pubs.acs.org/NanoLett
Experimental Evidence For The Nanocage Effect In Catalysis With Hollow Nanoparticles M. A. Mahmoud, F. Saira,† and M. A. El-Sayed* Laser Dynamics Laboratory, School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia ABSTRACT Five different hollow cubic nanoparticles with wall length of 75 nm were synthesized from platinum and/or palladium elements. The five nanocatalysts are pure platinum nanocages (PtNCs), pure palladium nanocages (PdNCs), Pt/Pd hollow shell-shell nanocages (NCs) (where Pd is defined as the inner shell around the cavity), Pd/Pt shell-shell NCs, and Pt-Pd alloy NCs. These are used to catalyze the reduction of 4-nitrophenol with sodium borohydride. The kinetic parameters (rate constants, activation energies, frequency factors, and entropies of activation) of each shell/shell NCs are found to be comparable to that of pure metal NCs made of the same metal coating the cavity in the shell-shell NCs. These results strongly suggest that the catalytic reaction takes place inside the cavity of the hollow nanoparticles. Because of the nanoreactor confinement effect of the hollow nanocatalysts, the frequency factors obtained from the Arrhenius plots are found to be the highest ever reported for this reduction reaction. This is the reason for enhanced rate of this reaction inside the cavity. The importance of mechanism of the homogeneous and the heterogeneous nanocatalytic reactions occurring on the external surface of a solid nanoparticle are contrasted with those occurring on the nanocavity surface. KEYWORDS Nanoshell, alloy, platinum, palladium, nanocatalysis, nanocage
A
s a result of the large surface-to-volume ratio, nanoparticles opened up the new active field of nanocatalysis.1-4 That is a rapidly growing field ever since the publication of the Science report5 on the synthesis of colloidal metallic nanoparticles of different shapes and the proposal that different shapes might catalyze different chemical reactions. As a result, the field of shape dependent catalysis became a very active one.6-9 The enhanced catalytic activity of the nanoparticles having sharp corners and edges has been demonstrated previously by our group.8-10 In addition to the dependence of catalytic properties of nanocatalyst on the shape and size,10 the electrochemical properties of the nanoparticles are sensitive to the particle size and its interface. The reduction potential shifts to more negative values as the size of metal particle approaches the nanoscale.11 The change in reduction potential could affect the catalytic properties especially if the reaction involves electron transfer processes.12 Because of the very small size of the nanoparticles, confinement effect introduces new important properties to some nanoparticles. Gold, silver, and copper nanoparticles acquire their surface plasmonic properties as a result of confining the electromagnetic fields of much larger photons of resonance frequencies.11 As a result, they acquire enhanced surface electromagnetic fields. Because of the hollow design, plasmonic nanocages and nanoframes have two different types of surfaces, and thus two coupled plasmon
surface fields (inside and outside the nanoparticle walls) that can couple to give much enhanced fields, which extend the use of their radiative properties to many sensing applications.13,14 How about the confinement of reactants within the cavity of hollow nanocatalysts? Could this lead to more efficient catalytic properties? Previously, we have studied15 the nanocatalytic photodegradation of methyl orange using gold nanocages of various sizes and different cavity volumes. We found them to be more efficient than the well-studied TiO2 catalyst. The catalysis was proposed15 to result from the cavity-confined reaction of the hydroxyl radicals produced from the reaction of the photoexcited Ag2O semiconductor on the inside surface of the cavity with water (the Ag2O