Journal of Colloid and Interface Science 394 (2013) 401408

Contents lists available at SciVerse ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Hydroxyapatite supported cobalt catalysts for hydrogen generation

Justyn Wayne Jaworski a,b, Sunghwa Cho b, Yeoungyong Kim b, Jong Hwa Jung b, Hyo Sang Jeon c,d, Byoung Koun Min c,d,, Ki-Young Kwon b,
a

Department of Chemical Engineering, Hanyang University, Seoul 133-791, South Korea Department of Chemistry, Gyeongsang National University and RINS, Jinju 660-701, South Korea c Clean Energy Research Center, Korea Institute of Science and Technology, 39-1 Hawolgok-dong, Seongbuk-gu, Seoul 136-791, South Korea d University of Science and Technology, 176 Gajung-dong, 217 Yuseong-gu, Daejeon 305-350, South Korea
b

a r t i c l e

i n f o

a b s t r a c t
The controlled generation of H2 from storage materials by using an efcient catalytic support is a highly sought after technology; however, the majority of successes utilize expensive materials considered unfeasible. In our report on the creation of a novel, durable, and inexpensive catalytic support material for hydrogen generation, we examine a critical surface modication of hydroxyapatite (HAP) with cobalt ions to provide the necessary catalytic transition metal for the fast hydrolysis of the hydrogen storage material, sodium borohydride (NaBH4). By altering the morphology and composition of the HAP crystal supports, we revealed novel methods for enhancing the hydrogen generation rates. Particularly, lowering the Ca composition during synthesis of the HAP crystals afforded a Ca decient HAP capable of exhibiting a higher surface coverage of cobalt, thereby eliciting faster hydrolysis reaction rates in comparison with the amorphous HAP control having the characteristic Ca content for HAP. A more signicant increase in hydrogen generation was observed when using single crystal HAP in comparison with amorphous and calcium decient HAP supports. Despite the smaller surface area of the hydrothermally prepared single crystal HAP, it provided signicantly faster hydrogen generation. Each of the HAP supports exhibit repeatability with catalytic efciency decreasing by approximately 25% over 3 weeks upon repeated daily exposure to solutions of the hydrogen storage material NaBH4. Through these experiments, we proved that altering the composition and morphology of cobalt ion exchanged HAP supports can offers a useful means for increasing the rate of controlled hydrogen generation. Crown Copyright 2012 Published by Elsevier Inc. All rights reserved.

Article history: Received 31 July 2012 Accepted 15 November 2012 Available online 3 December 2012 Keywords: Hydroxyapatite support Hydrogen generation Calcium decient Cobalt Ion exchange

1. Introduction Unlike fossil fuels, hydrogen provides a green, sustainable fuel, which has been considered an ideal future energy reservoir [1,2]. The use of hydrogen fuel is environmentally benign, since water is the only by-product of the combustion process [35]. Moreover, hydrogen is readily available from water and a variety of other sources, thereby removing the need for competition over energy resources [6]. Researchers have made an incredible effort to generate hydrogen in an environmentally clean and cost-effective manner [7,8]. Recent works in Co containing synthetic ligands have shed new light on some of the catalytic mechanism of molecular hydrogen evolution [9]. In pursuing this goal, water splitting into hydrogen and oxygen using homo/heterogeneous catalysts has been the most comprehensively studied method for acquiring hydrogen [2,10,11]. Once hydrogen is efciently generated, another challenge arises for its effective, safe, and convenient storage

for practical hydrogen fuel usage [12,13]. Hydrogen storage methods have been proposed using high pressure, cryogenics, and various physisorbing compounds for reversible uptake and release of hydrogen. Because of its high ammability, material safety under ambient conditions is an important consideration. For this reason, chemical hydrides have received well-deserved interest due to their hydrogen generation capabilities and safe transportability [4,14,15].

NaBH4 aq 2H2 Ol ! NaBO2 aq 4H2 g

catalyst

Corresponding authors. Fax: +82 55 761 0244.

E-mail addresses: bkmin@kist.re.kr (B.K. Min), kykwon@gnu.ac.kr (K.-Y. Kwon).

One chemical hydride, sodium borohydride (NaBH4), has been explored as a model material for the effective storage of hydrogen, due to its safety for transport and high hydrogen storage capacity (2.4 L of H2 per gram of NaBH4) [16,17]. In addition to being a high energy density hydrogen carrier, NaBH4 has proven to have several other advantages including the following: stability in air, ability to generate hydrogen at low temperatures, non-hazardous reaction products, and the ability to control the hydrogen generation rate [18,19]. Control of the hydrogen generation rate must be done under alkaline solution for stability, where a solution of 110% NaOH

0021-9797/$ - see front matter Crown Copyright 2012 Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2012.11.036

