Smart hydrogel-microsphere embedded silver nanoparticle catalyst with high activity and selectivity for the reduction of 4-nitrophenol and azo dyes

A simple method is reported for the preparation of silver nanoparticle (AgNP) embedded pH-responsive hydrogel microparticle catalyst via Michael addition gelation and in-situ silver nitrate (AgNO 3 ) reduction. The AgNP-hydrogel microsphere exhibited an efficient reduction of pollutants like 4-Nitrophenol (4-NP) and Congo red (CR) under acidic medium with turn over frequency (TOF) of ~170 h – 1 and ~124 h – 1 respectively. Interestingly, the activity of the catalysts was turned-OFF under a basic medium ( ≥ pH 12) due to the deswelling pH-responsive matrix surrounding the AgNPs. On the contrary, turning-OFF the hydrogenation of a cationic pollutant like methylene blue (MB) using high pH ( ≥ 12) was not possible, due to ionic interaction of MB molecules with the negatively charged catalyst at this pH. This feature was used to demonstrate selective hydrogenation of only MB from a mixture of 4-NP and MB. Finally, five recycling steps confirmed the reusability and practical application potential of the catalyst.


Introduction
Accumulation of degradation-resistant pollutants such as nitroaromatics and azo dyes in water bodies poses a threat to the aquatic ecosystem as well as human health. The high water solubility of certain nitroaromatics like 4-nitrophenol (4-NP) and azo dyes can make their removal quite challenging. This triggered the development of methods based on efficient adsorbents (Yagub et al., 2014;Gupta et al., 2013;Dias and Petit, 2015;Parida et al., 2021), photocatalytic degradation (Dias and Petit, 2015), bio-degradation (Marvin-Sikkema and de Bont, 1994;Ju and Parales, 2010), and catalytic conversion Zeng et al., 2013) as viable purification strategy to maintain good water quality. Among these options, catalytic conversion is the preferred method as it offers a possibility to convert the pollutants to valuable products and less harmful counterparts with high efficiency. For example, upon hydrogenation, carcinogenic 4-NP can be converted to 4-aminophenol (4-AP), which is an intermediate for corrosion inhibitors (Thenmozhi et al., 2014;Guenbour et al., 2000), pharmaceutical molecules, and dyes (Buschmann, 2007). Transforming nitro pollutants to useful amino compounds is a green and commercially beneficial approach to get rid of pollutants.
Among various methods, noble metal nanoparticles (NMPs) catalyzed hydrogenation continues to draw considerable attention owing to their high activity and oxidative stability Zheng and Zhang, 2007;Qin et al., 2019;Lu et al., 2006;Sogukömerogulları et al., 2019;Kästner and Thünemann, 2016). With the rise in demand for active catalysts, the use of noble metals like Au, Pt, and Pd-based catalysts have grown recently Sogukömerogulları et al., 2019;Ansar and Kitchens, 2016;Nguyen et al., 2019;Sun et al., 2014;Johnson et al., 2013;Goepel et al., 2014;. However, the high cost of these metals is a major drawback and thus triggered the development of supports for enhanced recoverability. Supports containing magnetic particle is one such approach for easy recovery and reuse of expensive NMP catalysts (Zeng et al., 2013;Yang et al., 2020;Xu et al., 2020). The development of bimetallic nanoparticle catalysts is another approach to reduce the cost along with improvement in the activity compared to their monolithic counterparts Fu et al., 2018). Despite these attempts, the overall price of such catalysts remains higher than catalysts prepared from metals like Ni, Cu, and Ag.
