Article pubs.acs.org/Biomac
Separation of Hemicellulose and Cellulose from Wood Pulp by Means of Ionic Liquid/Cosolvent Systems Carmen Froschauer,*,† Michael Hummel,‡ Mikhail Iakovlev,‡ Annariikka Roselli,‡ Herwig Schottenberger,† and Herbert Sixta‡ †
Faculty of Chemistry and Pharmacy, University of Innsbruck, 6020 Innsbruck, Austria Department of Forest Products Technology, Aalto University, 00076 Aalto, Finland
‡
S Supporting Information *
ABSTRACT: Pulp of high cellulose content, also known as dissolving pulp, is needed for many purposes, including the production of cellulosic fibers and films. Paper-grade pulp, which is rich in hemicellulose, could be a cheap source but must be refined. Hitherto, hemicellulose extraction procedures suffered from a loss of cellulose and the non-recoverability of unaltered hemicelluloses. Herein, an environmentally benign fractionation concept is presented, using mixtures of a cosolvent (water, ethanol, or acetone) and the cellulose dissolving ionic liquid 1-ethyl-3-methylimidazolium acetate (EMIM OAc). This cosolvent addition was monitored using Kamlet−Taft parameters, and appropriate stirring conditions (3 h at 60 °C) were maintained. This allowed the fractionation of a paper-grade kraft pulp into a separated cellulose and a regenerated hemicellulose fraction. Both of these exhibited high levels of purity, without any yield losses or depolymerization. Thus, this process represents an ecologically and economically efficient alternative in producing dissolving pulp of highest purity.
■
INTRODUCTION The constant growth of the global population inevitably leads to an increasing demand for food and textiles, while the availability of arable land is limited and even decreasing because of soil degradation (desertification, salinization, etc.). This will result not only in a shortage of fertile farmland and a foreseeable food crisis1,2 but also in a shrinking of the cotton growing area.4,3 Since the production of cotton is stagnant at 26 million tons per year and the supply of cotton fibers is no longer able to meet the growing demand, prices have spiked significantly over the past few years.5 For several decades, the growing fiber demand was covered by different petroleumbased synthetic fibers, such as polyester, polyamide, polypropylene, and polyacrylonitrile. However, the increasing requirement for a high physiological performance, associated primarily with moisture management and absorbency, can be favorably provided by natural fibers such as cellulose. As a consequence, it is currently impossible to cover the additional demand for cellulose, and assuming today’s growth rates, by 2050 there will be a shortage of 15 million tons per year of cellulosic fibers. Thus, a cellulosic-fiber gap is projected, which can be most likely closed by expanding the capacities of manmade cellulosic fiber production, such as viscose (Modal) or Lyocell (Tencel). The production of man-made cellulosic viscose fibers in relevant quantities already started in the 1930s and grew quickly until it peaked in the 1970s. Due to environmental regulations and more and more competitive synthetic products, the stable cellulosic fiber market experi© XXXX American Chemical Society
enced a steady decline for several decades. However, from 2001 on, cellulosic fibers underwent a renaissance mainly due to the expansion of viscose production in China and Southeast Asia. This is also reflected in an increased demand for hemicelluloselean dissolving pulp, which is suitable for the production of regenerated cellulose products and cellulose derivatives because of its cellulose content >90%, a high level of brightness, and a uniform molecular-weight distribution.6 The exploitation of a raw material source, which can be grown in forests on marginal land generally using natural irrigation, is characterized by a much smaller carbon footprint and a vast range of environmental benefits compared to the conventional production of cotton fibers, novel biobased fibers, and fossil fuel-based fibers.4,7 Wood-derived dissolving pulps are produced according to the acid sulfite and prehydrolysiskraft processes, which require further steps such as hot and cold caustic extractions to reach an adequate degree of purification, specified as the content of residual hemicelluloses and alkali resistance.8 The application of such harsh conditions in the current hemicellulose-removing technologies results in severe losses of cellulose (15−30%), mainly due to peeling-off reactions. Classic procedures to remove xylan, such as hydrolysis using steam and elevated pressure, enzymatic treatments15−17 or Received: January 22, 2013 Revised: April 19, 2013
