Article pubs.acs.org/est
Development of Pd−Cu/Hematite Catalyst for Selective Nitrate Reduction Sungyoon Jung, Sungjun Bae, and Woojin Lee* Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, 291 Daehak-ro, Yuseong-Gu, Daejeon 305-701, Korea S Supporting Information *
ABSTRACT: A new hematite-supported Pd−Cu bimetallic catalyst (Pd−Cu/hematite) was developed in order to actively and selectively reduce nitrate (NO3−) to nitrogen gas (N2). Four different iron-bearing soil minerals (hematite (H), goethite (G), maghemite (M), and lepidocrocite (L)) were transformed to hematite by calcination and used for synthesis of different Pd−Cu/hematite-H, G, M, and L catalysts. Their characteristics were identified using X-ray diffraction (XRD), specific surface area (BET), temperature programed reduction (TPR), transmission electron microscopy with energy dispersive X-ray (TEM-EDX), H2 pulse chemisorption, zetapotential, and X-ray photoelectron spectroscopy (XPS). Pd− Cu/hematite-H exhibited the highest NO3− removal (96.4%) after 90 min, while a lower removal (90.9, 51.1, and 30.5%) was observed in Pd−Cu/hematite-G, M, and L, respectively. The results of TEM-EDX, and TPR analysis revealed that Pd−Cu/hematite-H possessed the closest contact distance between the Cu and Pd sites on the hematite surface among the different Pd−Cu/hematite catalysts. The high removal can be also attributed to the highly active metallic sites on its positively charged surface. The XPS analysis demonstrated that the amount of hydrogen molecules can have a pivotal function on NO3− removal and a ratio of nitrogen to hydrogen molecule (N:H) on the Pd sites can critically determine N2 selectivity.
■
Since Vorlop and Tacke demonstrated that catalytic NO3− reduction actively and selectively degrades NO3− to harmless nitrogen gas (N2) without further NO3− treatment using a secondary process,8 numerous efforts have been made to develop novel bimetallic catalysts.6,9−12 According to previous bimetallic catalyst studies, NO3− can be converted to NO2− on promoter metal sites (Cu, Sn, and In) and NO2− can be further reduced to both ammonium (NH4+) and N2 on noble metal sites (Pd, Pt, and Rh). A variety of bimetallic combinations have been extensively investigated to date, and a combination of Pd−Cu is widely accepted as the most active and selective combination for catalytic NO3− reduction.5,13,14 Studies using different materials such as alumina,5 titania,6 silica,15 activated carbon (AC),16 and polymers17 have been also carried out to find out a proper support material for the catalytic denitrification. However, most support materials used in the studies are not of natural origin, requiring an artificial process for their synthesis. Recently, Jung et al. have used maghemite (γ-FeIII2O3) as an eco-friendly support material for the catalytic NO3− reduction and achieved a remarkable NO3− removal
INTRODUCTION
Nitrate (NO3−) contamination of surface and groundwater has gradually increased due to the excessive consumption of fertilizer and poor disposal of animal waste1 and industrial wastewater2 in both developing and developed countries. NO3− is a potentially harmful contaminant because it can be converted to nitrite (NO2−) in the human body and can cause methemoglobinemia, which results in a decrease in the oxygen-carrying capacity of hemoglobin, particularly in infants under 6 months of age.3 Furthermore, NO2− can be transformed into potential stomach cancer carcinogens, for example, n-nitroso compounds, via nitrosation.4,5 Due to the toxic effect of NO3− and its transformation products, the World Health Organization set the maximum concentration of NO3− in drinking water to 10 NO3−-N mg/L (nitrogen concentration in NO3−).6 Diverse environmental technologies, for example, reverse osmosis, ion exchange, and biochemical treatments, have been developed and applied to treat the groundwater contaminated by NO3−.2,3,5,7 However, their drawbacks have been reported frequently. For example, the reverse osmosis and ion exchange processes require a secondary treatment process to decrease the high NO3− concentration in their effluents.3,5,7 The formation of surplus sludge during the biochemical treatments has also been frequently reported.5 © 2014 American Chemical Society
