Rch

Sa

b

a

ARRAA

KRMAD

1

redsbcataa

(

h1

Journal of Molecular Catalysis A: Chemical 394 (2014) 253–261

Contents lists available at ScienceDirect

Journal of Molecular Catalysis A: Chemical

jou rn al hom epage: www.elsev ier .com/ locate /molcata

uthenium(0) nanoparticles supported on magnetic silica coatedobalt ferrite: Reusable catalyst in hydrogen generation from theydrolysis of ammonia-borane

erdar Akbayraka, Murat Kayab,∗, Mürvet Volkana, Saim Özkara,∗∗

Department of Chemistry, Middle East Technical University, 06800 Ankara, TurkeyDepartment of Chemical Engineering and Applied Chemistry, Atılım University, 06836 Ankara, Turkey

r t i c l e i n f o

rticle history:eceived 17 March 2014eceived in revised form 27 May 2014ccepted 5 July 2014vailable online 22 July 2014

eywords:uthenium nanoparticlesagnetic separation

mmonia-boraneehydrogenation

a b s t r a c t

Ruthenium(0) nanoparticles supported on magnetic silica-coated cobalt ferrite (Ru(0)/SiO2-CoFe2O4)were in situ generated from the reduction of Ru3+/SiO2-CoFe2O4 during the catalytic hydrolysis ofammonia-borane (AB). Ruthenium(III) ions were impregnated on SiO2-CoFe2O4 from the aqueous solu-tion of ruthenium(III) chloride and then reduced by AB at room temperature yielding Ru(0)/SiO2-CoFe2O4

which were isolated from the reaction solution by using a permanent magnet and characterized by ICP-OES, XRD, TEM, TEM-EDX and XPS techniques. The resulting magnetically isolable Ru(0)/SiO2-CoFe2O4

were found to be highly reusable catalyst in hydrolysis of AB retaining 94% of their initial catalytic activ-ity even after tenth run. Ru(0)/SiO2-CoFe2O4 provide the highest catalytic activity after the tenth use inhydrolysis of AB as compared to the other ruthenium catalysts. The work reported here also includes theformation kinetics of ruthenium(0) nanoparticles. The evaluation of rate constants for the nucleation andautocatalytic surface growth of ruthenium(0) nanoparticles at various temperatures provides the estima-

tion of activation energy for both reactions; Ea = 116 ± 7 kJ/mol for the nucleation and Ea = 51 ± 2 kJ/molfor the autocatalytic surface growth of ruthenium(0) nanoparticles. The report also includes the activa-tion energy of the catalytic hydrogen generation from the hydrolysis of AB (Ea = 45 ± 2 kJ/mol) determinedfrom the evaluation of temperature dependent kinetic data and the effect of catalyst concentration onthe rate of hydrolysis of AB.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Hydrogen is regarded as one of the best alternative energy car-iers to satisfy the increasing demand for a sustainable and cleannergy supply [1–3]. A great deal of research efforts has beenevoted to find suitable hydrogen storage materials ensuring theafe and economical way [4,5]. Ammonia-borane (NH3BH3, AB) haseen identified as one of the most attractive candidates for chemi-al hydrogen storage due to its high hydrogen content, nontoxicitynd high stability at room temperature [6–11]. AB can release H2 byhermolysis [12] or solvolysis [13,14]. However, the search for suit-

ble catalysts for the release of hydrogen from the hydrolysis of ABt an appreciable rate is crucial (Eq. (1)). Therefore, a large number

∗ Corresponding author. Tel.: +90 312 586 8561; fax: +90 312 586 8091.∗∗ Corresponding author. Tel.: +90 312 210 3212; fax: +90 312 210 3200.

E-mail addresses: muratkaya@atilim.edu.tr (M. Kaya), sozkar@metu.edu.trS. Özkar).

ttp://dx.doi.org/10.1016/j.molcata.2014.07.010381-1169/© 2014 Elsevier B.V. All rights reserved.

of studies have been published using transition metal nanoparticlesas catalysts for the hydrogen generation from AB [15,16].

H3NBH3(aq) + 2H2O(l)catalyst−→ NH+

4 (aq) + BO−2 (aq) + 3H2(g) (1)

Among the transition metal nanoparticles, ruthenium is oneof the most active catalysts for the hydrogen generation fromthe hydrolysis of AB under mild conditions. Although there exista wealth of reports on the hydrogenation of AB using ruthe-nium catalysts such as Ru(0)NP/PSSA-co-MA [17], Ru(0)@Hap [18],Ru/Carbon [19], RuNPs@ZK-4 [20], Ru@Al2O3 [21], Ru(0)NP/laurate[22], Ru(0)@MWCNTs [23] and Ru/C [24], there is no exampleof magnetically recoverable ruthenium catalyst for the hydrogenrelease from AB. Recently, much attention has been paid to themagnetically recoverable catalysts in liquid phase reactions due totheir easy magnetic separation making the recovery of catalysts

much easier than by filtration and centrifugation [25].

Herein, we report ruthenium(0) nanoparticles supported onsilica-coated cobalt ferrite (SiO2-CoFe2O4) as magnetically isolableand recyclable catalyst for the hydrolysis of ammonia-borane.

