Fabrication of CNTs with controlled diameters and their applications as electrocatalyst supports for DMFC

Fabrication of CNTs with controlled diameters and their applications as

electrocatalyst supports for DMFC

i

An-Ya Lo

, Ningya Yu

, Shing-Jong Huang

, Chin-Te Hung

, Shou-Heng Liu

, Zhibin Lei

,

Cheng-Tzu Kuo

⁎

, Shang-Bin Liu

⁎⁎

a

Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 10617, Taiwan b

Department of Material Science and Engineering, National Chiao Tung University, Hsingchu 30010, Taiwan c

Department of Chemistry, National Taiwan University, Taipei 10617, Taiwan d

Institute of Materials and System Engineering, Ming Dao University, Changhua 52345, Taiwan e_{Department of Chemistry, National Taiwan Normal University, Taipei 11677, Taiwan}

a b s t r a c t

a r t i c l e i n f o

Available online 19 January 2011 Keywords:

Carbon nanotube Templating synthesis Chemical vapor deposition Hexagonal mesoporous silica Electrocatalyst

Direct methanol fuel cell

A facile synthesis procedure based on chemical vapor deposition (CVD) process has been developed to fabricate carbon nanotubes (CNTs) with controlled diameters and high yields utilizing Fe-containing ordered hexagonal mesoporous silicas (HMSs) such as MCM-41 and SBA-15 having varied pore sizes as the catalysts as well as the templates. It is found that unlike Fe/HMS catalysts prepared by co-precipitation method, samples prepared by the impregnation method gave rise to multi-wall CNTs with uniform diameters, which were largely dictated by the pore size of the Fe/HMS catalysts. Among these uniform MWCNTs, sample with a larger diameter (≥8 nm) was found to be more favorable as support for Pt catalyst, leading to a homogeneous dispersion of metal nanoparticles. Consequently, the Pt/CNT electrocatalysts so prepared gave rise to superior methanol oxidation activities as well as tolerances for CO poisoning compared to Pt supported on commercial single-wall CNT (Pt/SWCNT) and XC-72 activated carbon (Pt/XC-72) having a similar metal loading.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Recent developments in fabrication of porous carbon supports with high surface areas and controllable morphologies have received considerable attention in R&D of supported electrocatalysts for direct methanol fuel cells (DMFCs) and proton-exchange membrane fuel cells (PEMFCs), which have been considered as the most prominent

candidates for next-generation portable power sources[1–4]. Highly

dispersed noble metal (Pt, Ru) nanoparticles (NPs) supported on conductive materials with high surface areas, such as carbon blacks [5,6], ordered mesoporous carbons (OMCs) [7–13], and carbon

nanotubes (CNTs)[14–18], are pertinent anodic/cathodic

electrocata-lysts for DMFCs and PEMFCs. Among them, CNTs have received

considerable attention due to their unique one-dimensional nanostruc-ture and superior electrical conductivity. Aside from the most common electrocatalysts for DMFCs, such as Pt/Ru supported on commercial Vulcan XC-72 carbon black, many attempts have been made utilizing

CNTs as catalyst supports[14–23]. However, while majority of past

research efforts have been devoted in optimizing the metal dispersion on single-wall (SW) and multi-wall (MW) CNTs aiming to promote their electrocatalytic performances and durability, practically no report had been focused on the diameter size of the CNT supports.

In general, the methodologies invoked in controlling the diameter of CNTs during chemical vapor deposition (CVD) process may be

classified into two main categories, namely by controlling the

processing parameters and by employing an auxiliary template. For the former, it has been reported that parameters such as the carrier

gas/carbon source flow rate, plasma intensity, morphology of the

catalyst, precursor compositions, and duration of treatment etc. have

considerable effects on the diameter of thefinal CNT products[24–

28]. In contrast to such sophisticated adjustment of processing

parameters, the use of an auxiliary template appears to be more advantageous in fabricating CNTs with tailorable diameters. For

examples, zeolites[29]and anodic metal oxides[30–36]have been

utilized as hard templates during the CVD process to fabricate CNTs with uniform diameters. In this case, the diameters of the CNTs so synthesized are largely dictated by the pore size of template used.

However, since it is rather difficult to prepare anodic metal oxides

i Presented at NDNC 2010, the 4th International Conference on New Diamond and Nano Carbons, Suzhou, China.

⁎ Corresponding author. Fax: +886 4 8879050.

⁎⁎ Correspondence to: S.-B. Liu, Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 10617, Taiwan. Fax: + 886 2 23620200.

E-mail addresses:kuoct@mdu.edu.tw(C.-T. Kuo),sbliu@sinica.edu.tw(S.-B. Liu). 1

Present address: Institute of Fine Catalysis and Synthesis and Key Lab of Sustainable Resources Processing and Advanced Materials of Hunan Province, Hunan Normal University, Changsha 410081, China.

2

Present address: Department of Chemical and Materials Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung 80778, Taiwan.

