Article pubs.acs.org/IECR
Cellobiose Decomposition in Hot-Compressed Water: Importance of Isomerization Reactions Yun Yu, Zainun Mohd Shafie, and Hongwei Wu* School of Chemical and Petroleum Engineering, Curtin University, GPO Box U1987, Perth Western Australia 6845, Australia ABSTRACT: This paper reports an investigation on the fundamental reaction mechanism of cellobiose decomposition in hotcompressed water (HCW) using a continuous reactor system at 225−275 °C. The importance of isomerization reactions to form two cellobiose isomers (i.e., cellobiulose and glucosyl-mannose) as the primary reaction products is clearly demonstrated under the reaction conditions, using a high-performance anion exchange chromatography with pulsed amperometric detection and mass spectrometry (HPAEC-PAD-MS). The results also confirm another two primary reactions take place during cellobiose decomposition in HCW: retro-aldol condensation reaction to produce glucosyl-erythrose (GE) and glycolaldehyde, and hydrolysis reaction to produce glucose. The data show that isomerization and retro-aldol condensation are the dominant primary reactions while hydrolysis of cellobiose is only a minor primary reaction (accounting for ∼10−20% of cellobiose decomposition depending on reaction temperature). The results indicate that the reaction solution becomes acidic at the early stage of cellobiose decomposition, most likely due to the formation of organic acids, resulting in the subsequent reactions exhibiting more characteristics of acid-catalyzed reactions. The results further suggest that the formed acidic condition has little catalytic effects on the primary reactions of cellobiose decomposition, but is effective in catalyzing secondary reactions of various reaction intermediates such as hydrolysis and dehydration reactions to form glucose and 5-HMF, respectively.
1. INTRODUCTION Hot-compressed water (HCW) processing is an attractive approach for synthesizing and producing green biofuels and biochemicals from renewable lignocellulosic biomass.1−3 As a result of its high ion product and low dielectric constant, HCW is both an excellent reactant and an ideal solvent for biomass degradation reactions.4,5 In particular, HCW technology is suitable for the processing of wet biomass (e.g., algal biomass6), as it eliminates the energy-intensive process for inherent moisture evaporation. A thorough understanding on biomass decomposition in HCW is highly desired and essential to facilitating the development of advanced technologies for related applications. Cellulose as a main component in lignocellulosic biomass has been widely used as a model compound for studying biomass decomposition in HCW.7−11 It was reported that cellulose decomposition in HCW proceeds via a series of complicated reactions to produce glucose and its oligomers as intermediates,7,9 then to gas, oil, and char as final products.8 A series of recent studies 12−16 by this group have reported the fundamental mechanism for the primary reactions taking place on the surface of cellulose particles, with the minimization of secondary reactions during cellulose decomposition in HCW. It was revealed that the primary products of cellulose hydrolysis in HCW include various sugar oligomers and their derivatives with a wide range of degrees of polymerization (DPs).13 These primary products subsequently dissolve into HCW and experience secondary reactions hence decompose into other products.13 However, the secondary reactions of the primary liquid products from cellulose hydrolysis in HCW are still largely unknown. In fact, apart from that of the monomer (glucose), the decomposition mechanism of glucose oligomers with various DPs in HCW is still poorly understood. © 2013 American Chemical Society
Cellobiose is the dimer of glucose and represents the simplest glucose oligomer. There were various previous studies on cellobiose decomposition in HCW, considering different reactor systems, analytical instruments and reaction conditions. For example, an early study by Bobbleter and Bonn in 1980s17 showed that under batch conditions at 180−249 °C, cellobiose decomposition in HCW produces glucose at a yield up to ∼60%. Subsequent studies18,19 using a continuous reactor enabled the elucidation of the reaction pathways and the evaluation of the reactions kinetics of cellobiose decomposition in HCW at 300−420 °C. Two primary reaction routes were proposed:19,20 one is hydrolysis at the glycosidic bond of cellobiose to produce glucose; the other is retro-aldol condensation (pyrolysis) at the reducing end of cellobiose to produce glycosyl-erythrose (GE), which can be further decomposed to glycosyl-glycolaldehyde (GG) via retro-aldol condensation. However, a recent study by Kimura et al.21 suggested that at 100−140 °C, the decomposition of cellooligosaccharides (DP up to 4) may proceed with the terminal glucose unit initially transformed to fructose, followed by the breakage of the glycosidic bond to release fructose. Therefore, the reports in the existing literature have obvious discrepancies in both results and proposed reaction mechanisms. The study of Kimura et al.21 suggests that isomerization may be important primary reactions during cellobiose decomposition in HCW. Under alkaline conditions (1 M NaOH), isomerization reactions of cellobiose was also suggested to take place even at room temperature (22 °C).22 As a reaction medium, HCW Received: Revised: Accepted: Published: 17006
