NC State
Vera-Loor, A., Walger, E., Mortha, G., and Marlin, N. (2023). “Oxidation of fully bleached paper-grade kraft pulps with H2O2 activated by Cu(Phen) and the effect of the final pH,” BioResources 18(3), 5222-5241.


Hydrogen peroxide is an environmentally friendly bleaching chemical that is widely used in alkaline media in the pulp and paper industry. In this study, unconventional reaction conditions of the conventional bleaching P-stage were performed to evaluate its effect on cellulose. The objective was to “tailor” the cellulose degradation according to different applications, such as dissolving pulps. Different operating conditions were studied: pH variation (alkaline or acidic medium), as well as the addition of Cu(II) or the Cu(II)-phenanthroline complex. The hypotheses considered were (1) Using H2O2 in unconventional conditions will favor hemicelluloses removal after depolymerization and will lead to the desired cellulose degradation, and (2) H2O2 oxidation carried out under unconventional conditions will favor the dissolving ability of cellulose.

Download PDF

Full Article

Oxidation of Fully Bleached Paper-grade Kraft Pulps with H2O2 Activated by Cu(Phen) and the Effect of the Final pH

Andrea Vera-Loor, Elsa Walger, Gérard Mortha, and Nathalie Marlin

Hydrogen peroxide is an environmentally friendly bleaching chemical that is widely used in alkaline media in the pulp and paper industry. In this study, unconventional reaction conditions of the conventional bleaching P-stage were performed to evaluate its effect on cellulose. The objective was to “tailor” the cellulose degradation according to different applications, such as dissolving pulps. Different operating conditions were studied: pH variation (alkaline or acidic medium), as well as the addition of Cu(II) or the Cu(II)-phenanthroline complex. The hypotheses considered were (1) Using H2O2 in unconventional conditions will favor hemicelluloses removal after depolymerization and will lead to the desired cellulose degradation, and (2) H2O2 oxidation carried out under unconventional conditions will favor the dissolving ability of cellulose.

DOI: 10.15376/biores.18.3.5222-5241

Keywords: Cellulose oxidation; Hydrogen peroxide; Copper phenanthroline; Fully bleached paper grade pulp; Cotton linters; Dissolving ability; Cellulose depolymerization

Contact information: Univ. Grenoble Alpes, CNRS, Grenoble INP, LGP2, F-38000 Grenoble, France;

* Corresponding author:



Hydrogen peroxide has been used in alkaline medium as a bleaching agent for virgin wood chemical pulps since the 80’s. Nowadays, due to its versatility, it is also used as a reinforcement delignifying agent after oxygen delignification, and in alkaline extraction stages to avoid the yellowing induced by sodium hydroxide (Carvalho et al. 2008). The P-stage is considered a complementary stage in an elementary chlorine free (ECF) sequence, to enhance the final brightness and brightness stability; and it is also a major bleaching stage in totally chlorine free (TCF) bleaching sequences, for delignification. The P stage (H2O2) could be used as a delignifying and/or a color-stripping treatment, depending on the temperature applied during bleaching and can be performed alone or in combination with oxygen. The classic operation conditions of the P-stage are 10 to 15% pulp consistency, 60 to 90 °C, 30 to 240 min of retention time, 0.2 to 4% NaOH charge, 0.2 to 4% H2O2 charge, and pH range of 10 to 12 (Walger et al. 2017).

As lignin is responsible for the color in the pulp, bleaching can be obtained either by delignification and/or by color-stripping, resulting in an oxidized lignin and partially depolymerized, which renders lignin soluble in the bleaching medium. Later, lignin is removed during washing steps placed between bleaching stages. Parallel to delignification, cellulose may also be damaged. In the case of D stage, cellulose is preserved, which is the reason why chlorine dioxide in acidic conditions is widely used in bleaching sequences.

The structures responsible for color in the pulp are polarizable groups (conjugated carbonyl groups, enones, and quinones), able to react with hydrogen peroxide in alkaline medium by the action of the perhydroxyl anion (HOO) at moderate temperature between 60 to 75 °C. The reaction leads to selective oxidation into non-colored compounds (Castellan and Grelier 2016). The destruction of these chromophores results in the whitening of the pulp, without significant lignin removal. Such reactions typically occur during the bleaching of mechanical pulps (Dence and Reeve 1996). In addition, the Dakin (or pseudo-Dakin) reaction is involved in the bleaching of chemical pulps. In alkaline medium, HOO- can react on some of the free phenolic structures in lignin (the ones bearing a carbonyl group on the Cα carbon, or on the Cα carbon of quinone methide induced structures, obtained at higher temperature in alkaline medium from the free phenolic structures bearing a benzyl alcohol). Yet, depending on the alkalinity and temperature, the reactivity of HOO- with lignin-containing pulps remains rather moderate. Contrary to lignin, polysaccharides in this range of temperatures do not react with H2O2 and HOO- (pKa = 11.6 at 25 °C), except in a strongly acidic medium.

