
Research Article


Influence of Operating Variables on Soda Pulping of Musa paradisiaca MidRib


B.O. Ogunsile,
M.A. Omotoso
and
M.A. Onilude


ABSTRACT

The effect of time, temperature, liquor to solid ratio and concentration of cooking liquor on the pulp yield and residual lignin of Musa paradisiaca midrib was studied. The midrib chips were delignified in a laboratory pulp wood digester under the soda pulping process. Experimental values of the cooking variables were fitted into a secondorder polynomial regression equation which reproduced the pulp yield and residual lignin content with error less than 2 and 4%, respectively. Pulping at high temperature for a short cooking with low to medium soda concentration makes the best compromise for high pulp yields and low residual lignin contents. Ratios of the dissolved lignin to the weight losses were relatively constant irrespective of the duration of cooking.





INTRODUCTION Pulp and paper industries are economically important in developing and developed countries alike because their products are consumed nearly in all facets of human endeavours. The consumption of paper per capital does not only correlate with the gross domestic product but also with the standard of welfare and education (Diesen, 1998; Wardle, 1995). In Nigeria, one of the problems in pulp and paper industries is lack of adequate supply of long fiber pulpwood for paper production (RMRDC, 1996). The great value of the forest resources and the increasing demand for paper and paper board in Nigeria call for an overall plan for the utilization of these resources and for the development of plantation of species most suitable for the various sectors of the forest industry. Apart from raphia palms proposed by Odeyemi (1995), there exist potentials to produce long fiber pulp (even at smallscale basis) from certain annuals. Osadare (1995) suggested the use of bast fibers like kenaf (Hibiscus cannabinus), roselle (Hibiscus sabdariffa) and plantain (Musa paradisiaca) as a way of alleviating the present long fiber shortage. Towards this end, the pulp and paper characteristics of some Nigerian grown Musa cultivas were examined (Oluwadare, 1998; Omotoso and Ogunsile, 2000). The results revealed that the cultivars are potential sources of long fiber pulp. The average fiber length compared favourable well with those obtained from soft wood species. The lignin contents were desirably low, while the cellulose content was a reflection of an average but tolerable pulp yields. Similar results including pulping studies were also reported in the literature but not much has been done on modeling of pulping parameters (Sankia et al., 1997; Cordeiro et al., 2004; Kalpana et al., 2005). Kinetic modeling based on mathematical design have been used as a way of optimizing operating condition in pulping processes (Tjeerdsma et al., 1994). However, this model becomes too complex when more than two independent variables are involved. A more recent approach is the use of central composite factorial design (Aknazarova and Kafarov, 1982). The factor design is preferred to kinetic modeling because of its simplicity especially when severally variables are involved. It allows the development of empirical models involving several independent variables in order to identify patterns of variation in the dependent variables of various pulping processes as applied to diverse vegetable materials (Vacquez et al., 1995; Vega et al., 1997; Gilarianz et al., 1999; Jimenez et al., 2000).
In the present study, a 2nd order factorial design was used to evaluate the effects of operating variables, namely, temperature, time, liquor to solid ratio and concentration of cooking liquors on the pulp yields and residual lignin content of a Nigerian grown Musa paradisiaca under the soda pulping process.
MATERIALS AND METHODS
Musa paradisiaca plants were collected at the Department of Agronomy
of the University of Ibadan. The study was conducted between 2002 and 2004.
The part of the plant considered was the midrib which has high pulp yield.
The composition of the sundried sample was determined to be 53.1% cellulose,
12.4% lignin and 6.9% ash. The alcoholbenzene extractive, cold water solubles,
hot water solubles and 1% NaOH solubles were 8.2, 6.4, 9.8 and 24.9%, respectively.
The fiber was 4.0mm long, 31.7 μm wide and 13.2 μm thick (Ogunsile,
2005).
Pulping: The midrib chips were digested in a 10litre electrically
heated stainless steel digester design after the method of Grant (1961). The
plant chips were charged into the digester with the required amount of chemical
solution at Liquor to Solid (LS) ratios of 7:1 and 10:1. The digester was heated
to the operating temperatures (150 and 170°C), which was then maintained
throughout the experiment for the different cooking times (30150 min). At the
end of each cooking scheme, the pulp was washed several times with water and
dried at 102°C in the oven in order to determine the yield. The pulps were
analyzed for Kappa number as described in TAPPI T236 cm85 1993. The Residual
Klason Lignin (RKL) was estimated from the relationship:
The secondorder factorial design: The central composite factorial design was employed to evaluate and quantify the effect of the operational variables on the pulp yields and Residual Klason Lignin (RKL). The effect of the variables were quantified more precisely by choosing part of the experimental results and grouping them to form a first order full factorial design, with variables at two levels (2^{4}). The factorial design gives expressions that relate the effect of each independent variable and their interactions on the dependent variables. Using this design, some of the experimental data were fitted to a first order polynomial regression equation as implemented in the MATHCAD 2001 Professional statistical package. Individual and second order interaction influences over the response surface of the independent variables were evaluated (Aknazarova and Kafarov, 1982). The mathematical model was: The response variable Y represents the pulp yield and Residual Klason Lignin (RKL). The independent variables X_{1}, X_{2}, X_{3 }and X_{4} correspond to liquor to solid ratio, time, temperature and concentration of cooking liquor, respectively. The ranges of values for each independent variable were: LS, 7:1 to 10:1, time, 30150 min, temperature, 150170°C and concentration, 68% NaOH solution. The values of the independent variables were normalized from 1 to +1 by using the equation:
Where:
X_{n} 
is the normalized value of LS, time, temperature and concentration, 
X 
is the absolute experimental value of the variable. 