September 22, 2013 October 28, 2013 October 30, 2013 November 15, 2013 dx.doi.org/10.1021/ie403140q | Ind. Eng. Chem. Res. 2013, 52, 17006−17014
Industrial & Engineering Chemistry Research
Article
mL min−1 before the flow was delivered into the electrospray ionization (ESI) interface of the mass spectrometer for MS analysis.24 The concentration of glucosyl-glycoaldehyde (GG) in a liquid sample was quantified using an isocratic method which elutes 50 mM NaOH solution. The total carbon in a liquid sample was also determined by a total organic carbon (TOC) analyzer (Shimadzu TOC-VCPH). Under the reaction conditions in this study, the gas production from cellobiose decomposition is negligible because the total carbon in the liquid sample is close to 100% under all reaction conditions. The pH values of the solutions before and after the reaction were also measured as indications of the formation of organic acids. Posthydrolysis experiments, modified from a NREL method,25 were performed for selected liquid samples to identify the cellobiose isomers. 2.3. Data Acquisition and Processing. On a carbon basis, the conversion of cellobiose (X), the yield (Yi) and selectivity (Si) of a compound i at a reaction time of t can be calculated according to the following equations:
has a high ion product and promotes both acidic and alkaline catalyzed reactions. Under such unique conditions, the isomerization reactions may also be potentially important. Unfortunately, there has been little report on this important aspect in the literature and the roles of isomerization reactions during cellobiose decomposition in HCW are largely unknown, particularly under practical conditions (at temperatures above 200 °C) relevant to most HCW applications. Consequently, this study aims to clarify the fundamental reaction mechanism of cellobiose decomposition in HCW, focusing more on the formation of cellobiose isomers via isomerization reactions. A series of systematic experiments were performed using a continuous reactor under noncatalytic conditions at 225−275 °C. A state-of-the-art high performance anion exchange chromatography equipped with pulsed amperometric detection and mass spectrometry (i.e., HPAECPAD-MS) was used for the analysis of liquid products. The unique capabilities of the HPAEC-PAD-MS enable this study to provide some new insights into the fundamental mechanism of cellobiose decomposition in HCW under the prevailing experimental conditions.
2. EXPERIMENTAL SECTION 2.1. Materials and Reactor System. Cellobiose, various high-purity standards and regents required for experiments were purchased from Sigma-Aldrich. A continuous reactor system was used to investigate cellobiose decomposition in HCW. The detailed description of the reactor system can be found elsewhere.23 Briefly, a stream of ultrapure water delivered by an HPLC pump was preheated in a fluidized sand bath. The preheated water was then mixed with a stream of cellobiose water solution (3 g L−1) fed by another HPLC pump to rapidly heat up the feed solution to a desired reaction temperature (225−275 °C). The ratio of the flow rates of the preheated water to the cellobiose solution was 2:1 in each experiment, delivering a final solution at a reactant concentration of 1 g L−1 before entering the stainless tube reactor at the reaction temperature. To monitor the temperature of the feed solution after mixing, a thermocouple was located at the inlet of the reactor. The residence time of reactant solution was controlled by adjusting the length of the stainless steel tube reactor and the flow rate. In each experiment, the effluent from the tube reactor was immediately cooled using an ice water bath and then collected as liquid sample for subsequent analyses. The reactor pressure was controlled to be 10 MPa by a backpressure regulator in all experiments. 2.2. Sample Analysis. The liquid samples were analyzed by an HPAEC-PAD-MS. The detailed method for simultaneous HPAEC analyses with PAD and MS detectors can be found elsewhere.24 In this study, the HPAEC-PAD-MS analysis was conducted using a Dionex ICS-5000 ion chromatography (IC) system equipped with a CarboPac PA20 analytic column (3 × 150 mm) and a guard column (3 × 30 mm). This study has developed a gradient method that achieves the successful separation of the glucose isomers and cellobiose isomers in the liquid sample. The method elutes 5 mM NaOH solution in the first 10 min and then increases the NaOH concentration to 50 mM in 30 min. After separation, the flow was split into two streams for PAD and MS analyses, respectively. An in-line desalter (Dionex ASRS 300, 2 mm) was also installed to convert the eluent into a MS compatible solution. For efficient ionization of the eluted carbohydrates, a LiCl (0.5 mM) solution was pumped into the eluent flow at a flow rate of 0.05
