Supporting palladium metal on gold nanoparticles improves its catalysis for nitrite reduction

Abstract

Nitrate (NO₃⁻) and nitrite (NO₂⁻) anions are often found in groundwater and surface water as contaminants globally, especially in agricultural areas due to nitrate-rich fertilizer use. One popular approach to studying the removal of nitrite/nitrate from water has been their degradation to dinitrogen via Pd-based reduction catalysis. However, little progress has been made towards understanding how the catalyst structure can improve activity. Focusing on the catalytic reduction of nitrite in this study, we report that Au NPs supporting Pd metal ("Pd-on-Au NPs") show catalytic activity that varies with volcano-shape dependence on Pd surface coverage. At room temperature, in CO₂-buffered water, and under H₂ headspace, the NPs were maximally active at a Pd surface coverage of 80%, with a first-order rate constant (k_cat = 576 L g_Pd⁻¹ min⁻¹) that was 15x and 7.5x higher than monometallic Pd NPs (∼4 nm; 40 L g_Pd⁻¹ min⁻¹) and Pd/Al₂O₃ (1 wt% Pd; 76 L g_Pd⁻¹ min⁻¹), respectively. Accounting only for surface Pd atoms, these NPs (576 L g_surface-Pd⁻¹ min⁻¹) were 3.6x and 1.6x higher than monometallic Pd NPs (160 L g_surface-Pd⁻¹ min⁻¹) and Pd/Al₂O₃ (361 L g_surface-Pd⁻¹ min⁻¹). These NPs retained ∼98% of catalytic activity at a chloride concentration of 1 mM, whereas Pd/Al₂O₃ lost ∼50%. The Pd-on-Au nanostructure is a promising approach to improve the catalytic reduction process for nitrite and, with further development, also for nitrate anions.

---

1. Introduction

Nitrate (NO₃⁻) is found in the groundwater and surface water globally, especially in agricultural areas due to the use of nitrate rich fertilizers.^1–4 This anion, and its metabolic conversion to the nitrite anion (NO₂⁻), cause adverse human health effects, including methemoglobinemia or the ‘blue baby’ syndrome.⁵ Nitrite can be also converted to carcinogenic N-nitroso compounds in the body,⁶ which may result in cancer and hypertension.⁷ The maximum contaminant levels (MCL) of NO₃⁻ and NO₂⁻ have been set at 10 mg N per L and 1.0 mg N per L (calculated by nitrogen weight, denoted as mg N per L), respectively, by the United States Environmental Protection Agency (US EPA).⁸ For the European Union countries, the European Drinking Water Directive set concentration limits for NO₃⁻ and NO₂⁻ at 11.3 mg N per L and 0.15 mg N per L, respectively.⁹ The World Health Organization guideline values of NO₃⁻ and NO₂⁻ are 11.3 mg N per L and 0.91 mg N per L.¹⁰

Current methods to treat nitrate/nitrite contamination include ion exchange, reverse osmosis, biological denitrification, and catalytic reduction.^11–21 The EPA-approved treatment technologies for removing nitrates/nitrites in waters destined for drinking purposes are ion exchange and reverse osmosis, with the former being the most widely used.²² The ion exchange approach removes nitrate/nitrite from water by replacing them with equivalent amount of anions such as chloride or bicarbonate ions.¹¹ The main disadvantage is that, after the ion exchange material is regenerated for further use by contacting with a brine solution, the resulting brine which contains the nitrate/nitrite anions must undergo further treatment prior to disposal. Reverse osmosis also results in a nitrate/nitrite-containing secondary stream that needs further treatment. The biological denitrification and catalytic reduction routes have been studied at the laboratory and pilot scales but they are not yet commercially practiced. Biological denitrification can selectively remove nitrate/nitrite from water, but it is affected by pH and the presence of other salts in water; it can generate undesired by-products;¹⁵ and bacterial contamination of the treated water is a concern.¹¹

The catalytic reduction of nitrate/nitrite has received intense research interest ever since Pd-based catalysts (e.g., Pd–Cu, Pd–In and Pd–Sn) were shown capable of converting nitrate/nitrite to harmless nitrogen (N₂) using hydrogen.²³ This chemistry was first reported by Vorlop and Tacke in 1989 using supported Pd–Cu/Al₂O₃ catalysts.¹⁴ Vorlop and co-workers reported improved reduction activity using Pd–Sn/Al₂O₃ and Pd–In/Al₂O₃ catalysts.²⁴

