Pd–Au nanoparticles supported by TiO2 fibers for catalytic NO decomposition by CO

Pd–Au nanoparticles supported by TiO2 fibers for catalytic NO decomposition by CO

Journal of Industrial and Engineering Chemistry 33 (2016) 91–98 Contents lists available at ScienceDirect Journal of Industrial and Engineering Chem...

3MB Sizes 0 Downloads 23 Views

Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec

Pd–Au nanoparticles supported by TiO2 fibers for catalytic NO decomposition by CO Hyeon Ung Shin a, Dinesh Lolla a, Zhorro Nikolov b, George G. Chase a,* a b

Department of Chemical and Biomolecular Engineering, The University of Akron, OH 44325, USA College of Polymer Science and Polymer Engineering, National Polymer Innovation Center, The University of Akron, OH 44325, USA

A R T I C L E I N F O

Article history: Received 19 June 2015 Received in revised form 18 September 2015 Accepted 19 September 2015 Available online 28 September 2015 Keywords: Pd–Au Catalyst Fiber Electrospinning

A B S T R A C T

Pd–Au nanoparticles supported by TiO2 fibers were fabricated and assembled into a porous medium. The 2–16 nm Pd–Au nanoparticles were randomly dispersed in the titania nanofibers (da = 117  68 nm). Xray diffraction patterns and X-ray photoelectron spectroscopy verified the formation of Pd–Au particles. Binding energy shifts were detected due to chemical bonding and reduction with hydrazine monohydrate. The fiber media with Pd–Au nanoparticles had greater reactivity in conversion of NO and CO gases than Pd monometallic catalyst alone, attributed to a lower activation energy of the reaction on the Pd–Au catalyst particles. ß 2015 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

Introduction Catalyzed reactions are considered the most efficient and cost effective methods to simultaneously eliminate NO and CO pollutant molecules form air [1]. The growing concerns for the environment have resulted in increasingly restrictive emission standards [2]. Removal of NO and CO from the exhaust streams of automobiles, power plants, and other combustion processes is a challenging task [3,4]. These gases are byproducts of incomplete combustion of hydrocarbon fuels (HC) (gasoline and diesel) from engines [5]. Particulate matter (PM) (from diesel) is also another undesired product which also creates several toxic and hazardous substances by post reaction. During the combustion process, very high temperatures are reached due to the generation of heat results in thermal fixation of N2 in air to form NOx with concentrations in the range of 100–3000 ppm [6,7]. Thus, to solve this problem, the use of CO as a reducing agent have been considered and investigated. NO reduction by CO is one of the important model reactions describing the operation of three-way catalysis, which is attracting more attention of catalytic researchers due to the simultaneous elimination of NO and CO pollutant molecules [8]. It is reported that a three way catalytic converter has the capacity to simultaneously eliminate three pollutants such as (1) reduction

* Corresponding author. Tel.: +1 3309727943. E-mail address: [email protected] (G.G. Chase).

of nitrogen oxides to nitrogen and oxygen, (2) oxidation of carbon monoxide to less harmful carbon dioxide and (3) oxidation of unburnt hydrocarbons (HC) to carbon dioxide and water [9,10]. The amount of NOx, CO and hydrocarbon emission in combustion varies, depending on the air-to-fuel ratio (A/F ratio) [11]. Three way catalysts can effectively remove the pollutants when the air/fuel (A/F) is close to its stoichiometric ratio for complete conversion [12]. Appropriate amounts of reducing and oxidizing agents need to be present in the exhaust to conduct reactions [13,14]. If an engine could be held to operate at the stoichiometric condition for the fuel used, it is theoretically possible to obtain 100% conversion efficiencies. Nano-sized catalyst particles have high surface areas per unit mass that enhances heterogeneous reaction rates. Immobilized nanoparticles supported on electrospun fibers have high interfacial surface areas for the catalyst particles to contact the reactant gases [15]. Pd, Pt, and Rh doped sub-micron sized alumina fibers have been investigated and applied for manufacture of fibrous porous fiber media to perform NO reaction with metal groups in the presence of CO [16,17]. Shahreen et al. noted that NO decomposition temperature depends on the anatase phase concentration of Pd doped TiO2 fibers [18]. Furthermore, Pd–Au catalysts are used in the industrial production of vinyl acetate (VA) [19,20]. It has also been successfully used to catalyzed many reactions including direct H2O2 synthesis from H2 and O2 [21,22], hydrodechlorination of Cl-containing pollutants in underground water [23], hydrodesulfurization of S-containing organics [24], hydrogenation of

http://dx.doi.org/10.1016/j.jiec.2015.09.020 1226-086X/ß 2015 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

