Nanoalumina-supported rhodium(0) nanoparticles as catalyst in hydrogen generation from the methanolysis of ammonia borane

Molecular Catalysis 439 (2017) 50–59

Contents lists available at ScienceDirect

Molecular Catalysis journal homepage: www.elsevier.com/locate/mcat

Research Paper

Nanoalumina-supported rhodium(0) nanoparticles as catalyst in hydrogen generation from the methanolysis of ammonia borane Derya Özhava, Saim Özkar ∗ Department of Chemistry, Middle East Technical University, 06800 Ankara, Turkey

a r t i c l e

i n f o

Article history: Received 4 May 2017 Received in revised form 9 June 2017 Accepted 13 June 2017 Keywords: Hydrogen generation Ammonia borane Methanolysis Alumina nanopowder Rhodium Nanoparticles

a b s t r a c t Rhodium(0) nanoparticles were in situ formed from the reduction of rhodium(II) octanoate and supported on the surface of nanoalumina yielding Rh(0)/nanoAl2 O3 which is highly active catalyst in hydrogen generation from the methanolysis of ammonia borane at room temperature. The kinetics of nanoparticle formation can be followed just by monitoring the volume of hydrogen gas evolved from the methanolysis of ammonia borane. The evaluation of the kinetic data gives valuable insights to the slow, continuous nucleation and autocatalytic surface growth steps of the formation of rhodium(0) nanoparticles. Rh(0)/nanoAl2 O3 could be isolated and characterized by a combination of advanced analytical techniques including ATR-IR, PXRD, TEM, XPS, SEM, SEM-EDX and ICP-OES. The results reveal that rhodium(0) nanoparticles are highly dispersed on the surface of nanoalumina. The particle size of Rh(0)/nanoAl2 O3 increases with the initial rhodium loading of nanoalumina. Rh(0)/nanoAl2 O3 is highly active catalyst in hydrogen generation from the methanolysis of AB providing an exceptional initial turnover frequency of TOF = 218 min−1 at 25.0 ± 0.5 ◦ C, which is the highest value ever reported for rhodium catalysts in hydrogen generation from the methanolysis of ammonia borane. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Ammonia borane (AB) is one of the potential hydrogen storage materials because of its high hydrogen content (19.6 wt%), nontoxicity, and stability under ambient conditions [1]. Although intensive efforts have been devoted to the hydrolysis of AB, methanolysis of AB has been considered as another efficient way of releasing the H2 gas because of many advantages over the hydrolysis [2]. The most frequently encountered problems in hydrogen generation from hydrolysis of AB are the concomitant release of ammonia gas from concentrated AB solutions [3] and difficulties in recycling process of hydrolysis product, ammonium metaborate (NH4 BO2 ). Additionally, AB has not only high solubility in methanol, 23% wt. at 23 ◦ C [4] but also high stability with respect to self methanolysis at ambient conditions. In the presence of a suitable catalyst, methanolysis can release 3.0 equivalents of H2 per mole of AB (Eq. (1)): catalyst

NH3 BH3 + 4CH3 OH → NH4 B(OCH3 )4 + 3H2

∗ Corresponding author. E-mail address: [email protected] (S. Özkar). http://dx.doi.org/10.1016/j.mcat.2017.06.016 2468-8231/© 2017 Elsevier B.V. All rights reserved.

Up to now, methanolysis of AB has been investigated in the presence of many transition metals [5]. Among them, rhodium(0) nanoparticles can be incontestably pronounced the most active transition metal catalyst in the methanolysis of AB. However, rhodium(0) nanoparticles as other metal(0) nanoparticles are prone to agglomerate due to their high surface energy and they need to be stabilized against agglomeration [6]. Porous materials with large surface area have been used as support to obtain stable and catalytically active rhodium(0) nanoparticles [7]. The catalytic activity of nanoparticles depends on the particle size of support. We have recently used nanopowders of silica [8], hydroxyapatite [9], and titania [10] as support to make rhodium(0) nanoparticles highly active and long lived catalyst. Alumina is another widely used oxide support due to favourable combination of its textural properties such as surface area, pore volume and pore size distribution as well as chemical and hydrothermal stability [11]. Alumina has also been used to stabilize rhodium(0) nanoparticles catalyst in the hydrolysis of AB, if not the methanolysis [12]. In our ongoing research on the use of oxide nanopowders as support for rhodium(0) nanoparticles in hydrogen generation from the methanolysis of AB, herein we report the use of nanoalumina with an average particle size of 13 nm for stabilizing rhodium(0) nanoparticles. Nanoalumina supported rhodium(0) nanoparticles, hereafter referred to as Rh(0)/nanoAl2 O3 , were in situ generated

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

from the reduction of rhodium(II) octanoate, impregnated on the surface of nanoalumina, during the methanolysis of AB in methanol. Rh(0)/nanoAl2 O3 were isolated from the reaction medium and characterized by a combination of advanced analytical techniques including ATR-IR, PXRD, TEM, XPS, SEM, SEM-EDX and ICP-OES. The results reveal the formation of uniformly dispersed rhodium(0) nanoparticles with a mean particle diameter of 3.6–4.1 nm on the porous nanoalumina. These rhodium(0) nanoparticles were tested as catalyst in hydrogen generation from the methanolysis of AB and the kinetics of the nanoparticle formation and catalytic methanolysis of AB was studied in details. 2. Experimental 2.1. Materials Rhodium(II) octanoate, ammonia borane (97%), methanol (99%), and nanoalumina (Al2 O3 , particle size ≈ 13 nm) were purchased from Aldrich. Methanol was distilled over Mg metal and kept under nitrogen atmosphere until the use in methanolysis. 2.2. Instrumentation

