Grantee Research Project Results
Final Report: Synthesis, Characterization, and Catalytic Studies of Transition Metal Carbide Nanoparticles as Environmental Nanocatalysts
EPA Grant Number: R829624Title: Synthesis, Characterization, and Catalytic Studies of Transition Metal Carbide Nanoparticles as Environmental Nanocatalysts
Investigators: Shah, S. Ismat , Chen, Jingguang G.
Institution: University of Delaware
EPA Project Officer: Hahn, Intaek
Project Period: March 1, 2002 through February 28, 2006 (Extended to February 28, 2007)
Project Amount: $350,000
RFA: Exploratory Research: Nanotechnology (2001) RFA Text | Recipients Lists
Research Category: Nanotechnology , Safer Chemicals
Objective:
The objective of this research project is to explore the possibility of using alternative catalytic materials, transition metal carbides, and oxycarbides (defined as oxygen-modified carbides) to replace platinum (Pt)-group metals for the reduction of NOx.
Transition metal carbides are known to be extremely hard materials with very high melting points. Among these materials, WC has been shown to exhibit catalytic behavior like Pt group metals and very unlike metallic W [1,2]. Several workers have also shown that the surface electronic properties and density of states for both Pt and WC have distinct similarities [3,4,5]. This is particularly important for automobile industry where Pt-group metals are extensively used for NOx reduction, usually referred as DeNOx. With the increasingly stringent regulation on the NOx emissions [6] and the scarcity of Pt, the need for an alternate material has become crucial. Though bulk WC has been shown to be catalytically active, the efficiency of NOx reduction is low. Nanostructured materials bring another dimension into the quest for cheaper and more stable catalyst for DeNOx. The generic advantage of using nanoparticles as catalyst stems from the large surface area that nanoparticles typically provide for surface reactions. Surface areas as high as 800 m2/gm are possible with some nanoparticles. The second advantage with nanoparticles is their ease of distribution in polymeric or ceramic matrices [7]. Though most of the catalytic studies have been done on bulk or thin films of WC, very few studies have involved nano WC particles. In this chapter, we present the results of our investigation on the synthesis and characterization of WC nanoparticles. We also present the results of the measurement of the DeNOx activity of WC nanoparticles.
Summary/Accomplishments (Outputs/Outcomes):
Advantages of nanoparticles
The advantage of using nanoparticles for catalysis is from an increased surface area. However, in the case of WC it extends beyond that. Most importantly, it has been shown that the surface activity of the catalyst depends on the orientation, and therefore the atomic density of the surface. For example, for NOx reduction, the surface activity and selectivity of Pt (111) surface is essentially zero whereas Pt (100) has very high activity as well as selectivity [8]. Similarly, Rh (100) surfaces show better catalytic properties for NOx reduction [9]. In fcc Pt and Rh, the (111) surface has the highest atomic packing density. The basis for the increase in the surface activity lies in the fact that more open-structured surfaces show higher surface activity. The (100) planes in Pt and Rh are less densely packed and, therefore, show higher activities.
Chen et. al. [10] have observed similar differences in catalytic activity for NOx reduction in carborized W surfaces. W is a bcc metal with (110) being the closest packet plane. In carborized W surfaces, the (110) shows lower activity as compared to the more open (111) surface. The more open surface provides more locations on the surface where a particular reaction can take place thereby increasing the catalytic efficiency. However, the shape of a particle is governed by the consideration of the minimum overall free energy. In nanoparticles, the surface free energy has a large contribution to the total free energy of the particles. There are two ways in which the surface energy can be reduced for a crystal of fixed mass / volume:
- By minimizing the surface area of the crystallite
- By ensuring that only surfaces of low surface free energy are exposed.
If matter is regarded as a continuum then the optimum shape for minimizing the surface free energy is a sphere since this has the lowest surface area/volume ratio of any 3D object. However, we have to consider the discrete, atomic nature of matter and the detailed atomic structure of surfaces when considering particles of the size found in catalysts. If, for example, we consider an fcc metal (e.g. Pt) and ensure that only the most stable (111)-type surfaces are exposed, then we end up with a crystal, which is an octahedron formed by the (111) planes. There are 8 different, but crystallographically equivalent surface planes that have the (111) surface structure - the {111} faces. They are related by the symmetry elements of the cubic fcc system. A compromise between exposing only the lowest energy surface planes and minimizing the surface area is obtained by truncating the vertices of the octahedron - this generates a cubo-octahedral particle as shown in Figure 1, with 8 (111)-type surfaces and 6 smaller, (100)-type surfaces, and gives a lower surface area / volume ratio. The atoms in the middle of the {111} faces show the expected coordination number of 12, characteristic of the (111) surface. Similarly, those atoms in the center of the {100} surfaces have the characteristic coordination number of 8 of the (100) surfaces. However, there are also many atoms at the corners and intersection of surface planes on the particle, which show lower coordination numbers. The atoms at the intersection of a {100} surface plane with two {111} surfaces show the lowest coordination number, 6. A lower coordination number leads to a higher catalytic activity, as more open area is available for the reaction to occur.
