Environ. Sci. Technol. 2011, 45, 726–731
Catalytic Decomposition of Toxic Chemicals over Metal-Promoted Carbon Nanotubes L I L I L I , †,§ C H A N G X I U H A N , ‡ XINYU HAN,† YIXIAO ZHOU,† LI YANG,† B A O G U I Z H A N G , * ,† A N D J I A N L I H U | College of Environmental Science and Engineering, Nankai University, Tianjin 300071, China, Institute of New Catalytic Materials Science, College of Chemistry, Nankai University, Tianjin 300071, China, Department of Life Science, Zhoukou Normal University, Zhoukou 466000, China, and Black & Veatch Corporation, Overland Park, Kansas 66221, United States
Received July 16, 2010. Revised manuscript received November 22, 2010. Accepted November 22, 2010.
Effective decomposition of toxic gaseous compounds is important for pollution control at many chemical manufacturing plants. This study explores catalytic decomposition of phosphine (PH3) using novel metal-promoted carbon nanotubes (CNTs). The cerium-promoted Co/CNTs catalysts (CoCe/ CNTs) are synthesized by means of coimpregnation method and reduced by three different methods (H2, KBH4, NaH2PO2 · H2O/ KBH4). The morphology, structure, and composition of the catalysts are characterized using a number of analytical instrumentations including high-resolution transmission electron microscopy, X-ray diffraction, X-ray photoelectron spectroscopy, BET surface area measurement, and inductively coupled plasma. The activity of the catalysts in PH3 decomposition reaction is measured and correlated with their surface and structural properties. The characterization results show that the CoCe/CNTs catalyst reduced by H2 possesses small particles and is shown thermally stable in PH3 decomposition reaction. The activities of these catalysts are compared and are shown in the following sequence: CoCe/CNTs > Co/CNTs > CoCeBP/ CNTs> CoCeB/CNTs. The difference in reduction method results in the formation of different active phases during the PH3 decomposition reaction. After a catalytic activity test, only the CoP phase is formed on CoCe/CNTs and Co/CNTs catalysts, whereas multiphases CoP, Co2P, and Co phases are formed on CoCeBP/CNTs and CoCeB/CNTs. Results show that the CoP phase is formed predominantly on the CoCe/ CNTs and Co/CNTs catalysts and is found to likely be the most active phase for this reaction. Furthermore, the CoCe/CNTs catalyst exhibits not only highest activity but also long-term stability in PH3 decomposition reaction. When operated in a fixedbed reactor at 360 °C, single-pass PH3 conversion of about 99.8% can be achieved.
Introduction The abatement of toxic chemicals, such as PH3, presented in chemical plant exhaust gas has become an important research topic. PH3 is one of the highly toxic and flammable gases that can present significant safety hazards even at very low concentration. PH3 is generated from industrial processes such as pesticide, herbicides, and chemical agents production. The stringent environment regulation will soon require that PH3 in the exhaust gas from chemical plants to be reduced to the lowest possible level (1). One of the approaches to reduce PH3 is to catalytically decompose it to elemental phosphorus (P) and hydrogen gas (H2). It is well-known that both P and H2 are high value chemicals; therefore, the PH3 decomposition process possesses not only environmental benefits but also economic advantages. Due to the similarity of decomposition behavior, the method can be applied to decomposition of other toxic gaseous compounds such as arsine (AsH3) which is generated from coal gasification (2). It is well-known that catalysts plays an important role in pollution control; therefore, it is necessary to develop an appropriate catalyst system for complete PH3 decomposition under temperature lower than the thermal decomposition temperature of 600 °C. Rare earth metal oxides have been employed as promoters for many catalyst systems. The addition of small quantities of rare earth elements to a catalyst can alter particle size and/or the electronic property of the catalyst. The presence of a rare earth metal oxide can also influence the adsorption characteristics and improve reducibility and stability of the catalyst (3-5). However, the rareearth metal oxides can also cause a negative effect on catalytic performance (6). It is important to understand the interactions between promoters, metals, and support. The methods of making and activating a catalyst become very important in developing rare earth metal promoted catalyst systems (7). For a supported catalyst system, the support material is not merely a carrier. It may contribute to the activity of the catalyst as well. In last 10 years, CNTs have received considerable attention as the support materials for heterogeneous catalysts and shown exceptional catalytic properties (8-10). Compared to traditional catalyst support materials such as Al2O3, TiO2, SiO2, V2O5, and various zeolites, CNTs possess many extraordinary characteristics, such as high thermal conductivity, and resistance to acid and base corrosion damage (11). In addition, upon appropriate treatment, active sites consisting of highly dispersed nanoparticles can be created on CNTs (12-14). Previous studies have shown that FeP binary alloy (15), FeCuP ternary alloy (16), and amorphous CoP/TiO2 alloy exhibited greater than 95% PH3 conversion at 450 °C (17). In our recent study, CoNiBP/CNTs catalyst was synthesized by induced electroless-plating method, and about 99.7% PH3 conversion can be achieved over this catalyst at 380 °C (18). The objectives of this study are to investigate the effect of cerium (Ce) addition to Co/CNTs catalyst on PH3 decomposition and to ascertain the influence of reduction methods on the activity of Ce-promoted Co/CNTs catalyst.