The catalytic reduction of nitrate is generally accepted to occur in two stages: the reduction of NO₃⁻ to NO₂⁻, and then reduction of NO₂⁻ to N₂ or NH₃ (Scheme 1).^15,25–28 In greater detail, the second metal (i.e., Cu, In, Sn) reduces the chemisorbed NO₃⁻ species in the first step and the Pd then regenerates the oxidized metal via H₂ reduction (reaction (1) of Scheme 1). In the second stage, Pd by itself reduces chemisorbed NO₂⁻ species in several surface reaction steps, eventually to form N₂ (reaction (2)) or NH₃ (reaction (3)). Ammonia is an undesirable reaction product because it is a groundwater pollutant with a US EPA MCL of 0.66 mg N per L.⁸ The Pd–Cu composition provides N₂ selectivity values as high as 80–95%, depending on the reaction testing conditions used.^19,29 Ammonia formation is favored at high pH values, and so N₂ selectivity is improved by operating the reaction at non-alkaline conditions, for example.²⁹

Scheme 1 General reaction pathway for the reduction of nitrate or nitrite anions using a Pd-based catalyst. M is a promoter metal such as Cu, In, and Sn.²⁷

While the compositional effects on catalytic nitrate/nitrite reduction have been amply studied, the effect of nanostructure is not well addressed yet. We have previously demonstrated that Pd catalytic activity for hydrodechlorination (a type of reduction reaction) can be varied with great control by depositing Pd metal on Au nanoparticles (to form Pd-on-Au NPs) at a pre-determined Pd surface coverage.^30–33 Motivating this work, we hypothesized that Pd-on-Au NPs catalyze nitrite reduction (the second stage of nitrate reduction) and that catalytic activity is correlated to Pd surface coverage. We synthesized ∼4.3 nm Pd-on-Au NPs of varying Pd coverages, and quantified their reaction rate constants for nitrite reduction at room temperature, under H₂ headspace, in water and at pH 5–7 (buffered using CO₂ gas). We quantified the corresponding N₂ selectivities, and studied the effect of chloride anions, a common groundwater constituent.

2. Materials and methods

2.1. Materials

Tetrachloroauric(III) acid (HAuCl₄·3H₂O, 99.99%), tannic acid (C₇₆H₅₂O₄₆, >99.5%), potassium carbonate (K₂CO₃, >99.5%), palladium(II) chloride (PdCl₂, 99.99%), 1 wt% Pd/Al₂O₃, modified Griess's reagent, and Nessler's reagent (K₂HgI₄) were purchased from Sigma-Aldrich. Trisodium citrate (Na₃C₆H₅O₇, >99.5%, Fisher), sodium nitrite (NaNO₂, 99.7%) and sodium chloride (NaCl, 99.99%) were obtained from Fisher. Hydrogen gas (99.99%) was purchased from Matheson. 0.9 wt% Au/Al₂O₃ was obtained from Mintek Autek, South Africa. All experiments were conducted using Nanopure water (>18 MΩ cm, Barnstead NANOpure Diamond).

2.2. NP synthesis

4 nm Au NPs, Pd NPs and Pd-on-Au bimetallic NPs were synthesized as previously reported.³⁰ To synthesize Au NPs, 0.05 g tannic acid, 0.018 g K₂CO₃ and 0.04 g trisodium citrate were dissolved in 20 mL of water. In a second flask, 200 μL of HAuCl₄ solution (0.127 mol L⁻¹) was dissolved in 79.8 mL of water. Both solutions were heated to 60 °C, and the first solution was added to the second under vigorous stirring. The color of the resultant sol immediately changed from pale yellow to reddish-brown, indicative of the formation of Au NPs. The solution was heated to boiling, allowed to boil for 2 min and then cooled to room temperature (Scheme 2). The Au NP concentration was 1.26 × 10¹⁴ NP per mL.³⁰

Scheme 2 Two-stage synthesis of Pd-on-Au NPs with a specific Pd surface coverage.

