Impact of Kraft Process Modifications on ... - ACS Publications

The impact of effective alkali (EA) splitting and anthraquinone (AQ) addition on Eucalyptus globulus kraft pulping yield and pulp quality was investig...
1 downloads 0 Views 479KB Size
Ind. Eng. Chem. Res. 2008, 47, 7433–7440

7433

Impact of Kraft Process Modifications on Eucalyptus globulus Pulping Performance and Polysaccharide Retention Ana Sofia Santiago and Carlos Pascoal Neto* CICECO and UniVersity of AVeiro, Department of Chemistry, 3810-193 AVeiro, Portugal

The impact of effective alkali (EA) splitting and anthraquinone (AQ) addition on Eucalyptus globulus kraft pulping yield and pulp quality was investigated. AQ addition to kraft pulping enhances the rate of delignification, leading to alkali saving and reducing time at maximum temperature. A 1.9% total yield increase was observed, assigned to the enhancement of both cellulose and xylan retention but also due an increase of rejects content. Contrary to results described in the literature for other woods, the use of an even effective alkali concentration profile throughout the cook of E. globulus did not affect pulp yield or carbohydrate composition of screened pulp. However, EA profiling resulted in a higher pulp viscosity, indicating a lower degradation of cellulose chains along the pulping. Such results were interpreted taking into account the peculiar chemical structure of E. globulus wood components. Introduction The objective of the chemical pulping of wood is the removal of lignin and the separation of wood fibers with minimal damage of the other fiber wall components, namely, cellulose and hemicelluloses. However, current pulping processes, including the kraft process where delignification is carried out in a strongly alkaline solution composed mainly by OH- and HS- ions, remove significant amounts of carbohydrates along with lignin, thus decreasing pulp yield.1 Carbohydrate reactions that mainly affect kraft pulp yield include peeling, alkaline hydrolysis of glycosidic bonds, and hydrolysis of acetyl groups in hemicelluloses.2 In addition to these chemical phenomena, physical dissolution and, in some cases, reprecipitation of hemicelluloses are relevant for kraft yield.1 The increase of the wood pulping process yield is a major goal of a chemical pulp mill, allowing increased production and reduced specific wood costs. Process modifications aiming to increase kraft pulp yield may be implemented in order to retard or avoid the carbohydrate degradation reactions, to increase delignification rate, or to increase mass transfer phenomena.3 Among the process modifications with great impact in pulp yield are the addition of anthraquinone (AQ) and modified chemical charges profiling, together with liquor recirculation along the kraft pulping reactor, enhancing the process selectivity and allowing extending delignification without decreasing pulp yield.3,4 However, the relative impact of such process modifications on pulping performance is highly dependent on the morphology and structure of the wood species, thus requiring a detailed study and optimization for each specific raw material. In the case of Eucalyptus globulus, an important wood source for paper production with a growing interest and use worldwide, a great research effort has been made in order to assess the chemical structure of this wood and to understand its performance in kraft pulping.5-10 Despite the existence of some studies on the kinetics and generic aspects of kraft pulping of E. globulus,11-13 there is a lack of systematic studies on modified pulping processes as well as on the impact of the pulping performance and pulp properties. AQ constitutes a technically successful redox alkaline pulping catalyst.4 It is stable on hot and strong alkaline solutions and * To whom correspondence should be addressed. Tel.: +351234370693. Fax: +351234370084. E-mail: [email protected].

may catalyze the degradation of lignin and, simultaneously, stabilize carbohydrates. Besides this, typical pulping AQ dosages have minimal adverse environmental effects. AQ is insoluble in the cooking liquor, having no catalytic function at the beginning of the pulping process; however, as the reactor temperature increases, AQ solubilization occurs by electrochemical reduction by aldehyde groups on polysaccharide chain ends, converting them to carboxylic acid groups. This transformation may stabilize the polysaccharides against peeling reactions, while the AQ molecule converts it to a reduced form, probably anthrahydroquinone, AHQ. On the other hand, AHQ reacts with transient reactive lignin structures and, at the same time, is regenerated to the original AQ form.14 This redox cycle is responsible for the small AQ charges required to enhance both the delignification rate and pulp yield.15 Although AQ was already introduced in the pulp-and-paper industry in the seventies,16 there are only a few studies concerning its effect on eucalypt kraft pulping. The use of AQ as a pulping catalyst to increase the pulp yield of kraft and modified pulping of Eucalyptus was the subject of some works, carried out mainly by Brazilian researchers.17,18 The application of AQ as a strategy to decrease sulfidity level and, consequently, TRS emissions in Eucalyptus kraft pulping was also the subject of several works.19-21 The kraft pulping process may also be modified by using an effective alkali (EA) concentration as low and as even as possible throughout the cook, a sulfide concentration as high as possible in the initial part of the cook and during the transition from the initial to the principal phase of delignification, and by keeping a dissolved lignin concentration as low as possible during the latter part of the cook.22,23 These modifications allow a decrease of pulp lignin content without lowering pulp yield or, for a lignin content similar to that obtained on conventional kraft pulping, afford a higher pulp yield. On the other hand, the decrease of the lignin content of unbleached pulp improves the economy and reduces the environmental impact of chlorine dioxide based bleaching processes, due not only to the decrease of the required amount of chlorine dioxide for the bleaching but also to the reduction of the amount of organochlorine compounds in the bleaching effluents. Such reasons explain the widespread acceptance of modified kraft process technologies in the pulp-and-paper

