Catalytic chemical processes for biomass conversion: Prospects for future biorefineries

Deoxygenation reactions

Platform chemicals can have disparate chemical compositions when compared to hydrocarbon end products, such as higher alcohols, demonstrating the need for profound catalytic transformations during feedstocks processing, as illustrated in Fig. 1. For example, lactic acid (2-hydroxypropanoic acid) and levulinic acid (4-oxopentanoic acid) are highly oxygenated compounds that require deoxygenation to be converted to advanced biofuels, or alternative value-added chemicals (e.g., methyltetrahydrofuran).

Fig. 1

Petroleum vs biomass processing to chemicals and fuels. Reproduced by permission of the Royal Society of Chemistry from [2].

Oxygen removal increases energy density while simultaneously decreasing chemical reactivity. This results in compounds with lower reactivity and more straightforward conversion of intermediates to end products. Deoxygenation of biomass feedstocks generally occurs via catalytic processes that remove H₂ O and/or CO_x species (CO and CO₂) such as hydrogenations, C-O hydrogenolysis, dehydrations and decarboxylations/decarbonylations. However, deoxygenation often requires the significant consumption of expensive and fossil-fuel derived hydrogen gas, which increases the CO₂ footprint of the bioprocess. In some cases, the external hydrogen consumption can be circumvented by using renewable H₂ sources (i.e., insitu generated hydrogen from formic acid decomposition under temperature) [7, 8].

To illustrate the utilization of renewable in-situ generated hydrogen, we have recently proposed a novel methodology for lignin deconstruction based on a formic acidmediated microwave assisted approach using metal nanoparticles supported on aluminosilicates [9–11]. The methodology involves the novel preparation of supported nanoparticle systems based on noble (Pd, Pt, Ru, and Rh) and transition metals (Ni and Cu) using a dry milling methodology to achieve highly accessible active sites preferentially located on the external surface of the aluminosilicate support [12]. Selected hydrogen donating solvents (e.g., tetralin, isopropanol, formic acid) were also utilized to maximize in-situ hydrogen production. The Nibased system was found to provide optimum depolymerization results in combination with formic acid, which was found to act as the hydrogen donating solvent (via decomposition under microwave irradiation to CO+CO₂ and H₂ [13, 14]), in addition to promoting acidolytic bond cleavage of lignin. In this way, several milligrams of simple aromatic compounds including mesitol, syringaldehyde, synringol and aspidinol (Scheme 1) were produced from organosolv lignin under mild reaction conditions (typically 30 min under microwave irradiation, under 150^oC).

The ability to efficiently process highly recalcitrant lignin fractions of various molecular weights to simple aromatic compounds by a simple hydrogenolytic methodology was also demonstrated (Fig. 2).

Fig. 2

Lignin purification and fractionation followed by heterogeneously catalyzed depolymerization to simple aromatics. Copyright Wiley-VCH Verlag GmbH & Co.KGaA. Reproduced with permission from reference [11].

Dehydration reactions are also highly relevant for biomass processing, particularly for the production of furanic compounds (HMF and furfural) from sugars with solid acid catalysts, namely zeolites, sulfonated materials, as well as ionic liquids [15, 16]. These catalyst compounds have also been reported for the valorization of (hemi)cellulosic feedstocks [15, 16] and mostly for monosaccharides (e.g., glucose, fructose, xylose and related sugars) [15–18].

Bronsted-acidic ionic liquids (ILs) have been developed for the deoxygenation of biorefinery-derived syrups containing a large content (over 40 %) of C5 oligomers via microwaveassisted dehydration [19]. Using pure xylose to evaluate the efficacy of the method, an 85 % conversion to furfural was achieved in 1 h with 1(4Sulfonylbutyl)pyridinium methanesulfonate, the most active IL examined (Scheme 2).

Scheme 2

Brönsted acidic RT-ILs investigated in C₅ dehydration to furfural. Reproduced by permission of the Royal Society of Chemistry from reference [19].