The Pd NPs were synthesized by replacing the HAuCl₄ solution with a H₂PdCl₄ solution of the same molar concentration,³⁰ and instead of boiling for 2 min, the mixed solution was allowed to boil for 25 min. Pd-on-Au bimetallic NPs were prepared by adding specific amounts of H₂PdCl₄ solution to the as-synthesized Au NPs. The various surface coverage percentages (sc%) (0, 10, 30, 60, 80, 100, 150, 300 sc%) of Pd-on-Au NPs were obtained by mixing 9, 18, 55, 73, 92, 150, 351 μL H₂PdCl₄ solution (2.47 mM) and 2 mL of the as-synthesized Au NPs.³⁰ The mixed solution was stirred for 1 min and then bubbled with hydrogen gas for 2 min at a flow rate of ∼200 mL min⁻¹. For the calculation of surface coverage percentages (sc%), we modeled the 4 nm Pd-on-Au NPs as magic clusters, where the Au NP is a magic cluster of 7 shells, and the Pd forms the 8th shell or more (Scheme 2).³⁰

2.3. Characterization

UV-vis absorbance spectra of the NP solutions were measured on a Shimadzu UV-2401 PC spectrophotometer. Transmission electron microscopy (TEM) images were obtained using a JEOL 2010 transmission electron microscope operating at an accelerating voltage of 200 kV. The particle size distribution was calculated by counting around 200 particles. pH measurements were taken using a VWR sympHony SB20 meter with a standard pH electrode.

Nitrite ions were analyzed using the Griess test.³⁴ A stock solution of the Griess reagent was prepared by dissolving 10 g of the powder (Sigma-Aldrich) in 250 mL Nanopure water, such that the final concentrations are 0.1 wt% N-(1-naphthyl)ethylenediamine dihydrochloride, 1 wt% sulfanilamide, and 5% H₃PO₄.³⁵ In a typical colorimetric assay, the Griess reagent solution (0.2 mL), a nitrite-containing solution (0.2 mL), and water (1.6 mL) are mixed together and kept in room temperature for 10 min. The sulfanilamide reacts with a nitrite anion, and the resulting compound further reacts with the amine, forming a colored solution. The absorbance at 540 nm is measured via UV-vis spectroscopy, and the NO₂⁻ concentration is determined in the 0 to 2.0 ppm range using a standard curve (Fig. S1†).

Ammonia measurements were made using an ammonium (NH₄⁺) ion selective electrode (Cole-Parmer, detection limit 0.01 ppm, concentration range from 0.01 to 17 000 ppm) and using Nessler's test (concentration range from 0.02 to 2.5 ppm).^34,35 In a typical colorimetric assay, a Nessler's reagent solution (1 mL, 0.09 M of potassium tetraiodomercurate and 2.5 M potassium hydroxide, Sigma-Aldrich) and an ammonia-containing sample (1 mL) are mixed together and then kept at room temperature for 1 min. The ammonium reacts with the tetraiodomercurate at high pH, forming a colored solution. The absorbance at 420 nm is measured, and the NH₄⁺ concentration is determined in the 0 to 10.0 ppm range using a standard curve. (Fig. S2†). These concentrations were independently verified using an ammonium ion selective electrode (Cole-Parmer, detection limit 0.01 ppm, concentration range from 0.01 to 17 000 ppm) (Fig. S3†).

2.4. Catalytic experiments

Batch nitrite reduction experiments were conducted in a three-neck 250 mL round bottomed flask. The amount of Pd-on-Au NPs added was chosen such that the total Pd amount per reaction was 0.0365 mg, the final catalyst charge concentration was 0.365 mg-Pd per L, and the final liquid volume was 99.6 mL (Table S1†). For example, 3.75 mL of a Pd-on-Au NP (80 sc%) sol was combined with 95.85 mL of water in the flask. The solution was then bubbled simultaneously with hydrogen gas (120 mL min⁻¹, to serve as reductant) and carbon dioxide gas (120 mL min⁻¹, to buffer the solution to a pH value of 5–7) for 5 min. For the Pd/Al₂O₃ case, 36.5 mg was suspended in 10 mL of water, and then 1 mL of suspension was combined with 98.6 mL of water in the flask; the total Pd amount charged to the reactor was 0.0365 mg. For the Au/Al₂O₃ case, 28 mg was combined with 99.6 mL of water in the flask; the total Au amount charged to the reactor was 0.25 mg. Additionally, a set of catalytic experiments were carried out in the absence of carbon dioxide gas.

The catalytic reactions were conducted at room temperature under constant stirring (400 rpm) and continuous hydrogen gas (120 mL min⁻¹) and carbon dioxide gas (120 mL min⁻¹). The NaNO₂ solution (0.4 mL, 10 mg mL⁻¹ of NO₂⁻) was injected to start the reaction, such the initial solution NO₂⁻ concentration was 40 mg NO₂ per L (or 12.2 mg N per L). The reaction was monitored periodically by withdrawing 1 mL aliquots from the reaction flask. By the end of the reaction test, ∼8 to 12 mL of reaction solution was removed from the flask, leaving behind 88–92% of the initial solution. For the Pd-on-Au NP tests, no separation of the particles from the reaction medium was performed. For the Pd/Al₂O₃ test, the powder was separated from the reaction medium before nitrite and ammonium concentrations were determined.

