Hydrogenation and Hydrogenolysis with Ruthenium Catalysts and Application to Biomass Conversion

With the rising emphasis on efficient and highly selective chemical transformations, the field of ruthenium-catalysed hydrogenation and hydrogenolysis reactions has grown tremendously over recent years. The advances are triggered by the detailed understanding of the catalytic pathways that have enabled researchers to improve known transformations and realise new transformations in biomass conversion. Starting with the properties of ruthenium, this chapter introduces the concept of the catalytic function as a basis for rational design of ruthenium catalysts. Emphasis is placed on discussing the principles of dissociative adsorption of hydrogen. The principles are then applied to the conversion of typical biomolecules such as cellulose, hemicellulose and lignin. Characteristic features make ruthenium catalysis one of the most outstanding tools for implementing sustainable chemical transformations.


Introduction
Life on earth inherently depends on the element carbon creating the heart of a myriad of chemical compounds that, together with water and some inorganic compounds, build living matter. Over geologic periods, life has established a dynamic equilibrium of the flows of carbon through the different geo-habitats [1]. With the rise of mankind, this balance has been undermined through the exploitation of vast amounts of fossil resources for generating heat and materials. The carbon dioxide (CO 2 ) emissions from combustion of fossil resources have resulted in rising atmospheric CO 2 concentrations and an increasingly evident change in the climate worldwide. Replacing fossil resources that at present make up more than 90% of the energy demand and the feedstock for the chemical industry [2] is one of the most pressing challenges of mankind. All our primary energy demand of annually 12.5 TW a −1 could be covered by harnessing a fraction of the 8405 TW a −1 renewable energy available annually that comprises solar, wind, geothermal, tidal and wave energy [3]. Nevertheless, a sustainable energy supply will be needed for carbonbased compounds in order to close carbon recycle streams. Biomass is a globally available resource that is considered a suitable alternative feedstock for producing basic chemical building blocks, so-called platform molecules [4], that could substitute the current fossil-based platform chemicals [5].
Biomass largely consists of complex molecules comprising mostly oxygen and other heteroatoms. Lignocellulose, the structural component of plants and the largest fraction of plant biomass, is essentially composed of cellulose, hemicellulose and lignin. Break down of the structure by depolymerisation of the corresponding molecular entities, followed by oxygen removal, yields fuels and platform chemicals for the value-chain of the chemical industry. Sustainable conversion depends on efficient conversion steps obtained ideally via catalytic processes. In this context, the catalytically highly active element ruthenium provides unique properties. Despite ruthenium being counted among the noble metals, it resembles a non-noble metal in many aspects. In metallic form, ruthenium atoms are highly polarisable. Unlike the higher homologue platinum, e.g., that has similar atomic radius, ruthenium has a much higher average electric dipole polarizability. Consequently, distinct catalytic functions can be realised with ruthenium catalysts.
To help readers understand why ruthenium catalysts are so frequently employed in biomass conversion, this chapter will first investigate the properties of ruthenium. Here, the catalytic properties of ruthenium are linked with its propensity to adsorb certain molecular entities. After exploring the interaction of adsorbed molecules with ruthenium surfaces, we will discuss the nature of selected adsorption states, the corresponding binding energies and structures of the adsorption complexes including ordering phenomena observed for molecules co-adsorbed on the ruthenium metal surface. This sets the scene for rational design of catalysts that are specific for the conversion of chemical entities in biomass. Last but not least, we will discuss selected examples for intriguing transformations of biomolecules.
To note here is that this chapter does not aim to comprehensively review the available data on catalysis with ruthenium. Nor does it attempt to summarise all data on the conversion of biomass with ruthenium catalysts. The extensive interest in this field is reflected presently by the more than 800 articles published each year on catalysis with ruthenium, more than 110 of which focus on biomass conversion. Instead, this chapter aims to summarise the catalytic principles governing hydrogenation and hydrogenolysis reactions with heterogeneous ruthenium catalysts with particular focus on applications in biomass conversion. Cited data and papers were selected to exemplify the field and illuminate the discussion.

Ruthenium
Ruthenium, from Latin ruthenia ("Russia"), is one of the late transition metals and is located in the periodic table in the 5th period and group 8 (Figure 1). With an abundance of 7.0 ± 0.9 ng . g −1 in the silicate shell [6], ruthenium is one of the rarest non-radioactive elements on earth. Its low abundance is due to segregation of the platinum group elements in the core of Earth that was partially compensated by addition of 0.3-0.8% of chondritic material after core formation had been complete [6]. Ruthenium is found mostly in deposits associated with the other platinumgroup elements [7] and as the rare RuS 2 mineral called laurite [8]. Ruthenium is a silvery white, extraordinarily hard and brittle metal. With a density of 12.45 g . cm −3 [9], ruthenium is the second lightest platinum group metal after palladium. In the electronics industry, it is used in devices for perpendicular recording [10], a technology applied in hard disks that enables high-density data storage on magnetic media. Calculated heats of formation of the (unstable) binary hydrides M n H y (n > y) for group 8-10 transition metals [12] (left) and position of ruthenium (circle) in the periodic table (right) indicating the group of noble metals (bold, grey) and the platinum group metals (dark grey). different catalysts are involved before the oxidising equivalents are reacted with carbon compounds [25].
Thinking in terms of sequences of consequential reaction steps is a useful strategy to rationally design heterogeneous catalysts. A good starting point is considering the catalytic functions [26] necessary for realising the desired transformations. The dissociative adsorption of molecular hydrogen is one of the key steps for hydrogenation and hydrogenolysis reactions, the focus of this chapter. In the case of transfer hydrogenation, the concepts equally apply to suitable hydrogen surrogates. As such, dissociative adsorption of hydrogen, as one of the important steps of catalysis, will be elucidated below. With the Langmuir-Hinshelwood mechanism most prominent in catalysis with late transition metals, co-adsorption of the substrate and transfer of hydrogen atoms to an unsaturated substrate need to be considered next. Other catalytic functions important for biomass conversion are the ability of a catalyst to either cleave or form C-C, C-O or C-N bonds. This results in a list of complementary catalytic functions that are required for realising the desired transformation. Thereby it is useful to consider orthogonal catalytic functions that do not interfere with each other. Rather molecules ought to be passed from one catalytic function to the next, like in a molecular assembly line. Noteworthy, such assembly lines may involve a single material comprising different functions. Frequently the support plays an important role even when the actual transformation occurs on supported metal nanoparticles. One aspect to be considered regarding hydrogenation and hydrogenolysis reactions is spill-over of hydrogen to surface sites on the support. Another concept for realising such assembly lines involve mechanical mixtures of two or more materials that comprise different catalytic functions. An example is given below. Whereas heterogeneous ruthenium catalysts can accommodate many of these catalytic functions, homogeneous ruthenium catalysts enable unique, highly distinct catalytic transformations. Once the necessary catalytic functions have been identified, it is useful to derive the link to the desired active state and the structure of the pre-catalysts that is to be used. This provides a straightforward path for rationally designing a particular catalyst for the desired transformation.