Number of Rh atoms on the surface of NP, ns =

4

 D 2 2

(2r)2

r = atomic radius of rhodium (0.135 nm)” 2.4. Activation parameters for the methanolysis of AB catalyzed by Rh(0)/nanoAl2 O3 The in situ generation of Rh(0)/nanoAl2 O3 and catalytic methanolysis of AB are performed starting with 200 mM (64 mg) AB and 0.24 mM Rh (50 mg Rh(II)/nanoAl2 O3 , 0.5 wt.% Rh,) at 20, 25, 30 and 35 ◦ C following the procedure given in the previous section. 2.5. Catalytic lifetime of Rh(0)/nanoAl2 O3 catalyst in the methanolysis of AB A lifetime experiment is performed starting with 10 mL solution containing 50 mg Rh(II)/nanoAl2 O3 (0.50 wt.% rhodium, 2.4 mmol Rh) and 400 mM (128 mg) AB at 25.0 ± 0.5 ◦ C following the procedure described elsewhere [8].

All the instrumentations used in this work are the same as reported elsewhere [8] with the exception of taking the IR spectra and the SEM images. The ATR-IR spectra of the nanoalumina samples were recorded on a Vartex-70 spectrophotometer. SEM analyses were run on a JEOL JSM-5310LV operating at 15 kV and 33 Pa in a low-vacuum mode without metal coating on the aluminum support.

2.6. Heterogeneity test for Rh(0)/nanoAl2 O3 by carbon disulfide poisoning in the methanolysis of AB

2.3. In situ preparation of Rh(0)/nanoAl2 O3 and catalytic methanolysis of AB

2.7. Leaching test for Rh(0)/nanoAl2 O3

The preparation of Rh(II)/nanoAl2 O3 and the in situ generation of Rh(0)/nanoAl2 O3 during the catalytic methanolysis of ammonia borane were performed in the same medium following the procedure given elsewhere [8]. Briefly an aliquot of the stock solution of 4.9 mM rhodium(II) octanoate is transferred into a jacketed 50 mL flask with a Teflon-coated stir bar, containing alumina nanopowders, thermostated by circulating water through its jacket at 25.0 ± 0.5 ◦ C or at a certain temperature specified. After addition of 6.5 mL methanol, the resulting suspension is stirred for 1 h. Then, 64 mg (2.0 mmol) AB in 3.0 mL methanol is added to the flask and the reaction is launched. The volume of H2 gas released is recorded by measuring the displacement of water level in the glass tube connected to the reaction flask [8]. Note that all the experiments in this study were performed under stirring at 1000 rpm to ensure that the methanolysis in the presence of Rh(0)/nanoAl2 O3 catalyst is in the kinetic regime, not under mass transfer limitation [8]. Calculation of the turnover frequency (TOF): TOF =

mol of H2 mol of Rh × time

Corrected TOF = TOF × percentage of catalytically active Rh atoms Number of Rh atoms in nanoparticle,



N0  n=

 4   D 3  3



2

W

where n = number of Rh atoms N0 = 6.022 × 1023 mol−1 D = diameter of rhodium nanoparticles W = atomic weight of rhodium (102.91 g/mol) ␳ = room-temperature density of rhodium (12.41 g/cm3 )

51

Catalyst poisoning experiment is performed by injecting carbon disulfide to reaction solution containing Rh(0)/nanoAl2 O3 during methanolysis of AB following the procedure described elsewhere [8].

A leaching test is performed for the Rh(0)/nanoAl2 O3 in hydrogen generation from the methanolysis of AB by following the procedure described elsewhere [8]. Additionally, the filtrate solution obtained by filtration of the reaction mixture after the catalytic methanolysis reaction is also analyzed by ICP-OES for the rhodium content. 3. RESULTS and DISCUSSION 3.1. Formation of rhodium(0) nanoparticles and concomitant hydrogen generation from the methanolysis of AB The reduction of rhodium(II) octanoate, impregnated on the porous nanoalumina, Rh(II)/nanoAl2 O3 , provides the in situ formation of nanoalumina supported rhodium(0) nanoparticles, Rh(0)/nanoAl2 O3 , during methanolysis of AB. First, the rhodium(II) octanoate solution in methanol was added to the suspension of nanoalumina in methanol solution and stirred for 1 h. Before using this precatalyst mixture in the catalytic methanolysis of AB, a solid sample was isolated for taking the ATR-IR spectrum to check whether rhodium(II) ions are accompanied by the octanoate anion in the impregnation on the surface of nanoalumina. Fig. 1 depicts the ATR-IR spectra of bare nanoalumina and nanoalumina-impregnated Rh(octanoate)2 , Rh(II)/nanoAl2 O3 with a rhodium loading of 1.0% wt. Both spectra show a broad absorption at 3500 cm−1 and a weak band at 1638 cm−1 for the stretching and bending modes of OH groups on nanoalumina, in addition to the strong framework bands in the range 500–1000 cm−1 (not shown in the spectra to simplify the view) [13]. ATR-IR spectrum of Rh(II)/nanoAl2 O3 shows additionally two absorption bands at 1570 and 1420 cm−1 due to the asymmetric and symmetric stretching of carboxylate group along with the C H stretching bands at around 2900 cm−1 indicating the presence of octanoate anion on

52

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

Fig. 1. ATR-IR spectrum of (a) nanoalumina (b) Rh(II)/nanoAl2 O3 with a rhodium loading of 1.0% wt. Rh.