Materials and Methods * Corresponding author phone: +86 2223503592; fax: +862223508807; e-mail:
[email protected]. † College of Environmental Science and Engineering, Nankai University. § Zhoukou Normal University. ‡ College of Chemistry, Nankai University. | Black & Veatch Corporation. 726
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011
Catalyst Preparation. In order to create more defects and holes, increase BET surface area, and develop hydrophilic surface groups on the walls of the CNTs (19), the CNTs were pretreated before the active components were impregnated. The details about the pretreatment procedure were described in our previous publication (18) and the Supporting Infor10.1021/es1022416
2011 American Chemical Society
Published on Web 12/08/2010
FIGURE 1. TEM images of (a) CoCe/CNTs and (b) CoCeB/CNTs. mation. A series of Ce-promoted Co/CNTs catalysts were prepared by the coimpregnation method. The pretreated CNTs (BET surface area 257 m2/g) were immersed in 50 wt % ethanol solution containing desired amounts of cobalt chloride (CoCl2 · 6H2O), citric acid(C6H8O7, 50 wt % of CNTs), and ceric sulfate (Ce(SO4)2 · 4H2O) to attain a Co loading of 15 wt % and a Co:Ce weight ratio of 10:1. The above precursors were reduced by three different methods: (1) in H2 at 400 °C for 6 h, (2) in potassium borohydride solution (KBH4, 2.0 mol/L) at 20 °C for 2 h, and (3) in sodium hypophosphite (NaH2PO2 · H2O, 0.4 mol/L) and potassium borohydride (KBH4, 2.0 mol/L) at 50 °C for 2 h. The as-prepared catalysts reduced by three different methods were passivated in 1 mol % O2/N2 at 60 mL/min for 0.5 h and designated as CoCe/ CNTs, CoCeB/CNTs, and CoCeBP/CNTs, respectively. In the preparation of CoCeBP/CNTs catalyst, NaH2PO2 · H2O was premixed with CoCl2 · 6H2O (the molar ratio of NaH2PO2 · H2O: CoCl2 · 6H2O was 4:1), Ce(SO4)2 · 4H2O, and C6H8O7 to form a homogeneous solution. Subsequently KBH4 solution (2.0 mol/L, the molar ratio of KBH4:CoCl2 · 6H2O was 2:1) was added dropwise to the solution. The resulting black precipitate was washed to remove Cl-, K+, SO42-, Na+ ions and other soluble impurities with deionized water and ethanol until the pH was 7. CoCeB/CNTs material was prepared by a similar method to that for CoCeBP/CNTs in the absence of NaH2PO2 · H2O. The molar ratio of B/Co was 5/1 to ensure the complete reduction of Co2+ ions. In order to better understand the effect of Ce addition to Co/CNTs catalyst on the performance of PH3 decomposition, a baseline Co/CNTs catalyst was synthesized by following the same procedure of CoCe/CNTs. Catalyst Characterization. The elemental analysis of the catalysts was carried out by inductively coupled plasma (ICP-AES, 9000(N+M)). The crystalline and amorphous structures were measured by X-ray powder diffraction (Rigaku-D/MAX 2500). The Joint Committee on Powder Diffraction Standards (JCPDS) database was used in the identification of the crystalline structure from the diffraction pattern. The particle size was calculated using the Scherrer Equation. The measurements of BET surface area, pore volume, and average pore diameter of the catalysts were conducted under N2 adsorption at 77 K using a Quantachrome Autosorb-1 instruments. The particle size and surface morphology of the catalysts were determined using a high-resolution transmission electron microscope (JEM-2100). The surface composition of the samples was analyzed by a Kratos Axis Ultra DLD multitechnique X-ray photoelectron spectroscopy.