A
dx.doi.org/10.1021/bm400106h | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
■
oxidative and reductive treatments, are often accompanied by a certain degree of biopolymer degradation.18−22 A proposed alternative method of using metal complexes of the nitren family for the generation of pulp of high purity is handicapped by an ineffective removal of the residual nickel from pulp.23,24 Relatively pure, polymeric hemicelluloses can be recovered by cold caustic extraction treatment of paper-grade pulps; this is a process that could be realized for commercial use and is thus used as a reference in this study.25 Consequently, there is a need for novel, environmentally friendly, and economically attractive manufacturing processes that allow a noninvasive, concomitant recovery of hemicelluloses and dissolving pulps in high yields and of high purity. Such single biopolymer fractions can independently serve as low-cost and sustainable feedstock for polymeric composite materials, base chemicals, and biofuels.9 This would greatly benefit the major hemicellulose xylan, since a lot of applications have been developed anticipating the availability of lost-cost xylan on the market (e.g., gel forming,10 paint formulation,11 chiral separations,12 pharmaceuticals13).14 In the field of biorefinery of wood and the processing of cellulose, ionic liquids (ILs), which are polar solvents exhibiting a negligible vapor pressure, gained considerable interest.26−30 In the majority of reported cases, ILs are applied for the pretreatment of lignocellulose or for the complete dissolution of cellulose. These cellulose/IL solutions can be further used for homogeneous derivatization and for decrystallization and activation for hydrolysis to mono and oligo sugars or can be processed to regenerated fibers, films, or beads.31−36 Continuing work on the dissolution of cellulose,37 this paper presents a new methodology for the quantitative separation of bleached paper-grade pulps into cellulose and hemicellulose fractions, both of high purity, while conserving cellulose I throughout the whole procedure.38 The solvent system consists of a cellulose-dissolving IL and an IL-miscible hydrophilic cosolvent, which can be stripped off easily for IL regeneration. In this case, 1-ethyl-3-methylimidazolium acetate (EMIM OAc)39−41 was the IL of choice since it has been reported in the literature that hemicellulose and lignin can be partially removed by treating lignocellulose with EMIM OAc.32,40,42,43 Different cosolvent/EMIM OAc mixtures were tested separately for their dissolving capacity on pure xylan and cellulose. Thus, water turned out to be an ideal cosolvent since it is known to decrease the solubility of cellulose44 and to reduce the effectiveness of IL pretreatments of lignocellulose with EMIM OAc as well.31,45 A water concentration window in EMIM OAc could be defined where pure xylan is soluble, but cellulose fibers remain unaffected by the solvent mixture. The interaction of the EMIM OAc/water mixtures with solutes was described by the empirical Kamlet−Taft (KT) parameters, primarily the β (H-bond basicity) and the α (H-bond acidity) values and the resulting net basicity β-α.45,46 The present study reports the effects of the type of cosolvent and the cellulose dissolving agent (EMIM OAc, NMMO), the cosolvent concentration, the dissolution time, the treatment conditions, and the pulp load on the fractionation of commercial bleached birch kraft pulp. Based on the outcome of this study, a simple and energy-saving process concept was developed, which permits the conversion of cheap paper-grade pulp into a high-purity dissolving pulp with concomitant recovery of high purity xylan.
