Continuous Flow Synthesis of Furfuryl Ethers over Pd/C Catalysts via Reductive Etherification of Furfural in Ethanol

Abstract

Furfural has become one of the most promising building blocks directly derived from biomass. It can be transformed into numerous important biobased chemicals. Among them, furfuryl ethers such as furfuryl ethyl ether (FEE) and tetrahydrofurfuryl ethyl ether (THFEE) are considered to be attractive derivatives, notably as fuel components, due to their high stability and high octane numbers. Therefore, the production of furfuryl ethers from furfural via a hydrogenation route is an important academic and industrial challenge and requires the deployment of new catalytic processes under green and competitive reaction conditions. The existing processes are based on a two-step process combining hydrogenation and reaction with a strong Bronsted acid catalyst in batch conditions. For the first time, a continuous flow one-step process has been elaborated for the conversion of furfural directly into furfuryl ethers based on reductive etherification. The present work explores the catalytic performance in a continuous flow of commercial palladium catalysts supported on activated carbon for the catalytic reductive etherification of furfural with ethanol in the presence of trifluoroacetic acid. The chemical and engineering aspects, such as the mechanisms and reaction conditions, will be discussed.

1. Introduction

Lignocellulosic biomass is a renewable, low-cost material and a major challenge for the substitution of a portion of fossil fuel products, whether in the energy sector or in the production of high-value chemical building blocks [1,2]. Moreover, its transformation has a reduced impact in terms of gas emissions, such as carbon dioxide, compared to conventional fossil fuels. Indeed, lignocellulosic biomass is mainly composed of cellulose (40−50%), hemicellulose (20−30%), and lignin (10−35%). Cellulose, a partially crystalline polymer of glucose (C6), is already widely used in many industries as a platform chemical (organic acids, polyols) for plastics and fuels [3,4].

However, hemicellulose and lignin are still little used today, and their current valorization are generally as low added value energetics. Hemicellulose is a heterogeneous carbohydrate fraction that contains a polymer of glucose (C6), arabinose and xylose (C5) that represents renewable source molecules for many chemical applications in food and pharmaceutical industries [5]. Therefore, the use of sugars like glucose (C6) and xylose (C5) as raw materials for the production of important, basic nonpetroleum chemicals is of great interest [6]. For example, furfural is obtained from these resources by dehydration [7], isomerization and thermal dehydration in acidic conditions. Furfural is a key derivative in the production of fuels, plastics and many ranges of important nonpetroleum-derived chemicals for pharmaceuticals and cosmetic applications. With an annual production close to 300,000 t/year and at a reasonable price (250 €/t), furfural is one of the few biobased chemicals in competition with basic petrochemicals such as toluene and benzene (225−250 €/t) [2,5].

In the past few years, furfural has become one of the most promising building blocks of the biobased chemical industry. It can be transformed into furfurylamines via reductive amination [8,9], into maleic and fumaric acid by catalytic oxidation [10], or into monomers like 1,2-pentanediol and several valuable products by hydrogenation [11,12,13]. In recent years, a special attention to the hydrogenation route of furfural has been given by many researchers because it provides a wide range of bulk chemicals that are often used as intermediates to synthesize many products and fuels such as furfuryl alcohol, 2-methyl tetrahydrofuran, 2-methylfuran, furan, tetrahydrofurfuryl alcohol and tetrahydrofuran [5]. In addition, furfuryl ethers such as furfuryl ethyl ether (FEE) and tetrahydrofurfuryl ethyl ether (THFEE) are considered attractive kinds of furfural derivatives, notably as fuel components, due to their high stability and high octane numbers [5,14].

Nowadays, furfuryl ethers are produced in two stages: first, furfural is hydrogenated to furfuryl alcohol or tetrahydrofurfuryl alcohol, which is then etherified in the second step with a desired alcohol. This pathway requires a bifunctional hydrogenation catalyst and a strong Bronsted acid catalyst. A more appealing pathway would be to produce such ethers directly from furfural and alcohol in one pot by means of a reductive alkylation reaction [15].

