Separation and recovery of lignin from hydrolysates of lignocellulose with a polymeric adsorbent

Accepted Manuscript Separation and recovery of lignin from hydrolysates of lignocellulose with a polymeric adsorbent Jari Heinonen, Quentin Sanlaville, Henna Niskakoski, Juha Tamper, Tuomo Sainio PII: DOI: Reference:

S1383-5866(17)30674-3 http://dx.doi.org/10.1016/j.seppur.2017.06.001 SEPPUR 13776

To appear in:

Separation and Purification Technology

Received Date: Revised Date: Accepted Date:

1 March 2017 17 May 2017 1 June 2017

Please cite this article as: J. Heinonen, Q. Sanlaville, H. Niskakoski, J. Tamper, T. Sainio, Separation and recovery of lignin from hydrolysates of lignocellulose with a polymeric adsorbent, Separation and Purification Technology (2017), doi: http://dx.doi.org/10.1016/j.seppur.2017.06.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

SEPARATION AND RECOVERY OF LIGNIN FROM HYDROLYSATES OF LIGNOCELLULOSE WITH A POLYMERIC ADSORBENT Jari Heinonen1,*, Quentin Sanlaville1, Henna Niskakoski1, Juha Tamper2, Tuomo Sainio1 1

School of Engineering Science Lappeenranta University of Technology Skinnarilankatu 34, FIN-53850 Lappeenranta, Finland 2

UPM-Kymmene Corporation North European Research Center, Biochemicals Paloasemantie 19, FIN-53200 Lappeenranta, Finland

*

Corresponding author. Tel: +358-40-127 2920, e-mail: [email protected]

Abstract Separation and recovery of lignin from monosaccharide rich hydrolysates of lignocellulose by adsorption was investigated. Six polymeric adsorbents were compared by batch equilibrium experiments. XAD-16N was found to be the most efficient. Use of XAD-16N was studied in more detail in a column. A good separation efficiency could be obtained: with 95 % monosaccharide recovery yield limit, the lignin removal level was 80 %. Organic acids and furans were not adsorbed on XAD-16N. Practically all of the adsorbed lignin could be recovered efficiently with 50 wt.% ethanol. The lignin fraction had 99 % purity with respect the other solutes in the hydrolysates. Applicability of the process on large scale was demonstrated successfully by scaling up the process using data obtained with a laboratory scale unit (scale-up factor: 420). A 76 % lignin removal (target: 70 %) was achieved with higher than 95 % monosaccharide recovery yield (target: 95 %).

Keywords: lignocellulose; hydrolysate; lignin; adsorption; scale up

1

INTRODUCTION

Lignin is the second most abundant polymeric material after cellulose and the most abundant aromatic polymer on Earth [1–3]. It is found in lignocellulosic biomasses where it accounts for as much as 30 % of weight and 40 % of energy content [1,4,5]. Lignin is an amorphous three dimensional polymer network consisting of repeating randomly branched phenylpropanoid (C9) units bound together by carbon–carbon or ether bonds [1,4,6]. The structure of lignin depends on its origins [1,3]. Similarly as the other main components of lignocellulosic biomasses, cellulose and hemicelluloses, lignin is a valuable platform chemical for the production of a wide and continually expanding range of products. Lignin can be used as raw material for aromatic fine chemicals such as phenols, vanillin, and syringaldehyde, dispersing agents, aromatic polymers, activated carbon, carbon fibers, composites, polyphenolic glues, ion-exchangers, biofuels, low cost fillers, and additives [1,2,4–8]. Lignin is liberated from lignocellulosic biomasses during pulp production and during other biomass valorization processes such as hydrolysis [8–13]. In hydrolysis, the polysaccharides in the biomass are cleaved into monosaccharides with mineral acids (e.g. H2SO4) or enzymes as the catalysts. However, lignin is resilient towards chemical and enzymatic degradation and thus it remains mostly unchanged [6,14]. Thus, recovery of lignin and lignin derivatives from lignocellulose-derived solutions would create a huge renewable source for aromatic compounds. In addition, lignin removal from lignocellulose-derived solutions might be necessary due to limitations set by the downstream processing of these solutions. Lignin is known, for example, to inhibit fermentation of lignocellulose derived monosaccharides [15–17]. A number of methods for the separation of lignin from aqueous solutions has been presented. These include precipitation (change in pH, temperature or ionic strength of the solution) [12,18– 20], membrane filtration (ultrafiltration and nanofiltration) [10,21,22], centrifugation [23], ionic liquids [13,19], coagulation [24], bacteria [25], and adsorption [9,11,26–31]. Adsorption is an efficient and highly selective method for the purification of complex mixtures. Removal of lignin by adsorption on activated carbon has been investigated considerably. Liu et al. [28] studied removal of lignin from pre-hydrolysis liquor in a two-step adsorption process. Over 80 % of lignin was removed with approximately 30 % hemicellulose losses [28]. Montané et al. [29] studied the use of three commercially available activated carbons for the removal of lignin from a xylo-oligosaccharide rich solution produced by autohydrolysis of almond shells. Removal of lignin from dilute aqueous solutions by adsorption on activated carbon has been investigated by

Venkata Mohan and Karthikeyan [30] and Andersson et al. [31]. Gütsch and Sixta [27] investigated the removal of lignin from Kraft pulping prehydrolysate solutions with activated carbon at high temperature (150-170 °C): 75 % lignin removal was obtained. Shen et al. [9] have also studied the removal of lignin from pre-hydrolysis liquors by adsorption on activated carbon: 85 % lignin removal was obtained with less than 20 % and 0% losses of oligosaccharides and monosaccharides, respectively [9]. Strand et al. [11] obtained lignin removal of 68 % and 85 % for spruce and birch based pre-hydrolysis liquors, respectively, by adsorption on activated carbon. The sugar losses were 20 % and 14 %, respectively [11]. Activated carbon is an efficient adsorbent, however, the strong interactions between the adsorbent and the adsorbed compounds make the regeneration of the adsorbent, i.e. the recovery of the adsorbed compounds challenging. Desorption of lignin from activated carbon has been presented only by Venkata Mohan and Karthikeyan [30] with 1 N NaOH and 1 N HCl solutions with little success: only approximately 7 % desorption ratio was obtained. Regeneration of activated carbon by heat was studied by Gütsch and Sixta [27]: lignin is converted into activated carbon at 950 °C temperature. Polymeric adsorbents are also efficient for removal of aromatic compounds (e.g. lignin and furans). On the contrary to activated carbon, desorption of compounds from these can be done more efficiently (e.g. [32]). Use of XAD-8 polyacrylate adsorbent for the removal of aromatic substances including lignin from dilute water suspensions of thermo-mechanical pulps was investigated by Pranovich et al. [33]. Efficient desorption of the adsorbed compounds was achieved with methanol. Koivula et al. [34] have investigated the use of polyacrylate-based XAD-7HP and polystyrene-based XAD-16 adsorbents for the adsorption of foulants in woodbased autohydrolysates as a pretreatment step for ultrafiltration process. XAD-16 adsorbent removed at maximum 70 % of lignin from birch autohydrolysate (lignin amount as UV absorbance), but also 50 % of hemicelluloses present in the solution. The corresponding values for XAD-7 were 50 % and 30 %. Desorption of lignin from the adsorbents was not investigated in [34]. XAD-16 and XAD-7 have also been used to remove lignin and lignans from a sprucebased galactoglucomannan fraction [35] and pulp mill process water [36], respectively. Schwartz and Lawoko [37] achieved 90 % removal of acid-soluble lignin from hemicellulose rich hardwood extracts with polystyrene-based XAD-4 adsorbent. 85 % of the adsorbed lignin could be recovered with 75 % acetone [37].

