vitamin c

European Journal of Scientific Research ISSN 1450-216X Vol.30 No.2 (2009), pp.294-304 EuroJournals Publishing, Inc. 2009 http://www.eurojournals.com/ejsr.htm

Optimization Hydrogenation Process of D-glucose to D-sorbitol

Over Raney Nickel Catalyst

Muthanna J. Ahmed Chemical Engineering Department, College of Engineering, Baghdad University

Baghdad, Iraq

Anees Abdullah Khadom Corresponding Author Department of Chemical and Process Engineering

Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia, Bangi 43600, Selangor, Malaysia

E-mail: [email protected] Tel: +60-17- 876 9594

Abdul Amir H. Kadhum

Department of Chemical and Process Engineering , Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia

Bangi 43600, Selangor, Malaysia

Abstract

Hydrogenation of D-glucose in the presence of Raney nickel as a catalyst was employed for the preparation of D-sorbitol. The effects of the reaction time (15-90 min), reaction temperature (10-60 oC), and catalyst to D-glucose ratio (5-15 wt %) on the yield of D-sorbitol were studied. The experimental design of Box-Wilson method was adopted to find a useful relationship between the effecting variables and the D-sorbitol yield. The experimental data collected by this design was successively fitted to a second order polynomial mathematical model. The analysis of variance shows that the catalyst to D-glucose ratio had the greatest effect on the yield of D-sorbitol among other variables. An optimum operating conditions of 67 min reaction time, 46 C reaction temperature, and 14.5 % catalyst to D-glucose ratio gave 87.15 % D-sorbitol yield, 6.51 % D-mannitol yield, and 95.93 % D-glucose conversion. D-sorbitol of purity 99.8 % was obtained after its separation from D-mannitol, the byproduct of the hydrogenation process, using the fractional crystallization method. Keywords: D-Sorbitol, Catalytic Hydrogenation, D-glucose, Raney-Nickel, Box-Wilson

Design 1. Introduction D-Sorbitol, C6H14O6, is a hexahydric alcohol with a straight chain of six carbon atoms and six hydroxyl groups. As a pure solid, it is a white, odorless, nontoxic, crystalline material. D-Sorbitol has been widely used in the food, pharmaceutical, medicine, and chemical industries. An important use of D-

Optimization Hydrogenation Process of D-glucose to D-sorbitol Over Raney Nickel Catalyst 295

sorbitol is for the preparation of isosorbide dinitrate which is a well known coronary vasodilator used in the treatment of hypertension (Orourke and vasodilator, 2002).

D-Sorbitol can be extracted from many natural raw materials such as seaweed, edible fruits as apples and grapes, and plants. However, the extraction of D-sorbitol from these raw materials is not a good commercial source. Both electrolytic reduction and catalytic hydrogenation processes are used. The catalytic hydrogenation process seems to have supplanted the electrolytic process for the commercial production of D-sorbitol (Gorp et al, 1999).

D-Sorbitol can be made by the catalytic hydrogenation of three naturally occurring hexoses, D-glucose, D-fructose, and L-sorbose. However, D-glucose is the practical source because of its greater availability and low cost. The catalytic hydrogenation of D-glucose to D-sorbitol has been widely applied on an industrial scale, and strenuous efforts have been made to achieve quantitative conversions. Raney nickel catalysts, to which various promoters have been added, are widely used in the hydrogenation process. These catalysts, which are prepared from nickel -aluminum alloys, are approximately as effective as platinum catalysts for promoting many hydrogenations at low pressures and temperatures (Zeifert et al, 2008).

During the preparation of D-sorbitol from D-glucose by catalytic hydrogenation, D-mannitol is also formed as a result of isomerization of D-glucose to D-fructose. Different separation processes have been used for the separation of D-mannitol from D-sorbitol. D-Mannitol is readily separated from D-sorbitol by a fractional crystallization process from aqueous solutions, in which D-sorbitol is soluble (Makkee and Kieboom, 1985). Both D-mannitol and D-sorbitol can be separated chromatographically on a column of calcium poly styrenesulfonate, which professionally retains D-sorbitol (Melaja and Hamalainen, 1975).

