Structure-activity relationships of carbon-supported platinum-bismuth and platinum-antimony oxidation catalysts

Journal of Catalysis 348 (2017) 47–58

Contents lists available at ScienceDirect

Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat

Structure-activity relationships of carbon-supported platinum-bismuth and platinum-antimony oxidation catalysts Mabuatsela V. Maphoru a, Sreejarani Kesavan Pillai b, Josef Heveling a,c,⇑ a

Department of Chemistry, Tshwane University of Technology, Pretoria 0001, South Africa National Centre for Nano-Structured Materials, Council for Scientific and Industrial Research, Pretoria 0001, South Africa c Institute for NanoEngineering Research, Tshwane University of Technology, Pretoria 0001, South Africa b

a r t i c l e

i n f o

Article history: Received 28 November 2016 Revised 2 February 2017 Accepted 5 February 2017

Keywords: Nanocomposite Platinum Activated carbon Promoter Oxidative coupling Autocatalytic oxidation

a b s t r a c t Compositional and morphological studies on supported platinum are important for the improvement and expanded use of catalysts for oxidative coupling reactions. Nanocomposites consisting of 5% Pt supported on activated carbon and promoted with 5% Bi or Sb were prepared by electroless deposition and microwave-assisted (MW) methods. Addition of promoters significantly increases the dispersion of Pt. Bismuth reacts with residual phosphorus of the support to form various phases of BiPO4, while Sb cannot be detected by XRD. However, samples prepared by the MW method are unique in that they contain crystalline PtBi or PtSb alloys as part of the phase matrix. The thermal stability of the samples in air and the TOFs for the oxidation of 2-methyl-1-naphthol correlate with the metal dispersion. Since the oxidation reaction is understood to take place on the surface of metals with high standard electrode potentials, sufficient Pt exposure is one of the key performance parameters. Ó 2017 Elsevier Inc. All rights reserved.

1. Introduction The IUPAC nomenclature [1] distinguishes between amorphous carbon (no long-range crystalline order), non-graphitic carbon (two-dimensional long-range order of the carbon atoms in planar hexagonal networks), and graphitic carbon (three-dimensional hexagonal crystalline long-range order). In principal, activated carbon (AC) consists of non-graphitic carbon, as any tarry and amorphous carbon is removed during the manufacturing (activation) process, and graphite is normally absent [2,3]. AC has a highly porous texture and is characterized by high thermal stability (in the absence of air), a large surface area, and phenomenal adsorptive properties [2]. AC can intercalate metal nanoparticles (NPs) to form composite materials that are useful in water purification, fuel cells and general catalysis [3–6]. AC has the added advantage of supporting the product life cycle: carbon burns readily when heated in air, allowing supported metals to be easily recovered and recycled [3]. Metal/AC composites can be prepared with the help of chemical and physical methods. Chemical methods include impregnation ⇑ Corresponding author at: Department of Chemistry, Tshwane University of Technology, Pretoria 0001, South Africa. E-mail address: [email protected] (J. Heveling). http://dx.doi.org/10.1016/j.jcat.2017.02.003 0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

and electroless (reductive) deposition (ED). Physical methods make use of lasers, sonication, microwave (MW) and UV irradiation, as well as plasma technology and supercritical fluids [7]. The method of preparation has a decisive influence on the structure and morphology of the final product [8]. Depending on the synthesis conditions, different metal and oxide phases can be generated, which allows tuning of the sample properties to the requirements of the envisaged application. The shape and size distribution of NPs are also important, e.g. in catalysis the particle size has often a crucial influence on catalyst activity and selectivity [9,10]. Precipitation of metals onto supports by deposition methods results frequently in an extended particle-size distribution (PSD), accompanied by agglomeration [7]. In contrast, use of MW irradiation during synthesis can be expected to create homogeneous nucleation sites, resulting in a narrow PSD with a high metal dispersion and uniform particle shapes [10,11]. In addition, MW irradiation offers shorter reaction times and reduced energy consumption [10,12]. In combination with ethylene glycol it is a very convenient, fast and efficient method of metal deposition onto supports [13], as glycol can act both as the solvent and as the reducing agent for metal precursors [14]. Pt-Bi and Pt-Sb composites have applications in thermoelectronics and batteries [15–17]. Carbon-supported Pt-Sb is a promising material for direct formic-acid and methanol fuel cells

48

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

[6,17,18]. Pt-Bi2O3 was found by Moniz et al. [19] to facilitate the photolysis of water, which neither Pt nor Bi2O3 can accomplish on their own. Pt-Bi/AC is a commercially available catalyst for electrocatalysis and selective oxidation reactions. With respect to the latter application, it is speculated that Bi exerts a geometric blocking effect on Pt, thus decreasing the active site ensembles of Pt. These minimized sites still allow for the desired chemical reaction to proceed, but they are inactive for unwanted side reactions [20–23]. In our previous studies, 5%Pt-5%Bi/AC was used for the oxidation and oxidative coupling of 1-naphthols [24,25], while 5%Pt-5%Sb/AC was shown to have superior selectivity for the oxidation of cinnamyl alcohol to cinnamaldehyde in comparison with corresponding Bi-containing catalysts [26]. In the present study, the properties of 5%Pt-5%Bi/AC and 5%Pt-5%Sb/AC composite materials prepared by ED, including MW-assisted glycol reduction, are described. The different nanofeatures obtained under varying synthesis conditions have an influence on the autocatalytic oxidation of the composite materials. In addition, they have an impact on the catalytic activities and selectivities of chemical oxidation reactions; this is demonstrated by example of the oxidative coupling of 2-methyl-1-naphthol (Scheme 1). The oxidative coupling of naphthols is an important method for the formation of new CAC bonds [27,28]. Apart from other reaction parameters, the outcome depends on the composition and the properties of the catalyst, as well as on the electronic and steric nature of the substrate. The formation of coupled products occurs often concurrently with the formation of monoquinones. Only a limited amount of work described in the literature is dedicated to the heterogeneous catalytic oxidation of 2-methyl-1-naphthol (Scheme 1). Most of these studies are aimed at the conversion of 1 to menadione (4), and the coupled products (2 and 3) are considered to be by-products [29–32]. Menadione is a blood coagulating agent (vitamin K3) [33]. The development of a viable and environmentally sustainable catalytic route to this molecule is of great commercial interest. High yields of 4 can be obtained over oxide catalysts in combination with hydrogen peroxide, e.g. Ti-MMM-2 (mesoporous titanium-silicate): 78% [34], Nb2O5-SiO2: 42% [35], Ti-SBA-15: 93% [31], Nb-SBA-15: 97% [32]. However, isolated yields were not reported, and the mass of the catalyst in relation to the mass of the starting material (SM) is always high (SM/cat. = 0.9–2.0 m/m). In contrast to oxides, metallic-state catalysts favor the formation of coupled products [24,25]. Depending on the reaction conditions and the solvent used, 2 or 3 can be obtained in high yields over Pt supported on AC (96–97%, SM/cat. = 25 m/m). Previously, Kral and Laatsch obtained 3 with a yield of 72% by the coupling of 1 with stoichiometric amounts of AgO as the oxidant [36]. The maximum menadione yield obtained over Pt is 30.5% [25].