Although lignin removal from various aqueous solutions by activated carbon and polymeric adsorbents has been investigated to some extent, authentic hydrolysates of lignocellulose have never been used as the feed solution. Here we present a study in which the recovery of lignin from such hydrolysates by adsorption on polymeric adsorbents is investigated. Both the adsorption and desorption steps are taken into account in order to properly evaluate the separation efficiency. The aim of this work is to present an adsorption-based process that allows the separation and recovery of lignin from these monosaccharide rich solutions for further processing. However, it is not an aim of this work to present a method for the production of pure monosaccharides from the hydrolysates as further purification of the hydrolysates (separation of organic acids and furans) after the lignin removal can be accomplished using methods presented in [32,38–41]. The applicability of the lignin separation and recovery process on a large scale is demonstrated by scaling up the separation process using data obtained with a laboratory scale adsorption unit.

2

EXPERIMENTAL

2.1

Materials and methods

Authentic pentose (from here on C5) and hexose (C6) rich hydrolysates of lignocellulose were used as feed solutions. These were produced in a demonstration scale production plant using hardwood as raw material. C5 hydrolysate was produced during dilute acid pre-treatment step of the biomass followed by solid–liquid separation stage. The remaining solid fraction was treated with a commercial enzyme to produce C6 hydrolysate. A second solid–liquid separation stage followed the enzymatic hydrolysis. pH values of C5 and C6 hydrolysates were approximately 2 and 4.5, respectively. The compositions of the hydrolysates are given in Table 1. The mix hydrolysate (Table 1) was prepared by mixing C5 and C6 hydrolysates. Analysis grade ethanol (100 %; AnalaR Normapur, VWR Prolabo), sodium hydroxide (≥ 99.0 %, pellets for analysis, Merck KGaA) and ultrapure water (purified with CENTRA-R 60/120 water purifier; ELGA/VEOLIA water) were used in the experiments.

<< Table 1 >>

Six polymeric adsorbents were tested for lignin removal (Table 2). CS11GC is a gel type strong acid cation exchange resin with sulfonic acid functional groups. It was used in acid (H+) form. The other adsorbents were commercially available neutral macroporous polymers with no functional groups.

<< Table 2 >>

CS11GC was shipped in Na+ form and was changed into H+ form in a column by flushing the adsorbent with 20 bed volumes (BV) of 1 mol/L hydrochloric acid. Afterwards, the adsorbent was washed with purified water to remove the excess acid. The other adsorbents were rinsed with purified water for 10 BV to remove any preservatives in the adsorbents.

2.2

Adsorbent screening

The effectiveness of the chosen six adsorbents in the removal of lignin (Table 2) was tested by batch reactor experiments. Both the adsorption and desorption of lignin was investigated. Prior to these experiments, excess water was removed from each adsorbent by centrifugation (2600 1/min; 15 min; Megafuge 2.0, Heraeus/Thermo Scientific). In the adsorption experiments, 5 g of each adsorbent was contacted with C5 or C6 hydrolysates with a phase ratio of 1:10 (wt./wt.). These mixtures were equilibrated for 4 h at 50 °C temperature in a shaker equipped with a heating thermostat. Samples were taken for analyses from the hydrolysates before and after equilibrating them with the adsorbents. The lignin adsorption efficiency and monosaccharide losses were calculated as the ratio of the lignin and monosaccharide concentrations in the hydrolysates before and after contact with the adsorbents. Prior to the desorption experiments, the loaded adsorbents were separated from the hydrolysates using a centrifuge (2 600 1/min; 15 min). 5 g of each centrifuged adsorbent was contacted with

50 wt.% aqueous ethanol solution with a phase ratio of 1:5 (wt./wt.). These slurries were then placed into a shaker with a heating thermostat and kept at 50 °C for 1 h. Samples were taken from the solutions equilibrated with the adsorbents for analyses. The efficiency of the desorption was evaluated with Desorption ratio 

C

des Clignin,eq Vdesorbent

ads lignin,0

ads  Clignin,eq Vhydrolysate

 100% ,

(1)

where Clignin is the lignin concentration, Vdesorbent is the volume of the desorbent in contact with the adsorbent, Vhydrolysate is the volume of the hydrolysate in contact with the adsorbent, superscripts ads and des refer to adsorption and desorption steps, respectively, and subscripts 0 and eq refer to initial and equilibrium, respectively.

2.3

Determination of adsorption isotherms for lignin on XAD-16N

Adsorption isotherms of lignin in C5 and C6 hydrolysates were measured for XAD-16N adsorbent by batch experiments. Several points on the adsorption isotherms were determined by changing the solid-to-liquid ratios in the batches. The phase ratios (wt./wt.) used were in the range of 1:2 – 1:200 with C5 hydrolysate and 1:2 – 1:80 with C6 hydrolysate, respectively. Different ranges were used because of different lignin concentrations in the hydrolysates (see Table 1). Before the measurements, XAD-16N was centrifuged to remove excess water. The adsorbent–hydrolysate slurries were placed into a shaker with a heating thermostat and kept at 50 °C for 4 h. Samples were taken for analyses from the hydrolysates before and after the equilibration. The adsorbed concentration of lignin qlignin onto the XAD-16N adsorbent was calculated from qlignin 

C

ads lignin,0

ads  Clignin,eq Vhydrolysate

madsorbent 1   p  / adsorbent

,

(2)

where madsorbent is the mass of wet adsorbent, εp is the porosity of the adsorbent (a value of 0.6 was used [42]), and ρads is the density of the wet adsorbent (a value of 1.02 g/mL was used [43]).

2.4

Lignin removal in a batch column

Removal of lignin from C5 and C6 hydrolysates was studied in a column in fixed bed mode with XAD-16N adsorbent. Both adsorption and desorption steps were investigated. The experimental setup used in these experiments consisted of an HPLC pump (Intelligent pump AI-12-13, Flom), a glass column with 1.5 cm inner diameter and a water-heating jacket (ECO SR 15/200, YMC Europe), water circulation thermostat (Lauda B, Lauda), UV absorbance detector (8453, Agilent) with a flow-through cell (cell flow path length = 1 cm), and a fraction collector (Foxy, Teledyme Isco). The experiments were carried out at a temperature of 50 °C. The volume of the adsorbent bed was 26.5 mL (dbed = 1.5 cm; hbed = 15 cm). XAD-16N was packed to the column as an aqueous slurry. 6 mL samples were collected from the column runs except for the cyclic runs. The void volume of the bed was calculated from the total porosity of the adsorbent bed εtot

 tot   b  1   b   p ,

(3)

where εb is the bed porosity (a value of 0.46 was used [42]). A value of 0.78 BV was obtained for the void volume.

