In Situ 13C NMR Study - American Chemical Society

Article pubs.acs.org/JPCA

Solvent Effect on Pathways and Mechanisms for D‑Fructose Conversion to 5‑Hydroxymethyl-2-furaldehyde: In Situ 13 C NMR Study Hiroshi Kimura,† Masaru Nakahara,† and Nobuyuki Matubayasi*,†,‡,§ †

Institute for Chemical Research, Kyoto University, Uji, Kyoto 611-0011, Japan Japan Science and Technology Agency (JST), CREST, Kawaguchi, Saitama 332-0012, Japan § Elements Strategy Initiative for Catalysts and Batteries, Kyoto University, Katsura, Kyoto 615-8520, Japan ‡

S Supporting Information *

ABSTRACT: Noncatalytic reactions of D-fructose were kinetically investigated in dimethylsulfoxide (DMSO), water, and methanol as a function of time at temperatures of 30−150 °C by applying in situ 13C NMR spectroscopy. The products were quantitatively analyzed with distinction of isomeric species by taking advantage of siteselective 13C labeling technique. In DMSO, D-fructose was converted first into 3,4dihydroxy-2-dihydroxymethyl-5-hydroxymethyltetrahydrofuran having no double bond in the ring, subsequently into 4-hydroxy-5-hydroxymethyl-4,5-dihydrofuran-2-carbaldehyde having one double bond through dehydration, and finally into 5-hydroxymethyl-2furaldehyde (5-HMF) having two double bonds. No other reaction pathways were involved, as shown from the carbon mass balance. In water, 5-HMF, the final product in DMSO, was generated with the precursors undetected and furthermore transformed predominantly into formic and levulinic acids and slightly into 1,2,4-benzenetriol accompanied by polymerization. D-Glucose was also produced through the reversible transformation of the reactant D-fructose. In methanol, some kinds of anhydro-D-fructoses were generated instead of 5-HMF. The reaction pathways can thus be controlled by taking advantage of the solvent effect. The D-fructose conversion reactions are of the first order with respect to the concentration of D-fructose and proceed on the order of minutes in DMSO but on the order of hours in water and methanol. The rate constant was three orders of magnitude larger in DMSO than in water or methanol.

1. INTRODUCTION Biomass-derived carbohydrates such as cellulose and starch, which are abundant and renewable, have been expected to be a promising alternative for the sustainable supply of fuel and valuable chemicals.1−15 In particular, 5-hydroxymethyl-2-furaldehyde (5-HMF) has been recognized as a potential platform chemical with a wide application profile.16 For developing further applications of biomass resources to prepare feedstocks, the progress of the physical chemistry of the structures17−21 and reactions14,22,23 of biomass is urgently necessary. The reaction pathways are also influenced by solvent environment so that it is of interest to explore the solvent effect on biomass conversion processes. The physicochemical study on biomass is not well advanced as yet, however, because of the complexities of sugar reaction pathways and because of the presence of various intervening isomers.22,23 Model studies are thus needed as the first step on monosaccharides. In a previous work, the D-glucose reaction under hydrothermal conditions was investigated, and its mechanism was found to be identical to the Dfructose reaction mechanism in the sense that D-glucose transformations occur only through D-fructose as an intermediate.22,23 In the present work, thus taking D-fructose as a starting material, we systematically carry out mechanistic and kinetic studies on the effect of solvent on the conversion of D-fructose into valuable feedstocks, such as 5-HMF and levulinic and formic acids. © 2013 American Chemical Society

So far, the conversion of D-fructose was studied in such solvents as water,24−32 methanol,33 DMSO,34−40 organic/water mixtures,10,41−43 ionic liquids,44−47 and biphasic water/organic systems42,48−50 in the presence of catalysts. However, it has been difficult to clarify how the reaction proceeds in such complicated systems. This is due to the existence of a large number of product species including carbohydrate isomers. When a high reaction temperature is set, further complication and difficulty can arise as a result of the poor time resolution. The questions addressed here are how conjugated double bonds are generated in the furan ring of 5-HMF and how the pathways for the D-fructose conversion can be controlled by solvent. To elucidate the reaction pathways and mechanisms for D-fructose, spectroscopic analysis needs to be conducted at atomic resolution with carefully designed experimental conditions. In this study, a low-temperature range of 30−150 °C is taken in the absence of any catalysts so that early products generated through the D-fructose conversion can be identified in the NMR time and atomic resolutions.22,23 As typical aprotic and protic solvents, DMSO (dielectric constant εr = 47.2 at 20 °C),51 water (εr = 80.1 at 20 °C),51 and methanol (εr = 33.0 at 20 °C)51 are Received: December 6, 2012 Revised: February 13, 2013 Published: March 4, 2013 2102

dx.doi.org/10.1021/jp312002h | J. Phys. Chem. A 2013, 117, 2102−2113

The Journal of Physical Chemistry A

Article

employed to get insight into the solvent effect on the reaction path selectivity. The presence of a variety of reactive species including isomers makes it difficult to quantitatively analyze the reactions of carbohydrates. To overcome the difficulty, we apply in situ 13C NMR spectroscopy in combination with a site-selective labeling technique, as employed in our previous works.22,23 With this method, we can noninvasively identify and quantify the mixture of the carbohydrate isomers, intermediates, and final products as functions of reaction time. At the same time, we can check the total mass balance that is essential to establish the reaction pathways so that we can control the path using the solvent effect.

2. EXPERIMENTAL SECTION 13 D-Fructose (Nacalai, 99.9%), D-[1- C]-fructose (ISOTEC, 99.8 atom % 13C), D-[2-13C]-fructose (ISOTEC, 99.8 atom % 13C), 13 13 D-[5- C]-fructose (ISOTEC, 99.8 atom % C), D-[6-13C]13 fructose (ISOTEC, 99.8 atom % C), D-[1,2,3,4,5,6-13C]-fructose (ISOTEC, 99 atom % 13C), dimethylsulfoxide-d6 (ISOTEC, 99.8 atom % D), water-d2 (ISOTEC, 99.96 atom % D), and methanold4 (ISOTEC, 99.8 atom % D) were used without further purification.52 In the kinetic analysis, D-[2-13C]-fructose was employed as a reactant. To avoid possible contribution of solute−solute interactions, we adopted a low concentration of 0.02 M (mol dm−3); the solute was fully dissolved in the three solvents in the course of in situ 13C NMR measurements. The solution was loaded into a Pyrex NMR tube (SHIGEMI, 8.0-mm o.d.). The sample tube was flamesealed after the tube inside was purged with argon. The apparatus and the experimental procedures were the same as before.22,23 The temperature was set to 30−150 °C and controlled within ±1 °C. The reaction systems in water and methanol are on the liquid branch of the gas−liquid coexistence curve above the temperatures of 100 and 65 °C, respectively. In the in situ 13C measurements, the proton irradiation was turned off to keep the spectral intensity proportional to the number of carbon atoms. It took 1−30 min for one in situ measurement, which is short enough compared with the time scale of the reaction of interest described below. Ab initio MO calculations in vacuum and in PCM (continuum) DMSO, water, and methanol were also performed using the Gaussian 09 program53 for all chemical species involved in the D-fructose conversion reactions. The geometry was optimized using the hybrid density functional B3LYP with the correlation consistent polarized valence triple-ζ (cc-pVDZ) basis set. At the optimized geometry, single-point energy and NMR chemical shift calculations were carried out at the B3LYP level of theory with the augmented cc-pVDZ (aug-cc-pVDZ) basis set.

