Gold nanoparticles supported on carbon materials for cyclohexane oxidation with hydrogen peroxide

Applied Catalysis A: General 467 (2013) 279–290

Contents lists available at ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Gold nanoparticles supported on carbon materials for cyclohexane oxidation with hydrogen peroxide S.A.C. Carabineiro a,∗ , L.M.D.R.S. Martins b,c,∗∗ , M. Avalos-Borja d,1 , J.G. Buijnsters e , A.J.L. Pombeiro c , J.L. Figueiredo a a LCM – Laboratory of Catalysis and Materials, Associate Laboratory LSRE/LCM, Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal b Chemical Engineering Department, ISEL, Rua Conselheiro Emídio Navarro, 1959-007 Lisboa, Portugal c Centro de Química Estrutural, IST, Technical University of Lisbon, Av. Rovisco Pais, 1049-001 Lisboa, Portugal d Centro de Nanociencias y Nanotecnología, Universidad Nacional Autónoma de México, A. Postal 2681, Ensenada, Baja California 22800, Mexico e Department of Metallurgy and Materials Engineering, KU Leuven, Kasteelpark Arenberg 44, B-3001 Leuven, Belgium

a r t i c l e

i n f o

Article history: Received 30 April 2013 Received in revised form 12 July 2013 Accepted 19 July 2013 Available online xxx Keywords: Gold nanoparticles Carbon materials Cyclohexane oxidation Catalysis

a b s t r a c t Gold (1 wt.%) was loaded on several types of carbon materials (activated carbon, polymer based carbon xerogels, multi-walled carbon nanotubes, nanodiamonds, microdiamonds, graphite and silicon carbide) using two different methods (sol immobilisation and double impregnation). Samples were characterised by N2 adsorption at −196 ◦ C, temperature programmed desorption, high-resolution transmission electron microscopy, selected area electron diffraction, energy dispersive X-ray spectrometry, high-angle annular dark-field imaging (Z-contrast), X-ray photoelectron spectroscopy and atomic absorption spectroscopy. The obtained Au/carbon materials were used as catalysts for the oxidation of cyclohexane to cyclohexanol and cyclohexanone, with aqueous H2 O2 , under mild conditions. The most active catalyst was prepared by supporting gold nanoparticles on carbon nanotubes by the sol method, achieving an overall turnover number of ca. 171 and an overall yield of 3.6% after 6 h reaction time. These values are comparable to the industrial process (that uses Co catalysts and high temperature), but were obtained at ambient temperature with considerable low loads of catalyst (Au catalyst to substrate molar ratio always lower than 1 × 10−3 ), which is of relevance for establishing a greener catalytic process for cyclohexane oxidation. Moreover, a very high selectivity towards the formation of cyclohexanol and cyclohexanone was achieved, since no traces of by-products were detected. The promoting effect of pyrazine carboxylic acid was observed and an optimum peroxide-to-catalyst molar ratio was found to be 2 × 104 . Further increase of the oxidant amount results in decreased yield due to overoxidation reactions at higher H2 O2 amounts. Catalyst recycling was tested up to six consecutive cycles for the most active catalytic system (gold deposited on carbon nanotubes by sol immobilisation), and it was found that the catalyst maintains almost the original level of activity after several reaction cycles (there was only a 6% drop in activity after the sixth cycle) with a rather high selectivity to cyclohexanol and cyclohexanone and with no catalyst leaching. The differences in activity for the other samples can be explained in terms of gold nanoparticle size and the textural properties of the carbon support. © 2013 Elsevier B.V. All rights reserved.

1. Introduction

∗ Corresponding author. Tel.: +351 225081582; fax: +351 225081449. ∗∗ Corresponding author at: Chemical Engineering Department, ISEL, Rua Conselheiro Emídio Navarro, 1959-007 Lisboa, Portugal. E-mail addresses: [email protected] (S.A.C. Carabineiro), [email protected] (L.M.D.R.S. Martins). 1 On leave at Instituto Potosino de Investigación Científica y Tecnológica (IPICyT), Division de Materiales Avanzados, San Luis Potosi, S.L.P., Mexico. 0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apcata.2013.07.035

Hydrocarbons, particularly alkanes, are interesting compounds as main constituents of natural oil and gas and their C H bond(s) can be converted to C OH or C O functional groups leading to the production of more valuable products for fine chemical synthesis. However, activation of the former bonds in such stable compounds is difficult, which still prevents their generalised use in the direct synthesis of added value chemical products [1–3]. An example with high industrial significance concerns the oxidation of cyclohexane to cyclohexanol and cyclohexanone (Scheme 1) that are important reagents for the production of adipic

280

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

Scheme 1. Oxidation of cyclohexane to cyclohexanol and cyclohexanone.

acid and caprolactam, used for the manufacture of nylon. The industrial process uses a homogeneous cobalt species as catalyst and dioxygen as oxidant at a considerably high temperature (150 ◦ C). However, the oxidation products are formed in low yields (5–12%) to achieve a good selectivity (ca. 80–85%) [1,2], and thus the need for more effective systems under milder reaction conditions has been recognised [1,2,4–6]. Gold catalysts are currently a “hot topic” of research, as they show application in many reactions of industrial and environmental importance [7–11]. Several variables have been considered as important factors influencing the chemistry, structure and catalytic activity. Among them are the method of preparation, the nature of the support and, particularly, the gold nanoparticle size [7–11]. Gold catalysts have been used successfully for cyclohexane oxidation. Some common supports include metal oxides [12–15], metalorganic frameworks [16], mesoporous [13,17–20] and other silica [21–24] based materials, modified aluminas [25,26] and hydroxyapatite [27]. It has been controversially debated if gold catalysts do act as catalysts or as promotors of the autoxidation reaction [28]. Some authors support that gold acts as a real catalyst for this reaction, since a conversion of 10% with selectivity of 70% could be achieved with Au on TiO2 doped SiO2 , but not with the Au-free supports [21]. However, Hereijgers and Weckhuysen studied the reaction over Au/Al2 O3 , Au/TiO2 and Au/SBA-15, and concluded that it proceeds via a pure radical pathway with products typical of autoxidation, the reaction being fully inhibited in the presence of radical scavengers [13]. On the other hand, Liu et al. showed that gold clusters on hydroxyapatite were capable of displaying high activity and that no reaction was taking place in the absence of gold, although the reaction required the presence of radical initiators, typically tert-butylhydroperoxide [27]. Hutchings and co-workers, by studying the reaction over Au/MgO type catalysts, suggested that an intermediate situation should be considered, being gold capable of accelerating the reaction without the need for initiators (thus being a catalyst by definition) [15]. However, the acceleration occurred by increasing the concentration of species (through C6 H11 OOH or C5 H11 OO• ) which are chain carriers in the radical pathway of the reaction and therefore promote catalytic autoxidation processes via a radical chain mechanism [15]. In general, Au/oxide studies were performed at high temperature (above 100 ◦ C) at oxygen pressures ranging from 0.3 to 3 MPa, with no solvent. Comparatively less work has been done using carbon materials as supports [29,30]. With Au/graphite, a limited activity at 70 ◦ C and 0.3 MPa was reported by Hutchings and co-workers, with very high selectivities to cyclohexanol and cyclohexanone after 17 h, but only for very low conversions [29]. Aiming to contribute towards the development of the still little explored Au/carbon combination for the oxidation of cyclohexane, we decided to test several carbon materials as supports: activated carbon, carbon xerogels (two samples, with different textural properties), carbon nanotubes, microand nanodiamonds (the latter in two forms: powder and liquid dispersion), silicon carbide and graphite. Activated carbons are highly porous materials with very rich surface chemistry [31–34], presenting unique advantages due to their low cost, high adsorption capacity and easy disposal. Carbon xerogels are mesoporous polymer based carbon materials, whose