2.4.1 Effect of flow rate and dynamic lignin adsorption capacity The effect of flow rate in the loading of the XAD-16N bed was investigated with three flow rates 1 mL/min (corresponds to 2.3 BV/h in the ECO SR 15/200 column), 2 mL/min (4.5 BV/h), and 3 mL/min (6.8 BV/h). With each flow rate 8 BV of hydrolysate was fed through the column. Fresh adsorbent was used in each experiment. The dynamic capacity, i.e. the maximum amount of lignin that can be adsorbed to the adsorbent bed, of XAD-16N was investigated by feeding 28 BV of C5 and C6 hydrolysates through fresh adsorbent bed. The flow rate was 3 mL/min (6.8 BV/h).

2.4.2 Desorption of lignin from the column Desorption of lignin from the XAD-16N bed was investigated with 20 wt.% and 50 wt.% aqueous ethanol solutions and 0.5 mol/L NaOH. The adsorbent used in these experiments was

loaded with lignin (solid-to-liquid ratio = 2:5) in a batch reactor prior to the desorption experiments. This was done in order to have the same amount of lignin adsorbed on XAD-16N in the beginning of each test. Desorption of lignin was studied with C5 hydrolysate. The loaded resin was packed to the column as a slurry with the treated hydrolysate as the liquid phase. Before the desorption of lignin, the excess hydrolysate was washed from the column with 3 BV of purified water. 3 mL/min (6.8 BV/h) flow rate was applied in the washing and desorption steps. Samples were taken for analyses from both the washing and the desorption steps.

2.4.3 Cyclic operation of adsorption column In order to see if regeneration is sufficient or if lignin fouls the adsorbent, cyclic column runs were conducted with C5 and C6 hydrolysates. These runs consisted of four steps: 1.

Adsorption of lignin. C5 and C6 hydrolysates were fed through the adsorbent bed for 16 BV and 21 BV, respectively. Four samples (4 BV each) were collected from the loading steps with C5 hydrolysate and five (4.2 BV each) from the loading steps with C6 hydrolysate in each cycle.

2.

Washing. The excess hydrolysate was washed from the adsorbent bed with 3 BV of purified water.

3.

Desorption of lignin. Lignin was desorbed with 4.5 BV of 50 wt.% aqueous ethanol solution.

4.

Washing. The ethanol was washed from the adsorbent bed with 3 BV of purified water.

In each step, the flow rate was 3 mL/min (6.8 BV/h). In the washing and desorption steps, all the solution that came out of the column was collected as one sample in each cycle. Five cycles were conducted with the C5 hydrolysate and three with the C6 hydrolysate.

2.5

Separation of lignin in a pilot scale adsorption column

A pilot scale separation column (PI140/850V0E-AB-T, YMC Europe) with 14 cm inner diameter and water heating jacket was used in the pilot scale lignin separation runs. In addition to the column, the adsorption unit consisted of two magnetic drive gear pumps (H3F-MC, Liquiflo), two mass flow meters (Optimass 3000, Krohne), pneumatic 3-way valves (Parker), and solenoid valves (SV-2, Parker). The liquid streams were monitored with two refractive index detectors (before and after the column; PR-23M, K-Patents), as well as a conductivity detector (876 ETC, Foxboro) equipped with a flow through sensor (EP402, Foxboro), and a UV detector (Flash 10 DAD 800, Ecom) equipped with a flow through sensor. The unit was controlled and all the online data collected with LabView program (version 14.0, National Instruments). The adsorbent bed height was approximately 73 cm (Vbed ≈ 11.2 L). The temperature in the experiments was 50 °C. Top–down flow was used in each process step. In order to mimic industrial process conditions, tap water (conductivity approximately 110 µS/cm) was used instead of purified water in the washing steps. The pilot scale unit was used to treat approximately 250 L of mix hydrolysate (see Table 1). The hydrolysate was treated in two cycles to obtain 70 % lignin removal level with 95 % monosaccharide recovery yield. This limit was set by the subsequent process step into which the purified hydrolysate was fed. Additional details concerning this step cannot be given here. Similarly as in the laboratory scale experiments (see Section 2.4.3) one cycle consisted of four steps: 1.

Loading step: 11 BV of hydrolysate was fed through the column to remove lignin from the hydrolysate;

2.

washing step: 3 BV of water was fed through the column to wash the excess hydrolysate from the adsorbent bed;

3.

desorption step: 4 BV of 50 wt.% aqueous ethanol was led through the column to desorb the adsorbed lignin;

4.

washing step: 3 BV of water was led through the column to wash the excess desorbent from the adsorbent bed.

2.6

Chemical analyses

Lignin concentrations were determined from UV absorbance according to a standardized method described in TAPPI standard UM 250 (Acid-soluble lignin in wood and pulp) [44].

A

wavelength of 205 nm was used for the quantification [44,45]. Lignin has an absorptivity maxima at 205 nm and 280 nm wavelengths. Of these 205 nm is suited for the analysis of lignin in acidic solution due to low contribution of sugar degradation products formed in acidic conditions, i.e., furans (furfural and hydroxymethyl furfural), to the UV absorbance. This is not the case at a wavelength of 280 nm [45]. In addition to lignin, other wood-based phenolic compounds (extractives and lignin degradation products) may contribute to UV absorbance at 205 nm wavelength. However, the contribution should be negligible due to following reasons. The amount of extractives in wood is low (1-5 %, [45,46]). In addition, only a small part of extractives are phenolic compounds [46]. Lignin is also a highly resilient polymer [6,14] and thus the amount of lignin degradation products in the hydrolysates used here should be low. Turbid samples were filtered with a syringe filter (0.45 µm polypropene filter, VWR). The UV absorbance was measured with Agilent 8453 (UV/Vis) detector using a 1 cm diameter quartz cuvette. The samples were diluted with purified water so that the absorbance reading was between 0.2 and 0.7. Purified water was used as a blank sample except in the analysis of the samples taken from lignin elution. With these samples, the blank sample was prepared by diluting pure desorbent solution with pure water using the same dilution factor as with the actual sample. For example, if a sample from the desoption step with 50 wt.% ethanol was diluted by a factor of 1 000, the blank was pure 50 wt.% ethanol diluted by a factor of 1 000. The measured absorbance values were converted to mass concentration values with

C lignin 

A D  ,  l

(4)

where A is the absorbance value, α is the absorptivity factor 0.11 L/(mg cm) (average value for samples containing different wood species) [44], l is the length of the cuvette, and D is the dilution factor. The monosaccharides, formic acid, acetic acid, hydroxymethyl furfural, and furfural concentrations were determined with an offline HPLC (HP/Agilent 1100) equipped with a refractive

index

detector

and

a

UV

detector

(variable

wavelength

detector;

wavelength = 280 nm). With the monosaccharides, a sugar analysis column Shodex SP-810 (Waters) was used with a Shodex SP-G guard column (Waters). The column temperature was

85 °C, and purified water was used as an eluent with a flow rate of 0.5 mL/min. An organic acid analysis column Metacarb 87H (Varian/Agilent) was used for the analyses of the organic acids and furans. The column temperature was 65 °C, and 0.005 mol/L sulfuric acid was used as an eluent with flow rate gradients between 0.6 – 0.8 mL/min.

3

RESULTS AND DISCUSSION

3.1

Adsorbent screening

The lignin removal efficiency of the six polymeric adsorbents (see Table 2) were determined by batch equilibrium experiments with C5 and C6 hydrolysates (Fig. 1). The comparison of the adsorbents was done on the basis of lignin adsorption capacity, lignin desorption efficiency with 50 wt.% ethanol as the desorbent, as well as monosaccharide losses during adsorption.

