Hydrogen storage in cobalt-embedded ordered mesoporous carbon

Hydrogen storage in cobalt-embedded ordered mesoporous carbon

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2 Available online at www.sciencedirect.com j...

838KB Sizes 1 Downloads 26 Views

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Hydrogen storage in cobalt-embedded ordered mesoporous carbon Chen-Chia Huang*, Yi-Hua Li, Yen-Wen Wang, Chien-Hung Chen Department of Chemical and Materials Engineering, National Yunlin University of Science and Technology, 123, Section 3, University Road, Douliu, Yunlin, Taiwan, ROC

article info

abstract

Article history:

Ordered mesoporous carbons (OMCs) were synthesized by using ordered mesoporous silica

Received 1 July 2012

as a template, and chitosan as carbon precursors. A novel process of pre-impregnation is

Received in revised form

proposed to prepare cobalt-embedded OMC. This process is based on using cobalt chelated

31 December 2012

chitosan as carbon precursor. The surface functional groups and metal contents were

Accepted 12 January 2013

determined by X-ray photoelectron spectroscopy. The bulk cobalt contents in the cobalt-

Available online 12 February 2013

embedded OMCs were measured by an atomic absorption spectrometer. The morphology of the OMCs was observed by small angle X-ray scattering analysis and transmission

Keywords:

electron microscope. The OMC texture characteristics were determined by using nitrogen

Hydrogen storage

adsorption analysis. Hydrogen capacities of the OMCs were obtained by a volumetric

Ordered mesoporous carbon

method. The cobalt-embedded OMCs possess obviously higher hydrogen adsorption ca-

Cobalt

pacity than that of pure OMC. At 298 K and under 5.5 MPa, the hydrogen capacities of the

Spillover

OMC and OMCeCo-5 are 0.2 and 0.45 wt%, respectively. The H2/Co ratio of the hydrogen

Kubas-type interaction

adsorbed on the OMCeCo-5 is 1.54 indicating a Kubas-type interaction between Co and H2. In addition, the hydrogen spillover effect might occur in parallel. Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

Hydrogen is considered as a promising renewable nonpolluting alternatives to fossil fuels. One of the key problems for promoting a hydrogen economy is to obtain efficient and safe storage materials for hydrogen. Currently, the main methods of storing hydrogen have involved compressed hydrogen, liquefied hydrogen, and storage in solid materials. The two major technologies to store hydrogen in solid materials are chemisorption in the form of metal hydrides [1,2] and physisorption of hydrogen on porous materials with large specific surface areas and high micropore volumes [3e7]. However, no material explored to date meets the storage targets (gravimetric capacity and volumetric capacity) set by

the US Department of Energy (DOE) for vehicular storage systems [8]. Carbon-based materials are among the major candidates of physisorption. The hydrogen molecule weakly interacts with the surface of the carbonaceous material. Since the interaction energies are very low, the hydrogen uptake amount is limited at room temperature. To increase the binding energy and enhance the adsorption capacity on carbon adsorbent, different routes have been explored. First, theoretical and experimental investigations have shown that it is possible to improve the hydrogen adsorption capacity of carbon materials by heteroatom substitution, such as boron, nitrogen, and phosphorus substitution [9e12]. A second possible approach is the use of hydrogen spillover by doping precious metals,

* Corresponding author. Tel.: þ886 5 534 2601x4616; fax: þ886 5 531 2071. E-mail address: [email protected] (C.-C. Huang). 0360-3199/$ e see front matter Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.01.081

