Mixed-linker UiO-66: structure–property relationships revealed by a combination of high-resolution powder X-ray diffraction and density functional theory calculations †

The use of mixed-linker metal–organic frameworks (MIXMOFs) is one of the most effective strategies to modulate the physical–chemical properties of MOFs without affecting the overall crystal structure. In many instances, MIXMOFs have been recognized as solid solutions, with random distribution of ligands, in agreement with the empirical rule known as Vegard’s law. In this work, we have undertaken a study combining high-resolution powder X-ray diffraction (HR-PXRD) and density functional theory (DFT) calculations with the aim of understanding the reasons why UiO-66-based amino-and bromo-functionalized MIXMOFs (MIXUiO-66) undergo cell expansion obeying Vegard’s law and how this behaviour is related to their physical–chemical properties. DFT calculations predict that the unit cell in amino-functionalized UiO-66 experiences only minor expansion as a result of steric effects, whereas major modification to the electronic features of the framework leads to weaker metal–linker interaction and consequently to the loss of stability at higher degrees of functionalization. For bromo-functionalized UiO-66, steric repulsion due to the size of bromine yields a large cell expansion, but the electronic features remain very similar to pristine UiO-66, preserving the stability of the framework upon functionalization. MIXUiO-66 obtained by either direct synthesis or by post-synthetic exchange shows Vegard-like behaviour, suggesting that both preparation methods yield solid solutions, but the thermal stability and the textural properties of the post-synthetic exchanged materials do not display a clear dependence on the chemical composition, as observed for the MOFs obtained by direct synthesis.


Introduction
2][3][4] Mixed-linker metalorganic frameworks (MIXMOFs or MTVMOFs) are a subclass of MOFs incorporating two or more linkers having similar sizes and different functional groups. 5,6The resulting crystal structure is isoreticular to that observed when only one type of ligand is used, but the physical-chemical properties of the material are modulated by the relative amounts of functional groups attached to the framework. 7,85][36] Apart from a few exceptions, 25 this behaviour is generally observed at any relative ratio of ligands.Information on the random distribution of linkers in MIXMOFs is an important aspect, especially for enabling their use as catalysts for cascade reactions or processes where the vicinity of the active sites is key to ensure catalytic activity, 11 or to precisely tailor the pore size to improve the sieving properties. 32,33ne of the pieces of evidence normally used to support the solid solution hypothesis in MIXMOFs is the observation of a gradual shift of reflections in the powder X-ray diffraction (PXRD) patterns associated with varying ligand ratios. 16,20The shift is due to the adaptation of the structure to the presence of linkers with different properties, leading to either shrinkage or expansion of the unit cell.][39] In the case of heterogeneous distributions arising from scarce compatibility between monomers, such as core-shell architectures or segregation into single phase domains, the PXRD pattern would clearly display the simultaneous presence of two well-defined sets of reflections associated with each single crystalline phase. 6][42] The agreement to Vegard's law for inorganic materials has been related to factors such as the ionic radius, volume of valence electrons, Brillouin zone overlaps and electrochemical factor. 39Moreover, the regular variation of the unit cell is often accompanied by a similar behaviour of the physical properties of the materials, e.g.melting temperature 43 and band gap. 44n the case of MIXMOFs, little is known so far about the factors responsible for the shrinkage or expansion of the unit cell upon changes in the composition.This lack of insight can be ascribed to the fact that MOFs, being organic-inorganic coordination compounds, are much more complex systems to study than metal alloys or ionic compounds.However, the existence of characteristic structural trends in MIXMOFs might as well correlate with the modulation of physical-chemical properties and it is therefore of high interest to investigate whether these trends can be associated with specific structureproperty relationships.
Two series of amino-and bromo-functionalized samples were obtained by direct synthesis (DS_MIXUiO-66), whereas two other series containing ABDC and BDC were prepared by postsynthetic ligand exchange (PSE_MIXUiO-66) starting from UiO-66 and UiO-66-NH 2 , respectively.HR-PXRD was exploited to assess whether MIXUiO-66 obeyed Vegard's law and can be regarded as solid solutions and if the preparation method had any relevant influence on the distribution of ligands in the crystal structure.DFT calculations were essential for unraveling the causes of the observed structural trends and for relating them to the thermal and chemical stability of the MIXUiO-66.We chose to investigate MIXUiO-66 containing either ABDC or BBDC because these two linkers have complementary features: the amino group in ABDC is a small-sized electron-donor, and the bromo substituent in BBDC is instead a sterically demanding electron-withdrawer.[47] Experimental section

