Modulated self-assembly of three flexible Cr(iii) PCPs for SO2 adsorption and detection

Modulated self-assembly protocols are used to develop facile, HF-free syntheses of the archetypal flexible PCP, MIL-53(Cr), and novel isoreticular analogues MIL-53(Cr)-Br and MIL-53(Cr)-NO2. All three PCPs show good SO2 uptake (298 K, 1 bar) and high chemical stabilities against dry and wet SO2. Solid-state photoluminescence spectroscopy indicates all three PCPs exhibit turn-off sensing of SO2, in particular MIL-53(Cr)-Br, which shows a 2.7-fold decrease in emission on exposure to SO2 at room temperature, indicating potential sensing applications.


S2. Synthesis
MIL-53(Cr): Terephthalic acid (1.66 g, 1 mmol) and chromium (III) chloride hexahydrate (2.66 g, 1 mmol) were added to a 200 mL Teflon autoclave liner and water (100 mL) was added along with 12 M HCl (1.5 mL). The mixture was stirred at room temperature for ~10 mins before sealing in a stainlesssteel autoclave and placing in a solvothermal oven at 220 °C for 72 hours. The autoclave was removed and allowed to cool naturally to room temperature. The solids were collected by centrifuge and washing with water (1 x 100 mL), DMF (2 x 100 mL), then ethanol (2 x 100 mL), before drying in a desiccator under vacuum at room temperature overnight.
MIL-53(Cr)-Br: 2-Bromoterephthalic acid (2.45 g, 1 mmol) and chromium (III) chloride hexahydrate (2.66 g, 1 mmol) were added to a 200 mL Teflon autoclave liner and water (100 mL) was added along with 12 M HCl (1.5 mL). The mixture was stirred at room temperature for ~10 mins before sealing in a stainless-steel autoclave and placing in a solvothermal oven at 220 °C for 72 hours. The autoclave was removed and allowed to cool naturally to room temperature. The solids were collected by centrifuge and washing with water (1 x 100 mL), DMF (2 x 100 mL), then ethanol (2 x 100 mL), before drying in a desiccator under vacuum at room temperature overnight.
MIL-53(Cr)-NO2: 2-Nitroterephthalic acid (1.055 g, 0.5 mmol) and chromium (III) chloride hexahydrate (1.33 g, 0.5 mmol) were added to a 200 mL Teflon autoclave liner and water (50 mL) was added along with 12 M HCl (0.5 mL). The mixture was stirred at room temperature for ~10 mins before sealing in a stainless-steel autoclave and placing in a solvothermal oven at 220 °C for 72 hours. The autoclave was removed and allowed to cool naturally to room temperature. The solid was collected by centrifuge and washing with water (1 x 50 mL), DMF (2 x 50 mL), then ethanol (2 x 50 mL), before drying in a desiccator under vacuum at room temperature overnight.

Activation protocol
To exchange the pore-bound terephthalic acid, the samples were treated in DMF (100 mL/g) at 220 °C overnight in a stainless-steel autoclave. The solid was collected by centrifuge and washed once with DMF, then dried in a desiccator under vacuum at room temperature. This yielded a product with bound DMF, MIL-53(Cr)-X-DMF. S3 DMF was removed from the sample by refluxing in methanol (100 mL/g) overnight in an RBF and then washing with methanol (2 x 100 mL/g), before drying in a desiccator under vacuum at room temperature overnight.

S3.1. Indexing
Where possible, powder X-ray diffractograms were indexed and/or Pawley fits were obtained. A comparison of all crystallographic data obtained with some pertinent literature examples is given in Table S1.        [b] Data taken from structure derived from room temperature powder diffraction data and computational modelling. S4 [C] Data taken from single crystal structure collected at 293 K. S5 [d] Data taken from single crystal structure collected at 293 K. S5

S3.3. Comparisons Across Synthesis and Work-Up
Comparisons of powder X-ray diffraction data are given for the three Cr PCPs across each stage of synthesis and work-up: as-synthesised, DMF-treated, MeOH treated, and activated. These show the structural differences induced by the functional groups.

