Glucose reactions within the heating period and the effect of heating rate on the reactions in hot compressed water

Article abstract of DOI:10.1016/j.carres.2005.05.019

Glucose reactions were conducted in hot compressed water (473-773 K, 4-40 MPa) by means of a batch-type reactor. The reactions in the heating period (about for 60 s) were observed. More than 80% of the glucose was consumed in the heating period above 573 K. Gasification of glucose was promoted with increasing temperature. The effect of heating rate (from 4.2 to 15.8 K/s) on glucose conversion was also examined, and gasification of glucose was enhanced with increasing the heating rate.

Full text of DOI:10.1016/j.carres.2005.05.019

Carbohydrate
RESEARCH
Carbohydrate Research 340 (2005) 1931–1939
Glucose reactions within the heating period and the effect of
heating rate on the reactions in hot compressed water
Masaru Watanabe, Yuichi Aizawa, Toru Iida, Caroline Levy,
Taku M. Aida and Hiroshi Inomata*
Research Center of Supercritical Fluid Technology, Tohoku University, 6-6-11-403Aoba, Aramaki, Aoba-ku, Sendai 980-8579, Japan
Received 10 December 2004; accepted 28 May 2005
Available online 14 July 2005
Abstract—Glucose reactions were conducted in hot compressed water (473–773 K, 4–40 MPa) by means of a batch-type reactor.
The reactions in the heating period (about for 60 s) were observed. More than 80% of the glucose was consumed in the heating per-
iod above 573 K. Gasification of glucose was promoted with increasing temperature. The effect of heating rate (from 4.2 to 15.8 K/s)
on glucose conversion was also examined, and gasification of glucose was enhanced with increasing the heating rate.
Ó 2005 Elsevier Ltd. All rights reserved.
Keywords: Glucose; Batch reactor; Hot compressed water; Gasification; Heating period; Heating rate
1
. Introduction
used a flow apparatus, almost without a heating period,
to evaluate the reaction rate. As a result, the primary
reactions of glucose were found to be as follows (Fig.
1): (1) glucose isomerization into fructose via keto–enol
tautomerization, (2) glucose dehydration into 1,6-anhy-
droglucose, and (3) glucose decomposition into alde-
hyde and ketone via retro-aldol condensation. Further,
(4) dehydration of the tautomerization intermediate
and fructose produce 5-hydroxymethyl furfural
(HMF). This reaction mechanism can typically be
applied to the reaction that proceeds in a rapid heat–
Biomass conversion in hot compressed water is one of
the candidates for a biomass refinery technique because
biomass reactions such as the conversion of cellulose
1
,2
into glucose and its oligomers, the conversion of oligo-
3
mers and glucose into useful chemicals,
ification of glucose into H and CH ,
–13
and the gas-
effectively
1
4–35
2
4
proceed in the media. For biomass cascade utilization,
a compound from biomass has to be transformed into
important chemicals. Glucose can be selectively pro-
duced from cellulosic biomass, and it becomes the start-
ing material of the biomass refinery because glucose is
the monomer unit of cellulose, which is the most abun-
dant organic compound on the earth and is the major
component of plant biomass. Thus, the conversion of
glucose into chemicals will be an alternative process to
petroleum chemistry and is desirable for a sustainable
feedstock.
1
0,12
36
flow apparatus. Sinag et al.
and Dinjus and Kruse
studied glucose decomposition in SCW. They proposed
a relatively simple reaction mechanism for glucose as
shown in Figure 2. According to their reaction scheme,
glucose converts into furfurals and acids/aldehydes in
parallel. Further conversion of furfurals leads to acids/
aldehydes and phenol formation. Acids and aldehydes
are precursors of the gaseous compounds. Furfural
formed through dehydration is shown in Figures 1 and
2. Further dehydration of furfural gives phenols. In con-
trast, acids and aldehydes are produced by bond-break-
ing reactions. This mechanism is quite simple; however,
it tells us that the glucose reaction mainly proceeds via
dehydration or bond breaking. The predominance of
The reaction mechanism of glucose in supercritical
3
water (SCW) was studied by Kabyemela et al.
