Title Isomerization of Saccharides in Subcritical Aqueous Alcohols

Isomerization of Saccharides in Subcritical Aqueous Alcohols( Title Dissertation_全文 )

Author(s) Gao, Da-Ming

Citation 京都大学

Issue Date 2016-03-23

URL https://doi.org/10.14989/doctor.k19754

Right 許諾条件により本文は2016-10-01に公開

Type Thesis or Dissertation

Textversion ETD

Kyoto University

Isomerization of Saccharides in Subcritical Aqueous Alcohols

Da-Ming Gao

2016

Contents

General Introduction ··········································································1

Chapter 1 Kinetics of Sucrose Hydrolysis in a Subcritical Water-ethanol Mixture 1.1. Introduction ·····················································································4 1.2. Materials and Methods ········································································4 1.3. Results and Discussion ········································································5 1.4. Conclusions ··················································································· 11

Chapter 2 Kinetic Analysis for the Isomerization of Glucose, Fructose, and Mannose in Subcritical Aqueous Ethanol 2.1. Introduction ···················································································12 2.2. Materials and Methods ······································································12 2.3. Results and Discussion ······································································13 2.4. Conclusions ···················································································20

Chapter 3 Promotion or Suppression of Glucose Isomerization in Subcritical Aqueous Straight- and Branched-chain Alcohols 3.1. Introduction ···················································································21 3.2. Materials and Methods ······································································21 3.3. Results and Discussion ······································································22 3.4. Conclusions ···················································································27

Chapter 4 Kinetic Effect of Alcohols on Hexose Isomerization under Subcritical Aqueous Conditions 4.1. Introduction ···················································································28 4.2. Materials and Methods ······································································28 4.3. Results and Discussion ······································································28 4.4. Conclusions ···················································································37 I

Chapter 5 Production of Rare Sugars from Common Sugars in Subcritical Aqueous Ethanol 5.1. Introduction ···················································································38 5.2. Materials and Methods ······································································38 5.3. Results and Discussion ······································································39 5.4. Conclusions ···················································································46

Chapter 6 Solubility of D-Galactose, D-Talose, and D-Tagatose in Aqueous Ethanol at Low Temperature 6.1. Introduction ···················································································47 6.2. Materials and Methods ······································································47 6.3. Results and Discussion ······································································48 6.4. Conclusions ···················································································49

Chapter 7 Production of Keto-disaccharides from Aldo-disaccharides in Subcritical Aqueous Ethanol 7.1. Introduction ···················································································50 7.2. Materials and Methods ······································································50 7.3. Results and Discussion ······································································52 7.4. Conclusions ···················································································63

Concluding Remarks ·············································································64

References ·····························································································67

Acknowledgements ················································································74

List of Publications ················································································75

General Introduction

Saccharides, which are also interchangeably called “carbohydrates” or “sugars”, are a group of marvelous and abundant bestowals from nature and have a variety of vital functions to all of the living creatures. For human beings, a common use of saccharides is to provide energy since human appeared despite unknowing what saccharide is. Research on the molecular level about saccharides did not begin until the nineteenth century. Since then, the physiological functions of saccharides were gradually discovered beyond as sources and stores of energy. Saccharides were found as an important constituent of cells as supporting tissue or displaying on the surface of cells playing critical roles in cell interactions [1,2]. Saccharides also can conjugate with proteins to form glycoproteins to alter protein structure and function [3]. In recent decades, many previously unknown physiological functions of saccharides were discovered as supplements or sugar-based medicines for various illnesses. In particular, rare sugars (also called “rare saccharides”), which were defined as the sugars that are rare in nature, are found possessing many important physiological functions [4]. Among the rare saccharides, rare ketoses such as D-tagatose, D-xylulose, D-ribulose, cellobiulose, maltulose, lactulose, and melibiulose have attracted the most attentions in the scientific research fields of health and synthetic chemistry. Because D-tagatose has low energy (almost 0 kJ/g of food energy) and can be used as a drug to control type-2 diabetes and obesity, it can be used as a sweetener substitute for sucrose [5,6]. Besides, dietary restriction of energy by the intake of D-tagatose reduces the incidence of neoplastic lesions and significantly extends the maximum life span [5]. D-Xylulose is not only an energy resource for growth but also can be used in the pharmaceutical chemistry [4]. Intake of lactulose can promote the growth of Bifidobacterium [7,8] and has been proved to be used as drugs in treatment of many diseases such as hepatic encephalopathy and chronic constipation [8–10]. There are few reports on the physiological effects of D-ribulose, cellobiulose, and melibiulose because of the difficulty in their production and high cost. D-Ribulose is an important material to synthesize nucleosides [11]. Aldose-ketose isomerization is considered as the most important method to prepare these ketoses and has been significantly progressed despite that the ketose formation is not thermodynamically favorable because the formation of ketose is endothermic. Among the possible isomers, only D-glucose, D-mannose, D-galactose, D-xylose, D-ribose, L-arabinose, 1 lactose, and melibiose are naturally abundant. Free cellobiose and maltose are not abundantly found in nature, but they can be easily obtained by hydrolyzing cellulose and starch, respectively [12]. The conversion of these abundant saccharides to rare ketoses has become necessary and attracted much attention. These monosaccharides and disaccharides can be isomerized to the corresponding C-2 ketoses by alkali-catalyzed, metal-catalyzed, and enzymatic isomerizations. Biotransformations of abundant saccharides to rare saccharides have been performed primarily using aldose-ketose isomerases, epimerases, and polyol dehydrogenases [13]. Enzymatic isomerization can afford high selectively and even high yield of products. Nevertheless, the operating temperature cannot be largely elevated to accelerate the reaction and improve the reaction equilibrium to avoid denaturation of the enzymes. The metal-catalyzed isomerization is complex and the product distribution strongly depends on the type of substrate, cosolvent, metal ions, and carrier of metal ions in heterogeneous isomerization [14–20]. The post-treatment of homogeneous metal-catalyzed isomerization and the preparation and recovery of the heterogeneous catalyst are often tedious, which usually result in low final yield or high cost of the desired product. In alkali-catalyzed isomerization, glucose-type (2,3-threo-type) monosaccharides could be easily isomerized to the corresponding C-2 ketoses such as isomerization of glucose to fructose; however, mannose-type (2,3-erythro-type) monosaccharides were isomerized to the C-2 ketoses in low yields [21–24]. Among these methods, only alkali-catalyzed isomerization was considered as versatile and was exclusively applied to produce a limited number of ketoses for the commercial demand suffering from low yields and many by-products. This has the implication that many rare ketoses are not readily accessible, which consecutively hindered researches of this class of saccharides and restricted the potential applications. Therefore, it is necessary to develop a new method to synthesize these rare ketoses simply and efficiently, preferably by one-step isomerization. It was reported that the isomerization of monosaccharides occurred in subcritical water. Subcritical water, which is defined as the water remaining liquid state under pressurized conditions between the atmospheric boiling point and critical temperature, possesses properties of high ion product and low relative dielectric constant and, therefore, is versatilely used in chemical processes beneficially affecting many organic reactions in terms of reaction rate and selectivity [25]. When treating hexoses in subcritical water, the kinetic analysis showed that aldoses easily isomerized to the corresponding C-2 ketoses, although the yields of the derived monosaccharides were still low

2 despite varying the reaction temperature [26,27]. However, the reaction equilibrium constants of the monosaccharide isomerization of aldoses to ketoses became larger with increasing reaction temperature [19,28]. Adding an organic solvent often changes the apparent chemical equilibrium and selectively promotes the desired reactions [25]. Among the organic solvents, water-miscible alcohols are common cosolvent; in particular, ethanol can be safely used in food manufacturing processes. It was reported that addition of methanol or ethanol promoted the alkali-catalyzed, metal-catalyzed, and enzymatic isomerization of glucose [15,29,30]. Therefore, in this thesis, we revisited the saccharide isomerization in alcohol-modified subcritical water in an attempt to develop a new method to versatilely synthesize rare ketoses by one-step isomerization. In Chapter 1, the kinetics of sucrose hydrolysis was investigated in subcritical water and subcritical aqueous ethanol in order to examine the effect of the addition of ethanol which resulted in lowering water concentration on hydrolysis and the effect on formation of the constituent glucose and fructose during hydrolysis. In Chapter 2, the kinetics of mutual isomerization of glucose, fructose, and mannose in subcritical aqueous ethanol was studied to provide basic data of saccharide isomerization. In Chapter 3, influence of other water-miscible alcohols (methanol, 1-propanol, 2-propanol, and t-butyl alcohol) on glucose isomerization was studied. In Chapter 4, the kinetic effect of water-miscible alcohols on hexose isomerization was investigated to demonstrate that the relative dielectric constant of the aqueous alcohols affects the saccharide isomerizations. In Chapter 5, as an application of the results obtained in previous chapters, a novel method was developed to produce rare monosaccharides by using subcritical aqueous ethanol. In Chapter 6, solubility of D-galactose, D-talose, and D-tagatose in aqueous ethanol at low temperature was measured for the development of purification process. In order to reduce the load on the separation process, decreasing the temperature or increasing the ethanol concentration to precipitate desirable or undesirable components before chromatographic separation is considered a possible solution as the first step in the purification process. In Chapter 7, by applying the results that ethanol restricts the hydrolysis of disaccharide (Chapter 1) and promotes the isomerization of the reducing saccharides (Chapter 2), subcritical aqueous ethanol was used to produce keto-disaccharides from aldo-disaccharides, and the effect of the type of glycoside linkage of aldo-disaccharides on their isomerizations was investigated.

Chapter 1 Kinetics of Sucrose Hydrolysis in a Subcritical Water-ethanol Mixture

1.1. Introduction During subcritical water treatments of biological resources, chemical reactions occur simultaneously with extraction. Several reaction models including hydrolysis of disaccharides such as maltose and sucrose in subcritical water have been proposed [31–33]. Sucrose was hydrolyzed in an autocatalytic mode most easily among several tested disaccharides in subcritical water. Sucrose was first hydrolyzed to fructose and glucose, which further decomposed to small acidic compounds [31,32,34]. However, the reactions that occur during the extraction process in a water-ethanol mixture under subcritical conditions are still unclear. In this chapter, the influence of the addition of ethanol into water on the kinetics of sucrose hydrolysis and formation of glucose and fructose under subcritical conditions were studied.

1.2. Materials and Methods 1.2.1. Materials Sucrose (purity, >97%), D-fructose (>99%), and D-glucose (>98%) were purchased from Wako Pure Chemical Industries (Osaka, Japan).

1.2.2. Hydrolysis of sucrose in a subcritical water-ethanol mixture Sucrose was dissolved in distilled water and then mixed with ethanol to produce solutions with a final sucrose concentration of 0.5% (w/v) and water concentrations of 20‒100% (v/v). The solutions were sonically degassed under reduced pressure before the subcritical treatment. The solution reservoir was connected to a helium gasbag to prevent re-dissolution of atmospheric oxygen. The solution was delivered into a coiled stainless steel tubular reactor (0.8 mmϕ × 1.0 m) immersed in an SRX 310 silicone oil bath (Toray-Dow-Corning Silicone, Tokyo, Japan), with a residence time of 10‒240 s, by an LC-10ADVP HPLC pump (Shimadzu, Kyoto, Japan). The reaction was conducted in the temperature range of 160‒ 190°C. The reactor effluent was directly introduced to a stainless steel tube immersed in an ice bath to terminate the reaction. The pressure inside the tube was regulated at ca. 10 MPa by a back-pressure valve (Upchurch Scientific, Oak Harbor, USA). The effluent was collected in a test tube for HPLC analysis. The experiments were carried out in triplicate, and the obtained 4 values were averaged. The residence time was calculated from the volumetric flow rate measured at room temperature according to the inner diameter, length of the stainless steel tube, and density of the water-ethanol mixture under subcritical conditions. The density of the mixture was calculated according to the densities of water and ethanol [35,36], assuming that additivity of the volume holds even for the mixture of ethanol and water under subcritical conditions.

1.2.3. Analysis The concentrations of the residual sucrose and the glucose and fructose products were determined by HPLC. The system was consisted of an LC-10ADVP HPLC pump, an RI-101 refractometer (Showa Denko, Tokyo), and a Supelcogel Ca column (7.8 mmϕ × 300 mm, Sigma-Aldrich Japan, Tokyo) with a guard column (4.6 mmϕ × 50 mm, Sigma-Aldrich Japan). The columns were kept at 60°C in a CTO-10AVP column oven (Shimadzu). The pH of the reactor effluent was measured using a D-51 pH meter (Horiba, Kyoto, Japan) with a 6377-10D pH electrode (Horiba) at room temperature.

1.3. Results and Discussion 1.3.1. Effects of temperature and water concentration on the rate of sucrose hydrolysis Figure 1-1 shows the changes in the remaining fraction of sucrose with residence time at different water concentrations and temperatures. Sucrose was hydrolyzed faster at higher temperature for all the water concentrations. For instance, the half-lives of sucrose at 190 and 160°C in subcritical water were ca. 45 and 180 s, respectively. The sucrose hydrolysis rate decreased with decreasing water concentration. For example, the half-life of sucrose at 190°C was almost 4 times longer in 20% (v/v) water than in subcritical water alone. In addition, the half-life at 180°C in 40% (v/v) water was near to that in subcritical water at 160°C. Because sucrose consumes water during its hydrolysis, water can be considered as not only a solvent but also a substrate in the reaction. Therefore, sucrose would be hydrolyzed more slowly at lower water concentration.

1.0

0.8

0.6

0.4

0.2 (a) (b)

0 (c) (d) 0.8

0.6 Remaining fraction of sucrose fractionRemaining 0.4

0.2

0 0 100 200 0 100 200 Residence time [s]

Fig. 1-1. Change in the remaining fraction of sucrose with residence time at (a) 160°C, (b) 170°C, (c) 180°C, and (d) 190°C and different water concentrations (v/v): () 100%, () 80%, () 60%, () 40%, and () 20%. Curves were drawn based on Eq. (1-3).

1.3.2. Kinetic analysis of sucrose hydrolysis Since sucrose was hydrolyzed through the autocatalytic mode in subcritical water [31], the same mode was assumed for sucrose hydrolysis in the subcritical water-ethanol mixture. In the autocatalytic mode, the hydrolysis rate is proportional to the concentrations of sucrose, water and resulting products, such as acidic compounds. Therefore, the hydrolysis rate can be expressed as a function of the molar concentrations of water, CW; the remaining sucrose, CS; the total sucrose, CSt; and the consumed sucrose, CSt – CS: dC S  kC C (C  C ) (1-1) dt S W St S where k is the rate constant. Although water is consumed during sucrose hydrolysis, its concentration is significantly higher than that of sucrose. Therefore, CW can be considered as a constant in the kinetic equation. Equation (1-1) can be then rewritten as follows: dY  k'Y (1Y ) (1-2) dt where Y is the remaining fraction of sucrose, CS/CSt, and k′ is the apparent rate constant, which is equal to kCwCSt. Equation (1-2) can be integrated with the initial condition Y = Y0 at t = 0 as follows: 6

1Y 1Y ln  k't  ln 0 (1-3) Y Y0 Based on Eq. (1-3), ln[(1−Y)/Y] was plotted against t. Figure 1-2 shows typical plots for the case of sucrose hydrolysis in 40% (v/v) water in ethanol. The straight-line plots indicated that sucrose was hydrolyzed autocatalytically in the subcritical water-ethanol mixture, similarly to the hydrolysis in subcritical water. The apparent rate constant k′ could be calculated from the slope of the line. The k′ value declined not only with decreasing temperature but also with decreasing water concentration (Fig. 1-3). Due to the difficulty in estimating the concentration of ionized water, the k′ value was plotted against the water concentration in Fig. 1-3. The apparent rate constants k′ had linear relationships with the water concentrations, indicating that the addition of ethanol resulted in a dilution effect on sucrose hydrolysis. The curves in

Fig. 1-1 were drawn by substituting the estimated k′ and Y0 values in Eq. (1-3).

The rate constant k was then obtained from the equation k′ = kCStCW. The k values did not depend on the water concentration. To investigate the dependence of the rate constant on temperature, the k values at different water concentrations were averaged and plotted against the reciprocal of absolute temperature (Arrhenius plot, Fig. 1-4). The plot was linear with a high correlation coefficient (R2 = 0.985). From the line, the frequency factor and activation energy were evaluated to be 1.7 × 109 L2 mol-2 s-1 and 90 kJ/mol, respectively. The activation energy was similar to that of acid-catalyzed hydrolysis of sucrose (ca. 95 kJ/mol) [37], and

1

]

)/ Y

- 2

[(1 ln 3

4

5 0 50 100 150 200 250 Residence time [s]

Fig. 1-2. Estimation of the apparent rate constants, k′, of sucrose hydrolysis at () 160°C, () 170°C, () 180°C, and () 190°C in a 40% (v/v) water-ethanol mixture.