S5.2. CO2 Adsorption Isotherms (273 K and 298 K)
CO2 adsorption isotherms were collected at 298 K at the University of Glasgow to allow calculation of enthalpies of adsorption and direct comparison with the SO2 adsorption data. Data collected at 273 K are presented in Figure 1d of the manuscript. Calculations were carried out accordingly to reported literature, S2, S7 using a virial-type equation (Eq. S1) to fit the low coverage region of two adsorption isotherms at 273 and 298 K for the three materials ( Figure S21).
Ln(n/p) = A0 + A1η + A2η 2 + … (Eq. S1) Where p is the pressure, n is the amount adsorbed and A0, A1, … are the virial coefficients. The plot of Ln (n/p) give a straight line at low surface coverage ( Figure S22). From the linear fittings (using the Clausius-Clapeyron equation) the virial coefficients are used to estimate the enthalpy of adsorption.

S5.3. SO2 Adsorption Isotherms (298 K)
The present scheme seeks to outline the process of pore opening in our materials. It is noteworthy that it has been widely reported that upon heating MIL-53 materials, as in our activation process, they adopt the narrow pore (np) phase. S8 In the case of MIL-53(Cr), the gate opening occurs at a gas pressure of 0.4 bar, to give way to the large pore (lp) phase. On the other hand, for MIL-53(Cr)-Br the pore opening occurs at 0.72 bar, while for MIL-53(Cr)-NO2 the gate opening does not occur during the pressure range of our measurements (0 to 1 bar).
It should be noted that the gate opening process has been reported to depend largely on the chemical nature of the functional groups of the ligands, S9 because these have an impact on intraframework interactions, such as hydrogen bonds that may occur between the functional groups and the H-atom of the μ2-OH bridging ligands from the secondary building unit of the materials. In addition, steric hindrance of the ligand functionalization modifies the inclination of the phenyl ring, which prevents rotation of the ligand and hardens the pore. S9 That is why for MIL-53(Cr) which has no functionalization, gate opening occurs at lower pressures, while for MIL-53(Cr)-Br gate opening occurs at higher pressures, due to Brligand···HμOH intra-framework interactions and Br steric hindrance hindering the opening of the pore. Meanwhile, MIL-53(Cr)-NO2 shows no gate opening from 0 to 1 bar SO2, suggesting that the intra-framework ONO2···HμOH hydrogen bonds are stronger than those occurring in the brominated material, in addition to the fact that the NO2 group is larger than the Br group, which further impedes the rotation of the ligand and consequently prevents the pore opening.
The organic functionalization of the ligands is a key player in the SO2 adsorption process. Based on our previous work, S10 we propose that in the non-functionalized MIL-53(Cr) the initial jump in adsorption occurs because hydrogen bonds are forming between SO2 molecules via their O-atoms and the H-atom of the μ-OH bridging ligand (OSO2···Hμ-OH) in the np phase, these SO2/μ-OH interactions remain predominant throughout the adsorption, even in the lp phase, though they form with slightly longer distances. On the other hand, for the MIL-53(Cr)-Br material we propose that, in addition to hydrogen bonds, the initial adsorption is due to strong electrostatic interactions between the halogen and the S-atom of SO2 (SSO2···Brligand), which causes the first step in the adsorption isotherm to be larger than that of the non-functionalized material. S11 For MIL-53(Cr)-NO2, we propose that polar interactions occur between the O-atom of the NO2 group and the S-atom of SO2 (SSO2···ONO2-ligand), as described in other works, S12 in addition to SO2/μ-OH hydrogen bonds. For all three materials, in addition to the interactions already described, additional interactions between SO2 and the organic ligands (OSO2···Cligand) and interactions between the SO2 molecules adsorbed inside the pores (SSO2··· OSO2) are also expected to occur. S10 It is worth mentioning that for MIL-53(Cr) and MIL-53(Cr)-Br a hysteresis is observed after desorption, this is in good correlation with Schneeman et al., S13 who postulates that upon gas adsorption there is an energy penalty associated with an increase in interfacial area as the phase change from np to lp occurs, however, for the desorption, which is the reverse structural change from lp to np it is not necessary to overcome such an energy barrier. Finally, in the case of MIL-53-NO2 there is no hysteresis observed because the pressure necessary to achieve gate opening is not reached.

