Multiphase Porous Electrochemical Catalysts Derived from Iron-Based Metal-Organic Framework Compounds.

Herbicide use has attracted attention recently due to potential damage to human health and lethality to the honey bees and other pollinators. Fenton reagent treatment processes can be applied for the degradation of herbicidal contaminants from water. However, the need to carry out the normal Fenton reactions under acidic conditions often hinders their practical application for pollution control. Herein, we report on the synthesis and application of multiphasic porous electro-Fenton catalysts prepared from calcinated metal-organic framework compounds, CMOF@PCM, and their application for the mineralization of herbicides in aqueous solution at circum-neutral pH. CMOF nanoparticles (NPs) are anchored on porous carbon monolithic (PCM) substrates, which allow for binder-free application. H2O2 is electrochemically generated on the PCM substrate which serves as a cathode, while ·OH is generated by the CMOF NPs at low applied potentials (-0.14 V). Results show that the structure and reactivity of the CMOF@PCM electro-Fenton catalysts are dependent on the specific MOF precursor used during synthesis. For example, CMIL-88-NH2, which is prepared from MIL-88(Fe)-NH2, is a porous core-shell structured NP comprised of a cementite (Fe3C) intermediate layer that is sandwiched between a graphitic shell and a magnetite (Fe3O4) core. The electro-Fenton production of hydroxyl radical on the CMOF@PCM composite material is shown to effectively degrade an array of herbicides.

Introduction

Due to extensive application of herbicides, they are now considered to be pseudopersistent pollutants.[1] Among the widely used herbicides, glyphosate is considered to be carcinogenic.[2,3] Furthermore, several herbicides are reported to impact pollinators resulting in a dramatic reduction of bee populations worldwide.[4−7] Herbicides are also suspected of causing harm to crustaceans, which are an important component of freshwater and oceanic food webs.[8] In spite of these growing concerns, there is little research on their removal from the aquatic environment compared to other contaminants of emerging concern (CECs).

Advanced oxidation processes (AOP) including electrochemical oxidation are suitable for herbicide degradation and removal from contaminated water.[9−11] However, electrochemical degradation of organic compounds is often energy intensive with high capital expenditure (CAPEX) costs. This is due to use of expensive material such as platinum group metals (PGM) or boron-doped diamonds as anodes or in the case of platinum as cathodes. In contrast, alternative oxidative approaches employing Fenton-reagent reactions (H2O2 + Fe(II) → ·OH + Fe(III) + OH–) utilize iron, an earth abundant metal.[12,13] Fenton-like systems can achieve complete degradation of some of the most persistent contaminants such as perfluorinated surfactants and perfluorotelomer alcohols.[14,15] However, a major drawback of using the Fenton process is that above pH 5, Fe(II) is readily oxidized to Fe(III) in the form of Fe(OH)3 by dissolved oxygen, thus limiting the pH range to carry out the overall reactions. Electro-Fenton reaction, on the other hand, overcomes this problem by reducing Fe(III) to Fe(II) at cathode.[16,17] Unfortunately, Fe(II) concentrations as high as 0.01 M need to be applied for the effective removal of contaminants.[15]

We have shown previously that a composite Fe-NP and nitrogen codoped graphite material can be used as an electrode for heterogeneous electro-Fenton oxidation.[18] This approach only required a loading level of 0.36 wt % Fe in order to maximize ·OH production. However, the percentage of Fe that was accessible on the surface was limited.[18] In order to explore the full potential of a heterogeneous electro-Fenton process, a porous electrode where all potential Fe sites are accessible is desired. Given this objective, inorganic porous nanomaterials derived from MOFs, or carbonized MOFs (CMOFs), appear to provide a solution that would achieve this. For example, metal NPs porous catalysts derived from MOFs have been shown to be useful for organic synthesis.[19]

In order to be useful in practical applications, the electro-Fenton catalysts must also meet the following criteria: (1) the metal nanocatalyst supported on a substrate must be well dispersed in order to allow for sufficient contact of electrolyte and substrate with the active catalyst; (2) the metal catalyst must form a transient bond with the substrate; and (3) both the substrate and nanoparticulate metal catalysts must be attached to a highly porous material. The first criterion allows the utilization of the electrochemically active nanosized catalyst. The third criterion allows for the construction of 3D electrodes that have form-factor advantages, which can maximize the electrode surface to electrolyte volume ratio during application.