7 those of lactose and D-melezitose hydrolyses in subcritical water (ca. 85 kJ/mol for both saccharides) [32,33,38]. On the other hand, it was a little smaller than those of other disaccharides, such as maltose (ca.118 kJ/mol) [32], turanose (ca. 152 kJ/mol) [32] and melibiose (ca. 132 kJ/mol) [32], indicating that the hydrolysis rate constant of sucrose had similar sensitivity toward temperature to lactose and D-melezitose.

0.15

]

1 -

[s k’

0.10

0.05

rateApparent constant,

0 0 20 40 60 Water concentration [mol/L] Fig. 1-3. Dependence of apparent reaction rate constants, k′, of sucrose hydrolysis on water concentration in the 160‒190°C range. Symbols are the same as in Fig. 1-2.

Temperature [oC] 190 180 170 160

]

 s

1 2

- 10

mol 2

[L k

Rate constant,

102

2.15 2.20 2.25 2.30 2.35 103/T [K1] Fig. 1-4. Arrhenius plot for the rate constant, k, of sucrose hydrolysis. The k values at different water concentrations were averaged. 8

1.3.3. Yields of glucose and fructose during sucrose hydrolysis Figure 1-5 shows the typical dependence of the yields of glucose and fructose on the conversion of sucrose at water concentrations of 100, 60, and 20% (v/v). When only hydrolysis occurs, equimolar amounts of glucose and fructose should be produced. The glucose and fructose yields were lower than the conversion of sucrose, indicating that the monosaccharides also underwent decomposition. It was reported that, in subcritical water or in a water-alcohol mixture under alkali conditions at 60°C [27,29], the monosaccharides decomposed to various compounds, such as 5-hydroxymethylfurfural and acids [39–41], or isomerized into each other or another saccharide such as mannose [27]. However, mannose was not detected in this study. Therefore, it can be considered that sucrose was first hydrolyzed to glucose and fructose, followed by the decomposition and isomerization of these two monosaccharides. In subcritical water, fructose and glucose were formed in almost the same yields, and the yields in the 40‒80% (v/v) water mixtures were higher than those in subcritical water alone. This indicates that the decomposition rates of fructose and glucose would be faster in

1.0

0.8 (a) (b) (c) 0.6

0.4

0.2

0 0.2 0.4 0.6 0.8 1.0 0.8 (d) (e)

0.6 Yields glucose of fructose and Yields 0.4

0.2

0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1.0 Conversion of sucrose

Fig. 1-5. Typical relationships between the yields of glucose and fructose and conversion of sucrose at (, ) 160°C, (▲, ) 170°C, (, ) 180°C, and (, ) 190°C and different concentrations of water (v/v): (a) 100%, (b) 80% (c) 60%, (d) 40%, and (e) 20%. Closed and open symbols represent glucose and fructose, respectively. The lines were drawn under the assumption that decomposition of glucose and fructose did not occur.

9 subcritical water. On the other hand, 20% (v/v) water mixture gave the lower yield of fructose. Although its reason is unclear, the presence of ethanol at moderate concentration may promote the stabilization of glucose and fructose. In the subcritical water-ethanol mixtures, the fructose yields were lower than those of glucose, and these differences became greater with decreasing water concentration. Possible reasons for the difference are as follows: 1) the decomposition rate constant of fructose was greater than that of glucose; 2) isomerization between fructose and glucose occurred; 3) the addition of ethanol may have promoted the decomposition and isomerization of glucose and fructose and 4) a combination of more than one of these factors may be in play.

1.3.4. pH change during sucrose hydrolysis When sucrose was treated in subcritical water or water-ethanol mixtures, the pH decreased due to the decomposition of the glucose and fructose into acidic compounds [31,42,43]. Figure 1-6 shows the relationship between the pH of the effluent measured at room temperature and the conversion of sucrose in water concentrations of 100, 60, and 20% (v/v). Changes in the pH were larger at longer residence times with higher conversions of sucrose (Figs. 1-1 and 1-6). In subcritical water, the pH decreased further after the sucrose was completely hydrolyzed. This corresponded to the decreases in the fructose and glucose yields at the sucrose conversion of 1.0 in Fig. 1-5, indicating that weak acids were produced through the decomposition of the monosaccharides [38]. In 60 and 20% (v/v) water mixtures, changes in the pH were ca. 2, and these values were smaller than that in subcritical water (ΔpH = ca. 3). Therefore, the hydrolysis would be suppressed in 20–80% water mixture due to the lower concentration of hydrogen ion (Fig. 1-1). 8 (a) (b) (c)

6 pH

2 0.01 0.1 0.01 0.1 0.01 0.1 1 Conversion of sucrose Fig. 1-6. Typical relationships between pH and conversion of sucrose at () 190°C, () 180°C, () 170°C, and () 160°C at different water concentrations (v/v): (a) 100%, (b) 60%, and (c) 20%. 10

1.4. Conclusions Sucrose underwent autocatalytic hydrolysis in the subcritical water-ethanol mixtures. The rate of sucrose hydrolysis to glucose and fructose was reduced with increasing ethanol concentration, and ethanol showed a dilution effect on the conversion. The temperature dependence of the reaction rate constant for sucrose hydrolysis obeyed the Arrhenius equation. The fructose and glucose products underwent further decomposition, and the yield of fructose was much lower than that of glucose when the ethanol concentration increased. Thus, ethanol exerted other effects on the reaction in the subcritical water-ethanol mixture.

Chapter 2

Kinetic Analysis for the Isomerization of Glucose, Fructose, and

Mannose in Subcritical Aqueous Ethanol

2.1. Introduction Nowadays, fructose plays an important role in food and related industries. The current industrial isomerization process to produce fructose involves the utilization of an immobilized glucose isomerase. This enzymatic isomerization is reversible (the equilibrium constant, Keq, is ca. 1 at 25°C) and slightly endothermic, indicating that the maximum attainable yield of fructose is governed by the reaction temperature [19,28]. Chapter 1 showed that the degree of hydrolysis of sucrose changed by the addition of ethanol under subcritical conditions. The yield of fructose was lower than that of glucose, and this difference increased with increasing ethanol concentration, possibly because of the isomerization between the two saccharides. However, the mechanism of isomerization has not been clarified. Therefore, in this chapter, the kinetics of the isomerization of glucose, mannose, and fructose in subcritical aqueous ethanol was analyzed.

2.2. Materials and Methods 2.2.1. Materials Ethanol, D-fructose, and D-glucose were the same as those described in Section 1.2.1. D-Mannose (> 99 %) was purchased from Nacalai Tesque (Kyoto, Japan).

2.2.2. Isomerization among fructose, glucose, and mannose in subcritical aqueous ethanol The isomerization treatment of glucose, fructose and mannose was as the same as that described in Section 1.2.2. The residence time was in the range of 30‒500 s. The reaction temperature was in the range of 180‒200°C.

2.2.3. HPLC analysis The conditions of compositional analysis of the reaction mixture were also the same as those described in Section 1.2.3.

2.3. Results and Discussion 2.3.1. Effect of ethanol concentration on glucose isomerization Figure 2-1 shows the effect of ethanol concentration on glucose isomerization at 180°C. When the ethanol concentration exceeded 40% (v/v), glucose isomerization significantly increased with increasing ethanol concentration. Fructose and mannose were competitively produced from glucose both in subcritical water and in subcritical aqueous ethanol. The yield of fructose was almost sevenfold higher than that of mannose at 500 s in 60% (v/v) aqueous ethanol. Mannose was not detected when the ethanol concentration was below 40% (v/v). These results were in contrast to those of the hydrolysis of sucrose in subcritical aqueous ethanol. The hydrolysis of sucrose decreased with increasing ethanol concentration, and the isomer of sucrose was not obtained (Chapter 1). This indicates that only the isomerization of reducing sugars is promoted in subcritical aqueous ethanol. Figure 2-1 also shows that ethanol affected the isomerization and decomposition behaviors of glucose. The yield of fructose increased with increasing ethanol concentration at 500 s (Fig. 2-2). The selectivity of fructose also increased by the addition of ethanol and reached the highest value in 60% (v/v) aqueous ethanol, where the selectivity was defined as the molar ratio of the produced monosaccharide to the consumed substrate. Although the selectivity of mannose was lower than that of fructose, it reached ca. 10% when the ethanol concentration was >60% (v/v). The total saccharide concentration was maintained at a high level, which was

1.0 0.3 (a) (b)

0.9 0.2 0.8

0.1

0.7 Remaining fraction of glucose of fraction Remaining

mannoseand fructose of Yield 0.6 0 0 200 400 0 200 400 600 Residence time [s] Fig. 2-1. Changes in (a) the fraction of remaining glucose treated at 180C in () 0% (v/v) (subcritical water alone), () 20%, () 40%, (□) 60%, and () 80% subcritical aqueous ethanol and (b) the yields of the fructose (open symbols) and mannose (closed symbols) obtained with residence times. The symbols in (b) are the same as those in (a). The curves show the calculated results. 13

1.0

0.8

0.6

0.4

fructose of Yield 0.2

content saccharide Total

Selectivity of fructose and of fructose mannose Selectivityand 0 0 20 40 60 80 100 Concentration of ethanol [% (v/v)]

Fig. 2-2. Effect of ethanol concentration on the (, ) selectivity and (, ) yield of fructose, (, ) selectivity of mannose, and (, ) the total saccharide content when glucose was treated at 180C (closed symbols) and 190C (open symbols) for 500 s.

slightly lower than the feed glucose concentration and scarcely depended on the ethanol concentration at 180°C, regardless of the increase in the conversion of glucose. This indicates that the side reaction of glucose was not significantly accelerated and that most of the consumed glucose was converted to fructose and mannose in subcritical aqueous ethanol at 180°C. The contribution of ethanol in promoting glucose isomerization under subcritical aqueous conditions is still unclear. It is known that the pKa values of glucose in aqueous ethanol and methanol-d4 decrease with increasing alcohol concentration at ambient temperature and that the initial reaction rate of alkali-catalyzed glucose isomerization is positively related to the ionization constant [29]. Therefore, an increase in ethanol concentration would accelerate glucose isomerization. Another important role of ethanol is that it changes the anomeric equilibrium of glucose, thus changing the apparent chemical equilibrium and facilitating the isomerization of glucose [44]. Ethanol not only changes the physical properties of the sugar solution but also participates in hydrogen atom migration during the isomerization [17]. On the other hand, a recent study reported that the initial presence of water in aqueous alcohol suppressed the aluminum-containing zeolite-catalyzed isomerization of glucose to fructose at 120°C, and that the one-pot synthesis of fructose from glucose could not be performed using aqueous alcohol [45]. However, these reported results are different from our results. One of

14 the reasons may be that the reaction temperature was different in our study, and the role of ethanol in promoting glucose isomerization may change under subcritical conditions.

2.3.2. Temperature dependence of glucose isomerization Figure 2-3 shows the effect of temperature on glucose isomerization in 80% (v/v) subcritical aqueous ethanol. Increasing reaction temperature increased the conversion of glucose and the yields of fructose and mannose at a given residence time. The maximum yield of fructose was achieved faster at higher temperatures. However, the total saccharide concentration at a given residence time decreased with increasing reaction temperature (Fig. 2-2), indicating that the fraction of disappeared hexoses increased at higher temperatures. Another disadvantage of increasing the reaction temperature is that the selectivity of fructose decreased. A higher maximum yield of fructose was realized at higher temperatures even though the selectivity was lower. In contrast, the selectivity increased at lower temperatures even though a longer residence time was required. Therefore, an appropriate temperature should be selected for the efficient production of fructose from glucose. Similar temperature effects were also observed at other ethanol concentrations. To achieve the maximum yield of fructose within a short time, the isomerization was performed in 80% (v/v) aqueous ethanol at 200°C in the subsequent studies. 1.0 0.4 (a) (b)

0.8 0.3

0.6 0.2 0.4

0.1

0.2

Remaining fraction of glucose of fraction Remaining Yield of fructose and mannoseand fructose of Yield 0 0 0 200 400 0 200 400 600 Residence time [s] Fig. 2-3. Changes in (a) the fraction of remaining glucose at () 180C, () 190C, and () 200C in 80% (v/v) subcritical aqueous ethanol and (b) the yields of fructose (open symbols) and mannose (closed symbols) obtained with residence times. The symbols in (b) are the same as those in (a). Curves show the calculated results.

2.3.4. Mutual isomerization of hexoses To investigate the mutual isomerization among glucose, mannose, and fructose, the latter two saccharides were also treated under the same conditions as those for glucose. Figure 2-4 shows the isomerizations of fructose to glucose and mannose, glucose to fructose and mannose, and mannose to fructose and glucose in 80% (v/v) aqueous ethanol at 200°C. The consumption of fructose was the slowest among the three saccharides. Although mannose and glucose were produced from fructose almost in the same yields, the yields were low. These results indicated that fructose significantly decomposed under these conditions. The isomerization of mannose proceeded faster than that of fructose and glucose. Fructose was most produced from mannose. However, when mannose was treated with 80% (v/v) aqueous ethanol at 200°C, the yield of fructose first reached the highest value at a residence time of ca. 150 s and then decreased at longer residence times, indicating that the fructose obtained was consequently decomposed (Fig. 2-4(b)). The yield of fructose produced from mannose was higher than that of fructose produced from glucose within the investigated residence time. Glucose was produced from mannose in a relatively low yield; however, the yield was more than that of mannose produced from glucose.

1.0 (a) (b) 0.6 0.8

0.6 0.4

0.4

0.2 fructose from

glucose, and mannose and glucose, 0.2 Remaining fraction of fructose, of fraction Remaining

0 0 mannose and glucose and glucose,

Yield of fructose and mannose from from mannoseand fructose of Yield glucose, fructose and mannose from from mannose and fructose glucose, 0 200 400 0 200 400 600 Residence time [s]

Fig. 2-4. Changes in (a) the fractions of remaining () fructose, () glucose, and () mannose at 200C in 80% (v/v) subcritical aqueous ethanol and (b) the yields of the hexoses obtained with residence times. Symbols  and  represent fructose and glucose produced from mannose;  and  represent fructose and mannose produced from glucose; and  and  represent mannose and glucose produced from fructose, respectively.

Decomposition products kFd Fructose

kF→G kM→F kF→M kG→F

k Glucose M→G Mannose kG→M kGd kMd Decomposition products Decomposition products Scheme 2-1. Simplified reaction pathways for the isomerization of three hexoses in subcritical aqueous ethanol.

2.3.5. Kinetic analysis of the isomerization and decomposition of monosaccharides As shown above, glucose, mannose, and fructose were isomerized and decomposed in parallel with subcritical aqueous ethanol. The probable reaction pathways are shown in

Scheme 2-1, where ki (i = FG, FM, Fd, GF, GM, Gd, MF, MG, and Md) are the rate constants, and F, G, M, and d represent fructose, glucose, mannose, and the decomposition of these three monosaccharides, respectively. When the decomposition and isomerization of the monosaccharides were assumed to follow first-order kinetics [30], the reaction rate of each monosaccharide can be expressed as follows: dC F  (k  k  k )C  k C  k C (2-1) dt FG FM Fd F GF G MF M dC G  (k  k  k )C  k C  k C (2-2) dt GF GM Gd G FG F MG M dC M  (k  k  k )C  k C  k C (2-3) dt MG MF Md M FM F GM G where CF, CG, and CM are the concentrations of fructose, glucose, and mannose, respectively. The rate constants for the reaction of each substrate were evaluated by minimizing the sum of the residual square between the experimental and calculated Ci values using the Solver of Microsoft® Excel 2010. The curves in Figs. 2-1, 2-3, and 2-4 were drawn using the estimated rate constants. The rate constants obtained for the isomerization and decomposition of each substrate at 180, 190, and 200°C were plotted against the ethanol concentrations shown in Fig.

2-5. The values of kM→F, kG→F, and kFd were larger than the others under any condition, and the kM→F value was the largest, indicating that mannose was most easily isomerized to 17

10-1 ]

1 (a) (b) (c) - -2

[s 10 k

10-3

10-4 Rate constant, Rate

10-5 0 20 40 60 80 0 20 40 60 80 0 20 40 60 80 100 Ethanol concentration [% (v/v)]

Fig. 2-5. Dependencies of the rate constants of the respective reaction steps on the ethanol concentration at (a) 180C, (b) 190C, and (c) 200C. The rate constants are expressed as follows: () kMF, () kFd, () kGF, () kMd, () kGd, () kMG, () kFM, () kFG, and () kGM.

fructose. In contrast, the rate constants for the reverse reactions, kF→M and kF→G, were much smaller. The rate constants for the isomerizations between mannose and glucose, kM→G and kG→M, were also smaller than kM→F and kG→F. These results indicated that the isomerization of mannose and glucose to fructose was faster than the reverse reactions and the isomerization between mannose and glucose and that these isomerizations were accelerated by the addition of ethanol and increase in temperature. Although decompositions were also promoted by increase in temperature and ethanol concentration, they were less sensitive to temperature and ethanol concentration than isomerizations. The kM→F was the most sensitive to the change in ethanol concentration.