2 r Cata

RcRfmtuatfaTckhtst

2

2

ncaDScw

2

wSs(sm1yeAaBf

2(

c[t[2thtw8bsw

54 S. Akbayrak et al. / Journal of Molecula

uthenium(0) nanoparticles supported on magnetic silica-coatedobalt ferrite were in situ generated from the reduction ofu3+/SiO2-CoFe2O4 during the catalytic hydrolysis of AB. Cobalt

errite (CoFe2O4) nanoparticles were preferred as magnetic coreaterials due to the easy preparation procedure as compared to

he other magnetic core materials such as Fe3O4 [26]. SiO2 wassed to protect the magnetic core material against leaching andgglomeration and it also provides high surface area to stabilizehe ruthenium nanoparticles. Ru(0)/SiO2-CoFe2O4 were isolatedrom the reaction solution by using a permanent magnet and char-cterized by ICP-OES, XRD, TEM, TEM-EDX and XPS techniques.he results reveal that highly dispersed ruthenium(0) nanoparti-les were successfully supported on SiO2-CoFe2O4. The formationinetics of ruthenium(0) nanoparticles was studied by using theydrogen evolution from the hydrolysis of AB as reporter reac-ion. The work reported here also includes the results of kinetictudy of the hydrolytic dehydrogenation of AB depending on theemperature and catalyst concentration.

. Experimental

.1. Materials

Iron(III) chloride (FeCl3), tetraethylorthosilicate (TEOS), ammo-ium hydroxide (NH4OH), sodium hydroxide (NaOH), cobalt(II)hloride (CoCl2), Ruthenium(III) chloride trihydrate (RuCl3·3H2O)nd ammonia-borane (H3NBH3, 97%) were purchased from Aldrich.eionized water was distilled by water purification system (Milli-Qystem). All glassware and Teflon-coated magnetic stir bars wereleaned with acetone, followed by copious rinsing with distilledater before drying in an oven at 150 ◦C.

.2. Characterization

The ruthenium contents of the Ru(0)/SiO2-CoFe2O4 samplesere determined by Inductively Coupled Plasma Optical Emission

pectroscopy (ICP-OES, Leeman-Direct Reading Echelle). Transmis-ion electron microscopy (TEM) was performed on a JEM-2100FJEOL) microscope operating at 200 kV. A small amount of powderample was placed on a copper grid of the transmission electronicroscope. Samples were examined at magnification between

00 and 400 K. The X-ray photoelectron spectroscopy (XPS) anal-sis was performed on a Physical Electronics 5800 spectrometerquipped with a hemispherical analyzer and using monochromaticl K� radiation of 1486.6 eV, the X-ray tube working at 15 kV, 350 Wnd pass energy of 23.5 keV. 11B NMR spectra were recorded on aruker Avance DPX 400 with an operating frequency of 128.15 MHz

or 11B.

.3. Preparation of magnetic silica-coated cobalt ferriteSiO2-CoFe2O4)

The preparation of magnetic cobalt ferrite nanoparticles wasarried out by modification of previously established procedure27]. The detailed information on the preparation and charac-erization of silica-coated cobalt ferrite can be found elsewhere28]. In a typical experiment 25 mL of 0.4 M iron(III) chloride and5 mL of 0.2 M of cobalt(II) chloride solutions were mixed at roomemperature. Then, in a separate vessel 25 mL of 3.0 M sodiumydroxide solution was prepared and slowly added to the salt solu-ion. After complete addition of NaOH solution, a black suspensionas obtained. The mechanical stirring was continued for 1 h at

0 ◦C. Then the solution was cooled to room temperature and thelack precipitates were collected by using an external magnet. Theupernatant was removed and the particles were washed 3 timesith deionized water–ethanol solution and then the particles were

lysis A: Chemical 394 (2014) 253–261

dispersed in 50 mL of water. Silica coating was applied by using amodified version of Stober method [29]. For the preparation of sil-ica coating, 200 mL ethanol, 1 mL TEOS and 0.5 mL of NH4OH wereadded to the reaction mixture and subsequently 50 mL cobalt ferritecolloid was added to the mixture and the mixture was stirred for 4 hat room temperature. After the formation of the thick silica shell,particles were collected with a magnet and washed 3 times withdeionized water. The resulting silica-coated cobalt ferrite nanopar-ticles (SiO2-CoFe2O4) were separated by using a permanent magnetand washed with excess ethanol and dried at 120 ◦C for 12 h in theoven.

2.4. Impregnation of ruthenium(III) ions on magneticsilica-coated cobalt ferrite [Ru3+/SiO2-CoFe2O4]

SiO2-CoFe2O4 (100 mg) was added to a solution of RuCl3·3H2O(5.65 mg) in 20 mL H2O in a 50 mL beaker. This slurry was stirredat room temperature for 12 h and then, all supernatant solutionwas removed by using a permanent magnet. Next, the resultingparticles Ru3+/SiO2-CoFe2O4 were washed with 20 mL of deionizedwater and isolated by using a permanent magnet and the remnantwas dried at 120 ◦C for 12 h in the oven.

2.5. In situ formation of ruthenium(0) nanoparticles supportedon magnetic silica-coated cobalt ferrite [Ru(0)/SiO2-CoFe2O4] andconcomitant catalytic hydrolysis of AB

Ruthenium(0) nanoparticles supported on magnetic silica-coated cobalt ferrite were in situ generated from the reductionof Ru3+/SiO2-CoFe2O4 during the catalytic hydrolysis of AB.Before starting the catalyst formation and concomitant catalytichydrolysis of AB, a jacketed reaction flask (20 mL) containing aTeflon-coated stir bar was placed on a magnetic stirrer (HeidolphMR-301) and thermostated to 25.0 ± 0.1 ◦C by circulating waterthrough its jacket from a constant temperature bath. Then, a gradu-ated glass tube (60 cm in height and 3.0 cm in diameter) filled withwater was connected to the reaction flask to measure the volume ofthe hydrogen gas to be evolved from the reaction. Next, 10 mg pow-der of Ru3+/SiO2-CoFe2O4 (1.96 wt.% Ru) was dispersed in 10 mLdistilled water in the reaction flask thermostated at 25.0 ± 0.1 ◦C.Then, 31.8 mg AB (1.0 mmol H3NBH3) was added into the flaskand the reaction medium was stirred at 1000 rpm. After addingammonia-borane, ruthenium(0) nanoparticles were formed andthe catalytic hydrolysis of AB started immediately. The volume ofhydrogen gas evolved was measured by recording the displace-ment of water level every 30 s at constant atmospheric pressure of693 Torr. The reaction was stopped when no more hydrogen evo-lution was observed. In each experiment, the resulting solutionswere filtered and the filtrates were analyzed by 11B NMR and con-version of AB to metaborate anion was confirmed by comparing theintensity of signals in the 11B NMR spectra of the filtrates.