DS_MIXUiO-66
DS_MIXUiO-66-NH 2 was prepared according to the procedure reported by Chavan et al.: 27 ZrCl 4 (333 mg, 1.43 mmol) was dissolved in N,N-dimethylformamide (DMF, 37 mL) and H 2 O (0.033 mL, 1.83 mmol).Then, BDC and/or ABDC were added in a stoichiometric amount with respect to ZrCl 4 (see Table S1 for details on the amounts, ESI †) and the mixture was sonicated until it became clear.Five reaction solutions were prepared in this way, containing 0%, 25%, 50%, 75% and 100% ABDC, respectively.Each solution was introduced into a glass vial, sealed and kept for 72 h in an oven at 100 1C.After the reaction was completed, the content of every vial was centrifuged and the solid was washed with DMF (10 mL), soaked in DMF (10 mL) for 16 h, washed with acetone (2 Â 10 mL), soaked in acetone (10 mL) for 16 h and finally dried in a vacuum oven at 70 1C.The mixed-linker samples were named according to the amount of ABDC used in the synthesis as follows: DS_MIXUiO-66-NH 2 -25, DS_MIXUiO-66-NH 2 -50 and DS_MIXUiO-66-NH 2 -75.
DS_MIXUiO-66-Br was prepared according to the same procedure, just by replacing ABDC with the same molar amount of BBDC in the reaction mixture (Table S2, ESI †).The mixed-linker samples were named according to the amount of BBDC used in the synthesis as follows: DS_MIXUiO-66-Br-25, DS_MIXUiO-66-Br-50 and DS_MIXUiO-66-Br-75.

PSE_MIXUiO-66
PSE was performed by slightly modifying the literature procedure. 29BDC (18 mg, 0.1 mmol; 36 mg, 0.2 mmol; 72 mg, 0.4 mmol, respectively) was introduced into three 20 mL glass vials, 2.5 mL of 1 M KOH was added, and then the solution was neutralized with 1 M HCl.70 mg of UiO-66 (0.2 mmol BDC) was added to each vial.The vials were sealed and kept at 80 1C for 5 days.The content of each vial was then centrifuged and the solid was washed with water (5 mL), washed with DMF (5 mL), soaked in DMF (5 mL) for 16 h, washed with acetone (2 Â 5 mL), soaked in acetone (5 mL) and finally dried in a vacuum oven at 70 1C.The samples prepared using this procedure were named PSE_MIXUiO-66-NH 2 (42%), PSE_MIXUiO-66-NH 2 (66%) and PSE_MIXUiO-66-NH 2 (86%), where the numbers in brackets are the amounts of ABDC after exchange calculated by quantitative NMR analysis.BDC (17 mg, 0.1 mmol; 34 mg, 0.2 mmol; 68 mg, 0.4 mmol, respectively) was introduced into three 20 mL glass vials, 2.5 mL of 1 M KOH was added, and then the solution was neutralized with 1 M HCl.78 mg of UiO-66-NH 2 (0.2 mmol ABDC) was added to each vial.The vials were sealed and kept at 80 1C for 5 days.Workup was performed as described above.The samples prepared using this procedure were named PSE_MIXUiO-66-NH 2 (65%), PSE_MIXUiO-66-NH 2 (57%) and PSE_MIXUiO-66-NH 2 (47%), where the numbers in brackets are the amounts of ABDC after exchange calculated by quantitative NMR analysis.

