SO₂–ethanol–water (SEW) fractionation of spruce: kinetics and conditions for paper and viscose-grade dissolving pulps

Abstract

The study describes SO₂–ethanol–water fractionation of spruce as a promising basis for future Biorefineries. The dissolution kinetics of lignin and hemicelluloses as well as the depolymerisation kinetics of cellulose during the fractionation process are expressed in terms of the H-factor, a parameter combining fractionation temperature and duration. The raw material moisture content is shown to have a very small effect on the fractionation kinetics. The liquor-to-wood ratio has little effect on the process during the (delignification) bulk phase, while in the residual phase the effect becomes pronounced. During the latter phase, lower liquor-to-wood ratios lead to higher residual lignin content, lower residual hemicelluloses and pulp viscosity, but higher hemicelluloses removal selectivity. Higher delignification rates and selectivities are obtained at ethanol-to-water ratios close to 1 : 1 (w/w) and high SO₂ concentrations (≥12 w/w%). Based on the presented data, the optimum fractionation conditions for the production of paper and viscose-grade dissolving pulps are quantified.

---

1 Introduction

1.1 Background

We consider SO₂–organic solvent–water (SOW) fractionation a promising platform for future Biorefineries. The combination of these chemicals provides a unique ability to fractionate lignocellulosic materials into pure cellulose, hemicellulose sugars and lignin. The organic solvent can be represented by low-molecular weight alcohols (methanol, ethanol, propanols, etc.), ketones (acetone, methyl isobutyl ketone, etc.) and other compounds.¹ SO₂ is responsible for lignin sulfonation and depolymerisation leading to its dissolution promoted by the presence of the organic solvent. SO₂ solvates (pK_a1 for SO₂·H₂O in water is about 3.5 at 150 °C)^2a and especially lignosulfonic acids (pK_a in water is close to 1)³ are responsible for the increased acidity of the system, affecting the kinetics of all relevant reactions. The major functions of the organic solvent are: 1. Fast transport of SO₂ to the reaction sites (α-carbons in lignin) throughout the entire cell wall thickness driven by high surface tension differences, and 2. Moderating the acidity of the system to satisfy the kinetic and selectivity requirements. This unique chemistry leads to the distinct advantages of the process – very fast impregnation, near-full yields of the biomass components (cellulose, hemicellulose sugars and lignin), and relatively easy and complete recovery of the fractionation chemicals (SO₂ and organic solvent) by distillation.

Lignocellulosic biomass is known to exhibit substantial variation in its properties in terms of composition and structure of the individual polymers as well as their distribution in the cell wall. One consequence is that processes suggested for lignocellulosic Biorefineries are often biomass type-specific, with many not suitable to softwoods prevailing in the Northern hemisphere. SO₂–organic solvent–water process is among few (if any) exhibiting exceptional flexibility towards biomass type, with the following species and biomass wastes having been successfully utilized – Spruce,^4–7 Red pine,⁴ Jack pine,⁴ Loblolly pine,⁴ Balsam fir,⁴ Douglas fir,⁴ Larch,^5,6,8 Birch,^4,5 Beech,^7,9 Aspen,^4,5 Poplar,^9,10 Sugar maple,⁴ Wheat straw,⁷ Oil palm empty fruit bunch,¹¹ Sugarcane straw,¹² low-quality biomass such as forest residues.^13,14

Owing to this omnivorous nature of the process, two principal applications can be suggested. In the first approach, industrial wood chips and also possibly non-wood feedstocks are used to produce cellulose-based materials such as papermaking pulp,¹⁵ dissolving pulp,¹⁶ nanofibrillated cellulose, etc. Dissolved hemicellulose sugars are utilized as a source for chemical and biochemical transformations, for instance, yeast fermentation to bioethanol. The primary requirement in this approach is to produce pure cellulose with specified limits of impurities – lignin and (for dissolving pulp) hemicelluloses – while maintaining a high enough cellulose degree of polymerisation (reflected in high intrinsic viscosity of the solutions in cellulose solvents such as cupriethylenediamine) to achieve the required final product (paper and dissolving pulp) properties.

In the second approach, low-quality lignocellulosic biomass, for instance, forest residues and also herbaceous crops, is used. Here, cellulose amounts and quality are low, and in case of forest residues it is more challenging to ensure high cellulose purity. Therefore, the cellulose residues after the fractionation are (enzymatically) hydrolysed to glucose, which can be utilized alone or after combining with the hemicellulose sugar stream. At present, yeast fermentation to bioethanol seems to be the most feasible application, and the technology can be easily adopted from the acid sulfite industry practice. Successful acetone–butanol–ethanol (ABE) fermentations of SO₂–ethanol–water liquors by Clostridia have also been reported.¹⁷

In both approaches, lignin is produced in two fractions – a low-sulfonated fraction, obtained after ethanol evaporation from the spent liquor, as well as lignosulfonates which are separated from the solution remaining after utilization of the dissolved sugars.

SOW fractionation is currently used on a demonstration scale of several tonnes of biomass per day to hydrolyse lignocellulosics to sugar monomers in a patented Biorefinery process termed AVAP by American Process Inc.¹⁸

1.2 The content of the present paper

The present paper addresses the following aspects of SO₂–organic solvent–water fractionation of spruce with ethanol representing the organic solvent (SEW process) – the effectiveness of wood impregnation, kinetics of the major reactions in terms of the development of the solid residue composition, effects of the liquor-to-wood ratio and chemical charge, ethanol-to-water ratio, in particular. Based on this information, the conditions corresponding to the paper- and dissolving-grade pulps are recommended.

1.2.1 The effectiveness of wood impregnation. In the presence of organic solvent, fractionation chemicals (SO₂ hydrates and solvates) are said to penetrate the cell wall tissue very fast, and no separate impregnation stage is needed. The information on the amounts of so-called undercooked wood (or “rejects”) will be an important proof for this thesis. Also, moisture content of the supplied raw material may vary due to transportation and storage, and penetration of the cooking agents inside the raw material may be affected. However, a viable fractionation process should be able to digest raw materials with a wide range of moisture content. For SO₂-free ethanol–water cooks, Kleinert¹⁹ demonstrated that drying wood before fractionation results in decreased delignification rates (the cooking liquor composition was adjusted to take into account the dilution with water introduced from wood). The effect was increasing in the order: air-drying to 30% moisture (8% decrease in delignification rate) < freeze-drying (21% decrease in the rate) < drying at 90 °C for 24 h (68% decrease in the rate). The effect of the raw material moisture content on SEW fractionation is presented in Section 3.1 of this paper.

1.2.2 Fractionation kinetics. Studies on the kinetics of SEW fractionation were conducted earlier. The effects of temperature²⁰ and different biomass species⁷ were presented. In this paper (Section 3.2) we elaborate further on the temperature effect on delignification, cellulose hydrolysis and hemicellulose removal for three spruce chip batches (I, II and III) and introduce the usage of H-factor concept to combine the effects of temperature and time and allow for prediction of the cellulose residue properties at different fractionation intensities.

1.2.3 Liquor-to-wood ratio effect. Liquor-to-wood ratio is a crucial parameter directly affecting energy efficiency and operational costs, as well as water and organic solvent charges governing waste-water treatment load and costs. Acceptable higher limit for the liquor-to-feedstock ratios for a fractionation process in a lignocellulosic Biorefinery is close to 3–4 L kg⁻¹. To our knowledge, no information on the effect of liquor-to-feedstock ratio on the SO₂–organic solvent–water fractionation (at other conditions being constant) is available in literature. The effect is assessed in Section 3.3.

