Reductive Dechlorination of Chloroacetamides with NaBH 4 Catalyzed by Zero Valent Iron, ZVI, Nanoparticles in ORMOSIL Matrices Prepared via the Sol-Gel Route

: The e ﬃ cient reductive dechlorination, as remediation of dichloroacetamide and monochloroacetamide, toxic and abundant pollutants, using sodium borohydride catalyzed by zero valent iron nanoparticles (ZVI-NPs), entrapped in organically modiﬁed hybrid silica matrices prepared via the sol-gel route, ZVI@ORMOSIL, is demonstrated. The results indicate that the extent of the dechlorination reaction depends on the nature of the substrate and on the reaction medium. By varying the amount of catalyst or reductant in the reaction it was possible to obtain conditions for full dechlorination of these pollutants to nontoxic acetamide and acetic acid. A plausible mechanism of the catalytic process is discussed. The present work expands the scope of ZVI-NP catalyzed reduction of polluting compounds, ﬁrst reports the catalytic parameters of chloroacetamide reduction, and o ﬀ ers additional insight into the heterogeneous catalyst structure of M 0 @ORMOSIL sol-gel. The ZVI@ORMOSIL catalyst is ferromagnetic and hence can be recycled easily. 95.5 nm, and volume mean diameter (VMD) of 95.5 nm, measured for 25 nm ZVI commercial powder in 1.0 M ethanolic suspension, show that agglomeration of a similar degree occurs in the suspension.


Introduction
Halogenated hydrocarbons are known to be toxic pollutants [1]. More than half a century ago it was proposed to study each halogenated compound independently of its homologue hydrocarbon series [1]. While there are number of halogenated pollutants that are monitored by the health authorities, e.g., halo-acetic acids, others which appear on the EPA's (United States Environmental Protection Agency) contaminant candidate list await proper directives and legislation, e.g., halo-acetamides which are formed during chlorination of drinking water [2]. Halo-organic compounds can be reduced electro-chemically [3,4], photo-chemically [5], radiolytically [6], and by a variety of reducing agents [7]. These processes often require a catalyst [8]. M 0 -nano-particles (M 0 -NPs), e.g., Ag 0 -NPs and

1.0 M ZVI-NPs Ethanolic Suspension
The 1.0 M ZVI-NPs suspension, prepared from the 25 nm ZVI commercial powder, was analyzed using transmission electron microscopy to ensure that no major agglomeration had occurred in the suspension used for preparation of the catalyst (Figure 1). Although agglomerated macro-particles are observed, they largely consist of recognizable nanosized ZVI spheres that fall in the nanometer range, <100 nm. Hence, they are suitable to be used as nano-catalysts. This colloidal suspension was used in the ZVI@ORMOSIL catalyst preparation.
The ethanolic suspension shows magnetic properties that were observed by simple magnet-induced moving of the suspended ZVI powder within the solution.
Photon cross correlation spectroscopy measurements were performed to estimate the initial particle aggregation degree due to interaction of the particles in the suspension. The results obtained support the TEM findings indicating some agglomeration of the ZVI NPs ( Figure 2). Although it might be suspected that the agglomerated particles that are observed in Figure 1 might originate from solvent evaporation that occurred while placing the suspension sample on the grid, the average values of Sauter mean diameter (SMD), 95.5 nm, and volume mean diameter (VMD) of 95.5 nm, measured for 25 nm ZVI commercial powder in 1.0 M ethanolic suspension, show that agglomeration of a similar degree occurs in the suspension.   ZVI@ORMOSIL Synthesis via the Sol-Gel Route The differences between ZVI@ORMOSIL gel forms in different preparation steps are visualized in Figure S1. As the wet gel dries in the desiccator, the raw gel undergoes shrinkage due to cooperative action between hydrolysis and densification [35]. In the final stages of the preparation, the dried gel is crushed into powder using a mortar and pestle. The obtained powder has magnetic properties ( Figure 3). ZVI@ORMOSIL Synthesis via the Sol-Gel Route The differences between ZVI@ORMOSIL gel forms in different preparation steps are visualized in Figure S1. As the wet gel dries in the desiccator, the raw gel undergoes shrinkage due to cooperative action between hydrolysis and densification [35]. In the final stages of the preparation, the dried gel is crushed into powder using a mortar and pestle. The obtained powder has magnetic properties ( Figure 3).

