Effective Adsorption of Precious Metal Palladium over Polyethyleneimine-Functionalized Alumina Nanopowder and Its Reusability as a Catalyst for Energy and Environmental Applications.

Palladium is one of the widely used precious metals toward catalysis, energy, and environmental applications. Efficient recovery and reusability of palladium from the spent catalysts is not only highly desirable for sustainable industrial processing but also for preventing environmental contamination. Here, we present a facile citrate-mediated amine functionalization of alumina nanopowder (AO) in aqueous medium. The surface functionalization is probed using infrared (IR), X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis, surface area, and zeta potential measurements. The amine-functionalized sorbent is thoroughly studied for its vital palladium-sorption parameters such as amount of adsorbent, pH, adsorption capacity, thermodynamics, and kinetics. The palladium adsorption over amine-functionalized AO is further characterized with X-ray diffraction and XPS. IR analysis of palladium adsorbed over polyethyleneimine is performed to elucidate the mechanistic insight on the role of nitrogen in capturing palladium. The amine-functionalized sorbent after adsorbing palladium is studied for the catalytic reduction of 4-nitrophenol and Cr(VI) and hydrogen generation from ammonia borane, which demonstrated its excellent catalytic activity and reusability toward energy and environmental applications. The environmentally benign materials and all-aqueous reactions employed in this work demonstrate the potential of the strategy for efficient and economical industrial transformations and waste-stream management.

Introduction

Palladium is one of the important heavy metals that find extensive usage in various catalytic applications such as coupling, hydrogenation, hydrogen generation, oxidation, decarbonylation, and so forth.[1−4] Several organic transformations that find widespread implications in pharmaceutical and other chemical industries primarily use palladium as one of the main catalysts. Albeit enormous efforts to find alternatives are being pursued, palladium is still considered to be a dominant candidate in many catalytic reactions till date. Some of the limitations of palladium include its high cost, environmental toxicity at higher concentrations, and limited natural availability.[5,6] From both environmental and cost perspectives, it is necessary to bring the amount of palladium in the final effluents less than 5 mg L–1.[5,6] Although toxicity of palladium is not substantial, at higher concentrations, the kidney tissues are more likely to be affected and recent facts testify that it perturbs the mitochondrial respiratory chain, leading to cell death by depleting cellular glutathione levels.[7] Also, palladium has the ability to be transported through plant roots and later could enter into the food chain.[8] Thus, it is imperative to effectively utilize the available palladium in a sustainable fashion and recover it successfully. The recovery of palladium could be accomplished through liquid–liquid extraction and solid-phase extraction processes. Among these, the solid-phase extraction is known to be an economical and effectual strategy for the recovery of palladium and other metal species.[9−11]

Development of efficient adsorbents to recover palladium from the spent catalysts and industrial wastes provides economical and sustainable opportunities. In this regard, activated carbon and biopolymer-modified activated carbon are reported for palladium and platinum removal.[12] Precious metals such as Au(III), Pd(II), and Pt(IV) are shown to be well-adsorbed onto the native graphene oxide.[13] The interaction of sulfur ligands such as mercaptobenzothiazole[14] and mercaptobenzimidazole[15] with biopolymers such as cellulose and chitosan shows good ability to adsorb and to recover palladium. Impregnation of ionic liquid (Aliquat-336) onto a mesoporous silica matrix (SBA-15) is another versatile adsorbent that has been utilized to recover palladium from an industrial catalyst.[16]

Recently, there is an increasing trend in using amine functionality such as polyethyleneimine (PEI) toward environmental remediation as it is a biocompatible polymer that has been widely used in various biological applications including tissue engineering and drug delivery.[17−20] While the linear PEI is mainly consisted of secondary amines, the branched PEI possesses primary, secondary, and tertiary amine groups. Because of the abundant nitrogen, PEI is also known for its polycationic behavior. On a different note, alumina is an inert, cost-effective, and non-toxic substance. The surface of alumina can be easily functionalized with amines, carboxylic acid, thiols, and so forth to obtain the desired property. The commercial availability of high surface area alumina nanopowder (AO) makes it interesting to be used as a cost-effective support.[21] PEI coated onto alumina was used as a sorbent with an adsorption capacity of 13 mg g–1.[22] PEI-coated polysulfone/Escherichia coli has been studied recently as a biosorbent for the sorption of Pd(II). Here, polysulfone was used as a polymer matrix for immobilization of E. coli biomass and to enhance the adsorption capacity.[23] Polyallylamine hydrochloride-modified E. coli also has been studied for its Pd(II) sorption efficiency.[24] The utility of PEI-modified Corynebacterium glutamicum biomass as a sorbent[25] has been reported for palladium recovery. Alumina combined with (5-bromo-2-pyridylazo)-5-diethylaminophenol shows good potential for the adsorptive removal of Pd(II) from different water samples with the sorption capacity of 11.0 mg g–1.[26]N,N-Bis(salicylidene)1,2-bis(2-aminophenylthio)ethane anchored over mesoporous silica was found to be efficient in adsorbing and sensing Pd(II).[27] Amine-functionalized TiO2 nanofibers obtained through electrospinning has been found to effectively perform as a membrane for recovery of precious metals such as Pd, Pt, and Rh.[28] Palladium supported onto amine-functionalized montmorillonite has also been used as a catalyst toward regioselective synthesis of aurones and flavones.[29] In this work, we chose high surface area AO and functionalized it with PEI. It is with the aim that the nitrogen moieties present in PEI can effectively interact with palladium, thereby leading to efficient palladium recovery and its reusability in catalytic applications. Furthermore, the surface functionalization and the catalytic applications were carried out entirely in aqueous medium, thereby offering an environmentally benign methodology.