Experimental designs are frequently performed in the study of empirical relationships between one or more measured responses and a number of variables. Having such relations, it can specify a combination of variables that will achieve some practical benefit. In the chemical industry, experimental designs are particular applied to the study of process variables and how they affect the product. The basic types of experimental design are factorial design, fractional factorial design, and Box-Wilson design which are a common type of statistical experiment especially applicable to optimization analysis (Lazic, 2004).

The catalytic hydrogenation process was used by different researchers to prepare D-sorbitol from different feed stocks using different catalysts. Makkee et al. (1985) used a copper on silica support catalyst for the preparation of D-sorbitol by catalytic hydrogenation of D-glucose. Barbosa et al. (1999) prepared D-sorbitol by catalytic hydrogenation of sucrose over a ruthenium on zeolite support catalyst. Toukoniitty et al. (2005) used Raney nickel catalyst for the preparation of D-sorbitol by catalytic hydrogenation of D-fructose.

The aim of the present work is to study the catalytic hydrogenation of D-glucose to D-mannitol using Raney nickel catalyst, and the effect of reaction time, reaction temperature, and catalyst ratio on the yield of D-sorbitol. 2. Experimental Work 2.1. Materials

D-Glucose: D-Glucose (supplied by Hopkins and Williams, Searle Company) of purity 99% was used for the preparation of D-mannitol. Hydrogen: Hydrogen of purity 99.9% was used for D-glucose hydrogenation. Catalyst: Highly active Raney nickel catalyst (supplied by Aldrich Company) of pore size 50 and surface area 90 m2/g was used for the hydrogenation of D-glucose to D-Sorbitol. It was stored as 50% slurry in water. Its nickel content was 75.11%.

296 Muthanna J. Ahmed, Anees Abdullah Khadom and Abdul Amir H. Kadhum

2.2. Apparatus

A schematic diagram of laboratory experimental unit used for the hydrogenation of D-glucose is shown in Fig.1. This unit consists of a reaction flask which was a Pyrex three-necked 500ml glass flask. The feed was charged to the reaction flask through a glass dropping funnel with a capacity of 100ml, and the hydrogen gas was fed to the reaction flask by means of a special perforated bulb tube (Sparser) in order to keep the solution in considerable agitation and to prevent settling of the catalyst. The loss of vapor from the reaction flask was prevented by using a Pyrex double pipe glass condenser with an inner pipe of spiral shape. The reaction flask temperature was measured by a glass thermometer range from 0 to 100C, and maintained at the desired value by the use of a water bath. The hydrogen gas flow rate was controlled by a needle valve (Micro Hooke mite) and measured by a rotameter (GEC-Elliott). The composition of product was measured by a shimadzu Lc-6A high performance liquid chromatography (HPLC) system consist of: 6mm ID 150mm L column (Shim-Pack CLC-ODS) and spectrophotometric detector (SPD-6A at 63nm). 2.3. Experimental Procedure

D-glucose 20 wt% solution was prepared by dissolving 40 g D-glucose in 160 ml distilled water. This solution was brought to pH value of 9 by the addition of 0.01 g calcium hydroxide per 40 g D-glucose and mixed with 2-6 g Raney nickel catalyst to form the reaction slurry. The resultant slurry was charged through the dropping funnel into the reaction flask. When the required temperature was reached, hydrogen gas was fed continuously at a flow rate of 1.86 liter/min through the reaction slurry by means of the perforated glass bulb tube to keep the slurry in considerable agitation during the reaction time. At the end of duration time, the solution was cooled and the catalyst was allowed to settle at the bottom of the reaction flask. The supernatant solution was filtered and analyzed by HPLC method for its D-sorbitol, D-mannitol, and D-glucose content. The above procedure was applied at the optimum operating conditions and the D-sorbitol was separated from the product solution as follows:

The alkaline product solution (filtrate) was first neutralized with dilute sulfuric acid and then concentrated to about 33ml at a temperature of 45 C and under a pressure of 50 mmHg using a glass evaporation-vacuum system. To this solution, 50 ml of 90% ethanol was added and the precipitated salts were removed by filtration. The solution was cooled to 15 C and seeded with D-mannitol. After 2 hours, D-mannitol crystallized in the form of fine needle-like crystals whish were removed by filtration. Then 0.1 % of the amount of D-sorbitol was added to the solution as finely powdered crystals and the solution was stirred for 14 hours at 15 C, the D-sorbitol crystals formed were separated by filtration, washed with cold 80% ethanol, and dried at 40 C and 600 mmHg. The dried D-sorbitol crystals were recrystallized twice from ethanol, filtered and dried. Then the purity (assay) of D-sorbitol was 99.8.

The D-sorbitol yield, D-mannitol yield, and D-glucose conversion are defined mathematically as follows: D-sorbitol yield (%) 100=

gf

sWW , D-mannitol yield (%) 100=

gf

mWW , and D-Glucose

conversion (%) 100= gf

gpgf

WWW .


Figure 1: Schematic diagram of D-Glucose Hydrogenation Laboratory Unit

3. Mathematical Modeling A second order polynomial mathematical model was employed to represent the yield of D-sorbitol y as a function of reaction time X1, reaction temperature X2, and catalyst ratio X3. The general form of this model for three variables is represented by the following model:

239

228

217326

315214332211o

XBXBXB XXB

XXBXXBXBXBXBBy

+++++++++=

(Model 1)

An experimental design based on Box-Wilson central composite method was used to organize the experiments. In order to design the experiments, the operating range of the variables is first specified, thus:

X1=reaction time from 15 to 90 min. X2=reaction temperature from 10 to 60 C. X3=catalyst ratio from 5 to 15 %. The total number of experiments N is computed according to the following equation,

122 ++= PN P , then, ( ) 1513223 =++=N The relationship between the coded variable and the corresponding real variable is as follows: ( ) PX XX XXcoded center centeractual = min (Model 2)

4. Results and Discussion 4.1. Analysis of Box-Wilson Experimental Results

Table 1 shows the coded and real values of independent variables for the experiments to be conducted according to Box-Wilson method, and the experimental response represented by D-sorbitol yield.

A nonlinear least-squares regression program based on Gauss-Newton method was used to fit. (Model 1) to the coded data and experimental D-sorbitol yield given in Table 1. This fitting gave the predicted D-sorbitol yield y, the residual error e, and the coefficients B of this equation as shown in Table 2. The fitted response surface of (Model 1) is:

23

22

2132

3121321

X25.3X33.5X14.7X0.45X-

X0.58XX0.14XX23.10X17.8.31X9.8174y

++++=

(Model 3)


The analysis of variance (F-test) was used for testing the significance of each effect in (Model 3). The calculations are given in Table 3. An estimate of the variance Sb2 is obtained by dividing the experimental error variance Sr2 by the sum of squares of each effect X2 , as follows, = 2

22X

Sb

rS ,

where, = 22 erS , and nN = . The significance of effects may be estimated by comparing the

values of the ratio (B2/Sb2) with the critical value of the F-distribution at 95 % confidence level (F0.95=6.61). If the ratio B2/Sb2 > 6.61 then the effect is significant. Thus, according to the results shown in Table 3, it appears that the interaction effects are not significant. The best response function is then conveniently written as follows:

23

22

21321 .25X3.33X5.14X7.23X10.17X8.31X981.74y +++= (Model 4)

This equation represents the best form of the mathematical model that relates the D-sorbitol yield y to the three variables in terms of coded levels. An equivalent equation, in terms of the actual levels will be more useful in estimating the response for any desired conditions in the range of the independent variables. The new equation with a correlation coefficient of 97 % is obtained as follows:

23

22

4-

21

4-32

21

2

9.58045X3X10*.5742082X10*.5319971

X45814.11X10*377045.2X10*040103.2168413.1y

+++= (Model 5)