Because the H₂ is excess to nitrite, the observed reaction rate k_meas (with units of min⁻¹) was calculated by assuming first-order dependence on nitrite concentration:³⁶

with the observed reaction rate constant k_meas = k_catC_cat, k_cat is the Pd normalized reaction rate constant (with units of L g_Pd⁻¹ min⁻¹), C_NO2 is the concentration of nitrite (with units of mg L⁻¹), t is reaction time (with unit of min) and C_cat is the concentration of Pd (with units of g L⁻¹). Accounting for only surface Pd atoms (calculated using the magic cluster model), k^′_cat comes from k_meas = k^′_catC^′_cat, where C^′_cat is the concentration of surface Pd.

To check that the proper amount of catalyst was used, various aliquots (0, 1.25, 2.5, 5.0, 7.5, 10 mL) of 60 sc% Pd-on-Au NP sols were added to the reactor and the total reaction volume was set at 99.6 mL. The corresponding Pd amounts in reactor were 0, 0.00913, 0.0183, 0.0365, 0.0548, 0.073 mg, respectively. Reaction rates were determined as described above.

To determine the effect of chloride, a series of experiments were performed by adding different amounts of 1 M NaCl solution was added to reactor with catalyst before bubbling H₂ and CO₂ gas, giving the final concentration of chloride from 0 to 0.05 M. The total reaction volume was kept at 99.6 mL. The actual volume of 1 M NaCl and nanopure water added to reactor were shown in Table S2.† The following catalyst was added to the reactor: 5 mL 60 sc% Pd-on-Au NPs solution or 7.2 mg Pd/Al₂O₃ (1 wt%). The other conditions are the same with the typical catalytic experiment.

3. Results and discussion

3.1. Structure of Pd-on-Au NPs

The synthesis and characterization of Pd-on-Au bimetallic NPs for hydrodechlorination reactions has been previously reported by our group.^{30,32,37–39}Fig. 1 shows an example TEM image of as-prepared Pd-on-Au NPs. These NPs have a relative uniform size distribution, with mean diameter of 4.3 nm and a relative standard deviation of 12%. Our X-ray absorption spectroscopy (XAS) results reported in earlier publications indicated that all Pd atoms were located on the surface of the Au NPs.^32,40 At relatively low surface coverage (<30 sc%), Pd atoms were found mostly as scattered atoms on the Au surface. At higher surface coverages, some of Pd atoms formed non-oxidizable two-dimensional (2-D) ensembles, which we concluded to be especially the most active species for the hydrodechlorination (HDC) of trichloroethene (TCE). At surface coverage of 70% and higher, three-dimensional (3-D) ensembles appeared, in which oxidized Pd atoms were found on top of other Pd atoms, and the per-gram-Pd catalytic activity decreased.

Fig. 1 (a) Transmission electron microscopy (TEM) image and (b) size distribution of Pd-on-Au bimetallic NPs (60 sc%). Each bar represents the total number of NPs of a particle diameter ±0.25 nm.

3.2. Effect of Pd surface coverage

For all catalyst compositions, 100% nitrite conversion was reached within 30 min. Fig. 2 is representative of the evolution of nitrite, ammonia and nitrogen as a function of time. The decrease in nitrite concentration fit well to a pseudo first-order model, and observed reaction rate constant, k_meas, was obtained from the slope of the natural log of nitrite concentration versus reaction time using linear least squares fitting. The selectivity to N₂ exceeded 99% for all catalysts (Table 1). These near-100% selectivity values towards N₂ are similar to catalytic nitrite reduction studies from other research groups.⁴¹ The Pd-on-Au NP catalyst remained active for nitrite reduction after spiking the reactor with additional nitrite solution two more times (Fig. S4†).

Fig. 2 Concentration–time curves of NO₂⁻, NH₄⁺ and N₂. Reaction conditions: 60 sc% Pd-on-Au NPs with 0.365 mg L⁻¹ Pd in reactor, 120 mL min⁻¹ H₂, 120 mL min⁻¹ CO₂, 400 rpm stirring rate, 1 atm pressure. The initial nitrite/Pd molar ratio was 252/1.

