Jiayi Fu

Waste-derived adhesives exhibit excellent adhesion to stainless steel without the addition of any tackifier or plasticizer, typically used in the manufacturing of adhesives from petroleum. We synthesize adhesives that are just as competitive as commercial tapes. Applications range from packaging and plastic wraps to labels and sticky notes.

We have introduced a strategy toward the development of high-performance, pressure-sensitive adhesives (PSAs) from raw biomass. By depolymerizing poplar wood with minimal purification, we have produced substituted aromatic compounds such as 4-propylsyringol and 4-propylguaiacol. Their efficient functionalization with either acrylate or methacrylate groups yield monomers that could easily be polymerized by a scalable, reversible addition-fragmentation chain-transfer process into polymeric materials with high glass-transition temperatures and robust thermal stabilities, especially relative to other potentially bio-based alternatives.

Lubricants represent a ~$61 billion global market at an annual consumption of 40 million tons. They have numerous applications in industrial machineries, automotive engines, agricultural equipment, aviation machinery, hydraulic fluids, hydropower, and transformers. Petroleum-based lubricants consist of C20-C50 molecules and are formulated with additives. The high volatility of low-carbon fractions causes thickening over time, requiring frequent lubricant replacements. High-performance applications under extreme conditions (e.g., in gas turbines and nuclear reactors) demand rigid specifications (e.g., low pour point, high thermal stability, low volatility, and high viscosity index). Synthetic, branched poly-alpha-olefin lubricants produced by olefin oligomerization meet some specifications, but selectively tuning their size and architecture is challenging. Use of less expensive, sustainable, and versatile carbon feedstocks to produce lubricants with controlled size and molecular architecture, and a narrow molecular weight distribution can improve lubricant properties and avoid complex formulations and high cost.

We have developed novel synthesis strategies to produce three broad classes of bio-lubricants from inexpensive and sustainable non-food biomass and natural oils or waste cooking oils. Yields ranging from 80 to over 90% have been achieved. These molecules contain furan rings, or fully saturated furan rings, or fully deoxygenated branched alkanes. Importantly, we can control with precision the number of carbons, the molecular structure, the degree of aromaticity and the oxygen level. Our carbon-carbon coupling synthesis strategies include: (1) acylation of furan; (2) hydroxyalkylation/alkylation of 2-alkylfurans with aldehydes; (3) conjugate addition-hydroxyalkylation/alkylation (CA-HAA) of 2-alkylfurans with enals; and (4) ketonic decarboxylation of fatty acids followed by aldol condensation. The derived furan-based lubricants can be further upgraded via hydrogenation or hydrodeoxygenation. New, inexpensive metal oxide on metal hydrodeoxygenation catalysts have been developed. The properties of these bio-lubricants are on par with or better than those of commercial mineral oils and synthetic lubricants.

We have introduced a novel, two-step strategy to synthesize benzene and branched cyclic alkane lubricant base-oils from lignin-derived guaiacol and lauryl aldehyde (produced from natural oils or biomass). The reaction pathway involves carbon–carbon coupling through Brønsted acid-catalyzed hydroxyalkylation/alkylation (HAA), followed by hydrodeoxygenation (HDO). We optimized the reaction conditions to achieve a maximum guaiacol conversion of 90%, with a product consisting of 70% benzene lubricant, 22% aldol condensation side products, and 8% unreacted guaiacols. Subsequent HDO of the products over Ir-ReOx/SiO₂ produced a lubricant-ranged mixture of cyclic and branched alkanes (C₂₄), at 82% yield, and small fractions of dodecyl cyclohexane and C-C cracked alkanes with carbon numbers between C₁₅ and C₁₀.  Quantification of viscosity properties (kinematic viscosity, viscosity index, and Noack volatility) indicated that the synthesized biobased lubricant base-oil was comparable to commercial petroleum-derived poly α-olefin Group III, IV, and refrigerant base oils. This approach provides a unique pathway for upgrading lignin-derived monomers into replacements of petroleum-derived base oils. We are expanding to additional lignin-based paths to lubricants and replace fatty acids with sugar-derived molecules.