Results and Discussion

The functionalization of the alumina nanosupport (AO) with PEI was accomplished in a two-step synthesis, as depicted in Scheme . First, citrate was chemically anchored over alumina nanosupport, followed by acidification with acetic acid to achieve carboxyl surface functionalization. It is envisaged that one of the carboxylate arms of the citrate gets anchored onto the alumina surface, whereas the remaining two carboxylates are free that can be converted to carboxylic acid upon acidification. In the second step, these carboxyl groups on the surface of AO were activated using well-known 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)/N-hydroxysuccinimide (NHS) coupling reagents, which facilitated covalent immobilization of PEI through amide bond formation. This was accomplished by the dropwise addition of citrate-functionalized AO (CA-AO) suspension to the PEI solution. The surface functionalization was monitored through infrared (IR), X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis (TGA), surface area, and zeta potential measurements.

Scheme 1

Pictorial Representation of Citrate-Mediated Amine Functionalization of AO

The IR measurements of pristine AO, AO treated with sodium citrate (SC-AO), CA-AO, and PEI-anchored AO (PEI-AO) were recorded to follow the surface functionalization, and the spectra are plotted in Figure . As seen from the figure, the pristine AO exhibited a broad band in the range of 400–1000 cm–1. A peak at 1636 cm–1 was found in the commercial AO, which may be attributed to the AlOOH layer on the surface.[30] After functionalization with SC (before acidification), the obtained SC-AO exhibited characteristic asymmetric νas(COO–) and symmetric νs(COO–) vibrational stretching frequencies of surface-bound carboxylate at 1638 and 1401 cm–1, respectively. The separation of 237 cm–1 between these two peaks indicates the bridging type of chelation of the −(COO–) groups over the adjacent metal centers in the AO support. After acidification, the peak maxima in CA-AO are observed at 1629, 1469, and 1417 cm–1. The narrowing of separation between the asymmetric νas(COO–) and symmetric νs(COO–) with an additional peak formation at 1468 cm–1 indicates successful acidification.[30−32] After immobilizing PEI onto CA-AO, the νas(COO–) and νs(COO–) stretching peaks were found to be slightly shifted to 1624 and 1405 cm–1, respectively. This is indicative of amide bond formation through free −COOH groups, whereas the remaining carboxylate groups remain chelated to the AO support. The new peak at 1574 cm–1 due to N–H stretching and a weak shoulder at 1075 cm–1 due to C–N stretching of PEI indicate that the surface is covered with amine functionalities because of the reaction between PEI and the alumina surface.[33] One may not exclude the possibility of PEI getting coated over CA-AO through electrostatic interaction. Thus, PEI could be immobilized partially through amide bond as well as electrostatic interaction over CA-AO.[22]

Figure 1

FT-IR spectra of the pristine TSC, AO, SC-AO, CA-AO, and PEI-AO.

To corroborate the observation from IR measurements, we performed XPS measurements on CA-AO and PEI-AO (Figure ). The survey scan spectrum of CA-AO revealed that the presence of characteristic Al 2p, Al 2s, C 1s, and O 1s peaks at the binding energies correspond to 73.5, 118.2, 285.0, and 530.7 eV, respectively.[34] The C 1s narrow scan further revealed that a significant portion of the carbon was possessing higher binding energy because of the bonding with the electronegative oxygen atoms. The deconvolution of the C 1s narrow scan showed a peak at 288 eV that can be attributed to the −C=O(O) of the carboxyl group.[35] This clearly indicated the presence of carboxylic acids on the surface of AO. The survey scan spectrum of PEI-AO, in addition to the peaks observed with CA-AO, revealed an additional N 1s peak at 399.1 eV, which was further resolved in the narrow scan. These observations additionally substantiated the successful surface functionalization of AO with PEI.

