Several works have been addressed to the production of DMF directly from biomass-derived hexoses, especially fructose. In this sense, the breakthrough of deriving DMF in good yields (76–79%) from fructose was first reported by Roman-Leshkov et al. (2007) in a two-step process. The process consisted of the production of 5-HMF starting from fructose (in an acid-catalyzed biphasic reactor to promote the simultaneous dehydration of fructose to 5-HMF and the solvent-extraction of the produced 5-HMF) followed by hydrogenation over a carbon-supported copper/ruthenium (Cu-Ru/C) catalyst using butanol as a solvent. From this pioneering work, similar works have been described in the literature using different solvents and catalysts. Recently, Upare et al. (2015) proposed an integrated process for the production of DMF from fructose with an outstanding yield of 92% in a two-step heterogeneous process in which fructose is firstly dehydrated to 5-HMF using Amberlyst-15 in butanol, and the 5-HMF in the resulting mixture is then converted to DMF by vapor-phase hydrogenolysis over an Ru-Sn/ZnO catalyst.

Another interesting approach to transform fructose into DMF in a one-pot process has been reported by Thananatthanachon and Rauchfuss (2010). The process is based on the use of formic acid, which can also be conveniently produced by biomass degradation, both as homogeneous acid catalyst for the dehydration of fructose into 5-HMF and hydrogen source to turn 5-HMF into the intermediate 2,5-dihydroxymethylfurfural (DHMF); finally acting again as a catalyst for the DMFH deoxygenation to give DMF. An overall yield of 51% was achieved, which is unsatisfactory for industrial applications. Other approaches based on hydrogen donors have been described in the literature but the DMF yields have been reduced.

Considering the use of glucose instead of fructose, a two-step approach for its conversion into DMF was also attempted in ionic liquids (ILs) in combination with acid catalysts (Chidambaram and Bell, 2010). This process involves the dehydration of glucose to 5-HMF in [EMIM]Cl and acetonitrile using 12-molybdophosphoric acid (12-MPA) as catalyst, followed by the subsequent conversion of 5-HMF into DMF over Pd/C in a one-pot method. However, such a system provides low 5-HMF conversion and poor DMF selectivity (13%), needing very high hydrogen pressures (62 bars) owing to its low solubility in ILs.

A more attractive reaction pathway is the one-pot combination of the dehydration of a cheap and renewable source, such as fructose, to 5-HMF followed by its etherification into EMF using ethanol as solvent and a heterogeneous catalyst. Both transformations are driven by acid catalysis, being feasible to optimize the selectivity toward EMF through the proper selection of the solid acid catalyst and the reaction conditions. In a pioneering work, Brown et al. (1982) evaluated the preparation of ethers of 5-HMF, together with 5-HMF itself and alkyl levulinates, from fructose using ion-exchange resins in nonaqueous solvents. However, both the selectivity to EMF and the reaction rates were low. More recently, different catalytic systems have been used in a similar way (Balakrishnan et al., 2012; Liu et al., 2012; Kraus and Guney, 2012; Li et al., 2014a; Yuan et al., 2015).

Although fructose has been mainly explored, it is clear that the large-scale sustainable use of EMF will require cellulosic biomass, ie, glucose, as the feedstock. However, catalytic studies report that no 5-HMF or EMF were detected, and the glucose instead reacts with ethanol to produce ethyl glucopyranoside (EDGP). Brönsted acid catalysts on their own appear not to be able to drive the transformation of glucose into EMF. Consequently, to produce EMF from glucose, isomerization of glucose into fructose appears as a particularly relevant reaction. Lew et al. (2012) have reported the one-pot synthesis of EMF from glucose using a combination of Sn-BEA (for the isomerization step) and Amberlyst 131 (for the dehydration-etherification steps) with acceptable yields.