Article
MATERIALS AND METHODS
Materials. 1-Ethyl-3-methylimidazolium acetate was purchased from BASF (quality >95%) and used without further purification (water content 0.130 wt %, according to Karl Fischer titration). Ethanol, acetone, and anhydrous NMMO were purchased from Sigma Aldrich and were used as received. Deionized water was obtained from a commercial deionizer and was used in all experiments. Dyes for KT parameter measurements (Reichardt’s dye, 4-nitroaniline, and N,Ndiethyl-4-nitronaniline) were gratefully received from the University of Helsinki. Cold caustic extracted (CCE) xylan from birch was used as a hemicellulose standard model. Therefore, cotton linters (524 mL/g, Milouban) were used as the representative cellulose sample. The paper-grade pulp sample used in this study was birch kraft pulp (965 mL/g, moisture content 5.2%, Stora Enso). The moisture content was based on oven-dry weight and was determined by differential weighing of the sample before and after drying overnight at 105 °C. The pulp samples were ground by means of a Willey mill. Prior to fractionation, the starting birch kraft pulp had the following carbohydrate compositions: cellulose 73.6%; xylan 25.4%; O-acetyl galactoglucomannan (GGM): 1.0%; Mn = 56.1 kg mol−1; Mw = 558 kg mol−1. Preparation of CCE Xylan. Xylan was isolated with 2.5 M aqueous NaOH at room temperature with a liquid-to-solid ratio of 20 L/kg for 60 min. After the extraction, the liquor was vacuum-filtered, and the pulp residue was washed until a neutral pH was obtained. The solubilized xylan was precipitated by the controlled addition of the filtrate to 4.5 M aqueous H2SO4. The charge of sulfuric acid was adjusted such that the final pH was 2.5. After storage for up to 10−12 h, the precipitated xylan was separated by centrifugation. In order to remove residual acid, the precipitate was washed with water in a centrifuge until the pH of the supernatant was around 5. Subsequently, the precipitate was freeze-dried, and the yield was determined gravimetrically. Preparation of EMIM OAc/Cosolvent Mixtures. These mixtures were prepared by combining water, acetone, or ethanol in ratios from 0 to 50 wt % with EMIM OAc. Small amounts (∼50 mL) of each mixture were prepared for the small scale experiments and used for the determination of KT parameters and rheology data (EMIM OAc/water mixtures). When preparing larger quantities of mixtures for upscaled fractionation experiments, the cosolvent was added slowly to EMIM OAc, while the flask was cooled with water to dissipate the exothermic mixing heat. Rheology. Shear rheology data of EMIM OAc/water mixtures was collected on an Anton Paar MCR 300 rheometer, with a plate and plate geometry as described in previous reports.47 The samples were subjected to a steady shear test over a velocity shear rate range of 0.1− 100 s−1 at different temperatures. Time stability tests showed no significant moisture uptake from the laboratory atmosphere or cosolvent evaporation at the plate edges within the required testing time. Thus, it was not necessary to seal the edges with paraffin oil as suggested by other research.48,49 All mixtures showed Newtonian behavior. For mixtures with high cosolvent content measured at high temperature, the testing device was close to its sensitivity limit and a stable flow curve was only observed at high shear rates. Determination of KT Parameters. The KT parameters (α, β, π*, ET(30)) were determined from the absorption peaks of the following three dyes: Reichardt’s dye (RD, range 518−585 nm), N,N-diethyl-4nitroaniline (DENA, range 402−414 nm), and 4-nitroaniline (NA, range 406−398 nm). Stock solutions of the dyes were prepared in acetone (60 mg RD, 2.4 mg NA, and 2.4 mg DENA in 8 mL of acetone, respectively). A measurement of 100-μL aliquots of a dye solution was transferred into a vial, where the acetone was evaporated by means of a nitrogen gas stream. Just 0.8 g of the IL/cosolvent mixture was added to the dried dye, mixed, and transferred into a cuvette (0.1 cm × 1 cm). The absorption peak of each of the three different dyes in each mixture was recorded using a PC-controlled Varian UV−vis spectrometer with a thermostat-regulated sample holder (25 ± 0.1 °C) according to reported procedures.37 The parameters were calculated using published equations.45 B
dx.doi.org/10.1021/bm400106h | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