To investigate the temperature dependence of the rate constants, the ki values were plotted against the reciprocal of absolute temperature (Arrhenius plot). Figure 2-6 shows the typical Arrhenius plots of the rate constants obtained in 80% (v/v) subcritical aqueous ethanol. The rate constants of the respective reaction steps separately lie on the straight lines. The frequency factors and activation energies were calculated to be in the ranges of 108–1012 s−1 and 90–130 kJ/mol, respectively. The activation energies of the isomerizations were similar to those of the monosaccharide decompositions in 100% (v/v) subcritical water [46], thus indicating that the isomerization and decomposition of monosaccharides would have a similar energy barrier.

2.3.6. Reaction equilibrium for the isomerization of monosaccharides Figure 2-5 also shows that the isomerizations of mannose to fructose and glucose to fructose have high reaction equilibrium constants, Keq. The Keq values were calculated from the rate constants (Keq,MF = kMF/kFM and Keq,GF = kGF/kFG); they were not affected by ethanol concentration. The equilibrium constants for each reaction at different ethanol concentrations were averaged and plotted against the reciprocals of absolute temperature according to van’t Hoff equation: H ln K   (2-4) eq RT where ΔH and R are the change in enthalpy and gas constant, respectively (Fig. 2-7). The equilibrium constants for the isomerization of mannose to fructose, Keq,M→F, were higher than those for the isomerization of glucose to fructose, Keq,G→F. The plot of the equilibrium constant for each isomerization lies on a straight line with high correlation coefficients (0.997 for Keq,M→F and 0.996 for Keq,G→F). The enthalpies for the isomerizations of mannose to fructose, ΔHM→F, and glucose to fructose, ΔHG→F, were calculated to be 18 and 24 kJ/mol, respectively. The positive enthalpies indicate that both the isomerizations were endothermic reactions. Therefore, the yield of fructose would increase with increasing temperature.

Temperature [K] 473 463 453

10-2

] 1

[s k k

10-3

-4

Rate constant, 10

10-5 2.10 2.14 2.18 2.22 103/T [K-1]

Fig. 2-6. Arrhenius plot for the rate constants of the respective reaction steps in 80% (v/v) subcritical aqueous ethanol. Symbols are the same as those shown in Fig. 2-5.

Temperature [K]

473 463 453

eq K

101 Equilibrium constants,

100 2.10 2.14 2.18 2.22 103/T [K-1]

Fig. 2-7. Temperature dependence of the equilibrium constants for the isomerization of mannose to fructose (, Keq,MF ) and glucose to fructose (, Keq,GF ).

2.4. Conclusions The isomerizations among glucose, mannose, and fructose were significantly promoted in subcritical aqueous ethanol. Mannose and glucose were easily isomerized to fructose. However, the isomerizations of fructose to glucose and mannose and that between glucose and mannose were not favorable in subcritical aqueous ethanol. Fructose mainly underwent decomposition when it was used as a substrate. The kinetic study showed that kM→F, kG→F, and kFd were larger than the other rate constants. Both the isomerizations of mannose to fructose and glucose to fructose had high equilibrium constants, indicating that subcritical aqueous ethanol may be a useful reaction medium to produce high fructose syrup.

Chapter 3

Promotion or Suppression of Glucose Isomerization in Subcritical

Aqueous Straight- and Branched-chain Alcohols

3.1. Introduction The influence of the addition of alcohols on the decomposition of disaccharides and the isomerization of monosaccharides has been investigated at ambient temperatures [29,47,48]. It was reported that methanol and ethanol can markedly accelerate the alkali-catalyzed saccharide isomerization; however, 1- and 2-propanol cannot give these satisfactory results [15,29,48]. Chapter 2 showed that subcritical aqueous ethanol could remarkably promote the isomerizations of glucose and mannose to fructose and that the isomerizations were accelerated with increasing ethanol concentration. However, effects of the kind of alcohols have not been clarified. Therefore, the influence of the type of the water-miscible alcohols (methanol, ethanol, 1-propanol, 2-propanol, and t-butyl alcohol) on the isomerization of glucose under subcritical conditions was investigated in this chapter.

3.2. Materials and Methods 3.2.1. Materials Straight-chain alcohols (methanol, ethanol, and 1-propanol) and branched-chain alcohol (2-propanol and t-butyl alcohol) were purchased from Wako Pure Chemical Industries (Osaka, Japan). D-Mannose, D-glucose, and D-fructose were the same as those described in Section 2.2.1.

3.2.2. Isomerization of glucose in subcritical aqueous alcohols The isomerization procedure was as the same as that described in Section 2.2.2. The temperature and residence time were set to 180°C and 30‒500 s, respectively, for the isomerization in aqueous methanol, ethanol, and 1- and 2-propanol. For the isomerization in subcritical aqueous t-butyl alcohol, they were set to 180°C or 200°C and 100‒1000 s. The density of methanol, 1- or 2-propanol under subcritical conditions was calculated based on the reported data [49–51]. However, because the density of t-butyl alcohol under

21 subcritical conditions was not reported, it was assumed to be the same as that of 2-propanol.

3.2.3. Saccharide analysis The reactor effluent was collected in a test tube for HPLC analysis. The HPLC system was consisted of an LC-10AD VP HPLC pump (Shimadzu, Kyoto, Japan), an RI-101 refractometer (Showa Denko, Tokyo, Japan), and a Cosmosil Sugar-D column (4.6 mmϕ × 250 mm, Nacalai Tesque, Kyoto, Japan). A mixture of water and acetonitrile (20:80, v/v) was employed as the mobile phase at a flow rate of 1.0 mL/min. The column was maintained at 30°C in a CTO-10A VP column oven (Shimadzu).

3.3. Results and Discussion 3.3.1. Isomerization of glucose in subcritical aqueous alcohols Figure 3-1 shows the typical change in the fraction of remaining glucose with residence time at 180°C in subcritical water and in 60% (v/v) subcritical aqueous alcohols. When glucose was treated for 500 s in subcritical water, the conversion was about 13%. However, the conversion of glucose at a residence time of 500 s was almost doubled and was 27–31% in the presence of any of the water-miscible primary or secondary alcohols. Chapter 2 showed that ethanol can promote the conversion of glucose. These facts show that addition of the primary and secondary alcohols will also achieve this conversion. In addition, there was no

1.0

0.9

0.8

0.7

of fractionglucose Remaining 0.6 0 200 400 600

Residence time [s] Fig. 3-1. Changes in the fraction of remaining glucose with residence time in () subcritical water and 60% (v/v) subcritical aqueous alcohols (() methanol, () ethanol, () 1-propanol, () 2-propanol, and () t-butyl alcohol) at 180°C.

0.3

0.2

0.1

mannose fructose of and Yield 0

0 200 400 600 Residence time [s] Fig. 3-2. Yields of fructose and mannose derived from glucose at various residence times in subcritical water and in 60% (v/v) subcritical aqueous alcohols at 180°C. Symbols are the same as those in Fig. 3-1, and the open and closed symbols represent the yields of fructose and mannose, respectively. obvious difference in the promotion ability among the four alcohols. Fructose was produced from glucose with high yield and selectivity, while mannose was produced with low yield and selectivity in subcritical aqueous methanol, and 1- and 2-propanol (Fig. 3-2). The yields of both fructose and mannose were increased by the addition of these alcohols. These facts indicate that the primary and secondary alcohols used can also promote isomerization. The type of alcohol slightly affected the yields of fructose and mannose. This is in contrast to the reported results, which showed that methanol promoted isomerization more efficiently than ethanol in the alkali-catalyzed isomerization of glucose at low temperature, and that 1- and 2-propanol could not promote the isomerization reaction. [15,48]. On the other hand, the addition of t-butyl alcohol suppressed the conversion of glucose as discussed in detail later.

3.3.2. Effect of the concentration of methanol, and 1- and 2-propanol on the isomerization of glucose Figure 3-3 shows the typical influence of the concentrations of methanol, 1- and 2-propanol on the selectivities of the derived fructose and mannose, yield of fructose, and fraction of

23 degraded hexoses, for the treatment of glucose at 180°C for 500 s. The yields of fructose increased with increasing concentration of the alcohols. In the 0–40% (v/v) concentration range, the yield of fructose showed a weak dependence on the concentration of the alcohols. However, in the higher concentration range, increasing the concentration of the alcohols, especially of 2-propanol, greatly raised the yield of fructose. Dependence of the selectivity for fructose on the alcohol concentration was different to that of the yield: The selectivity for fructose reached a maximum value in 60% (v/v) alcohol and then decreased when the concentration of the alcohols increased to 80% (v/v). One reason for this decrease in 80% (v/v) alcohol may be that the decomposition of fructose is promoted by the addition of alcohols (Chapter 2). Although the addition of alcohol can also promote the decomposition of hexoses, the fraction of degraded hexoses did not increase with an increasing concentration of methanol, 1-propanol, or, especially, 2-propanol. When the concentration of 2-propanol was increased to 80% (v/v), most of the consumed glucose was isomerized to fructose and mannose, and 2-propanol exhibited a better effect than methanol and 1-propanol on the isomerization of glucose to fructose. On the other hand, selectivity for mannose was kept at a low level in the presence of the three alcohols. The mechanism of the promotion of isomerization in subcritical aqueous alcohols is unclear. However, isomerization of glucose depended very little on the particular primary and secondary alcohols in subcritical aqueous conditions. Similarity in the behaviors of methanol, ethanol, 1- and 2-propanol in promoting isomerization would indicate that these alcohols promote the isomerization by the same mechanism. Glucose, mannose, and fructose can be mutually interconverted by acid, base, or acid-base catalysts [52–56]. Recent research

1.0 (a) (b) (c) 0.8

0.6

0.4

Yield of fructoseof Yield 0.2

0 20 40 60 80 0 20 40 60 80 0 20 40 60 80 100 Fraction of disappeared hexoses of disappeared Fraction Selectivity of fructose and mannose and fructose of Selectivity Concentration of alcohol [% (v/v)] Fig. 3-3. Effects of the alcohol concentration on the selectivities of () fructose and () mannose, () fraction of disappeared hexoses, and () yield of fructose for the treatment of glucose at 180°C for 500 s. (a) Methanol, (b) 1-propanol, and (c) 2-propanol. 24 demonstrated that, in aqueous alcohol with high alcohol concentration, the strength of the bond between the proton and oxygen atom of the alcoholic hydroxyl group was weaker than in pure alcohol, especially when the temperature exceeded 130°C [57]. It was also found that glucose could exchange its C-1 proton with the proton of the hydroxyl group of methanol to form fructose [17]. Accordingly, a possible mechanism under the subcritical conditions is that there is a higher dissociation of the alcohol (RO-H) to RO− and H+, either of which can catalyze the isomerization. Because the proton-accepting or electron-donating ability of RO− is higher than that of OH− [58], RO− could promote the isomerizations more effectively than OH‒ through alkali isomerization.

3.3.3. Isomerization of glucose in subcritical aqueous t-butyl alcohol The concentration of t-butyl alcohol also affected the conversion of glucose (Fig. 3-4). However, in contrast to the cases for methanol, ethanol, 1- and 2-propanol, the isomerization behavior was different in t-butyl alcohol: Glucose was converted more slowly with an increasing concentration of t-butyl alcohol compared to the conversion of glucose in subcritical water. When glucose was treated in 20% (v/v) t-butyl alcohol at 180°C for 1000 s, its conversion was almost the same as that at 180°C for 400 s in subcritical water.

1.0

0.9

0.8

of fractionglucose Remaining 0.7 0 400 800 1200 Residence time [s]

Fig. 3-4. Changes in the fraction of remaining glucose with residence time at 180°C in () 80% (v/v), () 60%, () 40%, () 20%, and () 0% t-butyl alcohol and at () 200°C in 60% (v/v) t-butyl alcohol.

Fructose was also formed from glucose in subcritical aqueous t-butyl alcohol (Fig. 3-5). However, the isomerization behavior was different from that in other subcritical aqueous alcohols, as described above: Only fructose was formed, and mannose was not detected in the presence of t-butyl alcohol. Besides, the yield of fructose was much lower than those obtained in subcritical water and in the presence of the other alcohols tested under the same reaction conditions. The formation of fructose was slower at a higher t-butyl alcohol concentration. However, the conversion and isomerization were accelerated by increasing the temperature (Figs. 3-4 and 3-5). When glucose was treated at 200°C for 1000 s in 60% (v/v) t-butyl alcohol, the conversion of glucose and the yield of fructose reached ca. 25% and 10%, respectively, which were about threefold and 2.5-fold higher, respectively, than those obtained at 180°C. Figure 3-6 shows the typical influence of t-butyl alcohol concentration on the selectivity for fructose, the yield of fructose, and the fraction of degraded hexoses, for the treatment of glucose at 180°C for 1000 s. The dependence of the selectivity for fructose on the t-butyl alcohol concentration was different to that of the other alcohols tested. The selectivity for fructose decreased with increasing t-butyl alcohol concentration. In particular, there was a large decrease in the selectivity when the concentration of t-butyl alcohol exceeded 40% (v/v). The yield of fructose almost linearly decreased with increasing t-butyl alcohol concentration.

0.12

0.08

0.04 Yield of fructoseof Yield

0 0 400 800 1200 Residence time [s]

Fig. 3-5. Yields of fructose derived from glucose at different residence times in subcritical aqueous t-butyl alcohol. The symbols are the same as those in Fig. 3-4.

0.8

hexoses 0.6

0.4

0.2

of fructose yieldand Selectivity Fraction of disappeared of disappeared Fraction 0 0 20 40 60 80 100 Concentration of t-butyl alcohol [% (v/v)]

Fig. 3-6. Effects of t-butyl alcohol concentrations on () the selectivity of fructose, () the fraction of degraded hexoses, and () the yield of fructose for the treatment of glucose at 180°C for 1000 s.

The fraction of degraded hexoses was also low in subcritical aqueous t-butyl alcohol and was only slightly affected by the t-butyl alcohol concentration. In line with the proposed mechanism regarding the promotion of the isomerization of glucose by primary and secondary alcohols, the reason for the suppression of the isomerization of glucose in subcritical aqueous t-butyl alcohol may be that t-butyl alcohol + − would be dissociated to H and the t-butoxy anion, (CH3)3CO , which possesses much stronger proton-accepting ability than the alkoxide anions of the other alcohols tested. However, the bulky t-butyl group gives the t-butoxy anion greater steric hindrance, which could suppress the isomerization. As a result, the addition of t-butyl alcohol only caused a diluting effect on the water concentration, which was similar to the addition of ethanol to subcritical water during the hydrolysis of sucrose, as described in Chapter 1.

3.4. Conclusions The isomerization of glucose was promoted with increasing alcohol concentration in subcritical aqueous primary and secondary alcohols. The type of alcohol slightly affected the isomerization. However, the addition of t-butyl alcohol suppressed the isomerization. These facts suggest a mechanism for the promotion of the monosaccharide isomerization in subcritical aqueous alcohols.

Chapter 4

Kinetic Effect of Alcohols on Hexose Isomerization under

Subcritical Aqueous Conditions

4.1. Introduction Changing the reaction solvent is an alternative method for changing the apparent reaction equilibrium and rate constants, in order to compensate for the problems that arise from varying the reaction temperature which may result in serious decomposition of substrates and products. Chapters 2 and 3 showed that the use of a different solvent such as subcritical aqueous methanol, ethanol, 1-propanol, and 2-propanol promoted the isomerization of glucose to fructose, whereas subcritical aqueous t-butyl alcohol suppressed the isomerization, in comparison with the case in subcritical water. It was also reported that the product distribution obtained from aldose isomerization using a combination of calcium chloride and sodium hydroxide in aqueous alcohols, depended both on the nature of the added alcohol and the substrates [15]. In this chapter, the effects of methanol, ethanol, 1-propanol, 2-propanol, and t-butyl alcohol on the kinetics of the mutual isomerization of mannose, glucose, and fructose were investigated to show the product distributions and reaction specificities of the hexoses.

4.2. Materials and Methods The materials, isomerization procedures, and the compositional analysis of the reaction mixture were all the same as those described in Section 3.2.

