A superhydrophobic MOF facilitating efficient solvent-free catalytic chemical fixation of CO₂ and oxidation of hydrocarbons and MOF@cotton@starch composite-based selective sensing of a herbicide

Abstract

Modern-day environmental pollution does not end with only air pollution by harmful gases. Issues also arise due to dust particles released by industry and motor vehicles. Regular use of herbicides, fungicides, and pesticides to increase the production of foodstuffs in agricultural fields also enhances the risk of soil and aquatic environmental pollution. Therefore, the systematic development of multi-functional materials for the catalytic conversion of toxic gases and hydrocarbons into value-added products under mild conditions, selective detection, and quantification of toxic organic pollutants in environmental water bodies are crucial to maintaining a sustainable environment. Herein, we have strategically engineered a highly robust, fluorine-rich, Zr(IV)–organic framework (1) and, desolvated (1′) to address three emerging environmental issues. This is the first recyclable metal–organic framework (MOF) catalyst (1′) where the Lewis basic nature of fluorine atoms and the Lewis acidic nature of the Zr(IV) sites are utilized for the mild pressure, solvent-free entrapping of CO₂ to epoxides with various kinetic dimensions in high yields. The hydrophobic cavity of 1′ also facilitated better conversions of hydrocarbons into their respective alcohols/ketones compared to non-hydrophobic MOF under aerobic oxidation conditions with recyclability. Notably, for the first time, selective, rapid (response time <5 s) and nanomolar (LOD = 1.1 nM) fluorescence-based detection of aclonifen (ACF) herbicide was performed using 1′ in various environmental water samples, at different pH conditions, and also in variety of food and vegetable extracts. Moreover, a reusable 1′@starch@cotton composite was fabricated for the on-site, naked-eye detection of ACF under UV light. Additionally, a detailed investigation of the mechanistic pathways of all the proposed applications was performed experimentally and theoretically.

---

Introduction

The exponential increase in global energy demand forces us to be more reliant on fossil fuels, which is a non-renewable source of energy. Burning of fossil fuels causes many adverse effects on nature, such as climate change, global warming, etc., due to the release of large amounts of CO₂.^1,2 Since the industrial revolution, extensive scientific efforts have been made for the selective capture and conversion of CO₂ into valuable fine chemicals. However, the chemical fixation of CO₂ is always challenging due to the kinetic inertness of CO₂ and high C=O bond enthalpy (805 kJ mol⁻¹).³ Various chemical fixation strategies⁴ have been proposed, including entrapping CO₂ in an epoxide ring to produce cyclic carbonate. This process gained huge importance due to the potential application of cyclic carbonates as fuel additives and electrolytes.⁵ Moreover, various homogeneous catalysts have been reported for selective conversion of CO₂ into cyclic carbonates with 100% atom economy.⁶ Despite all these advantages, the poor separation efficiency of the product and the lack of recyclability make these processes less economical. On the other hand, various MOF-based heterogeneous catalysts have gained considerable attention for the catalytic conversion of CO₂.³ The highly porous nature, as well as the presence of three-dimensional channels and Lewis acidic and basic functionalities in MOFs offer several advantages for the selective adsorption and activation of CO₂.^2,3,7 However, very few MOFs have been reported for conversion of CO₂ under mild pressure and temperature conditions; in most of the reported catalysts, the lone pair of ‘N’ atom acts as a Lewis base.^7,8 In this regard, for the first time, we have demonstrated the usefulness of fluorine-functionalized highly porous, superhydrophobic desolvated Zr(IV)-MOF (1′) for selective chemical fixation of CO₂. The presence of a strongly electron-dense pentafluorobenzene functionality provides two advantages. Firstly, the high fluorine content of the linker enhances the catalyst's hydrophobicity and aqueous/moisture phase stability. Secondly, highly electron-dense fluorine can act as a Lewis basic site for selective interaction and adsorption of CO₂ on the porous surface. Lewis basic activity of the fluorine and its influence on the catalytic conversion of CO₂ in cyclic carbonates is less explored.^9,10 Hence, scientific efforts should be made in this direction.

The presence of a highly electronegative –F group in the linker moiety may be helpful for selective hydrophobic–hydrophobic interaction with hydrocarbons. Hence, this advantage of hydrophobicity in 1′ was further explored in the catalytic aerobic oxidation of hydrocarbons to their corresponding alcohol (ol)/ketone (one) mixture with high selectivity (>90%), which is one of the important reactions in organic chemistry and fine chemical industry.¹¹ To achieve this target, a series of heterogeneous catalysts have been reported in which hydrocarbons are converted to their respective ol/one mixture. Among the various heterogeneous catalysts, MOF-based heterogeneous solids have received considerable attention due to the possibility of the preparation of MOF catalysts with different transition metals or active sites encapsulated within the pores of MOFs.^12–15 Among these different MOF catalysts, UiO-66-based solids have shown superior performance due to the presence of defective sites within the framework.¹⁶ Besides these conventional MOF catalysts, porphyrin-based MOFs have also been reported as oxidation catalysts.¹⁷ Due to these features of MOFs compared to other heterogeneous catalysts, wide ranges of MOFs have been reported for the aerobic oxidation of hydrocarbons, including Au-Pd/MIL-101,¹⁸ MIL-101,¹⁹ NHPI/Fe(BTC),²⁰ Fe(BTC)/TBAB,²¹ {[Cu_0.5La₂(HPDC)(PDC)₂(SO₄)(H₂O)₂]H₂O}_n,²² NHPI/Fe(BTC),²³ Cu-doped ZIF-8²⁴ and Ce-based MOF.²⁵ However, the catalytic performance of a MOF solid in the oxidation of hydrocarbon depends on many factors, such as the nature of active sites, the nature of the oxidant, reaction temperature, and solvent.^14,26 Although various types of MOFs are utilized to catalyse this reaction, very few of them are hydrophobic in nature.

The introduction of hydrophobic groups (–F atoms) in the organic linker of a MOF permits the accommodation of hydrophobic sites, which can be responsible for increasing the concentration of hydrocarbons within the framework of the MOF by providing a confined environment. This further provides closer proximity among the active sites of the oxidant, and the hydrocarbon to initiate a catalytic oxidation reaction. For instance, in a previous study, PCN-222(Fe)-F7 was employed as a heterogeneous hydrophobic MOF catalyst for the oxidation of cyclohexane and showed 50% conversion with 90% selectivity to ol/one mixture using t-butylhydroperoxide (TBHP) at 80 °C in acetonitrile.²⁷ In another report, PCN-224(Mn) MOF was exchanged with 2,2′-bis-(trifluoromethyl)-4,4′-diphenyl phthalate (L) to create hydrophobic environment in PCN-224(Mn)-L. The catalytic performance of PCN-224(Mn)-L in the aerobic oxidation of cyclohexane was observed with 51% conversion and 90% selectivity to ol/one using TBHP as the oxidant at 80 °C in acetonitrile.²⁸ In contrast, the conversion of cyclohexane in the presence of PCN-224(Mn) presented only 16% conversion with 79% selectivity to ol/one under similar conditions. Recently, we have reported that the conversion of cyclohexane to its respective ol/one mixture is significantly higher (21%) in the presence of DUT-52 MOF with trifluoroacetamido-functionalized linker in acetonitrile at 60 °C after 24 h, while the conversion of cyclohexane with DUT-52 alone was only 6% under similar conditions.²⁹ This lower activity in the absence of a hydrophobic environment is due to the poor interaction between the redox sites and cyclohexane, as cyclohexane remains at the exterior surface most of the time. In this study, by using superhydrophobic 1′, under mild reaction conditions, we achieved the maximum conversion of 69% and 100% selectivity for some hydrocarbons. Compared to these previous attempts, a superior activity was observed in the oxidation of cyclohexane with a high ol/one ratio and higher turnover number as a function of the framework pore diameter near the iron site in Fe₂(dobdc) (dobdc⁴⁻ = 2,5-dioxido-1,4-benzenedicarboxylate) MOF.³⁰ In another work, the selective oxidation of methane to CH₃OH was effectively altered by tuning the microenvironment through the hydrophobic modification around Fe-porphyrin within the MOF framework.³¹ Furthermore, a highly selective oxidation of methane to CH₃OH was reported by surface hydrophobic modification over Cu-BTC MOF at low temperature with a yield of 10.67 mmol g_cat⁻¹ h⁻¹ and 99.6% selectivity.³² These pertinent precedents clearly illustrate the important role of hydrophobic environments within the MOF structure to alter the reactivity and product selectivity compared to non-hydrophobic MOFs.