is the mean of all the experimental values for the variable in question 
X_{max}and X_{min} are the maximum and minimum value, respectively
of such a variable.
RESULTS AND DISCUSSION
The experimental design (2^{4} factorial designs) together with the
pulp yield and the residual klason lignin are presented in Table
1. The coefficients of the model equations and the statistical parameters
establishing their validity are summarized in Table 2. The
correlation graphs between experimental and calculated yield and Residual Klason
Lignin (RKL) content are shown in Fig. 1a and b
respectively, while the ratio of the dissolved lignin to weight loses are estimated
in Table 3.
Experiments 116 of Table 1 allowed the calculation of different parameters in the regression equations a_{i} and a_{ij}. These were subsequently subjected to a ttest to check their significance at 90 to 95% confidence level using the experimental error estimated from the replication at the central point of the design, that is, experiments 17, 18 and 19 of Table 1. The central point of the design corresponds to the following reaction conditions:
Liquor to solid ratio, LS = 8.5,
Time = 90 min,
Temperature = 160°C
Concentration = 7% soda
All normalized independent variables for the central points of the design are zero. The dependent variables (pulp yield and Residual Klason Lignin (RKL) were related to the independent variables through the following equations:
Table 1: 
Experimental design and result for yield and residual lignin 

Table 2: 
Signicative regression parameters 

The nonsignificant parameters at 0.1 level are in parenthesis 
Table 3: 
Ratio of the dissolved lignin to weight losses for the midribpulps 

Values calculated from the respective polynomial equations above were plotted
with the experimental results for the different response variables as shown
from Fig. 1a and b. The results showed very
high correlation between the experimental values and those predicted by the
models with errors less than 2% for the pulp yield and 4% for the Residual Klason
Lignin (RKL).
Pulp yield: Equation 3 allows the estimation of the
variation of the yield with changes in each independent variable over the range
considered while the other three remain constant. Nonsignificant parameters
were dropped to obtain simpler equations. The equation showed that there were
significant interactions between LS and time, temperature and concentration
and also between concentration and temperature.

Fig. 1: 
Correlation graph between experimental and calculated values
of (a) plup yield and (b) Residual Kalson Lignin (RKL) 
The lowest yield (25.8%) was obtained at large values of the four process variables
(i.e., +1 for all). If yield similar to chemical pulp is desired, then all the
process variables must be at normalized values of +1. However, this lowest yield
can be raised from 25.8 to 37.6% by using a lower concentration of cooking liquor
(1 normalized value). The use of a lower concentration of cooking liquor would
also result in some chemical savings in addition to increasing the pulp yield.
The greatest and smallest variations in the lowest pulp yield were caused by
changes in temperature (11.8 units) and LS (3.48 units) respectively while the
effect of time (4.58 units) and concentration (4.18 units) lie in between. The
highest obtainable yield was 49.13%, which occurs at low values of the process
variables (1 for all). The maximum and minimum variations in the highest pulp
yield were caused by changes in time (a change of 7.18 units) and concentration
(a change of 3.7 units), respectively. The effect of LS (6.52 units) and temperature
(5.32 units) lies in between the previous two.
Delignification: The minimum lignin content of the soda pulp according
to Eq. (4) was obtained at high values of cooking time, temperature
and concentration and a low value of the LS. Under this condition, the maximum
variation in the minimum RKL value is caused by a change in temperature (2.48
units) while the minimum is caused by a change in LS (0.5 units). Changes in
time (0.9 units) and concentration (0.7 units) lie in between. Pulping at high
temperatures for a short cooking time with low to medium concentration of pulping
liquor would give the best compromise for both pulp yield and Residual Klason
Lignin (RKL) content.
Ratio of dissolved lignin to weight losses: The ratio of the dissolved lignin to the weight losses as presented in Table 3 can be used to estimate the quantity of lignin present in the pulp provided the pulp yield is known. The results showed that the values are relatively constant irrespective of the duration or time of cooking. The average ratio of the dissolved lignin to the weight loss was 0.14. The constant values seem to suggest that solubilization and delignification take place at the same rate.