In this context, MNPs of moderately active and low-cost metals (Cu and Ag) are still attractive choices as catalysts (Li et al., 2015;Zhou et al., 2020;Das et al., 2019;Dong et al., 2014;Qian et al., 2020;Bhaduri and Polubesova, 2020;Sudhakar and Soni, 2018;Budi et al., 2021). To improve the activity of silver nanoparticles (AgNPs), supports have been designed to boost the catalytic activity via increasing the available surface area for catalysis. Using this principle, nanosheets (Li et al., 2015;Qian et al., 2020;Mao et al., 2018), conductive polymers (Das et al., 2019), and fibrous silica (Dong et al., 2014) were employed for the successful enhancement of the catalytic activity of AgNPs. On the other hand, porous supports prepared from carbon and organic polymers are also known to have a positive effect on the catalytic activity of NMPs Bhaduri and Polubesova, 2020;Gong et al., 2019;Xia et al., 2016;Budi et al., 2020). In the case of these supports, confined space reaction and enrichment of micro-environment abound NMPs by adsorption of substrate molecules enhance the activity Gong et al., 2019;Cárdenas-Lizana et al., 2013). Moreover, fast electron transfer from support like carbon black to NMPs is also known to improve catalytic efficiency . Although these supports are known to enhance the catalytic activity of NMP, a portion of the NMP surface is shielded by the support, thus, resulting in their underutilization.
To counter the underutilization of NMPs, relatively mobile crosslinked polymeric networks or hydrogels are investigated as supports (Lu et al., 2006;Li et al., 2010Li et al., , 2011Wang et al., 2010;Irene, 2018;Begum et al., 2019). These supports allow easy access to the NMP surface along with substrate and product exchange. Such polymeric supports also facilitate tuning the catalytic activity via an external stimulus such as temperature, pH, and salt concentration (Lu et al., 2006;Li et al., 2010;Li et al., 2011;Wang et al., 2010;Irene, 2018). However, complicated preparation methods of such stimuli-responsive catalysts make them unattractive (Li et al., 2010;Li et al., 2011). Lack of complete control over activity also hinders the wide acceptability of such responsive catalysts (Lu et al., 2006;Kästner and Thünemann, 2016;Li et al., 2011;Irene, 2018). Considering the potentials of responsive supports, a simplified method needs to be developed to improve the catalytic activity, controllability, and selectivity of embedded NMPs.
Herein, a simple water-in-oil emulsion route is reported for the preparation of AgNP embedded responsive hydrogel microsphere catalyst. Such catalyst was prepared by the emulsification of an aqueous solution of Trivinylphosphine oxide (TVPO), Piperazine, and AgNO 3 followed by simultaneous Michael addition crosslinking of TVPO-Piperazine and simultaneous in-situ formation of AgNPs within the emulsified micro-droplets. The composite microspheres were characterized by X-Ray diffraction (XRD), Scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS). Dynamic light scattering (DLS) analysis confirmed the pH-responsive behavior of the composite hydrogel-microsphere. pH-responsive swelling-deswelling of the composite microsphere was utilized to control the access to AgNPs and modulate their catalytic activity (Fig. 1). This feature was investigated using 4-NP, Congo red (CR), and Methylene blue (MB) as model pollutants. Switching ON-OFF of the catalytic activity was demonstrated by changing the pH of the reaction medium. Substrate selectivity of the novel catalyst was also investigated using a mixture of 4-NP and MB. Finally, the reusability of the catalysts was demonstrated to highlight their practical application potential.

Hydrogel preparation
Hydrogel was prepared by Michael addition of TVPO (64.0 mg, 0.50 mmol) and piperazine (64.6 mg, 0.75 mmol) in 2.5 mL of water (40 • C for 16 h) (Nazir et al., 2020b). pH-responsive swelling of the transparent hydrogel was determined by measuring the swelling ratio (SR) by the procedure reported in our previous publication (Nazir et al., 2020b) and details can also be found in Sec. S1.1.

In-situ AgNP-hydrogel preparation
2, 5, and 10.5 mol% of AgNO 3 was added to 2.5 mL aqueous solution of TVPO (64.0 mg) and piperazine (64.6 mg) kept in an ice bath. The solutions were then transferred to cuvettes, sealed, and transferred to an oven (at 40 • C). All operations were carried out in dark and separate cuvettes were used for each duration. UV-vis spectra of solutions were recorded at intervals.