Activity Testing. The PH3 decomposition reaction was carried out in a continuous flow fixed-bed reactor. In each test, 300 mg (60-80 mesh) of catalyst was placed at the center of the reactor. Prior to the reaction, the catalyst was reduced in continuous flow H2 at 400 °C for 2 h. The reaction temperature was measured with an interior placed thermocouple in direct contact with the catalyst bed. The feedstock consisted of 5.0 mol % PH3 in N2 and was measured by glass tube rotameter. Catalytic activity tests were conducted under the conditions of 300-460 °C, ambient pressure and PH3 gas hourly space velocity of 2520 mL/h.gcat. The flow rate at the reactor inlet was 252 mL/min. Reactor off gas was condensed in cold trap at 0 °C wherein gaseous P was condensed to form P4 solid. Noncondensable gaseous products (PH3 and H2) were analyzed by an online gas chromatograph (Shimadzu) equipped with a flame photometric detector (FPD) and Porapak Q column. After activity testing, the catalysts were cooled to room temperature in flowing nitrogen and then passivated in a flow of 1.0 mol % O2/N2 at 60 mL/min for 0.5 h. In addition, the off gases were absorbed by 0.6-0.8 wt % NaClO aqueous solution to remove unreacted PH3.
Results and Discussion Catalyst Characterization. The main technique used for direct morphology characterization is the transmission electron microscope (TEM). The images of the prepared Co/ CNTs, CoCe/CNTs, CoCeB/CNTs, and CoCeBP/CNTs catalysts are shown in Figure 1 and Figure S1. The TEM study reveals that catalyst particles are uniformly dispersed on the external surface of CNTs. On each TEM image, the spherical particles consist of either Co or the alloy form of Co-Ce, Co-Ce-B, Co-Ce-B-P, respectively. Based on measurements with a sample size of 100 particles, the averaged sizes of the spherical particles over Co/CNTs, CoCeBP/CNTs, and CoCeB/CNTs catalysts are about 11 nm, 10.1 nm, and 13.6 nm, respectively (Figure S1a, S1b and Figure 1b), whereas the particles over CoCe/CNTs are relatively smaller having the diameter of about 8.2 nm (Figure 1a). Figure 2a-d depicts X-ray powder diffraction (XRD) patterns of the prepared catalysts. For Co/CNTs and CoCe/ CNTs catalysts, the diffraction peaks at around 2θ ) 41.7°, 44.8°, 47.6°, 62.8°, and 76° are ascribed to the hexagonal cell of the Co phase (PDF 00-005-0727). No clear diffraction signals attributable to Ce particles on CoCe/CNTs catalyst are observed, possibly due to the low content of the Ce promoter or the formation of CoCe particles where the Ce is dispersed in the Co particles. When Ce is added to Co/CNTs, the intensity of Co diffraction peak decreases, suggesting that VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
727
FIGURE 2. XRD patterns of (a) prepared Co/CNTs, (b) prepared CoCe/CNTs, (c) prepared CoCeB/CNTs, (d) prepared CoCeBP/ CNTs, (e) prepared CoCeB/CNTs treated at 500 °C in H2 flow for 2 h, and (f) prepared CoCeBP/CNTs treated at 500 °C in H2 flow for 2 h. Co particle size decreases in the presence of cerium. This phenomena is confirmed by the TEM analysis as discussed in the previous section. These results demonstrate that the addition of cerium is beneficial for reducing Co particle size. For CoCeB/CNTs and CoCeBP/CNTs catalysts, only broadening peaks are observed at around 2θ ) 45°, indicating that the alloy particles over these two catalysts present as amorphous structure (20-22). The literatures has reported that the amorphous structure is thermodynamically metastable, and crystallization can occur spontaneously during the reaction (23, 24). Therefore, it is necessary to examine the crystallization behaviors of the prepared CoCeB/CNTs and CoCeBP/CNTs catalysts at high temperature. Figure 2e-f shows XRD patterns of CoCeB/CNTs and CoCeBP/CNTs catalysts treated at 500 °C in continuous H2 flow for 2 h. On calcination of the CoCeB/CNTs catalyst, only cubic Co crystalline (PDF 00-015-0806) is observed, whereas calcined CoCeBP/CNTs catalyst exhibits XRD patterns associated with cubic Co and Co2P (PDF 00-006-0595). The appearance of large metallic Co particles in the calcined CoCeB/CNTs and CoCeBP/CNTs catalysts are due to the transformation from amorphous structure to the crystalline structure under high temperature, causing agglomerization (25). Based on XRD analysis, there is no evidence for the formation of crystalline B-containing phases and Ce-containing phases in the calcined CoCeB/CNTs and CoCeBP/CNTs catalysts. High dispersion of B in the crystallization of amorphous alloy has been reported in previous work (26, 27). Our previous study also revealed that P atoms migrate into the alloy particles to form crystalline metal phosphide which acts as dominant active phase in the PH3 decomposition reaction (18). After catalytic activity testing at 300-460 °C, spent catalysts are analyzed by XRD and results are shown in Figure 3. The presence of the CoP phase (PDF 00-0290497) is