Sequential reactions
Rational and straight-forward catalyst design is the foundation of systematic conceptualisation of highly active catalysts that provide extraordinary specificity for a given transformation. Such specificity is essential upon designing catalysts for biomass transformations, because the chemist typically encounters many different molecules or molecular entities rather than single types of molecules that are to be converted. If chosen in the appropriate way, the catalyst will adsorb and convert only one type of molecule or chemical entity while leaving all other molecules and chemical entities untouched. This concept is also valuable for devising catalysts for sequentially connected, mutually exclusive catalytic reactions. To develop such catalysts, the chemist needs to fundamentally understand the nature and catalytic role of active sites to guide the design of new and improved catalysts. Two examples are described here. The general principle is exemplified for a radical reaction with a MOF catalyst; the potential is then demonstrated for the hydrogenation of a multifunctional substrate over a Ru/CNT-Pt/CNT catalyst mixture.
Metal organic framework (MOF) compounds are porous materials commonly obtained by hydrothermal reaction of metal ions and bridging organic DOI: http://dx.doi.org /10.5772/intechopen.97034 ligands [27]. MOFs combine the high porosity of a heterogeneous catalyst with the tunability of molecular functional groups. This combination of features has been exploited for the sequential oxidation of alcohols to carboxylic acids with molecular oxygen in the presence of TEMPO modified MOF UiO-68 [28]. The conversion involves two sequential oxidation steps, i.e., the aerobic oxidation of alcohols to aldehydes, and the consequential autoxidation of the aldehydes to carboxylic acids. Whereas the first step is a radical reaction, the second step is inhibited by radicals. Thus, the two reactions are mutually exclusive. Complete removal of the MOF catalyst after the first radical-catalysed aerobic oxidation step by filtration provides the radical scavenger-free conditions that are necessary for the second radical-inhibited autoxidation step. This is a beautiful example of the use of a functional heterogeneous catalyst for a sequential organic transformation.
The concept of connecting consecutive one-pot reactions with a "molecular assembly line" has been explored for the hydrogenation of bifunctional substrates A-B to products A H -B H [29]. Two catalysts were chosen in such a way that one catalyst (M 1 ) preferentially adsorbs one of the substrate moieties, and the other catalyst (M 2 ) preferentially adsorbs the second substrate moiety (Figure 2). In this case both catalysts function optimally, thereby yielding improved rates and selectivities compared to single or conventional bimetallic catalysts [29]. Moreover, substrate inhibition can be avoided. By adjusting the relative quantity of the two catalysts, the relative rates of the two sequential transformations can be adjusted to be equal, because this results in the highest overall rate at the lowest catalyst concentration.
This concept has been applied successfully to the full hydrogenation of nitroaromatics to cycloaliphatic amines over a mechanical mixture of carbon nanotube (CNT)-supported Ru/CNT -Pt/CNT catalysts [29]. Noteworthy is that the aromatic ring, considered to be "soft" due to the aromatic π-system delocalised over six carbon atoms, preferentially adsorbs on ruthenium that is readily polarizable. The nitro group, considered to be "hard" due to the negative charge which is delocalised over only two oxygen atoms, preferentially adsorbs on platinum with highly shielded d-electrons. Notably, metallic Ru and Pt have similar atomic radii of 133 and 137 pm, while differing in the static average electric dipole polarizability of 9.6 and 6.4 10 −24 cm 3 , respectively. A 95:5 mixture of the Ru/CNT (M 1 ) and Pt/CNT (M 2 ) catalysts provides the required equal rates for hydrogenation of the two respective moieties and optimum selectivity to the target product cyclohexylamine (Figure 3).

Catalytic transformations with ruthenium catalysts
Based on the unique catalytic functions given by heterogeneous and homogeneous ruthenium catalysts, a large number of important transformations have been realised. Many of these transformations are applied on an industrial scale. For hydrogenation and hydrogenolysis reactions, in particular, heterogeneous ruthenium catalysts are among the most frequently applied catalysts, because they provide outstanding activities and excellent selectivities.