Herein, we use binder-free multiphase porous electro-Fenton catalysts made from pyrolyzed MOFs for the electrochemical degradation of an array of herbicides. The resulting catalysts are anchored onto porous carbon monoliths (PCM), which are denoted as CMOF@PCMs. In contrast to a bulk-phase catalyst, the narrow distribution of the pores within the CMOF@PCMs allows for facile access to the active Fe(II)/Fe(III) catalytic centers. In addition, the PCM substrate allows in situ generation of hydrogen peroxide (H2O2) that is needed for the electro-Fenton reaction. The CMOF@PCMs are shown to be efficient at removing herbicides at neutral pH while Fe leaching is negligible. Furthermore, we discovered the resulting structures of CMOFs are independent of the topology of parent MOF. But the resulting structure appears to depend on the specific functional group and on the identity of the organic linker. A core–shell structured CMOF@PCM is obtained from a MIL-88(Fe)–NH2 precursor. The results provided herein provide insights into the development of stable and porous materials for use as sustainable electro-Fenton catalysts.

Materials and Methods

Materials

All chemicals were obtained from Sinopharm Chemical Reagent Co. Ltd. and were used without further purification. Solvents were HPLC grade unless otherwise stated. Pesticides with chemical purity ≥97.0% (Sinopharm Chemical Reagent Co. Ltd.) used in this study are listed in Table S1.

Preparation of Porous Carbon Substrate

The PCM substrate was synthesized based on a previously reported procedure.[20] Briefly, resorcinol (3.0 g, 27.3 mmol) and Pluronic F127 (1.25 g) were dissolved in a solvent mixture of ethanol (11.4 mL) and deionized water (9 mL). 1,6-Diaminohexane (0.078 g, 0.67 mmol) was added to the above solution and stirred for 30 min. Subsequently, formalin (4.42 g, 54.5 mmol) was added. The reaction mixture was stirred for 40 min before being sealed and heated at 90 °C for 4 h. The polymer monolith, poly(benzoxazine-co-resol), was dried at 90 °C for 48 h. Pyrolysis was performed at 800 °C for 2 h under a nitrogen protection. Crack-free PCM was obtained after cooling down to room temperature.

Preparation of MOFs

A microwave-assisted synthesis of Fe-MOFs was based on a similar method using microwave synthesizer (CEM, Discover SP).[21] In brief, FeCl3 (0.88 g, 0.54 mmol) and terephthalic acid (0.9 g, 0.54 mmol) were dissolved in N,N-dimethylformamide (DMF) (15 mL) inside a 30 mL microwave tube. The reaction was carried at 150 °C for predetermined time before being cooled to room temperature. The nanocrystals of MIL-88(Fe) were separated by centrifugation at 11 000 rpm and washed with DMF and methanol before drying in vacuo at 150 °C. MIL-100(Fe), MIL-101(Fe), MIL-88(Fe)–NH2, and MIL-101(Fe)–NH2 were synthesized using the same method.

Preparation of CMOF@PCM

PCM was mixed with the Fe-MOF powder at a given mass ratio. The powdered MIL-88(Fe)@PCM was carbonized at 600 °C for 0.5 h under a N2 atmosphere to produce CMIL-88@PCM. The same procedure was applied for the preparation of the other CMOF@PCMs using the other MOFs precursors. The CMOF@PCMs used in the current study were designated as CMIL-88@PCM, CMIL-100@PCM, CMIL-101@PCM, CMIL-88-NH2@PCM, and CMIL-101-NH2@PCM for the hybrid materials prepared from MIL-88(Fe), MIL-100(Fe), MIL-101(Fe), MIL-88(Fe)-NH2, and MIL-101(Fe)-NH2, respectively. Fe-MOF powder was replaced with Fe3O4 nanoparticle to produce Fe3O4@PCM using similar method.

Preparation of Electrodes

Evaluation of electrocatalytic property of CMOF@PCMs required identical geometrical surface areas for the working electrode. As the result, rigid carbon paper was coated with these electrocatalysts and used as the working electrode in order to ensure identical electroactive areas. The ground CMOF@PCM substrates (50 mg) were suspended in 4.47 mL of milli-Q water, 1.5 mL of isopropyl alcohol, and 100 μL of a Nafion solution. The resulting ink was spray-coated onto carbon paper (HCP030P Hesen) at a catalyst loading of 0.6 mg/cm2. Both sides of carbon paper were uniformly coated (with total effective area of 8 cm2) to avoid exposing the supporting carbon paper substrate to the electrolyte.

Characterization of Catalysts

The surface morphology of the PCM substrate was examined using a Hitachi SU8010 field-emission scanning electron microscope (FE SEM), while the surface morphologies and elemental compositions of the CMOF@PCMs were examined using a ZEISS Merlin FE SEM equipped with an Oxford INCA X-MAX50 energy dispersive spectrometer (EDS). TEM micrographs and TEM EDS measurements were obtained using a FEI Tecnai G2 F20 (200 kV) transmission electron microscope (TEM). X-ray powder diffraction patterns were collected using a Bruker D8 A25 Advance diffractometer with Cu–Kα radiation (λ = 1.5418 Å). The element composition of catalysts was characterized by X-ray photoelectron spectroscopy (XPS) using Surface Science Instruments Thermo Scientific ESCALAB 250Xi surface spectrometer. Monochromatic Al–Kα radiation (1486.6 eV) was used.

