Hydrogenolysis of alkanes: reactions of n-butane on Ru/zeolite catalysts

Geoffrey C. Bond * and Juan J. Garcia
Department of Chemistry, Brunel University, Uxbridge UB8 3PH, UK. E-mail: geoffrey10bond@aol.com

Received 7th April 2017 , Accepted 15th June 2017

First published on 19th June 2017


Ru/13X zeolite catalysts containing 0.6, 2 and 5.7% Ru prepared by reduction or decomposition of the hexammine complex are active and stable for n-butane hydrogenolysis between 400 and 460 K, and show similar activation energies (150–160 kJ mol−1) and orders of reaction (−1.5 to −2 in H2; 0.5–0.65 in n-butane); product selectivities show weak dependence on temperature and H2 pressure, with preference for terminal bond scission. Oxidising either the decomposed or the reduced catalysts gave a marked increase in activity, a lower activation energy (120–130 kJ mol−1), higher orders of reaction in H2 (∼−0.9), lower orders in n-butane (∼0.2) and product selectivities that showed greater change with temperature and H2 pressure; terminal bond scission was >90%. The change in character with oxidation was attributed to weaker H2 chemisorption and a higher concentration of partially dehydrogenated C4 species; the active phase may consist of very small particles of RuO2 confined within the supercages of the zeolite.


1. Introduction

The work of François Gault and his associates opened a new chapter in our knowledge of the transformations that hydrocarbons undergo at the surface of metallic catalysts, and of the mechanisms by which they take place. Their contributions have been summarised in the following way:1 “The mechanisms of skeletal isomerisation have been illuminated by experiments designed with high intelligence and performed with consummate skill, using alkanes labelled with either 13C or 14C, the products being analysed by mass-spectrometry, radiochemical methods or magic-angle spinning NMR (MASNMR).” For much of their work they used Pt catalysts in various forms, but the dependence of mechanism on catalyst structure was not at the centre of their enquiries, so that features such as particle size and morphology, and metal–support interaction, were not deeply studied. The variation of rate and product selectivity on operating conditions such as temperature and reactant pressures was also not a focus of attention, so that the determination of reaction kinetics, e.g. of orders of reaction and activation energy, was not much employed to resolve mechanistic questions.

An alternative approach to the study of reactions of this class is to concentrate on the behaviour of small alkanes, containing 2–4 carbon atoms, where the only or main reaction with H2 is C–C bond scission, i.e. hydrogenolysis. Skeletal isomerisation is of course possible with the butanes, but is largely restricted to Pt catalysts, which function only at quite high temperatures;1 metals such as Ru and Rh which are effective at much lower temperatures (typically 373–473 K) catalyse isomerisation only in very special circumstances.2 Quantitative study of these reaction is bedeviled by their tendency to lose activity through the formation of ‘carbonaceous residues’ that are probably formed by excessive dehydrogenation of the chemisorbed alkane. This tendency is exacerbated by the use of low H2/alkane ratios, so that the range of reactant ratios accessible for determining the kinetic equation is limited, and this restricts the accuracy of numbers derived from mathematical modelling of a reaction. A useful way of minimising this difficulty is the ‘short reaction pulse’ technique3 where a 1 min reaction period is followed by a longer period of H2 flow to cleanse the surface. The longer the alkane chain the worse the problem becomes, so that the relatively uninteresting ethane and propane are quite easy to examine; interest grows with n-butane, where either central or terminal scission can occur.

These quite simple reactions have attracted enormous interest, and there is an extensive literature describing the results and discussing their interpretation.4 The purpose of this article is to demonstrate that the use of a conventional flow system applied to Ru catalysts that have been exposed to various pretreatments is capable of yielding useful and indeed startling information without the need for full mathematical analysis; a little simple logic is all that is needed.

This paper reports a study of the formation of Ru/13X zeolite catalysts by ion exchange, followed by reduction or decomposition to metal. This support was chosen with a view to creating small Ru0 particles within the supercages ∼1.2 nm in size, so that we might contribute to the study of particle size effects and structure sensitivity in metal catalysis. The work being on a subject close to his interests, it is dedicated to the memory of François Gault, excellent scientist, good friend, taken from us far too soon.