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

such as palladium and platinum, on porous carbon adsorbents [13e15]. The hydrogen spillover is defined as the transport of an active hydrogen species from metal nanoparticles onto adjacent surfaces of a receptor via spillover and surface diffusion. Another possible method for increasing the hydrogen storage is to apply the Kubas interaction [16]. The Kubas-type interaction is based on molecular polarization and multiple s-bonding of hydrogen molecules with d orbitals of transition metal atoms or ions [17]. The Kubas interaction strength stands midway between hydrides and physisorption in binding strength and may be ideal in hydrogen storage systems designed for room temperature applications [18]. Recently, Contescu et al. [17] synthesized Pd-modified activated carbon fibers with 18 wt% of total Pd as isolated single atoms (observed by an atomic resolution scanning transmission electron microscopy, STEM) and 82 wt% of total Pd in nanoparticles. They suggested that four mechanisms contributed to the hydrogen uptake capacity: physisorption on carbon support, hydride formation in large Pd particles, H spillover, and multiple Kubas-type bonding to isolated Pd atoms. For application using either the hydrogen spillover or the Kubas interaction, the most important factor is to disperse completely the transition metal nanoparticles or atoms in porous carbon materials. However, it is difficult to prepare uniform metal nanoparticles or atoms into nanostructure carbon. By a conventional impregnation method, most of transition metals tend to cluster or aggregate into nanoparticles rather than remain as isolated atoms. Clustering lowers the capacity for hydrogen storage. Recently, ordered mesoporous carbon (OMC) materials, due to their super characteristics of high specific surface area, porosity, and chemical and thermal stability, have been studied for their applications as adsorbents [19], catalyst supports [20], hydrogen storage materials [21e24], and electrode materials [25]. Chitosan, consisting of b-(1 / 4)-linked Dglucosamine, is one of the most abundant biopolymers in nature. Because of its amino groups and secondary hydroxyl groups on the chain, chitosan is widely used to adsorb transition metal (TM), precious metal and rare metal ions through the chelation mechanism. It has been shown that addition of cobalt in carbon nanotubes can enhance the hydrogen adsorption [26]. In this paper, a novel process of preimpregnation is proposed. The ordered mesoporous carbon (OMC) was synthesized by replication using SBA-15 as the template and chitosan as the carbon source. To synthesize the cobalt-embedded OMCs, the cobalt ions were first chelated by chitosan. The cobalt-chelated chitosan solution was used as the carbon precursor. The aim of this work is to examine the texture properties as well as hydrogen storage capacity of the pure OMC and the cobalt-embedded OMC samples.

2.

Material and methods

The mesoporous silica template SBA-15 was synthesized using triblock copolymer Pluronic P123 as a template and tetraethyl orthosilicate (TEOS) as a silica source [27]. The pure OMC was synthesized according to the process described in the literature [28]. In a typical synthesis of OMC, 0.9 g chitosan powder was dissolved in a 9 mL solution of acetic acid (3%) and

3995

hydrochloric acid (0.3%). 1 g of calcined SBA-15 was added to a mixture of the chitosan solution and 0.12 g sulfuric acid. Then the mixture was placed in an oven at 373 K for 6 h. Subsequently, the oven temperature was raised to 433 K and maintained for another 6 h. The templateepolymer composite was then pyrolyzed in a furnace under nitrogen atmosphere up to 1173 K by a heating rate of 5 K/min, at the holding temperature for 6 h for carbonization. After cooling to room temperature, the silica framework was removed by dissolution in 2 M NaOH solution. The pure OMC was obtained after filtration, washing several times with de-ionized (DI) water and ethanol and then dried at 393 K. By a preliminary thermogravimetric analysis, about 70 weight percent of chitosan lost due to thermal degradation. To synthesize 1, 2.5, and 5 wt% of cobalt embedded in the OMCs, 2.73, 6.92, and 14.21 mg of cobalt nitrate, respectively, were dissolved in 2 mL DI water and then were added drop-wise into the chitosan solution. The prepared SBA-15 (1 g) was then added in the cobalt chelated chitosan solution with sulfuric acid. The following thermal treatments were the same as those for the synthesis of pure OMC. The cobalt-embedded ordered mesoporous carbons are denoted as OMCeCo-n, where n is the nominal Co embedded (in weight percent). The bulk contents of cobalt incorporated in the OMCeCo samples were determined by atomic absorption spectroscopy. Parts of the OMCeCo-5 samples were further submersed in potassium hydroxide solution (KOH/OMC ¼ 4 and 8) for 24 h. Then washed with large amounts of DI water and dried at 378 K for 24 h. The activated OMCeCo with different ratios of alkaline and carbon (KOH:OMC ¼ 4:1 and 8:1) are denoted as OMCeCo5-ac-4 and OMCeCo-5-ac8, respectively. The morphologies and nanostructures of the OMC samples were analyzed by using a small angle X-ray scattering analyzer (SAXS, Rigaku D/MAX-2500) and a transmission electron microscope (HRTEM, JEOL, JEM-2010 type). The Co-embedded OMCs were characterized by X-ray diffraction (XRD), using a GBC MMM X-ray diffractometer with Cu Ka radiation. The XRD patterns were recorded for 2q between 10 and 80 with 0.042 2q spacing and 5 s exposure times. The local surface element contents of the OMC with and without cobalt were determined by X-ray photoelectron spectroscope (XPS). The bulk cobalt contents in the cobalt-embedded OMCs were measured by an atomic absorption spectrometer (Perkin Elmer AAnalyst 400). The porous texture of the prepared OMC samples was measured with a gas adsorption apparatus (Quanta Chrome Autosorb-1) using nitrogen as adsorbate at 77 K. The specific surface area (SBET) was calculated by the BrunauereEmmetteTeller (BET) equation with nitrogen adsorption isotherms at relative pressure between 0.005 and 0.3. The pore size was calculated using the BJH model in the range 1.4e200 nm and the DFT method in the micropore range. The micropore volume was determined by the DeR model. The hydrogen adsorption in all OMC samples was accurately measured by a volumetric adsorption apparatus. Ultrahigh purity grade hydrogen (99.9995%) was used through a moisture trap (packed SICAPENT powder, Merck) to eliminate moisture impurities. About 200 mg of OMC samples were used for each measurement. The samples were outgassed under vacuum at 573 K for 3 h prior to the experiment. After the pretreatment, adsorption measurements were carried out