was formed from the oxidization of the left unexchanged Ag atoms on the internal surface of the gold nanocage formed from the galvanic replacement technique).15 The catalytic activity was found to increase and then decrease as the nanocavity increases in size. This behavior was proposed to result from the interplay between the size effect and the importance of the role played by the pore size on the cavity walls. One of the useful applications of hollow nanoparticles in the nanocatalysis field is the presence of two types of surfaces that increase the surface area available for the catalytic reaction.16 Recently, the reactions involving oxygen reduction and methanol oxidation in hollow Pt nanospheres have shown higher catalytic activity than on Pt solid nanoparticles. Furthermore, Pt cubic nanoboxes are found to have 1.5 times more activity than hollow Pt nanospheres for the same catalytic reaction.16 Recently, Zeng et al.17 have compared the catalytic activity of Au nanocages, partially hollow nanoboxes, and solid nanocubes in the reduction
* To whom correspondence should be addressed. E-mail:
[email protected]. † Permanent address: Department of Chemistry, Quaid-i-Azam University, 45320 Islamabad, Pakistan. Received for review: 07/16/2010 Published on Web: 08/11/2010
© 2010 American Chemical Society
3764
DOI: 10.1021/nl102497u | Nano Lett. 2010, 10, 3764–3769
reaction of 4-nitrophenol (4NP) by NaBH4 (SBH) and found that Au nanocages exhibited the highest catalytic activity.17 Reduction of 4-nitrophenol by NaBH4 is a thermodynamically allowed reaction, since the reduction potential of 4NP to 4-aminophenol (4AP) is 0.76 V, while for borate-borohydride (H3BO3/BH4-) it is -1.33 V; both potentials were measured versus NHE. This reaction is kinetically very slow in the absence of a catalyst (it takes two days).18 This reduction reaction proceeds in the presence of colloidal nanoparticles as a catalyst.19,20 Therefore, the reduction of 4NP to 4AP using SBH has become a model reaction for the evaluation of catalytic activity of metals (Au, Ag, Pt, Pd, and Ni).17,20,21 In this letter, we prepared five nanocatalysts with cubic hollow structures of Pt and Pd elements of comparable sizes. Two are synthesized of pure elements, two are of the shell/ shell NCs structure, and one has a Pt-Pd alloy NCs walls. These are used to catalyze the reduction of 4NP by SBH. A comparison is made of the reaction rates and the values of different kinetic parameters (such as the rate constants at room temperature, frequency factors, activation energies, and entropies of activation) in the presence of different types of catalysts. It is found that all the kinetic parameters of the catalysis by the shell/shell NCs are comparable to those by the pure hollow nanoparticles of the same metal covering their cavity walls (inner shells). This strongly supports the proposal that catalytic reaction with hollow nanoparticles takes place in the confined volume of the cavity. A brief discussion is given contrasting the importance of heterogeneous and homogeneous catalysis mechanisms of reactions catalyzed on the external surface of solid nanoparticles with thosecatalyzedonthecavitysurfaceofthehollownanoparticles. Experimental Section. Five nanocages were prepared from silver nanocubes by galvanic replacement of silver atoms by platinum and/or palladium atoms.22-25 Silver nanocubes were prepared by heating 35 mL of ethylene glycol (EG) at 150 °C for 1 h, followed by the addition of a solution of 0.25 g polyvinyl pyrrolidone (PVP) (molecular weight of ∼55 000 g) dissolved in 5 mL EG. After 5 min, 0.4 mL sodium sulfide (3 mM) dissolved in EG was added. AgNCs produced after 15 min from adding 0.12 g AgNO3 dissolved in 5 mL of EG. AgNCs cleaned from EG solvent and excess PVP by dilution with water-acetone mixture and centrifugation at 14 000 rpm for 10 min. The AgNCs precipitated down and dispersed in 20 mL of deionized