Organocatalytic Upgrading of Furfural and 5-Hydroxymethyl Furfural to C10 and C12 Furoins with Quantitative Yield and Atom-Efficiency

There is increasing interest in the upgrading of C5 furfural (FF) and C6 5-hydroxymethyl furfural (HMF) into C10 and C12 furoins as higher energy-density intermediates for renewable chemicals, materials, and biofuels. This work utilizes the organocatalytic approach, using the in situ generated N,S-heterocyclic carbene catalyst derived from thiazolium ionic liquids (ILs), to achieve highly efficient self-coupling reactions of FF and HMF. Specifically, variations of the thiazolium IL structure have led to the most active and efficient catalyst system of the current series, which is derived from a new thiazolium IL carrying the electron-donating acetate group at the 5-ring position. For FF coupling by this IL (0.1 mol %, 60 °C, 1 h), when combined with Et3N, furoin was obtained in >99% yield. A 97% yield of the C12 furoin was also achieved from the HMF coupling by this catalyst system (10 mol % loading, 120 °C, 3 h). On the other hand, the thiazolium IL bearing the electron-withdrawing group at the 5-ring position is the least active and efficient catalyst. The mechanistic aspects of the coupling reaction by the thiazolium catalyst system have also been examined and a mechanism has been proposed.


Introduction
Biofuran aldehydes (or furaldehydes), particularly furfural (FF) and 5-hydroxymethyl furfural (HMF) [1][2][3][4][5][6][7][8][9][10][11], which are primarily derived from dehydration of C5 and C6 (poly)sugars, have emerged as two key renewable biomass building blocks for biorefining towards addressing global challenges to develop technologically and economically feasible routes for converting nonfood lignocellulosic biomass into feedstock chemicals, sustainable materials, and liquid fuels [12][13][14][15][16][17]. However, such C5 and C6 furaldehydes lead to, upon upgrading, relatively low carbon-number and low energy-density bio-products or fuels. Hence, there is increasing interest in upgrading of FF and HMF into higher molecular weight and higher energy-density kerosene/jet (C8 to C16) or diesel (up to C22) intermediates or fuels. This upgrading can be accomplished by chain extension involving new C-C bond formation through coupling with other organic compounds [18][19][20][21]. Considering the fact that such furaldehydes cannot undergo self-aldol condensation due to lack of α-hydrogen, Dumesic and co-workers utilized cross-aldol condensation of HMF with enolizable organic compounds such as acetone in the presence of an alkaline catalyst, followed by hydrodeoxygenation (HDO) processes, to upgrade HMF into C9 to C15 liquid alkane fuels [22]. Alternatively, opening the furan rings first under mild conditions, followed by HDO, produced alkanes more selectively [23]. Direct (reductive) coupling of two FF molecules by metal catalysts has also been utilized, leading to formation of Pinacol coupling products as a mixture of C10 alcohols, which can be subsequently converted into C8-14 linear and branched alkanes via the HDO process [24].
ILs exhibit the unique ability to dissolve lignocellulosic biomass [39,40] and related carbohydrates [41,42] under relatively mild conditions; this feature, plus several other concurrent advantages (e.g., as designable and recyclable solvents with low volatility and toxicity), has attracted much interest [43][44][45], particularly in the pursuit of renewable energy and sustainable chemicals from plant biomass [46,47]. Pertinent to the organocatalytic upgrading of furaldehydes, ILs are also precursors to the NHC catalysts that can efficiently promote benzoin condensation, a direct coupling reaction between two aldehydes [26][27][28][29][30][31][32]48]. Scheme 2 outlines the mechanism for the NHC-catalyzed benzoin condensation of benzaldehyde originally proposed by Breslow [49][50][51]. In this mechanism, thiazolium salt A is deprotonated by a strong base (e.g., Et3N) to form thiazolin-2-ylidene NHC catalyst B. Nucleophilic attack of the aldehyde by B generates the carbene-aldehyde zwitterionic adduct C, followed by subsequent protonation/deprotonation or proton transfer, which affords the Breslow intermediate, D. This amino enol intermediate functions as an acyl anion equivalent and attacks a second aldehyde, forming zwitterionic adduct E [49][50][51]. After the proton transfer and the elimination of the benzoin product, the NHC catalyst B is regenerated. This process is also termed aldehyde umpolung (polarity inversion) that converts the electrophilic carbonyl carbon to a nucleophilic center as an acyl anion equivalent. As for the biomass-derived furaldehydes, benzoin condensation of FF is catalyzed by NHCs in a similar fashion [52][53][54]. For example, the methanol adduct of TPT, TPT(MeOH), was used to catalyze condensation of FF to give furoin in 78% yield [55]. On the other hand, 3-benzyl-5-(2-hydroxyethyl)-4-methylthiazolium bromide, HO [TM]Br, in combination with Et3N, was employed for benzoin condensation of FF, achieving 90% yield of furoin [56].
To explore inexpensive alternatives to the discrete NHC catalysts such as TPT for furaldehyde umpolung coupling, we arrived at thiazolium IL HO  nucleophilicity of the corresponding NHC catalyst derived from deprotonation could be modulated, thus providing an opportunity to discover more activity and/or selective catalysts for coupling the biomass-derived furaldehydes.

