Synthesis and Catalysis of Location-Specific ... - ACS Publications

Apr 23, 2012 - Cobalt nanoparticles located on the concave internal surface of multiwalled carbon nanotubes (Co-in-MW-CNTs) and the convex external su...
0 downloads 8 Views 3MB Size
Download PDF
Article pubs.acs.org/Langmuir

Synthesis and Catalysis of Location-Specific Cobalt Nanoparticles Supported by Multiwall Carbon Nanotubes for Fischer−Tropsch Synthesis Yuan Zhu, Yingchun Ye, Shiran Zhang, Mark E. Leong, and Franklin (Feng) Tao* Department of Chemistry and Biochemistry, University of Notre Dame, Notre Dame, Indiana 46556, United States ABSTRACT: Cobalt nanoparticles located on the concave internal surface of multiwalled carbon nanotubes (Co-in-MW-CNTs) and the convex external surface of MW-CNTs (Co-on-MW-CNTs) were synthesized. Their catalytic performances in Fischer−Tropsch synthesis (FTS) were investigated. A correlation between the location, pretreatment, and surface chemistry of the cobalt nanoparticles and the catalytic selectivity in FTS was built. It is found that the selectivity in production of C5+ molecules through FTS on cobalt catalysts supported by MW-CNTs depends on activation temperatures and surface chemistry of the cobalt nanoparticles. A pretreatment at 300 °C in H2 flow results in a different surface chemistry for Co-in-MW-CNTs than for Coon-MW-CNTs, which leads to a difference in selectvity to the production of C5+ molecules. Pretreatment at a relatively high temperature, 400 °C, in H2 flow produces completely reduced Co nanoparticles in Co-in-MW-CNTs and Co-on-MW-CNTs. There is no signifcant difference in catalytic selectivity between the two catalysts upon pretreatment at 400 °C. The absence of a significant difference in catalytic selectivity of metallic Co-on-MW-CNTs and metallic Co-in-MW-CNTs suggests that the electronic effect of the MW-CNT support does not significantly affect the C5+ selectivity of cobalt catalysts in FTS.

1. INTRODUCTION In Fischer−Tropsch synthesis (FTS), syngas (a mixture of CO and H2) can be catalytically converted to liquids, clean fuels, and chemical feedstock through a surface polymerization reaction. Since the pioneering work of Franz Fischer and Hans Tropsch on iron- and cobalt-based catalysts,1 efforts in synthesis technology have essentially synchronized with the crests and troughs of petro-economics. Designing FTS catalysts for the production of transportation fuels has been an important approach in the development of sustainable energy sources.2 For carbon nanotubes (CNTs), there is a difference in πelectron density of bent graphene layers between the concave internal wall and the convex external surface.3 Early theorertical studies suggested a correlation between the electronic structure of CNTs and the transition states of chemical reactions performed on them.4 Recently, CNTs were used to support catalysts for FTS,5 alcohol synthesis,5f,6 and other reactions.7 Bao et al. studied the redox properties of iron oxide nanoparticles by encapsulating iron catalysts into CNTs with different inner diameters.5d They demonstrated that reduction of iron oxide to iron carbide in FTS is enhanced when iron oxide is located on the internal wall of a multiwalled carbon nanotube (MW-CNT). The active phase, iron carbide, was identified using in situ X-ray diffraction (XRD) when CO contacts Fe2O3 during FTS.5d This study suggested a promotion effect in the selectivity of FTS when iron catalysts © 2012 American Chemical Society