Catalyst Characterization
To investigate the surface morphology and to determine the elemental composition of 1.0% ZVI@ORMOSIL, STEM/EDAX analysis was performed ( Figure 4).
To investigate the surface morphology and to determine the elemental composition of 1.0% ZVI@ORMOSIL, STEM/EDAX analysis was performed ( Figure 4).
The results indicate that the amorphous ORMOSIL surface comprises many cracks, probably due to non-uniform shrinkage and densification of the matrix in the drying stage of the sol-gel process. The elemental composition spectra were taken from two regions of the monolith, marked in Figure 4a. Near the cracks and breaks formed in the structure, the composition shows the presence of elemental iron (spectrum 10). Conversely, signals collected from the uniform part of the bulk (spectrum 11) exhibit no significant elemental iron. Nevertheless, traces of elemental iron might be present on the monolith surface. It is possible that ZVI will not be detected if it is at a low concentration level and below the sensitivity level of the specific apparatus. A surface composition sensitive method, XPS, was applied to ensure more precise iron detection, as shown below.
(a) The results indicate that the amorphous ORMOSIL surface comprises many cracks, probably due to non-uniform shrinkage and densification of the matrix in the drying stage of the sol-gel process. The elemental composition spectra were taken from two regions of the monolith, marked in Figure 4a.
Near the cracks and breaks formed in the structure, the composition shows the presence of elemental iron (spectrum 10). Conversely, signals collected from the uniform part of the bulk (spectrum 11) exhibit no significant elemental iron. Nevertheless, traces of elemental iron might be present on the monolith surface. It is possible that ZVI will not be detected if it is at a low concentration level and below the sensitivity level of the specific apparatus. A surface composition sensitive method, XPS, was applied to ensure more precise iron detection, as shown below.
Discrimination between successful entrapment in the voids of the bulk material and surface immobilization is possible by measuring the surface composition. To characterize the surface composition of the ZVI@ORMOSIL catalyst, XPS analysis was used. The survey scan of the whole range of 1.0% load ZVI@ORMOSIL is presented in Figure 5. Silicon and oxygen were successfully observed. The carbon detected is probably that of the methyl groups in the matrix originating from the methyltrimethoxysilane precursor. Focused scans for iron content led to detection of almost negligible traces, close to the baseline noise level, as shown in the insert of Figure 5. Using manual sample manipulation, a successful integration extracted a sample area that was calculated as 0.2% atomic of the whole sample area that was scanned in the analysis. Although iron was found as a part of the surface composition, as summarized in Table S1, the almost negligible amounts confirm that most of the catalyst was entrapped within the bulk of the hybrid silica host structure. These findings indicate that the ZVI is trapped in the pores, and they exclude the unwanted possibility of significant surface immobilization of the catalyst. Three types of samples were analyzed using powder X-ray diffraction, PXRD: the original 1.0 M suspension of ZVI used in the preparation of matrices, 1.0% ZVI-NP load ZVI@ORMOSIL, and 10% load ORMOSIL, which was prepared to aid the identification of ZVI phases. Incorporation of low amounts of ZVI nanoparticles in the bulk structure might result in diffraction patterns that are obscured due to the amorphous silica baseline signal. Figure 6 shows a comprehensive overlap among phases found in the original ZVI suspension that was used in the preparation of 1.0% and 10% load ZVI@ORMOSIL, as summarized in Table 1. Major phases were seen at 44.7 • and 65.1 • , fitting the 1,1,0 and 2,0,0 plains of Fe 0 phases. The slight presence of a Fe 3 O 4 phase was also confirmed with peaks at 35.4 • and 62.5 • . These findings fall in line with the expected observations since an almost neglectable catalyst presence was found in the surface layers of the gel and considerable presence was confirmed for the penetrative XRD analysis of ZVI@ORMOSIL.  The Scherrer equation was used to calculate the average crystallite size (Dp, nm). Reflection parameters at 44.7 • and 65.1 • were used for the calculation; the results were averaged and are summarized in Table 2 (all the relevant phases and card numbers files are given in the Supplementary Materials (SI-3)). A certain agglomeration extent might be attributed to the influence of ammonium hydroxide [36] introduced in the preparation of ZVI@ORMOSIL (2%NH 3 , 1.5 mL). Similarly prepared catalysts previously reported [31] had an average crystallite size of 34.5 nm, as calculated from the powder XRD results. It was noted that a ZVI secondary particle might comprise many primary crystallites. A descriptive ORMOSIL sol-gel pore system and framework study of the catalyst structure was performed using N 2 adsorption-desorption studies. Surface area and pore volumes for 1.0% ZVI@ORMOSIL and for the blank@ORMOSIL matrix without nanoparticles are summarized in Table 3. Both the host matrix without particles and the entrapped catalyst matrix have large surface areas and narrow distributions of the inner pores that are in the mesoporous range. The pore size distributions (diameter) are presented in Figure S4. Surprisingly, the doped matrix has a larger pore volume and surface area; this result suggests that that the agglomeration of the gel is affected by the presence of ZVI-NPs added during the gelation stage. The larger surface area of ZVI@ORMOSIL induces the reaction substrates to be adsorbed more onto its surface. The N 2 adsorption-desorption isotherms presented in Figure 7 can be classified as type IV isotherms with H2A hysteresis loops. The asymmetrical shape of these H2A hysteresis loop types (as per current IUPAC convention [37]) are typical for porous glass [38]. These loops describe a distribution of pore sizes and shapes with bottleneck constrictions [39]. They are characteristic of interconnected ink-bottle pores rather than isolated, individual ink-bottle pores, as an "assembly of cavities connected by constrictions" [39]. These findings correspond to previously reported organically modified silica (ORMOSIL) matrices prepared by mixing a large variety of precursors [40,41]. Similar hysteresis loops were previously reported for ORMOSIL entrapped metal nanoparticles used in the reductive dehalogenation of halo-acetic acids [13].