Aclonifen (ACF) is a herbicide of the diphenyl ether family often used in agricultural fields to control weeds in various crops.³³ It is predominantly used to control grass and broadleaf weeds around crops, such as peanuts, soybeans, and sunflowers.³³ ACF selectively interferes with the growth and development of unwanted grass and broadleaf weeds by disrupting their photosynthetic ability or by inhibiting certain metabolic processes.³⁴ This selective action allows crops to thrive, while minimizing the impact on non-target plants.³⁵ The widespread use of this herbicide all over the world for improvement of the efficiency of agricultural production increases the risk of environmental pollution.³⁵ Rural water bodies are mainly contaminated by the extensive use of ACF. Endocrine disruption, potential teratogenicity, limited mutagenic activity, and even some uncertain suspected environmental impacts are few of the hostile effects of ACF. Therefore, extensive exposure to ACF may adversely impact aquatic ecosystems, as well as human health. Hence, regular vigilance regarding the concentration of such toxic herbicides is crucial in natural aquatic systems. Therefore, the development of an economical, reliable sensor for fast, precise, ultrasensitive probing of ACF is required.

Collectively, based on these three points, we have successfully employed 1′ to promote two important organic reactions; (i) mild pressure cycloaddition of CO₂ with epoxides and (ii) aerobic oxidation of hydrocarbons under mild reaction conditions. Moreover, ultrasensitive, selective probing of ACF by 1′ in an aqueous medium is also explored. Some of the notable novelties of 1′ are the lowest ever reported limit of detection (LOD) value (1.1 nM), fast response time (<5 s), selective recognition of ACF in water, high sensitivity in wide pH conditions, and real field detection of ACF from various fruits and vegetable extracts. Additionally, fabrication of 1′ over the surface of starch@cotton bio-composite is also demonstrated for the on-site, naked eye nanomolar level detection of ACF under UV light. Further, a detailed mechanistic investigation of all the applications is executed both experimentally and theoretically.

Experimental section

Synthesis of [Zr₆O₄(OH)₄(C₁₅H₄F₅NO₅)₆]·5H₂O·6DMF (1)

A conventional solvothermal MOF synthesis procedure was utilized for the synthesis of 1. MOF with highest crystallinity was obtained after heating a reaction mixture (at 120 °C) containing 20 mg (0.05 mmol) of 2-(perfluorobenzamido)terephthalic acid linker and 13 mg (0.05 mmol) of ZrCl₄ in the presence of 115 μL (28 mmol) of trifluoroacetic acid as a modulator in 3 mL of DMF (N,N-dimethylformamide) medium in a glass tube for 24 h. After the completion of the scheduled time, the reaction mixture was allowed to cool isothermally to room temperature. A white colour solid was obtained after filtration followed by washing (with acetone) of the reaction mixture. Finally, the white solid was dried at 60 °C for 6 h. The obtained yield of 1 was 55 mg (0.01 mmol, 65% with respect to the metal salt used). Anal. calc. for C₁₀₈H₇₈F₃₀N₁₂O₄₅Zr₆ (3370 g mol⁻¹): C, 38.45; N, 4.98; H, 2.25. Found C, 38.34; N, 4.86; H, 2.12%.

Desolvation of 1

The unreacted metal salts, linkers, and the encapsulated solvent (DMF) molecules of the as-prepared MOF were removed by stirring 50 mg of 1 in excess (50 mL) dichloromethane for 20 h at room temperature. The solvent-exchanged solid was then collected via vacuum filtration and dried in a hot air oven. Finally, the remaining solvent molecules in the pores of the MOF were eliminated by heating the MOF under a high vacuum. Thus, we obtained a pure and completely guest-free MOF (1′) which was used for catalysis and sensing applications.

Results and discussion

Crystal morphology and elemental analysis of 1

The field emission scanning electron microscopy (FE-SEM) method was utilized to know the crystal morphology of 1. The obtained high-resolution image (Fig. S7, ESI†) displays nearly uniform-sized microcrystals with a circular shape. Such images demonstrate the crystalline nature and phase purity of 1′. The EDX elemental analysis (Fig. S8, ESI†) verified the presence of all the desired elements in 1′. Experimental evidence from powder X-ray diffraction (PXRD) further supported the phase purity and crystallinity of 1. The PXRD pattern of the newly synthesized MOF was quite similar to the previously published simulated framework of Zr-UiO-66 MOF (Fig. S6, ESI†).³⁶

ATR-IR spectroscopy

The successful formation of bonds between Zr(IV) and the carboxylate groups of the linker, as well as the complete removal of impurities after solvent exchange, were confirmed by the IR-spectroscopic analysis of the free linker, 1 and 1′ (Fig. S5, ESI†). The IR spectra of the linker displayed two characteristic peaks at 1585 and 1411 cm⁻¹, which originated from the asymmetric and symmetric stretching of free –COOH groups. These bands shifted to 1578 and 1387 cm⁻¹ in 1 and 1′. Such shifts in carboxylate stretching frequencies confirm the formation of coordination bonds between Zr(IV) and the carboxylate groups. One common peak at 1202 cm⁻¹ was observed in all three compouands. This peak is generated due to the stretching of the –C–F bonds. The carbonyl stretching of the amide functional group of the linker shows a strong peak at 1660 cm⁻¹ for 1 and 1′ which ensures the preservation of the functional group during the synthesis of MOF.

Structural description

Previously reported literature suggested that Zr(IV)/Hf(IV) salts can readily react with terephthalic acid linkers to construct UiO-66 framework structures.^36–38 Here, the experimentally obtained peak locations and intensities are closely similar to the theoretically obtained PXRD pattern of the aforementioned MOF. The indexed cell parameters and cell volume of the experimental PXRD profile are similar to the simulated framework (Table S1, ESI†). Again, the modest value of R_wp (2.6%) and R_p (1.8%) obtained after the Pawley refinement of the experimental PXRD pattern supported the structural closeness between the experimental and theoretical PXRD patterns (Fig. 1(a)). Similar to the pristine MOF, [Zr₆O₄(OH)₄]¹²⁺ clusters in 1′ exist as secondary building units that are interconnected by the (perfluorobenzamido)terephthalate linkers (Fig. 1(b) and (c)). Such interconnection results in two types of structural voids, i.e., larger (octahedral) and narrow (tetrahedral) voids, which are interconnected by triangular windows.

Fig. 1 (a) Pawley plot for the PXRD profile of 1′. (b) Cubic framework structure of 1′. (c) and (d) Structures of the free linker and ACF herbicide.