water (DI). To prepare the nanocage catalysts, 3 mL of AgNCs solution was diluted with 20 mL of DI water, then potassium tetrachloropalatinate (II) (0.05 g per 10 mL DI water) or potassium tetrachloropalladate (II) (0.02 g per 10 mL DI water) were added (1 mL by 1 mL) every 5 min with constant shaking. The Pt or Pd atoms were deposited first on the edges of the silver nanocube. However each two silver atoms are oxidized by one Pt or Pd ion, as the amount of Pt or Pd ions increases, the size of the AgNCs template shrinks from inside generating a larger cavity inside the nanocage. © 2010 American Chemical Society
Platinum nanocages (PtNCs) were synthesized by adding 7.5 mL of potassium tetrachloropalatinate (IV) to the AgNCs solution. While PdNCs were prepared by mixing of 5 mL potassium tetrachloropalladate (II) with the AgNCs solution. The Pd/Pt shell/shell NCs were prepared by adding 3 mL potassium tetrachloropalladate (II) to the AgNCs, the resulting solution was shaken for 30 min until a palladium layer builds on the outer surface of AgNC, then a 3 mL potassium tetrachloropalatinate (II) was injected, which replaced the rest of the silver atoms inside the nanocage forming the inner shell of Pt. On the other hand, the Pt/Pd shell/shell NCs were prepared in the opposite way. The 3 mL of platinum salt was added first, followed by 3 mL palladium salt. Ultimately, the Pt-Pd alloys NCs were prepared by adding 3 mL of platinum salt at once, and then the palladium salt was added, 0.5 mL by 0.5 mL every 10 min, with gentle shaking. The reason for the careful addition of the Pd salt is that the reaction of Pd ions with silver atoms is much faster than the reaction of Pt ions with silver atoms. The prepared nanocatalyst was cleaned from the byproduct before use in the reaction; the cleaning involves 30 min sonication of the catalyst solution, and then the solution was left for two days. The pure catalyst solution was decanted from AgCl, which precipitated out. The nanocatalysts precipitated down by centrifugation at 14 000 rpm and dispersed on 20 mL DI water; the resulting solution was sonicated for 30 min and left for a day to precipitate AgCl out (if any still left). Finally, the nanocatalysts decanted from the precipitate, precipitated-out by centrifugation, and then dispersed in 20 mL DI. The catalysis experiments were carried out in a 4.5 mL long neck quartz cuvette. Two hundred and fifty microliters (2 mM) aqueous solution of 4-nitrophenol was diluted with 1.6 mL DI water, and then 200 µL cleaned colloidal catalyst solution after 100 times dilution. The resulting solution was shaken gently and 2 mL of ice cold (0.06 M) sodium borohydride was added at once. The reduction in the optical absorption peak of 4-nitrophenol was determined from the UV-vis spectrum by using an Ocean optics HR4000Cg-UVNIR. A JEOL 100C transmission electron microscope (TEM) was used to characterize the colloidal nanocages synthesized. Thermo Scientific K-alpha X-ray photoelectron spectroscopy (XPS) with an Al anode used for the elemental analysis of the nanocatalyst. Results and Discussion. Imaging of the Nanocatalysts. In addition to the shape and size of the nanocatalyst, morphology also governs its chemical activity. However, for hollow-shaped nanocatalyst, the inside cavity size and the pore size on their walls are proposed to control the activity of the nanocatalyst15 in the photodegradation of methyl orange. To compare the catalytic properties of the prepared nanocatalysts, their shape, size, and morphology should be studied. Figure 1A-E shows the TEM images of the prepared NCs catalysts. All the catalysts have cubic hollow structure with outer wall length of 75 nm. From the panels of Figure 3765
DOI: 10.1021/nl102497u | Nano Lett. 2010, 10, 3764-–3769
FIGURE 1. TEM images of the different nanocages (A) PtNCs, (B) Pd/Pt shell-shell NCs, (C) PdNCs, (D) Pt/Pd shell-shell, and (E) Pt-Pd alloy NCs.