Figure 1. Structures and isomerization pathways of the D-fructose isomers.

respectively. By this, for example, β-furanose of D-fructose is denoted as β-Frcf, which corresponds to F5‑β in our nomenclature. Because a single-letter abbreviation is more convenient to see the carbohydrate configuration within the present context, a simplified set of notations is adopted in this article. 3.1. Solvent Effect on Isomer Population. To discuss the solvent effect on the D-fructose conversion pathways and mechanisms, it is of importance to first know how the pre-equilibrium populations of the D-fructose isomers vary with solvent. As previously shown, the open-chain and ring forms of D-fructose are all observed in water: F6‑α (98.5 ppm), F6‑β (98.7 ppm), F5‑α (105.3 ppm), F5‑β (102.3 ppm), and Fopen‑chain (211.1 ppm), where the parenthesized numbers are the 13C chemical shifts of C2 carbon atom of interest. In DMSO and methanol, which are less polar than water, all isomers are also detected with negligible solvent effect on the 13 C chemical shifts; the 13C spectra before the D-fructose conversion are given in Figure SI-1 of the Supporting Information. According to the integrated peak intensities, the isomer populations of D-fructose at pre-equilibrium are determined at temperatures of 30−150 °C, as shown in Figure 2. When the temperature dependencies of the isomer populations in DMSO, water, methanol are compared, the following tendencies are observed; (i) the six-membered fructose of type β, F6‑β, dramatically decreases with increasing temperature, and correspondingly, the other isomeric forms increase, (ii) the fivemembered fructose of type β, F5‑β, is the major form at high temperatures, and (iii) the relative population of F5‑β is higher in DMSO than in water and methanol. Points (i)−(iii) can be explained in terms of the solvation (or OH orientation) effect, as found previously.22,23 Thus, the enhanced formation of 5-HMF, belonging to the family of F5‑β, is suggested to be in DMSO or at high temperatures. This has been found, as shown below. 3.2. Spectral Analysis of D-Fructose Conversion. Now we discuss the solvent effect on the D-fructose conversion pathways and mechanisms by focusing not only on the products but also on the isomers; the isomeric species are treated distinctly as much as possible. To the end, the in situ spectra are to be analyzed as a function of time with the mass balance checked; see the in situ 13C spectra for D-[2-13C]-fructose in DMSO in Figure 3 and those in water and methanol in Figure 4. 3.2.1. In DMSO. Let us first show what precursors are observed during the conversion of D-fructose into 5-HMF in DMSO. As seen in Figure 3, in the early reaction stage (3 min), there are detected five peaks that come from the open-chain and ring

3. RESULTS AND DISCUSSION The in situ 13C NMR measurements are carried out by taking advantage of D-[2-13C]-fructose where the anomeric carbon site C2 (Figure 1) is enriched to be selectively monitored. To help one understand the complicated structures of carbohydrates,54 hereafter we express the open-chain form, the pyranoses (sixmembered ring), and the furanoses (five-membered ring) of α- and β-types for D-fructose (F) as Fopen‑chain, F6‑α, F6‑β, F5‑α, and F5‑β; see the carbon atom configurations and conformations shown in Figure 1. The similar notation and numbering systems are adopted for the D-glucose isomers (that is, Gopen‑chain, G6‑α, G6‑β, G5‑α, and G5‑β). Usually, the monosaccharides of D-fructose and D-glucose are expressed as Glc and Frc, respectively; moreover, the pyranose and furanose forms are indicated by p and f, 2103



Article

Figure 3. 13C spectral evolution for 0.02 M D-[2-13C]-fructose in DMSO examined at 90 °C. 5-HMF″ and 5-HMF′ were assigned on the basis of the experimental (using 13C-labeled D-fructose listed in Experimental Section) and theoretical (calculating NMR chemical shifts) approaches, as mentioned in Section 3.2.1. For the assignment of the D-fructose isomers and 5-HMF, see ref 22.

at 330 min, only one peak assigned to 5-HMF is present as a target product. At this reaction time, a small amount (∼5%) of insoluble polymers is also obtained as fine dark-brown particles.55 This is attributed to the condensation polymerization of 5-HMF;22,56 the uncontrolled loss of 5-HMF occurs when the concentration grows at later reaction times because the reaction order is higher than the first. By employing site-selectively 13C-labeled D-fructose, we determine the structures of the two species denoted above as 5-HMF″ and 5-HMF′. On the basis of the common rule connecting the 13C chemical shift with the presence of double bond,57 it is found that 5-HMF″ has no double bond and 5-HMF′ has one double bond between the C2 and C3 sites. It also turns out that C1 carbons of 5-HMF″ and 5-HMF′ are methine carbon of dihydroxymethyl group and carbonyl carbon of aldehyde group, respectively, as represented in Figure 5a. The comparison of the experimental chemical shifts in Figure 3 with the quantum-chemically computed values is also in favor of the structures of 5-HMF″ and 5-HMF′ in Figure 5a; comparison of the experimental and computed 13C chemical shifts is shown in Table SI-1 of the Supporting Informaitonl. However, the confirmed precursor structures does not support a putative reaction scheme,37 in which D-fructose is considered to be converted first into the 2-DMSO-substituted or the 1-enol intermediates, subsequently into the dihydrofuran-2aldehyde intermediate, and finally into 5-HMF. The secondary precursor, 5-HMF′ obtained here is identical to the dihydrofuran-2-aldehyde intermediate proposed, but the primary precursor, 5-HMF″ is neither the 2-DMSO-substituted nor the 1-enol intermediates. 5-HMF″ and 5-HMF′ as well as 5-HMF are five-membered

Figure 2. Temperature dependence of the pre-equilibrium isomer populations of D-fructose in (a) DMSO, (b) water, and (c) methanol. They were all examined at a concentration of 0.02 M. The vertical axis shows the concentration normalized by the initial concentration of D-fructose.

isomers of D-fructose shown in Figure 1. No conversion of D-fructose except for the isomers is observed at 3 min at this temperature (90 °C). At 15 min of reaction time, however, two more peaks are observed. The strong peak at 109 ppm comes from 3,4-dihydroxy-2-dihydroxymethyl-5-hydroxymethyltetrahydrofuran having no double bond (denoted as 5-HMF″) as a primary precursor. The weak one at 157 ppm comes from 4-hydroxy-5-hydroxymethyl-4,5-dihydrofuran-2-carbaldehyde having one double bond (denoted as 5-HMF′) as a secondary precursor. The assignment of these precursors of 5-HMF will be described below. To the best of our knowledge, the presence of 5-HMF″ is elucidated for the first time in this study. The detection of 5-HMF″ was possible only at a low-reaction temperature of 90 °C without any catalyst. At 45 min of reaction time, one more peak emerges at 152 ppm, and this is assigned to 5-HMF having two conjugative double bonds in the ring. The 5-HMF′ peak largely increases, and the peaks assigned to the D-fructose and its family isomers decrease correspondingly. After the reaction for 100 min, 5-HMF and 5-HMF′ become dominant. Finally, 2104