1.2.4 Ethanol-to-water ratio effect. Fractionation behaviour can be tuned considerably by adjusting the organic solvent-to-water ratio in the fractionation liquor. This is in part due to the fact that changing organic solvent-to-water ratio results in a change in the apparent acidity of the fractionation system (also referred to as Hammett acidity function or proton availability²¹). Many studies at moderate temperatures (generally at 25–50 °C) indicate a monotonous decrease in dissociation constant of weak acids, K_a, with increasing organic solvent-to-water ratio (for organic acids, see for example ref. 22 and 23; for solvated SO₂, see ref. 24). This is presumably due to the fact, that dissociation of acids is less energetically favourable in the less polar environment of organic solvent-rich solutions,²⁵ as well as due to competition for protons between organic solvent and water. Nevertheless, strong acids (HCl) were shown to exhibit a minimum acidity in the region 30–70 mole % organic solvent (in water).^26,27 Unfortunately, no information is available for the temperatures used in fractionation of lignocellulosics. However, cellulose hydrolysis can be used as a measure of the apparent acidity. The monotonous increase in cellulose degree of polymerisation (DP) with increasing ethanol-to-water ratio was observed for ethanol–water fractionation of poplar (L/W ratio 8 L kg⁻¹, 195 °C, 210 min).²⁸

In addition to the acidity, the organic solvent-to-water ratio affects the solubility of lignin and hemicellulose fragments adding complexity to the delignification and hemicellulose removal patterns. For instance, ethanol–water lignin exhibits maximum solubility in solutions containing about 70% ethanol and 30% water, while at 30% ethanol and 70% water, the solubility is only about one fifth of the maximum.²⁹

The (limited) available data on the fractionation response to organic solvent-to-water ratio indicates that delignification rate exhibits a maximum at a certain organic solvent content presumably at the point where the hydrolysis-to-condensation reaction rates ratio is at its maximum. For ethanol, the maximum is observed at EtOH/H₂O ratio of about 1 : 1 (w/w), both for SO₂-free (poplar, L/W ratio 8 L kg⁻¹, 195 °C, 210 min)²⁸ and SEW (spruce, L/W ratio 5 L kg⁻¹, 15% SO₂, 145 °C, 15 min;⁶ larch heartwood, L/W ratio 6 L kg⁻¹, 15% SO₂, 135 °C, 25–105 min)⁸ systems. For the SO₂-free acetic acid–water system, the maximum is shifted to AcOH/H₂O ratio of about 7 : 3 (birch and aspen, 145 °C, 3 h)³⁰ presumably due to lower pK_a of acetic acid, compared to ethanol.

It should be noted that the organic solvent assumes a role similar to that of a base in acid sulfite pulping in mitigating the acidity of the system. Therefore, similar patterns between the effects of organic solvent-to-water ratio (in SOW fractionation) and amount of bound SO₂ (in acid sulfite pulping) can be expected.³¹ The effect of ethanol-to-water ratio is discussed in Section 3.4.

1.2.5 SO₂ concentration effect. The effect of SO₂ concentration is pronounced and was earlier examined at a L/W ratio of 6 L kg⁻¹ for spruce wood³¹ and larch heartwood.⁶

At SO₂ concentrations lower than 12% the bulk delignification rate is proportional to SO₂ concentration. The following equation was suggested to express the effect of SO₂ concentration on delignification kinetics:³¹

where A_Lig – pre-exponential factor for the bulk delignification rate constant, k_Lig, L mol⁻¹ min⁻¹; [Lig] – residual lignin content, % on wood; [SO₂] – SO₂ concentration in the fresh liquor, mol L⁻¹; R – gas constant, 8.314 J mol⁻¹ K⁻¹; T – temperature, K; E_A,Lig – activation energy of delignification, J mol⁻¹.

Despite the warnings of Schorning (the pioneer of SO₂–organic solvent–water process, 1957;⁹ beech, 50 w/w% methanol–water, L/W ratio 10 L kg⁻¹), increasing temperature even at relatively low SO₂ concentration (down to 5%) does not lead to a decrease in delignification rate, the latter expected due to lignin condensation. Interestingly, the activation energy even increases from 93 kJ mol⁻¹ at 15% SO₂ to 115 kJ mol⁻¹ at 5% SO₂ (larch heartwood, 50 w/w% ethanol–water, L/W ratio of 6 L kg⁻¹, 125–155 °C).⁶

The acidity of SEW systems both in the absence of lignin (expectedly proportional to ) and in its presence seem to change very moderately with increasing SO₂ concentration. However, the rates of carbohydrate reactions in the initial phase of fractionation are strongly dependent also on delignification rates as part of hemicelluloses is removed as ligno–carbohydrate complexes (LCCs) while cellulose is protected by lignin.³¹

In Section 3.5 we present some results on SO₂ concentration effect at a low liquor-to-wood ratio of 3 L kg⁻¹.

Finally, Sections 3.6 and 3.7 provide the information on the selectivity of hemicelluloses removal and brightness of the pulps.

2 Materials and methods

2.1 Raw material

Three batches of Norway spruce (Picea abies) chips (denoted as I, II and III) were screened using the screens O45; //8; //6; //4 and//2 mm, and fractions 2–4 mm and 2–6 mm thick were collected. Table 1 provides the composition of the raw material.

Table 1 Raw material

Spruce chips batch	Chip thickness, mm	Drying method	Dry matter content, %	Composition, % based on o.d. wood
				Lignin	Cellulose	GGM^b	X^b	Acetone extractives	Ash	1% NaOH soluble “lignin”
a Used for most of the experiments. b GGM = Galactoglucomannan; X = Arabino-4-O-methylglucuronoxylan. c Without 4-O-MeGlcA. n.m. – not measured.
I	2–4	Air-dried^a	93.3	n.m.
II	2–6	Fresh	48.9	27.7	39.9	18.1	7.11	1.81	0.44	0.88
		Air-dried^a	92.8
III	2–6	Fresh^a	67.3	28.9	39.9	15.7	6.77^c	1.31	0.29	1.27
		Air-dried	99.5

---

1% NaOH extraction was performed on acetone-extracted Wiley-mill ground wood at a solution-to-wood ratio of 100 L kg⁻¹, 25 °C and 1 h, at constant stirring. The residue was washed 4 times with deionised water (each at 100 L kg⁻¹, 25 °C and 5 min; final pH about 8.5). The amount of 1% NaOH soluble phenolic structures (“lignin”) was determined using UV absorbance at 280 nm (extinction coefficient 23 L g⁻¹ cm⁻¹, Iakovlev and van Heiningen 2011; spectrophotometer Shimadzu UV-2550) after 2 times dilution with 0.1 M NaOH (blank solution – 0.1 M NaOH).

2.2 Fractionation liquor composition

The fractionation liquors were prepared by injecting gaseous sulfur dioxide into a cold (0 °C, ice bath) mixture of ethanol ETAX A (96.1 v/v%, density 807 kg L⁻¹) and deionised water (see Table 2).