The optimum operating conditions was determined by differentiating both sides of (Model 5) for each independent variable and equating the derivative to zero (Edgar and Himmelblau, 2001). Thus, the optimum conditions corresponding to a maximum D-sorbitol yield are 67 min reaction time, 46 C reaction temperature, and 14.5 % catalyst ratio. At these conditions the optimum theoretical D-sorbitol yield calculated from (Model 5) is 88.87 % and the experimental yield is 87.15 %. Also the experimental D-mannitol yield is 6.51 % and D-glucose conversion is 95.93 %. 4.2. Effect of Reaction Time

Figures 2 and 3 shows the effect of reaction time on the theoretical D-sorbitol yield at different reaction temperatures and catalyst ratios, respectively.

Figure 2 shows that the yield of D-sorbitol increases with increasing reaction time from 15 to 67 min. beyond such optimum reaction time, the yield of D-sorbitol decreased due to the isomerization of D-glucose to D-fructose, as explained by Pigman and Hortorn (1972). Table 3 shows that an increase in reaction time from 52.5 to 90 min at 35 C and 10 % catalyst ratio leads to a decrease in D-sorbitol yield from 74.8 to 66.6 %, an increase in D-mannitol yield from 6.24 to 12 %, and an increase in D-glucose conversion from 86.15 to 89.55 %. Figure 3 together with Fig. 2 show that there is no interaction between reaction time and catalyst ratio, and also between reaction time and temperature as summarized in Table 3. 4.3. Effect of Reaction Temperature

The effect of reaction temperature on the theoretical yield of D-sorbitol at different reaction times and catalyst ratios is shown in Figs. 4 and 5, respectively. These figures indicate that at lower temperature, longer reaction time and larger amounts of catalyst are necessary to obtain comparable values of D-sorbitol yield. At 45 C and 14.5 % catalyst ratio, 67.5 % D-sorbitol yield is achieved in 30 min, while at 17 C and 14.5 % catalyst ratio, 67 min is needed to achieve the same yield, as shown in Fig. 4. Figure 5 shows that at 35 C and 67 min, 69 % D-sorbitol yield is obtained using 8 % catalyst ratio, while at 19 C and 67 min it is necessary to use 14.5 % catalyst ratio in order to reach the same yield.

Figures 5 and 6 also show that the use of high temperatures (above 46 C) reduces D-sorbitol yield, because at high temperatures, Raney nickel catalyst losses its activity as explained by Jianping et al. (2004).


4.4. Effect of Catalyst Ratio

Figures 6 and 7 shows the effect of catalyst ratio on the theoretical D-sorbitol yield at different reaction times and temperatures, respectively. In Fig. 6, the yield of D-sorbitol yield increases with the increase of catalyst ratio. An increase in the catalyst ratio from 5 to 13 % at 46 C and 67 min causes an increase in the yield of D-sorbitol from 53 to 87.5 %. This figure also shows that the time required to reach 67.5 % yield can be reduced from 67 to 30 min by increasing the catalyst ratio from 7 to 13 %. The use of low catalyst ratio increases D-mannitol yield, because the rate of isomerization of D-glucose will be greater than its hydrogenation rate, as explained by Makkee et al. (1985).

Figure 7 shows that there is no interaction between the catalyst ratio and the reaction temperature, as explained in Table 3. This table shows that the catalyst ratio is the most important variable since it has the greatest effect on D-sorbitol yield. Table 1: D-Sorbitol Yields, D-Mannitol Yields, and D-Glucose Conversions of Box-Wilson Method

Experiments

Coded Variables Real Variables D-Sorbitol Yield (%) D-Mannitol Yield (%)

