Enzyme Technology for Pulp Bleaching and Deinking
Pulp and paper processing is one of the largest users of biomass today. The per capita consumption of paper in the United States exceeds 300 kg annually[1], and paper consumption increases with the standard of living, so one can expect the paper industry to increase in other countries as emerging economies continue to grow. However, the established pulping processes are relatively inefficient and environmentally costly, so new processes are needed. Enzymes will play roles in those developments. Moreover, greater emphasis is being placed on extending the wood resource through fiber recycling, and enzymes will help us solve some of the problems there as well.
Biotechnology will affect pulp and paper processing in increasing ways. We've already seen the commercialization of enzyme enhanced bleaching of chemical pulps, and enzymatic pitch removal is emerging as a commercial process in Japan. Enzymatic fiberization during beating has been long recognized as a practical process, and the use of enzymes to increase drainage rates for recycled fibers likewise is a useful and economic means to increase paper machine throughput. In the laboratory and at the pilot plant scale, we've demonstrated that enzymatic removal of ink from recycled fibers is more effective and economical than traditional chemical deinking methods. The technology for enzyme applications in the pulp and paper industry has recently been covered in a text[2].
In the longer term, we will see the introduction of enzyme-based processes that when combined with mechanical and extractive pulping will result in much higher yields with better pulp properties than what can be obtained with the kraft process today.
The kraft process renders wood into pulp with an overall yield of about 45 to 55%. Yields from mechanical and chemo-mechanical pulping can be as high as 97%, but the resulting pulp is of lower quality. In converting wood into pulp, it is most important to remove the lignin. That might be accomplished through a combination of chemo-mechanical pretreatment, enzymatic hydrolysis, and extractive processes. In the end, one could obtain a higher pulp yield and recover much more in terms of higher value products than is possible with contemporary technologies.
That, however, is a brief vision of the future for enzymatic pulp processing. The present chapter is concerned mainly with the development of current technology for enzymatic bleaching and deinking.
There are at least two explanations for how xylanases and mannanases enhance pulp bleaching. The first model suggests that they increase access of bleaching chemicals to pulp fibers by removing precipitated xylan[10]. The uncoated fibers are then more susceptible to bleaching chemicals and lignin extraction. Essentially, this model proposes that xylan physically entraps lignin and chromophores in the pulp matrix. The second model suggests that hemicellulases release chromophores and lignin from the cellulosic pulp matrix by breaking covalent linkages between the hemicellulose and lignin moieties. Rather than postulating a physical entrapment, it holds that residual lignin and chromophores are chemically bound in the pulp. Recent evidence supports the role of xylanase in breaking lignin-carbohydrate bonds[11], and Suurnäkki et al.[12] recently found that no extensive relocation of xylan to the outer surface occurs during pulping, so the occlusion model might not have a sound premise.
There is increasing evidence that chromophores arise at least in part from the degradation of hemicellulose rather than lignin[13], [14]. During kraft pulping, methylglucuronic acid and other hemicellulosic components are degraded to acidic chromophoric entities that remain bound to the xylan backbone. There are many different degradation and condensation products, and they are not fully characterized or documented. Degradation products of lignin and hemicellulose can cross-react with xylan and become bound into the hemicellulosic matrix. Hemicellulose hydrolysis can release the bound chromophores and lignin, but xylan removal, per se, is not good because it decreases the pulp yield, and if carried to an extreme, xylan removal can decrease pulp strength. So the objective is not to take out the xylan, but rather to remove the chromophores and lignin. The reason for using enzymes is to do that in a more specific, economical and environmentally benign manner.
Let us consider the molecular weight requirement in the light of the first mechanistic model for xylanase bleaching. Xylanases are proposed to remove precipitated xylan thereby enhancing the access of chemicals to pulp fibers and increasing bleachability. That is to say, xylanases -- with approximate diameters ranging from 10 to 30 angstroms -- are proposed to increase the accessibility of chlorine, which has an approximate diameter of 1 angstrom. Given this size comparison, it seems likely that chlorine would penetrate more readily into pulp than any enzyme, and that xylanases could not significantly enhance its access. The micropore structure of kraft pulp is on the order of 30 to 50 angstroms and is sufficient to allow enzyme diffusion.