3996

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

under pressure below 3 MPa and at 77 K and up to 6 MPa at 298 K. The determined values were additionally cross-checked using the hydrogen adsorption data of a known amount of AX21 [29]. The hydrogen adsorption data are reproducible. Typically, the variation in uptake capacity is less than 1.5% throughout the entire pressure range.

3.

Results and discussion

3.1.

Characterization of OMCs

Fig. 1 shows the TEM images of the OMC and OMCeCo-5 samples. It is very clear from Fig. 1(a) that the OMC sample shows hexagonal mesopore channels, and the average pore size observed is about 8 nm. From Fig. 1(b), the ordered structure is shown to be slightly damaged by embedded cobalt in the OMCeCo-5. Although the TEM image of Co atoms embedded in the OMCeCo-5 was ambiguous, it was definitely observed by energy dispersive X-ray (EDX) analysis (as shown in Fig. 1(c)). This observation is different from that reported by Chen and Huang [26]. They applied post-impregnation method to modify

carbon nanotubes (CNT) by using Co(NO3)2 as the precursor. Co nanoparticles with size of w10 nm were clearly observed over the external surface of CNT samples by TEM imaging analysis. In this work, Co(NO3)2 was also used as the precursor. The cobalt ions were chelated by chitosan polymers and retarded to cluster into nanoparticles. The structures of the synthesized materials are characterized by SAXS patterns as shown in Fig. 2. As can be seen, the pattern of the OMC exhibits three peaks at 0.98 , 1.71 , and 1.84 , assigned to (100), (110) and (200), respectively, which are reflections of a highly ordered 2D hexagonal p6mm space group [30]. The high order (110) and (200) diffractions become less resolved for the samples with the higher Co content (OMCeCo-5 and OMCeCo-8). It suggests a gradual degradation of mesostructural regularity for the OMC composites with increasing Co content. The hexagonal unit cell parameter ao of the OMC is calculated to be 10.96 nm, and the d100 spacing is 9.49 nm. This result is consistent with the observation of the TEM analysis. The nitrogen adsorptionedesorption isotherms and the corresponding pore size distribution curves by the DFT methods of the OMC samples are shown in Fig. 3. It is apparent

Fig. 1 e TEM images for OMC (a) and OMCeCo-5 (b), and EDX of OMCeCo-5(c).

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

Fig. 2 e The SAXS peaks of the OMC and Co-embedded OMCs.

that the adsorption isotherms of nitrogen on the OMCs exhibited a steep initial uptake, indicating the microporous character of the carbons. The isotherms of the OMC samples (except for OMCeCo-8) are of type IV isotherms (based on IUPAC classification) and exhibit the typical H1 hysteresis loop with capillary condensation at a relative pressure that ranged between 0.4 and 0.95. From the shape and hysteresis of the isotherms, it is plausible to suggest that the OMCs show a mesoporous structure. However, in the case of OMCeCo-8, nitrogen isotherm looks like type I isotherm. The mesoporous structure characteristic is not obvious. The reason will be discussed later. As shown in Fig. 3(b), it is clear that the pore size distributions of the OMCeCo samples are with tri-modal pores that range from 3 to 5 nm (mesopores), from 1 to 2 nm, and less than 0.9 nm (micropores). Though the pore volume is reduced, the pore distribution of the Co-embedded OMC samples was only slightly changed from that of the OMC. The specific surface area SBET, the total pore volume, the micropore volume of the OMC and the Co-embedded OMC samples were calculated from nitrogen isotherm data and listed in Table 1. The specific surface area and total pore volume of the OMC are 1011 m2/g and 1.12 cm3/g, respectively.

