Biomimetic Photodegradation of Glyphosate in Carborane-Functionalized Nanoconfined Spaces

The removal of organophosphorus (OP) herbicides from water has been studied using adsorptive removal, chemical oxidation, electrooxidation, enzymatic degradation, and photodegradation. The OP herbicide glyphosate (GP) is one of the most used herbicides worldwide, leading to excess GP in wastewater and soil. GP is commonly broken down in environmental conditions to compounds such as aminomethylphosphonic acid (AMPA) or sarcosine, with AMPA having a longer half-life and similar toxicity to GP. Metal–organic frameworks (MOFs) are excellent materials for purifying OP herbicides from water due to their ability to combine adsorption and photoactivity within one material. Herein, we report the use of a robust Zr-based MOF with a meta-carborane carboxylate ligand (mCB-MOF-2) to examine the adsorption and photodegradation of GP. The maximum adsorption capacity of mCB-MOF-2 for GP was determined to be 11.4 mmol/g. Non-covalent intermolecular forces between the carborane-based ligand and GP within the micropores of mCB-MOF-2 are thought to be responsible for strong binding affinity and capture of GP. After 24 h of irradiation with ultraviolet–visible (UV–vis) light, mCB-MOF-2 selectively converts 69% of GP to sarcosine and orthophosphate, following the C–P lyase enzymatic pathway and biomimetically photodegrading GP. Circumventing the production of AMPA is desirable, as it has a longer half-life and similar toxicity to GP. The exceptional adsorption capacity of GP by mCB-MOF-2 and its biomimetic photodegradation to non-toxic sarcosine make it a promising material for removing OP herbicides from water.

Gas sorption-desorption of CO2 at 273-313 K, N2 at 77 and 313 K, H2 at 77K, CH4, and H2O at 298 K measurements were performed using IGA001, ASAP2020, and Belsorp Max II surface area analyzer, respectively. The sample was first degassed at 130 °C for 12 h.
Crystals for X-ray Diffraction (XRD) were prepared under inert conditions immersed in perfluoropolyether or paratone as protecting oil for manipulation. Suitable crystals were mounted on MiTeGen Micromounts TM and used for data collection at BL13 (XALOC) 1 at the ALBA synchrotron (Spain) with an undulator source and channel-cut Si(111) monochromator and Kirkpatrick-Baez focusing mirrors with a selected wavelength of 0.72932 Å. An MD2M-Maatel diffractometer fitted with a Dectris Pilatus 6M detector was employed. The sample was kept at 100 K with an Oxford Cryosystems 700 series Cryostream. The structure was solved using the ShelXT 2014/5 (Sheldrick, 2014) structure solution program using the direct phasing methods solution method and by using Olex2 as the graphical interface. 2 The model was refined with version 2016/6 of ShelXL using Least Squares minimization. 3 Highly disordered solvent, identified as six ethanol per formula unit, was treated using a solvent mask (Squeeze). A summary of crystal data is reported in Table S1. Powder X-ray Diffraction (PXRD) was recorded at room S4 temperature on a Siemens D-5000 diffractometer with Cu Kα radiation (λ = 1.54056 Å, 45 kV, 35 mA, increment=0.02 o ).
Morphological features were first examined by optical microscopy and then by Scanning Electron Microscopy (SEM) with a QUANTA FEI 200 FEGESEM microscope.
Water contact angles were measured using a Krüss DSA 100 device at room temperature using water as the probe fluid (9 μL).