Digestion of samples for NMR analysis
Quantitative NMR analysis was performed on a Bruker Avance III spectrometer (500 MHz).20 mg of solid were digested for 24 h in 1 mL of 0.1 M NaOH in D 2 O.The NMR tubes were then loaded with the solution, which was filtered to avoid the presence of solid particles in dispersion.

Thermogravimetric analysis
Thermogravimetric analysis was performed on a Mettler-Toledo instrument.About 10 mg of samples was loaded in an alumina crucible and heated up to 700 1C at a rate of 5 K min À1 with an 80 : 20 N 2 /O 2 mixture.UiO-66 and UiO-66-NH 2 were also analyzed under a pure N 2 stream.

Gas sorption analysis
Nitrogen sorption isotherms at 77 K were recorded on a Micromeritics 3Flex analyzer.The samples (about 30 mg) were activated overnight under vacuum at 150 1C prior to analysis.The BET numbers were calculated in the 0.003-0.035P/P 0 range, where all the criteria defined by Gomez-Gualdron et al. 48were fulfilled.

HR-PXRD data collection and treatment
HR-PXRD data were collected at the Material Science (MS) beamline at Swiss Light Source (SLS), 49 using 16 keV radiation for the amino-functionalized samples and 12.4 keV for the bromofunctionalized samples.Glass capillaries with a diameter of 0.5 mm were used for all the standards and samples.Evacuation was performed on capillaries loaded with powders and glass wool as a stopper.The samples were inserted in a Schlenk tube and kept under vacuum at 120 1C for 72 h.Then, argon was flowed inside the Schlenk tube and the capillaries were sealed with a flame and wax without exposing them to moist air.In situ evacuation was performed on a capillary connected to a vacuum pump via a sample stage equipped with a Swagelok connection.
Whole powder pattern fitting on all standards and samples was performed by the Pawley method using the program TOPAS 5.0.Si-640d was employed as a standard to determine the exact wavelength, zero error, and horizontal and vertical displacement parameters.Na 2 Ca 3 Al 2 F 14 (NAC) was employed as a standard to determine the instrumental profile function.For MIXUiO-66 samples, the lattice parameter, crystal size and strain were refined.

Computational studies
Quantum Espresso v5.3 50 was used for all calculations.The geometry optimizations were performed at the gamma point using the DFT (density functional theory) PBEsol 51 normconserving pseudopotential 52 at a cut-off energy of 100 Ry (see Fig. S1 in the ESI † for the convergence).The convergence threshold was set to 10 À6 and 10 À5 for the energy and force, respectively, whilst the pressure threshold was set to 10 À4 .The electronic structure was converged up to 10 À9 .To reduce the computational cost, the calculations were performed using the UiO-66 primitive unit cell (114 atoms).