Article

the carbonyl carbon is 8.2 ppm. Note that the peak intensity of 5-HMF generated decreases drastically with time. This clearly indicates that 5-HMF serves as a precursor for further conversions to 1,2,4-benzenetriol, formic, and levulinic acids at long reaction time. As for the above point (i), it is not shown whether D-fructose is dehydrated to 5-HMF through 5-HMF″ and 5-HMF′ in water as in DMSO. If the dehydration mechanisms are common in water and DMSO, then the absence or unsuccessful observation of 5-HMF″ and 5-HMF′ in water can be attributed to the weaker solvation of the hydrophobic species in water than in DMSO due to the less favored orientation and the loss of the polar OH bonds in 5-HMF″ and 5-HMF′. According to the structures of 5-HMF″ and 5-HMF′, these precursors are considered to become too unstable in a protic (higher-polar) solvent, water, to be observed within NMR time resolution. Point (ii) is due to the reversible transformation of D-fructose into D-glucose via the hydrothermal keto−enol tautomerization.22 D-Glucose is not formed in DMSO because the proton (or hydroxide) is usually required for the tautomerization between D-fructose and D-glucose except under high-temperature conditions with water. Point (iii) shows that 5-HMF generated is further decomposed to levulinic and formic acids in equal amounts through hydrolysis and transformed into 1,2,4-benzenetriol through additional dehydration. It is to be noted that these organic acids produced can act as an acid catalyst to cause the condensation polymerization of 5-HMF, as mentioned in Section 3.2.1. As shown here, water is better than DMSO only when the target product(s) is not 5-HMF but levulinic and formic acids; recall the pathway complexity. 3.2.3. In Methanol. In methanol, as seen in Figure 4b, the product distribution in the temperature range covered here is quite different from those in DMSO and water. At the reaction time of 10 h, there are newly detected three peaks except for those derived from the D-fructose isomers; one strong peak at 108 ppm comes from 2,6-anhydro-β-D-fructofuranose (denoted as 2,6-ahrF5-β), one medium one at 104 ppm comes from 3,6anhydro-α-D-fructofuranose (denoted as 3,6-ahrF5-α), and one weak one at 100 ppm comes from 1,6-anhydro-α-D-fructofuranose (denoted as 1,6-ahrF5-α). The peak assignments for these anhydrosugars were conducted using 13C labeled compounds and calculating the free energies and NMR chemical shifts;23,65,66 see the Supporting Information (Table SI-1). Only these anhydrosugars with the saturated five-membered ring continue to increase, and no other products come out. 5-HMF and its precursors are not produced in this solvent at the low temperature, although there has been reported the generation of 5-HMF in sub- and supercritical methanol with sulphuric acid as a catalyst.33 3.3. Pathways and Mechanisms. Here we will establish the pathways and mechanisms for the D-fructose conversion in DMSO, water, and methanol on the basis of the spectral observations before and during the conversion reaction mentioned in Sections 3.1 and 3.2. In DMSO (Figure 5a), D-fructose is converted first into the primary precursor, 5-HMF″ with no double bond, subsequently into the secondary precursor, 5-HMF′ with one double bond, and finally into 5-HMF with two double bonds through dehydration. No other reaction pathways are present during these processes. In water (Figure 5b), D-fructose is converted into 5-HMF probably through 5-HMF″ and 5-HMF′, as in DMSO. In contrast with the reaction in DMSO, 5-HMF generated is further hydrolyzed into formic and levulinic acids, transformed in part into 1,2,4-benzenetriol, and also polymerized. In addition, D-fructose is transformed into

Figure 4. 13C spectral evolution for 0.02 M D-[2-13C]-fructose in (a) water and (b) methanol. These spectra were all obtained at 150 °C. Levulinic acid and 1,2,4-benzenetriol were assigned using authentic samples. The assignment of the anhydrosugars was done by using all and site-selectively labeled compounds and referencing the calculations of NMR chemical shifts, as described in Section 3.2.3. For the assignment of the D-glucose isomers, see ref 22.

cyclic compounds consisting of the furanose (or furan) ring. Therefore, the D-fructose conversion is considered to start from the furanose forms of F5‑α and F5‑β. The presence of these precursors is associated with the relatively favored solvation of the dehydrated species in the aprotic solvent, DMSO, as discussed in Section 3.4.2 and the Appendix. 3.2.2. In Water. How does the reaction pathway control change with solvent? Figure 4a shows the in situ 13C spectra after the reaction for 0.5−50 h in water. When the spectra in Figure 4a are compared with those in Figure 3, the following differences in water and DMSO are observed; (i) the precursors with no and one double bond (5-HMF″ and 5-HMF′, respectively) are not observed in water, (ii) D-glucose is generated in water together with 5-HMF in the early reaction stage (∼10 h), and (iii) equal amounts of levulinic and formic acids are produced in water together with a small amount of 1,2,4-benzenetriol in the later reaction stage (∼50 h); levulinic acid is a versatile building block for the synthesis of various organic chemicals and liquid transportation fuels,58−60 and formic acid is of potential use as hydrogen storage.61−64 The formic acid generation was explored using 1 H NMR because its carbon atom was derived from the nonlabeled C1 of 5-HMF; the chemical shift of 1H atom attached to 2105



Article

Figure 5. Noncatalytic conversion pathways for D-fructose in (a) DMSO, (b) water, and (c) methanol. The curly brace denotes a set of all isomers of each monosaccharide; only the representative isomers are shown for brevity. The k values are the first-order rate constants for the corresponding reaction paths shown by the arrows. D-glucose via the keto−enol tautomerization. In methanol (Figure 5c), some kinds of anhydro-D-fructose instead of 5-HMF are generated through the dehydration. The reaction schemes shown in Figure 5 will be demonstrated by analyzing the time evolution of the concentrations of the chemical species involved; to theoretically complement our experimental insights, solution free-energy calculations are performed in the Appendix. We show that all of the isomers can be treated collectively in analyzing the time evolution. As seen in Figure 6, the isomer fractions of D-fructose and D-glucose in all solvents are found to be constant over the reaction time. This clearly indicates that the isomerization rate constants are significantly larger than the conversion (or dehydration) rate constants shown in Figure 5. Therefore, the collective treatment of the isomers is valid in the analysis of the time evolution done below.

Then, we see how the reactant and products are evolved with time. For the reaction in DMSO, see Figure 7a. It is found that only 5-HMF″ is generated in the early reaction stage (≤3 min). After that, 5-HMF′ rises until 45 min; then, 5-HMF is produced and finally yielded up to a high value of 95%. Because the carbon mass balance is kept during the reaction time until 4 h, no sidereactions are present, and only the reaction pathways of Figure 5a are to be kinetically analyzed in this time region; in the later reaction stage (>4 h), the condensation polymerization of 5-HMF begins to take place due to the high concentration of 5-HMF and the appearance of formic acid in a trace amount, accompanied by the mass balance loss of 5%. It is to be noted that the concentration of 5-HMF″ reaches the maximum at 0.6 h and then begins to decrease with increasing time. Also, for 5-HMF′, the corresponding aspects of the time dependence are observed 2106



Article

Figure 7. Time evolution of the concentrations of the reactant and products for the D-fructose conversion in (a) DMSO, (b) water, and (c) methanol. Panel a was examined at 90 °C, and panels b and c were at 150 °C. The vertical axis on the left shows the concentration normalized by the initial concentration of D-fructose, and that on the right shows the mass balance. In panel b, the plots of formic and levulinic acids are nearly overlapped, and their data are hardly distinguishable from each other.