Inductively Coupled Plasma -Mass Spectrometry (ICP-MS) measurements were carried out
on an Agilent ICP-MS 7700x instrument. ICP-OES measurements were also collected on a Horiba JY Ultrex instrument.
Wide Angle X-ray Scattering (WAXS) patterns were recorded on the NCD-SWEET beamline at the ALBA synchrotron light source (Spain). An X-ray beam of 8 keV (λ = 1.54 Å) was set using a Si (111) channel-cut monochromator. The scattered radiation was recorded using a Rayonix LX-255HS area detector. The sample-to-detector distance and the reciprocal space calibration were obtained using Cr2O3 as a standard calibrant. The MOF was introduced in a borosilicate capillary and heated to 300 ºC using a Linkam TMS-350 capillary stage (10 ºC/min from 25 ºC; resting 30 minutes after every 50 ºC increase) under dynamic vacuum. Data was reduced from 2D images to 1D profiles via azimuthal integration using PyFAI. 4
The synthesis of the 1,7-di(3,5-dicarboxyphenyl)-1,7-dicarba-closo-dodecaborane ligand (mCB-H4L2) was adapted from the literature procedure. 5 The experiment was conducted under a nitrogen atmosphere in a round-bottomed flask equipped with a magnetic stir bar. 1.00 g (6.93 mmol) of m-carborane (mCB) was added to an oven-dried Schlenk flask. The flask was evacuated and backfilled with N2 three times, then 1,2-dimethoxyethane (50 mL) was added to the flask. Once the mCB was totally dissolved, 10.2 mL (1.6 M in hexane, 16.32 mmol) of n-BuLi was added dropwise at 0ºC. The mixture was then stirred at room temperature for 20 min, and then CuCl (2.38 g, 24.04 mmol) was added to the solution. The mixture was stirred for 20 min, and then 1.11 mL (0.010 mmol) of pyridine and 2.01 mL (13.86 mmol) of 5-I-m-xylene were added. The solution was heated and refluxed at 85ºC until the TLC showed the original compound was almost completely consumed. The cooled mixture was diluted with 200 mL ether and allowed to stand for 2 h. The precipitate was filtered off, and the solution was extracted three times with an HCl (3 M) solution. The diethyl ether was removed by rotatory evaporation, providing a sticky solid that was filtered through a silica gel column (ethyl acetate: petroleum ether = 1:10). The filtrate was concentrated using a rotary evaporator to obtain mCB-L1 as a white solid (1.54 g; 63.04%). Synthesis of 1,7-di(3,5-dicarboxyphenyl)-1,7-dicarba-closo-dodecaborane (mCB-H4L2).
The procedure was adapted from a literature procedure. 5 6.93 g (69.3 mmol) of CrO3 was added in small portions to a stirred mixture of 1.54 g (4.37 mmol) of mCB-L1, 60 mL glacial acetic acid, 30 mL acetic anhydride, and 6.23 mL of concentrated H2SO4. The dark green mixture was stirred at room temperature for 2 hours and then poured into 100 mL of distilled water. A precipitate was filtered off and washed with distilled water to remove the green chromium residues. The off-white solid was recrystallized by dissolving it in a Na2CO3 solution, filtering it, and then acidifying it with an HCl (12M) aqueous solution. The white precipitate that appeared was filtered off to obtain 1.89 g (57.6%) of pure mCB-H4L2. 1

Synthesis of [Zr6(μ3-O)4(OH)4(OH)4(H2O)4(mCB-L2)2] (mCB-MOF-2).
mCB-H4L2 (20 mg, 0.0423 mmol) and ZrCl4 (29 mg, 0.1269 mmol) were dissolved in DMF (5 mL). Formic acid (2.0 mL) was added to this solution in an 8-dram vial. The vial was closed and heated at 120 °C in an oven for 48 h, followed by slow-cooling to room temperature for 10 h. After centrifugation, the amount of adsorbed GP or GF was measured from the difference between the initial (C0) and equilibrium (Ce) concentrations in the supernatant. The equilibrium uptake was calculated by the equation: Where qe (mmol g -1 ) is the equilibrium adsorption capacity of GP or GF on mCB-MOF-2', V is the volume of OP solution (L), and W is the weight of the used adsorbents (g).
The adsorption isotherms for GP and GF using both the Langmuir and Freundlich models on mCB- Figures S8 and 4b where Ce is the equilibrium concentration of OP herbicide (mmol L -1 ), qe is the amount of OP adsorbed at equilibrium (mmol g -1 ), KL represents the Langmuir constant (L mmol -1 ) that relates the adsorption energy and affinity of binding sites, and qm denotes the maximum adsorption capacity (mmol g -1 ). Freundlich adsorption mathematical expression is as follows:

MOF-2' are provided in
where KF (mmol 1−n L n g −1 ) represents the Freundlich constant, which is related to the adsorption capacity of the adsorbent, and n is a parameter that indicates the adsorption intensity. The value of n reflects the type of isotherm to be favorable (0 < n < 1), irreversible (n = 0), or unfavorable (n > 1).
The adsorption data were also studied using the Temkin and Dubinin-Radushkevich adsorption models (Figures S9, S10 and Table S4)  Computational Studies.
DFT calculations were carried out using the CP2K code. 7  Photodegradation studies.
All photodegradation experiments were conducted using a 300 W Newport xenon arc lamp utilizing a Newport-Oriel Instruments OPS-A500 power supply. Irradiation was conducted using   Table S1. Single Crystal X-ray Diffraction.
a Based on the formula without uncoordinated solvent molecules. b R1 = (F0−FC)/F0.   Figure S4. FT-IR spectra of as-made (black) and acetone exchanged mCB-MOF-2 (blue). S17 Figure S5. TGA diagrams of as-made (black) and acetone exchanged mCB-MOF-2 (blue).  c. d. When mCB-MOF-2' was immersed in acidic (pH 1) or basic (pH 11) conditions for 24 h, the N2 uptake was mostly retained, with uptakes of 261 and 248 STP cm 3 /g, respectively, as compared to the pristine sample (288 cm 3 /g; Table S2). As expected, a similar trend was observed with the BET surface areas of the treated samples. mCB-MOF-2' yielded a surface area of 1095 m 2 /g, and the treated MOFs showed surface areas ranging from 1088 to 829 m 2 /g (Table S2).   Figure S9. Temkin Model Plots for mCB-MOF-2. The red line represents GP adsorption; the blue line represents GF adsorption.   Figure S11. PXRD patterns of mCB-MOF-2 after GP adsorption. c. d.  Figure S17. 1  there is no evidence of AMPA in the 31 P spectrum for mCB-MOF-2, while in b., the 31 P peak corresponding to AMPA can be seen at ~10.82 ppm. The large peak at approximately 0.00 ppm for corresponds to orthophosphate, which helps support the C-P Lyase type degradation by mCB-MOF-2 and our hypothesis that the shifting peak positions are due to pH changes throughout the reaction. Additionally, a large peak for orthophosphate is observed for mCB-MOF-2 (a.), while the same peak for TiO2 (b.) is much smaller, demonstrating again that mCB-MOF-2 is selective in forming only sarcosine and orthophosphate.

a.
b.

S32
Upon thoroughly investigating the 1 H NMR data for the photodegradation of GP via mCB-MOF-2, we noticed that longer UV-Vis irradiation times resulted in a downfield shift of the peaks corresponding to the α-protons for both GP and sarcosine. As sodium 3-(trimethylsilyl)-1-propane sulfonate (DSS) was used as an internal standard ( Figure S13) and to calibrate the axes, the shifting peak positions are likely due to other changes, such as pH (i.e., the production of orthophosphate).
Changes in pH throughout the reaction would result in changes in the chemical shift environment (electronic environment) of the α-protons. Previous work has shown that these changes can either be through-bond or through-space. 28      b., c., d., e., and f. are the corresponding mass spectra for each peak with the expected major ions for each product.

d. c.
b. a.
f. e. Figure S26. PXRD patterns of mCB-MOF-2 after UV-Vis irradiation shows that crystallinity is maintained after UV-Vis exposure.  Table S7. Summary of retention times and major ions in mass spectra of derivatized standards for GCMS.