2. Experimental 2.1. Materials Activated carbon (Darko KB-G 174), bismuth(III) nitrate pentahydrate (98%), antimony chloride (99%), hydrazine monohydrate (98%), nitromethane (95%) and silica gel for column chromatography (pore size 60 Å, 70–230 mesh) were purchased from SigmaAldrich. Hexachloroplatinic acid (metal content 40.31%) was bought from SA Precious Metals. Hydrogen peroxide (30%), dichloromethane (99.5%), ethyl acetate (99%), methanol (99.5%) and petroleum ether (90%) were obtained from SMM Instruments, ethylene glycol (99.5%) from Promark Chemicals, and Triton X-100 (98–102%) from BDH. 2-Methyl-1-naphthol was prepared as described previously [24].

2.2. Synthesis of nanocomposites Electroless deposition (ED): for the preparation of 5%Pt-5%Bi on AC, H2PtCl6∙6H2O (0.1328 g, 0.2564 mmol Pt) and Bi(NO3)3∙5H2O (0.1162 g, 0.2396 mmol Bi) were dissolved in 50% H2O/50% CH3OH (v/v, 100 ml). Triton X-100 (3 ml) was added, and the mixture was placed in an ultrasonic bath for 2 h to completely dissolve the bismuth nitrate. After addition of the AC support (0.9005 g), the mixture was stirred using a magnetic stirrer. After 10 min a 98% aqueous solution of N2H4∙H2O (0.2 ml, 4.041 mmol) was added dropwise with a syringe to reduce the metal ions. Stirring continued for 28 h. Each sample was filtered and washed with deionized water to pH 7, and finally washed with CH3OH. By testing with AgNO3, the final aqueous filtrate was found to be free of chloride. Samples were dried under different conditions, i.e. at room temperature (r.t.) under vacuum for 4 h (VD), or in an oven at 140 °C for 24 h (OD). The same procedure was used to prepare VD 5%Pt/AC, as well as VD and OD 5%Pt-5%Sb/AC; SbCl3 (0.1228 g, 0.4108 mmol) was the precursor for Sb. Microwave-assisted loading (MW): for the preparation of MW 5%Pt-5%Bi/AC, H2PtCl6
6H2O (0.1328 g, 0.2564 mmol) and Bi(NO3)3
5H2O (0.1161 g, 0.2393 mmol) were transferred into a 100 mL beaker containing ethylene glycol (20 mL), followed by ultrasonication for 10 min to completely dissolve the salts. To this mixture, AC (0.9001 g) was added, followed by ultrasonication for 1 h prior to MW irradiation. The MW-assisted metal loading was conducted in a Mars Xpress MW system at 600 W and 190 °C in a closed Teflon vessel. The time to reach the final temperature was set at 5 min, and irradiation continued for 10 min at 190 °C to allow for the complete reduction of Pt(IV) to Pt(0). Samples were centrifuged, washed with deionized water, and left to airdry (AD) for 24 h. The same method was followed to prepare

Scheme 1. Oxidation and oxidative coupling of 2-methyl-1-naphthol (1) using H2O2 as the oxidant.

49

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58 Table 1 Preparation methods, drying conditions, metal loadings and physisorption data of samples. Sample

Activated carbon VD 5%Pt/AC VD 5%Pt-5%Bi/AC OD 5%Pt-5%Bi/AC MW 5%Pt-5%Bi/AC VD 5%Pt-5%Sb/AC OD 5%Pt-5%Sb/AC MW 5%Pt-5%Sb/AC

Sample designation

AC VD Pt VD Bi OD Bi MW Bi VD Sb OD Sb MW Sb

Prep. method

– ED ED ED MW ED ED MW

Drying method

– VD VD OD AD VD OD AD

MW 5%Pt-5%Sb/AC; SbCl3 (0.1228 g, 0.4111 mmol) was used as the precursor for Sb. The designations used for the different samples (depending on the preparation/drying methods) are given in Table 1. 2.3. Characterization of nanocomposites Elemental analyses (Pt, Bi, Sb and P) were performed on a Spectro Arcos FHS ICP-OES instrument. The amount of metal loaded onto the AC support was quantified indirectly by determining the amount of metal that remained in solution during sample preparation. A Bruker D8 Advance powder X-ray diffractometer was employed with Ni-filtered Cu Ka radiation (k = 0.15418 nm) at 40 kV and 40 mA to analyze the crystallographic structure of the composites. The samples were scanned at r.t. at a scan speed of 0.1 s/step in the 2h range of 10–100°. The phases were identified using the Bruker DIFFRACplus evaluation software (EVA) in combination with the ICDD powder diffraction database (International Centre for Diffraction Data). Nitrogen-physisorption measurements (BET) were carried out using a Micrometrics ASAP 2020 surface area and porosity analyzer. Approximately 0.05 g of each sample was degassed for 6 h at 150 °C and analyzed at 196 °C (77 K) for 5 h in the pressure range 0.866–87.1 kPa using nitrogen as the adsorbent. Pore volumes and average pore diameters were calculated using the BJH method. Sample morphologies and elementary compositions were analyzed by means of a JEOL JSM7500F field-emission scanning electron microscope, equipped with a secondary electron detector and a Thermo Scientific ultradry X-ray detector. An electron accelerating voltage of 2.00 kV or 3.00 kV was used for imaging, and 15.0 kV was used for imaging, energy dispersive X-ray (EDX) analysis and mapping. All samples were sputter coated with carbon to avoid charging. Highresolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) were performed on a JEOL JEM-2100 HRTEM instrument. The electron acceleration voltage was 200 kV. The samples were dispersed in ethanol for 1 min using an ultrasonic bath, and supported on carbon-coated copper grids (200 mesh). Particle sizes were measured with the help of the ImageJ software. For each sample 500 particles were measured using four to six different images per sample. The thermal behavior of materials was assessed using a Perkin Elmer TGA 4000 thermogravimetric analyzer at an air flow of 20 ml/min and a heating rate of 10 °C/min in the temperature range of 50–1000 °C. 2.4. Oxidation reactions, product separation, product characterization and determination of catalyst performance Oxidative coupling reactions were carried out at r.t. and at 60 °C by placing 2-methyl-1-naphthol (6.34 mmol) and the catalyst (0.0401 g) into a 100 ml round-bottomed flask, equipped with a condenser. A solvent (25 ml) was added. The mixture was stirred with a magnetic stirring bar. While stirring, 30% aqueous H2O2

Drying temp.