<< Figure 1 >>

XAD-7HP, XAD-16N, and MN 200 adsorbents were able to remove over 60 % of soluble lignin from C6 hydrolysate (Fig. 1A). These adsorbents have aromatic PS–DVB matrices, and thus lignin as an aromatic macromolecule has a strong affinity towards them. It is noteworthy that the adsorption of lignin of MN270 and CS11GC, that also have aromatic matrices, was considerably weaker than on the three other PS–DVB base adsorbents. CS11GC as a cation exchange resin has charged functional groups attached to the resin matrix. These groups make the resin more hydrophilic than the unfunctionalized PS–DVB adsorbents and thus the adsorption of lignin was weaker (Fig. 1A). With MN 270, the weaker adsorption of lignin, when compared to for example MN 200, results most probably from the smaller pore sizes (high surface area, see Table 2). XAD-4 adsorbent is a polyacrylate based adsorbent and has a higher hydrophilicity than the PS– DVB based adsorbents, and thus lignin has a low affinity towards this adsorbent (Fig. 1A). Due to the results obtained with C6 hydrolysate, only the four PS–DVB based adsorbents were tested for lignin removal from C5 hydrolysate (Fig. 1B). The lignin removal efficiency of these adsorbents followed the same order as with C6 hydrolysate (pH ~ 4.5) even though C5

hydrolysate was considerably more acidic (pH ~ 2) and contained approximately 4-times more lignin than C6 hydrolysate. In the desorption step with 50 wt.% aqueous ethanol, all the lignin adsorbed from C5 and C6 hydrolysates could not be recovered from any of the adsorbents under the experimental conditions used here (Fig. 1). It was observed that higher than 1:5 (wt./wt.) phase ratio should be used in the desorption step in order to achieve higher lignin recovery. Highest desorption efficiencies could be obtained with XAD-7HP and XAD-16N adsorbents that both had approximately equal desorption efficiencies. Interestingly, the lowest desorption efficiency was obtained with CS11GC in H+ form that also adsorbed the lowest amount of lignin (Fig. 1A). As the aim in the lignin removal is to produce lignin-free monosaccharide rich hydrolysate, the adsorption of monosaccharides on the adsorbents should also be investigated (Fig. 1). The adsorption of monosaccharides was weak with each adsorbent: less than 5 % with C6 hydrolysate and slightly more than 5 % with C5 hydrolysate. On the basis of the results presented here, XAD-16N was chosen for further studies. This adsorbent had a high lignin removal efficiency and the bound lignin could be recovered most efficiently in the experimental conditions used here.

3.2

Adsorption isotherms for lignin on XAD-16N

Adsorption of soluble lignin from C5 and C6 hydrolysates on XAD-16N adsorbent was investigated by batch equilibrium experiments. The concentration of lignin in the solid phase was calculated with Eq. (2). With the phase ratios used in the batch reactor during the lignin adsorption experiments, the equilibrium concentrations of lignin (Fig. 2) were considerably lower than the concentrations in C5 and C6 hydrolysates (Table 1) before contact with the adsorbent. This clearly indicates very strong adsorption of lignin on XAD-16N. In addition, the adsorption of lignin was found to increase considerably with increasing pH (Fig. 2). With both hydrolysates, the shape of the lignin isotherm was concave upward shaped. Such behavior could be caused by binding of lignin from solution to lignin molecules already adsorbed on the adsorbent. From the theory of adsorption column dynamics [47] it is expected that during loading of the adsorbent bed with lignin such a shape of the isotherm results in a dispersed or elongated front of the lignin profile rather than a sharp shock layer.

<< Figure 2 >>

3.3

Separation of lignin in a batch adsorption column

Separation of lignin by adsorption can be accomplished in a simple batch reactor as was seen above. The separation efficiency in that case depends on the phase ratio used. However, a more efficient way to carry out adsorptive separations is in a fixed bed column. Here, the separation of lignin from C5 and C6 hydrolysates was investigated in a fixed bed column with XAD-16N adsorbent.

3.3.1 Adsorption of lignin

<< Figure 3 >>

The removal of lignin from both C5 and C6 hydrolysates was investigated in a XAD-16N column (Fig. 3). Breakthrough of lignin from the column occurred immediately after the void volume of the adsorbent bed with both hydrolysate (Fig. 3). The breakthrough curves were strongly elongated. This atypical shape of the curve results from the heterogeneous nature of the lignin: it consists of a wide variety of molecules of different sizes and shapes (see e.g. [48–52]). The largest lignin molecules elute from the column first at the void volume of the adsorbent bed as they do not fit into the pores of the resin. Similar behavior has been observed earlier by Montané et al. [29] during adsorption of lignin on activated carbon. In addition to the effect of heterogeneity of the lignin molecules, the shape of the breakthrough curve is affected by the concave upward shape of the lignin adsorption isotherm (see Fig. 2). With C6 hydrolysate the outlet concentration of lignin did not reach the feed level in 28 BV (Fig. 3). However, with C5 hydrolysate, containing approximately 4-times more lignin than C6 hydrolysate (see Table 1), the feed level was reached in approximately 22 BV.

The effect of the flow rate on the lignin adsorption on XAD-16N in a fixed bed was investigated with C5 hydrolysate by feeding approximately 5–7.5 BV of the hydrolysate through the fixed bed column (Fig. 4). The flow rate was found to have no effect on the behavior of lignin.

<< Figure 4 >>

The monosaccharides eluted similarly with both hydrolysates (Fig. 3). Breakthrough of the monosaccharides occurred right after the void volume indication very weak adsorption on XAD16N. The monosaccharide concentration increased to the feed level approximately in 1.3 BV with both hydrolysates, thus pH of the hydrolysate did not have any effect on the monosaccharide adsorption. The adsorption of the organic acids (formic acid and acetic acid) was also very weak on XAD16N (Fig. 3). The breakthrough of the acids also occurred close to that of the monosaccharides. Again, pH of the hydrolysate did not affect the adsorption of the organic acids on XAD-16N. The furans, HMF and furfural, elute from the column few bed volumes after the monosaccharides (Fig. 3). With C6 hydrolysate, the adsorption of the furans was slightly stronger than with C5 hydrolysate. This is more evident for furfural with which the feed concentration level was reached in approximately 13.6 BV with C6 hydrolysate and in 8.3 BV C5 hydrolysate. The weaker adsorption of the furans in C5 hydrolysate on XAD-16N results from the higher lignin concentration. The furans and lignin compete during the adsorption, and as the lignin concentration in C5 hydrolysate is higher than in C6 hydrolysate, the furans are displaced by lignin earlier in case of C5 hydrolysate. In addition, the differences in the pH of the hydrolysates might affect the adsorption of the furans. HMF and furfural also compete with each other during the adsorption (Fig. 3). The concentration of HMF increases first above the feed level before attaining it. This results from displacement of HMF from the solid phase by furfural. On the basis of the experimental data obtained from the breakthrough curve experiments (Fig. 3), the efficiency of the adsorption process with respect to lignin removal and monosaccharide recovery yield can be evaluated (Fig. 5). It should be noted that a relatively short column (hbed = 15 cm) was used in this stage of the study. The results shown here thus correspond to

operation of the adsorption column at very low plate number (NTP). In practice, this is the case at very high feed flow rate. The adsorbed amount of monosaccharides used in the determination of the monosaccharide recovery yield values was calculated as the area above the monosaccharide loading curves from the void volume to the point at which the monosaccharide concentration reached the feed level.