Amine mediated in-situ method (Method 1)
For the preparation of the responsive catalyst, the process was developed to prevent the AgNP formation on the surface of microspheres. Therefore, AgNO 3 was not premixed with the Michael adducts, rather it was added after the initiation of gelation. As shown in Fig. 2a, 0.5 g of Span 80 was mixed with 20 mL of cyclohexane using Ultraturax (19,000 rpm, 5 min). Then, a freshly prepared 5 mL aqueous solution of TVPO (128 mg, 1 mmol) and piperazine (129.2 mg, 1.5 mmol) was added dropwise to this mixture under stirring (19,000 rpm) to obtain a milky emulsion. Then, the centrifuge tube was covered with aluminum foil and transferred to a water bath under stirring (40 • C, 500 rpm). After 15 min, 0.1 mL aqueous solution of AgNO 3 (48 mg/mL) was added to the emulsion and stirred for 16 h to obtain a light brown emulsion (Fig. S2). Then, the emulsion was dialyzed (24 h) using ethanol as dialysate and 15 KD RC dialysis tubes. The dialysate was replaced 3 times followed by dialysis in deionized water (24 h) to replace ethanol with water. This sample is named E1 (Fig. 2a) and the emulsion prepared without the addition of AgNO 3 is named E0.

H 2 -assisted in-situ method (method 2)
Emulsion of TVPO (128 mg, 1 mmol) and piperazine (129.2 mg, 1.5 mmol) was prepared in a pressure reactor by the procedure as described in Section 2.3.1, and the pressure reactor was transferred to a water bath at 40 • C under stirring (15 min). 0.1 mL aqueous solution of AgNO 3 (48 mg/mL) was added to the emulsion and stirred for 2 h (Fig. 2b). Then, the reactor was pressurized with H 2 (~2.5 bar) and the stirring was continued for 14 h at 40 • C. The emulsion was then purified by dialysis as reported in Method 1 and the sample was named E2.

Fig. 1.
Synthesis of AgNPs few nanometers below the surface hydrogelmicrosphere and controlling the access to AgNP surface by swelling-deswelling.

Characterizations of composite microspheres 2.4.1. UV-vis analysis
UV-vis spectra for in-situ AgNP formation and catalytic reaction were recorded using a Varian Cary 50 UV-Vis Spectrophotometer. Catalytic reduction of 4-NP, MB, and CR were analyzed by recording the UV-vis spectra of the reaction solution as a function of time and the residual concentration was determined using their corresponding UV-vis calibration curves.

Inductively coupled plasma optical emission spectrometry (ICP-OES) analysis
ICP-OES analysis was used to determine the silver content (Ag-content) of samples. ICP-OES 5110 (Agilent, Basel, Switzerland) apparatus was used for these experiments. Samples preparation for ICP-OES consisted of mixing 10 mg of the sample with 3 mL HNO 3 , followed by the digestion at 250 • C for 30 min using microwave heating.

XRD analysis
XRD analysis was carried out in a Stoe IPDS-II instrument, operating at a voltage of 50 kV and a current of 40 mA with Mo Kalpha radiation (λ = 0.71073 Å) at an angular range (2θ) of 5-50 • . The instrumental contribution was taken into consideration by measuring the diffraction pattern of LaB6 as the reference material and used within Topas software (Coelho, 2018).

DLS analysis
DLS analysis was carried out in a Malvern Zetasizer Nano ZS90 at 25 • C to determine the particle size of hydrogel-microspheres. Before analysis, the pH of the dispersion was adjusted to the desired value and kept for 1 h at 25 • C to achieve an equilibrium swelling. The same samples were used to measure the ζ-potential using folded capillary Zeta Cell DTS1070.

SEM analysis
SEM analysis was carried out in a Hitachi S-4800 SEM equipment operating in scanning and transmission mode (30 kV). For SEM images, samples were prepared by putting a drop of emulsion on a silicon wafer and evaporating the water at room temperature for 16 h followed by 7 nm Au/Pd coating. For transmission images, the sample was prepared by putting a drop of emulsion on a lacey carbon grid and evaporating the water over 16 h at room temperature.