Figure 2

XPS survey scan measurements on CA-AO and PEI-AO (A), C 1s narrow scan measurements on CA-AO and PEI-AO (B), and N 1s narrow scan measurement on PEI-AO (C).

The TGA of pristine AO, CA-AO, and PEI-AO was performed to understand the degradation profile of the organics present in the functionalized alumina nanoparticles (Figure ). The pristine AO exhibited a gradual mass loss up to 600 °C, the temperature at which the total mass loss was found to be ∼6%. The mass loss study of the functionalized materials revealed that the CA-AO possessed ∼11% of organics, whereas the amount of organics present in the PEI-AO was in the range of ∼15%. In both the cases, the major degradation started from ∼250 °C, which was nearly completed at ∼600 °C. The difference in mass loss between the pristine AO and the functionalized AO indicates the increment in the overall proportion of organics with each functionalization step. Thus, the final proportion of citrate and PEI in the PEI-AO is ∼5 and ∼4%, respectively.

Figure 3

Mass loss study on AO, CA-AO, and PEI-AO using thermogravimetric analysis.

The particle size analyses of AO, CA-AO, and PEI-AO were performed by the dynamic light scattering (DLS) technique to understand the effect of surface functionalization in the particles’ hydrodynamic radius. As seen in Figure , the pristine AO possessed an average size of ∼80 nm (with the peak maxima of ∼72 nm), which was increased to ∼135 nm for CA-AO. This increase in particle size indicates some aggregation among the CA-AO. An additional peak centered at ∼420 nm indicates that a significant population of the particles is in the aggregated form possibly because of the interaction between surface-bound carboxylic groups. After functionalization of CA-AO with PEI, the hydrodynamic radius of the resultant PEI-AO was found to be ∼79 nm, which indicates that the particles regained their original surface after this step. Zeta potential measurements of these materials were carried out to understand the change in the surface charge with each surface functionalization (Figure S1). It is well-known that the high acidic conditions would protonate the surfaces that would give rise to positive zeta potential, whereas the high alkaline conditions would render the surfaces negatively charged and therefore lead to negative zeta potential. As expected, in all three samples, at ∼pH 2, the surfaces possessed a positive zeta potential that was switched to negative at ∼pH 12; however, their respective magnitudes were found to be different, additionally corroborating the functionalization. The isoelectric pH values of pristine AO, CA-AO, and PEI-AO were found to be 6.3, 6.5, and 7.3, respectively.

Figure 4

Size distribution analyses of AO, CA-AO, and PEI-AO using DLS.

The Brunauer–Emmett–Teller (BET) surface area of the pristine AO, CA-AO, and PEI-AO was measured to follow the change in surface area as a function of surface functionalization. The surface area of the pristine AO was found to be ∼141 m2/g, whereas that of the CA-AO was found to be ∼104 m2/g. The reduction in surface area could be attributed to the aggregation of carboxyl-functionalized AO during the acidification step. The surface area of PEI-AO was found to be ∼155 m2/g. The high value of surface area obtained after the PEI-functionalization step indicates that the PEI chains did not suffer aggregation, possibly because of the electrostatic repulsions between the amine groups. Furthermore, it can also be attributed to the additional adsorption sites provided by the surface-anchored organics. These results are in line with the DLS observation.

The PEI-AO thus obtained was studied for its palladium-sorption properties. The Pd(II) adsorption as a function of pH and amount of adsorbent revealed the optimal values to be 6.0 and 0.1 g, respectively (Figure S2). Figure S2 shows that with the increase in the amount of adsorbent, the % adsorption of Pd(II) significantly decreases. This unique behavior could be attributed to the following reason. Generally, it is well-known that adsorption–desorption is an equilibrium process. When an adsorbate is strongly anchored over an adsorbent, the equilibrium will be shifted toward adsorption, and therefore desorption will be suppressed. The nature of adsorption is determined through the magnitude of enthalpy of adsorption. In our case, the ΔH0 was found to be −103.63 kJ mol–1 (vide infra), which lies in the boundary between physisorption and chemisorption. With an increase in the adsorbent dosage, there exists a competition for Pd(II) between the PEI-AO nanopowders that shift the equilibrium toward desorption. Hence, at higher adsorbent dosage, the % Pd adsorption decreases.

For the isotherm studies, the initial concentrations of Pd(II) were varied from 40 to 300 mg L–1. The equilibrium adsorption data were applied to two important isotherms (Langmuir and Freundlich) from which we could obtain the various isotherm parameters.