Fig. 2. Plots of mole H2 evolved per mole of AB versus time for the catalytic ) 50 mg methanolysis of AB starting with 200 mM AB plus either (red circles, ) 1.9 mg of ofRh+2 /nanoAl2 O3 (1.0 wt.% Rh, [Rh] = 0.49 mM) or (black squares, rhodium(II) octanoate (0.49 mM Rh) in 10 mL methanol at 25.0 ± 0.5 ◦ C. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Plots of mole H2 evolved per mole of AB versus time for the catalytic methanolysis of AB starting with 200 mM AB and 50 mg of Rh(II)/nanoAl2 O3 (0.50 wt.% Rh, [Rh] = 0.24 mM) in 10 mL methanol at 25.0 ± 0.5 ◦ C. The red curve shows the fit of data to the F-W 2-step mechanism of nanoparticle formation [16]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the surface of nanoalumina [14]. Herein, the observed frequency difference of 150 cm−1 between asymmetric and symmetric C O vibrations indicates the existence of chelating or bridging carboxylate group [15]. As control two experiments were performed under the same conditions starting with rhodium(II) octanoate either alone or impregnated on nanoalumina to compare the activity and stability of rhodium(0) nanoparticles. Fig. 2 depicts the plots of mole of H2 evolved per mole of AB versus time for the catalytic methanolysis of 200 mM AB plus either 50 mg Rh+2 /nanoAl2 O3 (1.0% wt. Rh, [Rh] = 0.49 mM) or 1.9 mg rhodium(II) octanoate (0.49 mM Rh) in 10 mL methanol at 25.0 ± 0.5 ◦ C. The results of these two control experiments show that both catalysts have similar initial catalytic activity, however sole rhodium(0) nanoparticles in the absence of support lose their catalytic activity in the course of methanolysis reaction. It is noteworthy that bulk rhodium metal precipitates in the absence of support at the end of methanolysis reaction. Therefore, it can be concluded that nanoalumina is an efficient support in obtaining rhodium(0) nanoparticles which are stable and catalytically active. Rhodium(0) nanoparticles were produced from the reduction of rhodium(II) octanoate by adding AB solution to the reaction flask involving a suspension of Rh(II)/nanoAl2 O3 precatalyst and found

to be active catalyst in hydrogen generation from the methanolysis of AB. The volume of H2 evolved during the reaction was monitored and converted into the equivalent H2 per mole of AB using the known 3:1 H2 /AB stoichiometry (Eq. (1)). Evaluation of the kinetic data gives valuable insights to the formation of rhodium(0) nanoparticles and concomitant H2 generation from the methanolysis of AB. Fig. 3 displays the graph of mole H2 per mole of AB versus time for the catalytic methanolysis of 200 mM AB in the presence of 50 mg of Rh(II)/nanoAl2 O3 (0.50 wt.% Rh, [Rh] = 0.24 mM) in 10 mL methanol at 25.0 ± 0.5 ◦ C. This plot shows the release of 3.0 equivalents H2 per mole of AB within 15 min. Also the complete conversion of AB to the tetramethoxyborate [B(OCH3 )4 ]− anion is confirmed by the 11 B NMR spectrum taken from reaction solution after the catalytic methanolysis (d = 8.6 ppm) [2]. In such an experiment, the release of 3.0 equivalents H2 per mole of AB could be achieved with almost linear hydrogen generation curve after a short induction period less than 1 min. The reaction medium changes its colour from the initial pale blue to black during the induction period. In addition to the alteration in colour, the sigmoidal shape of the hydrogen generation versus time plot in Fig. 3 indicates the reduction of rhodium(II) to rhodium(0) and formation of rhodium(0) nanoparticles with Finke–Watzky 2step mechanism consisting of slow, continuous nucleation (A → B,

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

Scheme 1. Illustration of hydrogen generation from the catalytic methanolysis of AB as reporter reaction: A is the precursor Rh(II)/nanoAl2 O3 and B is the growing Rh(0)n nanoparticles on the surface of nanoalumina support.

Fig. 4. Powder X-ray diffraction (PXRD) patterns of (a) nanoAl2 O3 , (b) Rh(0)/nanoAl2 O3 with rhodium loading of 0.50 % wt. Rh.

rate constant k1 ) and autocatalytic surface growth (A + B → 2B, rate constant k2 ) given in Scheme 1, where A stands for the added precursor Rh(II)/nanoAl2 O3 and B for the growing Rh(0)n nanoparticles [16,17]. The kinetic data in Fig. 3 fit well to Eq. (2) derived from the F-W 2-step mechanism [16] providing the rate constants k1 = 6.6 × 10−2 min−1 and k2 = 7.8 × 101 M−1 min−1 . [AB]t = [AB]o −

k1 k2

1+

+ [AB]o

k1 exp (k1 k2 [AB]0

+ k2 [AB]0 ) t

(2)

These rate constants are for the slow, continuous nucleation and autocatalytic surface growth of Rh(0) nanoparticles, respectively, starting with 200 mM AB and 50 mg of Rh(II)/nanoAl2 O3 (0.50 wt.% Rh, [Rh] = 0.24 mM) in 10 mL methanol at 25.0 ± 0.5 ◦ C. The large value of k2 /k1 ratio (1.2 × 103 ) implies on the high level of kinetic control of the nanoparticles formation [16]. 3.2. Characterization of Rh(0)/nanoAl2 O3 After in situ formation of Rh(0)/nanoAl2 O3 from the reduction of precursor Rh(II)/nanoAl2 O3 during the methanolysis of AB at room temperature, the solid materials were separated from reaction solution and characterized by advanced analytical techniques. Powder X-Ray diffraction patterns for nanoalumina (nanoAl2 O3 ) and nanoalumina supported rhodium(0) nanoparticles (Rh(0)/nanoAl2 O3 ) are given in Fig. 4. The characteristic diffraction peaks of nanoalumina are seen in each of the powder XRD patterns, which indicate that alumina remains unchanged after the preparation of Rh(II)/nanoAl2 O3 and the reduction of rhodium(II) to rhodium(0) during the methanolysis of AB. No additional diffraction peaks are observed which would be ascribable to