X=
C2(0) − C2(t ) C2(0)
(1)
Yi =
ci × ai C2(0) × a 2
(2)
Si =
Yi ci × ai = X [C2(0)‐C2(t )] × a 2
(3)
where ci is the mass concentration of a compound i in the liquid sample after reaction; C2(0) is the mass concentration of cellobiose in the solution before reaction; C2(t), is the mass concentration of cellobiose in the solution collected at a reaction time t; ai and a2 are the carbon contents (wt%) of the compound i and cellobiose, respectively. For those compounds which can be identified by MS but cannot be quantified due to the unavailability of standards, a relative selectivity was defined based on the peak area as shown in eq 4: Sj =
Aj × aj [C2(0)‐C2(t )] × a 2
(4)
where Aj is the peak area of an unquantified compound j in the liquid sample after reaction; and aj is the carbon content (wt%) for an unquantified compound j. The validity of eq 4 requires the HPAEC-PAD detection system to operate in the detector’s linear response range for the compounds that can be identified but cannot be quantified due to the lack of standards. A method (the detailed description can be found elsewhere12) based on peak area is followed to ensure the relative selectivity is comparable for a compound in different liquid samples. Briefly, each liquid sample needs to be diluted to make sure that the concentration of the compound in the diluted sample is within the linear response range of the HPAEC-PAD detection system. To verify this, a sample which has the largest peak was chosen to be diluted by 2-fold. If the peak area ratio between the undiluted and diluted samples is 2, the sample diluted by 2-fold will be used to compare the relative selectivity for all the samples. If the peak area ratio is less than 2, higher dilution factors will be tested till a suitable dilution factor is found. 17007
dx.doi.org/10.1021/ie403140q | Ind. Eng. Chem. Res. 2013, 52, 17006−17014
Industrial & Engineering Chemistry Research
Article
3. RESULTS AND DISCUSSION 3.1. Identification of the Compounds in the Liquid Products from Cellobiose Decomposition in HCW. A typical IC chromatogram of liquid sample from cellulose decomposition at 275 °C is shown in Figure 1. It can be seen
Figure 1. Typical IC chromatogram and SIM scans of a liquid sample from cellobiose decomposition in HCW at 275 °C. Peaks: 1, levoglucosan; 2, cellobiosan; 3, glycolaldehyde; 4, 5-HMF; 5, glucose; 6, mannose; 7, fructose; 8, erythrose; 9, cellobiose; 10, isomer of cellobiose (i.e., cellobiulose); 11, isomer of cellobiose (i.e., glucosylmannose); 12, GE.
that based on available standards, a series of compounds can be identified by PAD, including sugar products (i.e., glucose, fructose, and mannose), dehydration products (i.e., cellobiosan, levoglucosan, and 5-HMF), and fragmentation products (i.e., glycolaldehyde, erythrose, and GG). Most of these compounds have been reported in previous literature.18,19 However, there are still several main peaks which can be detected by PAD but cannot be identified due to the unavailabilities of the standards, including peaks 10−12 as shown in Figure 1. Based on the simultaneous analysis of the same sample by MS, the selected ion monitoring (SIM) scan at m/z 349 (corresponding to a molecular weight of 342, i.e., that of cellobiose and its isomers) shows three individual peaks that match exactly the peaks 9−11 detected by PAD. Peak 9 is confirmed to be cellobiose by the standard. For peaks 10 and 11, the detailed mass spectra are then presented in Figure 2a and b, each of which clearly shows a single major peak at m/z 349. Therefore, peaks 10 and 11 must be two different isomers of cellobiose. Such identifications are significant to the understanding on the fundamental reaction mechanism of cellobiose decomposition in HCW. To the best of the knowledge of the current authors, this is the first report on the formation of two different cellobiose isomers during from cellobiose decomposition in HCW. Although the formation of cellobiulose (glucosyl-fructose, an isomer of cellobiose) from cellobiose decomposition has been recently reported by Kimura et al.21 at low temperatures of 100−140 °C, the presence of two cellobiose isomers was not reported previously. In fact, the formation of cellobiose isomers have never been identified by other works on cellobiose decomposition at higher temperatures, possibly due to the instrument used for sample analysis. Kimura et al.21 employed 13C NMR spectroscopy which is able to detect cellobiulose. However, other studies18−20 relied on HPLC systems which are difficult to separate the cellobiose and its isomers. Therefore, the results presented in this study demonstrate that the HPAEC-PAD system is powerful to separate and detect the isomers of cellobiose.