3

Present address: Department of Chemical & Biomolecular Engineering, National University of Singapore, Singapore.

0925-9635/$– see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2011.01.002

Contents lists available atScienceDirect

Diamond & Related Materials

with uniform pore sizes less than 25 nm, majority of the CNTs so

fabricated possess diameters exceeding 25 nm [36]. On the other

hand, CNTs fabricated by using microporous zeolites as templates mostly have diameters less than 1 nm. In view of the fact that the diameter of CNT is one of the key parameters affecting their physical properties, it is highly desirable to develop a facile synthesis route to

fabricate CNTs with tunable diameters within the range of 1–25 nm.

Ordered hexagonal mesoporous silicas (HMSs), especially those possessing straight mesoporous channels with uniform pore sizes in

the range of 2–50 nm, such as MCM-41[37]and SBA-15[38,39]seem

to represent the ideal templates to meet the aforementioned demand

[40]. It is worth mentioning that although various ordered

mesopor-ous silicas have been invoked for the preparation of CNTs, they were mostly employed as supports to disperse or to limit the size and

structure of the catalytic metal nanoparticles[41–48], overlooking the

templating function of the porous substrates. We report herein a facile synthesis route to synthesize CNTs with tunable diameters and high yields by using Fe-containing HMSs as catalyst templates. The resultant CNTs with varied diameters were used as supports to prepare various Pt/CNT anodic electrocatalysts for DMFC applications and their catalytic performances during methanol oxidation reaction (MOR) were evaluated and compared with a Pt/XC-72 catalyst (12.5 wt.% Pt on Vulcan XC-72).

2. Experimental

2.1. Preparation of Fe-containing HMSs

Three types of ordered HMSs, namely MCM-41, SBA-15, and pore expanded SBA-15 (denoted as PE-SBA-15) were synthesized by

known recipes reported previously [37–39,49]. Iron catalyst was

loaded onto the HMSs via either co-precipitation or impregnation

methods. For the former, typically ca. 0.4 g of Fe(NO3)3was stirring

with 1.0 g of the target HMS material for 0.5 h in deionized water

(20 mL), followed byfiltering and drying at 373 K, then subjected to

reduction treatment under H2 at 773 K for 3 h. The Fe-containing

HMSs so obtained from the siliceous MCM-41 and SBA-15 are denoted as Fe(co)/MCM-41 and Fe(co)/SBA-15, respectively. In the case of loading Fe catalyst by the impregnation method, proper amount of Fe

(NO3)3(ca. one-half pore volume of the corresponding support) was

dissolved in 20 mL deionized water, followed by adding 1.0 g of the target HMS. After being stirred for 0.5 h, the suspension was dried under vacuum. The obtained product was further stirred in presence of dichloromethane (CH2Cl2) to facilitate migration of Fe precursors

into the hydrophilic channels of the HMSs[50], followed by removal

of the CH2Cl2 solvent by evacuation. The above procedure was

repeated once and the final product was subjected to reduction

treatment carried out by first slowly ramping (2 K/min) the

temperature to 373 K under dried Ar, kept at the same temperature

for 3 h, followed by reduction under H2environment before a mixture

of acetylene (C2H2) and hydrogen (H2) was injected for CNTs growth. The Fe-containing HMSs so derived from the siliceous MCM-41, SBA-15, and PE-SBA-15 are denoted as Fe(im)/MCM-41, Fe(im)/SBA-SBA-15, and Fe(im)/PE-SBA-15, respectively.

2.2. Fabrication of CNTs with uniform diameters

CNTs with varied diameters were prepared by a CVD method similar to that reported earlier for the nano-sized tubular carbons (i.e., CMTs)

[51]. In brief, the syntheses were carried out in a home-made quartz

reactor using various Fe-containing HMSs as templates. Typically, after loading ca. 0.5 g of the fresh Fe-containing HMS in the reactor, the

system wasfirst gradually heated (1 K/min) to 873 K under vacuum,

followed by injecting a stream of C2H2/H2gas mixture at aflow rate of

50/50 sccm/sccm for 40 min under a pressure of ca. 2 kPa. The resultant

product was stirred with excess aqueous HF solution (1 M, 50% ethonal–

50% H2O) for 24 h to remove the silica template and Fe species, followed

byfiltering and drying under vacuum to obtain the final multi-wall

carbon nanotube (MWCNT) materials. 2.3. Preparation of Pt/CNT electrocatalysts

To explore the effect of tube diameter on the performances of various CNTs as catalyst supports for DMFC at anode, typically ca. 0.2 g of the selected home-made CNT was individually suspended in 10 mL

of H2PtCl6 aqueous solution (0.04 M) at room temperature. After

removing water under reduced pressure, the obtained solid was

treated at 523 K for 0.5 h under H2atmosphere to provoke reduction

of PtCl62−to Pt nanoparticles. For comparison purpose, two commer-cial products, namely the Vulcan XC-72 carbon black (Cabot Corp.) and single-wall CNTs (SWCNTs; Nano-C Inc.) were also adopted as supports to fabricate Pt-containing electrocatalysts with similar metal loading via the aforementioned procedures.