are located inside MW-CNTs, as opposed to being located on the external surfaces of MW-CNTs.5c Iron, cobalt, nickel, ruthenium, and rhodium are active for FTS. Cobalt is 2−3 times more active than iron catalysts without promotion and even 10 times more active with promotion.2,8 In this work, we selected cobalt to investigate how the internal concave and external convex surfaces of MWCNTs can be correlated to or even shape the catalytic selectivity in the production of C5+ molecules in FTS. In addition, this investigation was motivated by the fact that cobalt differs from iron in that its active phase is metallic instead of cobalt carbide. There is no phase transformation from cobalt oxide to cobalt carbide during FTS,8 although carbon is immisible in cobalt. Typically, CNTs have higher thermal and chemical stabilities than amorphous carbon supports. Co does not form cobalt carbide at temperatures of 200 °C or lower. This could make CNTs a good support for Co catalysts. In this work, we studied whether the internal surface of a MW-CNT could electronically promote the reduction of cobalt oxide and whether the internal concave surface of cobalt catalysts enhances the C5+ selectivity of FTS. We synthesized locationspecific cobalt nanoparticle catalysts supported by MW-CNTs as schematically shown in Figure 1 and investigated their C5+ catalytic selectivities in FTS. Received: February 10, 2012 Revised: March 27, 2012 Published: April 23, 2012 8275

dx.doi.org/10.1021/la300607k | Langmuir 2012, 28, 8275−8280

Langmuir

Article

Figure 1. Schematic showing the location-specific carbon-nanotubesupported cobalt catalysts for Fischer−Tropsch synthesis.

2. EXPERIMENTAL SECTION Synthesis of Cobalt Nanoparticle Catalysts Supported on the Internal Surface of MW-CNTs. A 0.5-g capped multiwalled carbon nanotubes (MW-CNTs, purity > 95%, i.d. = 10−20 nm, o.d. = 20−30 nm, US Research Nanomaterials, Inc.) was suspended in 20.0 g of concentrated nitric acid (68−70%, Sigma-Aldrich Co.) solution containing 1.0 g of cobalt(II) nitrate hexahydrate (>98%, SigmaAldrich Co.) and refluxed at 140 °C for 4.5 h. The loading amount was ∼5 wt % as measured by inductively coupled plasma atomic emission spectroscopy (ICP-AES). The suspension was allowed to settle, and the red-brown supernatant was carefully removed with a polytetrafluoroethylene (PTFE) pipet. The resulting insoluble black product was subjected to mild vacuum filtration. The product was scraped off the filter paper and dried at 120 °C overnight in an oven to form the catalyst precursor. Before FTS reaction, the precursor was activated by reduction in a pure H2 flow (50 mL min−1) at 300 or 400 °C for 5 h at a ramp rate of 2 °C min−1, forming the reactive Co-in-MW-CNTs catalyst. The protocol of synthesis of colbalt catalysts supported on internal walls of MW-CNTs is shown schematically in Figure 2. Synthesis of Cobalt Nanoparticle Catalysts Supported on External Surfaces of MW-CNTs. A 1.0-g capped MW-CNTs (the same MW-CNTs as used for the synthesis of Co-in-MW-CNTs) was dispersed in 50 mL of diluted nitric acid (35%) and refluxed at 110 °C for 5 h. The black solid was separated by solution filtration and washed with deionized water several times until the pH value was around 7. It was then dried in an oven at 100 °C overnight and redispersed in aqueous cobalt(II) nitrate solution (the loading amount of Co was ∼10 wt % as identified by ICP-AES). After ultrasonication for 1 h, extensive stirring was performed to obtain a homogeneous dispersion of Co nanoparticles on the external surface of carbon nanotubes. As for the Co-in-MW-CNTs, the catalyst precursor was activated by reduction in a pure H2 flow (50 mL min−1) at 300 or 400 °C for 5 h at a ramp rate of 2 °C min−1 before FTS. Measurement of Catalytic Performances of Co-in-MW-CNTs and Co-on-MW-CNTs. A 0.1-g cobalt-supported MW-CNTs catalysts was loaded into a fixed-bed single-pass high-pressure FTS microreactor. The catalytic performances of Co-in-MW-CNTs and Co-on-MW-CNTs were examined and evaluated at 205 °C in syngas (67% H2 and 33% CO) with a pressure of 20 bar and a flow rate of 33.3 mL/min. The products were analyzed online by gas chromatography using a capillary column (Agilent HP-PONA, 50 m × 0.2 mm × 0.5 μm) and two packed columns. Both a thermal conductivity detector (TCD) and a flame ionization detector (FID) were used. The C5+ selectivity of a catalyst in FTS is defined as the molar ratio of the C5+ hydrocarbons to the total of all products.