In a previous work on the production of biodiesel candidates from 5-(hydroxyme-thyl)furfural (HMF), Balakrishnan and co-workers showed the possibility of transforming HMF in batch conditions directly into 2,5-bis(alkoxymethyl)furan at 60 °C and 1.37 MPa pressure using a mixture of Amberlyst-15 and a Pt-supported alumina catalyst [16]. Consecutively, Lewis JD and co-workers developed a continuous flow strategy for the coupled transfer hydrogenation and etherification of HMF using Lewis Acid Zeolites (Hf-, Zr- and Sn-Beta zeolites), without using external hydrogen sources or precious metals [17]. On the other hand, Wang and co-workers used Pd nanoparticles supported on active carbon for the transformation of furfural into furfuryl ethyl ether (FEE) in ethanol. They achieved up to 81% yield at a low reaction temperature (60 °C) and H₂ pressure of 0.3 MPa [18]. However, this process suffers from catalyst degradation and substrates limitations.

Similarly, approaches for the Pd-catalyzed synthesis of alkyl glycerol or aromatic ethers from carboxylic acids or ketones were described by Sutter et al. [19]. In addition, Dan Wu and co-workers developed the reductive alkylation of aldehydes among which furfural was tested with several alcohols under mild reaction conditions involving in situ Brönsted acidity generation in the presence of hydrogen over a heterogeneous Pd-I catalyst [20]. Therefore, it appears clearly that the reductive alkylation of furfural with alcohol is feasible in one step. Nevertheless, no available study permits us to clearly rationalize the formation of ethers from furfural with commercial catalysts or how the reaction conditions could help to select one product to another, and especially how to select the production of the tetrahydrofurfuryl over the furfuryl ether. Another aspect that could be of interest in designing such a one-step process is the development of a continuous flow methodology that allows for control over the reaction process, and, therefore, fine product tuning. In fact, continuous flow processes are considered as promising tools for the transformation of biomass into value-added chemicals with great potential for industrial application [21,22].

In this context and in order to elaborate a more sustainable approach, we decided to investigate the reductive alkylation of furfural using ethanol in a continuous flow mode with a commercially available Pd supported on an active carbon catalyst with external hydrogen sources.

2. Results and Discussion

In this work, the main objective was to achieve the highest selectivity to furfuryl ethers, such as ethyl furfuryl ether (FEE) and tetrahydrofurfuryl ethyl ether (THFEE), through simultaneous etherification of an aldehyde with ethanol and the ring hydrogenation of furfural in a continuous flow condition. This was possible by choosing a suitable acid to first form the acetal. Here, we chose to study Trifluoro acetic acid (TFA) as a homogeneous catalyst for the in situ production of the acetal. Secondly, we evaluated the choice of catalyst for the reduction and its load, along with reaction conditions such as H₂ pressure, temperature and residence time, that could affect the formation of intermediates during this process.

2.1. Reductive Etherification of FF with Ethanol over the Pd/C Catalysts

The reactions were performed in continuous flow mode with the supply of H₂ by using a fixed-bed plug-flow reactor to study the product distributions, long-term stability of the Pd/C catalysts and deduce their deactivation mechanism in the hydrogenation of furfural. The commercial catalyst Pd/C (5% Escat™ 1431) was used for the first time in the continuous flow for reductive alkylation by Bruniaux and co-workers as well as by Sutter et al. [19,23]. So far, this type of Pd/C-catalyzed reaction has only been studied in batch reactors. As an example, catalyst studies by Wang and co-workers show that under mild reaction conditions, they obtained the highest selectivity to furfuryl ethyl ether (FEE) without the over-hydrogenation of the furan ring to obtain the tetrahydrofurfuryl ethyl ether (THFEE) [18]. Their results show that the conversion and selectivity of furfuryl ethyl ether (FEE) decreased within 4 consecutive runs. Therefore, they supposed that the decline in the activity of the catalyst was due to the change in the nature PdH, or rather the deposition of carbonaceous species at the surface of the catalyst.

