Conversion of Cellulose into Formic Acid by Iron(III)-Catalyzed Oxidation with O2 in Acidic Aqueous Solutions.

The conversion of abundant renewable cellulose into versatile formic acid (FA) is a potential process for efficient energy storage and application. Vanadium(V)-catalyzed oxidation with O2 in acidic aqueous media now is the most common method to realize the FA production from cellulose with both high yields and high purity. However, vanadium-based catalysts are difficult to synthesize and expensive. Thus, the seeking for cheaper catalysts with the same high efficiency is expected. In this work, after testing a variety of metal salts in acidic aqueous solution for the conversion of cellulose under O2, iron(III) was found as a cheaper and readily available catalyst for FA formation, with a comparable yield (51.2%, based on carbon) with that of vanadium(V). The effect of reaction parameters was studied. The competition between oxidation and hydrolysis was found and discussed in detail. FeCl3 and H2SO4 can accelerate oxidation and hydrolysis, respectively, whereas suppress the other. The effects can reflect on the product distribution. Intermediates were found and the pathway from cellulose to products was reasonably proposed. The reusability of the catalytic system shows good performance after four runs. The mechanism study suggests a catalytic ability by a mutual transformation between iron(III) and iron(II), where iron(III) oxidizes substrates to iron(II) that is reoxidized by O2.

Introduction

Conversion of biomass into chemicals has drawn much attention as a potentially alternative methodology for a sustainable strategy.[1,2] In view of avoiding the competition with food, the feedstock comes to nonfood lignocellulosic materials. Among lignocelluloses, cellulose is the most abundant resource, thus leading to extensive studies. Formic acid (FA) is an important chemical used widely in traditional industries.[3] In recent years, FA is considered as a potential high-performance material in the fields of hydrogen storage[4−6] and fuel cell.[7−9] Therefore, it is significant to realize an efficient conversion of cellulose to FA.

Wet air oxidation has long been regarded as an effective degradation method of organic wastes, by which CO2, water, and small molecular carboxylic acids are the main products. Jin et al.[10] employed this method to conduct the cellulose oxidation (using H2O2 as the oxidant) at a much shorter reaction time (just few minutes) and higher pH (using alkali). FA was formed in considerable yield (75%) as the formate form. However, the harsh conditions (high temperature, high pressure, and excess alkali consumption) limit further application. Thereafter, a series of vanadium(V) catalysts were found to be very selective on aerobic oxidation of cellulose to FA in a much milder conditions without alkali consumption. Wasserscheid et al.[11−14] and Fu et al.[15] developed a vanadium(V)-containing heteropoly acid (H5PV2Mo10O40)-catalyzed oxidation of cellulose to generate FA yields from 19% to 35% without any other liquid byproducts. Then, a series of vanadium(V)-contained heteropoly acids with different vanadium atom numbers were synthesized. Liu et al.[16] employed H4PVMo11O40 and PVMo11O404–-based ionic liquids for cellulose oxidation, and the maximum FA yield can reach 49.7%. Wang et al.[17] and Wu et al.[18−21] used much simpler vanadium(V) catalysts, VOSO4 and NaVO3, to improve the FA yield to 39% or higher. However, vanadium-based catalysts are difficult to synthesize and are expensive.

In all vanadium(V) systems, the observation of mutual transformation between vanadium(V) and vanadium(IV) suggests a process in which the substrate is first oxidized by vanadium(V) and reduced vanadium(IV) is then oxidized by O2. This fact indicates that other metal salt catalysts with (1) changeable valence and (2) lower reduction potential than O2 can also catalyze cellulose oxidation to form FA similar to vanadium(V) catalysts. Thus, in the present work, we have tested several high valence metal salts with the requirements of the above two as catalysts of cellulose oxidation, and have found iron(III) is a high-efficiency catalyst to catalyze the generation of FA from cellulose aerobic oxidation with both excellent yield and excellent reusability. Compared with vanadium(V), iron(III) is easy to obtain and synthesize, and has a low cast, which is suitable for application. The relationship between oxidation and hydrolysis in this iron(III)-catalyzed cellulose conversion, together with the transformation pathway, was studied in detail. The exchange between iron(III) and iron(II) was found to be the main transformation form of the catalyst.