The reaction followed a simple dehydration of xylose to furfural catalyzed by the Brönsted-acidic IL. For the C5 oligomer-enriched biorefinery syrup, the conversion to furfural was surprisingly up to 45–60 % in <4 h under similar reaction conditions, albeit via a more complex reaction process than for xylose (Scheme 3). In the first step, the complex C5 oligomers undergo IL promoted hydrolysis to simple sugars (mostly xylose), followed by dehydration to furfural (Scheme 3). These ILs have been immobilized on activated carbonaceous materials to provide a highly stable and reusable support with enhanced hydrolysis/dehydration activity, in addition to related chemistries [20].

Scheme 3

Reaction scheme for the conversion of C₅ oligomers into furfural. Reproduced by permission of the Royal Society of Chemistry from reference [19].

For further information, readers are kindly referred to recent overviews on all these topics reported by leading authors in the field [15–18, 21–23].

C–C bond-forming reactions

The high oxygen content in biomass platform molecules (as compared to petroleum feedstocks) is not the only limitation to overcome when advanced fuel production technologies are envisaged. Platform molecules are typically derived from biomass sugars that are compounds with a maximum number of carbon atoms limited to 6 (derived from glucose). Consequently, if targeted products are liquid hydrocarbon fuels (e.g., C₅–C₁₂ for gasoline, C₉–C₁₆ for jet fuel, and C₁₀–C₂₀ for diesel applications), deoxygenation needs to be combined with additional reactions aimed to increase the molecular weight in the molecule (e.g., C–C coupling reactions) [24].

Well known C–C bond-forming reactions including aldol condensations of carbonyl compounds, catalytic ketonic decarboxylation or ketonization of carboxylic acids, oligomerization of alkenes, and alkylations of hydrocarbons are particularly useful for increasing the molecular weight and fine tuning the structure of biofuel products [25, 26].

Aldol condensation and ketonization represents two of the most relevant reactions for C–C coupling of biomass derivatives [25]. The former achieves effective and high-yield coupling between carbonyl-containing biomass intermediates under moderate reaction conditions. Aldol condensations have shown to be effective at generating C–C bonds from biomass-derived furanics (containing an aldehyde group) and ketones to produce diesel and jet fuel [27–29].

Ketonization processes involve condensation of two carboxylic acid molecules to produce a larger (2n-1) symmetric ketone [30]. This reaction possesses great potential for the catalytic upgrading of biomass since C–C coupling takes place with simultaneous oxygen removal (i.e., the reaction involves the removal of CO₂ and water) from carboxylic acids, the latter of which are common intermediates in biomass conversion processes [31, 32]. This reaction is typically catalyzed by inorganic oxides such as CeO₂, TiO₂, Al₂ O₃ and ZrO₂ at moderate temperatures (573–698 K, Scheme 4).

Scheme 4

Ketonization mechanism of carboxylic acids over ceria-manganese mixed oxide catalysts as proposed by Nagashima et al. [35].

Last, but not least, oligomerization and alkylation reactions are particularly useful to upgrade deoxygenated petroleum feedstocks (e.g., alkenes and hydrocarbons). Relevant work by the Dumesic and Corma groups have recently shown these chemistries can be successfully employed for advanced biofuels production, particularly in the latter stages of biomass conversion with bio-derived feedstocks having very low oxygen content [33, 34].

Cascade-type combined deoxygenation processes

An important advantage of utilizing catalytic processing for biomass conversion is the ability to design multifunctional catalysts to facilitate cascadetype reactions (e.g., hydrolysis/hydrogenation, dehydration/hydrogenolysis, etc.). The design of multi-step reactions has the additional possibility to translate optimized conditions into a more scalable continuous flow process in which different parameters can be controlled to maximize feedstock conversion.

Along these lines, our group has been recently focused on the development of a simple and efficient methodologies for the conversion of (hemi)celluloses and polysaccharides (starting from glucose as model compound) to value-added products including HMF, furfural and furan derivatives (Scheme 5) [36].