At higher temperatures (above 80 °C), HOO- can decompose by the well-known Fenton reaction mechanism, giving rise to the formation of highly reactive intermediate oxygenated radicals, namely HO°, HOO° and O2 (pKa (HOO°/O2°)~ 4.8 at 25 °C). Such a decomposition involves the catalysis by metal cations (Mn) that are generally present as traces in the pulp. The global reaction for peroxide decomposition can be given as follows:


HO• is a strong electrophilic oxidant (E° > 2 V) with the capacity of electron abstraction on the alcoholic -OH groups of polysaccharides, leading to cellulose oxidation (carbonyl formation) and further cellulose depolymerization. Contrary to HO•, O2• is a nucleophilic species that can be oxidized into HO•, while HO• is reduced into H2O. In the presence of organic reducers, O2• can behave as a weak oxidant and react with them, since it is a long-life radical compared to HO•, and it can diffuse at a certain distance from the point where it was generated.

It is generally desired to monitor the effect of oxygenated radicals during alkaline bleaching or delignification stages applied on chemical pulps involving hydrogen peroxide (P stages). This is usually obtained by controlling the alkalinity, the temperature, and the metal cations profile in the pulp, with the possibility to apply chelating stages (Dence and Reeve 1996).

Not only lignin, but also dyes in cellulosic pulps can be fully removed or discolored by hydrogen peroxide. But depending on the dye’s molecular structure and the applied reaction conditions, peroxide’s efficiency can be poor with some recalcitrant dyes. Walger et al. (2016) focused on the possible ways of improving dye decolorizing of a fully bleached chemical pulp, impregnated by recalcitrant red dyes, which are not easily removed using conventional alkaline P stages. The study focused on the possible ways of monitoring and regulating the intensity of the hydroxyl radical flow generation, by the addition of copper additives at different pH values: either free copper (CuSO4), or the copper-phenanthroline complex (referred as Cu(Phen) in this paper). The addition of Cu(Phen) in near-neutral or alkaline aqueous solution promotes the formation of different complexes, which depend on the phenanthroline-to-copper ratio and on the concentration of OH- (Korpi et al. 2007). Walger et al. (2018) listed the abundance of the main copper species, either free, hydroxylated, or complexed with phenanthroline, according to different pH and Cu:Phen molar ratios (Walger et al. 2018). In the absence of phenanthroline, some of the obtained speciation results are given in Table, where it is evidenced that free copper dominates at near-neutral pH, whereas various hydroxylated copper complexes dominate at alkaline pH. In the presence of phenanthroline, various complexes are introduced, depending both on the alkalinity and the phenanthroline dose. In Table, speciation results for a (Cu:Phen) ratio of (1:2) are presented.

Table 1. Speciation of Copper at Different pH, in the Absence of Phenanthroline (Walger 2016)

Table 2. Speciation of Copper at Near-neutral and Alkaline pH, in the Presence of Phenanthroline at a Cu:Phen Molar Ratio of 1:2 (Walger 2016).

From Table it is apparent that the higher the pH, the more dominant the CuPhen(OH)2 complex, but also, traces of hydroxylated copper complexes arise with higher amounts of OH ligands at higher pH. At near-neutral pH, free copper is completely replaced by the CuPhen22+ complex. As free Cu2+, the latter is also involved in the catalysis of peroxide decomposition, as shown by Korpi et al. (2007) and Walger et al. (2016). Vladut et al (2011) demonstrated that HO• generation during P stages (at alkaline pH) was much more intense with Cu(Phen), when compared to a classical non-activated stage. Walger et al. (2016) showed that copper II alone, and the copper II-phenanthroline complex, both significantly improved the reactivity of H2O2. The results were a better color-stripping of dyed pulps, while cellulose became also more depolymerized during pulp decolorizing. Similar results about the effect on pulp viscosity had already been observed in previous studies on delignification (Vladut et al. 2011; Das 2012). delignification was improved using copper and phenanthroline species but the cellulose DPv was significantly affected.

In a detailed mechanistic study, Walger et al. (2021) showed the presence of free hydroxyl radicals in the pH conditions given in Tables 1 and 2, with the use of EPR spectroscopy and the spin-trapping technique, using DMPO (5,5-dimethylpyrroline-N-oxide). When present, the latter reacts almost instantaneously with the hydroxyl radicals generated in the medium to form the DMPO-OH adduct, which can be unambiguously identified by its characteristic four-lines EPR spectrum. Figure 1 presents two spectra, one at near-neutral pH with a strong signal, and the other one with a moderate signal at slightly alkaline pH.