Table 1 Rate constants and N₂ selectivity for Pd-on-Au NPs, Pd NPs, Pd/Al₂O₃, and Au/Al₂O₃

Sample name	k cat (L gPd^−1 min^−1)	k ^′cat (L gsurface-Pd^−1 min^−1)	N2 Selectivity (at 90% conversion)
a First-order rate constant kmeas = 0 min^−1.
0 sc% Pd-on-Au NPs	Not active^a	Not active^a	Not available
10 sc% Pd-on-Au NPs	32	32	99.3
30 sc% Pd-on-Au NPs	174	174	99.5
60 sc% Pd-on-Au NPs	479	479	99.8
80 sc% Pd-on-Au NPs	576	576	99.6
100 sc% Pd-on-Au NPs	466	466	99.5
150 sc% Pd-on-Au NPs	421	584	99.1
300 sc% Pd-on-Au NPs	188	460	99.3
Pd NPs	40	160	99.9
Pd/Al2O3 (1 wt%)	76	361	99.5
Au/Al2O3 (1.2 wt%)	Not active^a	Not active^a	Not available

---

The observed reaction rates determined for Pd-on-Au NPs (60 sc%) at various catalyst charges showed a linear dependence, showing that the reaction rate increased proportionately to the catalyst amount used (Fig. 3). Here, the implication is that the reaction rate constant determined at the standard catalyst charge using eqn (1) (the middle point of the 5-point line) is equal to the rate constant determined from 5 catalyst amounts (the slope of this 5-point line, which is equivalent to k_cat = k_meas/C_cat), This linearity was seen for other catalyst compositions. Fig. 3 gave a slope of 416 L g_Pd⁻¹ min⁻¹, smaller in value than 479 L g_Pd⁻¹ min⁻¹ (Table 1, data point circled in Fig. 3), which we attributed to experimental uncertainty. If the reaction rate does not increase proportionately to catalyst amount, then too much catalyst has been added and the mass transfer effect becomes noticeable.⁴²

Fig. 3 Observed reaction rate k_meas determined at different catalyst charges of Pd-on-Au NPs (60 sc%). The circled data point (blue color) indicates the standard Pd amount used for all other catalysts in this study.

The Pd-on-Au NP catalysts with various Pd surface coverage (sc%) were prepared by reducing appropriate amounts of Pd salt precursor in the presence of Au NPs using H₂ gas. The Pd surface coverages were calculated by modeling a magic cluster structure for the NPs,³⁰ such that all atoms could be accounted for. A ∼4 nm Au NP with surface coverages of 10, 30, 60, 80, 100, 150, and 300% had Pd weight loadings of 2.4, 5.8, 10.9, 15.5, 18.1, 19.7 wt%, respectively. The convenience and accuracy of the magic cluster model relies on the assumptions that all Pd precursor is fully reduced onto the Au surface and that the Au particles are perfectly monodisperse (which is not the case). Inductively coupled plasma-optical emission spectroscopy results of ∼4 nm Pd-on-Au NPs from an earlier study indicated the metals content was >95% of the expected values.⁴³ We note that the uncertainties in the amount of Pd salt added to the synthesis volume and in the amount of Au NPs used, lead to an uncertainty in the calculated Pd surface coverage of 0.01 × sc% (at one standard deviation).³³

Nitrite reduction activity of Pd-on-Au NPs varied significantly with Pd surface coverage (Table 1 and Fig. 4). Monometallic Au NPs (0 sc%) and Au/Al₂O₃ did not show activity for the reaction. Since Pd was the active metal, catalytic activity was quantified by normalizing k_meas to total Pd content (k_cat) or to total surface Pd content (k^′_cat). The rate constants k_cat increased monotonically from 10 sc% to 80 sc% and decreased from monotonically from 80 sc% to 300 sc%, characteristic of a "volcano-shape" structure–activity plot.

Fig. 4 Experimentally determined nitrite reaction rate constants for Pd-on-Au NPs as a function of Pd surface coverage and for monometallic Pd catalysts.