Determination of a Cosolvent Concentration Window (small scale). Separate dissolution trials for pure xylan and pure cellulose were conducted on a small scale. At that point, 3.0 wt % of CCE xylan was added to 3 g of each IL/cosolvent mixture (0, 2, 4, 6, 8, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55 wt % of water, ethanol, or acetone) in small screw-capped flasks, equipped with a magnetic stir bar. The suspensions were stirred at different temperatures (50, 60, 70, 80 °C) in an oil bath for varying time periods. The maximum temperature chosen for the experiments was 80 °C because of the boiling point of the cosolvents and since it is known that cellulosic constituents and EMIM OAc suffer from thermal degradation at elevated temperatures.40 An extra flask was prepared for each solvent dilution stage and each temperature. The dissolution time (homogeneous solution, no undissolved particles) was determined by optical microscopy. Afterward, the same experiments were conducted for 0.5 wt % of cotton linters in all EMIM OAc/cosolvent mixtures, where xylan was already known to be soluble. Since the specific focus of these experiments was whether cellulose could be dissolved, a low charge of cellulose (0.5 wt %) and a vague time record were chosen. In these preliminary tests, acetone and ethanol proved unsuitable for fractionation, and therefore no further pulp dissolving experiments were carried out using these two cosolvents. Fortunately, a specific cosolvent concentration window could be determined for water, which was further examined in small-scale hemicellulose-rich pulp dissolution experiments. For this purpose, 3.75 wt % of Stora Enso birch kraft pulp was treated with solvent systems containing 10−30 wt % of water at various temperatures (60 and 80 °C) and mixing periods (0.5−6.0 h). The resulting suspension was separated by means of centrifugation or filtration. The dissolved hemicellulose was precipitated from the solution by further addition of water and collected by centrifugation. Both fractions, cellulose and hemicellulose, were washed with water, dried, and analyzed. After completion of these experiments, the procedure was scaled up using the conditions that were identified as the most suitable in these preliminary tests. Selective Extraction of Hemicellulose (larger scale). For pulp fractionation experiments of a larger scale, the EMIM OAc/water mixture was put in a beaker, and 3.4 wt % of milled birch kraft pulp was added. The suspension (∼170 mL) was transferred into a heated vertical kneader (b+b Gerätetechnik, Germany) and stirred for 3 h at 60 °C. The dissolution state was controlled by monitoring the shaft torque, indicating maximal hemicellulose dissolution after approximately 75 min. The suspension was transferred to a heated pressure filtration unit, equipped with a metal fleece (pore diameter 1−5 μm) for subsequent phase separation. An appropriate filtration of the generated suspension ensured a quantitative phase separation. The filter cake was washed once with the corresponding EMIM OAc/water mixture of the same consistency as used for the dissolution and filtrated again at 60 °C. The solid white fibrous residue was washed several times with deionized water, followed by filtration and drying overnight at 40 °C for subsequent analysis. The hemicelluloses were precipitated from the combined filtrates by further addition of the cosolvent, which then acted as an antisolvent.50 The hemicellulose suspension was stored at room temperature for 1 h. Afterward, the hemicelluloses were centrifuged (4500 rpm, 25 °C, 15 min), washed several times with deionized water, repeatedly centrifuged (10,000 rpm, 4 °C, 25 min), and either dried in a vacuum-oven overnight at 40 °C or freeze-dried. The resultant off-white powder was prepared for further analysis. A flowchart of the fractionation process is given in Scheme 1, and a pictorial illustration is presented in the Supporting Information. The effect of the pulp consistency (3.4 and 10.5 wt %) and the cellulose solvent (EMIM OAc, NMMO) on the hemicellulose extraction were examined in subsequent experiments following the same procedure as described above. The EMIM OAc/water mixture can be recycled to the fractionation process after the water content has been adjusted to the desired value. Determination of Pulp Viscosity. The intrinsic viscosity of the isolated cellulose fraction was determined by a standard procedure using cupriethylenediamine (ISO 5351-1). Carbohydrate Analysis. The carbohydrate composition of the treated pulp was determined by quantitative saccharification upon acid