XPS analysis
XPS analysis was carried out on a Physical Electronics (PHI) Quantum 2000 X-ray photoelectron spectrometer equipped with a monochromatized AlKα source (at 15 kV, 28.8 W), and a hemispherical electron energy analyzer fitted with a channel plate and a positionsensitive detector. The sample was analyzed with an electron take-off angle of 45 • and spectra were recorded with constant pass energy mode (46.95 eV and energy resolution of 0.95 eV). Spectra were processed with PHI MultiPak.

4-NP hydrogenation
14.4 mg of 4-NP (0.1 mmol) and 35.4 mg of NaBH 4 (0.94 mmol) were dissolved in 12.0 mL water (40 • C) maintained at different pH. The reduction was initiated by adding 1.5 mL of E2 (at the same pH) to the freshly prepared 4-NP-NaBH 4 solution under stirring (500 rpm). The reduction of 4-NP was monitored by measuring the UV-vis intensity of 4-NP centered around 400 nm after required dilution. Separate samples were prepared for each duration. TOF of the reaction was calculated as the moles of 4-NP reduced by a mole of Ag-atom in an hour.

CR and MB hydrogenation
For MB hydrogenation, 60 μL of MB solution (10.0 g/L) and NaBH 4 (35.4 mg, 0.94 mmol) were added to 12.7 mL of water at desired pH. Then, 0.75 mL of E2 was added to the solution and the reaction was monitored by UV-vis spectroscopy after required dilution.
For CR reduction, 60 μL of CR solution (20.0 g/L) and NaBH 4 (35.4 mg, 0.936 mmol) were added to 12.75 mL of water at desired pH followed by the addition of E2 (0.75 mL) under stirring. Then the catalytic reduction of CR was monitored by UV-vis spectroscopy. Both MB and CR reduction were carried out at 40 • C and separate samples were prepared for each duration.

In-situ AgNP-hydrogel composites preparation
The hydrogels were synthesized via previously reported Michael addition crosslinking of TVPO and piperazine (at 40 • C) in presence of 2, 5, and 10.5 mol% AgNO 3 (Nazir et al., 2020a(Nazir et al., , 2020b. Colorless solutions turned dark red with time, indicating the formation of AgNPs ( Fig. 3 and S1). The appearance of a UV-vis band of nanosilver at λ max 380 nm just after 5 min confirmed the formation of Ag 0 nuclei at the early stage (Fig. S1d). The redshift of the band with time indicated the growth of nuclei to AgNPs (Agnihotri et al., 2014;Zhao et al., 2013). The presence of AgNO 3 led to faster gelation (Fig. S1a-c) and the solution containing 10.5 mol% of AgNO 3 formed the gel within 45 min compared to 16 h for pristine hydrogel (at 40 • C). Fast gelation in the case of precursor solution containing AgNO 3 can be attributed to physical reasons. In-situ nanoparticle formation (Mishra et al., 2014;Hoogesteijn von Reitzenstein et al., 2016) and crosslinking of precursors lead to a rapid rise in viscosity compared to pristine precursor solution. A combination of these factors reduces the flow behavior and the solutions behave like a hydrogel.
As explained earlier (Fig. 1), the swelling-deswelling behavior of the composite hydrogel is important to control the catalytic activity. As a measure of responsiveness, the swelling ratio (SR) of AgNP-hydrogel composites was determined (Eq.S1, Sec. S1.1). From Fig. 3b it can be seen that a slight decrease in deswelling was observed in composites prepared with 1.0 and 2.0 mol% of AgNO 3 . However, composite prepared with 10.5 mol% AgNO 3 displayed only limited deswelling, which can be attributed to the formation of the rigid matrix due to the presence of a large number of AgNPs. Finally, gel-phase NMR analysis of composite hydrogel (Fig. S2) confirmed the absence of any side reaction in presence of AgNO 3 . Details of NMR analysis can be found in Sec. S2.
Direct synthesis of AgNP-hydrogel composite from the precursors (TVPO and piperazine, silver salts) has simplified the catalyst preparation. The use of bulk composite as a catalyst can lead to poor catalytic activity due to diffusion limitations. To overcome such drawbacks and facilitate easy access to catalytic sites, the AgNP-hydrogel composite was prepared in the form of microspheres by an emulsion process (Fig. 2). Considering moderate Ag-content and minimum loss in swelling behavior, hydrogel-microspheres were prepared using 2.0 mol% of AgNO 3 .