textural properties can easily be tuned by changing the parameters of the preparation procedure [35–37]. Graphite is an allotrope of carbon that has electrical conductor and semimetal properties. Carbon nanotubes are nanofilamentous carbon tubes, with a hollow cavity with one or several concentric graphitic layers [38,39]. Silicon carbide is an inert, semiconductor, abrasive material. Nanodiamonds (NDs) are monocrystalline diamond particulates originated from a detonation process [40,41]. As a result, NDs exhibit a narrow particle size distribution and a small particle size (<10 nm) and they have a highly developed chemically active surface [40]. On the other hand, microdiamonds (MDs) are small diamonds (1–2.5 ␮m) extracted from rock samples by drill core or outcrop and that are widely used in grinding and abrasive technology. NDs have smaller particle size (typically 4–5 nm) than MDs and are rapidly becoming one of the most widely studied nanomaterials, since diamonds on the nanoscale have a higher surface area than MDs (BET surface areas around 300 m2 /g for NDs, in contrast to 5 m2 /g for MDs) which helps to create a larger amount of reactive chemical surface groups [42,43]. With the development of new environmentally friendly purification techniques, NDs are nowadays produced in large volumes at a low cost, which has stimulated increased interest on the subject material. There are some examples in literature of gold nanoparticles supported on activated carbon [44–48] and graphite [46,49] used for different oxidation reactions with very good results. Carbon xerogels [50,51] and carbon nanotubes [39,52] were less explored for this purpose. The use of gold/nanodiamond supports in catalysis is still scarce [53–55] as well as their use in oxidation [53,56]. To the best of our knowledge, neither MDs nor silicon carbide were reported for that purpose. For the cyclohexane oxidation, in particular to what concerns carbon materials, only graphite has been used as a support so far [29,30]. In our work, gold was loaded by two different methods (double impregnation and sol immobilisation) on the different above mentioned nine carbon supports. The obtained materials were tested for the oxidation of cyclohexane to cyclohexanol and cyclohexanone, under mild conditions (room temperature and atmospheric pressure), using an environmentally friendly oxidant (H2 O2 ). 2. Experimental 2.1. Carbon supports Different types of carbon materials were used in this work (as shown in Table 1): (1) Commercial activated carbon Norit ROX 0.8 (sample AC), which is an extruded acid washed activated carbon with cylindrical pellets of 0.8 mm diameter and 5 mm length, prepared from peat by steam activation, crushed into powder; (2) two carbon xerogels, with different textural properties, were synthesised by the polycondensation of resorcinol and formaldehyde, as described in earlier publications [36,37,57]. Sample CX refers to a xerogel with smaller mesopore width, prepared at pH = 6, while CXL is a sample with larger width, prepared at pH = 5.5; (3) commercial Nanocyl-3100 multi-walled carbon nanotubes were supplied by Nanocyl, Belgium (sample CNT). This material has an average diameter of 9.5 nm, an average length of 1.5 mm (with average inner diameter of 4 nm) and carbon purity higher than 95%. Further details on this material can be found elsewhere [58]; (4) two types of nanodiamonds produced by detonating carboncontaining explosives in a closed chamber and immediately cooled at a rate ≥3000 K min−1 , as described elsewhere

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

281

Table 1 Description and characterisation of (powder) carbon samples: surface area (SBET ), total pore volume (Vp ), average mesopore width (L), micropore volume (Vmicro ), external area (Sexternal ), obtained by adsorption of N2 at −196 ◦ C, and amounts of CO and CO2 desorbed, as determined by TPD. Sample name a

AC CNTa NDPW MD CX CXL GR SC a

Sample material

SBET (m2 /g)

Vp (cm3 /g)

L (nm)

Vmicro (cm3 /g)

Sexternal (m2 /g)

CO (␮mol/g)

CO2 (␮mol/g)

Activated carbon Carbon nanotubes Nano-diamonds (powder) Micro-diamonds Xerogel (pH = 6) Xerogel (pH = 5.5) Graphite Silicon carbide

974 257 295 4 611 614 5 ∼0

0.67 2.89 1.08 0.01 0.90 1.29 0.02 ∼0

– – 10.6 3.8 13.6 32.3 4.5 –

0.348 ∼0 ∼0 ∼0 ∼0 0.012 ∼0 ∼0

260 257 295 4 611 538 5 ∼0

740 194 780 51 564 379 40 –

205 70 260 25 156 125 21 –

Data from [57]

[41], namely dry powder of nano-sized diamonds (sample NDPW) (<10 nm) from Sigma Aldrich, and an aqueous monodispersion (4–5 nm) of 5% nanodiamonds (sample NDLIQ) from NanoCarbon Research Institute Co. Limited, Japan, known as NanoAmando®; (5) micro-sized diamond powder (1–2.5 ␮m) from Technodiamant, The Netherlands (sample MD), obtained by extraction from rock samples by drill core or outcrop of about 25–100 kg, which were crushed and dissolved in acid or a hot caustic solution; (6) graphite powder (<20 ␮m, synthetic), supplied by Sigma Aldrich (sample GR); silicon carbide (SiC, Carborundum < 0.2 mm) from VWR (sample SC). 2.2. Preparation of gold catalysts Gold (1 wt.%) was loaded on the carbon supports by two different methods: sol immobilisation or colloidal method (COL) [47,50,52,59] and double impregnation (DIM) [60–66]. The first procedure consists in dissolving the gold precursor, HAuCl4 ·3H2 O (Alfa Aesar), in water, adding polyvinyl alcohol (Aldrich) and NaBH4 (Aldrich), resulting in a ruby red sol to which the powder support is added under stirring (in the case of NDLIQ, the ND suspension was added to the solution). After a few days the solution starts to lose colour, as Au is deposited on the support. The colourless solution is filtered, the catalyst washed thoroughly with distilled water until the filtrate is free of chloride and dried at 110 ◦ C overnight. The organic scaffold is removed from the support by a heat treatment under nitrogen flow for 3 h at 350 ◦ C (shown by elemental analysis to be efficient for this purpose), and then, the catalyst is activated by further treatment under hydrogen flow for 3 h also at 350 ◦ C. In previous studies, colloidal procedures have been used with good success for the deposition of Au on carbon materials [47,50,52,59,67]. The DIM method consists in impregnating the support with an aqueous solution of the gold precursor and then with a solution of Na2 CO3 (Sigma Aldrich), followed by washing with water and drying. This step removes chloride, which is well known to cause sintering of Au particles, thus turning them inactive [9,10,68]. 2.3. Characterisation of carbon supports 2.3.1. Textural characterisation The materials were analysed by adsorption of N2 at −196 ◦ C in a Quantachrome NOVA 4200e apparatus. In a typical experiment, around 100 mg of sample was used and degassing was carried out for 3 h at 160 ◦ C. The specific surface area (SBET ) was calculated by the Brunauer–Emmett–Teller (BET) equation [69], the total pore volume (Vp ) was determined at P/Po = 0.98, the average mesopore width (L) by the Barrett–Joyner–Halenda (BJH) method [70], and the micropore volume (Vmicro ) and external area (Sexternal )

were determined by the t-method using an appropriate standard isotherm [71]. 2.3.2. Surface characterisation Temperature programmed desorption (TPD) experiments were performed in a fully automated AMI-200 Catalyst Characterisation Instrument (Altamira Instruments), equipped with a quadrupole mass spectrometer (Dymaxion 200 amu, Ametek). In a typical experiment, around 100 mg of sample was placed in a U-shaped quartz tube located inside an electrical furnace and subjected to a 5 ◦ C/min heating rate up to 1100 ◦ C, under helium flow of 30 cm3 /min. Desorbed CO and CO2 were monitored by mass spectrometry. 2.4. Characterisation of gold on carbon materials 2.4.1. Nanoparticle sizes Selected samples were imaged by scanning electron microscopy (SEM), transmission electron microscopy (TEM), high resolution transmission electron microscopy (HTREM) and high-angle annular dark-field (HAADF) with energy-dispersive X-ray spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS). SEM was performed on a FEI Quanta 200 apparatus. Conventional TEM measurements were performed with a JEOL 2010 microscope. HRTEM, EDS and Z-contrast measurements were done on a FEI Tecnai F30 instrument. The Z-contrast images were collected using an HAADF detector in scanning transmission mode (STEM). The presence of gold was confirmed by EDS. The samples were mounted on a carbon polymer supported copper micro-grid. A few droplets of a suspension of the ground catalyst in isopropyl alcohol were placed on the grid, followed by drying under ambient conditions. The average gold particle size was determined from measurements made on about 300–500 particles, depending on the sample. The metal dispersion was calculated by DM = (6ns M)/(Ndp ), where ns is the number of atoms at the surface per unit area (1.15 × 1019 m−2 for Au), M is the molar mass of gold (196.97 g mol−1 ),  is the density of gold (19.5 g cm−3 ), N is Avogadro’s number (6.023 × 1023 mol−1 ) and dp is the average particle size (determined by HRTEM, assuming that particles are spherical). 2.4.2. Oxidation state XPS analyses of the Au 4f region were carried out for selected samples in order to confirm the expected reduced state of gold. A VG Scientific ESCALAB 200A spectrometer was used, with Al K␣ radiation (1486.6 eV). 2.4.3. Metal loading Samples were incinerated at 600 ◦ C and the resulting ashes were dissolved in a concentrated HNO3 and H2 SO4 mixture. The resulting solution was diluted and analysed by atomic absorption spectroscopy (AAS) using a Unicam 939 atomic absorption spectrometer and a hollow cathode lamp Heraeus 3UNX Au.