1A-E, the cavity sizes were determined to be 52, 51, 53, 57, and 47 nm for PtNCs, Pd/Pt shell-shell NCs, PdNCs, Pt/ Pd shell-shell NCs, and the Pt-Pd alloy NCs, respectively. The XPS elemental analysis (Supporting Information Figure S1) shows that the ratio between Pt to Pd atoms is 88:11.55, 91.40:6.19, and 93.81:8.60% for Pd/Pt shell-shell NCs, Pt/ Pd shell-shell NCs, and Pt-Pd alloy NCs, respectively. A tracer of silver was also observed. Kinetics of the Catalytic Reaction. The rate of a chemical reaction depends on factors like concentration of reacting materials, temperature at which the reaction is carried-out, and the surface area of the catalyst if the reaction is catalyzed. The rate of the reaction generally increases with increasing any of these parameters. For nanocatalysis, the rate of reaction also depends on the shape, size, and the activity of nanocatalyst material. The reduction of 4NP with SBH in solution has been shown to be catalyzed with both platinum and palladium.19-21 To compare the catalytic properties of the prepared five different hollow NCs nanocatalysts, we studied the efficiency of these catalysts in catalyzing the above reduction reaction. The concentration ratio of SBH to 4NP was kept high during © 2010 American Chemical Society
the reaction to ensure that the reaction follows pseudo first order kinetics for 4NP. The rate of the reaction was calculated from the decrease in the concentration of 4NP from its UV-vis optical spectrum. 4NP has an optical absorption peak at ∼400 nm in basic solutions. Figure 2A-E shows the linear relationship between the natural logarithm of the absorption peak intensity at 400 nm and time in minutes. The slope of the straight line gives the rate constant of the catalytic reaction. The rate constants of the reaction carried out at temperatures of 25, 30, 35, and 40 °C were determined in the presence of different catalysts and are given in Table 1. Using these values, the activation energies are calculated from the plots shown in Figure 3A. Figure 3B is a plot of the activation energies versus the frequency factors of the reaction catalyzed with five different catalysts used. This clearly shows that the compensation law holds and probably the reaction mechanism is independent of the catalyst. Table 2 gives a comparison of rate constants at 25 °C, activation energies, entropies of activation, and frequency factorsforthisreactionwhencatalyzedbydifferentnanocages. 3766
DOI: 10.1021/nl102497u | Nano Lett. 2010, 10, 3764-–3769
FIGURE 2. The rate constants of the catalytic reduction reaction of 4-nitrophenol with sodium borohydride catalyzed by (A) PtNCs, (B) Pd/Pt NCs, (C) PdNCs, (D) Pt/Pd NCs, and (E) Pt-Pd alloy NCs. TABLE 1. Comparison of the Rate Constants at Different Temperatures for Pt-Pd Alloy NCs, PdNCs, Pt/Pd NCs, Pd/Pt NCs, and PtNCs catalyst Pt-Pd alloy NCs PdNCs Pt/PdNCs Pd/PtNCs PtNCs
25 °C (min-1)
30 °C (min-1) -4
-0.0133 ( 1.1 × 10 -0.019 ( 8.7 × 10-4 -0.019 ( 2.0 × 10-4 -0.0035 ( 1.3 × 10-4 -0.0036 ( 2.0 × 10-4
-0.0270 ( 7.3 × 10 -0.031 ( 5 × 10-4 -0.0270 ( 7.0 × 10-4 -0.0051 ( 1.6 × 10-4 -0.0048 ( 1.0 × 10-4
From this comparison, the following conclusions can be made: (1) PdNCs and Pt/Pd NCs (which have Pd covering © 2010 American Chemical Society
35 °C (min-1) -4
40 °C (min-1) -3
-0.0612 ( 2.0 × 10 -0.076 ( 1.0 × 10-4 -0.050 ( 1.0 × 10-3 -0.0082 ( 4.2 × 10-4 -0.008 ( 1.4 × 10-4
-0.099 ( 3.0 × 10-3 -0.100 ( 3.0 × 10-4 -0.095 ( 1.6 × 10-3 -0.015 ( 5.0 × 10-4 -0.0125 ( 3.7 × 10-4
their cavity surfaces, making inner shell) have similar kinetic parameters; (2) PtNCs and Pd/Pt shell-shell NCs (which have 3767
DOI: 10.1021/nl102497u | Nano Lett. 2010, 10, 3764-–3769
FIGURE 3. (A) Arrhenius plots for the reduction reaction of 4-nitrophenol with sodium borohydride carried out at temperatures of 25, 30, 35, and 40 °C, and catalyzed by PtNCs (black), Pd/Pt NCs (red), Pt/Pd NCs (green), PdNCs (blue), and Pt-Pd alloy NCs (cyan). (B) The compensation law plot of the frequency factors and the activation energy obtained from the slopes and intercepts of the Arrhenius plot. The straight line confirms that the reaction belongs to the compensation law. TABLE 2. Comparison of the Rate Constants at Room Temperature, Activation Energies, Entropies of Activation, and Frequency Factors for the Reduction of 4NP with SBH When Catalyzed with Pt-Pd Alloy NCs, PdNCs, Pt/Pd NCs, Pd/Pt NCs, and PtNCs nanocatalyst