Figure 2. Schematic showing the process of synthesis of Co-in-MWCNTs. (a) Breaking of MW-CNTs through oxidation of defect sites of MW-CNT surfaces using concentrated HNO3. (b) Deposition of cobalt ions on both external and internal walls of MW-CNTs through refluxing in concentrated HNO3. Blue spots: drops of cobalt ion solution. Steps a and b are performed simultaneously. (c) Vacuum filtration to purge most of the cobalt solution from the external surface of MW-CNTs and remove a small portion of the adsorbed solution from the internal concave surface of MW-CNTs. (d) Nucleation of cobalt and formation of precursor cobalt catalysts in an oven at 120 °C. Red spots: cobalt catalyst precursors.

of iron catalysts in the MW-CNTs.5a A similar protocol was used in their syntheses of rhodium5f and ruthenium5b catalytic materials. In addition, other methods were developed for synthesizing iron catalysts supported by MW-CNTs.9 However, it is still quite challenging to synthesize location-specific nanoparticle catalysts supported by MW-CNTs. There is certainly not a universal recipe to follow for the synthesis of metal nanoparticles supported on the concave internal surface of MW-CNT. Synthesis of cobalt catalysts supported on the internal surface of MW-CNTs with a high spatial selectivity has been quite challenging until now. This is probably due to the kinetic nature of the growth of cobalt nanoparticles in CNTs, because the synthesis involves cap opening and mass transfer of the precursor in solution, filtration in vacuum, and nucleation during calcination. The number of Co nanoparticles in MWCNTs and the total number of Co nanoparticles in transmission electron microscopy (TEM) images collected from different windows of the same TEM grid of the same sample (Co-in-MW-CNTs) were counted. The ratio of the number of Co nanoparticles in MW-CNTs to the total number is defined as the selectivity of location to the internal wall of MW-CNTs. This selectivity falls to 80−85%. The same statistical counting was done for other TEM grids prepared from other Co-in-MWCNT samples synthesized with the same protocol. The ratios were in the range of 80−85%. Thus, our protocol described in the Experimental Section allowed for the repeated synthesis of cobalt nanoparticle catalysts supported on MW-CNTs, in which 80−85% of the cobalt nanoparticles were located on the internal walls of the MW-CNTs (Figure 3a).

3. RESULTS AND DISCUSSION It is challenging to synthesize site-specific metal nanoparticles supported by MW-CNTs. Bao et al. developed a protocol involving a Ag-catalyzed oxidation to break MW-CNTs, ultrasonic treatment, and stirring, followed by impregnation 8276

dx.doi.org/10.1021/la300607k | Langmuir 2012, 28, 8275−8280

Langmuir

Article

Figure 3. TEM images of the synthesized Co-in-MW-CNTs: (a) large-scale image, (b) high-resolution image. Inset of b: diffraction pattern.

Figure 4. TEM images of the synthesized Co-on-MW-CNTs: (a) large-scale, (b) high-resolution image.

In general, the Co nanoparticles located in or on MW-CNTs can be distinguished by their different shapes and locations related to MW-CNTs. Their shapes and locations can be identified using TEM images such as in Figures 3 and 4. As Figure 4 shows, the Co nanoparticles on MW-CNTs have a spherelike shape and are always located near the outermost edges of the MW-CNTs. In this case, the edges of the Co nanoparticles on the MW-CNTs align with the outermost surface of the MW-CNTs. However, the edges of the Co nanoparticles in MW-CNTs did not align with the outermost edges of the MW-CNTs because the wall thickness of the carbon nanotubes was ∼5 nm. In contrast to Co nanoparticles on MW-CNTs, Co particles in MW-CNTs have a cylinderlike or rodlike shape (Figure 3). This is due to the limited growth along the direction perpendicular to the long axis of the MW-CNTs. Thus, Co nanoparticles in MW-CNTs preferentially grow along the long axis of the MW-CNTs, forming a rodlike shape (Figure 3b). The size of the cobalt nanoparticles in the MW-CNTs in this work (Figure 5) was found to be ∼4.4 ± 0.9 nm. The shapes and relative locations of these Co nanoparticles largely aided in the identification of the spatial selectivity of cobalt nanoparticles supported on or in MW-CNTs. If nanoparticles are supported on single wall carbon nanotubes (SW-CNTs), it will be necessary to pursue other methods as well to identify the locations of Co nanoparticles in or on SW-