Experimental Section

Chemicals

Microcrystalline cellulose (96%), d-mannose (99%), 1,3-dihydroxyacetone dimer, glycolic acid (98%), glyoxylic acid anhydrous (98%), methylglyoxal (solution, 40%), 1,2-propanediol (99.5%), 1,3-propanediol (99.5%), ethylene glycol (98%), 5-hydromethyl furfural (HMF, 98%), furfural (99%), levulinic acid (99%), ferric(III) sulfate (Fe2(SO4)3, 99.95%), ferric(III) chloride (FeCl3, 98%), copper(II) chloride (CuCl2, 98%), molybdenum(V) chloride (MoCl5, 99.6%), chromium(III) chloride (CrCl3·6H2O, 98%), aluminum(III) chloride (AlCl3, 98%), calcium(II) chloride (CaCl2, 96%), and magnesium(II) chloride (MgCl2·6H2O) were purchased from Aladdin Reagent Inc. (Shanghai, China). d-(+)-Glucose anhydrous (99%) was purchased from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China). Glycolaldehyde dimer was purchased from Sigma-Aldrich (Germany). Glyceraldehyde (90%) was purchased from J & K Scientific (Beijing, China). Sulfuric acid (H2SO4, 98%) was purchased from Beijing Modern Oriental fine chemistry Co., Ltd. (Beijing, China). FA (98%), formaldehyde (solution, 40%), and glyoxal (solution, 40%) were purchased from Tianjin Fuchen chemical reagents factory. Acetic acid (99.5%) and methanol anhydrous (99.5%) were purchased from Beijing Tongguang Fine Chemical Company. Oxalic acid (99.5%) was purchased from Beijing Chemical Plants (Beijing, China). Oxygen (O2, 99.995%) and nitrogen (N2, 99.999%) were supplied by Beijing Haipu Gases Co., Ltd. (Beijing, China). All reagents were of analytical grade and used without further purification.

Conversion of Substrates

A series of aqueous solutions with various metal salts and H2SO4 were prepared before the conversion. A certain amount of metal salt was dissolved into distilled water with ultrasonic heating and then mixed with a certain amount of diluted H2SO4. The mixed solution was fixed to a constant volume with distilled water.

The conversion was carried out in a 25 cm3 batch reactor of Hastelloy alloy (HC 276) with a magnetic stirrer. In a typical procedure, a certain amount of substrate, and 6.0 cm3 of prepared aqueous solution were loaded into the reactor. Then, the reactor was sealed and purged with O2. After that, O2 was charged into the reactor to a desired pressure. Next, the reactor was put into a heating furnace with a heating rate of 8–10 °C/min and stirred at a speed of 1000 rpm. The pressure and the temperature of the reactor were measured by using a pressure transducer with an uncertainty of ±0.025 MPa and a thermocouple with an uncertainty of ±0.5 °C, respectively. When the desired reaction temperature was reached (∼15 min after being put into the furnace), the reaction time was recorded. After the reaction, the reactor was quenched by using cold water. When the reactor reached room temperature, the gas was released and the liquid mixture was filtered. The residue was washed with distilled water and dried in an oven at 60 °C for 24 h before further use. The liquid sample and gas sample were analyzed as follows.

Analysis of Products

The liquid sample was analyzed by a high-performance liquid chromatography (Waters 2695, USA) with a Shodex SH 1011 column (Shodex, Japan). A diode array detector (Waters 2998, USA) was employed to analyze the furan compounds (HMF and furfural). A differential refractive index detector (Waters 4110, USA) was employed to analyze other products. The column oven temperature was 55 °C, and the mobile phase was diluted H2SO4 aqueous solution with a concentration of 0.01 mol/dm3 and a flow rate of 0.5 cm3/min. The gas sample was detected using a GC (Agilent 7890A) using a TCD detector (200 °C) with a Porapak Q column (230 °C), with helium as the carrier gas (40 cm3/min). The detected CO2 in the gas phase was quantified by a total organic carbon analyzer (TOC-L CPN, Shimadzu, Japan) after absorption in NaOH aqueous solution. All the yields of products were calculated on a carbon base. The conversion of cellulose was calculated by the difference in solid weight before and after the reaction. We performed three-time parallel experiments at each set of conditions, and the results reported herein represent the mean values. The reproducibility of yields of FA and acetic acid (AA) was estimated better than an average relative deviation of 3.6%.