92

H.U. Shin et al. / Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

phenol [25], and coupling of acetylene to benzene [26]. Goodman et al. also tested Pd–Au catalyst CO oxidation as a Langmuir– Hinshelwood type of reaction [20,27–30]. They reported that at near-atmospheric pressures, a sufficient amount of Pd segregates to the surface and generates contiguous Pd sites, which are capable of dissociating O2 to allow the CO2 formation reaction [27,29]. Moreover, the addition of Au can weaken the CO binding energy due to the charge transfer from Pd to Au [29,31]. Thus, Pd– Au alloys can perform the CO oxidation more efficiently than pure Au and pure Pd catalyst at low temperature. However, no studies have been reported to date on fiber supported Pd–Au for the NO and CO reaction though Pd–Au alloy catalyst was proven to be superior to single metal catalyst for the hydrogen evolution [32,33] and CO oxidation [34] reactions. In this work Pd–Au, Pd, Au composite nanoparticles were dispersed on and within TiO2 submicron sized fibers to form Pd–Au/TiO2, Pd/TiO2, and Au/TiO2 composite fiber structures. These fibers were characterized by SEM, TEM, EDX, XRD, BET and XPS. Reaction performances of the Pd–Au and Pd particles on titania fibers were compared.

Experimental Preparation of electrospinning solutions A 16 wt% solution of polyvinylpyrrolidone (PVP, Aldrich, MW: 1,300,000) was prepared by dissolving PVP in 10 ml ethanol (AAPER alcohol, 200 proof), 10 ml N,N-dimethylformamide (Fisher Scientific, 99.8%), and 10 ml acetone (Sigma–Aldrich, 99.5%) and stirred for 12 h. A titanium isopropoxide solution was prepared by blending 6 ml titanium isopropoxide (Sigma–Aldrich, 97%), 12 ml acetic acid (Sigma–Aldrich, 99.7%), and 12 ml ethanol. The PVP and titanium solutions were combined well stirred for 4 h at room temperature to form the electrospinning solution for synthesizing fibers. To form the fibers with catalyst particles the above polymer mixture was combined with appropriate powder of the catalyst precursors. The powers consisted of palladium chloride (Pressure Chemical, 60% Pd), gold chloride hydrate (Sigma–Aldrich, 99.999% metal basis), and a blended mixture of palladium and gold powder precursors in the amounts of 86.4 mg (0.49 mmol) of palladium chloride and 57.6 mg (0.17 mmol) of gold chloride hydrate. The final polymer solution contained 2.5 wt% PdCl2 + HAuCl4 with respect to titanium isopropoxide. This solution was stirred overnight at 40 8C using a magnetic stirrer.

Preparation of Pd–Au/TiO2 fibers and catalytic fiber medium The electrospinning solution was electrospun into PVP fibers containing catalyst and titanium precursors using a laboratory electrospinning setup as depicted in Fig. 1(A) and described elsewhere [15,35,36]. Two plastic syringes (BD 5 ml Syringe, Fisher Scientific) were filled with the prepared electrospinning solution and attached to the syringe pump and the power supply. The syringe pump (SP101I, WPI) was adjusted to provide a 3 ml/min of the electrospinning solution at the tips of the 21 gauge steel needles. A 25 kV of electric voltage potential was applied to the needles from the power supply (Gamma High Voltage, Ormond Beach, FL). The electric field overcame the surface forces on the droplets at the tips of the needles and caused jets to launch from the droplets toward the grounded collector [37–40]. The collected electrospun fibers were calcined in air at 550 8C to remove the polymer and to convert the sol precursor to ceramic form. The PdCl2 and HAuCl4 were reduced to metallic form by reaction with hydrazine monohydrate [41,42]. The reduction was carried out by immersing the electrospun and calcined PdCl2 + HAuCl4/TiO2 fiber in the hydrazine monohydrate solution. The treated fibers were separated from the hydrazine solution by filtration, washed using ethanol and vacuum-dried at 80 8C for 24 h to remove the residual solvents. The TiO2 fibers containing catalyst particles were fabricated into a high permeability fibrous mat structure by blending the submicron fibers with chopped alumina microfibers into an aqueous slurry and vacuum molding the slurry with the apparatus depicted in Fig. 1(B). The aqueous slurry was formed by blending into 4 l of deionized water 0.5 g of chopped alumina microfibers (Saffil HA bulk fibers, Thermal Ceramics), 0.05 g of Pd–Au supported by titania submicron fibers (synthesized in this work), 0.02 g of corn starch, 0.5 ml of binder (Megasol S50, Wesbond Corporation), and dilute H2SO4 acid (enough to bring the slurry to a pH of 6.0). The starch was added in order to enhance the binding capability of the Megasol S50 to bind the fibers together at junction points between the fibers. The liquid was drawn through the mold by vacuum into the collection tank with aid of a vacuum pump. The fibers were retained on the steel mesh covered by a Whatman filter paper to form a fibrous wet cake and the wet cake was heated to 500 8C to dry the cake and bind the fibers together [16]. Fig. 2 shows examples of the dry fiber media. The vacuum mold had an inside diameter of 2.3 cm resulting in the fibrous media to have approximately the same 2.3 cm diameter.