Electrochemical Experiments

Electrochemical experiments were performed using a CH CHI660E C17171 potentiostat with a three-electrode beaker cell at room temperature. A platinum (Pt) wire and saturated calomel electrode (SCE) were used as the counter electrode and the reference electrode, respectively (Figure S1). Recorded potentials were converted to the reversible hydrogen electrode (RHE) scale (RHE = SCE + 0.059 × pH + 0.241 V). Cyclic voltammetry (CV) was conducted in Ar- or O2-saturated electrolyte (0.1 M Na2SO4) at a scan rate of 10 mV/s. Constant potentials were applied to determine H2O2 yields obtained with the PCM substrate and electro-Fenton efficiencies of herbicide degradation by CMOF@PCMs. Na2SO4 (0.1 M) was used as electrolyte, and solution pH was adjusted to 4–10 using H2SO4 and NaOH. H2O2 yield measurements were performed in triplicate. Electro-Fenton degradation experiments were performed in duplicate with the average value being reported.

Analytical Methods

To quantify the H2O2 produced, samples collected at certain time intervals were analyzed by titanium oxalate spectrophotometric method.[22] Briefly, reaction sample was added into 3 M sulfuric acid solution before reaction with potassium titanium oxalate solution (0.5 wt % in water) to produce a yellow pertitanic acid (TiO(H2O2)2+) complex. The absorbance of the solution was measured using a JASCO V-750 spectrophotometer at 400 nm. H2O2 solutions with known concentration were used to construct a calibration curve.

During the electro-Fenton process, the concentrations of herbicides were analyzed using a high-performance liquid chromatography (Waters Acquity UPLC) system coupled to e2998 photodiode array detector and an e2695 separation module. Sunfire C18 column (4.6 × 50 mm; 5 μm particles) was used. (A) 0.1% Phosphoric acid and (B) methanol were used as mobile phase with flow rate of 0.8 mL min–1.

Results and Discussion

Synthesis of E-Fenton Catalysts from Fe-MOFs

The iron-containing MOFs MIL-88, MIL-100, and MIL-101 were used as templates in the preparation of electro-Fenton catalysts. MIL-88(Fe) and MIL-101(Fe) were synthesized by using the same organic ligand, terephthalic acid (BDC), with FeCl3. MIL-88(Fe) consists of an infinite framework with three-dimensional channels formed by [Fe3(μ3-O)(COO)6(H2O)2Cl]6 secondary building units (SBUs) and BDC linker (Figure ). MIL-101(Fe), however, consists of quasi-spherical cages formed by Fe3O–carboxylate trimers SBU and BDC linkers (Figure ). The mesoporous cages with dimensions of 2.9 by 3.4 nm are accessible via microporous windows or openings (1.2 and 1.6 nm). Comparison between MIL-88(Fe) and MIL-101(Fe) allows us to study the impact of MOFs topology on the subsequent electrocatalytic properties of the CMOFs. In addition, MIL-100(Fe) was synthesized by replacing the BDC ligand with 1,3,5-benzenetricarboxylic acid (BTC) in order to investigate the effect of the ligand on the electrocatalytic properties of the CMOFs. MIL-101(Fe) and MIL-100(Fe) have mesoporous cages formed by the combination of iron octahedra trimers of [(Fe3(m3-O))(OH)(H2O)2]6 SBUs and BTC linkers (Figure ). The MIL-100(Fe) cages (2.5 and 2.9 nm) have microporous apertures of 0.55 and 0.86 nm, which are significantly smaller than that of MIL-101(Fe)’s. In order to investigate the functional group effect, MIL-88(Fe)-NH2 and MIL-101(Fe)-NH2 were synthesized by using an aminated ligand, BDC-NH2. Since the MOF crystal size is directly proportional to the overall reaction time, the microwave reaction conditions needed to be optimized in order to produce nanoscaled MOF crystals. Thus, MIL-88(Fe) was synthesized as a function of microwave reaction time. The SEM images of the various MIL-88(Fe) products are summarized in Figure S2. Initially, the size of the MIL-88(Fe) crystals increased as the microwave irradiation time was lengthened. After 15 min, the average diameter of the MIL-88(Fe) crystals was stabilized at 500 nm which is most likely due to the depletion of starting materials. Since irradiation for 40 min produced the highest yield, we then used a 40 min microwave irradiation time for the synthesis of the other MOFs. The PXRD patterns of synthesized MOFs were consistent with the theoretically predicted PXRD patterns (Figure S3); thus, the identities of MOFs used in this study were confirmed.