–6
They
*
Corresponding author. Tel.: +81 22 795 7283; fax: +81 22 795
282; e-mail: inomata@scf.che.tohoku.ac.jp
7
0
008-6215/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2005.05.019

1932
M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
HOH C
2
O
HOH2C
CH OH
2
O
HO
OH
HO
OH
OH
OH
HO
Glucose
Fructose
H2C
OH
O
O
HO
OH
1
,6-anhydro
glucose
O
CHO
OH
HOH2C
CHO
CHO
OH
CH OH
2
CHO
H
H
H
O
+
+
HMF
OH
CH OH
CH OH
2
CH OH
CH OH
2
2
2
O
Glyceraldehyde Dihydroxyacetone
CHO
Furfural
CHO
O
O
CHO
CHO
+
CH3
CH OH CH OH
2
2
OH
HO OH
Erythrose
Pyruvaldehyde
Retro-Aldol condensation
Glycolaldehyde
Cannizzaro
reaction
Dehydration
COOH
Isomerization
H
OH
CH3
Lactic acid
Figure 1. Primary reaction pathways of glucose in supercritical water.
and cool down of the reactor. There are several studies
on the effect of this non-steady-state stage on the glucose
and cellulose reactions.
Glucose
2
,13,37
2
Ehara and Saka con-
ducted the experiments of cellulose conversion in super-
critical water with a batch and a flow apparatus. In the
batch system, the heating period was only 10 s from
room temperature to 653 K, and the reaction time was
Acids /Aldehydes
Furfurals
2
–8 s. Even at the shortest reaction time, they obtained
a lot of decomposition products such as dihydroxyace-
tone, glycolaldehyde, levoglucosan, and so on. This
means glucose reactions rapidly proceed at 653 K. They
also confirmed the rapid reaction of cellulose and glu-
cose using the flow apparatus at the reaction time below
Phenols
Gaseous
Bond breaking
Dehydration
0
.5 s. From this study, the effect of the heating period on
Figure 2. Simple reaction pathways of glucose reaction including
secondary or higher reactions in supercritical water.
37
glucose reaction was not clear. Matsumura et al. inves-
tigated the reaction within the heating period. They
focused on the effect of heating rate (how fast the tem-
perature of the reaction mixture increased) on gasifica-
tion of glucose in supercritical water. They confirmed
that higher heating rates brought higher gasification
and higher cold gas efficiency. In this report, they only
focused on gasification, and thus the effect of the heating
period on the other glucose reactions was not shown.
Sinag et al. observed the effect of heating rate on the
product distribution of glucose reaction in supercritical
water. They conducted experiments at 1 K/min
(0.0167 K/s) and 3 K/min (0.05 K/s) and reported that
the acids production and gasification were enhanced
the two pathways (dehydration and bond breaking) can
be estimated from the yields of the key compounds such
as furfurals, acids, aldehydes, phenols, and gaseous
products.
Most of the biomass is solid under ambient condi-
tions. Therefore, understanding biomass reactions in
batch and semi-batch reactors is important for the
establishment of a biomass refinery process. One of
the characteristics for batch and semi-batch systems is
that they have a non-steady-state stage, namely heating
up (this period is called Ôheating periodÕ in this study)
1
3

M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
1933
by rapid heating, while dehydration to furan was sup-
pressed by rapid heating. The final temperature of the
studies by Sinag et al. was 773 K. Since the glucose
reaction was quite fast even at 653 K, this temperature
Table 1. Experimental conditions
Run no.
Heating bath
Temperature of
the bath, K
Reaction
time, s
1
3
a
1
2
3
4
5
6
7
8
9
Sand bath
473
573
623
673
723
773
673
673
673
673
673
673
773
60
60
60
60
60
60
30
45
70
90
30
90
30
a
(
773 K) was too high to know the effect of heating rate
Sand bath
a
Sand bath
on the primary reactions of glucose. In order to develop
a biomass refinery batch or semi-batch process for real
biomass, the effect of the heating period and heating rate
on the reaction must be investigated in more detail at
various temperatures and heating rates. Therefore, in
this study, glucose conversion in a batch-type reactor
was conducted focusing on the effect of heating period
and heating rate.
a
Sand bath
a
Sand bath
a
Sand bath
a
Sand bath
a
Sand bath
a
Sand bath
a
1
1
1
1
0
1
2
3
Sand bath
Tin bath
b
Sand bath
Tin bath
a
With shaking.
Without shaking.
b
2. Experimental
Glucose was obtained from Wako Pure Chemical Co.
and used as received. Pure water, which was distilled
after deionization, was obtained with a water distillation
apparatus (Yamato Co., model WG-220).