402

J.W. Jaworski et al. / Journal of Colloid and Interface Science 394 (2013) 401408

is typically used to stabilize the NaBH4 to prevent its uncontrolled self-hydrolysis [20]. By using an effective catalyst such as that described in this work, the base-stabilized NaBH4 solution can be controllably hydrolyzed to activate the production of hydrogen in a directed manner as indicated in Eq. (1) [19]. The generation of hydrogen by a catalyst requires a stable catalyst-support structure to provide controlled starting and stopping of the reaction with high efciency. Previously, our group as well as others have shown that support structures bearing catalysts with high price noble metals (including Ru, Pt, and Rh) perform with high reusability [2123]; however, a more abundant, low-cost alternative is necessary. In order to be practical, researchers have attempted alloying with inactive metals such as Cu, Ag, or Pd, though the results revealed decreased performance [24]. To solve this problem, Ni and Co based catalysts have been explored due to their low cost and high performance in hydrolysis of NaBH4 [4,2528]. In the past, Co powders have had a tendency to travel with bubbles produced from the hydrolysis reaction and have been found to clog the reactors of hydrogen generation systems [24,26]. For the purpose of real-world applications, a catalyst-support structure can improve catalyst dispersion. Recent works have found supported Co catalysts to provide high activity, long-lifetime, and low cost for supply of hydrogen from the hydrolysis of sodium borohydride [26,29]. An often noted drawback of many supported Co systems for hydrogen generation is the lack of reusability in NaBH4 hydrolysis [30]. While the exact mechanism as to how the surface is becoming deactivated remains unclear, the formation of borate by-product has often been observed to block the catalytic surface and furthermore has shown to be removable by washing to afford a fully recovered system [31]. If given high activity, reusability, and economic feasibility, a supported Co catalyst may provide a viable prospect for controlled hydrogen generation. In the following study, we demonstrate improved reusability in a supported catalysts produced by Co ion exchange on the surface of hydroxyapatite (HAP). While the development of an improved catalyst material is of key importance, system level designs which reduce by-product and regenerate used NaBH4 are critical for the future of NaBH4 based hydrogen storage systems [32]. A range of catalysts for sodium borohydride hydrolysis including Ru/anion-exchange resins, Pt/LiCoO2, Ru/carbon, PtRu/LiCoO2 among many others that have been proposed and approach 100% hydrolysis, typically generating hydrogen at a rate of 0.12.8 L/(min g of catalyst) [33]. Having performed an in-depth look at a variety of other catalytic support structures, including nanoclusters, polymeric supports, and even foams [4,18,34,35], we can appreciate that certain material characteristics can be determining factors for the efciency of the support. Specically, the support structure should not agglomerate, should maintain a high active surface area, minimize deactivation due to borate by-product formation, and should not deteriorate due to surface stress resulting from rapid hydrogen bubble generation. Aside from supported catalysts systems, approaches for mobile hydrogen storage systems have also successfully demonstrated the use of aqueous solutions of cobalt as part of a dual-solid-fuel system known to have advantages in terms of hydrogen storage density and fuel conversion [36]. In such a system, Co2B particulates are formed which are catalytically active but also subject to rapid deactivation. Creating a supported Co catalyst with stability and reusability is thereby seen as a technical challenge for hydrogen generation. In this work, we examine several morphological and compositional forms of hydroxyapatite (HAP) supported Co catalysts to address the challenging issues in the use of NaBH4 for controlled hydrogen generation. Aside from being inexpensive, durable, easy to manufacture, environmentally benign, and stable in basic solution, we have previously found that HAP is an effective material