282

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

2.5. Catalytic tests Cyclohexane (Merck), acetonitrile (Riedel-de-Haën), hydrogen peroxide (30%) (Fluka), nitric acid (65%) (Riedel-de-Haën), pyrazinecarboxylic acid (Aldrich), cyclohexanol (Aldrich), cyclohexanone (Aldrich), cycloheptanone (Riedel-de-Haën), diethyl ether (Riedel-de-Haën), triphenylphosphine (Merck), diphenylamine (Fluka), bromotrichloromethane (Fluka) and nitromethane (Aldrich) were used as received from the supplier. The oxidation reactions were carried out in Schlenk tubes, initially under molecular nitrogen. In typical conditions, the reaction mixtures were prepared as follows: 0.1–9.9 ␮mol of Au deposited on each carbon material (2–200 mg) was added to acetonitrile, (3.00 mL). 5.00 × 10−3 mol of cyclohexane (0.54 mL) and 10.0 × 10−3 mol of H2 O2 in a 30% H2 O solution (1.02 mL) were then added (in this order) and the reaction mixture was stirred typically for 6 h, at room temperature and normal pressure. It was found that Au/CX and Au/CXL were the materials that better dispersed on the reaction media, followed by Au/GR, Au/NDLIQ and Au/MD. The most difficult to disperse was Au/AC, followed by Au/CNT, Au/NDPW and Au/SC. In the experiments with HNO3 or pyrazinecarboxylic acid, the acid was added immediately before the addition of the substrate. In the experiments with radical traps, CBrCl3 (5.00 mmol) or NHPh2 (5.00 mmol) was added to the reaction mixture. For the products analysis, 90 ␮L of cycloheptanone (internal standard) and 10.0 mL of diethyl ether (to extract the substrate and the organic products from the reaction mixture) were added. The obtained mixture was stirred during 10 min and then a sample (1 ␮L) was taken from the organic phase and analysed by gas chromatography (GC) by the internal standard method. Subsequently, an excess of solid triphenylphosphine was added to the final organic phase (to reduce the cyclohexyl hydroperoxide, if formed, to the corresponding alcohol, and hydrogen peroxide to water) and the mixture was analysed again to estimate the amount of cyclohexyl hydroperoxide, following a method developed by Shul’pin [72–75]. In our studies for determination of oxygenate concentrations only data obtained after treatment of the reaction sample with PPh3 were used. Authentic samples of all oxygenated products were used to attribute the peaks in chromatograms. Blank experiments confirm that no cyclohexanol or cyclohexanone are formed in the absence of a metal catalyst or by using only carbon materials. GC measurements were carried out using a FISONS Instruments GC 8000 series gas chromatograph with a FID detector and a capillary column (DB-WAX, column length: 30 m; internal diameter: 0.32 mm) and the Jasco-Borwin v.1.50 software. The temperature of injection was 240 ◦ C. The initial temperature was maintained at 100 ◦ C for 1 min, then raised 10 ◦ C/min to 180 ◦ C and held at this temperature for 1 min. Helium was used as the carrier gas. GC–MS analyses were performed using a Perkin Elmer Clarus 600C instrument. The ionisation voltage was 70 eV. GC was conducted in the temperature-programming mode, using a SGE BPX5 column (30 m × 0.25 mm × 0.25 ␮m). Reaction products were identified by comparison of their retention times with known reference compounds, and by comparing their mass spectra to fragmentation patterns obtained from the NIST spectral library stored in the computer software of the mass spectrometer.

3. Results and discussion 3.1. Characterisation of the carbon supports 3.1.1. Textural characterisation Table 1 shows the textural characterisation results of the different carbon samples. AC shows the highest BET surface area among

Fig. 1. TPD spectra of the carbon supports: CO (top) and CO2 (bottom) evolution.

the starting materials. NDPW has a BET surface area of 295 m2 /g, while for MDs a value of 4 m2 /g is derived, the large difference already being expected from literature [42,43,76] and from the large difference in average size of the diamond particulate supports (as seen ahead). GR shows a similarly low surface area of 5 m2 /g. The xerogels (CX and CXL) exhibit mesoporous properties with large pore sizes, just like the NDPW sample. These two samples, although having similar values of surface area, have different pore sizes. CXL was prepared at pH = 5.5 and CX was prepared at pH = 6. As expected from results reported in the literature [37], the former has a larger average mesopore size (32.3 nm) than the latter (13.6 nm). CNTs have a cylindrical structure and the pores result from the free space in the bundles. On the other hand, SC is not porous at all. NDLIQ was not characterised by these techniques, since it was in a liquid dispersion, which makes it unsuitable for this kind of analysis. 3.1.2. Surface characterisation Fig. 1 shows the TPD spectra of the different carbon materials. The total amounts of CO and CO2 released, determined by curve integration, are presented in Table 1. NDPW has the largest amount of surface groups, as it is known that diamond nanoparticulates have a highly developed chemically active surface [40]. Activated carbon follows second and it is well known that this material has a very rich surface chemistry as well [31–34]. The amounts of CO and CO2 released from the xerogels are much lower than those from AC, and CNT shows even less groups, as we also reported previously for similar samples [77]. CXL released less amounts of CO and CO2 than CX, as expected from results found in literature for xerogel samples with different pore sizes [37]. GR and MD have the lowest amount of surface groups of the set of samples studied. Indeed, for a low surface area graphite, low amounts of desorbed CO and CO2 species are also expected, as well as that MD releases far less amounts than AC [76]. Finally, the amounts released from SC were negligible. Being nano- and microdiamonds the less common kind of supports used in this study, they were imaged by electron microscopy. An SEM image of MDs is displayed in Fig. 2a. Little “rocks” of 1–2.5 ␮m can be clearly observed with well-defined and sharp

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

283

Fig. 2. SEM images of MD (a) and NDPW (b) samples. Bright field (c) and dark field (d) TEM images of NDPW.

edges. The respective EDS spectrum (not shown) confirmed the carbon nature of this material (the largest constituent being the carbon K transition). Fig. 2b shows an SEM image of the NDPW sample. It can be seen that the nanoparticles are strongly agglomerated. A regular TEM (bright field) image of NDPW is shown in Fig. 2c, whereas Fig. 2d shows a dark-field TEM image of the same portion of sample. This latter technique can be used to demonstrate that particles are crystalline. Nanoparticles with approximate sizes between 2 and 10 nm could be observed.