Drying time

Pt

Bi

(°C)

(h)

(mass%)

– r.t. r.t. 140 r.t. r.t. 140 r.t.

– 4 4 24 24 4 24 24

– 5.0 4.9 4.9 5.0 5.0 4.8 5.0

– – 4.8 4.9 4.9 – – –

Sb

S

PV 2

– – – – – 4.8 4.8 4.9

APD 3

(m /g)

(m /g)

(nm)

1195 865 655 535 860 625 620 855

0.82 0.66 0.45 0.55 0.56 0.44 0.54 0.54

4.6 5.5 5.7 8.1 5.5 5.8 5.7 5.4

(3.1 ml, 22.1 mmol) was added dropwise over 40 min directly into the solution using a peristaltic pump. The catalyst was filtered off, washed under suction with CH2Cl2, followed by MeOH, and left to air dry. Column chromatography was used to separate the reaction mixtures [25]. Product selectivities were calculated as mole percentages based on the moles of SM incorporated into each product. The full characterization of 2-methyl-1-naphthol (1), 3,30 -dimethyl-1,10 -binaphthalenyl-4,40 -diol (2), 3,30 -dimethyl-1,10 binaphthalenylidene-4,40 -dione (3) and 2-methyl-1,4-naphthoquinone (4) was reported previously [24,25]. Turnover frequencies (TOFs) were calculated as moles of converted SM per mole of Pt present on the catalyst over the total reaction time (min–1). Catalysts recycling was tested under conditions identical to those used for fresh catalysts; that is, the molar ratios of all reaction ingredients were kept constant in relation to the amount of the reused catalyst. 3. Results and discussion 3.1. Characterization of the nanocomposites The synthesis methods, the drying conditions, and the physisorption data obtained for all samples are shown in Table 1. Independent of the synthesis method used, the metal loadings are always near to the targeted values. The BET surface area of the AC support is close to 1200 m2/g. Lower total surface areas (S) and lower pore volumes (PV) are always found after metal loading, probably due to some pore blockage. Obviously, smaller pores are blocked preferentially, resulting in increased average pore diameters (APD) of the composite materials. Restructuring of AC during sample preparation could also influence the surface area. For instance, Zhu and coworkers reported the efficient use of MW treatment for the exfoliation of graphite, accompanied by a considerable increase in surface area [37]. Similarly, our metalloaded composites prepared by the MW method distinguish themselves by relatively higher surface areas (855–860 m2/g) compared to the VD and OD samples. XRD patterns of AC, VD Pt and the bimetallic composites are shown in Fig. 1a and b. The XRD trace of carbon (a) is typical for activated carbon [38,39]; reflections for graphite [38–40] are absent. The peaks at 2h = 39.9, 46.9, 68.4, 81.6 and 86.3° correspond to the (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2) planes of face-centered cubic (fcc) Pt(0), corresponding to the powder diffraction file (PDF) 01-089-7382. Bi has a low reduction potential (E0 = 0.320 V) and cannot be expected to be present in a fully reduced state. On VD Bi (Fig. 1a) a hexagonal phase of bismuth phosphate [BiPO4(h), ximengite, PDF 01-080-0208] is found, while for the OD and MW Bi samples monoclinic polymorphs of space group P21/m [BiPO4(mm), PDF 01-077-2208] and P21/n [BiPO4(nm), monazite, PDF 01-080-0209] are formed, respectively. According to the product specification of the supplier, the chosen AC support contains 3.5% PO3 4 . In our batch 0.77% P was found, which corresponds to 2.4% PO3 4 . This means that the support would contain

50

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

Fig. 1. XRD patterns of AC, VD Pt and Pt-Bi nanocomposites (a); and AC, VD Pt and Pt-Sb nanocomposites (b).

sufficient phosphate to convert all the Bi of the 5%Pt-5%Bi samples into BiPO4. This is possibly the case for the VD and OD composites, as no other Bi peaks are present. A study on BiPO4 crystals conducted by Mooney-Slater [41] found that the BiPO4 phase formed depends on the preparation and treatment temperature. Ximengite has a highly symmetrical arrangement of chains, which forms an open framework. It is stable at relatively low temperatures and loses half of its water molecules between 75 and 175 °C. Between 200 and 250 °C ximengite undergoes an exothermic chain reorientation to form monoclinic monazite of space group P21/n. Monazite has a less symmetrical chain arrangement, resulting in a compact spacefilling network. When monazite is exposed to temperatures between 600 and 700 °C, it transforms into another monoclinic modification (space group P21/m). In our case, the preparation conditions used during synthesis affect the BiPO4 phases formed. The composite dried at r.t. under vacuum (VD Bi) produces ximengite, the composite prepared in the same way, but dried at 140 °C (OD Bi) forms monoclinic BiPO4 of space group P21/m, and the MW method results in the formation of monazite. As observed for the Pt-Bi samples, the Pt-Sb composites (refer Fig. 1b) show broad XRD reflections of fcc Pt(0). In this regard, SAED (Fig. S1, Supporting information) confirms the presence of crystalline material in the VD Sb sample; a few diffraction spots are visible, which are attributed to Pt, since Sb is amorphous. By XRD, Sb is not detectable on the VD and OD samples. This is not exceptional for supported Sb, e.g. Yu and Pickup [6,17] could not find any diffraction peaks for Sb-containing phases on carbonsupported Pt-Sb prepared by reduction with NaBH4. Likewise, Dailly et al. [15] did not see any Sb peaks on an antimonygraphite composite material prepared by reduction of SbCl5 with NaH. Of our three synthesis procedures used, the MW-assisted method is unique in that it creates hexagonal PtBi (PDF 01-0725643) and PtSb (PDF 03-065-3432, stumpflite) alloys as part of the phase matrix. The alloy peaks are small and broad, suggesting that the alloys are almost amorphous. MW-assisted polyol reduction simultaneously reduces and deposits metals onto the support. In the case of bimetallic systems, this can indeed result in alloy formation [42]. In contrast to the Pt-Bi samples (Fig. 1a), phosphates (SbPO4) are not observed in the Pt-Sb composites (Fig. 1b). SbPO4 can be prepared at temperatures >650 °C by the reaction of metaphosphoric acid with Sb to give antimony meta-phosphate, which thermally decomposes to SbPO4 at 900 °C [43]. This means, the formation of SbPO4 is favored at high temperatures. Evidently, our conditions of sample preparation are not conducive for the formation of SbPO4.