2.6. Determination of activation energy for hydrolysis of ABcatalyzed by Ru(0)/SiO2-CoFe2O4

In a typical experiment, the hydrolysis reaction was performedstarting with 10 mL of 100 mM (31.8 mg) AB solution and 10 mgRu3+/SiO2-CoFe2O4 (1.96 wt.% ruthenium, [Ru] = 0.186 mM) at vari-ous temperatures (25, 30, 35, 40 ◦C) in order to obtain the activationenergy.

2.7. Reusability of Ru(0)/SiO2-CoFe2O4 in hydrolysis of AB

After the complete hydrolysis of AB started with 10 mL of100 mM AB (31.8 mg H3NBH3), and 60 mg Ru(0)/(SiO2-CoFe2O4)

S. Akbayrak et al. / Journal of Molecular Catalysis A: Chemical 394 (2014) 253–261 255

Fig. 1. The evolution of equivalent H2 per mole of AB versus time plot for theha

(lwawApA

3

3d

cRnaCdptAtapmimR2

aepttkurri

w

Scheme 1. Illustration of hydrolysis of ammonia-borane as reporter reaction: P isthe precursor ruthenium(III) impregnated on silica-coated cobalt ferrite and Q is thegrowing Ru(0)n nanoparticles.

Table 1The rate constants k1 of the slow, continuous nucleation, P → Q, and k2 of the autocat-alytic surface growth, P + Q → 2Q for the formation of ruthenium(0) nanoparticlescatalyst from the reduction of ruthenium(III) ions during the hydrolysis of AB atvarious temperatures.

Temp. (◦C) k1 (s−1) k2 (M−1 s−1) k2/k1

25 6.95 × 10−5 ± 8.73 × 10−6 1.41 ± 4.59 × 10−2 2.03 × 104

30 9.33 × 10−5 ± 1.02 × 10−5 2.32 ± 6.26 × 10−2 2.48 × 104

35 2.25 × 10−4 ± 3.12 × 10−5 3.01 ± 1.21 × 10−1 1.34 × 104

ydrolysis of AB starting with Ru3+/SiO2-CoFe2O4 (0.194 mM Ru) and 100 mM ABt 25.0 ± 0.1 ◦C.

1.96 wt.% ruthenium, [Ru] = 0.744 mM) at 25.0 ± 0.1 ◦C, the cata-yst was isolated using a permanent magnet. Ru(0)/SiO2-CoFe2O4

ere magnetically attracted to the bottom of the reaction vessel by magnet, and the upper solution was removed and the catalyst wasashed with 10 mL of water before every run in the reusability test.fter washing, the catalyst was isolated again and the isolated sam-le of Ru(0)/SiO2-CoFe2O4 redispersed in 10 mL solution of 100 mMB for a subsequent run of hydrolysis at 25.0 ± 0.1 ◦C.

. Results and discussion

.1. Formation kinetics of ruthenium(0) nanoparticles catalysturing hydrolysis of AB

Ruthenium(0) nanoparticles supported on magnetic silica-oated cobalt ferrite were in situ generated from the reduction ofu3+/SiO2-CoFe2O4 during the catalytic hydrolysis of AB. Ruthe-ium(III) ions were first impregnated on SiO2-CoFe2O4 from thequeous solution of ruthenium(III) chloride yielding Ru3+/SiO2-oFe2O4 and then reduced by AB at room temperature. Asescribed elsewhere [18], when AB solution is added to the sus-ension of Ru3+/SiO2-CoFe2O4, both reduction of ruthenium(III)o ruthenium(0) and hydrogen release from the hydrolysis ofB occur concomitantly. The progress of ruthenium(0) nanopar-

icles formation and concomitant hydrolytic dehydrogenation ofmmonia-borane was followed by monitoring the change in H2ressure which was then converted into the equivalent H2 perole of AB, using the known 3:1 H2/AB stoichiometry as given

n Eq. (1). Fig. 1 shows the evolution of equivalent hydrogen perole of AB versus time plot for the hydrolysis of AB starting with

u3+/SiO2-CoFe2O4 precatalyst (0.194 mM Ru) and 100 mM AB at5.0 ± 0.1 ◦C.

After a short induction period of 4.0 min the hydrogen gener-tion starts and continues almost linearly until the release of 3quivalents H2 per mole of AB. The observation of an inductioneriod and a sigmoidal shape of dehydrogenation curve indicatehe formation of ruthenium(0) nanoparticles with a 2-step, nuclea-ion and autocatalytic growth mechanism [30,31]. The formationinetics of the ruthenium(0) nanoparticles catalyst can be obtainedsing the hydrogen release from the hydrolysis of AB as reportereaction [30–32] given in Scheme 1 whereby P is the precursor

uthenium(III) impregnated on silica-coated cobalt ferrite and Qs the growing Ru(0)n nanoparticles.

