New evidence for the role of the borohydride pretreatment on the hydrogen peroxide bleaching of kraft pulp

Abstract

A study was conducted to clarify the role of the borohydride pretreatment on the hydrogen peroxide bleaching process, based on eucalyptus chemical pulps from a laboratory kraft pulping process. The results showed that the borohydride pretreatment is effective at suppressing the effect of the Fe(III) on the oxygen formation from the peroxide decomposition. This paper for the first time quantitatively proves that the carbonyl group content in the pulp can be significantly reduced (∼51.3%). Correspondingly, this paper is also the first to find that the dissolved lignin in the effluents became more stable, since there was no significant change in its UV spectrum. There is also strong evidence that the borohydride pretreated bleaching process produced a lower amount of methanol and oxalate (∼24.3% and ∼11.8% of decrease respectively), and had more residual hydrogen peroxide in the effluents (38.4% at 180 min), compared to the non-borohydride bleaching system. The study also showed that the effect of the borohydride pretreatment on the carbohydrate preservation is not significant.

---

1. Introduction

Hydrogen peroxide (H₂O₂), as an environment-friendly bleaching chemical, has been widely used in the pulp and paper industries.¹ The application of H₂O₂ in pulp bleaching can effectively reduce the use of chlorine related bleaching chemicals and thus minimize the formation of organic halogen related toxic species in the effluents.² However, the presence of transition metal ions, e.g., manganese, copper and iron, in the pulps can not only cause the decomposition of the hydrogen peroxide but also lead to the formation of strong radicals, i.e., hydroxyl and superoxide anions,^3–6 that are responsible for the severe degradation of carbohydrates.² Therefore, it is important to minimize the effect of these transition metals in the H₂O₂ bleaching process.

There are three main methods for pulp pretreatment in mill operations before the H₂O₂ bleaching process, i.e., the use of a chelating agent (e.g., ethylene diaminetetra-acetic acid or diethylenetriaminopenta-acetate), acid washing, and sodium borohydride pretreatment.¹ The use of a chelating agent can displace insoluble metals with soluble ones (chelated compounds), which can then be removed from pulps by the additional washing step.^2,7 In acid washing pretreatments (at a pH range of 1.5–3.0) the removal of the transition metals is more efficient compared to the chelation method.² However, since significant amounts of the alkaline earth metals (calcium and magnesium) are also removed from the pulps during the acid washing, cellulose degradation is more severe without the protection from these “good” metals.^8–10 Borohydride pretreatment is also considered an effective method for the removal of transition metals because only a small dosage of agent and a short time are required. Moreover, the pretreated pulp can be directly entered into the peroxide bleaching process without the additional washing step.^11–13

Previous studies have tried to clarify the mechanism of the borohydride pretreatment in the peroxide bleaching process. The assumptions include: the reduction of the transition metal ions from high to low valence, e.g., reduction of Mn³⁺ to Mn²⁺, because the low-valence metal ions have little or no catalytic effect on the decomposition of hydrogen peroxide;^12,14 that the borohydride can reduce the carbonyl groups in the cellulose to avoid the de-polymerization reactions during the peroxide bleaching process;^15,16 that the borohydride pretreatment may also modify the part of the lignin precursors through Dakin reactions (i.e., an ortho- or para-hydroxylated phenyl aldehyde or ketone reacts with the hydrogen peroxide in the base to form a benzenediol and a carboxylate¹⁷) so that the formation of new chromophores can be suppressed in the alkaline peroxide bleaching stage.^18,19 Although these explanations provide a possible route to understanding the indicated mechanism, they were not fully confirmed by the concrete evidence obtained based on the analysis of the chemical information of the pulp and liquor from the process, e.g., the change of the carbonyl group content in the pulps.

This paper tried to confirm the role of the borohydride pretreatment on the hydrogen peroxide bleaching process, based on the evidence obtained from the analysis of the compositions of interest in the pulps and the effluents. From these, the changes of major chemical parameters such as the peroxide decomposition (to form oxygen), the carbonyl groups in the pulp, the dissolved lignin structure, and the oxalate in the hydrogen peroxide bleaching processes with and without borohydride pretreatment can be compared. These comparisons could provide a new vision regarding the role of the borohydride pretreatment on the hydrogen peroxide bleaching, which is important for the further improvement of peroxide bleaching efficiency in mill operations.

