are present in UiO-66-FA with significant peaks above 8 Å radius attributed to the large pores created due to missing cluster moieties. In contrast, UiO-66 only shows pores of less than 9 Å radius [30]. The calculated micropore volume also supports the formation of larger pores with 0.30 cm 3 g −1 in UiO-66 and 0.73 cm 3 g −1 in UiO-66-FA. PXRD, CFM and BET results confirm the presence of defects in UiO-66-FA.
I2 adsorption was carried out in a sealed flask by heating for 3 days the activated, desolvated solid MOF sample with solid I2, each inside an open glass vessel. The MOF sample was then removed from the vessel for further analysis. TGA analysis of the I2loaded samples showed a drop in mass between 100 and 200 °C attributed to loss of I2 as monitored by mass spectrometry (Figure 3). The I2 uptake over three repeat cycles gave an average uptake for UiO-66 and UiO-66-FA of 1.17 and 2.25 g g −1 , respectively, consistent with the increased porosity due to defects in the structure of UiO-66-FA. The weight drop observed at around 500 °C is linked to decomposition of the ligand and the percentage drop for UiO-66-FA is less than that of UiO-66, reflecting fewer ligands present in the UiO-66-FA. The Raman spectrum of solid I2 shows a peak at 180 cm −1 assigned to the υI-I stretching vibration. This peak is shifted in both I2-loaded UiO-66 samples ( Figure  4), and the presence of this single peak rules out the presence of triiodide that would produce a peak between 110 and 140 cm −1 [33]. Other observable features in the complete Raman spectra ( Figure S6) include peaks at 1600 cm −1 , 1150 cm −1 and 850 cm −1 assigned to the C-C bonds in the aromatic ring of the ligand [34]. The overlapping peaks seen at 1400 cm −1 are due to the COO stretching vibration of the linker overlapping with another aromatic ring Raman peak. There is no change to these peaks upon I2 adsorption in both UiO-66 and UiO-66-FA, confirming that the linker remains intact after I2 adsorption. I2 has strong interaction with unsaturated zirconium clusters and increased access to the framework of UiO-66-FA results in a high I2 uptake.  The cycling of I 2 adsorption was carried out to show that even when clusters are removed from UiO-66 the structure remains stable on the adsorption and desorption of I 2 ( Figure 5). The decrease in I 2 uptake observed after each cycle in derivatives of UiO-66 has been observed previously [25,35]. However, this drop in uptake appears to be less for UiO-66-FA, which could be due to the openness of the structure reducing the impact of I 2 removal. Samples of MOF were monitored by PXRD on removal of I 2 from loaded material ( Figures S7 and S8). The PXRD patterns show little change in the structure or crystallinity throughout the cycling experiments, confirming that defects within UiO-66-FA do not decrease its stability on I 2 loading.  Figure S5.
The cycling of I2 adsorption was carried out to show that even when clusters are removed from UiO-66 the structure remains stable on the adsorption and desorption of I2 ( Figure 5). The decrease in I2 uptake observed after each cycle in derivatives of UiO-66 has been observed previously [25,35]. However, this drop in uptake appears to be less for UiO-66-FA, which could be due to the openness of the structure reducing the impact of I2 removal. Samples of MOF were monitored by PXRD on removal of I2 from loaded material ( Figures S7 and S8). The PXRD patterns show little change in the structure or crystallinity throughout the cycling experiments, confirming that defects within UiO-66-FA do not decrease its stability on I2 loading. In summary, the impact of structural defects on I2 uptake capacity has been studied in UiO-66 and UiO-66-FA. The presence of defects in UiO-66-FA was verified by BET surface area, TGA and CFM analysis. Defects caused by missing clusters within the structure   Figure S5.
The cycling of I2 adsorption was carried out to show that even when clusters are removed from UiO-66 the structure remains stable on the adsorption and desorption of I2 ( Figure 5). The decrease in I2 uptake observed after each cycle in derivatives of UiO-66 has been observed previously [25,35]. However, this drop in uptake appears to be less for UiO-66-FA, which could be due to the openness of the structure reducing the impact of I2 removal. Samples of MOF were monitored by PXRD on removal of I2 from loaded material ( Figures S7 and S8). The PXRD patterns show little change in the structure or crystallinity throughout the cycling experiments, confirming that defects within UiO-66-FA do not decrease its stability on I2 loading. In summary, the impact of structural defects on I2 uptake capacity has been studied in UiO-66 and UiO-66-FA. The presence of defects in UiO-66-FA was verified by BET surface area, TGA and CFM analysis. Defects caused by missing clusters within the structure In summary, the impact of structural defects on I 2 uptake capacity has been studied in UiO-66 and UiO-66-FA. The presence of defects in UiO-66-FA was verified by BET surface area, TGA and CFM analysis. Defects caused by missing clusters within the structure of UiO-66 results in an increase I 2 adsorption from 1.17 to 2.25 g g −1 for UiO-66 and UiO-66-FA, respectively, and the overall increased porosity of the latter also contributes to higher I 2 uptake. Cycling of I 2 loading in UiO-66-FA confirms that I 2 uptake can be increased without compromising the stability of the MOF, an approach that can be applied potentially to other capture systems.
Author Contributions: J.M. and X.K.: all experimental work; J.M. and L.L.: validation experiments; B.H., S.Y. and M.S.: design of project, supervision and overall strategy of project. All authors have contributed to the preparation of the manuscript. All authors have read and agreed to the published version of the manuscript.