Physicochemical stability

A robust catalyst and user-friendly sensor must be stable in various hazardous conditions. Thus, the determination of the thermal and chemical stability of a catalyst is very necessary. To verify the chemical tolerance of 1′, firstly, it was stirred in various solvent media (CHCl₃, CH₂Cl₂, hexane, EtOAc, toluene, DMF, cyclohexane and 1 M HCl), and a wide range of acidic and basic pH solutions (pH = 2–10) for 24 h and the PXRD measurements of the recovered catalysts were executed. The collected PXRD patterns are summarized in Fig. S9 and S10 (ESI†), which suggests the stability of the framework in the above-mentioned media. The stability of 1′ in different solvent media creates a scope for its practical use in different solvent systems.

Along with chemical stability, determining the thermal stability of 1′ is an important experiment. Thus, we executed the TG analysis under an O₂ atmosphere at the temperature range of 30–700 °C with a continuous heating rate of 4 °C min⁻¹. From Fig. S11 (ESI†), it can be concluded that both 1 and 1′ were thermally stable up to 300 °C. In the TG analysis curve of 1, initially, in between 30–125 °C, 3.1% weight loss occurs due to the removal of 5 water molecules per formula unit. Within the temperature range of 125–300 °C, the loss of 6 DMF molecules resulted in a 12.5% loss of weight. At last, after 300 °C, the collapse of 1 begins because of the cleavage of the Zr–O bonds which resulted in the final weight loss from 1. The single-step weight loss after 300 °C was perceived in the case of 1′ due to the breakdown of Zr–O bonds of the MOF.

Surface area analysis

N₂ sorption experiment at −196 °C was performed to measure the surface area and pore volume of the newly constructed MOF. The sorption measurement resulted in type-I isotherms with a pore volume of 0.35 cc g⁻¹ for 1′ (Fig. S12, ESI†). The DFT method for obtaining the porosity distribution curve suggested that most pores of the MOF are centered at 11.5 Å, which implies the microporous nature of the material (Fig. S13, ESI†). The measured specific BET surface area of 1′ was 626 m² g⁻¹. The measured surface area was lesser than to the BET surface area of the unfunctionalized MOF (S_BET = 1147 m² g⁻¹).³⁹ The accessible pore widths of the MOF decreased due to the incorporation of bulky functional groups, limiting the free diffusion of N₂ molecules across the pores of the framework. The obtained surface area is comparable to that of other functionalized MOFs of the same class.^40,41 A CO₂ sorption experiment was also executed which showed 1′ having CO₂ uptake capability of up to 20 cc g⁻¹ at room temperature under the pressure of 1 bar (Fig. S14, ESI†).

Hydrophobicity measurement

Here, the pentafluorobenzene group was strategically incorporated into the structure of MOF to introduce hydrophobicity. The hydrophobicity of 1′ was again confirmed by the water contact angle (WCA) measurement. The obtained WCA is 157 ± 1° which implies a superhydrophobic nature of 1′ (Fig. S15, ESI†).

Catalytic cycloaddition of CO₂

High physicochemical stability and the presence of Lewis acidic Zr(IV) site and Lewis basic –F groups warrant studying the catalytic activity in the catalytic cycloaddition of CO₂ with epoxides. Initially, epichlorohydrin was selected as a model substrate and catalytic activity of 1′ was tested at 80 °C at 1 bar pressure of CO₂ in the presence and absence of tetrabutylammonium bromide (TBAB) as a co-catalyst. In the absence of a co-catalyst, no catalytic activity was observed (Table 1, entry 1 and Fig. S17, ESI†). The activity of 1′ with 4% co-catalyst loading was enhanced and >99% conversion of epichlorohydrin to the corresponding cyclic carbonate product with 100% selectivity was observed (Table 1, entry 2 and Fig. S18, ESI†). When the co-catalyst loading was reduced up to ten-fold (0.4%), the conversion of epichlorohydrin remained almost unchanged (>98%) (Table 1, entries 3–5 and Fig. S19–S21, ESI†). When we further decreased the co-catalyst loading, the percentage conversion of epichlorohydrin was decreased (Table 1, entry 6 and Fig. S22, ESI†). Hence, the rest of the study was continued with a co-catalyst loading of 0.4%. A series of catalytic experiments were performed at different temperatures to check the influence of temperature on the catalytic performance of 1′ (Table 1, entries 8, 9 and Fig. S24, S25, ESI†). Even at room temperature, the conversion of epichlorohydrin reached 27%. The maximum conversion of epichlorohydrin was observed at 80 °C (Fig. S16, ESI†), and hence, this temperature was considered as the optimum temperature. Moreover, the evolution of reaction progress between CO₂ and epichlorohydrin was monitored as a function of time by NMR spectroscopy. The highest conversion of epichlorohydrin was achieved after 24 h (Fig. 2). Therefore, the substrate scope experiments were performed up to 24 h. Furthermore, another control experiment was performed in the absence of 1′ to assess the rate of conversion of epichlorohydrin (other experimental parameters were kept similar). Only 52% conversion was observed (Table 1, entry 7 and Fig. S23, ESI†). These catalytic results establish a synergistic effect between 1′ and TBAB, which leads to the high yield of cyclic carbonates.

Table 1 Optimization of reaction conditions for catalytic cycloaddition of CO₂ with epichlorohydrin with 1′^a

Entry	Catalyst	Co-catalyst (loading)	Temperature (°C)	Conversion^b (%)
a Reaction conditions: epichlorohydrin (20 mmol), catalyst (15 mg), TBAB (in mol%), CO2 pressure (1 bar), 24 h. b Determined by ^1H-NMR.
1	1′	—	80	—
2	1′	TBAB (4%)	80	>99
3	1′	TBAB (1%)	80	>99
4	1′	TBAB (0.8%)	80	>99
5	1′	TBAB (0.4%)	80	98
6	1′	TBAB (0.1%)	80	98
7	—	TBAB (0.4%)	80	52
8	1′	TBAB (0.4%)	60	85
9	1′	TBAB (0.4%)	25	27
10	ZrCl4	—	80	—
11	H2L linker	—	80	—
12	ZrCl4	TBAB (0.4%)	80	68
13	H2L linker	TBAB (0.4%)	80	49
14	ZrCl4 + H2L linker	TBAB (0.4%)	80	66

---

Fig. 2 Time conversion plot for the cycloaddition of CO₂ with epichlorohydrin employing 1′ under optimized conditions (entry 5, Table 1).

The high catalytic activity of 1′ inspired us to further explore its catalytic activity under optimized conditions with other epoxides possessing different substituents in the side chain. Experimental observations showed that the conversion efficiency almost remained unaltered with smaller epoxides (epichlorohydrin and epibromohydrin) (Table 2, entries 1, 2 and Fig. S21, S31, ESI†). These catalytic results reveal that the pore size of 1′ fits perfectly with the size of the substrate. Moreover, this enables easy adsorption, diffusion, and desorption of the product from the porous catalyst. Hence, a higher conversion rate was obtained. With the increase in the alkyl side chain length, the percentage conversion of epoxides decreased (Table 2, entries 3–7 and Fig. S32–S36, ESI†). For instance, 1,2-epoxydodecane has a larger size compared to the pore of 1′, and diffusion of the substrate to the active site is strongly hindered by the bulky side chain. Hence, the rate of conversion also decreased. Interestingly, 100% selectivity to cyclic carbonate products was observed for all the substrates tested. Further, we also checked the catalytic efficiency of unfunctionalized and amino-functionalized MOFs under similar optimized conditions (Fig. S38–S39, ESI†). A comparatively lower rate of conversion was obtained. Hence, active participation of Lewis basic –F functionality in the catalytic cycle could be established. The cumulative results indicate that 1′ can act as a highly stable and active heterogeneous catalyst for ambient pressure solvent-free fixation of CO₂.