10.1021/ie071488g CCC: $40.75  2008 American Chemical Society Published on Web 08/27/2008

7434 Ind. Eng. Chem. Res., Vol. 47, No. 19, 2008 Table 1. E. globulus Wood Composition (% on o.d. Wood) extractives (ethanol/toluene) lignin neutral anhydrous monosaccharides glucose xylose rhamnose arabinose mannose galactose

1.3% 21.5% 54.1% 15.0% 0.3% 0.4% 1.2% 1.5%

industry in the past few years. However, in spite of this interest in modified pulping technology, there is still a lack of knowledge, namely, on the effect of alkali profiling on E. globulus pulping behavior. The aim of the present work is to investigate the effect of anthraquinone addition and effective alkali profiling on E. globulus kraft pulping performance and polysaccharides retention. Experimental Section

Figure 1. Effective alkali and temperature profiles of kraft cookings.

Wood Chips. Eucalyptus globulus industrial chips were classified according to the standard method SCAN-CM 40:88 on a laboratory screen. Only accepted chips were used, and all visually observed pieces of bark and knots were removed by hand sorting before pulping. The chemical composition of E. globulus wood is presented in Table 1. The quantification of wood extractives in ethanol/ toluene solution was made according to the Tappi T204 cm-97 standard method, with minor modifications. Kraft Pulping Experiments. Three types of kraft pulping experiments were carried out: (i) standard kraft cooking (kraft std, for short), performed with total chemical charge introduced in the beginning of the cook; (ii) EA profiling kraft cooking (EA profiling, for short), performed with four EA charge addition points (see Figure 1); and (iii) kraft cooking with AQ addition (kraft + AQ, for short), also performed with total chemical charge introduced in the beginning of the cook. All the pulping experiments were carried out with 1 kg of oven dry (o.d.) wood and using a liquor-to-wood ratio equal to 4 L/kg o.d. wood chips which were cooked in a 6.5 L laboratory batch digester with liquor circulation (M/K system, Inc.). The temperature profile, shown in Figure 1, was the same for all kraft pulping experiments and involved the rise of temperature from 40 to 120 °C in about 45 min, keeping at 120 °C for 45 min (impregnation phase), heating to 160 °C (cooking temperature) in about 15-20 min, and keeping at that temperature for the time necessary to reach the required degree of delignification of pulp (kappa number of 14). All the pulping experiments were performed with constant total active alkali charge (AA) of 14% as Na2O and sulfidity of 30%. Therefore, the total effective alkali charge was also constant and equal to 11.9%, expressed as Na2O. Synthetic white liquors were prepared from analytical grade reagents and were analyzed by ABC titration (SCAN-N2:88). EA profiling cooking was performed with four liquor additions: beginning of the cook (65% of total EA charge); ending of the impregnation phase (15% of total EA charge); when temperature reached cooking temperature (15% of total EA charge); and after 30 min at cooking temperature (5% of total EA charge). The total Na2S charge was introduced at the beginning of the pulping experiment along with the first EA charge. The liquor volume corresponding to the first addition was such that the liquor-to-wood ratio was equal to 4 L/kg. The other three additions of EA charge along the cooking were made with concentrated liquor, corresponding each to 100 mL, in order to keep the liquor-to-wood ratio roughly constant.

In the kraft + AQ experiments, AQ powder was dispersed in cooking liquor and was added in the beginning of the pulping experiment. The AQ charge was equal to 0.1% on o.d. wood. The three types of kraft pulping experiments were interrupted at times corresponding to the EA charge addition points of the EA profiling cook (see Figure 1). Partially delignified wood, issued from interrupted cooks, was washed until the washing water conductivity was similar to that of tap water. After that, partially delignified wood was centrifuged for about 15 min and weighted, and the moisture content was determined in order to obtain total yield values. The partially delignified wood was then milled in a hammer mill (Restsch SK1) and sieved to 40-60 mesh fraction used for Klason lignin determination and carbohydrate analysis. Black liquor samples were collected for residual alkali concentration determination24 and for precipitation and isolation of black liquor dissolved polysaccharides, BLP, which were characterized in terms of neutral monosaccharides. Polysaccharide yield was calculated by subtracting lignin content of partially delignified wood from total yield. Cooked chips were washed and disintegrated, and the resulting pulp was screened through a 0.15 mm slot flat screen. Rejects were retained on the screen, and its o.d. weight was determined. Kappa number (parameter related with residual lignin content in pulp) and intrinsic viscosity (parameter related with the degree of depolymerization of pulp polysaccharides) were determined using standard methods (ISO 302-1981 and SCAN-cm 15:88, respectively). Screened yield is defined as the ratio between the mass of o.d. screened pulp and the mass of o.d. initial wood. Total yield is obtained by the sum of screened yield and rejects content. Polysaccharide yield of final pulps was calculated on a screened yield basis (the polysaccharides of rejects were not considered). The kraft pulping terminology used in this work was defined according to TAPPI Standard T1203 os-61: Active alkali charge, AA, defined as [mass(NaOH) + mass(Na2S)]/mass(o.d. wood); Effective alkali, EA, charge defined as [mass(NaOH) + 1/2mass(Na2S)]/mass(o.d. wood); and sulfidity, S, defined as mass(Na2S)/ [mass(NaOH) + mass(Na2S)]. All NaOH and Na2S masses are expressed as equivalent masses of Na2O: mass(NaOH as Na2O) ) mass(NaOH) · 0.775 and mass(Na2S as Na2O) ) mass(Na2S) · 0.795. Precipitation and Isolation of Black Liquor Dissolved Polysaccharides. Black liquor dissolved polysaccharides were isolated following a procedure described by Engstro¨m et al.,