2.4. Materials characterization and electrocatalytic tests

Powder X-ray diffraction (XRD) patterns of various samples were

recorded on a Philips X' PERT-Pro-MPD diffractometer using the CuKα

radiation (λ=1.542 Å). All nitrogen adsorption/desorption isotherm

measurements were carried out at 77 K on a Quantachrome Autosorb-1 sorptometer. Prior to the measurement, each sample was outgassed

at 473 K under vacuum (10− 6Torr) overnight. The surface area of

each sample was derived by the BET method, and its pore size distribution (PSD) was calculated from the adsorption branch of the

isotherm using the Barrett–Joyner–Halenda (BJH) method.

Transmis-sion electron microscopy (TEM) studies were performed on a JEOL

JEM-2100 F microscope operated at 200 kV. Each sample was ultrasonicated for 15 min in ethanol before introducing it onto the carbon-coated copper grids. The metal (Fe and Pt) contents in various samples were determined by thermal gravimetric analyses (TGA; Netzsch TG-209).

The electrochemical active surfaces (EASs), electrocatalytic per-formances, and stability of various supported Pt/C catalysts were evaluated on an Autolab PGSTAT30 galvanostat/potentiostat at room

temperature by using a glassy carbon thin-film as the working

electrode, Pt wire as the counter electrode, and Ag/AgCl as the

reference electrode. Typically, the glassy carbon thin-film electrode

was prepared byfirst adding ca. 10 mg of Pt-loaded CNTs sample into

5 mL of deionized water, followed by ultrasonic treatment for 0.5 h.

Next, ca. 20μL of the resultant suspension mixture was withdrawn

and injected onto the glassy carbon electrode (diameter ca. 5 mm),

followed by drying in air at 333 K for 1 h. Finally, ca. 20_{μL of 1% Nafion}

(DuPont) solution was added as a binder under N2 environment.

Prior to each cyclic voltammetry (CV) measurement, the electrolytic

solution was purged with purified N2gas (99.9%) for at least 0.5 h to remove the dissolved oxygen. The EAS and electrooxidation of

methanol (MeOH) was carried out with an electrolyte of H2SO4

(1.0 M) and H2SO4 (0.5 M)/MeOH (1 M) solution, respectively,

between−0.2 and 1.0 V at a 10 mV/s scanning rate. Prior to each

measurement, the corresponding working electrode was scanned by

CV in H2SO4(1.0 M) at room temperature till it reached a steady state.

In addition, chronoamperometry (CA) measurements were also carried out for assessment of catalyst stability. The corresponding working electrode was measured under a constant (0.6 V) anodic

potential with an electrolyte of H2SO4(0.5 M)/MeOH (1 M) during

these experiments, Note that all of the above electrochemical

experiments were conducted under purging N2condition.

3. Results and discussion

3.1. Effect of Fe loading method for Fe/HMS on CNT diameter control Fig. 1displays the XRD patterns of the parent siliceous MCM-41 and SBA-15 mesoporous silicas as well as their corresponding Fe-containing counterparts prepared by different methods. Both the siliceous SBA-15 and MCM-41 showed three well-resolved

diffrac-tion peaks that may be indexed as (100), (110), and (200) reflections

associated with the well-ordered hexagonal arrays of mesopores

[37–39]. The method applied in loading Fe onto the mesoporous

silica appears to have signi_{ficant effects on the mesostructure of the}

Fe-containing HMSs. Clearly, the Fe(co)/MCM-41 and Fe(co)/SBA-15 samples prepared via the co-precipitation method tend to retain the ordered structures of their parent silica supports, although substan-tial decreases in intensities of the diffraction peaks were evident. On the other hand, Fe-containing HMSs prepared by the impregnation method, i.e., Fe(im)/MCM-41 and Fe(im)/SBA-15, exhibited a main

d100diffraction peak with nearly diminishing higher order (d110and

d200) peaks. Thus, unlike Fe(co)/HMSs, incorporation of Fe catalyst by impregnation method tends to lose the integrity of long-range structural ordering of the siliceous HMSs, as observed for Fe(im)/

HMSs inFig. 1. The same conclusion may be drawn from the TEM

results inFig. 2. It is worth pointing out that a parallel alignment of

pore channels should be observed when the electron beam was

introduced perpendicular to the channel axis (Fig. 2a and d),

whereas hexagonal packing of cylindrical mesopores prevails when the electron beam was introduced along the channel axis

(see insets in Fig. 2a and d). It is clear that the Fe(co)/MCM-41

(Fig. 2b) and Fe(co)/SBA-15 (Fig. 2e) samples retained the hexagonal pore systems of their respective parent silicas. However, it is also evident that the co-precipitation process invoked led to a

(a) (b)

(c) (d)

Volume adsorbed

(cm

STP/g)