DS_MIXUiO-66
DS_MIXUiO-66 samples were prepared starting from reaction mixtures containing 0, 25, 50, 75 and 100% of functionalized ligands.Quantitative 1 H-NMR analysis was performed to determine the relative amounts of ligands in the MIXUiO-66, showing that the incorporation of both ABDC and BBDC into the solid was slightly preferential with respect to BDC (Tables S3  and S4, ESI †).Fig. 2 shows the HR-PXRD patterns for both the series of samples.
UiO-66 crystallizes in the cubic Fm% 3m space group, therefore a single lattice parameter a defines its unit cell and all the reflections in the PXRD pattern shift upon cell expansion or contraction.The data clearly showed that both the series of samples underwent cell expansion upon increasing the amount of functionalized ligands in the framework, with the reflections being shifted to a lower 2y angle (Fig. 2).Table 1 and Fig. 3 report the results of the Pawley refinement of the HR-PXRD patterns.
The unit cell parameter a for the amino-functionalized MIXUiO-66 increased from 20.7419(2) Å to 20.774(2) Å when the amount of ABDC in the framework increased from 0% to 100%.A similar regular trend was observed for bromofunctionalized MIXUiO-66, although in this case the cell expansion from UiO-66 to UiO-66-Br was larger, with a reaching a value of 20.795(2) Å for the fully functionalized material.Based on these results, the samples appear to obey Vegard's law, suggesting that any potential local inhomogeneity does not play a significant role.Recently, Chavan and co-workers synthesized for the first time a series of MIXUiO-66 using both BDC and ABDC as ligands. 27They assessed the solid solution nature of the materials by means of coupled thermogravimetricdifferential scanning calorimetric (TG-DSC) analysis, which displayed only one heat signal related to framework decomposition shifting to lower temperature with an increasing amount of ABDC in the framework.Our findings strengthen the solid solution hypothesis, expanding it to bromo-functionalized MIXUiO-66.
This journal is © the Owner Societies 2017 In order to separate the effect of the solvation on the unit cell expansion from the real effect of the functionalization, the amino-functionalized samples were evacuated for 72 h at 120 1C in a vacuum.Under these conditions, most of the molecules of solvent trapped in the pores were removed, while avoiding dehydroxylation of the clusters, which occurs at higher temperature and is known to cause cell contraction. 15he cell underwent expansion upon solvent removal, but the trend observed for the as-synthesized samples was maintained (Fig. 4 and Table S5, ESI †).In this case a increased from 20.7520(4) Å for UiO-66 to 20.798(2) Å for UiO-66-NH 2 , evidencing how the extent of the cell expansion was slightly larger than that in the as-synthesized samples (0.046 Å vs. 0.032 Å, respectively).An in situ HR-PXRD experiment was performed while evacuating a capillary containing non-functionalized UiO-66 to further assess the effect of solvent removal (Fig. S2, ESI †).The sample was initially kept at room temperature and vacuum was applied to the capillary: after 30 minutes the cell parameter increased from 20.750    expansion, the sample was then cooled down to room temperature under vacuum and a shrank to 20.763(2) Å.The expansion upon complete removal of the solvent was 0.013 Å, in good agreement with the value observed when the sample was evacuated ex situ (0.010 Å).The slightly lower expansion observed for the sample evacuated ex situ was attributed to the noncomplete removal of solvent.This result confirms the reliability of the data collected on the evacuated samples.Computational studies were performed to investigate the reason of the unit cell expansion.Starting from the experimental structure of pristine UiO-66, NH 2 -and Br-groups were consecutively added to each single BDC linker (six BDC linkers are present in the UiO-66 primitive cell) and fully optimized (i.e.relaxing the atomic positions and lattice parameters, see the ESI † for the optimized geometries).Fig. 5 shows the calculated highest occupied crystalline orbitals (HOCOs) for UiO-66, UiO-66-NH 2 and UiO-66-Br.UiO-66 is composed of Zr 6 O 4 (OH) 4 clusters and BDC organic ligands; its HOCO is a combination of non-bonding O p-orbitals which resembles that found in ZrO 2 (Fig. S3, ESI †).However, in UiO-66 also ''organic'' oxygen atoms of the carboxylate are present which contribute to the orbital (Fig. 5, top) resulting in a ''hybrid organic-inorganic'' orbital that stabilizes the framework.Functionalizing the ligand with amino groups significantly changed the electronic features of the compound, as recently described also by Hendrickx et al. 53 and Flage-Larsen et al. 54 As the BDC ligands were replaced with ABDC ligands, new orbitals were populated; the new HOCOs were localized on the phenyl ring and on the amino group of the functionalized ligands (Fig. 5, middle) and the nature of the HOCO changed becoming mainly ''organic'' in nature.As the new orbital got more and more populated increasing the percentage of the functionalized ligand, the band gap shrank and the interaction between the organic ligands and the clusters became weaker, explaining the loss of stability of UiO-66-NH 2 compared to that of the pristine UiO-66 found by thermogravimetric analysis. 54,55The mean Zr-O bond lengths in UiO-66-NH 2 and UiO-66 were virtually the same (2.269Å vs. 2.270 Å), excluding a straightforward correlation between the strength of the linker-metal interaction and the Zr-O bond length.Steric effects between NH 2 and the O lone pairs seemed not to be relevant at any ABDC loading, resulting in a very small expected cell expansion (Table 2).As found by experiments, a larger cell expansion for the bromo-functionalized MIXUiO-66 was predicted by calculations (Table 2).Upon functionalizing the ligand with Br instead of NH 2 , new HOCOs were populated but a different situation occurred.In this case the new populated HOCOs were closer to the configuration of pristine UiO-66 (i.e. of ''hybrid'' nature) than they were in the amino-functionalized samples.The presence of Br, which has a large atomic volume due to its lone pairs, yielded steric repulsion between the Br and O p-orbitals (Fig. 5, bottom), inducing an increase of the cell volume without affecting the stability, as revealed by thermogravimetric analysis, discussed herein.Fig. 6 compares the relative change in the volume (V/V 0 ) of the predicted and experimental unit cells with the increase of the functionalized ligand.The calculations seemed to underestimate the cell expansion, especially for the MIXUiO-66-NH 2 series, and the discrepancy between experimental and calculated trends became larger with increasing functionalization degree.This can be explained remembering that calculations are performed at 0 K and do not include any vibrations.Hence we can attribute the discrepancies to the thermal motion which adds some steric repulsion between the functionalized ligand and the framework.This steric contribution seems to be increasingly larger as the functionalization degrees increase.Nevertheless, a good agreement in the expansion trend (i.e.linear expansion with increasing functionalization, larger cell expansion for the Br-functionalized UiO-66) is found, corroborating the idea that the functionalized ligands are in fact randomly incorporated into the framework and that the solid solution hypothesis is reliable.This also suggested that defects, in the form of either missing-linkers (as revealed by TGA) 57 and/or missing-clusters (as revealed by weak superlattice reflections in the PXRD patterns), 58 were not the major contributors to the observed unit cell expansion.
TGA of the materials is shown in Fig. 7.As already reported by Chavan et al., 27 the stability of the framework of DS_MIXUiO-66-NH 2 samples decreased as the loading of ABDC increased (Fig. 7a).A similar behaviour was also observed for other topologies. 16,20In the case of MIXUiO-66, the loss of stability can be imputed to the decreasing strength of the metal-linker interaction when moving from pure UiO-66 to pure UiO-66-NH 2 , as evidenced by the calculations discussed above.Performing the analysis under a 100% nitrogen stream (Fig. S4, ESI †), the same trend was observed, ruling out the possibility that the stability loss was due to easier combustion of the organic part when more ABDC was embedded in the framework.On the other hand, Fig. 7b shows that the thermal stability of DS-MIXUiO-66-Br samples was minimally affected by the presence of BBDC, suggesting that steric effects did not have a significant impact on the strength of the ligand-metal interaction.Again, this is in good agreement with the prediction of our calculations.The amount of defects calculated from TGA was about 9% for all the MIXUiO-66-NH 2 samples (except for MIXUiO-66-NH 2 -25, for which we found 0%), whereas for MIXUiO-66-Br it ranged from 5% for MIXUiO-66-Br-50 to 13% for MIXUiO-66-Br-75 and UiO-66-Br up to 20% for MIXUiO-66-Br-25.The very similar decomposition temperatures of MIXUiO-66-Br samples, despite defectivity ranging between 5 and 20%, do not display a significant influence of defects on the stability.Our findings unravel the reasons behind the observed thermal stability of amino-and bromofunctionalized UiO-66 materials, which were previously the only object of speculations by Kandiah et al. 59 The observations by the same authors on the lower stability to hydrolysis of UiO-66-NH 2 if compared to UiO-66 and UiO-66-Br can also be rationalized based on the weaker interaction between ABDC and the clusters.It is indeed known that in most MOFs hydrolysis occurs at the metal-linker bonds, which are the chemically weakest points. 60itrogen sorption isotherms (Fig. S5, S6 and Tables S6, S7, ESI †) display ideal type 1 behaviour and a regular decrease of accessible volume with increasing loading of either ABDC or BBDC.