3997

Xia et al. [21] synthesized CMK-3 using SBA-15 as the template and sucrose as the carbon precursor. They reported a specific surface area and pore volume of 984 m2/g and 1.09 cm3/g, respectively. From Table 1, it is noteworthy that both BET specific surface areas and micropore volumes of the cobaltembedded OMCs decreased significantly with the increase in the amount of Co. As aforementioned, cobalt ions arecoordinated with two chitosan residues via amine and hydroxyl functional groups. During the carbonization process, glycosidic linkages and pyranose rings of the chitosan polymer were partially destructed resulting in micropores. The viscosity of chitosan solution increases along with the amount of cobalt ions added. The excessive introduction of Co, e.g. OMCeCo-8, could increase the degree of difficulty to feed in the SBA-15, resulting in the structural defects of the prepared OMC. Furthermore, when the amounts of embedded Co increased, the micropores were widened and the BET specific area and micropore volume were decreased. The Co loading in the OMC samples are also listed in Table 1. The bulk contents of cobalt embedded in the OMCs were 2.89, 4.78, and 7.95 wt%, superior to the expected values of 1, 2.5, and 5 wt%, respectively. This could be attributed to the partial gasfication of chitosan during carbonization process. As can be seen in Table 1, the Co content values determined by AA are very close to that observed by XPS, indicating the well dispersion of Co nanoparticles in the OMC carbons. Wide-angle XRD patterns of the pure OMC and the Coembedded OMC samples are shown in Fig. 4(a). As expected, the two broad diffraction peaks at 26.7 and 43.7 corresponded to the graphite structures (002) and (100). Furthermore, the XRD patterns also confirm that the embedded Co in the OMCs did not lead to any structural changes in the mesoporous carbons. No diffraction peaks of cobalt were observed from the XRD patterns of the Co embedded OMCs, even for the OMCeCo-8 with high cobalt loading as 8 wt%. This suggests that the crystalline size of Co on Co embedded OMCs is below the lower limit for XRD detection ability (5 nm). Sun et al. [31] did not observe any Co diffraction peak for a Co/ SiO2 sample (10 wt% Co) prepared from cobalt acetate, and concluded that Co should highly disperse on the silica surface. Wang et al. [32] did not find any diffraction peaks in calcined Co/SBA-15 catalysts prepared from cobalt acetate with Co loading up to 20 wt%, but XRD peaks of Co3O4 were detected

Fig. 3 e (a) N2 adsorption/desorption isotherms at 77 K and (b) pore-size distribution for the OMCs.

3998

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

Table 1 e Textural characters and metal contents of the OMCs. Sample

OMC OMCeCo-2 OMCeCo-5 OMCeCo-8 a b c d

SBET (m2/g)

1011 789 776 554

Vtot (cm3/g)

Vmica (cm3/g)

1.12 0.99 0.93 0.39

0.39 0.31 0.29 0.24

Vmesob (cm3/g)

0.73 0.68 0.64 0.15

Co content (wt%) AAc

XPSd

e 2.89 4.78 7.95

e 2.06 5.00 8.26

Micropore volume, calculate by the DeR equation. Mesopore volume, Vmeso ¼ Vtot  Vmic. Measured by atomic absorption spectrometer. Measured by XPS analysis.

when cobalt nitrate was used as a precursor. To confirm cobalt atoms were embedded in the carbon matrix, the OMCeCo-5 was thermal reduced in a furnace under hydrogen up to 800  C for 5 h. The XRD patterns of the OMCeCo-5 before and after thermal reduction are shown in Fig. 4(b) for comparison. It is clear that three peaks at 44.2 , 51.5 , and 75.8 , corresponding to (111), (200), and (220) crystal of cobalt, were observed in the XRD pattern of the OMCeCo-5 after heat treatment. The Co crystallite size in the OMCeCo-5 after thermal reduction treatment was calculated as 11.7 nm by the Scherrer equation. The XRD patterns of the OMCeCo-5 before and after thermal reduction treatment are significantly different. During the thermal reduction process, smaller Co nanocrystals migrated and merged to form larger Co nanocrystals. A similar process was reported by Jin et al. [33]. By using scanning transmission electron microscopy in situ, they observed the Pd nanoparticles began to migrate, coalesce, and agglomerate to form larger particles on the graphene sheets as the composite was heated to 1073 K. Therefore, based on the comparison of XRD patterns shown in Fig. 4(b), that cobalt ions were chelated by chitosan and were highly dispersed in the mesoporous carbon matrix after being pyrolyzed at 1173 K under nitrogen flow. X-photoelectron analysis was used in an attempt to attain more insight into the surface composition of cobalt embedded in mesoporous carbon. The XPS spectra of the pure OMC and the OMCeCo-5 are shown in Fig. 5(a). The peak at 780 eV of the OMCeCo-5 sample is attributed to Co 2p, indicating cobalt was embedded in the carbon matrix. The contents of cobalt in the OMCeCo samples determined by XPS analysis are listed in