Relative pressure (p/p

)

Relative pressure (p/p

)

Pore size (nm)

dV/dD (cm

g

nm

)

Pore size (nm)

dV/dD (cm

g

nm

)

Fig. 3. Nitrogen adsorption/desorption isotherms and pore size distributions of siliceous and Fe-containing HMSs (a, b) MCM-41 and (c, d) SBA-15. The isotherms of Fe(co)/MCM-41, Fe(im)/MCM-41, Fe(co)/SBA-15, and Fe(im)/SBA-15 were shifted vertically by 300, 600, 300, and 600 cm3

g− 1STP, respectively, whereas the pore size distribution curves for Fe(co)/ MCM-41, Fe(im)/MCM-41, Fe(co)/SBA-15, and Fe(im)/SBA-15 were shifted vertically by 1.5, 3.0, 0.3, and 0.6 cm3

heterogeneous dispersion of Fe particles on the external surfaces of the mesoporous silica supports.

On the other hand, while the TEM images of Fe(im)/MCM-41 (Fig. 2c) and Fe(im)/SBA-15 (Fig. 2f) seemingly showed well ordered hexagonal pore systems, incorporation of Fe onto HMSs by impreg-nation method led to partial formation of Fe nanorods within the pore channels. In this context, the diameter of these Fe nanorods should be constrained by the pore aperture of the HMS supports, leading to

substantial reductions in their scattering contrasts[52–54]. Thus, the

notable decrease in the intensity of the main d100diffraction peak and

the disappearance of the higher order peaks observed for Fe(im)/

MCM-41 and Fe(im)/SBA-15 in Fig. 1may be ascribed due to the

presence of Fe nanorods in the mesochannels of the silica supports rather than degradation of their mesostructures.

Nitrogen adsorption/desorption isotherms obtained from

vari-ous samples are shown inFig. 3together with their corresponding

PSDs. The textural properties of various samples were also depicted in Table 1. Comparing with their parent silica materials, Fe-containing HMSs typically showed smaller pore volumes and broader PSDs regardless of the method adopted in incorporating the Fe catalyst. While the Fe(co)/MCM-41 and Fe(co)/SBA-15 samples showed similar pore sizes (DBJH) compared to their respective silica supports, slight decreases in pore volumes (Vp) and BET surface areas (SBET) were observed after loading the Fe

catalyst (Table 1). On the other hand, notable decreases in DBJH, Vp,

and SBETwere observed for the Fe(im)/MCM-41 and Fe(im)/SBA-15

compared to their respective siliceous counterparts before loading

the Fe catalyst (Table 1). These observations are in line with the

aforediscussed XRD and TEM results, that is, incorporation of the Fe catalyst by impregnation method led to the formation of Fe nanorods within the pore channels of the HMSs, whereas samples prepared using the co-precipitation method resulted in an inhomogeneous dispersion of Fe nanoparticles (typically, ca. 20 nm in size) on the external surfaces of the HMSs.

Typical TEM micrographs of CNTs synthesized by CVD process

using various Fe-containing HMSs as catalysts are depicted inFig. 4. In

the cases of using the Fe(co)/MCM-41 and Fe(co)/SBA-15 supported catalysts, the obtained CNTs typically possessed an average diameter

greater than 5 (Fig. 4a) and 10 nm (Fig. 4c), respectively, which were

apparently greater than the pore sizes of their corresponding catalyst

templates (2.8 and 9.2 nm, respectively; seeTable 1). These

observa-tions are in parallel to the existing literature reports[42,43,47]. In

contrast, the CNTs synthesized using the Fe(im)/MCM-41 and Fe(im)/ SBA-15 catalysts were found to exhibit a uniform diameter of ca. 3 (Fig. 4b) and 8 (Fig. 4d) nm, respectively, in close resemblance with the average pore sizes of the corresponding Fe-containing HMSs (2.6

and 8.6 nm, respectively; seeTable 1). The above results indicate that

Fe-containing HMSs so designed and prepared can be employed not only as catalysts to fabricate CNTs but also as templates to manipulate the diameters of the synthesized CNTs. To further verify these points, an additional PE-SBA-15 sample with expanded pore size of 17.5 nm (Table 1 and Fig. 5a) was synthesized. After incorporating the Fe

catalyst by impregnation method, the resultant Fe(im)/PE-SBA-15 sample was employed as catalyst (and template) to produce CNTs. As

shown in Fig. 5b, the CNTs so fabricated also possessed a rather

uniform diameter (ca. 17 nm) comparable to the pore size of the Fe (im)/PE-SBA-15 catalyst (17.5 nm). It is noteworthy that the CNTs produced by CVD process using the Fe-containing HMSs reported

herein also exhibited superior high yields in terms of atom efficiency.

