Synthesis of α,β‐ and β‐Unsaturated Acids and Hydroxy Acids by Tandem Oxidation, Epoxidation, and Hydrolysis/Hydrogenation of Bioethanol Derivatives

Abstract We report a reaction platform for the synthesis of three different high‐value specialty chemical building blocks starting from bio‐ethanol, which might have an important impact in the implementation of biorefineries. First, oxidative dehydrogenation of ethanol to acetaldehyde generates an aldehyde‐containing stream active for the production of C4 aldehydes via base‐catalyzed aldol‐condensation. Then, the resulting C4 adduct is selectively converted into crotonic acid via catalytic aerobic oxidation (62 % yield). Using a sequential epoxidation and hydrogenation of crotonic acid leads to 29 % yield of β‐hydroxy acid (3‐hydroxybutanoic acid). By controlling the pH of the reaction media, it is possible to hydrolyze the oxirane moiety leading to 21 % yield of α,β‐dihydroxy acid (2,3‐dihydroxybutanoic acid). Crotonic acid, 3‐hydroxybutanoic acid, and 2,3‐dihydroxybutanoic acid are archetypal specialty chemicals used in the synthesis of polyvinyl‐co‐unsaturated acids resins, pharmaceutics, and bio‐degradable/ ‐compatible polymers, respectively.

Abstract: We report a reaction platform for the synthesis of three different high-value specialty chemical building blocks starting from bio-ethanol, which might have an important impact in the implementation of biorefineries. First, oxidative dehydrogenation of ethanol to acetaldehyde generates an aldehyde-containing stream active for the production of C 4 aldehydes via base-catalyzed aldol-condensation. Then, the resulting C 4 adduct is selectively converted into crotonic acid via catalytic aerobic oxidation (62 % yield). Using a sequential epoxidation and hydrogenation of crotonic acid leads to 29 % yield of b-hydroxy acid (3-hydroxybutanoic acid). By controlling the pH of the reaction media, it is possible to hydrolyze the oxirane moiety leading to 21 % yield of a,b-dihydroxy acid (2,3-dihydroxybutanoic acid). Crotonic acid, 3-hydroxybutanoic acid, and 2,3-dihydroxybutanoic acid are archetypal specialty chemicals used in the synthesis of polyvinyl-counsaturated acids resins, pharmaceutics, and bio-degradable/ -compatible polymers, respectively.
Polymers play a key role in the creation of a myriad of materials that have contributed to the development of our modern society. The current use of plastics, however, is not sustainable in the long term due to its dependence on non-renewable fossil fuels and the environmental pollution caused by plastic waste. [1] This dilemma has triggered intensive research in the development of environmentally friendly and sustainable bio-based and biodegradable polymers. [2] Polyhydroxyalkanoates (PHAs) are a class of biodegradable isotactic polymers synthesized by bacteria. [3] Until now, these family of polymer building blocks have been synthesized using genetically modified micro-organisms or enzymes. [4] However, the production cost of PHAs is nearly four times larger than its petroleum-based counterparts (circa PHAs [5] and polypropylene [6] prices are 5-6 and 1-2 $/kg, respectively). This is due to the high cost of raw materials, low conversion rates, the complex purification of the fermentation broths, and the large amounts of biomass waste generated (circa 5 kg of raw material per 1 kg of product), and low conversion rates. [7] Catalytic conversion routes of biomass-derived feedstocks to short b-hydroxy acids (e.g. lactic acid) have shown higher productivities and atom-efficiency at industrially relevant operating conditions, [8] but their application has been limited to short chain (C 3 ) molecules. Inspired by nature, we have developed a new catalytic cascade process that mimics the step-wise coupling of C 2 units that occur during the biosynthesis of PHA in bacteria. Accordingly, we have used basecatalyzed aldol-condensation followed by tandem oxidation, epoxidation and hydrogenation or hydration (Scheme 1). This catalytic cascade route has allowed us to generate mediumchain a,b-unsaturated acids (crotonic acid), a,b-dihydroxy acids (2,3-dihydroxybutanoic acid), and b-hydroxy acids (3hydroxybutyric acid), which are emerging specialty chemicals and building blocks.