Preparation of AgNP embedded hydrogel-microspheres
The absence of AgNPs on the surface of the microsphere is important to achieve complete control over the catalytic activity of AgNPs by swelling-deswelling. Therefore, the method was developed to prevent the formation of AgNPs on the surface of hydrogel-microspheres. Particularly, premixing of AgNO 3 with the Michael addition precursors was avoided, rather AgNO 3 was added to the emulsion after initiation of gelation within droplets to facilitate the formation of AgNPs a few nanometers below the surface. Hydrogel-microspheres without AgNPs (E0) and with AgNPs (E1) were prepared by this method (Section 2.3.1). The light brown color of E1 indicates the presence of AgNPs within microspheres (Fig. S3). The solid content of the purified E1 in water was found to be 0.9% with an Ag-content of ~0.35 wt% on dried E1 (Table S1), which is significantly lower than the calculated Agcontent of 1.1%. This can be attributed to the fast gelation of Michael precursors within the droplets. This leads to the formation of tertiary amines before the complete reduction of AgNO 3 . The reduction potential of tertiary amines is known to be lower than the secondary amines (Piao et al., 2011). As a result, unreduced AgNO 3 is removed from E1 during the subsequent purification step.
To enhance the Ag-content in the hydrogel-microspheres, H 2 was introduced in the reactor (~2.5 bar) 2 h after the addition of AgNO 3 (E2, Fig. 2b), and the emulsion was stirred for a further 14 h. H 2 -assisted method (E2) resulted in a much darker emulsion than E1 (Fig. S3), indicating higher Ag-content in E2. UV-vis spectra of purified E0, E1, and E2 were recorded after equal dilution in water (Fig. 4a). In the case of E0, UV-absorption was observed only below 320 nm, which is consistent with our previous report (Nazir et al., 2020b). This makes easy detection of embedded AgNPs within E1 and E2 hydrogel microspheres. The absorption band of AgNPs was visible at ~415 nm Sirohi et al., 2019), owing to the transparency of the hydrogel matrix within this wavelength range. The higher intensity of the band at 415 nm for E2 indicates a higher concentration of AgNPs within microspheres.
The formation of AgNPs within microspheres was also confirmed by powder XRD of dried E1 and E2 (Fig. 4b). Diffused diffraction pattern of hydrogel (E0) was present in the XRD patterns of both E1 and E2. Diffraction peaks at 2θ of 17.7 • , 20.4 • , 29.0 • and 33.5 • in case of E1 and E2 can be assigned to (111) Fig. S1. (b) Swelling ratio of pristine and composite hydrogels prepared using different AgNO 3 loading. centered cubic Ag 0 crystals . The size of AgNPs determined using 111-plane was found to be ~8 nm for both E1 and E2. Then, Ag-content in dried E2 was found to be 0.7 ± 0.04 wt%, which is close to the calculated value (i.e. 1.1%, Table S1). Therefore, only E2 was selected for further characterization and catalytic study. From the solid content and ICP-OES analysis (Table S1), the Ag-content of E2 was calculated to be 0.059 mg/mL.