Ind. Eng. Chem. Res., Vol. 47, No. 19, 2008 7435 Table 2. Summary of Pulping Results

kraft std EA profiling kraft + AQ

time at Tmax (min)

Kappa no.

total yield (% o.d. wood)

screened yield (% o.d. wood)

rejects (% o.d. wood)

viscosity (mL/g)

EA consumption (g of Na2O/kg of o.d. wood)

80 80 40

14.1 14.0 14.1

56.8 57.0 58.7

56.1 56.0 57.2

0.7 1.0 1.5

1420 1518 1390

105.5 104.0 97.5

with minor modifications.25 About 200 mL of 1,4-dioxane was slowly added to 100 mL of black liquor, followed by addition of glacial acetic acid until a pH ∼ 5 was reached. The solution was kept at 5 °C for two days, and after that time, the precipitated black liquor polysaccharides were separated by centrifugation and the clear solution decanted off. The precipitate was sequentially washed with 1,4-dioxane-water solution (2: 1), 1,4-dioxane, methanol, and acetone. The residue designated as black liquor dissolved polysaccharides, BLP, was dried under vacuum with phosphorus pentoxide. Carbohydrate Analysis. Wood, partially delignified wood, final pulps, and black liquor dissolved polysaccharides (BLP) were submitted to Saeman hydrolysis. The neutral monosaccharides were quantified by gas chromatography as alditol acetates.26 Lignin Content. Total lignin content of wood and partially delignified wood includes Klason and soluble lignin wood contents. Klason lignin was determined according to the Tappi T 222 om-98 standard, while soluble lignin was determined according to the useful method of Tappi 250, with minor modifications. GPC Analysis. The BLPs were dissolved in N,N-dimethylacetamide (DMA) containing roughly 10% LiCl and further diluted with DMA to a BLP concentration of about 4 mg/mL. The GPC analysis of BLP was performed on a PL-GPC 110 apparatus with an RI detector (Polymer Laboratories, Ltd.) using two Plgel 10 µm MIXED B 300 mm - 7.5 mm columns protected by a Plgel 10 µm precolumn. The temperature of the precolumn and columns, the injector, and the detector was kept constant at 70 °C. The eluent was DMA with 0.1 M LiCl at a flow rate of 0.9 mL/min. The calibration of the GPC columns was made with pullulan reference materials (Polymer Laboratories, Ltd.). Results and Discussion Three different types of pulping experiments were carried out, using the same total effective alkali charge (EA ) 11.9% as Na2O): (i) a standard kraft cooking; (ii) a kraft cooking with EA charge profiling, performed with EA charge splitted along the cooking; and (iii) standard kraft cooking with the addition of anthraquinone in the beginning of the cook. The effective alkali concentration profiles of the three types of pulping are shown in Figure 1. EA concentration profiles of standard and AQ cookings are very similar. On the other hand, as expected, the effective alkali concentration of the EA profiling cooking is significantly lower until the temperature reaches 160 °C; after that, the EA concentration profile is similar to the standard and AQ cookings. At the end of pulping, total alkali consumptions of standard and EA profiling cookings are similar and slightly higher than for the AQ cooking (see Table 2). The major results from the three different pulping experiments are shown in Table 2. Time at maximum temperature was adjusted in order to have pulps with the same delignification degree, corresponding to a kappa number of about 14. The delignification rate of AQ pulping is faster, and thus, the shorter time at maximum temperature led to an alkali saving of 8%, when compared to the AE charge consumption of the kraft std cook.

Results from Table 2 demonstrate that the addition of AQ to kraft pulping leads to an increase of both screened yield and rejects. The rejects increase may be explained by the reduced time at maximum temperature hindering the diffusion of pulping liquor into larger or denser wood chips or knots. On the other hand, effective alkali profiling does not affect either total or screened yield, in agreement with our previous results on profiling kraft cooking of E. globulus.27 Delignification. The delignification curves of the three types of pulping experiments are presented in Figure 2. The dissolution of lignin along the cooking may be divided into the three delignification phases typically observed in wood kraft pulping.1 The initial phase of delignification coincides with the impregnation phase. Only 19-22% of total lignin is removed, while about 50% of total EA consumption is used in this phase (see Figure 1). It is known that the main responsible reactions for the high EA consumption during the initial phase of pulping are not lignin reactions but others including the hydrolysis of the acetyl groups in the wood, solubilization/neutralization of wood extractives, neutralization of acidic degradation products from polysaccharides, and solubilization of low molecular weight polysaccharides.28 Above 120 °C, the delignification rate increases remarkably and remains high during this principal, or bulk, delignification phase, until about 90% of lignin has been removed. During this period, the delignification selectivity reaches the maximum value. The final phase, called the residual phase, is characterized by a slow delignification rate. The EA profiling cook performed with a lower EA initial concentration leads to a lower lignin removal in the initial stages of the process (Figure 2). However, when the temperature reaches the maximum value, the rate of delignification increases, and the EA profiling cook achieves, within the same lapse of time, a delignification degree similar to that obtained in the standard kraft cooking (∼95%). Our results show that the catalytic effect of AQ on delignification rate is not detected until temperature reaches 160 °C, becoming noticeable only when the maximum cooking temperature is reached. As Figure 2 shows, at 160 °C the delignification rate of AQ kraft cooking is higher than for the other two pulping experiments, allowing us to obtain the targeted

Figure 2. Delignification curves of kraft cookings.