Considering the octahedral and cubo-octaherdral geometries, the change in energy due to the formation of cubo-octaherdral compared to a regular octahedral is:
(1)
Here, G is the Gibbs free energy of formation; E111 and E110 are energies of the two surfaces. d and p are distances as shown in Figure 1. The distance at which the octahedral cuts from the center can be obtained by minimizing energy with respect to p.
(2)
Figure 1: (a) (100)-(111) Cubo-Octahedron from a (111) octahedron. (b) The (100) and (111) surfaces of the cubo-octahedron.
The expression for p, has the dimensions of length and depending on the material system and reaction condition will range from a few Å to a few nm. On evaluating p from the above equation and substituting in the expression for surface area term, we get the increase in A100 as a function of the particle size. The area of the (100) and the (111) plane is given as:
(3)
The relative area of (100) surface, A100/(A100+A111) of as a function of “d” and “p” is shown in figure 2. The figure clearly indicates that for fcc structures, the more open (100) surface area increases rapidly when the particle size decreases, particularly below 70 nm. If we remove the constraint that a thermodynamic equilibrium has to be achieved and that a metastable structure is possible, then the change in the A100 can be evaluated solely on the basis of the particle geometry.
Figure 2: Area ratio R=A100/(A100+A111) as a function of d and p.
It can be concluded from the previous discussion that a small crystalline particle, in addition to providing a larger surface area, also offers a higher probability of exposing a more reactive surface.
Synthesis of WC nanoparticles Experiment
Laser Assisted Inert Gas Condensation (LAIGC) technique was used for making nanoparticles. Fig 3 shows the schematics of the LAIGC system. In the Inert Gas Condensation (IGC) [11] technique, resistive heating typically evaporates the material and the resulting flux is rapidly condensed by a carrier gas, which has high thermal conductivity. Rapid condensation of atomic vapors limits the possibility of phase separation and also restricts chemical diffusion. Other methods of evaporation might also be employed, depending on the material that needs to be evaporated, and includes sputtering and electron beam evaporation. A KrF Excimer Laser (Lamda Physik LPX 301, wavelength 248 nm) is used to ablate the target. The stainless steel ablation chamber was pumped down to a base pressure of 1x10-6 Torr. Nanoparticles were synthesized at a pressure of 500 mTorr of He. The laser was used in a constant energy mode (1.0J) with a repetition rate of 30Hz. Commercially available WC and pressed W (2 inch) targets were used. These targets were mounted on a rotating target holder to avoid pitting of the targets.
Figure 3: Schematic diagram of the IGC system. (1. Evaporation boat, 2. Laser ablation target, 3. Stainless steel filter, 4. Hopper for the collection of particles, 5. Wire feeding unit, 6. Laser source, 7. Power supply, 8. Inert gas cylinder, 9. Turbo pump, 10. Roots blower, 11. Mechanical Pump, 12. Gas circulation line).
Two types of samples were prepared; one was pure WC nanoparticles in which case a pure WC target was used for ablation. The second type sample was prepared from pure W nanoparticles, which were subsequently carburized in a CH4 atmosphere at 1000 °C for 1 hour at ambient pressure. These particles will be referred to as carburized W sample in this chapter.
Density of States comparison
The valence band was studied to probe the Density of States (DOS) at the Fermi level for WC, WO3 and Pt (Figure 4). The information about the DOS at the Fermi level was obtained by performing a high resolution XPS scan in the lower binding energy region (8 to -6 eV). The similarities between WC and Pt DOS at the Fermi level are evident from Fig. 4, while WO3 shows a different profile and is also catalytically inactive. The calculated total density of states for Pt and WC and the projected DOS of different bands are shown in figure 5. A full potential linearized augmented plane wave (FPLAPW) method within the framework of density functional theory (DFT) was used [12]. The obtained results from the calculations include a fully optimized ground state structure calculation. Details of this method and code can be found elsewhere [13].