It is noted, that the presence of phosphorus (or nitrogen) can be an advantage for certain applications. For our intended use (catalysts for selective oxidation reactions) significant improvements in reaction rate or selectivity were observed after addition of P- or N-containing compounds that adsorb strongly on the Pt surface [21]. In addition, combinations of Pt-group metals with phosphates are more promising catalyst formulations for oxidation reactions than, for example, Cu2(OH)PO4 or Cu4O(PO4)2, which are only of low activity [21]. SEM and TEM images of VD Pt and the bimetallic Pt-Bi nanocomposites are shown in Fig. 2. For monometallic VD Pt the SEM image (Fig. 2a0 ) indicates the presence of highly agglomerated spherical Pt particles. This is confirmed by HRTEM (Fig. 2a00 ). An average particle size (APS) of 36.5 nm with a relatively broad PSD (18–80 nm) was determined (Fig. 2a000 ). The agglomerated NPs leave vacant spaces on the carbon carrier, contributing to a relatively high PV and a total surface area of 865 m2/g (Table 1). The addition of Bi to VD Pt significantly improves the dispersion of Pt 000 (Fig. 2b00 ). By adding 5% Bi, the APS decreases to 8.03 nm (Fig. 2b ), and some of the Pt NPs are isolated and others are associated with the Bi containing features, which appear as nanorods (Fig. 2b0 ). The elemental composition of the prominent features with respect to Pt, Bi, P and oxygen was examined by ‘‘point and shoot” EDX analysis. Although the EDX peak for Pt at 2.048 keV overlaps with the peak for P at 2.013 keV, the results (Fig. 3a) confirm that P is present in significant amounts. Points 1–5, 7 and 10 are positioned on rods and contain a high amount of Bi, associated with substantial amounts of P and O. Points 6, 8 and 9 are rich in Pt and do not contain P. This suggests that the rods consist of the BiPO4(h) phase found by XRD (Fig. 1a). The association and presence of Bi and P on the nanorods are also confirmed by SEM-EDX elemental mapping (Fig. 3b), while Pt is seen to be more evenly distributed over the catalyst surface. Hexagonal BiPO4 nanorods were previously prepared from BiNO3 and Na2PO4 in nitric acid and characterized by several methods, including XRD, SEM and HRTEM [44,45]. The images of VD Bi in Figs. 2b0 –b00 and 3 show that there is some variation with respect to the location of Pt relative to Bi. In a somewhat simplified approach, the structures of twocomponent supported composite materials can be categorized according to the models shown in Fig. 4. Using this system, VD Bi would fit Model B. The images of Figs. 2c0 –c00 and S2 point to the presence of particles of different sizes and shapes on OD Bi. In addition to agglomerates, SEM shows isolated NPs, which contain mainly Pt (Fig. S2). The NPs are highly dispersed and have an APS of 3.09 nm (Fig. 2c000 ). Small particles are possibly more effective in blocking some of the narrow pores of the AC support. Accordingly, OD Bi has the lowest surface area (535 m2/g) of all the materials shown in Table 1 and an

51

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

a'

Number of particles (%)

a'' Agglomerated Pt spheres Agglomerated Pt spheres 100 nm

Size d (nm)

BiPO4 rods

b''

80

Number of particles (%)

b'

a'''

24 36.5 ±1.02 nm 22 20 18 16 14 12 10 8 6 4 2 15 20 25 30 35 40 45 50 55 60 65 70 75 80

100 nm

b'''

8.03 ±0.06 nm

70 60 50 40 30 20 10 0 0

4

8

12 16 20 24 28 32 36 40

Size d (nm)

c''

Embedded NPs

3.09 ±0.04 nm

Number of particles (%)

c'

1 µm

c'''

140 120 100 80 60 40 20 0 0

2

4

6

8 10 12 14 16 18 20 22 24

Size d (nm)

d'

Pt NPs Bi species 10 µm

Number of particles (%)

d''

5.74 ±0.13 nm

110 100 90 80 70 60 50 40 30 20 10 0 0

2

4

6

8

10

12

14

d'''

16

18

Size d (nm) Fig. 2. SEM images, HRTEM images and PSDs for VD Pt (a), VD Bi (b), OD Bi (c), and MW Bi (d).

exceptionally high APD (8.1 nm). Rod-like features of BiPO4 (like those in Fig. 2b0 ) are not observed, but the agglomerates are rich in both Bi and P (Fig. S2). Probably, the agglomerates contain the monoclinic BiPO4(mm) phase found by XRD (Fig. 1a). In contrast to VD Bi, the particles are closely attached to or engraved into the surface of the AC. Hence, oven-drying seems to promote the anchoring of small particles onto the surface or into the pores of the support. Overall, the structure of OD Bi corresponds to Model C of Fig. 4.

000

For MW Bi an APS of 5.74 nm was determined (Fig. 2d ). The particles are less uniform compared to the particles observed for OD Bi. Like OD Bi, MW Bi shows no (or rather few) rod-like features. NPs and larger agglomerates are seen by SEM (Fig. 2d0 ). The NPs are rich in Pt, while the agglomerates consist mainly of Bi and P associated with Pt (Fig. S3). Most likely, the agglomerates contain the monazite phase BiPO4(nm) that was identified by XRD (Fig. 1a). The overall structure is best described by Model C of Fig. 4.

52

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

5

a 4

1

10

Point 1 2 3 4 5 6 7 8 9 10

C (%) 76.0 77.1 80.0 79.5 71.2 77.1 80.6 69.1 81.9 74.9

O (%) 15.7 13.3 13.3 14.1 17.5 7.44 12.3 4.45 9.16 15.7

Al (%) 0.17 0.15 0.15 0.18 0.76 0.97 0.14 0.33

Pt (%) 0.66 2.44 1.12 0.67 0.88 11.7 3.33 25.7 7.76 0.91

Bi (%) 4.73 4.94 3.50 3.21 7.19 2.46 0.11 0.02 0.03 6.45

P (%) 1.12 0.93 0.89 0.94 1.55 1.89 1.41

b

Fig. 3. ‘‘Point and shoot” EDX analysis (a) and SEM-EDX elemental mapping (b) of VD Bi.

Fig. 4. Models for the particle distribution in two-component supported systems, A phase segregated, B partially segregated (or partially associated), C full association of only one component (in our case Bi or Sb = red spheres), D full association of both components. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

For VD Sb, SEM (Fig. 5a0 –a00 ) points to homogeneously dispersed NPs that consist of Pt and Sb in somewhat varying proportions (Fig. S4). For carbon-supported Pt-Sb (35–39% Pt and 5–16% Sb) Yu and Pickup [6] also found that Pt and Sb are always associated.