For examples, a CNT yield as high as 0.3 g was obtained during a CVD duration time of 20 min using ca. 1.0 g of Fe(im)/SBA-15 as catalyst/ template. Likewise, a respective CNT yield of 0.1 and 0.4 g/20 min were attained when Fe(im)/MCM-41 and Fe(im)/PE-SBA-15 were employed. 3.2. Effect of CNT diameter on Pt dispersion and electrochemical properties

As shown above, the MWCNTs synthesized by using the Fe(im)/ MCM-41, Fe(im)/SBA-15, and Fe(im)/PE-SBA-15 catalyst possess an

Fig. 4. TEM images of CNTs prepared using various Fe-containing HMSs; (a) Fe(co)/ MCM-41, (b) Fe(im)/MCM-41, (c) Fe(co)/SBA-15, and (d) Fe(im)/SBA-15.

Table 1

Textural properties of various siliceous and Fe-containing HMSs.a

Sample d100(nm) DBJH(nm) SBET(m2/g) Vp(cm3/g) MCM-41 4.0 2.9 1153 0.97 Fe(co)/MCM-41 4.0 2.8 949 0.79 Fe(im)/MCM-41 4.2 2.6 540 0.50 SBA-15 9.4 9.1 823 1.30 Fe(co)/SBA-15 9.4 9.2 739 1.27 Fe(im)/SBA-15 8.7 8.6 498 0.64 PE-SBA-15 – 18.0 709 1.64 Fe(im)/PE-SBA-15 – 17.5 464 0.63 a

d100: d spacing measured in the (100) plane; DBJH: pore diameter derived by BJH method; SBET: BET surface area; Vp: pore volume.

Fig. 6. (Left) TEM images and (Right) Pt particle size distribution of various Pt/C electrocatalysts; (a) Pt/CNT-d3, (b) Pt/CNT-d8, (c) Pt/CNT-d17, (d) Pt/SWCNT, and (e) Pt/XC-72. Fig. 5. (a) Pore size distribution of Fe(im)/PE-SBA-15 and (b) the TEM image of the resultant CNTs.

average diameter of ca. 3, 8, and 17 nm, respectively. These CNTs with uniform diameters, together with a commercial SWCNT were subsequently used as supports for Pt catalyst, the resultant supported Pt/C catalyst samples are donated as d3, d8, Pt/CNT-d17, and Pt/SWCNT, respectively. The mean Pt particle sizes of these Pt/C samples as well as Pt/XC-72 were preliminary estimated by

well-known Scherrer formula[55,56]based on the large-angle XRD (220)

diffraction peak of Pt metal (not shown) occurring at 2θ of ca. 68°.

Accordingly, a similar Pt particle size of ca. 2.4 nm were observed for the Pt/CNT-d17, Pt/CNT-d8, and Pt/XC-72 samples, whereas slightly

larger Pt nanoparticles (N3 nm) was found for the Pt/CNT-d3 and the

Pt/SWCNT samples. The Pt particle sizes in different samples were

further evaluated by TEM analyses.Fig. 6displays the resultant TEM

images of various Pt/C catalysts and their corresponding histograms of Pt particle size distribution. Clearly, majority of the Pt/C samples with CNTs as supports show Pt particle size predominately in the range of

1–2 nm. However, unlike the Pt/CNT-d8 (Fig. 6b) and the Pt/CNT-d17

(Fig. 6c) samples, which showed uniform Pt dispersions typically with

Pt particle size_{≤2 nm, the Pt/CNT-d3 sample appeared to have lower}

Pt dispersion similar to that of the Pt/SWCNT and the Pt/XC-72

samples. As revealed byFig. 6a, d, and e, Pt particles exceeding 10 nm

may be identified in the latter three samples. A closer examination of

the TEM profile for the Pt/CNT-d3 sample revealed that the CNTs tend

to bundle together, leading to an inhomogeneous dispersion of Pt

nanoparticles (see inset inFig. 6a). Thus, it is indicative that, in terms

of Pt dispersion, CNT supports with larger diameters (preferably exceeding 3 nm) are loath to bundling and hence are more preferable as supports for the dispersion of metal catalyst, as evidenced by the

particle size distribution profiles inFig. 6.

The electrocatalytic performances of various Pt/C catalysts during

MOR are shown inFig. 7. All Pt/C samples displayed voltammograms

associated with forward (If) and backward (Ib) anodic peak current

densities within 0.4–0.7 V, which represent mass activity of catalyst

and resistance toward catalyst deactivation (by coking) over the

catalyst during MOR, respectively [11,12,57,58]. For comparison

purpose, the observed Ifand Ibpeak values are display in Table 2

along with the corresponding If/Ibratio for various electrocatalysts.

That the Pt/CNT-d3, Pt/CNT-d8, and Pt/CNT-d17 samples exhibited If/

Ibratios exceeding that of the Pt/SWCNT and the Pt/XC-72 indicates

that Pt/CNT electrocatalysts with uniform diameters have higher

electrocatalytic activities albeit all Pt/C samples have rather similar Pt

loading as well as degree of graphitization (Table 2). The latter being

inferred by the relative peak intensities of the D- and G-bands (i.e., the

ID/IG ratio) derived from their respective Raman spectra.