observed in all four spent catalysts. Diffraction peaks at 2θ ) 40.7° and 43.3° are ascribed to the Co2P phase (PDF 00-032-0306) which is present in spent CoCeB/CNTs and CoCeBP/CNTs catalysts. In addition, a diffraction peak corresponding to the cubic Co crystalline plane at 2θ ) 44.2° is present in the spent CoCeB/CNTs catalyst. The XRD patterns show that the reduction method employed in the preparation of catalysts determines the formation of active phases. When Co/CNTs and CoCe/CNTs catalysts are reduced by H2, only the CoP phase is formed. However, when reduced by KBH4, or KBH4/NaH2PO2 · H2O, besides the CoP phase, the Co2P phase is formed on spent CoCeBP/CNTs catalyst. Three different phases (CoP, Co2P, and cubic Co phases) are present on the spent CoCeB/CNTs. The particle size of CoP phase was calculated based on the diffraction 728
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011
FIGURE 3. XRD patterns of spent catalysts (a) Co/CNTs, (b) CoCe/CNTs, (c) CoCeB/CNTs, and (d) CoCeBP/CNTs. peak at 2θ ) 48.1°. The calculated CoP particle size for spent CoCe/CNTs catalyst is 12.4 nm, which is smaller than that of 16.7 nm for spent Co/CNTs. This implies that the addition of Ce can improve the thermal stability of supported metal catalysts, preventing them from sintering. On the contrary, CoP particles on spent CoCeB/CNTs and CoCeBP/CNTs catalysts are much larger, being 32.8 and 24 nm, respectively. This is an indication of poor thermal stability during the reaction. The elemental analyses of prepared and spent catalysts are listed in Table 1. In all four catalysts, the concentration of Co varies from 14.5-15.2 wt %, and the ratio of Ce/Co varies from 0.094 to 0.101. Compared to the prepared catalysts, the spent catalysts have significant P content. Table 1 shows that the molar ratio of P/Co in the spent Co/CNTs, CoCe/CNTs, CoCeB/CNTs, and CoCeBP/CNTs catalysts are 1.6, 1.5, 2.1, and 1.6, respectively. It can be concluded that the catalysts contain P in excess of that expected from the stoichiometry of CoP and Co2P. The molar ratio of P/Co in spent CoCeB/CNTs and CoCeBP/CNTs catalysts are 2.1 and 1.6. The excess P is not incorporated into the catalysts to form CoP. Instead, CoP and Co2P phases are formed on the CoCeBP/CNTs catalyst. In addition to CoP and Co2P phases, the cubic Co phase is observed on CoCeB/CNTs. Based on these results, it can be inferred that the coexistence of Ce and B species can inhibit the formation of the CoP phase during the PH3 decomposition reaction. The physical properties of spent catalysts are measured, and the results are summarized in Table 2. The loss of both surface area and pore volume with regard to the prepared catalysts is observed. The decrease in surface area varies between 23 and 50%. The extent of decrease in surface areas is in the following order: CoCeB/CNTs > CoCeBP/CNTs > Co/CNTs > CoCe/CNTs. These results have revealed the pore blockage due to the formation of phosphide and the presence of small amount of P. The pore volumes of the spent catalysts also decrease. The extent of decrease in pore volume is in the following order: CoCeB/CNTs > CoCeBP/CNTs > Co/ CNTs > CoCe/CNTs. It appears that the CoCeB/CNTs catalyst prepared by KBH4 reduction exhibits significant loss in surface area and pore volume during the reaction. It is postulated that the crystallization of amorphous alloys occurs at high temperature, forming large Co crystallites and eventually forming large phosphide particles during the reaction. Compared to the CoCeB/CNTs catalyst, the spent CoCeBP/ CNTs catalyst shows less decrease in surface area and pore volume. It seems that, for Co-Ce type of catalyst system, reduction by using NaH2PO2 · H2O/KBH4 can suppress the agglomerization of the amorphous particles on the support. The spent CoCe/CNTs catalyst has larger surface area and
TABLE 1. Composition of Ce-Promoted Co/CNTs Catalysts Prepared by Different Reduction Methods composition (at %) of prepared catalyst catalyst Co/CNTs CoCe/CNTs CoCeB/CNTs CoCeBP/CNTs
Co loading (wt %) of Ce:Co (weight ratio) prepared catalyst 15.1 14.7 15.2 14.5
0.0964 0.0942 0.1005
composition (at %) of spent catalyst
Co
P
Ce
B
Co
P
Ce
B
100 96.1 63.1 61.5
11.3
3.9 2.5 2.6
34.4 24.6
38.5 39.2 27.3 33.0
61.5 59.2 56.7 52.3
1.6 1.1 1.4
14.9 13.3
TABLE 2. Physical Properties of Ce-Promoted Co/CNTs Catalysts Prepared by Different Reduction Methods BET surface area (m2/g)
pore volume (cm3/g)
catalyst
prepared
spent
prepared
spent
prepared
spent
prepareda
spentb
Co/CNTs CoCe/CNTs CoCeB/CNTs CoCeBP/CNTs
129 126 113 120
86 97 57 72
0.78 0.74 0.67 0.71
0.59 0.63 0.42 0.54
21.1 27.6 18.1 23.2
23.3 34.9 31.1 28.2
11 8.2 13.6 10.1
16.7 12.4 32.8 24
average pore diameter (nm)
particle size (nm)
a Particle size of prepared catalysts observed by TEM. b CoP crystallite size derived from catalytic reaction (tested at 300-460 °C) calculated from XRD patterns using the Scherrer equation.