7436 Ind. Eng. Chem. Res., Vol. 47, No. 19, 2008

Figure 3. Lignin-free yields along the pulping processes.

delignification degree (95%) in half the time (at 160 °C) required in standard and EA profiling cooks. This result is in agreement with the explanation proposed by Samp and Li,29 based on the soluble and insoluble cycle of AQ. Initially, the insoluble AQ particles are suspended in the cooking liquor or deposited at the chips surface. As the temperature rises and some lignin and polysaccharides dissolve in the black liquor, the reduction of AQ begins by the reaction with reducing end units of dissolved carbohydrates leading to soluble AHQ (according to the authors, AHQ concentration in the liquor reaches its maximum value at 120 °C). The mobile AHQ molecules can then diffuse into the chips and there react with lignin, being oxidized to AQ again. As more AHQ diffuses into the chips, more pronounced is its catalytic effect, and more AQ remains in wood chips. This retained AQ may react with reducing end groups of carbohydrates in the solid cell wall, leading to their higher retention in the wood matrix. Our results also show that in the impregnation phase and rise temperature period (120-160 °C) AHQ molecules, formed by AQ reduction, are probably diffusing into the chips, and therefore, there is no AQ catalytic effect until 160 °C. Only at the maximum temperature period AHQ reacts with lignin structures, in a way similar to hydrosulfide toward the cleavage of phenolic β-aryl ether bonds,14 enhancing delignification rate (see Figure 2) and promoting, at the same time, the regeneration of the original form of AQ. Carbohydrate Retention on the Fibrous Matrix along the Pulping Processes. The three types of kraft pulping experiments were interrupted at times corresponding to the EA charge addition points of the EA profiling cook (see Figure 1). Partially delignified wood, issued from interrupted cooks, was exhaustively washed and then milled and sieved in order to be used for total lignin (Klason plus acid soluble lignin) determination and carbohydrate analysis. Results of lignin-free yield are presented in Figure 3. Figure 3 shows that EA profiling cook provides a lignin-free yield higher than kraft std for the stages performed with a lower EA charge (end of impregnation, with 65% of total EA charge added, and T ) 160 °C, with 80% of total EA charge added). For the experiment interrupted after 30 min at maximum temperature, when 95% of total EA charge was added, the lignin-free yield of EA profiling cook is no longer higher but rather similar to the kraft std one. Therefore, it may be concluded that EA profiling has no effect on polysaccharide yield increase. In addition, results from Table 2 indicate that this cook produces higher rejects content than the kraft standard as a consequence of a lower EA charge during the initial phase of pulping.

As was expected, Figure 3 also shows that the effect of AQ on the polysaccharide stabilization is notable only at maximum temperature period. The higher delignification rate of this cook allowed shortening of pulping time at maximum temperature, which also contributed to the pulp free lignin yield increase of 1.1%. Carbohydrate composition (calculated as homopolysaccharides) of partially delignified wood and screened pulps was normalized in order to have the sum of total polysaccharides and lignin equal to 100%. This approach considers that the major part of extractives, acetyl groups, and uronic acids was completely removed during the impregnation phase. Glucose and xylose contents of the samples are presented in Figure 4. Figure 4A shows that total glucose removal during the pulping processes was about 15-16% of total glucose content of wood, corresponding to 8-9% of the total wood weight. Native cellulose in wood is organized in a fibrilar assembly where crystalline regions alternate with amorphous zones, possessing an average high degree of polymerization and a relatively low accessibility into the crystalline domain.2 During the impregnation phase, glucose removal constitutes about 3-4% of the total wood weight, corresponding to 35-50% of total glucose loss. Removal of glucose during this initial phase may be essentially attributed to dissolution of some easily accessed amorphous regions and, also, to dissolution and/or peeling reactions of glucan chains. A recent study of our research group indicates that glucans represent the second most abundant polyoses after glucuronoxylan in E. globulus wood, constituting about 4.5% of the total weight of wood. These glucans are comprised essentially by linear (amylose, 20%) or ramified at O-6 (amylopectin, 80%) R-(1f 4)-linked glucans.30 Another study based on E. globulus kraft black liquor analysis indicated the presence, since the initial phase of pulping, of significant amounts of these glucans.31 Therefore, it is expected that glucans constitute the foremost responsibility for the high removal of glucose during this initial pulping phase. However, in latter stages of the pulping process, glucose removal should be mostly associated with cellulose degradation. Throughout the temperature rise period (B to C, Figure 4A), there are no significant losses of glucose. On the other hand, when the pulping temperature is reached (C to D, Figure 4A), glucose removal is significant for all pulping experiments, constituting about 35-50% of the total glucose loss. The glucose losses during the maximum temperature period may be essentially attributed to secondary peeling reactions of cellulose chains that were subject to random alkaline hydrolysis.32 During