3.2. Characterisation of gold on carbon materials 3.2.1. Nanoparticle sizes Fig. 3 shows HAADF micrographs and HRTEM images of gold nanoparticles on selected carbon supports, namely CNT, AC, MD, NDPW, NDLIQ, CX and CXL, respectively. The first technique allows detection of the gold particles through the Z-contrast micrographs, as they are seen as bright spots, while on HRTEM images gold nanoparticles are visible as darker areas. The presence of gold was confirmed by EDS. Fig. 4 shows the size distribution histograms

Table 2 Average gold nanoparticle size, HRTEM/HAADF measurements). Sample

Au average size (nm)

Au/CNT-COL Au/CNT-DIM Au/AC-COL Au/NDPW-COL Au/NDLIQ-COL Au/MD-COL Au/CX-COL Au/CXL-COL

5.1 6.5 6.8 7.6 8.2 8.2 4.4 4.4

± ± ± ± ± ± ± ±

1.8 1.6 1.5 1.4 1.4 1.4 1.0 1.0

range

and

dispersion

(calculated

from

Au range (nm)

Metal dispersion (%)

1–15 2–20 1–20 2–18 1–20 1–20 1–14 1–14

23.0 17.7 16.9 15.1 14.0 14.0 26.2 26.2

obtained and Table 2 shows a summary of the values of the gold nanoparticle sizes and dispersion. It can be seen that depositing gold on CNT by different methods (COL and DIM) yields nanoparticles with different sizes, being COL the most efficient method, leading to a larger dispersion and lower average particle size. In another work, gold was deposited on carbon nanotubes following the same experimental procedures, and it was shown that COL was the most efficient method in comparison with DIM [52]. In fact, DIM is a very good method to prepare gold on metal oxides [60,62,63,65] and rarely used on carbon materials [52]. The smallest nanoparticles were obtained on CX materials (4.4 ± 1.0 nm average size). There is no difference in size or range for the gold nanoparticles deposited on the two xerogels, as the results obtained are very similar, suggesting that the different mesopore size of these materials does not influence the gold deposition process or gold distribution. According to other authors [55], a large number of defects in the support structure is beneficial for gold deposition, as they act as anchoring sites, limiting the increase in nanoparticle size that occurs during further agglomeration of gold nanoparticles, before they stick onto the support [55]. Therefore, the defective structure of CX and CXL is expected to provide small gold nanoparticle sizes [55], as found in our work. A larger gold nanoparticle size was obtained on AC (6.8 ± 1.5 nm). This material is also porous, but with pores of smaller size (micropores). Therefore, mesoporosity seems to be more important for the formation of relatively small gold nanoparticle sizes than microporosity. NDPW acted as a better support for gold loading than NDLIQ, as it allowed smaller nanoparticles to be formed (7.6 ± 1.4 nm, against 8.2 ± 1.4 nm). Although the complete mono-dispersed NDLIQ liquid was added to the gold solution during the preparation of the gold catalysts (instead of a powder, as in the case for all other materials), a strong in situ agglomeration of the core ND particles upon gold formation might explain the different sizes. Similar to NDLIQ, an average size of

284

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

Fig. 3. HAADF micrograph of Au/CNT-COL (a). HRTEM images of Au/CNT-DIM (b), Au/AC-COL (c) and Au/MD-COL (d). HAADF micrograph of Au/NDPW-COL (e). HRTEM images of Au/NDLIQ-COL (f) and Au/CX-COL (g). HAADF micrograph of Au/CXL-COL (h). Gold nanoparticles are seen as darker spots on HRTEM images and as bright spots on HAADF micrographs.

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

285

Fig. 4. Size distribution histograms of gold nanoparticles deposited on CNT by COL (a) and DIM (b) methods, and by COL on AC (c), MD (d), NDPW (e), NDLIQ (f), CX (g) and CXL (h) supports.

286

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

8.2 ± 1.4 nm was obtained in the case of the MD support. Note that NDPW might be a more adequate support in comparison with MD also due to its mesoporosity. Apparently, the gold nanoparticle size does not seem to be related with the BET surface area of the supports, for the studied samples. 3.2.2. Oxidation state Since the COL samples are directly reduced with hydrogen, Au 4f XPS was carried out only in the Au/CNT-COL and Au/AC-COL samples for confirmation. As for the DIM method, the preparation method includes a second step of impregnation with Na2 CO3 , which can dissociate and, by reaction with water, generate hydroxyl ions, which hydroxylate the Au3+ species of the gold precursor [52]. Subsequently, Au3+ can be reduced to metallic gold by electron transfer from coordinated OH¯ ions on the surface of the support [78]. Therefore, it was also expected that materials prepared by this technique to be in the reduced state. XPS showed that this was also the case for the Au/CNT-DIM and Au/AC-DIM materials. Other authors also reported gold in the metallic state (on silica based materials [17,19,20,23] and other oxide materials [14]) for cyclohexane oxidation. 3.2.3. Metal loading The gold content obtained by AAS analysis was around ∼0.8 wt.% Au, which is lower than the nominal value (1 wt.% Au). The two preparation methods did not influence substantially the gold content of the catalysts, as the correspondingly prepared samples presented similar values. So, both COL and DIM techniques used in this work were efficient in producing catalysts with gold contents reasonably close to the expected (nominal) loading. 3.3. Catalytic activity The gold nanoparticles supported on carbon materials act as catalysts or catalyst precursors for the peroxidative oxidation of cyclohexane with H2 O2 , at room temperature, to cyclohexanol and cyclohexanone via formation of cyclohexyl hydroperoxide (primary product) according to Scheme 1. The best results were obtained with the Au/CNT samples (Table 3). In particular, Au/CNTCOL is the most active catalyst, achieving overall turnover numbers (TON, moles of product/mole of Au) up to ca. 171 (Table 3, entry 4) and overall yields up to 3.6% (Table 3, entry 5) after 6 h reaction time. The reason for the COL sample to be more active than the DIM material could be related to the smaller particle size found for the former (5.1 nm), in comparison with the latter (6.5 nm), as seen in Table 2, since it is well known that the gold nanoparticle size also plays an important role in this reaction [14,22], as it does in many others [8–10]. The yield of 3.6% obtained for the Au/CNT-COL is very similar to that reported in the literature (3.7%) for 1% Au on graphite by Xu et al. [29]. However, these authors used 17 h reaction time and a temperature of 70 ◦ C, and the total selectivity for cyclohexanol and cyclohexanone obtained was 23.1%. Higher values of selectivity (up to 91.6%) were obtained only for very low conversions (around 1%), applying the same conditions but using a 0.5% Au catalyst with an additive [29]. Therefore, our results of a higher selectivity and a similar yield are more favourable, since we used only 6 h reaction time at room temperature, which are more environmentally friendly and adequate reaction conditions for industrial purposes. AC, that is a microporous material, was the second best performing support tested. The nanoparticle size of Au/AC-COL (6.8 nm) is larger than those of CNT samples, possibly explaining the lower activity of AC in comparison with CNT. Surprisingly, the poorest catalysts found in this study, Au on CX and CXL supports, have the smallest gold nanoparticle sizes. This suggests that particle size alone cannot be accountable for the catalytic activity. Most

likely, mesoporous supports, like xerogels, are not adequate for this reaction, since also relatively poor results were obtained with the nanodiamond materials. Very recently, it was found that well developed mesoporosity of carbon supports was beneficial for the activity of Au/carbon catalysts in hydroamination of alkynes [55]. However, that does not seem to be the case for the peroxidative oxidation of cyclohexane, as evidenced in our work. In fact, it has been reported in the literature that mesoporous materials, like modified silicas, when used as supports for gold, produce catalysts with good selectivities for this reaction (up to 98%), but only for temperatures above 100 ◦ C and pressures higher than 0.6 MPa [17,20,24]. The only report concerning milder conditions (70 ◦ C, atmospheric pressure) deals with doped SiO2 , and the selectivity was reported to be high despite a conversion of only 1.2% [22]. As cyclohexanol and cyclohexanone are commonly much more reactive than cyclohexane, it is difficult to achieve high conversion and high selectivity simultaneously under milder conditions [22]. The mesoporous support systems described in literature achieve conversions up to 34.5%, but selectivities to the desired products decrease down to ∼80% [19]. Much better selectivity values were obtained on Au/MgO (98%), but at the expense of conversion (5%), for 17 h reaction at 140 ◦ C and 0.3 MPa in the case of a 1 wt.% Au loading [15]. Our Au/CNT-COL catalyst had 3.6% yield, but showed no trace of other products besides cyclohexanol and cyclohexanone, and required less reaction time (6 h) at room temperature and atmospheric pressure, which are added advantages. Seral-Ascaso et al. [55] showed that larger surface areas of carbon supports increase the catalytic activity for the hydroamination of alkynes. In our case, we cannot relate the catalytic activity with the surface area, since the second best support (AC) has indeed a high surface area, whereas xerogel materials, that also have relatively large surface areas (although not as high as AC), showed the worst results. The same authors also showed that dispersibility of the supports in water plays an important role during the gold loading process [55]. However, dispersibility in the solvent (toluene) is not mentioned, although the reaction medium is referred to as being a “suspension”. In our work, we found that, in general, dispersibility of the Au/carbon catalyst in the reaction medium (acetonitrile) was inversely related with catalytic activity, since the best dispersed materials (namely CX, CNL, GR, NDLIQ and MD) showed the worse catalytic results. On the other hand, catalysts that had worse dispersibility, like AC and CNT, showed the best results. In all cases, catalysts obtained by the COL method are more active than those formed by the DIM preparation, except for SC, where the opposite is the case. This might be due to the fact that DIM is a very good method to prepare gold on metal oxides [60,62,63,65], and SiC is usually covered by a passivation layer of silicon dioxide formed at its surface. Apparently, the existence of surface groups does not play a determining role in the peroxidative oxidation of cyclohexane. CNT is one of the best supports and has low amounts of oxygenated groups (Fig. 1 and Table 1). On the other hand, NDPW has the largest amount of surface groups and is a relatively poor support. Nevertheless, good results were obtained with AC that has a large amount of surface groups and is a microporous material. The absence of mesoporosity might increase its activity and balance a possibly detrimental effect of the surface groups. Regarding the diamond based materials, NDPW showed a higher activity than MD, which could be explained by the lower gold size (7.6 nm on NDPW and 8.2 nm on MD). However, gold nanoparticles on MD have the same size and range as those on NDLIQ. As referred above, the liquid dispersion of nanodiamonds was not analysed by characterisation techniques that require a powder sample. However, considering that the liquid dispersed nanodiamonds display similar core particle sizes as in dry powder form, again the existence