2. Catalyst preparation and characterisation

13X zeolite (Linde Air Products) is an aluminosilicate of the faujasite family, in which the sodalite units are linked through their hexagonal faces, creating large ‘supercages’ 1.2 nm in size, connected by channels 0.74 nm wide; the negative charge on the framework is balanced by Na+ ions. Ru catalysts having this as support were made by partial exchange of the Na+ ions by [RuIII(NH3)6]3+ ions. The compound [RuIII(NH3)6]Br3 was made by conventional methods:5,6 RuCl3 reacted with Zn and conc. NH4OH to form [RuII(NH3)6]ZnCl4, and this was oxidised by Br2–H2O to give the desired complex, which was characterised by UV spectroscopy. Its aqueous solution is slightly unstable at room temperature, significantly so at 320 K, and especially at pH >7. The zeolite, obtained as 1/16 inch pellets, was crushed and a solution of the complex acidified to pH 5.4 with HCl was used for the ion exchange, which was complete in <1 h; 31% of the Na+ ions were replaced, being those located in the supercages. For the catalytic work the pellets were crushed and sieved to obtain a 52–36 mesh fraction; exchange was then restricted to the exterior parts due to limited diffusion to the inside. Levels of Ru were altered by changing the concentrative of the complex solution, and were determined by extraction with H2SO4; catalysts containing 0.67, 2.0 and 5.7% Ru were subsequently studied.

The Ru complex ions were reduced to Ru0 by a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 H2[thin space (1/6-em)]:[thin space (1/6-em)]N2 flow for 2 h at 573 K; TPR gave three main peaks equal in size at 500–530 K, the H2 consumption in each corresponding to a decrease of one in the oxidation state. The reduced material was grey in colour. Their decomposition was effected by N2 flow (2 h at 673 K); the NH3 ligands acted as reductant:

[Ru(NH3)6]Br3 → Ru0 + 0.5N2 + 2NH3 + 3NH4Br

The colour of the material changed during this process from white to purple to white to light green to grey; some HBr and Br2 were also detected.

Experience showed that extreme care was needed to purify the gases used, as traces of O2 caused the appearance of a blue-black ring at the top of the catalyst bed (see below). N2 (O2 free) was passed over Cu/SiO2 at ∼650 K and then over activated C at 175 K; H2 flowed over 0.5% Pd/Al2O3, any H2O being retained by the support.

Catalysts formed by reduction (R) and decomposition (D) were examined by selective gas chemisorption using O2, H2 and CO. With O2, measurements were made at 296 K, and also at 195 K to minimise the risk of O penetration into the particle, but no difference was seen; in this case isotherms sloped significantly above the ‘knee’, especially in the R series. H2 chemisorption at 296 K was notably slow, taking up to 45 min to equilibrate. Results of a number of measurements of monolayer volume converted into mean sizes by conventional methods (Table 1) showed no regular variation with Ru content: they suggest that particles formed by reduction are probably too large to remain within the supercages, while the smaller ones formed by decomposition may do so. TEM on a 2% Ru (R series) revealed particles of ∼3 nm, but none could be seen in a D series catalyst. CO chemisorption was also very slow, polycarbonyl species perhaps being formed.

Table 1 Estimates of mean particle size
Catalyst d/nma d/nmb d/nmc
a By H2 chemisorption. b By O2 chemisorption. c By TEM.
R series 4–6 3.7–7 3
D series 1.5–2 2.5


3. n-Butane hydrogenolysis: method and preliminary observations

Reactions were performed in a simple flow system, the standard gas composition being H2 0.447 atm, n-butane 0.045 atm, balance N2; this ratio was selected to minimise the risk of activity loss by ‘carbon’ deposition, and it allowed flexibility for studying the effects of changing reactant concentrations without altering the flow rate. Product analysis was by gas chromatography using a 3 m SiO2 column. Activation energies were measured over a range of 60–90 K; measurements on the R series were made at 415 ± 5 K, and those for the D series at 421 ± 4 K.