cavity size (nm)
rate constant at 25 °C (min-1)
activation energy (kcal/mol)
entropy of activation (cal/mol.K)
frequency factor (min-1)
Pt-Pd alloy NCs PdNCs Pt/Pd NCs Pd/Pt NCs PtNCs
47 57 51 53 52
-0.0133 ( 1.1 × 10-4 -0.019 ( 8.7 × 10-4 -0.019 ( 2.0 × 10-4 -0.0035 ( 1.3 × 10-4 -0.0036 ( 2.0 × 10-4
26.2 ( 1.8 22.6 ( 1.5 18.5 ( 1.3 20.7 ( 1.8 16.2 ( 1.1
79.4 ( 6.0 67.8 ( 5.0 50.4 ( 4.2 61.4 ( 6.0 43.2 ( 3.6
1.76 × 1017 5.10 × 1014 8.80 × 109 2.13 × 1013 2.31 × 109
Pt covering their cavity surfaces) have similar kinetic parameters; (3) from observations 1 and 2, one concludes that the catalytic reaction is taking place within the cavity confines; (4) platinum on the nanometer scale is a better catalyst for this reaction than palladium as observed on the metallic surface; (5) the frequency factors reported in Table 2 are the highest ever reported for this reaction. Values reported in the literature for this reaction are 2 × 102, 1.3 × 105, 2.3 × 107, and 1.0 × 109 for Pt nanocubes,21 gold nanocages,17 gold nanoboxes17 and gold partially hollow nanoboxes,17 respectively; this again confirms the confined nature of the cavity mechanism in hollow nanoparticles. Contrasting the Mechanisms Involved in Hollow and on Solid Nanoparticles. The question is often raised about the mechanism involved in nanocatalysis. In the heterogeneous mechanism, the catalysis occurs on the nanoparticle surface.4 In the homogeneous mechanism, the catalysis takes place in solution by complexes involving dissolved metal atoms or ions of the nanocatalyst material. The stability of the metallic atoms on the nanosurface is thus important in determining whether the mechanism is homogeneous or heterogeneous. If the surface atoms are stabilized by complexing with the reactants (or even with solvent molecules), they are most likely to dissolve into the solution and could become active in homogeneous type catalysis.12 This is also the explanation for the observations that active catalysts are those with structure having many corners and edges.8 This is because atoms on corners/edges have unsat© 2010 American Chemical Society
urated valency with less number of bonds around them than those in the interiors or on the faces. These situations are expected to be found on the exterior surfaces of nonspherical solid nanoparticles. Cavity surfaces of hollow nanoparticles do not have sharp corners or edges (unless at surface defects). In a smooth nanocavity, atoms on the corners are better protected than on faces of nanocages. This discussion might then suggest that more of the smooth cavity nanocatalysis reactions most likely involve heterogeneous, rather than homogeneous, type mechanisms. Possible Conclusion. The reduction of 4-nitrophenol by sodium borohydride was used as a test reaction for the proposal, that the reactions catalyzed with hollow nanoparticles (nanocavity) occur inside the cavity. The rate constants of the catalyzed reaction at room temperature, activation energies, entropies of activation, and frequency factors of the reactions catalyzed by the Pd/Pt shell/shell NCs are found to have similar values to those of the pure metal present on the inner walls of the cavity. Furthermore, the frequency factors were found to have larger values than ever reported for this reaction or other reactions catalyzed by solid or partially hollow nanoparticles. This is most likely a result of the confinement effect of the reactions occurring within the cavity. All these results strongly support the proposal that the catalysis by hollow nanoparticles is taking place within the cavity. A discussion was given to suggest that heter3768
DOI: 10.1021/nl102497u | Nano Lett. 2010, 10, 3764-–3769
ogonous type mechanisms are most likely to be dominant in cavity type nanocatalysis if the cavity surface is not defected.