Electro-catalysts with a smaller ID/IGvalue thus have a better graphitization

degree (i.e., more carbon with sp3 _{than sp}2 _{structure), and}

presumably should have a better electrical conductivity. In other words, in view of the fact that all Pt/C samples have similar Pt loading

and ID/IG ratios, the mass activity observed for various Pt/C

electrocatalysts are mainly due to their diversities in Pt dispersion rather than the electrical characteristics (e.g., conductivities) of the carbon supports.

The EAS values obtained for various Pt/C catalysts are also

summarized inTable 2. These EAS values were derived based on the

following equation:

EAS = QH

½Pt × 0:21

where [Pt] represents the Pt loading (in unit of mg/cm3_{) in the}

electrode, the value 0.21 represents the charge required to oxidize a

monolayer of H2on a fresh Pt surface[56], and QHis the Coulombic

charge for hydrogen sorption in 1.0 M H2SO4 solution at room

temperature between−0.2 and 1.0 V at a scanning rate of 10 mV/s.

That higher EAS values were observed for the Pt/CNT-d3, Pt/CNT-d8, and Pt/CNT-d17 compared to the Pt/SWCNT and the Pt/Xc-72 samples indicates that these Pt/CNT catalysts with uniform diameters of MWCNT supports not only possess higher electrocatalytic perfor-mances during MOR but also better Pt dispersion, in line with the results obtained from TEM analyses. Indeed, it is known that a high dispersion and a narrow particle size distribution of noble metal nanoparticles are prerequisite in ensuring a high electrocatalytic

performance of the catalyst[59]. Note that the If/Ibratios obtained

from Pt/CNTs appear to increase with increasing diameter of the CNT

supports (Table 2), indicating that CNTs with larger diameters are not

only more favorable for dispersion of Pt nanoparticles with uniform sizes, but also less vulnerable to deactivation and more tolerable towards CO poisoning during MOR.

Finally, the stability and durability of various Pt/C electrocatalysts were accessed by CA measurements under prolonged operation time (Fig. 8). The CA curves of various electrocatalysts were recorded with

aqueous solution (1 M MeOH+ 0.5 M H2SO4) at afixed anodic potential

of 0.6 V[60,61]. Typically, all electrocatalysts exhibit a rapid decrease in

electrochemical activity during the initial period, followed by a gradual decrease with prolonged operation. Among them, the Pt/CNT-d8 and the Pt/CNT-d17 catalysts show the anticipated higher electrocatalytic

-0.2 0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 Pt/SWCNT Pt/CNT-d3 Pt/CNT-d8 Pt/CNT-d17

Mass Activity (mA/mg Pt)

Pt/XC-72

E (V) vs. Ag/AgCl

Fig. 7. Cyclic voltammograms of methanol oxidation for various Pt/C electrocatalysts.

Table 2

Physicochemical properties and catalytic performances of various Pt/C electrocatalysts. Catalyst Metal content (wt.%)a ID/IGb Ifc (A/g of Pt) Irc (A/g of Pt) If/Ir EAS (m2_/g Pt) Pt particle sized (nm) Fe Pt Pt/CNT-d3 2.6 12.9 1.25 308 294 1.20 60 3.7 Pt/CNT-d8 4.8 9.3 1.25 413 321 1.22 68 2.4 Pt/CNT-d17 5.2 10.5 1.29 416 316 1.31 75 2.4 Pt/SWCNT 4.5 12.9 1.29 298 262 1.18 52 3.2 Pt/XC-72 – 12.5 1.39 353 349 1.01 47 2.5 a_{Pt loading measured by TGA analysis.}

b _{Peak intensity ratio of D-band (sp}2_{carbon) vs. G-band (sp}3_{carbon) obtained from} the Raman spectrum.

c

Maximum current density of the forward (Ir) and reversed (Ir) scans during CV analyses.

d

Pt particle size estimated by Scherrer formula[55]based on the large-angle XRD (220) diffraction peak of Pt metal at 2θ–68°.

activity than the Pt/XC-72, in excellent agreement with the results obtained from EAS and CV measurements.

4. Conclusions

In summary, a series of Fe-containing hexagonal mesoporous silicas with varied pore sizes, which were prepared either by co-precipitation or impregnation method, have been employed to catalyze formation of CNTs via CVD process. It is found that by

properly confining the metal catalyst in the mesoporous channels of

the silica supports, multi-wall CNTs with uniform and controllable diameters may be facilely synthesized with high yield. The diameter of the MWCNTs is found to play a crucial role in the dispersion of Pt nanoparticles and the electrocatalytic performances of the resulting supported Pt/CNT catalysts during methanol oxidation reaction. These Pt/MWCNTs electrocatalysts, particularly those with diameter ex-ceeding ca. 8 nm were found to exhibit electrocatalytic performances surpassing that of the conventional Pt/XC-72 and Pt/SWCNT catalysts having a similar Pt loading and electrochemical properties of the carbon supports.