4.3. Results and Discussion 4.3.1. Isomerization of mannose and fructose in subcritical aqueous alcohols Mannose isomerization was greatly promoted in subcritical aqueous methanol, ethanol, 1-propanol, and 2-propanol compared to subcritical water (Fig. 4-1). This was exemplified by the fact that the conversion of mannose in the 80% (v/v) aqueous alcohol at the residence time of 500 s was >60% (ca. 75% in subcritical aqueous primary alcohols) compared to ca. 30% in subcritical water. The nature of the primary alcohols used had little effect on the conversion of mannose; however, the yield of fructose depended on the type of alcohol used. Fructose was

1.0 0.6 (b) 0.8

0.4 0.6

0.4 0.2

0.2 (a)

Yields of fructose and glucoseandfructoseof Yields Fraction of remaining mannose remaining of Fraction 0 0 0 200 400 0 200 400 600 Residence time [s] Fig. 4-1. Changes in (a) the fraction of remaining mannose treated at 180C in () subcritical water and 80% (v/v) aqueous alcohols ((□) methanol, () ethanol, () 1-propanol, () 2-propanol, and () t-butyl alcohol) and (b) the yields of fructose (open symbols) and glucose (closed symbols) as a function of residence time. The symbols in (b) represent similar quantities as (a). The curves for mannose isomerization in subcritical water and subcritical aqueous primary and secondary alcohols are constructed from the calculated values, whereas the broken curves for subcritical aqueous t-butyl alcohol were empirically drawn. The curves for mannose isomerization in subcritical aqueous ethanol are cited from Chapter 2.

obtained at higher yields in subcritical aqueous 1-propanol compared to the other aqueous alcohols. On the other hand, subcritical aqueous methanol and 2-propanol afforded almost similar and relatively low yield of fructose (ca. 45%) at the residence time of 500 s. The differences in the fructose yield for the various aqueous alcohols may be ascribed to the competition in the formation of fructose from mannose and its decomposition in the presence of subcritical aqueous alcohols. The conversion of fructose was also promoted by the addition of methanol, ethanol, 1-propanol, and 2-propanol (Table 4-1).

Table 4-1. Effects of the concentrations of methanol, ethanol, 1-propanol, 2-propanol, and t-butyl alcohol on fructose isomerization under subcritical conditions at 180C.

Total Solvent Residence Fructose Yield of Yield of content of [%, v/v] time [s] conversion [%] mannose [%] glucose [%] hexoses [%]

Water 500 22.1 ND* ND 77.9 Methanol (20%) 500 30.1 4.9 4.1 78.9 Methanol (40%) 500 36.6 7.1 6.3 76.9 Methanol (60%) 500 43.8 8.0 7.0 71.2 Methanol (80%) 500 60.7 8.4 9.6 57.4 Ethanol (20%) 500 20.5 ND ND 79.5 Ethanol (40%) 500 25.3 4.6 4.2 83.5 Ethanol (60%) 500 29.4 5.1 5.4 81.1 Ethanol (80%) 500 34.3 5.7 6.5 77.9 1-Propanol (20%) 500 31.6 7.5 5.2 81.2 1-Propanol (40%) 500 35.1 9.1 6.6 80.6 1-Propanol (60%) 500 39.9 10.3 7.8 78.2 1-Propanol (80%) 500 52.3 10.7 9.9 68.3 2-Propanol (20%) 500 25.3 6.4 4.9 86.0 2-Propanol (40%) 500 32.6 7.5 5.7 80.5 2-Propanol (60%) 500 42.5 10.3 6.4 74.3 2-Propanol (80%) 500 50.7 7.6 7.7 64.7 t-Butyl alcohol (20%) 1000 32.0 ND ND 68.0 t-Butyl alcohol (40%) 1000 20.0 ND ND 80.0 t-Butyl alcohol (60%) 1000 15.0 ND ND 85.0 t-Butyl alcohol (80%) 1000 11.0 ND ND 89.0 *Not detected

In contrast to the case of treating mannose, the nature of the alcohol used affected the conversion of fructose, as indicated by the results obtained in 80% (v/v) subcritical aqueous alcohols at a residence time of 500 s. Methanol (80% (v/v)) yielded a ca. 61% conversion of fructose, whereas ethanol, 1-propanol, and 2-propanol (every 80% (v/v)) afforded fructose

30 conversion of only ca. 34, 52, and 51%, respectively. Fructose was mainly decomposed and barely isomerized in these subcritical aqueous alcohols. Chapter 3 showed that the isomerization of glucose to fructose was also promoted by these primary and secondary alcohols under subcritical aqueous conditions. However, a significant difference was the slower conversion of glucose in aqueous ethanol (80% (v/v)) compared to other aqueous alcohols, which, therefore, resulted in low fructose yield. It was reported that aqueous methanol was a better solvent than aqueous ethanol for the mannose-to-fructose isomerization catalyzed by the combination of calcium chloride and sodium hydroxide [15]. Besides, in the alkali-catalyzed isomerization process glucose was abundantly accessible from fructose and vice versa both in aqueous methanol as well as in aqueous ethanol, whereas this was not the case in aqueous 1- and 2-propanol [29,48]. However, these results are different from the results obtained in this chapter. One of the reasons may be that the reaction temperature used in the present study was different from that used in the previous studies, and the role of alcohols in promoting reducing sugar isomerizations may change under different subcritical aqueous conditions. The substrate-dependent chemoselective isomerization of the investigated monosaccharides in subcritical aqueous primary and secondary alcohols was unique compared to the alkali- and metal-catalyzed isomerizations [15,17,59]. In alkaline aqueous solutions, fructose isomerized to glucose with the highest rate and excellent selectivity. However, the isomerization of mannose was the slowest, and both glucose and fructose were produced almost in the same yield [59]. Moreover, mannose isomerized slower than glucose in many ethanolic and methanolic solutions of metal ions [15,17]. Many studies reported that glucose was isomerized to fructose more easily than mannose under alkaline conditions [44,52,60]. These studies were carried out at low reaction temperatures. The unique product distribution among the investigated saccharides in subcritical aqueous primary and secondary alcohols can be attributed to the difference in the temperature dependences of the rate constants of the reactions and to the anomeric equilibrium of the hexoses. The difference in the temperature dependence of the rate constants for the decomposition and isomerization of fructose would result in its preferential decomposition at subcritical conditions. On the other hand, the shift in anomeric equilibrium of mannose can be considered to affect its product distribution. Chapter 3 postulated the following mechanism for the shift in anomeric equilibrium and the final product distribution. The primary and secondary alcohols are likely to dissociate to alkoxides to catalyze the isomerization reactions under subcritical conditions. Mannose prefers the

α-configuration in aqueous solutions at ambient temperature [61], while the β-configuration is required and kinetically favored for the isomerization of mannose to fructose [52]. Since the use of high temperature facilitated the endothermic transformation of α to β-configuration [61], mannose showed high isomerization reactivity under subcritical conditions. However, although glucose prefers the β-configuration, which is also required for isomerization to fructose in aqueous solutions at ambient temperature, the use of high temperature shifted the glucose anomeric equilibrium to the side of the α-configuration [61]. Thus, high temperature decreased the fraction of glucose in the β-configuration, which decelerated the glucose isomerization reaction. Elevating the temperature of the glucose isomerization reaction primarily changed the isomerization rate constant. Therefore, the change in anomeric equilibrium obtained by elevating the temperature may be the reason for the unique product distributions of mannose and glucose. Changing the reaction solvent to subcritical aqueous t-butyl alcohol, however, suppressed the isomerization reactions of mannose and fructose (Fig. 4-1 and Table 4-1). The conversions of mannose and fructose did not exceed 15% in 80% (v/v) aqueous t-butyl alcohol, even when the residence time was extended to 1000 s. When mannose was used as the substrate, fructose was produced at low yields, and glucose was not detected. On the other hand, when fructose was used as the substrate, glucose and mannose were not detected, i.e., fructose predominantly decomposed.

4.3.2. Effects of alcohol concentration on the isomerization reactions of mannose and fructose Figure 4-2 shows the effect of the concentrations of methanol, ethanol, 1-propanol, 2-propanol, and t-butyl alcohol on mannose isomerization performed at 180°C with a residence time of 500 s. The yield of fructose increased with increase in the concentrations of methanol, ethanol, 1-propanol, and 2-propanol. On the other hand, change in the t-butyl alcohol concentration negatively affected hexose isomerization, which will be discussed later. In the primary and secondary alcohol concentration ranges of 0‒60% (v/v), the yields of fructose were almost the same. The selectivity of fructose formation did not monotonically depend on the concentration of methanol, ethanol, 1-propanol, and 2-propanol, and different kinds of alcohols showed different fructose selectivity patterns. The differences in the molecular structure of alcohols and in the alcohol-water molar ratio at a given volumetric concentration of alcohol may cause differences in the dependence of fructose selectivity on

32 alcohol concentration. Chapter 2 reported that the promotion of the decomposition of hexoses by increasing the alcohol concentration caused the decrease in the selectivity of fructose at high alcohol concentrations. The selectivity of glucose was low (<0.15) and weakly depended on the primary and secondary alcohol concentrations. The onset of decomposition decreased the total hexose content with increase in the concentrations of methanol, ethanol, and 1-propanol, especially in subcritical aqueous methanol. The total hexose content decreased to ca. 83% in 80% (v/v) methanol compared to ca. 90% in 80% (v/v) 1-propanol at a residence time of 500 s. Subcritical aqueous 2-propanol was relatively unfavorable for the decomposition of hexoses since the total hexose content was relatively high in 40‒80% (v/v) aqueous 2-propanol. Increase in the concentrations of the primary and secondary alcohols slightly promoted fructose isomerization and mainly decreased the total hexose content, especially in aqueous methanol (Table 4-1). The yields of mannose and glucose from fructose increased only by ca. 3 and 5%, respectively, whereas the total hexose content decreased by >21% when the methanol concentration was changed from 20% to 80% (v/v) at the residence time of 500 s.

0.6 1.0 (a) (b) 0.8 0.4

fructose 0.6

0.2 0.4

of Yield Selectivity of fructose of Selectivity 0 0.2 (c) (d)

content 0.1 0.9

Total hexose Total glucose of Selectivity 0 0.8 0 20 40 60 80 0 20 40 60 80 100 Concentration of alcohol [% (v/v)]

Fig. 4-2. Effects of () methanol, () ethanol, () 1-propanol, () 2-propanol, and () t-butyl alcohol concentrations on (a) the yield of fructose, (b) selectivity for fructose, (c) selectivity for glucose, and (d) the total hexose content for the treatment of mannose at 180C for 500 s. The results for subcritical aqueous ethanol are also cited from Chapter 2.

On the other hand, the conversion of mannose to fructose barely changed in 20–80% (v/v) subcritical aqueous t-butyl alcohol (data not shown). In subcritical aqueous t-butyl alcohol, mannose was converted faster than glucose, and with higher yield of fructose from the former than from the latter (Chapter 3). In other words, mannose was isomerized more easily than glucose in subcritical aqueous alcohols regardless of the nature of alcohol used. In contrast to mannose, the conversion of fructose was markedly suppressed depending on the t-butyl alcohol concentration (Table 4-1). Isomerization of glucose was suppressed in subcritical aqueous t-butyl alcohol depending on its concentration (Chapter 3). These results show that subcritical aqueous t-butyl alcohol suppressed the hexose isomerization differently for different hexoses, depending on the molecular structure. Since hexose isomerization was suppressed in subcritical aqueous t-butyl alcohol, hexose isomerization in aqueous t-butyl alcohol was not kinetically analyzed.

4.3.3. Kinetics of isomerization of the hexoses in subcritical aqueous methanol, 1-propanol, and 2-propanol Kinetics of hexose isomerization in subcritical primary and secondary alcohols was investigated for determining the relationship between the rate constants and the properties of the aqueous alcohols, according to the method described in Chapter 2. The curves for mannose isomerization in subcritical water and subcritical aqueous primary and secondary alcohols shown in Fig. 4-1 were drawn using the estimated rate constants. The rate constants obtained for the isomerization and decomposition of each monosaccharide are

10-2

] 1

- (a) (b) (c)

[s k k 10-3

10-4 Rate constant, Rate

10-5 0 20 40 60 80 0 20 40 60 80 0 20 40 60 80 100 Alcohol concentration [% (v/v)] Fig. 4-3. Dependence of the rate constants of the respective reaction steps on the (a) methanol, (b) 1-propanol, and (c) 2-propanol concentrations at 180C. The rate constants are expressed as follows: () kMF, () kFd, () kGF, () kMd, () kGd, () kMG, () kFM,

() kFG, and () kGM. 34 plotted against the alcohol concentrations in Fig. 4-3. In all the subcritical aqueous alcohols, the values of kMF, kGF, and kFd were larger than the others at any alcohol concentration, with kMF being the largest, indicating that mannose was most easily isomerized to fructose.

In contrast, the rate constants for the reverse reactions, kFM and kFG, were lowest. The rate constants for the isomerization reactions involving mannose and glucose, kMG and kGM, were also much smaller than kMF and kGF. These results indicate that the isomerization reactions of mannose and glucose to fructose were faster than the reverse reactions and the isomerization between mannose and glucose, and that these isomerization reactions were accelerated by the addition of the primary and secondary alcohols. It has been reported that the isomerization rate constant is affected by the relative dielectric constant of the solvent [62]. Figure 4-4 shows the change in the relative dielectric constant of each aqueous alcohol as a function of its volumetric concentration, which was determined by extrapolation using a method reported in the literature [63,64]. The relative dielectric constant of the aqueous alcohols decreased with increase in the alcohol concentration as well as with increase in the number of carbon atoms contained in the alcohol. A linear relationship between the logarithmic value of each rate constant and the relative dielectric constant of the solution was determined as follows:

constant

dielectric Relative 0

0 20 40 60 80 100 Concentration of alcohol [% (v/v)]

Fig. 4-4. The relative dielectric constants of the aqueous alcohol solutions at various alcohol concentrations ((□) methanol, () ethanol, () 1-propanol, and () 2-propanol) at 180C and 10 MPa.

-2.0

-2.5

G→F k

log log

, -3.0

F M→ k log -3.5

-4.0

0 10 20 30 40 50 Relative dielectric constant of aqueous alcohols

Fig. 4-5. Typical relationships between the rate constants ((closed symbols) kMF and (open symbols) kGF) and the relative dielectric constants of aqueous (, ) methanol, (, ) ethanol, (, ) 1-propanol, and (►,) 2-propanol at 180C and 10 MPa.

log ki  log ki0  αiε (4-1) where ε is the relative dielectric constant of the solvent, αi is the sensitivity of the reaction rate constant to the change in the relative dielectric constant of the solvent, and ki0 is the reaction rate constant at ε = 0. Figure 4-5 shows typical linear relationships between kMF and kGF and the relative dielectric constant of the aqueous alcohols. Change in the reaction solvent from water to 80% (v/v) aqueous alcohol resulted in an increase in the kMF and kGF values by >0.5 logarithmic unit. It was found that the isomerization rate constant could be expressed as a function of the ionization constant of glucose, which increased with increase in the concentrations of methanol or ethanol in water [29]. This, in turn, accelerated the isomerization rate.

Table 4-2. αi and log ki0 values from Eq. (4-1) for each rate constant. i F→G G→F F→M M→F M→G G→M Fd Gd Md

αi 0.0232 0.0214 0.0242 0.0261 0.0264 0.0182 0.0145 0.0022 0.0075

log ki0 -3.08 -2.75 -2.88 -2.17 -3.30 -3.77 -2.84 -3.97 -3.65

The α and log k0 values of each reaction rate constant were calculated from the slope and intercept of the lines constructed according to Eq. (4-1), and these values are shown in Table 4-2. The isomerization rate constants were more sensitive to the change in the relative dielectric constant of the solvent than the decomposition rate constants, implying that the addition of any alcohol, except for t-butyl alcohol, mainly promoted the isomerization reactions. In particular, the rate constants of mannose isomerization were more sensitive than those of other monosaccharides. Increase in the alcohol concentration was more favorable for fructose decomposition than for glucose and mannose decompositions. Moreover, the largest log k0 value was obtained for kMF.

4.4. Conclusions Mutual isomerization reactions among mannose, glucose, and fructose were promoted in subcritical aqueous primary and secondary alcohols, and were suppressed in subcritical aqueous t-butyl alcohol. The aldose-to-ketose isomerization was more favorable than the ketose-to-aldose isomerization. These facts indicated that subcritical aqueous primary and secondary alcohols are promising to be used for industrial production of ketoses from the corresponding aldoses by one-step reaction. A simple relation between the rate constant and relative dielectric constant of the aqueous alcohols was established to understand the reaction kinetics in subcritical aqueous alcohols. This would have potential applications to design other type reactions.

Chapter 5

Production of Rare Sugars from Common Sugars in Subcritical

Aqueous Ethanol

5.1. Introduction Chapters 2–4 showed a new approach in saccharide isomerization, in which aldoses can be converted to the corresponding C-2 ketoses in high yield and selectivity. In particular, mannose is most easily isomerized. These results indicated a possible method to produce rare sugars by treating aldoses with subcritical aqueous alcohols. Among the alcohols investigated, ethanol can be safely used in food manufacturing processes. Therefore, the objective of this chapter is to investigate the isomerization of two glucose-type aldoses, D-galactose and D-xylose, and a mannose-type aldose, D-ribose, in subcritical aqueous ethanol for efficiently producing the corresponding rare C-2 ketoses, D-tagatose, D-xylulose, and D-ribulose.