PSE_MIXUiO-66
PSE is an alternative method to prepare MIXUiO-66 by soaking phase-pure materials in a solution of the desired ligand in a polar solvent. 613][64][65] Kim and co-workers 29 used aerosol time-of-flight mass spectrometry (ATOFMS) to prove that single UiO-66 microcrystals that underwent PSE contained both the ligands.This evidence, together with the observation that crystallinity was not affected, ruled  out the possibility that the MOF underwent dissolution and recrystallization into two distinct crystalline phases.However, the process of PSE is still not fully understood and it is not clear whether the degree of substitution is uniform throughout the microcrystals or if there is a gradient due to increased diffusion limitations in the inner part of the crystals that might generate domains with considerably different compositions.
We prepared a series of MIXUiO-66-NH 2 (hereafter PSE_MIXUiO-66-NH 2 _A) by soaking UiO-66 in aqueous solutions containing 0.5, 1.0 and 2.0 equivalents of ABDC, and another series (hereafter PSE_MIXUiO-66-NH 2 _B) by soaking UiO-66-NH 2 in aqueous solutions containing 0.5, 1.0 and 2.0 equivalents of BDC.Quantitative 1 H-NMR analysis revealed that the degree of exchange of ABDC into UiO-66 (42%, 66% and 86%, when 0.5, 1.0 and 2.0 equivalents of ABDC were used, respectively) was higher than that of BDC into UiO-66-NH 2 (35%, 43%, 53%, when 0.5, 1.0 and 2.0 equivalents of BDC were used, respectively).This confirmed once again the higher coordinating capability of ABDC with respect to BDC (Table S8 and Fig. S7, ESI †).Complete exchange was not achieved even when a two-fold excess of free ligand was used, suggesting that an equilibrium state was reached as the concentration of the ligand expelled from the MOF in the solution increased.
HR-PXRD patterns of all the samples clearly displayed the presence of a single phase (Fig. S8 and S9, ESI †).As for DS_MIXUiO-66-NH 2 , there was no evidence of the presence of significant local inhomogeneity.Table 3 and Fig. 8 show the results of the Pawley refinements on the HR-PXRD patterns.
The lattice parameter of PSE_MIXUiO-66-NH 2 _A grew larger as more ABDC was introduced into the framework.In the case of PSE_MIXUiO-66-NH 2 _B, the lattice parameter decreased as the framework incorporated more BDC, although sample PSE_MIXUiO-66-NH 2 _B(47%) slightly deviated from the ideal Vegard behaviour.Comparison of the absolute values of the lattice parameter for PSE_MIXUiO-66-NH 2 _A, PSE_MIXUiO-66-NH 2 _B and DS_MIXUiO-66-NH 2 reveals that the samples prepared by PSE have larger a than expected from analysis of the as-synthesized materials obtained by DS, but still smaller than the evacuated ones.This difference can be mainly attributed to the effect of the solvent: TGA of PSE_MIXUiO-66-NH 2 _A and PSE_MIXUiO-66-NH 2 _B (Fig. S10 and S11, ESI †) displays on average a smaller weight loss than DS_MIXUiO-66-NH 2 at temperature lower than 250 1C, where the removal of solvent molecules from the pores occurs.This is most likely a result of the different preparation procedures.DS was carried out in DMF and the workup involved the same solvent and acetone.
As a result, two distinct weight loss steps below 250 1C can be seen in the TG curves of these samples (Fig. 7a).PSE was performed in water and the workup procedure involved soaking in both DMF and acetone.A single weight loss step below 250 1C suggested that water was preferentially adsorbed inside the pores (Fig. S10 and S11, ESI †).
The thermal stability of the framework of all materials obtained by PSE was between those of pure UiO-66 and pure UiO-66-NH 2 (Fig. S10 and S11, ESI †).However, there was no clear dependence of the thermal stability on the amount of ABDC.Similarly, the gas sorption isotherms did not display any evident trend for the surface area and porosity (Fig. S12 and S13, ESI †).These results suggest that MIXUiO-66 prepared by PSE, although being phase-pure and seemingly obeying Vegard's law, could undergo partial amorphization or occlude free linkers inside the pores, which were not completely removed during the workup procedure, thus affecting their thermal stability and textural properties.