Table 2 Fractionation liquors (for the hypothetical 100% dry wood)

Composition characterisation		Example of preparation of the solution
Mixed^a	As weight ratios SO2/EtOH/H2O	Ethanol ETAX A (96.1 v/v%), mL	Deionised H2O, mL	SO2, g
a This characterisation was used in our previous publications.^7,20,31,32 In the first column v/v% means volumetric percentage based on the sum of the volumes of ethanol and water.
12 w/w% SO2 in 40.0 v/v% ethanol	12 : 30.6 : 57.4	416	584	125
12 w/w% SO2 in 47.5 v/v% ethanol	12 : 36.9 : 51.1	494	506	123
12 w/w% SO2 in 55.0 v/v% ethanol	12 : 43.5 : 44.5	572	428	121
12 w/w% SO2 in 62.5 v/v% ethanol	12 : 50.3 : 37.7	650	350	119
12 w/w% SO2 in 70.0 v/v% ethanol	12 : 57.3 : 30.7	728	272	117
6.0 w/w% SO2 in 55.0 v/v% ethanol	6.0 : 46.4 : 47.6	572	428	56.8
18 w/w% SO2 in 55.0 v/v% ethanol	18 : 40.5 : 41.5	572	428	195

---

It should be noted that due to the absence of base, practically all SO₂ is present in the fractionation liquor in a free form. The concentration is assumed to be nearly constant throughout fractionation, as about 97% of the initially charged SO₂ is recovered from the spent fractionation suspension by distillation.^31,32

2.3 Fractionation conditions

25.0 o.d. g of chips and fresh liquor at a liquor-to-wood ratio of 1–9 L kg⁻¹ (3 and 6 L kg⁻¹ for most experiments) were placed in 220 mL bombs. The bombs were put into a silicon oil bath at 125–165 °C (±1 °C). The treatment duration was 10–420 minutes. The heat-up time of about 8–13 minutes was included in the total reported duration. At the end of the treatment, the bombs were rapidly removed from the bath and put into cold water. After cooling, the liquid phase was separated from the solid residue using a nylon washing bag. The solid residue obtained from each bomb was washed 2 times with 50 mL of 40 v/v% ethanol–water solution at 60 °C and finally 2 times with 500 mL deionised water at room temperature.²⁰ Some experiments were performed in 2–6 bombs, and all the amounts of wash liquors were accordingly increased.

H-factor (for delignification and hemicelluloses removal) was calculated according to eqn (3) (see Section 3.2), using the activation energy value of E_A,Lig = 106.8 kJ mol⁻¹ (equivalent heat-up time 8 min), while H_C-factor (for cellulose hydrolysis) is based on the value E_A,Cel = 164.9 kJ mol⁻¹ (equivalent heat-up time 13 min).

2.4 Analyses of raw material and pulps

The raw material and the pulps were analysed for ash content,^33a lignin and carbohydrate content (using acetone extraction^34a and double-stage acid hydrolysis/HPAEC-PAD;^32,33b 4-O-MeGlcA using acid-methanolysis/GC-FID, as described in ref. 32). The pulps were analysed for yield, reject content (i.e. particles not passing 0.35 mm slots, during the wet screening of pulp), kappa number,^34b viscosity in cupriethylenediamine (CED,^34c pulps with kappa number higher than 35 were pre-subjected to chlorite delignification³⁵), sulfur content.^34d,e Lignin content for some pulps (spruce batch III) was calculated from kappa number using the published equation.³⁶ Viscosity-average degree of polymerisation (DP_v) of cellulose was calculated from intrinsic viscosity in CED based on:³⁷where [η] – intrinsic viscosity of pulps in CED, mL g⁻¹; [Hemi]_pulp – hemicelluloses content of pulp, unit fraction; [Cel]_pulp – cellulose content of pulp, unit fraction.

The details of the procedures are given elsewhere.^7,31,32

The term “delignification selectivity” refers to the extent of delignification related to the extent of polysaccharide reactions – hemicelluloses removal and cellulose hydrolysis. Delignification selectivity is higher when higher residual hemicelluloses content/cellulose DP is observed at the same residual lignin content.

The term “selectivity of hemicelluloses removal” refers to the extent of hemicelluloses removal related to the extent of cellulose hydrolysis.

2.5 Analyses of spent fractionation liquors

Selected spent liquors were analysed for total carbohydrate content using acid-methanolysis/GC-FID as described earlier.³²

3 Results and discussion

3.1 Impregnation of wood: share of rejects and effect of raw material moisture content

Efficient impregnation of the raw material with the cooking liquor is crucial for successful fractionation. SO₂–ethanol–water process does not employ a separate impregnation stage due to excellent penetration of lignocellulosics by ethanol–water solution. The efficiency of impregnation may be characterised by rejects content of pulps and by the effect of raw material moisture content on fractionation.

All produced SEW pulps which passed defibration point (i.e. the point along the fractionation, when fibres become liberated from each other; occurring at an H-factor of about 13 h, at SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight, see the next section) exhibited very low reject content, generally in the range 0.001–0.01%.

Effect of moisture content was studied using two batches of spruce chips (II and III) which were cooked both fresh and air-dried. The amount of water in the green chips was considered in the calculation of the fresh liquor composition and charge. Table 3 shows that both the green and air-dried chips are successfully fractionated, and the initial moisture content has little effect on the properties of the resulting solid and liquid phases provided the composition and amount of the fresh liquor are corrected considering the amount of water in the chips.

Table 3 Influence of spruce chips dry matter content on SEW fractionation (SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight, 135 °C)^a

Raw material, fractionation conditions	Spruce II, L/W 6 L kg^−1, 80 min (H = 23.1 h)		Spruce III, L/W 6 L kg^−1, 100 min (H = 29.5 h)		Spruce III, L/W 3 L kg^−1, 100 min (H = 29.5 h)
a n.m. – not measured.
Dry matter content of the raw material, %	48.9	92.8	67.3	99.5	67.3	99.5
Solid residue yield, %	50.1	51.2	47.0	49.2	46.8	48.8
Kappa number	32.0	34.1	24.0	27.4	30.4	34.7
Viscosity in CED, mL g^−1	n.m.	n.m.	1016	991	877	835
Solid residue composition, g per 100 g solid residue
Cellulose	82.1	80.1	n.m.
Hemicelluloses	11.7	12.5	9.6	9.1	7.2	7.3
Lignin	5.9	6.3	4.6	5.1	5.6	6.4
Spent liquor carbohydrates composition, g L ^−1 (as monosaccharides)
Mannose	14.3	13.3	n.m.
Xylose	6.7	6.6
Glucose	3.8	3.5
Total carbohydrates	34.4	32.6

---

Therefore ethanol is able to impregnate very fast both dried and green chips. The reason for this phenomenon is the so-called Marangoni effect whereby a bulk flow (convection) occurs from fluid in a low surface area region (ethanol–water mixtures having relatively low surface tension, close to 30 × 10⁻³ N m⁻¹ for 50 v/v% ethanol–water at 20 °C)³⁸ to that in a high surface area region (cell-wall capillaries and water, 72.5 × 10⁻³ N m⁻¹ at 20 °C).^39,40 However, yields and kappa numbers are somewhat lower for the green chips compared to air-dried chips, which agrees with the results of Kleinert for ethanol–water fractionation.¹⁹ It can be explained by hornification, i.e. irreversible decrease in the accessibility of wood matrix resulting from drying.

Summary: Both green and air-dried spruce chips are impregnated very fast with the SEW fractionation liquor, and produce pulps with negligible amount of rejects. Both phenomena are due to the presence of ethanol.