Ind. Eng. Chem. Res., Vol. 47, No. 19, 2008 7437

Figure 4. Carbohydrate composition of partially delignified woods and screened pulps of the pulping processes. Table 3. Carbohydrate Composition of Final Pulps with Kappa Number ∼ 14 (% on o.d. Wood)

kraft std EA profiling kraft + AQ

rhamnose

arabinose

mannose

galactose

xylose

glucose

screened yield

0.1 0.1 0.1

0.1 0.1 0.2

0.1 0.1 0.2

0.2 0.2 0.4

8.8 8.9 9.2

45.7 45.5 46.0

56.1 56.0 57.2

this period, the major glucose loss occurs for EA profiling pulping which is related with the addition of the third EA charge. Cellulose loss during the remaining time at maximum temperature (D to E, Figure 4A) is reduced and accounts for 8-12% of the total glucose loss. These minor losses may be attributed to the fairly small (∼6 g of Na2O/L) pulping liquor EA concentration. On the other hand, the termination of the peeling reaction of cellulose chains may also occur by physical stopping.33 In the final period of pulping, the amorphous cellulose regions are much less abundant, and the physical inaccessibility of the crystalline regions of cellulose contributes to its retention in the fiber wall. The main hemicellulose component of E. globulus, as for other hardwoods, is the O-acetyl-4-O-methylglucuronoxylan, commonly designated as xylan, which comprises a partially acetylated xylan backbone and methylglucuronic acid, MeGluA.2 Xylan is more sensitive to alkali than cellulose due to its amorphous state, accessibility to the pulping liquor, and low polymerization degree. Results from Figure 4B indicate that the major xylan losses take place during the impregnation phase (A to B, Figure 4B), corresponding to 17-23% of total xylan content in wood, and that xylan removal is dependent on EA concentration. Results published elsewhere have, however, demonstrated that for EA charges higher than ∼15% at the impregnation phase xylan retention on wood was roughly constant.27 Xylan losses throughout the temperature rise period (B to C, Figure 4B) and the first 30 min at maximum temperature (C to D, Figure 4B) are significant. During these periods, xylan loss occurs mainly through alkaline hydrolysis and consequent secondary peeling which, due to the lower polymerization degree of hemicelluloses, may originate sufficiently small chains that become readily soluble in the aqueous alkaline medium, as will be discussed later in this work.

During the remaining time at maximum temperature (D to E, Figure 4B), the xylan content is roughly constant, with only 4-6% of the total xylan loss occurring during that period. The roughly constant xylan content during this phase, as well as the high total retention of this polysaccharide in screened pulp (59-61% of total xylan content) when compared to other woods,7 could be attributed, at least partly, to the structural features of Eucalyptus xylan. The main hemicellulose of Eucalyptus are linked to other cell wall components (namely, glucans and rhamnoarabinogalactans9) through 4-O-methylglucuronic acid moieties substituted at O-2 with galactose or glucose, 8,9 hindering its dissolution on pulping liquor and, also, preventing the isomerization of the reducing moieties that precede the peeling reaction. On the other side, other xylan substituents at O-2, namely, terminal methylglucuronic acid and hexenuronic acid moieties, also minimize the peeling reaction for the same reason. In addition, xylan of E. globulus has a high average molecular weight when compared to softwoods and other hardwoods, which also contributes to the high hemicellulose retention.6,7 Carbohydrate Composition of Pulps. Detailed carbohydrate compositions of final pulps calculated on a screened yield basis are shown in Table 3. The carbohydrate compositions of kraft std and EA profiling pulps are very similar, meaning that alkali splitting had no effect either on xylan or on cellulose pulp retention. However, the viscosity values of final pulps presented in Table 2 indicate that the EA profiling pulp has a 100 unit higher viscosity than the kraft std. Since this increase for EA profiling pulping is not derived from higher cellulose retention, it is expected that cellulose chains are more preserved in this pulp. Previous works concerning the impact of EA profiling concentration on carbohydrate dissolution of birch wood have demonstrated that lowering the EA concentration in the impregnation and early cooking stages of the cook significantly

7438 Ind. Eng. Chem. Res., Vol. 47, No. 19, 2008

increased both total yield and viscosity, which was mainly attributed to higher cellulose retention.34,35 However, in our work, no differences were pointed out concerning carbohydrate composition and screened yield between kraft std and EA profiling pulps of E. globulus. At this point, it is important to emphasize that the total effective alkali used in the present work was very low (11.9% as Na2O, compared to EA charges of 15.5% and higher used in the above-mentioned works). In fact, our results presented elsewhere show that when the cook is performed with a higher EA charge (14.5%) with the same EA split (65%/15%/15%/5%) there is a slight increase of total yield (0.5%).36 The use of low EA charge is assigned to the peculiar lignin of E. globulus wood which, due to its high syringyl/guaiacyl unit ratio, high amount of β-O-aryl ether structures, and low degree of condensation, is extremely reactive in the kraft pulping medium.7,10,37 So, it may be concluded that when a low EA charge is used, the effective alkali splitting has no effect on E. globulus polysaccharide retention. AQ screened pulp presents a screened yield higher than final kraft pulp (57.2% vs. 56.1%), both xylan and cellulose retention being the major contributors (∼35% and ∼27%, respectively) for that increase (see Table 3). The higher retention of xylan in AQ final pulp is also coherent with the higher retention of arabinose and galactose which are monomer constituents of rhamnoarabinogalactans, appointed as one of the polysaccharides linked at xylan by O-2 of its 4-O-methylglucuronic acid moieties.8,9 Carbohydrate composition of AQ pulp, presented in Table 3, is in agreement with the results obtained by other authors that have demonstrated that total yield increases with AQ addition, both the cellulose and hemicellulose contents being responsible for that pulp yield increase.38 The viscosity values of kraft standard and AQ pulps, presented in Table 2, are similar. A higher viscosity value of AQ pulp could be expected from the higher cellulose retention on this pulp. This is not observed, certainly, because of the higher xylan content in this pulp (xylan possesses low molecular mass when compared to cellulose molecular mass) and the possible increase of random hydrolysis of glycosidic bonds of cellulose, as a result of oxidation by AQ of the primary and secondary alcohol groups to carbonyl groups.39 It should be noted that this last hypothesis does not meet consensus, and some authors defend that reducing additives does not affect pulp viscosity.40,41 Black Liquor Dissolved Polysaccharides. Polysaccharide fraction of BLP was composed mainly by xylan since xylose mass, after hydrolysis, represented about 78-84% of total polysaccharide mass fraction. This result is coherent with other studies with birch, which have demonstrated that an appreciable portion of the dissolved xylan appears in the cooking liquor as oligo- or polysaccharides.34,42 Our research group reported the existence of rhamnoarabinogalactan oligomers linked at O-2 of MeGluA residues of dissolved xylans which indicates that dissolved xylan in black liquor keeps some features of its particular structure.31 Glucose content of BLP represented about 10-14% of BLP mass fraction, indicating the presence of some glucans that are resistant to alkaline degradation. Structural analysis of the BLP of E. globulus kraft standard cooks made by Lisboa et al. allowed concluding that amylopectin oligosaccharides are the main source of the BLP glucose content.31 Xylose contents of BLP obtained in different stages of the pulping experiments are shown in Figure 5. Figure 6 presents the xylose content dissolved in black liquor not precipitated as BLP, which was estimated by mass balance: % dissolved xylose

