Oxidation of Furfural and Furan Derivatives to Maleic Acid in the Presence of a Simple Catalyst System Based on Acetic Acid and TS-1 and Hydrogen Peroxide.

Titanium silicate molecular sieve (TS-1) and acetic acid efficiently catalyze the oxidation of furan and furan derivatives to the corresponding maleic acid (MA) in very good yields using hydrogen peroxide as an oxidizing agent. The effect of various solvents, the effect of temperature, reaction time, concentration of hydrogen peroxide, and quantities of the catalyst on the MA yield was studied. With the best conditions, MA is the sole product obtained after a fast and simple purification by filtration and evaporation. Compared to the previously reported methods, this work is a good compromise between the different reaction parameters and offers a good alternative to the production of biosourced MA.

Introduction

Four carbon 1,4-diacids are at the center of intensive research. Maleic (MA), malic, fumaric (FA), and succinic (SA) acids are among the top 12 biobased molecules of the DOE’s list. They are widely used in the chemical industry, through an important panel of applications, for example, production of paints and coating, (un)saturated polyester resins, plasticizers, copolymers, lubricant additives, and preservatives in food and beverages.[1,2] They can be used to replace products from the petroleum-derived feedstock, such as phthalic anhydride or adipic acid. MA is also particularly interesting as a key building block because of its ability to be transformed into 1,4-diacids mentioned previously and other products like maleic anhydride, butanediol, and tetrahydrofuran (Scheme ).[3,4]

Scheme 1

Outline of Potential Chemical Derivatives from Maleic Acid

As MA is the hydrated form of maleic anhydride the most widespread derivative, the MA market can be studied through maleic anhydride production. Currently the production of maleic anhydride is estimated to be more than 2000 kt per year. Industrial production of maleic anhydride primarily needs petrochemical fossil resources. The most important processes are a selective vapor-phase oxidation of butane,[5,6] insaturated C4 (butene),[7,8] and benzene,[9] using O2 as an oxidant. The main byproducts of these transformations are CO and CO2. Then maleic anhydride is converted into MA by hydration processes. In previous years, the price of maleic anhydride ranges from 1.5 to 2.00 $ per kg.

Furfural and 5-hydroxymethylfurfural (HMF) are biomass-derived molecules. These compounds are obtained by dehydration under acidic conditions of sugars such as xylose or arabinose contained in hemicelluloses and lignocellulosic agro residues. Currently, annual production of furfural is estimated around 250–400 ktons, mainly produced in China.[10]

Many research groups have explored oxidation of furfural and HMF to MA or maleic anhydride under various conditions with homogeneous or heterogeneous catalysis in a gas or liquid phase.

In the liquid phase of several oxidation processes, it is often hydrogen peroxide which is used as an oxidant.[11−14] Hydrogen peroxide is regarded as an ideal oxidant in terms of atom economy, availability, and green metrics, so it is particularly relevant as a green oxidant for safe industrial processes in a continuous flow mode.[15] The main disadvantage of hydrogen peroxide is its instability and the risk of an explosion from concentrated solutions by disproportionation. By comparison, when oxygen is used as an oxidant, the temperature between 363 and 383 K and a high pressure of 10–20 bar will be needed.[16−22]

The specifications of the transformation are very complex. This still remains a challenge for the industrialization of a process. Selectivity, yield, reaction time, energy, stability, and recyclability of the catalyst, weight percentage of furan derivatives, and industrial hygiene are among the criteria that need to be adequately controlled. Previous results only partially achieved the objectives; there remains a lot of work to do.