Scheme 2.
Outlined catalytic cycle of the benzoin condensation through the Breslow intermediate (D) and the three thiazolium ILs investigated in this study as the precatalysts for coupling of furaldehydes.

Synthesis of Two New Thiazolium ILs AcO [TM]Cl and Ac [TM]I
The parent thiazolium salt HO [TM]Cl carries the 2-hydroxyethyl group at the 5-position of the methylthiazolium ring. To examine if the protic OH group facilitates or hampers the coupling reaction of furaldehydes, we designed the synthesis of its acetate derivative AcO [TM]Cl. Alkylation of the corresponding substituted methylthiazole with benzyl chloride produced AcO [TM]Cl in 81% isolated yield (Scheme 3). 1 H NMR (DMSO-d6, Figure 1) and 13 C NMR (DMSO-d6, Figure 2) spectra confirm the formation of the clean, desired product, with all the resonances assigned to the expected protons and carbons, respectively (see Experimental Section). Noteworthy is the chemical shift of the C2-H (i.e., NCHS) at 10.26 ppm, indicative of the acidic nature of this proton.
As both HO [TM]Cl and AcO [TM]Cl carry an electron-donating group at the 5-ring position, we designed the synthesis of Ac [TM]I with the electron-withdrawing methoxycarbonyl (acetyl type) group attached to the 5-ring position (Scheme 3). Initial attempts to alkylate the corresponding methylthiazole with benzyl chloride failed to produce the anticipated thiazolium chloride salt, Ph

Thiazolium ILs for FF and HMF coupling examined in this study
as a result of the electron-deficient nature of the thiazole ring with the methoxycarbonyl attached to it. In turn, we employed the stronger alkylating reagent, MeI, successfully producing the corresponding thiazolium iodide salt, Ac [TM]I. As can be seen from 1 H NMR (DMSO-d6, Figure 3) and 13 C NMR (DMSO-d6, Figure 4) spectra, the formation of the clean, desired product is achieved. As anticipated, the acidic C2-H (i.e., NCHS) is largely downfield-shifted to 10.02 ppm, consistent with the chemical shift of the same proton observed for HO