Catalyst stability is essential for obtaining sound kinetic information. The use of three Ru loadings each given three types of pretreatment (including oxidation, to be discussed later) provided nine sets of results; in six cases activity over a period of 2 or 3 h was constant from the outset, and in the other cases slight initial deactivation was followed by prolonged stability. This allowed a single charge of a catalyst to be used for all the work on it reported below, and accounts for its general consistency as shown in Tables 2 and 3 and the figures. Conversions were linear functions of contact time up to 30–35% conversion, but more importantly product selectivities changed only imperceptibly up to about 50% conversion (Fig. 1). It was important to establish this, so that the effects of operating variables that changed conversion could be recognised as valid within this range.

Table 2 Rates, activation energies and orders of reaction
r = kPHxPBy
Catalyst T/K r E x y
a r in mol min−1 gRu−1 × 106 at PH = 0.94 atm, PB = 0.059 atm. b In kJ mol−1. c At the cited temperature ±5 K.
0.6R 415 685 156 −1.97 0.50
2R 649 148 −1.78 0.65
5.7R 681 150 −1.70 0.50
0.6D 421 139 154 −1.90 0.48
2D 249 154 −1.97 0.65
5.7D 168 159 −0.84 0.45
0.6DO 404 2010 120 −0.84 0.23
2DO 2550 128 −0.92 0.18
5.7DO 1290 124 −0.90 0.45


Table 3 Effect of varying operating conditions on selectivity parameters F and T3; 1, flow rate T3: 1, flow rate; 2, temperature; 3, H2 pressure; 4, n-butane pressure
1 2 3 4
Catalyst F T 3 F T 3 F T 3 F T 3
Temperatures as used in Table 2; where duplicate entries are shown, they are for the lowest and highest levels of the variable.
0.6R 0.24 0.72 0.25 0.79 0.27 0.75 0.24 0.74
0.18 0.62 0.25 0.71
2R 0.19 0.62 0.19 0.82 0.16 0.71 0.16 0.71
0.13 0.84 0.16 0.63
5.7R 0.32 0.74 0.35 0.77 0.34 0.80 0.21 0.66
0.27 0.67 0.30 0.71
0.6D 0.22 0.74 0.23 0.84 0.21 0.75 0.22 0.79
0.21 0.70
2D 0.35 0.75 0.36 0.86 0.30 0.71 0.31 0.72
0.27 0.66
5.7D 0.53 0.87 0.54 0.76 0.57 0.77 0.53 0.74
0.37 0.59 0.47 0.72
0.6DO 0.05 0.41 0.05 0.80 0.05 0.63 0.08 0.66
0.02 0.33 0.05 0.54
2DO 0.05 0.41 0.01 0.75 0.06 0.64 0.04 0.61
0.04 0.36 0.07 0.48
5.7DO 0.07 0.42 0.01 0.79 0.06 0.66 0.06 0.65
0.05 0.28 0.06 0.40



image file: c7cy00677b-f1.tif
Fig. 1 Dependence of conversion and selectivities on reciprocal flow-rate for catalyst 0.6DO at 433 K (flow-rate in cm3 min−1).

4. Reaction network

The constancy of selectivities as the reaction proceeds (Fig. 1) immediately tells us that the three products (methane, ethane and propane) all result from n-butane during one residence on the surface, and that the last two of these are less reactive than n-butane, and cannot react further once they have vacated the surface, until its removal is well advanced. The reaction network7 therefore shows that adsorbed n-butane is either split into two C2 fragments (chance F) or into a C3 and a C1 (chance 1 − F). The C3 species may either desorb as propane (chance T3) or be further split into C2 + C1 (chance 1 − T3); similarly the C2 species may desorb as ethane (chance T2) or react to give two C1 species (chance 1 − T2). At this stage we do not need to specify the exact composition of each adsorbed species, or the number of H atoms involved in each step; some light is shed on these matters later.