For example, Shi et al. reported the optimum yield (49%) and selectivity of MA (52%) obtained with the mixture of heteropolyacids and copper nitrate H3PMo12O40 + Cu(NO3)2 (ratio: 2/1) as catalysts. The oxidant is oxygen, with a pressure of 20 atm. The solvent is water, temperature is 371 K, and reaction time is 14 h. However, MA extraction of the reaction mixture and recyclability of the catalyst were not reported.[18]

In 2014 and 2017 Fagundez et al. reported the selective aqueous-phase oxidation of furfural to MA using hydrogen peroxide as an oxidant and titanium silicalite as a catalyst.[23,24] Under optimal reaction conditions, MA was obtained in a high quantity. For example, a 78% yield of MA was reported using an H2O2/furfural molar ratio of 7.5, with a furfural/TS-1 wt % ratio of 1.0 at 323 K during a 24 h period of the reaction time. Furfural (4.6 wt %) was used. Moreover, 92% of MA was reported, using a two-step sequence of reactions conducted with two catalysts TS-1 and Amberlyst 70 consecutively. In this case, the H2O2/furfural molar ratio was 4.4, at 323 K and the reaction time was 52 h. Although Ti leaching was observed, the catalyst was reused for six runs without noticeable deactivation. In addition, interesting results have been reported by Wang et al. who demonstrated efficiency and selectivity of furfural oxidation with formic acid to get 2(5H)-furanone with a small amount of SA and MA.[11] Recently, with a different protocol and conditions, Zhang et al. have demonstrated the possibility to get an excellent selectivity and yield of MA by oxidation of furfural using formic acid as a solvent and hydrogen peroxide as an oxidant. However, a small amount of furfural (1.5 wt %) was used, and the effect of the furfural concentration was not reported. Best conditions were observed in a sealed tube with a H2O2/furfural molar ratio of 6 at 60 °C in 4 h which yielded 95% MA.[12] Among the latest contributions in this line of research, this year, Yang et al. reported the synthesis of MA from furfural in the presence of a KBr/graphitic carbon nitride catalyst and hydrogen peroxide under optimized conditions a very good yield of 70% MA is obtained.[25]

Other research teams have reported MA production with an acidic catalyst in a liquid phase, but MA was produced in lower yields with a mixture of products.[13,14] Various representative oxidation methods are summarized in Table , clearly showing the advantages and disadvantages of each.

Table 1

Various Oxidative Conditions of Furan Derivatives Reported in Research Papers

furan derivatives	catalyst wt %	wt % fur.a	solv.	m.r. O/Fb	temp. (°C)	time (h)c	product(s) %d	refs
furan	TS-1 2%	10	AcCN	2.4	0–RT	8	5-HFO 98%	(34)
furaldehyde	TS-1 4.6%	4.6	water	7.5	50	24	MA 78%	(23,24)
	Amberlyst-15 1.4%	2.7	water	4	80	24	SA 74%	(13,14)
							MA 11%
							FA 1.9%
	HCOOH 4.6%	11.5	1,2 DCE	2.4	60	3	FO 60.3%	(11)
							SA 12%
							MA 6.3%
	HCOOHe	1.5	HCOOH	10.1	100	4	MA 91%	(12)
	BHCd 16%	4	water	10	100	0.5	MA 61%	(35)
							FA 31%
	KBr/g-C3N4d 2%	2.9	water	f	100	3	MA 70%	(25)
	AcOH and TS-1 2%	4.3	AcOH	8	80	4	MA 60%	this work

Furan derivative loading in wt % with respect to the mass of all the constituents of the reaction mixture.

m.r. O/F: mole ratio oxidant and furan derivative, H2O2/furan derivative.

Reaction time.

BHC: bétaine hydrochloride, 5-HFO: 5-hydroxy-2-(5H)-furanone, FO: 2-(5H)-furanone, MA: maleic acid, SA: succinic acid, FA: fumaric acid, g-C3N4: graphitic carbon nitride.

Solvent and catalyst.

Aq. % of H2O2 used is not reported.