Product yield (mol %) of carbon compound was eval-
uated from carbon base as shown below:
Product yield ½mol %Š
The reaction was carried out in a SS 316 stainless steel
3
tube bomb reactor with an inner volume of 6 cm . The
amount of carbon atom in a product
¼
amount of carbon atom in the loaded glucose
loaded amount of glucose was 0.1 g and that of water
was 1.0 g. Ar gas was loaded at 2.5 MPa to ease recov-
ery of the gaseous product after the reaction. After the
loading, the reactor was submerged into the heating
bath. The set-point of the sand bath where the reactor
was submerged was from 473 K to 773 K. A tin bath
was also used to change the heating rate inside the reac-
tor at 673 K or 773 K. The reaction was stopped by
quenching the reactor in a water bath. Table 1 lists the
heating bath and the reaction time of each experiment.
Reaction time ranged from 30 to 90 s.
 100
Hydrogen yield (H yield, mol %) was evaluated from
ð1Þ
2
hydrogen base of a loaded amount of samples as
follows:
H Yield ½mol %Š
2
amount of hydrogen atom in H produced
2
¼
amount of hydrogen atom in the loaded glucose
 100
ð2Þ
The product gas was recovered in gas samplers con-
nected with GC–TCD to analyze its composition and
a syringe to measure the product volume. Among the
gaseous products, small hydrocarbons, namely CH4,
C H , C H , C H , C H , and so on, were also detected
at higher temperatures. However, the total yields of
these hydrocarbons were always below 1 mol %. In this
study, because hydrocarbon formation was not dis-
cussed, the yields of the hydrocarbons are not shown.
After collecting the gas samples, the reactor was opened
and washed with pure water to recover the liquid sam-
ples. Sometimes char formation was observed; however,
the amount of char was quite small and not analyzed
further.
3. Results and discussion
Before conducting experiments of glucose conversion,
the temperature profile inside the batch-type reactor
was measured. Water (1.0 g) was loaded into the reactor
and then sealed with a thermocouple to monitor the
temperature. Then, the batch reactor was shaken inside
the fluidized sand bath previously set at 573, 673, and
773 K. The heating profiles under these conditions are
shown in Figure 3. In this case, it took almost 60 s to
reach the desired temperature inside the reactor. Cool-
ing the reactor down below 373 K was achieved within
30 s for all three conditions. Next, we also employed a
tin bath in addition to the fluidized sand bath as the
heating media. In order to compare the temperature
profile in the tin bath with that in the sand bath, we
measured the temperature inside the reactor in the tin
bath and the fluidized sand bath. In the sand bath, the
temperatures inside the reactor were measured with
and without shaking the reactor. The temperature of
2
4
2
6
3
6
3
8
The identification and quantification of the gaseous
products was conducted by GC–TCD (Shimadzu, model
GC-7A, and Hitachi, model GC163). The amount of
carbon in the water solution was evaluated using TOC
(
5
total organic carbon detector, Shimadzu, model TOC-
000A). The liquid products in the recovered solution
were analyzed by HPLC–RI and to UV (JASCO) using
a KS-802 column (Shodex).

1934
M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
bath, with shaking, it took 60 s to reach the tempera-
ture, and (3) using tin bath, it took only 30 s to reach
the temperature.
All experimental results of glucose reactions are listed
in Table 2.
7
6
5
4
73
73
73
73
773 K
73 K
6
573 K
First, we examined glucose reactions within the heat-
ing process at various temperatures using the fluidized
sand bath. In this examination, the reactor was shaken
in the bath (i.e., it took 60 s to reach the temperature,
as shown in Fig. 3). The identified and quantified prod-
ucts in the recovered aqueous solution were glucose,
anhydroglucose (AHG), HMF, and furfural. Figure 5
shows the yield of these water-soluble products and
the TOC. The difference between the TOC value and
the sum of the quantified products indicates that the
other products that could not be identified were also
produced. For gaseous products, H , CO, and CO were
3
2
73
73
Loadings: Water 1.0 g
Fluidized sand bath with shaking
0
30
60
90
Time, s
Figure 3. Temperature profiles inside the reactor (in the fluidized sand
bath with shaking the reactor).