Figure 5. Cobalt oxide nanoparticle size distribution according to TEM images of as-synthesized Co-in-MW-CNTs.

CNTs because the ultrathin wall (one atomic layer) of SWCNTs cannot help identify the locations of nanoparticles. We must note that the vacuum (∼1 × 10−2 Torr) and time (2−3 min) in filtration to purge extra solution on the external wall of CNTs through vacuum filtration are two crucial parameters controlling the loading of cobalt in CNTs, the homogeneity of the size distribution, and the location of the cobalt nanoparticles. Before vacuum filtration, cobalt ions are impregnated on the external wall and diffused into the internal 8277

dx.doi.org/10.1021/la300607k | Langmuir 2012, 28, 8275−8280

Langmuir

Article

This suggests that the reduction of cobalt oxide at 300 °C was not complete for Co-in-MW-CNTs. Interestingly, the cobalt oxide of Co-on-MW-CNTs was completely reduced to the metallic state upon the same pretreatment (Figure 6a,b) in contrast to Co-in-MW-CNTs (Figure 6c,d). The influence of the incomplete reduction on the catalytic selectivity is discussed in the following paragraphs. Figure 7 presents the XRD patterns of Co-on-MW-CNTs and Co-in-MW-CNTs upon pretreatment at 400 °C and upon

wall. Nearly complete removal of the solution of Co2+ from the external surface along with controlled removal of a portion of the solution encapsulated in the MW-CNTs is the key allowing the precursor of cobalt oxide to be located specifically on the internal wall. Upon calcination, 80−85% of the cobalt oxide nanoparticles are encapsulated in the MW-CNTs (Figure 3). Cobalt nanoparticles supported on the external surface of MW-CNTs (Co-on-MW-CNTs) were synthesized with wellcontrolled impregnation, as described in the Experimental Section. Notably, a distinctly different process was used to synthesize Co-on-MW-CNTs compared to that used to synthesize Co-in-MW-CNTs. The synthesis of Co-on-MWCNTs requires only a mild treatment of MW-CNTs in HNO3 to avoid the potential opening of CNT caps and breakage of MW-CNTs at higher concentrations. Thus, nitric acid is typically diluted. The size of the synthesized cobalt nanoparticles was 5−10 nm. Figure 4a shows a TEM image of synthesized Co-on-MW-CNTs. For Co-on-MW-CNTs, most cobalt nanoparticles are located on the external surface of the MW-CNTs. Before evaluation of catalytic performance, the catalysts were pretreated in H2 at 2 bar at 300−450 °C for 4.5 h. Catalytic performances were measured in a fixed-bed single-pass flowing reactor at 205 °C in syngas at a pressure of 20 bar. Figure 6a

Figure 7. XRD patterns of Co-on-MW-CNTs and Co-in-MW-CNTs (a,c) upon pretreatment at 400 °C and (b,d) after both pretreatment at 400 °C and FTS at 205 °C.