In the liquid phase, these transformations often follow Baeyer–Villiger’s oxidation mechanisms, which were recently demonstrated with high efficiency on levoglucosenone derived from biomass with different biocatalysts in an aqueous or organic and aqueous phase. The catalysts are based on organic acid such as m-CPBA or acetic acid[26,27] enzyme[28−30] or metal-zeolite[31] and even without organic solvents and without catalysts.[32]

Because Taramasso and co-workers have reported the preparation of a titanium silicalite-1 (TS-1, MFI structured) molecular sieve, this catalyst has exhibited very interesting properties in order to facilitate a series of oxidation reactions.[33] TS-1 enables oxidation under mild conditions, and it is produced in an industrial scale and is robust and relatively cheap. After use, pyrolysis of TS-1 at 823 K regenerates its catalytic properties. TS-1 is one of the most efficient solid catalysts, useful in developing sustainable chemical processes.

In this work, we report that TS-1 combined to acetic acid is a very efficient method for oxidation of biomass derivatives such as furfural, hydroxymethylfurfural and furan, to produce MA. The conditions of the double catalytic system have been optimized with furfural to get the highest yield of MA. The use of acetic acid is an advantage because it is a less corrosive acid than other organic acids such as formic acid or p-toluene sulfonic acid, AcOH pKa = 4.7 versus HCOOH pKa = 3.7 or APTS pKa = −6.5. In addition, acetic acid is more cost effective economic and can be biobased. Parameters such as the H2O2/furfural ratio, the catalyst amount, and the temperature have been studied. Finally, a good compromise has been reached between the criteria like reaction time, yield, and furfural concentration leading to MA as the sole product (Table ).

Results and Discussion

The solvent effect is an important parameter during chemical transformations. The solubility of the reagents and the products as well as the miscibility of the different liquid phases play a significant role on the selectivity, kinetics, and the yield of the products. The effect of different physicochemical properties such as the boiling temperature, polarity, ability to form hydrogen bonds, and so forth, was tested via a selection of solvents. As previously reported, and in addition to the role of the solvent in TS-1, this chemistry is crucial in relation to its effects on kinetics. It assists the sorption/desorption of reagents/products in TS-1 (passive role) and participates in the catalytic cycle, through interactions with polar species involved in it (active role).[36] To compare the solvent effect, all the other parameters of the reaction were kept constant (concentrations, temperature, stirring). To do this, we used a Radleys Carousel Station device which allows a six batch simultaneous reaction in parallel under the same conditions (Table ).

Table 2

Catalytic Oxidation of Furfural in Different Solventsa

entry	solvents	conversion of furfural	yield of 5-HFOe	yield of MA
1	H2O	54	32	16
2	EtOH	59	10	4
3	AcOH	100	b	59
4	DMSO	d	0	c
5	acetone	55	28	7
6	CH3CN	61	30	6
7	THF	d	b	b
8	AcOEt	76	36	11
9	CH2Cl2	82	19	12
10	cyclohexane	100	0	43

Reaction conditions: 2.6 mmol of furfural, 14.0 mmol of H2O2 (35% aq. sol.), equivalent to H2O2/furfural mol ratio = 5.4, 0.1 g of catalyst TS-1, 5 mL solvent, 80 °C or reflux, 4 h.

Only traces were observed, yield < 3%.

DMSO was oxidized in dimethylsulfone.

Almost no conversion was measured <3%.

5-OHF = 5-hydroxy-furan-2(5H)one.

