Delignification and cellulose degradation kinetics models for high lignin content softwood Kraft pulp during flow-through oxygen delignification Vahid Jafari1*, Kaarlo Nieminen1, Herbert Sixta1 and Adriaan van Heiningen1,2 1
Department of Forest Products Technology, Aalto University, FI-00076, AALTO, Finland
2
Department of Chemical and Biological Engineering, University of Maine, 5737 Jenness Hall, Orono, ME, 04469-5737, USA Keywords: Power-law kinetic models, cellulose degradation, flow-through reactor, oxygen delignification, , high lignin content softwood pulp Corresponding author:
[email protected] Chemical and Biological Engineering 5737 Jenness Hall, Room 117 Orono, Maine 04469-5737 Phone: (207) 581-2277 | Fax: (207) 581-2323
Abstract The kinetics of oxygen delignification and cellulose degradation of a high kappa softwood Kraft pulp (kappa 65) are studied in a “flow-through” reactor where the pulp was held immobile within the reactor and a fresh oxygenated alkaline solution passes through the pulp mat at a set temperature. The feed solution maintains the alkali and dissolved oxygen concentration inside the pulp mat at nearly constant level equal to that in the feed stream. The rate of lignin removal is monitored by measuring the UV absorption in the outflow stream. The kinetics of delignification are obtained by fitting the experimental data to the power-law model with a lignin order of 3.5 in order to describe the rapid initial and subsequent slow delignification phases. The activation energy for lignin removal is 47 kJ/mol. The cellulose degradation kinetics were described by different models. The best cellulose model takes the cellulose degradation to be proportional to delignification due to attack of cellulose by oxygen based radicals generated via the phenolate anion lignin degradation pathway. The novelty of this model is that it also assumes that the fraction of the generated radicals which reach and attack cellulose is inversely proportional to the amount of residual lignin. This model leads to surprisingly simple kinetics and gives an excellent prediction of cellulose degradation kinetics.
1
Introduction The objective of chemical wood pulping is to remove lignin from wood while preserving cellulose and hemicelluloses. Cooking and bleaching are the two main chemical processes in commercial pulp production. About 90% of the lignin inside the fiber wall is removed during cooking, while the remaining lignin (2-5% on wood) is removed in several bleaching stages to obtain a desired pulp brightness and quality (Pekkala, 1984). Environmental concerns have compelled the pulp and paper industry to adopt the environmental friendly oxygen (O) delignification process as a stage located between cooking and bleaching starting in the 1970’s. The goal of O-delignification is to decrease the consumption of chlorine-based bleaching chemicals and decrease effluent pollution by integrating the oxygen filtrate into the regeneration system of the pulping chemicals (Parsad et al., 1996). Lignin removal from softwood Kraft pulp during O-delignification is limited to 50-60%, because further delignification results in unacceptable yield losses and cellulose degradation (Tao et al., 2011). In conventional softwood Kraft pulping, the cooking process is terminated at 25 ~ 30 kappa, followed by O-delignification to a bleachable grade (kappa 12-15) with minimal loss in pulp strength. Kappa number is an indication of residual lignin of pulp and is measured using a standardized analysis method (Jafari et al., 2013). It is well known that the selectivity of Kraft cooking (delignification/cellulose degradation) decreases during the final delignification phase (Teder and Olm, 1981, Hart and Connel, 2006). Therefore, if Kraft cooking is terminated at kappa 40-70 and delignification is continued using the more selective oxygen-alkali process to kappa 15, it may be possible to achieve a pulp yield advantage relative to that of conventional pulping and oxygen delignification (Kleppe et al., 1972, Danielewicz and Surma-Ślusarska, 2006, Jafari et al., 2013, 2014b). This has motivated a number of high kappa O-delignification studies (Iribarne and Schroeder, 1997, Backstrom and Jensen, 2001, Danielewicz and SurmaŚlusarska, 2006, Luo et al., 2012, Vehmaa et al., 2012), including our recent study which showed that a bleachable grade pulp, i.e. a pulp of about 15 kappa and adequate intrinsic viscosity (a measure of cellulose degree of polymerization), could be produced at 2% higher yield (on wood) from a high kappa number pulp (about kappa 60) oxygen delignified in a flow-through (FT) reactor as compared to a conventional oxygen delignified Kraft pulp (Jafari et al., 2014a, Jafari et al., 2014b). The most important parameters characterizing the results of O-delignification are pulp yield, kappa number and intrinsic viscosity. They are dependent on the properties of the incoming pulp and the operating conditions of O-delignification. A number of investigations have been carried out to determine the kinetics of O-delignification on Kraft pulps of high (Parsad et al., 1994, Danielewicz and Surma-Ślusarska, 2006) and standard lignin content levels (Iribarne and Schroeder, 1997, Agarwal et al., 1999, Nguyen and Liang, 2002, Tran, 2004, Ruuttunen and Vuorinen, 2005, Rubini and Yamamoto, 2006). These studies have been undertaken in batch mode in which the chemical concentrations change during reaction. For example, the sodium hydroxide concentration decreases by neutralization with generated acids while the dissolved lignin concentration increases significantly over time (Iribarne and Schroeder, 1997, Agarwal et al., 1999, Nguyen and Liang, 2002, Danielewicz and Surma-Ślusarska, 2006). As a result, it is difficult to obtain intrinsic chemical kinetics in these experiments unless they are performed at very low pulp consistency as was done by Olm and