The 80 sc% Pd-on-Au NPs had a rate constant 15x and 7.5x times greater than those of monometallic Pd NPs and Pd/Al₂O₃, respectively (Table 1). Except for the 10 sc% case, all Pd-on-Au NPs were also more active, indicating a positive and direct interaction of the Au with the Pd metal. The volcano plots and higher catalytic activity of Pd-on-Au NPs observed here are consistent with observations for HDC of TCE and tetrachloroethene (PCE).³⁹ Even though nitrite is not structurally related to TCE (or PCE), the similarity in the volcano-shape plots suggests the reduction reactions occur on a common set of catalytically active sites. We attribute the promotional effect of Au on Pd catalysis of nitrite reduction to the high Pd dispersion on the Au NP surface and the presence of the non-oxidized 2-D Pd ensembles.

At surface coverages of 150 and 300 sc%, the per-gram-Pd catalytic activity was lower due to a lower Pd dispersion and the topmost atoms of the 3-D Pd ensembles being oxidized. Accounting for surface Pd atoms, these Pd-on-Au NPs had higher rate constant k^′_cat values than those of monometallic Pd NPs and Pd/Al₂O₃ (which also increased, Table 1). However, the trend in k^′_cat (which differ from k_cat above 100 sc%) was not as clear. This is due to the over-simplication of the metal-on-metal NP structure when using the magic cluster model for calculations, the presence of oxidized Pd and the lack of experimental data quantifying the amount of surface Pd atoms (e.g., through titration methods).³² That the surface-Pd-atom-normalized rate constants were larger than those for Pd NPs and Pd/Al₂O₃ may be the result of the electronic effect, in which there is change in valence electron density of states in the Pd metal.^44–46

In the absence of the CO₂ buffer, nitrite reduction was extremely slow. For example, the k_cat for 60 sc% Pd-on-Au NPs was 1.0 L g_Pd⁻¹ min⁻¹ (compared to 479 L g_Pd⁻¹ min⁻¹). In a sampling of 4 different compositions, the final conversion for all was ∼1 to 2%, and the final pH increased from an initial pH of 5–7 to >7. The low catalytic activity resulted from the alkaline condition, which was observed by other researchers.^28,29,47 The reduction of nitrite generates OH⁻ species, which block the active sites of palladium based catalyst. Thus, the use of a buffer like CO₂ (ref. 29) and formic acid,¹⁵ or a water-solubilized conducting polymer⁴¹ with buffering capacity is important to maintain high activities and accurate rate constant measurements.

Recently, Werth and coworkers used isotopically labeled nitrate/nitrite to study the nitrate/nitrite reduction pathways using a Pd–In/Al₂O₃ catalyst (5 wt% Pd, 0.5 wt% In).²⁷ Surface-bound NO and N₂O were found as reaction intermediates, providing greater mechanistic insight into the general reaction pathway of nitrite/nitrate catalytic reduction (Scheme 1). Derived from the further reduction of NO₂⁻, the NO surface intermediate reduces in the presence of H₂ to form NH₄⁺.²⁷ The NO intermediate also reacts with another NO to form the nitrous oxide surface intermediate, N₂O, in a parallel reduction step. This N₂O then reduces to form N₂. The high N₂ selectivity values observed for all catalysts in this study suggest that the NO-to-N₂O surface reaction step dominates the NO-to-NH₄⁺ step under the reaction conditions used, independent of the bimetal nanostructure.²⁷

3.3. Effect of chloride concentration

Chloride is one of the most common ions in the drinking water, with a typical concentration of ∼50 mg L⁻¹ (= ∼50 ppm = ∼0.001408 M).⁴⁸ Werth and coworkers reported that chloride drastically inhibited the activity of alumina supported Cu–Pd catalysts for nitrate/nitrite reduction.⁴⁸ We assessed the deactivation resistance of Pd-on-Au NPs to various chloride concentrations, and compared it to that of Pd/Al₂O₃.

In the presence of 35.5 ppm chloride (0.001 M), the observed reaction rate of Pd/Al₂O₃ decreased by approximately 50%, whereas the reaction rate of Pd-on-Au NPs decreased by ∼2% (Fig. 5). Pd/Al₂O₃ deactivated completely in the presence of 1775 ppm chloride (0.05 M). In contrast, Pd-on-Au NP activity decreased by ∼20%. Consistent with our previous study on chloride deactivation effects during TCE HDC,³⁸ Pd-on-Au NPs showed significantly enhanced resistance to chloride deactivation for nitrite remediation.

Fig. 5 The reaction rate constants k_cat of 60 sc% Pd-on-Au NPs and Pd/Al₂O₃ at various chloride concentrations.