3.2 Fractionation kinetics: effect of temperature and time

3.2.1 Delignification kinetics. Earlier we reported the preliminary data on the effects of temperature and time on SEW fractionation of spruce²⁰ where first-order bulk delignification followed by slower residual phase was presented. Since then more insight into delignification, including precise correlation between residual lignin content and kappa number,³⁶ has been attained. Now, based on these and additional experiments, the activation energy value of spruce bulk delignification can be updated – 106.8 kJ mol⁻¹. This value is similar or slightly higher than that of acid sulfite delignification (67–105 kJ mol⁻¹).^41–43 Based on this value, H-factor concept⁴⁴ combining the effects of temperature and time on delignification can be introduced to SEW process:where R – gas constant, R = 8.314 J mol⁻¹ K⁻¹; E_A,Lig – activation energy of delignification, J mol⁻¹; T(t) – temperature development, K, with cooking time, t, h.

Note, that H-factor 50 h means 156 min of isothermal fractionation at 135 °C, or 51 min at 150 °C, or 18 min at 165 °C. Fig. 1a presents the delignification of three spruce chips batches as a function of H-factor. It is seen that all three batches are delignified with the same rate during the first, faster, phase, called “bulk” delignification. Thus, the residual lignin content can be predicted down to about 1–2% (based on wood) using the following equation:

[Lig] = [Lig]₀ exp(−0.095H), R² = 0.98, (4)

where [Lig] – lignin content, % on original wood, [Lig]₀ – lignin content in original spruce wood, %; in this work [Lig]₀ = 27.7% (spruce batch II) and 28.9% (spruce batch III).

Fig. 1 SEW fractionation kinetics for spruce batches I, II and III at SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight (see Table 2) and liquor-to-wood ratio 6 L kg⁻¹. (a) delignification, (b and c) cellulose hydrolysis, (d–f) hemicelluloses removal, (g and h) delignification selectivity. H-factor (for delignification, (a), and hemicelluloses removal, (f)) is based on activation energy value of E_A,Lig = 106.8 kJ mol⁻¹, while H_C-factor (for cellulose hydrolysis, (b and c)) is based on the value E_A,Cel = 164.9 kJ mol⁻¹. Equivalent heat-up time for delignification and hemicelluloses removal is 8 min, while that for cellulose hydrolysis is assumed to be 13 min. Fig. 1a–c include some of the data corrected and recalculated from ref. 20 (batch I) and ref. 31 (batch II). Mannan and xylan represent the anhydrosaccharides alone.

Kappa number 30 typical for unbleached softwood paper pulps is achieved at an H-factor of about 24 h (still within bulk delignification phase, for all examined spruce batches).

However, in the residual delignification phase (starting from H = 25–30 h), chips batch III is delignified substantially slower compared to batches I and II. The lowest kappa number values obtained for batch III are 9–10, while for batches I and II, kappa number values of 5.5–8 were common (L/W ratio 6 L kg⁻¹). This may be related to the presence of bark containing polyphenolic acids which are not removed by any solvent, other than alkali.⁴⁵Table 1 shows that spruce III contains considerably higher amounts of 1% NaOH soluble phenolic substances compared to spruce II (1.27 vs. 0.88%).

The residual phase becomes very important when the process is meant for production of highly pure cellulose, such as viscose-grade dissolving pulps. In acid sulfite process commonly kappa number values of close to 5 are obtained for unbleached pulp (less than 0.5% lignin, based on wood) at a relatively low liquor-to-wood ratio of 3 L kg⁻¹. Fig. 1 shows that residual SEW delignification seems to be less successful, at least in case of chips batch III.

It should be noted that although the H-factor values were calculated based on bulk delignification rates, H-factor appears to describe well also the residual phase (for every spruce batch), which indicates similar activation energy values for the residual delignification.

3.2.2 Cellulose hydrolysis kinetics. No detectable cellulose dissolution occurs during SEW fractionation of spruce, in line with the earlier observations.³² Nonetheless, glycosidic bonds in cellulose chains are randomly cleaved through acid-catalysed hydrolysis. This process can be described in terms of zero-order kinetics assuming that acidity does not change with time at each temperature:⁴⁶where DP_n – number-average DP of cellulose at time t; DP_n,0 – DP of the original spruce cellulose; k^′′_Cel(T) and k^′_Cel(T) – real and composite cellulose hydrolytic cleavage rate constants at temperature T.

Assuming a constant polydispersity index, P = DP_w/DP_n, the equation can be rewritten for weight-average DP values, DP_w:

The DP_w values are known to be close to the viscosity-average DP values.⁴⁷ In this work the latter were calculated from the intrinsic viscosity of pulp solutions in cupriethylenediamine (CED). We used the correlation between the viscosity of cellulose tricarbanilates in CED and their DP (determined by light scattering) obtained by Evans and Wallis⁴⁸ and corrected for the presence of hemicelluloses³⁷ (see eqn (2)).

Similarly to delignification, H_C-factor concept can be applied to cellulose hydrolysis (activation energy 164.9 kJ mol⁻¹: H_C-factor = 500 h translates into 313 min of isothermal fractionation at 135 °C, 55.7 min at 150 °C or 11.2 min at 165 °C).

It is seen that all the available data (both cellulose DP and viscosity) is described well by the H_C-factor (Fig. 1b and c). In fact, in the whole range of conditions (from wood to H_C-factor of about 1000 h) cellulose hydrolysis can be represented as a single-stage process (Fig. 1b), although spruce batch III deviates somewhat from the line obtained for spruce batch I. Extrapolating to H_C = 0 leads to DP_w,0 of about 10 000 corresponding roughly to DP of native wood cellulose. DP_w values can be thus determined as follows:

Pulp viscosity can be roughly predicted down to about 500 mL g⁻¹ using the equation:

[η] = −1.73H_C + 1230, R² = 0.96, (8)

It should be noted that at SO₂ concentration lower than 12%, initial delays in cellulose SEW hydrolysis were reported and explained by the protection of lignin, which is removed increasingly slower with decreasing SO₂ concentration (see eqn (1)).³¹

The recommended lower limits of the viscosity for the unbleached softwood paper and dissolving grade pulp, about 700 and 500 mL g⁻¹, correspond to H_C-factors of 300 and 420 h, respectively.

3.2.3 Hemicelluloses removal kinetics. Hemicelluloses removal proceeds in three phases – initial, bulk and residual (see Fig. 1d and e). More than half of hemicelluloses are removed in the initial, relatively short, phase. It was shown earlier that in this phase hemicelluloses are likely to be removed together with lignin as ligno–carbohydrate complexes (LCCs).³¹ It should be noted that at higher temperatures the transition point (initial-to-bulk phases) occurs at lower cooking time and at lower residual hemicellulose content.

During the second phase (called “bulk” phase) the removal is substantially slower and is first order in mannan and xylan (i.e. the linear region in Fig. 1d and e). The lower removal rate of glucomannan and xylan during the bulk phase may be related to the morphology of this residual fraction. In effect this fraction consists of glucomannan closely associated with cellulose already in the original wood and that which is “crystallised” onto cellulose during the initial phase. Also in the course of the initial phase most of the labile side units of the wood polysaccharides as well as pectins are removed.

The first-order behaviour during the bulk phase is consistent with S_N1 acid hydrolysis, and the kinetics can be expressed as follows:

where [Hemi] – mannan and xylan content of the pulps, % based on wood; k^′_Hemi and k_Hemi – real and composite rate constants for hemicelluloses removal; E_A,Hemi – activation energy of hemicelluloses removal, J mol⁻¹; A_Hemi – pre-exponential factor.