2
Teder (1979) ( 0.3% consistency). A new approach to determine oxygen delignification kinetics was introduced by van Heiningen and Ji (2012) who maintained uniform NaOH and dissolved oxygen concentrations by dissolving oxygen in an alkali media in a feed tank, and then passed the oxygenated alkali solution through a pulp fiber bed using a flowthrough (FT) reactor. This allowed identification of operating conditions which provide maximum selectivity and delignification efficiency, as well as the formulation of delignification kinetics which provided a fundamental understanding of the reaction pathways for a 24 kappa Kraft pulp (Danielewicz and Surma-Ślusarska, 2006, van Heiningen and Ji, 2012). There are only a few kinetic oxygen delignification studies of high kappa number pulps, and all are performed at medium consistency in batch reactors (Iribarne and Schroeder, 1997, Backstrom and Jensen, 2001, Danielewicz and Surma-Ślusarska, 2006, Luo et al., 2012, Vehmaa et al., 2012). The objective of the present study is to determine the intrinsic kinetics of oxygen delignification and cellulose degradation for a high kappa (65) number softwood pulp based on data obtained in an earlier study which used the FT reactor.(Jafari et al., 2014b) The data were fitted to simple mathematical models to describe the effect of process variables i.e. temperature, alkali concentration and oxygen pressure on lignin removal and cellulose degradation as a function of time.
Kinetic modeling The kinetics of pulp bleaching is complex due to the heterogeneous structure of the pulp fiber cell wall, and the complexity of its structural components(Lindholm, 1986). The type of wood species and conditions of Kraft pulping also affect the kinetics. Four different categories of mathematical models have been used to describe oxygen-alkali delignification kinetics:
One-stage model or power-law rate model (Iribarne and Schroeder, 1996, Agarwal et al., 1999, Ji et al., 2006, Ji et al., 2007).
Two-stage model comprising two parallel rate equations with fast and slow phases and different reaction orders for lignin (Olm and Teder, 1979, Hsu and Hsieh., 1988, Iribarne and Schroeder, 1997).
Model based on dimensionless parameters (Dogan and Guruz, 2008).
Nuclei growth model according to Avrami-Erofeev (Chandranupap and Nguyen, 1998, Nguyen and Liang, 2002).
The most widely used equation for the lignin removal reactions is the power-law model earlier used to describe polymer degradation (Schoeoen. and Nils, 1982). In this model the delignification rate, -dL/dt (mg/g pulp.min) is described by an Arrhenius-type rate constant, kl (min
-1
), which is a function of temperature, and reaction orders in alkali
-
concentration [OH ], oxygen pressure PO2 and lignin content L. (See Eq.1 and 2). The pre-exponential factor Aq (min1
), and reaction orders m, n, and q are obtained by least-square fitting of the experimental data to the model. In Eq.2, EA
is the activation energy (J/mol), R the universal gas constant (8.314 J/mol.K) and T the absolute temperature (K). kl
2
kl
3
The kinetic parameters using the power-law model obtained in a number of previous O-delignification studies are listed in Table 1. It can be seen that the reaction orders and activation energies of the different studies vary widely, most likely because of differences in wood species and lignin content. Specifically, the reaction order in kappa number ranges from 2 to 5.23, and the activation energy from 36 to 100 kJ/mol.
Table. 1 Power-law kinetic parameters of several oxygen delignification studies Reference
Origin
Hsu & Hsie(1988) Iribarne & Schroeder(1996) Argawal et al(1999) Yun Ji et al( 2009) Teder(1981)
Southern pine
Initial kappa 29.5
Pre-exponential factor (s-1) 3.20E+10
Reaction order m n §q 0.68 1.28 5.23
Ea (kJ/mole)
Pinus Taeda
20.3–58
3.00E+06
0.7
0.7
2
51
Southern hardwood Southern pine Nordic softwood
48 24 (FT Reactor) 19
4.42E+07 4.9E+04 -
1.2 0.41 0.6
0.23 0.3 0.5
5.15 1 3.2
98.9 54.5 70
97.2
Experimental Berty “flow-through” reactor The schematic of the Berty type “flow hrough” reac or (FT reactor) is depicted in Fig .1. The FT reactor has a nominal volume of 280 mL containing a 100 mL stationary basket which holds 5g oven dry pulp at medium consistency. The amount of lignin removed from the pulp was determined from the lignin concentration in the effluent passing through a UV flow cell at the reactor outlet. The UV absorption wavelength set at 280 nm was converted to lignin concentration using an extinction coefficient of 22 L-1 g-1 cm based on lignin Indulin AT, (Mead-Westvaco).