<< Figure 5 >>

The lignin removal level and monosaccharide recovery yield depend strongly on each other (Fig. 5). For example, if the monosaccharide yield must be at least 95 %, the maximum amount of lignin that could be removed is 82 % with C6 hydrolysate and 81 % with C5 hydrolysate. Interestingly these lignin removal levels are approximately the same although C5 hydrolysate contained 4-times more lignin than C6 hydrolysate (see Table 1). To reach such lignin removal levels with 95 % monosaccharide yield 9.2 BV (Fig. 5A) and 8.4 BV (Fig. 5B) of C6 and C5 hydrolysates could be fed through the column, respectively. Higher than 80 % lignin removal level would decrease the monosaccharide yield significantly with both hydrolysates (Fig. 5). On the other hand, if higher than 95 % monosaccharide yield is needed, this can be obtained at the cost of the amount of lignin separated from the hydrolysate.

3.3.2 Desorption of lignin Desorption of lignin from the loaded XAD-16N adsorbent bed was investigated with 20 wt.% and 50 wt.% ethanol and with 0.5 mol/L NaOH with C5 hydrolysate (Fig. 6). The adsorbent contained approximately 670 mg of adsorbed lignin in each test. Excess hydrolysate was removed from the column with 3 BV of purified water. It was found that practically no lignin was desorbed with water. 50 wt.% aqueous ethanol solution was found to be the most efficient of tested desorbents (Fig. 6, Table 3). Practically all of the adsorbed lignin was desorbed in less than 4.5 BV with 50 wt.% ethanol. Notable focusing of the lignin was observed during the desorption step with 50 wt.%

ethanol: the maximum lignin concentration was approximately 1.4 times higher than the value in the C5 hydrolysate (Fig. 6). As no tailing of lignin was observed in the desorption step with 50 wt.% ethanol (Fig. 6), use of desorbent with higher ethanol concentration was not investigated. Desorption of lignin could not be carried out efficiently with 20 wt.% ethanol as with 50 wt.% ethanol (Fig. 6 and Table 3). In addition, the elution took a considerably longer time than with the 50 wt.% ethanol and the lignin fraction was less concentrated (Fig. 6). Solubility of lignin is high at alkaline pH range (e.g. [20,53]). Thus 0.5 mol/L aqueous NaOH solution was also tested as a desorbent (Fig. 6 and Table 3). However, with NaOH, only approximately 40 % lignin desorption level could be achieved.

<< Figure 6 >>

<< Table 3 >>

Due to the behavior of the main components of the hydrolysates (monosaccharides, organic acids, and furans) in the studied system, the lignin fraction obtained in the desorption step had approximately 99 % purity with respect to these components. Thus, XAD-16N adsorbent can be used to both remove lignin from lignocellulosic hydrolysates and to produce lignin with high purity. Separation of lignin from the desorbent (50 wt. % ethanol) could be done by evaporation or by nanofiltration. Recovery of the desorbent in such a way that it can be reused in the desorption step is important for low chemical consumption. The separation of lignin and ethanol was not investigated in this study. 3.3.3 Effect of repeated cycles on separation efficiency In order to assess the changes in the performance of the adsorption system for example by fouling caused by irreversible (or very strong) adsorption of lignin on XAD-16N, repeated operating cycles were carried out (Fig. 7). A 66 % lignin removal level target was set.

The amount of lignin removed from both hydrolysates remained practically constant from cycle to cycle (Fig. 7). This means that 50 wt. % aqueous ethanol solution desorbs practically all of lignin from the adsorbent and fouling of XAD-16N is not a major issue. On the other hand, it was visually observed that some degree of fouling occurs as the color of the adsorbent bed changed from white to brownish during the process. Interestingly, the monosaccharide yield was found to increase slightly from the first cycle to the last cycle (Fig. 7). This was attributed to adsorption of lignin-type compounds that further reduce the affinity of the monosaccharides on the polymer. The XAD-16N adsorbent works better in the separation of lignin from monosaccharides after few cycles than as a fresh, but longer cyclic runs are needed in order to more thoroughly evaluate the influence of the minor fouling of the adsorbent on process performance.

<< Figure 7 >>

3.4

Separation of lignin in a pilot scale adsorption column

The results shown above were obtained using a small laboratory scale adsorption column (Vbed = 26.5 mL). However, the aim of this study is to present an industrially viable process for the adsorptive removal and recovery of lignin from hydrolysates of lignocellulose. Thus, the scale up of the process on the basis of the obtained results was investigated. This was done by scaling up the process from laboratory scale to pilot scale (Vbed = 11.2 L; scale up factor of approximately 420). In addition to the significantly larger volume of the pilot scale adsorption column, also the adsorbent bed length was five times longer than in the laboratory scale column. With the longer column, a better separation efficiency can be expected because the column efficiency, usually quantified as the number of theoretical plates (NTP), is proportional to the column length [47]. The scale up was done in such a way that the reduced flow rate was kept constant (6.8 BV/h) and the linear flow rate was increased due to increase in column dimensions. The feed solution in the scale up trial was a mix hydrolysate (see Table 1). A 70 % lignin removal and 95 % monosaccharide recovery yield targets were set for the purification. The duration of the loading step was determined using the data shown in Fig. 5. Two cycles with 11 BV of hydrolysate treated on each were done to purify the 250 L batch. This is a rather conservative design, but it was deemed appropriate in order to guarantee 70 % lignin removal

from the mix hydrolysate, the properties of which are somewhere between those of C5 and C6 hydrolysates studied in laboratory scale. The profiles of lignin and monosaccharides in the loading step obtained with the pilot scale unit (Fig. 8) were qualitatively similar as those obtained with the laboratory scale unit (see Figs. 3 and 6). As the adsorbent bed in the pilot unit was longer than that of the laboratory scale unit, a higher lignin removal level was expected. However, the lignin removal level obtained was approximately the same: 76.4 % on the first cycle and 75.6 % on the second cycle. The overall removal was approximately 76 %. With the short laboratory scale column, the lignin removal levels at 11 BV feed volume were 78 % and 76 % with C6 and C5 hydrolysates, respectively (see Fig. 5). The monosaccharide recovery yield was approximately 95.5 % on the first cycle and 96.4 % on the second cycle which are also close to the values obtained with the laboratory scale unit. Thus, the higher NTP value (long column) of the pilot scale adsorbent bed did not improve the separation efficiency. This is most likely due to due to less perfect packing of the adsorbent bed in the pilot scale column. However, as the achieved lignin removal and monosaccharide recovery yield levels were on the same level with the laboratory scale and the pilot scale units, the scale up approach where the reduced flow rate (BV/h) is held constant by increasing the linear and volumetric flow rate in the larger columns is useful. Desorption of the lignin was done with 50 wt.% aqueous ethanol solution (data not shown) with good efficiency. Practically all of the adsorbed lignin (630 g) could be recovered from XAD-16N with 4 BV of 50 wt.% ethanol. No other compounds present in the hydrolysates were observed in significant quantities in the obtained lignin fraction.