Ammonia synthesis and methanation with ruthenium catalysts
Analogous to iron and osmium, ruthenium catalyses the formation of ammonia from nitrogen and hydrogen (Eq. 1). Ruthenium has superior catalytic activity compared to iron [30] and results in enhanced NH 3 yields at lower pressures. A ruthenium catalyst, which is supported on a carbon matrix and improved by barium and caesium as promoters, has been in industrial use in two production sites in Trinidad since 1998 [31]. As the slow methanation of the carbon support [32] interferes with the process, alternative supports are preferred for ruthenium catalysts applied in ammonia synthesis. Efficiencies as close as possible to the theoretical limit are highly relevant for decentralised, islanded ammonia production plants [33,34], where round-trip efficiencies of up to 61% can be reached [35]. An example for a highly active and stable low-temperature ammonia catalyst are ruthenium nanoparticles on a Ba-Ca(NH 2 ) 2 support [36]. At a weight hourly space velocity (WHSV) of 36 L g −1 h −1 , a rate of 23.3 mmol NH3 g −1 h −1 is obtained at 300 °C and 9 bar. Such catalytic activity is about 6 times higher than that of industrial iron-based benchmark catalysts (at 340 °C) and 100 times higher than that of industrial rutheniumbased benchmark catalysts (Cs-doped Ru/MgO, at 260 °C) [36]. In addition, for the reverse reaction of ammonia cleavage, high activities are likewise important [37,38] and imply the use of ruthenium catalysts for the upcoming production of CO x -free hydrogen by ammonia cleavage in energy applications.
Analogous to nickel, ruthenium catalyses methanation, the production of methane from hydrogen and carbon dioxide (Eq. 2) or carbon monoxide (Eq. 3), the so-called Sabatier reaction. Water is obtained as by-product. Carbon dioxide methanation could be seen as the combination of the reverse water gas shift reaction that converts a mixture of carbon dioxide and hydrogen to carbon monoxide and water (Eq. 4), and methanation. Over ruthenium catalysts, such as Ru/Al 2 O 3 , the coproduction of CO is negligible [39]. This suggests a different reaction pathway not involving the intermediate formation of CO. Both reactants, H 2 and CO 2 , are strongly adsorbed on the surface [39] giving rise to a Langmuir-Hinshelwood mechanism. Ruthenium catalysts are highly selective to methane and provide a very low fraction of side products, such as higher hydrocarbons, alcohols, or formic acid. Due to the exothermicity and volume reduction, the reaction is thermodynamically favoured at low temperatures and high pressures. Typical operation conditions are 200-500 °C and pressures of 10-30 bar [40]. Since ruthenium catalysts have a higher activity than nickel catalysts, they enable higher conversions at low temperature. Methanation has long been used for removing CO x from the hydrogen-nitrogen syngas mixture used in ammonia production [41]. Carbon dioxide methanation is an option for biogas upgrading that constitutes an alternative to the removal of carbon dioxide [42]. Carbon dioxide methanation has also been discussed in the context of storing intermittent energy generated as a result of electricity production from renewable resources. Methane can be transported and stored in the existing natural gas grid. Therefore, methanation of carbon dioxide is being discussed as one of the promising Power-to-X technologies [43].

Hydrogenation with ruthenium catalysts
Ruthenium is an efficient catalyst for hydrogenating aromatics, acids, ketones and unsaturated nitrogen compounds. The selective hydrogenation of aromatic amines to cycloaliphatic primary amines is an industrially relevant transformation, but is impaired by formation of secondary amines and other side products. Modification of carbon nanotube (CNT)-supported ruthenium catalysts Ru/CNT catalysts with a base (LiOH) significantly improves selectivity in toluidine hydrogenation [44,45] without decreasing the activity of the catalysts. LiOH-modified Ru/CNT catalysts can efficiently convert also other challenging substrates, such as methylnitrobenzenes [46]. The effect of LiOH is understood as (i) LiOH reducing acidic sites on the catalyst support, (ii) enhancing hydrogen dissociation and reducing hydrogen spillover from ruthenium to the support (vide infra) and (iii) shifting the adsorption mode of the substrate on the ruthenium metal nanoparticles from binding of the amine group to the aromatic ring. In a similar manner, nitro compounds are able to change the binding mode of aromatic amines to the ruthenium surface [47,48].

Hydrogenolysis with ruthenium catalysts
The hydrogenolysis of alkanes is an important unit operation in refineries for reducing the chain length of acyclic alkanes. It also serves as a model for the hydrogenolysis of C-O and C-N bonds in various applications relevant for oil refining and biofuel generation. Cleavage of the C-C bond is preceded by a series of quasiequilibrated dehydrogenation steps (see Figure 4 for ethane hydrogenolysis [49]). Desorption of two chemisorbed hydrogen atoms generates the necessary adsorption sites on the surface. Physisorbed ethane dissociates stepwise via CH 3 CH 2 *, *CH 2 CH 2 *, *CH 2 CH* to form *CHCH*. Activation of the C-C bond in *CH-CH* has a lower intrinsic barrier in further dehydrogenation. Cleavage of the C-C bond in the *CH-CH* surface intermediate is thought to be the rate limiting step. During the entire process, the surface is covered to a large extent with chemisorbed hydrogen (H*). The high hydrogen coverage also enhances the re-hydrogenation of the unsaturated fragments to produce methane that is desorbed from the surface.
Similar to Ru, C-C bond cleavage in more deeply dehydrogenation intermediates is preferred for Os, Rh, Ir, and Pt relative to cleavage of the C-C bond in more saturated intermediates (Figure 4, right). Cleavage of the C-C bond in more saturated intermediates starts to compete as one moves more to the right of the periodic table. For the group 10 metals (Ni, Pd, Pt), the most favourable mechanism is C-C activation in *CHCH*, while other intermediates have activation energies of about 40 kJ mol −1 suggesting that multiple routes may coexist. For the coinage metals (Cu, Ag, Au), there is a preference for cleavage of the C-C bond in the most saturated intermediate CH 3 CH 2 *. The overall free-energy barrier for C-C bond activation is lowest for Ru providing the highest turnover rate for *CHCH* bond cleavage. Thus, the less noble metal Ru is more active than the more noble metals. This is also consistent with experimental data that show a decrease in the turnover rate in the sequence Ru > Rh > Ir > Pt [49].