Table 1. The XPS spectra of Co 2p regions for the OMCeCo samples are illustrated in Fig. 5 (b). Signals at 780.9 and 797.0 eV binding energy and distinct satellite structures at 787.0 and 804.0 eV binding energy were found, corresponding to Co 2p3/2 and Co 2p1/2 structures of CoO [34]. It is known that a low intensity of shake-up satellite peak observed at ca. 787 eV is considered to be high-spinning Co2þ as the main form of cobalt species in the composites, due to the absence of a shake-up process for low-spinning Co3þ ions. Co3O4, a mixed-valence oxide, shows a weak satellite structure symptomatic of shake-up from the minor Co2þ component. Therefore, the cobalt embedded in the OMCeCo samples primarily performs in the form of CoO or Co (II) atoms.

3.2.

Hydrogen storage

Table 2 lists the comparison of hydrogen adsorption capacity of various OMCs at different temperatures and pressures. The hydrogen adsorption capacity of the OMC samples at 77 K ranged from 0.98 to 1.39 wt% and 2.3 to 2.8 wt% at 0.1 MPa and 3 MPa, respectively. The hydrogen adsorption capacity at higher temperatures dramatically decreased and ranged from 0.12 to 0.14 wt% at 298 K and 3 MPa. The difference of hydrogen adsorption capacity on various OMCs prepared by different laboratories might be caused by various carbon precursors and synthesis conditions which resulted in different texture characteristics. The hydrogen adsorption capacity of this work is comparable to the previous works reported in the literature [11,21,23,24,35].

Fig. 4 e (a) XRD patterns of OMC and Co-embedded OMCs, (b) comparison of XRD patterns of OMCeCo-5 before and after thermal reduction.

3999

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

Fig. 5 e (a) XPS spectra of pure OMC and Co-embedded OMCs (b) Co 2p XPS spectra of Co-embedded OMCs.

Table 2 e Comparison of hydrogen adsorption capacity on various OMCs at 77 K, 298 K and different pressures. SBET (m2/g)

Vt (cm3/g)

Vmic (cm3/g)

825 852 984 1011 1301

1.36 e 1.09 1.12 1.16

0.22 e 0.37 0.39 0.19

1330

1.084

0.013

H2 adsorption capacity (wt%) 77 K (P) 298 K (P) 2.3 1.03 0.98 3.3 1.39 2.85 1.24 2.8

The hydrogen isotherms of pure OMC along with various cobalt loadings of the OMC samples at 298 K with hydrogen pressure up to 900 Torr are plotted in Fig. 6. As shown in Fig. 6, a linear isotherm for the pure OMC via the original point was found. The adsorption behavior indicated low coverage physisorption onto the OMC receptor in the Henrys law region. It was observed from the isotherms that the hydrogen adsorption amounts on the Co-embedded OMCs are significantly higher than that on the pure OMC. The hydrogen adsorption capacities on Co-embedded OMCs increase noticeably with cobalt loading. It is noted that two isotherm adsorption behaviors for Co-embedded OMCs are developed:

Fig. 6 e Low pressure hydrogen adsorption isotherms on OMC and Co-embedded OMCs at 298 K.

(3 MPa) (0.1 MPa) (0.1 MPa) (3 MPa) (0.1 MPa) (3 MPa) (0.1 MPa) (3 MPa)

References

0.12 (3 MPa) 0.14 (3 MPa) e 0.11 (3 MPa) e

[35] [11] [21] This work [24]

0.24 (4.5 MPa)

[23]