5.2. Materials and Methods 5.2.1. Materials D-Xylose, D-lyxose, D-galactose, D-tagatose, D-talose, and D-arabinose were purchased from Wako Pure Chemical Industries (Osaka, Japan). D-Xylulose and D-ribose were purchased from Sigma-Aldrich Japan (Tokyo, Japan). Because the isomerization of only the D-enantiomers was focused in this study, the prefix, D-, of all the saccharides was omitted hereafter.

5.2.2. Isomerization of common sugars to rare sugars in subcritical aqueous ethanol The isomerization procedures of each aldose was as the same as those described in previous chapters. The reaction temperature was set at 180C for the treatment of xylose and ribose and at 160–200C for the treatment of galactose. Xylose and ribose were treated in 80% (v/v) aqueous ethanol, and the ethanol concentration was changed from 0 to 80% (v/v) for the treatment of galactose. Both the concentrations of xylose and ribose were adjusted at 0.5% (w/v), and the galactose concentration was 0.5–8.5% (w/v).

5.2.3. HPLC analysis The HPLC analysis of the components of the reaction mixtures was the same as that in Chapters 3 and 4. Lyxose was analyzed by a combination of two columns, a Cosmosil Sugar-D column and a Ca2+ ion-exchange column, Supelcogel Ca (7.8 mm I.D. × 300 mm length, Sigma-Aldrich Japan, Tokyo), under the same conditions as for the other saccharides.

5.2.4. Purification and NMR confirmation of rare sugars The reactor effluent (ca. 200 mL) was concentrated to ca. 1 mL using a rotary evaporator. The concentrated solution was pretreated using a Strata C18-E cartridge column (55 μm, 70 Å, Shimadzu). A solution of 50% (v/v) aqueous methanol was used as the eluent. The solvent in the effluent from the cartridge was evaporated using a rotary evaporator. The purification of the target saccharide was performed using a Cosmosil Sugar-D column (10 mm I.D. × 250 mm length, Nacalai Tesque, Kyoto) using 95% (v/v) acetonitrile as the eluent at a flow rate of 3.0 mL/min. The 1H NMR analyses of the purified saccharides were carried out using an Ascend 400

NMR spectrometer (400 MHz, Bruker Japan, Osaka) with D2O as the solvent. Acetonitrile (δH = 2.06 ppm) was used as the internal standard. The 1H NMR spectra of the purified tagatose, xylulose, and ribulose were compared with those of the commercial samples or with literature data [65]. Confirmation of arabinose, lyxose, and talose was performed by comparing their retention times in the HPLC chromatograms with the commercial ones, as the low yields of purified products obtained prevented NMR analysis.

5.2.5. Solubility of galactose in aqueous ethanol The solubility of galactose in aqueous ethanol of different concentrations was measured at 25C. A sufficient amount of galactose was added to aqueous ethanol, and the mixture was stirred at 25C using a digital stirring plate (Koike Precision Instruments, Kyoto) to reach the equilibrium. The concentration of galactose in the supernatant was then measured using the abovementioned HPLC system.

5.3. Results and Discussion 5.3.1. Production of tagatose, xylulose, and ribulose from the corresponding common sugars Isomerization of the common sugars was performed in subcritical aqueous ethanol. Figure

(a)

Gal RI Tag

Tal

0 5 10 15

Xyl (b)

Xylu RI

Lyx

0 5 10 15 Rib (c)

Ribu RI Ara

0 5 10 15 Retention time [min]

Fig. 5-1. HPLC chromatograms of the reaction mixtures containing rare sugars formed from (a) galactose, (b) xylose, and (c) ribose in 80% (v/v) aqueous ethanol at the residence time of 300 s. The feed concentration of each substrate was 0.5% (w/v).

5-1 shows the typical HPLC chromatograms of the reaction mixtures containing rare sugars formed from the corresponding common sugars. By the treatment of common sugars in subcritical aqueous ethanol, several peaks, which would correspond to the respective rare sugars, were observed. By comparing the NMR spectra, it was seen that tagatose and xylulose were formed during the treatment. Formation of ribulose was confirmed by comparing the NMR spectrum with that in the literature [65]. Chapter 2 showed that mannose and glucose can isomerize each other in subcritical aqueous ethanol, and that arabinose, lyxose, and talose can be also formed in subcritical water from ribose, xylose, and galactose, respectively [26]. In addition, the retention times of these rare sugars in the HPLC chromatograms corresponded with those of the commercial ones. These results suggested that arabinose, lyxose, and talose would be formed in subcritical aqueous ethanol. Figure 5-2 shows the time courses of the isomerization of galactose to tagatose and talose, xylose to xylulose and lyxose, and ribose to ribulose and arabinose in 80% (v/v) aqueous ethanol at 180C. The conversions of galactose, xylose, and ribose increased with increasing residence time and reached ca. 32%, 60%, and 75%, respectively, at a residence time of 500 s. The yield of tagatose increased with the consumption of galactose and reached ca. 24% at a residence time of 500 s. In contrast, the selectivity of tagatose at first remained high (>80%) until the residence time of 200 s and then decreased at longer residence times, where the selectivity was defined as the molar ratio of the obtained product to the consumed substrate. Overall, the selectivity of tagatose was >70% within the treatment time. Talose, the C-2 epimer of galactose, was also synthesized from galactose; however, its yield was <5%. Xylulose and lyxose were synthesized from xylose, and their yields at the residence time of 500 s were ca. 38% and 2%, respectively. The selectivity of xylulose gradually decreased with the consumption of xylose; however, it was >60% at the investigated residence times. Ribose decreased most rapidly among the three substrates. Some peaks were observed in the HPLC chromatogram of the reaction mixture (Fig. 5-1), and arabinose was obtained in <10% yield at the tested residence times. On the other hand, ribulose was obtained in high yields. A maximum yield of ca. 40% was first obtained at a residence time of 200 s; however, the yield then decreased at longer residence times. At the residence time of 500 s, only 30% of the ribulose remained, indicating that the obtained ribulose was further decomposed to other compounds and that prolonging the reaction time promoted the decomposition reaction. Therefore, the optimal residence time to synthesize ribulose was set at ca. 200 s.

1.0

0.8

0.6

remaining substrateremaining 0.4

0.2 Yield isomerization of Yield product

Fraction

0 0 200 400 600

Residence time [s]

Fig. 5-2. Isomerization processes of () galactose to () tagatose and () talose, () xylose to (▲) xylulose and () lyxose, and () ribose to () ribulose and () arabinose in 80% (v/v) subcritical aqueous ethanol at 180C. The concentration of each substrate was 0.5% (w/v).

The isomerization behaviours of these three aldoses in subcritical aqueous ethanol were unique in the following four aspects compared to the alkali-catalyzed, metal-catalyzed, and enzymatic isomerizations [21,66,67]. 1) Subcritical aqueous ethanol converted both glucose-type (2,3-threo) and mannose-type (2,3-erythro) aldoses to the corresponding C-2 ketoses in high yield and selectivity. This indicates that subcritical aqueous ethanol showed different substrate-dependent chemoselective aldose-ketose isomerizations. In particular, ribulose could be synthesized in high yield from ribose rather than from arabinose. 2) The isomerizations between aldoses and ketoses were difficult. 3) The apparent reaction equilibrium shifted towards the formation of ketoses at high temperatures. 4) The production process is one-step and only requires a very simple reactor and safety solvent. In addition, recovery of catalysts was not required during the post-treatment, and ethanol can be easily recovered by evaporation for its reuse. However, there are some disadvantages. Because ethanol is contained in the reaction mixture, saturated concentration of sugars is lower than in water, indicating that the product concentration would not be high.

5.3.2. Temperature dependence of galactose isomerization Figure 5-3 shows the effect of reaction temperature on the isomerization of galactose in 80% (v/v) aqueous ethanol. The conversion of galactose increased with increasing reaction 42 temperature. When the temperature exceeded 180C, galactose underwent rapid conversion with a low yield of tagatose. The increase in the reaction temperature from 160 to 180C increased galactose conversion only by ca. 13% at a residence time of 500 s. However, when the reaction temperature was increased from 180 to 190C, the conversion of galactose also increased from ca. 34% to 71%. On the other hand, the initial reaction rates for the formation of tagatose and talose increased with increasing reaction temperature. The yield of tagatose did not exceed 8% at 160C, but increased to ca. 24% at 180C at a residence time of 500 s. However, increasing the temperature to 190 and 200C increased the yield of tagatose at a residence time shorter than 300 s, albeit with lower selectivity. The selectivity of tagatose obtained at 200C for 200 s was approximately half of that obtained at 180C at a residence time of 500 s. The results indicate that increasing reaction temperature beyond 180C favored the decomposition reaction and caused a large loss of the total saccharide yield. The maximum yields of tagatose at 180, 190, and 200C were almost the same; thus, a reaction temperature higher than 180C did not increase the yield of tagatose.

1.0

0.8

0.6

0.4

Yields tagatoseof and talose Yields 0.2 Fraction of galactoseremaining

0 600 0 200 400 Residence time [s]

Fig. 5-3. Effect of temperature, () 160C, () 180C, ( ) 190C, and () 200C, on the isomerization behavior of galactose in 80% (v/v) aqueous ethanol with a galactose feed concentration of 0.5% (w/v). Symbols , , , and  represent tagatose and symbols , , , and  represent talose synthesized at 160, 180, 190, and 200C, respectively.

5.3.3. Effect of ethanol concentration on the isomerization of galactose The isomerization of galactose significantly increased with increasing ethanol concentration (Fig. 5-4). When the reaction medium was changed from subcritical water to 80% (v/v) subcritical aqueous ethanol for the isomerization of galactose at 180C for 500 s, the conversion of galactose and the yield of tagatose increased to ca. 34% and 24%, respectively. However, the change in ethanol concentration barely affected the selectivity of tagatose and the fraction of decomposed galactose. This result indicates that increasing the concentration of ethanol promoted both the isomerization and decomposition of galactose. In contrast, the yield and selectivity of talose were hardly affected by the change in ethanol concentration. It was also reported that ethanol promoted both the isomerization and decomposition of monosaccharides in an aqueous alkaline solution [29].

5.3.4. Isomerization of galactose at different feed concentrations It is desirable to feed galactose in high concentrations without decreasing the yield and selectivity of tagatose during the isomerization. The solubility of galactose in aqueous ethanol was measured at 25C (Fig. 5-5). The addition of ethanol significantly decreased the solubility of galactose from ca. 36% (w/v) to 1.0% by changing the solvent from water to 80% (v/v) aqueous ethanol. Considering the yield of tagatose and the solubility of galactose, the reaction was carried out in 60% (v/v) ethanol at 180C to investigate the effect of galactose feed concentration on its isomerization. The solubility of galactose was ca. 9.3% (w/v) in 60% (v/v) aqueous ethanol at 25C. To avoid the precipitation of galactose, the maximum feed concentration was selected as 8.5% (w/v). Other feed concentrations were 0.5% (w/v), 1.0%, 2.0%, and 5.0%. The conversion of galactose decreased when its concentration was at 5.0% and 8.5% (w/v) (Table 5-1). For the treatment at 180C for 500 s, the yield and selectivity of tagatose decreased to ca. 13% and 57%, respectively, and the conversion of galactose decreased to ca. 23% by changing the galactose feed concentration from 0.5% (w/v) to 8.5%. However, the maximum productivity of tagatose, ca. 80 g/(L·h), was realized at a feed concentration of 8.5% (w/v). In the enzymatic isomerization, the highest productivity of tagatose was reported to be 15.3 g/(L·h) using an L-arabinose isomerase [68]. The results indicate that the subcritical aqueous ethanol treatment is promising for the practical synthesis of tagatose. The reasons for the decrease in the galactose conversion at high feed concentrations, in subcritical aqueous ethanol at high temperature, are presently unclear.

0.8

0.6

0.4

galactose, yield of tagatose 0.2

Conversion of galactose, fraction of decomposed talose,or selectivity of tagatosetalose or 0 0 20 40 60 80 100

Concentration of ethanol [% (v/v)]

Fig. 5-4. Effect of ethanol concentration on () the conversion of galactose, () yield of tagatose, () selectivity of tagatose, () yield of talose, () selectivity of talose, and () fraction of decomposed galactose. Galactose with a feed concentration of 0.5% (w/v) was treated at 180C and at a residence time of 500 s. The fraction of decomposed galactose was defined as the ratio of the concentration of the side products to that of the consumed galactose.

[% (w/v)]

Solubility of galactose 0 0 20 40 60 80 100 Concentration of ethanol [% (v/v)]

Fig. 5-5. Solubility of galactose in aqueous ethanol at 25C.

Table 5-1. Effect of galactose feed concentration on its isomerization. The isomerization reaction was carried out in 60% (v/v) subcritical aqueous ethanol at 180C and at the residence time of 500 s. Feed Galactose Yield of Yield of Selectivity of Productivity Total yield concentration conversion talose tagatose tagatose of tagatose of sugars of galactose [%] [%] [%] [%] [g/(L·h)] [%] [%, w/v] 0.5 32 7 21 66 8 96 1.0 29 5 20 69 14 96 2.0 29 4 20 69 29 95 5.0 24 4 14 58 50 93 8.5 23 3 13 57 80 93

5.4. Conclusions Three rare ketoses, tagatose, xylulose, and ribulose, were preferably synthesized by one-step isomerization from the corresponding aldoses using subcritical aqueous ethanol. In contrast, the rare aldoses, talose, lyxose, and arabinose, were synthesized in low yields. These facts clearly demonstrate that ketoses can be selectively synthesized from the corresponding aldoses using subcritical aqueous ethanol.

Chapter 6

Solubility of D-Galactose, D-Talose, and D-Tagatose in Aqueous

Ethanol at Low Temperature

6.1. Introduction Chapter 5 showed that D-tagatose can be produced in high yield by treating D-galactose with subcritical aqueous ethanol, along with the production of a small amount of D-talose. This method is also considered promising for the production of D-tagatose from D-galactose because of its high productivity and short reaction time. This method produces unreacted substrates, desired products, and some byproducts in the reaction effluent. Therefore, to reduce the load on the separation process, decreasing the temperature or increasing the ethanol concentration for precipitating desirable or undesirable components before chromatographic separation is considered a possible solution as the first step in the purification process. In this chapter, the solubility of D-galactose, D-tagatose, and D-talose in aqueous ethanol in the temperature range of –30°C to 20°C was measured.

6.2. Materials and Methods 6.2.1. Materials Ethanol, D-galactose, D-tagatose, and D-talose were the same as those used in Section 5.2.1.

6.2.3. Solubility measurement The solubility of galactose, tagatose, and talose in aqueous ethanol (20–80%, w/w) was measured in the temperature range of –30°C to 20°C. The temperature was controlled using an FP89-HL ultralow refrigerated-heating circulator (Julabo Japan, Osaka). A solution of galactose, tagatose, or talose was prepared in a vial by adding an excess amount of hexose to aqueous ethanol. The solution was kept at the measuring temperature for 24 h to ensure equilibrium solubility. The supernatant of the hexose solution was diluted to a detectable concentration for HPLC quantification with aqueous ethanol. The HPLC system was the same as that described in Section 5.2.3.

6.3. Results and Discussion Figures 6-1(a), (b), and (c) show the solubility of galactose, talose, and tagatose in 20–80% (w/w) aqueous ethanol at various temperatures according to the van’t Hoff equation: dlnS H i   i (6-1) d(1/T ) R where Si is the solubility of hexose i; ΔHi, the dissolution enthalpy of hexose i; R, the gas constant; T, the absolute temperature; and subscript i, galactose, talose, or tagatose. The solubility of all hexoses decreased upon increasing the ethanol concentration and decreasing the temperature. For example, at 20°C, the solubility of galactose, talose, and tagatose decreased from 0.093 to 0.003, 0.502 to 0.057, 0.527 to 0.048 mol/kg-solvent, respectively, upon changing the solvent from 20% (w/w) aqueous ethanol to 80% (w/w) aqueous ethanol. The solubility of galactose and tagatose in 40% (w/w) aqueous ethanol both decreased by ca. 40% upon decreasing the temperature from 20°C to –30°C, and by ca. 66% from 20°C to –20°C for talose. Further, the hexose solubility was more sensitive to temperature change at lower ethanol concentration [69,70], indicating that the solvent with low ethanol concentration was favorable for obtaining a solution with high tagatose content by hexose crystallization, which would be similar to the separation of fructose and glucose by freezing crystallization [71]. The solubility of talose and tagatose was much higher than that of galactose under all measured conditions. Hexoses exist as equilibrium mixtures of pyranose and furanose with α- or β-configurations. Therefore, the temperature dependence of the saccharide solubility cannot be simply explained. However, it was reported that the solubility

of mannose and fructose in aqueous ethanol was higher than that of glucose at the same ]

100 solvent

- (a) (b) (c)

10-1

10-2

10-3 3.0 3.5 4.0 3.0 3.5 4.0 3.0 3.5 4.0 4.5

Solubilityof each hexose [mol/kg 103/T [K-1]

Fig. 6-1. Temperature dependence of solubility of (a) galactose, (b) talose, and (c) tagatose in () 20% (w/w), () 40%, () 60%, and () 80% ethanol.

