Kraft pulping process has been known for decades. It has higher market value
and compatibility with most of the wood species and gives high strength pulp
and better operation in chemical recovery system. It is desirable for many paper
products and produces a stronger sheet of paper thus being able to compete with
the other pulping processes such as sulfite (Yusup, 2004).
Many parameters can affect the pulp properties but there are also some disadvantages
in Kraft pulping for instance; pulp must be bleached and low yield due to carbohydrate
losses (Kusuma, 2003).
From 400 species of Eucalyptus in Australia, only 17 species are more
useful for pulping. However, growth rate is a main factor in chemical properties
of wood (Rashidi, 2002). Eucalyptus can grow well
on various soils. Eucalyptus camaldulensis is easily planted in tropical
zone and has good suitability with different climate and ecological conditions
especially in Thailand. It contains higher holocellulose (72-76.8%) which makes
it one the most competitive hardwoods to be used for pulping (Ona
et al., 1996).
There is a general agreement that high alkali concentration in the initial
phase and a leveled out effective alkali concentration in the bulk phase result
low lignin concentration in the residual stage of cooking (Saucedo
and Kishanagopalan, 2002). The higher residual alkali charge is required
to prevent the recondensation of lignin on the fibers. It has been found that
the amount of residual alkali charge in the case of high alkali charge was high
(Minh, 2000). Previous study on pulping characteristics
of Eucalyptus camaldulensis showed that the highest cooking yield can
be obtained at low alkali charge, lower temperature, shorter cooking time, higher
liquor to wood ratio (L:W). The delignification rate slowed down due to decrease
in Effective Alkali (EA) charge in the cooking process. This effect was more
pronounced at high L:W ratio (Kusuma, 2003). In the Kraft
pulping, the rate of delignification strongly depends on temperature and time
so that the most part of delignification is done in the bulk delignification
phase. H-factor, area under relative reaction rate curve against time, is a
valuable parameter not only including both cooking time and temperature, but
it also helps to control pulping process for a target kappa number (Gullichsen
and Fogelholm, 1998). In this study all experiments were conducted at the
The viscosity of pulp depends on the hemicelluloses content of pulp and it
has been found that the pulp viscosity at a given kappa number increases with
increasing sulfide sorption in the wood (Heaing, 2000;
Olm et al., 2000). It has been found that the ECF bleaching sequence
(D0E0D1D2) of delignified Kraft
pulps made from Eucalyptus gave higher pulp brightness (>90% ISO)
(Camilla and Ulf, 1991).
The sulfidity and active alkali charge are two effective parameters that influence
the pulp properties (Gullichsen and Fogelholm, 1998).
From the delignification in Kraft pulping, sulfide and hydrosulfide ions facilitate
the reaction of lignin removal and sulfur works mainly as a catalyst which is
not consumed or transformed much (Bhowick, 1993). It has
been found that the efficiency of lignin removal increases with increasing sulfidity.
In practice, sulfidity is most often determined by chemical balance of the mill:
higher in closed cycle and lower in open mills. The optimum sulfidity also depends
on several factors, such as wood species, alkali charge, cooking temperature
and properties desired in the final product. The upper limit is typically determined
by odor release from the plant (Mirshokraee, 2003). Thus
both active alkali charge and sulfidity play an important role in the Kraft
Modeling strategy allows developing empirical models as a function of various
independent variables (Navaee-Ardeh et al., 2004).
Such a modeling can aim improving the pulping operation as well as minimizing
extra experimental attempts thus saving a lot of time and costs. It has been
reported that the strength properties of Kraft made pulps can be improved under
optimized cooking conditions (Akgul et al., 2007).
Proper mathematical models can be used to control the pulping process for a
given product or specific pulp mill. In spite of extensive experimental investigations
on Kraft pulping process, the modeling of pulp characteristics of Eucalyptus
camaldulensis versus pulping variables is yet missing. In this study, the
effect of various sulfidity and active alkali charges of cooking liquor were
investigated on the produced pulp properties. Furthermore, the multiple nonlinear
modeling of experimental data was performed to illustrate the effect of each
independent variable on pulping properties.
MATERIALS AND METHODS
Pulping and Screening
Fresh screened (according to standard SCAN-CM 40: 94) mill chips of E.
camaldulensis collected from a local pulp mill near Pulp and Paper Center
at Asian Institute of Technology (AIT) were used for the pulping process. The
dry matter content of the chips was determined according to the standard SCAN-CM
39:94. Wood chips were cooked in an air heated 6-2.5 L autoclave digester with
400 g o.d. chips charge in each autoclave. The sulfidity and active alkali were
changed in the range of 20-40% (as NaOH) and 19-25% (on o.d. wood), respectively.
The pulps were prepared at a liquor to wood ratio of 4:1. The time from room
temperature to 80°C was 20 min and from 80°C to the cooking temperature
(160°C) was 55 min followed by a further 2 h at the cooking temperature.
After cooking the pulp of each sample was washed using 8 L tap water for 5 times.
The washed pulp then was centrifuged and homogenized. The homogenized pulp was
weighted and dry matter content of pulp was determined (SCAN-C 3:78) and yield
calculated. Then it was disintegrated and followed by a screening with a 0.2
mm slotted flat screen. Kappa number was measured according to the standard
Fifty gram of each screened pulp was used for bleaching by a D0ED1
sequence in plastic bag in water bath. Pulp and distilled water before bleaching
were preheated in microwave oven. The bleaching conditions have been summarized
in Table 1. After each stage, the pulp was washed with distilled
water using suction bottle. Brightness and viscosity were measured according
to the standards SCAN- C 11:75 and SCAN-CM 15:88, respectively.