Aromatics constitute a large market, with BTX (benzene-toluene-paraxylene) being one of the highest-volume and considerably energy-intensive chemicals made from crude oil. As steam cracker operators are capitalizing on the shale-gas revolution and inexpensive natural gas liquids are pushing naphtha feedstocks aside, the development of alternative routes for the production of aromatics from renewables, such as biomass, is receiving considerable attention. Of the possible routes garnering interest, CCEI scientists have proposed the dehydrative aromatization of the Diels-Alder product between bio-based furans with an appropriate dienophile; a tandem scheme where the cycloadduct obtained from the Diels-Alder is catalytically dehydrated to an aromatic. The strategy was showcased by the formation of p-xylene from 2,5-dimethylfuran and ethylene over the Brønsted-acidic catalysts H-Y and H-BEA. The same strategy has been used to convert methylfuran and ethylene to toluene; ethylene and methyl 5-(methoxymethyl)-furoate to methyl 4-(methoxymethyl)-benzoate; ethylene and the dimethyl ester of 2,5-furandicarboxylic acid to dimethyl terephthalate; furan and maleic anhydride to phthalic anhydride; and furan and methyl acrylate to methylbenzoate. and furan and ethylene to benzene.

In order to curtail undesired side reactions promoted by Brønsted acids, CCEI researchers turned their attention to Lewis-acidic zeotypes, specifically isomorphically substituted zeotypes with Lewis-acidic centers, such as Sn-BEA, Zr-BEA, Ti-BEA, Hf-BEA and Zn-BEA. This multi-institution effort led to the synthesis of numerous aromatics: methyl 4-(methoxymethyl)-benzoate from ethylene and methyl 5-(methoxymethyl)-furoate; dimethyl terephthalate from ethylene and the dimethyl ester of 2,5-furandicarboxylic acid; as well as p-xylene. These efforts led to the remarkable finding that Lewis-acidic zeotypes are as effective as the Brønsted-acidic H-BEA in influencing the rate of the tandem scheme, and in particular the dehydration of oxanorbornene derivatives.

The fleeting existence of the Diels-Alder cycloadduct does not allow one to decouple the two steps of the tandem scheme and thus elucidate whether the zeotypic Brønsted and Lewis acids have any effect on the rate of the Diels-Alder step itself. With furans being among the less reactive dienes, accelerating the cycloaddition reaction will be a pivotal step in the success of producing aromatics by dehydrative aromatization of the Diels-Alder product of biomass-derived furans and appropriate dienophiles. Calculations have asserted that when ethylene is the dienophile, neither Brønsted nor Lewis acid zeotypes can influence the rate of the Diels-Alder step, because the furanic molecule binds more strongly to the active site of the catalysts.

CCEI catalysis experts have achieved to synthesize methylbenzoate by Diels-Alder aromatization of furan and electrophilically activated methylacrylate over the zeotypic Lewis acids Sn-BEA, Zr-BEA, and Hf-BEA. With the aid of computational chemistry, they were able to show that these Lewis acids are capable of accelerating the Diels-Alder reaction heterogeneously.

In order to curtail by-product formation, CCEI researchers have also been exploring phosphorous-containing siliceous zeolites. They were able to synthesize zeolites with BEA (P-BEA) and MFI topology (P-SPP). These catalysts are highly selective and stable for this reaction, reaching an unprecedented p-xylene yield of 97%. The superior properties of phosphorous-containing siliceous zeolite catalysts are attributed to the nature of the acid sites which exclusively catalyze the dehydration of cycloadduct intermediate, and remarkably minimize the alkylation and oligomerization reactions.

The HDO Thrust has aimed to develop science and catalysts for the upgrade of furan-based compounds.

Furans, obtained from dehydration of sugars or pyrolysis of lignocellulose and oxygenated fuel precursors obtained from chain growth chemistry of furans are important intermediate platforms of lignocellulosic biomass.

However, owing to their oxygen-rich content, their transformation to fuels and chemicals requires selective oxygen removal. Upgrade of furans requires transformative technology to improve product yields and introduce new chemicals and to reduce catalyst cost.

We discovered that moderately reducible metal oxides, such as RuOx, can contain multifunctional sites that can selectively hydrogenate carbonyl groups, compared to the ring, via the MPV mechanism over Lewis acidic sites, and then hydrodeoxygenate the side OH groups on vacancies. The latter behave like redox centers, where a single concerted mechanism occurs, entailing radical formation and ring conjugation to facilitate the C-OH bond scission. While RuOx is an effective and selective catalyst, it is being reduced during catalysis. In order to improve catalyst stability, this concept was extended further. Computational studies allowed for the first time to establish catalyst design principles, which can ultimately lead to the discovery of the next-generation of C-O activation catalysts. First, we have explored the activity and stability tradeoff of single metal oxides and demonstrated a volcano curve where the activity/stability is correlated with the bulk Gibbs free energy as well as surface properties of the oxides. RuOx and IrOx are the best single metal oxide catalysts. Second, we used colloidal synthesis and atomic layer deposition (ALD) of bimetallics to create metal core/oxide monolayer thick shell, e.g., cobalt-oxide-overcoated Pt nanoparticles and Cu/Ni catalysts. These catalysts activate hydrogen and importantly do not get reduced; they can excitingly provide nearly quantitative yield in converting HMF to dimethyl furan (DMF).