for the immobilization of metal ions onto the surface [21]. Recently, through the cation exchange of protons of the P-OH groups with Co2+ ions, researchers have determined that Co is capable of being quickly immobilized onto the HAP surface with equilibrium occurring after only 5 min [37]. Moreover, the amount of Co content on HAP can be easily controlled with capacities of up to 0.130.30 mmole of Co per gram of HAP, depending on the variety of HAP support [38]. In this study, we created single crystal HAP (s_HAP), amorphous HAP (a_HAP), and calcium decient amorphous HAP (cd_HAP) catalytic supports to examine the effect of composition and morphology on the generation of hydrogen. Our analysis reveals the amount of Co immobilization depends on the degree of calcium composition of the HAP support, wherein calcium decient HAP could provide a higher content of surface Co and hence provide a more effective catalytic support for hydrolysis of sodium borohydride. Interestingly, we nd that the single crystal HAP, despite having a lower surface area, results in a greater rate of hydrogen generation per weight of catalytic support (Co and HAP). Examination of the long-term stability by measurement of total hydrogen generation demonstrated that each of the three catalytic support structures were reusable over the course of 20 days. These main results and supporting experiments are reported and discussed herein. 2. Materials and methods 2.1. Synthesis of calcium decient HAP Puriss grade reagents were obtained from Sigma Aldrich. Preparation of Ca decient HAP was carried out by adjusting the molar ratio of Ca/P to 1. 5.59 g of (NH4)2HPO4 dissolved in 420 mL of water is added dropwise into 10 g of Ca(NO3)24H2O dissolved in 420 mL of water while stirring under ambient temperature and pressure. After the solution of pH is adjusted to 10 using concentrated ammonia solution, the solution is reuxed for 2 h under a 1 atm N2 condition while stirring. After cooling to room temperature, the mixture was ltered and washed three times with approximately 200 mL distilled water (the nal ltrate solution has a pH  6.7). Finally, the white precipitate is dried overnight in a 70 C oven under static air. 2.2. Synthesis of amorphous HAP Preparation method of amorphous HAP is same to that of calcium decient HAP except for a Ca/P molar ratio of 1.67 was used. 2.3. Synthesis of single crystal HAP Puriss grade reagents were obtained from Sigma Aldrich. To the 30 mL solution of Ca(NO3)24H2O (7.79 g), 50 mL solution of Na2HPO42H2O (3.56 g) is added dropwise. The solution is adjusted to pH 10 by adding 2 M NaOH solution. The solution is transferred to a Teon lined autoclave reactor and is aged at 200 C for 24 h without stirring (while heating, the pressure reached approximately 10 bar). After cooling the solution mixture to room temperature, white precipitates were obtained. The precipitate is washed three times using approximately 200 mL of DI water and ltered. Finally, the white precipitate is dried overnight in a 70 C oven under static air. 2.4. Hydroxyapatite characterization The crystalline phases of the Ca decient HAP, single crystal, and amorphous HAP were evaluated by XRD (3 kW Cu X-ray Diffractometer, D8 Advance, Bruker AXS Germany) before and after

J.W. Jaworski et al. / Journal of Colloid and Interface Science 394 (2013) 401408

403

Co ion exchange. The morphologies of the samples were observed using a transmission electron microscope (Jeol 200 kV JEM-2010). 2.5. Ion exchange of cobalt on hydroxyapatite ACS grade cobalt(II) nitrate was obtained from Sigma Aldrich. Cobalt(II) nitrate, Co(NO3)2, was dissolved in 20 mL of distilled water to concentrations of 0.01, 0.05, 0.1, 0.14, 0.18, 0.22, and 0.25 mM. These solutions were then exposed to 2 g of single crystal HAP, amorphous HAP, or Ca decient HAP. Initially, the samples were shaken vigorously and then sonicated for 10 min. The samples were then mixed moderately for another 20 min on a rocking platform. After ion exchange, the solution was ltered and then washed twice with 200 mL of distilled water. The samples were dried overnight in a 70 C oven under static air for 24 h before undergoing elemental analysis and hydrogen generation experiments. 2.6. Analysis of elemental composition ACS grade nitric acid was obtained from Sigma Aldrich. In order to monitor the elemental composition of the ion exchanged catalytic supports, a small amount of each support (5 mg) was separately dissolved in 1 mL of 70% nitric acid followed by the addition of 30 mL of distilled water. The samples underwent elemental analysis using an Inductively Coupled Plasma (ICP) spectrometer (OPTIMA 4300DV/5300DV, Perkin Elmer) to observe the elemental composition of cobalt, calcium, and phosphorous. 2.7. Catalytic generation of hydrogen In order to measure the generation of hydrogen by the supported catalyst systems, the volume of water displaced from an inverted graduated cylinder was measured from a closed system with the catalyst and substrate in a connected round bottom ask at 22 C and 1 atm. Using 0.1 g of HAP supported catalyst (s_CoHAP, a_CoHAP, or cd_CoHAP having particle sizes as indicated by TEM) placed in the round bottom ask, 0.2 g of sodium borohydride dissolved in 20 mL 1% NaOH (0.25 M) was added to the ask with stirring at 1000 rpm using a Teon-coated stir bar and a magnetic stirrer. (See Supplementary Fig. 4 for analysis of the effects of stirring speed on the hydrogen generation rate). The closed system was created after addition of the sodium borohydride by replacement of the stopcock. The volume of water displaced from the cylinder, and the time was simultaneously recorded until the reaction was complete. Blank reactions were carried out with only sodium borohydride present in the sodium hydroxide solution revealing no hydrogen generation (negative controls). Other controls which yielded no hydrogen generation were control samples of 0.1 g s_HAP, a_HAP, or cd_HAP without undergoing cobalt ion exchange but tested in the same manner as above. Additional control experiments using 0.03 mmole of pure Co(NO3)2 without any HAP have also been conducted using the methods above to obtain the rate of hydrolysis using standalone cobalt in an amount equivalent to that present on some of the supported CoHAP catalysts. It should be noted that due to particulate formation when using pure Co(NO3)2, a proportion of the cobalt remains unavailable for reaction thereby rendering them as inactive Co sites. (See Supplementary Fig. 3 for data and additional analysis). 2.8. Catalyst stability experiments In these experiments, s_CoHAP, cd_CoHAP, or a_CoHAP catalyst supports were generated by a 0.1 M Co(NO3)2 ion exchange reaction as described in Section 2.5. Using the previously described experimental setup for the catalytic generation of hydrogen (Sec-