Figure 5. Xylose content of black liquor dissolved polysaccharides (precipitated).

Figure 6. Xylose not precipitated in black liquor (obtained by mass balance).

in BL ) % xylose in wood - % xylose in BLP - % xylose in the solid phase (partially delignified woods or screened pulps). Results depicted in Figure 5 demonstrated that, during the impregnation phase and temperature rise period from 120 to 160 °C, only 10-20% of total removed xylose from wood could be precipitated as black liquor dissolved xylan. Therefore, the major part of removed xylan during the initial stages of pulping is present in black liquor as dissolved or degraded monomers or oligomers. During the first 30 min at maximum temperature, the content of xylose in BLP almost duplicates for all pulping experiments, meaning that alkaline hydrolysis of xylan bonds is intensified during this period. These random degradation reactions promote xylan bond scission, originating xylan chains sufficiently small to be soluble in the alkaline pulping liquor. On the other hand, the xylose dissolved content in BL only slightly increases (see Figure 6), indicating that secondary peeling reactions (which lead to monomer release to solution) are less significant. During the remaining time at maximum temperature (50 min for kraft std and profiling cooks and 10 min for AQ kraft cook), the xylan retention in pulp is roughly constant, and therefore, the contents of both precipitated xylan in BLP and dissolved xylose in black liquor are almost constant for all three types of cookings.

Ind. Eng. Chem. Res., Vol. 47, No. 19, 2008 7439

Figure 7. Average molecular weight (Da) of xylans precipitated from BLP.

In spite of the lower extent of xylan removal from pulp in AQ pulping (see Table 3), Figure 5 shows that the amount of xylan recovered by precipitation in BLP is, trendily, slightly higher than the corresponding amounts found in BLP of kraft std. This fact could be due to AQ reaction with the terminal reducing ends of the dissolved xylan, hampering the occurrence of peeling reactions in the aqueous alkaline medium. On the other hand, the amounts of xylan precipitated in BLP of EA profiling pulping are lower than the corresponding amounts found in BLP of kraft std. This is assigned to the lower EA concentrations along the pulping, leading to lower xylan degradation and dissolution rates from wood (see Figure 4B). Gel permeation chromatography of BLP allowed an estimation of the average molecular weight of dissolved xylan in black liquor for kraft std and EA profiling cooks (Figure 7). Figure 7 shows that the xylan precipitated from black liquor drained off at the end of the impregnation phase of kraft std possesses the highest average molecular weight. This is due to the higher initial EA concentration at the impregnation phase of the standard kraft pulping, thus leading to high xylan dissolution from the wood (see Figure 4B) having, certainly, a higher molecular weight than its counterpart in the EA profiling pulping experiment. In general, the molecular weights of BLP xylan of kraft std are lower than the ones of BLP xylans of EA profiling cook. This observation could be explained by the higher EA concentration of kraft std pulping liquor that intensifies the degradation reactions of dissolved xylans in black liquor. A remarkable feature for both types of cookings is the relative stability of xylans in the alkaline medium. This may be assigned to the improved stability of this xylan due to the reducing terminal fragment constituted by rhamnose and galacturonic acid, as well as by the glucuronic acid moieties on the xylan backbone. The blockage of the OH-2 position by a glycosidic linkage of galacturonic acid and xylose (to rhamnose and 4-O-methylglucuronic acid, respectively) prevents the isomerization of the reducing moieties that precedes the β-elimination reaction (peeling reaction).43 These high molecular weight xylans dissolved in black liquor represent a great potential for the improvement of polysaccharide retention in E. globulus kraft pulping, by selective reprecipitation on fiber surfaces. Conclusions The EA splitting and AQ addition revealed important improvements in the performance of E. globulus wood toward kraft pulping. The use of an even EA concentration profile