Monitoring the hydrogen evolution from the hydrolysis of ABill accurately report on and amplifies the amount of Ru(0)n

40 6.39 × 10−4 ± 7.87 × 10−5 3.88 ± 1.78 × 10−1 6.06 × 103

nanoparticles catalyst, Q, present if the hydrogen generationrate is fast in comparison to the rate of nanoparticle forma-tion. The observation of a sigmoidal dehydrogenation curve inFig. 1 and its curve-fit to the slow, continuous nucleation, P → Q(rate constant k1) followed by autocatalytic surface growth,P + Q → 2Q (rate constant k2) kinetics are very strong evidencefor the formation of metal(0) nanoparticles catalyst from a solu-ble transition–metal complex in the presence of reducing agent[32]. The rate constants determined from the nonlinear leastsquares curve-fit in Fig. 1 are k1 = 6.95 × 10−5 ± 8.73 × 10−6 s−1 andk2 = 1.41 ± 4.59 × 10−2 M−1 s−1 (the mathematically required cor-rection has been made to k2 for the stoichiometry factor of 515 asdescribed elsewhere [31], but not for the “scaling factor”; that is, nocorrection has been made for changing the number of rutheniumatoms on the growing metal surface). These rate constants are forthe continuous slow nucleation and autocatalytic growth of Ru(0)n

nanoparticles, respectively, starting with Ru3+/SiO2-CoFe2O4 pre-catalyst (0.194 mM Ru) and ammonia-borane (100 mM) in aqueoussolution at 25.0 ± 0.1 ◦C.

Fig. 2a shows the evolution of equivalent hydrogen per moleof AB versus time plot for the hydrolysis starting with Ru3+/SiO2-CoFe2O4 precatalyst (0.194 mM Ru) and 100 mM AB at fourdifferent temperatures. For each temperature, experimental datacurve-fit well to the 2-step mechanism, giving the rate constantsk1 of the slow, continuous nucleation, P → Q, and k2 of the autocat-alytic surface growth, P + Q → 2Q for the formation of ruthenium(0)nanoparticles catalyst from the ruthenium(III) ions during thehydrolysis of ammonia-borane (Table 1). The large value of k2/k1ratio (Table 1) is indicative of the high level kinetic control inthe formation of ruthenium(0) nanoparticles from the reductionof the precursor ruthenium(III) ions on the surface of silica-coatedcobalt ferrite [32]. From the Arrhenius plots constructed by usingthe values of rate constants k1 and k2 at various temperatures inFig. 2b and c, respectively, one can obtain the activation energyEa = 116 ± 7 kJ/mol for the nucleation and Ea = 51 ± 2 kJ/mol for theautocatalytic surface growth of ruthenium(0) nanoparticles. It isnoteworthy that the k1 and k2 values can give an idea on theenergy barrier for the slow nucleation and autocatalytic surfacegrowth of metal(0) nanoparticles catalyst and that the large values

of k2/k1 ratio is indicative of the kinetic control of nanoparticlesformation.

256 S. Akbayrak et al. / Journal of Molecular Catalysis A: Chemical 394 (2014) 253–261

Fig. 2. (a) The evolution of equivalent hydrogen per mole of AB versus time plotfor the hydrolysis of AB starting with Ru3+/SiO2-CoFe2O4 precatalyst (0.194 mMRu) and 100 mM AB at various temperatures. (b) The Arrhenius plot fornucleation of ruthenium(0) nanoparticles (In k1 = −13947.9(1/T) + 37.02). (c) TheArrhenius plot for the autocatalytic surface growth of ruthenium(0) nanoparticles(In k2 = −6129.25(1/T) + 20.98).

Fig. 3. Powder XRD patterns of (a) silica-coated cobalt ferrite, (b) Ru3+/SiO2-CoFe2O4

and (c) Ru(0)/SiO2-CoFe2O4, in situ generated during the hydrolysis of AB, with1.96 wt.% Ru loading.

3.2. Isolation and characterization of ruthenium(0) nanoparticlessupported on magnetic silica-coated cobalt ferrite

Ruthenium(0) nanoparticles supported on magnetic silica-coated cobalt ferrite, Ru(0)/SiO2-CoFe2O4, in situ generated duringthe hydrolysis of AB, could be isolated from the reaction solutionas powder by using a permanent magnet and characterized by ICP-OES, XRD, TEM, TEM-EDX, and XPS techniques. Ruthenium contentof Ru(0)/SiO2-CoFe2O4 was determined by ICP-OES. XRD pattern ofRu3+/SiO2-CoFe2O4 and Ru(0)/SiO2-CoFe2O4 in Figs. 3b and c givepeaks at 30.3◦, 35.8◦, 43.3◦, 57.4◦ and 62.5◦ assigned to the (220),(311), (400), (511) and (440) reflections of SiO2-CoFe2O4, respec-tively (PDF Card #22-1086). The comparison of the XRD patternsof SiO2-CoFe2O4, Ru3+/SiO2-CoFe2O4 and Ru(0)/SiO2-CoFe2O4 witha ruthenium loading of 1.96 wt.% Ru, given in Fig. 3a–c, respec-tively, clearly shows that there is no change in the characteristicdiffraction peaks of silica-coated cobalt ferrite, (SiO2-CoFe2O4). Thisobservation indicates that the host material remains intact afterimpregnation and reduction of Ru3+ ions without noticeable alter-ation in the framework lattice or loss in the crystallinity. There is noobservable peak attributable to ruthenium nanoparticles in Fig. 3aand b, probably as a result of low ruthenium loading of silica-coatedcobalt ferrite nanoparticles.