Table 2 Catalytic efficiency of 1′ toward cycloaddition of CO₂ with various epoxides^a

Entry	Catalyst	Co-catalyst	Epoxide	Conversion (%)
a Reaction conditions: epoxide (20 mmol), 1′ (15 mg), TBAB (0.4 mol%), CO2 pressure (1 bar), 80 °C, 24 h.
1	1′	TBAB (0.4%)		98
2	1′	TBAB (0.4%)		95
3	1′	TBAB (0.4%)		78
4	1′	TBAB (0.4%)		95
5	1′	TBAB (0.4%)		90
6	1′	TBAB (0.4%)		68
7	1′	TBAB (0.4%)		33

---

Moreover, a detailed comparison between the catalytic activity of 1′ towards the cycloaddition of CO₂ to epoxide ring was performed with other reported MOF catalysts in Table S9 (ESI†). A comparative literature survey reveals the superiority of the catalytic performance of 1′ over other reported catalysts in terms of the mild pressure and temperature required for its function.

Recyclability of cycloaddition reaction

To enhance the applicability of catalyst 1′, recyclability experiments were executed up to four consecutive cycles. The catalyst recovered from the 1^st cycle was suspended in 10 mL of methanol and left for 1 h. The catalyst was recovered by centrifuging the suspension at 10 000 rpm for 5 min. The process was repeated thrice and the recovered catalyst was dried at 80 °C for 6 h. Finally, it was activated under vacuum at 80 °C for 12 h. The activated catalyst was evaluated for the next catalytic cycle under a similar optimized condition with epichlorohydrin. With each cycle, a slight decrease in catalytic activity was observed (Fig. S42, ESI†). Up to the 4^th cycle of recyclability, 1′ had efficiently converted 82% (Fig. S37, ESI†) of epichlorohydrin to its corresponding cyclic carbonate. Furthermore, PXRD (Fig. S43, ESI†), ATR-IR (Fig. S44, ESI†), and FE-SEM (Fig. S45, ESI†), analysis were performed on the recycled 1′, and the similarity in results between fresh 1′ and recycled 1′ was assessed. Moreover, the obtained BET surface area of reused 1′ (598 m² g⁻¹) (Fig. S46, ESI†) and fresh 1′ are quite comparable. Furthermore, ICP-MS analysis was performed on the filtrate obtained after catalytic cycloaddition of CO₂ with epichlorohydrin. The results (Table S7, ESI† entry no. 1) revealed that the Zr(IV) ion concentration in the filtrate was insignificant, which nullified the possibility of the Zr(IV) ion leaching from the MOF during the catalytic cycle. These results suggest that structural integrity, crystallinity, and functionality were retained during the catalytic cycles.

Plausible mechanism of cycloaddition

Various control experiments and detailed characterizations were executed to better understand the catalyst's active sites and the cycloaddition mechanism. Initial observations showed that 1′ undergoes size-selective catalytic cycloaddition of the substrates. Smaller substrates have a higher conversion rate, but the rate gradually decreases with an increase in the alkyl chain length (Table 2). This observation indicates that the well-defined pores of 1′ play an important role in the catalytic cycle. Hence, catalytic cycloaddition steps occur inside the pores of 1′. Improved mass transfer of smaller substrates and CO₂ to the active site may increase leading to a higher conversion rate. During the recyclability test, parent and recycled catalysts showed similar FE-SEM images, PXRD, and IR spectral patterns (Fig. S43 and S45, ESI†). These results make it evident that no loss or breaking of functionality occurred during catalysis. Moreover, 1′ showed excellent recyclability with minimal loss in catalytic activity. Furthermore, the effects of fluorine functionality and Lewis acidic Zr(IV) metal and the co-catalyst TBAB were examined by various control reactions, where epichlorohydrin was considered the model substrate under the same optimized conditions. The presence of only ZrCl₄ salt or the linker molecule had a small effect on the cycloaddition process (Table 1, entries 10–11, Fig. S26 and S27, ESI†). Hence, no catalytic conversion was observed in NMR spectra. Moreover, the percentage of conversion obtained in the presence of only co-catalyst TBAB was 52%. With the addition of ZrCl₄ or linker or both of them, the conversion percentage showed little variations (Table 1, entries 12–14, Fig. S28–S30, ESI†). Hence, these results substantiate that the metal salt or the linker does not have any catalytic activity individually. These data indirectly prove that the highly heterogeneous, fluorine-rich, porous framework structure of 1′ plays an important role in adsorption and the activation of CO₂ during the catalytic cycle. Under similar optimized reaction conditions, the catalytic activity of un-functionalized and amino-functionalized MOFs was also found to be less (Fig. S38 and S39, ESI†). Hence, from these observations, it is clear that the presence of Lewis acidic (Zr(IV)) or Lewis basic (–F functionality) sites alone cannot initiate the cycloaddition reaction. Moreover, a well-defined three-dimensional framework, along with the –F functionality unveils an active role in the selective interaction and harvesting of CO₂ with the active site of the catalyst (Scheme 1). This statement was further supported by a hot-filtration experiment. We considered epichlorohydrin as a model substrate and studied cycloaddition experiments for up to 3 h. Under optimized conditions, only 45% of reactant conversion was observed (Fig. S40, ESI†). After that 1′ was removed by filtering out from the reaction medium and the reaction was continued for another 9 h. The observed conversion of the reactant was 53% (Fig. S41, ESI†). Hence, after the removal of 1′, the extent of the reaction was much less. Henceforth, synergistic cooperativity of 1′ with TBAB was established to achieve a higher rate of conversion.

Scheme 1 A plausible mechanism of fluorine functionalized porous 1′ MOF-mediated catalytic cycloaddition of epoxide and CO₂.

By taking the experimental outcomes and former literature reports into account,^42,43 we described a plausible mechanism for the cycloaddition in Scheme 1. In the first step, the epoxide will be diffused through the pores of 1′ and undergo weak reversible interaction with the Lewis acidic Zr(IV) sites of the octahedral oxo cluster, followed by adsorption and activation of CO₂ molecule in the near vicinity of the catalytic site by Lewis basic –F functionality. In the next step, the ring opening of the epoxide takes place via a nucleophilic attack by the Br⁻ ion of the TBAB co-catalyst on the less hindered site of the epoxide in an S_N² manner. Furthermore, coupling between activated CO₂ and the ring-opened epoxide causes the formation of an alkyl carbonate anion. In the last step, an intermolecular ring-closing reaction occurs via a nucleophilic attack by the alkyl carbonate anion on the most electrophilic β-carbon atom, followed by elimination of Br⁻. In the last step, reductive elimination of five-membered cyclic carbonate occurs to generate 1′ for the subsequent cycle.