Fig. 1. Comparison between near-neutral and alkaline pH (Walger et al. 2021; CC BY 4.0)

Figure 1 provides evidence that the presence of copper stimulated an intense generation of HO• radicals at near-neural pH, while the effects was more moderate but still present at alkaline pH. The Cu:Phen molar ratio was 1:2. Comparison of the results were done at near-neutral pH (between 7.5 and 9) or at alkaline pH (pH around 12.5).

In the presence of H2O2, copper II is the activating species, and phenanthroline only acts as a stabilizing agent to adjust the solubility, the stability, and the redox potential of copper II. The activation of H2O2 is mainly due to its decomposition into HO• by the copper species, and the extent of the activation is strongly dependent on the pH of the medium. Finally, the mentioned study concluded that the H2O2/Cu(II) system is an interesting generator of HO• radicals with adjustable flow. All details about the conditions for controlling HO• flow intensity, qualitatively identified and measured by ESR spectroscopy, are described in the studies cited. Overall, the use of free copper (CuSO4) at acidic or neutral pH promotes a non-inhibited, strong radical flow, together with fast peroxide decomposition. Such a flow can be strongly attenuated at alkaline pH, or even regulated by adding adjusted doses of NaOH and phenanthroline.

The main application of bleached pulps with low degree of polymerization considered in this study is the dissolving pulps for viscose applications. When referring to a high-quality dissolving pulps for viscose applications, the intrinsic viscosity should be in the range of 400 to 600 mL/g; a uniform molecular weight distribution is desirable to ensure homogeneous reactions, as well as high alpha-cellulose content (> 92%) (Sixta 2006). Pulps with high surface area, high porosity, and large pore size often can have high accessibility and reactivity. In summary, the key parameters to monitor in dissolving pulps include high α-cellulose content, alkali solubility, intrinsic viscosity, uniform molecular weight distribution (MWD), and good accessibility and reactivity.

In the present study, the results published by Walger et al. (2016, 2018, 2021) have been the main guide to be directly reused (without further measurement of HO• radical flow), in the context of a new focused objective, i.e., modifying a lignin-free cellulosic pulp by hydrogen peroxide to increase its dissolving ability and the control of its degree of polymerization, for further applications of chemical celluloses.


All the oxidation trials in the present section were carried out at high temperature (90 °C) to favor the formation of hydroxyl radicals. The effects on the final pH, chemical charge of H2O2 and NaOH have been explored.

Cellulosic Substrates

The screening of the most interesting conditions for cellulose oxidation with hydrogen peroxide was carried out on paper-grade kraft pulps from hardwood and softwood. The cellulosic samples selected were a softwood (SW) and a hardwood (HW) with low intrinsic viscosity values (IV) of 563 and 496 mL/g, respectively. For comparison with a substrate without hemicelluloses, cotton linters (CL, IV of 695 mL/g) were also included in the study. Other initial conditions are presented along the document in tables and figures as REF, referring to the non-treated sample.

Oxidation trials using H2O2 alone were carried out, as well as trials with copper (II) activators, either Cu(II) from copper sulfate or from the Cu(Phen) complex. The pH of the oxidation reaction varied with the addition of different amounts of sodium hydroxide: 0, 0.2 or 2% NaOH, for a constant H2O2 charge (2%). The aim was to investigate the behavior of the three H2O2-based systems in acidic, neutral, and alkaline medium.

Table 3. Conditions using 2% H2O2 and 0.02% Cu or Cu(Phen) with Different Doses of NaOH

After oxidation, the pulps were analyzed through intrinsic viscosity (IV), alkali resistance (R18), crystallinity, and brightness (%ISO). The final pH and the consumption of H2O2 were also monitored for the purpose of optimizing the conditions. Finally, the dissolving ability of the pulp was assessed using the methods proposed by Vera-Loor et al. (2023), where the derivatization of cellulose was performed while measuring the particle size by DLS; as well as by the dissolution in the NaOH:urea:water system (Vera-Loor et al. 2023). The chemical doses used for the oxidation trials are detailed in Table. These conditions have been extracted from previous work done at LGP2 by E. Walger and S. Das. Note: the percentage values mentioned are considered as weight of the chemical per weight of dry pulp.