D-Glucose Conversion (%) Run No. X1 X2 X3 X1 X2 X3

1 -1 -1 -1 6.05 26 0.0469 31 4.8 40.91 2 1 -1 -1 10.29 26 0.0469 50.14 8.31 62.38 3 -1 1 -1 8.86 54 0.0469 47.76 6.91 59.72 4 1 1 -1 13.01 54 0.0469 69.37 10.32 86.8 5 -1 -1 1 1.89 26 0.0931 47.6 1.12 52.71 6 1 -1 1 6.42 26 0.0931 72.08 5.56 81.85 7 -1 1 1 5.07 54 0.0931 65.61 3.01 72.48 8 1 1 1 9.07 54 0.0931 86.5 7.13 93.06 9 -1.732 0 0 2.57 40 0.07 41.06 2.00 47.12

10 1.732 0 0 11.25 40 0.07 66.6 12.00 89.55 11 0 -1.732 0 5.09 15 0.07 46 1.00 49.73 12 0 1.732 0 7.76 65 0.07 72.51 6.05 81.83 13 0 0 -1.732 11.52 40 0.03 45.36 11.5 62.75 14 0 0 1.732 2.29 40 0.11 85.62 1.99 92.07 15 0 0 0 9.30 40 0.07 74.8 6.24 86.15

Table 2: Statistical Analysis Results of Fitting (Model 1)

No. X1 X2 X3 Y(%) y(%) e(%) Coefficient B 1 -1 -1 -1 31.00 31.37 -0.37 B0 74.81 2 1 -1 -1 50.14 49.12 1.02 B1 9.31 3 -1 1 -1 47.76 48.87 -1.11 B2 8.17 4 1 1 -1 69.37 66.06 3.31 B3 10.23 5 -1 -1 1 47.60 51.57 -3.97 B4 -0.140 6 1 -1 1 72.08 71.63 0.45 B5 0.578 7 -1 1 1 65.61 67.29 -1.68 B6 -0.445 8 1 1 1 86.50 86.79 -0.29 B7 -7.14 9 -1.732 0 0 41.06 37.26 3.80 B8 -5.33

10 1.732 0 0 66.60 69.52 -2.92 B9 -3.25 11 0 -1.732 0 46.00 44.67 1.33 12 0 1.732 0 72.51 72.96 -0.45 13 0 0 -1.732 45.36 47.33 -1.97 14 0 0 1.732 85.62 82.77 2.85 15 0 0 0 74.8 74.80 0.00


Table 3: Variance Analysis of Variable Effects*

Effect 2X Coefficient B Variance = 222X

Sb

rS F-value

2b

2 /SB= F0.95 =6.61 15 B0 74.81

X1 14 B1 9.31 0.988 87.73 S X2 14 B2 8.17 0.988 67.56 S X3 14 B3 10.23 0.988 105.92 S

X1X2 8 B4 -0.140 1.73 0.011 NS X1X3 8 B5 0.578 1.73 0.193 NS X2X3 8 B6 -0.445 1.73 0.114 NS X12 26 B7 -7.14 0.532 95.83 S X22 26 B8 -5.33 0.532 53.40 S X32 26 B9 -3.25 0.532 19.85 S

* S= Significant, NS= Not Significant

Figure 2: D-Sorbitol Yield vs. Reaction Time at Different Temperatures

0

20

40

60

80

100

0 20 40 60 80 100Reaction Time (min)

D-S

orbi

tol Y

ield

(%)


Figure 3: D-Sorbitol Yield vs. Reaction Time at Different Catalyst Ratios

0

20

40

60

80

100

0 20 40 60 80 100Reaction Time (min)

D-S

orbi

tol Y

ield

(%)

Figure 4:D-Sorbitol Yield vs. Reaction Temperature at Different Times

0

20

40

60

80

100

0 20 40 60 80

Reaction Temperature (C)

D-S

orbi

tol Y

ield

(%)


Figure 5: D-Sorbitol Yield vs. Reaction Temperature at Different Catalyst Ratios

0

20

40

60

80

100

0 20 40 60 80

Reaction Tempertaure (C)

D-S

orbi

tol Y

ield

(%)

Figure. 6: D-sorbitol Yield vs. Catalyst Ratio at Different Times

0

20

40

60

80

100

0 5 10 15 20

Catalyst Ratio (%)