Aerobic oxidation of hydrocarbons

As discussed earlier in the introduction, we aimed to investigate the catalytic performance of a superhydrophobic MOF constructed through the covalent attachment of a pentafluorobenzene unit to the organic linker of UiO-66 MOF through amino linkage (1′). This MOF certainly possesses high BET surface area and porosity which can favour the catalytic reaction. The catalytic performance of UiO-66-NH₂ (without the pentafluorobenzene unit) was tested under identical conditions for the aerobic oxidation of hydrocarbons to focus specifically on the influence of hydrophobic groups. To compare the activity between these two MOFs, we have selected a series of hydrocarbons, such as cyclooctane, ethylbenzene, diphenylmethane, fluorene, and indane, as model substrates using TBHP as a radical initiator and molecular oxygen as an oxidant in acetonitrile at 60 °C (Fig. S47–S58, ESI†). The observed catalytic data are summarized in Table 3. A series of control experiments indicated that poor conversions are observed in the absence of O₂ or TBHP. Hence, further experiments were conducted in the presence of O₂ and TBHP. The aerobic oxidation of cyclooctane using 1′ afforded 28% conversion with 95% selectivity to the respective ol/one mixture at 60 °C after 24 h. Under similar conditions, 15% of cyclooctane was converted to the corresponding ol/one mixture with 86% selectivity using UiO-66-NH₂. Then, via aerobic oxidation, ethylbenzene was oxidized to acetophenone with 10% conversion using 1′ as a solid catalyst, while poor conversion of ethylbenzene was observed with UiO-66-NH₂ under identical conditions. A similar trend was observed with diphenylmethane, where higher conversion (14%) was observed with 1′ as compared to UiO-66-NH₂ (6%). On the other hand, aerobic oxidation of fluorene led to the conversion of 17 and 9% using 1′ and UiO-66-NH₂, respectively, with complete selectivity to fluorenone under identical conditions. Finally, an effort to oxidize indane with 1′ exhibited 69% conversion, while UiO-66-NH₂ showed only 12% conversion under identical experimental conditions. Fig. 3 shows the time conversion plots for cyclooctane and indane using 1′ as a solid catalyst. These catalytic data clearly explain the strong influence of hydrophobic groups in 1′ to enhance the conversion of all tested substrates compared to UiO-66-NH₂ without possessing a hydrophobic environment. The presence of hydrophobic groups in 1′ favours facile penetration of hydrocarbons, thus providing a confined environment for the reactive species and hydrocarbons, promoting higher conversions. In contrast, UiO-66-NH₂ fails to provide such intimacy between reactive species and hydrocarbons. Thus, the presence of these hydrophobic groups in 1′ significantly increases the concentration of substrates near the hydrophobic sites in 1′. Furthermore, the observed catalytic data with 1′ are in good agreement with the previously reported hydrophobic MOFs.^31,32 Although reactivity may not be directly comparable due to the difference in the reaction conditions, such as substrates used, the present catalytic data prove the influence of hydrophobicity of 1′. In addition, the conversion of a substrate under these optimized conditions mainly depends on its kinetic dimension. The stability of the catalyst was assessed by performing reusability experiments. Thus, the solid catalyst was recovered after the reaction by centrifugation and washed with acetonitrile. Later, this was dried at 100 °C for 6 h and used for subsequent cycles. The conversion of cyclooctane was 28, 24, and 22 for the 1^st, 2^nd and 3^rd cycles, respectively. Moreover, the PXRD patterns, FE-SEM images, and IR spectra of the catalyst that had been reused twice supported the structural and functional groups integrity compared to the fresh solid catalyst (Fig. S59–S61, ESI†). ICP-MS analysis of the filtrate obtained after catalytic aerobic oxidation (Table S7, ESI† entry no. 2) revealed that the Zr(IV) ion concentration in the filtrate was negligible. These data support that the Zr(IV) ions did not leach from the MOF during the catalytic cycle. On the other hand, the BET surface area of the reused 1′ was 591 m² g⁻¹ (Fig. S62, ESI†), which is similar to that of the fresh 1′. The obtained surface area of the reused catalyst further supported the structural integrity of the catalyst during the aerobic oxidation reaction of cyclooctane.

Table 3 Aerobic oxidation of hydrocarbons using 1′ and UiO-66-NH₂ to their respective products^a

Sl. no.	Hydrocarbon	Catalyst	Conv.^b (%)	Selectivity^b (%)		Others (%)
				ol	one
a Reaction conditions: hydrocarbon (1 mmol), 1′ (10 mg), TBHP (15 μL), acetonitrile (2.5 mL), O2 purged through balloon, 60 °C, 24 h; hydrocarbon (1 mmol), UiO-66-NH2 (10 mg), TBHP (15 μL), acetonitrile (2.5 mL), 60 °C, O2 purged through balloon, 24 h. b Conversion and selectivity were determined by GC–MS. Selectivity to ol/one represents the corresponding alcohol and ketone.
1		1′	28	35	60	5
2		UiO-66-NH2	15	23	63	14
3		1′	10	—	100	—
4		UiO-66-NH2	2	—	100	—
5		1′	14	—	100	—
6		UiO-66-NH2	6	—	93	7
7		1′	17	—	100	—
8		UiO-66-NH2	9	—	100	—
9		1′	69	24	67	9
10		UiO-66-NH2	12	18	75	7

---

Fig. 3 Time conversion profile for the aerobic oxidation of (a) cyclooctane and (b) indane using 1′ as a solid catalyst. The error bars indicate the standard deviations obtained from three independent experiments.

Considering the observations on the aerobic oxidation of hydrocarbons using 1′, a reaction mechanism is proposed as shown in Scheme S2 (ESI†). The conversion of hydrocarbon (indane) was almost negligible using 1′ as a catalyst without TBHP, while the conversion of indane was appreciable with TBHP. Thus, the Lewis acidic sites (Zr(IV)) in 1′ activate TBHP to afford t-BuOO˙ radicals,^44,45 which subsequently abstracts hydrogen from indane to provide carbon radical (indanyl radical). This further reacts with molecular oxygen to give indanylperoxy radical, followed by hydrogen abstraction to provide indanehydroperoxide. Finally, this intermediate affords the expected ol/one mixture with the assistance of Zr(IV) in 1′.

Table S10 (ESI†) compares the activity of 1′ with other hydrophobic MOFs reported in the literature for the oxidation of hydrocarbons. Although this work tested the reactivity of five different hydrocarbons in the presence of 1′, the activity of 1′ may not be compared with previous catalytic data as the reaction conditions, such as the substrate, reaction temperature, and oxidants, are different. In addition, the number of studies on the use of hydrophobic MOFs in the aerobic oxidation of hydrocarbons is limited. In any case, the present work clearly illustrates the importance of hydrophobicity in achieving higher activity compared to non-hydrophobic MOFs. Furthermore, this observation is in good agreement with previous reports.^28,29

Fluorescence sensing of ACF by 1′

High physicochemical stability and excellent fluorescence signal output of 1′ motivated us to detect ACF in environmental water bodies. Initially, the MOF suspension displayed strong fluorescence intensity in the sensing medium when it was excited with a wavelength of 350 nm (Fig. S63, ESI†). A dramatic quenching in fluorescence was observed after a gradual increase in the concentration of ACF (from 0 μL to 300 μL) (Fig. 4). There was around 98% quenching in the initial fluorescence of the MOF within 5 s of the addition of 300 μL (10 mM) solution of ACF. Hence, it can be concluded that the response time of the MOF to sense ACF is 5 s (Fig. 5(a) and (b)). Moreover, the ACF sensing experiment was also conducted by using 1 as a sensor under identical sensing conditions. The sensing experiment indicated that ACF can quench up to 96% of the fluorescence intensity of 1 within 5 s of the addition of 300 μL of 10 mM ACF (Fig. S75, ESI†). Therefore, it can be concluded that both 1 and 1′ have similar efficiency to sense ACF.

Fig. 4 Diminution in fluorescence of 1′ with increasing concentration of ACF.

Fig. 5 (a) Changes in fluorescence intensity of 1′ as a function of time. (b) Time-dependent fluorescence saturation plot (the standard deviations of three individual measurements are displayed as the red error bars). The luminescence colour change of 1′ under UV-lamp after injection of ACF is displayed in the inset of Fig. 5(b).