Characterization of AgNP embedded hydrogel-microspheres
The pH responsiveness of E2 was determined by DLS analysis at different pH values. At pH 4, the particle size was 980 nm and ζ-potential of + 34 mV indicates its good dispersion stability at this pH (Fig. 5a, Table 1). A decrease in the particle size (807 nm) and ζ-potential (+13 mV) was observed with an increase in the pH to 7. This led  to poor dispersion stability of E2 and precipitation after ~2 h (Fig. S4). To compare the swelling behavior of E2 with the bulk composite hydrogel, the ratio of particle volume at pH 4 and pH 7 (V pH4 /V pH7 ) was calculated using the average particle size obtained in DLS experiments. V pH4 /V pH7 of E2 was found to be 1.8, which is in agreement with the ratio of SR of the bulk hydrogel at pH 4 and pH 7 (SR pH4 /SR pH7 =1.9).
Increasing the pH to 10 resulted in a decrease of the ζ-potential to − 6 mV (Table 1), indicating the unstable nature of the dispersion (Fig. S4), and two populations of particles were observed during DLS analysis (Fig. 5a). Although the decrease in particle size of E2 is expected at pH 10, the larger particle size observed at this pH is due to the aggregation of smaller particles. Two particle populations make it difficult to compare the change in volume at pH 10 ( Fig. 5a). Further increasing the ≥ pH 12, an extensive aggregation was observed with unreliable DLS results.
SEM analysis of E2 was carried out both in scanning and transmission mode to determine the location of AgNPs in microspheres. As expected, no sign of AgNPs was observed during SEM analysis of E0 (Fig. 5b). SEM images of E2 showed the presence of AgNPs a few nanometers below the surface of the microsphere (Fig. 5c). Transmission images also confirmed the presence of AgNPs (13 ± 3 nm, Fig. 5d, and e) few nanometers within the microsphere. The size of AgNPs determined by transmission images was found to agree with XRD analysis (~8 nm). It is worth mentioning here that, absence of AgNPs on the surface of microspheres prevents any catalytic activity at its deswollen state, which offers excellent control over the activity just by swelling and deswelling of the E2.
XPS analysis was carried out to determine the oxidation state of the silver in E2 (Sirohi et al., 2019;Ruíz-Baltazar et al., 2018;Cheng et al., 2015). The high-resolution scan in Fig. 5f shows the presence of two peaks between 365 and 377 eV due to spin-orbital splitting to Ag 3d 5/2 and Ag 3d 3/2 core levels. Both the peaks can be resolved into two components, with major peaks at 368.2 (3d 5/2 ) and 374.2 eV (3d 3/2 ) assigned to Ag 0 . The small peaks at 369.6 and 375.3 eV were assigned to oxidized silver (Ag + ). Based on the peak area, the Ag 0 component was found to be ~95%. Analyzing the N1s spectra, an overall shift towards lower binding energy was observed in E2 as compared to E0 (Fig. 5g). This can be attributed to an increased number of secondary amines in E2 than E0, which is in agreement with the NMR analysis. Analysis of N1s spectra confirms the absence of any AgNPs-hydrogel chemical interactions, and AgNPs are physically stabilized within the hydrogel.