Figure 6. Time dependence of the populations of the pre-equilibrium isomers of D-fructose (blue) and D-glucose (red) in (a) DMSO, (b) water, and (c) methanol. Panel a was determined at 90 °C, and panels b and c were determined at 150 °C.

carbon mass balance is kept during the reaction time until 10 h, it subsequently decreases and finally drops to 52%. The decrease in the mass balance corresponds well to that of the generation of formic and levulinic acids. This clearly indicates that these organic acids serve as acid catalysts and enhance the condensation polymerization of 5-HMF. In DMSO, in contrast, the polymerization is suppressed as a result of the absence of the hydrolysis of 5-HMF to formic and levulinic acids. In methanol (Figure 7c), 1,6-ahrF5-α, 2,6-ahrF5-β, and 3,6-ahrF5-α are produced through the dehydration, and 5-HMF is not observed over the reaction time examined (≤50 h). The concentration ratio of 1,6-ahrF5-α/2,6-ahrF5-β/3,6-ahrF5-α is 0.08:0.56:0.36 through the reaction at 150 °C and is almost independent of temperature; the ratio is 0.07:0.57:0.36 even at the lowest temperature of 100 °C. The carbon mass balance is preserved, and no other reaction paths than those in Figure 5c are involved in this case. The ether ring formation between the C2 and C6 sites is most favored, probably due to the lowest ring strain and steric hindrance. Thus it is concluded that DMSO is the best for controlling the

in the evolution. These observations clearly show that 5-HMF″ and 5-HMF′ are the precursors in the process of the conversion of D-fructose into 5-HMF. There are a few previous studies on the conversion of D-fructose into 5-HMF. Some10,67,68 proposed that the furanoses (five-membered ring) play a key role in the generation of 5-HMF in view of the possibility of dehydrating to form 5-HMF, and others50,69 proposed that the open-chain form is a key. The experimental results obtained here (Figure 5a) suggest that the furanoses can be the starting compound for the 5-HMF synthesis. In water (Figure 7b), 5-HMF is seen to be produced in the early reaction stage (10 h), a small amount of 1,2,4-benzenetriol is also produced as a byproduct. Although the 2107



Article

Table 1. Rate Constants (kHMF″, kHMF′, and kHMF) for the D-Fructose Conversion in DMSO at Temperature Range of 90−150 °C T/°C

kHMF′′/10−3 s−1

kHMF′/10−3 s−1

kHMF/10−3 s−1

90 100 110 120 130 140 150

0.58 ± 0.06 1.4 ± 0.2 2.7 ± 0.3 4.1 ± 0.3 8.4 ± 0.5 17.2 ± 1.5 31.3 ± 2.2

0.95 ± 0.23 1.5 ± 0.5 1.7 ± 0.5 2.9 ± 0.5 5.1 ± 0.8 8.5 ± 1.2 11.2 ± 1.4

0.26 ± 0.07 0.39 ± 0.12 0.71 ± 0.27 0.74 ± 0.23 1.3 ± 0.2 1.9 ± 0.5 3.3 ± 0.5

reaction pathways to the selective production of the valuable product, 5-HMF, with the conjugative double bonds. 3.4. Kinetic Analysis. Before going to discuss the rate laws, we have confirmed that there have been no differences in the reactive species fractions when varying the initial concentration of D-fructose from 0.02 to 0.05 and 0.1 M. Because the D-fructose conversion is independent of the initial concentration of D-fructose, the reactions represented in Figure 5 are all of the first order with respect to the concentrations of D-fructose and D-glucose. According to the time evolution of all the chemical species, therefore, we can establish the rate laws, solve the rate equations, and determine the kinetic parameters defined in the reaction schemes in Figure 5. 3.4.1. Rate Laws. First let us see the rate laws for the D-fructose conversion in DMSO. According to the scheme in Figure 5a, the first-order rate equations with respect to the concentrations of D-fructose, 5-HMF″, and 5-HMF′ are introduced as follows: d[{D‐fructose}] = −kHMF ″[{D‐fructose}] dt

(1)

d[5‐HMF″] = kHMF ″[{D‐fructose}] − kHMF ′[5‐HMF″] dt (2)

d[5‐HMF′] = kHMF ′[5‐HMF″] − kHMF[5‐HMF′] dt

(3)

d[5‐HMF] = kHMF[5‐HMF′] dt

(4)

Figure 8. Arrhenius plots of the rate constants for the D-fructose conversion in (a) DMSO (kHMF″, kHMF′, and kHMF), (b) water (k+1, k−1, and k2), and (c) methanol (k2,6‑ahr, k3,6‑ahr, and k1,6‑ahr) as functions of 1/T. They were linearly fitted to the solid line, and the slopes of the solid lines give the activation energies (Ea) as follows: Ea,HMF″ = 167, Ea,HMF′ = 91, Ea,HMF = 74, Ea,+1 = 141, Ea,−1 = 132, Ea,2 = 102, Ea,2,6‑ahr = 137, Ea,3,6‑ahr = 135, and Ea,1,6‑ahr = 133 kJ mol−1, respectively.

where t is the time, kHMF″, kHMF′, and kHMF are the rate constants for the successive reaction pathways shown, and the square and curly braces denote the concentration and the set of all of the fastexchanging isomers of D-fructose, respectively. To determine the values of kHMF″, kHMF′, and kHMF, we fit the time-dependent concentrations of D-fructose, 5-HMF″, and 5-HMF to quadratic polynomial functions. Substituting the quadratic expressions for D-fructose, 5-HMF″, and 5-HMF into the left-hand side of eqs 1, 2, and 4, respectively, we numerically solved these differential equations to determine kHMF″, kHMF′, and kHMF. In the time region examined for the kinetic analysis, the effect is negligible of acidic products like formic and levulinic acids, which serve as acid catalysts. The time region was divided into eight segments with equal intervals of 1−6 min depending on the reaction temperature. In each time region, we simultaneously determined kHMF″, kHMF′, and kHMF using eqs 1, 2, and 4 and confirmed the selfconsistency of the parameters obtained among the time segments. As for the D-fructose conversion in water, according to the reaction scheme given in Figure 5b, the following first-order rate equations with respect to the concentrations of D-fructose, D-glucose, and 5-HMF are established:

d[{D‐fructose}] = k+1[{D‐glucose}] − k −1[{D‐fructose}] dt − k 2[5‐HMF]

(5)

d[{D‐glucose}] = k −1[{D‐fructose}] − k+1[{D‐glucose}] dt (6)

d[5‐HMF] = k 2[{D‐fructose}] dt

(7)

Here k+1, k−1, and k2 are the rate constants for the corresponding reaction pathways. These differential equations are common to the case of the hydrothermal reaction of D-glucose studied previously.22 We numerically solved these differential equations and determined k+1, k−1, and k2, as done for the reactions in DMSO. The time region in which organic acids are absent was divided into eight segments with equal intervals of 1 to 2.5 h 2108