However, the above-discussed routes would necessitate a source of high-purity levulinic acid (nontrivial to obtain due to contamination by polymeric humins). Therefore, it would be more advantageous if levulinates could be obtained in high yields from C6 carbohydrate-based biomass, ie, fructose, glucose, sucrose, cellulose and the like, in alcohol medium without the need for first isolating levulinic acid. As an additional advantage, the use of a nonaqueous alcoholic reaction medium for the treatment of lignocellulosic biomass minimizes the formation of undesired byproducts, since polymeric humins formation greatly diminishes in alcohol. Although homogeneous catalytic systems, such as sulfuric acid, have been employed, the research interest is currently focused on heterogeneous catalysts that can avoid the known problems of homogeneous acids as well as improve the selectivity to EL. Thus, Peng et al. (2011) reported the production of EL from glucose using different sulfated catalysts, including SO₄/ZrO₂, SO₄/TiO₂, SO₄/ZrO₂-TiO₂, and SO₄/ZrO₂-Al₂O₃. They showed that different components of the sulfated metal oxides have markedly different catalytic effects on the ethanolysis of glucose. An optimized EL yield of above 30 mol% was obtained for sulfated zirconia at 200°C, leading to the side formation of diethyl ether (coming from the autoetherification of ethanol). On the other hand, SO₃H-SBA-15 catalyst has shown a better catalytic activity than zeolites and sulfated zirconia catalysts for the conversion of fructose to ethyl levulinate (Saravanamurugan and Riisager, 2012). However, this type of Brönsted acid catalyst cannot drive the direct conversion of glucose to EL, instead leading to ethyl glucopyranoside (EDGP). This evidences that the activity of sulfated zirconia reported by Peng et al. (2011) for glucose should be attributed not only to the presence of sulfate groups but also of zirconium oxide, which is able to isomerize glucose into fructose. In the same line, a work by Tominaga et al. (2011) has showed that mixed-acid systems consisting of both Lewis (metal triflates such as In(OTf)₃) and Brönsted (sulfonic) acids is an efficient combination for the direct synthesis of methyl levulinate from cellulose. In this case, the reaction proceeds in two steps; cellulose is first solvolyzed to sugars, which are readily converted to methyl levulinate. The former step is mainly catalyzed by sulfonic acids, and the latter by the metal triflates.

Current efforts are focused on the preparation of bifunctional catalysts capable of directing the Lewis acid-catalyzed isomerization of alkyl glucoside intermediates to alkyl fructosides, and their subsequent Brönsted acid-catalyzed dehydration to 5-alkoxymethylfurfural and esterification to form alkyl levulinate and formate (Fig. 13.2).

The correct balance of Lewis and Brönsted acid sites is critical to the success of this complex tandem transformation. In this context, we have recently reported the preparation of conformal sulfated zirconia (SZ) monolayers throughout an SBA-15 architecture that confers efficient acid-catalyzed one-pot conversion of glucose to ethyl levulinate (Morales et al., 2014). In this work, we have shown that conformal SZ monolayers with tunable surface acid strength and site density can be dispersed over a mesoporous SBA-15 framework through a simple wet chemical grafting/hydrolysis protocol. A bilayer SZ/SBA-15 material exhibits the maximum surface acidity and balance of Lewis:Brönsted sites, and exhibits good performance in the one-pot conversion of glucose to alkyl levulinates under mild conditions. A clear advantage of these catalysts is the capability of reaching relatively high EL yields under moderate temperatures, while avoiding ethanol losses as diethyl ether. Furthermore, the sulfated ZrO₂ monolayers grafted on SBA-15 are highly stable, overcoming the extended leaching problems of commercial sulfated zirconias.

