Efficient utilization of renewable feedstocks: the role of catalysis and process design

Abstract

Renewable carbon feedstocks such as biomass and CO₂ present an important element of future circular economy. Especially biomass as highly functionalized feedstock provides manifold opportunities for the transformation into attractive platform chemicals. However, this change of the resources requires a paradigm shift in refinery design. Fossil feedstocks are processed in gas phase at elevated temperature. In contrast, biorefineries are based on processes in polar solvents at moderate conditions to selectively deoxygenate the polar, often thermally instable and high-boiling molecules. Here, challenges of catalytic deoxygenation, novel strategies for separation and opportunities provided at the interface to biotechnology are discussed in form of showcases.

This article is part of a discussion meeting issue ‘Providing sustainable catalytic solutions for a rapidly changing world’.

1. Introduction

Driven by the desire for an entirely renewable energy supply, the share of wind, water and solar power is steadily increasing. Despite these technologies for the production of ‘green’ electrons, today's infrastructure relies mainly on fossil resources [1,2]. The latter provide chemical feedstocks and liquid fuels as energy carriers of high volumetric and gravimetric energy density, respectively. Aiming for full sustainable energy and feedstock supply, renewable carbon sources are indispensable. In this regard, biomass as a renewable carbon source has attracted intensive attention in recent years. Plants harvest CO₂ during their growth which is again released upon combustion of biomass-based products such as biofuels. Therefore, biomass as feedstocks bears the potential to achieve overall CO₂-neutral value chains and products. Also a direct use of CO₂ as a renewable feedstock is possible, but remains challenging for others than point sources such as exhausts of power plants and steel or concrete production. Alternatively, Nature's ability to concentrate low CO₂ levels in the atmosphere in plant material may be used. The public discussion on food versus fuel has led to awareness that a careful analysis of suitable biomass sources is mandatory for a sustainable industrial utilization of biomass feedstocks. This relates not only to the CO₂ footprint but also to societal and environmental aspects such as food production, land use, biodiversity and water supply.

Lignocellulose comprehends 70% of all biomass annually produced by land (170–200 × 10⁹ tons yr⁻¹) [3] is non-edible and presents the major constituents of agricultural waste streams and food industry. In contrast to fossil feedstocks including natural gas, crude oil and coal, lignocellulose is composed of highly functionalized building blocks. This results in a significantly higher oxygen content of biomass feedstocks and several biomass-based products when compared with hydrocarbons derived from fossil feedstocks [4]. The major constituents of lignocellulose are cellulose, hemicelluloses and lignin. Cellulose is a linear polymer of glucose, whereas hemicelluloses are branched polymers composed of different C₅ and C₆ monosaccharides such as xylose, arabinose, galactose and mannose. Lignin presents a polymer of randomly polymerized p-coumaryl-, coniferyl and synapyl alcohols as major building blocks. The chemical nature of these natural resources entails the need for suitable process technologies. Indeed, current refineries are based on the functionalization of non-polar compounds in gas-phase processes at elevated temperatures. By contrast, biorefineries mainly comprehend liquid-phase processes of high-boiling and highly functionalized biopolymers and their monomers at moderate temperatures and in polar solvents such as water. Alternatively, biomass can be converted into synthesis gas and subsequently introduced into chemical value chains via established processes covering methanol synthesis, methanol-to-olefins and Fischer–Tropsch technology. Although technologically attractive due to the high technology readiness level of the involved processes, biomass valorization via synthesis gas as a building block does by no means make use of the chemical synthesis provided by nature. Indeed, several promising platform molecules available based on the carbohydrate fraction of lignocellulose have been identified and are intensively discussed in the literature with regard to their synthesis and follow-up chemistry [4–8].

In this opinion piece, selected challenges associated with the selective transformation of renewable feedstocks into these platform chemicals and their further processing will be discussed. The showcases are considered based on the authors' expertise and not meant to provide a comprehensive overview of the field. Instead, research interfaces are highlighted offering vast opportunities for fruitful interdisciplinary collaboration across various research fields. Figure 1 illustrates the selected challenges accompanying the design of biorefinery processes from the prospective of heterogeneous catalysis and material design.

Figure 1.

Selected challenges in the biomass valorization from the perspective of heterogeneous catalysis and material design. (Online version in colour.)

2. Discussion

(a). Catalytic defunctionalization

The high-oxygen content of (poly)saccharides together with the need for their efficient and selective transformations in aqueous reaction systems has promoted intensive research in recent years. Especially, dehydration and hydrodeoxygenation or hydrogenolysis were considered but also decarbonylation and deoxydehydration found increasing attention (figure 2) [9].