Enzyme penetration into pulp, however is not simply determined by molecular weight or porosity. Some xylanases possess cellulose binding domains that promote substrate adhesion and limit the rate of enzyme diffusion. Other enzymes, once absorbed, do not readily dissociate from the substrate, and their actions become localized. Both of these factors favor maximal enzyme action on the surface of the fibers, and various experiments suggest that enzyme action is greatest on the outer layers.
Perhaps the better question is whether precipitated hemicellulose forms a physical barrier that prevents the subsequent extraction of residual lignin. If, however, the residual lignin cannot be extracted from pulp under the much more extreme hot alkaline conditions prevailing during kraft cooking, it seems unlikely that they are any more likely to be extracted following chemical oxidation -- unless the oxidation itself is breaking bonds between the bulk lignin and carbohydrate moieties.
What we know from chemical analysis of enzymatically digested kraft pulps is that covalently-bound lignin or chromophores generated during the kraft cook[15] cross-react with the hemicellulose (and in some cases between the lignin and cellulose), and that those linkages must be broken in order to extract the residual lignin and to remove the chromophores. Hemicellulases liberate residual lignin and kappa by breaking covalent lignin-carbohydrate linkages, and thereby increasing extraction. The more residual lignin and chromophores one can remove by extraction, the less chlorine (or other oxidant) is required.
Enzyme penetration throughout the cellulose fiber is essential in order to fully access and extract residual lignin. Complete penetration might not be achieved in a single enzyme application. Enzyme penetration is determined in part by molecular weight, but other factors such as substrate binding may be more critical. High substrate affinity may be a positive factor when enzymes are applied to pulps in a dilute suspension, (because they would keep the enzyme in the pulp rather than the free solution) but at the relatively high consistencies used in bleaching plants (>10% total solids), they could work against enzyme diffusion to the interior of the fiber.
The second important characteristic, at least with respect to kraft pulp, is that the enzymes should have an alkaline pH optimum. If one uses a xylanase with an acidic pH optimum, it is possible to wash most of the alkali out of the pulp then neutralize the residual with sulfuric acid. Not much acid is required, but alkali keeps leaching out of these pulps even after extensive washing and pH adjustment. In fact, enzyme activity actually enhances the leaching of the alkali, and so one can observe a constantly increasing pH in the bulk solution. Moreover, neutralization may be uneven, and local pH may be critical. In order to maximize effectiveness, it is essential for the enzyme to act in the interior of the fiber where the residual alkali concentration is highest, so an alkaline pH optimum is very important.
A third feature that's useful in a xylanase is an alkaline isoelectric point. With an alkaline pI, the enzyme binds more readily to the fibers under the pH prevailing in the pulp, and this is important for determining its ability to attack the substrate. Pulp fibers are negatively charged due to the presence of sugar acids, and if an enzyme has an alkaline pI, and hence a positive charge at the operational pH, it will bind more effectively.
The fourth factor, thermal stability is not generally a problem for xylanases. Most microbial xylanases are stable at 50-60deg.C, and this is within the range of prevailing temperatures in pulp mills. The combination of alkaline and thermal stability, however, is more difficult to come by.
The final factor, substrate specificity, is not well understood with respect to pulp bleaching and is the principal subject of ongoing studies. We know that the best prebleaching enzymes are xylanases, but these enzymes differ significantly in their effectiveness.
Classifications based on molecular weight and pI are necessarily related to those based on sequence, and sequence analysis can reliably predict crystal structure, but few studies have been performed that relate sequence or structural family to action patterns and substrate specificity. Even fewer studies have explored the specificity of hemicellulases with respect to substrate branching patterns or substitution.
Family 10 xylanases occasionally exhibit endocellulase activity; they generally have a higher molecular weight, and they occasionally will possess a cellulose binding domain. Members of family 10 will act on both PNP-xylobiose and PNP-cellobiose, however, the overall catalytic efficiency on PNP-xylobioside is about 50 times higher[17]. This suggests that family 10 enzymes act mainly on xylan.
Even though all xylanases are endo acting, they can show variations in their product profiles. Some enzymes form predominantly xylose and xylobiose and others predominantly (or exclusively) form xylotriose and other higher oligosaccharide products. This difference appears to result from the number of substrate-binding subsites on the enzyme surface. The number of pyranose rings that the enzyme will bind effectively determines the nature of the oligoproducts.