Interestingly, in our sample agglomerated chain-like structures with a diameter of 11.6 nm are observed by HRTEM (Fig. 5a000 ). EDX analysis (Fig. S5) indicates that the nanochains contain both Pt and Sb. No fringes were detectable on the HRTEM images. This

53

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

a'

a''

a''' Agglomerated nanochains

10 µm

b'

100 nm

b'''

b''

Agglomerated nanochains Embedded NPs 10 µm

Pt-Sb NPs

c''

Pt-Sb agglomerated

10 µm

7.88 ±0.19 nm

90

Number of particles (%)

c'

1 µm

80

c'''

70 60 50 40 30 20 10 0 0

2

4

6

8 10 12 14 16 18 20 22 24 26

Size d (nm) Fig. 5. SEM and HRTEM images (and PSD) for VD Sb (a), OD Sb (b) and MW Sb (c). (Due to the nanochain-like morphology, PSDs could not be determined for a and b.)

suggests the presence of almost amorphous Pt embedded in amorphous Sb or vice versa, thus forming a nearly amorphous matrix, corresponding to Model D of Fig. 4. The morphology of OD Sb (Fig. 5b) is similar to that observed for VD Sb. In the SEM image, spherical particles are seen on the surface of AC (Fig. 5b0 –b00 ), while a chain-like morphology emerges by 000 HRTEM (Fig. 5b ). As for VD Sb, the nanochains contain both Pt and Sb (Fig. 6). Compared to VD, OD leads to particles that are more firmly attached to or embedded into the surface of AC. This was also observed for OD Bi (Fig. 2c0 ). The SEM image of MW Sb (Fig. 5c0 ) indicates the presence of NPs and agglomerates. According to Fig. S6, all features contain Pt and Sb (corresponding to Model D of Fig. 4). Uniform crystalline cubic particles containing Pt and Sb are seen by HRTEM (Fig. 5c00 ), and an APS of 7.88 nm was calculated (5c000 ). Nanochain-like structures are not observed. As for MW Bi, the total surface area (855 m2/g) remains relatively high (Table 1). The structural properties of all composites are summarized in Table 2. 3.2. Thermal analysis of the AC support and the nanocomposite materials The thermal stabilities of AC and of the composite materials were assessed by TGA in air (Fig. 7). The initial temperatures Ti

(temperature at which the materials start losing carbon), the 1stderivative-peak temperatures Tp (minimum of the first derivative curve), and the final temperatures Tf (temperature at which the mass loss is complete) are listed in Table 3. It is well known that different structural forms of carbon exhibit different oxidation behaviors. Higher oxidation temperatures are associated with purer, less defective samples [46,47]. Less ordered or ‘‘amorphous” carbons tend to oxidize around 350 °C, and there seems to be a consensus on an oxidation temperature of ca. 400 °C [47]. An even lower decomposition temperature (ca. 190 °C) was reported by Ko et al. [12]. (Unfortunately, the literature does not clearly distinguish between amorphous carbon and non-graphitic carbon as defined by IUPAC [1].) A well graphitized structure starts to oxidize at relatively higher temperatures of 600–800 °C [38,40,47,48]. In any case, in the presence of oxidation catalysts, such as Pt, the oxidation temperatures must be expected to shift to lower values [49,50]. Therefore, as seen in Fig. 7, the decomposition temperatures for the non-graphitic carbon of our Pt-loaded composite materials are never higher than that of AC itself. The TGA curve of AC is indicative of a high quality, nongraphitic material that contains no amorphous carbon. This finding is in agreement with the XRD (Fig. 1) and the Raman spectrum shown in Ref. [40]. After oxidation of all the combustibles, AC leaves a residue of 5.1%, corresponding to the ash content. The residues found for the metal containing composites must be expected

54

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

Point

6

1 2 3 4 5 6 7

5

1

C (%) 40.1 58.3 71.0 84.8 82.1 79.9 67.6

7 4 2 3

O (%) 4.76 8.13 10.7 8.30 6.61 7.88 6.19

Al (%) 0.08 0.09 0.06 0.08 0.04

Pt (%) 49.4 28.1 12.8 4.74 8.38 9.08 22.3

Sb (%) 5.69 5.50 4.57 1.36 1.98 2.73 3.89

P (%) 0.43 0.17 0.38 -

2.5 µm

2.5 µm

2.5 µm

Fig. 6. ‘‘Point and shoot” EDX analysis (top) and SEM-EDX elemental mapping (bottom) of OD Sb.

Table 2 Structural properties of the AC supported composites. Sample designation

Bi phasesa

Sb phasesa

APSb (nm)

Morphologyc

Particle distributiond

VD Pt VD Bi

– BiPO4(h)

– –

36.5 8.03

– B

OD Bi MW Bi

BiPO4(mm) BiPO4(nm) PtBi – – –

– –

3.09 5.74

amorph. amorph. PtSb

– – 7.88

Pt: agglomerated spheres Pt: partially agglom. NPs; Bi: partially agglom. nanorods; Pt and Bi partially segregated Pt: NPs; Bi: agglomerates; Pt mostly ass. with Bi and Bi always ass. with Pt Pt: partially agglom. NPs; Bi: agglomerates; Pt partially ass. with Bi and Bi always ass. with Pt agglom. Pt-Sb nanochains agglom. Pt-Sb nanochains Pt-Sb NPs and agglomerates; Pt and Sb always associated

VD Sb OD Sb MW Sb

a

100

b

100

90

90 70

60

MW Bi

60 50

Temperature ( C)

o

900

700

600

900

1000

o

800

700

600

500

0

400

10

0

300

10

200

20

500

VD Sb

20

400

MW Sb

30

300

40

200

OD Bi

30

100

VD Bi

40

VD Pt OD Sb

1000

VD Pt

50

AC

80

70

800

AC

80

100

c d

Determined by XRD, Pt is always present as fcc Pt(0). Determined by HRTEM. Determined by SEM, HRTEM and EDX. According to Fig. 4.

Weight %

a b

Temperature ( C)

Fig. 7. TGA curves of AC, VD Pt, and Pt-Bi nanocomposites (a); and AC, VD Pt, and Pt-Sb nanocomposites (b).

C C D D D

55

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58 Table 3 Thermogravimetric data of AC and metal-loaded AC.

AC VD Pt VD Bi OD Bi MW Bi VD Sb OD Sb MW Sb

Non-graphitic carbon

Tia (°C)

Tfa

Tin (°C)