As an initial study, we started the reaction without (Figure 1a) and with (Figure 1b), adding TFA at 80 °C and 50 Bar of H₂ with a flow rate of 0.5 mL min⁻¹. For both experiments, 0.2 M of furfural in ethanol solution was flowed into a cartridge (0.88 mL of internal volume), filled with the palladium catalyst (300 mg, Pd/C, 5% Escat 1431) and heated at 80 °C. The total time on stream was 2 h continuously to test the long-term behavior of the catalyst and the product distribution. All samples were taken from the output of the reactor and were analyzed by GC-FID.

The furfural conversion was 100% for both cases. One must note that only the three main product yields were represented in both figures; other side products were produced with a cumulative yield < 10%. In the first reaction without TFA, the predominant products were tetrahydrofurfuryl alcohol (THFA) and 2-(methyltetrahydro)furan (2-MTHF), with a yield of 43% and 41%, respectively. The rest of the product consisted of mainly tetrahydrofurfuryl ethyl ether (THFEE, yield of 6.5%). When using TFA as an acid in the second reaction, the dominating product was THFA for the first 15 min on the stream and the yield was 32–36%, but then it gradually dropped over time, while the THFEE increased to achieve a yield of 41%. However, 2-MTHF remained almost constant during the 2 h on stream (28–33%). It should be noted that by adding the TFA in the solution that contained furfural and ethanol while keeping the rest of the reaction conditions unchanged, the THFEE yield could have been increased from 6.5% to 41%. This result was consistent with the previous report in the literature showing that the reductive alkylation is only efficient in acidic conditions [18]. Indeed, the acid is necessary to catalyze the reaction between furfural and ethanol to form the acetal 2-(diethoxymethyl) furan (DOF), which was continuously reduced in the second step to form furfuryl ethers.

2.2. Study of the Reaction Mechanism

In order to comprehend the mechanism of this reaction, different commercially purchased substrates were used as starting reactants under the same continuous flow conditions, using ethanol as a solvent and with the addition of TFA (300 mg of Pd/C (5%), 70 °C, 50 Bar H₂, 0.5 mL min⁻¹). We analyzed the different product distributions during the reaction on stream and summarized the results in

Table 1. Conversion and product yields of different starting reactants over 45 min continuous flow hydrogenation.

Entry	Substrate	Conv. (%)	2-MTHF	THFA	FEE	THFEE	Others
1	DOF	100	16	-	-	65	19
2	FEE	100	18	-	-	69	13
3	FA	100	39	53	-	-	8
4	THFA	0	-	100	-	-	-

.

We started the reaction with DOF to verify if it was the precursor to form furfuryl ethers (FEE) and (THFEE). As shown in Table 1, the predominant products were THFEE (entry 1). Even when starting from DOF or FEE, the results were similar (entry 2). These results demonstrate that THFEE was formed by the hydrogenolysis of DOF to FEE in the first stage, which is consistent with the prior reports, as noticed in the synthesis of other ethers [24,25]. In the second stage, ring hydrogenation of FEE to obtain TFFE occurred, which continued towards 2-MTHF production via over-hydrogenation.

In the case of FA as a starting reactant (entry 3), no THFEE was formed and we obtained only two main products THFA and 2-MTHF with a yield of 53% and 39%, respectively. When THFA was used as the reactant in the same conditions, no conversion was achieved.

Accordingly, we can conclude that as it was previously reported, two different and competitive routes are possible for the hydrogenation of furfural (Scheme 1). One will lead to FA and THFA with a production of 2-MTHF from the FA. The second will lead to THFEE via the consecutive formation of DOF and FEE. This reaction pathway shows the importance of the in situ formation of DOF via the acetalization of furfural with ethanol for the production of furfuryl ethers FEE and THFEE, the latest being then partially over-hydrogenated in MTHF.

One possible improvement in the production of the ethers would be to conduct the reaction directly onto the DOF acetal, though it is always in equilibrium with furfural. Therefore, we decided to evaluate its formation from furfural and ethanol.

2.3. Study of the TFA-Catalyzed Furfural Acetalization with Ethanol

Acetalization is one of the most effective methods for the protection of aldehydes, which is commonly come across in organic synthesis [26,27]. The acetals are generally produced by combining aldehydes with alcohols in the presence of catalysts such as p-toluenesulfonic acid, H₂SO₄, TFA and dry HCl [28,29]. Mikhail Yu’s group reported their preparation of furfural diethyl acetal via the direct interaction of the furfural with a slight excess of ethanol in the presence of calcium oxide (CaO) at room temperature [30].