larger pore volume compared to the other three spent catalysts, which is attributed to the formation of small CoP particles during the reaction. The addition of Ce can suppress metal crystallites from sintering and subsequently increase the dispersion of supported phosphide particles. It is believed that large surface area and pore volume are beneficial to mass transport for catalytic reaction (28). Spent catalysts CoCe/CNTs and CoCeB/CNTs are characterized by TEM. On spent CoCe/CNTs (Figure S2a), small spherical particles are still uniformly distributed on CNTs, whereas particles agglomeration is observed on the CoCeB/ CNTs catalyst (Figure S2b). A small amount of P atoms randomly distributed on the surface are also observed. X-ray photoelectron spectroscopy (XPS) measurement was carried out for the spent catalysts. The results are shown in Figure S3 and is summarized in Table S1. Essentially, Co 2p3/2 core level spectra include two bands. The first band centered at 781.9 eV is attributed to Co2+ ions (24), presumably in the form of cobalt phosphate (29, 30). The second band centered at 779.1 eV is attributed to the formation of CoP and Co2P (29). The P 2p3/2 core level spectra of all the spent catalysts are observed at binding energy of around 134.4 and 129.4 eV. The band centered at 129.4 eV is due to the presence of reduced phosphorus in the form of CoP and Co2P (29), and the band centered at 134.4 eV represents typical phosphate species formed on the surface of phosphide during passivation (29, 30). All the Ce species are present in the oxidized state. The peak at 882.3 eV is attributed to CeO2 (4), whereas the peak at 885.4 eV is attributed to Ce2O3 (31). In contrast, the B species in the spent CoCeB/CNTs and CoCeBP/CNTs catalysts are present in both the elemental state (B, 187.6 eV) and the oxidized state (BO2-, 192.4 eV) (23). After the catalytic test, the spent catalysts were subjected to passivation, so that a thin layer of oxide is formed on the surfaces of the particles, preventing deep oxidation of the catalysts upon air exposure. In the present study, the four spent catalysts have different particle sizes. The XRD results showed that the CoP particle size on spent CoCe/CNTs and Co/CNTs is smaller than that on CoCeB/CNTs and CoCeBP/ CNTs catalysts. Small particles are more easy to be oxidized. As a result, the intensity of reduced Co and P species (CoP, Co2P) on CoCe/CNTs and Co/CNTs catalysts are not as strong as that of CoCeBP/CNTs and CoCeB/CNTs catalysts. Catalytic Performance in PH3 Decomposition Reaction. The decomposition of PH3 on CoCe/CNTs is a dynamic process: first the phosphidation takes place to form CoP and
then PH3 adsorbs and reacts on the active phase CoP to form P and H2. The boiling point of P is 280.5 °C; therefore, under reaction conditions of 360 °C and 0.1 MPa, P desorbs from catalyst surface to gas phase. Phosphorus is collected in a condenser installed at outlet of the reactor. The accumulation of free P atoms on catalyst surface is very minimum, which is supported by the data from XRD and XPS analysis. The formation of cobalt phosphides (CoP) active phase and its catalytic behavior are very similar to the formation of metal nitride and metal sulfide on ammonia decomposition and hydrodesulfurization catalysts (28, 32). The catalytic performance of all four catalysts in PH3 decomposition reaction was studied by varying the reaction temperature. Since PH3 decomposition is an endothermic reaction (33), an increase in reaction temperature would result in an enhancement of PH3 conversion. As shown in Figure S4, the CoCe/CNTs catalyst reduced by H2 exhibits the highest activity. On raising temperature, conversion of PH3 on the CoCe/CNTs catalyst increases, reaching 99.8% conversion at 360 °C. Apparently, H2 reduction is the most suitable method for preparing a highly active catalyst. The CoCeB/CNTs is the least active catalyst, achieving 99.7% conversion at 440 °C. The CoCeBP/CNTs achieves 99.8% conversion at 420 °C. Both CoCeB/CNTs and CoCeBP/CNTs are less active than Co/CNTs. Referring to the XRD results shown in Figure 3, the spent CoCe/CNTs and Co/CNTs catalysts only contain the CoP phase; no Co2P and Co phases are observed in these two catalysts. The Co and Co2P phases present in spent CoCeB/CNTs and CoCeBP/CNTs catalysts seem to be less active than the CoP phase. That is why CoCeB/CNTs and CoCeBP/CNTs catalysts are less active in PH3 decomposition. In other words, KBH4 reduction causes a decrease in catalytic activity. When KBH4 is partially replaced by NaH2PO2 · H2O in the reduction process, the effect on inhibiting catalytic activity is reduced. Referring to the XRD patterns shown in Figure 2, at high temperature, P in the CoCeBP/CNTs is leached out from amorphous alloy to form Co2P. As shown in Table S1, XPS analysis indicates that B in CoCeB/CNTs and CoCeBP/CNTs catalysts presents in the form of elemental state and oxidized state in the catalyst, respectively. It has been reported that Co2P could react with PH3 to form the CoP phase (29). As a result, it is speculated that the B compound is responsible for the suppression of phosphidation of Co crystallite. Phosphidation of Co to form CoP is important for the PH3 decomposition reaction, because CoP is the most active phase. Moreover, the dramatic decrease VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