Cellulose Oxidation

20 g of oven-dried pulp were weighed and later disintegrated in deionized water using a Lombardy apparatus for 1 min, following the guidelines for pulp disintegration of ISO 5263-1995. The pulp was then concentrated by filtration in a sintered glass crucible of porosity No. 2 and introduced in a polyethylene (PE) bag used as reactor. The total weight was completed up to 5% consistency (400 g of fiber suspension, considering the chemicals to be added). The pulp was preheated in a water bath at 95 °C for 20 min, and the chemicals were added and mixed by hand in each step in the following order: Cu(Phen) or CuSO4, NaOH, and H2O2. The pH was measured before closing the PE bag, right before putting it in the water bath for 1 h to start the oxidation. After the reaction, the fiber suspension was filtered, and the first effluent collected for pH measurement and determination of H2O2 consumption. Finally, the pulp was washed with deionized water until neutral pH. 2 g of oven-dried pulp were separated for brightness measurements, which were performed the same day of the oxidation. The remaining pulp was air-dried for further measurements such as intrinsic viscosity and functional group quantification.

The assumed reaction is the partial oxidation of the hydroxyl groups (-OH) of the cellulose into carbonyl (C=O) and / or carboxyl groups (COOH). The addition of such groups may enhance the dissolving ability of cellulose, but in parallel the oxidation may also depolymerize the cellulose chain.

The Cu(Phen) complex was prepared using CuSO4·5H2O and 1,10-phenanthroline, as detailed by Walger et al (2017). It was prepared in aqueous solution by dissolving phenanthroline and adding CuSO4, using 2 moles of phenanthroline for 1 mole of CuSO4. The complex is referred to in the document as Cu(Phen).


The pH was measured at the beginning of the reaction, after the oxidative chemical addition, and at the end of the reaction before washing.

Hydrogen Peroxide Titration

The consumption of H2O2 was determined immediately after each oxidation trial, and it was performed by iodometric titration in the presence of sulfuric acid. The titration chemical was sodium thiosulfate (0.2 M).

Pulp Brightness

The brightness was determined according to the standard ISO 2470-1:2016. Two grams (oven-dried basis) of pulp were homogenized in deionized water, and handsheets were formed the same day after the oxidation trial was performed. Once the sample was dried in the sheet former, the brightness was measured 4 times per sample, and its average was reported.

Alkali Resistance

The alkali resistance of pulp is commonly addressed as R18, and it was calculated following the standard method ISO 699-1982, which is applicable only for bleached and delignified pulps. It is used for the determination of the alkali-insoluble fraction of pulps using sodium hydroxide solution of fixed concentration. The analysis requires 2.5 g of pulp (oven-dried basis), a solution of 18% NaOH for the reaction, and a 10% acetic acid solution for washing. To calculate the R18, the weight of cellulose is expressed per gram of dry pulp sample. This method assumes that only hemicelluloses are soluble in 18% NaOH solution.

COOH Quantification

Carboxyl group quantification was estimated following the standard method TAPPI T237 cm-08.

Viscosity-average degree of polymerization of the pulp in CuED

The viscosity-average cellulose degree of polymerization approximately indicates the weight-average number of monomeric glucose units in the cellulose polymer. A viscosimetric method can be used. The determination was done using 50 mg of sample (oven-dried basis), following the procedure of the standard method ISO 5351:2010.

Cellulose Accessibility

The accessibility of the cellulosic samples can be assessed based on morphological changes, such as fiber length, crystallinity, pore structure, and specific surface area. The fiber length and width were measured using a MorFi analyzer to monitor the physical changes of the fibers before and after treatments. For the estimation of the crystallinity index (CI), the samples were cryo-ground to obtain a fine powder, which was analyzed by X-Ray Diffraction. The analyses were performed using a diffractometer X’Pert Pro MPD (PANalytical, Netherlands) equipped with a copper anode (Kα λ = 1.5419 Å). The diffractometer was working in the Bragg-Brentano geometry, with the angle 2θ ranging from 6° to 60° (0.05° interval). The measurements were analyzed with FULLPROF software to obtain the crystallinity profiles. For the calculation of the CI, the Segal peak height method was used. The method uses a maximum intensity value (I200) found between the scattering angles 2ϴ=22° and 23°. The minimum value (Imin) is typically between 2ϴ= 18° and 19°. The CI is simply the difference between these two intensities, divided by the intensity of the highest (Segal et al. 1959; French and Santiago Cintrón 2013). The CI was calculated as follows:


Carbohydrate Composition

The carbohydrate composition was determined following the guidelines of TAPPI T-249 cm-00. A specimen of 350 g (oven dried basis) was put in a glass test tube, and 3 mL of 72% H2SO2 was added. The tube was then placed in a water bath at 30 °C for one hour, and the sample was mixed often with a stirring rod. When the hour was over, the solution was transferred into a 100 mL glass resistant flask, and exactly 84 mL of distilled water was added. The flask was closed and then put in an autoclave for 1 h at 120 °C. At the end of the hydrolysis, and when the flask was cooled down, the solution was filtered and later used for the preparation of dilutions for further analysis in the DIONEX-HPLC.