the tin bath was 673 K. The results are shown in Figure
. As shown in Figure 4, it took about 30 s to reach
73 K with the tin bath. In the case of the fluidized sand
bath, it took 90 s to reach 673 K without shaking the
reactor. On the other hand, it took about 60 s to reach
the same temperature when the reactor was shaken dur-
ing heating. In summary, we found as follows: (1) in the
sand bath, it took 90 s to reach the temperature inside
the reactor without shaking the reactor, (2) in the sand
4
6
2
2
obtained. Water-insoluble products such as char was
obtained in all the experiments. The amount of char
was small, and therefore no further investigation was
done. Most of the glucose remained during the heating
period to 473 K. On the other hand, 20% of the glucose
was mainly converted into dehydrogenated products
during the heating to 573 K. Above 623 K the glucose
was completely consumed within the heating period.
Beyond 673 K the yields of the dehydration products
Fluidized sand
bath with
shaking: 6.3 K/s Fluidized sand
bath: 4.2 K/s
(AHG, HMF, and furfural) and TOC values did not
significantly change with the temperature.
Tin bath: 12.5 K/s
Figure 6 shows the gaseous products of the glucose
reaction within the heating period at the various temper-
atures. With increasing temperature, the yield of gaseous
products increased. According to the mechanism pro-
6
5
4
3
2
73
73
73
73
73
1
3
posed by Sinag et al., the increase of the gaseous prod-
ucts at higher temperature indicates that the bond-
breaking reaction is predominant. Figure 7 shows the
Loadings:
Water 1.0 g
CO/CO ratio in the gaseous products. As shown in this
2
graph, the CO/CO ratio increased with increasing tem-
2
0
30
60
Time, s
90
120
150
perature. This indicates that CO formation is favored at
higher temperature. The yield of H2 increased with
increasing temperature as shown in Figure 6. There
Figure 4. Temperature profiles inside the reactor.
Table 2. Experimental results
Run no. TOC, mol % Glucose, mol % AHG, mol % HMF, mol % Furfural, mol %
H
0
2
, mol % CO, mol % CO
2
, mol % CO/CO
2
1
2
3
4
5
6
7
8
9
1
1
1
1
102.8
106.4
79.4
62.6
60.6
61.6
99.2
90.9
66.3
58.6
69.3
59
87.2
64.5
0
17.4
0
6.28
0
0
0
0
0.77
6.56
6
0.008
0.016
0.18
0.68
1.37
0
0.008
0.611
1.32
3.56
4.12
0
0.21
2.26
3.82
4.6
0.04
0.27
0.35
0.77
1.06
0
1.25
9.56
16.2
11.1
9.67
0
0
6.94
6.59
4.71
4.85
4.75
2.02
6.43
4.9
0
8.07
7.15
6.58
0.29
1.76
2.9
51.7
0
16.3
13.4
5.75
1.83
12.9
14.2
10.8
8.15
0.023
0.33
1.08
0.46
0.21
4.36
0.49
1.12
1.42
1.73
1.12
0.28
0.39
0.38
0.36
0.28
1.44
0
0
1
2
3
0
3.29
4.21
5.92
2.06
3.73
4.85
3.98
8.35
0
8.05
8.9
0
3.3
7.151
63
0
2.98
12

M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
1935
increased with increasing temperature. At lower temper-
atures, CO formation was favored, while CO formation
Glucose 0.1 g
Water 1.0 g
Time 60 s
1
00
2
8
6
4
2
0
0
0
0
0
TOC
was enhanced with temperature. These results show that
the decarboxylation of acids is assumed to be the main
pathway of gas formation, at lower temperature and
the C–C bond breaking of an organic compound
(glucose, aldehydes, and so on) to form CO is enhanced
at higher temperature.
The reaction time dependence on glucose reactions
was also observed at 673 K. As shown in Figures 3
and 4, in the case of the temperature of the sand bath
at 673 K the temperature inside the reactor was about
Furfural
HMF
AHG
Glucose
4
73
573
623
673
723
773
Temperature of the sand bath, K
Figure 5. Liquid products of glucose reactions within heating period.
573 K at 30 s. At 45 s, the temperature inside the reactor
was 653 K. After 45 s, the temperature inside the reactor
was constant at 673 K.