Figure 4: High resolution XPS spectra of the valence band region for Pt, WC and WO3
Figure 5: Calculated DOS for Pt and WC.
XRD Results
The XRD results for both pure WC and carburized W are shown in Fig 6. All peaks could be indexed with the diffraction lines of WC in the powder diffraction diagram obtained from the International Center for Diffraction Data [14,15]. However, the XRD of the carburized sample shows W (110) and (200) reflections in addition to the WC reflections indicating that the samples were not completely carburized. Based on the intensity of W (110) reflection, the extent of carburization is found to be approximately 48%. The particle size was determined from the XRD data using the Scherer’s formula
(4)
Where λ is the x-ray wavelength, β is the FWHM in radians and θ is the Bragg angle. The particle size of the as-prepared WC was found to be around 60nm. However, the particle size of the carburized sample was greater than 100nm, perhaps due to sintering during carburization.
Figure 6: XRD patterns for pure WC and Carburized W samples.
DeNOx Results
A mixture of He with 1% NO was used for DeNOx experiments. The exhaust gas concentration was analyzed by a Hewlett-Packard model 5890 series II gas chromatograph equipped with a thermal conductivity detector (TCD). A 1/8” custom-made packed column (Hayesep D 100/120, part number 16166, Alltech Associates, Inc., IL) was used to allow moderate separation of NO, N2, and other intermediate compounds. Helium was chosen as the carrier gas at flow rate of 30 sccm and injection head pressure of 54 psi. Injection temperature was kept at 120 °C. Oven temperature was set at 35 °C. This GC system was modified with a 6-port inline valve system for the real time measurement (Model A46UWE, Valco Instruments). The volume of the sampling loop was 200 μL and additional helium source was used for the waste venting.
Both as-prepared and carburized WC particles were tested for catalytic activity for NOx reduction. Equal weight (2 gms) of the material was loaded in the reactor. The reactor had three stages with different temperature controllers for each of the stage. The temperature of the reactor ranged from room temperature to 1000°C. In order to check the stability and the reproducibility of the catalyst 5 consecutive DeNOx cycles were performed with both the samples. In each of the cycle, the temperature was ramped from 100 to 800°C with a ramp rate of about 20°C per minute and the composition of the exhaust gas from the reactor was sampled into the GC at systematic temperature intervals of 50°C. To avoid statistical fluctuation, the gas was sampled several times at a given temperature. All the plots are normalized with respect to the maximum area of NO peak obtained without the catalyst and at room temperature. The average standard error in the measurement was around 0.05. Fig 7 shows the DeNOx results of pure WC and carburized W samples. For the pure WC, it can be seen that the onset of the DeNOx activity is at temperature around 350°C and it reaches a conversion efficiency of over 90% at temperatures above 600°C. Also, the results for 5 cycles are repeatable. For the carburized W a slightly higher onset of activity temperature (400°C) was measured compared to that of pure WC.
Figure 7: Normalized conversion of NO on pure WC and carburized W
It can also be noticed that the activity through the first cycle is different than that for the subsequent cycles. In the first cycle the onset of activity was around 600°C. This is most probably due to presence of metallic W left over from incomplete carburization. However the subsequent cycle the results are repeatable and also the conversion efficiency reaches over 90% at temperatures similar to that of pure WC nanoparticles.
Structural stability of the catalyst
In order to investigate the structural stability of the samples during the DeNOx process, XRD and XPS analyses were carried out on samples before and after the DeNOx cycles. For XRD structural investigations, the samples were held in a He + NO atmosphere at three different temperatures for 2 hours. The gas flow rate was fixed at 50sccm for all the samples. Figure 8 shows the XRD results of the structural evolution of pure WC and carburized W samples, respectively. In both cases, the formation of oxide can be easily noticed. The intensity of the oxide peaks increase with the temperature, suggesting continued oxidation during the DeNOx process. Based on the intensity of WO3 (100) peak, over 85% carburized W gets oxidized. However, in the pure WC the extent of oxidation is only around 22%.