Fig. 1. (A) Schematic diagram of the double-needle electrospinning setup and (B) vacuum molding setup.

H.U. Shin et al. / Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

93

Fig. 2. Fiber media of 2.3 cm diameter formed of alumina microfibers and (A) Pd–Au/TiO2 (B) Pd/TiO2, (C) Au/TiO2, and (D) TiO2 submicron fibers.

Catalytic reaction set up for NO decomposition by CO The test sample medium was placed in a steel reactor with a 2.3 cm diameter hole. The reactor was wrapped with a heating tape (Omega FGS051-060) connected to a temperature controller (Micromega CN77000). Gas concentrations before and after the rector were measured with gas chromatography (TCD GC, SRI 8610C). Peak Simple software (Version 3.29) was used to calculate and display the peaks and concentrations of the chemical components of the gas. The reaction gases consisted of 1 mol/ mol% NO, 1% CO, and 98% He, and the inlet volumetric flow rate of 100 sccm. Each experiment was repeated three times with the same inlet flow rates and same inlet concentration ratio of the gases. A flow diagram of the laboratory set up is shown in Fig. 3. The NO and CO conversions were calculated based on the following equations and definitions. The rate of conversion is defined by Xa ¼

F a;0  F a F a;0

(1)

where Fa,0 and Fa are the inlet and outlet molar flow of gas species a (mol/s), respectively. The conversion, X (Eq. (1)), is the total number of moles of gas a, that reacted per mole of a, fed to the reactor. The disappearance rate (ra0 ) of NO and CO per mass of catalyst, was calculated using [43] ra0 ¼

F a;0  X a C a;0  v0  X a v0 PT  Xa ¼ ¼  DW cat DW cat DW cat RT

(2)

where Xa is the fractional conversion of gas a, DWcat is the amount of catalyst (g), Ca,0 is the inlet concentration of gas a (mol/cm3), and v0 is the inlet volumetric flow rate (cm3/s).

The inlet concentration, Ca,0, can be calculated from the inlet stream temperature and pressure, molar flow rates, and volumetric flow rate via the ideal gas law. The disappearance rate (ra00 ) of NO and CO was calculated using (ra00 ¼ ra0 =Smetal ), where Smetal is the specific surface area of the Pd and Pd–Au nanoparticle catalysts. The turnover frequency (TOF) for NO and CO was calculated by [43] TOF ¼

ra00  NA S0

(3)

where NA is Avogadro’s number, and S0 is the surface atom density of Pd–Au nanoparticles. Material characterization of the fibers Scanning electron microscope (FEI Quanta 200 at 30 kV and HITACHI TM3000 at 15 kV) images were analyzed to study the fiber morphology and fiber diameter distribution. Energy dispersive Xray analysis, EDX (Bruker Quantax 70) was used for elemental quantification analysis. It was designed to run on the SEM TM3000 simultaneously. Multiple spectra were displayed for quantitative comparison of the elements. Transmission Electron Microscope (TEM, JEM 1200XII and HRTEM, FEI Tecnai G2 F20) images were used to study catalytic particle size and morphology of Pd–Au nanoparticles supported by TiO2 nanofibers. X-ray diffractograms (Bruker AXS Dimension D8) were used for determination of the crystal structure of the fibers. The Brunauer, Emmett and Teller (BET) surface area of materials was measured using nitrogen gas adsorption. Each sample was dried in a clean dry tube at 100 8C for four hours in a degas station (Micromeritics Vac Prep 061) to remove the moisture from the samples. The combined masses of the sample tube and stopper were determined to 0.1 mg precision using a calibrated microbalance (AX205, Mettler Toledo). Surface areas were measured using Surface Area and Porosity Analysis Instrument (Micromeritics Tristar II). A value of 0.162 nm2 was used for the molecular crosssectional area of N2 at 77 K. High resolution X-ray photoelectron spectroscopy was used to investigate the chemical composition and chemical bond strengths. The XPS spectra were obtained using a VersaProbe II Scanning XPS Microprobe from Physical Electronics (PHI), under ultrahigh vacuum conditions with a pressure of 1  106 Pa. Automated charge neutralization was used during the analysis of the samples to provide accurate data. The analyzer pass energy was 23.5 eV for the high-resolution Pd 3d and Au 4f scans. The XPS spectra were recorded at a takeoff angle of 458. Each spectrum was collected using a monochromatic (Al Ka) X-ray beam (E = 1486 eV) over a 200 mm diameter probe area. Curve fitting of the peaks was performed by the PHI MultiPak software that applied the Shirley background correction method. Results and discussion