The rate constants for the hemicelluloses removal are given in Table 4. It is noted that at each temperature the ratio of the removal rates for glucomannan and xylan is constant, about 1.2, indicating the same temperature dependence for both. However, the opposite, i.e. xylan faster hydrolysed than glucomannan, is often observed in acid sulfite cooking.^2b,49a Also methyl-β-D-xylopyranoside is hydrolysed 1.6 times faster than methyl-β-D-mannopyranoside.⁵⁰ However, a similar behaviour as seen in the present study was observed for spruce acid sulfite dissolving pulp.^49b Primarily, this highlights that the hydrolysis rates of model compounds do not necessarily correspond to those of the related wood polysaccharides. The presence of 4-O-methylglucuronide side units considerably stabilises wood xylan in both SEW and acid sulfite cooking. Regarding glucomannan it was shown that its retention is highly dependent on the conditions at the beginning of cooking. For instance, mild neutral or slightly alkaline treatment (in a two-stage sulfite cooking) leads to complete deacetylation and galactose units removal while the DP is not significantly reduced. The polymeric state of glucomannan (i.e. linear and long-chain) is said to be the most favourable for crystallisation onto cellulose fibrils and subsequent stabilisation against acid hydrolysis.⁵¹ In acid sulfite cooking the conditions in the beginning are milder due to long impregnation at low temperature compared to SEW cooking where the temperature is increased within 15 minutes to the cooking level. Therefore glucomannan stabilisation should decrease in the order: two stage sulfite > acid sulfite > SEW. On the other hand, crystallisation of xylan is less likely⁵¹ due to the presence of uronic acid side units, and because of the morphology of softwood cell wall featuring intimate coverage of cellulose by glucomannan, rather than xylan.⁵² The present results are consistent with these considerations.

Table 4 Hemicelluloses removal kinetics (spruce I and II, bulk phase, SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight)

Temperature, °C	Mannan	Xylan
	k Man × 10^3, min^−1	k Xyl × 10^3, min^−1
135	8.46	6.91
145	16.4	12.7
155	37.6	31.0
165	72.9	60.7
E A,Hemi, kJ mol^−1	108.4	110.1
A Hemi, min^−1	6.07 × 10^11	8.09 × 10^11

---

By applying the Arrhenius equation, the activation energy values for glucomannan and xylan removal in the bulk phase were obtained, 108.4 and 110.1 kJ mol⁻¹, respectively, with excellent correlation coefficients (R² = 0.99). These values are very close not only to each other but also to that for delignification, 106.8 kJ mol⁻¹. Nevertheless, the straight lines in Fig. 1d and e do not intersect at the equivalent heat-up time (8 minutes) which otherwise would have resulted in a constant delignification selectivity at all temperatures. This is especially important for glucomannan removal (Fig. 1d), and leads to substantially different selectivities at different temperatures as shown in Fig. 1g (apparent activation energy for glucomannan removal is therefore higher and approaching that of cellulose hydrolysis). This phenomenon can be also related to the lower stabilisation degree of glucomannan at higher temperature. In case of xylan the difference in selectivity is smaller (Fig. 1h) as the lines in Fig. 1e approach each other closer at the equivalent heat-up time (8 minutes). The behaviour is also reflected in the H-factor plot (using E_A,Lig = 106.8 kJ mol⁻¹, Fig. 1f): H-factor is (conditionally) applicable only to xylan removal. It would be appropriate here to cite Rusten,⁴⁷ who also argues that the apparent activation energy value for hemicelluloses hydrolysis in spruce acid sulfite pulping is higher than that for cellulose hydrolysis, with the reason being higher stabilisation of glucomannan at lower temperature due to better crystallisation on cellulose.

Therefore, the delignification selectivity is better at lower temperatures mostly due to the fact that more glucomannan is preserved after the initial removal phase (Fig. 1g and h). This is coherent with the earlier observations^20,47 that at lower temperature, higher pulp yield is obtained at the same kappa number.

Finally, the third (“residual”) phase is observed: 240 and 320 min at 135 °C points are relatively far from the line corresponding to the bulk phase (see Fig. 1d and e). The difference is especially high for mannan which can be again related to its stabilization through crystallisation on cellulose.⁵¹ The evidence for the residual phase is, however, limited to only these three points.

Paper-grade pulps correspond to the bulk hemicellulose removal phase, while dissolving pulps span from the end of the bulk phase to the residual phase. For unbleached softwood viscose-grade acid sulfite pulps, common combined amounts of anhydromannose and anhydroxylose are close to 5% (based on pulp). This value is obtained in SEW fractionation (SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight, L/W ratio 6 L kg⁻¹) at an H-factor of about 50–75 h (higher H-factors are needed at lower temperatures).

Summary: H-factor concept can be applied to bulk delignification (E_A,Lig = 106.8 kJ mol⁻¹) and cellulose hydrolysis (E_A,Cel = 164.9 kJ mol⁻¹). The activation energy values for bulk glucomannan and xylan removal are E_A,Man = 108.4 kJ mol⁻¹ and E_A,Xyl = 110.1 kJ mol⁻¹. Nonetheless, the apparent activation energy values are higher, especially for glucomannan, because initial removal is higher at higher temperatures. This is attributed to higher glucomannan stabilisation at lower temperatures, possibly through “crystallisation” onto cellulose.

Residual delignification proceeds slower for spruce batch III, compared to batches I and II, for unknown reason. A possible explanation is the presence of non-lignin phenolic substances in batch III.

3.3 Effect of liquor-to-wood ratio

The reduction of the liquor-to-wood ratio is most important for mill operation because of the energy and chemicals savings. The effect of this parameter on the fractionation results was studied for the fractionation conditions corresponding both to bulk and residual delignification phases.

Table 5 includes the properties of the solid and liquid phases obtained from the fractionation of spruce at liquor-to-wood ratio values of 6 to 1 L kg⁻¹ (SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight, H-factor = 23.1 h, bulk delignification phase). No difference in the fractionation rate and composition of the resulting streams is observed when decreasing the liquor-to-wood ratio to 3 L kg⁻¹ indicating that the process is governed by SO₂ concentration in the liquid phase rather than by SO₂ charge. Further reduction in liquor-to-wood ratio leads to somewhat slower delignification due to SO₂ depletion and/or due to increase in the acidity because of higher lignosulfonic acid concentration. The carbohydrate retention in the solid phase is somewhat lower at lower L/W ratios due to the increased acidity, while their concentration in the liquid phase is practically constant. At a liquor-to-wood ratio of 1 L kg⁻¹ delignification is significantly reduced possibly due to the pronounced condensation favoured by high acidity and/or because of lignin solubility limitation. Dehydration reactions may be responsible for a somewhat lower carbohydrates yield in the liquid phase.

Table 5 Effect of liquor-to-wood ratio on spruce fractionation (spruce II, SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight, 135 °C, 80 min, H-factor = 23.1 h)^a

Liquor-to-wood ratio, L kg^−1	6	3	2	1
a n.d. – not defibrated solid residue.
SO2 charge, % on wood	69	34	23	11
Solid residue yield, % on wood	51.5	50.0	49.8	55.4
Kappa number	33.5	34.8	44.7	n.d.
Lignin content, % on wood	3.14	3.12	3.94	9.01
Lignin-free yield, % on wood	48.4	46.9	45.9	46.4
Spent liquor carbohydrates composition, g per 100 g wood (as anhydrosaccharides)
Mannose	9.3	9.4	9.2	9.0
Xylose	4.0	4.0	4.0	3.7
Glucose	2.2	2.3	2.3	2.4
Total carbohydrates	20.5	20.6	20.4	19.5

---

Nevertheless, Fig. 2 shows that in the region of more severe fractionation conditions (residual delignification), the effect of L/W ratio on delignification becomes pronounced resulting in 5–10 units difference in kappa number (L/W ratios 3 vs. 6 L kg⁻¹).