The acetalization of furfural is a reversible reaction. Moreau et al. reported that the in situ generated water hydrolyzed the acetal formed via a pseudo-first order mechanism [31]. To eliminate the backward hydrolysis reaction and increase furfural conversion to shift the acetalization equilibrium (Scheme 2), a large excess of ethanol should be used in comparison to the furfural, up to 50 to 100 times more than the stoichiometry ratio (molar ratio EtOH:FF = 100:1) [32].

In order to produce DOF in the highest quantity from furfural, we investigated the yield of furfural acetalization with ethanol using different TFA loading at ambient temperature. The results obtained in the screening of TFA loading are reported in Figure 2.

We first noticed that after 90 min of reaction, we reached an equilibrium in DOF/FF concentration and that without TFA in this time length, the production of DOF was negligible. It appeared that the maximum ratio (DOF/FF) 2.30 was obtained for a TFA loading of 0.03%; further increase in the TFA amount did not really improve the ratio (2.31). Thus, to further study our production of ethers in ethanol we decided to feed our continuous flow with a pre-acetalized solution that consisted of a mixture of DOF and furfural.

Before studying the variation of the reaction conditions, we evaluated the variation of TFA amount on THFEE production by feeding it with the DOF/FF mixture. Figure 3 shows a plot of product distribution at 90 min from the stream for the catalytic reduction in the different balanced solutions (FF and DOF) in different concentrations of TFA. As can be seen in Figure 3, TFA presence is still necessary in the media to improve the THFEE yield and decrease the formation of THFA. Adding more TFA than the initial 0.02% did not improve the yield. The only other major compound found was the 2-MTF.

2.4. Hydrogen Pressure and Temperature Effects on the Reductive Etherification of FF

Having established the pathway of the reaction and proposed a feeding of DOF/FF solution in the presence of TFA, we tried in the next step to optimize the reaction conditions by studying temperature and pressure effects on product selectivity, especially for producing THFEE. First, we investigated the effect of different hydrogen pressures on the THFEE yield (Figure 4).

The reactions were operated under different pressures of H₂ 1, 6, 10 and 15 Bar at 70 °C with a flow rate of 0.5 mL min⁻¹, using Pd/C (5%) as catalyst and an addition of 0.01% of TFA. All samples were taken from the output of the reactor at 30 min from stream and were analyzed by GC-FID. Interestingly, FEE was the dominant product with 53% of yield, and just 6% of THFEE could be obtained under a low pressure of hydrogen (1 Bar). Nevertheless, when the reaction was conducted at higher pressure (6–15 Bar), the yield of FEE decreased rapidly at 3–7% whereas the THFEE yield increased gradually to reach a maximum of 41% at 15 Bar.

For a better understanding of the reaction process, we decided to investigate the temperature effects on the reductive etherification. In this step, we increased the temperature from 40 °C to 110 °C while keeping the hydrogen pressure unchanged (10 Bar). In this case, THFEE yield was almost the same at 40 °C (37%) and at 100 °C (39%), with a slight increase at 90 °C (42%). The best THFEE yield of 45% was obtained at 70 °C.

2.5. Variation of the Quantity of Pd/C Catalysts

As the temperature variation had a small impact, we then evaluated the catalyst loading effect.

Table 2. Continuous reductive etherification of equilibrium solution (FF-DOF) in ethanol at different Pd/C percents.

Entry	Pd/C (%)	Yield (%)
		FA	THFA	2-MTHF	FEE	THFEE	Others
1	0.7	0	16	0	57	4	23
2	1.25	9	3	6	38	15	29
3	2.5	12	4	8	20	22	34
4	3	7	6	25	9	34	19
5	5	3	9	26	0	46	16

Reaction conditions: furfural (4 mmol) and DOF (10 mmol) with TFA 0.01% in ethanol (70 mL); 70 °C; 15 Bar H₂; 100 mg catalyst; 60 min.

displays the catalytic performance of different catalysts percentages in the continuous flow system for the reductive etherification of an equilibrium solution that contains furfural and DOF in ethanol. Variation of the Pd percentage was obtained by mixing designated amounts of carbon, without Pd, together with the catalyst. Absence of the catalytic activity of carbon without Pd was verified in advance.