2. Experimental

2.1 Raw materials

All of the chemicals including hydrogen peroxide, magnesium sulfite, sodium hydroxide, sodium borohydride and iron trichloride were of analytical grade and received from commercial sources. A eucalyptus kraft pulp from a laboratory kraft pulping process was prepared in this study. The pulping process conditions are as follows: effective alkali: 20% (as NaOH), sulfidity: 25%, ratio of liquor to wood: 4 : 1, highest temp.: 160 °C, and time at 160 °C: 150 min. The kappa number, viscosity, and brightness of the resultant kraft pulp are 14.0, 866 mL g⁻¹, and 42.36% ISO, which were tested according to Tappi method T236om-13, SCAN-C 15:62 and Tappi T452 om-02, respectively.

2.2 Hydrogen peroxide bleaching process with and without borohydride pretreatment

Peroxide bleaching experiments were conducted in plastic bags (23 mm × 33 mm) under the following conditions and in a strict order of chemical additions as follows: pulp and water (10% pulp consistency), MgSO₄ (0.05%), NaOH (0.7%), Na₂SiO₃ (1.5%), and H₂O₂ (3.0%). All chemical dosages were based on 15 g of oven dried pulp. The pH of the starting bleaching medium was adjusted to 11.00(±0.02). After adding each chemical, the sample was mixed for 5 min by kneading. Before placing the flattened plastic bag into a thermal water-bath (at 90 °C), the air in the plastic bag was vented out and then it was sealed with a heat-sealing machine. When the reaction was terminated, the sealed plastic bags were cooled to room temperature and both the pulp and bleaching effluent were collected for further property and chemical analysis. The so-called P_R bleaching was separated into two stages, i.e., the borohydride pretreatment (R) and the peroxide bleaching (P). Firstly, each chemical, as those described above, was added, except in this case using 0.1% NaBH₄ instead of H₂O₂, and the system was reacted at 90 °C for 10 min. Then, 3% H₂O₂ was added into the reaction system and a hydrogen peroxide bleaching process was followed as in the aforementioned procedures.

2.3 Pulp properties and liquor compositions analysis

Pulp viscosities and brightnesses were determined according to SCAN-C 15:62 and Tappi T452om-02 using Technidyne color touch PC CTP-ISO, USA. The amount of carbonyl in the pulps was tested using a headspace gas chromatographic method.²⁰

The amount of oxygen produced from the hydrogen peroxide decomposition was determined using a volumetric method.

The contents of residual hydrogen peroxide were tested according to the following procedure: 9 mL of sample solution and 0.5 mL of H₂SO₄ solution (2 mol L⁻¹) were added to a headspace sample vial (20 mL), which was sealed by an aluminum cap with a rubber septum. Then, 0.5 mL of KMnO₄ solution was injected to the sealed vial which was then placed in the headspace auto-sampler (DANI HS 86.50, Italy) and allowed to equilibrate at 60 °C for 10 min with shaking. Finally, a headspace sample from the vial was automatically withdrawn and measured by gas chromatography (Agilent HP-7890, Palo Alto, CA, USA) which was operated at 45 °C with nitrogen as the carrier gas at a flow rate of 3.1 mL min⁻¹, with the thermal conductivity detector set at 220 °C.²¹

The amounts of methanol in the process liquors were determined by directly adding 10 μL of black liquor into a sealed headspace sample vial (20 mL) followed by heating to a temperature of 105 °C, in which the equilibrated vapor was determined by a full evaporation headspace gas chromatographic technique developed previously.²²

The determination process for the oxalate in the process liquor was as follows: 5 mL of sample solution was directly mixed with 5 mL of sulfuric acid solution (1.0 mol L⁻¹). After one minute of ultrasonic agitation, ∼3 mL of the supernatant was filtered by a membrane filter (0.45 μm). 300 μL of the filtered solution was added into a headspace vial containing 3 mL of KIO₃ solution (10 g L⁻¹) and 1.7 mL of sulfuric acid solution ([H⁺] = 0.1 mol L⁻¹). The vial was immediately sealed and placed in an oven for 45 min at 95 °C. Then, the carbon dioxide in the vials converted from the reaction was determined by headspace gas chromatography.²³