pretrement at 400 °C followed by FTS at 205 °C. Notably, there are no significant differences in the XRD patterns between Co-in-MW-CNTs and Co-on-MW-CNTs. Upon pretreatment at 400 °C in H2 (Figures 7a and 7c), only metallic cobalt was identified. There is no identifiable diffraction pattern of cobalt oxide in Figure 7, showing that the cobalt oxide in both Co-in-MW-CNTs and Co-on-MWCNTs was completely reduced at 400 °C. It is noted that the (111) diffraction peak of Co nanoparticles is a combination of a sharp and broad peaks (Figure 7). This suggests a bimodal distribution of metallic cobalt nanoparticle catalysts upon pretreatment at 400 °C in H2. Based on the Scherrer equation, the appearance of a sharp peak indicates the aggregation of nanoparticles. The coexistence of the sharp and wide peaks at 44.2° showed that 400 °C is a relatively high pretreatment temperature for Co nanoparticle catalysts supported by MWCNTs. The TEM image of Co nanoparticles on MW-CNTs is shown in Figure 8. Both large and small particles coexist, which is consistent with a bimodal distribution. However, there is no such bimodal distribution of iron nanoparticle catalysts supported on MW-CNTs under a very similar pretreatment in H2 followed by FTS.5c From this point of view, MW-CNTs seem to be a better support for iron catalysts than for cobalt catalysts. The differences in aggregation between Co-on-MWCNTs and Fe-on-MW-CNTs and between Co-in-MW-CNTs and Fe-in-MW-CNTs probably result from the fact that an iron carbide phase is formed. It is expected that iron carbide could strongly bond to the internal carbon surface of MW-CNTs through Fe−C bonds. However, reduced metallic cobalt particles could aggregate readily due to the absence of strong chemical binding between cobalt atoms and CNTs. Thus, the absence of strong chemical binding results in a bimodal distribution of metal cobalt particles supported by MW-CNTs, in contrast to iron carbide supported by MW-CNTs. Table 1 lists the measured catalytic selectivities of Co-inMW-CNTs and Co-on-MW-CNTs in the production of C5+

Figure 6. XRD patterns of Co-on-MW-CNTs and Co-in-MW-CNTs (a,c) upon pretreatment at 300 °C and (b,d) after pretreatment at 300 °C followed by FTS at 205 °C.

shows XRD patterns of Co-on-MW-CNTs pretreated at 300 °C. The XRD pattern of Co-on-MW-CNTs after pretreatment followed by FTS is shown in Figure 6b. The peak at 44.2° is attributed to the (111) face of metallic cobalt (JCPDS 150806). Clearly, Co-on-MW-CNTs remains in the metallic state upon FTS, a synthesis performed at 205 °C for 34 h in syngas (67% CO, 33% H2) at 20 bar (Figure 4b). The size of cobalt nanoparticles supported by CNTs can be calculated by the Scherrer equation [d = 0.89λ/β(cos θ)]. For Co-on-MW-CNTs and Co-in-MW-CNTs upon pretreatment at 300 °C followed by FTS, the sizes of the Co nanoparticles are both ∼5 nm, which is similar to those obtained upon only pretreatment at 300 °C (both ∼5 nm). TEM studies suggested a similarity in the sizes of Co nanoparticles in Co-on-MW-CNTs that experienced pretreatment followed by FTS and those that experienced only pretreatment. Traces c and d of Figure 6 are the XRD patterns of Co-inMW-CNTs upon pretreatment at 300 °C and after pretreatment followed by FTS, respectively. One extra peak of Co-inMW-CNTs in Figure 6c,d was identified at 36.9°, which can be assigned to the (311) face of cobalt oxide (JCPDS 42-1467). 8278