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

287

Table 3 Peroxidative oxidation of cyclohexane with H2 O2 (selected data).a Entry

Catalyst

n(H2 O2 )/n(catalyst) × 10−4

n(Hpca)/n(catalyst)b

Yield (%)c

TONe

Conv. (%)f

Sel. (%)g

d

-ol

-one

Total

1 2 3 4 5 6 7

Au/CNT-DIM Au/CNT-DIM Au/CNT-COL Au/CNT-COL Au/CNT-COL Au/CNT-COL Au/CNT-COL

2 2 2 2 0.2 1 4

0 50 0 50 50 50 50

1.0 1.4 0.5 1.1 2.0 1.6 1.9

0.2 0.8 0.6 0.8 1.6 1.1 0.4

1.2 2.2 1.1 1.9 3.6 2.7 2.3

14 101 13 171 43 106 126

1.2 (5) 2.3 (9) 1.1 (2) 1.9 (4) 3.6 (4) 2.7 (3) 3.3 (8)

96 92 98 98 99 99 68

8 9 10 11 12 13 14

Au/AC-DIM Au/AC-DIM Au/AC-COL Au/AC-COL Au/AC-COL Au/AC-COL Au/AC-COL

2 2 2 2 0.2 1 4

0 50 0 50 50 50 50

1.3 1.5 1.4 1.8 1.8 1.3 0.4

0.0 0.9 1.3 1.2 0.1 0.4 0.3

1.3 2.4 2.7 3.0 1.9 1.7 0.7

16 54 34 93 23 46 43

1.5 (1) 2.4 (2) 2.9 (0) 3.0 (6) 2.2 (1) 1.7 (7) 1.3 (2)

86 99 93 98 86 96 53

15 16 17

Au/NDPW-DIM Au/NDPW-DIM Au/NDPW-COL

2 2 2

0 50 0

0.3 0.4 0.5

0.5 0.3 0.8

0.8 0.7 1.3

10 19 16

0.8 (8) 0.7 (3) 1.4 (9)

91 95 87

18 19 20 21 22

Au/NDPW-COL Au/NDLIQ-DIM Au/NDLIQ-DIM Au/NDLIQ-COL Au/NDLIQ-COL

2 2 2 2 2

50 0 50 0 50

0.9 0.4 0.4 0.6 0.7

0.7 0.1 0.5 0.3 1.0

1.6 0.5 0.9 0.9 1.7

50 6 12 11 21

1.7 (2) 0.5 (9) 1.0 (2) 1.1 (4) 2.0 (0)

93 84 88 79 85

23 24 25 26

Au/MD-DIM Au/MD-DIM Au/MD-COL Au/MD-COL

2 2 2 2

0 50 0 50

0.3 0.3 0.7 0.8

0.0 0.2 0.5 0.9

0.3 0.5 1.2 1.7

4 6 15 34

0.3 (3) 0.5 (6) 1.2 (2) 1.7 (7)

92 90 98 96

27 28 29 30

Au/SC-DIM Au/SC-DIM Au/SC-COL Au/SC-COL

2 2 2 2

0 50 0 50

0.1 0.0 0.0 0.1

0.2 2.4 0.0 0.0

0.3 2.4 0.0 0.1

15 29 – 4

0.3 (4) 2.5 (3) – 0.1 (1)

88 95 – 93

31 32 33 34

Au/CX-DIM Au/CX-DIM Au/CX-COL Au/CX-COL

2 2 2 2

0 50 0 50

0.0 0.2 0.1 0.4

0.0 0.2 0.0 0.1

0.0 0.4 0.1 0.5

– 5 1 6

1.2 (3) 0.4 (6) 0.1 (3) 0.4 (1)

– 87 76 81

35 36 37 38

Au/CXL-DIM Au/CXL-DIM Au/CXL-COL Au/CXL-COL

2 2 2 2

0 50 0 50

0.0 0.2 0.0 0.2

0.1 0.1 0.0 0.2

0.1 0.3 0.0 0.4

1 4 – 5

0.1 (3) 0.3 (5) 2.1 (9) 0.5 (5)

79 85 – 73

39 40 41 42

Au/GR-DIM Au/GR-DIM Au/GR-COL Au/GR-COL

2 2 2 2

0 50 0 50

0.0 0.4 0.0 0.8

0.3 0.5 0.0 0.6

0.3 0.9 0.0 1.4

4 12 – 17

0.3 (3) 0.9 (3) 3.3 (4) 1.5 (1)

91 97 – 93

a Reaction conditions (unless stated otherwise): acetonitrile (3.0 mL), cyclohexane (5.0 mmol), r.t., under dinitrogen; 0.1 − 9.9 ␮mol of Au catalyst (2 − 200 mg of Au/carbon material), H2 O2 (10 mmol; 1:1 to 4 × 104 :1 molar ratio of oxidant to Au catalyst). Percentage of yield, TON determined by GC analysis (upon treatment with PPh3 ). b Hpca – pyrazine carboxylic acid. c Molar yield (%) based on substrate, i.e. moles of products (cyclohexanol (-ol) and cyclohexanone (-one)) per 100 mol of cyclohexane. The molar yield (%) based on oxidant can be calculated as moles of products (cyclohexanol (-ol) and cyclohexanone (-one)) per 100 mol of hydrogen peroxide. d Moles of cyclohexanol + cyclohexanone per 100 mol of cyclohexane. e Turnover number (moles of product per mol of Au catalyst). f Conversion (Conv.) = Moles of converted (reacted) cyclohexane per mole of cyclohexane determined by GC–MS analysis. g Selectivity (Sel.) = Moles of product per mole of converted cyclohexane.

of (different) surface groups and mesoporosity of nanodiamonds (when in powder form) seem to be detrimental for the catalytic activity. Finally, GR showed a relatively low activity as well. This is likely ascribed to both the low porosity and high dispersibility of this support. The yield values (Table 3) are comparable with those of the industrial process [1,2] and were obtained at ambient temperature, atmospheric pressure and with considerable low loads of catalyst (Au catalyst to substrate molar ratio always lower than 1 × 10−3 ). These features are of utmost importance for the establishment of

a greener catalytic process for cyclohexane oxidation. Moreover, a high selectivity towards the formation of cyclohexanol and cyclohexanone is exhibited by our systems (Table 3), since no traces of by-products were detected by GC–MS analysis of the final reaction mixtures for the optimised conditions. The previously recognised promoting effect of pyrazine carboxylic acid on the peroxidative oxidation of alkanes catalysed by homogeneous [4,6,72,79–83] or heterogeneous [84] metallic species is also observed for all the present systems (Table 3). The effect of the peroxide-to-catalyst molar ratio was also investigated and is depicted in Fig. 5. The increase of the peroxide

288

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

Fig. 5. Dependence of the overall turnover number (moles of cyclohexanol + cyclohexanone per mole of Au nanoparticles loaded on the carbon material) of the products on the amount of oxidant (H2 O2 , molar ratio relatively to Au/AC-COL ( ) and Au/CNT-COL ( )) in the oxidation of cyclohexane. Reaction conditions: n(H2 O2 )/n(catalyst) (0–4 × 104 ), CH3 CN (3.0 mL), cyclohexane (5.0 mmol), n(Hpca)/n(catalyst) (50), room temperature, 6 h.