tion 2.7), the volume of hydrogen generated from a controlled amount (0.2 g in 20 mL basic solution of 1% NaOH) of sodium borohydride exposure was measured. The same 0.1 g of CoHAP catalyst supports was repeatedly exposed to sodium borohydride daily over a 3 week period, and there generation rates were measured to assess the stability. At 24 h intervals, measurements were taken, and the solution was replaced with a new 20 mL solution containing 0.2 g sodium borohydride in 1% NaOH. The catalytic support solution was not removed or washed during this process. Control experiment using 0.03 mmole of pure Co(NO3)2 without any HAP was also conducted in this manner to observe the reusability of standalone cobalt. 2.9. Surface characterization of catalytic supports before and after prolonged hydrogen generation Samples of the prepared s_CoHAP, a_CoHAP, and cd_CoHAP were used for repeated daily hydrolysis of sodium borohydride for 3 weeks as explained in Section 2.8. The resulting reacted samples were washed three times with approximately 200 mL distilled water and then subjected to FTIR as were also the unreacted samples of s_CoHAP, cd_CoHAP, and a_CoHAP. IR spectra of the CoHAP samples were obtained as KBr pellets, using 32 scans in the range of 4004000 cm1, with a Shimadzu FTIR 8400S instrument. Samples of s_CoHAP before and after sodium borohydride exposure were subjected to X-ray photoelectron spectroscopy (XPS). The XPS analysis was conducted with a PHI 5000 VersaProbe (UlvacPHI) spectrometer using monochromatic Al Ka lines of Al (1486.6 eV) as an X-ray source. 3. Results 3.1. Characterization The crystallite shape of Ca decient and amorphous HAP prepared by precipitation resulted in crystals exhibiting a plate-like structure, Fig. 1, consistent with that found in the literature for HAP synthesized by wet-chemical precipitation [39]. The average crystal size distribution of the Ca decient and amorphous HAP samples was determined to be approximately 78 nm in length. The aspect ratio of the crystals did not appear homogeneous, and the shape of the major plane was consistently asymmetric. In contrast, the single crystal HAP samples were signicantly larger with a length of approximately 155 nm and possessing an elongated hexagonal rod shape. From the TEM images, we can also see that no morphological changes can be observed due to ion exchange in Co solution for a_HAP, cd_HAP, and s_HAP. By XRD qualitative analysis (Fig. 2), we veried that the synthesis produced hydroxyapatite structures, without the presence of impurity phases such as CaO, CaHPO4, or b-Ca(PO4)2, and the XRD data were found to be consistent with the XRD patterns known in literature [4042]. The sharp peaks of the XRD patterns for s_HAP and s_CoHAP reveal that the single crystal was not altered in its bulk crystal structure due to ion exchange. For the other HAP samples as well, ion exchange was found to not affect the bulk crystal structure as determined by comparison of the XRD data patterns. This indicates the ion exchange is taking place on the outmost several layers of the crystals. From chemical analysis, Ca decient hydroxyapatite was found to have a Ca:P ratio of 1.448, and the hexagonal crystal system of the Ca decient hydroxyapatite, Ca10x(HPO4)(PO4)6x(OH)2x, was determined to have 0.3036 vacancies per 10 calcium ions. In addition to these comparisons, the amorphous and Ca decient HAP were found through nitrogen adsorption experiments to have a similar BET surface area of approximately 56.4 3 m2/g, while the BET surface area of single

404

J.W. Jaworski et al. / Journal of Colloid and Interface Science 394 (2013) 401408

Fig. 1. TEM images of (A) amorphous HAP (a_HAP), (B) Co ion exchanged amorphous HAP (a_CoHAP), (C) Ca decient HAP (cd_HAP), (D) Co ion exchanged Ca decient HAP (cd_CoHAP), (E) single crystal HAP (s_HAP), (F) Co ion exchanged single crystal HAP (s_CoHAP). All scale bars in the gure represent 100 nm.

Fig. 2. XRD pattern of amorphous HAP (a_HAP), Co ion exchanged amorphous HAP (a_CoHAP), Ca decient HAP (cd_HAP), Co ion exchanged Ca decient HAP (cd_CoHAP), single crystal HAP (s_HAP), and Co ion exchanged single crystal HAP (s_CoHAP),

crystal HAP was found to be 6.7 0.2 m2/g. 1. We also nd that the surface area was not affected by ion exchange as is consistent with our TEM images revealing no morphological changes after ion exchange. 3.2. Effect of hydroxyapatite calcium composition on Co ion exchange Ion exchange of Co(NO3)2 in the presence of Ca decient HAP and amorphous HAP revealed the amount of cobalt that could be incorporated onto the supports as related to Ca composition. In order to identify the Co content of the catalytic supports after ion exchange,