Coupling Reaction of FF to Furoin by HO [TM]Cl, AcO [TM]Cl, and Ac [TM]I
The self-coupling of FF to 1,2-Di(furan-2-yl)-2-hydroxyethanone (furoin) by the thiazolium salt in combination with a base was carried out under neat conditions (no solvent), the results of which were summarized in Table 1. The acetate derivative, AcO [TM]Cl, when combined with 2 equiv of Et3N, was found to be highly effective to catalyze the FF-to-furoin coupling reaction. Specifically, with a loading of 1 mol % pre-catalyst, furoin was achieved in 99% yield (by NMR) at 80 °C in 3 h (entry 1, Table 1). Lowering the pre-catalyst loading to 0.5 and 0.1 mol % while keeping the temperature and time the same still afforded furoin in quantitative yield (entries 2 and 3, Table 1). When carried out at 60 °C for 1 h, the furion yield with a low pre-catalyst loading of 0.1 mol % depended on the amount of the base added, varying from 81% (entry 4), 86% (entry 5), to >99% (entry 6) when the amount of Et3N was increased from 1, 2, and 4 equivalents (relative to the pre-catalyst), respectively. Other bases such as KO t Bu and 1,8-diazabicyclo [5.4.0]undec-7-ene (DBU) were also examined but found to be much inferior to Et3N (entries 7 and 8 vs. entry 5, Table 1).  Table 1). Increasing the pre-catalyst loading to 0.5, 5, and 10 mol %, the furoin yield was enhanced gradually to 16, 49, and 84%, respectively (entries 12, 13, and 14, Table 1). This large reduction in the catalytic performance by Ac [TM]I could be attributable to the electron-withdrawing substitute that renders low nucleophilicity of the corresponding NHC catalyst. Overall, for the current three thiazolium ILs investigated for the FF coupling reaction, the catalytic performance follows this order: AcO

Coupling Reaction of HMF to DHMF by HO [TM]Cl, AcO [TM]Cl, and Ac [TM]I
The self-coupling of HMF to DHMF by the thiazolium salt in combination with a base was also carried out under neat conditions (no solvent), the results of which are summarized in Table 2. The coupling of HMF to DHMF is a much more challenging reaction than the self-coupling of FF, requiring the catalyst in higher loadings. Specifically, with 1 and 5 mol % loadings of AcO [TM]Cl in combination with 2 equivalents of Et3N at 80 °C for 3 h, DHMF was achieved in only 34% and 50% yield, respectively (entries 1 and 2, Table 2). Increasing the catalyst loading to 10 mol % significantly enhanced the DHMF yield to 94% (entry 3, Table 2). Increasing the reaction temperature to 100 and 120 °C improved the DHMF yield somewhat to 96% and 97%, respectively (entries 4 and 5, Table 2). Again, other bases such as KO t Bu and DBU gave inferior catalytic performances relative to Et3N (entries 6, 7, and 8 vs. entries 3 and 4, Table 2).

Proposed Mechanism for FF and HMF Coupling Reaction
To provide further insight into the mechanism of the coupling reaction of FF by the thiazolium/Et3N system, we monitored the reaction of Ac [TM]I + Et3N + FUR (1:1:1) at 80 °C by 1 H NMR ( Figure 5). The 1 H NMR spectrum of the initial mixture showed a mixture of starting reagents (Figure 5a). As the reaction progresses, formation of a new species (from spectrum (a) to (b) to (c)) became apparent; the NMR characteristics of this species (δ 8.05 (dd, 1H), 7.45 (dd, 1H), 6.73 (dd, 1H) ppm for the three furan ring protons, δ 6.21 (s, 1H) for the methide proton (CHOH),  The above results and the identification of intermediate II led to the coupling mechanism proposed in Scheme 4, which is consistent with the mechanism proposed for the self-coupling of HMF by [EMIM]OAc [36]. Specifically, the NHC catalyst (I) is generated by in situ deprotonation of Ac

Materials, Reagents, and Methods
All syntheses and manipulations of air-and moisture-sensitive materials were carried out in flamed Schlenk-type glassware on a dual-manifold Schlenk line, on a high-vacuum line, or in an inert gas (Ar or N2)-filled glovebox. NMR-scale reactions were conducted in Teflon-valve-sealed J. Young-type NMR tubes. HPLC-grade organic solvents were first sparged extensively with nitrogen during filling 20 L solvent reservoirs and then dried by passage through activated alumina (for Et2O, THF, and CH2Cl2) followed by passage through Q-5 supported copper catalyst (for toluene and hexanes) stainless steel columns. DMSO-d6 was first degassed and dried over CaH2, followed by vacuum distillation (CaH2 was filtered off before the distillation). NMR spectra were recorded on a Varian Inova 300 (FT 300 MHz, 1 H; 75 MHz, 13 C) or a Varian Inova 400 MHz spectrometer. Chemical shifts for 1 H and 13 C spectra were referenced to internal NMR solvent residual resonances and reported as parts per million relative to SiMe4 (TMS). High-resolution mass spectrometry (HRMS) data were collected on an Agilent 6220 Accurate time-of-flight LC/MS spectrometer.