A UV/Vis spectrophotometer (Agilent 8453, USA) equipped with a 10 mm quartz cuvette was used for determining the lignin content in the cooking liquor and full spectra of the samples were recorded.²⁴

3. Results and discussion

3.1 Effectiveness of the pretreatment processes on transition metal ion removal

3.1.1 Effect of pH on metal ion removal. After alkaline pulping and the oxygen delignification process, the transition metals remaining in the pulps are in the precipitation form of hydroxides, such as Fe(OH)₂, Fe(OH)₃, Mn(OH)₂ and so on, which are hard to remove by a pulp washing process. Therefore, acidic washing, chelation or a borohydride pretreatment are required in order to minimize the negative effects of these transition metals on the hydrogen peroxide bleaching.

Based on eqn (1), i.e.,

M(OH)_n + H⁺ → Mⁿ⁺ + nH₂O (1)

acidic washing is an effective way to dissolve the hydroxide precipitates. By merging eqn (1) and (2), i.e.,

K_sp = [Mⁿ⁺][OH⁻]ⁿ (2)

we can obtain

With the K_sp of the corresponding metal hydroxides, the pH required for dissolving the given amount of the metal ions at the thermodynamic equilibrium can be calculated. In Table 1, we have listed the typical metal hydroxides present in a pulp system and the critical pH required for dissolving 100 ppm of metal ions from the hydroxide precipitation.

Table 1 Typical metal hydroxides and critical pH for dissolving 100 ppm of metal ions from the hydroxide precipitation

Name of metal ion	Hydroxide	Ksp at 25 °C	pH
Fe(II)	Fe(OH)2	8 × 10^−16	7.93
Fe(III)	Fe(OH)3	4 × 10^−38	2.54
Mn(II)	Mn(OH)2	2.1 × 10^−13	9.29
Cu(II)	Cu(OH)2	4.8 × 10^−20	5.81

---

It can be seen from the table that the majority of metal ions can be dissolved from their hydroxide precipitations at a pH > 5.0. Fe(III) is an exceptional case because of the extremely small K_sp of its hydroxide. However, it can be effectively removed from pulps by an acidic washing at pH < 3. In general, a higher temperature leads to a lower K_sp for the hydroxide precipitations.^25,26 Therefore, the actual temperature (e.g., 60 °C) in the mill practice is unfavorable for the removal of Fe(III) ions from its hydroxide precipitation with acidic washing. Also, there are some bonding forces between the metal ions and the functional groups (e.g., carboxyl) in the pulps. Because of all these effects, the actual pHs required for the removal of metal ions from their hydroxide precipitations are lower than those listed in Table 1. On the other hand, the acidic washing performed at a very low pH (<3) is not recommended in mill operations because it not only causes a corrosion problem for the equipment but also as a significant loss of the calcium and magnesium ions from the pulps will cause severe carbohydrate degradation during the pulp bleaching.² Therefore, Fe(III) ions remaining in the pulps are a difficult species to be effectively removed in the current mill acidic washing conditions.

3.1.2 Effect of the chelation agent on Fe(III) removal. It is well known that many insoluble metal compounds can be converted into soluble metal-ion complexes with a chelating agent, such as EDTA or DTPA,²⁷ which can then be easily removed by water washing. Chelating agents, such as EDTA, can displace the metals, making them soluble and easily removable by washing. Since EDTA is a hexaprotic acid: H₆Y²⁺, the formation constant K_f, for the following reaction:

Fe³⁺ + Y⁴⁻ ⇔ FeY⁻ (4)

should be:

Under various pHs, the principal species of EDTA is not only Y⁴⁻, and there exists the below solubility equilibrium:

H₆Y²⁺ ⇔ H⁺ + H₅Y⁺ ⇔ 2H⁺ + H₄Y ⇔…⇔ 6H⁺ + Y⁴⁻ (6)

so for the above dissociation of EDTA, there is:

[Y⁴⁻]_t = α[EDTA] (7)

where [Y⁴⁻]_t is the concentration of Y⁴⁻ only in the solubility equilibrium of EDTA, [EDTA] is the total concentration of all 7 EDTA species, α is the fractional composition of EDTA, and it can be written as below and calculated based on the stepwise formation constants.