REFERENCES 
Cordeiro, N., M.N. Belgacem, I.C. Torres and J.C. Mourad, 2004. Chemical composition and pulping of banana pseudostems. Ind. Crops Prod. (NLD), 19: 147154. Direct Link 
Diesen, M., 1998. World Paper Markets. In: Papermaking Science and Technology, Diesen, M. (Ed.). Vol. 1, Fapet Oy, Helsinki, Finland, pp: 6165.
Gilarranz , M.A., M. Oliet, F. Rodriguez and J. Tijero, 1999. Methanol based pulping of Eucalytus globulus. Canadian J. Chem. Eng., 77: 515521.
Grant, J., 1961. Digestion Experiments. 2nd Edn., Edward Arnold Ltd., London, pp: 4953.
Iglesias G., M. Bao, J. Lamas and A. Vega, 1996. Soda pulping of Miscanthus sinensis: Effects of operational variable on pulp yield and lignin solubilization. Bioresour. Technol., 58: 1723. Direct Link 
Jimenez, L., I. Perez, M.J. De la Torre, F. Lopez and J. Arisa, 2000. Use of formaldehyde for making wheat straw cellulose pulp. Biores. Technol., 72: 283288. Direct Link 
Kalpana, S., S. Sathiamoorthy and V. Kumar, 2005. Evaluation of commercial cultivars of banana (Musa) for their suitability for the fiber industry. Plant Gen. Res. Newslett., 142: 2935. Direct Link 
Odeyemi, F., 1995. Nigerian raphia palm speciesfuture pulp and paper raw materials. Proceedings of the 1st Making and Paper Machine Technology, (MPMT' 1995), Helsinki, Finland, pp: 279289.
Ogunsile, B.O., 2005. Chemical analysis, fibre morphology, soda and sodaethanol pulping of some Nigerian grown Musa species. Ph.D. Thesis, Chemistry Department, University of Ibadan.
Oluwadare, A.O., 1998. Evaluation of the fiber and chemical properties of some selected Nigerian wood and nonwood species for pulp production. J. Trop. For. Res., 14: 110119.
Omotoso, M.A. and B.O. Ogunsile, 2000. Preliminary evaluation of the pulp and paper potential of Musa paradisiaca. Bull. Sci. Assoc. Nig., 23: 207211.
Osadare, A.O., 1995. Strategies toward selfsufficiency in long fiber pulp production in Nigeria. Nig. J. For., 24: 1620.
Parajo, J.C., J.L. Alonso, D. Vazquez and V. Santos, 1993. Optimization of catalysed acetosolv frationation of pine wood. Holzforschung, 47: 188196.
Raw Material Research and Development Council (RMRDC), 1996. Report on the technoeconomic survey of the multidisciplinary task force on pulp. Paper Products Printing and Publishing Sector, Lagos, Nigeria, pp: 132.
Saikia, D.C., T. Gaswami and C.N. Saikia, 1997. Wild banana plants (Musa spp) as source of fibre for paper and cordage industries. J. Sci. Ind. Res., 56: 408413. Direct Link 
Tjeerdsma, B.F., F.H.A. Zomers, E.C. Wilkizon and R. SierraAlvarez, 1994. Modeling organosolv pulping of hemp. Holzforschung, 48: 415422.
Vazquez, G., G. Antorrena and J. Gonzalex, 1995. Acetosolv pulping of Eucalyptus globulus wood. 1. The effect of operational variables on pulp yield, pulp lignin content and pulp potential glucose content. Holzforschung, 49: 6974. Direct Link 
Vega, A., M. Bao and J. Lamas, 1997. Application of factorial design to the modelling of organosolv delignification of Miscanthus sinensis (elephant grass) with phenol and diluteacid solutions. Biores. Technol., 61: 17. Direct Link 
Wardle, P., 1995. World outlooks for supply and demand of fibrous raw materials for paperan Asian perspective. Proceedings of the Paperex 952nd International Conference on Pulp and Paper Industry, Dec. 911, New Delhi, India, pp: 18.