Figure 1

Structure of (a) MIL-88(Fe), (b) MIL-100(Fe), and (c) MIL-101(Fe).

The electro-Fenton catalysts were prepared by direct pyrolysis of the MOFs at 873 K for 30 min under a N2 atmosphere. The synthesized catalysts are denoted as CMIL-88, CMIL-100, CMIL-101, CMIL-88-NH2, and CMIL-101-NH2, for MIL-88(Fe), MIL-100(Fe), MIL-101(Fe), MIL-88(Fe)-NH2, and MIL-101(Fe)-NH2, respectively. SEM micrographs show that CMOFs retained crystal morphologies similar to that of parent MOFs (Figure S4). PXRD pattern showed that the main iron-containing species in the CMOFs is Fe3O4 (Figure ), except for CMIL-88-NH2 for which cementite (Fe3C) was also identified as an iron component. The compositions of the CMOFs were confirmed by energy dispersive X-ray spectroscopy (EDS). The results confirmed the relatively even distribution of Fe within the CMOFs (Figure S5).

Figure 2

PXRD spectrum of CMOFs used in the current study.

The MOFs precursor and CMOFs were also analyzed by the FTIR; the results are shown in Figures S6 and S7. For MIL-88(Fe)–NH2, peaks at 3414 and 3370 are assigned to the symmetric and asymmetric stretching of the primary amine, while peaks at 1620 and 1392 cm–1 are assigned to the asymmetric and symmetric vibrations of the carboxyl groups, respectively. The absence of peak at 1700 cm–1 indicates that no uncoordinated BDC is present in MIL-88(Fe). In addition, doublet at 572, 487 cm–1 is most likely due to the Fe–O vibration in MIL-88(Fe). For CMIL-88, a broad peak at 3397 cm–1 is assigned to the O–H stretching vibrations of C–OH group. A newly formed peak at 1644 cm–1 is assigned to the vibration of C=C bonds in the graphitic carbon skeleton. A strong peak at 522 cm–1 is associated with the stretching vibration of Fe–O from the Fe3O4-NPs. These peaks are assigned based on previously reported assignments.[23,24] MIL-101(Fe)–NH2 and CMIL-101–NH2 gave similar FTIR spectra (Figure S7). This result implies that the formation of graphitic carbon and Fe3O4 during MOFs pyrolysis is independent of the specific MOF topology.

To prepare the PCM-supported bifunctional electrochemical catalysts (CMOF@PCM), resorcinol–formaldehyde (RF) polymers were synthesized as a carbon precursor. The RF polymer was pyrolyzed at 1073 K for 2 h under N2 atmosphere to produce PCM. After formation, the product PCM was impregnated with MOFs before carbonization at 873 K for 30 min under N2. The MOFs loading amount was adjusted by changing the amount of the specific MOF mixed with PCM during impregnation. The MOFs loading of MOF@PCM catalyst precursor was fixed at 25 wt %. Pyrolysis of MOF@PCM at 873 K for 30 min under a N2 atmosphere produced mesoporous material designated as CMOF@PCM. SEM analysis indicates that the PCM substrate has an open porous network that is fully accessible to both electrolytes, substrates, and dissolved gas molecules (Figure ). In addition, SEM micrograph shows that the CMOF s were deposited inside the macropores of PCM material (Figure S8). Based on these analyses, the loaded CMOFs have mean particle sizes ranging between 150 to 200, 100 to 200, 100 to 150 nm for CMIL-88@PCM, CMIL-100@PCM, and CMIL-101@PCM, respectively.

Figure 3

SEM micrographs of (a) CMIL-88@PCM, (b) CMIL-100@PCM, and (c) CMIL-101@PCM.

Electrogeneration of H2O2 by Porous Carbon Substrates

The electrochemical activities of the PCM support for the oxygen reduction reaction (ORR) were investigated. Cyclic voltammetry (CV) was used to examine the oxygen reduction potential. As shown in Figure S9, the working electrode prepared from the PCM substrate was featureless for an Ar-saturated electrolyte solution but showed a pronounced reduction peak in an O2-saturated electrolyte solution. The current density measured at the ORR peak was 2.3 mA/cm2. A similar value was found at pH 4, 7, and 10. The onset potentials at pH 4, 7, and 10 were 0.49, 0.75, and 0.94 V vs RHE, respectively. These results suggest the PCM substrate can electrochemically catalyze the ORR on site over a wide range of solution pH. Such a result bodes well for promoting the coupled Fenton reactions occurring under these conditions.