729
FIGURE 4. Stability of different catalysts tested at 360 °C (GHSVPH3 ) 2520 mL/h.gcat). in surface area and pore volume of CoCeB/CNTs and CoCeBP/CNTs catalysts during the PH3 decomposition reaction is another cause of low activity. In order to further understand the effect of reduction methods on the Ce promoted Co catalysts, the stability of the catalysts was investigated under the conditions of T ) 360 °C and GHSVPH3 ) 2520 mL/h.gcat for a duration of 32 h. According to the results shown in FigureS4, at 360 °C, decomposition of PH3 on the best catalyst CoCe/CNTs is near completion. Therefore, activity of all other catalysts can be compared with CoCe/CNTs under this temperature. As shown in Figure 4, PH3 conversion on all four catalysts increases initially, reaching the maximum. After reaching the maximum activity, CoCeB/CNTs catalyst undergoes a gradual deactivation after 4 h time-on-stream. Higher PH3 conversion is achieved with CoCeBP/CNT and Co/CNTs. However, deactivation is inevitable due to agglomeration of the phosphides particles. The catalytic behaviors of these three catalysts are speculated as follows: On all these catalysts, phosphidation of Co and agglomeration of CoP are two competing processes which occurred simultaneously, having an opposite effect on catalyst performance. Cobalt phospidation is a dynamic process, whereas agglomeration is irreversible. For CoCeB/CNTs, the presence of high concentration B on the CoCe particles prevents phosphidation and that is why the maximum conversion of CoCeB/CNTs is lower than other catalysts. In the meantime, on CoCeB/ CNTs, before Co is fully phosphidized, the CoP agglomeration rate increases and exceeds the phosphidation rate. Large Co particles become even more difficult to phosphidize, which creates a negative impact on continued phosphidation of Co. As a result, rapid deactivation after 4 h time on stream is observed on CoCeB/CNTs. In contrast, for CoCeBP/CNTs and Co/CNTs, because the inhibition effect on Co phosphidation at the beginning of the reaction is less intensive, the CoP agglomeration does not exert negative impact until after more than 12 h. When the CoP agglomeration rate is further increased, the Co phosphidation rate decreases, and then deactivation of these two catalysts becomes inevitable. Different from the other three catalysts, conversion of CoCe/CNTs reaches about 100% after 3 h time-on-stream. No evidence of deactivation is observed after 32 h time-onstream operation. The superior performance of CoCe/CNTs catalyst is attributed to the high dispersion of CoP particles formed during reaction and the resistance to thermal degradation. In summary, a series of Co/CNTs catalysts promoted by Ce oxides is prepared by the coimpregnation method and then reduced by different methods. The catalytic activities of these catalysts in the PH3 decomposition reaction are 730
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011
compared with the baseline Co/CNTs catalyst. For the CoCe/ CNTs catalyst, hydrogen reduction results in the formation of large surface area and large pore volume catalyst with reduced tendency of particle agglomeration during catalytic reaction. The CoCe/CNTs catalyst exhibits stable activity in the PH3 decomposition reaction, whereas CoCeB/CNTs and CoCeBP/CNTs catalysts are less active than the baseline Co/ CNTs catalyst; both catalysts undergo gradual deactivation. The results demonstrate that the reduction method employed in the preparation of these catalysts determines the type of phases formed in the catalytic reaction. It seems likely that the CoP phase is more active than the Co2P phase and the cubic Co phase. For the catalyst reduced by KBH4 (CoCeB/ CNTs), the decrease in activity and stability is probably caused by the formation of large CoP alloy particles and less active Co2P and cubic Co phases in the reaction. Overall, this study has demonstrated that the addition of Ce oxides to Co/CNTs can improve the catalytic activity and stability of Co/CNTs catalyst in the PH3 decomposition reaction, and H2 reduction is the most suitable method for preparing highly active CoCe/ CNTs catalyst. Such a catalyst system has a great potential to both environmental and economic benefits.