1
1
1
4
2
0
8
6
4
2
0
Glucose 0.1 g
Water 1.0 g
Time 60 s
H2
Figure 8 shows the liquid products of the glucose reac-
tion when the sand bath was set to 673 K. At 30 s of
reaction time (573 K inside the reactor), half of the
loaded glucose was mainly converted into liquid prod-
ucts. After 45 s glucose completely disappeared. The
yield of dehydrogenated products and the values of
TOC decreased with time. Figure 9 shows the gaseous
CO
CO2
4
73
573
623
673
723
773
products. CO and H formation increased with increas-
2
Temperature of the sand bath, K
ing reaction time. As well as the case of the effect of
temperature (Figs. 6 and 7), we consider that all the
H formation was due to the water–gas shift reaction.
Figure 6. Gaseous products of glucose reactions within the heating
period.
2
The CO/CO ratio implies the gas formation pathway.
2
Figure 10 shows the CO/CO ratio. At the earlier reac-
2
tion time (namely, below 673 K), CO formation was
2
1
.2
1
dominant, and this implies decarboxylation was favored
at a lower temperature as same as the results in Figure 7.
With increasing reaction time, CO formation linearly
increased after 60 s (i.e., the reaction at 673 K). In this
region (after 60 s), glucose completely disappeared.
Thus, the product of glucose decomposed further into
CO by breaking C–C bonds. Considering Figure 2, the
precursor of the gaseous compounds are acids or alde-
hydes. As shown in Figures 7 and 10, at the lower
WGS correction
Without WGS correction
0
0
0
0
.8
.6
.4
.2
0
Glucose 0.1 g
Water 1.0 g
Time 60 s
temperatures, CO formation was preferred and this im-
2
plies that the formation of acids occurs with ease. In
contrast, CO formation became dominant at higher
4
73
573
673
773
Temperature of the sand bath, K
2
Figure 7. CO/CO ratios within the heating period.
1
00
80
Temperature 673 K
Glucose 0.1 g
Water 1.0 g
TOC
are a lot of possibilities for the formation of H with the
2
glucose reaction, for example, the dehydrogenation of
the substances that have a hydroxyl group, decomposi-
tion of small compounds, among others. However, H2
formation was assumed to relate to the water–gas shift
6
0
Furfural
HMF
40
AHG
2
0
0
Glucose
reaction here. That is, if all the H formation was caused
2
by the water–gas shift reaction, the H originated from
2
3
0
45
60
Time, s
70
90
CO. In Figure 6, the CO/CO ratio with correction of
2
the contribution of the water–gas shift reaction on CO
and CO formations is also shown. From this figure,
the contribution of the water–gas shift reaction
2
Figure 8. Liquid products of glucose reactions at 673 K including the
heating period.

1936
M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
8
6
4
2
0
temperature profiles at the different heating procedures
Temperature 673 K
Glucose 0.1 g
Water 1.0 g
are shown in Figure 4. In the case of the immersion of
the reactor into the tin bath, the heating rate inside
the reactor was 12.5 K/s. The slowest heat-up procedure
was obtained by use of the fluidized sand bath without
shaking the reactor. In this case, it takes 90 s to reach
H2
CO
6
73 K, which means that the heating rate was 4.2 K/s.
The heating rate was 6.3 K/s in the sand bath with shak-
ing the reactor. As shown in Figure 11, the shortest reac-
tion time (highest heating rate in the tin bath) showed
the highest yield of water-soluble products (TOC and
the dehydration products). Figure 12 shows the gaseous
products at the different heating times. The total yield of
gaseous products was not significantly different; how-
CO2
3
0
45
60
Time, s
70
90
Figure 9. Gaseous products of glucose reactions at 673 K including the
heating period.
ever, yields of CO and H at the highest heating rate
2
were the highest among all three experimental results.
The heating rate also affected the gas formation reac-
tion. Figure 13 shows the CO/CO ratio with and with-
1
WGS correction
2
out correction of the water–gas shift reaction. Although
the maximum temperature was the same at 673 K for all
the experimental results in Figure 13, the CO yield in-
creased with increasing the heating rate. This result
Without WGS correction
0
0
0
0
.8
.6
.4
.2
0
Temperature 673 K
Glucose 0.1 g
Water 1.0 g
1
00
80
Glucose 0.1 g
Water 1.0 g
TOC
6
4
0
0
Furfural
HMF
2
0
40
60
80
100
AHG
Time, s
1
2.5 K/s
6.3 K/s
4
.2 K/s
Figure 10. CO/CO
2
ratios at 673 K including the heating period.