However, as previously discussed for alkyl levulinates, this approach would necessitate a source of high-purity levulinic acid. Therefore, research studies have lately focused on obtaining GVL from carbohydrates in a one-pot approach through the combined action of acid and hydrogenation catalysts, avoiding the need for isolating LA. Heeres et al. (2009) reported a one-step strategy for producing γ-valerolactone directly from glucose and fructose, as well as sucrose and cellobiose, by combining a homogeneous acid catalyst (trifluoroacetic acid) with a heterogeneous hydrogenation catalyst (Ru/C). Reactions were performed in water at 180°C in the presence of hydrogen or formic acid as the hydrogen source. However, this one-pot approach still faces a serious problem of deactivation of the hydrogenation metal catalyst due to the presence of strong acids in the reaction medium. This deactivation is more severe when levulinic acid comes from the sulfuric acid-catalyzed deconstruction of cellulose. In order to increase the stability of the catalyst, the addition of Re as an alloy with Ru improved the catalyst stability in the presence of sulfuric acid, although the TOF was comparatively low. Another way to minimize this problem is the use of different systems for extracting levulinic acid from the acid aqueous solutions. For instance, alkylphenol solvents are able to extract up to 80% of levulinic acid from these aqueous feedstocks (Sen et al., 2012a). Also, the reactive extraction with different alcohols or with olefins; to produce levulinate esters which can be easily separated from the aqueous feedstock has been proposed (Gurbuz et al., 2011). These extracting systems allow obtaining levulinic acid (or an ester thereof) pure enough to be hydrogenated to GVL and the recycling of the mineral acid. Going a step further, the use of two heterogeneous catalysts can avoid the deactivation associated with mineral acids. Thus, the direct catalytic conversion of cellulose to levulinic acid (LA) by niobium-based solid acids and further upgrading to γ-valerolactone (GVL) on a Ru/C catalyst were performed through sequential reactions in the same reactor (Ding et al., 2014). Firstly, using aluminum-modified mesoporous niobium phosphate as a catalyst, cellulose can be directly converted to LA with as high as a 52.9% yield in aqueous solution, even in the presence of the Ru/C catalyst. Then, after replacing N₂ with H₂, the generated LA in the reaction mixture can be directly converted to γ-valerolactone through hydrogenation over the Ru/C catalyst without further separation of LA.

As a recent attractive alternative to the reduction of LA to GVL using H₂ or formic acid over metal catalysts, the group of Dumesic has explored the reduction of LA by catalytic transfer hydrogenation (CTH) through the Meerwein–Ponndorf–Verley (MPV) reaction. In this reaction, a sacrificing secondary alcohol such as 2-butanol is used as a hydrogen source (Chia and Dumesic, 2011; Assary et al., 2013). This approach offers important advantages, such as an increased chemoselectivity for the reduction of carbonyl groups under milder reaction conditions, even in the presence of other functional groups; or the fact that the MPV reaction does not require precious metal heterogeneous catalysts. Moreover, the MPV hydrogen donor, usually 2-propanol, can be recycled after hydrogenation over base-metal catalysts such as nickel or copper, or even sold as a commodity chemical in its oxidized form (ie, ketones). Dumesic's group demonstrate in their work that CTH via MPV reaction is a viable means for the hydrogenation of LA and its esters (levulinates) over inexpensive, metal oxide heterogeneous catalysts that are easily recovered (ZrO₂ was demonstrated to be a highly active material), with attainment of close to quantitative yields of GVL under appropriate reaction conditions. In the same direction, Román-Leshkov's group (Luo et al., 2014) have investigated the reaction kinetics of the MPV reduction of methyl levulinate (ML) to 4-hydroxypentanoates and subsequent lactonization to GVL catalyzed by Lewis acid Beta zeolites (modified with Hf, Ti, Zr, and Sn metal species). All catalysts generated GVL with selectivities >97%, with Hf-Beta exhibiting the highest activity. Likewise, Geboers et al. (2014) have demonstrated the feasibility of the CTH approach to convert levulinates into hydroxypentonaotes and GVL, using Raney Ni as catalyst and 2-propanol as H-donor and solvent.