Oxygen reduction taking place on graphitic materials is selective toward a two-electron reduction to produce H2O2 with a selectivity over 90%.[25,26] As a baseline, the amount of H2O2 produced via O2 reduction on the PCM support was determined (Figure ). These results show that PCM substrate can produce H2O2 at three pH values which are consistent with the cyclic voltammetry results at each pH (Figure S9). In addition, the rate of formation of H2O2 increases with increasing applied potential. However, once the applied potential exceeds that of ORR peak potential, there is a declining increase in H2O2 productivity for all solution pH tested. This is due to a diffusion-controlled limit for H2O2 release from the porous structure of PCM substrate. Thus, with the porous structure there is an elevated concentration of H2O2 that works against a further increase in H2O2 production rate as measured in solution. The rates of H2O2 production on 6 mg of the PCM substrate at pH 4, 7, and 10 were 9.06, 9.41, and 9.48 mmol/L/h, respectively, which suggests solution pH has limited effect on its H2O2 productivity. In contrast, nonporous carbon material produced H2O2 at a rate of 4.41 mmol/L/h at pH 7 under similar reaction conditions.[27] Since the oxygen reduction performance of carbon-based materials is known to be unaffected by buffer,[15] this suggests solution pH can be easily maintained during the oxygen reduction reaction using PCM material.

Figure 4

Concentration of H2O2 produced by 6 mg of PCM substrate at (a) pH 4, (b) pH 7, and (c) pH 10.

Electro-Fenton Degradation of Napropamide

In order to demonstrate the feasibility of catalytically enhancing heterogeneous electro-Fenton reactions on CMOF@PCM, napropamide was selected as a model compound for testing the electro-Fenton process using CMIL-88@PCM. The initial concentration of napropamide was 10 ppm in 50 mL of 0.1 M Na2SO4; the effective surface area of the CMIL-88@PCM electrode was 6 cm2. For comparison, we investigated the removal efficiency of napropamide via the E-Fenton process under multiple sets of conditions; these results are summarized in Figure . H2O2 alone was unable to degrade napropamide. The kinetics of napropamide removal using the CMIL-88(Fe)@PCM electro-Fenton process, electrosorption on CMIL-88(Fe)@PCM, electrosorption on PCM in an Ar-saturated 0.1 M Na2SO4 electrolyte, and electrosorption on PCM in an O2-saturated 0.1 M Na2SO4 electrolyte resulted in loss percentages of 75.5%, 35.4%, 14.6%, and 16.0%, respectively. The electrosorption of napropamide by the PCM support material rapidly reaches equilibrium within 30 min due to the limited surface area of the PCM material. On the other hand, CMIL-88(Fe)@PCM removed substantially more napropamide by electrosorption due to the additional CMOFs embedded in the PCM support. The removal efficiency of napropamide on CMIL-88(Fe)@PCM in Ar-saturated electrolyte is similar to that of the O2-saturated electrolyte at the presence of isopropyl alcohol (IPA) as an ·OH trap. This result suggests that ·OH is the primary oxidant generated during the electro-Fenton process.

Figure 5

Napropamide (10 ppm) removal by H2O2 oxidation (400 ppm), PCM substrate (−0.14 V), and CMIL-88@PCM (−0.14 V). All experiments were performed in 0.1 M Na2SO4 electrolyte at pH 7.

CMOF@PCM is a unique, multifunctional electrochemical catalyst in which the electro-Fenton reaction is catalyzed by CMOF-NPs while H2O2 is generated on site by the PCM support substrate. Because the electrochemical performance of CMOFs may be controlled by the MOFs precursors, various MOFs with different topology including MIL-88(Fe), MIL-100(Fe), and MIL-101(Fe) were synthesized to investigate the MOFs topology effects on CMOF’s electro-Fenton performance. Among the MOFs, MIL-88(Fe) and MIL-101(Fe) were synthesized using an identical metal salt and a bidentate linker, BDC. In contrast, a tridentate linker, BTC, was used to synthesize MIL-100(Fe). The relative napropamide removal efficiency decreased in the following order: CMIL-100@PCM > CMIL-88@PCM > CMIL-101@PCM, resulting in removal efficiencies of 82.3%, 64.1%, and 60.56%, respectively, after 60 min of electrolysis (Figure ). The pseudo-first-order kinetic plots versus time for napropamide degradation are shown in Figure S10. The corresponding pseudo-first-order constants for napropamide removal were determined to be 1.70, 0.99, and 0.87 h–1 for CMIL-100@PCM, CMIL-88@PCM, and CMIL-101@PCM, respectively (Table S2).

Figure 6

Napropamide (10 ppm) removal by electro-Fenton using CMOF@PCMs prepared from MIL-88(Fe), MIL-100(Fe), and MIL-101(Fe). All experiments were performed in 0.1 M Na2SO4 electrolyte at pH 7 (−0.14 V).