Fig. 3. Flow diagram for measuring catalytic performance.

Fig. 4(A) shows a sample SEM image and a sample fiber diameter distribution calculated from multiple images of the Pd– Au/TiO2 fibers. The morphologies of the fibers were continuous,

94

H.U. Shin et al. / Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

Fig. 4. (A) SEM image of Pd–Au/TiO2 fibers, (B) fiber diameter distribution and (C) SEM image with (D) the corresponding EDX analysis of elemental composition of Pd–Au/TiO2 fibers.

smooth and without beads. Fibraquant 1.3 software (nanoScaffold Technologies LLC) was used to interpret the images to determine the fiber size distributions. The fiber diameters were in the range of 30–300 nm with an average diameter of 117  68 nm as indicated in Fig. 4(B). Fig. 4(C) shows the highlighted region (within the circle) that was evaluated by EDX. The EDX spectra and the elemental compositions of the fibers are listed in the table in Fig. 4(D). The data show the fibers in that particular sample had 2.73 wt% of Pd–Au metal which is close to the expected 2.5 wt% based on the material compositions of the electrospinning solutions. The atomic molar ratio of Pd:Au = 2.6:1 (EDX data) is also a reasonable comparison with the Pd–Au molar ratio of 2.88(Pd):1(Au) expected from composition of the electrospinning solution. Pd–Au particle composition is approximately similar in the molar ratios of the metal precursors, indicating that reduction of both precursors was successfully carried out after calcination process. The percentages of total metal loading were 2.73% for Pd–Au (particle composition: Pd68Au32), 2.82% for Pd, 3.16% for Au. Fig. 5 shows the TEM images and size distribution of catalyst particles in or on the fibers. Fig. 5(B) shows a good dispersion of Pd–Au particles within and near the surface of the TiO2 nanofibers. The statistically calculated average particle size was 7 nm with a standard deviation of 3 nm. Fig. 5(D) and (E) was magnified to study the lattice fringe areas and the selected-area electron diffraction (SAED). The HRTEM image in Fig. 5(D) shows clear lattice spacing throughout the Pd– Au nanoparticles, suggesting a single crystallinity and alloy formation. However, some of the Pd and Au in the fibers may exist as a distribution of monometallic crystallites or as bimetallic particles. More extensive analysis in future work is needed to determent the full extent of the presence of these compositions.

The crystallinity of the Pd–Au/TiO2 fibers was confirmed by SAED, which revealed two broad diffraction rings corresponding to the (101) for anatase-TiO2 and (111) for Pd–Au particles reflections in Fig. 5(E). The relatively faint unlabeled diffraction rings may be indications of some other crystal structures of the titania component of the Pd–Au/TiO2 fibers. Fig. 6 shows the XRD peaks of Pd–Au/TiO2 and Pd/TiO2 fibers after reduction with the hydrazine monohydrate. The characteristic peaks for Pd–Au, marked by the indices (111), and (200), revealed that the resultant particles were essentially face-centered cubic (fcc) structures. For comparison of peaks position among Au/TiO2, Pd–Au/TiO2 and Pd/TiO2, it was also found in the inset of Fig. 6 that the peak positions slightly shifted toward a lower diffraction angle. The shift in XRD peaks is consistent with an increase of Au content, showing a successful fabrication of some alloyed Pd–Au particles [44,45]. The surface textural characteristics for Pd–Au nanoparticles supported by TiO2 nanofibers before and after reduction process were estimated via gas adsorption measurements by studying the adsorption and desorption isotherms with the inset corresponding BJH pore size distribution (Fig. 7). Both of the samples had similar profiles with classical type IV adsorption model performances [46]. The capillary condensation steps occurred at a relative pressure of about 0.45. In plots in the inset in Fig. 7 shows the BJH pores had narrow size distributions in the range of 1 to 13 nm. The estimated specific BET surface areas of the Pd–Au/TiO2 fibers before and after the reduction process were 109 m2/g and 115.2 m2/g respectively with the corresponding BJH average pore diameters of 4.4 nm and 4.9 nm. To further verify the above analyses, X-ray photoelectron spectroscopy was used to evaluate the chemical bonding before