20
0
temperature and longer reaction time at 673 K. These
indicate that aldehyde was preferentially formed. Goto
et al. conducted experiments of glucose conversion
using a flow apparatus at various temperatures and pres-
sures. They confirmed that the retro-aldol condensation
to form aldehydes mainly proceeded at higher tempera-
Tin bath
Sand bath
(with shaking)
60 s
Sand bath
90 s
3
0 s
3
8
Figure 11. Effect of heating rate on the liquid products of glucose
reactions at 673 K.
8
ture and lower pressure. Sasaki et al. also confirmed
this phenomenon. As a consequence, at higher tempera-
ture, the formation of aldehyde resulted in CO forma-
tion. At 673 K the retro-aldol condensation also
proceeded, and the product aldehydes gradually con-
verted into CO; thus CO formation was dominant at
longer reaction time. For confirmation of these specula-
tions, the yields of acids and aldehydes must be mea-
sured and are now being identified and quantified.
Glucose 0.1 g
Water 1.0 g
6
1
2.5 K/s
6.3 K/s
4.2 K/s
H2
CO
4
2
0
CO2
3
7
As mentioned in the Introduction, Matsumura et al.
1
3
and Sinag et al. reported that higher heating rates lead
to higher gasification efficiency. Therefore, finally, we
also examined the effect of heating rate on the glucose
reactions.
Tin bath
Sand bath
(with shaking)
Sand bath
90 s
3
0 s
6
0 s
Figure 11 shows the effect of the heating rate on the
yield of the liquid products and the TOC value. The
Figure 12. Effect of heating rate on gaseous products of glucose
reactions at 673 K.

M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
1937
0
0
0
0
0
.6
.5
.4
.3
.2
30
WGS correction
Glucose 0.1 g
Water 1.0 g
1
5.8 K/s
2
2
5
0
Without WGS correction
15
7
.9 K/s
H2
1
0
5
0
CO
CO2
0
5
10
15
Tin bath
0 s
Sand bath
(with shaking)
Heating rate, K/s
3
6
0 s
Figure 13. Effect of heating rate on CO/CO
2
ratios at 673 K.
Figure 15. Effect of heating rate on gaseous products of glucose
reactions at 773 K.
suggests that the longer time at lower temperature re-
sulted in inhibition of gas formation, in particular CO
formation.
5
We also investigated the effect of heating rate on gas
formation at a higher temperature (at 773 K). The
experiments at 773 K were also conducted by use of
the tin bath and the fluidized sand bath (with shaking).
The heating time required to achieve 773 K and 673 K in
the tin bath was same, namely both 30 s. Figure 14
shows the yield of the liquid products and the TOC val-
ues. Increasing the heating rate suppressed HMF and in-
creased anhydroglucose formation. The TOC value and
furfural yield was not affected by this change in heating
rate. Figure 15 shows the effect of the heating rate on the
gaseous product yield at 773 K. Experiments using the
tin bath significantly increased the gaseous yield. Figure
WGS correction
4
3
2
1
0
773 K
WGS correction
73 K
Without WGS
correction 773 K
6
Without WGS
correction 673 K
0
5
10
15
20
Heating rate, K/s
1
6 shows the CO/CO ratio at different heating rates.
2
For comparison, the results obtained at 673 K are also
shown in this figure. The maximum ratio was about
Figure 16. Effect of heating rate on CO/CO
2
ratios at 773 K.
0
.5 at 673 K, where it reached up to 4.5 at 773 K. The
heating rate in the tin bath (15.8 K/s) was almost twice
as high than that in the sand bath (7.9 K/s). With twice
the increase of heating rate, the CO/CO ratio increased
2
four times. Thus, heating has a great influence for gasi-
fication of glucose at higher temperatures, as suggested
TOC
60
40
20
0
37
by Matsumura et al. and Sinag et al.
13
Glucose 0.1 g
Water 1.0 g
Through this study, we observed the effect of heating
period and rate on glucose conversions in hot com-
pressed water. The concentration of glucose was always
constant in this study. Chemical reactions are affected
by concentration of a reacting species. Naturally, the
effect of heating on the glucose reaction is also changed
by the concentration of glucose. Also, the ratio of glu-
cose to water is an interesting topic for glucose refinery
in hot compressed water. The amount of water is closely
related to the pressure and the concentration of water as
a reactant. The next step for developing the batch and
semi-batch biomass refinery process by use of
hot compressed water will be studied on the effect of
the factors listed above.