Figure 2. Mass spectra of main unidentified peaks. (a) Peak 10; (b) Peak 11; (c) Peak 12.
Although it was indicated that one of cellobiose isomer could be cellobiulose according to a recent work by Kimura et al.,21 the other cellobiose isomer is still unknown. To further identify the other cellobiose isomer, posthydrolysis of a liquid sample was performed using a diluted acid at ∼110 °C for 1 h. It was found that three disaccharide peaks all decrease after posthydrolysis, resulting in the increases in the peaks of three monomers including glucose, fructose, and mannose. The increased formation of fructose after posthydrolysis clearly confirms that one celobiose isomer is cellobiulose, whereas the increased formation of mannose after posthydrolysis indicates that the other isomer of cellobiose is glucosyl-mannose. For the other unknown peak 12, the mass spectrum (see Figure 2c) via the SIM scan of MS at m/z 289 match the position of the peak well. Therefore, the peak corresponds to glucosyl-erythrose (GE), which has been widely reported18,19 as a main primary product of cellobiose decomposition in HCW. A careful comparison of the compounds published in other works18−20 indicates that glucosyl-glycolaldehyde (GG) was 17008
dx.doi.org/10.1021/ie403140q | Ind. Eng. Chem. Res. 2013, 52, 17006−17014
Industrial & Engineering Chemistry Research
Article
Figure 3. Yields of various products as a function of residence time at 225−275 °C, expressed on a carbon basis. (a) Yields of sugar products at 225 °C; (b) yields of dehydrated products at 225 °C; (c) yields of fragmented products at 225 °C; (d) yields of sugar products at 250 °C; (e) yields of dehydrated products at 250 °C; (f) yields of fragmented products at 250 °C; (g) yields of sugar products at 275 °C; (h) yields of dehydrated products at 275 °C; and (i) yields of fragmented products at 275 °C.
residence time of ∼263 s (corresponding to a cellobiose conversion of ∼79%). At higher temperatures (e.g., 275 °C), the glucose yield initially increases, achieves a maximum, and then decreases as decomposition proceeds. It can be seen that the highest glucose yield achieved under the experimental conditions in this study is reduced at higher temperatures from ∼31% at 225 °C (at a residence time of ∼263 s, corresponding to a cellobiose conversion of ∼79%), to ∼28% at 250 °C (at a residence time of ∼99 s, corresponding to a cellobiose conversion of ∼90%) and then to ∼24% at 275 °C (at a residence time of ∼33 s, corresponding to a cellobiose conversion of ∼92%). This indicates that glucose decomposition is more pronounced at higher temperatures so that high temperature conditions are not favored for the production of glucose from cellobiose. For the dehydration products, 5HMF has the highest yield, and its formation continuously increases as cellobiose decomposition proceeds. The maximal yield of 5-HMF achieved in this study is ∼18% at 275 °C so that 5-HMF is one of main compounds from cellobiose decomposition in HCW. This is in consistence with the previous reports on 5-HMF being the product of fructose dehydration.26,27 For the fragmentation products, glycolaldehyde has the highest yields compared to GG and erythrose. The yield of glycolaldehyde also continuously increases as the residence time increases, indicating that it is also a main product from cellobiose decomposition. Under the experimental conditions in this study, the maximal yield of glycolaldehyde is ∼9% at 275 °C, lower than that of 5-HMF.
not identified by PAD in Figure 1. To confirm the presence of GG in the liquid sample, the SIM scan at m/z 247 was performed. Indeed, a large peak appeared at a position close to peak 10. While separation of cellobiose and its isomers can be effectively achieved, the data suggest that PAD analysis using the designed gradient method has a very low response to GG. This is confirmed by the tiny peak of the PAD analysis for standard GG solutions. Therefore, this study has developed an isocratic method which can effectively identify and quantify GG in the liquid samples. 3.2. Yields and Selectivities of Products during Cellobiose Decomposition in HCW. It has been clearly demonstrated in the above sections that HPAEC-PAD-MS is powerful to identify the compounds in the liquid products from cellobiose decomposition in HCW. The compounds were further quantified with available standards. The yields of various products are then calculated on a carbon basis, and the results are shown in Figure 3. It can be found that cellobiose is easily decomposed in HCW as the conversion is over 90% for a residence time of ∼30s at 275 °C. It can also been seen that various products, including hydrolysis, dehydration and fragmentation products, have been identified in the liquid products from cellobiose decomposition in HCW. For the hydrolysis products, glucose has the highest yield, compared to other sugar products including fructose and mannose. At 225 °C, the glucose yield continuously increases as cellobiose decomposition proceeds. Under the range of residence time studied, the highest yield of glucose at 225 °C reaches ∼31% based on total carbon in cellobiose after a 17009
dx.doi.org/10.1021/ie403140q | Ind. Eng. Chem. Res. 2013, 52, 17006−17014
Industrial & Engineering Chemistry Research
Article
Figure 4. Selectivities of various products as a function of cellobiose conversion at 225−275 °C, calculated on a carbon basis. (a) Glucose; (b) fructose; (c) mannose; (d) 5-HMF; (e) cellobiosan; (f) levoglucosan; (g) glycoaldehyde; (h) erythrose; and (i) GG.