ICP elemental analysis was perform on both Ca decient as well as on amorphous HAP support structures for various ion exchange reaction concentrations of Co(NO3)2. We identied that Co saturation on the surface of the HAP supports occurs at a 0.1 M concentration of Co(NO3)2 in solution (Fig. 3a). Under these saturation conditions, the amorphous HAP supports contained approximately 0.2 mmoles of Co per gram of CoHAP support. A slightly higher Co content of 0.3 mmoles per gram of cd_CoHAP is observed from Fig. 3b. By lowering the Ca composition during synthesis of the HAP crystals, it is evident that the Ca decient HAP support was capable of exhibiting approximately 50% higher surface coverage of cobalt.

rates of 1.0 mL min1, 1.45 mL min1, and 2.54 mL min1, respectively, per gram of catalytic support (weight of Co and HAP). As can be expected, due to the higher Co content capacity achievable for the Ca decient HAP support, the cd_CoHAP samples exhibited faster hydrogen generation rates in comparison to a_CoHAP. 3.4. Inuence of morphology on hydrogen generation

Using the three HAP samples synthesized in this work, a_HAP, cd_HAP, and s_HAP, Co ion exchange was performed at 0.1 M Co concentration resulting in a_CoHAP, cd_CoHAP, and s_CoHAP for assessment of the morphological dependence on hydrogen generation. We measured the hydrogen generation proles from 0.2 g of sodium borohydride in 20 mL aqueous solution of 1% NaOH exposed to either a 0.1 g sample of Ca decient CoHAP catalyst, amorphous CoHAP catalyst, or the larger single crystal CoHAP catalyst. As can be observed in (Fig. 5), a signicant increase is evident when using Co immobilized single crystal HAP supports as compared to amorphous and calcium decient HAP supports. Given that the same weight amount of 0.1 g of catalytic support was used in each case, the larger sized single crystal HAP supports will have represented a sample with less surface area. Despite this, the single crystal HAP provided signicantly higher hydrogen generation. This is reected from the hydrogen re moree tonditions. a_C-323.6(morAP)]TJ0-1Tf1.4199-2.6247TD[(Usi-30

3.3. Inuence of Co content on hydrogen generation As mentioned in the previous section, catalytic cobalt was incorporated onto the support structure of Ca decient HAP and amorphous HAP samples by ion exchange under several different Co(NO3)2 concentration conditions. Each resulting CoHAP catalytic support was tested to determine the hydrogen generation capabilities. We measured the hydrogen generation proles from 0.2 g of sodium borohydride in 20 mL aqueous solution of 1% NaOH exposed to either a 0.1 g sample of Ca decient CoHAP catalyst (cd_CoHAP) or amorphous CoHAP catalyst (a_CoHAP). Catalysts obtained from ion exchange reactions under higher Co(NO3)2 concentrations exhibited increased catalytic activity for the case of both a_CoHAP (Fig. 4a) and cd_CoHAP (Fig. 4b). We believe the observed induction periods to merely represent the 30 mL deadvolume present in the system, which must be lled before the generated hydrogen could be measured. From Fig. 4, higher Co(NO3)2 concentrations in the ion exchange reaction resulted in catalytic supports that exhibited faster hydrogen generation rates as a result of higher loading of Co active sites on the surface. For a_CoHAP with Co(NO3)2 ion exchange concentrations of 0.01 M, 0.05 M, and 0.1 M (Fig. 4a), we measured, by observing the linear reaction rate over three samples, average hydrogen generation rates of 0.71 mL min1, 1.25 mL min1, and 1.73 mL min1, respectively, per gram of catalytic support, that is the weight of cobalt and HAP combined. This rate represents the average linear region, excluding the induction period or reduced rate phases in the beginning and end of the reaction, respectively. By providing the rates of hydrolysis per weight of the entire catalytic support (Co and HAP), it is clear to see the effect of the calcium decient HAP catalytic support as it offers a proportionally higher amount of Co content as compared to a_HAP. Examining Ca decient CoHAP (cd_CoHAP) catalysts from ion exchange concentrations of 0.01 M, 0.05 M, and 0.1 M Co(NO3)2 (Fig. 4b), we found average hydrogen generation

406

J.W. Jaworski et al. / Journal of Colloid and Interface Science 394 (2013) 401408

Fig. 4. Hydrogen generation measurements showing a dependence on Co content for (a) amorphous HAP (a_CoHAP) ion exchanged with Co(NO3)2 concentrations of (solid) 0.1 M, (triangle) 0.05 M, and (circle) 0.01 M, as well as for (b) Calcium decient HAP (cd_CoHAP) with ion exchange concentrations of (solid) 0.1 M, (triangle) 0.05 M, and (circle) 0.01 M Co(NO3)2. Average values (N = 3) are provided without error bars for the sake of clarity. Experiments were conducted as outlined in Section 2.7.