However for
o
experiments at temperatures above 100 C and delignification times larger than 60 minutes, the extinction coefficient was raised to 26 L-1 g-1 cm to maintain a consistent mass balance between the amount of lignin removed and the Klason lignin removed from the pulp. It is likely that the increase in extinction coefficient is caused by UV absorption of carbohydrate degradation products. For more details regarding the operation of the Berty reactor see reference (Jafari et al., 2014b).
Fig.1 Process Flow Schematic for flow- through (FT) Reactor
Pulp properties A commercial softwood pine Kraft pulp (Pinus sylvestris) with a kappa number of 65, intrinsic viscosity of 1295 mL/g and yield of 51% on oven dry wood was used. Kappa number (Scan-C 1:00), intrinsic viscosity in cupriethylenediamine
4
(Scan-CM 15:99), and carbohydrate composition using HPAEC according to Janson (1970) were determined. Pulps with a kappa number higher than 35 were subjected to a fast bleaching reaction with chlorite delignification prior to viscosity measurement to remove lignin (1 g pulp in 40 mL water + 1 g NaClO 2 + 0.5 mL acetic acid at 70°C for 5 min) (Jafari et al., 2013). The hexenuronic acid (HexA) content of pulp before and after oxygen treatment was measured according to Vuorinen et al, (2003). The kappa number in this study includes HexA content unless stated otherwise. It must be noted that HexA was not removed during the oxygen alkali process and its contribution to the kappa number corresponds to 2 kappa units i.e. 20 µmol HexA/g pulp. The cellulose D is calcula e fro
he i ri sic viscosi y [η] of
the pulp in mL/g, and the mass fractions of hemicellulose (H) and cellulose (G) in the pulp according to Eq. 3 (Ji, 2007)
(
[ ]
)
The average number of moles of cellulose per tonne of pulp, mn, is also used as a measure of cellulose degradation. It can be calculated according to Eq. 4: ra s er olecular
e ric o gra s ole
06 62D
06 62D
ra oles e ric o ul
Data analysis procedure The reactor setup was calibrated and validated in the Berty FT reactor by changing the pulp weight, mixing speed and feed flow rate. It was found that the kinetics of O-delignification were not affected by the weight of pulp at less than 5 g, a flow rate between 70-130 ml/min and mixer rotational speeds of 1200 rpm or higher. Therefore, the flow rate was fixed at 100 mLmin-1 (corresponding to a reactor mean residence time of about 3 minutes), a mixer speed of 1200 rpm and a pulp weight of 4 g. More details about the FT reactor operation and validation can be found in our earlier study (Jafari et al., 2014b). A typical delignification rate development and also temperature profile is depicted in figure 2. It can be seen that the delignification rate initially decreases rapidly and finally reaches a low value after 20 minutes of reaction. Also shown is the temperature inside the reactor which remains close to the target temperature from the beginning of reaction with variation of ± 0.3 oC throughout the entire experiment.
104
3.0
o
3.3g NaOH/L-7 bar O2- 95 C
2.5
102
2.0
100
1.5
98
1.0 0.5
96 0
10
20
30
40
time (min)
5
50
60
0.0
r(t) (mg/g pulp/min)
Target temperature (oC)
Temperature Delignification rate
Fig.2 Delignification rate and temperature variation vs. time
The final kappa number was calculated using Eq.5. which incorporates the conversion of lignin content into kappa by dividing the lignin content by 0.15% (van Heiningen and Ji, 2012, Jafari et al., 2014b).
ka
a
i ial ka a
v
∫
r
0
g 0
Results and discussion Several experiments were conducted at different operational conditions, i.e., 1.1-3.3 g NaOH/L, 95-105 oC and 5-10 bar oxygen pressure for saturation at room temperature and reaction times of 5-90 minutes. The reference experiment was 2.2 g NaOH/L, 95oC and 7 bar oxygen pressure saturation at room temperature, while only one variable was changed in the other experiments relative to that of the reference experiment. A MATLAB software package was used for curve fitting the kinetic data according to the power law model (equation 1 and 2) by the method of least-square error minimization. For delignification the optimal model fit was:
kl r
000
e c
kl
0
2
0
c
A comparison of the experimental and predicted delignification rate based on equations 6 and 7 plotted against HexAfree kappa number at different alkali concentration (1.1-3.3 g NaOH/L) but fixed temperature of 95 oC and oxygen pressure of 7 bar is shown in Figure 3a. The residual lignin content of the pulp, Lc (i.e. the HexA-free kappa number on the horizontal axis), is obtained by integration of the time-dependent delignification rate (Eq.7). It can be seen that the model gives a good prediction of the delignification rate and the residual lignin content of the pulp. Close agreement between the model and experimental delignification rate is also evident at different temperatures (Figure 3 b) and pressures (Figure 3c).