of glucose selectivity at X → 0 (i.e., limX→0(Yglucose/X)) is nonzero when the glucose selectivity is extrapolated to X → 0 (see extrapolated lines in Figure 4a). On the other hand, the intercept of glucose selectivity at X → 0 is only ∼10% so that at the initial stage only ∼10% of cellobiose is converted into glucose via hydrolysis reaction. Such observation in turn suggests that glucose is not a main primary product from cellobiose decomposition. An increase in the reaction temperature increases the intercept of glucose selectivity at X → 0, indicating that high temperatures enhance the selectivity of hydrolysis reaction during cellobiose decomposition (i.e., from ∼10% at 225 °C to ∼20% at 275 °C). The data in Figure 4 also show that the evolution of glucose selectivity largely depends on the reaction temperature. Although the initial glucose selectivity is low at a low temperature, it increases rapidly as cellobiose conversion increases. At cellobiose conversions >25%, the glucose selectivity at 225 °C is actually higher than that at 275 °C. Such an increase in glucose selectivity could be due to two possible reasons. One is that glucose can be produced from some intermediates from cellobiose decomposition, and such reactions are promoted at low temperatures. The other is that some products from cellobiose decomposition may catalyze the hydrolysis reaction to produce glucose from cellobiose. As shown in Figure 4b and c, the selectivities of fructose and mannose increase with cellobiose conversion up to a high conversion (i.e., ∼90%), but are much lower than that of glucose. The highest selectivities are ∼5.8% for fructose and ∼1.2% for mannose at 250 °C, respectively. An increase in reaction temperature has slight inhibition effect on the formation of fructose and mannose. It is also clear that fructose and mannose are not the primary products from cellobiose
For GG, its yield is even lower, with a maximal yield of only ∼0.9% at 275 °C. Based on the product yields at various temperatures, the selectivities of different products were further estimated (on a carbon basis) and plotted as a function of cellobiose conversion in Figure 4. It should be noted that at a high cellobiose conversion (e.g., >95% at 275 °C, corresponding to a residence time longer than 50 s, see Figure 3g−i), the reactions are dominant by secondary reactions of primary products produced from cellobiose decomposition because the primary reactions of cellobiose have nearly completed. Yet, the data in Figure 3 clearly suggest that such secondary reactions lead to significant changes in the product selectivities. Figure 4a shows that the selectivity of glucose initially increases with cellobiose conversion, achieves a maximum, and then decreases as the conversion further increases. The maximal selectivity of glucose is ∼41%, which is achieved at 225 °C and a cellobiose conversion of ∼70%. An increase in reaction temperature leads to a reduction in the maximal selectivity of glucose, from ∼41% at 225 °C, to ∼32% at 250 °C and then to ∼29% at 275 °C. It should be noted that the condition where the maximal glucose selectivity is achieved is not same as the condition where the maximal glucose yield is achieved. For example, at low temperatures (i.e., 225 °C), although the glucose selectivity starts to decrease at conversions >70%, the glucose yield still increases as cellobiose conversion increases (see Figure 3), because of more glucose formation than glucose decomposition. Another interesting observation from Figure 4a is that glucose is indeed a primary product from cellobiose decomposition. This is concluded based on the delplot method.28,29 It is clearly shown in Figure 4a that the intercept 17010
dx.doi.org/10.1021/ie403140q | Ind. Eng. Chem. Res. 2013, 52, 17006−17014
Industrial & Engineering Chemistry Research
Article
Figure 5. Relative selectivities of unquantified products as a function of cellobiose conversion at 225−275 °C, calculated on a basis of peak area. (a) Peak 10; (b) Peak 11; (c) GE.