Table 6-1. Dissolution enthalpies of galactose, talose, and tagatose, ∆Hi, in aqueous ethanol.

Ethanol concentration ∆Hi [kJ/mol] [%, w/w] Galactose Talose Tagatose 20 10 16 8.6 40 3.9 17 6.7 60 3.1 11 3.9 80 3.0 6.9 1.3 temperature and ethanol concentration [72]. Because talose, tagatose, and galactose were the C-4 epimers of mannose, fructose, and glucose, respectively, this tendency may be also applicable to the difference in solubility of talose, tagatose, and galactose. The difference in solubility increased with increasing ethanol concentration. In particular, the solubility of talose was the most sensitive to the temperature change compared with that of galactose and tagatose.

The dissolution enthalpy, ∆Hi, of each hexose was evaluated from the slope of each line drawn in Fig. 6-1 according to Eq. (6-1). The estimated ∆H values are listed in Table 6-1. The dissolution enthalpy of any hexose decreased upon increasing the ethanol concentration, and the dissolution enthalpy of talose was larger than those of galactose and tagatose. These results indicate that talose and tagatose can be separated from galactose by adding ethanol, decreasing the temperature of the reaction mixture, or by the combined use of these two means.

6.4. Conclusions The solubility of each sugar decreased upon increasing the ethanol concentration and decreasing the temperature. The solubility of D-talose and D-tagatose was higher than that of D-galactose under all conditions. The dissolution enthalpy of each hexose decreased upon increasing the ethanol concentration. The dissolution enthalpy of D-talose was higher than those of D-galactose and D-tagatose.

Chapter 7

Production of Keto-disaccharides from Aldo-disaccharides in

Subcritical Aqueous Ethanol

7.1. Introduction Chapters 2–5 showed that subcritical aqueous ethanol could be used as an excellent solvent for aldo-monosaccharide-to-keto-monosaccharide isomerizations. The kinetic analysis showed that, compared to the decomposition steps, the isomerization steps accelerated when changing the reaction solvent from subcritical water to subcritical aqueous ethanol (Chapter 2). Besides, the reaction equilibrium constant was high (ca. 3 at 200°C for glucose to fructose isomerization), which suggested a possibly high yield of keto-monosaccharides. Importantly, sucrose hydrolysis was strongly restricted in subcritical aqueous ethanol than in subcritical water (Chapter 1); this may be also applicable to other disaccharides. These results suggest that high yields of keto-disaccharides can be produced from the corresponding aldo-disaccharides using subcritical aqueous ethanol. This chapter was in attempts to efficiently produce keto-disaccharides, maltulose, palatinose, cellobiulose, lactulose, and melibiulose, from the corresponding aldo-disaccharides, maltose, isomaltose, cellobiose, lactose, and melibiose, respectively, using subcritical aqueous ethanol, and to investigate the effect of the type of glycoside linkage of aldo-disaccharides on their isomerizations.

7.2. Materials and Methods 7.2.1. Materials Maltose (α-D-Glc-(1→4)-D-Glc) monohydrate; melibiose (α-D-Gal-(1→6)-D-Glc); lactose (β-D-Gal-(1→4)-D-Glc); sucrose (α-D-Glc-(1→2)-β-D-Fru), where Glc, Gal, and Fru represent D-glucose, D-galactose, and D-fructose, respectively; D-glucose; D-fructose; D-galactose; D-tagatose; D-talose and ethanol were purchased from Wako Pure Chemical Industries (Osaka). Cellobiose (β-D-Glc-(1→4)-D-Glc), palatinose (α-D-Glc-(1→6)-D-Fru), syrupy isomaltose (α-D-Glc-(1→6)-D-Glc) and trehalose (α-D-Glc-(1→1)-D-Glc) were obtained from Hayashibara Biochemical Laboratories (Okayama, Japan). Lactose and D-mannose were purchased from Nacalai Tesque (Kyoto). Maltulose (α-D-Glc-(1→4)-D-Fru) monohydrate and

50 isomaltose for calibration were purchased from Tokyo Chemical Industry (Tokyo). Cellobiulose (β-D-Glc-(1→4)-D-Fru) and melibiulose (α-D-Gal-(1→6)-D-Fru) for calibration were prepared and purified in this study, according to the methods described later. Because the present study only focused on the isomerization of the D-enantiomers, the prefix, D-, of all the saccharides is omitted hereafter.

7.2.2. Isomerization or hydrolysis of saccharides in subcritical aqueous ethanol The isomerization treatment was the same as that described in Chapter 2. The residence time was set in the range of 50–500 s. The reaction temperature was set at 180–220°C. Ethanol concentration was adjusted from 0 to 80% (w/w) and the substrate feed concentration ranged from 0.5 to 5% (w/w) for maltose isomerization. For other saccharides, the reaction temperature was set at 200°C; ethanol concentration was adjusted at 0 (subcritical water) and 60% (w/w) and the feed concentration was 0.5% (w/w).

7.2.3. Saccharide analysis The HPLC system for saccharide analysis was the same as that described in Section 3.2.3. However, to separate the yielded glucose and galactose clearly, another column, Asahipak NH2P-50 4E (4.6 mmϕ × 250 mm length) was used for compositional analysis of the reaction mixture of lactose and melibiose, also using an 80% (v/v) aqueous acetonitrile solution as the mobile phase at a flow rate of 1.0 mL/min. The calibration curves were prepared using commercially available saccharides or saccharides prepared and purified in this study.

7.2.4. Product purification and NMR confirmation The purification procedures for the produced keto-dissacharides were the same as those described in Chapter 5. 1H and 13C NMR analyses were performed using an Ascend 400 NMR spectrometer (400 13 MHz, Bruker Japan, Osaka) with D2O as the solvent. Acetonitrile (δH = 2.06 ppm, δC ( CH3) 13 1 = 1.47 ppm, δC ( CN) = 119.68 ppm) was used as the internal standard. Further, the H and 13C NMR spectra of maltulose, palatinose, and lactulose were compared with those of the commercial samples and reported data [73–75]. Because cellobiulose and melibiulose are not commercially available, we used 1H and 13C NMR spectra of suspected cellobiulose and melibiulose for comparison. Cellobiose and melibiose have been shown to undergo isomerization to their ketoses in subcritical water or in aqueous alkaline solutions [76,77], and

Chapters 2–5 showed the promotion of aldose to ketose isomerization in subcritical aqueous ethanol. Therefore, the formation of cellobiulose and melibiulose could be assured.

7.2.5. Solubility of maltose monohydrate in aqueous ethanol The solubility of maltose monohydrate in aqueous ethanol was measured at 25°C according to the method described in Chapter 5.

7.3. Results and Discussion 7.3.1. Effect of ethanol concentration on maltose isomerization Maltose underwent slow conversion below 220°C and rapid conversion at temperatures higher than 220°C in subcritical water [33]. Further, the conversion of reducing hexose was increased (Chapter 2) and hydrolysis of sucrose was restricted (Chapter 1) by increasing ethanol concentration. Therefore, the reaction temperature of 220°C was selected to investigate the effect of ethanol concentration on hydrolysis and isomerization of maltose. Figure 7-1 shows the isomerization and hydrolysis processes of maltose in subcritical water and 60% (w/w) subcritical aqueous ethanol at 220°C. To easily calculate the mass balance during the reaction, the yield of each produced monosaccharide was defined as the molar ratio of produced monosaccharide to glucose residues of maltose. Glucose, fructose, mannose, and maltulose were produced from maltose in subcritical water and 60% (w/w) aqueous ethanol. The formation of maltulose was assured by comparing the 1H and 13C NMR spectra of the 1.0

(a) (b) 0.8

0.6

0.4

0.2 saccharide each of Yield 0 0 200 400 0 200 400 600

Residence time [s] Fig. 7-1. Isomerization and hydrolysis processes of () maltose to () maltulose, () glucose, () fructose, and () mannose in (a) subcritical water and (b) 60% (w/w) subcritical aqueous ethanol at 220°C. The feed concentration of maltose monohydrate was

0.5% (w/w). 52

(a)

(b)

5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 ppm

(c)

(d)

130 120 110 100 90 80 70 60 50 40 30 20 10 0 ppm

Fig. 7-2. 1H and 13C NMR spectra of the standard and purified sample of maltulose. (a) 1H NMR of standard sample of maltulose, (b) 1H NMR of purified sample of maltulose, (c) 13C NMR of standard sample of maltulose, and (d) 3C NMR of purified sample of maltulose.

53 suspected maltulose with those of the commercial maltulose (Fig. 7-2). However, epimaltose, a C-2 epimer of glucose in the reducing side of maltose, was not detected. It has also been reported that epimaltose is not detected in alkali-catalyzed maltose isomerization [24]. This may be because of its low yield controlled by thermodynamic equilibrium and high isomerization reactivity under subcritical aqueous conditions, since the reducing side of epimaltose is a mannose residue. Chapter 2 showed that among the mutual isomerization of glucose, mannose, and fructose, mannose was the most easily isomerized, and its formation from glucose and fructose was difficult in subcritical aqueous ethanol. In subcritical water, glucose was the most producible and its maximum yield was ca. 45% at the residence time of ca. 400 s. The yield of maltulose increased with the residence time until 100 s and then decreased, indicating that the maltulose formed had decomposed. Fructose, which can be directly obtained through hydrolysis of maltulose, as well as through isomerization of glucose and mannose, was gradually produced by prolonging the residence time; however, the yield of fructose was <15%. On the other hand, <0.5% mannose was obtained within the investigated residence time. Changing the reaction solvent from subcritical water to 60% (w/w) subcritical aqueous ethanol significantly suppressed the hydrolysis of maltose, and all the hexoses were produced in <10% yields. Chapter 1 showed the dilute effect of ethanol on water concentration during sucrose hydrolysis. It may be also suggested that ethanol affects the maltose hydrolysis in the same manner. The yield of maltulose in 60% (w/w) ethanol was higher than that in subcritical water at the same residence time. Rapid conversion of maltose occurred in 60% (w/w) subcritical aqueous ethanol until ca. 200 s, after which the conversion

Table 7-1. Effect of ethanol concentration on maltose isomerization in subcritical aqueous ethanol at 220°C. Concentration Residence Maximum Selectivity Residence Maximum Selectivity of ethanol time yield of of maltulose time for yield of of glucose [%, w/w] [s] maltulose [%] glucose glucose [%] [%] [s] [%] 0 100 14 24 400 44 45 20 200 19 26 500 30 31 40 200 19 25 500 15 16 60 100 24 39 400 9 10 80 50 22 37 300 4.4 4.6

54 decreased lesser than that in subcritical water. This could be because the addition of ethanol promoted the isomerization of maltulose to maltose and because the isomerization dominantly affected maltose conversion after 200 s (Chapter 2). The abovementioned results clearly showed that hydrolysis and isomerization of maltose were more suppressed and promoted, respectively, in subcritical aqueous ethanol than in subcritical water. In order to investigate the effect of ethanol concentration on maltose isomerization in detail, maltose isomerization at different ethanol concentrations was further examined (Table 7-1). The maximum yield of maltulose and its selectivity at that yield, ca. 24% and 39%, respectively, were higher at higher ethanol concentration. The rate constant for isomerization of maltose to maltulose, which was calculated from the initial slope of time course for maltulose formation, almost doubled when changing the solvent from water to 80% (w/w) aqueous ethanol.

7.3.2. Temperature dependence of maltose isomerization Effect of reaction temperature on the isomerization of maltose to maltulose was examined in 60% (w/w) aqueous ethanol at the feed substrate concentration of 0.5% (w/w) (Table 7-2). The maximum yield of maltulose was markedly increased by ca. 10% by increasing the reaction temperature from 180 to 200°C. However, reaction temperatures above 200°C were not suitable for producing maltulose within the residence time of 500 s, because relatively low maximum yields were obtained at such temperatures. On the other hand, at the temperatures above 200°C, the maximum yield of maltulose was achieved at short residence times. Thus, the disadvantages of increasing the reaction temperature were significant decreases in the selectivity of maltulose and total sugar content obtained at specific residence times, for example, a residence time of 500 s. These results indicated that increasingly significant decompositions of di- and monosaccharides occurred at high temperature.

Table 7-2. Effect of temperature on maltose isomerization in 60% (w/w) aqueous ethanol. Temperature Residence Maximum yield of Selectivity of Total sugar [°C] time [s] maltulose [%] maltulose [%] content [%] 180 500 18 48 85 190 500 23 40 71 200 500 28 36 60 210 300 26 37 64 220 100 24 40 72

Temperature [K] 493 473 453

10-2 ]

- [s

10-3

Rateconstant,

10-4 2.00 2.05 2.10 2.15 2.20 2.25 103/T [K-1]

Fig. 7-3. Arrhenius plot for the rate constants of maltose to maltulose isomerization in 60% (w/w) subcritical aqueous ethanol.

The rate constant of maltose-to-maltulose isomerization increased by increasing the reaction temperature, and the activation energy was calculated to be ca. 110 kJ/mol according to the Arrhenius plot (Fig. 7-3), which was similar to that of glucose-to-fructose isomerization in subcritical aqueous ethanol (Chapter 2) and that of maltose-to-maltulose isomerization in alkaline solution [24].

7.3.3. Effects of the type of glycoside linkage and constituent monosaccharides of the disaccharides on their isomerization and hydrolysis To investigate the effects of the type of glycoside linkage and constituent monosaccharides on the reaction behaviours of disaccharides, we assessed the effects of seven other disaccharides (isomaltose, cellobiose, lactose, melibiose, palatinose, trehalose, and sucrose) in subcritical water and 60% (w/w) aqueous ethanol at 200°C (Figs. 7-4–7-10). In contrast to maltose, the other tested aldo-disaccharides (isomaltose, cellobiose, lactose, and melibiose) predominantly underwent isomerization rather than hydrolysis in subcritical water and 60% (w/w) aqueous ethanol. The addition of ethanol suppressed the hydrolysis of disaccharides and promoted their conversion to the corresponding keto-disaccharides. The formation of palatinose from isomaltose (Fig. 7-4) and lactulose from lactose (Fig. 7-6) were also assured by comparing the 1H and 13C NMR spectra of the suspected products with those of the 56 commercial ones (data not shown). The formation of cellobiulose from cellobiose (Fig. 7-5) and melibiulose from melibiose (Fig. 7-7) were assured by comparing the change in the 1H and 13C NMR spectra between the substrates and the purified samples and by the fact that ketose can be produced in high yield from the corresponding aldose in subcritical aqueous ethanol (Chapters 2, 3, and 5). The yields and selectivities of cellobiulose, lactulose, and melibiulose increased in 60% (w/w) aqueous ethanol than in subcritical water. Moreover, these results also showed that the type of glycoside linkage affected the isomerization of the aldo-disaccharides, i.e. the aldo-disaccharides linked by β-glycosidic bond (cellobiose and lactose) were more easily isomerized than those linked by α-glycosidic bond (isomaltose and melibiose) in subcritical aqueous ethanol. However, for maltose and isomaltose with α glycoside linkage, it was showed that the yield of the produced keto-disaccharides was not significantly affected by the binding position of glucose on reducing side. Palatinose was used as a typical substrate to investigate the isomerization of keto-disaccharide in subcritical water and 60% (w/w) subcritical aqueous ethanol (Fig. 7-8). Monosaccharides were the dominant products in subcritical water. Changing the reaction solvent to 60% (w/w) ethanol improved the conversion of palatinose, yield of isomaltose, and suppressed hydrolysis. The maximum yield of isomaltose was ca. 10%. Compared with isomaltose, palatinose decreased rapidly at short residence times. However, the yield of isomaltose formed from palatinose was lower than that of palatinose formed from isomaltose under the same reaction conditions, indicating that increasing difficulty in the isomerization of palatinose to isomaltose. Chapter 2 showed that ketohexose had a low ability to isomerize than aldohexose. These results show that reducing keto-disaccharides mainly underwent decomposition rather than isomerization in subcritical aqueous ethanol. In order to compare the isomerization of aldo-disaccharides with that of monosaccharides, isomerizations of glucose and galactose were also investigated in 60% (w/w) ethanol at 200°C (data not shown). Fructose and tagatose were obtained from glucose and galactose at the maximum yields of 29% and 26%, respectively. No significant difference in the initial reaction rate was observed between reducing aldo-disaccharides and monosaccharides. This observation differed from that in alkaline solution. Previous studies showed that reducing disaccharides isomerized considerably faster than did monosaccharides in alkaline solution [29,60]. We believe that the reason for no difference in the initial reaction rate between reducing aldo-disaccharides and monosaccharides in this study can be attributed to the change in temperature dependence of the isomerization rate constants under subcritical conditions.