Reduction of 4-NP
To study the possible reduction of 4-NP in the absence, aqueous solutions of 4-NP and NaBH 4 at different pH were prepared (Section 2.5.1) and the solutions were monitored by UV-vis spectroscopy (Fig. S6a). No change in the intensity of 4-NP absorption peak (at 400 nm) was observed over time. This indicates a lack of 4-NP reduction in absence of E2 at all pH ranges (Table S2). Adding the required quantity of E2 to the reaction solution, the yellow color started to disappear (Fig. 6a) along with a decrease in the intensity of the 4-NP UV-absorption peak and the appearance of a new band at ~300 nm corresponding to 4-AP (Fig. 6b). The conversion and rate of reaction were determined from the intensity of the UV-absorption peak at 400 nm and given in Fig. S6b and Fig. 6c. C 0 and C are the initial and residual concentrations of 4-NP at a given Table 1 Particle size and ζ-potential of E2 at different pH. pH Particle size (nm) ζ-potential (mV) 4 980 +34 7 807 +13 10 --6 12 --Note: blank space in case of pH 10 and 12 indicate unreliable values due to precipitation of microspheres. time, respectively. Linear correlation between − ln(C/C 0 ) vs. time indicates the pseudofirst-order reduction of 4-NP at all pH values (Fig. 6c), with a visible effect of pH on the rate of hydrogenation of 4-NP (Fig. 6c). The highest k value at pH 4 (0.11 min − 1 ) followed by a decrease with an increase in pH (Fig. 6c) indicates the effect of deswelling of hydrogel matrix surrounding the AgNPs. Increasing the pH ≥ 12, a very low k value (0.0006 min -1 ) was observed, which signifies the lack of 4-NP hydrogenation ( Fig. 6c and S6b). These observations, highlight the ability to control the activity of AgNPs by controlling the swelling of the microspheres. TOF of the 4-NP hydrogenation (at pH 4) was determined at different reaction duration (Fig. S6c) and was found to decrease with time, which is in agreement with earlier reports (Kozuch and Martin, 2012). TOF at full conversion was found to be ~170 h − 1 and ~100 h − 1 for pH 4 and 7 respectively ( Table 2). TOFs achieved by E2 at full conversion are higher than the recently reported state of art Au Fu et al., 2018) and Ag Mao et al., 2018) based catalysts.
Based on the swelling behavior of composite microsphere and catalytic activity at different pH, a mechanism to explain the tunable behavior is proposed in Fig. 7. As observed during DLS analysis, E2 reached a swollen state around pH 4, which coincides with the pH of the highest catalytic activity. This signifies the easy accessibility of AgNPs for the hydrogenation of 4-NP. Positive ζ-potential of E2 at this pH (Table 1) also favors the adsorption of 4-NP and BH 4 on the catalyst. As a result, the local concentration of 4-NP and BH 4 around the microenvironment of AgNPs remains high and facilitates high catalytic activity. A similar strategy has been used to enhance the catalytic activity of Palladium/MOF catalysts for styrene hydrogenation (Huang et al., 2016). Additionally, pores of the hydrogel can also act like nano-reactors to enhance the activity via the well-known confined space reaction (Fig. 7) (Gong et al., 2019;Cárdenas-Lizana et al., 2013). A combination of these factors led to fast hydrogenation of 4-NP.
Increasing the pH of the reaction medium decreased the access to AgNPs due to the deswelling of the E2 (Fig. 7a). The decrease in ζ-potential (Table 1) with pH also reduces the local concentration of 4-NP. As a result, the catalytic reduction of 4-NP slowed down. Complete deswelling along with negative ζ-potential of E2 above pH ≥ 12 block the access to AgNPs and, prevents any catalytic reduction (Figs. 6c, 7a).

Reduction of CR and MB
E2 was also used for the reduction of pollutants of anionic (Congo red) and cationic (methylene blue) nature. The reduction pathway of carcinogenic azo dye like CR by AgNPs in the presence of NaBH 4 and AgNPs is well established and shown in Fig. S7a (Rajesh et al., 2014;Nasrollahzadeh et al., 2020). The residual CR in water is quantifiable by recording the UV-vis intensity of CR solution at different times ( Fig  S7b). Hydrogenation of CR under acidic medium (pH 4) was fast (k = 1.8 min -1 ) with a TOF of ~124 h -1 (Fig. 8a, Table 2) highlighting the excellent activity of E2. The activity decreased with an increase in the pH and activity was turned OFF at pH 13 (Fig. 8a, S7c). Such behavior resembles the reduction of 4-NP and can be explained by Fig. 7a. In our previous study, we have shown that under an acidic medium this hydrogel favors adsorption of anionic dyes (Nazir et al., 2020b). Changing the pH to alkaline results in the repulsion of dye molecules from the hydrogel matrix (Nazir et al., 2020b). This behavior favors the adsorption of CR and faster degradation under an acidic medium. Repulsion between CR and hydrogel microsphere along with deswelling of E2 prevents any degradation above pH 12.
During MB reduction, the highest catalytic activity was observed at pH 4 (Fig. 8b, S8) with a k value of 0.52 min -1 , which is lower than the kvalue observed during the reduction of 4-NP and CR. It is due to the decreased adsorption of cationic MB by the positively charged E2 (Fig. 7b). As a result, the TOF of 27 h -1 was observed during MB reduction at pH 4 (Table 2). On the contrary to 4-NP and CR hydrogenation, it was not possible to completely stop the hydrogenation of MB at pH 13. This can be explained as the adsorption and slow diffusion of MB molecules to reach AgNPs for catalysis owing to the negative surface charge of E2 at pH 13 (Figs. 7b and 8b).