The significant deactivation of Pd/Al₂O₃ catalyst was attributed to the chemisorption of chloride to the Pd surface, blocking the active sites from reaction.³⁸ Chloride anions have been observed to oxidatively absorb to a Pd(111) surface in water under acid conditions.⁴⁹ Pd-on-Au NPs was much less affected by chloride, presumably due to a lesser extent of chloride chemisorption onto the Pd supported on the Au surface.

4. Conclusion

This study demonstrates the application of Pd-on-Au NP catalysts on the reduction of nitrite, a common groundwater contaminant known to cause adverse health effects in humans. While monometallic Au NPs were inactive for the reaction, bimetallic Pd-on-Au NPs exhibited a volcano-shape dependence of activity on Pd surface coverage. 80 sc% Pd-on-Au NPs had maximum activity among all Pd-on-Au compositions, and was more active than monometallic Pd in both NP and Al₂O₃-supported forms. All catalyst compositions showed near-100% selectivity to N₂ over ammonia. The buffered reaction conditions were important to eliminate the pH rise arising from the formation of hydroxide anions during nitrite reduction. The Pd-on-Au catalyst was much more resistant to chloride deactivation compared to Pd/Al₂O₃, at concentrations typical of groundwater and at higher concentrations. These results suggest that bimetallic Pd-on-Au NPs have catalytic properties amenable for water pollution abatement and that the metal-on-gold nanostructure is a useful direction towards designing improved catalysts for the rapid and selective reduction of NO₃⁻ and other oxyanions like perchlorate (ClO₄⁻) and bromate (BrO₃⁻).

Acknowledgements

The authors gratefully acknowledge financial support from the J. Evans Attwell-Welch Postdoctoral Fellowship Program of the Smalley Institute of Rice University (to H.Q.), the National Science Foundation (CBET-1134535), the Welch Foundation (C-1676), and GSI Environmental, Inc. We thank Ms. S. Gullapalli for the assistance in TEM characterization.