Scheme 1. Schematic Illustration of the Fractionation of Hemicellulose-Rich Pulp into Cellulose and Hemicellulose
hydrolysis according to standard procedures reported by Sluiter et al.51 The monosaccharides were determined by high performance anion exchange chromatography with pulse amperometric detection (HPAEC-PAD) in a Dionex ICS-3000 system. Based on the amount of individual monosaccharides, cellulose and xylan content in pulp samples was calculated by the Janson formula.52,53 Using this formula, cellulose is defined as the content of anhydroglucose in the sample after subtracting the contribution of glucose to glucomannan, and xylan is defined as the content of anhydroxylose including uronic acid substituents. Size Exclusion Chromatography. Molar mass distribution of pulps was determined by gel permeation chromatography (GPC) as reported by other research.54 All samples were first activated by water−acetone−N,N-dimethylacetamide (DMAc) sequence. The activated samples were dissolved in 90 g L−1 lithium chloride (LiCl) containing DMAc at room temperature and under gentle stirring. The samples were then diluted to 9 g L−1 LiCl/DMAc, filtered with 0.2 μm syringe filters, and analyzed in a Dionex Ultimate 3000 system with a guard column, four analytical columns (PLgel Mixed-A, 7.5 mm × 300 mm), and RI-detection (Shodex RI-101). Flow rate and temperature were 0.75 mL min−1 and 25 °C, respectively. Narrow pullulan standards (343 Da−2500 kDa, PSS GmbH) were used to calibrate the system. The molar masses (MM) of the pullulan standards were modified to correspond to those of cellulose (MMcellulose = q × MMppullulan), as proposed by Berggren et al.55 The coefficients q = 12.19 and p = 0.78 were found by a least-squares method, using the data published in their report. Crystallinity. The wide-angle X-ray scattering (WAXS) measurements were carried out under perpendicular transmission geometry, with a setup consisting of a Seifert ID 3003 X-ray generator (voltage 36 kV, current 25 mA) equipped with a Cu tube (wavelength 1.54 Å), a Montel multilayer monochromator, and a MAR345 image plate detector. The samples were placed inside metallic sample holder rings with a thickness of 1 mm and sealed on both sides with Mylar foil. The calculation of crystal size and crystallinity was carried out as previously reported by Penttilä et al.56
■
RESULTS AND DISCUSSION Differences in the Solubility Behavior of Cellulose and Xylan. For preliminary studies, the dissolution behavior of isolated hemicellulose and cellulose dependent on the cosolvent concentration in EMIM OAc was investigated separately. The varying solubility of different pulp components may be correlated to the molar mass of the pulp components, morphology, supramolecular structure, degree of crystallinity, and other chemical properties. Therefore cold caustic extracted (CCE) xylan and cotton linters were used as model standards for the wood components. EMIM OAc was the IL of choice for these investigations, since it was reported that pure EMIM OAc
C
dx.doi.org/10.1021/bm400106h | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
can dissolve cellulose in high concentrations. Additionally, EMIM OAc shows low toxicity (LD50 >2000 mg kg−1), low corrosiveness, a low melting point (←20 °C), and low viscosity, as well as reasonable thermal stability (up to ∼200 °C).40,57 By screening technically relevant cosolvents (water, ethanol, or acetone) at different concentrations in EMIM OAc, an appropriate concentration window, in which the hemicelluloses are completely soluble while cellulose fibers stay unaffected and filtratable, was sought. A consistency of 3.0 wt % of CCE xylan was found to be soluble up to a water concentration of 40 wt % (Figure 1), an ethanol concentration of 30 wt %, or an acetone concentration of 50 wt % in EMIM OAc.
Figure 2. Viscosity of EMIM OAc/water mixtures as a function of water content at various temperatures.
It is worth noting that the dissolution time was only marginally influenced by the temperature when EMIM OAc mixtures containing more than 15 wt % water were used (Figure 1). This can be attributed to a comparably small further decrease of the viscosity upon heating once this threshold is passed (Figure 2). The subsequent dissolution experiments with cotton linters as cellulose model were conducted using only cosolvent ratios that were already known to allow complete dissolution of CCE xylan (water