6
3.5
r(t) (mg/g lignin/min)
3.0 2.5
a: different alkali data model
1.1
2.2
3.5
3.3
3.0
b:different temperature data model
105 oC
100 oC
95 oC
2.5
o
95 C- 7 bar O2 after 90 min
2.0
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.0 15 20 25 30 35 40 45 50 55 60 65
0.0 10 15 20 25 30 35 40 45 50 55 60 65
2.2 g NaOH/L, 7 bar O2 after 90 min
HexA-free kappa
c: different pressure r(t) (mg/g lignin/min)
3.0 2.5
data model
10 bar
7 bar
2.0
2.2 g NaOH/L, 95 oC, after 90 min
5 bar
1.5 1.0 0.5 0.0 15 20 25 30 35 40 45 50 55 60 65
HexA-free kappa
Fig. 3 Comparison of experimental results and model predictions of delignification rate vs. HexA-free kappa at different alkali concentrations (a); temperatures (b) and oxygen pressures (c) The order in lignin concentration of 3.5 obtained in this study is in reasonable agreement with the value of 3.2 determined by Olm and Teder (1979), but differs significantly from 5.23 obtained by Hsu and Hsieh (1988), 5.15 reported by Agarwal et al (1999), and 1.0 and 2.0 reported by Ji (2007) and Iribarne and Schroeder(1996) respectively. The present value for the activation energy of 47 kJ/mol is in close agreement with Iribarne and Schroeder (51 kJ/mol) (1996) and Ji (2006) (54.5 kJ/mol), while much higher values were obtained by the other studies listed in table 1 (Hsu and Hsieh., 1988, Agarwal et al., 1999). These differences may be explained by the different initial kappa number and pulp species used. It is interesting to note that a comparison between the present power model obtained for high kappa (65) pine pulp and that by Ji (2007) also achieved in a FT reactor but for a conventional Kraft (24 kappa) Southern pine pulp, shows a close agreement between the reaction order in alkali concentration (0.47 against 0.42), oxygen pressure (0.47 against 0.305) as well as activation energy value (47kJ/mol vs 53kJ/mol). However, the order in lignin content was found to be 1.0 for the conventional Kraft pulp in the latter study, whereas a much higher value of 3.5 is obtained for the present high kappa pulp. The model predicts that the initial delignification rate decreases rapidly during the first 10-20 minutes (see Figure 2) when the kappa number drops to about 30-35, but then the rate slows down to very low values so that it becomes difficult to reach kappa values of about 15 unless the temperature is increased to 105 °C (see Figure 3). Thus it seems that high kappa number pulps cannot be delignified below a certain limiting level, and that this limiting value is affected by the operating conditions. Thus the present delignification kinetics are clearly different from those obtained by Ji (2007) for a conventional 24 kappa softwood pulp. The results in the latter study implied that all lignin could be removed irrespective of operating conditions if the reaction time was long enough. Therefore the present results suggests that upon Kraft cooking to a high kappa number pulp, less reactive lignin structures remain after a high degree of lignin removal during oxygen delignification treatment. In a separate study of the residual lignin in the oxygen delignified pulp, we have shown that refractory lignin compounds such as 5-5´condensed-OH groups are twice
7
as prevalent in extended oxygen delignified high kappa Kraft pulp than in original Kraft pulp(Jafari et al., 2014a). Therefore, this study shows that at the studied operational conditions the starting kappa number for oxygen delignification could not be higher than about 65 in order to obtain a bleachable pulp grade kappa number of about 15.
Selectivity and carbohydrate degradation We tried to model the cellulose degradation with different rate expressions such as the standard power-law model, a power-law model with zero reaction order in cellulose, and finally a time dependent cellulose degradation model earlier used by Jarrehult and Samuelson (1987). However, fitting the experimental data to the first two models was unsuccessful. For the standard power-law model the activation energy was unrealistically high. The zero order in cellulose model fails to describe the initial fast degradation of cellulose. The model by Jarrehult was successful in describing the cellulose degradation data, but the absence of a mechanistic explanation of the time dependent rate is unsatisfactory. Therefore we tested the cellulose degradation data using two other kinetic models. The first was earlier presented by Ji and van Heiningen (2007) based on a dual mechanism of cellulose degradation by oxygen-based radicals generated by phenolic delignification and additional by direct alkaline attack. The second model also assumes that the cellulose degradation rate is proportional to the delignification rate due to attack by oxygen based radicals generated by the phenolate anion lignin degradation pathway. However the novelty of latter model is that it also assumes that the fraction of the generated radicals which reach and attack cellulose is inversely proportional to the amount of residual lignin. This is because most of the radicals react with the residual lignin where they are generated. This latest model leads to surprisingly simple kinetics which gives excellent prediction of cellulose degradation.