Cellulose Tricarbanilation

This reaction was done following the publication of Vera-Loor et al. (2023). An amount 250 mg of oven dried basis ground cellulosic substrate was weighed in a 250 mL glass bottle with a screw cap and mixed with 25 mL of the reaction solvent, DMSO. Then it was left under stirring at 250 rpm overnight at room temperature. After 18 h, with continuous stirring, the temperature was raised to 70 °C (approximately 1 hour). Then 5 mL of phenyl isocyanate was added slowly under stirring (t0), followed by 20 mL of DMSO. The reaction proceeded for 48 h.

To monitor the particle size evolution during the tri-carbanilation by Dynamic Light Scattering (DSL), 1 mL of the sample was collected from the reacting system and diluted 10 times as follows: 5 mL of DMSO and 1 mL of acetone into a 10 mL volumetric flask, followed by 1 mL of the cellulosic substrate undergoing dissolution was sampled, and the 10 mL volume was completed with DMSO. This sampling procedure was performed for the first 6 h of the reaction. The repeatability was evaluated by triplicate testing. As for the DLS information input, a DMSO refractive index of 1.4772 and a viscosity of 1.99 mPa-s were considered at 25 °C. The instrument used was a VASCOTM particle size analyzer with a light source from a diode laser emitting at 657nm. Z-average and PDI from the Cumulants method, as the parameters recommended by ISO 22412:2017, were used in this study to evaluate the particle size.

Size-exclusion Chromatography

At the end of the carbanilation, with 48 mL left when 8 samples of 1 mL were taken, 16 mL of acetone were added carefully in the hot reactional mixture, and it was left to cool down slowly under stirring. 1 mL of this solution was diluted in 10 mL of THF, and 100 µL of the diluted solution was injected in the SEC system for MWD analyses. A multi-detector SEC system (The Viscotek-Malvern TDA 302-GPCmax system) was used for the analyses, and the MWD profiles were obtained using the software OmniSEC 4,5 (Malvern Co.). Some of the parameters studied included peak integration limits (based on RI signal), and MW versus elution volume fit, both influenced by the dispersity index (PD) and MW calculations (Mortha et al. 2011).

Dissolution in NaOH:urea:water Mixture

This experiment was performed following the guidelines proposed by Vera-Loor et al. (2023). A mass 0.25 g (1 wt%) of cellulosic substrates was dispersed in 24.75 g of pre-cooled (-12 °C) solvent made of NaOH:urea:water with a ratio of 7:12:81 wt%, The liquid fraction was pre-cooled in the system under agitation at 400 rpm for 15 min before adding the sample in a 250 mL glass bottle with a screw cap. After the pre-cooling time, the bottle was opened, and the sample transferred in very slowly with the help of a tweezer. The transfer duration was exactly 5 min, while maintaining 400 rpm agitation. After the addition of the sample, it was verified that there was no sample left on the walls of the bottle. The system was closed and allowed to dissolve for 15 min. Then, the solution was transferred to centrifuge tubes and centrifuged for 30 min at 4000 rpm. The supernatant was removed with a pipette, and the undissolved fraction was washed with distilled water until reaching neutral pH (~600 mL water was needed). The pre-weighted crucible (pore No.2), which had been used for washing, was then placed in the oven at 105 °C for 24 h. Before measuring the final weight, the hot crucible was placed in a desiccator for 15 min. The cellulose solubility (Sa) in NaOH:urea:water solution was calculated as presented in Eq. 2, where w0 is the weight of the sample and wi the undissolved fraction.



Figure 2 presents the final pH and intrinsic viscosity for the studied cellulosic substrates and the above-described oxidation treatments.

In the initial trials, only H2O2 was used with the combination of different doses of NaOH. It was observed that there were similar tendencies for SW and HW but different for CL. While SW and HW reached the final pH below neutral, there was a significant reduction of the intrinsic viscosity (IV). However, when the final pH was as high as 11, the IV reduction was much attenuated. It was observed that for CL, the higher the pH, the less affected was the cellulose (with limited IV reduction), as presented in Fig. 2(A). This can be related to the type of fiber, the presence or absence of hemicelluloses, crystallinity, and accessibility. However, they all obtained the highest intrinsic viscosity value when the oxidation was carried out at alkaline pH. When comparing the kraft pulps, IV values decreased more for HW than SW, maybe as an indication of the better accessibility of reagents to cellulose in the nanostructure of the fiber wall.