Dehalogenation of Di-and Mono-Chloroacetamides
The reductions of monochloroacetamide (MAcAm), dichloroacetamide (DAcAm), and of solutions containing both substrates at a 1:1 molar ratio were performed in different reaction media. The reactions were performed at an initial pH of 8.0 with a constant substrate:NaBH 4 molar ratio of 1:20, if not otherwise stated. The products distributions obtained under each reaction condition are summarized in Table 4.
To check the reduction rates of ZVI alone as a reducing agent, typical dehalogenation reactions of both DAcAm and MAcAm were performed, the absence of the sodium borohydride addition step being the only difference. These reactions were continued for 24 h with no reduction yields worth mentioning. Thus, the NaBH 4 addition is required. NaBH 4 was added to the homogenized suspension of the substrates and catalysts and 2.0 mL of water was added shortly afterward to ensure the full dissolution of NaBH 4 . The products analysis was performed 15 min later. This time frame choice was designed to assist in the mechanistic study displaying the distribution of the reaction products under various conditions. The plausible reduction products for DAcAm; MAcAm, acetamide (AcAm), Acetic acid (AA), and ammonia are given in reaction (1) and the product distributions obtained by its reduction are presented in Figure 8.
of both DAcAm and MAcAm were performed, the absence of the sodium borohydride addition step being the only difference. These reactions were continued for 24 h with no reduction yields worth mentioning. Thus, the NaBH4 addition is required. NaBH4 was added to the homogenized suspension of the substrates and catalysts and 2.0 mL of water was added shortly afterward to ensure the full dissolution of NaBH4. The products analysis was performed 15 min later. This time frame choice was designed to assist in the mechanistic study displaying the distribution of the reaction products under various conditions. The plausible reduction products for DAcAm; MAcAm, acetamide (AcAm), Acetic acid (AA), and ammonia are given in reaction (1) and the product distributions obtained by its reduction are presented in Figure 8.  The results in Table 4 suggest that: I. The BH 4 − will first reduce the surface hydroxides/oxides of the ZVI-NPs. Then, it is expected to form (ZVI-NP)-H n+m (n−m)− via reactions (2) and (3) [31]. In the absence of an oxidizing agent, these reactions are followed by reactions (4) and/or (5) [31]. In the presence of an oxidizing substrate, reactions (4) and (5) II. The dehalogenation of the chloro-acetamides by BH 4 − is expected to follow an analogous mechanism to the dechlorination of Cl 3 CCO 2 − [31]. As B(OH) 3 formed in reaction (2) is a buffer with a pK b = 4.9 and as OH − is formed in reactions (3) and (5), the reaction media are always slightly alkaline. The first step in the de-halogenation process is expected to follow either reaction (6), which is a hydrogen transfer process, or reaction (7), which is an electron transfer process; in both, the radical ClCHCONH 2 . is formed.
The MAcAm product formed in the dechlorination of DAcAm occurs via one or more of reactions (9)- (11).
IV. The results indicate that the nature of the solvent, though enough water is always present, affects the mechanism of the dehalogenation considerably. At present, the data available do not enable us to determine at what stage of the process this is crucial.
The relative difficulty in performing MAcAm dechlorination in comparison with the dechlorination of DAcAm is attributed to the relative stabilities of the C-Cl bonds.
It was subsequently decided to check whether increasing the amount of the catalyst or the [BH 4 − ] will enable full dechlorination.