53

rhodium(0) nanoparticles, most likely due to the low metal loading of nanoalumina. Inspection of TEM images of Rh(0)/nanoAl2 O3 samples with different rhodium loadings (0.50, 1.0, 1.5 and 2.0% wt. Rh) given in Fig. 5 reveals the uniform dispersion of rhodium(0) nanoparticles on the surface of nanoalumina. HRTEM image of Rh(0) nanopaticles in Fig. 5e is used to measure the (111) lattice fringes of 0.222 nm [18]. Fig. 5f gives the plot of particle size versus the percent rhodium loading of Rh(0)/nanoAl2 O3 samples showing the anticipated increase in the particle size with the increasing metal loading of the nanomaterials. The oxidation state and chemical environment of deposited rhodium nanoparticles can be analyzed by X-Ray photoelectron spectroscopy. The survey scan XPS of Rh(0)/nanoAl2 O3 with a rhodium loading of 2.0% wt. Rh in Fig. 6a shows the presence of rhodium in addition to the framework elements of alumina (Al and O) in agreement with the SEM-EDX results (see later). The deconvolution of high resolution Rh 3d XPS spectrum of the same Rh(0)/nanoAl2 O3 sample in Fig. 6b depicts two prominent peaks at 307.2 and 312.2 eV, readily ascribed to the Rh(0) 3d5/2 and 3d3/2 , respectively, by comparing to the values of metallic rhodium [19]. Additionally, higher energy bands observed at 308.9, 314.9 eV and 310.3, 315.2 eV are attributable to the Rh+3 and Rh+4 3d5/2 and 3d3/2 , respectively, likely in the form of oxide, which may originate from surface oxidation of rhodium(0) nanoparticles during the XPS sampling [20]. Surface area of nanoAl2 O3 and Rh(0)/nanoAl2 O3 with a loading of 1.5% wt. Rh were determined by the nitrogen adsorption measurement to be 91.6 and 96.1 m2 × g−1 , respectively. The slight increase in the surface area of nanoalumina can be considered as an additional evidence for the formation of rhodium(0) nanoparticles on the surface of supporting materials. SEM image of Rh(0)/nanoAl2 O3 with a rhodium loading of 2.0% wt. in Fig. 7a shows that there is no bulk rhodium metal formation in detectable size on the surface of nanoalumina and SEM-EDX spectrum in Fig. 7b confirms the presence of rhodium in addition to the framework elements of alumina (Al, O). 3.3. Catalytic activity of Rh(0)/nanoAl2 O3 in the methanolysis of AB An excellent catalytic performance of Rh(0)/nanoAl2 O3 is observed in H2 generation from the methanolysis of AB providing 3.0 equiv. H2 per mole of AB at room temperature. Fig. 8a shows the plots of mol H2 evolved per mole of AB versus time during the catalytic methanolysis performed starting with 200 mM AB in various Rh loading, thus in various Rh concentrations at 25.0 ± 0.5 ◦ C. In all cases after a short induction period less than 1.0 min, a H2 evolution starts and continues almost linearly until the complete consumption of AB present in the solution. It is noteworthy that the rate of hydrogen generation remains practically constant in linear portion of the plots. This implies on the structure insensitivity of catalytic methanolysis of AB. All the kinetic data in Fig. 8a fit well to the FW 2-step mechanism of nanoparticle formation [16] yielding the rate constants k1 of the slow, continuous nucleation and k2 of the autocatalytic surface growth for the formation of rhodium(0) nanoparticles from rhodium(II) during the methanolysis (Table 1) [16]. The large values of k2 /k1 ratio in Table 1 indicate that the formation of rhodium(0) nanoparticles from the reduction of rhodium(II) precursor during the methanolysis of AB is kinetically controlled. Turnover frequency (TOF) values were calculated from the hydrogen generation rate in the linear part of each plot given in Fig. 8a for the methanolysis of 200 mM AB at 25.0 ± 0.5 ◦ C and listed in Table 1 along with the corrected TOF values by considering the number of rhodium sites on the surface of nanopar-

54

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

Fig. 5. TEM images of nanoalumina stabilized rhodium(0) nanoparticles with different rhodium loading (a) 0.5, (b) 1.0, (c) 1.5, and (d) 2.0 wt.% Rh after catalytic methanolysis reaction performed starting with 200 mM AB catalyzed by desired amount of nanoalumina stabilized Rh(0) nanoparticles at 25.0 ± 0.5 ◦ C. (e) HRTEM images of nanoalumina stabilized rhodium(0) nanoparticles with rhodium loading of 1.0 wt.% (f) the plot of percentage rhodium loading versus particle size.

Table 1 The rate constants k1 of the slow, continuous nucleation, A → B, and k2 of the autocatalytic surface growth, A + B → 2B, for the formation of rhodium(0) nanoparticles catalyst from the reduction of rhodium(II) ions during the methanolysis of AB (200 mM) starting with Rh(II)/nanoAl2 O3 with different rhodium loading, rhodium concentrations, hydrogen generation rates, turnover frequency (TOF) values, and the TOF values corrected for the number of rhodium atoms on the surface of nanoparticles for the methanolysis of AB at 25.0 ± 0.5 ◦ C. wt.%Rh Particle size (nm) 0.5 1.0 1.5 2.0

3.6 ± 0.4 3.8 ± 0.7 3.9 ± 0.7 4.1 ± 0.6

Number of Rh atoms in NP

Number of Rh atoms on surface

[Rh] (mM)

k1 (min−1 )

k2 /k1 k2 (M−1 min−1 )

H2 generation rate (mmol/min)

TOF (min−1 )

TOFcorrected (min−1 )