Conclusions
We performed a thorough fundamental investigation on aminoand bromo-functionalized MIXMOFs based on the UiO-66 topology prepared by either direct synthesis or post-synthetic ligand exchange.The samples obtained by direct synthesis displayed the ideal behaviour of solid solutions.HR-PXRD analysis revealed that they obeyed Vegard's law, undergoing cell expansion with increasing relative amount of ABDC and BBDC.No clear evidence of important short-range inhomogeneity was detected in the diffraction patterns.This proved that the ligands were randomly embedded in the framework.DFT calculations were exploited to support the solid-solution hypothesis and to identify the factors responsible for cell expansion, with the aim of disclosing specific structure-property relationships.The experimental trends of cell expansion were correctly predicted by theory and the thermal  This journal is © the Owner Societies 2017 and chemical stability of the framework could be explained on the basis of the electronic features of the MOFs.The unit cell of MIXUiO-66-NH 2 slightly expanded as a consequence of a small steric contribution, while the metal-linker interaction was significantly weakened, causing loss of stability.MIXUiO-66-Br underwent a larger unit cell expansion due to the presence of bulky bromine atoms, however the electronic features remained very similar to those of pristine UiO-66, without affecting the stability of the framework.MIXMOFs prepared by post-synthetic exchange did not display the same behaviour, suggesting that different synthesis methods lead to materials with different properties.The findings reported in this work highlight the intimate correlation existing between the fine structural details, electronic features and physical-chemical properties of MIXMOFs, paving the way for a rational design of these materials with precise features.