Figure 6. Summarized reaction pathways during cellobiose decomposition in HCW under the reaction conditions, considering the reports in the literature18,19 and the current study.
decomposition as the two compounds are found at cellobiose conversions >10%. The selectivity of 5-HMF, which is a main product from cellobiose decomposition, increases as cellobiose conversion increases (see Figure 4d). A decrease in reaction temperature promotes the formation of 5-HMF, as the selectivity of 5-HMF is always higher at lower temperatures at the same cellobiose conversion level. It is also clear that 5-HMF is also not a primary product from cellobiose decomposition as it is found at cellobiose conversions >20%. Compared to 5-HMF, other dehydration products such as cellobiosan and levoglucosan have very low selectivities (i.e., 10%. Compared to glycolaldehyde and erythrose, the selectivity of GG is even lower, and it increases with cellobiose conversion and starts to decrease at conversions >70% (see Figure 4i). Similarly, GG is also not primary product as it can only be found at conversions >25%. Further analyses of carbon balances indicate that the total quantified products only account for 18−61% of total carbon in converted cellobiose depending on reaction temperature and 17011
dx.doi.org/10.1021/ie403140q | Ind. Eng. Chem. Res. 2013, 52, 17006−17014
Industrial & Engineering Chemistry Research
Article
residence time. Particularly, the carbon balances are much lower at lower cellobiose conversions. Obviously, there are significant amount of unquantified products in the liquid sample from cellobiose decomposition, especially at low cellobiose conversions. Figure 4 also indicates that glucose and glycolaldehyde are only the minor primary products from cellobiose decomposition. To find out other primary products from cellobiose decomposition, a relative selectivity method was used to evaluate other unquantified peaks based on their peak area. The relative selectivities of unquantified compounds are plotted as a function of cellobiose conversion in Figure 5. Peaks 10 and 11 are the isomers of cellobiose based on their mass spectra, and their relative selectivities are found to decrease continuously with increasing the cellobiose conversion. Clearly, the data in Figure 5 show that the two cellobiose isomers are indeed primary products from cellobiose decomposition. Similarly, the relative selectivity of GE (peak 12) also shows a decrease trend as cellobiose conversion increases. This clearly indicates that GE is also the primary products from cellobiose decomposition. The formation of GE as a primary product from cellobiose decomposition is in consistence with the conclusions drawn in previous studies.18,19 As shown in Figure 5, the extrapolation of the relative selectivities of cellobiose isomers to X → 0 shows that a higher percentage of cellobiose decomposes via isomerization reactions to produce isomers at lower temperatures. As cellobiose decomposition proceeds, the selectivities of the two cellobiose isomers decrease more rapidly at lower temperatures, indicating that the decomposition of cellobiose isomers is promoted as reaction temperature decreases. In contrast, the GE selectivity is significantly increased with increasing reaction temperature, because an increase in reaction temperature leads to the decomposition of a higher percentage of cellobiose via retro-aldol condensation reactions to produce GE. 3.3. Reaction Pathways and Mechanism of Cellobiose Decomposition in HCW. The results presented in this paper so far provide some new insights into the reaction pathways of cellobiose decomposition in HCW. Figure 6 summarized the state-of-the-art understanding in the reaction pathways of cellobiose decomposition in HCW under the reaction conditions, considering the reports in previous publications18,19 and this study. The results in this study show that there are at least three primary reactions: (i) hydrolysis reaction to produce glucose; (ii) retro-aldol condensation reaction to produce GE and glycolaldehyde, and (iii) isomerization reactions to produce cellobiose isomers. The first two primary reactions, that is, the hydrolysis and retro-aldol condensation reactions were previously reported.18,19 The data in this study has clarified at least three important points related to the primary products of cellobiose decomposition in HCW. First, this study confirms that GG is not a primary product of cellobiose decomposition. This clarifies the discrepancies in the literature. For example, in the study by Kabyemela et al.,18 GG was considered as an primary product of cellobiose decomposition in HCW via retro-aldol condensation reactions. However, in another study by Sasaki et al.,19 GG is not considered as a primary product of cellobiose decomposition, rather, GG is considered as a secondary product of GE via retro-aldol condensation reactions. The experimental data in this study clearly show that GG is not a primary product, because GG can only be found at cellobiose conversions higher than 25% (see Figure 4i), supporting the proposed production pathway proposed by Sasaki et al.19
Second, the data in this study reveal that glucose as a primary product of cellobiose decomposition in HCW has a very low selectivity (