In order to investigate the effects of MOF functional groups on CMOF@PCM electro-Fenton kinetics, amine-functionalized MOFs using an aminated linker were synthesized. CMOFs were prepared from BDC and BDC-NH2. However, due to the steric hindrance of the tridentate trismic acid, CMIL-100-NH2@PCM was not prepared in the same fashion. A comparison of Napropamide electro-Fenton removal rates using CMOF@PCM and CMOF-NH2@PCM is given in Figure . The removal percentages after 60 min were 73.42%, 64.1%, 60.56%, and 53.63% for CMIL-88-NH2@PCM, CMIL-88@PCM, CMIL-101@PCM, and CMIL-101-NH2@PCM, respectively. Again a pseudo-first-order kinetic decay of napropamide was observed (Figure S11) with first-order constants of 1.26, 0.99, 0.87, and 0.74 h–1 for CMIL-88-NH2@PCM, CMIL-88@PCM, CMIL-101@PCM, and CMIL-101-NH2@PCM, respectively (Table S2).

Figure 7

Napropamide (10 ppm) removal by electro-Fenton using CMOF@PCMs and CMOFs-NH2@PCMs (0.1 M Na2SO4, pH 7, −0.14 V).

The effect loading levels of CMIL-100 on the observed electro-Fenton kinetics of napropamide degradation were studied for samples that were prepared from mixing 10, 25, 50, and 75 wt % of CMIL-100 into the PCM support. These sample were designated as CMIL-100@PCM10, CMIL-100@PCM25, CMIL-100@PCM50, and CMIL-100@PCM75, respectively. Napropamide degradation was carried out with CMIL-100 alone without PCM substrate for comparison. As illustrated in Figure S12, the napropamide degradation percentage after 60 min is 11.8, 89.2, 97.0, 94.7, 92.6, and 91.3% for PCM, CMIL-100@PCM10, CMIL-100@PCM25, CMIL-100@PCM50, CMIL-100@PCM75, and CMIL-100, respectively. Napropamide removal also follows apparent pseudo-first-order kinetics (Figure S13). The first-order kinetic constants are summarized in Table S2. The rate of napropamide degradation increased with the loading amount of CMIL-100 from 10 to 25 wt % but declined at higher doping concentrations. The maximum degradation efficiency was obtained using CMIL-100@PCM25 (i.e., 97.0% after 60 min of electrolysis). In traditional applications of the electro-Fenton process, contaminant removal efficiencies have been found to increase with increasing amount of heterogeneous electro-Fenton catalyst until a maximum efficiency is reached at higher loading levels.[28] In these aforementioned systems, nonporous bulk-phase Fe3O4 was used, but there are limits in that nonsurface accessible Fe(II) does not participate in the Fenton reaction. Thus, a relatively low utilization efficiency of electro-Fenton catalyst normally requires large amounts of Fe3O4 to be used in order to fully consume the added hydrogen peroxide. In contrast, for CMOF@PCM, the main electro-Fenton active species is porous CMOFs. The active form in our case is Fe3O4-NP that is embedded in graphitic carbon derived from the specific organic linker. The highly uniform crystalline structure of the MOF precursors allows even distribution of the Fe3O4-NPs throughout the CMOF-NPs after carbonization (Figure S5). Therefore, the embedded Fe3O4-NPs in the graphitic molecular support have similar local coordination environments, which appear to be a prerequisite for maximizing the number of electrocatalytically active sites.[29] Since both the PCM support substrate and CMIL-100 NPs are porous materials, they are able to maximize the interactions of H2O2 with the Fe3O4-NPs. However, at a low loading level of CMOFs (e.g., CMIL-100@PCM5), the electro-Fenton reaction rate is limited by the amount of Fe3O4 that was embedded in the multifunctional electrocatalysts, CMOF@PCM. Low loading of Fe3O4 leads to an insufficient amount of Fe(II) available for the reaction with H2O2 to produce a sufficient amount of ·OH radical. However, the electrochemical reduction of O2 as catalyzed by the PCM substrate is the rate-limiting step at higher loading levels of CMOF in the form of CMIL-100@PCM75.

The impact of pH on napropamide degradation using CMIL-100@PCM was minimal as shown in Figures S14 and S15. The corresponding pseudo-first-order rate constants as given Table S2 suggest the overall rate of degradation is somewhat higher at circum-neutral pH than at low or high pH. The applied potential was set at −0.14 V, which is a potential that produces the highest concentration of H2O2 (Figure ). Any pH dependence may be due to impact on the sorption of napropamide onto the CMIL-100@PCM electrode.

Because actual wastewater has a wide range of conductivity, we have evaluated napropamide removal efficiency at different electrolyte concentration. Figure S16 shows the napropamide removal efficiency was slightly reduced at reduced electrolyte concentration. In addition, Figure S16 suggests napropamide can be successfully removed by CMIL100@PCM in simulated river water prepared according to Table S5.