Typical Procedure for Coupling Reaction of FF into Furoin
The coupling reaction was carried out under solvent-free (neat) conditions. In a typical coupling experiment, FF (0.77 g, 8.0 mmol), AcO [TM]Cl (25 mg, 0.08 mmol), and Et3N (2.23 μL, 0.16 mmol) were loaded into a 20 mL vial in the glovebox. The vial was sealed, taken out of the glovebox, and then put in a temperature-controlled orbit shaker. The reaction mixture was shaken (300 rpm) at 80 °C for 3 h. After the reaction, the solidified product was washed with 10 mL hexanes and filtered. After drying in vacuum, furoin was obtained in 99% yield as a bright-yellow powder. 1

Typical Procedure for Coupling Reaction of HMF into DHMF
The coupling reaction was carried out under solvent-free (neat) conditions. In a typical coupling experiment, HMF (0.126 g, 1.00 mmol), AcO [TM]Cl (0.0311 g, 0.1 mmol), and Et3N (27.4 μL, 0.2 mmol) were loaded into a 20 mL vial in the glovebox. The vial was sealed, taken out of the glovebox, and then put in a temperature-controlled orbit shaker. The reaction mixture was shaken (300 rpm) at 80 °C for 3 h. After the reaction, the solidified product was washed with 10 mL toluene to remove the residual catalyst and any unreacted HMF. DHMF (94% yield) was obtained as white powder after filtration and vacuum drying. 1  [TM]I (29.9 mg, 0.01 mmoL) and Et3N (10.0 mg, 0.01 mmoL) were dissolved in 0.5 mL DMSO-d6 and transferred into a J. Young-type NMR tube, to which FF (9.6 mg, 0.01 mmoL) in 0.5 mL DMSO-d6 was added and fully mixed. The mixture was heated to 80 °C on an NMR spectrometer, and the reaction was followed by taking 1 H NMR spectra of the reaction mixture at predetermined time intervals. After 3 h at 80 °C, the formation of intermediate II was indicated by 1 H NMR spectrum. 1

Conclusions
This work has synthesized, and examined catalytic performance of, two new electronically modified thiazolium ILs as inexpensive alternatives to the discrete NHC catalysts for umpolung coupling of bio-derived furaldehyde FF and HMF into chain-extended C10 and C12 furoins. Comparing to the current thiazolium benchmark catalyst based on HO [TM]Cl, we found that AcO [TM]Cl bearing an electron-donating group at the 5-ring position, when combined with Et3N, is the most active and efficient catalyst in this series for coupling of FF, being 4.6 times faster than the current benchmark catalyst system. Thus, the catalyst system with a low catalyst loading of only 0.1 mol % of AcO [TM]Cl and 0.4 mol % Et3N, >99% yield of furoin was obtained from the coupling reaction of FF at 60 °C for 1 h. On the other hand, the catalyst system based on Ac [TM]I carrying an electron-withdrawing group at the 5-ring position is the least active and efficient system of the series. For the more challenging coupling reaction of HMF, AcO [TM]Cl and HO [TM]Cl performed equally well, with both achieving 97% yield of DHMF at 120 °C after 3 h, albeit requiring a high catalyst loading of 10 mol %.
The study also yielded the proposed mechanism for the coupling reaction (c.f., Scheme 4). The catalyst responsible for the coupling reaction is the in situ generated NHC I, upon deprotonation of the thiazolium IL by Et3N, which attacks the carbonyl group of FF to form the NHC-FF adduct, intermediate