The process of converting ethanol into unsaturated and mono-and di-hydroxy acids starts with the catalytic dehydrogenation of ethanol to acetaldehyde (Scheme 1 a). This step was accomplished in a flow reactor at 250 8C using SiO 2supported 10 wt. % Cu and 5 wt. % Ni catalysts (see Table S1 in the Supporting Information). When the reaction was performed at conversions below 20 % in the presence of oxygen, the observed rates were 0.67 AE 0.06 and 0.46 AE 0.02 g EtOH g cat À1 h À1 on the Cu and Ni catalysts, respectively. At these conditions the selectivity on both catalysts in the presence of oxygen was circa 100 %, which is in agreement with previous kinetic studies performed on Cu-SiO 2 , Cu-ZnO and Cu-Al 2 O 3 catalysts that indicated that selectivities to acetaldehyde 80 % can be obtained below 20 % conversion. [9,10] The unreacted ethanol could either be recycled back for further reaction or kept in the product mixture for conducting the aldol condensation step ( Figure S9).
To selectively convert the C 2 aldehydes into the corresponding C 4 coupling products (Scheme 1 b) we have developed a number of catalytic systems based on basic K/Na-X and K/Na-Y, [11] and MgO-K/NaX, [12] zeolites as well as Bdoped MgO oxides. [13] The Mg-Al Hydrotalcite, Al-Beta zeolite, which along with the Mg-B catalyst previously employed have shown the highest C 4 productivities among all the catalysts investigated, with productivities of 8.3, 23.4, and 24.2 mmol of C 4 g cat À1 h À1 , respectively, at 50-60 % conversion level and 180 8C (Table S2). Notably, on the 7.5 wt. % B-MgO and Al-Beta catalysts the yield of C4 unsaturated aldehydes reached 60-70 mol C % at levels of conversion (80-90 %).
The subsequent conversion of the C 4 adduct to the corresponding b-hydroxy acid is rather challenging as the dehydration towards the C 4 a,b-unsaturated acid is thermo-dynamically and kinetically more favorable under the reaction conditions required for the oxidation. [14] To circumvent this limitation, we employed a tandem oxidation-epoxidation sequence. First, to accomplish aerobic oxidation step we employed two distinct types of Ru catalysts supported on reducible metal-oxides (5 wt. % RuCo x Ce y O z ) and a microporous titanosilicate catalysts (ETS-4) ( Table 1). When the reaction was performed on the Co x Ce y O z catalyst without Ru the conversion of crotonaldehyde was % 66 % and the yield to crotonic acid was 54 % after 6 hours of reaction in decalin.
Addition of 5 wt. % Ru to the Co x Ce y O z led to a significant improvement in the selectivity and activity of the catalyst that resulted in conversion values of circa 92 % and yield to crotonic acid of 100 %. The high activity and selectivity of this reaction could be attributed to the cooperative interaction between Ru and Ce sites on the CoO(OH) surface that facilitate the oxidation of the Co 2+ to Co 3+ , which are believed to be the active sites for the aerobic oxidation of aldehydes to acids via radical chain reaction. [15][16][17] Notably, when the reaction was performed in water using the RuCo x Ce y O z the crotonaldehyde conversion and crotonic acid yield decreased substantially to % 78 % and 9 %, respectively ( Figure 1). Kinetic and mechanistic studies have shown that the oxidation of aliphatic aldehydes starts with the reaction of the carbonyl aldehyde group and the Co 3+ sites to form acyl radicals, protons with the concomitant reduction of the cobalt sites to Co 2+ . The oxidation of the acyl radical with oxygen leads to an acylperoxy that abstracts a hydrogen from another aldehyde to yield a peracid and another acyl radical. This peracid undergoes protonation to carboxylic acid, water, and oxidized Co 3+ sites. [15,16,18] Due to the Lewis character of the Co centers it is possible that water competes with the acylperoxy radicals for the coordination sphere of the Co 3+ cations. This competition in combination with the quenching of chain-propagating radicals in aqueous phase, could explain the lower rates of reaction observed in water. Decreasing temperature from 100 8C to 80 and 60 8C did not improve the selectivity of the reaction in aqueous environments (Table 1). Instead, the selectivity to crotonic acid monotonically decreased as the temperature decreased, indicative of low activation barriers for the secondary reactions in the aqueous phase.  