To verify the potential applicability of this sensor, the selectivity of 1′ to detect ACF was evaluated in the presence of various metal ions and herbicides that are commonly present in real samples. The results demonstrated that 1′ exhibited high selectivity for ACF sensing (Fig. 6(a) and Fig. S64–S74, ESI†). However, the addition of 300 μL of ACF solution in the presence of other competitive analytes resulted in similar quenching of fluorescence of the MOF (Fig. 6(b)). Therefore, 1′ demonstrates a strong preference for detecting ACF, even in the existence of its competitive congeners (Fig. S64–S74, ESI†). The error bars in the bar plots depict standard deviation values of three individual measurements which indicate the reproducibility of these measurements. The accuracy of sensing in the intra-day and inter-day (on three distinct days), precision (Table S6, ESI†), and the error of all the measurements (Table S6, ESI†) were assessed to validate the originality and repeatability of the measurements.

Fig. 6 (a) Alteration in emission response of 1′ after exposure to various analytes for ACF sensing. (b) Reduction in luminescence of 1′ following the inclusion of ACF in the presence of diverse interfering analytes (the red bars represent the standard deviations of three distinct observations).

The determination of the Stern–Volmer (S–V) quenching constant (K_sv) is crucial to understanding the sensing mechanism responsible for the quenching of fluorescence in the presence of ACF. The determined K_sv also provides valuable information on the sensitivity of the probe for targeted analytes. Here, the K_sv value was computed using the equation: I₀/I = K_sv[Q] + 1, where I₀ and I represent the emission intensity of 1′ in the presence and absence of the ACF, respectively. [Q] signifies the concentration of the analyte. For ACF, the calculated K_sv value is remarkably high (5.38 × 10⁶ M⁻¹) (Fig. S76 and S77, ESI†). This high K_sv value indicates a strong affinity and sensitivity of 1′ for ACF.

The assessment of a sensor's performance heavily depends on its LOD value. Therefore, calculating the LOD of a sensor is very important. To determine the LOD of 1′ for the sensing of ACF, initially, the standard deviation (σ) of luminescence intensities emitted by 1′ in the absence of any analyte was calculated. Subsequently, a series of fluorescence titration experiments were conducted by progressively diluting the ACF solution. Plotting the concentrations of ACF against the resulting fluorescence emission intensities yielded a linear relationship with a negative slope (k) (Fig. S78, ESI†). Finally, the LOD value was computed using the formula, 3σ/k. The computed LOD value of the MOF to sense the ACF is 1.1 ± 0.2 nM. This LOD value is significantly lower than all the other sensors of ACF (Table S8, ESI†). Additionally, the limit of quantification for this measurement was 3.8 ± 0.6 nM. The average regression equation for the LOD curve was calculated as 15871.6x + 427206.7 with an outstanding average regression coefficient (R²) of 0.996 (Table S4, ESI†). All the statistical results supported the excellent sensitivity of the probe to sense ACF.

A reliable, cost-effective sensor must maintain its quenching effectiveness even after undergoing multiple operational cycles. In this context, we conducted the recyclability test of 1′ for the sensing of ACF up to five cycles. After each cycle of sensing, the MOF material was thoroughly washed with methanol and activated. The quenching efficiency of the probe remained virtually unchanged even after the fifth cycle of sensing (Fig. S79, ESI†). This experiment conclusively demonstrated that 1′ can be effectively recycled to sense ACF. The fluorescence lifetime measurements of the sensor before and after the treatment of ACF were also executed. The obtained lifetime of the MOF before and after the sensing of ACF were 4.6 and 2.4 ns, respectively (Fig. S80 and Table S2, ESI†).

Sensing of ACF in environmental water samples

The main aim of the development of this sensor is to monitor the concentration of ACF in environmental aquatic bodies. The extraordinarily low LOD, high selectivity, and reusability for five times without loss of efficiency inspired us to conduct the sensing experiment in environmental water samples (e.g., lake, river, and tap water). The sensing results are summarised in Fig. S81 (ESI†), which shows that the MOF is equally efficient in determining the presence of ACF in various environmental waters. With regular variation in the concentration of ACF, the quenching efficiency of the MOF varied systemically. Moreover, the concentration of ACF in environmental water was also quantified. The observed ACF concentrations closely mirrored the known (spiked) concentrations of ACF (Table S5, ESI†). Furthermore, the recovery percentages of ACF in various water samples were ∼100%. These findings unequivocally support the ability of 1′ to accurately sense ACF concentrations in real aquatic systems.

Sensing of ACF in various pH media

The pH of environmental water may vary widely in different sources of water. Therefore, to validate the broad-scope effectiveness of 1′ to sense ACF in various pH media, sensing experiments were executed in the pH range of 2–12. For this experiment, 300 μL of ACF solution (10 mM) was injected in the suspension of MOF in various pH media. The quenching efficiency of the MOF remained intact in all the tested pH media (Fig. S82, ESI†). Therefore, the experiment confirms the sensing ability of the MOF in different pH media.

Sensing of ACF in fruits and vegetable extracts

ACF is a commonly used herbicide in agricultural fields. Therefore, the detection of ACF in various fruit and vegetable extracts is very important from a food safety point of view. Thus, we executed a sensing study of ACF in the extract of carrot, pointed gourd, cabbage, and orange. For this experiment, 300 μL of a 10 mM solution of ACF-spiked extract of the aforementioned fruits and vegetables was added to the suspension of MOF. The fluorescence intensity of the MOF diminished rapidly after the addition of the ACF-spiked extract of fruits and vegetables. The obtained quenching efficiencies of all the samples were above 97%, which suggests the extraordinary sensing capability of the MOF in the above-mentioned extracts (Fig. S83, ESI†). This observation also enlarges the scope of application of this sensor for vegetable and fruit safety purposes.

MOF@starch@cotton composite-based sensing of ACF

The dependence on using costly instruments and professional and technical experts can be overcome by developing an on-site paper strip-like device. A portable, cost-effective device is suitable for the on-field detection of ACF. The impressive reusability of MOF with its consistent efficiency inspired us to develop a 1′@starch@cotton composite for rapid, on-field detection of ACF. Utilizing cotton as the composite material provides the necessary mechanical strength for sustainability and reusability experiments, which are typically challenging to perform with traditional paper strip-based composites. In previous studies, paper strip-based composites were fabricated by the physisorption of the sensor on filter paper.^37,46 There was a serious issue of leaching of sensors during sensing which eventually hampers the effectiveness of a reusable sensor. The adherence of a bio-polymeric support (starch) can prevent the possibility of leaching. Starch not only acts as a barrier against leaching but also enhances the mechanical durability of the composite.

The details of the preparation procedure of the composite are discussed in the ESI† of this manuscript. After synthesis, the successful immobilization of 1′ on the fabric surface was confirmed by PXRD, IR, and FE-SEM analysis of the cotton fabric (Fig. S84–S86, ESI†). Next, various concentrations of ACF were injected into the composites, and after drying, they were placed inside a UV chamber. It was observed that the fluorescence intensity of the composites was quenched to achieve up to the nanomolar concentrations of ACF (Fig. 7). These observations ensure that the composite is capable of detecting the presence of ACF at nanomolar concentrations. Additionally, this robust composite can be reused up to three times without a significant loss in its detection efficiency (Fig. S87, ESI†).

Fig. 7 Digital images of 1′@starch@cotton composites under UV-lamp after treatment with different concentrations of ACF solutions.