Figure 2.

Products of catalytic defunctionalization of sorbitol.

Focusing on the developments in hydrodeoxygenation, often also named hydrogenolysis, sugar alcohols were considered as promising feedstocks for production of highly valuable products, such as ethylene glycol. In 2014, the global demand for ethylene glycol reached 25 000 ktons yr⁻¹, with an estimate market growth of 5% each year. More than 58% of ethylene glycol is used in the polymer industry for manufacturing poly(ethylene terephthalate) (PET) as fibres, containers and packaging films. Nowadays, ethylene glycol is industrially produced mainly based on fossil feedstocks, though some biomass-based manufacturing methods have already been commercialized. For example, more than 650 ktons yr⁻¹ of ethylene glycol are produced based on bioethanol. Hydrogenolysis of saccharides and sorbitol has also been industrialized with a market capacity of more than 200 ktons yr⁻¹ [10]. Cutting-edge contributions illustrated the possibility to transform cellulose and even raw biomass directly into polyols [11–14]. However, a selective transformation into individual target products remained challenging, and mixtures of sugar alcohols were formed. In addition, undesired dehydration to anhydro sugar alcohols such as sorbitan and isosorbide as well as decarbonylation and C–C cleavage via retro aldol reactions occurred. Analyses of the complex reaction networks [15–18] and careful design of supported metal catalysts enabled distinct advances in the selective formation of, e.g., ethylene and propylene glycol based on sorbitol or cellulose [19]. Li et al. [20] even demonstrated ethylene and propylene glycol formation from raw biomass using cellulose as a feedstock. In particular, a careful balance of acidic or basic sites and metal sites proved essential [21–23]. Nevertheless, if molecular acids or bases are used, neutralization of the product solution is necessary and associated with salts formed as by-products. Few studies emphasized the possibility to use solid acids and bases in combination with metal catalysts [24–26].

The need for careful adjustment of the nature and relative ratio of active sites presents a recurrent element in catalysis with special pronounced importance for the selective hydrodeoxygenation and ring-opening hydrogenation of biogenic compounds (figure 3). Starting from furfural or 5-hydroxymethylfurfural (HMF) tailored catalysts facilitated direct access to 1,5-pentanediol and 1,6-hexanediol [27]. These diols represent α,ω-diols, which are in great demand for production of polyesters, elastic fibres and polyurethanes. The current market volume of 1,6-hexanediol is approximately 138 ktons yr⁻¹. 1,5-Pentanediol is produced with a much smaller capacity of 3 ktons annually due to limited readily accessible C₅ petroleum feedstocks [28]. However, an increase of production can be envisaged upon shift towards renewable feedstocks. While first catalyst concepts were based on, e.g., Pt/CoAl₂O₃ allowing 35% yield of 1,5-pentanediol from furfural [27], subsequent contributions elucidated the potential of bi- and multimetallic systems. In this regard, Pd-doped Ir–ReO_x/SiO₂ enabled 1,5-pentanediol yields above 71% [29]. Tuteja et al. [30] applied palladium supported on zirconium phosphate in the ring-opening hydrogenation of HMF to 1,6-hexanediol with up to 43% yield. Again, a combination of metal catalysts, Pd/SiO₂ and Ir–ReO_x/SiO₂, proved superior facilitating both hydrogenation and selective ring-opening to yield up to 57.8% 1,6-hexanediol [31]. Interestingly, a recent study presented a comprehensive process concept for highly efficient 1,5-pentanediol production from furfural again using rather simple catalysts such as alumina and Ru/C relying on an elaborated control of the individual reaction sequences [28]. Also access to longer alcohols became feasible, proceeding via aldol condensation of furfural with acetone followed by subsequent hydrogenation and ring-opening hydrogenation over ruthenium supported on activated carbon together with an acidic ionic liquid yielding up to 93% 1-octanol/1,1′-dioctyl ether [32].

Figure 3.

Examples of selective hydrodeoxygenation and ring-opening hydrogenation of biogenic compounds.