Catalysis with molecular ruthenium catalysts
Some very active molecular ruthenium (pre)catalysts were developed for catalytic hydrogenation and transfer hydrogenation. Selected examples are shown in Figure  5. Ruthenium hydride complexes [50] with phosphine or diamine ligands are active for the hydrogenation of many substrates. Transfer hydrogenation with ruthenium catalysts is frequently used for the reduction of ketones to alcohols [51] and amides, imines and nitriles to amines [52,53]. Isopropanol is commonly employed as hydrogen donor [54]. The hydrogenation and transfer hydrogenation can be stereoselective if the starting material is prochiral and a chiral complex is employed [52,55]. However, chiral BINAP catalysts can reduce only functionalised ketones in a stereoselective manner. Whereas Noyori precatalysts of the type [RuCl 2 (diphosphane) (diamine)] enable the asymmetric hydrogenation of ß-keto esters as well as the  [49]. Energies are relative to a surface covered with chemisorbed hydrogen (H*); * denotes coordination to the ruthenium surface.
Hydrogenation and Hydrogenolysis with Ruthenium Catalysts and Application to Biomass… DOI: http://dx.doi.org/10.5772/intechopen.97034 reduction of prochiral ketones and aldehydes, olefins are usually not converted. The stereoselectivity is enhanced, when the substituents on the ligands differ in size. The concept of bifunctional asymmetric catalysis with ruthenium complexes has later been transferred to a variety of C-C, C-O and C-N forming reactions [56].
Ruthenium is also the central metal in the Grubbs catalysts [57], which are among the most important precatalysts for olefin metathesis. There are two generations of Grubbs catalysts (Figure 5). The first generation is often employed for ring-opening polymerisation (ROMP [21]) and for the synthesis of large rings by metathesis. The second generation [58] has a much higher activity. In Grubbs-Hoveyda catalysts, one of the tricyclohexylphosphine (PCy 3 ) ligands of the Grubbs catalysts is replaced by an aromatic ether that links to the carbon substituent. There is a wide field for ruthenium-catalysed cyclisation reactions [59]. Ruthenium N-heterocyclic carbene (NHC) complexes based on the second-generation Grubbs catalysts have also been applied in a variety of related transformations, such as hydrogenation [60], hydrosilylation, and isomerization [61]. Metathesis can also be combined with a second chemical transformation to tandem reaction sequences [61]. Likewise, living free radical polymerizations are feasible with ruthenium complexes [62]. An example is the polymerisation of methyl methacrylate with [RuCl 2 (PPh 3 ) 3 ] as a catalyst [63,64].

Hydrogen adsorption on metallic ruthenium
As for the other platinum group elements, metallic ruthenium is characterised by excellent catalytic results for a variety of transformations. The interaction of molecular hydrogen with the surface of ruthenium is particularly interesting as far as catalytic hydrogenation or hydrogenolysis reactions are concerned; it will be analysed in further detail here. Accordingly, the fundamental concepts discussed here likewise are valid for transfer hydrogenation reactions.
Dissociative chemisorption of hydrogen on the surface is a pivotal step of the transformation and is often rate-limiting. The adsorption of hydrogen may be considered as competing molecular and dissociative adsorption of hydrogen (Figure 6) [65]. Molecular adsorption is governed by the van der Waals interactions between molecular dihydrogen and the ruthenium surface [66]. On a Ru(0001) surface with point group symmetry C 3v [67], the four high-symmetry adsorption sites involve binding of the hydrogen molecule to a single ruthenium atom (on top), a position bridging two ruthenium atoms (brg) or three-fold coordination at fcc or hcp sites (Figure 6) [67,68]. Physisorption attracts a charge of −0.04 electrons to the hydrogen molecule [66]. This is consistent with promotion of dissociative adsorption of hydrogen in the presence of alkaline metal cations [69]. Electron transfer to the anti-binding orbitals of hydrogen and upward shift of the transition metal d-band centre towards the Fermi level are likely explanations.
For coordination of hydrogen in the molecular state [70], the on top site provides the highest adsorption energy of −20.3 kJ . mol −1 and the lowest dissociation barrier of 16.4 kJ . mol −1 [66]. Consequently, an entrance channel barrier is missing, and this dissociation channel appears to be active even for dissociation of H 2 molecules with negligible incident energy. Nevertheless, a suitable approach of the H 2 molecules to the ruthenium surface is essential for such a low dissociation barrier. Dissociation of molecular hydrogen on ruthenium is a rather slow process [69], and equilibrium is obtained only after several hours [71]. Point-like defect structures, like Ru vacancies or Ru adatoms on the surface do not seem to provide comparably low dissociation barriers. Other defects that are present at finite temperatures on the surface include steps, kinks and adatom islands [72,73]. Low coordinated defect sites may be the preferential sites for a direct dissociative adsorption pathway on ruthenium nanoparticles [74]. Due to the low barrier, the on top site is likely the most reactive site for hydrogen dissociation on extended ruthenium surfaces [67]. For supported ruthenium catalysts, a rapid H 2 /D 2 isotopic equilibration reaction has been reported [69]. Even so, the isotope exchange is slowed down considerably in the presence of alkaline metal cations that prevent spillover [70,75] of hydrogen atoms to the support.
Dissociative adsorption occurs when the bonds formed between the two hydrogen atoms and the ruthenium surface are stronger than the strength of the hydrogen-hydrogen bond (460 kJ . mol −1 ). This is the case when the hydrogen atoms adsorb at either  [66], chemisorption GGA with periodic plane-wave basis set, 1 monolayer coverage, no correction for zero point energy [78].