On the other hand, GVL can also be used as an intermediate for the production of valeric biofuels. The process consists of the catalytic hydrogenation of GVL to valeric acid (VA) over hydrogenation metal catalysts and subsequent acid-catalyzed esterification to alkyl (mono/di) valerate esters. Lange et al. (2010) showed that although GVL is a relatively stable product under hydrogenation conditions, it can be hydrogenated to VA in the presence of bifunctional catalysts containing both hydrogenation and acidic functions. Using a continuous high-pressure plug-flow reactor they identified Pt-loaded SiO₂-bound H-ZSM-5 as a very effective catalyst for the production of VA, though other zeolites and hydrogenation metals also gave good yields. The process can be intensified by converting LA to ethyl valerate (EV), the most interesting valeric biofuel, in a single step (going through the intermediate sequential production of GVL and VA). Thus, cofeeding ethanol with LA as a physical or chemical mixture (in the form of ethyl levulinate) over a Pt-modified zeolite-based catalyst leads to the efficient coproduction of VA and EV. More recently, Chan-Thaw et al. (2013) have proposed the production of EV and pentyl valerate (PV) in a one-pot one-step reaction from GVL. The bifunctional catalyst used consisted of Cu supported on an amorphous weakly acidic material, therefore representing an interesting alternative to Pt/zeolite catalysts.

Furfuryl alcohol is an important chemical intermediate in the transformation of furfural into energy-dense oxygenated biofuels. It is one of the most common products in the hydrogenation of furfural. It has been estimated that 62% of the furfural produced globally each year is converted into FA (Yan et al., 2014). The hydrogenation of furfural to FA is relatively easy to achieve and become more mature over the last several decades development. However, at present, furfuryl alcohol may only be produced from xylose by a two-step process based on catalysts with different features and operating at different conditions. Recently, Perez and Fraga (2014), studied the one-pot production of furfuryl alcohol via xylose dehydration followed by furfural hydrogenation over a dual catalyst system composed of Pt/SiO₂ and sulfated ZrO₂ as metal and acid catalysts, respectively. They found that the presence of both acid and metal sites is compulsory in order to promote both reaction steps. Likewise, selectivity toward furfuryl alcohol is strongly dependent on the solvent, which can inhibit its polymerization to some extent.

The production of MF (sylvan) by furfural hydrogenation through the FA as an intermediate has been reported over various supported noble metal and bimetallic catalysts. Cu-based catalysts such as Raney-Cu, Cu/alumina, and carbon-supported Cu chromite have shown selective conversion of furfural to MF but often operated at high temperature and low pressure. However, catalyst deactivation is an important drawback of using those catalysts. Hence, increasing the catalyst stability and the development of an effective regeneration procedure is required (Yan et al., 2014). Some authors have recently performed the hydrogenation of furfural in different solvents under milder reaction conditions using supported Pd complex, achieving 100% yield of MF after 1 h of reaction (Climent et al., 2014).

The conversion of furfural into MTHF is achieved by a multistep process which includes hydrogenation–deoxygenation of furfural to MF and further hydrogenation of MF to MTHF in separated reaction systems. High H₂ pressure is needed and the multistep conversion requires multicomponent catalysts, at least two different reactors, and the isolation systems for the intermediates. This limits the large-scale production of MTHF. There are few reports dealing with the one-step conversion of furfural to MTHF. Recently, one-step direct conversion of furfural to MTHF was carried out under atmospheric pressure over a dual solid catalyst based on two-stage-packed Cu–Pd in a reactor. This strategy provides a successive hydrogenation process, which avoids high H₂ pressure, uses the reactor efficiently, and eliminates the product-separation step. Therefore, it could enhance the overall efficiency because of low cost and high yield of MTHF (97.1%) (Dong et al., 2015).

Finally, tetrahydrofuran (THF) can be obtained by decarbonylation of the carbonyl group of furfural under reductive conditions using Pd-based catalysts followed by hydrogenation of furan formed in the presence of a variety of metal catalysts (Sitthisa and Resasco, 2011).