Model 1: Dual attack of cellulose by radicals and alkali Attack by radicals (generated by phenolic delignification) as well as by alkaline hydrolysis has been proposed by Ji and van Heiningen 33 as the degradation pathways of cellulose during O-delignification (Ji, 2007). The kinetic model is: [
]
(8)
where kc is the rate constant of radical attack, kh the reaction rate coefficient of alkaline hydrolysis, K the kappa number and [OH-] the alkali concentration. Integration of equation (8) gives: [
]
(9)
In equation (9) the initial kappa number, K0, is 65. The values of the two constants were determined by least square error fitting the expression for mn to the experimental data. Initially both coefficients were allowed to vary depending on the operating parameters. However, it was found that the best fit was obtained with kh unaffected by the operating variables, and kc only dependent on temperature. The constant alkaline hydrolysis coefficient kh is 0 000262 li er∙ ol cellulose/ o
ul ∙g Na
∙ i u e , while he estimated values of kc at the three temperatures studied are given in
Table 2.
8
Table. 2 kc values at different temperatures Temperature (°C)
kc (moles/ton pulp ∙kappa)
95 100 105
0.0126 0.0168 0.0216
The graphical display in Figure 4 shows that model 1 provides a good description of the experimental data at the different operating conditions.
1.1 2.2 3.3
mn development
1.5
c: effect of pressure
b: effect of temperature
a: effect of alkali Experiments
95 °C 100 °C 105 °C
2.1
2.2 g NaOH/L+10 bar O2
5 bar 7 bar 10 bar
1.5
1.8
1.2
1.5
1.2
1.2
95 oC+10 bar O2
Model 0.9
0.9
0.9 0
20
40
60
80
2.2 g NaOH/L+95 oC
0
100
20
40
60
80
0
100
20
40
Time (min)
Time (min)
60 Time (min)
80
100
Fig.4 Comparison of predictions ion of cellulose degradation model 1 with experimental data of the moles of cellulose per tonne of pulp (mn) during O-delignification at different operational conditions. The relative contribution of radical and alkaline attack on cellulose degradation at the different reaction conditions is seen in Figure 5. The y axis of the graph indicates the increase in moles of cellulose per ton of pulp compared to that of starting pulp (kappa 65), i.e. mn(t) – mn0,. Figure 5a shows that with increasing alkali charge the impact of radical attack increases slightly with increasing alkali concentration due to increased delignification. The effect of direct alkaline hydrolysis increases linearly with alkali concentration as follows from equation (10). In all cases the contribution of alkaline hydrolysis of cellulose is small compared to that by radical attack, but its relative contribution increases with time when the delignification rate slows down while alkaline attack continues at the same rate. a: effect of alkali
Change in mn
0.6
1.1 2.2 3.3
b: effect of temperature
1.2 0.9
0.4
2.2 g NaOH/L + 95 oC 0.4
o
95 C + 10 bar O2
c: effect of pressure
0.6
95 oC 100 oC 105 oC OHcontribution
5 bar 7 bar 10 bar OH-
0.6
0.2
0.2
0.3
Effect of alkali hydrolysis 0.0 0
20
40 60 Time (min)
80
contribution
2.2 g NaOH/L+10 bar O2 100
0.0 0
20
40 60 Time (min)
80
100
0.0 0
20
40
60 Time (min)
80
100
Fig. 5 Contribution of two different mechanisms (harmful radical attacks and alkali hydrolysis) in cellulose degradation during O-delignification at different operational conditions (the unit of mn is moles of cellulose per tonne of pulp).