Article

where k1,6‑ahr, k2,6‑ahr, and k3,6‑ahr are the rate constants for the corresponding pathways. The total value (denoted as ktotal) of k1,6‑ahr, k2,6‑ahr, and k3,6‑ahr for the competing paths is obtained from the slope of the D-fructose concentration against time t. Because the population ratio of 1,6‑ahrF5-α, 2,6‑ahrF5-β, and 3,6‑ahrF5-α remains 0.08:0.56:0.36 over the reaction time examined, as seen in Figure 7c, we were able to determine the values of k1,6‑ahr, k2,6‑ahr, and k3,6‑ahr by using each population ratio. 3.4.2. Rate Constants. The rate constants for the D-fructose conversion reactions in DMSO, water, and methanol determined at the temperatures examined are summarized in Tables 1−3, respectively, and the Arrhenius plots are shown in Figure 8 with the activation energies (Ea).70 In DMSO (Table 1), it is found that kHMF″ is larger by a factor of one to three than kHMF′ and kHMF is smaller by a factor of two to four than kHMF′ with factors depending on the reaction temperature. In water (Table 2), it turns out that the value of k+1 is one order of magnitude smaller than that of k−1, and that the equilibrium

depending on the reaction temperature, and the self-consistency of the rate parameters was confirmed. For the reactions in methanol, as seen in Figure 5c, the first-order rate equation with respect to the concentration of D-fructose is given as d[{D‐fructose}] = −(k1,6 ‐ ahr + k 2,6 ‐ ahr + k 3,6 ‐ ahr) dt [{D‐fructose}]

(8)

Table 2. Rate Constants (k+1, k−1, and k2) for the D-Fructose Conversion in Water at Temperature Range of 100−150 °C T/°C

k+1/10−6 s−1

k−1/10−5 s−1

k2/10−6 s−1

100 110 120 130 140 150

0.31 ± 0.07 0.78 ± 0.18 1.6 ± 0.4 4.0 ± 0.9 8.5 ± 1.5 11.8 ± 2.4

0.23 ± 0.03 0.60 ± 0.08 1.3 ± 0.2 3.1 ± 0.4 5.1 ± 0.6 5.9 ± 0.6

0.35 ± 0.04 0.65 ± 0.07 1.4 ± 0.2 2.6 ± 0.3 4.1 ± 0.4 9.3 ± 0.9

Figure 9. (a) Diagrams of the electronic (Evac) and solution free energies (Gsol) for the conversion of D-fructose into 5-HMF computed in vacuum (black) and PCM DMSO (red). Δμ is the free energy change of reactive species brought from vacuum to solvent. In the figure, the solution free energies computed in PCM water and methanol are not shown for brevity, since they agree with the computed values in DMSO within ±3 kJ mol−1. (b) Diagrams of Gsol for hydrolysis of 5-HMF to formic and levulinic acids computed in DMSO (red) and water (blue). ΔGslv corresponds to the free energy change upon transfer from DMSO to water. These energy calculations were done at B3LYP/aug-cc-pVDZ level. 2109



Article

Table 3. Rate Constants (k1,6‑ahr, k2,6‑ahr, k3,6‑ahr, and ktotal) for the D-Fructose Conversion in Methanol at Temperature Range of 100−150 °C T/°C

k1,6‑ahr/10−7 s−1

k2,6‑ahr/10−6 s−1

k3,6‑ahr/10−6 s−1

ktotal (= k1,6‑ahr + k2,6‑ahr + k3,6‑ahr)/10−6 s−1

100 110 120 130 140 150

0.38 ± 0.04 0.69 ± 0.07 1.7 ± 0.2 4.4 ± 0.6 9.9 ± 1.0 15.5 ± 1.7

0.32 ± 0.03 0.54 ± 0.06 1.2 ± 0.1 3.1 ± 0.3 7.0 ± 0.7 10.5 ± 1.1

0.20 ± 0.02 0.35 ± 0.04 0.80 ± 0.09 2.0 ± 0.2 4.6 ± 0.5 6.9 ± 0.7

0.56 ± 0.06 0.95 ± 0.11 2.2 ± 0.2 5.6 ± 0.6 12.6 ± 1.3 18.9 ± 2.0

■

APPENDIX: SOLUTION FREE ENERGIES FOR D-FRUCTOSE CONVERSION We show whether the experimental findings described in the main text are supported theoretically using quantum-chemical calculation with continuum solvent model.53 The changes in the electronic energy (ΔEvac) and solution free energy (ΔGsol) calculated in vacuum and in PCM DMSO are plotted in Figure 9 for each elementary reaction step of the D-fructose conversion to 5-HMF shown in Figure 5. In panel a, the free-energy diagrams for the following conversion reactions of D-fructose into 5-HMF are summarized:

lies in favor of D-glucose (the equilibrium constant defined as k+1/k−1 is 0.13 to 0.20), in good agreement with the result obtained for the noncatalytic hydrothermal reactions of 22 D-glucose. When k2 is compared with kHMF″, kHMF′, and kHMF in DMSO, the rate constant for the dehydration of D-fructose into 5-HMF is overwhelmingly (three orders of magnitude) larger in DMSO than in water. This can be interpreted in terms of the solvation free energy. The precursors, 5-HMF″ and 5-HMF′, and 5-HMF are more hydrophobic than D-fructose (F5‑β), so that water (more polar than DMSO) can stabilize 5-HMF″, 5-HMF′, and 5-HMF less strongly. That is to say, the generation of 5-HMF″, 5-HMF′, and 5-HMF is more preferable in DMSO than in water, which results in the large difference in the rate for the D-fructose conversion. In methanol, as seen in Table 3, it is found that k2,6‑ahr and k3,6‑ahr are on the same order and k1,6‑ahr is one order of magnitude smaller than k2,6‑ahr and k3,6‑ahr. The rate constant for the dehydration of D-fructose is larger by a factor of 1.5 to 3 in methanol (ktotal) than in water (k2).

F5 ‐ β → 5‐HMF″

(A1)

5‐HMF″ → 5‐HMF′ + 2H 2O

(A2)

5‐HMF′ → 5‐HMF + H 2O

(A3)

It is found that ΔGsol of these double-bond-forming dehydration processes are all smaller than the corresponding ΔEvac due to the relatively favored solvation of the dehydrated species by aprotic DMSO; in particular, ΔGsol of the dehydration step from 5-HMF″ to 5-HMF′ drastically decreases. It also turns out that the solution free energy of the final product state (5-HMF + 3H2O) is smaller than that of the initial reactant state (β-D-fructofuranose, F5‑β). In contrast, the final product state is less stable in vacuum. Thus a key is played by the solvation in controlling the direction of the reaction. Note that three water molecules are finally generated during the dehydration of the monosaccharide F5‑β to 5-HMF. Thus, the solvation free energy of water contributes greatly to the stabilization of the product state as well as the enhancement of the reaction. Actually, as seen in Figure 8a, the two product states of 5-HMF′ + 2H2O and 5-HMF + 3H2O are stabilized more strongly than the other two states of F5‑β and 5-HMF″. In PCM water and methanol, the energy diagram similar to Figure 9 is obtained. The fact that the dehydration rate is overwhelmingly larger in DMSO than in water and that the generation of anhydro-D-fructoses instead of 5-HMF predominantly occurs in methanol cannot be interpreted by the calculation. This is probably due to the limitation of the PCM calculation for the solvation free energies, in particular, in protic (hydrogen bonding) solvents, water, and methanol. Further hydrolysis of 5-HMF to formic and levulinic acids is observed not in DMSO but in water, as mentioned in Section 3.3. The free-energy diagrams in PCM DMSO and water for the following hydrolysis fragmentation reaction are shown in Figure 9b:

4. CONCLUSIONS By applying the time-resolved in situ 13C NMR spectroscopy, we have elucidated the solvent effect on the pathways and mechanisms for the D-fructose conversion in DMSO, water, and methanol as solvents. Thus the pathway control can be made possible by tuning the solvent. In DMSO, we have shown that D-fructose is converted first into 3,4-dihydroxy-2-dihydroxymethyl5-hydroxymethyltetrahydrofuran (containing no double bond), subsequently dehydrized into 4-hydroxy-5-hydroxymethyl-4,5dihydrofuran-2-carbaldehyde (containing one double bond), and finally into 5-HMF (containing two double bonds). These precursors with no and one double bond as well as 5-HMF are fivemembered cyclic compounds consisting of furanose (or furan) ring, so that the conversion reaction is considered to take place via the β-furanose form of D-fructose. The characteristics observed for the water reactions are essentially common to those in DMSO, except that 5-HMF generated is further hydrolyzed into formic and levulinic acids or transformed into 1,2,4-benzenetriol and that D-glucose is generated through the reversible transformation of D-fructose. The product distribution in methanol is quite different from those in DMSO and water; only some anhydroD-fructoses are present, and 5-HMF is absent during the reaction. According to the kinetic analysis, the rate constant for the conversion of D-fructose into 5-HMF is overwhelmingly (three orders of magnitude) larger in DMSO than in water. In view of the path selectivity and reaction rate, DMSO is advantageous to controlling the reaction pathways to the selective production of the valuable compound, 5-HMF. If the purpose of conducting the D-fructose reaction is to produce useful organic acids such as formic and levulinic acids, then it will be accomplished by employing a hydrothermal method.

5‐HMF + 2H 2O → formic acid + levulinic acid

(A4)

Although the hydrolysis of 5-HMF to formic and levulinic acids is predicted to take place easily not only in water but also in DMSO according to the calculation, the generation of formic and levulinic acids is observed in experiment only in the case of the water system. 2110



Article

In fact, the stabilization of the final fragmentation in water is favored by a very small amount, but the difference is too small to explain the drastic differences in the reaction pathways. Similarly, as mentioned above, this is probably due to the limitation of the PCM calculation for the solvation free energies for such hydrogen bonding species as water, formic, and levulinic acids. More precise liquid theory or MD simulation is needed for functional molecules derived from carbohydrates.

■

(10) Huber, G. W.; Chheda, J. N.; Barrett, C. J.; Dumesic, J. A. Production of Liquid Alkanes by Aqueous-Phase Processing of BiomassDerived Carbohydrates. Science 2005, 308, 1446−1450. (11) Girisuta, B.; Jassen, L. P. B. M.; Heeres, H. J. Kinetics Study on the Acid-Catalyzed Hydrolysis of Cellulose to Levulinic Acid. Ind. Eng. Chem. Res. 2007, 46, 1696−1708. (12) Matson, T. D.; Barta, K.; Iretskii, A. V.; Ford, P. C. One-Pot Catalytic Conversion of Cellulose and of Woody Biomass Solids to Liquid Fuels. J. Am. Chem. Soc. 2011, 133, 14090−14097. (13) Whistler, R. L.; Smart, C. L. Isolation of Crystalline D-Glucose and Cellobiose from an Enzymatic Hydrolyzate of Cellulose. J. Am. Chem. Soc. 1953, 75, 1916−1918. (14) Huber, G. W.; Iborra, H.; Corma, A. Synthesis of Transportation Fuels from Biomass: Chemisty, Catalysts, and Engineering. Chem. Rev. 2006, 106, 4044−4098. (15) Bose, S.; Armstrong, D. W.; Petrich, J. W. Enzyme-Catalyzed Hydrolysis of Cellulose in Ionic Liquids: A Green Approach toward the Production of Biofuels. J. Phys. Chem. B 2010, 114, 8221−8227. (16) Zakrzewska, M. E.; Bogel-Łukasik, E.; Bogel-Łukasik, R. Ionic Liquid-Mediated Formation of 5-Hydroxymethylfurfural−A Promising Biomass-Derived Building Block. Chem. Rev. 2011, 111, 397−417. (17) Duus, J. Ø.; Gotfredsen, C. H.; Bock, K. Carbohydrate Structural Determination by NMR Spectroscopy: Modern Methods and Limitations. Chem. Rev. 2000, 100, 4589−4614. (18) Agarwal, V.; Huber, G. W.; Conner, C.; Auerbach, S. M. Simulating Infrared Spectra and Hydrogen Bonding in Cellulose I at Elevated Temperatures. J. Chem. Phys. 2011, 135, 134506 (13 pages). (19) Klemm, D.; Heublein, B.; Fink, H. -P.; Bohn, A. Cellulose: Fascinating Biopolymer and Sustainable Raw Material. Angew. Chem., Int. Ed. 2005, 44, 3358−3393. (20) Damager, I.; Engelsen, S. B.; Blennow, A.; Møller, B. L.; Motawia, M. S. First Principles Insight into the α-Glucan Structures of Starch: Their Synthesis, Conformation, and Hydration. Chem. Rev. 2010, 110, 2049−2080. (21) Hansen, P. I.; Larsen, F. H.; Motawia, S. M.; Blennow, A.; Spraul, M.; Dvortsak, P.; Engelsen, S. B. Structure and Hydration of the Amylopectin Trisaccharide Building Blocks−Synthesis, NMR, and Molecular Dynamics. Biopolymers 2008, 89, 1179−1193. (22) Kimura, H.; Nakahara, M.; Matubayasi, N. In Situ Kinetic Study on Hydrothermal Transformation of D-Glucose into 5-Hydroxymethylfurfural through D-Fructose with 13C NMR. J. Phys. Chem. A 2011, 115, 14013−14021. (23) Kimura, H.; Nakahara, M.; Matubayasi, N. Non-catalytic Hydrothermal Elimination of the Terminal D-Glucose Unit from Malto- and Cello-oligosaccharides through Transformation to DFructose. J. Phys. Chem. A. 2012, 116, 10039−10049. (24) Watanabe, M.; Aizawa, Y.; Iida, T.; Nishimura, R. Catalytic Glucose and Fructose Conversions with TiO2 and ZrO2 in Water at 473 K: Relationship between Reactivity and Acid-Base Property Determined by TPD Measurement. Appl. Catal., A 2005, 295, 150−156. (25) Watanabe, M.; Aizawa, Y.; Iida, T.; Aida, T. M.; Levy, C.; Sue, K.; Inomata, H. Glucose Reactions with Acid and Base Catalysts in Hot Compressed Water at 473 K. Carbohydr. Res. 2005, 340, 1925−1930. (26) Asghari, F. S.; Yoshida, H. Acid-Catalyzed Production of 5Hydroxymethyl Furfural from D-Fructose in Subcritical Water. Ind. Eng. Chem. Res. 2006, 45, 2163−2173. (27) Asghari, F. S.; Yoshida, H. Dehydration of Fructose to 5Hydroxymethylfurfural in Sub-Critical Water over Heterogeneous Zirconium Phosphate Catalysts. Carbohydr. Res. 2006, 341, 2379−2387. (28) Yao, C.; Shin, Y.; Wang, L.-Q.; Windisch, C. F., Jr.; Samuels, W. D.; Arey, B. W.; Wang, C.; Risen, W. M., Jr.; Exarhos, G. J. Hydrothermal Dehydration of Aqueous Fructose Solutions in a Closed System. J. Phys. Chem. C 2007, 111, 15141−15145. (29) Asghari, F. S.; Yoshida, H. Kinetics of the Decomposition of Fructose Catalyzed by Hydrochloric Acid in Subcritical Water: Formation of 5-Hydroxymethylfurfural, Levulinic, and Formic Acids. Ind. Eng. Chem. Res. 2007, 46, 7703−7710.