/3) and hcp (1/3,2/3) (top) [67] and changes in adsorption energy for physisorption of hydrogen E ads (H 2 ), dissociation barrier ΔE diss and chemisorption of hydrogen E ads (H) (physisorption vdW-DF2 + PBE level,
the fcc or the hcp hollow site (−258.6 and − 258.2 kJ . mol −1 , respectively). Noteworthy is the relatively small difference in energy between the fcc and hcp hollow sites. As for extended surfaces of other late transition metals, hydrogen, thus, has a pronounced preference for binding to multi-fold coordination sites [76,77]. As far as metal clusters and nanoparticles are concerned, the number of adsorption sites can differ, whereby specific 2-, 3-, and 4-fold coordination to surface atoms has been reported [76]. The barrier for surface diffusion [70] of hydrogen is rather small and was estimated to 13-21 kJ . mol −1 . There is a small decrease of −12.1 kJ . mol −1 in adsorption energy with coverage θ increasing from partial (1/3) to monolayer coverage.
At low temperature, the catalytically active ruthenium surface is normally covered to a large extent with hydrogen. The surface coverage remains incomplete under reaction conditions even at elevated pressures. Thus, at 100 bar, a coverage θ of ca. 85% was calculated at room temperature, whereby it decreased to ca. 70% at increasing temperature (500 °C) [78]. Temperature-programmed desorption of hydrogen from ruthenium catalysts shows two distinct desorption peaks as a characteristic feature [71,74]. The peaks represent strongly and weakly chemisorbed hydrogen, consistent with distinct NMR signals at −60 and − 30 ppm [79]. The corresponding heats of adsorption were determined to be 40-70 kJ . mol −1 (α H ) and 10 kJ . mol −1 (β H ), respectively, by microcalorimetry [69]. This suggests that part of the hydrogen is not dissociated over real samples. Consequently, a chemisorption stoichiometry x M exceeding unity is frequently considered (x M = 1.4 [74]; 2 [71,80]; 5 [79]). Although surface processes dominate, subsurface hydrogen cannot be ruled out [70,81]. Furthermore, the support can act as a reservoir for hydrogen [69].
Under catalytic conditions, surfaces are saturated by hydrogen or one or more adsorbed intermediates. This leads to strong co-adsorbate interactions. These interactions are not accounted for in kinetic models built on Langmuir isotherms. In real catalysts, however, mostly supported metal nanoparticles are employed, where these co-adsorbate interactions are lessened. The curvature of the nanoparticles allows for adlayer relaxation [82]. Thus, CO hydrogenation rates on Ru clusters are much higher at high CO coverage than predicted based on a Langmuir approach [83]. Activation of adsorbed CO by reaction with surface hydrogen results in transition states that occupy less space than [82] the pair of surface moieties that they replace. This causes the overall activation energy to decrease with increasing CO* coverage.
Interestingly, species co-adsorbed on a ruthenium surface may show a strong tendency to segregate. Thus, with carbon monoxide and hydrogen co-adsorbed on a Ru(0001) surface, the carbon monoxide molecules form islands that are surrounded by hydrogen atoms [84]. At cryogenic temperatures, the carbon monoxide molecules form triangular islands of up to 21 molecules located on the on top sites. Through this type of island formation, long-range lateral CO-H repulsive interactions are minimised. With an increase in temperature, the carbon monoxide molecules shift to the hcp sites and the island size decreases to 3-6 molecules [84]. Through this decrease in domain size, repulsive CO-CO interactions that become more prominent upon increasing the temperature are reduced. The proximity of the carbonyl and hydride adsorbate species to one another (3.0-3.7 Å distance) [84] explains the propensity of ruthenium surfaces for Fischer-Tropsch reactions. The ensuing CO bond cleavage is facilitated by the formation of partially hydrogenated CHO and COH intermediates.