(10) Mahmoud, M. A.; Tabor, C. E.; El-Sayed, M. A.; Ding, Y.; Wang, Z. L. J. Am. Chem. Soc. 2008, 130, 4590–4591. (11) Henglein, A. Chem. Rev. 1989, 89, 1861–73. (12) Mahmoud, M. A.; El-Sayed, M. A. J. Phys. Chem. C 2007, 111, 17180–17183. (13) Mahmoud, M. A.; El-Sayed, M. A. Nano Lett. 2009, 9, 3025–3031. (14) Mahmoud, M. A.; Snyder, B.; El-Sayed, M. A. J. Phys. Chem. C 2010, 114, 7436–7443. (15) Yen, C. W.; Mahmoud, M. A.; El-Sayed, M. A. J. Phys. Chem. A 2009, 113, 4340–4345. (16) Peng, Z. M.; You, H. J.; Wu, J. B.; Yang, H. Nano Lett. 2010, 10, 1492–1496. (17) Zeng, J.; Zhang, Q.; Chen, J. Y.; Xia, Y. N. Nano Lett. 2010, 10, 30–35. (18) Saha, S.; Pal, A.; Kundu, S.; Basu, S.; Pal, T. Langmuir 2010, 26, 2885–2893. (19) Pradhan, N.; Pal, A.; Pal, T. Colloids Surf., A 2002, 196, 247–257. (20) Ghosh, S. K.; Mandal, M.; Kundu, S.; Nath, S.; Pal, T. Appl. Catal., A 2004, 268, 61–66. (21) Mahmoud, M. A.; Snyder, B.; El-Sayed, M. A. J. Phys. Chem. Lett. 2010, 1, 28–31. (22) Xiong, Y.; Cai, H.; Wiley, B. J.; Wang, J.; Kim, M. J.; Xia, Y. J. Am. Chem. Soc. 2007, 129, 3665–3675. (23) Xiong, Y.; Wiley, B.; Chen, J.; Li, Z.-Y.; Yin, Y.; Xia, Y. Angew. Chem., Int. Ed. 2005, 44, 7913–7917. (24) Sun, Y.; Mayers, B.; Xia, Y. Adv. Mater. (Weinheim, Ger.) 2003, 15, 641–646. (25) Chen, J.; Wiley, B.; McLellan, J.; Xiong, Y.; Li, Z.-Y.; Xia, Y. Nano Lett. 2005, 5, 2058–2062.
Acknowledgment. This work was supported by the NSF Grant 0957335. Supporting Information Available. The XPS spectra shown on page S1. This material is available free of charge via the Internet at http://pubs.acs.org. REFERENCES AND NOTES (1) (2) (3) (4) (5) (6) (7) (8) (9)
Eppler, A.; Rupprechter, G.; Guczi, L.; Somorjai, G. A. J. Phys. Chem. B 1997, 101, 9973–9977. Freund, P. L.; Spiro, M. J. Chem. Soc., Faraday Trans. 1 1986, 82, 2277–82. Freund, P. L.; Spiro, M. J. Phys. Chem. 1985, 89, 1074–7. Freund, P. L.; Spiro, M. J. Chem. Soc., Faraday Trans. 1 1983, 79, 481–90. Ahmadi, T. S.; Wang, Z. L.; Green, T. C.; Henglein, A.; El-Sayed, M. A. Science (Washington, DC) 1996, 272, 1924–1926. Tian, N.; Zhou, Z.-Y.; Sun, S.-G.; Ding, Y.; Wang, Z. L. Science (Washington, DC) 2007, 316, 732–735. Burda, C.; Chen, X.; Narayanan, R.; El-Sayed, M. A. Chem. Rev. (Washington, DC) 2005, 105, 1025–1102. Narayanan, R.; El-Sayed, M. A. Nano Lett. 2004, 4, 1343–1348. Narayanan, R.; El-Sayed, M. A. J. Am. Chem. Soc. 2004, 126, 7194– 7195.
© 2010 American Chemical Society
3769
DOI: 10.1021/nl102497u | Nano Lett. 2010, 10, 3764-–3769