The Langmuir isotherm model was used to comprehend the monolayer adsorption on the homogeneous surface containing a limited number of identical sites and to calculate the maximum adsorption capacity of the adsorbent.[36,37] This isotherm explains the relation between the equilibrium concentration of palladium (Ce) and the amount adsorbed at equilibrium (qe) on the surface of PEI-AO. The maximum adsorption capacity q0 (amount of Pd(II) adsorbed per unit weight of the adsorbent) and constant b were obtained from the linear plot (Figure A). The linearized form for the Langmuir isotherm is shown in eq .Using the above equation, the maximum adsorption capacity q0 was found to be 97.7 mg g–1 (R2 = 0.91). Suitability of a particular adsorption isotherm could be inferred through an important parameter called the dimensionless separation factor RL. This is given by the equation RL = 1/(l + bC0), where C0 is the initial concentration of Pd(II) in mg L–1 and b is the Langmuir constant (L mg–1). The value of b was found to be 0.03 L mg–1. The value of RL less than 1 relates to satisfactory adsorption, whereas the same when greater than 1 indicates reduced or unfavorable adsorption.[38] The value of RL for the adsorption of palladium over the PEI-AO adsorbent was found to be 0.77, indicating the usefulness of the Langmuir isotherm in this system.

Figure 5

Langmuir (A) and Freundlich (B) isotherms for the adsorption of Pd(II) over PEI-AO. Pseudo-first-order (C) and pseudo-second-order (D) kinetics for the adsorption of Pd(II) over PEI-AO.

The Freundlich isotherm, based on sorption on a heterogeneous surface, is an alternative useful model to gauge the adsorption phenomena occurring in dilute solutions.[39,40] The linearized form for the Freundlich isotherm is given in eq In this particular model, Ce is the equilibrium concentration of the Pd(II) in mg L–1, qe is the amount of Pd(II) adsorbed at equilibrium in mg g–1, and KF and n are the Freundlich constants for adsorption capacity and adsorption intensity, respectively. KF and n were obtained from the slope and intercept of the Freundlich isotherm plots, respectively (Figure B). The regression coefficient R2 was found to be 0.90. The Freundlich exponent n should fall in the range of 1–10 so as to predict the suitability of this isotherm[16] for a particular adsorption process. The respective values of n and KF were calculated to be 1.61 and 5.12, respectively, implying the favorable adsorption process. The Langmuir and Freundlich isotherm parameters are given in Table .

Table 1

Parameters of Langmuir and Freundlich Isotherm Models

isotherm model	parameters	values
Langmuir	q0 (mg g–1)	97.7
	b (L mg–1)	0.03
	RL	0.77
	R2	0.91
Freundlich	KF (mg1–1/n g–1 L1/n)	5.12
	n	1.61
	R2	0.90

The kinetics of the palladium adsorption onto the PEI-AO surface was studied using first-order[41] and pseudo-second-order[42] models, and the linearized equations are expressed as given by eqs and 4.where qe and q refers to the amount of palladium adsorbed at equilibrium and time t with the first- and second-order rate constants k1 and k2, respectively. By fitting the experimental data through the plots of log(qe – q) and t/q against t (Figure C,D), we obtain the kinetic parameters for the above two models. The adsorption data follow well with the pseudo-second-order model as indicated by the higher regression coefficient (Table ). Further, the qe values obtained experimentally and from the pseudo-second-order kinetic model were found to be 2.6362 and 2.5159 mg g–1, respectively. With the experimental and calculated qe values being quite agreeable, it proves the applicability of the pseudo-second-order model in understanding the adsorption kinetics of palladium onto the PEI-AO adsorbent. Also, the kinetics data revealed the optimal equilibration time to be 50 min.

Table 2

Kinetic Parameters for the Adsorption of Pd(II)

		pseudo-second-order kinetic model
concentration of Pd(II) solution (mg L–1)	qe (mg g–1)	k2 (g mg min–1)	q2 (mg g–1)	R2
10	2.6362	0.1004	2.5159	0.99