H.U. Shin et al. / Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

95

Fig. 5. (A) TEM image of Pd–Au/TiO2 fibers, (B) magnified view of TEM image, (C) particle size distribution, (D) HRTEM image of a single Pd–Au particle on a TiO2 fiber, and (E) SAED patterns of Pd–Au/TiO2 fibers.

and after reduction with the hydrazine monohydrate. The high resolution XPS spectra corresponding to Pd 3d and Au 4f binding energy regions (330–350 eV and 80–90 eV, respectively) are shown in Fig. 8. The data were fitted with asymmetric peak shapes after Shirley background correction, assuming the Pd was present in 3 different chemical environments. The Pd 3d spectra before reduction revealed that multiple oxidation states of palladium existed in the samples. The Pd 3d doublet peaks were fitted with the three components corresponding to metallic Pd and oxidized PdO and PdO2. The energies for the 3d5/2 peaks were 334.7, 335.9, and 337.8 eV respectively. Further, the data show that after calcination of the electrospun fibers, the PdO and PdO2 spectral peaks were dominant, while a smaller amount of metallic Pd was also present. These values are in good agreement with the reported binding energies [29,42,47,48]. After reduction using hydrazine monohydrate, the PdO and PdO2 presence was not detected and only palladium metallic peaks were detected at 334.5 and 340 eV, Fig. 8(A). The XPS peaks in the in the 80–90 eV range were attributed to Au 4f7/2 at 83.0 eV and 4f5/2 at 86.8 eV. As a result of the reduction

process, the Au 4f peaks shifted toward lower binding energies by 0.26 and 0.25 eV, respectively. In comparison with pure Au XPS peaks (84.0 and 88.0 eV) [49,50], the shift in binding energy toward the lower energies were more pronounced. The negative shift of about 1.00 eV of the Au 4f peaks from monometallic Au 4f peaks may be due to the addition of Pd to the catalyst particles [24]. Conceivably these shifts in binding energies could be due to the charge transfer from the Pd atoms to the Au atoms, which is indicative of alloy formation [31]. The noticeable negative shift in the Au 4f peak energies observed in our nanofibers could be due to many agglutinations between the Au and Pd atoms in the nanoparticles in the TiO2 fibers, which increases the opportunity for electron transfer from Pd to Au atoms [51,52]. To compare the catalytic performance, four different fiber media were tested in the NO + CO reaction. The reaction results are plotted in Figs. 9 and 10. In Fig. 9 the outlet stream species concentrations are plotted as a function of the reactor temperature. The inlet compositions for the reactions were essentially the same as the outlet concentrations at 100 8C. All of the data points represent the averages of at least three replications and the error bars are one standard

Fig. 6. XRD diffraction peaks of (upper curve) Au/TiO2, (center curve) Pd–Au/TiO2, and (lower curve) Pd/TiO2 after reduction with hydrazine monohydrate.

Fig. 7. (A) N2 adsorption–desorption isotherms (top left inset: BJH pore size distribution curves of Pd–Au/TiO2 nanofibers).

96

H.U. Shin et al. / Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

Fig. 8. XPS peaks of Pd 3d (A) and Au 4f (B) before reduction (lower plots) and after reduction (upper plots).