Acknowledgment

Thefinancial support of this work by the National Science Council,

Taiwan (NSC98-2113-M-001-007-MY3 to SBL) is gratefully acknowledged.

References

[1] A. Roucoux, J. Schulz, H. Patin, Chem. Rev. 102 (2002) 3757.

[2] J.M. Thomas, B.F. Johnson, R. Raja, G. Sankar, P.A. Midgley, Acc. Chem. Res. 36 (2003) 20.

[3] W.M. Chen, G.Q. Sun, Z.X. Liang, Q. Mao, H.Q. Li, G.X. Wang, Q. Xin, H. Chang, C.H. Pak, D. Seung, J. Power Sources 160 (2006) 933.

[4] H. Chang, S.H. Joo, C. Pak, J. Mater. Chem. 17 (2007) 3078.

[5] P.V. Shanahan, L.B. Xu, C.D. Liang, M. Waje, S. Dai, Y.S. Yan, J. Power Sources 185 (2008) 423.

[6] F.J. Nores-Pondal, I.M.J. Vilella, H. Troiani, M. Granada, S.R. de Miguel, O.A. Scelza, H.R. Corti, Inter. J. Hydrogen Energy 34 (2009) 8193.

[7] H. Chang, S.H. Joo, C.H. Pak, J. Mater. Chem. 17 (2007) 3078.

[8] H.T. Kim, D.Y. You, H.K. Yoon, S.H. Joo, C.H. Pak, H. Chang, I.S. Song, J. Power Sources 180 (2008) 724.

[9] B. Fang, J.H. Kim, C. Lee, J.S. Yu, J. Phys. Chem. C 112 (2008) 639.

[10] E.P. Ambrosio, C. Francia, M. Manzoli, N. Penazzi, P. Spinelli, Inter. J. Hydrogen Energy 33 (2008) 3142.

[11] S.H. Liu, R.F. Lu, S.J. Huang, A.Y. Lo, S.H. Chien, S.B. Liu, Chem. Commun. 32 (2006) 3435.

[12] S.H. Liu, W.Y. Yu, C.H. Chen, A.Y. Lo, B.J. Hwang, S.H. Chien, S.B. Liu, Chem. Mater. 20 (2008) 1622.

[13] S.H. Liu, C.C. Chiang, M.T. Wu, S.B. Liu, Inter. J. Hydrogen Energy 35 (2010) 8149. [14] X. Wang, W.Z. Li, Z.W. Chen, M. Waje, Y.S. Yan, J. Power Sources 158 (2006) 154. [15] S.J. Guo, S.J. Dong, E. Wang, J. Phys. Chem. C 112 (2008) 2389.

[16] R.I. Jafri, N. Sujatha, N. Rajalakshmi, S. Ramaprabhu, Inter. J. Hydrogen Energy 34 (2009) 6371.

[17] H.Y. Du, Y.T. Tsai, C.P. Chen, C.J. Huang, L.C. Chen, K.H. Chen, H.C. Shih, J. Power Sources 171 (2007) 55.

[18] H.Y. Du, C.H. Wang, H.C. Hsu, S.T. Chang, U.S. Chen, S.C. Yen, L.C. Chen, H.C. Shih, K.H. Chen, Diamond Relat. Mater. 17 (2008) 535.

[19] Y. Mu, H. Liang, J. Hu, L. Jiang, L. Wan, J. Phys. Chem. B 109 (2005) 22212. [20] W.Z. Li, C.H. Liang, W.J. Zhou, J.S. Qiu, Z.H. Zhou, G.Q. Sun, Q. Xin, J. Phys. Chem. B

107 (2003) 6292.

[21] J. Kong, M. Chapline, H. Dai, Adv. Mater. 13 (2001) 1384.

[22] H.C. Choi, M. Shim, S. Bangsaruntip, H. Dai, J. Am. Chem. Soc. 124 (2002) 9058. [23] T. Nelson, K. Vinodgopal, G.G. Kumar, P. Kamat, Electrochem. Soc. Proc. 12 (2004)

152.

[24] C.J. Lee, S.C. Lyu, Y.R. Cho, J.H. Lee, K.I. Cho, Chem. Phys. Lett. 341 (2001) 245. [25] C.H. Kuo, A. Bai, C.H. Huang, Y.Y. Li, C.C. Hu, C.C. Chen, Carbon 43 (2005) 2760. [26] Y.C. Choi, Y.M. Ghin, Y.H. Lee, B.S. Lee, G.S. Park, W.B. Choi, N.S. Lee, J.M. Kim, Appl.

Phys. Lett. 76 (2000) 2367.

[27] Y.Y. Wei, G. Eres, V.I. Merkulov, D.H. Lowndes, Appl. Phys. Lett. 78 (2001) 1394. [28] W.Z. Li, D.Z. Wang, S.X. Yang, J.G. Wen, Z.F. Ren, Chem. Phys. Lett. 335 (2001) 141. [29] N. Wang, Z.K. Tang, G.D. Li, J.S. Chen, Science 408 (2000) 50.