The sorption of Pd(II) onto the PEI-AO surface was evaluated through the thermodynamic studies involving Gibbs free energy (ΔG0), standard enthalpy change (ΔH0), and standard entropy change (ΔS0). These parameters were calculated (Table ) from the temperature-dependent adsorption data using the Van’t Hoff equations,[43,44] as shown in eqs and 6.The Kc values were obtained from the ratio of concentration of tetrachloropalladate(II) anion adsorbed onto the PEI-AO adsorbent to that in the liquid phase at equilibrium. The standard enthalpy and entropy changes were obtained through the ln K against 1/T (Figure S3) plot. The negative ΔG0 substantiates the spontaneity of adsorption of Pd(II) onto PEI-AO. The decrease in ΔG0 values with an increase in temperature shows that palladium adsorption does not proceed well at higher temperatures. The negative ΔH0 (−103.63 kJ mol–1) points the exothermic adsorption behavior, and also its magnitude gives vital information on the adsorption type, which could be either physical or chemical. When there is physical adsorption, ΔH0 would generally be less than 80 kJ mol–1, and for chemical adsorption, ΔH0 has a range of 80–400 kJ mol–1.[45] The negative ΔS0 (−319.67 J mol–1 K–1) is indicative of decreased randomness at the Pd(II)-PEI-AO interface. The negative activation energy (Ea) at different temperatures (Ea = ΔHads0 + RT) is also an indicator of the exothermic nature of adsorption of Pd(II) onto the PEI-AO adsorbent surface.[14]

Table 3

Thermodynamic Parameters for the Adsorption of Pd(II) ions

temperature (kelvin)	ΔG0 (kJ mol–1)	ΔS0 (J mol–1 K–1)	ΔH0 (kJ mol–1)	Ea (kJ mol–1)
298	–8.3683	–319.67	–103.63	–101.02
308	–5.1716
318	–1.9749
328	–1.2217

For the recovery of adsorbed palladium, the fixed bed column studies were performed on a glass column packed with 1.0 g of the PEI-AO adsorbent. A 10 mg L–1 Pd(II) was poured onto the adsorbent column (flow rate 5 mL min–1), and the concentration of palladium in the solution phase was measured using atomic absorption spectroscopy (AAS). Thiourea that contains sulfur and nitrogen as coordinating atoms was used as a desorbing agent. Desorption of palladium was found to be effective with a 9 mL volume of 0.2 mol L–1 aqueous thiourea. This could be elucidated through the formation of palladium–thiourea yellow colored complex, which considerably lowers the interaction between the adsorbent (PEI-AO) and Pd(II) ions, thereby facilitating palladium desorption from the adsorbent surface. After 3 cycles of regeneration, the percentage adsorption of palladium was slightly decreased from 91.4 to 90.8%, whereas the 4th cycle showed the percentage adsorption as 84.6%. This shows that the adsorbent could be regenerated and reused for three cycles using thiourea without any significant loss in adsorption capability. This has additionally proven the robustness and stability of the PEI surface functionalization.

The interaction between Pd(II) and PEI was probed using IR spectroscopy (Figure ). The IR spectrum of PEI shows characteristic peaks at 3359 cm–1 (−N–H stretching), 2947, 2849 cm–1 (−C–H stretching), 1570 cm–1 (−N–H bending), 1476 cm–1 (−C–H bending), and 1312, 1116 cm–1 (−C–N stretching).[46−48] After loading Pd(II) onto PEI, the N–H bending peak at 1570 cm–1 and the −C–N stretching at 1312 cm–1 were vastly suppressed, whereas new peaks were prominent at 1631, 1514, and 1466 cm–1. This shows that the electrostatic interaction through the N–H group is certainly accountable for palladium sorption. In the case of PEI-AO, the N–H stretching peak at 1567 cm–1 was suppressed after palladium sorption, confirming the role of nitrogen in the adsorption of palladium (Figure S4).

Figure 6

FT-IR spectra of neat PEI before and after adsorption of Pd(II).

The Pd-PEI-AO and rPd-PEI-AO were characterized using XPS to gain additional insight into the interaction between palladium and the adsorbent (Figure ). In both the cases, the survey scan revealed the presence of N 1s and Pd 3d peaks. The N 1s narrow scan revealed that the nitrogen peak of PEI-AO at 399.1 eV was shifted to 399.8 eV after Pd adsorption, indicating that the electron density surrounding the nitrogen was decreased because of the interaction with Pd(II).[48] After reduction of palladium, the N 1s peak at 399.5 eV was appeared to be broader and less in intensity, which indicates that the palladium linked to the nitrogen is reduced; as a consequence, the nitrogen has partially regained its electron density. The Pd 3d narrow scan revealed that the Pd-PEI-AO exhibited the characteristic Pd 3d5/2 and Pd 3d3/2 peaks at 335.5 and 340.7 eV, respectively. The deconvolution of the curve reveals that the palladium mostly exists in its +2 oxidation state when adsorbed over PEI-AO. After reduction, the peaks have become much sharper and shifted to a lower binding energy value, which substantiated that the Pd present in rPd-PEI-AO was successfully reduced to zero-valent oxidation state.[49]

Figure 7

XPS survey scan (A), N 1s narrow scan (B), and Pd 3d narrow scan (C) of Pd-PEI-AO and rPd-PEI-AO.

