Glucose hydrogenation in a trickle-bed reactor

Article abstract of DOI:10.1135/cccc19971423

Catalytic hydrogenation of 40% aqueous solutions of D-glucose to D-glucitol was studied in a high-pressure trickle-bed reactor. The reactions were performed on a supported nickel catalyst at temperatures ranging from 115 to 165°C and in the pressure range 0.5 to 10 MPa. The order of the reaction with respect to hydrogen is 0.65 and apparent activation energy 23.8-48.5 kJ mol^-1, the latter depending on initial molar glucose concentration and density and viscosity of the solution. The influence of external diffusion is necessary to take into account for scaling-up the process.

Full text of DOI:10.1135/cccc19971423

Glucose Hydrogenation
1423
GLUCOSE HYDROGENATION IN A TRICKLE-BED REACTOR
Vratislav TUKAC
Department of Organic Technology, Prague Institute of Chemical Technology, 166 28 Prague 6,
Czech Republic; e-mail: vratislav.tukac@vscht.cz
Received February 20, 1997
Accepted June 2, 1997
Catalytic hydrogenation of 40% aqueous solutions of D-glucose to D-glucitol was studied in a high-
pressure trickle-bed reactor. The reactions were performed on a supported nickel catalyst at tempera-
tures ranging from 115 to 165 °C and in the pressure range 0.5 to 10 MPa. The order of the reaction
–
1
with respect to hydrogen is 0.65 and apparent activation energy 23.8–48.5 kJ mol , the latter de-
pending on initial molar glucose concentration and density and viscosity of the solution. The in-
fluence of external diffusion is necessary to take into account for scaling-up the process.
Key words: D-Glucose; Sorbitol; Hydrogenation; Nickel on kieselgur catalyst; Trickle-bed reactor.
D-Glucitol (D-Sorbitol) has a variety of applications in pharmaceutics, cosmetics and
textile and pulp industry, from the synthesis of vitamin C to food products. Its esters
and derivatives are used in plasticizers, emulsifiers and detergents. Production of sorbi-
1
–3
tol is mainly based on glucose hydrogenation on Raney nickel in batch stirred tank
reactors or continuous bubbled towers followed by catalyst separation and ion-ex-
change purification. Ruthenium appears as a new promising catalyst studied both sup-
4
5
4,6,7
ported on carrier and in a form of homogeneous phosphine complex . The use
of
trickle-bed reactor with supported metal catalyst allows to eliminate catalyst separation
and to improve the economy of the process.
6
0
The catalytic hydrogenation of glucose to glucitol is an exothermic reaction (∆H
57.36 kJ mol ), which proceeds according to the Eq. (1).
=
98
2
–
1
–
CHO
OH
CH OH
2
OH
H2
HO
HO
OH
OH
OH
OH
(1)
CH2OH
CH2OH
D-glucose
D-glucitol
The back reaction, dehydrogenation of sorbitol, can be neglected due to high hydroge-
nation equilibrium constant, amounting 470 at 150 °C. Undesirable side and consecu-
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
424
Tukac:
1
2
tive reactions like isomerization of D-glucose to D-galactose and hydrogenolysis of
glucitol to glycerol and propylene glycol play an important role outside of the tempera-
ture range used. The kinetics
8
–10
of hydrogenation on Raney nickel have been studied
in detail. The reaction is first order with respect to glucose and its order with respect to
hydrogen ranges from 0.5 to 1.
The aim of the work was to test feasibility of D-glucose hydrogenation on commer-
cial supported nickel catalysts in continuous trickle-bed reactor. The influence of reac-
tion conditions and glucose solution properties on the rate of glucose conversion was
investigated.