tion results remains that the HAP possessing a lower Ca content allowed for increased uptake of Co during ion exchange. By comparison of the amorphous HAP and Ca decient HAP as Co catalyst-support structures, we prove the importance in controlling the composition of the crystal support in order to create sufcient active sites for the Co catalyzed hydrolysis of sodium borohydride. As mentioned above, XRD data in conjunction with TEM images reveal the Ca decient HAP and amorphous HAP to be structurally similar, while ICP data provide a signicantly higher degree of Co incorporation on the Ca decient HAP support. From hydrogen generation experiment, we can conclude that this additional Co present for the case of cd_CoHAP is present on the surface of the catalytic support, thereby allowing the higher degree of catalytic activity. After close analysis of various Co ion exchange conditions, we identied the extent of Co incorporation was higher for the calcium decient HAP catalytic support until a point of saturation was reached in which the nal Co loading capacity was in the

range of 0.20.35 mmoles of Co per gram of HAP, which is in agreement with existing Co ion exchange studies [38,46]. This distinction in saturation levels provided a 50% higher number of catalytic Co sites on the Ca decient HAP in contrast to the amorphous HAP allowing its higher catalytic activity. The key point that the cd_HAP supports allowed a higher amount of Co loading makes this catalytic support formulation a signicant improvement as is evident by its higher hydrogen generation rates, which are attributed to the higher loading capacity of Co as compared to the a_HAP. Since examining the hydrogen generation capabilities for the HAP supports as a function of their Co loading conditions reveals a clear increase in the catalytic activity for increased Co content on the support, we can conclude that the increase in the hydrogen generation rate is a direct result of increasing the number of catalytic Co sites available on the surface of the support for sodium borohydride hydrolysis. This result correlates well with an existing

J.W. Jaworski et al. / Journal of Colloid and Interface Science 394 (2013) 401408

407

Fig. 5. Two successive hydrogen generation measurements (a and b) showing a dependence of HAP morphology on the generation rate. a_CoHAP: Amorphous HAP previously ion exchanged with Co(NO3)2 concentrations of 0.1 M (squares), cd_CoHAP: calcium decient HAP ion exchanged with Co(NO3)2 concentrations of 0.1 M (circles), s_CoHAP: single crystalline HAP ion exchanged with Co(NO3)2 concentrations of 0.1 M (triangles). Data are provided for individual samples. Experiments were conducted as outlined in Section 2.7 with time point zero positioned as the onset of measurable hydrogen production.

dehydrogenation study that used Co2+ loaded hydroxyapatite as a heterogeneous catalyst [47]. An interesting assessment of HAP supported catalyst from our prior study using Ru ion exchanged HAP reveals a clear advantage for using Ru over Co with respect

to the hydrogen generation rate from hydrolysis of sodium borohydride. Ru supported HAP reveals a factor of 4 increase in the hydrogen generation rate when normalized by the amount of catalyst (337 mL/(min mmole of Ru)) in comparison with Co (80 mL/ (min mmole of Co)) [21]. Looking into the morphological differences among our synthesized HAP supports, the larger sized single crystal HAP provides a catalytic support with an almost 10 times smaller surface area per gram as compared to the cd_HAP and a_HAP. It is worth noting that ion exchange did not result in any noticeable change in surface area as determined by BET measurements and no observable change in morphology from TEM. While the morphologies of a_HAP and cd_HAP may be indistinguishable, the single crystal HAP, owing to the different synthesis method, possesses a unique morphology of hexagonal rods elongated along the [0 0 1] direction, which results in six equivalent (1 0 0)-like surfaces being exposed to provide the majority of the crystal with a relatively at surface [48,49]. Despite the larger size and correspondingly lower surface area by weight, the single crystal CoHAP catalysts offered signicantly higher hydrogen generation rates over the other supports. It is still unclear if this predominant (1 0 0) surface provides any preferential sites for higher density Co loading or improved arrangement of active catalyst sites as compared to the exposed planes of the amorphous or calcium decient HAP surface. Recent studies have shown that, by providing a sufciently short distance between active sites, reaction rates of heterogeneous catalysts can be altered by morphological control of the exposed reactive crystal planes with preferential transition-metal active sites [21,50,51]. While we cannot assume the (1 0 0) s_CoHAP surface is more suitable for hydrogen formation, we can conclude that having a particular crystal morphology as well as composition control can provide a successful means for increasing the hydrogen generation rates among CoHAP catalytic supports. From our long-term stability measurements, we observed a pattern of gradual deactivation of the catalyst over repeated daily use, falling to nearly 75% over a 3 week period. Each of the three catalytic support structures exhibited similar deactivation over the course of the 20 day experiment. To attempt to determine the cause of the deactivation, we observed the samples before and after prolonged sodium borohydride exposure using a variety of

Fig. 6. Stability experiment of 0.1 g of amorphous CoHAP, Ca decient CoHAP, and single crystal CoHAP catalyst (previously ion exchanged with 0.1 M Co(NO3)2 and using the experimental conditions in Section 2.8). The percentage completion of the hydrolysis reaction from repeated exposures to 20 mL of 1% sodium borohydride in 1% NaOH decreased over time during the 3 week experiment. Data are provided for individual samples.