Supports for heterogeneous ruthenium catalysts
For applicable heterogeneous catalysts, metallic ruthenium is supported in form of ruthenium nanoparticles on a suitable support. This ensures a high dispersion and a large surface area of ruthenium. Carbon supports, in particular active carbons and carbon nanotubes, and oxidic supports are frequently employed. To ensure that the ruthenium nanoparticles are immobile on the support surface under the catalytic conditions, there has to be a sufficiently strong interaction between metal nanoparticles and the support. Otherwise, there would be pronounced sintering of the ruthenium nanoparticles that would lead to gradual loss of the catalytic activity. The support also influences the electron density in the ruthenium nanoparticles, thereby lowering or increasing the Fermi level. For oxidic supports, the interaction between nanoparticles and the support cannot be too strongly pronounced, because ruthenium cations tend to diffuse into the bulk of the support material.
For carbonaceous materials anchoring sites have to be generated on the surface to anchor the ruthenium nanoparticles. Providing high surface area, active carbons and carbon nanotubes thus usually undergo an oxidative pre-treatment. As a result, oxygenated moieties are generated to which the ruthenium nanoparticles strongly bind. In this aspect, the property of ruthenium being at the borderline between noble and non-noble metals is exploited. Under more driving reductive conditions of a hydrogen atmosphere, however, the susceptibility of carbon carriers to methanation is challenging for carbon-supported ruthenium catalysts, because it leads to degradation of the carrier and sintering of the ruthenium clusters. Compared to active carbons, carbon nanotubes lend a more defined support and higher stability.
Carbon nanotubes combine physicochemical properties that make them interesting as support for ruthenium, such as high surface area, good mechanical strength, chemical and thermal stability, high heat and electric conductivity. So far, the high costs incurred by elaborate synthesis procedures [85][86][87][88][89] hinder their more widespread use as well-defined catalyst supports [90]. For immobilisation of metal nanoparticles, anchoring sites need to be generated on the surface of the carbon support. A method of preparing a Ru/CNT catalyst with supported ruthenium nanoparticles involves treatment of the CNT in refluxing nitric acid [91]. Deposition-precipitation of the ruthenium precursor Ru(NO)(NO 3 ) x (OH) y followed by reduction of the precursor to the metal with molecular hydrogen provides well-dispersed surface-anchored Ru nanoparticles (Figure 7) [29]. Such catalysts are excellent hydrogenation and hydrogenolysis catalysts (see below).
Oxidic supports that are frequently employed comprise silica, alumina (mostly γ-Al 2 O 3 ), zirconia, ceria and the corresponding mixed oxides. Even though amorphous materials provide the necessary high surface area, they often are associated with certain distribution of surface functions. Yet as surface groups, they may be harmful in catalysis. The presence of different surface sites often leads to alternative Particle size distribution of the ruthenium nanoparticles for a typical Ru/CNT catalyst and representative transmission electron microscopy images [29]. The carbon nanotubes are Baytubes C 150 P. DOI: http://dx.doi.org /10.5772/intechopen.97034 catalytic pathways that result in reduced selectivity of the transformation. Instead, more defined support materials are nanoporous zeolites, such as zeolite Y, Beta, and ZSM-5, or mesoporous materials, such as MCM-41. The internal pore system ( Table 1) provides a uniform environment for the catalytic transformation. Nevertheless, many biomolecules are too large to enter the pore system and need to be cut to molecular entities first. Catalysis at the pore mouth or using molecular catalysts is an option for the depolymerisation step.
The hydrophobicity index (HI) is a good measure for assessing the internal hydrophobicity of porous materials. HI can be determined by the competitive adsorption of a toluene-water mixture. The hydrophobicity index is defined by the ratio of the adsorption capacity for toluene (Q Toluene ) to that of water (Q Water ). For comparison, the reported HIs of some activated carbons, microporous zeolites, and mesoporous materials are listed in Table 1 [27]. The hydrophobicity index (HI) of typical zeolites, such as beta, Y and ZSM-5 is low (HI = 1.4, 0, 8, respectively), which is consistent with the hydrophilic nature of the pore walls. This is attributed to a certain polarity of the zeolite walls that results from the aluminium atoms substituting a certain part of the silicon atoms. All-silica zeolites, such as Silicalite-1, are clearly more hydrophobic (HI = 15.2) and more resemble activated carbons which are commonly regarded as hydrophobic adsorbents.
Unsupported metal nanoparticles can be employed as quasi-homogeneous catalysts but need to be stabilised by ligation or generation of an electric double layer to prevent agglomeration of the nanoparticles [94]. Upon decreasing the size of the metal nanoparticles, the boundary of the metallic state is obtained for two-shell clusters of about 1.5 nm in diameter [95]. Ruthenium  , the selectivity changes to the reaction intermediate aniline. This is attributed to the relative binding strength of ionic liquid and intermediates to the ruthenium surface. Thus, the use of ionic liquids as stabiliser lends a ready method to tailor the properties of the catalyst. Interestingly, ionic liquid-stabilised nanoparticles are readily supported on a mesoporous support [96,97] thus turning the quasi-homogeneous catalyst into a true heterogeneous catalyst. Noteworthy, the catalytically active site remains in the flexible environment of the ionic liquid [98] which imparts beneficial properties to the catalyst [99]. During the chemical transformation,  [27]; the pore size of zeolites (Beta, Y, ZSM-5, Silicalite-1 [92]) and mesoporous materials (MCM-41 [93]) is the maximum diameter of a sphere that can diffuse through the channels.
the active species can easily adapt to the geometry changes that occur during the path from reactant to transition and product state. Moreover, the equivalence of all catalytically active sites is readily maintained which can render enhanced selectivity. The ionic liquid then again provides a polar medium for tailoring the adsorption of the substrate molecules and desorption of the product molecules [100] that precede and succeed the catalytic reaction, respectively. Interestingly, in supported films of ionic liquid. Rates as well as chiral induction can be enhanced, as was demonstrated for the hydrogenation of the prochiral substrate acetophenone over [Ru((R)-BINAP) (PPh 3 ) n Cl 3-n ], n = 0, 1 [101]. A useful feature of such supported catalysts is that fixed-bed reactor technology common in continuous chemical processes can be employed [97].

Biomass conversion with ruthenium catalysts
About 1% of the incoming solar radiation on earth is captured for generating biomass [102]. This energy is utilised in photosynthesis [103] to build a myriad of complex molecules [104] such as carbohydrates, lignin, proteins, fats and oils, and terpenes. In this way about 170 x 10 9 t/a of complex substances are produced annually [105]. In plants, the radiation use efficiency is controlled by the netphotosynthetic capacity and the canopy structure [106,107]. Cultivars with a heavy canopy and long growth period are able to harness more solar radiation [108]. A large fraction of the produced biomass is characterised by a high oxygen content ( Table 2). Cellulose, a polymeric carbohydrate, and lignin a randomly linked phenolic polymer constitute a major fraction of plant biomass (around 95% [109]). Their oxygen content is much higher than that of fossil resources such as crude oil, natural gas and coal ( Table 2). About 56% of the oil extracted from the resources is utilised to make liquid fuels (70.6%) for transport purposes [2]. About 14% of the oil and 8% of the gas extracted from these resources is utilised to make petrochemicals. Both fuels and many petrochemical products are characterised by a low oxygen-to-carbon ratio. Some examples are given in Table 2. Consequently, in order to exploit biomass, a controlled de-functionalisation is necessary. In particular, efficient strategies are needed to decrease the oxygen-tocarbon ratio.
At present, biorefinery routes [118,119] have been improved to more efficiently exploit biomass feedstock. In the production of bioethanol from lignocellulosic biomass, e.g., by hydrolysis of wood with dilute acids, hexoses are obtained that are good feedstock for fermentation [4]. The target product then needs to be separated from the aqueous fermentation broth. By producing ethanol in this way, about 8.7% of the mass and 11% of the energy contained originally in the wood are found in the product [109]. The remainder are 37% by-products and 40% waste products, mostly carbon dioxide (36%) that need to be utilised or disposed. Green chemistry metrics [120], notably the E-factor and atom economy, clearly need to be improved further. One option is the direct chemical conversion of lignocellulosic biomass in a single reaction step over a multifunctional catalyst as outlined below. Such transformation follows the principles of a molecular assembly line. Thus, efficient and frequently multistep catalysis is one of the keys for realising fast and highly selective conversion of biomass [109]. Before the particular aspects of ruthenium catalysts in biomass conversion are considered, the general architecture and the availability of biomass is analysed briefly. Lignocellulose makes up the structural components of plants and a large fraction of the plant biomass available for producing platform chemicals. Wood, e.g., is essentially composed of cellulose (39-45%), hemicelluloses (27-32%) and lignin (22-31%) [121]. DOI: http://dx.doi.org/10.5772/intechopen.97034