As we can see in Table , after a reaction time of 4 h at 80 °C, almost all the solvents gave a low yield of MA between 4 and 16%. By contrast, a very good result was observed with acetic acid, which gave 59% of yield, followed by cyclohexane which gave 43%. It is remarkable that after 4 h of reaction most of the solvents with the exception of acetic acid and cyclohexane gave a higher yield of 5-hydroxy-furan-2(5H)one (5-HFO) than MA, meaning that the oxidation reaction is slower, less selective, and partial. With the exception of water aprotic solvents like acetone, acetonitrile and ethylacetate seem to favor the formation of 5-HFO (entries 5, 6, 8, 9). In the particular case of dimethylsulfoxide, the latter is in competition with furfural in the oxidation reaction and dimethylsulfone was obtained exclusively. Note that with organic solvents immiscible with the aqueous phase of hydrogen peroxide, a very good or total conversion of furfural was observed (entries, 8, 9, 10). MA is very soluble in the aqueous phase[37] (443 g L–1 at 25 °C); therefore, a complete oxidation of furfural into MA leads to the enrichment of this phase. However, the intermediate oxidation products are more soluble in the organic phase. In the case of 5-HFO, its migration and its concentration in the organic phase explains the measured yields for this product when an immiscible solvent is used. The production of 5-HFO is evidence of the oxidation reaction mechanism, which certainly starts with a Baeyer–Villiger oxidation. We will see later further evidence in favor of this hypothesis.

Acetic acid and acetonitrile are two solvents often reported and described as the best in the conversion and oxidation of HMF and furfural to maleic anhydride under homogeneous or heterogeneous catalysis conditions,[17,19,20] and acetic acid is found to be a particularly effective solvent in the presence of a vanadium catalyst to obtain MA.[21,22]

On the basis of these initial results, we investigated the optimal conditions for the transformation of furfural in the presence of acetic acid and TS-1 using hydrogen peroxide as an oxidant.

Several oxidant species are present under these conditions. Hydrogen peroxide reacts with acetic acid to give peracetic acid.[38,39]

The mixture of hydrogen peroxide and peracetic acid should be a stronger oxidizing system than hydrogen peroxide alone because the oxidation potential is pH-dependent following the Nernst equation (E = E0 – 0.059 pH). Moreover, it has been established that Ti(IV) center in the TS-1 catalyst is oxidized by hydrogen peroxide in several Ti-hydroperoxo and Ti-peroxo species which act as strong oxidizing agents (Scheme ).[40]

Scheme 3

Titanium Silicalite Activation with Hydrogen Peroxide and Ti-Hydroperoxo, Ti-Peroxo Equilibrium

Note that formic acid is released during processing. The peracid of formic acid can therefore, also partially participate in the oxidation process (Scheme ).

Scheme 2

Acetic Acid and Formic Acid, Peracid Equilibrium in the Presence of Hydrogen Peroxide

The comparative study of MA obtained in the presence or absence of the catalyst TS-1 in acetic acid, clearly demonstrates the gains obtained by this double catalyst system. The result observed after 1 h of the reaction is particularly interesting; the yield of MA is almost twice as high as the reaction carried out without TS-1, and it is slightly higher than the yield of MA obtained in 4 h of the reaction without TS-1. TS-1 in the presence of acetic acid has an effect on both the kinetic parameters and global yield of the reaction. Under these conditions, the catalytic contribution of TS-1 is essential. The conversion of furfural is very fast in all cases, after half an hour of reaction time almost all the furfural was consumed, 83% in AcOH and 96% in the mixture of AcOH and TS-1. There is no more furfural after 1 h of the reaction. Traces of 5-hydroxy-furan-2(5H)-one and other secondary intermediates were observed but not quantified.

In comparison with other oxidation methods, it is important to note that under our conditions, no cis trans isomerization of MA was observed.[41] It is well known that acidic conditions with heating catalyze the reversible addition of protons which induce a free rotation of the central C–C bond and the formation of more stable FA.

First the effect of the amount of the TS-1 catalyst was studied at 80 °C, in the presence of 8 equiv of 35% aqueous hydrogen peroxide in acetic acid (Table , entries 1–4). Without TS-1, 42% of MA was produced after a reaction time of 4 h. When an amount of 0.9 wt % of TS-1 is added, the yield of MA increases by 7 or 49% MA. Doubling the amount of the catalyst to 1.8 wt % further increases the yield of MA to 60%. In view of these results, the mass quantity of the catalyst was further doubled to 3.6 wt %, but after 4 h of reaction a near equivalent yield of MA at 61% was observed. As we have seen in the comparative study of oxidation of furfural in the presence or absence of catalyst TS-1 in acetic acid. The conversion of furfural and the yield of MA are more rapidly achieved in the presence of TS-1 (Figure and Table , entry 4). However, increasing the amount of the catalyst above 1.8 wt % does not make it possible to increase the yield of MA beyond 60%. We can conclude that a small amount of the 1.8 wt % catalyst is sufficient to achieve the maximum MA yield in a reasonable time.