<< Figure 8 >>

4

CONCLUSIONS

Separation and recovery of lignin from monosaccharide rich hydrolysates of lignocellulose was investigated. Of six tested commercial polymeric adsorbents, polystyrene-based XAD-16N was found to be most efficient. Use of XAD-16N was investigated in a detailed manner in a column: 80 % lignin removal level can be achieved simultaneously with 95 % monosaccharide yield. Further purification of the monosaccharides in the hydrolysates (i.e. separation from organic

acids and furans) cannot be accomplished simultaneously with the lignin separation. However, methods for such purification of the monosaccharides have been presented elsewhere. Practically all of the lignin adsorbed on XAD-16N can be recovered with 99 % purity (with respect to the solutes in the hydrolysates) using 50 wt.% ethanol. Thus the presented adsorption process can be used to both separate lignin from the hydrolysates of lignocellulose and to produce pure lignin fractions for use as a raw material. In addition, the adsorption process is applicable on a large scale. This was demonstrated by successfully scaling up the adsorption process using data obtained in laboratory scale separation experiments (scale up factor of 420).

Acknowledgements Support from K-Patents Oy (Finland) in on-line refractive index analyses for the pilot adsorption column and financial support from Academy of Finland (grant SA/298548) are gratefully acknowledged.

Nomenclature Clignin D dbed hbed l qlignin Vbed Vdesorbent Vhydrolysate madsorbent

concentration of lignin in liquid phase, g/L dilution factor, adsorbent bed diameter, cm adsorbent bed length, cm length of the cuvette, cm concentration of lignin in solid phase, g/L adsorbent bed volume, mL or L desorbent volume, mL hydrolysate volume, mL mass of wet adsorbent, g

Greek letters α εb εp ρads

absorptivity factor, L/(mg cm) bed porosity, porosity of the adsorbent, density of the wet adsorbent, g/mL

Subscripts and superscripts 0

initial

ads des eq

adsorption step desorption step equilibrium

References [1]

[2]

[3] [4]

[5] [6]

[7]

[8] [9]

[10]

[11]

[12]

[13]

[14] [15]

[16]

[17]

J.E. Holladay, J.F. White, J.J. Bozell, D. Johnson, Top Value-Added Chemicals from Biomass Volume II - Results of Screening for Potential Candidates from Biorefinery Lignin, Pacific Northwest Natl. Lab. II (2007) 87. doi:10.2172/921839. E.A.B. da Silva, M. Zabkova, J.D. Araújo, C.A. Cateto, M.F. Barreiro, M.N. Belgacem, A.E. Rodrigues, An integrated process to produce vanillin and lignin-based polyurethanes from Kraft lignin, Chem. Eng. Res. Des. 87 (2009) 1276–1292. doi:10.1016/j.cherd.2009.05.008. F.G. Calvo-Flores, J.A. Dobado, Lignin as renewable raw material, ChemSusChem. 3 (2010) 1227–1235. doi:10.1002/cssc.201000157. J. Zakzeski, P.C.A. Bruijnincx, A.L. Jongerius, B.M. Weckhuysen, The Catalytic Valorization of Ligning for the Production of Renewable Chemicals, Chem. Rev. 110 (2010) 3552–3599. doi:10.1021/cr900354u. IEA, Bio-based Chemicals Value Added Products from Biorefineries, (2011) 1–36. S. Constant, M. Robitzer, F. Quignard, F. Di Renzo, Vanillin oligomerization as a model of side reactions in lignin fragmentation, Catal. Today. 189 (2012) 123–128. doi:10.1016/j.cattod.2012.03.056. H.M.N. Iqbal, G. Kyazze, T. Keshavarz, Advances in the valorization of lignocellulosic materials by biotechnology: An overview, BioResources. 8 (2013) 3157–3176. doi:10.15376/biores.8.2.3157-3176. H. Lange, S. Decina, C. Crestini, Oxidative upgrade of lignin - Recent routes reviewed, Eur. Polym. J. 49 (2013) 1151–1173. doi:10.1016/j.eurpolymj.2013.03.002. J. Shen, I. Kaur, M.M. Baktash, Z. He, Y. Ni, A combined process of activated carbon adsorption, ion exchange resin treatment and membrane concentration for recovery of dissolved organics in pre-hydrolysis liquor of the kraft-based dissolving pulp production process, Bioresour. Technol. 127 (2013) 59–65. doi:10.1016/j.biortech.2012.10.031. A. Arkell, J. Olsson, O. Wallberg, Process performance in lignin separation from softwood black liquor by membrane filtration, Chem. Eng. Res. Des. 92 (2014) 1792–1800. doi:10.1016/j.cherd.2013.12.018. E. Strand, M. Kallioinen, M. Kleen, M. Mänttäri, K. Spruce, Activated carbon treatment to improve ultrafiltration performance in recovery of hemicelluloses from wood extracts, 30 (2015). A. Dieste, L. Clavijo, A.I. Torres, S. Barbe, I. Oyarbide, L. Bruno, F. Cassella, Lignin from Eucalyptus spp. Kraft Black Liquor as Biofuel, Energy & Fuels. 30 (2016) 10494– 10498. doi:10.1021/acs.energyfuels.6b02086. R. Duan, B.S. Westerlind, M. Norgren, I. Anugwom, P. Virtanen, J. Mikkola, com Fibre Stress-Strain Response of High-Temperature Chemi- Thermomechanical Pulp Treated with Switchable Ionic Liquids, 11 (2016) 8570–8588. M.J. Taherzadeh, K. Karimi, Enzyme-based hydrolysis processes for ethanol from lignocellulosic materials: A review, 2007. doi:10.15376/BIORES.2.3.472-499. B. Hahn-Hägerdal, K. Karhumaa, C. Fonseca, I. Spencer-Martins, M.F. Gorwa-Grauslund, Towards industrial pentose-fermenting yeast strains, Appl. Microbiol. Biotechnol. 74 (2007) 937–953. doi:10.1007/s00253-006-0827-2. R.C. Sun, Detoxification and separation of lignocellulosic biomass prior to fermentation for bioethanol production by removal of lignin and hemicelluloses, BioResources. 4 (2009) 452–455. doi:10.15376/biores.4.2.452-455. H. Shi, P. Fatehi, H. Xiao, Y. Ni, A process for isolating lignin of pre-hydrolysis liquor of kraft pulping process based on surfactant and calcium oxide treatments, Biochem. Eng. J. 68 (2012) 19–24. doi:10.1016/j.bej.2012.06.017.