Scheme 5

Reproduced by permission of the Royal Society of Chemistry from reference [36].

Possible reaction products obtained from hydrogenation of HMF.

For this purpose, Cu-based aluminosilicates were utilized as catalysts under microwave irradiation and compared with analogous Pd systems and commercial catalysts as depicted in the pictorial representation in Fig. 3 [36]. The reactions involved in this methodology include a preliminary dehydration of sugars to furfural (C5 sugars) and HMF (glucose and fructose) followed by subsequent hydrogenation/hydrogenolysis steps to different products as depicted in Scheme 5. Rapid promoted chemistries lead to the formation of 5-methylfurfural (MF) which then is further hydrogenated to 5methylfurfuryl alcohol (MFA, major product obtained under the investigated conditions) and dimethylfuran (DMF) or 5methyltetrahydrofuran (MTHFA). Traces of products including angelicalactone (AL), MF-derived dimers and oligomers as well as other side products including dihydroxymethylfuran (DHMF) or hexadione (HD) were also observed under the investigated conditions (Scheme 5). The process generally involves relatively mild reaction conditions – reaction temperature typically under 180°C, reaction time on the order of 15–30 min, and the use of formic acid both as co-catalyst to promote the initial dehydration step of the sugar to the furan (HMF or furfural) as well as the hydrogen donating solvent (generating in-situ hydrogen for the hydrogenation/hydrogenolysis step in a similar way to that of lignin depolymerization [9–11]). Quantitative conversion could be obtained in the systems with a high selectivity to reduced products (>85 %) [36].

Fig. 3

Transformation of glucose and related biomass-derived polysaccharides into valuable furanics under Cu-mediated catalysis and microwave irradiation.

Recent research endeavors have also been focused on C4 and C5 organic acids as these constitute a promising family of chemicals from which relevant end products can be derived including polymers, solvents, flame retardants, food additives, cosmetics and related commodities [5, 6, 18, 21, 22].

Levulinic acid (LA) has been identified as a key platform chemical for the production of high value-added derivatives including 1,4-pentanediol (as intermediate for biopolymers [37]), γ valerolactone (as key compound for the production of biofuels and biopolymers [33, 38]) and 2-methyltetrahydrofuran (as a potential biofuel [39–41] and alternative green solvent for the pharmaceutical industry [42]).

Following our recent findings related to the use of formic acid as hydrogendonating media in heterogeneously-catalyzed lignin depolymerizations [9–11] and furanics production from sugars, we have proposed an innovative methodology and demonstrated the selective conversion of LA to MTHF in high yields, avoiding the use of molecular hydrogen in the system. This was made possible by the use of formic acid (FA) as hydrogen-donating solvent combined with a suitably designed nanocatalytic system. In this way, we have been able to mimic a plausible biomass scenario in which cellulose from forestry residues (e.g., lignocellulosics) are hydrolyzed to glucose and then to LA and FA via dehydration. This FA co-generated in the dehydration process could be then utilized in a similar way to that proposed in our approach for the sustainable and selective production of GVL, PDO, or MTHF using appropriately designed catalysts (Fig. 4).

Fig. 4

Proposed approach as compared to a theoretical biorefinery concept on cellulose valorization.

The final goal of the proposed concept will be the translation of optimized results into a more scalable continuous flow process currently under investigation in our laboratories. The use of designer Cu-based catalysts was particularly found to provide optimum results for MTHF production (>90 % yield) under microwave irradiation in a short period of time (typically 30 min reaction, Table 1). Two different type of materials were utilized in the reaction: Noble metal supported Starbons® [43] as compared to Cu-MINT materials (prepared using a one-pot microwave assisted approach) [44, 45]. Starbon® materials comprise a novel family of mesoporous carbonaceous materials derived from polysaccharides [43].