Fig. 2 Liquor-to-wood ratio effect on SEW fractionation of spruce batch III at SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight: (a and b) kappa number; (c and d) pulp intrinsic viscosity in CED; (e and f) residual hemicelluloses. Residual hemicelluloses in (e) are calculated as a sum of anhydromannose and anhydroxylose.

Realizing the importance of liquor-to-wood ratio for fractionation processes, a set of liquor-to-wood ratios (at two sets of conditions) was tested: 3, 4.5, 6 and 9 L kg⁻¹ (Fig. 2b, d and f).

From Fig. 2b, it can be seen that increasing liquor-to-wood ratio (for both sets of conditions) leads to higher delignification because of two possible reasons. Firstly, at higher L/W ratio the acidity of the cooking liquor is lower leading to the decrease in harmful lignin condensation reactions. Secondly, at lower L/W ratios lignin solubility may appear as a limiting factor in delignification. Table 6 provides sulfonation degree of residual lignin of the pulps. It is seen that sulfonation degree at a L/W ratio of 3 L kg⁻¹ is somewhat higher than that at higher L/W ratios which also suggests the lower dissolution-to-sulfonation rates ratio for L/W ratio of 3 L kg⁻¹. It has to be also noted that Mg-based acid sulfite cooking results in much lower residual lignin content even at low L/W ratios. This can be explained by much higher sulfonation degrees as compared to SEW fractionation (S/C9 ratios of about 0.3–0.4 for the residual lignin^2c).

Table 6 Sulfur content of residual SEW lignin (spruce III, SO₂/EtOH/H₂O = 12 : 43.5 : 44.5, by weight, 135 °C, 240 min, H-factor = 74.3 h)

L/W ratio, L kg^−1	Sulfur, % on pulp	Lignin, % on pulp	S/C9^a
a Lignin monomeric unit is assumed to possess a molecular weight of 190 g mol^−1.^2d
3	0.0714	4.01	0.106
4.5	0.0394	2.70	0.0866
6	0.0292	2.53	0.0685
9	0.0310	2.15	0.0856

---

The liquor-to-wood ratio has the opposite effect on carbohydrate reactions compared to delignification (Fig. 2): the lower is the L/W ratio, the higher carbohydrate hydrolysis is observed, and therefore, the lower is the hemicelluloses (both mannan and xylan) retention and the lower is pulp viscosity. The possible reason for the enhanced cellulose depolymerisation and hemicelluloses hydrolysis is the Donnan effect: the distribution of free H₃O⁺ ions between the two phases will be such that the concentration of free H₃O⁺ ions in the cell wall will be higher than the concentration in the liquid phase. Since the lignosulfonic acid group concentration is higher in the case of low L/W ratios, more protons will diffuse inside the cell wall. According to Donnan's theory, the difference in the concentration of H₃O⁺ ions in the two phases can be minimized by the addition of a diffusible salt.

The aforementioned behaviour results in a pronounced negative effect of lower L/W ratios on the delignification selectivity.

Summary: Bulk delignification is unaffected down to a liquor-to-wood ratio of 3 L kg⁻¹. L/W ratio effect becomes pronounced in the later stages of fractionation, corresponding to residual delignification. There, lower L/W ratios lead to higher residual lignin content (possibly due to higher condensation and limited solubility) and lower residual hemicelluloses content and cellulose DP (possibly due to Donnan effect).

3.4 Effect of ethanol-to-water ratio

3.4.1 Delignification. The effect of ethanol concentration on SEW fractionation was studied at three sets of conditions covering different temperatures and L/W ratios (all at 12 w/w% SO₂). The results are shown in Fig. 3.

Fig. 3 Effect of ethanol content (v/v% based on sum of ethanol and water volumes, see Table 2 for clarification; spruce batch III; 12 w/w% SO₂). (a) kappa number; (b) residual lignin; (c) pulp intrinsic viscosity in CED; (d) cellulose DP; (e) residual mannan (as anhydromannose); (f) residual xylan (as anhydroxylose).

For all the condition sets, the obtained curves “kappa number vs. ethanol concentration” have the same character exhibiting a minimum close to 55 v/v% ethanol concentration. The effect of ethanol concentration is in accordance with the previous results obtained for both SEW (spruce, 15% SO₂, L/W ratio 6 L kg⁻¹,⁶) and ethanol–water (poplar, L/W ratio 8 L kg⁻¹,²⁸) fractionation. This could be explained as follows: at high ethanol concentrations the acidity is too low and the hydrolytic cleavage of lignin is relatively slow, while at low ethanol concentrations the acidity is too high which leads to pronounced lignin condensation.

However, it should be noted that at L/W ratio of 3 L kg⁻¹ the concentration of formed lignosulfonic acids is about twice as that at L/W ratio of 6 L kg⁻¹. This leads to about twice higher H₃O⁺ concentration at L/W ratio of 3 L kg⁻¹, which would require higher ethanol concentration in order to bring the acidity to the same level as at L/W 6 L kg⁻¹. The location of the minimum is nevertheless the same.

As discussed above, the difference between the lowest values at L/W ratios of 3 and 6 L kg⁻¹ comprise over 5 kappa number units, which is possibly related to the lignin solubility and Donnan effects.

3.4.2 Hemicelluloses removal and cellulose hydrolysis. Both hemicelluloses (mannan and xylan) removal (Fig. 3e and f) and cellulose hydrolysis (Fig. 3c and d) decrease monotonously with increasing ethanol concentration at all three sets of conditions. This is again in line with the previous results for SEW⁶ and ethanol–water²⁸ fractionation; and explained by the decrease in the acidity of the system with increasing ethanol concentration.

However, it is apparent that both hemicelluloses retention and viscosity are considerably more sensitive to ethanol concentration at L/W ratio of 6 L kg⁻¹ compared to L/W ratio of 3 L kg⁻¹ (Fig. 3c–f). This can be related to the solubility effects (note again considerably lower kappa number values obtained at L/W ratio of 6 L kg⁻¹).

The described patterns result in C-shaped delignification selectivity plots.

Summary: Residual lignin content exhibits a minimum close to 55 v/v% ethanol, while both residual hemicelluloses content and cellulose DP increase monotonously with increasing ethanol concentration. Both effects are presumably due to the decrease in the acidity with increasing ethanol concentration.

3.5 Effect of SO₂ concentration

The effect of SO₂ concentration on fractionation at a L/W ratio of 3 L kg⁻¹ (150 °C, 55 v/v% ethanol–water) observed in the present study is in general similar to that reported earlier for L/W ratio of 6 L kg⁻¹.^6,31

In particular, the delignification rate is significantly lower at 6% SO₂ compared to 12% SO₂ (see ESI, Fig. S1†) presumably due to increase in condensation due to insufficient sulfonation. It is supported by the observation that at 6% SO₂, the pulps and spent liquors have more intense colour compared to those obtained at higher SO₂ concentrations. It is noted that at 6% SO₂ kappa number and lignin content even increase towards the last point, which is another evidence for condensation. This phenomenon of lignin reaccumulation in pulp has not been observed earlier for SEW process (spruce, L/W ratio 6 L kg⁻¹, 135 °C;³¹ larch heartwood, L/W ratio 6 L kg⁻¹, 125–155 °C).⁶ On the other hand, increasing SO₂ concentration from 12 to 18% leads to only moderate increase in delignification rate (similar to ref. 31).