S6. Custom ex situ SO2 Adsorption System
The system ( Figure S4) contains two principal parts: SO2 gas generator (A) dropping funnel with H2SO4 conc.
[1] connected to a Schlenk flask with Na2SO3 (s) under stirring [2]; and the saturation chamber (B), constructed from a round flask with distilled water [3], connected to a sintered glass filter adapter [4] and to a vacuum line [5]. To begin with, a sample of about 60 mg in a 1.5 mL glass vial was activated in a sand bath at 220°C under vacuum for 24 h. Then, the vial was quickly placed on the glass adapter, and the system was evacuated with a vacuum line. Next, SO2 gas was generated by dripping concentrated sulfuric acid over Na2SO3, which passed through the sample continuously for 3 hours. Figure S24. Ex situ SO2 homemade system.

S7. Characterisation after SO2 Adsorption
Samples were analysed by powder X-ray diffraction and infrared spectroscopy after exposure to SO2. Figure S25. Comparison of the PXRD patterns of synthesized and washed MIL-53(Cr) material (bright blue), the one exposed to SO2 (blue), and the one exposed to CO2 (blue). Figure S26. Comparison of the PXRD patterns of synthesized and washed MIL-53(Cr)-Br material (brown), exposed to SO2 (mustard), and exposed to CO2 (yellow). Figure S27. Comparison of the PXRD patterns of synthesized and washed MIL-53(Cr)-NO2 material (brown), exposed to SO2 (red), and exposed to CO2 (melon).

S9. Ideal Adsorbed Solution Theory (IAST)
Predictions of the co-adsorption of SO2:CO2 mixtures on MIL-53(Cr), MIL-53(Cr)-Br, and NIL-53(Cr)-NO2 were performed assuming the Ideal Adsorbed Solution Theory (IAST) assumptions as valid and using the Python package pyIAST. S14 The analytical Langmuir isotherm was fitted to the experimental CO2 isotherms for all three materials and to the experimental SO2 isotherm of MIL-53(Cr)-NO2 with non-significant root mean square errors ( Figure S32). None of the analytical models available in pyIAST fitted the experimental SO2 isotherms of MIL-53(Cr) and MIL-53(Cr)-Br. Therefore, these SO2 adsorption data were linearly interpolated, and the distributed pressures were calculated by numerical quadrature implemented in pyIAST ( Figure  S14). Therefore, the adsorption selectivity was calculated as: where xi and yi are the mole fraction of components i = SO2 and CO2 in adsorbed and gas phase, respectively.  The selectivities of SO2 versus CO2 of different SO2:CO2 mole fractions at 1 bar pressure were calculated. The results are shown in Table S2.

S11. Fluorescence Spectroscopy
After SO2 saturation in our homemade saturation system, we moved the samples out of the system and immediately packed them into solid quartz holders which were inserted inside an Edinburgh FS5 Spectrofluorometer coupled to an SC-10 solid-state sample holder. Figure S35. Comparison of the fluorescence spectra of BDC linker (purple) and MIL-53(Cr) material (blue).  It can be noted that, after a time following activation and SO2 saturation, the fluorescence signal of the activated and saturated MIL-53(Cr)-Br samples approaches the signal of the washed and evacuated MIL-53(Cr)-Br sample. The same was observed for MIL-53(Cr) and MIL-53(Cr)-NO2.