dx.doi.org/10.1021/la300607k | Langmuir 2012, 28, 8275−8280

Langmuir

Article

number of 7 at step edges compared to those on terraces with a coordination number of 9.10 In addition, this hypothesis seems to be consistent with the thermodynamically favorable oxidation of transition-metal nanoparticles that have large fractions of undercoordinated metal atoms on their surfaces.11 The oxide nanoclusters or oxygen atoms strongly bound at the corners of cobalt nanoparticles could act as termination or poisoning sites for the chain propagation of FTS on the surface of cobalt nanoparticles. In fact, XRD patterns (Figure 6) show that the pretreatment at 300 °C in H2 does not lead to aggregation of cobalt nanoparticles. The absence of aggregation of cobalt nanoparticles upon pretreatment at 300 °C is consistent with the existence of oxide nanoclusters at the corners of the cobalt nanoparticles and the above hypothesis (termination of chain propagation at oxide nanoclusters on cobalt nanoparticles) because oxides typically have much smaller surface free energies than metals and thus resist to aggregate. Thus, the difference in catalytic selectivity between Co-on-MW-CNTs and Co-in-MWCNTs upon pretreatment at 300 °C (Table 1) likely results from the different surface chemistries rather than a potential electronic effect of MW-CNTs in this case. Upon pretreatment at 400 °C in H2, the cobalt nanoparticles were completely reduced. There was no significant difference in XRD patterns between Co-on-MW-CNTs and Co-in-MWCNTs (Figure 7). Catalytic measurements showed that the selectivity of Co-on-MW-CNTs in the production of C5+ was only 3% higher than that of Co-in-MW-CNTs. There was no significant difference in selectivity in producing C5+ molecules. As described above, the XRD patterns of Co-on-MW-CNTs and Co-in-MW-CNTs upon pretreatment at 400 °C and FTS at 205 °C both had a wide peak and a sharp one (Figure 7). Both large and small cobalt nanoparticles of Co-on-MW-CNTs were observed in TEM studies (Figure 8). The complete reduction upon pretreatment at 400 °C and the absence of chemical binding between cobalt nanoparticles and MW-CNTs make metal cobalt nanoparticles aggregate partially. Compared to cobalt catalysts, iron catalysts of Fe-onMW-CNTs and Fe-in-MW-CNTs do not aggregate.5c Again, this difference could result from the formation of Fe−C bonds between Fe atoms of iron catalysts and MW-CNTs. Cobalt does not form cobalt carbide at the reaction temperature of 205 °C because Co-rich cobalt carbide alloy forms only at temperatures higher than 200 °C; however, iron carbides can be formed even at 0 °C.12 Thus, the absence of chemical binding of cobalt nanoparticles to MW-CNTs makes the diffusion barrier of metallic cobalt nanoparticles on MW-CNTs much lower than that of iron carbide nanoparticles.

Figure 8. TEM image of Co on MW-CNT after 400 °C pretreatment followed by FTS. Both large and small particles coexist.

Table 1. Catalytic Selectivities of C5+ Products from Cobalt Nanoparticles Supported by MW-CNTs in Fischer−Tropsch Synthesisa) cobalt nanoparticle catalysts

pretreatment at 300 °C

pretreatment at 400 °C

Co-in-MW-CNTs Co-on-MW-CNTs

15.7% 31.5%

23% 26%

a Catalyst of Co supported by CNTs, 0.1 g; flow rate, 33.3 mL/min; syngas, 67% CO + 33% H2.

molecules through FTS upon two representative pretreatments. A difference in C5+ selectivity was identified for Co-on-MWCNTs and Co-in-MW-CNTs. Co-on-MW-CNTs pretreated at 300 °C had a selectivity of 31.5%, much higher than the 15.7% value obtained for Co-in-MW-CNTs. This difference (31.5% versus 15.7%) is even larger than that reported recently between Fe-in-MW-CNTs and Fe-on-MW-CNTs.5c Analysis of the XRD patterns suggests that the large difference in catalytic selectivity between Co-in-MW-CNTs and Co-on-MW-CNTs (pretreated at 300 °C in H2) results from the incomplete reduction of Co-in-MW-CNTs (Figure 6c,d) and the complete reduction of Co-on-MW-CNTs (Figure 6a,b) at 300 °C. This is consistent with X-ray photoelectron spectroscopy (XPS) studies. Experimental observation of cobalt oxide existing on the cobalt catalyst of Co-in-MW-CNTs pretreated at 300 °C suggests that 300 °C is a high enough reduction temperature for Co-on-MW-CNTs but not for Co-in-MW-CNTs. Oxide remaining on the cobalt catalyst following pretreatment at 300 °C in H2 (Figure 6c) could result in the low C5+ selectivity. How the small portion of cobalt oxide on cobalt nanoparticles coexists with the metallic cobalt of the catalysts remains unknown, as it is challenging to access the internal surface of MW-CNTs directly using microscopy techniques. Our hypothesis is that the cobalt oxide exists at the corners of the cobalt nanoparticles or oxygen atoms chemically bonded to cobalt atoms at the corners. As the corner atoms of metal nanoparitcles have lower coordination numbers than those on the nanocrystallite faces, they typically have a large binding energy to oxygen molecules/atoms. Thus, from the point of view of thermodynamics, oxygen atoms preferentially bind to metal atoms at the corners. Related experimental evidence is the preferential oxidation of Pt atoms with a low coordination