In a parallel effort, we developed efficient and highly selective catalysts for (1) ring-opening catalytic pathways to convert biomass-derived furans and tetrahydrofurans into adipic acid, an essential monomer in 6,6-nylon; (2) deoxydehydration of vicinal diols to olefins, a class of compounds with a multitude of uses. A first generation of catalysts was obtained by adopting transition metals and structural motifs from homogeneous catalysts, e.g., oxide-supported oxo-rhenium; (3) non-precious metal oxide catalysts for the ring-opening of furanics. We have demonstrated that the use of reduced Cu-Co-Al mixed metal oxides achieves the formation of 1,5-pentanediol with upward yields of 45% and 1,2-pentanediol with upward yields of 17%. This is the highest reported yield for 1,5-pentanediol when compared with other catalysts with a similar mechanism; (4) production of mono and di-ethers of furans suitable for detergents and fuels via the MVP reaction over Lewis acid centers, e.g., Sn-containing zeolites, and etherification over acid sites.

Furans are potentially the most versatile platform molecules for converting lignocellulose into various bio-products. CCEI has demonstrated a number of new bio-products derived from furans. Yet, furans production has been hindered by low yields due to various byproducts, such as humins. The furans thrust aimed at paradigm changing approaches for the selective transformations of sugars to furans and their further conversion to value-added chemicals and fuels.

Inspired by an earlier demonstration of intermolecular Meerwein-Ponndorf-Verley (MPV) carbonyl reduction and Oppenauer alcohol oxidation (MPVO) in Sn-BEA, we demonstrated sugar isomerization by an analogous intramolecular MPVO cycle. We proposed that the Sn⁴⁺ centers catalyze the intramolecular H-shift to fructose similarly to the reaction catalyzed by D-xylose isomerase metalloenzymes. Using ¹¹⁹Sn MAS NMR spectroscopy and isotopic labelling (²H, ¹³C), we fully characterized the Sn-BEA active site and showed the Sn centers in their “open” configuration ((-SiO)₃-Sn-OH) can be distinguished from those in their “closed” configuration [(-SiO)₄-Sn]. Using DFT calculations, we elucidated the glucose-fructose isomerization mechanism and showed that the “open” framework stannanol (Sn-OH) sites are catalytically more active than the “closed” ones and suggested that group-cooperativity—often encountered in enzymatic catalysis—determines selectivity to fructose, via 1,2-hydride shift, or to mannose (epimerization), via 1,2-carbon shift.

We contributed to the selective formation of 5-hydroxymethylfurfural (HMF) from fructose, with emphasis on understanding why certain organic solvents (e.g., DMSO, neat or as co-solvent) accelerate fructose dehydration and minimize HMF loss to levulinic and formic acids and humins. Using quantum mechanics/molecular mechanics molecular dynamics simulations we showed that rate-limiting elementary steps of the dehydration are profoundly influenced by solvent dynamics and re-organization. We developed the COSMO-SAC modeling platform to predict solvents for reactive extraction to increase the yield of furans during dehydration. We introduced reactive adsorption and suitable adsorbents as an alternative to reactive extraction and showed that this can be a competitive process at lower temperatures. Finally, we developed the first kinetic models for dehydration of fructose to HMF and byproducts and for glucose isomerization to fructose and used these to optimize operating conditions.

We employed speciation modeling, ab initio molecular dynamics and a suit of experimental methods to elucidate the mechanism of HMF formation in the glucose-fructose and xylose-xylulose isomerization using metal salts as homogenous Lewis acid catalysts. We showed that the homogeneous aldol-ketose isomerization mechanism is similar to that over Sn-BEA. Hydrolysis of the salt produces Brønsted acidity that drives fructose dehydration to HMF and further to formic and levulinic acids.

We developed methodologies that enable the synthesis of hierarchical zeolites. We introduced two representative hierarchical architectures (self-pillared MFI and 3-dimensionally-ordered-mesoporous-imprinted (3DOm-i) zeolites) with nanometer-sized microporous domains accessible through larger pores, ensuring coexistence of micropore activity and selectivity with fast mass transport. These hierarchical zeolites can be synthesized in pure silica or aluminosilicate form and now with metals (e.g., Sn) incorporated in the frameworks.

We achieved moderate-temperature (around 100 °C) retro-aldol reactions of various hexoses in aqueous and alcoholic media with catalysts traditionally known for their capacity to catalyze 1,2-intramolecular carbon shift reactions of aldoses, for example, various molybdenum oxide and molybdate species and alkali-exchanged stannosilicate molecular sieves, paving the road to alkyl lactates from renewables.