Etherification of furfuryl alcohol is performed in the presence of strong acid catalysts in an alcoholic medium. In this transformation, it is important to control the reaction time and conditions. Severe reaction conditions (increasing of catalyst acidity and temperature) leads to heavy byproducts production coming from condensation reactions of furfuryl alcohol. Likewise, although alkyl levulinates can be easily obtained by direct esterification of levulinic acid with the corresponding alkyl alcohol in the presence of an acid catalyst (see Section 13.3.1), they can be advantageously prepared from furfuryl alcohol, without the need for isolating levulinic acid. The reaction mechanism involves the quickly etherification of furfuryl alcohol into the alkyl-furfuryl ether followed by a slowly transformation into alkyl levulinate. It is important to take into account that the transformation also involves the formation of byproducts such as humins and ethers (dialkyl ethers from the alcohol used as reaction media). Lange et al. (2009) reported a decrease in catalytic activity in the following order: H₂SO₄ > macroreticular resins > gel resins > zeolites, finding that the two major factors that settle the activity of catalysts are the acid strength of active sites and their accessibility.

Another approach is the esterification of furfuryl alcohol in the presence of carboxylic acids. Particularly, a process that can be considered a promising route is the direct production of furfuryl acetate in a one-step hydrogenation-esterification of furfural with acetic acid. Under moderate reaction conditions, a bifunctional catalytic system composed of an acid functionality for esterification reaction and a hydrogenating functionality (5% Pd/Al₂ (SiO₃)₃ and 5% Pd/Al-SBA-15) has been proposed with good selectivity to the desired products (Yu et al., 2011a,b).

Recently, Liu and Chen (2014) have developed an integrated catalytic process for the conversion of 5-HMF into alkane fuels. The integrated catalytic process involves three different steps: (1) 5-HMF production from fructose and glucose; (2) self-coupling of 5-HMF catalyzed by n-heterocyclic carbine (NHC) to yield furoin intermediates; and (3) linear alkanes production by hydrodeoxygenation using metal-acid tandem as catalysts system consisting of Pd/C + La(OTf)₃ + acetic acid. Alkanes were produced in 78% yield with a 64% selectivity to n-C₁₂H₂₆.

Opposite to 5-HMF, high yields of single and double condensation products are achieved in the aldol-condensation of furfural with acetone in the presence of an aqueous phase with NaOH catalyst. Mixed oxides with different basic strength have also been used for this reaction, getting more activity with those catalysts with higher concentration of strength basic sites, ie, Mg-Zr > Mg-Al > Ca-Zr (Faba et al., 2012). Moreover, the basic site distribution can be improved supporting the Mg-Zr mixed oxide on mesoporous carbons which leads to a higher interaction of the reactants with the carbon surface achieving 96% conversion of furfural with 88% selectivity for C₁₃ and C₈ adducts (Faba et al., 2013).

More interesting is the sequential strategy developed by the Dumesic group. Similar to 5-HMF, the production of C₁₀ alkanes is carried out by a cascade reaction aldol-condensation of furfural with acetone followed by hydrogenation/dehydration using a bifunctional catalyst Pd/MgO-ZrO₂ with high overall carbon yield (>80%) (Barret et al., 2006). Another sequential process of two consecutive steps to obtain a mixture of long-chain alkanes with excellent properties as a diesel fuel (cetane number and flow properties at low temperature) was recently reported by Corma et al. (2012) named the “Sylvan process.” The first step consists of a hydroxyalkylation/alkylation of three molecules of 2-methylfuran (sylvan) or hydroxylation of sylvan with aldehydes or ketones catalyzed by organic and inorganic acids to yield oxygenated intermediate molecules (butanal is considered to be the most promising molecular linker). In the second step, a complete hydrodeoxygenation of the previous products catalyzed by platinum metal supported on nonacidic materials leads to the desired mixture of alkanes within the diesel range namely 6-alkylundecane.

More recently, different solid acid catalysts were studied for the alkylation of MF with mesityl oxide (Li et al., 2014b). Among the investigated candidates, Nafion-212 resin exhibited the highest catalytic efficiency, which can be explained by its higher acid strength. For the second step of hydrodeoxygenation, Ni–Mo₂C/SiO₂ exhibited an evident advantage at the cost and the selectivity to diesel range alkanes (77% yield).