9
Comparison of Figure 5b with 5a and 5c shows that the effect of temperature on cellulose degradation by radicals is significantly larger than that of alkali concentration or oxygen pressure. For example, with temperature increase from 95 oC to 105 oC, mn increases by about 0.6 moles/tonne at 90 minutes and the reference condition (2.2 g NaOH/L and 10 bar), while the increase in mn value at an alkali concentration increase from 1.1 to 3.3 g/L is only about 0.1 moles/tonne, similar to that seen for an increase in oxygen pressure from 5 to 10 bar. The increase in the value of kc with temperature can be described by an activation energy of 63 kJ/mol. The values reported by Ji33 at 90 °C, 3.3 g/L NaOH and 75 psig oxygen pressure (at 90 °C) are 0.0360 (moles/ton pulp.kappa) for kc and 0.00107 for kh li er∙ ol cellulose/ ton pulp∙g Na
∙ i u e , i.e. about 4 times higher than the present
calculated value of 0.0092 for kc at 90 °C, and 3 times higher for kh. The interpretation of kc is that it is a measure of the fraction of the radicals generated by phenolic delignification which are able to reach cellulose by diffusion and then attack cellulose. Thus the lower value of kc for the present high lignin content pulp means that the effective radical yield per amount of lignin removed is smaller at larger lignin content. This may be explained that at higher lignin content a larger fraction of radicals are captured by lignin due to the higher residual concentration of lignin in the pulp. In other words it would be reasonable to assume that kc is inversely proportional to the residual lignin concentration.
Model 2: Initial cellulose degradation by alkali then radicals If this dependency of Kc on lignin content is introduced in equation (8) and it is assumed that the alkaline hydrolysis term in equation (8) can be neglected, then the second model is obtained described by equation (10): (10) In equation 10 we have replaced dmn/dt by d(1/DP)/dt, a more fundamental description of the cellulose degradation, and Kc is now the new rate constant of radical attack which is independent of the lignin concentration, and thus also applicable to pulps with different initial lignin content. Integration of equation (10) leads to: (11)
where DPs is the degree of polymerization of cellulose at the start of oxygen delignification when the pulp has reached the operating temperature. Based on equation (11) the natural logarithm of the residual lignin at time t, Lc, is plotted against 1/DP in Figure 6 with the different operating parameters as variable. It is evident from figure 6 that all the data are well described by straight lines and that these lines have about the same slope. Actually, the data at 95 oC obtained at different alkali concentrations as well as at different oxygen pressures collapse to a single straight line within experimental error. However, the data at 100 and 105 °C are represented by two different approximately straight lines with the same slope but shifted further to the right at higher temperature.
10
a: effect of alkali
2.4
10 bar O2+ 2.2 g NaOH/L
95 oC+ 10 bar O2
2.0 Ln Lc
b: effect of temperature
2.4 2.0
1.6
1.6
1.2
1.2 1.1 2.2 3.3 Origin pulp
0.8 1.5 2.4
95 oC 100 oC 105 oC Origin pulp
0.8
1.8 2.1 2.4 2.7 c: effect of pressure
2.1
3.0
3.3
1.5
1.8
2.1 2.4 2.7 1/DP*104
3.0
3.3
95 oC+ 2.2 g NaOH/L
Ln Lc
1.8 1.5 1.2 5 bar 7 bar 10 bar Origin pulp
0.9 0.6 1.5
1.8
2.1 2.4 1/DP*104
2.7
3.0
3.3
Fig. 6 Natural logarithm of kappa number versus 1/DP with [NaOH], Temperature and Oxygen pressure as parameter The results in Figure 6 are in agreement with equation (11) with Kc being a constant not affected by [NaOH], temperature, oxygen pressure or delignification time. However the shift to the right of the data in Figure 6b suggests that higher temperatures lead to additional cellulose degradation. This could be interpreted as alkaline attack on labile groups on cellulose which become more accessible at higher temperatures. However, after this increased initial alkaline attack the cellulose degradation is only dependent on the oxygen-based radicals generated by phenolic delignification. Therefore DPs in equation (11) was represented by DPs= DPo –α D
o
loss in DP0 ue o alkali e a ack
o
he value for α is o ly e e
e
whereby α is a constant that quantifies the initial e
oxygen pressure. Thus the final model has two adjustable parameters; Kc a
era ure bu
o o [
-
] concentration and
α where he for er is i e e
e
of all
o era i g variables, while α is only a function of temperature. The values of the two adjustable parameters were determined by least square error fitting of the data with equation (11). They are a value for Kc of 7.48 x 10-5 a
α of 0.098, 0.21 and 0.31 at 95, 100 and 105 °C respectively. The comparison
of the model predicted and experimental DP values are presented in figure 7 for all the present experiments. It can be seen that the agreement is extraordinary good, providing very strong support for the soundness of this cellulose degradation model during oxygen delignification. It is interesting to note the strong and approximately linear increase in α wi h i creasing temperature. This suggests that temperatures above 95 °C should be avoided to minimize the additional initial cellulose degradation during oxygen delignification of high kappa Kraft pulp unless this can be mitigated by the presence of viscosity protectors or catalysts. This is the topic of our further research. Unfortunately without these additional measures a lower temperature decreases the delignification rate and this is especially problematic for high kappa pulps which require a high degree of delignification prior to entering the bleach plant.