EXPERIMENTAL
The hydrogenation catalyst used was nickel on kieselgur (12% NiO, 2% Cr O ) Cherox 37-03, Che-
2
3
mopetrol Group, a.s., Litvinov, Czech Republic, its original 8 × 8 mm pellets were crushed to size
–3 mm. Two pretreatment methods were used; (i) the reduced and stabilized form of the catalyst
2
–
1
was reactivated with hydrogen in the reactor at 175–200 °C for 3 h at space velocity 500 h (catalyst a),
and (ii) the catalyst was calcinated for two hours at 420 °C and after that it was reduced for 12 h at
3
–
1
70 °C and of hydrogen space velocity 500 h (catalyst b).
The hydrogenated aqueous solutions of glucose contained 40% D-glucose (I), 40% D-glucose and
5% D-glucitol (II), or 40% D-glucose and 27% D-glucitol (III).
1
The experiments were carried out in a trickle-bed micropilot reactor which consisted of a 540-mm
long stainless steel tube with an inner diameter of 30 mm. The volume of the catalyst bed, which was
located in the middle zone of the isothermal region, was 200 ml. The reactor was equipped with a
thermocouple probe in the axis of the bed and an electronically controlled heating jacket. Hydrogen
–
1
was fed from a gas cylinder and depressurized gas flow was set up at 5 l h . An aqueous solution
–
1
of glucose was fed with a high pressure pump (50–400 ml h ) to a preheater and mixed with hy-
drogen at the inlet of the reactor. At the outlet of the reactor, liquid and gas flows were cooled and
separated in the phase separator.
6
r
4
2
0
0.2
0.4
0.6
W/F
FIG. 1
–
1
–1
–1
Mean reaction rate r (mmol g h ) versus residence time W/F (h g ml ) at various temperatures:
✚) 115 °C, (∆) 130 °C, (●) 145 °C, (■) 165 °C, solution I, catalyst a
(
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Glucose Hydrogenation
1425
Determination of glucose in reaction mixtures was performed by spectrophotometry at 520 nm using
–
1
an alkaline solution of 3,5-dinitrosalicylic acid. The recommended concentration range is 0.05–0.60 g l
glucose. The soluble nickel was determined as a red complex with diacetyl dioxime.
RESULTS AND DISCUSSION
In the trickle-bed reactor, kinetic experiments can be influenced by internal and exter-
nal diffusion, inside and outside the particle of the catalyst. The internal diffusion limi-
tation is generally accepted in reactor packed by pellets of porous catalyst. On the other
hand, elimination of influence of external diffusion on kinetic results is necessary. In
the kinetic regime, the rate of glucose hydrogenation is independent of the residence
time of the liquid in the bed. In Fig. 1 is shown the mean reaction rate vs residence time
dependence in the catalyst bed for various temperatures. The range of experimental
conditions for kinetic measurements was found for space hour velocities of the liquid
ranging from 1 to 2, and at temperatures from 115 to 165 °C. Therefore, further evalu-
ation of kinetic parameters was carried out in this region.
The effect of partial pressure of hydrogen on the rate of glucose hydrogenation is
8
–10
presented in Fig. 2. The published
reaction orders with respect to hydrogen partial
TABLE I
Properties of reaction mixtures
–1
–1
r, mmol g
T, K
h
2.46
388
3.00
403
5.26
418
6.32
438
10
r
1
0.1
0
.02
0.01
1
10
20
p
, MPa
H
2
FIG. 2
–
1
Effect of partial pressure of hydrogen pH2 on the hydrogenation rate of glucose r; liquid feed 0.4 l h ,
45 °C, solution I, catalyst a
1
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
426
Tukac:
pressure range from 0.5 to 1. In the presented experiments the straight line follows the
rate order of 0.65. pH value of the resulting reaction mixture was between of 3.9 and 5.1.
From the mean reaction rates measured at 115–165 °C, under 10 MPa of total pressure
–
1
and the feed of 400 ml h , the apparent activation energy was obtained.
The values for three glucose solutions and two catalyst pretreatment methods are
presented in Table I. The catalyst a with pretreatment according to (i) exhibited higher
activity; in addition, the activation energy is almost half of that for the catalyst b pre-
treated according to (ii). Therefore, the reaction rate was probably influenced by exter-
nal mass transfer limitations.