The family 10 catalytic domain is a cylindrical [[alpha]]/b barrel resembling a salad bowl with the catalytic site at the narrower end, near the C-terminus of the [[beta]]-barrel[18,19]. There are five xylopyranose binding sites. Catalytic domains of these enzymes belong to a "super family" that includes Family A cellulases, [[beta]]-glucosidase, [[beta]]-galactosidase, [[beta]]-(1-3)-glucanases, and [[beta]]-(1-3, 1-4)-glucanases[20]. Family 10 xylanases have relatively high molecular weights, and they tend to form oligosaccharides with a low degree of polymerization (DP).
Family 11 xylanases are true xylanases. They don't have cellulase activity; they consistently exhibit a low molecular weight, and they can have either a high or low pI. They are formed by both bacteria and fungi. Family 11 catalytic domains consist principally of [[beta]]-pleated sheets formed into a two layered trough that surrounds the catalytic site[21,22]. Protruding down into the trough, and located toward one side of the protein is a long loop terminating in an isoleucine. Törrönen and Rouvinden have likened the trough to the palm and fingers and the loop to the thumb of a right hand.[23] The positions of many amino acids are essentially identical in the family 11 xylanases from bacterial (Bacillus circulans) and fungal (Trichoderma harzianum) origins. Thus, there has been a tremendous conservation of the basic structure of the catalytic site of family 11 xylanases during evolution.
Two family 11 xylanases are produced by Trichoderma. Xyn1 has an acidic pI (5.5), possesses a smaller, tighter groove than Xyn2, and a lower pH optimum[24]. It also exhibits a fifteen-fold higher turnover number[25] and a three-fold lower Km than Xyn2. The latter has a basic pI (9.0), a more open structure, and a wider pH range. Xyn2 tends to produce larger oligo-saccharides. Both Xyn1 and Xyn2 release xylobiose in the retained [[beta]]-configuration, indicating that the product is transiently attached to the enzyme surface[26]. The difference in pI between Xyn1 and Xyn2 is attributable to the presence of more lysine an arginine residues on the sides of the isoleucine "thumb" of the enzyme. The function of these charged groups is not well established, but they could assist in binding to acidic side chain substituents on the xylan backbone. Binding of Trichoderma xylanases to polysaccharides is affected by the pH and the ionic strength[27]. Enzymes are totally bound to xylan when the pH is below their pI, but are mainly unbound at pH values above the pI.
The pH optimum depends on properties of the acid/base catalyst[28]. In Xyn1, this is Glu164; in Xyn2, it is Glu177. In Xyn1, Asp33 makes a strong hydrogen bond (2.9 angstroms) to Glu164, thereby lowering the pKa. In Xyn2, an asparagine residue (Asn44) takes the place of Asp33; the hydrogen bond is much longer (3.7 angstroms), and the interaction is weaker. All acidic pI xylanases of family 11 have an aspartic residue in this position; all basic xylanases have an asparagine residue.
Figure 1 shows the pH activity of an alkaliphilic xylanase from Bacillus sp. V1-4 which we isolated from a pH 11 kraft pulp[29]. The organism will grow on xylan at pH 10. The enzyme has a rather broad pH optimum extending well into the alkaline region. Other enzymes have even more alkaline optima. Pulpzyme HC, has a distinct optimal activity around pH 9 to 9.5.
Figure 1. pH activity curves for alkaline-active xylanase from Bacillus sp. V1-4 and a cloned preparation, Pulpzyme(TM) HC.
See Biochemistry and Genetics of Microbial Xylanases for a more complete discussion of the molecular properties of xylanases.
Chromophore release correlates with enzyme dose over the effective range of enzyme addition (Fig. 2). In fact we can also see that the kappa number following enzyme treatment and alkali extraction decreases with the enzyme dose. Absorptivity is much greater in the UV region than in the viable region, so monitoring UV absorptivity is a sensitive means to determine the extent of enzyme action on pulp.
Figure 2. Chromophore release (a) and kappa reduction (b) from hardwood pulp as a function of treatment with crude xylanase from Streptomyces roseiscleroticus.
Effective xylanases show a high ability to release chromophores. Figure 3 illustrates the relative chromophore release and kappa reduction following enzyme treatment with four different specific isozymes that we've isolated from Streptomyces roseiscleroticus.
Figure 3. Correlation of chromophore release and kappa reduction by several purified isoenzyme fractions.