the H2 chemisorption by Co nanoparticles and the H2 physisorption by porous carbon receptors. Generally, the hydrogen chemisorption capacity of Co is obtained from the linear isotherm portion in a pressure range above 80 Torr extrapolated to zero pressure [36]. The increase in the amount of hydrogen adsorbed onto the metal-doped carbons is commonly attributed to the spillover effect [37]. High pressure hydrogen adsorption isotherms at 298 K for the pure OMC and the Co-embedded OMC samples are presented in Fig. 7(a). As shown in Fig. 7(a), the pure OMC had a hydrogen storage capacity of 0.2 wt % at 298 K and under 5.5 MPa. For the OMCeCo-8 sample, the hydrogen uptake at 5.5 MPa was enhanced to 0.45 wt %. It can be seen that the cobalt-embedded OMCs had obviously higher hydrogen adsorption capacities. The enhanced hydrogen storage capacity could not be attributed to the differences in surface areas or pore volumes because the Co-embedded OMC sample had less surface areas and micropore volumes (as listed Table 1). In addition, the hydrogen capacity on the 5.0 wt % Co postimpregnated OMCePIeCo-5 is also shown in Fig. 7 for comparison. The BET specific surface area and micropore volume of the OMCePIeCo-5 are 515 m2/g and 0.24 cm3/g, respectively. The hydrogen storage capacity of OMCeCo-5 (by preimpregnation method) is much higher than that of OMCePIeCo-5 (by post-impregnation method). In other words, well dispersed cobalt atoms were responsible for this enhanced capacity. According to XRD and XPS analysis, the embedded cobalt was highly dispersed in the OMCeCo samples as Co atoms. The H2/Co ratio of the hydrogen adsorbed on the OMCeCo-8 is 1.54, indicating the Kubas-type interaction between Co and H2. The maximum number of H atoms (Nmax)

4000

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

Fig. 7 e (a) High pressure hydrogen adsorption capacity on OMC and Co-embedded OMCs at 298 K, (b) hydrogen adsorption and desorption isotherms of OMCeCo-5.

a TM atom can adsorb satisfies the 18-electron rule [38,39], Nmax ¼ 18  nv  5 where nv is the number of valence electrons of the TM atom, and 5 is the number of electrons contributed by the cyclopentadiene rings. By using the 18-electron rule, a maximum of 2 dihydrogen molecules can be bound on each Co atom. The hydrogen uptake is less than what might be expected. This can be attributed to part of the Co atoms being buried in the carbon that is not exposed to the experiment [40]. Hydrogen adsorption and desorption isotherms for OMCeCo-5 are presented in Fig. 7(b). The desorption branch isotherms were measured by decreasing pressure down to 0.1 MPa. The desorption branch nearly followed the adsorption branch although there appeared to be a slight hysteresis. A similar result is reported by Wang and Yang [41]. The OMCs were further activated with KOH to create more micropores [42]. The ratio of KOH and OMCeCo-5 was fixed at 4:1 or 8:1. The activation was conducted at 973 K for 1 h under nitrogen flow of 500 mL/min. The activated products were washed with diluted HCl solution to remove potassium species. After filtration the samples were dried at 378 K. The OMCeCo-5 samples, after KOH activation, are referred to as OMCeCo-5-ac4 or OMCeCo-5-ac8 for different KOH/ carbon ¼ 4 or 8, respectively. After being mixed with KOH, the

Fig. 8 e High pressure hydrogen adsorption capacity on OMCeCo-5 and activated OMCeCo-5 at 298 K.

surface areas and micropore volumes of the activated OMCeCo-5-ac4 increased from 776 to 1083 m2/g and 0.29 to 0.46 cm3/g, respectively. The specific surface area of OMCeCo5-ac8 decreased to 968 m2/g, though the micropore volume increased to 0.65 cm3/g. The incremental surface area of OMCeCo-5 after KOH activation implies that defects on the graphite structure were formed by carbonaceous gasification [26]. Additionally, excessive defect sites on the graphite walls of the OMCeCo-5 caused structural collapse after the activation process with too much of the KOH. At 298 K and under 5.5 MPa, as shown in Fig. 8, the hydrogen capacity of the OMCeCo-5-ac8 was 0.50 wt%, which was enhanced by 15% compared to that of the OMCeCo-5. The hydrogen storage capacity enhancement of the OMCeCo-5-ac8 after KOH activation was attributed to the increases of BET surface area and micropore volume.

4.

Conclusions

An ordered mesoporous carbon (OMC) was synthesized by using ordered mesoporous silica SBA-15 as a template, and chitosan and cobalt chelated chitosan as carbon precursors. By TEM and SAXS analyses, the ordered 2D hexagonal structure was well retained for the synthesized OMC samples even when Co was embedded. Both BET specific surface areas and micropore volumes of the cobalt-embedded OMCs are less than those of the OMC. By XPS and XRD analyses, cobalt atoms were highly dispersed in the OMCeCo samples. The cobaltembedded OMC obviously possesses higher hydrogen adsorption capacity than that of the OMS. The H2/Co ratio of the hydrogen adsorbed on the OMCeCo-5 is 1.54, indicating a Kubas-type interaction between Co and H2. At 298 K and under 5.5 MPa, the hydrogen capacity of the OMCeCo-5-ac8 is 0.50 wt%, which is enhanced by 15% compared to that of the OMCeCo-5. The hydrogen storage capacity enhancement of the OMCeCo-5-ac8 after KOH activation was attributed to the increases of BET surface area and micropore volume. In short, the Kubas interaction along with hydrogen spillover and physisorption to carbon support play the roles of the mechanism of hydrogen adsorption in cobalt-embedded mesoporous carbons.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