Received: Revised: Accepted: Published: 9651
February 18, 2014 July 29, 2014 July 30, 2014 July 30, 2014 dx.doi.org/10.1021/es502263p | Environ. Sci. Technol. 2014, 48, 9651−9658
Environmental Science & Technology
Article
(99.5%) in 90 min.18 However, their highest N2 selectivity with Pd−Cu/maghemite catalyst was only 47% at diverse environmental conditions. Therefore, it is timely to develop a new natural eco-friendly bimetal support material for higher selective catalytic NO3− reduction. Hematite (α-FeIII2O3), one of the most common natural iron-bearing soil minerals in the earth,19 has been extensively studied for treating environmental contaminants due to its abundance, stability, and nontoxicity.20,21 It has been recently reported that the Au/hematite catalyst, which have uniformly distributed Au particles, can efficiently reduce α, β-unsaturated ketones.22,23 The catalyst exhibited significantly higher reduction efficiencies and selectivities toward α, β-unsaturated alcohols than the Au/alumina catalyst.23 Although these studies were performed using monometallic catalysts, they indicate that hematite is a potential support material that can provide appropriate sites for deposition of metals and can treat NO3− actively because the well-dispersed metals on the support surface can significantly improve the NO3− removal.24 However, significant foundational knowledge has not yet been provided for developing a bimetallic catalyst supported by nontoxic eco-friendly hematite for active NO3− reduction with high N2 selectivity. Furthermore, influential research has not yet been conducted to appropriately use and apply this process. The objectives of this study are to develop a bimetallic Pd− Cu catalyst supported by hematite, to investigate the NO3− reduction mechanism using catalysts, and to investigate the effect of significant environmental factors on the catalytic NO3− reduction and N2 selectivity. Four different Pd−Cu/hematite catalysts were synthesized from different iron-bearing soil mineral supports (hematite, goethite (α-FeIIIOOH), maghemite, and lepidocrocite (γ-FeIIIOOH)) and their characteristics were identified using X-ray diffraction (XRD), specific surface area (BET), temperature-programmed reduction (TPR), transmission electron microscopy, and energy dispersive X-ray (TEM/EDX), H2 pulse chemisorption, zeta-potential, and Xray photoelectron spectroscopy (XPS).
2PdCl 2 + NaBH4 + 3H 2O → 2Pd(0) + B(OH)3 + NaCl + 3HCl + 2H 2
(1)
2CuCl 2· 2H 2O + NaBH4 + H 2O → 2Cu(0) + B(OH)3 + NaCl + 3HCl + 2H 2O + 2H 2
(2)
The reduced Pd−Cu/hematite catalysts were vacuum-filtered using a 0.2 μm membrane filter (Advantac, Japan) and washed with deaerated deionized water (DDI water, purged with Ar gas for 4 h) twice to remove the residual chemicals. The washed catalysts were used immediately for the catalytic NO3− reduction test. Pd−Cu/SiO2, Pd−Cu/CeO2, and Pd−Cu/ Al2O3 catalysts were prepared using same method described above. A complete list of chemicals used in this study is provided in the Supporting Information (SI). Catalytic NO3− Reduction Test. The catalytic NO3− reduction test was performed in a 500 mL glass batch reactor equipped with a mechanical stirrer, gas inlet, gas outlet, and two ports for injection and sampling. DDI water (199.7 mL) containing 0.25 g of Pd−Cu/hematite was introduced to the reactor and ultrasonicated for 1 min for complete dispersion of the catalyst particles in the suspension. Prior to the initiation of the denitrification, H2 was bubbled through the diffuser for 30 min in order to remove the dissolved