In this part, the reaction conditions were set at 70 °C and 15 Bar H₂ pressure with a 0.5 mL min⁻¹ flow rate. For each experiment, a pure ethanol solution was pumped through a cartridge (length 30 mm, internal diameter 4 mm, internal volume 0.38 mL). As we started from the FF/DOF mixture, we decided to fill it with less catalyst powder (100 mg). When the pressure and temperature were stabilized, the feed was switched to an ethanol solution (70 mL) containing furfural (4 mmol) and DOF (10 mmol) at equilibrium with 0.01% of trifluoroacetic acid. The first sample was collected and analyzed after 60 min under the flow. Furfural and DOF conversion were 100% for all the percentages of Pd/C catalysts.

It was noticeable that the five different percentages of palladium in the Pd/C catalyst had different product distributions (Table 2). For entry 1 (0.7% of Pd), the dominating product was furfuryl ethyl ether (FEE) with a yield of 57%. The other products were tetrahydrofurfuryl alcohol (THFA, yield of 16%) and tetrahydofurfuryl ethyl ether (THFEE, yield of 4%). When 1.25 of the palladium percentage was used (entry 2), the main products contained 38% of FEE, 15% of THFEE and 9% of furfuryl alcohol FA. When the palladium loading was increased to 2.5% (entry 3), the major product was THFEE with a yield of 22%. Other by-products included FA (12%), THFA (4%), 2-(methyltetrahydro)furan (2–MTHF, yield of 8%) and FEE (22%). Interestingly, the predominating product was THFEE in entries 4–5 with a yield of 34 and 46%, respectively. We noticed that by increasing the catalyst percentage from 0.7% to 3 or 5%, the THFEE yield increased from 4% to 47%, and the yield of FEE decreased simultaneously from 57% to not detected. Alongside the increase in the THFEE yield, a considerable increase of 2-MTHF yield from 6% to 26% was also pointed out, resulting in the over-hydrogenation of the intermediates. Thus, it should be noted that palladium loading significantly affects the selectivity of the reaction. This result clarified the mechanism hypothesis that we increase the selectivity towards the formation of the second ether THFEE by increasing the catalyst percent to a certain amount, and under specific temperature and pressure conditions.

2.6. Stability of Pd/C Catalysts

Finally, to evaluate the catalyst stability over a longer lifetime we performed a reaction in similar conditions over 8 h. In this reaction, conditions were set at 70 °C and 15 Bar H₂ pressure with a 0.5 mL min⁻¹ flow rate. A cartridge (length 30 mm, internal diameter 4 mm, internal volume 0.38 mL) was filled with Pd/C (1%) catalyst powder (100 mg). After stabilizing the pressure and temperature with pure ethanol, the feed was switched to an ethanol solution (70 mL) containing furfural (4 mmol) and DOF (10 mmol) at equilibrium with 0.01% trifluoroacetic acid. The samples were collected every 30 min under the flow and analyzed. Furfural and DOF conversion were 100% throughout all the experiments.

Over an 8 h period of utilization, the catalyst activity was relatively stable, with a slight decrease of the yield of THFEE going from 56 to 52% (Figure 5). FEE yield also slowly decreased, and simultaneously a slight increase in the THFA production was observed.

3. Experimental

3.1. Materials and Methods

Alcohol EtOH was purchased from Fisher Scientific (Waltham, MA, USA). Furfural, 2-methyltetrahydrofuran, trifluoroacetic acid, furfuryl alcohol, and 2-(ethoxymethyl)furan (FEE) were purchased from Acros Organics (Geel, Belgium). Pd/C powder (5% Escat™ 1431) was purchased from STREM Chemicals (Bischheim, France)). LTo prepare lower Pd containing catalyst the commercial catalyst was simply mixed with C powder. Tetrahydrofurfural alcohol, 2-(diethoxymethyl) furan and tetrahydrofurfuryl ethyl ether (THFEE), were purchased from Merck Sigma-Aldrich (Saint Quentin Fallavier, France). All materials were used as received without further purification.