Surface morphology of residues after conversion of cellulose was studied by scanning electron microscopy (SEM, Zeiss Supra 55) with an accelerating voltage of 20 kV.

The humin yield was calculated as follows: humin yield = cellulose conversion – yield of the detectable products – yield of CO2.

Recycle and Reuse of the Catalytic System

After completion of the conversion, the solid was separated by filtration. The liquid products were extracted by n-butanone with the same volume (for three times). Then, the n-butanone dissolved in the liquid was swept by N2 at 50 °C. Water was added to the recovered liquid system to fit the original volume before the next run.

Results and Discussion

Screening of Catalysts

At the beginning of the study, we tested various metal salts for the cellulose conversion in H2SO4 aqueous solution under O2. The results are shown in Figure . The conversion of cellulose exhibits no significant difference, whether the metal salts were used or not (nearly every experiment provides a cellulose conversion >90%). In the absence of any metal salt, an FA yield of 26.4% was formed. When the metal salt was added to the reaction mixture, the FA yield was generally increased. This indicates that metal salts can change the product selectivity but cannot significantly catalyze the degradation of cellulose. Among these selected metal salts, two ferric salts, FeCl3 and Fe2(SO4)3 showed the best performance for giving the highest FA yield (48.3% and 47.8%, respectively). CuCl2 showed a lower FA yield (40.8%) after the ferric salts, followed by AlCl3 (35.7%), MoCl5 (32.2%), CaCl2 (31.9%), CrCl3 (30.9%), and MgCl2 (27.3%). Therefore, we employed the highly efficient ferric salt, FeCl3, as the catalyst for cellulose oxidation in further investigation.

Figure 1

Screening of metal salt catalysts in cellulose conversion with O2. Reaction conditions: cellulose, 0.100 g; catalyst, 0.00287 mol/dm3 (0.00144 mol/dm3 for Fe2(SO4)3); H2SO4, 2.0 wt %; H2O, 6.0 cm3; O2, 3.0 MPa; temperature, 160 °C; time, 80 min.

Conversion of Cellulose in FeCl3–H2SO4 Aqueous Solution

Effect of FeCl3 Concentration on the Conversion

The effect of FeCl3 concentration on cellulose conversion and products yields is shown in Figure a. The cellulose conversion was not changed obviously and maintained at the range of 93–96% when the FeCl3 concentration was increased from 0 to 0.1 wt %. The FA yield was increased from 26.4 to 48.3% when the FeCl3 concentration was increased from 0 to 0.0466 wt %. When the FeCl3 concentration was further increased, the FA yield showed a slight decrease (45–48%), probably because of the accelerating decomposition of FA under higher FeCl3 concentration (Figure S1a). Besides FA, AA and glycolic acid (GA) were formed as the main byproducts in liquid mixture. Both yields of the two acids were <4.5%. It is worth noting that, even the yields were low, the decrease of both GA yield and AA yield started when relatively low FeCl3 addition (0.0233 wt %) was added. Especially, the AA yield was decreased in the whole range of FeCl3 concentration employed in this series of experiments. The reason for the decrease in the AA yield is discussed later in section .

Figure 2

Effects of (a) FeCl3 concentration and (b) H2SO4 concentration on cellulose conversion and products yields. Reaction conditions for (a): cellulose, 0.100 g; H2SO4, 2.0 wt %; H2O, 6.0 cm3; O2, 3 MPa; temperature, 160 °C; time, 80 min. Reaction conditions for (b): cellulose, 0.100 g; FeCl3, 0.0466 wt %; H2O, 6.0 cm3; O2, 3.0 MPa; temperature, 160 °C; time, 80 min.

Effect of H2SO4 Concentration on the Conversion

The effect of H2SO4 concentration on cellulose conversion and products yields is shown in Figure b. The cellulose conversion was increased from 18 to 100%, when H2SO4 concentration was increased from 0 to 2.5 wt %. The increase of cellulose conversion is due to the enhancement of the acidic hydrolysis in H2SO4 solution (hydrolysis of cellulose in H2SO4 solution was accelerated rapidly by the increase of H2SO4 concentration, as shown in Figure S2). The FA yield was also increased (from 8 to 48.3%) with increasing H2SO4 concentration, probably because soluble carbohydrates derived from cellulose by hydrolysis were generated faster as the substrate in higher H2SO4 concentrations. The yields of GA and AA both were increased with increasing H2SO4 concentration.