oxygen in the suspension. The catalytic NO3− reduction was initiated through the addition of 0.3 mL of NO3− stock solution (20,000 mg/L as NO3−-N) in order to obtain an initial concentration of 30 mg/L as NO3−-N at room temperature (25 °C). H2 and CO2 gases were continuously introduced to the reactor in order to reduce NO3− and adjust the suspension pH to 6, respectively. Control experiments with different stirring speed (80, 100, 140, and 200 rpm) showed that more than 98% of NO3− was removed at all experimental conditions in 120 min, indicating that external mass transfer by mixing did not affect the NO3− reduction in this study. The NO3− reduction tests were also conducted with three different catalysts (Pd−Cu/SiO2, Pd−Cu/CeO2, and Pd−Cu/Al2O3) and controls (hematite only, Cu/hematite, and Pd/hematite) in order to compare their NO3− removal with that using the Pd−Cu/hematite catalyst. Unless stated otherwise, the experimental conditions were fixed at 2.8 wt % Pd, 1.6 wt % Cu, 30 mL/min H2, and 40 mL/min CO2. Five different loadings of Pd and Cu (0.4, 1.0, 1.6, 2.2, and 2.8 wt %) and five different flow rates of H2 (10, 20, 30, 40, and 50 mL/ min) were used to investigate their effects on the NO3− removal and transformation product selectivity. All experiments were conducted in duplicate. Characterization of Catalysts. The characteristics of synthesized catalysts were identified with XRD, BET surface area, TEM-EDX, TPR, ICP-AES, H2 pulse chemisorption, zetapotential, and XPS. More details on instrumentation, sample preparation, and analytical procedures are provided in the SI. Analytical Methods. In order to measure the concentrations of NO3−, NO2−, and NH4+, 4 mL samples was removed from the reactor using a 5 mL disposable syringe (Korea Vaccine Corp., Korea) at each sampling time. The sample was immediately filtered using a 0.2 μm membrane filter and 1 mL of the filtrate was used to measure the concentrations of NO3− and NO2− using a HPLC. The remaining filtrate (3 mL) was used to measure the NH4+ concentration via IC. The amount of nitrous oxide (N2O) and N2 were measured in closed batch
■
EXPERIMENTAL METHODS Catalyst Preparation. Four bimetallic Pd−Cu/hematite catalysts were synthesized using hematite (Pd−Cu/hematiteH), goethite (Pd−Cu/hematite-G), maghemite (Pd−Cu/ hematite-M), and lepidocrocite (Pd−Cu/hematite-L) through the impregnation method18 and they were evaluated to determine the most efficient Pd−Cu/hematite catalyst among them. The support materials (1.5 g of each hematite, goethite, maghemite, and lepidocrocite) were mixed with 200 mL of deionized water (DI water, 18 MΩ·cm) and underwent ultrasonic vibration for 6 min prior to the addition of the Cu (CuCl2·2H2O) and Pd (PdCl2) precursors. Each Cu precursor solution (0.4−2.8 wt %, prepared with DI water) and Pd precursor solution (0.4−2.8 wt %, prepared with 0.5 M HCl) was sequentially introduced into the support suspensions and they were mixed under continuous stirring for 2 h. The mixed solutions with Cu and Pd precursors were dried in an oven at 105 °C for 24 h and then calcinated for stabilization of the precursors on the surface of support at 350 °C for 2 h. Next, the calcinated bimetallic catalysts were reduced through the dropwise addition of an excessive amount of 0.01 M NaBH4 (the molar ratios of NaBH4 to Cu and Pd were higher than 2:1 and 3:1, respectively) in order to completely reduce the Cu and Pd on the catalyst surface through the following reactions:25 9652
dx.doi.org/10.1021/es502263p | Environ. Sci. Technol. 2014, 48, 9651−9658
Environmental Science & Technology
Article
Figure 1. TEM images of (a) Pd−Cu/hematite-H, (b) Pd−Cu/hematite-G, (c) Pd−Cu/hematite-M, and (d) Pd−Cu/hematite-L, and their EDX spectra at 2.8 Pd wt % and 1.6 Cu wt %.
system with a gas chromatograph (GC) equipped with a thermal conductivity detector. Further details for analytical methods are provided in the SI.