Acknowledgments This work was financially supported by the National Science Council, Taiwan, Republic of China, under contract (NSC 1002221-E-224-035).

references

[1] Yoo JH, Shi G, Park CN, Kim WB, Cho SW. Influence of Mn or Mn plus Fe on the hydrogen storage properties of the Ti-Cr-V alloy. Int J Hydrogen Energy 2009;34:9116e21. [2] Huang Z, Cuevas F, Liu X, Jiang L, Wang S, Latroche M, et al. Effects of Si addition on the microstructure and the hydrogen storage properties of Ti26.5V45Fe8.5Cr20Ce0.5 BCC solid solution alloys. Int J Hydrogen Energy 2009;34:9385e92. [3] Chen C-H, Huang C-C. Hydrogen storage by KOH-modified multi-walled carbon nanotubes. Int J Hydrogen Energy 2007; 32:237e46. [4] Stro¨bel R, Garche J, Moseley PT, Jo¨rissen L, Wolf G. Hydrogen storage by carbon materials. J Power Sources 2006;159: 781e801. [5] Yu¨ru¨m Y, Taralp A, Veziroglu TN. Storage of hydrogen in nanostructured carbon materials. Int J Hydrogen Energy 2009;34:3784e98. [6] Huang C-C, Chen H-M, Chen C-H. Hydrogen adsorption on modified activated carbon. Int J Hydrogen Energy 2010;35: 2777e80. [7] Guo H, Gao Q. Cryogenic hydrogen uptake of high surface area porous carbon materials activated by potassium hydroxide. Int J Hydrogen Energy 2010;35:7547e54. [8] Satyapal S, Petrovic J, Read C, Thomas G, Ordaz G. The US department of energy’s national hydrogen storage project: progress towards meeting hydrogen-powered vehicle requirements. Catal Today 2007;120:246e56. [9] Wang L, Yang FH, Yang RT. Hydrogen storage properties of Band N-doped microporous carbon. AIChE J 2009;55(7): 1823e33. [10] Xia Y, Walker GS, Grant DM, Mokaya R. Hydrogen storage in high surface area carbons: experimental demonstration of the effects of nitrogen doping. J Am Chem Soc 2009;131: 16493e9. [11] Jiang J, Gao Q, Zheng Z, Xia K, Hu J. Enhanced room temperature hydrogen storage capacity of hollow nitrogencontaining carbon spheres. Int J Hydrogen Energy 2010;35: 210e6. [12] Jin Z, Sun Z, Simpson LJ, O’Neill KJ, Parilla PA, Li Y, et al. Solution-phase synthesis of heteroatom-substituted carbon scaffolds for hydrogen storage. J Am Chem Soc 2010;132: 15246e51. [13] Lueking AD, Yang RT. Hydrogen spillover to enhance hydrogen storage e study of the effect of carbon physicochemical properties. Appl Catal A 2004;265:259e68. [14] Li Y, Yang RT. Significantly enhanced hydrogen storage in metal-organic frameworks via spillover. J Am Chem Soc 2006;128:726e7. [15] Saha D, Deng S. Hydrogen adsorption on ordered mesoporous carbons doped with Pd, Pt, Ni, and Ru. Langmuir 2009;25(21):12550e60. [16] Kubas GJ. Molecular hydrogen complexes: coordination of a s bond to transition metals. Acc Chem Res 1988;21(3):120e8. [17] Contescu CI, van Benthem K, Li S, Bonifacio CS, Pennycook SJ, Jena P, et al. Single Pd atoms in activated carbon fibers and their contribution to hydrogen storage. Carbon 2011;49:4050e8.