D-S

orbi

tol Y

ield

(%)


Figure 7: D-Sorbitol Yield vs. Catalyst Ratios at Different Temperatures

0

20

40

60

80

100

0 5 10 15 20

Catalyst Ratio (%)

D-S

orbi

tol Y

ield

(%)

5. Conclusion The reaction conditions of d-glucose conversion to d-sorbitol were studied using a second order mathematical model. The result calculated using this model is in a good agreement with the experimental data. Model and experiment result exhibited a significant effect due to catalyst ratio to d-glucose affecting the yield of d-sorbitol. The highest yield of d-sorbitol (87.15%) was observed when the catalyst ratio is 14.5% at 45 C and reaction time within 67 min. 6. Acknowledgement We gratefully acknowledge Universiti Kebangsaan Malaysia for assist and support of this work


References [1] Orourke S.T., Nitro vasodilator, 2002. Pharmacology and Use in the Treatment of Myocardial

Ischemia, Am. J. Pharm. Educ., 66, pp.177-180. [2] Gorp V.K., E. Boerman, and C. V. Cavenaghi, 1999. Catalytic Hydrogenation of fine

Chemicals, sorbitol Production Catalysis Today, 52, pp.349-361. [3] Zeifert B., J. S. Blasquez, and J.G.C. Moreno, 2008. Raney-Nickel Catalysts produced by

Mechanical Alloying, Rev. Adv. Mater. Sci., 18, pp. 632-638. [4] Makkee M. and A.P. Kieboom, 1985. Production Methods of D-Mannitol , Starch, 37(4), pp.

136-141. [5] Melaja A.J. and L. Hamalainen, 1975. Process for the production of mannitol and Sorbitol,

U.S. Patent 3864406, Feb. 4. [6] Lazic Z.R., 2004. Design of Experiments in Chemical Engineering. A practical Guide,

Wiley-VCH Velrag Gmbh & Co. KGaA Weinheim. [7] Makkee, M., A.P.G. Kieboom, and H.Van Bekkum,1985. "Production Methods of D-

Mannitol", Starch, 37, 4, pp. 136-141. [8] Barbosa C.M.B.M., E. Falabella, and M. J.Mendes , 1999. Kinetic Evaluation of

Hydrogenation of Sucrose Over Ruthenium Containing Y Zeolites , Reaction Kinetics & Catalysis Letters, 68, 2, pp.291-298.

[9] Toukoniitty, B., T., Kuusisto, J., and J. P. Mikkola, 2005. "Effect of Ultrasound on Catalytic Hydrogenation of D-Fructose to D-Mannitol", Ind. & Eng. Chem. Res.,44, 25, pp. 9370-9375.

[10] Edgar T. F. and D. M. Himmelblau, 2001. Optimization of Chemical Processes, 2nd Ed., McGraw-Hill Chemical Engineering series.

[11] Pigman W. and D. Hortorn, 1972. The carbohydrates , 2nd Ed., NewYork, London. [12] Jianping, W., W. Changlin, and L. Yanxin, 2004. "The Preparation of D-Glucitol in A Multi-

Tube Airlift Loop Reactor with Low H/D Ratio, Chem. Biochem. Eng. Q., 18, 3, pp. 273-277. Nomenclature

B Coefficient of estimated model (%) C Initial concentration of D-glucose solution (wt. %) e Residual error (%) n Number of coefficients in the model N Number of experiments NS Not significant P Number of variables QH Volumetric flow rate of hydrogen (liter/min) S Significant Sb2 Variance of coefficient B Sr2 Residual variance Wgf Weight of D-glucose in the feed (g) Wgp Weight of D-glucose in the product (g) Wm Weight of D-mannitol in the product (g) Ws Weight of D-sorbitol in the product (g) X1 Reaction time (min) X2 Reaction temperature (C) X3 Catalyst ratio (%) y Theoretical or predicted value of D-sorbitol yield (%) Y Experimental or observed value of D-sorbitol yield (%) Number of degree of freedom

vitamin c

Documents