1773 2085 2254 2619

558 622 655 724

0.244 0.491 0.761 0.991

(6.64 ± 0.25) × 10−2 (7.46 ± 0.57) × 10−2 (8.54 ± 0.64) × 10−2 (8.21 ± 0.96) × 10−2

78.1 ± 3.1 117.5 ± 5.4 104.1 ± 4.2 118.9 ± 6.2

0.54 1.17 2.38 3.37

218 188 158 154

692 630 544 557

ticle. Rh(0)/nanoAl2 O3 sample with rhodium concentration of 0.24 mM Rh provides the highest catalytic activity with an initial TOF = 218 min−1 (692 min−1 corrected for surface rhodium atoms, Table 1) ever reported for hydrogen generation from the

1.18 × 103 1.57 × 103 1.22 × 103 1.48 × 103

methanolysis of AB at 25.0 ± 0.5 ◦ C [21]. Comparison of the values in Table 1 shows that the TOF decreases as the rhodium loading in the Rh(0)/nanoAl2 O3 catalyst increases. The inverse relation between the TOF and the catalyst concentration (Fig. 8b) can be cor-

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

55

Fig. 6. (a) X-ray photoelectron (XPS) survey scan spectrum, (b) high resolution Rh 3d XPS spectrum of Rh(0)/nanoAl2 O3 with a rhodium loading of 2.0% wt. Rh.

related to the increasing size of rhodium(0) nanoparticles (Fig. 5f) as precedented in literature for various catalytic reactions [8,22]. However, only our previous three reports have provided the experimental evidence showing that the increasing particles size with the increasing catalyst concentration is the reason of the decrease in catalytic activity [8,23,24]. In the present work, the inverse correlation between the catalytic activity and the size of nanoparticles was investigated by measuring both quantities accurately. The catalytic activity of Rh(0)/nanoAl2 O3 in H2 generation from the methanolysis of AB decreases as the particle size of the rhodium(0) nanoparticles increases with the increasing Rh loading (Figs. 5f and 8b). However, when TOF values are corrected for the number of Rh atoms on the surface of nanoparticles (Table 1) the inverse dependence of catalytic activity on the size of nanoparticles becomes less pronounced. The disappearance of inverse dependence for the corrected TOF values is also indicative of that the inverse relation between the catalytic activity and Rh concentration is indeed a size issue. The hydrogen generation rate was determined from the linear portion of graph for each rhodium concentration in Fig. 8a and plotted versus the initial Rh concentration, both axes in logarithmic scale (Inset of Fig. 8a). This plot gives a straight line with a slope of 0.90 ≈ 1.0 indicating that the methanolysis of AB catalyzed by Rh(0)/nanoAl2 O3 is first order with respect to catalyst. It is noteworthy that hydrogen generation starts immediately after a short induction time less than 1.0 min and proceeds almost linearly in course of methanolysis reaction; therefore it is conceivable that methanolysis of AB is zero order with respect to substrate.

Fig. 7. a) SEM image of Rh(0)/nanoAl2 O3 b) SEM-EDX spectrum of Rh(0)/nanoAl2 O3 with a rhodium loading of 2.0% wt. Rh.

Fig. 9a gives the plots of mole H2 evolved per mole of AB versus time for the methanolysis of AB starting with 0.24 mM Rh catalyst (0.50 wt.% Rh, 50 mg Rh(II)/nanoAl2 O3 ) plus 200 mM AB substrate at various temperatures in the range of 20–35 ◦ C. The rate constants k1 of the slow, continuous nucleation and k2 of the autocatalytic surface growth for the formation of rhodium(0) nanoparticles catalyst during the methanolysis of AB can be calculated by curve-fitting the experimental data to the F-W 2-step mechanism [16]. Table 2 gives the list of the rate constants k1 and k2 as well the k2 /k1 ratio. A high level kinetic control in the formation of rhodium(0) nanoparticles from the reduction of rhodium(II) precursor in the presence of nanoalumina can readily be seen from the large value of k2 /k1 ratio at any temperature in the range 20–35 ◦ C [16]. Since we have the rate constants k1 and k2 for the individual steps of the rhodium(0) nanoparticles formation at various temperatures we are now able to calculate the activation parameters for the slow, continuous nucleation and autocatalytic surface growth of the rhodium(0) nanoparticles on the surface of nanoalumina, as well as for the catalytic methanolysis of AB catalyzed by Rh(0)/nanoAl2 O3 from the temperature dependent kinetic data. Activation energy for the nucleation and autocatalytic surface growth were determined from the Arrhenius plots by using the values of rate constants k1 and k2 at different temperatures in Fig. 9b and c, respectively: Ea = 67 ± 2 kJ/mol for the nucleation and Ea = 48 ± 2 kJ/mol for the autocatalytic surface growth. Activation energy values give an idea

56

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

Fig. 8. a) Plots of mole H2 evolved per mole of AB versus time for methanolysis of 200 mM AB in different metal concentrations ([Rh] = 0.24, 0.49, 0.74 and 0.99 mM, each prepared using metal loading of 0.50, 1.0, 1.5, 2.0% wt. Rh, respectively) at 25.0 ± 0.5 ◦ C. Inset: Plot of hydrogen generation rate versus the concentration of rhodium (both in logaritmic scale.). b) Inverse dependence of initial TOF value on the initial precatalyst Rh concentration.