References

1. F. T. Wakida and D. N. Lerner, Water Res., 2005, 39, 3–16.
2. G. Jacks and V. Sharma, Environ. Geol., 1983, 5, 61–64.
3. E. Lagerstedt, G. Jacks and F. Sefe, Environ. Geol., 1994, 23, 60–64.
4. B. R. Scanlon, I. Jolly, M. Sophocleous and L. Zhang, Water Resour. Res., 2007, 43, W03437.
5. H. H. Comly, JAMA, J. Am. Med. Assoc., 1987, 257, 2788–2792.
6. S. Shephard, Health aspects of nitrate and its metabolites (particularly nitrite), Proceedings of an international workshop, Bilthoven, Netherlands, 1995.
7. P. Issenberg, Fed. Proc., 1976, 35, 1322–1326.
8. US.EPA, National Primary Drinking Water Regulations and Contaminant Candidate List, ed. US.EPA, 2008.
9. EU, Counicl Directive 98/83/EC, Official Journal of the European Communities, Brussel, 1998.
10. WHO, Guidelines for drinking-water quality, 3rd, World Health Organization 2006.
11. A. Kapoor and T. Viraraghavan, J. Environ. Eng., 1997, 123, 371–380.
12. B. P. Chaplin, M. Reinhard, W. F. Schneider, C. Schuth, J. R. Shapley, T. J. Strathmann and C. J. Werth, Environ. Sci. Technol., 2012, 46, 3655–3670.
13. M. Duca and M. T. M. Koper, Energy Environ. Sci., 2012, 5, 9726–9742.
14. K.-D. Vorlop and T. Tacke, Chem. Ing. Tech., 1989, 61, 836–837.
15. U. Prusse, M. Hahnlein, J. Daum and K.-D. Vorlop, Catal. Today, 2000, 55, 79–90.
16. I. Dodouche and F. Epron, Appl. Catal., B, 2007, 76, 291–299.
17. J. Sa and H. Vinek, Appl. Catal., B, 2005, 57, 247–256.
18. A. Pintar and J. Batista, Appl. Catal., B, 2006, 63, 150–159.
19. F. Zhang, S. Miao, Y. Yang, X. Zhang, J. Chen and N. Guan, J. Phys. Chem. C, 2008, 112, 7665–7671.
20. A. Garron and F. Epron, Water Res., 2005, 39, 3073–3081.
21. F. A. Marchesini, S. Irusta, C. Querini and E. Miro, Appl. Catal., A, 2008, 348, 60–70.
22. A. Kapoor and T. Viraraghavan, J. Environ. Eng., 1997, 123, 371–380.
23. N. Barrabes and J. Sa, Appl. Catal., B, 2011, 104, 1–5.
24. U. Prusse, S. Horold and K.-D. Vorlop, Chem. Ing. Tech., 1997, 69, 93–97.
25. A. Pintar, J. Batista, J. Levec and T. Kajiuchi, Appl. Catal., B, 1996, 11, 81–98.
26. K. Daub, G. Emig, M. J. Chollier, M. Callant and R. Dittmeyer, Chem. Eng. Sci., 1999, 54, 1577–1582.
27. R. Zhang, D. Shuai, K. A. Guy, J. R. Shapley, T. J. Strathmann and C. J. Werth, ChemCatChem, 2012, 5, 313–321.
28. K. A. Guy, H. Xu, J. C. Yang, C. J. Werth and J. R. Shapley, J. Phys. Chem. C, 2009, 113, 8177–8185.
29. U. Prusse and K.-D. Vorlop, J. Mol. Catal. A: Chem., 2001, 173, 313–328.
30. M. O. Nutt, K. N. Heck, P. Alvarez and M. S. Wong, Appl. Catal., B, 2006, 69, 115–125.
31. H. Qian, L. A. Pretzer, J. C. Velazquez, Z. Zhao and M. S. Wong, J. Chem. Technol. Biotechnol., 2013, 88, 735–741.
32. L. A. Pretzer, H. J. Song, Y.-L. Fang, Z. Zhao, N. Guo, T. Wu, I. Arslan, J. T. Miller and M. S. Wong, J. Catal., 2013, 298, 206–217.
33. Z. Zhao, Y.-L. Fang, P. J. J. Alvarez and M. S. Wong, Appl. Catal., B, 2013, 140–141, 468–477.
34. D. F. Boltz and J. A. Howell, Colorimetric Determination of Nonmetals, John Wiley & Sons, New York, 1978.
35. L. C. Green, D. A. Wagner, J. Glogowski, P. L. Skipper, J. S. Wishnok and S. R. Tannenbaum, Anal. Biochem., 1982, 126, 131–138.
36. D. Shuai, J. K. Choe, J. R. Shapley and C. J. Werth, Environ. Sci. Technol., 2012, 46, 2847–2855.
37. M. O. Nutt, J. B. Hughes and M. S. Wong, Environ. Sci. Technol., 2005, 39, 1346–1353.
38. K. N. Heck, M. O. Nutt, P. Alvarez and M. S. Wong, J. Catal., 2009, 267, 97–104.
39. M. S. Wong, P. J. J. Alvarez, Y. L. Fang, N. Akcin, M. O. Nutt, J. T. Miller and K. N. Heck, J. Chem. Technol. Biotechnol., 2009, 84, 158–166.
40. Y.-L. Fang, J. T. Miller, N. Guo, K. N. Heck, P. J. J. Alvarez and M. S. Wong, Catal. Today, 2011, 160, 96–102.
41. I. Dodouche, D. P. Barbosa, M. d. C. Rangel and F. Epron, Appl. Catal., B, 2009, 93, 50–55.
42. B. C. Gates, Catalytic Chemistry, Wiley, New York, 1992.
43. Y.-L. Fang, K. N. Heck, P. J. J. Alvarez and M. S. Wong, ACS Catal., 2011, 1, 128–138.
44. B. E. Koel, A. Sellidj and M. T. Paffett, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 46, 7846–7856.
45. D. I. Enache, J. K. Edwards, P. Landon, B. Solsona-Espriu, A. F. Carley, A. A. Herzing, M. Watanabe, C. J. Kiely, D. W. Knight and G. J. Hutchings, Science, 2006, 311, 362–365.
46. L. Kesavan, R. Tiruvalam, M. H. A. Rahim, M. I. bin Saiman, D. I. Enache, R. L. Jenkins, N. Dimitratos, J. A. Lopez-Sanchez, S. H. Taylor, D. W. Knight, C. J. Kiely and G. J. Hutchings, Science, 2011, 331, 195–199.
47. A. Pintar, M. Vetinc and J. Levec, J. Catal., 1998, 174, 72–87.
48. B. P. Chaplin, E. Roundy, K. A. Guy, J. R. Shapley and C. J. Werth, Environ. Sci. Technol., 2006, 40, 3075–3081.
49. A. Carrasquillo, J. J. Jeng, R. J. Barriga, W. F. Temesghen and M. P. Soriaga, Inorg. Chim. Acta, 1997, 255, 249–254.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3nr04540d

---