Sustainable feedstock from biomass
Cellulose is an important structural component of the cell wall of green plants, many forms of algae and the oomycetes. Many bacteria secrete it to form biofilms [122]. Plants build about 10 11 -10 12 t/a of cellulose annually mostly in combination with hemicelluloses and lignin [123]. This makes cellulose the most abundant organic polymer on Earth [124]. Cellulose is a polysaccharide, a linear chain with the formula (C 6 H 10 O 5 ) n consisting of 7,000-15,000 of β(1 → 4) linked D-glucose units [125].
Oxygen content of typical components of biomass in comparison to fossil resources and selected derived products.
The components of plant biomass are normally fractionated using biochemical [129,130], thermochemical [131] and/or catalytic methods [132]. Lignin is particular is that respect that it is highly resistant to depolymerisation. Consequently, at present the lignin fraction is often used to a large extent as fuel for heat generation. Methods have been developed to utilise the phenolic structure for producing polymers, resins, additives, fuels and chemicals. Common methods for depolymerisation of lignin into monomeric phenolic compounds involve pyrolysis [133][134][135][136], enzyme [137,138], acid or base [139,140] catalysed hydrolysis, and hydrogenolysis [141][142][143]. Catalysts based on metallic ruthenium are frequently employed in hydrolysis and hydrogenolysis of the ether linkages or hydrodeoxygenation [110,144] of the phenol products (vide supra).
In subsequent downstream processing, the biomass fractions are converted to platform chemicals. Based on the generally accessible biomass, platform molecules (Figure 9 [4]) include organic acids, such as propionic acid, 3-hydroxypropionic acid, succinic acid, fumaric acid, itaconic acid, and levulinic acid [145], fat and oil-derived polyols, in particular glycerol, as well as sugar-derived polyols such as sorbitol and xylitol. Additional platform chemicals are alcohols such as methanol, ethanol, and propanol, cyclic ethers, such as furfural and 5-hydroxymethylfurfural, and terpenes, such as isoprene. Such platform molecules can be exploited as fuels  and industrially relevant chemicals or are readily transformed into such fuels and chemicals. Ruthenium-based catalysts are frequently employed in key transformations such as hydrogenation, hydrogenolysis, and oxydehydrogenation [146]. Compared to nickel-based catalysts, ruthenium-based catalysts provide higher activity and better stability that result in lower catalyst loadings, longer lifetimes and less pronounced deactivation. Although ruthenium-based catalysts are more expensive, these costs are offset by their higher activity and their lower tendency to leach.

Ruthenium catalysts in cellulose conversion
While first-generation bioethanol is produced on the million t/a scale, production of second-generation bioethanol from cellulosic biomass is still in its infancy [4]. The challenge is the enzyme-or acid-catalysed hydrolysis of lignocellulosic materials to simple sugars that can be fed into fermentation, from which ethanol is separated by distillation [147]. A one-step catalytic conversion of cellulosic biomass (bagasse and corn stalk) to bioethanol has been realised with a ruthenium-based catalyst [148]. The catalyst comprises well-dispersed Ru and WO x nanoparticles on a H-ZSM-5 solid acid support. Under catalytic conditions, also highly dispersed Ru 3 W 17 alloy nanoparticles are formed. In a cascade reaction cellulose undergoes hydrolysis on moderately acidic sites of the H-ZSM-5 support, followed by glucose retro-aldol condensation to glycolaldehyde over WO x and hydrogenation over Ru to yield ethylene glycol that is dehydrated and finally hydrogenated to ethanol on the Ru 3 W 17 alloy sites.
Interestingly, subcritical water is an efficient reaction medium for cellulose conversion [149,150]. Thus, cellulose is converted to polyols over ruthenium supported on crosslinked polystyrene [149,151]. Swelling of the polymer [152] thereby facilitates access of the substrate to the catalytic sites.
A carbon-supported ruthenium hybrid catalyst with a specific surface area of 1200 m 2 g −1 was employed for the direct hydrogenolytic cleavage of cellulose to sorbitol [153]. High microporosity and low acidity of the carbon support favour high dispersion of the metallic ruthenium. Interestingly, ball-milling of cellulose with carbon supported ruthenium provides enhanced conversions and selectivities to sorbitol [154,155]. Selective conversion of cellulose to sorbitol is achieved i.a. by use of bi-functional ruthenium catalysts supported on sulphated zirconia and sulphated silicazirconia [156]. Tetragonal zirconia, associated with generation of superacidity, is the active phase for cellulose depolymerisation that accompanies the hydrogenation function of ruthenium. Also, zeolite- [146,157] and silica- [158] supported ruthenium nanoparticles are suitable for the hydrogenation of glucose to the sugar alcohol sorbitol.