Catalyst Dosing Dehalogenation Experiments
The results presented in Figure 9 indicate that the increase of the catalyst amount suspended in the reaction solution considerably improves the dechlorination process. The catalyst amount introduced in the reaction has considerable impact on the degree of dechlorination. The ability to reduce the yield of MAcAm to 24% in the reduction of DAcAm, Figure 9a, and the dechlorination by 86% of MAcAm, Figure 9b, is significant. A plausible explanation is that with the decrease in the ratio [BH 4 − ]/[catalyst], the charge on the ZVI-NPs decreases, which slows down reactions (4) and (5), which compete with the dechlorination process. Surprisingly, at the high catalyst concentration, some acetic acid is formed during the dechlorination of MAcAm. This contradicts the discussion above. A plausible explanation is that the intermediates formed during the dechlorination of MAcAm are {(ZVI-NP)-H n+m−k }-(CH 2 CONH 2 ) k (n−m)− and not {(ZVI-NP)-H n+m−1 }-CH 2 CONH 2 (n−m)− . The number of -CH 2 CO-NH 2 bound to a given ZVI-NP affects its properties and a smaller number slows down the ZVI-C bond heterolysis facilitating the amide hydrolysis.

NaBH 4 Dosing Dechlorination Experiments
The dependence of the distribution of the products on [BH 4 − ] is presented in Figure 10.  However, here too at high [BH 4 − ], the dechlorination of MAcAm yields some acetic acid. This is probably due to the increase in the number of H-atoms/hydrides, n+m, bound to the same ZVI-NP coherently. The effect of these hydrogens on the properties of the NPs is similar to that of the -CH 2 CONH 2 groups discussed in the previous section.
In an effort to fully dechlorinate DAcAm and MAcAm, a combination of use of high catalyst amounts, 0.70 g of ZVI@ORMOSIL, and high [BH 4 − ]/[substrate] = 40:1 was performed. For DAcAm as substrate, the yields were only 7% of MacAm, 57% of AcAm, and 36% of AA. For MAcAm, the yields were 10% of MAcAm, 78% of AcAm, and 12% of AA, i.e., full dechlorination was not achieved. In a last set of experiments, full dechlorination of MAcAm was achieved by performing an identical experiment but delivering the same amount of NaBH 4 in two doses. The first dose was allowed to react for 15 min and then the second portion was added for another 15 min. In this experiment, the final products consisted of 63% of acetamide and 37% acetic acid, i.e., 100% dechlorination.