Fig. 4 shows the TEM images of silica-coated cobalt ferrite andRu(0)/SiO2-CoFe2O4 with a 1.96 wt.% Ru loading taken with differ-ent magnifications. The size of SiO2-CoFe2O4 nanoparticles used assupport is around 15 nm (Fig. 4a and b and Figure S1) and highlydispersed ruthenium nanoparticles are formed on the silica-coatedcobalt ferrite (Fig. 4d) as seen from the comparison of the images inFig. 4b and d taken from the area indicated with an arrow in Fig. 4aand c, respectively. TEM-EDX spectrum (Fig. 4e and Figure S2) takenfrom the Ru(0)/SiO2–CoFe2O4 and SiO2–CoFe2O4 shown in Fig. 4dand b indicates that ruthenium is the only element detected inthe sample in addition to the framework elements of silica-coatedcobalt ferrite (Si, O, Co, Fe). Impregnation of ruthenium(III) followedby reduction to ruthenium(0) causes no change in the framework

lattice of the silica-coated cobalt ferrite in agreement with the XRDresults.

The composition of Ru(0)/(SiO2-CoFe2O4) formed in situ dur-ing the hydrolysis of AB and the oxidation state of ruthenium was

S. Akbayrak et al. / Journal of Molecular Catalysis A: Chemical 394 (2014) 253–261 257

Fig. 4. TEM images of silica-coated cobalt ferrite with the scale bar of (a) 10 nm,(b) 5 nm and TEM images of Ru(0)/SiO2-CoFe2O4 with 1.96 wt.% Ru loading with thescale bar of (c) 10 nm, (d) 5 nm and (e) TEM-EDX spectrum of Ru(0)/SiO2-CoFe2O4.

Fig. 4. (Continued ).

also studied by XPS technique. The survey-scan XPS spectrum ofRu(0)/(SiO2-CoFe2O4) with 1.96 wt.%. Ru loading (Fig. 5a) showsall the framework elements of ruthenium(0) nanoparticles sup-ported on magnetic silica-coated cobalt ferrite in agreement withthe TEM-EDX result. High resolution X-ray photoelectron spectrumof a Ru(0)/(SiO2-CoFe2O4) sample given in Fig. 5b shows two promi-nent bands at 484.9 eV and 462.5 eV which can readily be assignedto Ru(0) 3p1/2 and 3p3/2, respectively [33].

3.3. Hydrolysis of ammonia-borane catalyzed by ruthenium(0)nanoparticles supported on magnetic silica-coated cobalt ferrite

Before starting with the investigation on the catalytic activityof Ru(0)/SiO2-CoFe2O4 in the hydrolysis of AB, a control experi-ment was performed to check whether SiO2-CoFe2O4 shows anycatalytic activity in the hydrolysis of AB at temperatures in therange 25–40 ◦C. In a control experiment starting with 1.0 mmolof AB and 10 mg of powder of SiO2-CoFe2O4 (the same amountas the one used in catalytic activity tests) in 10 mL of water, nohydrogen generation was observed in 1 h. This observation indi-cates that the hydrolysis of AB does not occur in the presence ofSiO2-CoFe2O4 in the temperature range used in this study. On theother hand, ruthenium(0) nanoparticles supported on magneticsilica-coated cobalt ferrite are highly active catalyst in the hydrol-ysis of ammonia-borane generating 3.0 equivalent H2 gas per molof AB. It is noteworthy that the release of ammonia was checkedduring the hydrolysis of AB in all the experiments performed inthis study following the procedure described elsewhere [34] and noammonia evolution was detected in any of the experiments. Fig. 6ashows the plots of equivalent H2 gas generated per mole of H3NBH3versus time during the catalytic hydrolysis of 100 mM AB solutionstarting with Ru(0)/SiO2-CoFe2O4 with 1.96 wt.% Ru loading in dif-ferent catalyst concentration at 25.0 ± 0.1 ◦C. In each experiment,hydrogen evolution starts after a short induction period of less than12 min and continues almost linearly until the complete conversionof the substrate giving 3.0 equivalents of H2 per mole of AB. All theexperimental data fit well to sigmoidal curve according to the two-step mechanism, which provides the rate constants k1 of the slow,continuous nucleation, P → Q, and k2 of the autocatalytic surfacegrowth, P + Q → 2Q for the formation of ruthenium(0) nanoparti-cles catalyst from the ruthenium(III) ions during the hydrolysis of

ammonia-borane (Table 2) [32]. The rate constant k1 for nucleationincreases while the rate constant k2 for surface growth decreaseswith the increasing concentration of ruthenium as seen in Table 2.

258 S. Akbayrak et al. / Journal of Molecular Catalysis A: Chemical 394 (2014) 253–261

Table 2The rate constants k1 of the slow, continuous nucleation, P → Q, and k2 of the autocatalytic surface growth, P + Q → 2Q for the formation of ruthenium(0) nanoparticles catalystfrom the ruthenium(III) ions during the hydrolysis of AB at various concentration of ruthenium.

[Ru] mM k1 (s−1) k2 (M−1 s−1) Induction time (s) TOF (min−1)

0.097 1.20 × 10−5 ±1.15 × 10−6 1.65 ± 3.11 × 10−2 720 172.50.194 6.95 × 10−5 ±8.73 × 10−7 1.41 ± 4.59 × 10−2 240 143.540.291 9.70 × 10−5 ±1.14 × 10−5 1.19 ± 3.73 × 10−2 120 119.620.388 1.52 × 10−4 ±2.02 × 10−4 0.98 ± 3.84 × 10−2 60 100.48

F1

ptgac

gdopcsc

Fig. 6. (a) mol H2/mol H3NBH3 versus time graph depending on the ruthenium con-centration in Ru(0)/SiO2-CoFe2O4 for the hydrolysis of AB (100 mM) at 25.0 ± 0.1 ◦C.

ig. 5. (a) X-ray photoelectron (XPS) spectrum of Ru(0)/SiO2-CoFe2O4 sample with.96 wt.% Ru loading, (b) The high resolution scan Ru 3p bands.