Driving hydrodeoxygenation even further, fully deoxygenated products are interesting targets. In 2005, Huber and co-workers demonstrated the production of C₇–C₁₅ alkanes from carbohydrates proceeding via dehydration to HMF, aldol condensation with acetone and hydrodeoxygenation to the corresponding alkanes [33,34]. Aiming for branched alkanes, the group of Corma evidenced efficient catalytic pathways with 2-methyltetrahydrofuran as a substrate, available via hydrogenation of levulinic acid [35,36]. Also, the transformation of sorbitol and cellulose into hexane could be achieved. Supported platinum catalysts were suitable [37,38], though Ir–ReO_x/SiO₂ combined with HZSM-5 allowed a direct transformation of cellulose to n-hexane with yields of 83% and 78% for ball-milled and microcrystalline cellulose, respectively [39]. Likewise, tungstosilicic acid with Ru/C facilitated the hydrodeoxygenation of microcrystalline cellulose to 82% n-decane-soluble products, mainly n-hexane [40]. Interestingly, the dominant pathway to liquid alkanes proceeded via HMF, whereas the route via sorbitol appeared to be less efficient, emphasizing again the major importance of a deep understanding of the involved reaction mechanisms for targeted catalyst and process design.

(b). Novel separation strategies

Despite the vast amount of studies addressing selective transformations of biomass-based platform chemicals, the challenges associated with their separation find less attention [41]. In recent years, HMF has been highlighted as a very promising biomass-derived platform molecule due to its low-oxygen content and a straightforward synthetic procedure. Since 2013, AVA Biochem BSL AG has produced HMF commercially with an annual capacity of 20 tons [42]. HMF can be synthesized directly from carbohydrates obtained after depolymerization of cellulose and isomerization of glucose into fructose (figure 4). HMF is readily formed in aqueous media via dehydration in the presence of simple acidic catalysts, such as H₂SO₄ or HCl [43]. The preparation and separation of HMF depicts some of the challenges associated with the production of polar and high-boiling platform chemicals. First of all, the selective aqueous-phase synthesis of HMF is hampered by undesired consecutive reactions. From the viewpoint of reaction engineering, this presents an all acid-catalysed reaction sequence: fructose → HMF → levulinic acid, where the intermediate presents the target product. In line, in situ extraction of HMF is attractive suppressing a subsequent dehydration to levulinic acid and separating HMF from the aqueous reaction environment. Several studies followed this approach and emphasized the potential of reactive extraction to enhance the HMF yield [44–46]. Nevertheless, it has to be mentioned that the partition coefficient of HMF from water into most organic solvents is only around 1 [45].

Figure 4.

Scheme of HMF synthesis based on cellulosic feedstocks. The reactions in the box depict the processes which take place during the HMF formation in aqueous media in the presence of an acidic catalyst. Two possibilities of HMF reactive removal include extraction and adsorption. (Online version in colour.)

An alternative separation can be provided by selective adsorption. Activated carbons showed potential as adsorbents, but do not possess sufficient selectivity for HMF compared with sugars [47–51]. Surprisingly, very hydrophobic hyper-cross-linked polymers (HCPs) enable a very selective adsorption of HMF from aqueous solutions containing mixtures of fructose and HMF [52]. The capacity is related to the high specific surface area of HCPs, while hydrophobicity appears to determine selectivity for HMF adsorption rather than fructose. In the system water–fructose–HMF, the latter exhibits the least hydrophilicity and is consequently preferentially adsorbed. As demonstrated in the following section, the same strategies could also be applied for itaconic acid, a platform chemical derived from fermentation with glucose as a substrate [53].

Along with HMF production, an efficient isomerization of glucose into fructose (the second reaction in figure 4) poses a major challenge in valorization of cellulosic feedstocks. As the yield of fructose is thermodynamically restricted, isomerization naturally gives rise to a mixture of both isomers—glucose and fructose---that are present in approximately 1 : 1 ratio. The current production of fructose relies on the fermentative isomerization followed by chromatographic separation of the mixture to produce high-fructose syrups (HFSs) for the food industry [54]. Noteworthy, more than 10 000 ktons of HFSs is produced annually on an industrial scale [55]. Chemo-catalysts potentially outperform biocatalysts in terms of thermal stability, robustness and requirements for purity of substrates. In this regard, numerous investigations have focused on developing a suitable chemo-catalyst for selective isomerization of glucose into fructose [55,56]. We have recently proposed a method to produce fructose using a combination of catalysis and anionic extraction, as shown in figure 5 [57]. The first step presents the aqueous-phase isomerization of glucose into fructose in the presence of NaH₂PO₄ + Na₂HPO₄ as a catalyst. In the next step, fructose is separated from NaH₂PO₄ + Na₂HPO₄ and remaining glucose by anionic extraction with phenylboronic acids. The third step of the proposed method includes recovery of fructose from the organic medium by means of back-extraction into an acidic solution. After back-extraction of fructose, the organic phase can be reused for extraction. The proposed method presents a process integration allowing simple product recovery as well as the recycling of the catalyst and the rest of the substrate. Importantly, a separation strategy based on reversible esterification with boronates can also be applied for separation of not only saccharides but also vic-diols [58]. For example, we have recently proposed a method to recover microbially produced 2,3-butanediol from complex fermentation broths using anionic extraction with phenylboronic acid [59]. 2,3-Butanediol can be applied as a renewable platform molecule for synthesis of commodities, such as butadiene or methyl ethyl ketone [60]. Technologies to produce fine chemicals, for example diacetyl [61], based on 2,3-butanediol are also known. Application of a reactive adsorption instead of reactive extraction would further improve the sustainability of the suggested approach as a potential element of future biorefinery.