For reaction (4) combined with the solubility equilibrium in reaction (6), there is:

[Y⁴⁻] = [Y⁴⁻]_t − [FeY⁻] = α[EDTA] − [FeY⁻] (8)

Through integrating eqn (5) to eqn (8) and together with eqn (3), [FeY⁻] can be calculated by:

So, the total soluble metal ions concentration under various pH values can be calculated by:

C^T_Fe = [Fe³⁺] + [FeY⁻] (10)

Based on eqn (10), the applicable pH range for dissolving the given amount (100 ppm) of Fe(III) ions, i.e., Fe³⁺ and FeY⁻, can be calculated using various dosages of chelating agent (see Table 2).

Table 2 The applicable pH range for dissolving 100 ppm Fe(III) from hydroxide precipitation by chelation

Concentration of chelating agent, mol L^−1	Applicable pH range
	EDTA	DTPA
a Fe(III) in hydroxide precipitate cannot be released by chelating.
1.0	7.5–10.7	4.0–12.0
0.1	8.5–10.2	4.0–11.5
0.01	—^a	4.0–11.2
0.001	—^a	—^a

---

From the table, it can be seen that DTPA has a wider applicable pH range than EDTA for chelating Fe(III) and increasing the dosage of the chelating agent is helpful for the removal of Fe(III) ions. However, due to the high chemical cost, only a small amount of chelating agent, e.g., 0.38 mmol L⁻¹ EDTA, is applied in mill operations.²⁷ Therefore, the current chelation pretreatment is not sufficient to remove Fe(III) ions from pulps. However, since it can effectively remove manganese (>90%) from pulps without significantly losing the calcium and magnesium existing in the pulps, the process is commonly applied in many pulp mills.²

3.2 Effect of borohydride on reducing Fe(III) catalysis

The above calculations show that both acidic washing and chelation treatment are not effective to remove Fe(III). Because of its reduction capability for the metal ions, borohydride pretreatment was introduced in mill operations.¹³ However, the effectiveness of the borohydride pretreatment for transition metal removal was typically evaluated by comparing the amount of residual hydrogen peroxide in the effluent at a given pulp kappa number (or brightness).^12,28,29 Since the hydrogen peroxide is also consumed by the reactions with lignin and the carbohydrates in the bleaching process, it is very difficult to provide a quantitative decomposition of hydrogen peroxide by the transition metal ions. In this work, the effect of Fe(III) on the peroxide decomposition was evaluated by measuring the amount of oxygen generated, which is the eventual product of the invalid decomposition of hydrogen peroxide. As shown in Fig. 1, the formation of oxygen caused by the decomposition of hydrogen peroxide due to Fe(III) catalysis (in the bleaching conditions but without pulps) can be significantly minimized after the borohydride pretreatment. Since the Fe(II) reduced by the borohydride can be re-oxidized to Fe(III) by the oxygen from hydrogen peroxide decomposition, the system contains both Fe(II) and Fe(III) ions. Therefore, the degree of hydrogen peroxide decomposition is greater than that of the peroxide self-decomposition in the alkaline medium, as shown in Fig. 1.

Fig. 1 Oxygen formation (a) from peroxide (3.33 g L⁻¹) self-decomposition in an alkaline solution (2.32 g L⁻¹); (b) in the presence of Fe(III) (92.1 ppm); (c) after the addition of borohydride (0.11 g L⁻¹) for 3 min; and (d) for 10 min.

Fig. 2 shows the oxygen formation from the hydrogen peroxide decomposition during the peroxide bleaching process of a kraft pulp with and without the borohydride pretreatment. Clearly, borohydride pretreatment is very effective for suppressing the oxygen formation from hydrogen peroxide decomposition during the bleaching process. It should be pointed out that an additional washing after the pretreatment was not applied after the borohydride pretreatment, which is cost-effective compared to the chelation pretreatment.

Fig. 2 Oxygen formation during the hydrogen peroxide bleaching processes.

3.3 Effect of borohydride pretreatment on the lignins and the carbohydrates

3.3.1 Changes of the carbonyl group content in the pulps during the peroxide bleaching process. It is well known that borohydride can react with carbonyl groups (ketones and aldehydes) in organic compounds to form hydroxyl groups.³⁰ Fig. 3, shows the change of the carbonyl group content in the pulps with and without borohydride pretreatment during the hydrogen peroxide bleaching process. It can be seen that the amount of carbonyl groups in the pulps is significantly reduced from 70 to 34 mol kg⁻¹ (i.e. 51.3% of decrease) after the borohydride pretreatment. During the bleaching process, the amount of carbonyl groups in the pulps is slightly decreased in the non-borohydride system. However, the amount of carbonyl groups in the pulps has some increase in the early stages of the bleaching process with borohydride pretreatment.