The hydrogen generation rate was determined from the linearortion of each plot in Fig. 6a and plotted versus the initial concen-ration of ruthenium, both in logarithmic scale, in Fig. 6b, whichives a straight line with a slope of 0.62 indicating that the cat-lytic hydrolysis of AB is 0.6 order with respect to the rutheniumoncentration.

The turnover frequency values, shown in Table 2, for hydrogeneneration from the hydrolysis of AB (100 mM) at 25.0 ± 0.1 ◦C wereetermined from the hydrogen generation rate in the linear portionf plots given in Fig. 6a for experiments starting with 100 mM AB

lus Ru(0)/SiO2-CoFe2O4 with 1.96 wt.% Ru loading in various Ruoncentration at 25.0 ± 0.1 ◦C. It is noteworthy that the rate con-tant k1 correlates well with the induction period while the rateonstant k2 correlates with the rate of hydrogen generation from

(b) The plot of hydrogen generation rate versus the concentration of Ru, both inlogarithmic scale; ln(rate) = 0.62 ln[Ru] + 2.98.

the hydrolysis of ammonia-borane [30]. The rate constant k1 for theslow, continuous nucleation is inversely proportional to the induc-tion period as seen from Fig. 7a. The correlation plots given in Fig. 7band c show that the rate constant k2 for the surface growth of ruthe-nium(0) nanoparticles increases linearly as both the TOF value ofthe ruthenium(0) nanoparticles catalyst and the rate of hydrogengeneration from the hydrolysis of AB increase, respectively.

The TOF value is as high as 172 min−1 (mol H2/mol Ru min).TOF values of the ruthenium catalysts used in hydrolysis ofammonia-borane in the literature are listed in Table 3 for com-parison. As clearly seen from the comparison of values listed in

Table 3, Ru(0)/SiO2-CoFe2O4 provide catalytic activity comparableto that of Ru(0)NP/PSSA-co-MA [17], Ru(0)@Hap [18], Ru/Carbon[19], RuNPs@ZK-4 [20], Ru@Al2O3 [21], Ru(0)NP/laurate [22] in

S. Akbayrak et al. / Journal of Molecular Catalysis A: Chemical 394 (2014) 253–261 259

Fig. 7. Correlation between (a) the rate constant k1 (s−1) for the slow, continu-ous nucleation and the reciprocal induction period (s−1), (b) the rate constant k2

( −1 −1

Tsf

hC

at

Table 3Catalytic activity of reported ruthenium catalysts used for the hydrolysis of AB.

Entry Catalyst Ea (kJ/mol) TOF (min−1) Ref.

1 Ru/C 34.8 429.5 [24]2 Ru(0)@MWCT 33 329 [23]3 Ni0.74Ru0.26 alloyNPs 37.18 194.8 [35]4 Ru(0)/SiO2-CoFe2O4 45.6 172 This study5 Ru(0)@PSSA-co-MA 54 172 [17]6 Ru(0)@HAp 58 137 [18]7 Ni@Ru 44 114 [36]8 Ru/Carbon 76 113 [19]9 Ru/Graphene 11.7 100 [37]

10 RuNPs@ZK-4 28 90.2 [20]11 Ru@Al2O3 48 83.3 [21]12 Ru/�-Al2O3 23 77 [38]13 Ru(0)NP/laurate 47 75 [22]14 RuCo(1:1)/�-Al2O3 47 32.9 [39]15 RuCu(1:1)/�-Al2O3 52 16.4 [39]

M s , corrected) for the surface growth of ruthenium(0) nanoparticles and theOF (min−1) value of the catalyst, (c) the rate constant k2 (min−1, uncorrected) for theurface growth of ruthenium(0) nanoparticles and the rate of hydrogen generationrom the hydrolysis of ammonia-borane.

ydrolysis of AB. However, the catalytic activity of Ru(0)/SiO2-

oFe2O4 is lower than that of Ru(0)@MWCNTs [23] and Ru/C [24].

Activation energy for the hydrolysis of ammonia-borane cat-lyzed by Ru(0)/SiO2-CoFe2O4 could be determined by evaluatinghe temperature dependent kinetic data presented in Fig. 2a. The

Fig. 8. The Arrhenius plot for the Ru(0)/SiO2-CoFe2O4 catalyzed hydrolysis of AB([H3NBH3] = 100 mM and [Ru] = 0.194 mM). ln k = −5490.61 (1/T) + 20.70.

rate constants for the hydrogen generation at different tempera-ture were calculated from the slope of linear portion of each plotgiven in Fig. 2a and used for the calculation of activation energy(Ea = 45 ± 2 kJ/mol) from the Arrhenius plot in Fig. 8. The activa-tion energy for the hydrolysis of ammonia-borane catalyzed byRu(0)/SiO2-CoFe2O4 is comparable to the literature values reportedfor the other ruthenium catalysts in the same reaction (Table 3).