■
Fe2O3 + 2HCl + H 2O → 2FeOCl + 2H 2O
(3)
2FeOOH + 2HCl + H 2O → 2FeOCl + 3H 2O
(4)
Hematite and goethite exhibited less significant change in the iron oxychloride formation than maghemite and lepidocrocite due to their remarkably stable and crystalline structures.26 After calcination at 350 °C, the peaks of the iron oxychloride disappeared completely in all samples and well-organized new peaks corresponding to the standard peaks of hematite were observed (SI Figure S1i−l), which indicates that the four different support materials were finally transformed into hematite through calcination at 350 °C. In contrast to these results demonstrating the transformation of lepidocrocite to hematite, Jung et al. demonstrated that lepidocrocite was transformed to maghemite under the same synthesis procedure, except with the addition of a Pd precursor prepared in 0.5 M HCl.18 This indicates that the four different iron-bearing soil mineral supports transformed into hematite through the formation of an intermediate (i.e., iron oxychloride). It has been reported that iron oxychloride can be transformed into hematite through thermal treatments at approximately 377 °C.27 Therefore, it can be concluded that the formation of iron
RESULTS AND DISCUSSION
Characteristic of the Pd−Cu/hematite Catalysts. The XRD analysis was conducted in order to investigate the transformation of four different iron-bearing soil mineral supports for Pd−Cu/hematite catalysts during the synthesis procedure. SI Figure S1a−d presents the well-ordered peaks of the initial iron-bearing soil minerals (hematite, goethite, maghemite, and lepidocrocite). After adding the Cu and Pd precursors, increases in the iron oxychloride (FeOCl) peaks were observed in the Pd−Cu/hematite-H (SI Figure S1e) and Pd−Cu/hematite-G (SI Figure S 1f), but the initial peaks of hematite and goethite remained dominant. For the Pd−Cu/ hematite-M and Pd−Cu/hematite-L, the initial peaks of maghemite and lepidocrocite almost transformed into those of iron oxychloride (SI Figure S1g and h). The formation of iron oxychloride might result from using HCl (0.5 M) during the catalyst synthesis as resultant reactions. 9653
dx.doi.org/10.1021/es502263p | Environ. Sci. Technol. 2014, 48, 9651−9658
Environmental Science & Technology
Article
oxychloride during the synthesis is a crucial process in producing the same final transformation product (i.e., hematite). Both Pd(0) and Cu(0) peaks were not observed in the diffractograms of the Pd−Cu/hematite catalysts due to the low bimetal loadings (2.8 wt % Pd and 1.6 wt % Cu), which were near the detection limit of the XRD analysis.28 Surface area in all catalysts decreased after the calcination (Pd−Cu/hematite-H (12.7 → 11.7 m2g−1), Pd−Cu/hematiteG (12.1 → 11.7 m2g−1), Pd−Cu/hematite-M (36.2 → 1.7 m2g−1), and Pd−Cu/hematite-L (15.9 → 2.2 m2g−1)) (SI Table S1). It has been reported that the specific surface area of the bimetallic catalyst decreased via thermal treatment due to shrinkage of the material surface by water evaporation.18,29 Pd− Cu/hematite-H and Pd−Cu/hematite-G showed a relatively small decrease due to their extremely stable structures as mentioned above.26 On the other hand, an increasing trend was observed in most catalysts after reduction with NaBH4 (Pd− Cu/hematite-H (11.7 → 15.9 m2g−1), Pd−Cu/hematite-G (11.7 → 11.4 m2g−1), Pd−Cu/hematite-M (1.7 → 6.9 m2g−1), and Pd−Cu/hematite-L (2.2 → 7.3 m2g−1)) (SI Table S1). This may be caused by the formation of zerovalent Cu(0) and Pd(0) by the addition of NaBH4. The TEM-EDX analysis was conducted to investigate the morphological characteristics of different Pd−Cu/hematite catalysts (Figure 1). The amorphous shape of the hematite particles from 40−500 nm were observed in all Pd−Cu/ hematite catalysts. It should be noted that many nanosized particles existed on the catalyst surfaces (Figure 1a-d). The EDX analysis revealed that these nanoparticles were Cu and Pd particles (Figure 1(inset)). Most of bimetal particles were ranged in 4−8 nm in all cases, except Pd−Cu/hematite-H catalyst in the range of 3−5 nm (SI Figure S2). Furthermore, STEM-EDX mapping also revealed that Cu and Pd were well dispersed on the surface of all catalysts (SI Figure