Product selectivities are defined as

 
S1 + 2S2 + 3S3 = 4(1)

Steady state analysis then leads to the following relations:

 
S3 = T3(1 − F)(2)
 
S2 = T2(1 + FS3)(3)

Mass-balance considerations and the lack of changes to S2 and S3 at low conversion then lead to

 
S2 + S3 = 1 + F(4)
that is to say, the adsorbed C2 has unit chance of desorbing as ethane at this stage. Application of a non-linear regression analysis to results in a wider range of conversions showed that the value of T2 was almost always >0.9, and values of the other constants closely matched those from (4). In the Tables following, constants obtained by regression analysis are shown for flow-rate variation; those shown for temperature and reactant concentration variations are given by (4). No values for T2 are quoted, as they do not impinge on the main conclusions.

5. Catalysts formed by reduction (R) and decomposition (D) of the precursor

Rates at about 420 K for the six catalysts formed in these ways, expressed per unit weight of Ru, are shown in Table 2, and are those observed at the highest H2 pressure used (PH = 0.94 atm; PB = 0.059 atm). For the R series they are about the same; for the D series they are considerably lower, notwithstanding the higher dispersion, and there are some differences. Temperature variation as described in Section 4 gave excellent Arrhenius plots in all cases, and both procedures and the three Ru contents all afforded apparent activation energies (Ea) close to 150 kJ mol−1 (Table 2); the differences are not thought significant, and do not reflect the lower rates shown by the D series. For determining orders of reaction, H2 pressures were varied over a range of at least 0.59 atm; for n-butane orders, the range of pressure variation was at least 0.11 atm. Dependence of rates on reactant concentrations is well described by the simple power rate law
 
r = kPHxPBy(5)

Fig. 2 shows examples. The substantially negative orders of reaction in H2 (−1.5 to −2) are a common feature of alkane hydrogenolysis;4 the orders in n-butane are about 0.5.


image file: c7cy00677b-f2.tif
Fig. 2 Dependence of rate (in mol g−1 min−1) on reactant pressures: (A) H2 pressure variation for catalysts 0.6DO at 418 K and 0.6D at 422 K; (B) n-butane pressure variation for catalyst 0.6DO at 418 K and 0.6D at 422 K.

The fact that H2 inhibits reaction at once implies that the n-butane is in competition with it for sites on which to adsorb. It can however only adsorb by the breaking of one or probably more than one C–H bonds, with the formation of σ-C–Ru and H–Ru bonds. This work does not permit further penetration into this matter, although the literature4,5 suggests that forming triadsorbed species having one C[double bond, length as m-dash]Ru bond may be needed to stress the C–C bond to the point where it can be broken by an attacking H atom. A quite large ‘landing site’ is therefore required for activating the n-butane molecule, and this is shown by the strongly negative orders in H2. The concepts of active centres and landing sites have been discussed:8 it is by no means certain that the ‘site’ at which a C–Ru bond can be formed is the same as that which can accommodate an H atom, but such nuances are currently beyond the capability of mathematical modelling of reaction kinetics. Furthermore the atomic/molecular description of reactions on particles so small that they may only be capable of permitting one reactive occurrence at a time when enclosed within a zeolite cavity has not yet been attempted.

Before discussing the mechanism further, we should consider the effect of the process variables on product selectivities (Table 3). Where duplicate entries are shown, they are for those for the lowest and highest levels of the variable. Again the behaviours of the R and D series are broadly similar; irrespective of whether the variable is temperature, H2 or n-butane pressure, most of the values of F lie between 0.15 and 0.35. Some of the occasional values outside this range are due to the limited accuracy of the analysis at conversions below 1%. In each set there is no systematic variation of F within the range employed; Fig. 3 shows the results for the 0.6D set. The constancy is particularly marked when the n-butane pressure is changed, so that only one value is shown in each set (Table 3). In most cases however the values of T3 decrease slightly with increasing temperature and are essentially independent of H2 pressure (Fig. 3); closely corresponding values are found when n-butane pressure is altered (Table 3). Fig. 3 shows that there is close consistency in values of F and of T3 regardless of the variable used, after allowance is made for the exact conditions used.


image file: c7cy00677b-f3.tif
Fig. 3 Effects of temperature and H2 pressure on selectivity parameters F and T3 for catalyst 2D: filled points, from flow-rate variation; half-filled points, from order of reaction measurements.