Reusability of Ru(0)/SiO2-CoFe2O4 catalyst was tested in suc-cessive experiments performed using the catalyst isolated from thereaction solution after a previous run of hydrolysis of AB. After thecompletion of hydrogen generation from the hydrolysis of AB start-ing with 0.744 mM Ru3+/SiO2-CoFe2O4 plus 100 mM AB in 10 mLaqueous solution at 25.0 ± 0.1 ◦C, the catalyst was isolated usinga permanent magnet (Fig. 9a and b) and washed with 10 mL ofwater. After washing, the isolated sample of Ru(0)/SiO2-CoFe2O4was redispersed in 10 mL solution containing 100 mM AB and a sec-ond run of hydrolysis was started immediately and continued untilthe completion of hydrogen evolution. This was repeated 10 times.After each run, the catalyst was isolated using a permanent magnetand the upper solution was separated. The resulting solutions aftereach subsequent runs were analyzed by ICP-OES and no leachingof ruthenium into the solution was detected. Therefore, the slightdecrease in the catalytic activity of Ru(0)/SiO2-CoFe2O4 after tenthrun in hydrolytic dehydrogenation of AB can be attributed to partialaggregation of nanoparticles on the surface of silica-coated cobaltferrite (see Figure S3, the TEM image of Ru(0)/SiO2-CoFe2O4 after

tenth run, in the Supplementary Content).

The reusability tests reveal that Ru(0)/SiO2-CoFe2O4 are stillactive in the subsequent runs of hydrolysis of AB providing 100%

260 S. Akbayrak et al. / Journal of Molecular Cata

Fig. 9. The pictures of (a) dispersed catalyst in water (b) isolated catalyst using apermanent magnet.

Fig. 10. mol H2/mol H3NBH3 versus time graph for the first and tenth use ofRu(0)/SiO2-CoFe2O4 in hydrolysis of AB.

Table 4The percentage of initial catalytic activity of various reported ruthenium catalystsafter the reuse for the hydrolysis of ammonia-borane.

Entry Catalyst Run Catalytic activity (%) Ref.

1 Ru(0)/SiO2-CoFe2O4 10 94 This study2 Ru@Al2O3 10 90 [21]3 Ru(0)@Hap 5 92 [18]4 RuNPs@ZK-4 5 85 [20]5 Ru(0)NP/laurate 5 53 [22]

cc

ercea

4

p

[8] H. Kim, A. Karkamkar, T. Autrey, P. Chupas, T. Proffen, J. Am. Chem. Soc. 131

6 Ru/C 5 43 [24]7 Ru(0)@MWCNT 4 41 [23]

onversion and Ru(0)/SiO2-CoFe2O4 preserve 94% of their initialatalytic activity even after tenth run (Fig. 10).

As shown in Table 4, the Ru(0)/SiO2-CoFe2O4 have the high-st reusability among the ruthenium(0) nanoparticles catalystseported in the literature for the same reaction. Since isolation of aatalyst by centrifugation or filtration results in material loss afterach subsequent runs, magnetically isolable Ru(0)/SiO2-CoFe2O4re highly advantageous.

. Conclusions

The main findings of this work as well as the implications orredictions can be summarized as follows:

[

lysis A: Chemical 394 (2014) 253–261

• Ruthenium(0) nanoparticles supported on magnetic silica-coatedcobalt ferrite were in situ generated from the reduction ofRu3+/SiO2-CoFe2O4 during the catalytic hydrolysis of ammonia-borane.

• Ru(0)/SiO2-CoFe2O4 were magnetically isolated from the reac-tion solution by using a permanent magnet.

• Ru(0)/SiO2-CoFe2O4 are highly active catalyst in hydrogen gen-eration from the hydrolysis of ammonia-borane providing aturnover frequency value up to 172 min−1 at room temperature.

• Ru(0)/SiO2-CoFe2O4 are reusable catalyst in hydrogen generationfrom the hydrolysis of ammonia-borane retaining 94% of theiroriginal catalytic activity after the tenth use providing the evolu-tion of 3.0 equivalent H2 per mole of ammonia-borane. Among theruthenium catalysts used in the hydrolysis of ammonia-borane,Ru(0)/SiO2-CoFe2O4 show the highest catalytic activity after thetenth use.

• The silica coated cobalt ferrite nanoparticles are suitablehost providing large surface area for the formation ruthe-nium(0) nanoparticles. Thus, Ru(0)/SiO2-CoFe2O4 nanoparticlesare highly active and durable catalysts.

• It is shown that the formation kinetics of ruthenium(0) nanopar-ticles can be studied by using the hydrogen evolution fromthe hydrolysis of ammonia-borane as reporter reaction. Allthe kinetic data, collected for the nanoparticles formation andconcomitant hydrolysis of AB catalyzed by Ru(0)/SiO2-CoFe2O4under various experimental conditions, fit well to the 2-stepmechanism for the nanoparticles formation: the nucleation(P → Q, rate constant k1) and then autocatalytic surface growth(P + Q → 2Q, rate constant k2).

• The activation energy values for the slow nucleation and autocat-alytic surface growth of metal(0) nanoparticles are 116 ± 7 kJ/moland 51 ± 2 kJ/mol, respectively, while the activation energy forhydrogen generation from the hydrolysis of ammonia-borane is45 ± 2 kJ/mol.

• Ru(0)/SiO2-CoFe2O4 are highly active, magnetically isolable andrecyclable catalysts in hydrogen generation from the hydrolysisof ammonia-borane.

Acknowledgement

Partial support by Middle East Technical University, METU Cen-tral Laboratory and Turkish Academy of Sciences is gratefullyacknowledged.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.molcata.2014.07.010.

References

[1] L. Schlapbach, A. Züttel, Nature 414 (2001) 353–358.[2] Z. Xiong, C.K. Yong, G. Wu, P. Chen, W. Shaw, A. Karkamkar, T. Autrey, M.O.