Catalytic processes involved in these processes comprise an initial hydrogenation/cyclization to gamma-valerolactone, that is then further converted to pentanediol via hydrogenation/ring opening, and then isomerization of pentanediol on acid sites leads to the formation of MTHF (Scheme 6). Other products available included angelicalactone (AL) from FA-catalyzed dehydration of LA as well as pentanoic acid (PA) and 4-hydroxyvaleric acid (HVA) from hydrogenation of LA.

Scheme 6

Reaction pathway for the production of MTHF from LA.

Deoxygenation of levulinic acid to advanced biofuels is another interesting catalytic process which has been developed in the absence of hydrogen by means of a pure thermal pyrolysis treatment. The process is denoted as thermal deoxygenation (TDO) and involves the formation and subsequent decomposition of calcium levulinate at temperatures between 350 and 450 °C in an inert atmosphere [46] (Scheme 7). Calcium levulinate simultaneously condenses via ketonization and deoxygenates (by internal cyclation and dehydration) leading to the production of a broad product distribution of cyclic and aromatic compounds with very low oxygen content. Production of aromatics, which are valuable components of gasoline and jet fuels, can be maximized by operating at higher temperatures. The addition of equimolar amounts of calcium formate as an in situ hydrogen source provided improved results due to the promotion of TDO processes and the formation of petroleumlike oils that could be further processed in existing refinery facilities [47].

Scheme 7

Main catalytic routes for the conversion of levulinic acid into advanced biofuels.

An alternative MTHF formation mechanism involves the selective conversion of GVL into pentanoic acid (PA) by means of ring-opening (on acid sites) and hydrogenation (on metal sites) reactions employing bifunctional heterogeneous catalysts at moderate temperatures and pressures (Scheme 7). Lange and co-workers have recently exploited this route to continuously produce the so-called valeric biofuels (i.e., alkyl valerates), a novel family of lignocellulosic biofuels with excellent energy density, polarity and volatility ignition properties to be used in conventional engines without any modifications [48]. Produced fuels can be adapted to fit in both gasoline and diesel engines by varying the alkyl chain length. However, such technology needs an external source of alcohol for esterification. Excellent yields of PA (92 %) could be achieved from aqueous solutions (e.g., 50 wt %) of LVA over Pd/Nb₂ O₅ operating in flow mode [49]. Remarkably, PA could be also ketonized to 5-nonanone over the niobic support with yields up to 70 % when the space velocity in the flow reactor was reduced to the level of WHSV = 0.1 h⁻¹. This yield is however limited by the large number of processes involved which prevent a proper control of intermediate steps (LVA to PA to 5nonanone, Scheme 7) [49, 50]. A proposed solution to overcome this limitation involves the utilization of double-bed flow reactors in which conditions of each bed are independently adjusted to maximize yields of intermediate steps.

For example, a double-bed configuration of Pd/Nb₂ O₅ (at 275 °C) + Ce_0.5 Zr_0.5 O₂ (at 425 °C) allows aqueous solutions of GVL to be transformed into 5-nonanone with yields of 90 % [50]. In this configuration, the first bed achieves conversion of PA to GVL that is subsequently converted into 5nonanone in the second bed in excellent final yields. Such C₉ ketones can serve as platform molecules for the production of hydrocarbon fuels with molecular weight and structures appropriate for gasoline, diesel and jet fuels applications [26].

Similar dual continuous flow reactor approaches have been designed to upgrade aqueous solutions of GVL into jet fuels through the formation of C₄ alkenes (Scheme 7) [33]. The first flow reactor is loaded with acidic silica-alumina that achieves decarboxylation of GVL to butenes that are subsequently oligomerized to a distribution of alkenes centered at C₁₂ using an acidic Amberlyst catalyst. In this case, the double bed configuration is not feasible since water has to be removed before entering the second reactor to achieve an effective oligomerization of butenes. The flow process is simple and achieves GVL deoxygenation and subsequent C−C coupling in a clean (CO₂ can be efficiently sequestrated at the system pressure) and cheap (no external hydrogen required) fashion.