Mechanistic interpretation of ACF sensing by 1′

The most likely sensing mechanism responsible for the selective sensing of ACF by 1′ was investigated with the help of various instrumental data and theoretical simulation results. Firstly, the structural integrity of the framework during the sensing experiment was confirmed by the PXRD pattern, IR spectra, FE-SEM images, and BET surface area measurements (S_BET = 605 m² g⁻¹) of the recovered MOF after the sensing experiment (Fig. S88–S91, ESI†). ICP-MS analysis of the filtrate obtained after sensing ACF (Table S7, ESI† entry no. 3) revealed that the Zr(IV) ion concentration in the filtrate was insignificant. This data nullifies the possibility of leaching of Zr(IV) ion from the MOF during the sensing experiment. These observations ruled out the possibility of disintegration of the MOF structure during the sensing process. Therefore, the collapse of the MOF structure during the sensing event did not account for the observed fluorescence change. The sensor's remarkable five-fold recyclability with minimal loss of efficiency prompted us to explore non-reaction-based sensing mechanisms.

Numerous phenomena have been documented in the literature, including ground-state complexation (GSC), photoinduced electron transfer (PET), the inner-filter effect (IFE), and Förster resonance energy transfer (FRET), all of which could be accounted for the diminution of fluorescence of 1′ after the addition of ACF in the sensing medium. Based on the change in the lifetime of the excited state, all the above-mentioned phenomena can be categorized into two groups: dynamic and static change in fluorescence. If the fluorophore's lifetime changes after the introduction of an external analyte, the process is considered dynamic, whereas if it remains unchanged, the process is deemed to be static. In the case of IFE and GSC processes, fluorescence lifetime generally remains constant. However, for PET and FRET processes, alterations in lifetime are typically observed.

In the present case, changes in the lifetime of the MOF after the addition of ACF were observed. The lifetime of the MOF dropped significantly from 4.6 ns to 2.4 ns after the addition of ACF in the sensing medium (Fig. S80 and Table S1, ESI†). Such decrease in lifetime values cancelled the possibility of IFE and GSC.

We further investigated the possibility of GSC between the MOF and ACF by examining the solid-state ATR-IR and solid-state UV-Vis spectra of 1′ before and after its treatment with ACF (Fig. S92 and S89, ESI†). Significant changes in the solid-state UV-Vis and IR spectra before and after treatment would be expected for GSC. However, the negligible changes in the solid-state UV-Vis and IR spectra of the analyte-treated 1′ dismissed the possibility of GSC in the sensing medium after the addition of ACF.

The interference of IFE in the fluorescence of a fluorophore can be verified by the shifting of the excitation wavelength of the fluorophore. If IFE is responsible for the quenching of fluorescence of the fluorophore, the quenching efficiency must be decreased after shifting the excitation wavelength of the fluorophore to a higher wavelength region. In the present case, the quenching efficiency remains similar even after shifting the excitation wavelength of the MOF to 390 nm from 350 nm (Fig. S93, ESI†). This experiment provided evidence that IFE is not responsible for the quenching of fluorescence of 1′ in the presence of ACF.

Here, the quenching occurs in a dynamic way which opens the possibility of both FRET and PET mechanisms. PET is a process involving the transfer of excited state electrons, in which the excited state electron of one molecule transfers to the low-lying vacant orbital of another molecule. Here, the quenching of fluorescence of the MOF occurred in the presence of ACF. Therefore, for a feasible PET process, electron transfer should occur from the lowest unoccupied molecular orbital (LUMO) of MOF to the LUMO of ACF. The theoretical location of the highest occupied molecular orbital (HOMO) and LUMO of the linker of the MOF and ACF indicate that the location of the LUMO of ACF is much higher as compared to that of the MOF (function B3LYP and Pople diffuse basis set 6-31G++(d, p) were utilized for all the calculations) (Fig. S94, ESI†). Therefore, the transfer of electrons from the LUMO of MOF to the LUMO of ACF is not energetically favourable. Therefore, PET is not the actual reason for quenching of fluorescence of the MOF in the existence of ACF.

To elucidate the possibility of FRET, we acquired UV-Vis absorption spectra for all tested analytes. Notably, the absorption spectrum of ACF displayed significant overlap with the emission spectrum of the MOF (Fig. S95, ESI†). This observation suggests that FRET could be a probable mechanism for energy transfer from the electron-rich MOF to the electron-deficient ACF; the change in the lifetime values also supports this hypothesis. The red shifting of the emission maxima of 1′ after the addition of ACF indicates the formation of a short-lived weakly interactive species by the close association of ACF and the MOF (required for the FRET process). The formation of such a species in the sensing medium by the close association between the MOF and ACF is responsible for the shift of emission maxima of 1′ by 35 nm.⁴⁷ Moreover, no other tested analyte's absorption spectrum displays any significant overlap with the emission maximum of the MOF. This could be the reason for the tremendous selectivity in ACF sensing by 1′ against the other competing analytes.

Conclusions

This study reports the solvothermal synthesis and physicochemical characterization of a pentafluorobenzene-functionalized, aqua-stable Zr-MOF and its potential applications to tackle three different environmental issues. This is the first ever reported Zr-MOF where the Lewis basic nature of the –F groups and Lewis acidic Zr(IV) sites were utilized for the mild pressure chemical fixation of one of the major contributors to greenhouse gas (CO₂) with excellent conversion efficiency and reusability. The size-selective catalytic activity, high conversion rate and wide substrate scope are the distinctive features of this fluorine-rich catalyst (1′). The hydrophobic cavity of 1′ was also successfully employed for the selective diffusion of hydrocarbons in the aerobic oxidation of hydrocarbons. The conversion of hydrocarbons was relatively higher with 1′ compared to unfunctionalized MOF due to the hydrophobic nature of the former solid. High ol/one selectivity with wide substrate scopes was achieved under mild reaction conditions using this hydrophobic catalyst. Moreover, the photoluminescence property of the MOF was utilized to monitor the presence of ACF in environmental wastewater, various fruits and vegetable extracts, and different pH media. This is the first ever reported MOF-based fluorescence probe for the detection of ACF. This sensor for ACF exhibited exceptional selectivity, reusability, sensitivity (LOD = 1.1 nM and K_sv = 5.38 × 106 M⁻¹) and rapid response (<5 s). The calculated LOD value of this MOF is the lowest of all the reported sensors for ACF. Additionally, an eco-friendly, bio-polymer grafted, portable 1′@starch@cotton composite was fabricated for the on-site, nanomolar detection of target herbicide. The mechanistic processes underlying both catalysis and ACF sensing were systematically investigated using appropriate analytical techniques and theoretical simulations. Overall, this MOF holds promise as an efficient catalyst for the chemical conversion of two major pollutants into valuable products and a highly effective sensor for an environmental pollutant herbicide, ACF, with potential applications in monitoring water pollution and food quality.

Author contributions

SG, SM, VK, and PB performed the experimental work. The manuscript was written by the contributions of all the authors, and the final version was submitted with their approval.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This project obtained funding from SERB grants, CRG/2021/000080 and EEQ/2021/000013. SG and SM are thankful to PMRF for the financial support. A. D. is the beneficiary of a grant from María Zambrano in Universitat Politècnica de València within the framework of the grants for retraining in the Spanish university system (Spanish Ministry of Universities, financed by the European Union, NextGeneration EU).