X-ray diffraction (XRD) patterns were recorded for pristine AO, Pd-PEI-AO, and rPd-PEI-AO and presented in Figure . The pristine AO exhibited broad peaks, indicating the amorphous nature of alumina possessing a certain amount of nanocrystalline domains. The XRD pattern of Pd-PEI-AO was found to be almost identical to the pristine AO (JCPDS # 77-0396) except for the peak at 39.9° corresponding to the (111) plane of palladium (JCPDS # 89-4897). After reduction, the intensity of the (111) plane of palladium in rPd-PEI-AO was increased, whereas that of the planes that correspond to AO were relatively suppressed. This could be attributed to the increased crystallinity of palladium, facilitated by the reduction process.

Figure 8

XRD patterns of pristine AO, Pd-PEI-AO, and rPd-PEI-AO (* and # indicate the planes corresponding to alumina and palladium, respectively).

Field-emission scanning electron microscopy (FE-SEM) imaging and energy-dispersive spectroscopy (EDS) analyses for PEI-AO, Pd-PEI-AO, and rPd-PEI-AO were performed, and the results are shown in Figure . In all the samples, the surface morphology of the AO was found to be almost identical. The samples were vastly found to possess flat platelike particles, which were comprised of nanogranules. The EDS analysis of PEI-AO revealed the presence of elements such as C, O, and Al. It is known that nitrogen is less sensitive and susceptible for interference in EDS and is therefore excluded. In the case of Pd-PEI-AO, in addition to C, O, and Al, about 0.8 at % of Pd and 1.3 at % of Cl were also detected. This could be attributed to the PdCl42– adsorbed over the Pd-PEI-AO. After reduction of palladium, the rPd-PEI-AO showed an increased amount of Pd (∼1.1 at %), whereas the Cl was completely absent. The loss of chlorine has additionally confirmed that the PdCl42– was successfully reduced to metallic palladium.

Figure 9

FE-SEM and EDS images of pristine AO (A,D), Pd-PEI-AO (B,E), and rPd-PEI-AO (C,F).

One of the popular catalytic performance evaluations of a noble metal catalyst is the conversion of 4-nitrophenol (4-NP) to 4-aminophenol (4-AP).[50] This conversion has a practical significance that an endocrine disruptor (4-NP) chemical is converted to its corresponding amino derivative. In this case, the product 4-AP is a precursor for the synthesis of paracetamol, thus signifying the valorization of a potential hazardous chemical. The conversion of 4-NP to 4-AP was followed using UV–visible (UV–vis) spectroscopy. The original yellow color of 4-NP was intensified to dark yellow upon addition of NaBH4 because of the formation of the phenolate ion, which exhibited an absorption maximum at 400.5 nm. When the catalyst was added, the color of the solution faded and became colorless on complete reduction of 4-NP. This was witnessed by the gradual decrease in the absorption band at 400.5 nm and formation of a new band at 307 nm, corresponding to the 4-AP. The representative UV–vis spectra and the recyclability data are shown in Figure . In the first cycle, 98% reduction of 4-NP was achieved in 3 min, whereas the fifth cycle required 11 min to accomplish the same. The gradual decrease in the activity could be due to the loss of small amounts of palladium during each cycle and/or catalyst poisoning. Therefore, we quantified the total palladium content leached into the solution phase over three cycles using AAS, and it was found that less than 1% of palladium got leached into the solution phase, whereas that desorbed using thiourea from the spent catalyst was found to be ∼96%. The rate of 4-NP reduction at 10% conversion in the first cycle was found to be 125 mmol L–1 min–1 g–1.

Figure 10

UV–vis absorption spectra of 4-NP reduction (4th cycle) by Pd-PEI-AO (top) and the recyclability data for the reaction up to five cycles (bottom).

A representative inorganic polluting species is Cr(VI), which is considered to be potentially carcinogenic, whereas Cr(III) is known to be an essential metabolite for the function of insulin.[51] Hence, it is essential to convert Cr(VI) to less toxic Cr(III) through a simple and facile process. The catalytic reduction of Cr(VI) to Cr(III) was chosen as a model system to explore the potential of Pd-PEI-AO toward environmental remediation. In an attempt toward Cr(VI) reduction, no significant reaction occurred when the Pd-PEI-AO was directly used as the catalyst. This is because the catalytic property of Pd(II) toward Cr(VI) reduction under the experimental conditions was poor. Therefore, the Pd-PEI-AO was subjected to reduction to yield rPd-PEI-AO, which was further used for Cr(VI) reduction. Oxalic acid was used as a sacrificial agent in this case. The reduction of Cr(VI) was followed using UV spectrometry at a wavelength maximum of 350 nm. About 95% reduction was attained within 3 min in the first cycle, whereas the fifth cycle required 7 min to accomplish the same (Figure ). This has revealed that the catalyst possessed adequate activity even after multiple times of regeneration. The rate of Cr(VI) reduction at 10% conversion in the first cycle was found to be 67 mmol L–1 min–1 g–1. The palladium leaching studies revealed that ∼7% of the metal was leached into the solution phase over three cycles of the reaction and ∼88% of palladium was desorbed using thiourea solution. The little excess of palladium leaching as compared to the 4-NP reduction could be due to the oxalic acid used in the catalytic reaction, which could have interacted with the basic amine groups in PEI and caused liberation of small quantities of the bound metal.