amount up to n(H2 O2 )/n(catalyst) molar ratio of 2 × 104 leads to the maximum products TON values (e.g., 171 and 93 for Au/CNTCOL and Au/AC-COL, Table 3, entries 4 and 11, respectively). Further increase of the oxidant amount results in a yield drop due to overoxidation reactions at higher H2 O2 amounts. In fact, some overoxidation products, such as 1,4-cyclohexanedione and 1,2epoxycyclohexane, were detected by GC–MS using the conditions of entry 7, Table 3. A comparison of the different supports and impregnation methods is depicted in Fig. 6. Catalyst recycling was tested up to six consecutive cycles for the most active catalytic system, Au/CNT-COL. On completion of each stage, the products were analysed and the catalyst was recovered by filtration, thoroughly washed, and then reused for a new set of cyclohexane oxidation experiments. The filtrate was analysed relative to the presence of Au by atomic absorption spectroscopy, thus excluding the hypothesis of catalyst leaching. Moreover, it was tested in a new reaction (by addition of fresh reagents), and no oxidation products were detected. Fig. 7 shows the recyclability of the system Au/CNT-COL: it maintains almost the original level of activity after several reaction cycles (in a second, third, fourth, fifth and sixth run, the observed activity was 98%, 97%, 96%, 96% and 94% of the initial one) with a rather high selectivity to cyclohexanol and cyclohexanone.

Fig. 6. Dependence of the overall turnover number (moles of cyclohexanol + cyclohexanone per mole of Au nanoparticles loaded on the carbon material) of the products on the type of support and impregnation method (COL ( ) and DIM ( )). Reaction conditions: CH3 CN (3.0 mL), cyclohexane (5.0 mmol), n(Hpca)/n(catalyst) (50), room temperature, 6 h.

Fig. 7. Effect of the catalyst recycling on the overall turnover number of the products for cyclohexane oxidation catalysed by Au/CNT-COL. Reaction conditions: CH3 CN (3.0 mL), cyclohexane (5.0 mmol), n(H2 O2 )/n(catalyst) (2 × 104 ), n(Hpca)/n(catalyst) (50), room temperature, 6 h.

Most tests of catalyst recycling for Au catalysts reported in the literature go up to 4 cycles [14,17,21,23,25,26] and some show a decrease in the activity after the first cycle [14,17,23]. Gold nanoparticles supported on mesoporous silica prepared in the presence of thioether functionality showed a slight decrease in conversion after the second run but no obvious activity loss was observed in the following 4 cycles [19]. On the other hand, functionalised mesoporous silica (organosilane mercapto-propyltrimethoxysilane, MPTMS) was recycled for 6 times and exhibited little activity loss [20], similar to our Au/CNT-COL catalyst. The conversions obtained on Au/MPTMS were higher (15.8–22.1%) than ours, but the selectivity values ranged from 91.5% to 94.5% and a temperature of 150 ◦ C was needed, along with 0.8–1 MPa pressure, unlike our system.

3.4. Reaction mechanism As stated in the Introduction, Hutchings and co-workers [15] demonstrated that cyclohexane oxidation catalysed by Au/MgO proceeds through a radical chain mechanism. The peroxidative oxidation of cyclohexane, catalysed by various metal complexes, is believed to proceed mainly via a radical mechanism which involves both carbon- and oxygen-centred radicals [62,64,80,83,85–90]. It appears to be also the case in the present study, in view of the inhibition effect observed (no products detected) when the peroxidative oxidation is carried out in the presence of either a carbon-radical trap such as CBrCl3 or an oxygen-radical trap such as Ph2 NH [91,92]. In the experiments performed in the presence of CBrCl3 , the formation of CyBr was detected (4.2% relative to CyH) by GC analysis. Hence, by analogy with the proposed mechanisms for various Mn+1/n (e.g., V, Re, Fe, Cu) systems [80,82–84,86–90] we can propose the following. Metal-catalysed (and Hpca-assisted) decomposition of H2 O2 can lead to the oxygen-centred radicals HO• and HOO• , upon reduction of H2 O2 by Au0 and oxidation by AuI , respectively (reactions (1) and (2)), as suggested by Quintanilla et al. for gold nanoparticle catalysts for phenol oxidation [93]. Water is believed to catalyse H+ -shift steps towards the formation of HO• from H2 O2 [94–96]. Cyclohexyl radical Cy• is then formed upon H-abstraction from cyclohexane CyH by HO• (reaction (3)). Reaction of Cy• with dioxygen leads to CyOO• (reaction (4)), and CyOOH can then be formed upon H-abstraction from H2 O2 by CyOO• (reaction (5)) or upon reduction of the latter to CyOO− by Au0 followed by protonation. Metal-assisted decomposition of CyOOH to CyO• and CyOO• (reactions (6) and (7)) would then afford the cyclohexanol (CyOH) and cyclohexanone (Cy-H O) products (reactions (8) and (9)) [97]. As suggested by Hutchings and co-workers [15], gold can

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

accelerate the reaction by increasing the concentration of species (through CyOOH or CyOO• ). Au0 + H2 O2 → HO• + AuI + HO−

(1)

AuI + H2 O2 → HOO• + H+ + Au0

(2)

HO• + CyH → H2 O + Cy•

(3)

Cy• + O2 → CyOO•

(4)

CyOO• + H2 O2 → CyOOH + HOO•

(5)

CyOOH + Au0 → CyO• + AuI + HO−

(6)

CyOOH + AuI → CyOO• + H+ + Au0

(7)

CyO• + CyH → CyOH + Cy•

(8)

2CyOO• → CyOH + Cy-H = O + O2

(9)

4. Conclusions Of the studied nine different Au/carbon catalysts, Au/CNT-COL was the most active catalyst for cyclohexane oxidation, with TON of ca. 171 and a yield of 3.6% after 6 h reaction. This is comparable to the industrial process (that uses Co catalysts and high temperature, like 150 ◦ C), but was obtained at ambient temperature with considerable low loads of catalyst (Au catalyst to substrate molar ratio always lower than 1 × 10−3 ), which is of large importance for establishing a greener catalytic process for this reaction. Moreover, a very high selectivity towards the formation of cyclohexanol and cyclohexanone was obtained. Catalyst recycling was tested up to six consecutive cycles for the Au/CNT-COL material, and it was found that the catalyst maintains almost the original level of activity after several reaction cycles (the activity dropped to 94% after the sixth cycle) with a rather high selectivity to cyclohexanol and cyclohexanone and with no catalyst leaching. The differences in catalytic performance of the studied Au/carbon catalysts can be explained in terms of gold nanoparticle size (the lower sizes generally showed improved activity, as expected), mesoporosity characteristics of the support and dispersibility in the reaction media (which were found to have a detrimental effect on the activity). No apparent relationship was found between catalyst performance and the support BET surface area or surface chemistry. The peroxidative oxidation of cyclohexane most likely proceeds through a radical chain mechanism. Acknowledgments The authors are grateful to Fundac¸ão para a Ciência e a Tecnologia (FCT) and FEDER for the financial support through projects PEst-C/EQB/LA0020/2011, PEst-OE/QUI/UI0100/2011, PTDC/EQUEQU/122025/2010, PTDC/QUI-QUI/102150/2008 and PTDC/QUIQUI/100682/2008 in context of COMPETE programme, and CIÊNCIA 2007 programme (SACC), as well as the Executive Research Agency of the European Union for funding under the Marie Curie IEF grant number 272448 (JGB). Dr. Nicolas Cayetano is acknowledged for ˜ for SEM work, help with TEM work, Gladis Labrada and Ana Iris Pena and LINAN for providing access to microscope facilities. The authors are also thankful to Dr. Carlos M. Sá (CEMUP) for assistance with XPS.