throughout the cook did not affect pulp yield or carbohydrate distribution on final pulp. However, EA profiling resulted in a higher pulp viscosity indicating that cellulose chains of this pulp possess a higher polymerization degree. AQ addition to kraft pulping enhanced the rate of delignification, leading to an alkali saving and reduced time at maximum temperature. Also, AQ addition promoted a screened yield increase due to both cellulose and xylan retention improvement. The carbohydrate analysis along the pulping processes allowed concluding that glucose removal is significant at the impregnation phase and at the beginning of the maximum temperature period. Glucans probably constitute the foremost responsibility for the high removal of glucose during the initial pulping phase, while in latter stages of pulping, glucose removal should be mostly associated with cellulose degradation. The xylan removal occurs predominantly during the impregnation phase and temperature rise period. The high retention of this polysaccharide in wood (∼60% of total xylan content) may be attributed to the peculiar features of E. globulus xylan, not found in other hardwood xylans. Black liquor dissolved polysaccharide analysis suggested that random alkaline hydrolysis of polysaccharides on fibers occurs mainly during the initial stages of the maximum temperature period, since the precipitable carbohydrate material increases significantly during this time. Precipitated xylans from black liquor possess high stability in the alkaline medium and have high molecular weight, constituting a high potential for the improvement of polysaccharide retention in E. globulus kraft pulping, by its selective reprecipitation on fibers surfaces. Acknowledgment The authors thank the Portuguese Foundation for Research FCT for the financial support of project POCTI/46124/EQU/ 2002 and PhD grant SFRH/BD/8166/2002, RAIZ for financial support and collaboration in pulping experiments, and Prof. Dmitry Evtuguin for the realization of the GPC analysis. Literature Cited (1) Grace, T.; Leopold, B.; Malcolm, E. Alkaline Pulping, in Pulp and paper manufacture, 3rd ed.; TAPPI/CPPA: Atlanta, GA, 1989; Vol. 5. (2) Sjo¨stro¨m, E. Wood Chemistry. Fundamentals and Applications; Academic Press: London, 1981. (3) Courchene, C. The tried, the true, and the new - getting more pulp from chips - modifications to the kraft process for increased yield. Proceedings of the Breaking the Pulp Yield Barrier Symposium; TAPPI Press: Atlanta, GA, 1998.

7440 Ind. Eng. Chem. Res., Vol. 47, No. 19, 2008 (4) Santiago, A.; Pascoal Neto, C. Assessment of potential approaches to improve Eucalyptus globulus kraft pulping yield. J. Chem. Technol. Biotechnol. 2007, 82, 424–430. (5) Pinto, P.; Evtuguin, D.; Pascoal Neto, C.; Silvestre, A. Behaviour of Eucalyptus globulus lignin during kraft pulping. I. Analysis by chemical degradation methods. J. Wood Chem. Technol. 2002, 22 (2-3), 93–108. (6) Pinto, P.; Evtuguin, D.; Pascoal Neto, C. Structure of hardwood glucuronoxylans: modifications and impact on pulp retention during wood kraft pulping. Carbohydr. Polym. 2005, 60, 489–497. (7) Pinto, P.; Evtuguin, D.; Pascoal Neto, C. Effect of structural features of wood biopolymers on hardwood pulping and bleaching performance. Ind. Eng. Chem. Res. 2005, 44 (26), 9777–9784. (8) Shatalov, A.; Evtuguin, D.; Pascoal Neto, C. (2-O-a-D-Galactopyranosyl-4-O-methyl-a-D-glucurono)-D-xylan from Eucalyptus globulus Labill. Carbohydr. Res. 1999, 320, 93–99. (9) Evtuguin, D.; Toma´s, J.; Silva, A.; Pascoal Neto, C. Characterization of an acetylated heteroxylan from Eucalyptus globulus Labill. Carbohydr. Res. 2003, 338 (7), 597–604. (10) Evtuguin, D.; Pascoal Neto, C.; Silva, A.; Domingues, P.; Amado, F.; Robert, D.; Faix, O. Comprehensive study on the chemical structure of dioxane lignin from plantation Eucalyptus globulus wood. J. Agric. Food Chem. 2001, 49, 4252–4261. (11) Santos, A.; Rodrı´guez, I.; Gilarranz, M.; Moreno, D.; Garcı´a-Ochoa, F. Kinetic modeling of kraft delignification of Eucalyptus globulus. Ind. Eng. Chem. Res. 1997, 36 (10), 4114–4125. (12) Fernandes, N.; Castro, J. Steady State Simulation of a Continuous Moving Bed Reactor in the Pulp and Paper Industry. Chem. Eng. Sci. 2000, 55 (18), 3729–3738. (13) Carvalho, G.; Martins, A.; Figueiredo, M. Kraft pulping of Portuguese Eucalyptus globulus: effect of process conditions on yield and pulp properties. Appita J. 2003, 56 (4), 267–274. (14) Landucci, L. Quinones in alkaline pulping. Characterization of an anthrahydroquinone-quinone methide intermediate. Tappi J. 1980, 63 (7), 95–99. (15) Blain, T. Anthraquinone pulping: fifteen years later. Tappi J. 1993, 76 (3), 137–146. (16) Holton, H. Soda additive softwood pulping: a major new process. Pulp. Paper Mag. Can. 1977, 78 (10), 218–223. (17) Bassa, A.; Sacon, V; Ju´nior, F.; Barrichelo, L. Modified cooking of E. grandis and hibrid E. grandis x E. urophylla. Proceedings of the 35th Annual Pulp and paper Congress & Exhibition; ABTCP: Brazil, 2002. (18) Almeida, J.; Gomide, J.; Colodette, J.; Silva, D. Estudo de alternativas te´cnicas para aumento de rendimento da polpac¸a˜o kraft continua de Eucalyptus. ReVista A´rVore 2000, 24. (19) Silva, F.; Gomide, J.; Colodette, J.; Filho, A. Effect of sulfidity reduction and addition of anthraquinone on pollutant emission and quality of Eucalyptus kraft pulp Proceedings of the 34th Annual Pulp and Paper Congress & Exhibition; ABTCP: Brazil, 2001. (20) Jeroˆnima, L.; Foelkel, C.; Frizzo, S. Anthraquinone addition in the alkaline pulping of Eucalyptus saligna. Cieˆncia Florestal 2000, 10 (2), 31– 37. (21) Yoon, S.-H.; Chai, X.-S.; Zhu, J.; Li, J.; Malcolm, E. In-digester reduction of organic sulfur compounds in kraft pulping. AdV. EnViron. Res. 2001, 5 (1), 91–98. (22) Hartler, N. Extended delignification in kraft cooking - a new concept. SVen. Papperstidn. 1978, 81 (15), 483–484. (23) Norde´n, S.; Teder, A. Modified kraft processes for softwood bleached-grade pulp. Tappi J. 1979, 62 (7), 49–51. (24) Paulonis, M.; Krishnagopalan, A. kraft liquor alkali analysis using an in-situ conductivity sensor. Tappi J. 1990, 73 (6), 205–211.