1.0

0.8 (a) (b)

0.6

0.4

0.2 Yield of each saccharide eachof Yield

0 0 200 400 0 200 400 600 Residence time [s]

Fig. 7-4. Isomerization and hydrolysis processes of () isomaltose to () palatinose, () epiisomaltose, () glucose, () fructose, and () mannose in (a) subcritical water and (b) 60% (w/w) subcritical aqueous ethanol at 200°C. The feed concentration of isomaltose was 0.5% (w/w).

1.0

0.8 (a) (b)

0.6

0.4

0.2 Yield of each saccharide eachof Yield

0 0 200 400 0 200 400 600 Residence time [s]

Fig. 7-5. Isomerization and hydrolysis processes of () cellobiose to () cellobiulose, () epicellobiose, () glucose, () fructose, and () mannose in (a) subcritical water and (b) 60% (w/w) subcritical aqueous ethanol at 200°C. The feed concentration of cellobiose was 0.5% (w/w).

1.0

0.8 (a) (b)

0.6

0.4

0.2 Yield of each saccharideeach of Yield

0 0 200 400 0 200 400 600 Residence time [s]

Fig. 7-6. Isomerization and hydrolysis processes of () lactose to () lactulose, () epilactose, () glucose, () fructose, () mannose, () galactose, () tagatose, and (►) talose in (a) subcritical water and (b) 60% (w/w) subcritical aqueous ethanol at 200°C. The feed concentration of lactose was 0.5% (w/w).

1.0

0.8 (a) (b)

0.6

0.4

0.2 Yield of each saccharide eachof Yield

0 0 200 400 0 200 400 600 Residence time [s]

Fig. 7-7. Isomerization and hydrolysis processes of () melibiose to () melibiulose, () epimelibiose, () glucose, () fructose, () mannose, () galactose, () tagatose, and (►) talose in (a) subcritical water and (b) 60% (w/w) subcritical aqueous ethanol at 200°C. The feed concentration of melibiose was 0.5% (w/w).

1.0

0.8 (a) (b)

0.6

0.4

0.2 Yield of each saccharideeach of Yield

0 0 200 400 0 200 400 600 Residence time [s]

Fig. 7-8. Isomerization and hydrolysis processes of () palatinose to () isomaltose, () glucose, () fructose, and () mannose in (a) subcritical water and (b) 60% (w/w) subcritical aqueous ethanol at 200°C. The feed concentration of palatinose was 0.5% (w/w).

1.0

0.8

0.6

0.4

0.2 Yield of each saccharideeach of Yield

0 0 200 400 600

Residence time [s]

Fig. 7-9. Hydrolysis processes of (,) trehalose to (,) glucose, in subcritical water (open symbols) and 60% (w/w) subcritical aqueous ethanol (closed symbols) at 200°C. The feed concentration of trehalose was 0.5% (w/w).

1.0

0.8

0.6

0.4

0.2

saccharideeach of Yield 0 0 200 400 600

Residence time [s]

Fig. 7-10. Hydrolysis processes of () sucrose to () glucose, () fructose, and () mannose in 60% (w/w) subcritical aqueous ethanol at 200°C. The feed concentration of sucrose was 0.5% (w/w).

Trehalose and sucrose, which are non-reducing disaccharides, only underwent hydrolysis and isomerization of their constituent monosaccharides. However, trehalose was hydrolyzed slowly with little effect of the ethanol concentration (Figs. 7-9 and 7-10). Treatment of trehalose in subcritical water and in 60% (w/w) ethanol for 500 s achieved a conversion of ca. 5%. On the other hand, in autocatalytic mode, sucrose was easily hydrolyzed to fructose and glucose as the main products, which further slightly isomerized to mannose (Chapters 1, 2, and 3). These results definitely showed that only reducing disaccharides can be isomerized in subcritical water and subcritical aqueous ethanol.

7.3.4. Productivity of maltulose at different feed maltose concentrations The results demonstrated the reliability of the process to produce rare keto-disaccharides using subcritical aqueous ethanol. Maltose to maltulose isomerization was typically selected to examine the effect of feed substrate concentration on the isomerization in 60% (w/w) aqueous ethanol at 200°C. Chapter 5 showed that high feed concentration of galactose improved productivity of the desired tagatose. This would be expected to be true for maltulose production. The solubility of maltose monohydrate in subcritical aqueous ethanol was

solvent]

- g

monohydratein /100

altose 20

ethanol [ aqueous

0 Solubility of m Solubility 0 20 40 60 80 100 Concentration of ethanol [wt%] Fig. 7-11. Solubility of maltose monohydrate in aqueous ethanol at 25°C. measured at 25°C (Fig. 7-11). The solubility of maltose monohydrate significantly decreased by the addition of ethanol from ca. 60% (w/w) in water to ca. 1% (w/w) in 80% (w/w) aqueous ethanol. The solubility of maltose monohydrate was ca. 5.5% (w/w) in 60% (w/w) aqueous ethanol at 25°C, which was lower than that of galactose (Chapter 5). To avoid the precipitation of maltose in the tubular reactor, we used a feed maltose concentration ranging from 0.5 to 5% (w/w). The maximum yields of maltulose were obtained at the residence time of 500 s within the tested residence time range at any feed maltose monohydrate concentrations (Table 7-3). The maximum yield of maltulose and corresponding conversion of maltose decreased from 29% and 79% at 0.5% (w/w) maltose monohydrate to 18% and 51% at 5% (w/w), respectively. However, the maximum productivity of maltulose increased almost sevenfold to ca. 41 g/(h·kg-solution). Besides, the selectivity of maltulose was little affected by the change in feed substrate concentration, and the total sugar content was >70% at a feed substrate concentration of 5% (w/w). These results suggest that subcritical aqueous ethanol is a suitable solvent for effectively producing the rare keto-disaccharides from their corresponding common aldo-saccharides.

Table 7-3. Isomerization of maltose in 60% (v/v) aqueous ethanol at 200°C at different feed concentration. Feed Residence Maltose Maximum Selectivity Total Productivity concentration of time conversion yield of of maltulose sugar of maltulose maltose [s] [%] maltulose [%] content [g/(h·kg- monohydrate [%] [%] solution)] [%, w/w] 0.5 500 79 29 36 59 6.6 1 500 70 27 39 67 16 2 500 67 24 35 65 29 5 500 51 18 35 74 41

7.4. Conclusions It was demonstrated that maltulose, palatinose, cellobiulose, lactulose, and melibiulose can be produced by isomerization of the corresponding aldo-disaccharides in subcritical aqueous ethanol. The hydrolytic and isomerization reactions of disaccharides were suppressed and promoted, respectively, by increasing ethanol concentration, which lead to high maximum yields of keto-disaccharides. The type of glycoside linkage and constituent monosaccharides affected the isomerization of disaccharides. Higher yields of keto-disaccharides linked by β-glycosidic bond than by α-glycosidic bond from the corresponding aldo-disaccharides were produced. On the other hand, the keto-disaccharide, palatinose, mainly underwent decomposition rather than isomerization.

Concluding Remarks

Chapter 1 The kinetics of sucrose hydrolysis was investigated in a water-ethanol mixture with ethanol concentrations of 0–80% (v/v) under subcritical conditions in the 160‒190°C range. Sucrose underwent autocatalytic hydrolysis in the subcritical mixtures. The rate of sucrose hydrolysis to glucose and fructose was reduced with increasing ethanol concentration, and ethanol showed a dilution effect on the conversion. The temperature dependence of the reaction rate constant for sucrose hydrolysis obeyed the Arrhenius equation. The yielded fructose and glucose underwent further decomposition, and the yield of fructose was much lower than that of glucose when the ethanol concentration increased.

Chapter 2 Fructose, glucose, and mannose were treated with subcritical aqueous ethanol for ethanol concentrations ranging from 0 to 80% (v/v) at 180–200°C. The aldose-ketose isomerization was more favorable than ketose-aldose isomerization and glucose-mannose epimerization. The isomerization of the monosaccharides was promoted by the addition of ethanol. In particular, mannose was isomerized most easily to fructose in subcritical aqueous ethanol. The apparent equilibrium constants for the isomerizations of mannose to fructose, Keq,MF, and glucose to fructose, Keq,GF, were independent of ethanol concentration and increased with increasing temperature. Moreover, the Keq,MF value was much larger than the Keq,GF value.

The enthalpies for the isomerization of mannose to fructose, ΔHMF, and glucose to fructose,

ΔHGF, were estimated to be 18 and 24 kJ/mol, respectively, according to van’t Hoff equation. Subcritical aqueous ethanol can be used to produce fructose from glucose and mannose efficiently.

Chapter 3 The influence of water-miscible alcohols (methanol, 1-propanol, 2-propanol, and t-butyl alcohol) on the isomerization of glucose to fructose and mannose was investigated under subcritical aqueous conditions (180–200°C). Primary and secondary alcohols promoted the conversion and isomerization of glucose to afford fructose and mannose with high and low selectivity, respectively. On the other hand, the decomposition (side-reaction) of glucose was suppressed in the presence of the primary and secondary alcohols compared with that in 64 subcritical water. The yield of fructose increased with increasing concentration of the primary and secondary alcohols, and the species of the primary and secondary alcohols tested had little effect on the isomerization behavior of glucose. In contrast, the isomerization of glucose was suppressed in subcritical aqueous t-butyl alcohol. Both the conversion of glucose and the yield of fructose decreased with increasing concentration of t-butyl alcohol. In addition, mannose was not detected in reactions using subcritical aqueous t-butyl alcohol.

Chapter 4 The kinetic effect of subcritical aqueous alcohols (methanol, ethanol, 1-propanol, 2-propanol, and t-butyl alcohol) on the isomerization of hexoses (mannose, glucose, and fructose) were examined at 180°C at the alcohol concentrations ranging from 0 to 80% (v/v). The results showed that increase in the concentration of primary and secondary alcohols markedly promoted the aldose-to-ketose isomerization; however, the ketose was scarcely isomerized to aldose and predominantly decomposed. Further, the kinetic analysis revealed that the rate constants of isomerization and decomposition strongly depended on the relative dielectric constant of the aqueous alcohols. The rate constants for isomerization were more sensitive to the change in the relative dielectric constant than the decomposition rate constants. On the other hand, the isomerization reactions of mannose and fructose were suppressed in subcritical aqueous t-butyl alcohol despite of the low relative dielectric constant of the solution.

Chapter 5 A new isomerization reaction was developed to synthesize rare ketoses. D-Tagatose, D-xylulose, and D-ribulose were obtained in the maximum yields of 24%, 38%, and 40%, respectively, from the corresponding aldoses, D-galactose, D-xylose, and D-ribose, by treating the aldoses with 80% (v/v) subcritical aqueous ethanol at 180C. The maximum productivity of D-tagatose was ca. 80 g/(L·h). Increasing the concentration of ethanol significantly increased the isomerization of D-galactose. Variation in the reaction temperature did not significantly affect the production of D-tagatose from D-galactose. Subcritical aqueous ethanol converted both 2,3-threo and 2,3-erythro aldoses to the corresponding C-2 ketoses in high yields. Thus, the treatment of common aldoses in subcritical aqueous ethanol can be regarded as a new method to synthesize the corresponding rare sugars.

Chapter 6 In order to reduce the load on the separation process, decreasing the temperature or increasing the ethanol concentration for precipitating desirable or undesirable components before chromatographic separation is considered as a possible solution the first step in the purification process of rare sugars. Therefore, the solubility of D-galactose, D-tagatose, and D-talose in aqueous ethanol (20–80%, w/w) was measured in the temperature range of –30°C to 20°C. The solubility of each saccharide decreased upon increasing the ethanol concentration and decreasing the temperature. The solubility of D-talose and D-tagatose was higher than that of D-galactose under all conditions. The temperature dependence of the solubility of each hexose can be expressed by the van’t Hoff equation. The dissolution enthalpy of each hexose decreased upon increasing the ethanol concentration. The dissolution enthalpy of D-talose was higher than those of D-galactose and D-tagatose.

Chapter 7 Isomerization of disaccharides (maltose, isomaltose, cellobiose, lactose, melibiose, palatinose, sucrose, and trehalose) was investigated in subcritical aqueous ethanol. A marked increase in the isomerization of aldo-disaccharides to keto-disaccharides was noted and their hydrolytic reactions were suppressed with increasing ethanol concentration. Under any study condition, the maximum yield of keto-disaccharides produced from aldo-disaccharides linked by β-glycosidic bond was higher than that produced from aldo-disaccharides linked by α-glycosidic bond. Palatinose, a keto-disaccharide, mainly underwent decomposition rather than isomerization in subcritical water and subcritical aqueous ethanol. No isomerization was noted for the non-reducing disaccharides trehalose and sucrose. The rate constant of maltose to maltulose isomerization almost doubled by changing the solvent from subcritical water to 80% (w/w) aqueous ethanol at 220°C. Increased maltose monohydrate concentration in feed decreased the conversion of maltose and the maximum yield of maltulose, but increased the productivity of maltulose. The maximum productivity of maltulose was ca. 41 g/(h·kg-solution).

References

[1] Disney, M.D. and Seeberger, P.H. (2004). The use of carbohydrate microarrays to study carbohydrate-cell interactions and to detect pathogens. Chem. Biol., 11, 1701–1707. [2] Nyholm, P.G. and Pascher, I. (1993). Steric presentation and recognition of the saccharide chains of glycolipids at the cell surface: favoured conformations of the saccharide-lipid linkage calculated using molecular mechanics (MM3). Int. J. Biol. Macromol., 15, 43–51. [3] Sears, P. and Wong, C.-H. (1998). Enzyme action in glycoprotein synthesis. Cell. Mol. Life Sci. CMLS, 54, 223–252. [4] Beerens, K., Desmet, T., and Eoetaert, W. (2012). Enzymes for the biocatalytic production of rare sugars. J. Ind. Microbiol. Biotechnol., 39, 823–834. [5] Levin, G.V., Zehner, L.R., Saunders, J.P., and Beadle, J.R. (1995). Sugar substitutes: their energy values, bulk characteristics, and potential health benefits. Am. J. Clin. Nutr., 62, 1161s–1168s. [6] Lu, Y., Levin, G.V., and Donner, T.W. (2008). Tagatose, a new antidiabetic and obesity control drug. Diabetes Obes. Metab., 10, 109–134. [7] Mizota, T., Tamura, Y., Tomita, M., and Okonogi, S. (1987). Lactulose as a sugar with physiological significance. Bull. Int. Dairy Fed., 212, 69–76. [8] Schuster-Wolff-Bühring, R., Fischer, L., and Hinrichs, J. (2010). Production and physiological action of the disaccharide lactulose. Int. Dairy J., 20, 731–741. [9] Smits, B.J., Whitehead, A.M., and Prescott, P. (1990). Lactulose in the treatment of symptomatic diverticular disease: a comparative study with high-fibre diet. Br. J. Clin. Pract., 44, 314–318. [10] Rossi-Fanelli, F., Riggio, O., Cangiano, C., Cascino, A., De Conciliss, D., Merli, M., Stortoni, M., and Giunchi, G. (1982). Branched-chain amino acids vs lactulose in the treatment of hepatic coma. Dig. Dis. Sci., 27, 929–935. [11] Marugg, J.E., Tromp, M., Kuyl-Yeheshiely, E., van der Marel, G.A., and van Boom, J.H. (1986). A convenient and general approach to the synthesis of properly protected D-nucleoside-3’-hydrogenphosphonates via phosphite intermediates. Tetrahedron Lett., 27, 2661−2664. [12] Gunther, W.R., Wang, Y.R., Ji, Y.W., Michaelis, V.K., Hunt, S.T., and Griffin, R.G. (2012). Sn-Beta zeolites with borate salts catalyse the epimerization of carbohydrates via an intramolecular carbon shift. Nat. Commun., 3, 1–8. 67

[13] Izumori, K. (2006). Izumoring: a strategy for bioproduction of all hexoses. J. Biotechnol., 124, 717–722. [14] Angyal, S.J. (1997). A short note on the epimerization of aldoses. Carbohydr. Res., 300, 279–281. [15] Yanagihara, R., Soeda, K., Shiina, S., Osanai, S., and Yoshikawa, S. (1993). C-2 epimerization of aldoses by calcium ion in basic solutions. A simple system to transform

D-glucose and D-xylose into D-mannose and D-lyxose. Bull. Chem. Soc. Jpn., 66, 2268– 2272. [16] Shukla, R., Verykios, X.E., and Mutharasan, R. (1985). Isomerization and hydrolysis reactions of important disaccharides over inorganic heterogeneous catalysts. Carbohydr. Res., 143, 97–106. [17] Tanase, T., Takei, T., Hidai, M., and Yano, S. (2001). Substrate-dependent chemoselective aldose-aldose and aldose-ketose isomerizations of carbohydrates promoted by a combination of calcium ion and monoamines. Carbohydr. Res., 333, 303–312. [18] Choudhary, V., Caratzoulas, S., and Vlachos, D.G. (2013). Insights into the isomerization of xylose to xylulose and lyxose by a Lewis acid catalyst. Carbohydr. Res., 368, 89–95. [19] Moliner, M., Román-Leshkov, Y., and Davis, M.E. (2010). Tin-containing zeolites are highly active catalysts for the isomerization of glucose in water. Proc. Natl. Acad. Sci. USA, 107, 6164–6168. [20] Bermejo-Deval, R., Gounder, R., and Davis, M.E. (2012). Framework and extraframework tin sites in zeolite beta react glucose differently. ACS Catal., 2, 2705– 2713. [21] Khadem, H.S.E., Ennifar, S., and Isbell, H.S. (1989). Evidence of stable hydrogen-bonded ions during isomerization of hexoses in alkali. Carbohydr. Res., 185, 51–59. [22] Mendicino, J.F. (1960). Effect of borate on the alkali-catalyzed isomerization of sugars. J. Am. Chem. Soc., 82, 4975–4979. [23] Tipson, R.S. and Brady Jr, R.F. (1969). Synthesis of the two D-2-pentuloses new

derivatives of D-erythro-pentulose. Carbohydr. Res., 10, 549–563. [24] Vuorinen, T. (1982). Kinetics of alkaline degradation of maltose in ethanol-water solutions. Carbohydr. Res., 108, 213–219.