4. SUMMARY Location-specific cobalt catalysts supported on the internal (Co-in-MW-CNTs) or external (Co-on-MW-CNTs) surface of MW-CNTs were synthesized. Catalytic measurements showed that selectivity in the production of C5+ molecules through FTS on Co-in-MW-CNTs depends on the activation temperature and surface chemistry of the cobalt nanoparticles. However, there is no significant difference between catalytic selectivity in the production of C5+ through FTS on Co-in-MW-CNTs and Co-on-MW-CNTs catalysts if they are pretreated at 400 °C in pure H2 and the cobalt nanoparticles are completely reduced. The absence of significant differences in the catalytic selectivities of metallic Co-on-MW-CNTs and metallic Co-inMW-CNTs suggests that the different electronic densities 8279

dx.doi.org/10.1021/la300607k | Langmuir 2012, 28, 8275−8280

Langmuir

Article

(7) (a) De Jong, K. P.; Geus, J. W. Carbon nanofibers: Catalytic synthesis and applications. Catal. Rev.Sci. Eng. 2000, 42 (4), 481− 510. (b) Serp, P.; Corrias, M.; Kalck, P. Carbon nanotubes and nanofibers in catalysis. Appl. Catal. A 2003, 253 (2), 337−358. (c) Li, L. J.; Khlobystov, A. N.; Wiltshire, J. G.; Briggs, G. A. D.; Nicholas, R. J. Diameter-selective encapsulation of metallocenes in single-walled carbon nanotubes. Nat. Mater. 2005, 4 (6), 481−485. (d) Shiozawa, H.; Pichler, T.; Gruneis, A.; Pfeiffer, R.; Kuzmany, H.; Liu, Z.; Suenaga, K.; Kataura, H. A catalytic reaction inside a single-walled carbon nanotube. Adv. Mater. 2008, 20 (8), 1443−1449. (8) Khodakov, A. Y.; Chu, W.; Fongarland, P. Advances in the development of novel cobalt Fischer−Tropsch catalysts for synthesis of long-chain hydrocarbons and clean fuels. Chem. Rev. 2007, 107 (5), 1692−1744. (9) Tessonnier, J. P.; Ersen, O.; Weinberg, G.; Pham-Huu, C.; Su, D. S.; Schlogl, R. Selective Deposition of Metal Nanoparticles Inside or Outside Multiwalled Carbon Nanotubes. ACS Nano 2009, 3 (8), 2081−2089. (10) Zhu, Z.; Tao, F.; Zheng, F.; Chang, R.; Li, Y.; Heinke, L.; Liu, Z.; Salmeron, M.; Somorjai, G. A. Nano Lett. 2012, 12, 1491. (11) Xu, Y.; Shelton, W. A.; Schneider, W. F. Effect of particle size on the oxidizability of platinum clusters. J. Phys. Chem. A 2006, 110 (17), 5839−5846. (12) Okamoto, H. Desk Handbook: Phase Diagrams for Binary Alloys; ASM International: Materials Park, OH, 2000.

between the internal concave and external convex surfaces of MW-CNTs do not significantly affect the catalytic selectivity in the production of C5+ molecules on cobalt catalysts supported by MW-CNTs, although this is demonstrated well in other FTS catalysts. Alternatively, the surface chemistry of cobalt nanoparticle catalysts supported in MW-CNTs affects their catalytic performance. These insights are important for developing FTS catalysts with high selectivities. In addition, an operando study using ambient pressure analytical techniques developed recently in our group is being planned to explore the surface chemistry in detail.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS F.T. acknowledges financial support from Center for Sustainable Energy at Notre Dame (cSEND). M.E.L. acknowledges financial support from cSEND through a Slatt fellowship.