Effect of Temperature and Reaction Time on the Conversion

The effect of reaction temperature on the transformation of cellulose at different reaction times was studied. As shown in Figure a, at a low temperature of 150 °C, the cellulose conversion was low (43%) after 40 min. The conversion was only increased to 74% after 140 min. When the temperature was increased, the cellulose conversion was dramatically increased (the conversion was completed in 5 min at 180 °C).

Figure 3

Effects of reaction temperature (a) on cellulose conversion and (b) on FA yield at different reaction times. Reaction conditions: cellulose, 0.100 g; FeCl3, 0.0466 wt %; H2SO4, 2.0 wt %; H2O, 6.0 cm3; O2, 3.0 MPa.

The effect of temperature on the FA yield shows a similar trend (Figure b). High temperature leads to quick formation of FA. For instance, a yield of 51.2% can be obtained at 170 °C for 50 min and a yield of 51.2% at 180 °C for 7 min. The byproduct is only CO2 on the basis of results of GC and TOC analyses of gas and liquid products. With a prolonged reaction time at higher temperatures (170 and 180 °C), a decrease of FA yield was observed. This decrease of FA can be explained by the instability of FA in high temperature under oxidative conditions (Figure S1b).

Oxidation and Hydrolysis in Cellulose Transformation in the FeCl3–H2SO4 System with O2

The transformation of cellulose in FeCl3–H2SO4 aqueous solution with O2 contains, apparently, two types of reactions, hydrolysis and oxidation. Both two reactions occur under the same conditions, and reaction parameters variation affects the rates of both reactions with different extents at the same time. Combination of hydrolysis and oxidation with different rates can probably lead to different product distributions. Therefore, we tried to investigate the relationship between various reactions (mainly hydrolysis and oxidation) by studying the effects of reaction parameters on product distributions.

Competition on FA Selectivity

The effect of FeCl3 concentration on FA selectivity is shown in Figure a. Compared with the reaction without FeCl3, FA selectivity was sharply increased in the presence of only slight amount of FeCl3. A change of FeCl3 concentration from 0.0699 to 0.0938 wt % exhibits a slight decrease in FA selectivity, possibly due to the instability of FA (discussed in Section ). Together with almost unchanged cellulose conversion under different FeCl3 concentrations (Figure a), one conclusion can be drawn that FeCl3 shows no activity on cellulose degradation, but catalyzes FA formation from the oxidation from soluble carbohydrates generated from cellulose hydrolysis.

Figure 4

Dependence of FA selectivity on FeCl3 concentration and H2SO4 concentration is shown in (a,b), respectively. The dependence of AA selectivity on FeCl3 concentration and H2SO4 concentration is shown in (a,b), respectively. Reaction conditions for (a,c): cellulose, 0.100 g; H2SO4, 2.0 wt %; H2O, 6.0 cm3; O2, 3.0 MPa; temperature, 160 °C; time, 80 min. Reaction conditions for (b,d): cellulose, 0.100 g; FeCl3, 0.0466 wt %; H2O, 6.0 cm3; O2, 3.00 MPa; temperature, 160 °C; time, 80 min.

The effect of H2SO4 concentration on FA selectivity is shown in Figure b. Unlike FeCl3, FA selectivity did not change much when H2SO4 was added. This independence of FA selectivity on H2SO4 concentration indicates that H2SO4 only catalyzes hydrolysis of cellulose (Figure S2) but does not involve closely in the oxidation for FA production. The selectivity decreasing at higher H2SO4 concentration suggests the instability of FA (Figure S1c).

Competition on AA Selectivity

The effect of FeCl3 concentration on AA selectivity is shown in Figure c. The AA selectivity, contrary to the FA selectivity, was decreased with increasing FeCl3 concentration. AA was proposed to be the main product of oxidation of levulinic acid (hydrolysis product of cellulose), evidenced by ∼50% yield of AA obtained from the conversion of levulinic acid. It suggests that there is a competition between oxidation (to form FA) and side reaction (probably hydrolysis, to form byproducts such as AA).