Furfural
HMF
AHG
7
.9 K/s
1
5.8 K/s
Tin bath
0 s
Sand bath
3
(with shaking)
6
0 s
Figure 14. Effect of heating rate on liquid products of glucose
reactions at 773 K.

1938
M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
. Conclusions
3. Kabyemela, B.; Adschiri, T.; Malaluan, R.; Arai, K. Ind.
Eng. Chem. Res. 1997, 36, 1552–1558.
4
4
5
. Kabyemela, B.; Adschiri, T.; Malaluan, R. M.; Arai, K.;
Ohzeki, H. Ind. Eng. Chem. Res. 1997, 36, 5063–5067.
. Kabyemela, B. M.; Adschiri, T.; Malaluan, R.; Arai, K.
Ind. Eng. Chem. Res. 1997, 36, 2025–2030.
Glucose conversions using a small batch reactor were
performed. Different types of heating bath were
employed: the fluidized sand bath, the fluidized sand
bath with shaking the reactor, and the tin bath. The
heating periods were 90 s, 60 s, and 30 s in the fluidized
sand bath, the fluidized sand bath with shaking the reac-
tor, and the tin bath, respectively.
6. Kabyemela, B. M.; Takigawa, M.; Adschiri, M.; Malal-
uan, R. M.; Arai, K. Ind. Eng. Chem. Res. 1998, 37, 357–
3
. Kabyemela, B. M.; Adschiri, T.; Malaluan, R. M.; Arai,
61.
7
K. Ind. Eng. Chem. Res. 1999, 38, 2888–2895.
8. Sasaki, M.; Furukawa, M.; Minami, K.; Adschiri, T.;
Within the heating period the fluidized sand bath with
shake (60 s), almost all the glucose disappeared above
Arai, K. Ind. Eng. Chem. Res. 2002, 41, 6642–6649.
. Sasaki, M.; Goto, K.; Tajima, K.; Adschiri, T.; Arai, K.
Green Chem. 2002, 4, 285–287.
9
6
23 K. The yields of TOC and the dehydration products
became lower at higher temperature and plateaued at
higher temperature; that is, almost 60 mol % of TOC
and 20 mol % of the dehydration products still exited
at 673 K, 723 K, and 773 K. On the other hand, gasifica-
tion significantly enhanced at higher temperature.
Among the gas products, CO formation was preferred
at the higher temperature.
1
0. Sinag, A.; Kruse, A.; Schwarzkopf, V. Eng. Life Sci. 2003,
, 469–473.
11. Kruse, A.; Gawlik, A. Ind. Eng. Chem. Res. 2003, 42, 267–
3
279.
1
1
2. Sinag, A.; Kruse, A.; Schwarzkopf, V. Ind. Eng. Chem.
Res. 2003, 42, 3516–3521.
3. Sinag, A.; Kruse, A.; Rathert, J. Ind. Eng. Chem. Res.
2004, 43, 502–508.
14. Yu, D.; Aihara, M.; Antal, M. J., Jr. Energy Fuels 1993, 7,
574–577.
At 673 K, the time dependence of the glucose reaction
was measured. The glucose was completely consumed at
1
5. Sealock, L. J., Jr.; Elliott, D. C.; Baker, E. G.; Butner, R.
S. Ind. Eng. Chem. Res. 1993, 32, 1535–1541.
6. Elliott, D. C.; Sealock, L. J., Jr.; Baker, E. G. Ind. Eng.
Chem. Res. 1993, 32, 1542–1548.
4
5 s, namely before reaching 673 K. The yields of TOC
and the dehydration products gradually decreased with
the reaction time. The gas formation, in particular,
CO formation continued to increase with time.
The effects of heating rate on the glucose reaction
were examined at 673 K and 773 K of the heating bath.
The heating rates at 673 K were 4.2 K/s, 6.3 K/s, and
1
17. Elliott, L. J.; Sealock, D. C., Jr.; Baker, E. G. Ind. Eng.
Chem. Res. 1994, 33, 558–565.
1
8. Elliott, D. C.; Phelps, M. R.; Sealock, L. J., Jr.; Baker, E.
G. Ind. Eng. Chem. Res. 1994, 33, 566–574.
1
9. Xu, X.; Matsumura, Y.; Stenberg, J.; Antal, M. J., Jr. Ind.
Eng. Chem. Res. 1996, 35, 2522–2530.