4001

[18] Hamaed A, Trudeau M, Antonelli DM. H2 storage materials (22 kJ/mol) using organometallic Ti fragments as s-H2 binding sites. J Am Chem Soc 2008;130:6992e9. [19] Hartmann M, Vinu A, Chandrasekar G. Adsorption of vitamin E on mesoporous carbon molecular sieves. Chem Mater 2005; 17(4):829e33. [20] Li L, Zhu ZH, Lu GQ, Yan ZF, Qiao SZ. Catalytic ammonia decomposition over CMK-3 supported Ru catalysts: effects of surface treatments of supports. Carbon 2007;45(1):11e20. [21] Xia K, Gao Q, Wu C, Song S, Ruan M. Activation, characterization and hydrogen storage properties of the mesoporous carbon CMK-3. Carbon 2007;45:1989e96. [22] Armandi M, Bonelli B, Karaindrou EI, Otero Arean C, Garrone E. Post-synthesis modifications of SBA-15 carbon replicas: improving hydrogen storage by increasing microporous volume. Catal Today 2008;138:244e8. [23] Giraudet S, Zhu Z, Yao X, Lu G. Ordered mesoporous carbons enriched with nitrogen: application to hydrogen storage. J Phys Chem C 2010;114:8639e45. [24] Kockrick E, Schrage C, Borchardt L, Klein N, Rose M, Senkovska I, et al. Ordered mesoporous carbide derived carbons for high pressure gas storage. Carbon 2010;48: 1707e17. [25] Zhou H, Zhu S, Hibino M, Honma I, Ichihara M. Lithium storage in ordered mesoporous carbon (CMK-3) with high reversible specific energy capacity and good cycling performance. Adv Mater 2003;15(24):2107e11. [26] Chen C-H, Huang C-C. Enhancement of hydrogen spillover onto carbon nanotubes with defect feature. Microporous Mesoporous Mater 2008;109:549e59. [27] Chareonpanich M, Nantangern A, Limtrakul J. Short-period synthesis of ordered mesoporous silica SBA-15 using ultrasonic technique. Mater Lett 2007;61:5153e6. [28] Ryoo R, Joo SH, Jun S. Synthesis of highly ordered carbon molecular sieves via template-mediated structural transformation. J Phys Chem B 1999;103:7743e6. [29] Lachawiec Jr AJ, DiRaimondo TR, Yang RT. A robust volumetric apparatus and method for measuring high pressure hydrogen storage properties of nanostructured materials. Rev Sci Instrum 2008;79:063906e12. [30] La´zaro MJ, Calvillo L, Bordeje´ EG, Moliner R, Juan R, Ruiz CR. Functionalization of ordered mesoporous carbons synthesized with SBA-15 silica as template. Microporous Mesoporous Mater 2007;103:158e65. [31] Sun S, Tsubaki N, Fujimoto K. The reaction performances and characterization of FischereTropsch synthesis Co/SiO2 catalysts prepared from mixed cobalt salts. Appl Catal A 2000;202:121e31. [32] Wang Y, Noguchi M, Takahashi Y, Ohtsuka Y. Synthesis of SBA-15 with different pore sizes and the utilization as supports of high loading of cobalt catalysts. Catal Today 2001;68:3e9. [33] Jin Z, Nackashi D, Lu W, Kittrell C, Tour JM. Decoration, migration, and aggregation of palladium nanoparticles on grapheme sheets. Chem Mater 2010;22:5695e9. [34] Natile MM, Glisenti A. Study of surface reactivity of cobalt oxides: interaction with methanol. Chem Matter 2002;14: 3090e9. [35] Giraudet S, Zhu Z. Hydrogen adsorption in nitrogen enriched ordered mesoporous carbons doped with nickel nanoparticles. Carbon 2011;49:398e405. [36] Prado-Burguete C, Linares-Solano A, Rodriguez-Reinoso F, Lecea CSMD. Effect of carbon support and mean Pt particle size on hydrogen chemisorption by carbon-supported Pt catalysts. J Catal 1991;128:397e404. [37] Wang Z, Yang RT. Enhanced hydrogen storage on Pt-doped carbon by plasma reduction. J Phys Chem C 2010;114: 5956e63.

4002

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 8 ( 2 0 1 3 ) 3 9 9 4 e4 0 0 2

[38] Huheey JE, Keiter EA, Keiter RL. Inorganic chemistry: principles of structure and reactivity. 4th ed. New York: Harper Collins College Publishers; 1993. [39] Kiran B, Kandalam AK, Jena P. Hydrogen storage and the 18electron rule. J Chem Phys 2006;124:224703. [40] Pocard NL, Alsmeyer DC, McCreery RL, Neenan TX, Callstrom MR. Nanoscale platinum(0) clusters in glassy carbon: synthesis, characterization, and

uncommon catalytic activity. J Am Chem Soc 1992;114(2): 769e71. [41] Wang L, Yang RT. Molecular hydrogen and spiltover hydrogen storage on high surface area carbon sorbents. Carbon 2012;50:3134e40. [42] Zheng Z, Gao Q, Jiang J. High hydrogen uptake capacity of mesoporous nitrogen-doped carbons activated using potassium hydroxide. Carbon 2010;48:2968e73.