By changing the solvent to a dipolar aprotic organic solvent such as g-valerolactone (GVL) we observed similar conversion as that obtained in water (67 %), while the yield to crotonic acid increased to 54 % ( Figure 1). The differences in yield in the two solvent environments was caused by the activation of acid catalyzed condensation reactions. In the aqueous phase crotonic acid can partly dissociate (crotonic acid pK a = 4.69 [19] and water proton affinity 681.9 kJ mol À1 [20] ), lowering the pH of the reaction media and activating the vinylic bonds of the crotonaldehyde and crotonic acid. Furthermore, acid catalyzed aldol condensation of the crotonaldehyde could also be enhanced at low pH. For this reason, often oxidation reactions in aqueous environments are performed in alkaline media. [17] The results of crotonaldehyde aerobic oxidation to crotonic acid using the ETS-4 indicated that on decalin it is possible to achieve 58 % yield at 70 % conversion, which is similar to the results obtained on the Co x Ce y O z catalyst. Notably, when the solvent was changed to acetic acid, the conversion and yield significantly increased resulting in quantitative production of crotonic acid (91 mol C %) in the presence of 50 mg of catalyst. This enhancement led to STY that were two orders of magnitude higher (11.3 mol g À1 h À1 ) than those obtained in the Ru Co x Ce y O z catalyst in decalin (0.09 mol g À1 h À1 ). The higher yield and activity of this reaction could be attributed to the formation of peracetic acid. During aerobic oxidation the partially uncoordinated Ti cations can act as Lewis acid sites that activate the formation of peroxide species. These species can be selective oxidizing agents for the conversion of the crotonaldehyde to crotonic acid. Detailed batch reaction experiments were performed for several reaction times, using only 50 mg of ETS-4 catalyst. About 84 % of conversion was obtained during the first three hours of reaction ( Figure S1). At this point the rate of the reaction plateaued. Notably, at this point the selectivities achieved were 92-100 %.
Then, we studied the selective epoxidation of crotonic acid over Sharpless-like catalysts based on tungsten oxide catalysts in the presence of diluted hydrogen peroxide as oxidant at 65 8C (step d in Scheme 1). We employed 13 C-NMR for product identification and HPLC for quantification. The results showed that before reaction only the crotonic acid can be detected by 13 C-NMR ( Figure S2). After 3 h of reaction at pH 2.9 additional chemical shifts appeared in the 13 C-NMR spectra that corresponded to the formation of the epoxide (3methyloxirane-2-carboxylic acid) and a,b-dihydroxy acid (2,3-dihydroxybutanoic acid). At pH 2.9 the molar fraction of epoxide was lower (22 mol %) than that of the a,bdihydroxy acid (37 mol %) (Figure 2). The high yields to the di-hydroxy acid (37 %) in the acidic environment can be ascribed to the hydrolysis of epoxide ring (Table S3). The exact mechanism of this reaction, step e in Scheme 1, is still under debate. Early work on kinetic isotope effects on the hydrolysis of ethylene oxide showed that there are two possible mechanisms (S N 1 and S N 2). [21][22][23] In both mechanisms, the first step is the protonation of the epoxide oxygen in a fast equilibrium step. [21,24] The second step is either the decomposition of the conjugated acid followed by rapid reaction of the carbonium ion with water (S N 1 mechanism) or the bimolecular substitution of the conjugated acid with water (S N 2 mechanism). [25] A similar strategy is applied industrially for the production of ethylene glycol from ethylene oxide, although at higher temperature (200 8C). [26] To reduce the rate of a,b-dihydroxy acid formation we decided to increase the pH to 6.4 by adding KOH to the reaction mixture (Figure 2). The 13 C-NMR results indicated that the two main species present in the reaction mixture after 3 h of reaction were crotonic acid and the epoxide. At these conditions, the conversion of crotonic acid increased to 89 % and the yield of 3-methyloxirane-2-carboxylic acid and 2,3dihydroxybutanoic acid were 76 and 22 %, respectively (see Table S3). Previous reports indicate that increasing the pH beyond 7 leads to faster rates of hydrolysis, [21,24] which in our  case would have led to higher selectivity to 2,3-dihydroxybutanoic acid. These results indicate that nearly neutral pH can effectively reduce the rate of epoxide hydrolysis increasing the selectivity towards the epoxide adduct.