Acknowledgments The authors gratefully acknowledge financial support from the National Natural Science Foundation of China under the Grant No. 20772116.
Supporting Information Available CNTs pretreatment procedure, detailed XPS spectra, TEM images of fresh and spent catalysts, effect of temperature, and the yield of phosporus over each catalyst. This material is available free of charge via the Internet at http:// pubs.acs.org.
Literature Cited (1) Yu, Q.; Yi, H.; Tang, X.; Ning, P.; Wang, C. Progress on Phosphine control technology. Environ. Sci. Technol. (Chinese) 2009, 32 (10), 87–91. (2) Panish, M. B.; Hamm, R. A. A mass spectrometric study of AsH3 and PH3 gas sources for molecular beam epitaxy. J. Cryst. Growth 1986, 78 (3), 445–452. (3) Zheng, W.; Zhang, J.; Ge, Q.; Xu, H.; Li, W. Effects of CeO2 addition on Ni/Al2O3 catalysts for the reaction of ammonia decomposition to hydrogen. Appl. Catal., B 2008, 80, 98–105. (4) Li, H.; Zhang, S.; Luo, H. A Ce-promoted Ni-B amorphous alloy catalyst (Ni-Ce-B) for liquid-phase furfural hydrogenation to furfural alcohol. Mater. Lett. 2004, 58, 2741–2746. (5) Daza, C. E.; Kiennemann, A.; Moreno, S.; Molina, R. Dry reforming of methane using Ni-Ce catalysts supported on a modified mineral clay. Appl. Catal., A 2009, 364, 65–74. (6) Mu, Z.; Li, J.; Duan, M.; Hao, Z.; Qiao, S. Catalytic combustion of benzene on Co/CeO2/SBA-15 and Co/SBA-15 catalysts. Catal. Commun. 2008, 9, 1874–1877. (7) Lucre´dio, A. F.; Jerkiewicz, G.; Assaf, E. M. Cobalt catalysts promoted with cerium and lanthanum applied to partial oxidation of methane reactions. Appl. Catal., B 2008, 84, 106– 111. (8) Iijima, S. Helical microtubules of graphitic carbon. Nature 1991, 354, 56–58. (9) Yin, S.; Xu, B.; Zhu, W. Carbon nanotubes-supported Ru catalyst for the generation of COx-free hydrogen from ammonia. Catal. Today 2004, 93-95, 27–38. (10) Nhut, J. M.; Pesant, L.; Tessonnier, J. P.; Wine´, G.; Guille, J.; Pham-Huu, C.; Ledoux, M. J. Mesoporous carbon nanotubes for use as support in catalysis and as nanosized reactors for one-dimensional inorganic material synthesis. Appl. Catal., A 2003, 254, 345–363. (11) Bahome, M. C.; Jewell, L. L.; Padayachy, K.; Hildebrandt, D.; Glasser, D.; Datye, A. K.; Coville, N. J. Fe-Ru small particle bimetallic catalysts supported on carbon nanotubes for use in Fischer-Tro¨psch synthesis. Appl. Catal., A 2007, 328, 243–251. (12) Wang, S.; Yin, S.; Li, L.; Xu, B.; Ng, C.; Au, C. Investigation on modification of Ru/CNTs catalyst for the generation of COx-
(13) (14)
(15) (16) (17) (18)
(19) (20) (21) (22) (23) (24)
free hydrogen from ammonia. Appl. Catal., B 2004, 52, 287– 299. Yin, S.; Zhang, Q.; Xu, B.; Zhu, W.; Ng, C.; Au, C. Investigation on the catalysis of COx-free hydrogen generation from ammonia. J. Catal. 2004, 224, 384–396. Tre´panier, M.; Tavasoli, A.; Dalai, A. K.; Abatzoglou, N. FischerTropsch synthesis over carbon nanotubes supported cobalt catalysts in a fixed bed reactor: influence of acid treatment. Fuel Process Technol. 