[25] Lindström, U.M. (2002). Stereoselective organic reactions in water. Chem. Rev., 102, 2751–2772. [26] Lü, X. and Saka, S. (2012). New insights on monosaccharides’ isomerization, dehydration and fragmentation in hot-compressed water. J. Supercrit. Fluids, 61, 146– 156. [27] Usuki, C., Kimura, Y., and Adachi, S. (2007). Isomerization of hexoses in subcritical water. Food Sci. Technol. Res., 13, 205–209. [28] Palazzi, E. and Converti, A. (1999). Generalized linearization of kinetics of glucose isomerization to fructose by immobilized glucose isomerase. Biotechnol. Bioeng., 63, 273–284 [29] Vourinen, T. and Sjöström, E. (1982). Kinetics of alkali-catalyzed isomerization of D-glucose and D-fructose in ethanol-water solutions. Carbohydr. Res., 108, 23–29. [30] Visuri, K. and Klibanov, A.M. (1987). Enzymatic production of high fructose corn syrup (HFCS) containing 55% fructose in aqueous ethanol. Biotechnol. Bioeng., 30, 917–920. [31] Haghighat Khajavi, S., Kimura, Y., Oomori, T., Matsuno, R., and Adachi, S. (2005). Kinetics on sucrose decomposition in subcritical water. LWT-Food Sci. Technol., 38, 297–302. [32] Oomori, T., Haghighat Khajavi, S., Kimura, Y., Adachi, S., and Matsuno, R. (2004). Hydrolysis of disaccharides containing glucose residue in subcritical water. Biochem. Eng. J., 18, 143–147. [33] Haghighat Khajavi, S., Kimura, Y., Oomori, T., Matsuno, R., and Adachi, S. (2004). Decomposition kinetics of maltose in subcritical water. Biosci. Biotechnol. Biochem., 68, 91–95. [34] Quintas, M., Brandao, T.R.S., and Silva, C.L.M. (2007). Modelling autocatalytic behavior of a food model system-sucrose thermal degradation at high concentrations. J. Food Eng., 78, 537–545. [35] Bazaev, A.R., Abdulagatov, I.M., Bazaev, E.A., and Abdurashidova, A. (2007). PVT measurements for pure ethanol in the near-critical and supercritical regions. Intl. J. Thermophys., 28, 194–219. [36] Sommer, D., Kleinrahm, R., Span, R., and Wagner, W. (2011). Measurement and correlation of the (p, ρ, T) relation of liquid cyclohexane, toluene, and ethanol in the temperature range from 233.15 K to 473.15 K at pressures up to 30 MPa for use as density reference liquids. J. Chem. Thermodyn., 43, 117–132.

[37] Rhim, J.W., Nunes, R.V., Jones, V.A., and Swartzel, K.R. (1989). Appearance of a kinetic compensation effect in the acid-catalyzed hydrolysis of disaccharides. J. Food Sci., 54, 222–223. [38] Haghighat Khajavi, S., Ota, S., Nakazawa, R., Kimura, Y., and Adachi, S. (2006). Hydrolysis kinetics of trisaccharides consisting of glucose galactose, and fructose residues in subcritical water. Biotechnol. Prog., 22, 1321–1326. [39] Srokol, Z., Bouche, A., Estrik, A., Strik, R.C.J., Thomas, M., and Peters, J.A. (2004). Hydrothermal upgrading of biomass to biofuel; studies on some monosaccharide model compounds. Carbohydr. Res., 339, 1717–1726. [40] Boisen, A., Christensen, T.B., Fu, W., Gorbanev, Y.Y., Hansen, T.S., Jensen, J.S., Klitgaard, S.K., Pedersen, S., Riisager, A., Tåhlberg, T.S., and Woodley, J.M. (2009). Process integration for the conversion of glucose to 2,5-furandicarboxylic acid. Chem. Eng. Res. Des., 87, 1318–1327. [41] Durham, D.G., Hung, C.T., and Taylor, R.B. (1982). Identification of some acids

produced during autoclaving of D-glucose solutions using HPLC. Int. J. Pharmaceut., 11, 31–40. [42] Girisuta, B., Janssen, L.P.B.M., and Heeres, H.J. (2006). Green chemicals: a kinetic study on the conversion of glucose to levulinic acid. Chem. Eng. Res. Des., 84, 339–349. [43] Aida, T.M., Tajima, K., Watanabe, M., Yuki, S., Kuroda, K., Nonaka, T., Hattori, H., Lee, R., Smith, J., and Arai, K. (2007). Reactions of D-fructose in water at temperatures up to 400°C and pressures up to 100 MPa. J. Supercrit. Fluids, 42, 110–119. [44] Angyal, S.J. (1969). The composition and conformation of sugars in solution. Angew. Chem. Int. Ed., 8, 157–166. [45] Saravanamurugan, S., Paniagua, M., Melero, J.A., and Riisager, A. (2013). Efficient isomerization of glucose to fructose over zeolites in consecutive reactions in alcohol and aqueous media. J. Am. Chem. Soc., 135, 5246–5249. [46] Haghighat Khajavi, S., Kimura, Y., Oomori, T., Matsuno, R., and Adachi, S. (2005). Degradation kinetics of monosaccharides in subcritical water. J. Food Eng., 68, 309– 313. [47] Moody, W. and Richards, G.N. (1981). Thermolysis of sucrose in the presence of

alcohols. A novel method for the synthesis of D-fructofuranosides. Carbohydr. Res., 97, 247–255. [48] Asaoka, H. (1985). Isomerization of D-glucose by disodium pentasilicate

(Na2Si5O11·xCH3·yH2O) in methanol-water solutions. Carbohydr. Res., 137, 99–109. [49] Bazaev, A.R., Abdulagatov, I.M., Bazaev, E.A., Abdurashidova, A.A., and Ramazanova, A.E. (2007). PVT measurements for pure methanol in the near-critical and supercritical regions. J. Supercrit. Fluids, 41, 217–226. [50] Abdulagatov, I.M., Safarov, J.T., Aliyev, F.Sh., Talibov, M.A., Shahverdiyev, A.N., and Hassel, E.P. (2008). Experimental densities and derived thermodynamic properties of liquid propan-1-ol at temperatures from 298 to 423 K and at pressures up to 40 MPa. Fluid Phase Equilib., 268, 21–33. [51] Alaoui, F.E.M., Montero, E.A., Bazile, J.-P., Aguilar, F., and Boned, C. (2012). Liquid density of oxygenated additive 2-propanol at pressures up to 140 MPa and from 293.15 K to 403.15 K. J. Chem. Thermodyn., 54, 358–365. [52] Topper, Y.J. and Stetten, D.Jr. (1951). The alkali-catalyzed conversion of glucose into fructose and mannose. J. Biol. Chem., 189, 191–202. [53] Garrett, E.R. and Young, J.F. (1970). Alkaline transformation among glucose, fructose, and mannose. J. Org. Chem., 35, 3502–3509. [54] De Wit, G., Kieboom, A.P.G., and van Bekkum, H. (1997). Enolisation and isomerization of monosaccharides in aqueous, alkaline solution. Carbohydr. Res., 74, 157–175. [55] Harris, D.W. and Feather, M.S. (1973). Evidence for C-2→C-1 intramolecular hydrogen-transfer during the acid-catalyzed isomerization of D-glucose and D-fructose ag]. Carbohydr. Res., 30, 359–365. [56] Román-Leshkov, Y., Moliner, M., Labinger, J.A., and Davis, M.E. (2010). Mechanism of glucose isomerization using a solid Lewis acid catalyst in water. Angew. Chem. Int. Ed., 49, 8954–8957. [57] Yoshida, K., Kitajo, A., and Yamaguchi, T. (2006). 17O NMR relaxation study of dynamics of water molecules in aqueous mixtures of methanol, ethanol, and 1-propanol over a temperature range of 283–403 K. J. Mol. Liq., 125, 158–163. [58] Rived, F., Rosés, M., and Bosch, E. (1998). Dissociation constants of neutral and charged acids in methyl alcohol. The acid strength resolution. Anal. Chim. Acta., 374, 309–324. [59] EI Khaden, H.S., Ennifar, S., and Isbell, H.S. (1987). Contribution of the reaction pathways involved in the isomerization of monosaccharides by alkali. Carbohydr. Res., 169, 13–21. [60] Isbell, H.S., Frush, H.L., Wade, C.W.R., and Hunter, C.E. (1969). Transformations of sugars in alkaline solutions. Carbohydr. Res., 9, 163–175.

[61] Ha, S., Gao, J., Tidor, B., Brady, J.W., and Karplus, M. (1991). Solvent effect on the anomeric equilibrium in D-glucose: a free energy simulation analysis. J. Am. Chem. Soc., 113, 1553–1557. [62] Hynes, J.T. (1986). Chemical reaction rates and solvent friction. J. Statist. Phys., 42, 149–168. [63] Akerlof, G. (1932). Dielectric constants of some organic solvent-water mixtures at various temperatures. J. Am. Chem. Soc., 54, 4125–4139. [64] Archer, D.G. and Wang, P. (1990). The dielectric constants of water and Debye-Hückel limiting law slopes. J. Phys. Chem. Ref. Data, 19, 371–411. [65] Vourinen, T. and Serianni, A.S. (1990). Synthesis of D-erythro-2-pentulose and D-threo-2-pentulose and analysis of the 13C- and 1H-n.m.r. spectra of the 1-13C- and 2-13C-substituted sugars. Carbohydr. Res., 209, 13–31. [66] Yamauchi, T., Fukushima, K., Yanagihara, R., Osanai, S., and Yoshikawa, S. (1990). Epimerization and isomerization of various monosaccharides using metal-diamine systems. Carbohydr. Res., 204, 233–239. [67] Jørgensen, F., Hansen, O.C., and Stougaard, P. (2004). Enzymatic conversion of D-galactose to D-tagatose: heterologous expression and characterisation of a thermostable L-arabinose isomerase from Thermoanaerobacter mathranii. Appl. Microbiol. Biotechnol., 64, 816–822. [68] Lim, B.C., Kim, H.J., and Oh, D.K. (2007). A high production of D-tagatose by the addition of boric acid. Biotechnol. Progr., 23, 824–828. [69] Alves, L.A., Almeida e Silva, J.B., and Giulietti, M. (2007). Solubility of D-glucose in water and ethanol/water mixtures. J. Chem. Eng. Data, 52, 2166–2170. [70] Flood, A.E., Addai-Mensah, J., Johns, M.R., and White, E.T. (1996). Refractive index, viscosity, density, and solubility in the system fructose + ethanol + water at 30, 40, and 50°C. J. Chem. Eng. Data, 41, 418–421. [71] Silva, A.T.C.R., Martinez, K.C.L., Brito, A.B.N., and Giulietti, M. (2010). Separation of glucose and fructose by freezing crystallization. Cryst. Res. Technol., 45, 1032–1034. [72] Hudson, C.S. and Yanovsky, E. (1917). Indirect measurements of the rotatory powers of some alpha and beta forms of the sugars by means of solubility experiments. J. Am. Chem. Soc., 39, 1013–1038. [73] Pfeffer, P.E. and Hicks, K.B. (1982). Characterization of keto disaccharides in solution by deuterium-induced, differential isotope-shift 13C-NMR spectroscopy. Carbohydr. Res.,

102, 11–22. [74] Low, N.H., Brisbane, T., Bigam, G., and Sporns, P. (1988). Carbon-13 nuclear magnetic resonance for the qualitative and quantitative analysis of structurally similar disaccharides. J. Agric. Food Chem., 36, 953–957. [75] Kimura, H., Nakahara, M., and Matubayasi, N. (2012). Noncatalytic hydrothermal elimination of the terminal D-glucose unit from malto- and cello-oligosaccharides

through transformation to D-fructose. J. Phys. Chem. A, 116, 10039–10049. [76] Yu, Y., Shafie, Z.M., and Wu, H. (2013). Cellobiose decomposition in hot-compressed water: importance of isomerization reactions. Ind. Eng. Chem. Res., 52, 17006–17014. [77] Corbett, W.M. and Kenner, J. (1954). The degradation of carbohydrates by alkali. Part VIII. Melibiose. J. Chem. Soc., 3281–3283.

Acknowledgements

Study as a Ph.D. student in Kyoto University was a dream for me, and has become a great part of my life. Because it was a splendid, fruitful as well as challenging experience. There are a number of people without whom this thesis could not be finished and to whom I must be sincerely indebted. First of all, I should express my sincere gratitude to my supervisor Prof. Shuji Adachi, for he gave me an opportunity to enter the Laboratory of Bioengineering, guided me toward the world of research, and sharped me to be a researcher with professional views, cheerful enthusiasm, and endless patience. I am grateful to Associate Professor Kyuya Nagakawa for his kind encouragement and valuable advices. My thanks must go to Assistant Professor Takashi Kobayashi who supported me many and many assistance, valuable advices, and discussions with great patience. I also thank Miss Rumiko Kamiya for her meticulous care and advices in affairs and my personal life to help me overcome various difficulties. I thank all the members of the laboratory for accompanying me with help and gaiety: Dr. Tai-Ying Chiou, Dr. Takenobu Ogawa, Dr. Boonnakhom Tangkhavanich, Dr. Teeraya Jarunglumlert, Intira Komyart, Chiayapat Incharoensakdi, Yukie Ohishi, Naho Mizuno, Masashi Yoshino, Yayoi Miyagawa, Hironori Nagamizu, Takao Roppongi, Souma Fukuzawa, Kazutaka Katsuki, Risako Yamamoto, Keisuke Shintani, Ango Tamura, Shinri Tamiya, and all the other members of this laboratory. I want to thank my Chinese friends who encouraged me a lot. I gratefully acknowledge the Monbukagakusho Scholarship from the Japanese government and the Tojuro Iijima Foundation for Food Science and Technology. Last but not the least; I am very grateful to all members of my family, they always gave me support and encouragement every moment no matter where I am and what I do.

List of Publications

1. Gao, D., Kobayashi, T., and Adachi, S. (2014). Kinetics of sucrose hydrolysis in a subcritical water-ethanol mixture. J. Appl. Glycosci., 61, 9–13.

2. Gao, D.-M., Kobayashi, T., and Adachi, S. (2015). Kinetic analysis for the isomerization of glucose, fructose, and mannose in subcritical aqueous ethanol. Biosci. Biotechnol. Biochem., 79, 1005–1010.

3. Gao, D.-M., Kobayashi, T., and Adachi, S. (2015). Promotion or suppression of glucose isomerization in subcritical aqueous straight- and branched-chain alcohols. Biosci. Biotechnol. Biochem., 79, 470–474.

4. Gao, D.-M., Kobayashi, T., and Adachi, S. Kinetic effect of alcohols on hexose isomerization under subcritical aqueous conditions. Chem. Eng. Res. Des., in press.

5. Gao, D.-M., Kobayashi, T., and Adachi, S. (2015). Production of rare sugars from common sugars in subcritical aqueous ethanol. Food Chem., 175, 465–470.

6. Gao, D.-M., Kobayashi, T., and Adachi, S. (2015). Solubility of D-galactose, D-talose, and

D-tagatose in aqueous ethanol at low temperature. Food Sci. Technol. Res., 21, 801–803.

7. Gao, D.-M., Kobayashi, T., and Adachi, S. Production of keto-disaccharides from aldo-disaccharides in subcritical aqueous ethanol. Biosci. Biotechnol. Biochem., in press.