Selectivity and reusability
Differential catalytic reduction of cationic and anionic molecules by E2 at pH ≥ 12 prompted the investigation of the substrate selective activity of the novel catalyst. A mixture of MB to 4-NP (3 × 10 -4 mmol) maintained at pH 13 was subjected to hydrogenation. The green color of the solution slowly turned yellow (Fig. 9c), suggesting the selective reduction of only MB. UV-vis spectra of the solution also confirmed the hydrogenation of only MB (Fig. 9a, b). Interestingly, the concentration of 4-NP remained unchanged even after 3 h. This can be attributed to the selective and competitive adsorption of cationic MB molecules on E2, due to the negative surface charge at this pH.
To demonstrate the practical applicability, the catalyst E2 was subjected to reuse cycles using 4-NP as a substrate. E2 maintained its activity even after the fifth cycle (Fig. 9d) and activity could be turned OFF at the third cycle by raising the pH to 12. E2 recovered its activity in the fourth cycle (at pH 4.5), highlighting the reversibility of its activity. No loss in catalytic activity after the fifth reuse cycle indicates easy recovery of E2 and low loss of silver. Ag-content of 0.65 ± 0.06% determined by ICP-OES analysis after the fifth cycle confirmed a very low Ag loss (Table S1) and minimizes the risk of secondary pollution due to silver leaching from E2. Transmission images indicate hydrogel matrix successfully prevents agglomeration of AgNPs (Fig. S9), although some degree of agglomeration of hydrogel microspheres can be observed in Fig. S9. Agglomeration of microspheres can take place during sample preparation on a TEM grid or during catalyst recovery. From the catalyst reuse study, it is clear that such agglomeration is temporary and has no effect on its catalytic activity.
TOF of the novel catalyst (E2) calculated after 4-NP reduction (at pH 4) is superior to most of the recently reported AgNP-catalysts (Table 3, Entry 2-6). In the case of entry 7, excellent catalytic activity was  reported for AgNP based catalysts, which is attributed to their ultrafine particle size and porous support. The catalyst prepared from an alloy of gold and silver also displayed lower TOF than E2 (entry 8). AuNP based catalyst reported in entries 9 and 10 displayed significantly higher activity than E2, because of porous support that provides easy access to AuNPs and inherently high activity of AuNPs. Interestingly, E2 requires a significantly lower quantity of NaBH 4 to achieve such high TOF compared to highly active catalysts listed in Table 3. Additionally, ease of controlling the activity and catalytic selectivity are distinct advantages of E2 compared to the reported catalysts.

Conclusion
In summary, this work demonstrates a simple method to prepare AgNP based pH-responsive catalysts with excellent catalytic reduction of pollutants like 4-NP and dyes (CR and MB). The pH-responsive hydrogel support offers the possibility to control the access to AgNPs and the micro-environment around them, thereby tuning the catalytic activity. At its swollen state, AgNP-hydrogel-microsphere (E2) displayed an excellent catalytic reduction with TOF of 170 h -1 and 124 h -1 for 4-NP and CR respectively. Increasing the pH of the reaction media resulted  in a decrease in activity and was turned OFF once the pH was above 12. The composite catalyst also displayed a selective reduction of only MB at pH 13 from a mixture of MB and 4-NP, owing to the selective adsorption of MB (cationic pollutant) due to negative ζ-potential at this pH. Successful recyclability of this novel catalyst makes it suitable for various practical applications. Additionally, the substrate selectivity of this catalytic system can be utilized for further development of selective catalysts using other metals.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Table 3
The catalytic activity of the various catalyst for reduction of 4-NP to 4-AP.   TOF = moles of 4-NP hydrogenated by one mole of Ag-atoms in 1 h. Note: TOF was calculated using the data provided in the literature and at the completion of 4-NP reduction. a PCP =multi-amino poly(cyclotriphosphazene-co-polyethyleneimine) (PCP) microspheres.