408

J.W. Jaworski et al. / Journal of Colloid and Interface Science 394 (2013) 401408

analytical tools. BET measurements revealed no signicant change in the surface area, indicating that surface erosion did not play a role. XPS analysis revealed that the Co(II) oxidation remained unchanged after deactivation with no relevant differences in the spectra before and after hydrogen generation, see Supplementary Fig. 2 for full details and analysis. We did however nd a noticeable difference in the FTIR spectra when comparing each of the catalytic supports (before and after 3 weeks of sodium borohydride hydrolysis). Specically, a characteristic BO stretching band was observed after hydrogen generation, as seen in Supplementary Fig. 1. Since boron oxide could not be observed by XPS or SEM, the exact cause of the gradual deactivation remains unclear. Although given the FTIR results, the formation of borate by-product may be plausible, as it is often known to block the catalytic surface in sodium borohydride hydrolysis reactions [31]. 5. Conclusion Since cobalt appears naturally, due to mining and coal combustion, it is readily available in large amounts. This attribute favors its use in the catalytic generation of hydrogen over expensive and rare noble metal based catalysts. By immobilizing Co onto a hydroxyapatite support, we proved that the CoHAP provides an effective catalytic support for the hydrolysis of sodium borohydride resulting in the useful generation of hydrogen. By further examining various forms of hydroxyapatite based on Ca composition and morphology, we revealed benecial strategies for increasing the catalytic activity of these support structures for the generation of hydrogen. Specically, we found that Ca decient HAP provides a 50% higher loading capacity of Co under saturated ion exchange conditions. Through this, we reveal a higher number of catalytically active surface sites useful for increasing the hydrogen generation rates. In addition, we nd that the single crystal HAP morphology affords increased hydrogen generation rates. Accordingly, improvements in the HAP support structures can be realized through altering the synthesis conditions of Ca composition in hydroxyapatite as well as the synthesis method to alter the morphology, which dramatically affects the activity of a CoHAP catalytic support. Through our 3 week course of repeatability studies, we found that each form of HAP possessed similar catalytic stabilities to repeated exposure to sodium borohydride solutions decreasing to approximately 75% of the original activity. Future endeavors to increase the reusability of a CoHAP based hydrogen generation system may prove valuable in creating a commercially applicable catalytic support given the high catalytic activity, the ease of fabrication, and the abundance of starting materials. Acknowledgments This work was supported by the National Research Foundation of Korea (NRF-2010-0006157) and by the research fund of Hanyang University (HY-2011-N). This work was also supported by the Priority Research Centers Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012R1A6A1029029). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2012.11.036.