Fig. 3 Change of carbonyl content in the pulps as a function of bleaching time in the P and P_R processes.

The amount of carbonyl groups in the pulps is contributed to by both the residual lignin and the carbohydrates. The amount of carbonyl groups in the residual lignin is reduced through the Dakin reaction,¹⁷ meanwhile some hydroxyl groups in the carbohydrates are converted to carbonyl groups by the oxidation of peroxide radicals during the bleaching process. If the rates of these two kinds of reactions are the same, the overall amount of carbonyl groups in the pulps shows no change. According to Fig. 3, the formation of carbonyl groups in the pulps is only observed in the borohydride pretreated bleaching process. This indicates that the oxidation of the hydroxyl groups in the carbohydrates is more significant than the Dakin reaction in the borohydride pretreated bleaching process.

3.3.2 Changes of the dissolved lignin structure during the peroxide bleaching process. To prove the effect of the borohydride pretreatment on lignin, we conducted UV spectral analyses for the dissolved lignin in the bleaching effluents for the different bleaching times. Fig. 4a and b show the UV spectral changes of the dissolved lignin between these two bleaching processes with and without borohydride pretreatment. For a better comparison, the original UV spectra are normalized (based on 280 nm (ref. 24)) to keep the same lignin concentration. Clearly, there is no significant change observed on the spectrum of the dissolved lignin during the borohydride pretreated bleaching process. However, the absorbance of the dissolved lignin decreased below 280 nm when the bleaching proceeded in the non-borohydride system. Such spectral changes can be attributed to the reduction of the auxochrome groups on the lignins, typically through demethoxylation.²⁷ The spectral evidence indicates that the dissolved lignin in the borohydride pretreated system is relatively stable, which prevents the further reaction between the peroxide and the dissolved lignin in the effluents, typically the demethoxylation of the dissolved lignin.

Fig. 4 Normalized UV spectroscopy of dissolved lignin in (a) P and (b) P_R.

3.3.3 Comparison of methanol formation between the peroxide bleaching processes with and without the borohydride pretreatment. In Fig. 5, the methanol formation during the bleaching processes is shown. Because methanol cannot be generated from the methoxyl groups on the uronic acids in the xylans under either oxygen delignification or hydrogen peroxide conditions,³¹ the methanol produced during the processes must be from the demethoxylation of lignin. There is less methanol produced (i.e., 24.3% of decrease) in the borohydride pretreated system than the non-borohydride system (see Fig. 5), which can be explained by the phenomenon observed in the UV spectral changes in the dissolved lignin measurement.

Fig. 5 Comparison of the methanol produced in the P and P_R processes.

3.3.4 Change of the dissolved lignin content and the residual peroxide during the peroxide bleaching process. In Fig. 6, the concentration profiles of the dissolved lignin in the effluents during the bleaching process with and without the borohydride pretreatment are shown. It is clear that the delignification took place in the early stage of the bleaching processes (within 20 min), and that the borohydride pretreated bleaching process removes more dissolved lignin (∼12%) than that of the non-borohydride process. However, the amount of dissolved lignin in the effluents decreases as the processes continue in both cases, which should be attributed to the further degradation of the dissolved lignin by the reaction with hydrogen peroxide which can be proven by Fig. 7.

Fig. 6 Comparison of dissolved lignin content between the effluents from the P_R and P processes.

Fig. 7 The time-dependent concentration profiles of the residual hydrogen peroxide in the bleaching effluents.

Because the reactions on the dissolved lignin can consume the hydrogen peroxide (see Fig. 7) and the invalid peroxide consumption will not be beneficial to increasing the pulp brightness, as it will result in a higher chemical cost, consequently the borohydride pretreated bleaching process is helpful for saving chemicals (i.e., 38.4% of peroxide at 180 min).