Tpn

Tfn

– 214 161 – – 229 276 –

– 434 411 – – 431 454 –

464 464 411 340 421 431 466 437

647 635 536 483 555 506 559 553

808 763 663 619 651 569 680 635

to vary with the total remaining ash content of the support, the metal loadings, and the final oxidation states. According to the thermograms of Fig. 7a the temperature for the loss of non-graphitic carbon increases in the order OD Bi < VD Bi < MW Bi < VD Pt < AC. Only VD Bi shows a significant mass loss in the temperature region that could be assigned to the oxidation of amorphous carbon. Obviously, the amorphous carbon is formed from non-graphitic carbon of AC under the catalyst synthesis conditions. Hardly any loss of amorphous carbon is seen for MW Bi. This is not surprising, as MW irradiation is an efficient method for the removal of ‘‘amorphous” carbon from, e.g. carbon nanotubes [51]. It can therefore be expected that during the MW-assisted synthesis any amorphous carbon formed is removed from the sample. This could also contribute to the high ash content observed for the MW composites (Fig. 7), as the selective removal of some carbon would result in an increase in the ash/carbon ratio. VD Pt shows the highest Tpn of all the metal-loaded samples (635 °C, Table 3). It contains large agglomerated Pt particles (Fig. 2a0 –a00 ), and it is conceivable that during the TGA run the relatively small amount of exposed Pt gets over-oxidized (poisoned) by oxygen. This would retard the further oxidation of carbon. Over-oxidation of the catalytically active metal surface is one of the major deactivation mechanisms anticipated for (unpromoted) precious-metal oxidation catalysts [21]. The thermograms of the Pt-Sb composites (Fig. 7b) display similar trends to those observed for Pt-Bi. Again, a considerable mass loss in the temperature region for the oxidation of amorphous carbon is mainly observed for the VD sample, and loss of amorphous carbon is particularly low for MW Sb. The temperature for the decomposition of non-graphitic carbon increases in the order VD Sb < MW Sb ffi OD Sb < VD Pt < AC. Fig. 8 shows a correlation between the total surface area and Tpn for all samples (including AC), and it is found that Tpn increases with the surface area. This is in contrast to the situation found on pure carbons, for which an inverse relationship is observed

700

Tpn (oC)

650

VD Pt

AC

600 OD Sb MW Bi

550 VD Bi

500 450 400 500

MW Sb

VD Sb OD Bi

600

700

Residue (wt%)

800

900

1000

1100

1200

Surface area (m2/g) Fig. 8. First-derivative-peak temperature of non-graphitic carbon (Tpn) as a function of the total surface area of the materials.

Tpn

5.10 9.40 8.65 10.2 15.9 8.65 12.9 13.4

TOF

650

16

(a) MW Bi

14

VD Pt

MW Sb

600

12 10 8

550

(b)

VD Bi

6

TOF (min-1)

Amorph. carbon

Tpn (°C)

Sample designation

4

500 OD Bi

2 0

450 0

5

10

15

20

25

30

35

40

Particle size (nm) Fig. 9. Influence of the particle size of the catalytically active component on the first-derivative-peak temperature of non-graphitic carbon (Tpn) determined by TGA (curve a), and on the TOF determined for the conversion of 2-methyl-1-naphthol at r.t. in MeOH (curve b).

[48,52]; that is, the temperature at which 15% carbon weight is lost (T15) decreases with an increase in surface area. However, the active surface area and not the total surface area is the important parameter controlling the oxidation rate of carbons [53]. Nevertheless, this situation must be expected to change fundamentally when catalytically active material (e.g. Pt) is present. In such a case, the genuine influence on the oxidation behavior is more likely exerted by the type and the structure of the catalyst. In fact, Fig. 9 (curve a) shows that the Tpn of our samples is influenced by the particle size (Table 2), and more highly dispersed Pt oxidizes carbon at lower temperatures. Therefore, apart from the sample composition, the particle size must be regarded as an important factor that contributes to the overruling of the surface area effect observed by Jiang and Charsley for pure carbons [48,52]. 3.3. Oxidation of 2-methyl-1-naphthol When 2-methyl-1-naphthol (1) is oxidatively coupled over Pt/AC, Pt-Bi/AC and Pt-Sb/AC catalysts in MeOH or MeNO2 as the solvent, 3,30 -dimethyl-1,10 -binaphthalenyl-4,40 -diol (2) and 3,30 dimethyl-1,10 -binaphthalenylidene-4,40 -dione (3) are the principal products, accompanied by varying amounts of 2-methyl-1,4naphthoquinone (menadione, 4) as a possible side product (Scheme 1). The binaphthol 2 is regarded as the intermediate to the binaphthone 3, while the formation of menadione is a parallel reaction, competing with the formation of 2 [25]. The reactions are supposed to be irreversible; in each step water is released. The specific outcome depends on the catalyst morphology, the promoter (Bi or Sb), the solvent, and the reaction temperature (Table 4). The TOFs are plotted in Fig. 10. These are based on the total amount (mol) of Pt present on each catalyst. Under comparable reaction conditions this should lead to TOFs reflecting the relative activities of the catalysts [54,55].

56

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

Table 4 Oxidative coupling of 2-methyl-1-naphthol catalyzed by AC-supported 5%Pt, 5%Pt-5%Bi and 5%Pt-5%Sb prepared by different methods.

a b

Entry

Sample designation

Solvent

T (°C)

Conv. (%)

1 2 3 4 5 6 7 8 9 10.1 10.2a 11 12 13 14 15 16b 17 18 19 20

VD Pt VD Bi OD Bi MW Bi VD Sb OD Sb MW Sb VD Pt VD Bi OD Bi OD Bi MW Bi VD Sb OD Sb MW Sb VD Bi OD Bi MW Bi VD Sb OD Sb MW Sb

MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeOH MeNO2 MeNO2 MeNO2 MeNO2 MeNO2 MeNO2

r.t. r.t. r.t. r.t. r.t. r.t. r.t. 60 60 60 60 60 60 60 60 r.t. r.t. r.t. r.t. r.t. r.t.

28.9 61.4 100 86.0 14.7 9.60 75.2 100 100 100 100 100 33.9 12.2 100 89.6 100 86.8 20.5 8.14 69.1

Selectivity (mol%) 2

3

4

67.1 91.5 96.0 75.1 89.1 93.8 85.6 11.4 23.3 – 1.4 5.7 96.8 89.1 20.1 98.3 62.1 98.3 91.2 89.7 99.7

30.6 4.89 – 24.3 – – 12.3 85.6 68.0 83.0 86.4 82.9 – – 71.6 – – – – – –

– – – – – – – 2.90 8.30 16.2 11.7 11.0 – – 8.20 Trace 22.5 – – – –

Recycled catalyst containing 4.70% Pt. Black deposits together with other by-products formed.

15.4 15.7 15.7 15.4

15.7

16

15.7 14.1

13.3

14

13.4

11.6

12

TOF, (min-1)

15.4

10.7 9.7

10 8

6 4 2

5.2

4.5 2.3

3.2 2.0

1.5

1.3

0

Catalyst, solvent Fig. 10. TOFs calculated for the catalytic oxidation of 2-methyl-1-naphthol in different solvents at r.t. or under reflux.