289

References [1] K. Weissermel, H.J. Arpe, Industrial Organic Chemistry, 2nd ed., VCH Press, Weinheim, 1993. [2] Ullmann’s Encyclopedia of Industrial Chemistry, 6th ed., Wiley-VCH, Weinheim, 2002. [3] R.A. Sheldon, I. Arends, U. Hanefeld, Green Chemistry and Catalysis, Wiley-VCH, Weinheim, 2007. [4] U. Schuchardt, D. Cardoso, R. Sercheli, R. Pereira, R.S. Da Cruz, M.C. Guerreiro, D. Mandelli, E.V. Spinacé, E.L. Pires, Appl. Catal. A: Gen. 211 (2001) 1–17. [5] G.B. Shulpin, Mini Rev. Org. Chem. 6 (2009) 95–104. [6] A.M. Kirillov, G.B. Shul’pin, Coord. Chem. Rev. 257 (2013) 732–754. [7] G.J. Hutchings, M. Haruta, Appl. Catal. A: Gen. 291 (2005) 2–5. [8] G.C. Bond, C. Louis, D.T. Thompson, Catalysis by Gold, Imperial College Press, London, United Kingdom, 2006. [9] S.A.C. Carabineiro, D. Thompson, in: U. Heiz, U. Landman (Eds.), Nanocatalysis, Springer, Berlin, Heidelberg, 2007, pp. 377–489. [10] S.A.C. Carabineiro, D. Thompson, in: C. Corti, R. Holliday (Eds.), Gold: Science and Applications, CRC Press, Taylor & Francis Group, Boca Raton, London, New York, 2010, pp. 89–122. [11] P. Wu, K.P. Loh, X.S. Zhao, Sci. Adv. Mater. 3 (2011) 970–983. [12] J.T. Carneiro, T.J. Savenije, J.A. Moulijn, G. Mul, J. Photochem. Photobiol. A: Chem. 217 (2011) 326–332. [13] B.P.C. Hereijgers, B.M. Weckhuysen, J. Catal. 270 (2010) 16–25. [14] A. Alshammari, A. Koeckritz, V.N. Kalevaru, A. Bagabas, A. Martin, ChemCatChem 4 (2012) 1330–1336. [15] M. Conte, X. Liu, D.M. Murphy, K. Whiston, G.J. Hutchings, Phys. Chem. Chem. Phys. 14 (2012) 16279–16285. [16] Z.G. Sun, G. Li, L.P. Liu, H.O. Liu, Catal. Commun. 27 (2012) 200–205. [17] G.M. Lu, D. Ji, G. Qian, Y.X. Qi, X.L. Wang, J.S. Suo, Appl. Catal. A: Gen. 280 (2005) 175–180. [18] L. Li, W. Ji, C.-T. Au, Progr. Chem. 21 (2009) 1742–1749. [19] P. Wu, Z. Xiong, K.P. Loh, X.S. Zhao, Catal. Sci. Technol. 1 (2011) 285–294. [20] P. Wu, P. Bai, Z. Lei, K.P. Loh, X.S. Zhao, Micropor. Mesopor. Mater. 141 (2011) 222–230. [21] L.-X. Xu, C.-H. He, M.-Q. Zhu, K.-J. Wu, Y.-L. Lai, Catal. Commun. 9 (2008) 816–820. [22] X. Jiang, H. Deng, X. Wang, Colloid Surf. A: Physicochem. Eng. Asp. 358 (2010) 122–127. [23] J. Xie, Y. Wang, Y. Li, Y. Wei, React. Kinet. Mech. Catal. 102 (2011) 143–154. [24] C. Wang, L. Chen, Z. Qi, Catal. Sci. Technol. 3 (2013) 1123–1128. [25] L.-X. Xu, C.-H. He, M.-Q. Zhu, S. Fang, Catal. Lett. 114 (2007) 202–205. [26] J. Zhao, M. Zhu, J. Chen, Y. Yang, Y. Tang, Z. Cai, Y. Shen, C. He, in: Z. Cao, L. Sun, X.Q. Cao, Y.H. He (Eds.), Fundamental of Chemical Engineering, Pts 1–3, 2011, pp. 254–259. [27] Y. Liu, H. Tsunoyama, T. Akita, S. Xie, T. Tsukuda, ACS Catal. 1 (2011) 2–6. [28] C. Della Pina, E. Falletta, M. Rossi, Chem. Soc. Rev. 41 (2012) 350–369. [29] Y.J. Xu, P. Landon, D. Enache, A. Carley, M. Roberts, G.J. Hutchings, Catal. Lett. 101 (2005) 175–179. [30] G.J. Hutchings, S. Carrettin, P. Landon, J.K. Edwards, D. Enache, D.W. Knight, Y.-J. Xu, A.F. Carley, Top. Catal. 38 (2006) 223–230. [31] J.L. Figueiredo, M.F.R. Pereira, M.M.A. Freitas, J.J.M. Órfão, Carbon 37 (1999) 1379–1389. [32] J.L. Figueiredo, M.F.R. Pereira, M.M.A. Freitas, J.J.M. Órfão, Ind. Eng. Chem. Res. 46 (2007) 4110–4115. [33] J.L. Figueiredo, M.F.R. Pereira, Catal. Today 150 (2010) 2–7. [34] S.A.C. Carabineiro, M.F.R. Pereira, J.J.M. Órfão, J.L. Figueiredo, in: J.F. Kwiatkowski (Ed.), Activated Carbon: Classifications, Properties and Applications, Nova Science Pub, Inc., New York, 2011, pp. 125–168. [35] N. Job, R. Pirard, J. Marien, J.-P. Pirard, Carbon 42 (2004) 619–628. [36] P.V. Samant, F. Gonc¸alves, M.M.A. Freitas, M.F.R. Pereira, J.L. Figueiredo, Carbon 42 (2004) 1321–1325. [37] N. Mahata, M.F.R. Pereira, F. Suárez-García, A. Martínez-Alonso, J.M.D. Tascón, J.L. Figueiredo, J. Colloid Interface Sci. 324 (2008) 150–155. [38] P. Serp, in: P. Serp, J.L. Figueiredo (Eds.), Carbon Materials for Catalysis, John Wiley & Sons, Inc., Hoboken, New Jersey, 2009, pp. 309–372. [39] P. Serp, E. Castillejos, ChemCatChem 2 (2010) 41–47. [40] A. Krueger, J. Mater. Chem. 18 (2008) 1485–1492. [41] A. Schrand, S. Ciftan Hens, O. Shenderova, Crit. Rev. Solid State Mater. Sci. 34 (2009) 18–74. ¯ [42] M. Ozawa, M. Inaguma, M. Takahashi, F. Kataoka, A. Krüger, E. Osawa, Adv. Mater. 19 (2007) 1201–1206. [43] V.N. Mochalin, O. Shenderova, D. Ho, Y. Gogotsi, Nat. Nanotechnol. 7 (2012) 11–23. [44] H. Okatsu, N. Kinoshita, T. Akita, T. Ishida, M. Haruta, Appl. Catal. A: Gen. 369 (2009) 8–14. [45] J. Zhu, S.A.C. Carabineiro, D. Shan, J.L. Faria, Y. Zhu, J.L. Figueiredo, J. Catal. 274 (2010) 207–214. [46] S. Gil, L. Munoz, L. Sanchez-Silva, A. Romero, J. Luis Valverde, Chem. Eng. J. 172 (2011) 418–429. [47] E.G. Rodrigues, S.A.C. Carabineiro, X. Chen, J.J. Delgado, J.L. Figueiredo, M.F.R. Pereira, J.J.M. Orfao, Catal. Lett. 141 (2011) 420–431. [48] Z. Li, D. Huang, R. Xiao, W. Liu, C. Xu, Y. Jiang, L. Sun, Curr. Nanosci. 8 (2012) 26–28. [49] H. Alshammari, P.J. Miedziak, S. Bawaked, D.W. Knight, G.J. Hutchings, ChemCatChem 4 (2012) 1565–1571.