It can be noted that the lower the SO₂ concentration in the SEW liquor, the lower is the rate of cellulose hydrolysis and hemicelluloses removal. The slower hydrolysis of the polysaccharides is explained by two factors.³¹ Firstly, wood polysaccharides are protected by lignin, and the latter is removed much slower at lower SO₂ concentration. Secondly, at lower SO₂ concentrations the acidity is lower which is related to the somewhat lower sulfonation degree of dissolved lignin.

Thus, delignification selectivity is substantially affected by SO₂ concentration, with acceptable levels achieved at concentrations 12 w/w% or higher.

Summary: SO₂ concentration has a pronounced effect on the rates of major reactions, especially when lower than 12 w/w%.

3.6 Selectivity of hemicelluloses removal

Fig. 4 shows that the hemicelluloses removal selectivity is quite insensitive to fractionation conditions. This phenomenon is observed also in acid sulfite process^49c and is explained by the fact that the same chemical reaction (i.e. acid-catalysed hydrolysis of glycosidic bonds) is responsible for both cellulose hydrolysis and hemicelluloses removal. However, it is apparent that somewhat higher selectivity is observed at lower liquor-to-wood ratios and (only for some selected conditions) at higher temperatures. The former could be possibly explained by protection of cellulose by lignin present in higher amounts at lower L/W ratios, while the latter is related to high apparent activation energy for glucomannan removal.

Fig. 4 Selectivity of hemicelluloses removal in SEW fractionation (spruce I and III, 6 and 12 w/w% SO₂ in 40–70 v/v% ethanol–water). For symbols see Fig. 1a, 2a and S1. Residual hemicelluloses are calculated as a sum of anhydromannose and anhydroxylose. Grey area corresponds to the target properties for the unbleached viscose-grade dissolving pulps.

However, the overall effect of temperature on the selectivity is unclear. There is no agreement on the effect of temperature on hemicelluloses removal selectivity in acid sulfite process, either. For example, Ulfsparre (in ref. 47) showed that at a particular cellulose DP the hemicellulose retention increases with decreasing temperature. Conversely, according to Sixta et al.^49c at the same DP value a higher hemicelluloses retention is achieved at higher temperature.

Fig. 4 shows that the purity requirements of viscose-grade dissolving pulp (grey area) are achieved in SEW fractionation.

3.7 Optical properties of unbleached SEW pulps

ISO brightness values obtained for the unbleached SEW pulps at different conditions are provided in Fig. 5. It can be seen that at L/W ratios of 4.5 and 6 L kg⁻¹, all pulps have ISO brightness of at least 37% and reach 48%, while at a L/W ratio of 3 L kg⁻¹, 37% brightness is the highest value, with some pulps having as low brightness as 22%. This suggests the importance of L/W ratio in terms of the pulp brightness. The fact that brightness is not affected so strongly by ethanol concentration suggests that acidity is not the (only) governing factor. Therefore, it is possible that at L/W ratio of 3 L kg⁻¹ some reprecipitation of lignin on fibres occurs which results in impaired brightness.

Fig. 5 ISO brightness of unbleached SEW pulps (spruce III, 12 w/w% SO₂).

4 Conclusions

SO₂–ethanol–water fractionation is only slightly affected by moisture content of the raw material provided that the fresh liquor composition is adjusted to include water originating from the raw material.

Kinetics of bulk delignification and cellulose hydrolysis can be linearized using H-factor concept, i.e. combining the effects of temperature and time, as ln([Lig]) = f(H) and 1/DP = f(H_C), respectively (using activation energy of 106.8 kJ mol⁻¹ for delignification and 164.9 kJ mol⁻¹ for cellulose hydrolysis). Activation energy of bulk glucomannan removal is 108.4 kJ mol⁻¹, and that of xylan is 110.1 kJ mol⁻¹. However, bulk hemicelluloses removal phase is preceded by the initial phase with higher apparent activation energy, and therefore applying H-factor concept for hemicelluloses removal, in particular glucomannan, is problematic. The observed patterns are in line with the fact that higher delignification selectivity is obtained at lower temperatures. No consistent effect on hemicelluloses removal selectivity (relative to cellulose hydrolysis) was observed.

Due to the absence of base in SEW process, SO₂ concentration should be relatively high to avoid lignin condensation. Absence of base also dictates the fact that the adjustment of the acidity has to be performed by changing the ethanol concentration. Delignification rate is the highest at about 55 v/v% ethanol (for both L/W ratios 3 and 6 L kg⁻¹), where the optimum ratio between hydrolytic cleavage and condensation is observed.

Bulk delignification rate is affected neither by the variation in spruce chips nor by the liquor-to-wood ratio within 3–6 L kg⁻¹. Nevertheless, at L/W ratios lower than 3 L kg⁻¹ delignification is impaired. On the other hand, residual delignification is more sensitive to these conditions. For spruce batches I and II, kappa numbers of 5.5–8 are easily obtained (L/W ratio 6 L kg⁻¹), while for spruce batch III, residual delignification is less successful (lowest achieved kappa number was 9 at L/W ratio of 6 L kg⁻¹ and 14 at L/W ratio of 3 L kg⁻¹). This behaviour is contrary to Mg acid-sulfite delignification of spruce where very low kappa number values (about 5) are achieved even at low L/W ratios. Therefore, the importance of residual SEW delignification, including the aspects of lignin solubility and diffusion rates, is apparent.

Hemicelluloses removal and cellulose hydrolysis are faster at lower ethanol concentration and lower liquor-to-wood ratio, due to the higher acidity. Interestingly, the effect of ethanol concentration on the carbohydrate reactions (both hemicelluloses removal and cellulose hydrolysis) is stronger at L/W ratio of 6 L kg⁻¹ than at 3 L kg⁻¹.

ISO brightness is considerably lower at L/W ratio of 3 L kg⁻¹ compared to L/W ratios of 4.5 and 6 L kg⁻¹.

Therefore, when aiming at successful and selective delignification which is generally required in a lignocellulosic Biorefinery, the recommendation is to use high SO₂ concentration (at least 12 w/w%), high liquor-to-wood ratio, and the ethanol-to-water ratio of about 1 : 1 (by weight).

For paper pulps, lower temperature is desirable (for instance, 135 °C) in order to preserve high residual hemicelluloses amounts and cellulose molecular weight, but the substantial increase in fractionation duration should be taken into account. H-factor corresponding to kappa number 30 is equal to about 24 h.

For dissolving pulps, where much lower lignin and hemicelluloses content is desired, higher H-factors are recommended, about 50–75 h. Choosing liquor-to-wood ratio should be considered carefully as lower values lead to higher residual lignin but lower residual hemicelluloses at the same cellulose DP.

Based on the fractionation data, SEW pulps can be recommended for paper/tissue and viscose production.

Acknowledgements

The discussions with Dr Simo Porras and Prof. Ernst Kenndler on the acidity in mixed solvents are deeply acknowledged. The analysis of dissolved sugars was performed in Åbo Akademi University with the help of Dr Andrej Pranovich, which we also are very thankful for. We are also grateful to Myrtel Kåll for the IC analysis of sulfate anions and to Prof. Raimo Alén for providing Schöniger combustion apparatus. Financial support of TEKES through the FiDiPro program is greatly appreciated.