Figure 5.

Scheme of extraction-assisted catalytic synthesis of fructose. Adapted with permission from Delidovich & Palkovits [57]. (Online version in colour.)

(c). The interface to biotechnology

Several promising platform chemicals are accessible via fermentation of carbohydrates including ethanol, 2-butanol, 2,3- or 1,4-butanediol and lactic, succinic or itaconic acid. These value-added compounds exhibit rather high polarity and are produced in aqueous medium—most frequently in low concentrations. In addition to the target molecules, the fermentation broths typically contain a number of other components including other metabolites and electrolytes. Accordingly, their separation is not straightforward as illustrated in the previous section [41]. Nevertheless, in most cases, follow-up chemistry requires exhaustive purification of the compounds before further processing.

Itaconic acid serves as an example illustrating the challenges of fermentation broths as feedstocks. Itaconic acid is produced currently mainly for polymer industry as a building block of water-soluble poly(itaconic acid) and as a monomer for co-polymers [62,63]. Additionally, utilization of itaconic acid as a platform chemical has attracted attention in recent years. For example, a number of studies has highlighted the possibility to hydrogenate itaconic acid into the corresponding lactones and diols up to the formation of 3-methyltetrahydrofuran (figure 6) [64–67]. The latter and the corresponding 2-methyltetrahydrofuran based on levulinic acid hydrogenation have found attention as potential biofuels with excellent combustion properties and energy density together with the potential of significantly reduced NO_x and soot emissions [68,69]. Efficient hydrogenation systems were mostly based on noble metal catalysts [64] and transformations in organic solvents such as THF or dioxane in the presence of acidic additives [65]. A hydrogenation of pure itaconic acid in aqueous phase is possible, e.g. using supported ruthenium catalysts, though lactones and diols are the major products [66]. Applying exactly the same reaction conditions to fermentation broths containing itaconic acid and residual glucose, no conversion could be observed [67]. Interestingly, optimization of the fermentation strategy to yield a glucose-free product solution improved direct further hydrogenation with a conversion of about 20%. Adsorption offers an energy efficient possibility for purification. Therefore, we tested the hydrogenation of itaconic acid derived via ad- and desorption from a real fermentation broth. With full conversion and significant lactone formation, the potential of this strategy was illustrated [66]. Nevertheless, a direct use of fermentation solution avoiding prior purification appears highly desirable for enhanced process intensification. This would allow for a decreased number of process steps and operational units. In addition, a partial loss of itaconic acid upon recovery would be avoided.

Figure 6.

Stepwise catalytic hydrogenation/dehydration of itaconic acid for production of potential biofuels.

From a general point of view, fermentation broths are prevailingly aqueous electrolyte containing substrate solutions. A technology heavily relying on such electrolyte solutions as solvents is electrocatalysis. Another important aspect relates to the targeted transition of our worldwide energy system which will also impact the chemical industry. Considering production of electrical energy via technologies such as wind, solar or water power, ‘green’ electrons become available. Indeed, electrocatalysis enables direct integration of such renewable electrical energy into chemical value chains. Applying electrocatalysis for the transformation of renewable carbon sources such as biomass and CO₂, alternative valorization schemes with near-to-zero use of fossil resources become feasible. Therein, electrons can be used as reduction equivalents; tailored oxidation processes of biomass-based platform chemicals are possible [70] and even C–C coupling reactions can be realized. Recent literature examples comprehend the formation of octane based on Kolbe electrolysis of valeric acid and the formation of 2,7-octanedione and 2,5-dimethyladipic acid via Kolbe electrolysis of carboxylic acids [71–76]. Tackling this challenge, we recently demonstrated a direct electrochemical reduction of itaconic acid in fermentation broth without the need for prior purification (figure 7) [67].