References
[1] J.A. Turner, Science 305 (2004) 972. [2] X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J.M. Carlsson, K. Domen, M. Antonietti, Nat. Mater. 8 (2009) 76. [3] X.H. Hao, L.J. Guo, X. Mao, X.M. Zhang, X.J. Chen, Int. J. Hydrogen Energy 28 (2003) 55. [4] H.B. Dai, Y. Liang, L.P. Ma, P. Wang, J. Phys. Chem. C 112 (2008) 15886. [5] N.S. Lewis, Nature 414 (2001) 589. [6] P.M. Grant, Nature 424 (2003) 129. [7] V.A. Goltsov, T.N. Veziroglu, L.F. Goltsova, M. Ni, D.Y.C. Leung, M.K.H. Leung, K. Sumathy, Int. J. Hydrogen Energy 31 (2006) 153. [8] M. Ni, D.Y.C. Leung, M.K.H. Leung, Int. J. Hydrogen Energy 32 (2007) 3238. [9] S.C. Marinescu, J.R. Winkler, H.B. Gray, Proc. Nat. Acad. Sci. 109 (2012) 15127. [10] H. Kato, M. Hori, R. Konta, Y. Shimodaira, A. Kudo, Chem. Lett. 33 (2004) 1348. [11] Z. Zou, J. Ye, K. Sayama, H. Arakawa, Nature 414 (2001) 625. [12] T. Hugle, M.F. Kuhnel, D. Lentz, J. Am. Chem. Soc. 131 (2009) 7444. [13] L. Schlapbach, A. Zuttel, Nature 414 (2001) 353. [14] U.B. Demirci, O. Akdim, J. Andrieux, J. Hannauer, R. Chamoun, P. Miele, Fuel Cells 10 (2010) 335. [15] T.J. Groshens, R.A. Hollins, Chem. Commun. (2009) 3089. [16] D.J. Collins, H.-C. Zhou, J. Mater. Chem. 17 (2007) 3154. [17] J.C. Walter, A. Zurawski, D. Montgomery, M. Thornburg, S. Revankar, J. Power Sources 179 (2008) 335. [18] S.C. Amendola, S.L. Sharp-Goldman, M. Saleem Janjua, M.T. Kelly, P.J. Petillo, M. Binder, J. Power Sources 85 (2000) 186. [19] A.M. Seayad, D.M. Antonelli, Adv. Mater. 16 (2004) 765. [20] D. Xu, H. Zhang, W. Ye, Catal. Commun. 8 (2007) 1767. [21] J.W. Jaworski, D. Kim, K. Jung, S. Kim, J.H. Jung, J.O. Jeong, H.S. Jeon, B.K. Min, K.Y. Kwon, J. Colloid Interface Sci. 358 (2011) 598. [22] M. Rakap, S. zkar, Int. J. Hydrogen Energy 36 (2011) 7019. [23] D. elik, S. Karahan, M. Zahmakran, S. zkar, Int. J. Hydrogen Energy 37 (2012) 5143. [24] B.H. Liu, Z.P. Li, J. Power Sources 187 (2009) 527. [25] D. Xu, P. Dai, X. Liu, C. Cao, Q. Guo, J. Power Sources 182 (2008) 616. [26] W. Ye, H. Zhang, D. Xu, L. Ma, B. Yi, J. Power Sources 164 (2007) 544. [27] B.H. Liu, Z.P. Li, S. Suda, J. Alloy. Compd. 415 (2006) 288. [28] C. Wu, F. Wu, Y. Bai, B. Yi, H. Zhang, Mater. Lett. 59 (2005) 1748. [29] M. Rakap, S. zkar, Catal. Today 183 (2012) 17. [30] O. Akdim, U.B. Demirci, P. Miele, Int. J. Hydrogen Energy 36 (2011) 13669. [31] U.B. Demirci, P. Miele, Phys.Chem. Chem. Phys. 12 (2010) 14651. [32] S.S. Muir, X. Yao, Int. J. Hydrogen Energy 36 (2011) 5983. [33] J.-H. Wee, K.-Y. Lee, S.H. Kim, Fuel Process. Technol. 87 (2006) 811. [34] M. Zahmakiran, S. -zkar, J. Mater. Chem. 19 (2009) 7112. [35] J.M. Planeix, N. Coustel, B. Coq, V. Brotons, P.S. Kumbhar, R. Dutartre, P. Geneste, P. Bernier, P.M. Ajayan, J. Am. Chem. Soc. 116 (1994) 7935. [36] H.-B. Dai, G.-L. Ma, X.-D. Kang, P. Wang, Catal. Today 170 (2011) 50. [37] M. Wakamura, K. Kandori, T. Ishikawa, Colloids Surf. A: Physicochem. Eng. Asp. 142 (1998) 107. [38] I. Smiciklas, S. Dimovic, I. Plecas, M. Mitric, Water Res. 40 (2006) 2267. [39] X. Deng, J. Hao, C. Wang, Biomaterials 22 (2001) 2867. [40] A. Siddharthan, S.K. Seshadri, T.S.S. Kumar, J. Mater. Sci. Mater. Med. 15 (2004) 1279. [41] Z. Honghui, L. Hui, G. Linghong (Eds.), Key Engineering Materials, 2007, p. 119. [42] H.-b. Zhang, K.-c. Zhou, Z.-y. Li, S.-p. Huang, J. Phys. Chem. Solids 70 (2009) 243. [43] A. Lpez-Macipe, R. Rodrguez-Clemente, A. Hidalgo-Lpez, I. Arita, M.V. Garca-Garduo, E. Rivera, V.M. Castao, J. Mater. Synth. Process. 6 (1998) 21. [44] A. Osaka, Y. Miura, K. Takeuchi, M. Asada, K. Takahashi, J. Mater. Sci. Mater. Med. 2 (1991) 51. [45] T. Tsuchida, J. Kubo, T. Yoshioka, S. Sakuma, T. Takeguchi, W. Ueda, J. Catal. 259 (2008) 183. [46] T. Suzuki, T. Hatsushika, M. Miyake, J. Chem. Soc., Faraday Trans. 1: Phys. Chem. Condensed Phases 78 (1982) 3605. [47] K. Elkabouss, M. Kacimi, M. Ziyad, S. Ammar, F.i. Bozon-Verduraz, J. Catal. 226 (2004) 16. [48] K.-Y. Kwon, E. Wang, A. Chung, N. Chang, S.-W. Lee, J. Phys. Chem. C 113 (2009) 3369. [49] K.-Y. Kwon, E. Wang, N. Chang, S.-W. Lee, Langmuir 25 (2009) 7205. [50] B. Lim, M. Jiang, P.H.C. Camargo, E.C. Cho, J. Tao, X. Lu, Y. Zhu, Y. Xia, Science 324 (2009) 1302. [51] X. Xie, Y. Li, Z.-Q. Liu, M. Haruta, W. Shen, Nature 458 (2009) 746.