Table 3

Optimization Parameters in Batch Mode

entry	H2O2 (equiv)	TS-1 (wt %)	T (°C)	time (h)	conv. (%)	yield MA (%)
1	8	0	80	4	100	42
2		0.9	80	4	100	49
3		1.8	80	4	100	60
4		3.6	80	4	100	61
5	8	1.8	60	1	100	22
6				4	100	39
7	8	1.8	100	1	100	48
8				4	100	62
9	4	1.8	100	2	100	31
10				4	100	32
11	6	1.8	100	2	100	52
12				4	100	59
13	10	1.8	100	1	100	53
14				2	100	57
15				4	100	62

Figure 1

Comparison of the MA yield (blue and red lines) and furfural conversion (green and purple lines) depends on the catalytic system AcOH vs AcOH and the TS-1 catalyst. Reaction conditions: 2.6 mmol of furfural, 14.0 mmol of H2O2 (35% aq. sol.), equivalent to H2O2/furfural mole ratio = 5.4, with and without 0.1 g of catalyst TS-1, 3.8 mL AcOH, 80 °C, 4 h.

The effect of the reaction temperature has also been studied. Increasing the temperature from 60 to 100 °C makes it possible to obtain the maximum yield of MA more quickly. For example, with 8 equiv of hydrogen peroxide at 100 °C in 1 h, 48% of MA will be obtained, more than double the amount of MA obtained at 60 °C in 1 h (Table , entries 5 and 7). Note that the maximum yield of MA is 62% even after 4 h of the reaction at 100 °C. The extension of the reaction time to 8 h does not increase the yield, but no byproduct or degradation of MA has been observed, which provides good information on the stability of MA under these conditions. No isomerization to FA occurs.

Finally, we studied the effect of the variation with the equivalent number of hydrogen peroxide (Table , entries 1–15). Based on the stoichiometry of the reaction, at least 3 equiv of hydrogen peroxide were required to oxidize furfural to MA. Nevertheless, when a little more than 4 equiv of hydrogen peroxide were used, a maximum of 32% of MA was obtained. We know that hydrogen peroxide begins to decompose quite rapidly by disproportionation into water and oxygen when the temperature reaches 60 °C. Therefore, we can expect a loss of the oxidant by degradation under our conditions. However, it should be noted that hydrogen peroxide under such conditions has finally been determined as being much more stable than expected. An excess of the oxidant is still necessary. In fact when the experiment is performed at 100 °C, 10 equiv H2O2 will be of an adequate amount and necessary to achieve the maximum possible yield of MA in 2 h (Table , entry 14). Working with 6 equiv of the oxidant leads to a good yield of MA, between 52 and 59%, which means that a decrease in the amount of hydrogen peroxide below this amount is not appropriate. Because hydrogen peroxide is relatively inexpensive, in order to be sure of getting the maximum MA yield, a good compromise is to use 8 equiv of hydrogen peroxide. Finally, we can conclude that the greater the amount of the TS-1 catalyst and the higher the amount of the oxidant and temperature, the faster the MA is produced to a maximum yield of 60%. Note that even under the harshest conditions, 100 °C and 10 equiv of hydrogen peroxide, the maximum yield is achieved without overoxidation and without isomerization of the MA formed.

Because titanium silicalite (TS-1) is a heterogeneous catalyst, the process for purifying the MA obtained consists of a simple filtration followed by evaporation of the reaction medium. The purification process is crucial when taking into account an industrial process. It must be the simplest and most economical possible. Our method is very relevant from this point of view.