The performance of the WO 3 catalyst prepared using the combustion method was compared with commercial WO 3 (Sigma-Aldrich), and SBA-15 doped with WO 3 prepared by incipient wetness impregnation (Figure 3). The molar distribution after 0.5 h of reaction indicates that with the WO 3 combustion catalyst it was possible to obtain up to 54 mol % of the epoxide (3-methyloxirane-2-carboxylic acid), while in the case of commercial WO 3 and WO 3 -SBA-15 catalysts these values decreased to 15 and 7 %, respectively. With all the catalysts, formation of the a,b-dihydroxy acid (2,3-dihydroxybutanoic acid) was low due to the near neutral pH employed during reaction.
Significantly lower conversions were obtained for the commercial WO 3 and WO 3 -SBA-15, which only yielded 16 and 7 %, respectively (see Table S4). Here, it was observed that WO 3 prepared using the combustion method achieved rates of 9 mol h À1 mol of WO 3 , while in the case of WO 3 supported on SBA-15 and commercial WO 3 catalysts the rates were 6.1 and 2.3 mol h À1 mol of WO 3 , respectively. Diffuse reflectance UV-vis characterization of the as-prepared catalysts indicated that in the WO 3 catalyst the UV-visible absorption edge was shifted towards higher wavelengths in comparison to the WO 3 commercial, which can be assigned to isolated tetrahedral and small oligomers WO x ( Figure S3). In contrast, the UV-vis spectrum of WO 3 -SBA-15 catalyst showed an absorption maximum at 230 nm in combination with a superimposed shoulder at 260 nm, which can be assigned to isolated tetrahedral and small oligomers WO x . Previous kinetic and spectroscopic studies reported by C.
Hammond et al. [28] on epoxidation of cyclooctene in 1-butanol using W-Zn/SnO 2 catalyst showed that the most active and stable species during epoxidation of cyclooctene was not the isolated W iv species, but instead the polymeric and bulk WO 3 phases. This could explain the higher activity of the WO 3 prepared by combustion.
The HPLC analysis of the product mixture obtained from the homogeneous reduction of isolated 3-methyloxirane-2carboxylic acid in aqueous solution indicated that this reaction step can be readily accomplished with very high chemo-selectivity ( % 100 %) at 40 % conversion. These promising results served as basis for the evaluation of the second set of experiments using the un-purified 3-methyloxirane-2-carboxylic acid in the presence of 34 bar of H 2 and 5 wt. % Ru/C catalyst. The reaction mixture was composed of crotonic acid, 3-methyloxirane-2-carboxylic acid, and 2,3dihydroxybutanoic acid ( Table 2).
In summary, the utilization of tandem oxidation, epoxidation, followed by epoxy-ring activation by either hydrolysis or hydrogenation offers a flexible platform for the production of specialty chemical building blocks from biomass-derived ethanol derivatives. This strategy allows for the selective conversion of bio-ethanol to crotonic acid, 3-hydroxy butanoic acid, and 2,3-dihydroxyacid at high carbon yields (62, 21, and 29 %, respectively).
We have successfully demonstrated that using an intermediate step of epoxidation of the unsaturated acid it is possible to selectively produce b-hydroxy acids and a,bdihydroxy acids, which are otherwise inaccessible via direct oxidation of the aldol-adduct. We envision that the heterogeneous catalytic cascade approach that we have developed here can serve as a basis for the production of numerous highvalue chemical building blocks from bio-ethanol. Furthermore, tuning the aldol condensation process to increase the yield towards C 6 and C 8 aldehydes could enable the production of long-chain hydro-acids for the production of longchain PHAs with enhanced mechanical properties.