References

1. M. Iakovlev, Doctoral thesis, Aalto University, Finland, 2011.
2. (a) S. A. Rydholm, Pulping Processes, John Wiley & Sons Inc., London, 1965, p. 456; (b) S. A. Rydholm, Pulping Processes, John Wiley & Sons Inc., London, 1965, p. 540; (c) S. A. Rydholm, Pulping Processes, John Wiley & Sons Inc., London, 1965, pp. 488–490; (d) S. A. Rydholm, Pulping Processes, John Wiley & Sons Inc., London, 1965, p. 186.
3. S. S. Vishnevskaya, R. K. Boyarskaya and M. N. Tsypkina, Koksnes Kimija, 1981, 1, 23–28.
4. R. Puumala, Master thesis, Michigan Technological University, USA, 1991.
5. S. F. Primakov, Nauchn. Tr. Vses. Nauchn.-Issled. Inst. Tsellyulozn.-Bumazhn. Prom., 1961, 47, 69–75.
6. M. G. Eliashberg, A. I. Parfenova and S. F. Primakov, Trudy Leningrad. Lesotekh. Akad. im. S. M. Kirova, 1960, 91, 235–245.
7. M. Iakovlev, H. Sixta and A. van Heiningen, J. Wood Chem. Technol., 2011, 31, 250–266.
8. S. F. Primakov, Nauchn. Tr. Vses. Nauchn.-Issled. Inst. Tsellyulozn.-Bumazhn. Prom., 1961, 46, 83–118.
9. P. Schorning, Faserforsch. und Textiltechnik, 1957, 8, 487–494.
10. S. F. Primakov, I. M. Tsarenko and V. P. Zaplatin, Koksnes Kimija, 1979, 6, 39–42.
11. E. Sklavounos, M. Iakovlev, S. Survase, T. Granström and A. van Heiningen, Bioresour. Technol., 2013, 147, 102–109.
12. X. You and M. Iakovlev, unpublished work.
13. M. Yamamoto, M. Iakovlev and A. van Heiningen, Holzforschung, 2011, 65, 559–565.
14. E. Sklavounos, M. Iakovlev and A. van Heiningen, Ind. Eng. Chem. Res., 2013, 52, 4351–4359.
15. M. Iakovlev, E. Hiltunen and A. van Heiningen, Nord. Pulp Pap. Res. J., 2010, 25, 428–433.
16. H. Sixta, M. Iakovlev, L. Testova, A. Roselli, M. Hummel, M. Borrega, A. van Heiningen, C. Froschauer and H. Schottenberger, Cellulose, 2013, 20, 1547–1561.
17. S. A. Survase, E. Sklavounos, G. Jurgens, A. van Heiningen and T. Granstroem, Bioresour. Technol., 2011, 102, 10996–11002.
18. T. Retsina and V. Pylkkänen, Paper 360°, 2007, February issue, pp. 18–19.
19. T. N. Kleinert, Tappi, 1975, 58, 170–171.
20. M. Iakovlev, T. Pääkkönen and A. van Heiningen, Holzforschung, 2009, 63, 779–784.
21. E. A. Braude, J. Chem. Soc., 1948, 1971–1975.
22. E. Grunwald and B. J. Berkowitz, J. Am. Chem. Soc., 1951, 73, 4939–4944.
23. R. G. Bates and G. Schwarzenbach, Helv. Chim. Acta, 1955, 38, 699–716.
24. B. L. Wedzicha and S. J. Goddard, Food Chem., 1991, 40, 119–136.
25. K. Sarmini and E. Kenndler, J. Chromatogr., A, 1998, 811, 201–209.
26. E. A. Braude and E. S. Stern, Nature, 1948, 161, 169–170.
27. A. R. Tourky, A.-A. A. Abdel-Hamid and I. Z. Slim, Z. Phys. Chem., 1972, 250, 49–60.
28. R. Sierra-Alvarez and B. Tjeerdsma, 8th International Symposium on Wood and Pulping Chemistry, Helsinki, 6–9 June 1995, 2, pp. 207–211.
29. Y. Ni and Q. Hu, J. Appl. Polym. Sci., 1995, 57, 1441–1446.
30. V. M. Reznikov, M. A. Zil'bergleit, B. S. Simkhovich, T. V. Korneichik, T. P. Shkirando, T. V. Sukhaya and I. L. Kats, Bumazhnaya Promyshlennost, 1988, 9, 12–13.
31. M. Iakovlev and A. van Heiningen, RSC Adv., 2012, 2, 3057–3068.
32. M. Iakovlev and A. van Heiningen, ChemSusChem, 2012, 5, 1625–1637.
33. NREL/TP: (a) 510–42622, Determination of Ash in Biomass; (b) 510–42618, Determination of Structural Carbohydrates and Lignin in Biomass.
34. Scandinavian Standards: (a) SCAN-CM 49:03, Content of acetone-soluble matter, 2003; (b) SCAN-C 1:00, Chemical pulp: Kappa number, 2000; (c) SCAN-CM 15:99, Viscosity in cupriethylenediamine solution, 1999; (d) SCAN-CM 57:99, Pulp: Total sulphur content, 1999; (e) SCAN-N 36:98, Sulphite, sulphate and thiosulphate ion concentrations, 1998.
35. Tappi standard, 1966, T 230, Om-66, Viscosity of pulp (capillary viscometer method).
36. M. Iakovlev and A. van Heiningen, J. Wood Chem. Technol., 2011, 31, 233–249.
37. D. da Silva Perez and A. R. P. van Heiningen, Seventh European Workshop on Lignocellulosics and Pulp (EWLP, proceedings), 2002, pp. 393–396.
38. R. D. Girard, Doctoral Thesis, University of New Brunswick, Canada, 1998.
39. C. V. Sternling and L. E. Scriven, AIChE J., 1959, 5, 514–523.
40. R. Marton and S. Granzow, Tappi, 1982, 65, 103–106.
41. G. Goldfinger, Pap. Trade J., 1941, 113, 27–30.
42. B. Hagberg and N. H. Schoon, Sven. Papperstidn., 1973, 76, 561–568.
43. T. Richards and A. van Heiningen, J. Pulp Pap. Sci., 2004, 30(10), 263–268.
44. K. E. Vroom, Pulp Pap. Can., 1957, 58, 228–231.
45. W. Jensen, K. E. Fremer, P. Sierilä and V. Wartiovaara, The chemistry of bark, in The chemistry of wood, ed. B. I. Browning, John Wiley & Sons, London, 1963.
46. A. M. Emsley and G. C. Stevens, Cellulose, 1994, 1, 26–56.
47. D. Rusten, Nor. Skogind., 1962, 16, 328–339.
48. R. Evans and A. F. A. Wallis, J. Appl. Polym. Sci., 1989, 37, 2331–2340.
49. (a) H. Sixta, A. Potthast and A. W. Krotschek in Handbook of Pulp, ed. H. Sixta, Wiley-VCH, Weinheim, 2006, vol. 1, pp. 109–509, p. 110; (b) H. Sixta, A. Potthast and A. W. Krotschek in Handbook of Pulp, ed. H. Sixta, Wiley-VCH, Weinheim, 2006, vol. 1, p. 449; (c) H. Sixta, A. Potthast and A. W. Krotschek in Handbook of Pulp, ed. H. Sixta, Wiley-VCH, Weinheim, 2006, vol. 1, pp. 461–463.
50. M. S. Feather and J. F. Harris, J. Org. Chem., 1965, 30(1), 153–157.
51. G. E. Annergren and S. A. Rydholm, Sven. Papperstidn., 1959, 62, 737–746.
52. L. Salmén and A.-M. Olsson, J. Pulp Pap. Sci., 1998, 24(3), 99–103.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra45573d

---