Although a small amount of the catalyst was used, rather than systematically discarding it, we tried to recycle it. However, during the reaction, the magnetic bar used for stirring crushed and divided the catalyst into very fine particles, which resulted in the end of the transformation to obtain a paste that could be filtered on Celite to separate the reaction mixture, but not completely on sintered glass. Nevertheless, at least 50% of the catalyst TS-1 was recovered after filtration, leaving us no choice but to add a new catalyst to complete and restart a new cycle. The centrifugation of the reaction mixture could have been a solution but it is not adapted to a simple industrial process, so we have discarded it. In addition, knowing that TS-1 tends to slowly lose some of its titanium, we gradually added a portion of the fresh catalyst during the cycles; this allowed us to maintain a maximum yield from the transformation by saving a complete new addition. This could be observed over 4 cycles while maintaining the maximum yield of MA using the optimum conditions at 80 °C, 8 equiv of hydrogen peroxide, 1.8 wt % of the catalyst for 4 h. Note that under these conditions the amount of hydrogen peroxide remaining is less than 0.015 mol dm–3 or 0.5% of the initial hydrogen peroxide, which was not a problem during small-scale evaporation. On an industrial installation, continuous evaporation should be provided to avoid concentration of residual hydrogen peroxide and add a control system.

Because the catalyst was being ground during the reaction its macroscopic appearance changed, and from this we wanted to know if morphological differences also appeared on the microscopic scale. We therefore, analyzed catalyst samples before and after the reaction by electron microscopy. In order to eliminate any traces of organic residues after the reaction, we carried out a heat treatment at 823 K of the catalyst used before.

As can be seen from the electron microscopy images, we cannot distinguish any difference in morphology between the catalyst before and after the reaction. No chemical degradation was detectable, so the structural integrity was therefore preserved (Figure ).

Figure 2

SEM images of TS-1 (a,b) before synthesis, (c,d) after recycling followed by pyrolysis at 823 K, 6 h.

The method was extended to other furan derivatives, under optimized oxidation conditions of furfural with 1.8 wt % TS-1, with 8 equiv of 35% aqueous hydrogen peroxide, at 80 °C for 4 h (Scheme ).

Scheme 4

Oxidation of Different Furan Derivatives under Optimized Reaction Conditions Obtained with Furfural

In all cases the oxidation leads to the formation of MA but in variable proportions, which is significant for the reaction mechanism. The best yield was obtained with 2(5H)-furanone which gave 76% of MA. The use of furan also resulted in a good MA yield of the order of 58%. With furfuryl alcohol only 49% of MA will be obtained. Considering that furfuryl alcohol must first be oxidized to furfural, it requires an additional oxidant equivalent compared to furfural, which could explain the difference in the yield observed with furfural. With HMF which has an additional carbon, it would require at least two additional oxidant equivalents, the yield of MA observed was even lower at 44%. However, with diformylfuran and furoic acid, which are, respectively, the oxidized species of HMF and furfural, the yields of MA were 36 and 19%. These latter results indicate that these compounds are most likely not oxidation intermediates in the reaction sequence while the 2(5H)-furanone is very likely to be one, signifying a mechanism for its formation (Scheme ).

Scheme 5

Possible Pathway and Mechanism of Oxidation of Furfural to MA

[O] is an oxidizing agent, for example, hydrogen peroxide, peracid or titanium silicalite (hydro) peroxo species. Pathway I: refs;[23,42−47] pathway II: refs;[34,48] pathway III: ref (23); and pathway IV: refs.[12−14]