In this study, all the independent variables were normalized according to
the following formula (Navaee-Ardeh et al., 2004):
The normalized independent variables and experimental data of pulping and bleaching processes were used for developing the best empirical models to fit the curve in which the dependent variables were evaluated by the following formulas:
Where, y1 and y2 are dependent variables, xi
is the normalized independent variable and Xmin and Xmax
are the minimum and maximum value of corresponding independent variables, respectively.
Unknown linear and non-linear coefficients (ai, bi and
cij) were found using experimental data (Table 2)
and were summarized in Table 3. Six conducted experiments
with the corresponding independent and dependent variables have been summarized
in Table 2.
|| Bleaching conditions
|| Summary of the white liquor components and corresponding
pulping, screening and bleaching processes results
|| Linear and non-linear coefficients of dependent variables
according to Eq. 2 and 3
|KN = Kappa Number, CY = Cooking Yield, RR = Reject Rate, UBPV
= Unbleached Pulp Viscosity, BPV = Bleached Pulp Viscosity, UBPB = Unbleached
Pulp Brightness, BPB = Bleached Pulp Brightness, : x1=
Normalized sulfidity (% as NaOH); x2 = Normalized active alkali
(% on o.d. wood), : x1 = Normalized kappa number; x2
= Normalized screened pulp yield, r2 = Piersons coefficient
RESULTS AND DISCUSSION
Normalization of the independent variables provides better estimates for the regression coefficients by reducing correlation between linear and quadratic interaction terms. Multiple nonlinear regression analysis to find the best model for experimental data was performed by MATLAB software (version 6.1) using least square method. Kappa number has two roles: independent and dependent variable. The results of this process have been summarized in Table 3.
Based on the experimental results, independent variables were changed over
the following ranges:
||Active alkali charge (on o.d. wood): 19-25%
||Sulfidity (as NaOH): 20-40%
||Kappa number: 20 -34.4
||Screened pulp yield: 44.3-47.9%
In these experiments, cooking time, temperature and liquor to wood ratio were
kept constant. The best advantage of multiple nonlinear regression is related
to considering the effects of both sulfidity and active alkali charge at the
same time on the dependent variables. Kappa number as a function of active alkali
charge and sulfidity has been shown in Fig. 1. It can be seen
that by increasing sulfidity, kappa number linearly increased, however kappa
number varied nonlinearly in respect to the active alkali charge. At low active
alkali charge, kappa number varied rapidly but its variation slowed down at
the high active alkali charge. It can be inferred that the active alkali charge
has reverse effect on kappa number, but sulfidity has positive effect (Fig.
1). It can also be seen that high kappa number was obtained at low active
alkali charge and high sulfidity whereas low kappa number is achieved at high
active alkali charge and low sulfidity.
The effect of sulfidity on cooking yield variation was marginal in comparison to the variation of cooking yield with active alkali charge. The variation of cooking yield was like the kappa number variation but it varied nonlinearly respect to the sulfidity. Moreover, the variation of cooking yield at low active alkali charge significantly increased (Fig. 2).
Insufficient cooking process generates higher rejects and results more shives, chop and debris in the screening process. There was a maximum reject rates at 33% sulfidity which needs to be avoided. Indeed it can be inferred that the kappa number, cooking yield and reject rate as dependent variables are more sensible to active alkali charge than sulfidity. However by increasing active alkali this sensibility decreased but there was a uniform change for these dependent variables with the variation of sulfidity.
At low screened pulp yield by increasing initial pulps kappa number,
the viscosity of unbleached pulp increased (Fig. 4). The main
reason can be explained that the higher the kappa number, the lower the removal
of lignin during pulping and therefore less hemicellulose dissolution occurs.
The variation of viscosity of unbleached pulp was completely fitted as a polynomial
function of the normalized screened pulp yield and kappa number.
|| Variation of Kappa number vs. active alkali and sulfidity
|| Variation of cooking yield vs. active alkali and sulfidity
|| Reject rate in pulp screening vs. active alkali and sulfidity
|| Unbleached pulp viscosity vs. screened pulp yield and kappa
|| Brightness of unbleached pulp vs. screened pulp yield and
There was a minimum viscosity and brightness for unbleached pulp showing lower
hemicellulose content. At low kappa number by increasing screened pulp yield,
the viscosity linearly increased. By increasing kappa number and decreasing
screened pulp yield both viscosity and brightness of unbleached pulp increased
In contrast, the variation of bleached pulp viscosity and brightness was completely different than unbleached pulp (Table 3 and Eq. 3).
Different chemical charges in the pulping process of E. camaldulensis chips affected pulping yield, kappa number, viscosity and brightness, significantly. The average values of cooking yield, unbleached and bleached pulp brightness and viscosity were 46.1%, 23.1, 83.5 ISO%, 769 and 719 mL g-1, respectively. It was possible to improve the brightness of the pulp using D0ED1 bleaching sequence 60.4 ISO% (261.47% improvement) compared to the unbleached pulp brightness whereas the reduction in viscosity (hemicellulose content) was only 50 mL g-1 (6.5% reduction). Generally the viscosity (hemicellulose content) of unbleached pulp was more than the viscosity of bleached pulp. The empirical models facilitated to investigate the impact of different parameters on pulping properties. The models are also able to help controlling the pulping process for better results as well as minimizing the number experimental trials thus cost saving in the mill scale. The developed models can also be used to optimize the pulping process according to the requirements of a particular pulp mill.
Financial support from Finland Government is greatly appreciated.