Hydrogen peroxide is used as the oxidizing agent. Ideally, the addition rate of the oxidant should match the rate of its consumption [56], as any excess of H2O2 must be expected to decompose on the catalyst surface [57,58] and may lead to over-oxidation of the catalytically active metal [21]. For each test 3.5 equivalents of the oxidant were added in a dropwise manner over 40 min, which is sufficient to achieve total conversions over the catalyst with the highest Pt dispersion (OD Bi) at r.t. and under reflux. This condition should be met, as the catalyst with the highest Pt surface area must also be expected to exhibit the highest decomposition rate for H2O2. In refluxing MeOH (Table 4, entries 8–14) all catalysts, except VD and OD Sb, convert 100% of 1 into various proportions of 2, 3 and 4. OD Bi is the most active sample; all of the intermediate binaphthol 2 is converted further to the binaphthone 3 (entry 10.1). VD and OD Sb are clearly the least active samples; conversions are 33.9 and 12.2%, respectively (entries 12 and 13). In contrast, MW Sb (entry 14) shows a reaction pattern similar to that of VD and MW Bi. Reactions carried out at r.t., that is under the condition of partial conversion (entries 1–7), confirm the activity trends seen under

reflux. Complete conversion of 1 is only obtained over the most active OD Bi, which forms 2 with a yield of 96% (entry 3); any further conversion to 3 is absent. At r.t. 2 is the main product over all seven catalysts and menadione (4) is not observed. The catalyst activity increases in the order OD Sb < VD Sb < VD Pt < VD Bi < MW Sb < MW Bi < OD Bi. In nitromethane at r.t. (entries 15–20), the catalyst activity increases in a similar order (OD Sb < VD Sb < MW Sb < MW Bi = VD Bi < OD Bi). As at r.t. in methanol, complete conversion is only observed for OD Bi, and again, OD and VD Sb are the least active samples. Except for OD Bi (entry 16), 2 is the sole product formed. Typically, over OD Bi high yields of menadione (4) can be obtained: 16.2% in MeOH at 60 °C (entry 10.1), and 22.5% in MeNO2 at r.t. (entry 16). With respect to the catalytic activity, the most relevant findings are as follows: the sample with the lowest APS (OD Bi, 3.1 nm) displays consistently high TOFs (Fig. 10), and the conversion of the SM over that catalyst is always complete (Table 4). OD Bi shows also the lowest temperature (Tpn) for the autocatalytic oxidation of the AC support (Table 3). As seen in Fig. 9, Tpn increases with the

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

particle size (curve a), while the TOFs determined for the oxidation of 1, decrease with the particle size (curve b). Composites with a chain-like morphology (VD and OD Sb) always exhibit inferior catalytic activities (TOF = 1.3–5.2 min–1, Fig. 10). These materials are low in exposed Pt. As the reaction is thought to take place on the surface of metals with high standard electrode potentials [24], sufficient metal exposure/dispersion is a key performance parameter. Therefore, the inferior catalytic activity of the VD and OD Sb composites is in agreement with the metallic-state reaction mechanism proposed earlier [24]. Entry 10.2 of Table 4 demonstrates catalyst recyclability by example of the most active OD Bi catalyst. Due to a small loss of Pt during the first run (entry 10.1) the TOF increases from 15.7 to 16.4 min1. In a control test no activity of leached metal was found [25]. The impact of the reaction conditions (temperature, substrate concentration, peroxide/2-methyl-naphthol ratio, and amount of catalyst used) on product selectivity, in particular on the formation of menadione, was described in a previous paper [25]. There, a detailed account of solvent effects (using the OD Bi catalyst) can also be found.

4. Conclusions

Acknowledgements M. V. Maphoru is grateful for financial support by a DAAD-NRF bursary. The authors thank Dr. Sabine Verryn (XRD Analytical and Consulting, Pretoria) for help with the interpretation of X-ray diffraction data, and Ms. Charity Maepa and Ms. Rirandzu Rikhotso (CSIR, Pretoria) for SEM and TEM characterization. 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.jcat.2017.02.003. These data include MOL files and InChiKeys of the most important compounds described in this article. References [1] [2] [3] [4] [5] [6] [7] [8]

The structure of the composites depends on the preparation method and the promoter used (Bi or Sb). On all samples platinum is present as fcc Pt(0). The addition of Bi or Sb significantly increases the dispersion of Pt. The lowest APS (3.09 nm) is found for OD Bi. On the Pt-Sb samples Sb is always associated with Pt, and Sb is not detectable by XRD. However, a PtSb alloy is seen on MW Sb. VD and OD Sb form Pt- and Sb-containing nanochains (HRTEM), which have little exposed Pt. The chosen AC contains residual phosphate, and on all the composites containing 5%Pt-5%Bi, bismuth phosphate is formed. The identified BiPO4 phase depends on the preparation method, and is different for VD, OD and MW. There is no indication for the formation of SbPO4 on Sb containing samples. The MW method is unique in that it creates crystalline PtBi and PtSb alloys as part of the phase matrix. In addition, amorphous carbon (that would burn in a TGA at low temperature) is not detected. Furthermore, the MW samples distinguish themselves by relatively high remaining total surface areas (855–860 m2/g). In contrast, oven-drying (OD) promotes the anchoring of small particles onto the surface or into the pores of the support, resulting in a reduced total surface area. The thermal stability of the composites at high temperature and the TOFs for the oxidation of 2-methyl-1-naphthol are strongly influenced by the metal dispersion. That is, the temperature for the autocatalytic oxidation of non-graphitic carbon (Tpn) increases with the particle size, while the TOFs decrease with the particle size. OD Bi (APS = 3.09 nm) shows the lowest and VD Pt (APS = 36.5 nm) the highest Tpn. Similarly, OD Bi displays a high TOF (15.7 min1) for the oxidation of 1, and the conversion of 1 over that catalyst is always complete. In contrast, the TOF measured for VD Pt is only 4.5 min1 (in MeOH at r.t.). Most interestingly, composites with a chain-like morphology (VD and OD Sb) exhibit the poorest catalytic activities of all samples (at r.t. TOF = 1.5–2.3 min1). These materials are low in exposed Pt. As the reaction is understood to take place on the surface of metals with high standard electrode potentials, sufficient metal exposure (dispersion) is a key performance parameter. Therefore, the inferior catalytic activity of VD Pt, VD Sb and OD Sb composites is in agreement with a reaction mechanism that proceeds on the Pt surface. The most active catalyst (OD Bi) was recycled, and the TOF of the reused catalyst remained close to 16 min1.