ASSOCIATED CONTENT

S Supporting Information *

In situ 13C spectra for D-[2-13C]-fructose in DMSO, water, and methanol obtained before reaction, IR spectra for D-fructose in DMSO after reaction, and experimental and computed 13C chemical shifts for the precursors (5-HMF″ and 5-HMF′) and anhydro-D-fructoses (1,6‑ahrF5-α, 2,6‑ahrF5-β, and 3,6‑ahrF5-α) obtained in DMSO and methanol, respectively. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone/Fax: 0774-38-3071. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Grants-in-Aid for Scientific Research (nos. 21300111 and 23651202) from the Japan Society for the Promotion of Science, by the Grant-in-Aid for Scientific Research on Innovative Areas (no. 20118002) and the Elements Strategy Initiative for Catalysts & Batteries from the Ministry of Education, Culture, Sports, Science, and Technology, and by the Nanoscience Program, the Computational Materials Science Initiative, and the Strategic Programs for Innovative Research of the Next-Generation Supercomputing Project. M.N. acknowledges the support for the Water Chemistry Energy Laboratory from Asahi Glass Co., Ltd.

■

REFERENCES

(1) Farrell, A. E. Ethanol Can Contribute to Energy and Environmental Goals. Science 2006, 311, 506−508. (2) Chheda, J. N.; Huber, G. W.; Dumesic, J. A. Liquid-Phase Catalytic Processing of Biomass-Derived Oxygenated Hydrocarbons to Fuels and Chemicals. Angew. Chem., Int. Ed. 2007, 46, 7164−7183. (3) Mascal, M.; Nikitin, E. B. Direct, High-Yield Conversion of Cellulose into Biofuel. Angew. Chem., Int. Ed. 2008, 47, 7924−7926. (4) Binder, J. B.; Raines, R. T. Simple Chemical Transformation of Lignocellulosic Biomass into Furans for Fuels and Chemicals. J. Am. Chem. Soc. 2009, 131, 1979−1985. (5) Sasaki, M.; Kabyemela, B.; Malaluan, R.; Hirose, S.; Takeda, N.; Adschiri, T.; Arai, K. Cellulose Hydrolysis in Subcritical and Supercritical Water. J. Supercrit. Fluids 1998, 13, 261−268. (6) Nagamori, M.; Funazukuri, T. Glucose Production by Hydrolysis of Starch under Hydrothermal Conditions. J. Chem. Technol. Biotechnol. 2004, 79, 229−233. (7) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek, G.; Cairney, J.; Eckert, C. A.; Frederick, W. J.; Hallett, J. P.; Leak, D. J.; Liotta, C. L.; et al. The Path Forward for Biofuels and Biomaterials. Science 2006, 311, 484−489. (8) Shin, Y.; Wang, L.-Q; Bae, I.-T.; Arey, B. W.; Exarhos, G. J. Hydrothermal Syntheses of Colloidal Carbon Spheres from Cyclodextrins. J. Phys. Chem. C 2008, 112, 14236−14240. (9) Lin, Y.-C.; Cho, J.; Tompsett, G. A.; Westmoreland, P. R.; Huber, G. W. Kinetics and Mechanism of Cellulose Pyrolysis. J. Phys. Chem. C 2009, 113, 20097−20107. 2111



Article

(30) Hansen, T. S.; Woodley, J. M.; Riisager, A. Efficient MicrowaveAssisted Synthesis of 5-Hydroxymethylfurfural from Concentrated Aqueous Fructose. Carbohydr. Res. 2009, 344, 2568−2572. (31) Assary, R. S.; Redfern, P. C.; Greeley, J.; Curtiss, L. A. Mechanistic Insights into the Decomposition of Fructose to Hydroxy Methyl Furfural in Neutral and Acidic Environments Using High-Level Quantum Chemical Methods. J. Phys. Chem. B 2011, 115, 4341−4349. (32) Qian, X. Mechanisms and Energetics for Acid Catalyzed β-DGlucose Conversion to 5-Hydroxymethylfurfural. J. Phys. Chem. A 2011, 115, 11740−11748. (33) Bicker, M.; Kaiser, D.; Ott, L.; Vogel, H. Dehydration of DFructose to Hydroxymethylfurfural in Sub- and Supercritical Fluids. J. Supercrit. Fluids 2005, 36, 118−126. (34) Nakamura, Y.; Morikawa, S. The Dehydration of D-Fructose to 5Hydroxymethyl-2-furaldehyde. Bull. Chem. Soc. Jpn. 1980, 53, 3705− 3706. (35) Musau, R. M.; Munavu, R. M. The Preparation of 5Hydroxymethyl-2-furaldehyde (HMF) from D-Fructose in the Presence of DMSO. Biomass 1987, 13, 67−74. (36) Seri, K.; Inoue, Y.; Ishida, H. Highly Efficient Catalytic Activity of Lanthanide(III) Ions for Conversion of Saccharides to 5-Hydroxymethyl-2-furfural in Organic Solvents. Chem. Lett. 2000, 22−23. (37) Amarasekara, A. S.; Williams, L. D.; Ebede, C. C. Mechanism of the Dehydration of D-Fructose to 5-Hydroxymethylfurfural in Dimethyl Sulfoxide at 150 °C: An NMR Study. Carbohydr. Res. 2008, 343, 3021− 3024. (38) Qi, X.-H.; Watanabe, M.; Aida, T. M.; Smith, R. L., Jr. Selective Conversion of D-Fructose to 5-Hydroxymethylfurfural by Ion-Exchange Resin in Acetone/Dimethyl Sulfoxide Solvent Mixtures. Ind. Eng. Chem. Res. 2008, 47, 9234−9239. (39) Shimizu, K.; Uozumi, R. Satsuma, A. Enhanced Production of Hydroxymethylfurfural from Fructose with Solid Acid Catalysts by Simple Water Removal Methods. Catal. Commun. 2009, 10, 1849− 1853. (40) Wang, F.; Shi, A. W.; Qin, X. X.; Liu, C. L.; Dong, W. S. Dehydration of Fructose to 5-Hydroxymethylfurfural by Rare Earth Metal Trifluoromethanesulfonates in Organic Solvents. Carbohydr. Res. 2011, 346, 982−985. (41) Bicker, M.; Hirth, J.; Vogel, H. Dehydration of Fructose to 5Hydroxymethylfurfural in Sub- and Supercritical Acetone. Green Chem. 2003, 5, 280−284. (42) Roman-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Phase Modifiers Promote Efficient Production of Hydroxymethylfurfural from Fructose. Science 2006, 312, 1933−1937. (43) Qi, X.-H.; Watanabe, M.; Aida, T. M.; Smith, R. L., Jr. Catalytic Dehydration of Fructose into 5-Hydroxymethylfurfural by IonExchange Resin in Mixed-Aqueous System by Microwave Heating. Green Chem. 2008, 10, 799−805. (44) Lansalot-Matras, C.; Moreau, C. Dehydration of Fructose into 5Hydroxymethylfurfural in the Presence of Ionic Liquids. Catal. Commun. 2003, 4, 517−520. (45) Moreau, C.; Finiels, A.; Vanoye, L. Dehydration of Fructose and Sucrose into 5-Hydroxymethylfurfural in the Presence of 1-H-3-Methyl Imidazolium Chloride Acting Both as Solvent and Catalyst. J. Mol. Catal. A: Chem. 2006, 253, 165−169. (46) Zhao, H. B.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Metal Chlorides in Ionic Liquid Solvents Convert Sugars to 5-Hydroxymethylfurfural. Science 2007, 316, 1597−1600. (47) Yong, G.; Zhang, Y.; Ying, J. Y. Efficient Catalytic System for the Selective Production of 5-Hydroxymethylfurfural from Glucose and Fructose. Angew. Chem., Int. Ed. 2008, 47, 9345−9348. (48) Rigal, L.; Gaset, A.; Gorrichon, J. P. Selective Conversion of DFructose to 5-Hydroxymethyl-2-furancarboxaldehyde Using a WaterSolvent-Ion-Exchange Resin Triphasic System. Ind. Eng. Chem. Prod. Res. Dev. 1981, 20, 719−721. (49) Roman-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Production of Dimethylfuran for Liquid Fuels from Biomass-Derived Carbohydrates. Nature 2007, 447, 982−986.