290

S.A.C. Carabineiro et al. / Applied Catalysis A: General 467 (2013) 279–290

[50] E.G. Rodrigues, M.F.R. Pereira, J.J.M. Orfao, Appl. Catal. B: Environ. 115 (2012) 1–6. [51] X. Xu, J. Zhu, J.L. Faria, J. Li, J.L. Figueiredo, Cent. Eur. J. Chem. 10 (2012) 1867–1874. [52] E.G. Rodrigues, S.A.C. Carabineiro, J.J. Delgado, X. Chen, M.F.R. Pereira, J.J.M. Orfao, J. Catal. 285 (2012) 83–91. [53] E.V. Golubina, E.S. Lokteva, A.G. Majouga, M.V. Lobanov, V.V. Lunin, Diamond Relat. Mater. 20 (2011) 960–964. [54] R. Martin, S. Navalon, M. Alvaro, H. Garcia, Appl. Catal. B: Environ. 103 (2011) 246–252. ˜ [55] A. Seral-Ascaso, A. Luquin, M.J. Lázaro, G.F. de la Fuente, M. Laguna, E. Munoz, Appl. Catal. A: Gen. 456 (2013) 88–95. [56] R. Martin, S. Navalon, J.J. Delgado, J.J. Calvino, M. Alvaro, H. Garcia, Chem. Eur. J. 17 (2011) 9494–9502. [57] S.A.C. Carabineiro, T. Thavorn-Amornsri, M.F.R. Pereira, J.L. Figueiredo, Water Res. 45 (2011) 4583–4591. [58] J.-P. Tessonnier, D. Rosenthal, T.W. Hansen, C. Hess, M.E. Schuster, R. Blume, F. Girgsdies, N. Pfänder, O. Timpe, D.S. Su, R. Schlögl, Carbon 47 (2009) 1779–1798. [59] Y. Onal, S. Schimpf, P. Claus, J. Catal. 223 (2004) 122–133. [60] M. Bowker, A. Nuhu, J. Soares, Catal. Today 122 (2007) 245–247. [61] S.A.C. Carabineiro, S.S.T. Bastos, J.J.M. Orfao, M.F.R. Pereira, J.J. Delgado, J.L. Figueiredo, Catal. Lett. 134 (2010) 217–227. [62] S.A.C. Carabineiro, S.S.T. Bastos, J.J.M. Orfao, M.F.R. Pereira, J.L. Delgado, J.L. Figueiredo, Appl. Catal. A: Gen. 381 (2010) 150–160. [63] S.A.C. Carabineiro, A.M.T. Silva, G. Drazic, P.B. Tavares, J.L. Figueiredo, Catal. Today 154 (2010) 21–30. [64] S.A.C. Carabineiro, B.F. Machado, R.R. Bacsa, P. Serp, G. Drazic, J.L. Faria, J.L. Figueiredo, J. Catal. 273 (2010) 191–198. [65] S.A.C. Carabineiro, N. Bogdanchikova, M. Avalos-Borja, A. Pestryakov, P.B. Tavares, J.L. Figueiredo, Nano Res. 4 (2011) 180–193. [66] S.A.C. Carabineiro, M. Avalos-Borja, J.G. Buijnsters, in: F.M.B.S.A. Méndez-Vilas (Ed.), Current Microscopy Contributions to Advances in Science and Technology, Formatex Research Center, Badajoz, Spain, 2012, pp. 1246–1251. [67] P. Beltrame, M. Comotti, C. Della Pina, M. Rossi, Appl. Catal. A: Gen. 297 (2006) 1–7. [68] S.A.C. Carabineiro, A.M.T. Silva, G. Drazic, P.B. Tavares, J.L. Figueiredo, Catal. Today 154 (2010) 293–302. [69] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319. [70] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380. [71] F. Rodríguez-Reinoso, A. Linares-Solano, Chemistry and Physics of Carbon, Marcel Dekker Inc., New York, 1989. [72] G.B. Shul’pin, J. Mol. Catal. A: Chem. 189 (2002) 39–66. [73] G.B. Shul’pin, C. R. Chim. 6 (2003) 163–178. [74] G.B. Shul’pin, Y.N. Kozlov, L.S. Shul’pina, A.R. Kudinov, D. Mandelli, Inorg. Chem. 48 (2009) 10480–10482.

[75] G.B. Shul’pin, Y.N. Kozlov, L.S. Shul’pina, P.V. Petrovskiy, Appl. Organomet. Chem. 24 (2010) 464–472. [76] A. Dandekar, R.T.K. Baker, M.A. Vannice, Carbon 36 (1998) 1821–1831. [77] S.A.C. Carabineiro, T. Thavorn-amornsri, M.F.R. Pereira, P. Serp, J.L. Figueiredo, Catal. Today 186 (2012) 29–34. [78] Y. Sunagawa, K. Yamamoto, H. Takahashi, A. Muramatsu, Catal. Today 132 (2008) 81–87. [79] G.B. Shul’pin, G.V. Nizova, Y.N. Kozlov, L.G. Cuervo, G. Süss-Fink, Adv. Synth. Catal. 346 (2004) 317–332. [80] G.S. Mishra, T.F.S. Silva, L.M.D.R.S. Martins, A.J.L. Pombeiro, Pure Appl. Chem. 81 (2009) 1217–1227. [81] R.R. Fernandes, J. Lasri, A.M. Kirillov, M.F.C. Guedes da Silva, J.A.L. Silva, J.J.R. Fraústo da Silva, A.J.L. Pombeiro, Eur. J. Inorg. Chem (2011) 3781–3790. [82] D.S. Nesterov, E.N. Chygorin, V.N. Kokozay, V.V. Bon, R. Boˇca, Y.N. Kozlov, L.S. Shul’Pina, J. Jezierska, A. Ozarowski, A.J.L. Pombeiro, G.B. Shul’Pin, Inorg. Chem. 51 (2012) 9110–9122. [83] M. Sutradhar, M.V. Kirillova, M.F.C. Guedes da Silva, L.M.D.R.S. Martins, A.J.L. Pombeiro, Inorg. Chem. 51 (2012) 11229–11231. [84] G.S. Mishra, E.C.B. Alegria, L.M.D.R.S. Martins, J.J.R. Fraústo da Silva, A.J.L. Pombeiro, J. Mol. Catal. A: Chem. 285 (2008) 92–100. [85] S. Contaldi, C. Di Nicola, F. Garau, Y.Y. Karabach, L.M.D.R.S. Martins, M. Monari, L. Pandolfo, C. Pettinari, A.J.L. Pombeiro, Dalton Trans. 0 (2009) 4928–4941. [86] C. Di Nicola, F. Garau, Y.Y. Karabach, L.M.D.R.S. Martins, M. Monari, L. Pandolfo, C. Pettinari, A.J.L. Pombeiro, Eur. J. Inorg. Chem. 2009 (2009) 666–676. [87] T.F.S. Silva, K.V. Luzyanin, M.V. Kirillova, M. Fátima Guedes Da Silva, L.M.D.R.S. Martins, A.J.L. Pombeiro, Adv. Synth. Catal 352 (2010) 171–187. [88] T.F.S. Silva, M.F. Guedes da Silva, G.S. Mishra, L.M.D.R.S. Martins, A.J.L. Pombeiro, J. Organomet. Chem. 696 (2011) 1310–1318. [89] M.N. Kopylovich, A.C.C. Nunes, K.T. Mahmudov, M. Haukka, T.C.O. Mac Leod, L.M.D.R.S. Martins, M.L. Kuznetsov, A.J.L. Pombeiro, Dalton Trans. 40 (2011) 2822–2836. [90] T.F.S. Silva, T.C.O. MacLeod, L.M.D.R.S. Martins, M.F.C. Guedes Da Silva, M.A. Schiavon, A.J.L. Pombeiro, J. Mol. Catal. A: Chem. 367 (2013) 52–60. [91] J.A. Howard, Free Radicals, Wiley, New York, 1973. [92] L.M. Slaughter, J.P. Collman, T.A. Eberspacher, J.I. Brauman, Inorg. Chem. 43 (2004) 5198–5204. [93] A. Quintanilla, S. García-Rodríguez, C.M. Domínguez, S. Blasco, J.A. Casas, J.J. Rodriguez, Appl. Catal. B: Environ. 111–112 (2012) 81–89. [94] M.L. Kuznetsov, A.J.L. Pombeiro, Inorg. Chem. 48 (2009) 307–318. [95] M.V. Kirillova, M.L. Kuznetsov, V.B. Romakh, L.S. Shul’pina, J.J.R. Fraústo da Silva, A.J.L. Pombeiro, G.B. Shul’pin, J. Catal. 267 (2009) 140–157. [96] M.V. Kirillova, M.L. Kuznetsov, Y.N. Kozlov, L.S. Shul’pina, A. Kitaygorodskiy, A.J.L. Pombeiro, G.B. Shul’pin, ACS Catal. 1 (2011) 1511–1520. [97] G.B. Shul’Pin, G. Süss-Fink, J. Chem. Soc. Perkin Trans. 2 (1995) 1459–1463.