Table 2 The rate constants k1 of the slow, continuous nucleation, A → B, and k2 of the autocatalytic surface growth, A + B → 2B, for the formation of rhodium(0) nanoparticles catalyst from the reduction of rhodim(II) ions during the methanolysis of AB, hydrogen generation rates, kobs values, turnover frequency (TOF) values, and the TOF values corrected for the number of rhodium atoms on the surface of nanoparticles for the methanolysis of AB starting with [AB] = 200 mM and [Rh] = 0.24 mM at different temperatures. Temperature (◦ C)

k1 (min−1 )

k2 (M−1 min−1 )

k2 /k1

H2 generation rate (mmol H2 min−1 )

kobs (mol H2 molRh−1 s−1 )

TOF (min−1 )

20 25 30 35

5.16 × 10−2 ± 1.36 × 10−3 6.63 × 10−2 ± 2.51 × 10−3 1.27 × 10−1 ± 5.65 × 10−3 1.83 × 10−1 ± 7.11 × 10−3

56.7 ± 0.17 78.4 ± 3.2 117.9 ± 6.3 142.7 ± 7.7

1.10 × 103 1.18 × 103 0.93 × 103 0.78 × 103

0.42 0.54 1.04 1.32

2.86 3.67 7.08 9.04

172 218 425 557

on the energy barrier for the slow nucleation and autocatalytic surface growth of metal(0) nanoparticles catalyst. Activation parameters for methanolysis of AB catalyzed by Rh(0)/nanoAl2 O3 can also be determined from the temperature dependent kinetic data in Fig. 9a. The slope of each plot in the linear part can be used to calculate the rate constant kobs and the TOF values at various temperatures (Table 2). It is noteworthy that also TOF value increases with increasing temperature reaching a remarkable value of 557 min−1 (1770 min−1 corrected for the number of rhodium atoms on the surface of nanoparticle) in hydrogen generation from the catalytic methanolysis of AB at 35 ◦ C. In addition, the values of reaction rate constant kobs were used for drawing of Arrhenius plot given in Fig. 9d. Thus, the activation energy was calculated to be 62 ± 2 kJ/mol which is the same as the value obtained by using the nanosilica-supported rhodium(0) nanoparticles [8]. The activa-

tion energy for H2 generation from the catalytic methanolysis of AB is comparable to the literature values reported for the methanolysis of AB using different catalysts (Table 3). A catalytic lifetime experiment was performed starting with 10 mL suspension containing 0.24 mM Rh (50.0 mg Rh(0)/nanoAl2 O3 , 0.5 wt.% Rh,) and 400 mM AB at 25.0 ± 0.5 ◦ C (Fig. 10). The reaction was sustained by adding a new batch of AB into reaction flask when the AB present in the solution was consumed in releasing hydrogen gas. This procedure was continued until no more gas evolution was observed. The catalytic lifetime experiment reveals a minimum of 16,300 turnovers in hydrogen generation from the methanolysis of AB at 25.0 ± 0.5 ◦ C before deactivation and provides the highest exceptional initial turnover frequency (TOF) value of 218 min−1 in hydrogen generation from the methanolysis of AB at 25.0 ± 0.5 ◦ C. The rate of hydrogen gen-

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

57

Table 3 Catalytic activity in TOF and catalyst life time of metal(0) catalysts used for the catalytic methanolysis of AB and activation energy of catalytic methanolysis. Catalyst

TOF (mol H2 mol catalyst−1 min−1 )

Activation energy, Ea (kJ mol−1 )

Total turnover numbers (TTO)

Ref.

RhCl3 RuCl3 CoCl2 NiCl2 PdCl2 Pd/C Nano Cu2 O Nano Cu@Cu2 O PVP-stabilized Ni PVP-stabilized Pd PVP-stabilized Ru Co-Co2 B Ni-Ni3 B Co-Ni-B Zeolite confined Rh MMT stabilized Ru Co48 Pd52 /C CuPd alloy mesoporous CuO Cu-Cu2 O-CuO/C Rh(0)/nanoHAP Ru/graphene AgPd alloy G-Cu36 Ni64 Rh/CC3-R-homo Rh/CC3-R-hetero Rh(0)/nanoSiO2 Rh(0)/nanoAl2 O3

100 150 3.7 2.9 1.5 1.9 0.2 0.16 12.1 22.3 47.7 7.5 5.0 10.0 30.0 90.9 27.7 53.2 2.41 24.0 147 99.4 366.4 49.1 215.3 65.5 168 (506)a 218 (692)a

– – – – – – – – 62 ± 2 35 ± 2 58 ± 2 – – – 40 ± 2 23.8 25.5 – 34.2 ± 1.2 67.9 56 ± 2 54 ± 2 37.5 24.4 – – 62 ± 2 62 ± 2

– – – – – – – – 14,500 23,000 71,500 – – – 74,300 – – – – – 26,000 35,600 – – – – 10,000 16300

[2] [2] [2] [2] [2] [2] [26] [26] [27] [28] [29] [30] [30] [30] [31] [32] [33] [34] [35] [36] [9] [37] [38] [39] [40] [40] [8] This study

a

The value corrected for the number of rhodium atom on the surface of nanoparticle.

eration decreases as the reaction proceeds and finally reaches the zero, may be, due to the increasing viscosity of reaction solution or deactivation effect of tetramethoxyborate concentration. It should be noted that this TTO value can be considered as a lower limit unless the increase in the viscosity and the deactivation of catalyst surface are avoided by removal of tetramethoxyborate product from the reaction medium. Thus much higher TTO values for hydrogen generation from methanolysis of AB catalyzed by Rh(0)/nanoAl2 O3 might be expected. 3.4. Carbon disulfide poisoning as heterogeneity test for Rh(0)/nanoAl2 O3 in the methanolysis of AB A CS2 poisoning experiment was performed to check whether the catalytic methanolysis of ammonia borane is heterogeneous or homogeneous. When the active metal center is blocked by poison molecule, it won’t be accessible by the substrate molecules and catalytic activity will be ceased [25]. In the present work, after the formation of catalytically active rhodium(0) nanoparticles and liberation of 40% hydrogen from methanolysis of ammonia borane, 0.20 equivalent CS2 per mole of rhodium was rapidly injected to reaction mixture containing 200 mM AB catalyzed by 0.24 mM Rh (0.50 wt.% Rh). The hydrogen evolution was stopped immediately (Fig. 11). This observation indicates that the Rh(0)/nanoAl2 O3 catalyst was poisoned by addition of 0.2 equivalent CS2 per mole of rhodium. Thus the CS2 poisoning experiment provides a compelling evidence for the heterogeneity of the methanolysis of ammonia borane catalysed by Rh(0)/nanoAl2 O3 . 3.5. Leaching test for Rh(0)/nanoAl2 O3 in hydrogen generation from the methanolysis of AB When the rhodium(0) nanoparticles are not strongly anchored on the surface of support, rhodium can leach into the solution during the catalytic methanolysis of ammonia borane. This can be checked by performing the widely used leaching test for