Figure 11

UV–vis absorption spectra of Cr(VI) reduction (5th cycle) by rPd-PEI-AO (top) and the recyclability data for the reaction up to five cycles (bottom).

Sustainable hydrogen production is considered to be one of the important avenues to meet the growing energy demand. Among the various hydrogen sources, ammonia borane (AB) is considered to be one of the safest and green compounds that possess high density of hydrogen.[52] On-demand catalytic hydrogen generation is therefore considered to be an important reaction. The Pd-PEI-AO was tested for its catalytic activity toward hydrogen evolution from AB in aqueous medium. The first cycle produced the maximum amount (∼52 mL) of hydrogen in about 2 min, whereas the successive cycles up to five times required 3 to 4 min of duration to produce similar quantity of hydrogen (Figure ). This has additionally proven the high catalytic performance of Pd-PEI-AO even after multiple cycles with a minimal to moderate loss in activity. The rate of hydrogen production at 10% conversion in the first cycle was found to be 10.5 mol min–1 g–1. The stability of the adsorbed palladium over three cycles was studied through the estimation of the palladium content in the solution phase and solid phase before and after the reaction. AAS studies revealed that the total palladium content leached into the solution phase over three cycles was less than 1%, whereas that desorbed using thiourea from the spent catalyst was found to be ∼96%.

Figure 12

Hydrogen generation studies from AB up to five cycles using Pd-PEI-AO.

Conclusions

Facile surface functionalization of PEI over AO was accomplished in a two-step approach, which was confirmed using IR, XPS, TGA, and zeta potential measurements. The present work showed the utility of the PEI-coated AO toward effective adsorption of palladium. The sorption parameters of palladium were studied through pH effect, contact time, amount of adsorbent, concentrations, and temperature. The optimum pH for the adsorption of Pd(II) was found to be 6.0, and the adsorption adhered to the pseudo-second-order kinetics. Although the native alumina showed an adsorption capacity of 35.08 mg g–1, the PEI-AO exhibited a high adsorption capacity of 97.7 mg g–1, obtained through the Langmuir adsorption isotherm. The exothermic nature of the adsorption process and the negative free energy obtained at various temperatures confirm the spontaneity of adsorption. XPS analyses of Pd-PEI-AO and rPd-PEI-AO revealed the favorable interaction of PdCl42– with nitrogen moieties of PEI. XRD studies on these samples revealed an increase in the crystallinity of palladium after reduction. The catalytic reduction of 4-NP and Cr(VI) using Pd-PEI-AO was found to proceed with rates as high as 125 and 67 mmol L–1 min–1 g–1, respectively. Hydrogen generation from AB was found to be efficient with Pd-PEI-AO with a rate of 10.5 mol min–1 g–1 at 10% conversion. Recyclability of Pd-PEI-AO was studied up to five cycles in all the three catalytic reactions, which revealed the potential of the sorbent for efficient reusability. The amine functionalization strategy presented in this work could also show potential in recovering and reusing other precious metals such as gold, platinum, and rhodium.[28]

Experimental Section

Materials

AO of ∼80 nm in average size was procured from Nanoshel Inc. and used as received. SC, acetic acid, PEI (branched, Mw ≈ 800), EDC, and NHS were procured from Sigma-Aldrich. Sodium borohydride, 4-NP, potassium dichromate, sodium hydroxide, sulfuric acid, and oxalic acid were purchased from SD Fine Chemicals, India Ltd. Palladium chloride was procured from HiMedia, India. The procured chemicals were used as received, unless otherwise mentioned. AB was synthesized following the literature procedure.[53]