■

REFERENCES

(1) (a) Fischer, F.; Tropsch, H. The direct synthesis of petroleum hydrocarbons with standard pressure. (First report). Ber. Dtsch. Chem. Ges. 1926, 59, 830−831. (b) Fischer, F.; Tropsch, H. Regarding the high molecular synthesis of paraffin hydrocarbon in carbon dioxide. Ber. Dtsch. Chem. Ges. 1927, 60, 1330−1334. (2) de Smit, E.; Weckhuysen, B. M. The renaissance of iron-based Fischer−Tropsch synthesis: On the multifaceted catalyst deactivation behaviour. Chem. Soc. Rev. 2008, 37 (12), 2758−2781. (3) (a) Haddon, R. C. Chemistry of the Fullerenes: The Manifestation of Strain in a Class of Continuous Aromatic Molecules. Science 1993, 261 (5128), 1545−1550. (b) Peralta-Inga, Z.; Lane, P.; Murray, J. S.; Body, S.; Grice, M. E.; O’Connor, C. J.; Politzer, P. Characterization of surface electrostatic potentials of some (5,5) and (n,1) carbon and boron/nitrogen model nanotubes. Nano Lett. 2003, 3 (1), 21−28. (4) Trepanier, M.; Dalai, A. K.; Abatzoglou, N. Synthesis of CNTsupported cobalt nanoparticle catalysts using a microemulsion technique: Role of nanoparticle size on reducibility, activity and selectivity in Fischer−Tropsch reactions. Appl. Catal. A 2010, 374 (1− 2), 79−86. (5) (a) Pan, X. L.; Bao, X. H. The Effects of Confinement inside Carbon Nanotubes on Catalysis. Acc. Chem. Res. 2011, 44 (8), 553− 562. (b) Guo, S.; Pan, X.; Gao, H.; Yang, Z.; Zhao, J.; Bao, X. Probing the Electronic Effect of Carbon Nanotubes in Catalysis: NH3 Synthesis with Ru Nanoparticles. Chem.Eur. J. 2010, 16 (18), 5379−5384. (c) Chen, W.; Fan, Z. L.; Pan, X. L.; Bao, X. H. Effect of confinement in carbon nanotubes on the activity of Fischer−Tropsch iron catalyst. J. Am. Chem. Soc. 2008, 130 (29), 9414−9419. (d) Chen, W.; Pan, X. L.; Bao, X. H. Tuning of redox properties of iron and iron oxides via encapsulation within carbon nanotubes. J. Am. Chem. Soc. 2007, 129 (23), 7421−7426. (e) Yu, G. B.; Sun, B.; Pei, Y.; Xie, S. H.; Yan, S. R.; Qiao, M. H.; Fan, K. N.; Zhang, X. X.; Zong, B. N. FexOy@C Spheres as an Excellent Catalyst for Fischer−Tropsch Synthesis. J. Am. Chem. Soc. 2010, 132 (3), 935−937. (f) Pan, X.; Fan, Z.; Chen, W.; Ding, Y.; Luo, H.; Bao, X. Enhanced ethanol production inside carbon-nanotube reactors containing catalytic particles. Nat. Mater. 2007, 6 (7), 507− 511. (6) Mestl, G.; Maksimova, N. I.; Keller, N.; Roddatis, V. V.; Schlogl, R. Carbon nanofilaments in heterogeneous catalysis: An industrial application for new carbon materials? Angew. Chem., Int. Ed. 2001, 40 (11), 2066−2068. 8280

dx.doi.org/10.1021/la300607k | Langmuir 2012, 28, 8275−8280