deviation of the averaged values. All of the media were the same in thickness and amounts of fibers, and all of the experiments were operated at the same flow rates, hence the relative reactivities of the catalysts corresponded to the reactor temperatures at which the reaction approached completion. The catalysts with higher activities had lower the reactor temperatures at which the CO and NO compositions approached zero. The Pd–Au/TiO2 fiber media showed the highest catalytic reaction activities (Fig. 9(A)) where the NO and CO concentrations approached zero at about 250 8C reactor temperature. NO–CO concentrations began decreasing at temperatures above 100 8C, indicating that the catalyst was active for the NO–CO reaction. As the reaction temperature increased, the NO was converted to nitrogen (N2) and nitrous oxide (N2O). N2O was formed as an intermediate product and underwent further reaction with CO to produce N2 and CO2, as indicated by the gradual decrease in N2O concentrations at higher temperatures. Cho et al. and Pisanu et al. reported that the formation of N2O was a significant intermediate step during the NO–CO reaction [53,54]. The Pd/TiO2 (Fig. 9(B)) and Au/TiO2 (Fig. 9(C)) fiber media showed similar trends, but the reactions were completed at higher reactor temperatures of about 350 8C, indicating lower reaction activities. Thus, the combined Pd–Au catalytic particles had improved performance over the Pd alone and Au alone catalysts. Fig. 9(D) shows the performance of the TiO2 fiber media without Pd or Au particles. The NO and CO concentrations approached zero at reactor temperatures above 450 8C showing low catalytic activity of the TiO2 by itself. As another way of presenting the reaction results, the percent conversions of NO and CO plotted as functions of reactor temperature in Fig. 10. The plots clearly show the Pd–Au/TiO2 media had conversions approaching 100% at lower temperatures than the Pd/TiO2, Au/TiO2 and TiO2 fiber media. The light-off temperatures of 50% NO conversion (T50) is another indication of

the relative reactivities of the media. The NO T50 values for Pd– Au/TiO2, Pd/TiO2, Au/TiO2 and TiO2 were 1898, 2058, 2588 and 393 8C respectively. Similarly, the CO T50 values were 2048, 2398, 2278 and 387 8C respectively. The turnover frequencies (TOF) of the NO and CO reaction activity are compared in Fig. 11. Initially, the Pd–Au/TiO2 fiber medium had the highest TOF values compared to Pd/TiO2 and Au/ TiO2 at 150 8C. At higher temperatures Au had a higher TOF. These data show the Pd–Au combined particles promoted the NO decomposition reaction kinetics by CO. The TOFs were calculated assuming bimetallic crystallinity of the Pd–Au particles, but as discussed previously, some of the metals may be present as monometallic crystallites or as bimetallic particles. Hence the TOF values are not reliable until more information is available on the crystalline states of the metals. The activation energies Ea were calculated from the Arrhenius plots (insets of Fig. 11) for the reaction of NO and CO. The Ea values were 31.5 kJ/mol, 24.2 kJ/mol, and 17.8 kJ/mol for Pd/TiO2, Au/ TiO2, and Pd–Au/TiO2 respectively. The reaction kinetic data shows that the activation energy decreased for the combined Pd and Au metals in the TiO2 fibers. This shows that the addition of Au is beneficial to the catalytic reaction of NO with CO and is consistent with data reported in literature for composition ratios Pd:Au = 4:1 [29,55]. In support of these observations our results show that the synergetic effect between Pd and Au leads to a higher activity catalyst, thus effectively promoting the decomposition of NO by CO. Future work The results reported here show a synergistic effect between Au and Pd on the titania support fibers. In some practical applications oxygen, water, or CO2 may be present in the inlet gas stream. The effects of these compounds on the performance should be

H.U. Shin et al. / Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

97

Fig. 9. Catalytic performance of (A) Pd–Au/TiO2, (B) Pd/TiO2, (C) Au/TiO2, and (D) TiO2 fiber media.

considered in future work. Stability of the catalyst is important for practical applications and should be explored. Future work should also compare performances of the imbedded catalyst particles via the electrospinning to the performance of supported catalysts formed using conventional methods such as wet coating. Au atoms in the vicinity of Pd atoms may prevent Pd aggregation and lead to higher stability of the Pd ensemble in Pd4Au1 catalysts as reported in literature [34]. The reverse may also be possible, that the Pd atoms may prevent aggregation of Au

atoms. The effect of aggregation should be explored in future work, particularly because the submicron fiber structure restricts the potential particle movement to aggregate to one-dimensional movement along the fiber axis in contrast to two-dimensional movements that are possible on larger structures. The aggregation mechanism is one of several that may affect the useful lifetime of the catalyst structure. Durability tests should be conducted and compared with other supported catalysts and structures as part of future work.

Fig. 10. Catalytic conversion of NO (%) and CO (%) for catalytic fiber media based on (A) Pd–Au/TiO2, (B) Pd/TiO2, (C) Au/TiO2, and (D) TiO2 fibers only (no added catalyst particles).

H.U. Shin et al. / Journal of Industrial and Engineering Chemistry 33 (2016) 91–98

98

Fig. 11. Turnover frequency (TOF) for NO (A) and CO (B) using Pd/TiO2, Au/TiO2, and Pd–Au/TiO2 (inset: Arrhenius plots for NO and CO).