Catalyst Stability Test
To check the recyclability of ZVI@ORMOSIL in the dechlorination of DAcAm, 8 consecutive reactions were performed. For each dehalogenation reaction cycle, the same catalyst was utilized. Before each cycle, the matrix was washed with distilled water and dried. No significant effect on the products distribution was observed ( Figure 11).

Preparation of ZVI-NP Suspension
An amount of 0.56 g of 25 nm iron nanoparticles powder was suspended in 10.0 mL of ethanol (AR) resulting in 1.0 M ZVI suspension sealed and stored under nitrogen at 4 • C, equilibrated to room temperature before use.
3.2.2. ZVI@ORMOSIL Synthesis via the Sol-Gel Route 1.0% mol load ZVI@ORMOSIL The catalyst was prepared by using the two steps acid/base sol-gel synthesis route. Briefly, 37% HCl (62 µL, 2.53 µmol) was dissolved in water (2.72 g, 0.151 mol), and the mixture was added slowly into a premixed solution containing MTMOS (1.556 g, 0.011 mol) and TEOS (5.6 g, 0.027mol), which were dissolved in ethanol (7.02 g, 0.152 mol). The resulting mixture was homogenized for 15 min. A 2.0% NH 3 solution (1.5 mL, 7.7 mmol) was then added to the mixture dropwise. When the gelation started, 380 µL of 1.0 M ethanolic suspension of ZVI was added and the mixture was stirred vigorously. The wet black gel was kept for 15 days for aging and drying at room temperature. The solid matrix obtained was crushed with a mortar and pestle into a powder and washed with water several times. The washed matrix was then dried and used for the catalytic tests. A tenfold volume of 1.0 M ZVI suspension was used to obtain 10% mol load ZVI@ORMOSIL catalyst.

Catalyst Characterization
Transmission microscopy images, electron microscopy, and elemental composition were obtained using a 3FE-Tescan ultra high resolution MAIA microscope with an AZTEC microanalysis EDX detector, Oxford Instruments, Colorado, USA. Suspension mean particle size was measured using photon cross correlation spectroscopy (PCCS) with nanophox/R (Sympatec, Germany). The data was analyzed using WINDOX 5. Sample suspensions were sonicated for 10 min in an ultrasonic water bath at room temperature. The measurement was done in a 4 mL plastic cuvette placed in a temperature-controlled water bath. The readings were performed at 632.8 nm. Ethanol (LCMS) RI 1.362, viscosity 1.071 mPas (25 • C).
X-ray diffraction (XRD) measurements were performed with a Bruker (Karlsruhe, Germany) AXS D8 ADVANCE Series II diffractometer equipped with a LynxEye detector (reflection θ-θ geometry, Cu Kα radiation (λ = 0.154 nm), divergence slit 0.60 mm, anti-scattering slit 8.0 mm). Diffraction data were collected in the angular range of 10 • < 2θ < 80 • , step size 0.05 • , and a step time of 0.5 sec/step. N 2 adsorption-desorption isotherms were measured using an APP gold instruments VSorb 2800 model surface area and porosity analyzer, sensitivity 0.010 m 2 /g, range 2-500 nm pore size. Specific surface area, average pore size distribution (PSD), pore volume, and adsorption-desorption isotherms were measured in analyses performed using Brunauer−Emmett−Teller (BET) and Barrett−Joyner−Halenda (BJH) methods on a BET surface analyzer. XPS surface analysis was performed using a Thermo Fisher Scientific (East Grinstead, United Kingdom) NEXSA XPS system with monochromatized Al Kα source (400 micron diameter). Pass Energy ("resolution") of 200 eV was used for survey scans to obtain a general surface composition profile and 50 eV for the high-resolution scans that were used for quantitative analysis.