Figure shows that about 90% mineralization of napropamide can be achieved within 2 h of electrolysis via the electro-Fenton reaction as catalyzed by CMIL-100@PCM at pH 7.

Figure 8

TOC removal efficiency of electro-Fenton removal of 10 ppm napropamide by CMIL-100@PCM (0.1 M Na2SO4, pH 7, −0.14 V).

The CMIL-100@PCM electro-Fenton was applied for the degradation of other herbicides beyond (Figure ). The degradation of metolachlor, glyphosate, atrazine, acetochlor, and dichloprop also followed apparent first-order kinetics (Figure S17) with degradation percentages of 83.0, 78.5, 70.9, 91.1, and 76.2% after 60 min of electrolysis and kobs values of 1.50, 1.34, 0.99, 1.90, and 1.27 h–1, respectively. On the basis of reported electro-Fenton and Fenton degradation mechanism,[16] we propose a generalized mechanistic pathway for ·OH-mediated degradation of pesticides using CMOF@PCM electrodes (Figure S18).

Figure 9

Electro-Fenton degradation of metolachlor, glyphosate, atrazine, acetochlor, and dichlorprop (10 ppm) by CMIL-100@PCM (0.1 M Na2SO4, −0.14 V, pH 7).

Reusability and Stability

The recyclability of the CMIL-100@PCM electro-Fenton catalyst has been assessed for three consecutive cycles. No significant change to the napropamide degradation efficiency has been observed (Figure S19), demonstrating the excellent stability of CMOF@PCM. Over time leaching of reactive Fe from CMIL-100@PCM may impact the durability of this system. However, Figure S20 shows that only approximately 6.5 ppb of the Fe was detected in the solution, and the Fe concentration remained unchanged throughout the reaction using CMIL-100@PCM. The initial Fe concentration may be due to the dissociation of CMIL-100@PCM that was loosely bonded to the carbon paper substrate by nafion binding agent. In comparison, significant Fe leaching has been observed for Fe3O4@PCM, which consists of Fe3O4-NP embedded on PCM substrate. The concentration of leached Fe in the electrolyte had increased linearly and reached about 200 ppb after 2 h reaction.

The Primary Basis for High E-Fenton Reactivity

In order to increase the electrochemical catalytic efficiency, it is important to be able to access the largest number of surface active sites that are capable of participating in an electrocatalytic reaction. However, the reactivity of the electrochemical active sites is affected by both the surface structure of the catalyst and its interactions with molecules or molecular moieties close to the site of catalysis. For example, heterogeneous catalytic reactions are dependent on the catalytically active surface area. In our system, it is difficult to accurately determine the actual electrochemically active surface area of the Fe3O4-NPs embedded within the porous carbon matrix due to the dielectric behavior of magnetite. Thus, Brunauer–Emmett–Teller (BET) surface area measurements are used as a surrogate. BET measurements of the CMOF surfaces were determined at 77 K by N2 adsorption with the results summarized in Table S3. For nonfunctionalized CMOFs, the BET surface areas ranged from 287.9, 340.9, and 361.6 m2/g for CMIL-88, CMIL-100, and CMIL-101, respectively. Hysteresis (type H3) between adsorption and desorption branches at medium pressure (P/P0 = 0.4) were observed for both CMIL-88 and CMIL-101 (Figure S21). This behavior is consistent with the presence of wide distribution of mesopores. In contrast, CMIL-100 gave a type H4 isotherm, which suggests the presence of narrow slitlike pores that are dominated by micropores with a limited number of mesopores. In fact, the pore size distribution (Figure S22) showed that the average pore diameters were 13.12, 4.85, and 13.65 nm for CMIL-88, CMIL-100, and CMIL-101, respectively. This result indicates that both the BET surface areas and the pore structures of CMOFs are dependent on the linker and core metal ion rather than the topology of parent MOFs. The BET surface areas of the CMOFs synthesized from the aminated MOFs are given in Table S3. These results show that amine functionalization resulted in slightly lower BET surface areas for CMIL-88-NH2 and CMIL-101-NH2 of 217.88 and 211.85 m2/g, respectively. Both CMIL-88-NH2 and CMIL-101-NH2 gave type H4 BET isotherms (Figure S23) that indicated that micropores control the pore structure. The average pore diameters for CMIL-88-NH2 and CMIL-101-NH2 were determined to be 8.50 and 11.10 nm, respectively (Figure S24). These results clearly suggest that amine functional groups in the linker lead to a reduction in pore diameters of the CMOFs. These results are also consistent with the location of amine functional group in the MOFs precursor, which is at the edge of the aperture (Figure S25). These observations highlight the importance of BET surface area in evaluating the reactivity of electro-Fenton catalysts. Furthermore, the results presented in Table S3 also show that the CMOF@PCMs have greater BET surface area than CMOFs without PCM substrate. This is consistent with the higher BET surface area of PCM. In light of these results, we have normalized the pseudo-first-order rate constants to the BET surface areas (kSA) to give surface area normalized heterogeneous rate constants. However, since the BET surface area of the PCM substrate is the same for all of the CMOF@PCMs, the BET surface areas of the CMOFs were used instead of CMOF@PCMs for normalization. These results are summarized in Table S4, which show that the normalized rate constants, kSA, follow the order of CMIL-88-NH2@PCM > CMIL-100@PCM > CMIL-101-NH2@PCM = CMIL-88@PCM > CMIL-101@PCM.