(50) Chheda, J. N.; Roman-Leshkov, Y.; Dumesic, J. A. Production of 5-Hydroxymethylfurfural and Furfural by Dehydration of BiomassDerived Mono- and Poly-Saccharides. Green Chem. 2007, 9, 342−350. (51) CRC Handbook of Chemistry and Physics, 93rd ed.; Haynes, W. M., Ed.; CRC Press: Boca Raton, FL, 2012; section 6. (52) D-Fructose as a solute was anhydrous, and it was almost completely dry when treated in a dry box under N2 atmosphere. The sample concentration (0.02 mol dm−3) is so low that the water content derived from the solute can be negligibly small. Although DMSO and methanol as solvents were also handled in a dry box, they may contain certain amounts of water as an impurity. The water contents in sample compounds (both solute and solvent) do not affect our insights obtained here, however. Actually, the pathways and kinetics in these solvents were clearly different from those in water, as described in detail in Sections 3.3 and 3.4; for example, (i) the dehydration rate for D-fructose is overwhelmingly faster in DMSO than in water and (ii) anhydrofructoses instead of 5-HMF is selectively generated in methanol. These differences are highlighted, not obscured, when water is absent. Impurity water of the sample is thus of no importance. (53) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. et al. Gaussian 09, revision A.02; Gaussian, Inc.: Wallingford, CT, 2009. (54) McMurry, J. E. Organic Chemistry, 8th ed.; Brooks/Cole: Belmont, CA, 2012; Chapter 25. (55) According to IR spectroscopy, the insoluble polymers precipitated were found to have double bonds, carbonyl groups, hydrogen atoms bonded to a carbon atom, and the carbonization did not proceed to completion; see Figure SI-2 in the Supporting Information. (56) A trace amount of formic acid is generated through a minor fragmentation path of the monosaccharides, D-fructose and D-glucose,22 at long reaction time, and the condensation polymerization of 5-HMF is catalyzed by the presence of formic acid. The fragmentation path is different from the hydrolysis of 5-HMF and gives no production of levulinic acid. (57) Atkins, P.; Paula, J. D. Atkins’ Physical Chemistry, 9th ed.; Oxford University Press: Oxford, U.K., 2009; Chapter 14. (58) Bozell, J. J.; Moens, L.; Elliott, D. C.; Wang, Y.; Neuenschwander, G. G.; Fitzpatrick, S. W.; Bilski, R. J.; Jarnefeld, J. L. Production of Levulinic Acid and Use as a Platform Chemical for Derived Products. Resour., Conserv. Recycl. 2000, 28, 227−239. (59) Manzer, L. E. Catalytic Synthesis of α-Methylene-γ-valerolactone: A Biomass-Derived Acrylic Monomer. Appl, Catal., A 2004, 272, 249− 256. (60) Lange, J. P.; Vestering, J. Z.; Haan, R. J. Toward ‘Bio-Based’ Nylon: Conversion of γ-Valerolactone to Methyl Pentanoate under Catalytic Distillation Condition. Chem. Commun. 2007, 3488−3490. (61) Yoshida, K.; Wakai, C.; Matubayasi, N.; Nakahara, M. NMR Spectroscopic Evidence for an Intermediate of Formic Acid in the Water-Gas-Shift Reaction. J. Phys. Chem. A 2004, 108, 7479−7482. (62) Yasaka, Y.; Yoshida, K.; Wakai, C.; Matubayasi, N.; Nakahara, M. Kinetic and Equilibrium Study on Formic Acid Decomposition in Relation to the Water-Gas-Shift Reaction. J. Phys. Chem. A 2006, 110, 11082−11090. (63) Enthaler, S. Carbon Dioxide―The Hydrogen-Storage Material of the Future? ChemSusChem 2008, 1, 801−804. (64) Majewski, A.; Morris, D. J.; Kendall, K.; Wills, M. A ContinuousFlow Method for the Generation of Hydrogen from Formic Acid. ChemSusChem 2010, 3, 431−434. (65) The 13C chemical shifts for these species emerge not at the low magnetic field of 120−160 ppm but around those for the furanoses of F5‑α and F5‑β, which implies that these species have no double bond and are structurally similar to these furanoses. In the previous studies (refs 22 and 66), the generation of anhydrosugars and their derivatives was found to be more favorable under the conditions lacking of water. In the present case, some kinds of anhydro-D-fructoses are thus expected to be generated as well. Taking the distance between the OH groups and the OH orientation into consideration, the possible structures of these anhydrosugars are given as follows: 1,2-anhydro-α-, 1,4-anhydro-β-, 1,62112



Article

anhydro-α-, 2,6-anhydro-β-, 3,6-anhydro-α-, and 3,6-anhydro-β-Dfructofuranose. According to the free energy and NMR chemical shift calculations, the possibilities for 1,2-anhydro-α-, 1,4-anhydro-β-, and 3,6-anhydro-β-D-fructofuranose are negated, and the three unknown species are assigned to be 1,6-anhydro-α-, 2,6-anhydro-β-, and 3,6anhydro-α-D-fructofuranose. (66) Choudhary, V.; Burnett, R. I.; Vlachos, D. G.; Sandler, S. I. Dehydration of Glucose to 5-(Hydroxymethyl)furfural and Anhydroglucose: Thermodynamic Insights. J. Phys. Chem. C 2012, 116, 5116− 5120. (67) Antal, M. J.; Mok, W. S. L.; Richards, G. N. Mechanism of Formation of 5-(Hydroxymethyl)-2-furaldehyde from D-Fructose and Sucrose. Carbohydr. Res. 1990, 199, 91−109. (68) Antal, M. J.; Mok, W. S. L.; Richards, G. N. Four-Carbon Model Compounds for the Reactions of Sugars in Water at High Temperature. Carbohydr. Res. 1990, 199, 111−115. (69) Moreau, C.; Durand, R.; Razigade, S.; Duhamet, J.; Faugeras, P.; Rivalier, P.; Ros, P.; Avignon, G. Dehydration of Fructose to 5Hydroxymethylfurfural over H-Mordenites. Appl. Catal., A 1996, 145, 211−224. (70) The pressures of these solvent systems in the relatively low temperature range of 90−150 °C are just close to the ambient. The activation energy (Ea) is thus defined as Ea = −R[∂ ln k/∂(1/T)]p.

2113


In Situ 13C NMR Study - American Chemical Society

Recommend Documents