3.3.5 Effectiveness of the borohydride on carbohydrate degradation. Compared to the non-wood pulp materials, the amount of transition metal ions in the eucalyptus chemical pulps is relatively low.³² Therefore, the effect of the borohydride pretreatment on the carbohydrates’ protection (excluding the impact from the transition metal ions) can be addressed. As can be seen in Fig. 8, although the borohydride pretreatment can effectively reduce the amount of carbonyl groups in the pulps, it seems not to directly associate with the carbohydrates’ protection when comparing the pulp viscosity data (proportional to the degree of carbohydrate polymerization¹) with the non-borohydride bleaching system. However, as can be seen in Fig. 8, the pulp viscosity has a tendency to decrease with the increase in the pulp brightness, and the pretreatment of borohydride didn’t change this tendency. This is possibly due to the fact that the reduced metal ions in the pretreatment step were re-oxidized to their corresponding high-valence states by the oxygen which resulted from the peroxide decomposition. This re-oxidation reaction may induce the decomposition of the peroxide³³ and the production of a hydroxyl radical which is capable of degrading the cellulose directly³⁴ and correspondingly decrease the viscosity.

Fig. 8 Change of pulp viscosity as a function of brightness.

It is well known that the oxalate is one of the degradation products from carbohydrates and lignin in the hydrogen peroxide bleaching process.^35–37 Fig. 9a and b show the formation of the oxalate and the relationship between the amounts of oxalate formed and the pulp brightness, during the bleaching processes with and without the borohydride pretreatment. Clearly, the process with the borohydride pretreatment produces less oxalate (i.e., 11.8% of decrease at maximum) than the non-borohydride bleaching process, indicating that the borohydride pretreatment has some effect on suppressing oxalate formation in the process. Because there are no significant differences in the pulp viscosity (i.e., cellulose polymerization) decreasing with its brightness increasing, it can be concluded that the suppression effect of the borohydride pretreatment should take place on the lignin side. Comparing Fig. 4 with Fig. 9a, it is noticeable that the amount of dissolved lignin in the bleaching effluent starts decreasing at >20 min, however the amount of oxalate in the effluent still increases, with or without the borohydride pretreatment. Since some of the lignin in the pulps is modified by the borohydride pretreatment (by converting carbonyl groups to alcohol groups) and becomes more stable, a lower amount of oxalate was formed from the dissolved lignin in the corresponding bleaching process.

Fig. 9 (a) Oxalate formation during the hydrogen peroxide bleaching processes; (b) relationship between the amount of oxalate in the effluent and the pulp brightness.

3.4 General economic evaluation of the peroxide bleaching processes with a borohydride pretreatment

Borohydride pretreatment showed several positive effects as discussed above, but for industrial applications, the cost is also a big concern. Although the market price of sodium borohydride is about 50 times that of hydrogen peroxide (30%, w/w), the total cost of the P_R process can still be comparable to the conventional P process because the dosage of the peroxide is 30 times that of the borohydride and because of the great performance in terms of the peroxide savings (38.2% as shown in Fig. 7). Nonetheless, further research on the cost-control will be necessary, such as finding an alternative reducing agent of low cost, or a catalytic agent to enhance the effectiveness of the borohydride.

4. Conclusions

In this work a study was conducted to clarify the role of borohydride pretreatment on the hydrogen peroxide bleaching process. Based on the evidence obtained from the analysis of the compositions of interest in the pulps and the effluents, we can conclude that the borohydride pretreatment is relatively effective at suppressing the effect of Fe(III) on the peroxide decomposition. With the borohydride pretreatment, the content of carbonyl groups in the pulp can be significantly reduced and the residual lignin becomes more stable. As a result, the reaction between the peroxide and the dissolved lignin in the bleaching effluents can be minimized. The study also indicates that the effect of the borohydride pretreatment on the carbohydrate preservation is not significant.

Acknowledgements

This work was supported by Foundation of State Key Lab of Pulp and Paper Engineering Team Fund (No. 2015ZD01), South China University of Technology, China, and National Natural Science Foundation of China (No. 21576105), China.