Fig. 1
Fig. 1 Molecular structures of the ligands employed in this work.
(3) Å to 20.756(3) Å, indicating that part of the solvent had already been removed.Upon continuing evacuation for further 30 minutes, a reached a steady value of 20.758(3) Å.When the temperature was increased to 70 1C, the cell parameter further expanded to 20.764(2) Å, reaching a steady value of 20.767(3) Å at 100 1C, suggesting that all the solvent had been removed.To get rid of the effect of thermal

Fig. 4
Fig.4Comparison of the refined lattice parameters for DS_MIXUiO-66-NH 2 samples as-synthesized (black) and after evacuation at 120 1C for 72 h under vacuum (red).

Fig. 5
Fig.5Calculated highest occupied crystalline orbitals (HOCOs) for UiO-66 (top), UiO-66-NH 2 (middle) and UiO-66-Br (bottom).The zoomed insets show how the orbital is spread between the inorganic and organic part for UiO-66 and UiO-66-Br, while for UiO-66-NH 2 it is predominantly of organic nature.For UiO-66-Br a steric repulsion (pointed out inside the circle in the inset) between the Br and O lone pair arises.Images of the crystalline orbitals were obtained using the VESTA software.56Adopted colour scheme: Br light brown, C dark brown, H light pink, N blue, O red, and Zr purple.Isovalue: 5 Â 10 À4 .

Table 2
PBEsol calculated unit cell volume This journal is © the Owner Societies 2017

Table 3
Refined lattice parameters for PSE-MIXMOFs