Similarly, the effect of H2SO4 concentration on AA selectivity was contrary to that on FA selectivity. AA selectivity was increased with increasing H2SO4 concentration (Figure d), which differs from the independence of H2SO4 concentration on the FA yield (Figure b). These results prove further that the side reaction, mainly hydrolysis, shows a competition relationship with the oxidation to FA.

Competition on Humin Formation

As we know, stronger acidic condition always leads to a larger extent of humin formation. Humin is a mixture of substances with unknown structures usually generated under acidic conditions via complicated reactions among hydrolysis products.[22] In the absence of FeCl3, humin was largely formed even with O2 (see Figure S3). When 0.0233 wt% FeCl3 was added, humin yield was sharply decreased to 1.4%. Further additions of FeCl3 all show very limited formation of humin acid. This result provided further evidence for the suppression of FeCl3 on deep hydrolysis.

Competition on Residue Agglomeration

We have observed that the residues after the conversion under different conditions exhibit different morphology. This difference is illustrated as SEM images of the residues under different H2SO4 concentrations in Figure . The unconverted cellulose appears as powder with the size <100 μm (Figure a). After cellulose was converted in FeCl3 aqueous solution with O2 (no H2SO4), the residue size was approximately the same as the original cellulose (Figure b). It displays a “peeling” degradation of cellulose in FeCl3-catalyzed oxidation. Interestingly, after H2SO4 was added, even though the cellulose conversion was increased, the size of the residue became markedly larger (Figure c). Higher H2SO4 concentration leads to the residue with larger size (Figure d–f). The change of residue size is presumably due to the cellulose agglomeration, which is caused by the reaction of hydrolysis products (possibly HMF and furfural) and unreacted cellulose. Therefore, it is another evidence for suppression of FeCl3 on hydrolysis.

Figure 5

SEM images of residues after conversion under different H2SO4 concentrations. (a) Untreated cellulose; (b) 0; (c) 0.5; (d) 1.0; (e) 1.5; (f) 2.0 wt %. Reaction conditions: cellulose, 0.100 g; FeCl3, 0.0466 wt %; H2O, 6.0 cm3; O2, 3.0 MPa; temperature, 160 °C; time, 80 min.

Pathway of the Cellulose Conversion in the FeCl3–H2SO4 System

In order to further investigate the pathway of cellulose conversion in FeCl3–H2SO4 aqueous solution with O2 as the oxidant, we attempted to detect the intermediates by the 13C NMR technique. Under acidic conditions, cellulose was supposed to be hydrolyzed to generate water-soluble glucose for further reactions. Therefore, 13C NMR detection of the mixture after conversion of glucose was carried out (shown in Figure S4). The 12 signals at 61.0, 61.2, 70.0, 70.1, 71.8, 71.9, 73.2, 74.6, 76.2, 76.3, 92.5, and 96.3 ppm were assigned to the six carbon in d-glucose (existing in water as α and β anomers). The signals for FA (166.2 ppm), AA (21.0 and 176.8 ppm), and GA (59.9 and 176.6 ppm) were also found in the spectrum.

Additional information was obtained as other signals appeared. The signals at 61.5, 67.4, 70.8, 71.3, 72.8, 94.2, and 94.6 ppm were attributed to carbons of d-mannose in α and β conformations (other signals overlapped with those of d-glucose). These results mean that epimerization of d-glucose to d-mannose occurs in FeCl3–H2SO4 aqueous solution. Besides, the signals at 28.3, 29.3, 38.2, and 177.2 ppm appeared, representing levulinic acid. The detection of levulinic acid further confirms the hydrolysis occurs in FeCl3–H2SO4 system, as discussed in Section .

Afterward, these observable intermediates were selected as model compounds for conversion (conducted in milder condition, see entry 1–3, Table ). d-Glucose and d-mannose were both converted into high yields of FA with byproducts GA and AA formation. By contrast, levulinic acid yielded limited FA (9.6%, entry 3, Table ), but gave a large amount of AA (49.7%). It indicates that levulinic acid may be the source of AA in the conversion of cellulose. This view is in accordance with the conclusion drawn in Section that AA was derived from hydrolysis products.