We may therefore summarise the performance of all members of the R and D series in the following way. The chance of converting the adsorbed C2 units to gaseous ethane (T3) varies only slightly, in a way that suggests this process is favoured by a higher coverage of the surface by H atoms; the small extent of the variation implies that they are quite strongly held. The position of scission appears to be essentially independent of all variables, meaning that the structure of the reactive species is determined by the manner in which the molecule is adsorbed and not by the availability of a suitable landing site. Thus a 2,3-diadsorbed or a 2,2,3-triadsorbed species is implicated. The lower and variable rates and differences in reaction orders shown especially in the D series are not reflected in the product selectivities, and are thus not of fundamental significance.

6. Catalysts formed by oxidising ‘decomposed’ catalysts (DO series)

The dramatic effect of oxidising a ‘decomposed’ catalyst was revealed by the following experiment (Table 4). A sample of 2D catalyst was prepared in the usual way; in the reactor its colour was mainly grey but there was a small blue-black ring at the top of the bed, comprising about 5% of the whole. This was attributed to an oxide phase, formed despite the rigorous precautions taken to purify the N2 (see above). The H/Ru ratio was 0.34. To obtain a thoroughly reduced material, it was then treated in H2 twice, first at 453 K for 1 h and then at 493 K for 0.5 h, the activity being measured at 407 K after each stage; selectivities (Table 4) were similar to those found with the R and D series, and although the black ring did not disappear there were minor changes possibly caused by its partial removal. After purging with H2 and then N2, the temperature was raised to 673 K, and pulses of air introduced via a septum cap into the N2 flow until the catalyst looked completely black; the activity was measured again at 407 K. There was a very marked increase in conversion and substantial changes to product selectivities; S1 and S3 both increased and S2 fell, and in consequence F dropped to about 0.1.
Table 4 Effect of reducing and oxidising a ‘decomposed’ catalyst
Pretreatments Conversion/%a S 1 S 2 S 3 F T 3
a As measured at 407 K after a short period for stabilising.
Decomposition 1.0 0.96 0.76 0.51 0.26 0.68
1st reduction 0.9 0.83 0.91 0.45 0.36 0.70
2nd reduction 1.3 0.79 1.00 0.40 0.40 0.67
Oxidation 20.1 1.19 0.51 0.60 0.11 0.67


Each of the three precursors differing in Ru content was then decomposed as usual, and after purging was oxidised at 673 K as before; these formed the DO series and gave the results shown in Tables 2 and 3. Activation energies were lowered to about 125 kJ mol−1, the H2 orders became much less negative and the n-butane orders less positive; the change therefore led to less strongly inhibiting H2, allowing easier formation of the reactive form of the alkane, and thus faster rates. Consistently with this, temperature had a much larger effect on T3, and rather more marked increase as H2 pressure was raised (Fig. 4). Mechanistic implications are discussed further below. Values of F were always less than 0.1, showing that terminal C–C bond scission was even more greatly favoured than before. Once again results obtained by variation of flow rate, temperature and attractant pressures were broadly consistent (Fig. 4). The catalyst with the highest Ru content (5.7D) was slightly different from the others in respect of activity and n-butane order, but otherwise showed parallel behaviour. A 2% reduced catalyst was briefly examined after oxidation as above, and behaved just as the 2DO catalyst.


image file: c7cy00677b-f4.tif
Fig. 4 Effects of temperature and H2 pressure on selectivity parameters F and T3 for catalyst 2DO: filled points, from flow-rate variation; half-filled points, from order of reaction measurements.