Figure 2(B), presents the results after the oxidation took place with 2% H2O2 and 0.02% Cu(II). IV values of HW were reduced significantly when the final pH was acidic. But it was only reduced to 396 mL/g at alkaline pH, which was a similar behavior as with H2O2 alone. Such behavior was also observed for SW, where the lowest IV value was reached with the lowest pH. Yet, CL behaved differently, since degradation only took place at alkaline pH, going from its initial value of 695 mL/g to 317 mL/g. Alkaline medium provided conditions for cellulose swelling and not too fast generation of radicals during peroxide decomposition, as copper ions are hydroxylated in alkaline medium, contrary to the acidic medium where the radicals flow is intense. Moreover, cellulose in CL is not protected by a hemicelluloses’ barrier around the microfibrils when it becomes more accessible by swelling.

When focusing on Fig. 2(C), a general observation can be made: with Cu(Phen), CuSO4-H2O2 or H2O2 alone, for SW and HW, the lowest cellulose degradation took place at a higher final pH (2% NaOH). This is in contrast to CL, which achieved its highest IV value at pH close to neutral (0.2% NaOH). In general terms, at acidic pH, the intrinsic viscosity values decreased down to three times the initial values, the hydrogen peroxide decayed down to 20% consumption, and the brightness %ISO was more affected compared with the other scenarios.

The H2O2 oxidation activated with CuPhen in CL at pH 7.8 allowed a slow and constant radical generation, which slightly degraded the sample, in comparison to the acidic condition with pH 3.6, where it is presumed that the radical generation was too fast, and thus the degradation was too strong. Both behaviors are likely to be explained by a Fenton-like decomposition of H2O2 in cellulose, where traces of Fe2+ promote the degradation of fibers in acidic medium (Walter 2009). Yet, in the scenario with pH 11.7, a slight degradation was observed, and it can be presumed that is due to the residual Fenton decomposition. The overall resistance of CL against degradation, when comparing to SW and HW, can be explained by its compact cell wall and highly crystalline structure.

Nevertheless, these were not the only parameters monitored. The final brightness and the H2O2 consumption, when no copper was involved, are presented in Fig. 3.

Fig. 2. Intrinsic viscosity (IV) in CuED and final pH after oxidations. (IVinitial: 563 ml/g for SW, 496 ml/g for HW, and 695 ml/g for CL)

Fig. 3. H2O2 consumption, brightness (%ISO), and final pH after oxidations with only 2% H2O2. (Initial ISO brightness: 87.9% for SW, 87.9 % for HW, and 90.2% for CL)

The results presented in Fig. 3 provide evidence that HW, SW, and CL behaved similarly. The H2O2 consumption increased when the final pH was higher. When considering intrinsic viscosity values higher than 400 mL/g, which is interesting for dissolving applications, the acidic pH did not give interesting results, as well as not achieving total consumption of H2O2 (see Fig. 2(A)). H2O2 consumption was not directly related to the decrease of IV, and the peroxide consumption was higher in alkaline medium despite a reduced bulk attack of cellulose, due to slower rate of HO• generation. In acidic medium, the bulk attack of cellulose was more important, but H2O2 consumption was lower.

The role of copper as an activated species was evaluated but is not presented in this paper. Indeed, for both scenarios with Cu(II) and Cu(Phen), it was possible to achieve the complete H2O2 consumption (>95%) because of copper catalysis, and the brightness (%ISO) was rather not affected or slightly increased.

Comparison of the Three H2O2-based Oxidation Systems

The section above showed similar reactivities for SW and HW, but significant differences with CL. For the latter, the cellulose structure was much more compact (high crystallinity) and hemicelluloses were absent, as shown in Table.

Cellulose swells in alkaline medium, which allows better bulk penetration of reagents, under conditions promoting a regulated flow of radicals (by copper complexation, Cu(OH)2 or Cu(Phen)). SW and HW are more accessible and thus reactive to the flow of radicals, despite the presence of hemicelluloses, especially in acidic medium. Finally, alkaline pH provides a better control of the radical flow, which is almost nonexistant in the absence of copper, leading to a light cellulose degradation and the highest brightness level.

Cellulose Oxidation by H2O2/Cu(Phen) – Towards Optimization

Additional tests were performed to determine the most interesting conditions to get a final pH close to neutral or slightly alkaline, and still reaching interesting final properties, with the aim of improving the dissolution ability of the studied substrates. For these tests, the concentrations of H2O2 and Cu(Phen) remained the same, 2% and 0.02% respectively, while the dose of NaOH was explored until reaching the desired final pH. The results are shown in Table 4.

Table 4. Characteristics of the Most Interesting Conditions When Using 2% H2O2 and 0.02% Cu(Phen) (REF=untreated pulp)

The determined conditions allowed full H2O2 consumption (≈ 100%) and high brightness values (≈ 90% ISO). Intrinsic viscosity values around 400 mL/g were reached for CL and SW, as shown in Table 4. However, HW did not achieve intrinsic viscosity values in the desirable range (400 to 600 mL/g) for viscose applications. Complementary alkali resistance tests were performed to assess the cellulose purity. No differences between the reference and the oxidized substrate were observed for CL (as only hemicelluloses are expected to dissolve in the R18 test). For HW, a slight improvement in the cellulose purity was observed, especially at 0.5% NaOH, reaching a good R18 value; and for SW, purification was slight but better with 0.3%NaOH than with 0.4%NaOH.