Table 1

Oxidation of Model Compounds in FeCl3–H2SO4 Aqueous Solution with O2a

			yield/%
entry	substrate	conversion/%	FA	AA	GA	others
1	d-glucose	100	53.3 ± 3.1	4.8 ± 0.5	1.9 ± 0.2	CO2
2	d-mannose	100	51.3 ± 3.0	3.7 ± 0.4	2.8 ± 0.3	CO2
3	levulinic acid	100	9.6 ± 0.6	49.7 ± 2.9		CO2
4	d-glyceraldehyde	100	51.7 ± 3.0	8.4 ± 0.9	5.2 ± 0.5	CO2
5	1,3-dihydroxyacetone	100	49.5 ± 2.9	11.0 ± 0.6	4.7 ± 0.5	CO2
6	glycolaldehyde dimer	100	66.2 ± 3.9		15.0 ± 1.6	CO2
7	methylgloxal	100	18.7 ± 1.1	36.8 ± 2.1		CO2
8	1,3-propanediol	68.3 ± 1.2	35.8 ± 2.1		1.8 ± 0.2	CO2
9	1,2-propanediol	76.0 ± 1.3	31.0 ± 1.8	17.2 ± 1.0		CO2
10	ethylene glycol	20.4 ± 1.0	9.4 ± 0.5		1.3 ± 0.1	CO2
11	glyoxal	100	29.3 ± 1.7		43.5 ± 2.5	CO2
12	glyoxylic acid	100	31.5 ± 1.8			CO2
13	oxalic acid	100				CO2
14	methanol	0
15	formaldehyde	0
16	furfural	96.6 ± 1.4	34.3 ± 2.0	4.6 ± 0.5		CO2
17	5-HMF	99.0 ± 1.4	32.0 ± 1.9	11.0 ± 1.1		CO2
18	glycolic acid	42.1 ± 1.1	20.4 ± 1.2			CO2
19	AA	0
20	FA	6.3 ± 1.0				CO2

Reaction conditions: substrate, 0.050 g; FeCl3, 0.0466 wt %; H2SO4, 2.0 wt %; H2O, 6.0 cm3; O2, 3.0 MPa; temperature, 170 °C; time, 50 min.

The transformation of d-glucose can possibly start with a retro-aldol condensation, by which C3–C4 bond cleavage proceeds to form d-glyceraldehyde and 1,3-dihydroxyacetone.[23] Therefore, the latter two substances can also possibly be the intermediates. The conversion of d-glyceraldehyde yielded FA in high selectivity with nearly no byproducts in the liquid phase (120 °C, detected in 13C NMR in Figure S5a). The conversion of d-glyceraldehyde at a higher temperature in which cellulose was converted (170 °C) produced a high yield of FA (51.7%, entry 4, Table ) with AA and GA formation. The conversion of 1,3-dihydroxyacetone was slower and produced FA, AA, GA, and glycolaldehyde (120 °C, detected in 13C NMR in Figure S5b), although conversion of it at 170 °C produced a slightly lower yield of FA (49.5%) with AA and GA formation (entry 5, Table ).

As conversion of 1,3-dihydroxyacetone produced glycolaldehyde, we tested glycolaldehyde for conversion. A much higher yield of FA (66.2%, entry 6, Table ) was obtained with considerable GA formation (15.0%). AA was derived from 1,3-dihydroxyacetone probably via a dehydration product of 1,3-dihydroxyacetone, methylglyoxal.[23−25] Conversion of the latter produced an AA yield of 36.8% and an FA yield of 18.7% (entry 7, Table ).

Additionally, other possible intermediates with 1–3 carbon atom(s) were tested (entries 8–15). These transformations result in either low conversions (1,3-propanediol, 1,2-propanediol, ethylene glycol, methanol, and formaldehyde) or low yields of FA with no AA formation (glyoxal, glyoxylic acid, and oxalic acid). Therefore, these were not the intermediates of the cellulose conversion. Hydrolysis products HMF and furfural may be the intermediates (entries 16–17), but FA and AA were generated much more possibly via levulinic acid.

The products FA, AA, and GA were at last tested (entries 18–20, Table ). GA was not fully converted and yielded FA. FA and AA were relatively stable with <7% decomposition.