2009, 90, 67–374. Matsubara, H.; Tabei, S.; Ichimura, S.; Iso, A. Production of elemental phosphorus. Japan Patent 01-313309, 1989. Han, C.; Ren, J.; Tang, X.; Zhang, B. Preparation of nanometer FeCuP alloy and its application in decomposition of PH3. Chin. Chem. Lett. 2007, 18, 1285–1288. Han, C.; Lin, X.; Zhang, B. Effect of support TiO2 on the characteristics of Co-P amorphous alloy. Acta Chim. Sin. 2007, 65 (9), 793–797. Li, L.; Han, C.; Yang, L.; Wang, X.; Zhang, B. The nature of PH3 decomposition reaction over amorphous CoNiBP alloy supported on Carbon nanotubes. Ind. Eng. Chem. Res. 2010, 49, 1658–1662. Chen, C.; Chen, X.; Yi, B.; Liu, T.; Li, W. Zinc oxide nanoparticle decorated multi-walled carbon nanotubes and their optical properties. Acta Mater. 2006, 54, 5401–5407. Wonterghem, J. V.; Morup, S.; Koch, C. J.; Charles, S. W.; Wells, S. Formation of ultra- fine amorphous alloys particles by reduction in aqueous solution. Nature 1986, 322 (14), 622–623. Li, H.; Zhao, Q.; Li, H. Selective hydrogenation of p-chloronitrobenzene over Ni-P-B amorphous catalyst and synergistic promoting effects of B and P. J. Mol. Catal., A 2008, 285, 29–35. Yan, X.; Sun, J.; Wang, Y.; Yang, J. A Fe-promoted Ni-P amorphous alloy catalyst (Ni-Fe-P) for liquid phase hydrogenation of mand p-chloronitrobenzene. J. Mol. Catal., A 2006, 252, 17–22. Li, H.; Li, H.; Deng, J. The crystallization process of ultrafine Ni-B amorphous alloy. Mater. Lett. 2001, 50, 41–46. Li, H.; Yang, P.; Chu, D.; Li, H. Selective maltose hydrogenation to maltitol on a ternary Co-P-B amorphous catalyst and the
(25)
(26)
(27)
(28) (29)
(30)
(31) (32) (33)
synergistic effects of alloying B and P. Appl. Catal., A 2007, 325, 34–40. Jiang, Z.; Yang, H.; Wei, Z.; Xie, Z.; Zhong, W.; Wei, S. Catalytic properties and structures of nano-amorphous Ni-B alloys affected by annealing temperatures. Appl. Catal., A 2005, 279, 165–171. Parks, G. L.; Pease, M. L.; Burns, A. W.; Layman, K. A.; Bussell, M. E.; Wang, X.; Hanson, J.; Rodriguez, J. A. Characterization and hydrodesulfurization properties of catalysts derived from amorphous metal-boron materials. J. Catal. 2007, 246, 277– 292. Rajesh, B.; Sasirekha, N.; Lee, S.; Kuo, H.; Chen, Y. Investigation of Fe-P-B ultrafine amorphous nanomaterials: influence of synthesis parameters on physicochemical and catalytic properties. J. Mol. Catal., A 2008, 289, 69–75. Li, L.; Zhu, Z.; Wang, S.; Yao, X.; Yan, Z. Chromium oxide catalysts for COx-free hydrogen generation via catalytic ammonia decomposition. J. Mol. Catal., A 2009, 304 (1-2), 71–76. Cecilia, J. A.; Infantes-Molina, A.; Rodrı´guez-Castello´n, E.; Jime´nez-Lo´pez, A. Dibenzothiophene hydrodesulfurization over cobalt phosphide catalysts prepared through a new synthetic approach: effect of the support. Appl. Catal., B 2009, 92, 100– 113. Sawhill, S. J.; Layman, K. A.; Van Wyk, D. R.; Engelhard, M. H.; Wang, C.; Bussell, M. E. Thiophene hydrodesulfurization over nickel phosphide catalysts: effect of the precursor composition and support. J. Catal. 2005, 231, 300–313. Liu, J.; Zhao, Z.; Wang, J.; Xu, C.; Duan, A.; Jiang, G.; Yang, Q. The highly active catalysts of nanometric CeO2-supported cobalt oxides for soot combustion. Appl. Catal., B 2008, 84, 185–195. Skrabalak, S. E.; Suslick, K. S. On the possibility of metal borides for hydrodesulfurization. Chem. Mater. 2006, 18, 3103–3107. Sun, Y.; Law, D. C.; Hicks, R. F. Kinetics of phosphine adsorption and phosphorus desorption from gallium and indium phosphide (001). Surf. Sci. 2003, 540, 12–22.
ES1022416
VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
731