20. Matsumura, Y.; Xu, X.; Antal, M. J., Jr. Carbon 1997, 35,
1
2.5 K/s in the fluidized sand bath, the fluidized sand
bath with shaking the reactor, and the tin bath, respec-
tively. At 773 K, the heating rates were 7.9 K/s and
1
tin bath, respectively. The effect of the heating rate on
the TOC and dehydration products were not observed
at 673 K and 773 K. In contrast, the gas production
was significantly affected by the heating rate. In particu-
lar, CO formation was promoted by rapid heating. The
819–824.
1. Minowa, T.; Ogi, T. Catal. Today 1998, 45, 411–416.
2. Minowa, T.; Zhen, F.; Ogi, T. J. Supercrit. Fluids 1998, 13,
2
2
5.8 K/s in the fluidized sand bath with shake and the
2
53–259.
3. Minowa, T.; Inoue, S. Renewable Energy 1999, 16, 1114–
2
1117.
24. Antal, M. J., Jr.; Allen, S. G.; Schulman, D.; Xu, X.;
Divilio, R. J. Ind. Eng. Chem. Res. 2000, 39, 4040–4053.
5. Kruse, A.; Meier, D.; Rimbrecht, P.; Schacht, M. Ind.
Eng. Chem. Res. 2000, 39, 4842–4848.
6. Schmieder, H.; Abeln, J.; Boukis, N.; Dinjus, E.; Kruse,
A.; Kluth, M.; Petrich, G.; Sadri, E.; Schacht, M.
J. Supercrit. Fluids 2000, 17, 145–153.
2
2
2
0 mol % of carbon in glucose could be gasified by the
quick heating at 773 K. In this case, the formation of
CO was yielded four times higher than that of CO2.
2
7. Yoshida, T.; Matsumura, Y. Ind. Eng. Chem. Res. 2001,
4
0, 5469–5474.
28. Watanabe, M.; Inomata, H.; Arai, K. Biomass Bioenergy
002, 22, 405–410.
9. Cortright, R. D.; Davda, R. R.; Dumesic, J. A. Nature
002, 418, 964–967.
0. Lee, I.-G.; Kim, M.-S.; Ihm, S.-K. Ind. Eng. Chem. Res.
002, 41, 1182–1188.
Acknowledgements
2
2
3
This research was supported by the Industrial Technol-
ogy Research Grant Program in 02A44001d from the
New Energy and Industrial Technology Development
Organization (NEDO) of Japan.
2
2
31. Hao, X. H.; Guo, L. J.; Mao, X.; Zhang, X. M.; Chen, X.
J. Int. J. Hydrogen Energy 2003, 28, 55–64.
3
3
2. Watanabe, M.; Inomata, H.; Osada, M.; Sato, T.; Ads-
chiri, T.; Arai, K. Fuel 2003, 82, 545–552.
3. Park, K. C.; Tomiyasu, H. Chem. Commun. (Cambridge)
References
2
003, 964–965.
1
2
. Sasaki, M.; Kabyemela, B.; Malaluan, R.; Hirose, S.;
Takeda, N.; Adschiri, T.; Arai, K. J. Supercrit. Fluids
34. Kruse, A.; Henningsen, T.; Sinag, A.; Pfeiffer, J. Ind. Eng.
Chem. Res. 2003, 42, 3711–3717.
35. Osada, M.; Sato, T.; Watanabe, M.; Adschiri, T.; Arai, K.
Energy Fuels 2004, 18, 327–333.
1998, 13, 261–268.
. Ehara, S.; Saka, S. Cellulose 2002, 9, 301–311.

M. Watanabe et al. / Carbohydrate Research 340 (2005) 1931–1939
1939
3
3
6. Dinjus, E.; Kruse, A. J. Phys. Condens. Matter 2004, 16,
S1161–S1169.
7. Matsumura, Y.; Harada, M.; Nagata, K.; Yamashita,
K.-I.; Shimamura, M.; Kikuchi, Y. Proceeding of
2003 AIChE Annual Meeting, Miami, USA, 2003;
p 165e.
38. Goto, K.; Tajima, K.; Sasaki, M.; Adschiri, T.; Arai, K.
Kobunshi Ronbunshu 2001, 58, 685–691, in Japanese.