Figure 4 demonstrates that the brightness increase attained following extraction and chlorine dioxide bleaching varies with the isozyme treatment. Following treatment with specific isoenzyme fractions applied to pulp at a dosing rate of 3 international units per gram, we attained a brightness of 80 -- up 12 points from the 68 brightness attained with the control pulp not treated with enzymes[32].
Figure 4. Enhanced brightness of a softwood pulp following treatment with purified isoenzyme fractions from Streptomyces B 12-2.
In conclusion, with respect to bleaching, we know now that xylanases reduce chemical bleaching; thermostable alkaline-active enzymes are commercially available, and we will have better enzymes in the future; the enzymes increase the pulp accessibility and chromophore release; our objective is to remove color, not the xylan; and that if we monitor the chromophore release we can greatly increase the implementation of this technology in pulp mills.
Enzymatic de-inking could become commercial because the technology fits easily into current fiber recycling practices; it is lest costly than chemical deinking, and it is more efficient.
Enzyme action is affected by the paper constituents in the de-inking condition. The chemically pulped fibers are more susceptible than mechanically pulped fibers. This is important because mechanical fibers have a lot of lignin left and hence they're much more resistant to the cellulases. Therefore, this technology works best in mills that are recycling stock paper such as office waste which is very high in chemical pulp content.
Office waste paper, unfortunately, is also high in laser and toner content, thus it has a very low value because you can't mix it with anything, and the technology for taking the toner particles out is not very good at the moment.
Enzyme deinking is complemented by maceration, because the major role of cellulase is to release the toner particles from the surface of the fiber. One depends upon the mechanical action in the pulping process to release the toner particles, and physical separation remove the toner particles by flotation.
They might also decrease the specific surface area of the fibers and thereby reduce interaction with contaminants. That is to say there might be microfibrils on the surface of these very frazzled, recycled fibers which could be trapping the ink particles, and by giving the fibers a "haircut", we reduce their adhesion.
Our results indicate that the most important factor with respect to toner removal is the increased flotation efficiencies imparted by cellulase activities. The biggest increase in toner removal that we observed occurred during the flotation stage. During flotation, air bubbles rise to the surface of the flotation tank through a relatively dilute pulp stock, approximately a 1% consistency. The surfaces of air bubbles are relatively hydrophobic and they carry the toner particles to the top where they are removed by a skimming action.
Figure 5. Toner particles with fewer fibers are more likely to be carried to the surface during flotation.
Figure 5, offers a cartoon notion of how air flotation works in removing ink particles from pulp fibers. The bubble on the left looks as though it is having a bad hair day. He's got a lot of cellulose fibers still attached to the toner particles that he is carrying along, and they're going to fall off before he gets to the surface. The bubble on the right has picked up two particles that are free of fiber, and as a consequence they stick to the bubble much better and experience less drag from interaction with the water. If you can give the toner particles a clean shave, they'll get carried to the top much faster.
The first stage, fiberization refers to the dissolution of paper particles in water prior to enzyme addition. During this phase of the pulping process, the amount of fiber surface area exposed to water increases dramatically. We want to have as much surface area exposed as possible because that determines the efficiency of enzyme action. Addition of surfactant in this stage is very important because it helps to wet the fiber surfaces and overcome some of the problems that one encounters with sizing. Sizings are chemicals that are added to paper prevent inks from running. The make the paper surface more hydrophobic and less susceptible to enzyme action.
After fiberization, we add the enzyme batch-wise in 10-15% of the total pulp volume. It is necessary to dilute the enzymes in large quantities of water in order for them to have good contact with all the fiber surfaces. Then we pulp at 14% consistency with the enzymes at 50deg.C for 30 minutes.
In a three-day, five-ton pilot plant trial, enzyme treated pulps came out with particle counts about one-tenth those attained with the heat-killed enzyme control[33]. Brightness likewise increased with most of the gain coming at the flotation stage. What was also interesting was that in the pilot plant trials, we also gained an improvement in the biological oxygen demand in the reject effluents. That is to say we decreased the biological oxygen demand in the final wastewater. In earlier experiments, we had shown that in the laboratory, enzyme treatments alone are more efficient than chemical deinking technologies or enzymes combined with chemicals[34].
For comments or further information write to Tom Jeffries: twjeffri@facstaff.wisc.edu