Ruthenium catalysts in lignin conversion
Hydrogenolysis of lignin involves reductive bond cleavage of C-O bonds linking the phenolic moieties, thereby generating hydrogenated and therefore less reactive monomeric species. For the reduction step, ruthenium catalysts are frequently employed. A variety of reducing agents have been suggested [141,160,161], such as hydrogen [142], carbon monoxide, formic acid (HCOOH/NEt 3 [53]), methanol, ethanol, isopropyl alcohol [54], acetonitrile, acetone. The energy needed for producing the reductant and the associated CO 2 -footprint ought to be taken into account when the lignin-derived products are utilised as biofuels [162]. Supercritical fluids as solvent have been claimed to produce fewer solid residues and provide higher biomass conversions [163,164]. Catalytic transfer hydrogenolysis of corn stover lignin in supercritical ethanol with a Ru/C catalyst yields bio-oil with a high fraction of monomeric moieties [163]. The key transformation is the reductive cleavage of ether linkages. Sequential extraction with a series of solvents differing in polarity results in monomer fractions that are enriched in alkylated phenols, guaiacols, syringols and hydrogenated hydroxycinnamic acid derivatives (Figure 10).
For using bio-oils as fuel, hydrotreating is necessary for lowering the oxygen content. Hydrotreating increases stability and energy density while decreasing the viscosity of the bio-oils. Ruthenium catalysts are often used in this hydrogenolytic upgrading of bio-oils. Even though zeolites are a good support material, substituted phenols cannot enter the micropores of typical zeolites. One concept for overcoming this challenge are catalysts comprising hierarchical pore systems. Thus, Ru supported on mesoporous ZSM-5 with a characteristic pore size of 4.5-4.7 Å of the MFI lattice channels (Figure 11) [92] and the mesopore system aligned to the b-axis was found to be effective for the hydrodeoxygenation of phenolic biomolecules [144]. For comparison, the Van-der-Waals radius of the syringol molecule is estimated to   (Figure 11) based on the distance of the outermost hydrogen atoms [165] and a Van-der-Waals radius for hydrogen of 1.04 Å [166]. Only at the channel entries do the open mesopores expose acid sites to the approach of bulky molecules necessary for catalysing the cleavage of the phenolic C-O bonds. This type of catalyst was found to effectively catalyse the hydrodeoxygenation of phenol and 2,6-dimethoxyphenol at 4.0 MPa H 2 -pressure and a temperature of 150 °C [144]. Conversions were > 99.5 and 97.5% after a 4 h reaction time, respectively; product selectivities to cyclohexane were accordingly 95.0 and 70.0%.

Conclusions on biomass transformation with ruthenium catalysts
Ruthenium, a late transition element, provides catalytic pathways that make it highly promising for catalysts applied for biomass conversion. Biomass, a globally available resource, is a sustainable feedstock for producing platform chemicals, that could substitute the current fossil-based platform chemicals in the chemical industry. However, in order to implement further processes in small and large-scale biorefineries, more efficient transformations will be required. Here, the distinct catalytic functions provided by ruthenium and ruthenium complexes could open new pathways. Biomass largely consists of complex molecules that comprise oxygen and other heteroatoms. Catalytic transformations need to accommodate for these heteroatoms, because molecules with heteroatoms tend to adsorb strongly to catalytic sites possibly causing substrate-or product inhibition. However, the preferential adsorption of chemical moieties associated with heteroatoms on the catalytically active site can be exploited for directing catalytic transformations. The principle has been explored for the consecutive hydrogenation of unsaturated moieties on a molecular assemble line. In this context, it is useful considering the concept of orthogonal catalytic functions, where a catalyst preferentially binds and transforms a selected chemical entity without hindering other catalysts that may be added for realising preceding or subsequent catalytic transformations.
In this context, the catalytically highly active element ruthenium embodies unique features. Ruthenium does not form binary hydrides that are stable under usual catalytic conditions. Nevertheless, metallic ruthenium can dissociate molecular hydrogen. Under an atmosphere of hydrogen, the surface of metallic ruthenium is covered with hydrogen atoms. Adsorption states and chemical reactivity of this hydrogen is well understood. Desorption of a fraction of the hydrogen provides the empty coordination sites necessary for co-adsorption of reactant molecules. Typically following a Langmuir-Hinshelwood-type mechanism, hydrogen atoms can be transferred to unsaturated moieties. Remarkably, ruthenium can also form and cleave C-C, C-O and C-N bonds. Combined with its strong propensity for hydrogenation, this ability gives rise to hydrogenation, hydrogenolysis and hydrodeoxygenation transformations that make ruthenium catalysts so interesting for biomass conversion. Noteworthy are the distinct catalytic transformations that can be realised with ruthenium catalysts. Selected examples for intriguing transformations of biomolecules and bio-derived molecules have been discussed above.
Understanding the interaction of adsorbed molecules with ruthenium surfaces, the nature of adsorption states, binding energies and structures of the adsorption complexes lies at the heart of rational design of catalysts that are specific for the conversion of the chosen chemical entity in biomass. It is anticipated that new transformations will be realised based on the unique catalytic functions provided by heterogeneous and homogeneous ruthenium catalysts. Serving as important tools for the synthetic chemist, these transformations will bolster the use of biomass as sustainable feedstock for the chemical industry.

Conflict of interest
There are no conflicts of interest to declare.

Notes
For compiling Figure 2, image 1961095 was obtained with the standard licence from Shutterstock.