Figure 8: XRD of carburized W (top) and pure WC (bottom) samples as a function of the annealing temperature under He-NO atmosphere
High resolution XPS analyses were carried out on the pure WC samples both in as-prepared condition and also right after 5 cycles of DeNOx. Figure 9 shows the high-resolution XPS spectra of W 4f photoelectron peak. For the pure sample, the spectrum was fitted with 4 peaks. The first two peaks correspond to W 4 f7/2 and 4 f5/2 at 31.5eV and 33.6eV respectively (the peak separation between 4 f7/2 and 4 f5/2, referred to as Δ4 f, is equal to 2.1eV). The other two are from the peak shifts due to WO3 oxide formation: 4 f7/2 and 4 f5/2 peaks are at about 35.2eV and 37.4eV (Δ4 f = 2.2eV), respectively. These peak positions are close to the reported values for WO3 [16]. All the peaks are charge corrected with respect to carbon 1s peak position (284.5 eV). The pure WC sample also shows small amount of oxide, which could be from surface contamination due to atmospheric exposure.
Figure 9: High Resolution XPS spectra of W4f region of pure WC samples before (top) and after 5 cycles of DeNOx (bottom)
The XRD of the pure WC did not show any WO3 peak, which confirms that the pure-WC nanoparticles do not have oxide phase in the bulk of the material and the XPS oxide shift is only due to surface contamination. The XPS spectrum of the sample after 5 cycles of DeNOx (Fig 9) was also fitted with four peaks following the same procedure as described earlier. Here, the intensity of the oxide peak is much higher. Both the XPS and XRD confirm oxidation of the catalyst however the oxidation is not complete and there is still some of the pure WC phase left which explains why the sample continues to be active for DeNOx. We have also confirmed that WO3 is not catalytically active for DeNOx within the temperature range employed in the experiments. These observations indicate that WC is unstable for DeNOx reaction. Although the exact reaction path still needs further investigated, it is possible that the available oxygen after the DeNOx reaction reacts with the WC to form carbon dioxide leaving behind WO3. In a real automobile exhaust there is also C containing species, e.g., CO, hydrocarbons, etc., present that can serve to replenish the C in WC and keep it from complete oxidation.
Effect of CO Addition
DeNOx experiments were also performed in the presence of a mixture of gases that resembles automobile exhaust, except for the addition of He, which serves as the carrier gas. The exact composition of the mixture was 1%CO, 0.9%O2, 0.01%C2H6 and He. The gas flow rate was fixed at 50 sccm of NO mixture (1% NO rest He) and 25 sccm of the exhaust mixture. Figure 10 shows the DeNOx results with the addition of C containing gasses. It can be seen that the onset of the activity remains similar to the non-C case, around 400°C.
Figure 10: Normalized conversion of NO for WC with CO mix
To compare with the earlier stability studies, the catalyst was held at 800 °C for 2 hours and subsequently analyzed by XRD. Figure 11 shows the XRD patterns from the pure WC sample, before and after the DeNOx process. The sample shows a significant reduction in the extent of oxidation, suggesting that CO mixture greatly regenerates the catalyst. A decrease in the oxide peak intensity by about 85% was observed at 800 °C. Even though we only have data for one combination of CO in the gas mixture with NO, it is conceivable that an appropriate ratio of the two gases could lead to more promising results.
Figure 11: XRD patterns showing the structural evolution of WC in the presence CO mixture
Summary
Through this research we have identified a unique method to synthesize WC nanoparticles. Both pure WC and carburized W were synthesized with Laser Assisted Inert Gas Condensation technique using a pure WC and W targets respectively. XRD confirms the formation of WC in the case of pure WC and shows 48% carburization in the carburized W sample. The prepared samples show DeNOx activity at temperature around 400 °C and the activity reaches a conversion efficiency of more than 90% at temperatures greater than 600 °C. Both the samples show oxidation and the extent of the oxidation increases with temperature. DeNOx experiments were performed in the presence of CO and hydrocarbon containing gases with one fixed ratio. The activity remains unchanged but the extent of oxidation decreases significantly.
Recommendations for further research:
In case of WC nanoparticles, the first aspect to look is to study the effect of particle size in the catalytic activity. WC nanoparticles are extremely difficult to synthesize using conventional PVD processes. However, there are some CVD processes, which seem promising. Also, the catalytic properties during De-NOx with a lean mixture (higher CO containing gas) should be investigated. The exact reaction pathway during catalysis needs further clarification, and in these lines, temperature program desorption studies are crucial. Our initial work has shown that the catalyst is efficient, but the activation temperature is a bit higher than conventional material used. This can only be addressed by gaining a better understanding on the exact reaction pathway. It is possible that particle size might play a crucial role in addressing this issue. Recently, there has been an interest in WC thin films as a replacement for Pt electrodes in fuel cell application. The electro-catalytic application of WC may also be of great interest in this regard.