Figure 7.

Electrocatalytic reduction of itaconic acid (IA) to 2-methylsuccinic acid (MS) over different electrode materials. FE denotes faradaic efficiency, and X and Y are conversion and yield, respectively (room temperature, 1 h, CE glassy carbon, E = −1.06 V versus SHE (standard hydrogen electrode), 16 ml water, 0.5 mol l⁻¹ H₂SO₄, 0.1 mol l⁻¹ IA (1.6 mmol)).*E = −1.41 V versus SHE (a) and comparison of the electrocatalytic reduction of a model solution with a fermentation broth using a Pb electrode (room temperature, 1 h, CE glassy carbon, I = 0.3 A, 16 ml water, 0.5 mol l⁻¹ H₂SO₄, 0.57 mol l⁻¹ IA (9.1 mmol)).*E = −1.41 V versus SHE (b). Adapted with permission from Holzhäuser et al. [67]. (Online version in colour.)

For the electrochemical hydrogenation of itaconic acid to methylsuccinic acid, the metals Cu, Ni, Fe, Pb and Cu–Pb (plastic bronze) have been investigated for their electrochemical activity using a fixed voltage (figure 7a) [67]. Ni allows the highest catalytic activity and faradaic efficiency, while Fe still revealed notable yield and conversion. Interestingly, when switching the applied potential from −1.06 to −1.41 V versus SHE, Pb electrodes meet the required overpotential for hydrogenation and start to clearly stand out over the other tested electrodes. Under optimized reaction conditions, an excellent faradaic efficiency can be reached and the conditions can be applied to real fermentation broth with only a minor change in efficiency. The applied potential has to be increased (from −1.4 to −4.4 V versus SHE) as the resistance of the fermentation broth is higher than of the model solutions. However, this problem is easily solved adding diluted sulfuric acid solution. Noteworthy, significant differences in the faradaic efficiencies in figure 7a,b can be explained by different initial concentrations of the substrate (0.1 versus 0.6 mol l⁻¹), as the substrate concentration was found to be crucial for the faradaic efficiency [67].

Indeed, electrocatalysis presents a promising approach for the transformation of biomass-based feedstocks enabling a direct further processing of raw feedstock streams. In view of the future energy system based on a significant fraction of renewable energy, electrocatalysis presents a crucial element at the interface of energy generation, storage and utilization in chemical industry.

3. Conclusion

Biomass presents a promising renewable carbon source with the potential to arrive at overall CO₂-neutral value chains and products. Until technologies for direct air capture of CO₂ are mature, nature's ability to concentrate low CO₂ levels in the atmosphere in plant material is a valid concept. In addition, biomass possesses well-defined and highly functionalized building blocks offering access to attractive platform chemicals for synthesis of biofuels, monomers and fine chemicals. Major challenges associated with a selective catalytic valorization of biomass relate to the required paradigm shift for biorefineries compared with today's refinery technology. Fossil feedstocks are mostly processed at elevated temperature in gas phase. By contrast, biorefineries rely on liquid-phase transformations, often in water, at mild conditions. Therein, highly polar, oxygen-rich molecules with high-boiling points are processed. This entails the need for advances in various fields. This contribution discusses aspects of efficient catalytic deoxygenation processes, possibilities of a redesign of separation technologies as well as the opportunities provided by the interfaces to biotechnology and electrocatalysis. Indeed, significant progress in the design of multifunctional catalysts for selective hydrodeoxygenation has been achieved allowing controlled access to building blocks such as ethylene and propylene glycol, 1,6-hexanediol or n-hexane based on cellulose. In separation, novel concepts including selective adsorption on high-performance HCPs and the opportunities offered by anionic extraction were elucidated. These strategies are also decisive at the interface to biotechnology were removal of platform chemical from aqueous electrolyte solutions presence a bottle-neck. Both the mentioned separation concepts and a direct combination with electrocatalytic transformations constitute important fundamental approaches for molecularly driven process intensification.

Data accessibility

This article has no additional data.

Author contributions

R.P. and I.D. contributed equally.

Competing interests

We declare we have no competing interests.

Funding

The authors thank the Cluster of Excellence ‘Tailor-Made Fuels from Biomass’ (DFG EXC 236) funded by the Excellence Initiative by the German federal and state governments to promote science and research at German universities.