According to the study on model compounds and hydrolysis–oxidation relationship (Section ) and the results reported by Wasserscheid et al.,[14] a proposed pathway of cellulose conversion in FeCl3–H2SO4 aqueous solution with O2 as the oxidant can be concluded, as shown in Scheme . Cellulose is first converted into d-glucose (in equilibrium with d-mannose),[21] which initiates two parallel reactions: retro-aldol condensation and hydrolysis. Retro-aldol condensation leads to C3–C4 bond cleavage, producing d-glyceraldehyde and 1,3-dihydroxyacetone.[17] These two C3 intermediates are further oxidized to FA and glycolaldehyde, which is further oxidized to GA. GA remains as the final product with partial degradation to FA and CO2. 1,3-Dihydroxyacetone can also be dehydrated to methylglyoxal, which is further oxidized to FA, AA, and CO2. Hydrolysis results in levulinic acid formation, which yields a large amount of AA, as the main pathway for AA formation.

Scheme 1

Proposed Pathway of Cellulose Conversion in FeCl3–H2SO4 Aqueous Solution with O2

Reuse of the Catalyst

It is necessary to test the reusability of the catalyst from the view point of further application. The catalyst can be recovered after organic solvent extraction of the products.[21] After four runs, the conversions of cellulose and yields of FA show no significant decrease, as shown in Figure . This result indicates that FeCl3–H2SO4 catalytic system exhibits good performance in the reusability.

Figure 6

Reuse of the FeCl3–H2SO4 catalytic system. Reaction conditions for each run: cellulose, 0.100 g; FeCl3, 0.0466 wt %; H2SO4, 2.0 wt %; H2O, 6.0 cm3; O2, 3.0 MPa; temperature, 170 °C; time, 50 min.

Iron(III) ⇌ Iron(II) Mutual Transformation in the Catalytic Process

During the transformation, iron(III) can possibly oxidize the substrate to generate iron(II) first. Subsequently, the formed iron(II) can reasonably be oxidized as iron(III) form by O2, according to the lower redox potential of iron(III)/iron(II) than O2/H2O (Eθ(O2/H2O) = 1.229 V, Eθ(iron(III)/iron(II)) = 0.771 V). Further study on the detection of iron(III) and iron(II) in reaction mixture confirms this assumption. Mixing of KSCN with FeCl3–H2SO4 aqueous solution gave a solution with a color of dark yellow, indicating the interaction between KSCN and iron(III). After the cellulose conversion under the FeCl3–H2SO4 system with O2, the mixing of reaction mixture with KSCN also provided the same color of the solution, suggesting the existence of iron(III) after reaction. In the absence of O2, no color change after mixing of KSCN and the reaction mixture, indicating no iron(III) detection. Iron(II) in the absence of O2 can be found by UV–vis spectrum after a complexation of iron(II) and phenanthroline (see Figure ). Therefore, during the oxidation, the catalyst FeCl3 shows its catalytic ability by a mutual transformation between iron(III) and iron(II).

Figure 7

UV–vis spectra of reaction mixtures after conversion and after conversion without O2 (replaced by N2 with the same initial pressure). Reaction conditions: cellulose, 0.100 g; FeCl3, 0.0466 wt %; H2SO4, 2.0 wt %; H2O, 6.0 cm3; O2 (or N2), 3.0 MPa; temperature, 160 °C; time, 80 min.

Conclusions

In this work, a variety of metal salts in acidic aqueous solution for the conversion of cellulose into FA under O2 were tested, and iron(III) was selected as a more environmentally benign and readily available catalyst for FA formation, with a comparable yield (51.2%, based on carbon) with that of vanadium(V). The effect of reaction parameters was studied, and the competition between oxidation and hydrolysis were found and discussed in detail. FeCl3 and H2SO4 can accelerate oxidation and hydrolysis, respectively, whereas suppress the other one. The effects can reflect on the product distribution. Intermediates were found and the pathway from cellulose to products was reasonably proposed. The reusability of the catalytic system shows no significant decrease in both conversion and FA yield after four runs. The mechanism study suggests the catalytic ability was derived by a mutual transformation between iron(III) and iron(II), in which iron(III) is reduced to iron(II) by substrates, and then iron(II) is reoxidized to iron(III) by O2.