Characterization

A Fourier transform infrared (FT-IR) spectrophotometer (JASCO FTIR-4200) was utilized to study the functionalization of alumina nanosupport and palladium adsorption. TGA was carried out using a Shimadzu DTG-60 differential thermal analyzer to follow the degradation profile of the surface-immobilized organics. XPS measurements were performed in a PHI 5000 Versa Prob II (FEI Inc.) instrument to study the surface chemical composition. A Zetasizer Nano-ZS instrument (Malvern) was employed to measure the particle size as well as the surface zeta potential, wherein the test samples were taken in the form of 20 μg/mL aqueous suspensions of the materials. The surface area analysis of the materials was examined using a BET (Smart Sorb 93) apparatus using the nitrogen adsorption/desorption method. The surface morphology and the elemental composition of the functionalized nanosupport were determined by FE-SEM fitted with EDS (Carl Zeiss ULTRA-55). The XRD pattern of the samples was measured using a Rigaku Ultima IV X-ray diffractometer with Cu Kα (λ = 1.5418 Å) at a scan rate of 1°/min. The palladium concentration in the aqueous phase was analyzed using flame atomic absorption spectrophotometry (Shimadzu AA 7000) at 247.6 nm with an acetylene–air flame combination.

Carboxyl Functionalization of AO

In a typical reaction, 1.0 g of AO was added to 100 mL of Millipore water in a 250 mL round bottom flask. To this, about 20 wt % (w.r.t. AO) SC was added, and the mixture was sonicated for 30 min in an ultrasonication bath to disperse the particles in the media. The temperature of the reaction medium was increased to 90 °C in an oil bath and allowed to stand for 8 h under stirring.[54] Subsequently, the mixture was cooled to room temperature and centrifuged. The resultant product (SC-AO) was acidified by adding 50% (v/v) aqueous acetic acid solution and incubated for 10 min. The CA-AO thus obtained was centrifuged, washed with Millipore water three times, and then dried in a vacuum oven for overnight at room temperature.

PEI Functionalization of CA-AO

The amine functionalization was carried out by covalently anchoring PEI onto CA-AO through amide coupling.[55] To achieve this, about 1.0 g of CA-AO in 20 mL of water was treated with a solution containing 0.6 g of EDC and 0.2 g of NHS. The resulting suspension was stirred for 3 h at room temperature. After this time, the suspension was added dropwise into a beaker containing 1.2 g of PEI dissolved in 500 mL of water and the pH of the mixture was adjusted to 5 and stirred for 24 h at room temperature. Finally, the PEI-AO was obtained by centrifugation, washing, and drying in a manner similar to CA-AO.

Batch Adsorption Experiments

The batch adsorption studies were carried out by taking 0.1 g of adsorbent (PEI-AO) in 30 mL of 10 mg L–1 Pd(II) solution prepared in HCl medium, and the parameters such as pH, adsorbent amount, adsorption time, and temperature were optimized at room temperature (25 °C) for the desired time interval in an orbital incubator shaker at 120 rpm. The optimized pH was 6.0. Various initial concentrations of Pd(II) ranging from 40 to 300 mg L–1 were used to study the isotherms. The adsorbed palladium(II) at equilibrium (qe) is calculated using the following relationwhere C0 and Ce are the initial and final liquid-phase concentrations of palladium, respectively, V is the volume (in L) of the solution, and W is the weight (in g) of the PEI-AO adsorbent used. The adsorption process was studied using first-order and second-order kinetics, and the isotherms were evaluated using Langmuir and Freundlich models.

Characterization and Catalysis Studies of Pd-Adsorbed PEI-AO

For the characterization studies, 300 mg L–1 palladium chloride solution was stirred with 0.1 g of PEI-AO to ensure maximum palladium adsorption. The resultant material was coded as Pd-PEI-AO and used for catalytic studies. The reduction of 4-NP was carried out by taking the catalytic quantity (∼5 mg) of Pd-PEI-AO in 10 mL of solution containing 0.143 mM 4-NP and 14.3 mM NaBH4. The reaction progress was monitored using a UV–vis spectrophotometer. After the complete reduction of 4-NP, the Pd-PEI-AO was recovered through centrifugation and reconstituted for recyclability studies. To perform Cr(VI) reduction, the Pd-PEI-AO was subjected to reduction using NaBH4 solution to reduce the adsorbed palladium, referred hereafter as rPd-PEI-AO. A stock solution of 20 mg L–1 potassium dichromate containing 0.33 M oxalic acid was prepared, and the pH was adjusted to 3.0. To a 3 mL of this solution, 5 mg of the rPd-PEI-AO was added and the UV–vis absorbance of the resultant solution was recorded as a function of time. After the complete reduction of Cr(VI), the solution was centrifuged to recover the catalyst, which was utilized for the recyclability studies. For hydrogen generation experiments, 100 mg of rPd-PEI-AO was added to a 20 mL 64.8 mM AB solution and stirred. The evolved hydrogen gas was measured using the gas burette method. In all the catalytic reactions, the reusability was studied up to five cycles.