The formation and properties of the black phase were explored further. Air dosing at room temperature also gave the change from grey to black, but the oxidation was then probably only superficial, as it was easily removed by H2; however oxidation at 323 K was more profound, and not removable by H2 at room temperature. The O2 uptake at 633 K was measured volumetrically; the isotherm showed a very sharp ‘knee’, the uptake corresponding exactly to O/Ru = 2.0. The constancy of behaviour during their study showed that no reduction occurred under reaction conditions; no H2 adsorption was detected by the pulse method on the DO catalysts. The only evidence for the particle size is negative: as with the D series, no particles were visible by TEM. It is therefore likely that there was no major change in size or location on oxidation.

7. 1% Ru/SiO2 catalyst

A 1% Ru/SiO2 catalyst prepared by impregnation and reduction of the chloride showed low activity at 425 K and deactivated quickly (0.1% conversion after 97 min); the F parameter was ∼0.5 and T3 0.6–0.7. Oxidation afforded a stabler and slightly more active catalyst, with F ∼ 0 and T3 ∼ 0.3. Other work on this system has shown that oxidation does not alter the dispersion,9 but these changes suggest some alteration to surface morphology of the type seen elsewhere.7 This catalyst was not studied further.

8. The reaction mechanism

Alkane hydrogenolysis has several unusual features that lead to the results presented above. Alkanes chemisorb dissociatively by losing one or more H atoms in an endothermic process that is favoured by rise in temperature, but the scission of a C–C bond and desorption of the fragments requires more H atoms than the alkane releases, and these are supplied by H2 molecules, the dissociative chemisorption of which is exothermic. Coverage by H atoms therefore falls with rising temperature. At low H2 pressure and H coverage, the alkane will adsorb freely, as landing sites are abundant; the reaction orders will therefore tend to x = 1 and y = 0, and the scission step is rate-determining. At high H2 pressure and H coverage, the rate is determined by the ability of the alkane to find a landing site, so the orders tend to x = 0 and y = 1. There will therefore be a maximum rate when the optimum balance exists between sites for alkane adsorption and H atoms to attack the adsorbed alkane. With our 10-fold excess of H2 as standard, and generally as reactant pressures are changed, we are well above the rate maximum, but well away from either limiting condition. However the changes in H2 and n-butane orders produced by oxidation of the catalysts point clearly to a lowering of the strength of H2 chemisorption, so the lesser inhibition of the reaction by H2 is the cause of the large increase in rate.

Two further consequences of this change can be traced. (1) The H2 adsorption being weaker on the oxidised catalyst, the H atom coverage falls more quickly as temperature is raised (Fig. 4). Further scission of the C3 fragment is therefore favoured by low H atom coverage, and conversely its desorption is favoured by high H atom coverage. This suggests that scission is initiated by a single H atom, while several are needed to break the Ru–C bonds that hold it to the surface.

(2) Values of activation energy are often disregarded as being uninformative about mechanism. The experimental value, properly termed the apparent activation energy Ea differs from the true value (Et) that would be found if the surface concentrations of reactants did not change. For exothermically adsorbed molecules the rate will increase less than it ought if coverage remained constant, and the apparent one will be less than the true one; the opposite will apply in the unusual case of endothermic adsorption, but here we have one of each, and we cannot know which will dominate. Detailed analysis of related systems7 shows that Et < Ea when a 10-fold excess of H2 is used, so that the greater ease of achieving chemisorption of the alkane on increasing temperature is more important than the loss of H2 adsorption. The lower value of a found after oxidation therefore reflects its easier chemisorption and activation. This situation is described formally by the Temkin equation:

 
Ea = Et − ∑niΔHi(6)
where ni are the orders of reaction and ΔHi are the adsorption enthalpies.10 In our situation this becomes
 
Ea = EtxΔHH + yΔHB(7)
so that increasing x and decreasing y both lower Ea providing Et does not change. A general trend between x and Ea has been noted for a number of metals;11 the least active metals (Pt and Pd) have Ea values of 200–250 kJ mol−1 and x is about −2.5, while for more active metals Ea is less than 150 kJ mol−1 and values of x are less negative or even positive.