[18] H. Kumar, R. Alén, G. Sahoo, Characterization of Hardwood Soda-AQ Lignins Precipitated from Black Liquor through Selective Acidification, BioResources. 11 (2016) 9869–9879. doi:10.15376/BIORES.11.4.9869-9879. [19] S.S. Mohtar, T.N.Z. Tengku Malim Busu, A.M. Md Noor, N. Shaari, N.A. Yusoff, M.A. Bustam Khalil, M.I. Abdul Mutalib, H.B. Mat, Extraction and characterization of lignin from oil palm biomass via ionic liquid dissolution and non-toxic aluminium potassium sulfate dodecahydrate precipitation processes, Bioresour. Technol. 192 (2015) 212–218. doi:10.1016/j.biortech.2015.05.029. [20] W. Zhu, H. Theliander, Precipitation of lignin from softwood black liquor: An investigation of the equilibrium and molecular properties of lignin, BioResources. 10 (2015) 1696–1714. [21] A. Toledano, A. García, I. Mondragon, J. Labidi, Lignin separation and fractionation by ultrafiltration, Sep. Purif. Technol. 71 (2010) 38–43. doi:10.1016/j.seppur.2009.10.024. [22] S. Hellstén, J. Lahti, J. Heinonen, M. Kallioinen, M. Mänttäri, T. Sainio, Purification process for recovering hydroxy acids from soda black liquor, Chem. Eng. Res. Des. 91 (2013) 2765–2774. doi:10.1016/j.cherd.2013.06.001. [23] K.A. Underkofler, R.E. Teixeira, S.A. Pietsch, K.G. Knapp, R.T. Raines, Separation of lignin from corn stover hydrolysate with quantitative recovery of ionic liquid, ACS Sustain. Chem. Eng. 3 (2015) 606–613. doi:10.1021/sc500731c. [24] C. Wu, L. Bing, S. Li, D. Yu, D. Wang, Effect of coagulating agents on lignin and oligosaccharide contents in pre-hydrolysis liquor obtained in the production of dissolving pulp from poplar residual slabs, BioResources. 11 (2016) 87–94. doi:10.15376/biores.11.1.87-94. [25] R. Priyadarshinee, A. Kumar, T. Mandal, D. Dasguptamandal, Unleashing the potential of ligninolytic bacterial contributions towards pulp and paper industry: key challenges and new insights, Environ. Sci. Pollut. Res. 23 (2016) 23349–23368. doi:10.1007/s11356-0167633-x. [26] X. Chen, Q. Yang, C.L. Si, Z. Wang, D. Huo, Y. Hong, Z. Li, Recovery of Oligosaccharides from Prehydrolysis Liquors of Poplar by Microfiltration/Ultrafiltration Membranes and Anion Exchange Resin, ACS Sustain. Chem. Eng. 4 (2016) 937–943. doi:10.1021/acssuschemeng.5b01029. [27] J.S. Gütsch, H. Sixta, Regeneration of spent activated charcoals used for lignin removal from prehydrolysis-kraft prehydrolyzates, Ind. Eng. Chem. Res. 51 (2012) 8624–8630. doi:10.1021/ie3006116. [28] X. Liu, P. Fatehi, Y. Ni, Removal of inhibitors from pre-hydrolysis liquor of kraft-based dissolving pulp production process using adsorption and flocculation processes, Bioresour. Technol. 116 (2012) 492–496. doi:10.1016/j.biortech.2012.03.069. [29] D. Montané, D. Nabarlatz, A. Martorell, V. Torné-Fernández, V. Fierro, Removal of lignin and associated impurities from xylo-oligosaccharides by activated carbon adsorption, Ind. Eng. Chem. Res. 45 (2006) 2294–2302. doi:10.1021/ie051051d. [30] S.V. Mohan, J. Karthikeyan, Removal of lignin and tannin colour from aqueous solution by adsorption onto activated charcoal, Environ. Pollut. 97 (1997) 183–187. doi:10.1016/S0269-7491(97)00025-0. [31] K.I. Andersson, M. Eriksson, M. Norgren, Removal of lignin from wastewater generated by mechanical pulping using activated charcoal and fly ash: Adsorption isotherms and thermodynamics, Ind. Eng. Chem. Res. 50 (2011) 7722–7732. doi:10.1021/ie200378s. [32] T. Sainio, I. Turku, J. Heinonen, Adsorptive removal of fermentation inhibitors from concentrated acid hydrolyzates of lignocellulosic biomass, Bioresour. Technol. 102 (2011) 6048–6057. doi:10.1016/j.biortech.2011.02.107. [33] A. V Pranovich, M. Reunanen, R. Sjoholm, B. Holmbom, Dissolved lignin and other

[34]

[35]

[36] [37] [38]

[39]

[40]

[41]

[42]

[43]

[44] [45]

[46] [47] [48] [49]

[50]

[51]

aromatic substances in thermomechanical pulp waters, J. Wood Chem. Technol. 25 (2005) 109–132. doi:Doi 10.1080/02773810500191575. E. Koivula, M. Kallioinen, T. Sainio, E. Antón, S. Luque, M. Mänttäri, Enhanced membrane filtration of wood hydrolysates for hemicelluloses recovery by pretreatment with polymeric adsorbents, Bioresour. Technol. 143 (2013) 275–281. doi:10.1016/j.biortech.2013.05.129. N. Westerberg, H. Sunner, M. Helander, G. Henriksson, M. Lawoko, A. Rasmuson, Separation of galactoglucomannans, lignin, and lignin-carbohydrate complexes from hotwater-extracted norway spruce by cross-flow filtration and adsorption chromatography, BioResources. 7 (2012) 4501–4516. S. Willför, P. Rehn, A. Sundberg, K. Sundbgerg, B. Holmbom, Recovery of water-soluble acetylgalactogluco-mannans from mechanical pulp of spruce, Tappi J. 2 (2003). T.J. Schwartz, M. Lawoko, Removal of acid-soluble lignin from biomass extracts using Amberlite XAD-4 resin, BioResources. 5 (2010) 2337–2347. R. Wooley, Z. Ma, N.-H.L. Wang, A Nine-Zone Simulating Moving Bed for the Recovery of Glucose and Xylose from Biomass Hydrolyzate, Ind. Eng. Chem. Res. 37 (1998) 3699– 3709. doi:10.1021/ie9800896. Y. Xie, D. Phelps, C.H. Lee, M. Sedlak, N. Ho, N.H.L. Wang, Comparison of two adsorbents for sugar recovery from biomass hydrolyzate, Ind. Eng. Chem. Res. 44 (2005) 6816–6823. doi:10.1021/ie049079x. J. Heinonen, T. Sainio, Chromatographic Recovery of Monosaccharides for the Production of Bioethanol from Wood, Ind. Eng. Chem. Res. 49 (2010) 2907–2915. doi:10.1021/ie901598z. J. Heinonen, S. Kukkonen, T. Sainio, Evolutionary multi-objective optimization based comparison of multi-column chromatographic separation processes for a ternary separation, J. Chromatogr. A. 1358 (2014) 181–191. doi:10.1016/j.chroma.2014.07.004. I. Turku, T. Sainio, Modeling of adsorptive removal of benzalkonium chloride from water with a polymeric adsorbent, Sep. Purif. Technol. 69 (2009) 185–194. doi:10.1016/j.seppur.2009.07.017. Dow Chemical Company, Amberlite XAD-16N - Product data sheet, (n.d.). http://msdssearch.dow.com/PublishedLiteratureDOWCOM/dh_08d3/0901b803808d30bf. pdf?filepath=liquidseps/pdfs/noreg/177-03120.pdf&fromPage=GetDoc. TAPPI standard UM 250 “Acid-soluble lignin in wood and pulp,” in: TAPPI Useful Methods, 1991: pp. 47–48. G. Brunow, K. Lundquist, G. Gellerstedt, Lignin, in: E. Sjöström, R. Alén (Eds.), Anal. Methods Wood Chem. Pulping, Papermak., Springer-Verlag, Berlin Heidelberg, 1999: pp. 77–124. E. Sjöström, Wood chemistry - Fundamentals and applications, 2nd ed., Academic Press, 1993. G. Guiochon, A. Felinger, D.G. Shirazi, A.M. Katti, Fundamentals of Preparative Chomatography, 2006. T. Hatakeyama, H. Hatakeyama, Lignin Structure, Properties, and Application, 2010. doi:10.1007/12_2009_12. J. Asikkala, T. Tamminen, D.S. Argyropoulos, Accurate and reproducible determination of lignin molar mass by acetobromination, J. Agric. Food Chem. 60 (2012) 8968–8973. doi:10.1021/jf303003d. M. Helander, H. Theliander, M. Lawoko, G. Henriksson, L. Zhang, M.E. Lindström, Fractionation of technical lignin: Molecular mass and pH effects, BioResources. 8 (2013) 2270–2282. R. Feistel, G.M. Marion, A Gibbs-Pitzer function for high-salinity seawater