3.2. Continuous Reductive Etherification

In an ordinary experiment, a commercial continuous flow system (ThalesNano H-cube Pro™) was used (ThallesNano Inc., Budapest, Hungary). This apparatus allowed for a variation of the pressure, the temperature and the residence time. First, a pure alcohol solution was pumped through the system (0.5 mL min⁻¹) until the pressure and temperature were stabilized. Then, the feed was switched to an alcohol solution (70 mL) containing furfural (1.159 mL, 14 mmol). The solution was passed through a cartridge (length 70 mm, internal diameter 4 mm, internal volume 0.88 mL) filled with Pd/C powder (300 mg, 5% Escat™ 1431) under the presence of H₂ (30 Bar) at (60–100 °C). The total time on flow was 2 h. The first sample was collected after 15 min under the flow. Each experiment was conducted at least 3 times, and no variation in yield higher than 5% was observed.

3.3. Product Analysis

Samples (0.5 mL) were collected at the outlet and mixed with ethyl acetate solution containing toluene as an internal standard. This mixture was analyzed through gas chromatography coupled to a flame ionization detector (GC-FID) using a Perkin Elmer Autosystem XL instrument (Perkin Elmer, Villebon-sur-Yvette, France) equipped with an autosampler and a 30 m AT™-1ht capillary column (0.25 mm diam, 0.1 μm film thickness), and using nitrogen as a carrier gas with a constant rate of 2 mL min⁻¹. An injection volume of 1 µL was performed at inlet temperature of 350 °C, split ratio (8:1), the oven was set at 40 °C (2 min) then ramped up to 250 °C (20 °C·min⁻¹), and finally held at 250 °C for 3 min. The recognition of the substrates was carried out by identification of the retention times with pure standards, and the quantitative analysis was achieved using the internal standard method. The mass balance, FF conversion, and the yields to each corresponding product are defined as follows:

$$X\%={\displaystyle \frac{C{FF}_0-C{FF}_n}{C{FF}_0}}\times 100$$

$$YieldP\%={\displaystyle \frac{C{P}_n}{C{FF}_0}}\times 100$$

CFF₀, CFF_n and CP_n refer to the initial and final concentration of furfural and the final concentration of the product (mol L⁻¹), respectively. The total volume in the system was supposed to remain constant during the reactions.

4. Conclusions

In summary, we successfully developed a novel catalytic method for the direct synthesis of furfuryl ethyl ether by the reductive etherification of furfural in ethanol using continuous flow as alternative technology. Usually, furfuryl ethers are obtained by two steps in batch mode, via procedures involving the reduction of furfural into furfuryl alcohol, which is further etherified [18]. Several authors have studied the utilization of various mixtures of heterogeneous catalysts (mostly synthetized) containing both hydrogenation and Brønsted acid sites in a one-step process [16,17,18,33]. All these reports are mainly focused on producing furfuryl ethers and seldom propose tetrahydrofurfuryl ether production. Usually, these methodologies allow good selectivity in furfuryl ethers (up to 80%) on a first batch, but with a significant decrease when reused and problematic increases in furfuryl alcohol by-product. Here, our process presents an advantage to use a simple commercial Pd/C catalyst that offers an eventual industrial application of this process and allows for the synthesis of furfuryl as well as tetrahydrofurfuryl ethers under mild conditions of temperature and hydrogen pressure with a slight concentration of TFA. The optimized conditions (pressure and Pd loading) can be efficiently tuned to allow for selecting the ether formed: FEE at low pressure and low Pd loading and THFEE at higher pressure and Pd loading. When producing THFEE, an over oxidation of 2-MTHF is observed, but it is worth noting that MTHF is also a valuable product and could easily be isolated from THFEE by simple distillation. Moreover, we show that the selectivity is relatively stable over a long period of time without an over-production of by-products. Finally, our process is run in continuous flow conditions which has been, to our knowledge, reported for the production of furfuryl ethers only once, very recently, by Nuzhdin et al. [34]. Our study therefore appears complementary, as it focuses on the production of ethyl ethers, which was not evaluated in their work.