[30] N.D. Hoa, N.V. Quy, Y. Cho, D. Kim, Sens. Actuators B 127 (2007) 447. [31] T. Kyotani, L.F. Tsai, A. Tomita, Chem. Mater. 8 (1996) 2109. [32] I. Eswaramoorthi, L.P. Hwang, Diamond Relat. Mater. 6 (2007) 1571. [33] S.H. Jeong, H.Y. Hwang, S.K. Hwang, K.H. Lee, Carbon 42 (2004) 2073. [34] W.S. Im, Y.S. Cho, G.S. Choi, F.C. Yu, D.J. Kim, Diamond Relat. Mater. 13 (2004)

1214.

[35] I. Eswaramoorthi, L.P. Hwang, Carbon 44 (2006) 2341. [36] B.D. Yao, N. Wang, J. Phys. Chem. B 105 (2001) 11395.

[37] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T.W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins, J.L. Schlenker, J. Am. Chem. Soc. 114 (1992) 10834.

[38] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548.

[39] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120 (1998) 6024.

[40] Y. Wan, D. Zhao, Chem. Rev. 107 (2007) 2821.

[41] Y. Chen, L. Wei, B. Wang, S. Lim, D. Ciuparu, M. Zheng, J. Chen, C. Zoican, Y. Yang, G.L. Haller, L.D. Pfefferle, ACS Nano 1 (2007) 327.

[42] X.Q. Wang, M. Wang, H.X. Jin, Z.H. Li, P.M. He, Appl. Surf. Sci. 243 (2005) 151. [43] T. Somanathan, A. Pandurangan, Appl. Surf. Sci. 254 (2008) 5643.

[44] Y. Chen, B. Wang, L.J. Li, Y. Yang, D. Ciuparu, S. Lim, G.L. Haller, L.D. Pfefferle, Carbon 45 (2007) 2217.

[45] Y. Murakami, S. Yamakita, T. Okubo, S. Maruyama, Chem. Phys. Lett. 375 (2003) 393.

[46] S. Lim, D. Ciuparu, C. Pak, F. Dobek, Y. Chen, D. Harding, L. Pfefferle, G. Haller, J. Phys. Chem. B 107 (2003) 11048.

[47] Y. Yang, J.D. York, J. Xu, S. Lim, Y. Chen, G.L. Haller, Micropor. Mesopor. Mater. 86 (2005) 303.

[48] D. Barreca, W.J. Blau, G.M. Croke, F.A. Deeney, F.C. Dillon, J.D. Holmes, C. Kufazvinei, M.A. Morris, T.R. Spalding, E. Tondello, Micropor. Mesopor. Mater. 103 (2007) 142.

[49] C.Y. Mou, P. Lin, Pure Appl. Chem. 72 (2000) 137.

[50] Y.J. Han, J.M. Kim, G.D. Stucky, Chem. Mater. 12 (2000) 2068.

[51] A.Y. Lo, S.J. Huang, W.H. Chen, Y.R. Peng, C.T. Kuo, S.B. Liu, Thin Solid Films 498 (2006) 193.

[52] M.H. Lim, A. Stein, Chem. Mater. 11 (1999) 3285.

[53] S. Jana, B. Dutta, R. Bera, S. Koner, Langmuir 23 (2007) 2492. [54] L. Mercier, T.J. Pinnavaia, Adv. Mater. 9 (1997) 500.

[55] B.D. Cullity, Elements of X-ray diffraction, Second Ed., Addison-Wesely Publishing Co., 1978, pp. 100–102.

[56] A. Pozio, M. De Francesco, A. Cemmi, F. Cardellini, L. Giorigi, J. Power Sources 105 (2002) 13.

[57] Z. Liu, X.Y. Ling, X. Su, J.Y. Lee, J. Phys. Chem. B 108 (2004) 8234. [58] J.N. Tiwari, F.M. Pan, R.N. Tiwari, S.K. Nandi, Chem. Commun. (2008) 6516. [59] Y. Shao, J. Liu, Y. Wang, Y. Lin, J. Mater. Chem. 19 (2009) 46.

[60] J. Xu, K. Hua, G. Sun, C. Wang, X. Lv, Y. Wang, Electrochem. Commun. 8 (2006) 982. [61] B. Rajesh, K. Ravindranathan Thampi, J.M. Bonard, A.J. McEvoy, N. Xanthopoulos,

H.J. Mathieu, B. Viswanathan, J. Power Sources 133 (2004) 155.

0 500 1000 1500 2000 2500 100 200 300 400

M

a

ss Activity (A/g Pt)

Time (s)

Fig. 8. CA curves for Pt/CNT-d17, Pt/CNT-d8, Pt/CNT-d3, Pt/SWCNT, and Pt/XC-72 samples in 0.5 M H2SO4+ MeOH at 0.6 V.