the catalytic methanolysis of ammonia borane. Such a control experiment was performed starting with 200 mM AB plus 50 mg Rh(II)/nanoAl2 O3 (0.50 % wt. Rh, [Rh] = 0.24 mM) in 10 mL methanol at 25.0 ± 0.5 ◦ C. After the liberation of 3.0 equiv. H2 gas per mole of AB, the reaction was filtered to separate the solid Rh(0)/nanoAl2 O3 phase from the solution in the inert nitrogen atmophere. The filtrate solution was transferred into another reaction flask. A new batch of 200 mM AB was added to the solution and hydrogen generation was followed at 25.0 ± 0.5 ◦ C. The filtrate solution was found to be catalytically silent in the methanolysis of AB (Fig. 11). The results of the leaching test provides a compelling evidence for the retaining of rhodium(0) nanoparticles within the nanoalumina matrix. In addition, the filtrate solution was also analyzed by ICP and no rhodium was detected in the solution. Taking the results of both poisoning and leaching tests together one can conclude that Rh(0)/nanoAl2 O3 are kinetically competent catalysts and the methanolysis of ammonia borane is heterogeneous catalysis. It is worth to mention that exact measurement of the catalytic activity of isolated solid materials in subsequent runs of methanolysis is not doable because of the noticeable materials loss during the isolation and redispersion processes. Therefore, a reliable reusability test could not be performed. 4. Conclusions In conclusion, nanoalumina stabilized rhodium(0) nanoparticles with tuneable particle size depending on the initial metal loading of the precursor can be in situ generated from reduction of rhodium(II) ions impregnated on the surface of porous nanoalumina during the methanolysis of AB at room temperature. The mean particle diameter of Rh(0)/nanoAl2 O3 increases with the increasing initial rhodium concentration. Evaluation of kinetic data collected by monitoring the volume of hydrogen released can give valuable insights to the formation kinetics of rhodium(0) nanoparticles. The kinetic data curve fits well to the F-W 2-step mechanism consisting of slow, continuous nucleation (rate constant k1 ) and

58

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

Fig. 10. Plot of total turnover number (TTO) or turnover frequency versus time for the methanolysis of AB with 10 mL solution of 0.24 mM Rh (50.0 mg Rh(0)/nanoAl2 O3 , 0.50 wt.% Rh,) and 400 mM AB (for each run) at 25.0 ± 0.5 ◦ C.

Fig. 11. Plots of mole H2 evolved per mole of AB versus time for the methanolysis of 200 mM AB catalyzed by Rh(0)/nanoAl2 O3 (0.24 mM Rh, 0.50 wt.% Rh(0)/nanoAl2 O3 ) with (square, 䊐) and without (circle, ) addition of 0.2 equiv. CS2 , and (triangle, ) the filtrate solution obtained from filtration of active catalyst after methanolysis of AB at 25.0 ± 0.5 ◦ C.

Fig. 9. (a) Plots of mole H2 evolved per mole of AB versus time for the catalytic methanolysis of AB at various temperatures in the range of 20–35 ◦ C keeping the concentration of substrate at [AB] = 200 mM and rhodium at [Rh] = 0.24 mM (0.50 wt.% Rh, 50 mg Rh(0)/nanoAl2 O3 ), (b) The Arrhenius plot for nucleation of rhodium(0) nanoparticles, (c) The Arrhenius plot for the autocatalytic surface growth of rhodium(0) nanoparticles, (d) The Arrhenius plot.

autocatalytic surface growth (rate constant k2 ) for the formation of rhodium(0) nanoparticles catalyst from the reduction of rhodium(II) ions during the methanolysis of AB yielding the values of both rate constants. Nanoalumina stabilized rhodium(0) nanoparticles are highly active and long lived catalyst in hydrogen generation from the methanolysis of AB providing a release of 3.0 equivalent of H2 per mole of AB with an exceptional initial turnover frequency of 218 min−1 (692 min−1 when corrected for the number of rhodium atom on the surface of nanoparticle) which is the highest value ever reported for rhodium catalysts and 16,300 turnovers in hydrogen generation from the methanolysis of AB at 25.0 ± 0.5 ◦ C before deactivation. In addition, the kinetic study reveals that methanolysis is first order with respect to the catalyst concentration and zero order with respect to substrate concentration. Performing the impregnation of rhodium(II) octanoate and then its reduction to Rh(0)/nanoAl2 O3 in the same medium provides an opportunity to avoid laborious catalyst preparation steps. One-pot preparation, high stability and high catalytic activity of nanoalumina stabilized rhodium(0) nanoparticles make them promising