9. Conclusions

We believe the oxidised phase comprises very small particles of RuO2, which in bulk is blue-black and shows metallic conductivity; the conditions of its formation and its stoichiometry clearly point to this. Although it is reluctant to chemisorb H2 at room temperature, the process must be activated, because in order to cause hydrogenolysis it must do so above ∼380 K, but in a weaker form than that shown by the metal. The formation of Ru0 atoms either by reduction or decomposition of the precursor must be followed by their aggregation into small particles, as the exchange is limited to one complex per supercage. Their invisibility in TEM suggests that they may remain in the supercages, and be restricted in size to about 1 nm. Recalling the problems experienced with the use of selective chemisorption, the size estimates by this method (Table 1) are not in serious disagreement with this conclusion.

Very similar changes in behaviour caused by oxidation have also been found with Ru/TiO2[thin space (1/6-em)]12 and Ru/Al2O3,7 although in these cases oxidation at 673 K was followed by reduction at 433 K, shown by TPR to effect reduction to Ru0. The changes in rate, Ea and product selectivities were however ascribed to an increase in particle size7 or formation of 2D patches12 due to unrestricted movement and aggregation of Ru4+ ions, and highly disordered surfaces after reduction. This ionic migration is not possible with SiO2 as support.9 With Ru/Al2O3 catalysts these alterations clearly led to weaker H2 adsorption,7 which in turn affected the kinetic parameters as seen in this work. There is however one distinction: neither of these systems showed the quite low scission factors we observed (Table 3), their values being mainly 0.5–0.75.7,12 Central C–C bond scission may be preferred in these cases because of the slightly weaker secondary C–H bonds favouring 2,3-diadsorption. The very low values (<0.1) we see after oxidation may result from our particles being inside supercages, into which the n-butane molecular has to enter head first, so that terminal scission is much preferred.

Acknowledgements

J. J. G. received financial support from CONICIT, Venezuela.

References

  1. G. C. Bond, Metal-Catalysed Reactions of Hydrocarbons, Springer, New York, 2005, ch. 14 Search PubMed .
  2. R. Burch, G. C. Bond and R. R. Rajaram, J. Chem. Soc., Faraday Trans. 1, 1986, 82, 1985–1998 RSC .
  3. G. C. Bond and J. C. Slaa, J. Chem. Technol. Biotechnol., 1996, 65, 15–20 CrossRef CAS .
  4. G. C. Bond, Metal-Catalysed Reactions of Hydrocarbons, Springer, New York, 2005, ch. 13 Search PubMed .
  5. F. M. Lever and A. R. Powell, J. Chem. Soc. A, 1969, 1477–1482 RSC .
  6. J. E. Fergusson, J. L. Love and J. N. Armor, Ruthenium Hexammines in Inorg. Synth., ed. F. A. Cotton, J. Wiley and Sons, Hoboken NY, USA, 1972, vol. 13 Search PubMed .
  7. G. C. Bond and J. C. Slaa, J. Mol. Catal., 1995, 98, 81–99 CrossRef CAS .
  8. G. C. Bond, Metal-Catalysed Reactions of Hydrocarbons, Springer, New York, 2005, ch. 5 Search PubMed .
  9. G. C. Bond and R. Yahya, J. Mol. Catal., 1991, 69, 75–93 CrossRef CAS .
  10. G. C. Bond, M. A. Keane, H. Kral and J. A. Lecher, Catal. Rev.: Sci. Eng., 2000, 42, 323–383 CAS .
  11. G. C. Bond, Catal. Today, 1999, 49, 41–48 CrossRef CAS .
  12. G. C. Bond and X. Yide, J. Chem. Soc., Faraday Trans. 1, 1984, 80, 3103–3118 RSC .

Footnote

This paper describes experimental work performed by Dr J. J. Garcia but he is unaware that it has been submitted for publication as we have no contact details for him. Dr J. J. Garcia, therefore, does not take any responsibility for the submission.

This journal is © The Royal Society of Chemistry 2017