57

[9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30]

[31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47]

https://goldbook.iupac.org/index-alpha.html (accessed 2017-01-14). F. Rodríguez-Reinoso, Carbon 36 (1998) 159. E. Auer, A. Freund, J. Pietsch, T. Tacke, Appl. Catal. A 173 (1998) 259. S. Zhang, R. Fu, D. Wu, W. Xu, Q. Ye, Z. Chen, Carbon 42 (2004) 3209. E. Antolini, Appl. Catal. B 88 (2009) 1. X. Yu, P.G. Pickup, Electrochim. Acta 55 (2010) 7354. J.M. Campelo, D. Luna, R. Luque, J.M. Marinas, A.A. Romero, ChemSusChem 2 (2009) 18. R.S. Amin, A.A. Elzatahry, K.M. El-Khatib, M. Elsayed Youssef, Int. J. Electrochem. Sci. 6 (2011) 4572. I.N. Francesco, F. Fontaine-Vive, S. Antoniotti, ChemCatChem 6 (2014) 2784. Z. Wu, E. Borretto, J. Medlock, W. Bonrath, G. Cravotto, ChemCatChem 6 (2014) 2762. W. Chen, J. Zhao, J.Y. Lee, Z. Liu, Mater. Chem. Phys. 91 (2005) 124. F. Ko, C. Lee, C. Ko, T. Chu, Carbon 43 (2005) 727. H. Li, S. Zhang, S. Yan, Y. Lin, Y. Ren, Int. J. Electrochem. Sci. 8 (2013) 2996. H. Kawasaki, Nanotechnol. Rev. 2 (2013) 5. A. Dailly, R. Schneider, D. Billaud, Y. Fort, P. Willmann, Electrochim. Acta 47 (2002) 4207. X. Ji, K.T. Lee, R. Holden, L. Zhang, J. Zhang, G.A. Botton, M. Couillard, L.F. Nazar, Nat. Chem. 2 (2010) 286. X. Yu, P.G. Pickup, J. Power Sources 196 (2011) 7951. C. Pan, Y. Li, Y. Ma, X. Zhao, Q. Zhang, J. Power Sources 196 (2011) 6228. S.J.A. Moniz, D. Bhachu, C.S. Blackman, A.J. Cross, S. Elouali, D. Pugh, R.Q. Cabrera, S. Vallejos, Inorg. Chim. Acta 380 (2012) 328. T. Mallat, Z. Bodnar, P. Hug, A. Baiker, J. Catal. 153 (1995) 131. T. Mallat, A. Baiker, Chem. Rev. 104 (2004) 3037. J.L. Haan, K.M. Stafford, R.I. Masel, J. Phys. Chem. C 114 (2010) 11665. F.J. Vidal-Iglesias, A. López-Cudero, J. Solla-Gullón, J.M. Feliu, Angew. Chem. Int. Ed. 52 (2013) 964. M.V. Maphoru, J. Heveling, S.K. Pillai, ChemPlusChem 79 (2014) 99. M.V. Maphoru, J. Heveling, S. Kesavan Pillai, Eur. J. Org. Chem. 331 (2016). S. Langa, B.C. Nyamunda, J. Heveling, Catal. Lett. 146 (2016) 755. J. Hassan, M. Sevignon, C. Gozzi, E. Schulz, M. Lemaire, Chem. Rev. (2002) 1359. T. Takeya, T. Otsuka, I. Okamoto, E. Kotani, Tetrahedron 60 (2004) 10681. K.I. Matveev, V.F. Odyakov, E.G. Zhizhina, J. Mol. Catal. A 114 (1996) 151. O.V. Zalomaeva, N.N. Trukhan, I.D. Ivanchikova, A.A. Panchenko, E. Roduner, E. P. Talsi, A.B. Sorokin, V.A. Rogov, O.A. Kholdeeva, J. Mol. Catal. A 277 (2007) 185. M. Selvaraj, M. Kandaswamy, D.W. Park, C.S. Ha, Catal. Today 158 (2010) 377. M. Selvaraj, D.-W. Park, I. Kim, S. Kawi, C.S. Ha, Dalton Trans. 41 (2012) 9633. F. Weber, A. Rüttimann, Ullmann’s Encyclopedia of Industrial Chemistry, vol. 38, Wiley-VCH, Weinheim, 2003, p. 155. O.A. Kholdeeva, O.V. Zalomaeva, A.N. Shmakov, M.S. Melgunov, A.B. Sorokin, J. Catal. 236 (2005) 62. G. Strukul, F. Somma, N. Ballarini, F. Cavani, A. Frattini, S. Guidetti, D. Morselli, Appl. Catal. A 356 (2009) 162. A. Kral, H. Laatsch, Z. Naturforsch. B 48 (1993) 1401. Y. Zhu, S. Murali, M.D. Stoller, A. Velamakanni, R.D. Piner, R.S. Ruoff, Carbon 48 (2010) 2118. N.J. Welham, J.S. Williams, Carbon 36 (1998) 1309. R. Hussain, R. Qadeer, M. Ahmad, M. Saleem, Turk. J. Chem. 24 (2000) 177. L.J. Malobela, J. Heveling, W.G. Augustyn, L.M. Cele, Ind. Eng. Chem. Res. 53 (2014) 13910. R.C.L. Mooney-Slater, Z. Kristallogr. 117 (1962) 371. H. Chen, Q. Tang, Y. Chen, C. Yan, Z. Guo, X. Jia, Y. Yang, Catal. Sci. Technol. 3 (2013) 328. W. Brockner, L.P. Hoyer, Spectrochim. Acta A 58 (2002) 1911. J. Geng, W.-H. Hou, Y.-N. Lv, J.-J. Zhu, H.-Y. Chen, Inorg. Chem. 44 (2005) 8503. Z. Li, S. Yang, J. Zhou, D. Li, X. Zhou, C. Ge, Y. Fang, Chem. Eng. J. 241 (2014) 344. S. Arepalli, P. Nikolaev, O. Gorelik, V.G. Hadjiev, W. Holmes, B. Files, L. Yowell, Carbon 42 (2004) 1783. R.B. Mathur, S. Seth, C. Lal, R. Rao, B.P. Singh, T.L. Dhami, A.M. Rao, Carbon 45 (2007) 132.

58

M.V. Maphoru et al. / Journal of Catalysis 348 (2017) 47–58

[48] W. Jiang, G. Nadeau, K. Zaghib, K. Kinoshita, Thermochim. Acta 351 (2000) 85. [49] J.E. Mudd, T.J. Gardner, A.G. Sault, Sand Report, Sandia National Laboratories, USA, SAND2001-3118, 2001. [50] K.B. Hong, A.A.B. Ismail, M.E.B.M. Mahayuddin, A.R. Mohamed, S.H.S. Zeiri, J. Nat. Gas Chem. 15 (2006) 266. [51] K. Chajara, C.-H. Andersson, J. Lu, E. Widenkvist, H. Grennberg, New J. Chem. 34 (2010) 2275. [52] E.L. Charsley, J.G. Dunn, Rubber Chem. Technol. 55 (1982) 382.

[53] [54] [55] [56]

N. Laine, F. Vastola, P. Walker, J. Phys. Chem. 67 (1963) 2030. S. Kozuch, J.M.L. Martin, ACS Catal. 2 (2012) 2787. G. Lente, ACS Catal. 3 (2013) 381. R. Anderson, K. Griffin, P. Johnston, P.L. Alsters, Adv. Synth. Catal. 345 (2003) 517. [57] L. Kreja, Z. Phys, Chem. (Munich) 140 (1984) 247. [58] M.A. Hasnat, M.M. Rahman, S.M. Borhanuddin, A. Siddiqua, N.M. Bahadur, M.R. Karim, Catal. Commun. 12 (2010) 286.