(25) Engstrom, N.; Vikkula, A.; Teleman, A.; Vuorinen, T. Structure of hemicelluloses in pine kraft cooking liquors. Proceedings of the 8th ISWPC, 1995. (26) Coimbra, M.; Waldron, K.; Selvendran, R. Isolation and characterization of cell wall polymers from olive pulp (Olea europaea). Carbohyd. Res. 1999, 320, 93–99. (27) Santiago, A.; Pascoal Neto, C.; Vilela, C. Impact of effective alkali and sulfide profiling on Eucalyptus globulus kraft pulping. Selectivity of the impregnation phase and its effect on final pulping results. J. Chem. Technol. Biotechnol. 2008, 83 (3), 242–251. (28) Rekunen, S.; Jutila, E.; Lonnberg, B.; Virkola, N.-E. Examination of reaction kinetics in kraft cooking. Paperi ja Puu 1980, 62 (2), 80–90. (29) Samp, J; Li, J. How does mass transfer affect the effectiveness of AQ? Appita J. 2004, 57 (2), 132–136. (30) Lisboa, S.; Evtuguin, D.; Pascoal Neto, C. Characterization of noncellulosic glucans in Eucalyptus globulus wood and kraft pulp. Holzforschung 2007, 61. (31) Lisboa, S.; Evtuguin, D.; Pascoal Neto, C.; Goodfellow, B. Isolation and structural characterization of polysaccharides dissolved in Eucalyptus globulus kraft black liquors. Carbohydr. Polym. 2005, 60, 77–85. (32) Rydholm, S. Pulping Processes; Interscience Publishers: New York, 1965. (33) Lai, Y.; Sarkanen, K. Kinetics of alkaline hydrolysis of glycosidic bond in cotton cellulose. Cell. Chem. Technol. 1967, 1, 517–527. (34) Jiang, J.; Kettunen, A.; Henricson, K.; Hankaniemi, T.; Vuorinen, T. Effect of alkali profiles on carbohydrate chemistry during kraft pulping of hardwoods. Proceedings of the 10th ISWPC; Yokohama: Japan, 1999. (35) Achre´n, S.; Hultholm, T.; Lo¨nnberg, B.; Kettunen, A.; Jiang, J.; Henricson, K. Improved pulp yield by optimized alkaline profiles in kraft delignification. Proceedings of the Breaking the Pulp Yield Barrier Symposium, TAPPI PRESS: Atlanta, GA, 1998. (36) Santiago, A.; Pascoal Neto, C. Eucalyptus globulus kraft process modifications: Effect on pulping and bleaching performance and papermaking properties of bleached pulps. J. Chem. Technol. Biotechnol. 2008, accepted. (37) Guerra, A.; Lucia, L; Argyropoulos, D. Isolation and characterization of lignins from Eucalyptus grandis Hill ex Maiden and Eucalyptus globulus Labill. by enzymatic mild acidolysis (EMAL). Holzforschung. 2008, 62, 24–30. (38) Zou, H.; Genco, J.; Heiningen, A.; Cole, B.; Fort, R. Effect of hemicellulose content in kraft brownstock on oxygen delignification Proceedings of the 2002 TAPPI Fall Conference, San Diego, CA, 2002. (39) Arbin, F.; Schroeder, L; Thompson, N; Malcolm, E. Anthraquinoneinduced scission of polysaccharide chains. Tappi J. 1980, 63 (4), 152–153. (40) Kubes, G.; MacLeod, J.; Fleming, B.; Bolker, H. The viscosities of unbleached alkaline pulps. J. Wood Chem. Technol. 1981, 1 (1), 1–10. (41) Kubes, G.; Fleming, B.; MacLeod, J.; Bolker, H. Viscosities of unbleached alkaline pulps. II: The G-factor. J. Wood Chem. Technol. 1983, 3 (3), 313–333. (42) Simonson, R. The hemicellulose in the sulfate pulping process. Part 3. The isolation of hemicellulose fractions from birch sulfate cooking liquors. SVen. Papperstidn. 1965, 68 (8), 275–280. (43) Johansson, M.; Samuelson, O. Reducing end groups in birch xylan and their alkaline degradation. Wood Sci. Technol. 1977, 11 (4), 251–263.

ReceiVed for reView November 2, 2007 ReVised manuscript receiVed July 9, 2008 Accepted July 17, 2008 IE071488G