60
x, %
40
2
0
0
0
.0
0.5
1.0
1.5
W/F
FIG. 3
Conversion x versus residence time W/F of various glucose solutions; 145 °C, 6 MPa total pressure,
solution: ■ I, ▲ II, ● III, catalyst b
6
0
0
c
4
2
0
0
110
130
150
170
T, °C
FIG. 4
–
1
Nickel leakage at various liquid feeds: ■ 50, ◆ 100, ● 250, ✚ 400 ml h , solution I, catalyst b
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Glucose Hydrogenation
1427
Thus, external diffusion was responsible for different effects of residence time on
glucose conversion for solutions with different water contents (see Fig. 3). The increase
in density and especially in viscosity due to the sorbitol content shifted the reaction
conditions to the diffusion region. The course of conversion of both sorbitol solutions
was hardly to distinguish, so the decrease in activity may have been caused by sorbitol
sorption.
The main problem of long-term utilisation of the supported catalyst is the leakage of
active metal, which follows in decreasing of catalyst activity. In Fig. 4, the dependence
of the nickel content in the reaction mixture on temperature is presented for various
liquid feeds. It can be concluded that the metal content depends mainly on the residence
time proportional to the reciprocal liquid feed and influence of temperature is not such
important.
The commercially stabilized catalyst of the type a exhibited somewhat higher activity in
comparison to the laboratory b type, probably due to possible metal sintration during
laboratory catalyst reduction; on the other hand, the higher activity shifts the reaction
to the diffusion-limited region.
CONCLUSIONS
Hydrogenation of glucose can be performed continually with more than 98% conver-
sion in a trickle-bed reactor with the fixed bed of supported nickel catalyst. The reac-
tion is first order with respect to glucose and of the order of 0.65 with respect to
hydrogen. Nevertheless, the kinetic data obtained at conditions unaffected by internal
and/or external diffusion can be misleading for scaling up the process. The real glucose
hydrogenation proceeds under strong diffusion limitations due to high substrate concen-
trations, viscosity of solutions, temperature and pressure, and therefore the process
must be investigated under these conditions.
SYMBOLS
c
concentration, mg l–1
activation energy, J mol
–
1
Ea
F
p
–
1
liquid feed, ml h
pressure, MPa
reaction rate, mmol gcat h–1
–1
r
T
W
x
temperature, K
weight of catalyst, g
glucose conversion, %
The author is grateful to Dr J. Soukup in memoriam for his help in the work.
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
428
Tukac:
REFERENCES
. Natta G., Beati E.: Chim. Ind. 25, 211 (1943).
1
2
3
4
5
6
7
8
. Mizuguchi J., Ohoshi M.: J. Chem. Soc. Jpn., Ind. Chem. Sect. 53, 205 (1950).
. Gallezot P., Cerino P. J., Blanc B., Fleche G., Fuertes P.: J. Catal. 146, 93 (1994).
. Arena B. J.: Appl. Catal., A 87, 219 (1992).
. Kolaric S., Sunjic V.: J. Mol. Catal., A 110, 189 (1996).
. Turek F., Chakrabarti R. K., Lange R., Geike R., Flock W.: Chem. Eng. Sci. 38, 275 (1983).
. Dechamp N., Gamez A., Perrard A., Gallezot P.: Catalysis Today 24, 29 (1995).
. Bizhanov P. H., Sokolskii D. V., Popov N. I., Malkhina N. Ya., Khisametdinov A. M.: Kinet.
Catal. 8, 620 (1967).
9. Verma R., Gehlavat J. K.: J. Chem. Technol. Biotechnol. 46, 295 (1989).
1
0. Brahme P. H., Chaudhari R. V., Ramachandran P. A.: Ind. Eng. Chem., Process Des. Dev. 23,
57 (1984).
8
Collect. Czech. Chem. Commun. (Vol. 62) (1997)