S3). Especially, both metals were much more closely and uniformly distributed on the surface of Pd−Cu/hematite-H than that of other Pd−Cu/hematite catalysts. The result indicates that Pd− Cu/hematite-H provides the most appropriate surface environment for the catalytic activity of bimetallic catalysts.24 TPR analysis was conducted in order to investigate the interaction of Pd and Cu on the hematite surface (Figure 2). A peak of negative hydrogen consumption near 60 °C caused by the decomposition of Pd β-hydride was not observed in all catalysts, indicating that the Pd particles were well distributed on the surface of all hematite catalysts.18,31,32 The TPR patterns of the Pd−Cu/hematite catalysts exhibited a positive hydrogen consumption peak between 250 and 360 °C, which was assigned to the reduction of the Cu oxide.11,30 In the TPR profiles of Pd−Cu/hematite-H, the reduction peak of Cu oxide occurred at the lowest temperature (292 °C) followed by Pd− Cu/hematite-M (293 °C), Pd−Cu/hematite-G (310 °C), and Pd−Cu/hematite-L (326 °C) (Figure 2). It is known that the peak demonstrating the reduction of Cu oxide tends to shift toward the lower temperature region due to an enhanced hydrogen spillover from Pd to Cu oxide sites when the Cu particles are closely associated with the Pd particles on the catalyst surface. Therefore, the Pd−Cu/hematite-H exhibiting a peak at the lowest temperature may have the closest contact distance between Cu and Pd among the Pd−Cu/hematite catalysts synthesized in this study. In addition, all catalysts exhibited a broad peak increase starting from approximately 360 °C, which is attributed to the thermal reduction of
Figure 2. TPR profiles of (a) Pd−Cu/hematite-H, (b) Pd−Cu/ hematite-G, (c) Pd−Cu/hematite-M, and (d) Pd−Cu/hematite-L at 2.8 Pd wt % and 1.6 Cu wt %.
hematite to magnetite (FeIIFeIII2O4) via H2 during the TPR analysis.18 Catalytic NO3− Reduction using the Pd−Cu/hematite Catalysts. Figure 3 presents the NO3− removal via the Pd−
Figure 3. NO3− removal by Pd−Cu/hematite-H, Pd−Cu/hematite-G, Pd−Cu/hematite-M, Pd−Cu/hematite-L, Pd−Cu/SiO2, Pd−Cu/ CeO2, and Pd−Cu/Al2O3 (1.25 g/L) at 90 min (2.8 Pd wt %, 1.6 Cu wt %, and 30 and 40 mL/min for H2 and CO2 flow rates).
Cu/hematite catalysts in order to select the most active Pd− Cu/hematite catalyst. Pd−Cu/hematite-H exhibited the highest NO3− removal (96.4%) at 90 min and was followed by Pd−Cu/ hematite-G (90.9%), Pd−Cu/hematite-M (51.1%), and Pd− Cu/hematite-L (30.5%). As shown in the TPR and STEMEDX mapping analysis, a closer contact distance between Pd and Cu on the surface of Pd−Cu/hematite-H could result in the highest NO3− removal. The closer contact distance could increase the NO3− removal due to the enhanced Cu oxide reduction through hydrogen molecules from Pd site.11,30,32 Based on the experimental results, Pd−Cu/hematite-H was the most active catalyst for NO3 − reduction and it was 9654
dx.doi.org/10.1021/es502263p | Environ. Sci. Technol. 2014, 48, 9651−9658
Environmental Science & Technology
Article
Figure S5), which indicates that the adsorption of NO3− on the hematite surface was negligible and that the NO3− reduction via hematite and monometallic catalysts did not significantly occur. The catalytic denitrification test using Pd−Cu/hematite under an experimental condition (2.8 wt % Pd, 1.6 wt % Cu, and 30 mL/min of H2 flow rate) was also conducted to investigate the transformation product distribution (SI Figure S5). During the catalytic NO 3 − reduction, NO 2 − increased slightly to approximately 4% after 10 min in the early stage of the reaction, but decreased to 0% at the reaction completion. In contrast, the NH4+ concentration increased continuously to 37% after 90 min. N2O and N2 were not detected throughout the experiment due probably to the high amount of H2 and CO2 provided into a semibatch reactor. To investigate the formation of N2O and N2 possibly produced in the reactor,34 we performed the NO3− reduction in a closed-batch reactor (SI Figure S6). Total N mass balance was kept at >95% throughout the test. A fair amount of NO3− (70%) was removed in 300 min. The concentrations of NH4+ and N2 increased to approximately 30 and 70%, however NO2− formation was very low (