Catalytic Tests
Dichloroacetamide and monochloroacetamide were taken as model substrates to study the catalytic activity of ZVI@ORMOSIL towards reductive dehalogenation. In a typical catalytic dehalogenation experiment, ZVI@ORMOSIL was weighed (ca. 0.10 g) in a glass vial, 16.90 mg of dichloroacetamide or 12.34 mg of monochloroacetamide (8.8 mM of each substrate) was accurately transferred to the vial. 13 mL of a solvent (deaerated water, ethanol, acetonitrile or 2-propanol) was added, then 0.10 g of NaBH 4 was accurately added to the vessel, composing a relative excess ratio of 1:20 (0.132 mmol of substrate/2.64 mmol NaBH 4 ). Finally, 2.0 mL of deuterated water was added to complete a total reaction volume to 15 mL. In a comparative dehalogenation experiment, half of the amount of each substrate was transferred to the same vial in the same preparation method to obtain 1:1 mole/mole mixture, 8.5 mg of DAcAm, and 6.2 mg of MAcAm, yielding 4.4 mM of each substrate in the final reaction solution. The resulting suspensions were stirred for 15 min, and after that, the catalyst was recovered using filtration, and the filtrate was analyzed using HPLC. The reaction samples, standards, and blanks were diluted and filtered before RP-HPLC monitoring for the degradation of haloacetamides. Ammonia amounts are equimolar to the monitored acetic acid for obvious reasons.
HPLC analysis was performed on a Dionex Ultimate 3000 equipped with a Diode Array Detector by Thermo Hypersil-gold C18 1504.6 mm 3 µm column. Acetic acid, dichloroacetamide, monochloroacetamide, and acetamide were eluted by (0.10% H 3 PO 4 ; ACN), (4:96) mobile phase, 0.6 mL/min flow, 23 • C column temp, with UV detection at 200 nm, 5.0 µL injection volume. The reaction samples were quenched with a few drops of diluted H 3 PO 4 , filtered through 0.22 µm PES or H-PTFE according to the reaction solvent, and all samples were diluted and adjusted to pH 3.0 if necessary, filtered with 0.22 µm PES filter membrane before analysis.
Spectroscopic analysis of ammonia was performed according to a reported procedure [45][46][47][48]. Briefly, 0.10 mL of 1.62 M ethanolic solution of phenol was added to 2.5 mL of the filtered reaction solution. After vigorous stirring, 0.10 mL of 0.50% w/v of hexacyanoferrate trihydrate in water was added. After additional stirring, 0.25 mL of the oxidizing solution comprising alkaline 0.64 M citrate trihydrate and 0.12% NaOCl was added, and the standard solutions were prepared likewise. The samples were covered with aluminum foil and kept in the dark at ambient temperature to let the color develop. Absorbances were recorded at 640 nm in 1.0 cm quartz cuvettes. A Varian Cary UV Bio 50 spectrophotometer with dual-beam, Czerny-Turner monochromator, Xenon pulse lamp single source, and dual Si diode detector was used (Agilent technologies, Middleburg, The Netherlands).

Conclusions
To conclude, a facile room temperature synthesis of 1% load ZVI@ORMOSIL catalyst via the sol-gel route was performed. The interconnected bottleneck constrictions may contribute to diffusion-controlled rates of dehalogenations. ZVI@ORMOSIL heterogeneous robust catalyst exhibits a good extent of dehalogenation of haloacetamides, producing fully dehalogenated reaction products. Thus, zero valent iron ORMOSIL immobilized nanoparticles prove to be a worthy replacement in reactions catalyzed with rare and expensive metals. The stability of M 0 -C formed as an intermediate is a crucial step for determining which solvent mixture may be best for performing the shown reactions.
This report broadens the scope of use of ZVI immobilized in sol-gel as a heterogeneous catalyst for environmental remediation applications. A previously reported use of ZVI as a reductant obtained a noteworthy dehalogenation of haloacetamides, but not to the full dehalogenation extent [25][26][27]. A complete dehalogenation of DAcAm and MAcAm is reported herein, obtaining fully detoxified acetamide and acetic acid products.