Students Supported Through EPA Grant Number R829624:
- Abdul Rumaiz PhD 2008
- Minghue Zhang PhD 2006
- J.R. McCormic PhD 2006
- I. Balditchev Undergraduate student
References:
- R. B. Levy, and M. Boudart, Science 81, 547 (1973).
- L.H. Bennet, J.R. Cuthill, A.J. McAlister, N.E.Erickson and R.E. Watson, Science, 181,547 (1974).
- P.M. Stefan, M. L. Shek, I. Lindau, W. E. Spicer, L. I. Johansson, F. Herman, R. V. Kasowski, and G. Brogen, Phys. Rev. B. 29, 5423, (1984).
- L.F. Mattheiss and D.R. Hamman, Phys. Rev. B. 30, 1731 (1984).
- G.C Koltssakis and A.M. Stamatelos, Prog. Energy Combust Sci. 23 (1997).
- A. A. Pradhan, S. Ismat Shah, Lisa Pakstis, Mat. Res. Soc. Symp. Proc. Vol. 740, 16.10.1 (2003).
- T. W. Root and L. D. Schmidt, G.B.Fisher, Surf. Sci. 134, 30 (1983).
- D.R.McIntyre, G.T.Burstein, A.Vossen, J. Power Sources 107, 67 (2002).
- M. J. P. Hopstaken and J. W. Niemantsverdriet, Phys. Chem. B, 104, 3058 (2000).
- M.Zhang, H.H.Hwu, M.T. Buelow, J.G.Chen, T.H.Ballinger, P.J.Anderson, Catal. Letters 77, 29 (2001).
- H. Gleiter, J. Weissmuller, et. al., Acta Mater., 49, 737 (2001).
- P. Blaha, K. Schwarz, G. K. H. Madsen, D. Kvasnicka, and J. Luitz, WIEN2K, an augmented plane wave plus local orbitals program for calculating crystal properties, Vienna University of Technology, Vienna, 2001.
- R. M. Dreizler and E. Gross, Density Functional Theory, An approach to Quantum Many Body Problem, Springer-Verlag, Berlin (1990).
- JCPDS WC Card 25-1047 International Center for Diffraction Data.
- JCPDS WO3 Card 88-0545 International Center for Diffraction Data.
- R.J. Colton and J.W. Rabalais Inorganic Chemistry. 15, 236 (1976).
Journal Articles on this Report : 2 Displayed | Download in RIS Format
Other project views: | All 16 publications | 3 publications in selected types | All 3 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Rumaiz AK, Lin HY, Baldytchev I, Shah SI. Nanosized tungsten carbide for NOx reduction. Journal of Vacuum Science & Technology B 2007;25(3):893-898. |
R829624 (Final) |
Exit |
|
Zhang MH, Hwu HH, Buelow MT, Chen JG, Ballinger TH, Andersen PJ, Mullins DR. Decomposition pathways of NO on carbide and oxycarbide-modified W(1 1 1) surfaces. Surface Science 2003;522(1-3):112-124. |
R829624 (2003) R829624 (Final) |
Exit |
Supplemental Keywords:
Tungsten carbide, WC, environmental catalysts, reactive sputtering, reactive gas condensation, RGC, mobile sources, nitrogen oxides, acid rain, environmental chemistry, extended x-ray adsorption fine structure, EXAFS, nanoparticles, decomposition, gas chromatography, GC, catalyst, bulk and theoretical spectra,, RFA, Scientific Discipline, Air, Waste, Sustainable Industry/Business, Sustainable Environment, air toxics, Environmental Chemistry, Remediation, Technology for Sustainable Environment, Civil/Environmental Engineering, Biochemistry, New/Innovative technologies, Chemistry and Materials Science, Environmental Engineering, Nitrogen dioxide, catalyst composition, waste reduction, detoxification, industrial wastewater, in situ remediation, air pollutants, automotive emissions, membranes, remediation technologies, nanotechnology, environmental sustainability, catalytic studies, nanocatalysts, reductive degradation of hazardous organics, catalysts, Nitric oxide, environmentally applicable nanoparticles, nanoparticles, sustainability, reductive dechlorination, hazardous organics, bimetallic particles, ultrafiltration, membrane-based nanostructured metals, pollution prevention, innovative technologies, membrane technology, reductive detoxification, recycle, transition metal carbidesProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.