11
Simulated DP
7000
y=x R² = 0.97
6000 5000
Kc= 7.48
4000
0.098 @ 95 oC 0.21 @100 oC 0.31 @ 105 oC
3000 3000
4000 5000 6000 Measured DP
7000
Fig. 7 Predicted versus measured DP.
Another aspect which is different in practical oxygen delignification compared to the present flow-through experiments is the presence of dissolved lignin in the former. In order to establish the effect of dissolved lignin on delignification and cellulose degradation, two additional Berty FT experiments were performed with post-oxygen wash filtrate obtained from the mill which also had supplied the Kraft pulp. Similarly as with the oxygenated alkaline water FT experiments, the filtrate solutions were saturated with oxygen at 10 bar and their alkalinity adjusted to 2.2 g NaOH/l. The TOC (total organic carbon) content in the filtrates was 6400 mg/L. Figure 8 compares the selectivity of Odelignification in the FT reactor obtained with alkaline water and post-oxygen wash filtrate solutions. Figure 8 clearly shows a lower viscosity at slightly increased kappa number for the filtrate solution compared to the alkaline water solution. This suggests that additional radicals are generated by the high concentration of dissolved organic compounds in the filtrate which lead to increased net attack on cellulose, while also capturing some of the radicals generated by phenolic degradation which attack residual lignin in the pulp.
915
Water Post-oxygen washing filtrate (TOC : 6400 mg/L)
Viscosity (mL/g)
900 885
95 oC
870 855
100 oC
2.2 g NaOH/L; 10 bar O2 60 min
840 18
19
20
21 Kappa
22
23
24
Fig. 8 Selectivity of O-delignification in the flow-through reactor with alkaline water and with alkaline oxygenated post-oxygen washing filtrate( initial kappa 65 and viscosity of 1295 mL/g)
12
Conclusion: The kinetic models of delignification and carbohydrates degradation of a high kappa (65) softwood pulp were obtained in a flow-through oxygen delignification set-up. The delignification kinetics were well represented by a power-law model with 3.5-order in residual lignin. The high lignin reaction order is very different from the first order in lignin kinetics for a standard kappa (24) softwood pulp reported in literature using a similar reactor. The higher lignin order for the present high kappa pulp is likely due to the different structure of the residual lignin compared to that in conventional Kraft pulp. However, the reaction orders in alkali and oxygen pressure as well as the delignification activation energy are similar for the high kappa and conventional Kraft pulps. The cellulose degradation kinetics were described by different models. The best cellulose model assumes that the cellulose degradation is proportional to delignification due to attack of cellulose by oxygen based radicals generated via the phenolate anion lignin degradation pathway. The novelty of this model is that it also assumes that the fraction of the generated radicals which reach and attack cellulose is inversely proportional to the amount of residual lignin. In another words, the higher lignin content provides protection to cellulose in softwood pulp towards degradation and explains the higher delignification/cellulose degradation selectivity of higher lignin content pulp. This model leads to surprisingly simple kinetics and extraordinary good prediction of the cellulose degradation data, providing very strong support for the soundness of the model. The model also shows that oxygen delignification above 95 °C needed to obtain a bleachable grade pulp from high kappa Kraft pulp requires additional protection of cellulose against degradation. In addition, the presence of dissolved lignin generates more radicals resulting in additional cellulose attack which must be considered.
Acknowledgements The Finnish Bioeconomy (FIBIC) is acknowledged for financial support. We would also like to thank Dr. Kari Kovasin for all his support and positive attitude toward this project.
13
References Agarwal, Genco J, M.Cole B, and W.Miller W (1999) Kinetics of Oxygen Delignification. Journal of Pulp and Paper Science 25, 361-6. Backstrom M and Jensen A (2001) Modified kraft pulping to high kappa numbers. Appita Journal 54, 203-9. Chandranupap P and Nguyen KL (1998) Kinetics of chlorine dioxide delignification of eucalypt kraft pulps. Pulping Conf. 1, 365-73. Danielewicz D and Surma-Ślusarska B 2006 Textiles in Eastern Europe 14, 89-93.