Jones, S.R. Johnson, P.P. Edwards, W.I.F. David, Nat. Mater. 7 (2008) 138–142.[3] K.J. Jeon, H.R. Moon, A.M. Ruminski, B. Jiang, C. Kisielowski, R. Bardhan, J.J.

Urban, Nat. Mater. 10 (2011) 286–290.[4] S. Orimo, Y. Nakamori, J.R. Eliseo, A. Züttel, C.M. Jense, Chem. Rev. 107 (2007)

4111–4132.[5] C. Fellay, P.J. Dyson, G. Laurenczy, Angew. Chem. Int. Ed. 47 (2008) 3966–3968.[6] A. Gutowska, L. Li, Y. Shin, C.M. Wang, X.S. Li, J.C. Linehan, R.S. Smith, B.D.

Kay, B. Schmid, W. Shaw, M. Gutowski, T. Autrey, Angew. Chem. 117 (2005)3644–3648.

[7] T. He, Z. Xiong, G. Wu, H. Chu, C. Wu, T. Zhang, P. Chen, Chem. Mater. 21 (2009)2315–2318.

(2009) 13749–13755.[9] M. Chandra, Q. Xu, J. Power Sources 156 (2006) 190–194.10] J.M. Yan, X.-B. Zhang, S. Han, H. Shioyama, Q. Xu, Angew. Chem. 120 (2008)

2319–2321.

r Cata

[[

[[[[

[[[

[[[[[

[

[

[

[

[[[[[

[[[

(2011) 6308–6310.

S. Akbayrak et al. / Journal of Molecula

11] Z. Li, K. Aranishi, Q. Xu, Chem. Commun. 48 (2012) 3173–3175.12] A. Gutowska, L. Li, Y. Shin, C.M. Wang, X.S. Li, J.C. Linehan, R.S. Smith, B.D. Kay,

B. Schmid, W. Shaw, M. Gutowski, T. Autrey, Angew. Chem. Int. Ed. 44 (2005)3578–3582.

13] Z.H. Lu, Q. Xu, Funct. Mater. Lett. 5 (2012) 1230001–1230009.14] M. Yadav, Q. Xu, Energy Environ. Sci. 5 (2012) 9698–9725.15] M. Zahmakıran, S. Özkar, Nanoscale 3 (2011) 3462–3481.16] S. Özkar, S.L. Suib (Eds.), Batteries, Hydrogen Storage and Fuel Cells, Elsevier,

2013.17] Ö. Metin, S. S ahin, S. Özkar, Int. J. Hydrogen Energy 34 (2009) 6304–6313.18] S. Akbayrak, P. Erdek, S. Özkar, Appl. Catal. B: Environ. 142–143 (2013) 187–195.19] S. Basu, A. Brockman, P. Gagare, Y. Zheng, P.V. Ramachandran, W.N. Delgass, J.P.

Gore, J. Power Sources 188 (2009) 238–243.20] M. Zahmakıran, Mater. Sci. Eng. B 177 (2012) 606–613.21] H. Can, Ö. Metin, Appl. Catal. B: Environ. 125 (2012) 304–310.22] F. Durap, M. Zahmakıran, S. Ozkar, Int. J. Hydrogen Energy 34 (2009) 7223–7230.

23] S. Akbayrak, S. Özkar, ACS Appl. Mater. Interfaces 4 (2012) 6302–6310.24] H. Liang, G. Chen, S. Desinan, R. Rosei, F. Rosei, D. Ma, Int. J. Hydrogen Energy 3

(2012) 17921–17927.25] Z.H. Lu, J. Li, A. Zhu, Q. Yao, W. Huang, R. Zhou, R. Zhou, X. Chen, Int. J. Hydrogen

Energy 38 (2013) 5330–5337.

[[[

lysis A: Chemical 394 (2014) 253–261 261

26] K. Maaz, A. Mumtaz, S.K. Hasanain, A. Ceylan, J. Magn. Magn. Mater. 308 (2007)289–295.

27] J. Choi, J.C. Kim, Y.B. Lee, I.S. Kim, Y.K. Parka, N.H. Hur, Chem. Commun. (2007)1644–1646.

28] M. Kaya, M. Zahmakıran, S. Özkar, M. Volkan, ACS Appl. Mater. Interfaces 4(2012) 3866–3873.

29] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1968) 62–69.30] M.A. Watzky, R.G. Finke, J. Am. Chem. Soc. 119 (1997) 10382–10400.31] J.A. Widegren, J.D. Aiken, S. Özkar, R.G. Finke, Chem. Mater. 13 (2001) 312–324.32] M.A. Watzky, R.G. Finke, Chem. Mater. 9 (1997) 3083–3095.33] K.C. Park, I.Y. Jang, W. Wongwiriyapan, S. Morimoto, Y.J. Kim, Y.C. Jung, T. Toya,

M. Endo, J. Mater. Chem. 20 (2010) 5345–5354.34] Ö. Metin, S. Ozkar, Int. J. Hydrogen Energy 36 (2011) 1424–1432.35] G. Chen, S. Desinan, R. Rose, F. Rose, D. Ma, Chem. Eur. J. 18 (2012) 7925–7930.36] G. Chen, S. Desinan, R. Nechache, R. Rose, F. Rose, D. Ma, Chem. Commun. 47

37] N. Cao, W. Luo, G. Cheng, Int. J. Hydrogen Energy 38 (2013) 11964–11972.38] M. Chandra, Q. Xu, J. Power Sources 168 (2007) 135–142.39] G.P. Rachiero, Ü.B. Demirci, P. Miele, Int. J. Hydrogen Energy 36 (2011)

7051–7065.