Conclusions This study introduced a practical method for the preparation of fibrous catalytic media based on Pd–Au nanoparticles supported by TiO2 nanofibers. The Pd–Au/TiO2 fibers morphology study was confirmed by SEM, TEM and HRTEM. The crystal structures and chemical states of the Pd–Au/TiO2 fibers were evaluated by XRD and XPS. The reactivities of the Pd–Au combined fiber media were greater than Pd alone and Au alone fiber media. The enhanced reaction activities of Pd–Au/TiO2 fiber media corresponded with decreased activation energies in the Arrhenius expression. Acknowledgements This work was financially supported by the Coalescence Filtration Nanofibers Consortium: Donaldson, Ahlstrom, Parker Hannifin, Cummins Filtration, Bekaert, Hollingsworth and Vose, and SNS Nanofiber Technology to fabricate the fibers. The authors specially thank Sangyoup Hwang and Jeongwoo Lee for their technical assistance with XRD and GC analysis. References [1] X. Yao, F. Gao, Y. Cao, C. Tang, Y. Deng, L. Dong, Y. Chen, Phys. Chem. Chem. Phys. 15 (2013) 14945. J.N. Armor, Appl. Catal. B: Environ. 1 (1992) 221. D. Fino, N. Russo, G. Saracco, V. Specchia, J. Catal. 242 (2006) 38. B. Courtemanche, Y.A. Levendis, Fuel 77 (1998) 183. V.I. Paˆrvulescu, P. Grange, B. Delmon, Catal. Today 46 (1998) 233. R.M. Heck, R.J. Farrauto, S.T. Gulati, Automotive Catalyst, John Wiley & Sons, Inc., 2009p. 101. [7] J.N. Armor, Catal. Today 26 (1995) 99. [8] Y. Fu, Y. Tian, P. Lin, J. Catal. 132 (1991) 85. [9] J. Kasˇpar, P. Fornasiero, N. Hickey, Catal. Today 77 (2003) 419. [10] Z. Hu, C.Z. Wan, Y.K. Lui, J. Dettling, J.J. Steger, Catal. Today 30 (1996) 83. [11] A. Cybulski, J.A. Moulijn, Catal. Rev. 36 (1994) 179. [12] M. Iwamoto, H. Hamada, Catal. Today 10 (1991) 57. [13] N. Miyoshi, S.I. Matsumoto, K. Katoh, T. Tanaka, J. Harada, N. Takahashi, K. Yokota, M. Sugiura, K. Kasahara, Development of New Concept Three-way Catalyst for Automotive Lean-burn Engines. SAE Technical Paper, 1995. [14] E.C. Su, W.G. Rothschild, J. Catal. 99 (1986) 506. [15] S.J. Park, S. Bhargava, E.T. Bender, G.G. Chase, R.D. Ramsier, J. Mater. Res. 23 (2008) 1193. [16] S. Swaminathan, G.G. Chase, in: T. Lin (Ed.), Nanofibers-production, Properties and Functional Applications, Intech Open Access Publisher, 2011, (Chapter 3), http://www.intechopen.com/articles/show/title/ electrospinning-of-metal-doped-alumina-nanofibers-for-catalyst-applications. [2] [3] [4] [5] [6]