The Crystallinity Index (CI) was affected by the oxidation for all the substrates, but to different extents. This seems to have been related to the final pH. For the SW, a high CI drop was observed as soon as the final pH became alkaline. A strong drop was also observed for the HW, even at near-neutral pH. This agrees with the fact that HW is more accessible and more degraded than SW at this pH. In contrast, CL was less degraded (lower IV variations) and did not show an important CI decrease, compared to the two kraft pulps. As already discussed, CL is the least accessible substrate, especially when it is not much swelled by NaOH. For CL, was also possible to observe that the cellulose recrystallization during the drying of the wet pulp fully washed after oxidation was not hindered by the presence of hemicelluloses, as CI also decreased.

Dissolving Ability Evaluation

Dissolution trials were performed in duplicate and analyzed by the carbanilation reaction and monitored by DLS. The particle size measurements during carbanilation for SW, HW, and CL are displayed in Fig. 4.

Only the oxidized samples CL 0.2%NaOH, SW 0.3%NaOH, and HW 0.4%NaOH dissolved more slowly than their corresponding untreated reference sample. When focusing on the final pH after the oxidations (Fig. 3), it can be presumed that a pH around 9 resulted in a better dissolution, at least for hardwoods, and for CL only at the beginning. Yet, the dissolution behavior in DMSO may be affected by the creation of functional groups during the oxidation of the substrates. It is well established that carboxyl groups react at a much slower rate with phenyl isocyanate than alcohols do. The detailed observation in Fig. 4 shows that for the pulps treated at the highest alkalinity, its dissolution was accelerated at the beginning but rapidly became slower after a certain time. This could be an indication of better accessibility to the bulk of the substrate and to the R-OH groups, but slower reactivity at the end with the presence of RCOOH groups.

MWD Profiles of the Oxidized Substrates

Further analyses of the carbanilated samples were done by HPSEC, as described in the methods section. As already shown by intrinsic viscosity values of Table 4, the degradation of the samples after the different oxidation treatments was confirmed by the MWD profiles (Fig. 5) and the acquired SEC data (Table 5), compared to the reference. The dispersity index (Mw/Mn) is an interesting indicator of the narrowness of the MWD profile. Intrinsic viscosity (weight-average), DPw and hydrodynamic radius Rh are also derived for all the samples. Figure 5 displays the MWD profiles of SW, HW, and CL, respectively obtained before and after oxidation.

Fig. 4. Average particle size during carbanilation (oxidation with 2% H2O2 and 0.02% Cu(Phen) using different NaOH doses. REF=untreated pulp)

Table 5. SEC Analysis of the Oxidized Substrates (oxidation with 2% H2O2 and 0.02% Cu(Phen) Using Varied NaOH Doses. REF=untreated pulp)

Fig. 5. Molecular weight distribution (oxidation with 2% H2O2 and 0.02% Cu(Phen) using different NaOH doses. REF=untreated pulp)

Softwood (SW) and hardwood (HW), both containing hemicelluloses and cellulose, displayed a classical bimodal MWD profile. In the reference samples (blue curves), hemicelluloses were present in the range of logM=4.5 to 5, whereas the cellulose fraction appeared at higher molar mass, centered around logM=5.8 to 6. After oxidation (red and green curves), MWD profiles were shifted towards lower molar mass. This is also evidenced by the drop of DPw, IV and Rh (Table). All oxidized kraft pulps (SW and HW) still presented a signal in the hemicelluloses range, revealing that hemicelluloses were not completely removed during oxidation. Additionally, in the ​​intermediate range, between 105.1 and 105.3, whose abundance is increasingly marked, the corresponding DP values, between 240 and 400 (considering a monomer molar mass for the CTC of 519 g/mol), correspond to the characteristic zone of short chains, degraded cellulose. The enlargement of this zone was evident for the three pulps, and in the case of the HW and SW pulps, it tended to gradually mask the initially well-marked separation between long chains of cellulose and short chains of hemicelluloses.

For both kraft pulps, degradation was related to the final pH. The most important degradation occurred at the lowest alkalinity dosage, producing either near neutral or very slightly alkaline medium. Again, this seems to have been under the control of the hydroxyl radical generation flux, which was less important in alkaline medium when copper was complexed. Moreover, the more accessible substrate, HW, tended to increase its dispersity value, showing that oxidation was less selective and operates on all chain sizes, compared to the less accessible SW pulp.