yge
elig ifica io of high-kappa number pine kraft pulp Fibres and
Dogan I and Guruz G (2008) Dimensionless Parameter Approach for Oxygen Delignification Kinetics. Ind. Eng. Chem. Res. 47, 5871-8. Hart P and Connel D (2006) The effect of digester kappa number on the bleachability and yield of EMCCTM softwood pulp Tappi Journal 67, 127-37. Hsu CL and Hsieh. JS (1988) Reaction kinetics in oxygen bleaching. AIChE Journal 34, 116-22. Iribarne J and Schroeder L (1997) High-Pressure Oxygen Delignification of Kraft Pulps,1: Kinetics. Tappi Journal 80, 134-250. Iribarne J and Schroeder LR (1996) High-pressure oxygen delignification of kraft pulps. II. Simulation. Int. Pulp Bleaching Conf. 1, 81-8. Jafari V, Labafzadeh S, King A, Kilpeläinen I, Sixta H, and van Heiningen A (2014a) Oxygen delignification of conventional and high alkali cooked softwood kraft pulps, and study of the residual lignin structure. RSC Advances DOI: 10.1039/c4ra00115j, Jafari V, Sixta H, and van Heiningen A (2014b) Kinetics of oxygen delignification of high-kappa pulp in a continuous flow through reactor. Submitted to Industrial & Engineering Chemistry Research Jafari V, Sixta H, and van Heiningen A (2013) Multistage oxygen delignification of high-kappa pine kraft pulp with peroxymonosulfuric acid (Px). Holzforschung 10.1515/hf-2013-0148, Janson J (1970) Calculation of polysaccharide composition of wood and pulp. Paperi Ja Puu 52, 323-6, 8-9. Järrehult B and Samuelson. O (1987) Oxygen bleaching of kraft pulps at low consistency. Svensk. Papperstidn 81, 533– 40. Ji Y (2007) Kinetics and mechanism of oxygen delignification[Thesis]. Type, Ji Y, Vanska E, and van Heiningen A ( 2009) New kinetics and mechanisms of oxygen delignification observed in a continuous stirred tank reactor. Holzforschung 63, 264-71 Ji Y, Wheeler C, and van Heiningen A (2007) Oxygen delignification kinetics: CSTR and batch reactor comparison. AIChE Journal 53, 2681-7. Ji Y, Wheeler MC, and van Heiningen A (2006) O2 delignification kinetics from CSTR and batch reactor data. AIChE Annu. Meet., Conf. Proc. 673e/1-e/14. Kleppe PJ, Chang H-M, and Eckert RC (1972) Delignification of high yield pulp with oxygen and alkali. I. Preliminary studies on southern pines. Pulp & Paper Magazine of Canada 73, 400-4. Lindholm CA (1986) Effect of heterogeneity in pulp bleaching with ozone. Paperi Ja Puu 68, 283-4, 6-8, 90.
14
Luo X, Hu H, Chai XS, Cao S, Huang L, and Chen L (2012) Improving bleached pulp yield and paper strength properties of eucalyptus through integrating kraft pulping to high kappa number and oxygen delignification. J. Biobased Mater. Bioenergy 6, 531-7. Nguyen KL and Liang H (2002) Kinetic model of oxygen delignification. Part 1 - effect of process variables. Appita Journal 55, 162-5. Olm L and Teder A (1979) The kinetics of oxygen bleaching. Tappi Journal 62, 43-6. Parsad B, Gratzl J, Kirkman A, Jameel H, Rost T, and Magnotta V (1994) High-kappa pulping and extended oxygen delignification decreases recovery cycle load. Tappi Journal 77, 135-47. Parsad B, Kirkman A, Jameel H, Tratzl J, and Magnotta V (1996) Mill Closure with High-Kappa Pulping and Extended Oxygen Delignification. Tappi Journal 79, 144-52. Pekkala O (1984) Reactivity of kraft pulp components in some chemical treatments. Part 2. Alkaline treatments of pulps after prolonged kraft cooking. Paperi Ja Puu 66, 229-34, 37-8, 41-2. Rubini BR and Yamamoto CI (2006) Development of predictive oxygen delignification models using kinetic expressions and neural networks. Tappi Journal 5, 3-6. Ruuttunen K and Vuorinen T (2005) Developing Catalytic Oxygen Delignification for Kraft Pulp: Kinetic Study of Lignin Oxidation with Polyoxometalate Anions. Ind. Eng. Chem. Res. 44, 4284-91. Schoeoen. and Nils H (1982) Interpretation of rate equations from kinetic studies of wood pulping and bleaching. Sven. Papperstidn. 85, R185-R93. Tao LM, Joseph G, Barbara JWCRC, and Fort JR (2011) Selectivity of oxygen delignification for southern softwood kraft pulps with high lignin content. Tappi Journal 10, 29-39. Teder A and Olm L (1981) Extended delignification by combination of modified kraft pulping and oxygen bleaching. Paperi ja Puu 63, 315-18, 21-2, 25-6. Tran AV (2004) The two-stage oxygen delignification of hardwood kraft pulp. Appita Journal 57, 440-3. van Heiningen A and Ji Y (2012) Southern pine oxygen delignified pulps produced in a Berty throughflow reactor: How to obtain the highest degree of delignification while maintaining pulp yield and quality Tappi Journal 11 9-18. Vehmaa J, Engstroem J, and Piira J (2012) Method and arrangement for treating filtrate after oxygen delignification of chemical pulp cooked to a high kappa number. ISSN-2012-FI50272-2012127111, 20120321.
15