The normalized rate constants for the electro-Fenton reactions show that the catalytic activity of the CMIL-88-NH2@PCM catalyst is almost twice that of the CMIL-88@PCM in spite of a lower BET surface area (Figure S26). The PXRD spectra indicate that the CMIL-88 NPs are present primarily as Fe3O4 without an additional phase (Figure S27). This result is consistent with the high-resolution TEM images that clearly identify nanoparticulate Fe3O4 in Figure S28. The observed lattice lines have a lattice spacing of 4.90 Å, which is consistent with the (110) plane of Fe3O4. In contrast, the CMIL-88-NH2 NPs TEM images show the presence of two iron phases, Fe3O4 and Fe3C. The high-resolution TEM images shown in Figure show that the CMIL-88-NH2 NPs are uniformly dispersed within the graphitic PCM substrate. CMIL-88-NH2 has a unique core–shell structure. The graphic carbon shell thickness is ∼2 nm, with a lattice spacing of the intermediate layer of 2.38 Å, which corresponds to the (210) plane of Fe3C. The core is Fe3O4 with a lattice spacing of 2.53 Å. These results are also consistent with the PXRD spectrum (Figure S27).

Figure 10

(a) TEM images of CMIL-88-NH2@PCM; (b) high-resolution TEM image of CMIL-88-NH2@PCM core–shell structure.

The difference in electro-Fenton reaction performance between CMIL-88-NH2@PCM and CMIL-88@PCM is most likely due to the difference in their structures. In the case of CMIL-88@PCM, Fe3O4 NPs are imbedded in the carbon substrate matrix. The unit cell for Fe3O4 has cubic inverse spinel structure, each containing 8 Fe atoms occupying Wyckoff position 8a and 16 Fe atoms occupying Wyckoff position 18d in space group Fd3m (Figure S29a). The Fe atoms are directly accessible by the H2O2 for electro-Fenton reaction. In terms of CMIL-88-NH2@PCM, core–shell structured NPs are imbedded in the graphitic carbon substrate. Within the core–shell nanoparticulate structures, the Fe3C interfacial layer is sandwiched between the graphitic carbon shell and Fe3O4 core. Fe3C has orthorhombic structure in a Pnma space group. Each unit cell contains 12 Fe atoms and 4 C atoms. Eight Fe atoms occupy Wyckoff position 4c, and the remaining Fe atoms occupy Wyckoff position 8d (Figure S29b). Compared to the base elemental metal, the formation of a metal carbide (Fe3C) leads to a lengthening of the metal–metal bond distance due to presence of carbon. As the result, a metal d-band contraction occurs, generating more density of states (DOS) near the Fermi level.[30−32] Therefore, metal carbides are quasi electron conductors. Incorporation of Fe3C as an intermediate layer within CMIL-88-NH2@PCM composite material improves the electro-Fenton reactivity by (1) facilitating the electron transfer between the graphitic carbon shell and Fe3O4 core, (2) increasing the catalytic reactivity by reducing the lattice mismatch between the graphitic carbon shell and Fe3O4 core,[33] and (3) fixing in three dimensions Fe(II) species within the Fe3O4 core as well as locking Fe atoms within Fe3C. The latter effect essentially prevents Fe(II) leaching into solution and provides for long-term stability of the electrode. This is confirmed by Fe leaching measurement from the composite catalyst (Figure S20).

In conclusion, the degradation of a set of herbicides has been achieved using an electro-Fenton process based on multifunctional CMOF@PCM catalysts. In this version of an electro-Fenton reaction system, porous carbonized Fe-based metal organic framework compounds (CMOFs) optimize the availability of Fe(II) sites to initiate the initial oxidation of Fe(II) by H2O2 to Fe(III) and ·OH. The reverse reaction, reduction of Fe(III), is achieved by the negative applied potential on the composite carbon substrate electrode (Figure S18). A low applied potential (−0.14 V) combined with a broad operational pH range may make this particle electro-Fenton reaction system cost-effective for the degradation and mineralization of a wide range of herbicides or other oxidizable organic chemical water contaminants.