The behavior of CL could be slightly distinguished from that of the kraft pulps. MWD profiles were still monomodal after oxidation, meaning a non-deep attack (which would create a bimodal profile, as already observed with other oxidants). Here, cellulose was more depolymerized at near-neutral pH, as seen by the higher shift of the MWD mode (curve peak), but the difference seemed less under the alkaline conditions. Moreover, at alkaline pH, a higher dispersity was observed. This shows the effect in alkaline medium, where the solids were more swollen and accessible. However, the reaction was more limited due to a lower flux of radicals, and it was less selective with respect to chain size and accessibility.

No modifications in terms of fiber morphology and water retention values were observed (results not presented). It seems that the oxidation treatments did not deeply affect the substrates at the fiber morphological level.

Comparison of the Oxidation Systems: H2O2/Cu(Phen) vs. H2O2 Alone

The presence of Cu(Phen) with peroxide generates a regular flow of radicals, with a strong dependence on the pH. H2O2 alone generates much less radicals, especially in alkaline medium (the usual conditions for pulp bleaching). Therefore, for the HW (the most accessible substrate), a pH close to neutral was selected, since at this pH radical’s formation takes place with or without Cu(Phen). For the SW, a slightly alkaline pH was selected since it favors the substrate accessibility while keeping some radical generation in the presence of Cu(Phen).

Table 6. Comparison of the Final Parameters When Using or Not 0.02% Cu(Phen) in the Oxidations with 2% H2O2 (REF= untreated pulp)

For CL, the choice of the pH was dictated by the poor substrate accessibility except at alkaline pH where swelling occurs, as discussed in the previous paragraphs. For the mentioned trials, all the samples showed an increased H2O2 consumption (higher than 97%), and with Cu(Phen), the final pH decreased, and the intrinsic viscosity was more affected, as shown in Table 6. It could also be noticed that the R18 values increased, and the CI was slightly affected. Table 7 presents the information extracted from SEC.

Table 7. Additional Information Extracted from SEC and HPLC Analyses (REF= untreated pulp)

As a general observation, cellulose depolymerization was accentuated in the presence of Cu(Phen), by observing IV values. For the three samples studied, the DPw decreased after both kinds of oxidations. Degradation was generally higher in the presence of Cu(Phen). In the case of the HW, this was revealed by the increase of the dispersity index, which caused an increase of DPw, being influenced by the higher proportion of long chains (as seen on the MWD profile, Fig. 5(B)). The same tendency can be observed looking at the intrinsic viscosity values (IV) in Table 7. The values of hydrodynamic radius confirm what was observed with the particle size measurements by DLS during carbanilation, showing an average particle size below 100 nm at the end of the derivatization. No significant changes were observed in the MWD, and the dissolution behavior of the samples oxidized in the mentioned conditions without Cu(Phen). (Results not included in the paper).

No important formation of carboxyl groups was observed on the three substrates (Table 6). The more significant enrichment could be observed for HW, possibly linked to its better accessibility/reactivity observed above, and possibly the type of hemicelluloses in each sample. Perhaps part of the polysaccharides was effectively carboxylated but also depolymerized and solubilized, and for these reasons no important changes could be observed.

It is important to highlight that for dissolving applications, a high purity is required. But for microfibrillated cellulose (MFC) applications, a fair amount of hemicelluloses may enhance the microfibrillation (Foster et al. 2018), and the target may be the creation of carboxyl groups to favor wettability, swelling, and cellulose final dissolution in water or in a polar medium (Vera-Loor et al. 2022). In such cases, the fastest dissolution measured by cellulose carbanilation may not be the right indication for the best dissolution behavior. Brightness values above 90% ISO were achieved, as well as cellulose content above 90%, when referring to R18 trials. However, the cellulose was strongly degraded by the oxidation conditions. The range of cellulose purity for the kraft pulps was not reached, and the intrinsic viscosity for viscose applications (dissolving pulps), ideally in the range of 400 to 500 mL/g, was only reached for some samples. Considering that IV initial values in the present study were low, it is fair to believe that dissolving-grade characteristics could be achieved in the case of higher IV initial values.

Dissolving Ability in the NaOH:urea:water System

To finish the analysis described in the methods section, the solubility in the NaOH:urea:water system was studied, and the results are presented in Fig. 6. In this solvent system, solubility was dominated by the cellulose DPv (measured in CuED), which is also the general tendency presented in Fig. 6. For the three reference samples, a higher DPv implies lower solubility, and solubility varied in the expected order of CL < SW < HW. Nevertheless, the two kraft pulps values were close to each other and quite far from the CL pulp, which was hardly dissolved. This shows the significant impact of the presence of hemicelluloses in the kraft pulps on the solubility in this system.