[17] S.J. Park, Y.C. Kang, R.D. Ramsier, K.U. Jeong, G.G. Chase, J. Nanosci. Nanotechnol. 10 (2010) 5225. [18] L. Shahreen, G.G. Chase, A.J. Turinske, S.A. Nelson, N. Stojilovic, Chem. Eng. J. 225 (2013) 340. [19] N. Macleod, J.M. Keel, R.M. Lambert, Appl. Catal. A: Gen. 261 (2004) 37. [20] Y.F. Han, J.H. Wang, D. Kumar, Z. Yan, D.W. Goodman, J. Catal. 232 (2005) 467. [21] B.E. Solsona, J.K. Edwards, P. Landon, A.F. Carley, A. Herzing, C.J. Kiely, G.J. Hutchings, Chem. Mater. 18 (2006) 2689. [22] J.K. Edwards, B.E. Solsona, P. Landon, A.F. Carley, A. Herzing, C.J. Kiely, G.J. Hutchings, J. Catal. 236 (2005) 69. [23] M.O. Nutt, K.N. Heck, P. Alvarez, M.S. Wong, Appl. Catal. B: Environ. 69 (2006) 115. [24] A.M. Venezia, V. La Parola, G. Deganello, B. Pawelec, J.L.G. Fierro, J. Catal. 215 (2003) 317. [25] X. Yang, L. Du, S. Liao, Y. Li, H. Song, Catal. Commun. 17 (2012) 29. [26] C.J. Baddeley, M. Tikhov, C. Hardacre, J.R. Lomas, R.M. Lambert, J. Phys. Chem. 100 (1996) 2189. [27] F. Gao, Y. Wang, D.W. Goodman, J. Am. Chem. Soc. 131 (2009) 5734. [28] C.W. Yi, K. Luo, T. Wei, D.W. Goodman, J. Phys. Chem. B 109 (2005) 18535. [29] F. Gao, D.W. Goodman, Chem. Soc. Rev. 41 (2012) 8009. [30] M. Chen, D. Kumar, C.-W. Yi, D.W. Goodman, Science 310 (2005) 291. [31] A. Samanta, T. Rajesh, R. Nandini Devi, J. Mater. Chem. A 2 (2014) 4398. [32] K. Qian, W. Huang, Catal. Today 164 (2011) 320. [33] J.K. Edwards, A.F. Carley, A.A. Herzing, C.J. Kiely, G.J. Hutchings, Faraday Discuss. 138 (2008) 225. [34] J. Xu, T. White, P. Li, C. He, J. Yu, W. Yuan, Y.-F. Han, J. Am. Chem. Soc. 132 (2010) 10398. [35] G.G. Chase, K. Nartetamrongsutt, H.U. Shin, Simple Device for Economically Producing Electrospun Fibers at Moderate Rates, US Patent 20,150,158,230 (2015). [36] H.U. Shin, A.B. Stefaniak, N. Stojilovic, G.G. Chase, Environ. Sci.: Nano 2 (2015) 251. [37] D.H. Reneker, A.L. Yarin, Polymer 49 (2008) 2387. [38] D.H. Reneker, A.L. Yarin, H. Fong, S. Koombhongse, J. Appl. Phys. 87 (2000) 4531. [39] H.U. Shin, Y. Li, A. Paynter, K. Nartetamrongsutt, G.G. Chase, Polymer 65 (2015) 26. [40] H.U. Shin, Y. Li, A. Paynter, K. Nartetamrongsutt, G.G. Chase, Data Brief 5 (2015) 41. [41] M.M. Demir, M.A. Gulgun, Y.Z. Menceloglu, B. Erman, S.S. Abramchuk, E.E. Makhaeva, A.R. Khokhlov, V.G. Matveeva, M.G. Sulman, Macromolecules 37 (2004) 1787. [42] M.-L. Wu, D.-H. Chen, T.-C. Huang, Langmuir 17 (2001) 3877. [43] S.Y. Hwang, C. Zhang, E. Yurchekfrodl, Z. Peng, J. Phys. Chem. C 118 (2014) 28739. [44] F. Li, Y. Guo, R. Li, F. Wu, Y. Liu, X. Sun, C. Li, W. Wang, J. Gao, J. Mater. Chem. A 1 (2013) 6579. [45] G. Zhan, J. Huang, M. Du, I. Abdul-Rauf, Y. Ma, Q. Li, Mater. Lett. 65 (2011) 2989. [46] P. Webb, C. Orr, Analytical Methods in Fine Particle Technology, Micromeritics Instrument Corp., Norcross, GA, 1997. [47] A. Babaei, S.P. Jiang, J. Li, J. Electrochem. Soc. 156 (2009) B1022. [48] E.H. Voogt, A.J.M. Mens, O.L.J. Gijzeman, J.W. Geus, Surf. Sci. 350 (1996) 21. [49] V. Jovic, W.-T. Chen, D. Sun-Waterhouse, M.G. Blackford, H. Idriss, G.I.N. Waterhouse, J. Catal. 305 (2013) 307. [50] P. Fuchs, K. Marti, S. Russi, Metrologia 49 (2012) 615. [51] Z. Li, F. Gao, Y. Wang, F. Calaza, L. Burkholder, W.T. Tysoe, Surf. Sci. 601 (2007) 1898. [52] S. Nishimura, Y. Yakita, M. Katayama, K. Higashimine, K. Ebitani, Catal. Sci. Technol. 3 (2013) 351. [53] B.K. Cho, J. Catal. 138 (1992) 255. [54] A.M. Pisanu, C.E. Gigola, Appl. Catal. B: Environ. 20 (1999) 179. [55] J. Zhang, H. Jin, M.B. Sullivan, F.C.H. Lim, P. Wu, Phys. Chem. Chem. Phys. 11 (2009) 1441.