thermodynamics, Prog. Oceanogr. 74 (2007) 515–539. doi:10.1016/j.pocean.2007.04.020. [52] A. Jacobs, O. Dahlman, Absolute molar mass of lignins by size exclusion chromatography and MALDI-TOF mass spectroscopy, Nord. Pulp Pap. Res. J. 15 (2000) 120–127. [53] P.S.B. dos Santos, X. Erdocia, D.A. Gatto, J. Labidi, Characterisation of Kraft lignin separated by gradient acid precipitation, Ind. Crops Prod. 55 (2014) 149–154. doi:10.1016/j.indcrop.2014.01.023.

FIGURE CAPTIONS Figure 1.

Comparison of adsorbents for the removal of soluble lignin from C6 (A) and C5 (B) hydrolysates in a batch reactor. Experimental details in adsorption step: madsorbent = 5 g; mhydrolysate = 50 g; T = 50 °C; teq = 4 h; hydrolysate compositions are given in Table 1. Experimental details in adsorbent regeneration step with 50 wt. % ethanol: madsorbent = 5 g; methanol = 50 g; T = 50 °C; teq = 1 h. Colors: black = amount of lignin removed; grey = monosaccharide loss; white with stripes = amount of lignin desorbed with 50 wt. % ethanol. Error bars mark the standard error.

Figure 2.

Adsorption of lignin from C6 (pH ~ 4.5) and C5 (pH ~ 2) hydrolysates on XAD16N adsorbent. Experimental details: batch experiments with varying phase ratios (1:2–1:200); T = 50 °C; teq = 4 h; hydrolysate compositions are given in Table 1. Symbols: black open ▲= C6 hydrolysate; grey open ●= C5 hydrolysate. Error bars mark the standard error.

Figure 3.

Breakthrough curves of the main compounds of A) C6 and B) C5 hydrolysates with XAD-16N adsorbent bed. Experimental details: Q = 6.8 BV/h (3 mL/min); hydrolysate compositions are given in Table 1; for other details, see caption to Fig. 4. Symbols: black ● = lignin; green ▲ = monosaccharides; red filled ◆ = formic acid; red open ◆ = acetic acid; blue filled ▼ = hydroxymethyl furfural ; blue open ▼ = furfural.

Figure 4.

Effect of flow rate on the breakthrough curve of lignin with XAD-16N adsorbent bed. Flow rate: black ● = 2.3 BV/h (1 mL/min); grey ▲ = 4.5 BV/h (2 mL/min); black ○ = 6.8 BV/h (3 mL/min). Experimental details: dbed = 1.5 cm; hbed = 15 cm; T = 50 °C; feed solution = C5 hydrolysate, composition is given in Table 1. Vertical dashed line represents the void volume of the adsorbent bed.

Figure 5.

Lignin removal level (black ●) and monosaccharide recovery yield (grey ▲) as a function of the volume of A) C6 and B) C5 hydrolysates fed through a XAD-16N column. Values were calculated from the data shown in Fig. 3.

Figure 6.

Desorption of lignin from XAD-16N adsorbent in a column. Adsorbent loaded from C5 hydrolysate. Desorbent composition: black ● = 20 wt. % ethanol; red ◆ = 50 wt. % ethanol; green ▲ = 0.5 mol/L NaOH. Experimental details: Q = 6.8 BV/h (3 mL/min); dbed = 1.5 cm; hbed = 15 cm; T = 50 °C.

Figure 7.

Lignin removal and monosaccharide recovery yield from A) C6 and B) C5 hydrolysates during cyclic runs using XAD-16N adsorbent. Feed volume per cycle: 21 BV with C6 hydrolysate and 16 BV with C5 hydrolysate. See caption to Fig. 4 for experimental conditions. Symbols: ● = lignin removal level; ▲ = monosaccharide recovery yield. Colors: black = 1st cycle, red = 2nd cycle; blue = 3rd cycle; green = 4th cycle; grey = 5th cycle.

Figure 8.

Separation of lignin from mix hydrolysate (Table 1) with XAD-16N adsorbent in a pilot scale adsorption column. Only the loading step is shown. Experimental details: Q = 6.8 BV/h; Vbed = 11.2 L; hbed = 73 cm; T = 50 °C. Symbols: black filled ● = lignin concentration on 1st cycle; black open ● = lignin concentration on 2nd cycle; grey filled ◆ = lignin removal level on 1st cycle; grey open ◆ = lignin removal level on 2nd cycle; green ▲ = monosaccharide concentrations on 1st cycle.

TABLES Table 1.

Compositions of the C5 and C6 hydrolysates used in this work. Concentration, g/L Component “C5 hydrolysate” * “C6 hydrolysate” * “Mix hydrolysate” ** Lignin 13.98 3.39 6.67 Glucose 10.59 74.94 75.26 Xylose 83.47 17.95 25.50 Galactose 6.32 0.88 1.90 Mannose 10.85 2.46 3.70 Total monosaccharides 111.23 96.23 106.36 Formic acid 1.33 0.52 n.a. Acetic acid 22.89 6.18 n.a. Furfural 5.51 2.2 n.a. 5-hydroxymethyl furfural 0.48 0.35 n.a. * Used in laboratory scale experiments. ** Used in pilot scale experiments.

Table 2.

Adsorbents used in the lignin separation and recovery experiments. Functional Adsorbent Provider Matrix dp, mm AS, m2/g group CS11GC Finex PS-DVB 0.32 n.a. Sulfonic acid XAD-16N Dow PS-DVB 0.56-0.71 ≥ 800 none XAD-7HP Dow Polyacrylate 0.51-0.71 ≥ 380 none XAD-4 Dow PS-DVB 0.49-0.69 ≥ 750 none Macronet MN200 Purolite PS-DVB 0.45-0.62 ≥ 900 none Macronet MN270 Purolite PS-DVB 0.6-0.85 ≥ 1200 none

Table 3.

Amount of lignin desorbed from loaded XAD-16N adsorbent in a column. The values are calculated from the data shown in Fig. 6. Desorbent Desorption ratio, % 20 wt. % ethanol 78.1 % 50 wt. % ethanol 100 % 0.5 mol/L NaOH 41.4 %

HIGHLIGHTS -

Separation of lignin from hydrolysates of lignocellulose by adsorption was studied

-

Efficient separation was achieved using polystyrene-based XAD-16N adsorbent

-

Adsorbed lignin could be desorbed with 100 % desorption ration with 50 wt.% ethanol

-

Lignin with high purity can be produced with the presented adsorption process

-

Applicability of the process on large scale was demonstrated successfully