Based on the oxidation mechanisms described in the literature, four types of reaction paths are conceivable (Scheme ). Routes 1[23,42−47] and 2[34,48] resulting from furoic acid as the first oxidized species of furfural are unlikely, in fact the yield of MA obtained from furoic acid is 19%. While following path 2 the yield obtained from furan (58%) made it possible to speculate on the hypothesis of a mechanism involving furan as an intermediate associated with the formation of singlet oxygen generated from peroxo and hydroperoxo titanium species. This path 2 is also excluded because furoic acid is an upstream precursor in this reaction sequence. On the other hand, paths 3[23] and 4[12−14] involving a Baeyer–Villiger oxidation step seem more in agreement with our results. The most convincing element in favor of path 4 being the very good yield obtained from 2(5H)-furanone (76%). Because no intermediate of path 3 between furfural and 5-hydroxy-furan-2(5H)-one has been tested or observed, we cannot exclude this reaction mechanism. At this stage of our investigation, we therefore, believe that MA obtained is essentially derived from reaction path 4.

Conclusions

We have developed an effective method for the production of MA from furfural and other furan derivatives, which are biobased products. The oxidant used was hydrogen peroxide, which is a green oxidant. The developed method is composed of a double catalytic system consisting of acetic acid and TS-1 titanium silicalite. Note that acetic acid is a cheap industrial product and less corrosive compared to other organic acids such as formic acid; moreover, acetic acid can be biosourced from the fermentation of sugar, ethanol, or cornstarch. Titanium silicalite (TS-1) is also produced industrially. TS-1 is advantageous because it is robust and selective for oxidation; moreover, it can be filtered and regenerated by pyrolysis, which facilitates the purification step and may allow some recycling. We have demonstrated that a small amount of this 1.8 wt % catalyst is sufficient with this method. This catalytic system is very effective in accelerating the reaction. In 4 h the total conversion of furfural was observed and a high yield of 62% MA was obtained as the only product without the mixing with FA. A furfural mass proportion of 4.3 wt % was used, which is relatively important compared to certain mass percentages reported in the literature. For all these reasons, we believe that this method is a good compromise between the different parameters, time, temperature, and wt % mass of furan derivatives and catalyst, while maintaining very good yields of MA which can meet the industrial challenge of the production of biosourced MA and is part of a sustainable development context. Based on these results, we are currently working on the development of a continuous flow of MA production.

Methodology of Experiments

All of the reagents are of analytical purity grade and were used without further purification. All products or solvents and reagents were purchased either from Fischer Scientific or Merck.

Preparation and Characterization of TS-1

Catalyst samples were prepared with a slight modification of the procedure described by Taramasso and others.[33,49 ,50] Optimum Ti insertion of [Ti]/([Ti] + [Si]) = 0.025 or 2.5% was targeted. Characterizations by X-ray diffraction, scanning electron microscopy (SEM), and Fourier transform infrared spectroscopy (FTIR) are in agreement with TS-1 characteristics of the literature data. Additional information for catalyst preparation and procedures for characterizations are provided in the Supporting Information.

Analytical Methods

All products were analyzed and quantified by high-performance liquid chromatography (HPLC). Calibration curves were obtained from a commercial sample with the exception of 5-hydroxyfuranone which was prepared following the procedure described by Kumar and Pandey.[34]

General Procedures for Furfural Oxidation

Typically, 2.6 mmol of furfural, hydrogen peroxide (aq. 35%) with a H2O2/furfural molar ratio between 4 and 10 (0.9–3.2 mL), and TS-1 as a catalyst (0–0.2 g) was stirred in 2.8–4.1 mL of solvent depending on the quantity of hydrogen peroxide used at a temperature between 333 and 373 K during 1 to 4 h. After this, the reaction mixture was cooled to room temperature, filtered and evaporated, or filtered and diluted with water in a volumetric flask prior to HPLC analyses.

Furfural conversion and product yields were calculated according the following formulas:

nsubstrate0 corresponds to the mole quantity of furfural initially loaded into the reactor, and nsubstrate and nproduct refer to the number of moles of furfural and products in the reaction mixture at a given time. The HPLC chromatographic factor of the organic products was calculated by analyzing solutions with known concentrations of the different organic products. Additional information is available in the Supporting Information.