D. Özhava, S. Özkar / Molecular Catalysis 439 (2017) 50–59

candidate to be exploited as catalyst in hydrogen generation from the methanolysis of AB for the fuel cell applications. Acknowledgements Partial support by Turkish Academy of Sciences and TUBITAK for a scholarship to DÖ is gratefully acknowledged. References [1] C.W. Hamilton, R.T. Baker, A. Staubitz, I. Manners, Chem. Soc. Rev. 38 (2009) 279–293. [2] P.V. Ramachandran, P.D. Gagare, Inorg. Chem. 46 (2007) 7810–7817. [3] A. Brockman, Y. Zheng, J. Gore, Int. J. Hydrogen Energy 35 (2010) 7350–7356. [4] P.A. Storozhenko, R.A. Svitsyn, V.A. Ketsko, A.K. Buryak, A.V. Ul’yanov, Russ. J. Inorg. Chem. 50 (2005) 980–985. [5] W.W. Zhan, Q.L. Zhu, Q. Xu, ACS Catal. 8 (2016) 6892–6905. [6] M. Zahmakıran, S. Özkar, Nanoscale 3 (2011) 3462–3481. [7] M.S. El-Shall, V. Abdelsayed, A.E.R.S. Khder, H.M. Hassan, H.M. El-Kaderi, T.E. Reich, J. Mater. Chem. 19 (2009) 7625–7631. [8] D. Özhava, S. Özkar, Appl. Catal. B Environ. 181 (2016) 716–726. [9] D. Özhava, S. Özkar, Int. J. Hydrogen Energy 40 (2015) 10491–10501. [10] S. Akbayrak, S. Genc¸türk, I˙ . Morkan, S. Özkar, RSC Adv. 4 (2014) 13742–13748. [11] M. Trueba, S.P. Trasatti, Eur. J. Inorg. Chem. 17 (2005) 3393–3403. [12] M. Chandra, Q. Xu, J. Power Sources 168 (2007) 135–142. [13] J. Yu, H. Bai, J. Wang, Z. Li, C. Jiao, Q. Liu, M. Zhang, L. Liu, New J. Chem. 37 (2013) 366–372. [14] S.J. Lin, T.N. Hong, J.-Y. Tung, J.H. Chen, Inorg. Chem. 36 (1997) 3886–3891. [15] G.B. Deacon, R.J. Phillips, Coord. Chem. Rev. 33 (1980) 227–250. [16] M.A. Watzky, R.G. Finke, J. Am. Chem. Soc. 119 (1997) 10382–10400. [17] J.A. Widegren, J.D. Aiken, S. Özkar, R.G. Finke, Chem. Mater. 13 (2001) 312–324.

59

[18] B.A. Kakade, S. Sahoo, S.B. Halligudi, V.K. Pillai, J. Phys. Chem. C 112 (2008) 13317–13319. [19] V. Mévellec, A. Nowicki, A. Roucoux, C. Dujardin, P. Granger, E. Payen, K. Philippot, New J. Chem. 30 (2006) 1214–1219. [20] Z.W. Sieh, R. Gronsky, A.T. Bell, J. Catal. 170 (1997) 62–74. [21] A.P. Umpierre, E. deJesús, J. Dupont, ChemCatChem 3 (2011) 1413–1418. [22] B.A. Steinhoff, S.R. Fix, S.S. Stahl, J. Am. Chem. Soc. 124 (2002) 766–767. [23] S. Akbayrak, S. Özkar, J. Mol. Catal. A Chem. 424 (2016) 254–260. [24] S. Akbayrak, Y. Tonbul, S. Özkar, Dalton Trans. 45 (2016) 10969–10978. [25] B.J. Hornstein, J.D. Aiken, R.G. Finke, Inorg. Chem. 41 (2002) 1625–1638. [26] S.B. Kalidindi, U. Sanyal, B.R. Jagirdar, Phys. Chem. Chem. Phys 10 (2008) 5870–5874. [27] D. Özhava, N.Z. Kilic¸aslan, S. Özkar, Appl. Catal. B: Environ. 162 (2015) 573–582. [28] H. Erdo˘gan, Ö. Metin, S. Özkar, Phys. Chem. Chem. Phys 11 (2009) 10519–10525. [29] H. Erdo˘gan, Ö. Metin, S. Özkar, Catal. Today 170 (2011) 93–98. [30] S.B. Kalidindi, A.A. Vernekar, B.R. Jagirdar, Phys. Chem. Chem. Phys 11 (2009) 770–775. [31] S. C¸alis¸kan, M. Zahmakiran, S. Özkar, Appl. Catal. B: Environ. 93 (2010) 387–394. [32] H.B. Dai, X.D. Kang, P. Wang, Int. J. Hydrogen Energy 35 (2010) 10317–10323. [33] D. Sun, V. Mazumder, Ö. Metin, S. Sun, ACS Catal. 2 (2012) 1290–1295. [34] P. Li, Z. Xiao, Z. Liu, J. Huang, Q. Li, D. Sun, Nanotechnology 26 (2014) 25401–25410. [35] Q. Yao, M. Huang, Z. Lu, Y. Yang, Y. Zhang, X. Chen, Z. Yang, Dalt. Trans. 44 (2015) 1070–1076. [36] M. Yurderi, A. Bulut, I.E. Ertas, M. Zahmakiran, M. Kaya, Appl. Catal. B: Environ. 165 (2015) 169–175. [37] S. Peng, J. Liu, J. Zhang, F. Wang, Int. J. Hydrogen Energy 40 (2015) 10856–10866. [38] D. Sun, P. Li, B. Yang, Y. Xu, J. Huang, Q. Li, RSC Adv. 6 (2016) 105940–105947. [39] C. Yu, J. Fu, M. Muzzio, T. Shen, D. Su, J. Zhu, S. Sun, Chem. Mater. 29 (2017) 1413–1418. [40] J.K. Sun, W.W. Zhan, T. Akita, Q. Xu, J. Am. Chem. Soc. 137 (2015) 7063–7066.