References

1. H. Sixta, Handbook of Pulp, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2006.
2. P. Bajpai, Environmentally benign approaches for pulp bleaching, Elsevier, Amsterdam, 2012.
3. J. Basta, L. Holtinger and J. Höök, 6^th International Symposium on Wood and Pulping Chemistry Note, Appita, Parkville, Victoria, Austrilia, 1991.
4. J. Basta, L. Holtinger, A. S. Haelstroem and P. Lundgren, Tappi Pulping Conference, Atlanta, GA, 1994.
5. F. Desprez, J. Devenyns and N. Troughton, Tappi Pulping Conference, Atlanta, GA, 1994.
6. M. Ek, P. Axegård, E. Bergnor and U. Ekholm, Tappi Pulping Conference, Atlanta, GA, 1994.
7. B. van Lierop, N. Liebergott and M. Faubert, J. Pulp Pap. Sci., 1994, 20, 193.
8. J. Basta, L. Holtinger, W. Hermansson and P. Lundgren, 1994 International Pulp Bleaching Conference, Vancouver, British Columbia, 1994.
9. J. Bouchard, H. Nugent and R. Berry, J. Pulp Pap. Sci., 1995, 21, 203.
10. S. Ghasemi, R. Behrooz, P. Fatehi and Y. Ni, Bioresources, 2010, 5, 2258.
11. Z. He, Y. Ni and E. Zhang, Appita J., 2005, 58, 72.
12. Y. Ni, Y. Ju and H. Ohi, J. Pulp Pap. Sci., 2000, 26, 90.
13. S. Wang, Z. Li, Y. Ni and E. Zhang, Appita J., 2004, 57, 377.
14. Z. Qiu, Y. Ni and S. Yang, J. Wood Chem. Technol., 2003, 23, 1.
15. C. Chirat, M. T. Viardin and D. Lachenal, Pap. Puu, 1994, 76, 417.
16. E. Ramos, S. Calatrava and L. Jiménez, Afinidad, 2008, 65, 366.
17. H. D. Darkin, Am. Chem. J., 1909, 42, 477.
18. G. Leary and D. Giampaolo, J. Pulp Pap. Sci., 1999, 25, 141.
19. C. Heitner, H. Bolker and H. Jones, Pulp Pap. Can., 1975, 76, 80.
20. Y. T. Song, H. C. Hu and X. S. Chai, China Pulp Pap., 2014, 33, 36.
21. H. C. Hu, H. J. Jin and X. S. Chai, J. Chromatogr. A, 2012, 1235, 182.
22. H. L. Li, H. Y. Zhan, S. Y. Fu, M. R. Liu and X. S. Chai, J. Chromatogr. A, 2007, 1175, 133.
23. H. C. Hu, H. J. Jin and X. S. Chai, J. Ind. Eng. Chem., 2014, 20, 13.
24. Q. L. Tu, F. Y. Fu, H. Y. Zhan, X. S. Chai and L. A. Lucia, J. Agric. Food Chem., 2008, 56, 3097.
25. F. J. Rossotti, The determination of stability constants and other equilibrium constants in solution, McGraw-Hill Book Co., New York, 1961.
26. K. Denbigh, The principles of chemical equilibrium, Cambridge University Press, Cambridge, 1966.
27. M. Ek, G. Gellerstedt and G. Henriksson, Pulp and paper chemistry and technology, Volume 2: pulping chemistry and technology, Walter de Gruyter GmbH & Co, Berlin, 2009.
28. Z. He, Y. Ni and E. Zhang, J. Wood Chem. Technol., 2005, 24, 1.
29. M. Wekesa and Y. Ni, Can. J. Chem. Eng., 2003, 81, 968.
30. S. W. Chaikin and W. G. Brown, J. Am. Chem. Soc., 1949, 71, 122.
31. T. Vuorinen, P. Fagerström, J. Buchert, M. Tenkanen and A. Teleman, J. Pulp Pap. Sci., 1999, 25, 155.
32. L. Lapierre, J. Bouchard, R. Berry and B. van Lierop, J. Pulp Pap. Sci., 1995, 21, 268.
33. J. Takagi and K. Ishigure, Nucl. Sci. Eng., 1985, 89, 177.
34. J. Gratzl, Papier, 1992, 46, 1.
35. M. A. Paulonis and G. A. Krishnagopalan, J. Pulp Pap. Sci., 1994, 20, 254.
36. Y. Yamamoto, E. Niki, H. Shiokawa and Y. Kamiya, J. Org. Chem., 1979, 44, 2137.
37. A. Elsander, M. Ek and G. Gellerstedt, Tappi J., 2000, 83, 73.

---