Notes and references

1. C. Anastasi, R. Hudson and V. J. Simpson, Energy Policy, 1990, 18, 936–944.
2. T. K. Pal, D. De and P. K. Bharadwaj, Coord. Chem. Rev., 2020, 408, 213173.
3. S. Singh Dhankhar, B. Ugale and C. Nagaraja, Chem. – Asian J., 2020, 15, 2403–2427.
4. H. Luo, J. Ren, Y. Sun, Y. Liu, F. Zhou, G. Shi and J. Zhou, Chin. Chem. Lett., 2023, 34, 107782.
5. R. Das, S. S. Dhankhar and C. Nagaraja, Inorg. Chem. Front., 2020, 7, 72–81.
6. L. Guo, K. J. Lamb and M. North, Green Chem., 2021, 23, 77–118.
7. F. Guo and X. Zhang, Dalton Trans., 2020, 49, 9935–9947.
8. P.-Z. Li, X.-J. Wang, J. Liu, J. S. Lim, R. Zou and Y. Zhao, J. Am. Chem. Soc., 2016, 138, 2142–2145.
9. H. Chen, T. Zhang, S. Liu, H. Lv, L. Fan and X. Zhang, Inorg. Chem., 2022, 61, 11949–11958.
10. P. M. Bhatt, Y. Belmabkhout, A. Cadiau, K. Adil, O. Shekhah, A. Shkurenko, L. J. Barbour and M. Eddaoudi, J. Am. Chem. Soc., 2016, 138, 9301–9307.
11. J. Tong, L. Bo, X. Cai, H. Wang, Q. Zhang and L. Su, Ind. Eng. Chem. Res., 2014, 53, 10294–10300.
12. A. Dhakshinamoorthy, A. M. Asiri, J. R. Herance and H. Garcia, Catal. Today, 2018, 306, 2–8.
13. A. Dhakshinamoorthy, A. M. Asiri and H. Garcia, Trends Chem., 2020, 2, 454–466.
14. A. Dhakshinamoorthy, A. López-Francés, S. Navalon and H. Garcia, ChemCatChem, 2022, 14, e202201036.
15. A. Dhakshinamoorthy, A. M. Asiri and H. Garcia, Chem. – Eur. J., 2016, 22, 8012–8024.
16. A. Dhakshinamoorthy, A. Santiago-Portillo, A. M. Asiri and H. Garcia, ChemCatChem, 2019, 11, 899–923.
17. W.-L. He, X.-L. Yang, M. Zhao and C.-D. Wu, J. Catal., 2018, 358, 43–49.
18. J. Long, H. Liu, S. Wu, S. Liao and Y. Li, ACS Catal., 2013, 3, 647–654.
19. A. Santiago-Portillo, S. Navalón, F. G. Cirujano, F. X. L. I. Xamena, M. Alvaro and H. Garcia, ACS Catal., 2015, 5, 3216–3224.
20. A. Dhakshinamoorthy, M. Alvaro and H. Garcia, Chem. – Eur. J., 2011, 17, 6256–6262.
21. N. Nagarjun and A. Dhakshinamoorthy, ChemistrySelect, 2018, 3, 12155–12162.
22. P. Cancino, A. Vega, A. Santiago-Portillo, S. Navalon, M. Alvaro, P. Aguirre, E. Spodine and H. García, Catal. Sci. Technol., 2016, 6, 3727–3736.
23. A. Dhakshinamoorthy, M. Alvaro and H. Garcia, J. Catal., 2012, 289, 259–265.
24. N. Nagarjun and A. Dhakshinamoorthy, New J. Chem., 2019, 43, 18702–18712.
25. G. Ji, L. Zhao, Y. Wang, Y. Tang, C. He, S. Liu and C. Duan, ACS Catal., 2022, 12, 7821–7832.
26. M. Caruso, S. Navalón, M. Cametti, A. Dhakshinamoorthy, C. Punta and H. García, Coord. Chem. Rev., 2023, 486, 215141.
27. L. Li, Q. Yang, S. Chen, X. Hou, B. Liu, J. Lu and H.-L. Jiang, Chem. Commun., 2017, 53, 10026–10029.
28. J. Ji, F. Liu, W. Yang, M. Tan, W. Luo and S. F. Yin, ChemCatChem, 2020, 12, 4331–4338.
29. C. Gogoi, N. Nagarjun, S. Roy, S. Mostakim, D. Volkmer, A. Dhakshinamoorthy and S. Biswas, Inorg. Chem., 2021, 60, 4539–4550.
30. D. J. Xiao, J. Oktawiec, P. J. Milner and J. R. Long, J. Am. Chem. Soc., 2016, 138, 14371–14379.
31. J. Sui, M.-L. Gao, B. Qian, C. Liu, Y. Pan, Z. Meng, D. Yuan and H.-L. Jiang, Sci. Bull., 2023, 68, 1886–1893.
32. W. Li, Z. Li, H. Zhang, P. Liu, Z. Xie, W. Song, B. Liu and Z. Zhao, Proc. Natl. Acad. Sci. U. S. A., 2023, 120, e2206619120.
33. P. K. Gopi, B. Mutharani, S.-M. Chen, T.-W. Chen, G. E. Eldesoky, M. A. Ali, S. M. Wabaidur, F. Shaik and C. Y. Tzu, Ecotoxicol. Environ. Saf., 2021, 207, 111285.
34. B. He, Y. Hu, W. Wang, W. Yan and Y. Ye, Agronomy, 2022, 12, 2792.
35. C.-C. Hong, M. Shimomura-Shimizu, M. Muroi and K.-I. Tanamoto, Biol. Pharm. Bull., 2004, 27, 1136–1139.
36. L. Valenzano, B. Civalleri, S. Chavan, S. Bordiga, M. H. Nilsen, S. Jakobsen, K. P. Lillerud and C. Lamberti, Chem. Mater., 2011, 23, 1700–1718.
37. S. Ghosh, A. Rana, S. Kumar, C. Gogoi, S. Mukherjee, U. Manna and S. Biswas, Mater. Chem. Front., 2022, 6, 2051–2060.
38. S. Mukherjee, S. Ghosh and S. Biswas, ACS Appl. Nano Mater., 2023, 6, 22231–22240.
39. C. Piscopo, A. Polyzoidis, M. Schwarzer and S. Loebbecke, Microporous Mesoporous Mater., 2015, 208, 30–35.
40. S. Ghosh, D. Mal, S. Mukherjee and S. Biswas, ACS Sustainable Chem. Eng., 2023, 11, 13179–13186.
41. S. Mukherjee, K. Sarkar and S. Biswas, Dalton Trans., 2023, 52, 5597–5605.
42. N. Wei, R.-X. Zuo, Y.-Y. Zhang, Z.-B. Han and X.-J. Gu, Chem. Commun., 2017, 53, 3224–3227.
43. H. Xu, B. Zhai, C.-S. Cao and B. Zhao, Inorg. Chem., 2016, 55, 9671–9676.
44. O. A. Kholdeeva, Catal. Today, 2016, 278, 22–29.
45. N. V. Maksimchuk, K. A. Kovalenko, V. P. Fedin and O. A. Kholdeeva, Chem. Commun., 2012, 48, 6812–6814.
46. S. Mukherjee, S. Ghosh and S. Biswas, Inorg. Chem. Front., 2022, 9, 6288–6298.
47. R. B. Sekar and A. Periasamy, J. Cell Biol., 2003, 160, 629.

---

Footnotes

† Electronic supplementary information (ESI) available: NMR spectra, PXRD patterns, IR spectra, EDX spectra, TGA curves, FE-SEM images, N₂ and CO₂ sorption isotherms, catalytic results, sensing plots, fluorescence lifetime plots, recyclability, UV-vis spectrum, comparison tables and statistical results. See DOI: https://doi.org/10.1039/d3tc04140a

‡ Equal contribution.

---