Continuous liquid – liquid extraction to recover lignin and furanics from lactic acid: Choline chloride deep eutectic solvent after cooking of spruce

Lactic acid: choline chloride mixed in a 10:1 ratio is known to be a suitable delignification solvent. This deep eutectic solvent (DES) can produce lignin and hemicellulose derivatives as byproducts to cellulose. Hemicellulose byproducts include furanic compounds like furfural and 5-HMF. Sustainable and economically viable fractionation of lignin and furanics from the DES is required. After removing the larger lignin ( > 2500 Da) through water precipitation, non-precipitated molecules such as smaller lignin molecules and furanics can be retrieved through liquid – liquid extraction (LLX). In a previous investigation on solvent extraction conducted in a batch mode, 2-MTHF and guaiacol have been identified as potential solvents. This study investigates the application of a continuous centrifugal contactor separator (CCCS) and Karr reciprocating plate column for LLX from aqueous DES-black liquor remaining after water-precipitation, was employed as the feed. Single stage CCCS operation exhibited extraction yields close to equilibrium yields, thus almost 100 % extraction efficiency. Performing a two-stage countercurrent LLX process via CCCS showed increased removal for furfural and 5-HMF extraction, while lignin extraction yield remained at about 90 %, comparable to single-stage extraction. Karr column countercurrent extraction with different solvent-to-feed ratios (0.5 and 1) yielded approximately 99 % furfural extraction. For 2-MTHF system, 5-HMF achieved 99 % extraction yield at a solvent-to-feed ratio of 1, while guaiacol yielded around 90 % at both solvent-to-feed ratios. However, 5-HMF extraction using 2-MTHF with a ratio of 0.5 required more stages. For guaiacol improvement of lignin extraction applying more stages was hindered by hydrophobic characteristic of residual lignin in raffinate. Conversely, 2-MTHF ’ s multi-stage approach enhanced smaller lignin extraction influenced by increased lactic acid leaching into solvent. Both CCCS and Karr column methods proved viable for counter-current extraction with both solvents.


Introduction
Effective and sustainable fractionation of biomass components is a major step needed in the process of replacing fossil-based chemical building blocks with bio-based ones, making use of biorefineries [1][2][3].Using lignocellulosic biomass, the most accessible and abundant biomass on earth, as feedstock for the replacement of petroleum derivatives is a viable option [4,5].Lignocellulosic biomass has a complex structure and is made up of cellulose (40-60 %), lignin (15-30 %), and hemicellulose (10-40 %) [6][7][8].Developing a sustainable method for breaking down lignocellulose components and utilize them all as effectively as possible is still challenging [9][10][11] due to the highly complex entanglement of all these biopolymers.This challenge also applies to the paper and pulp industry, the primary consumer of this biomass [12,13], especially those aiming to not only use the cellulose fibers but also find a way to make use of the other components, i.e. lignin and hemicellulose.
Nowadays, the kraft pulping method is the most well-known, welldeveloped and energy-efficient method used in the paper and pulp industry [14].In this method, cellulose is separated from lignocellulose biomass using aqueous sodium hydroxide and sulfides.After the cellulose has been separated, the majority of the waste stream, known as black liquor, is sent to the evaporators to recover volatiles from the solvent, and then the concentrated stream is sent to the boilers to produce heat by burning of the lignin [15,16].This waste stream contains next to the lignin also hemicellulose, and both of which have the potential to be converted into chemicals with a higher value [13,16,17].To address this issue, researchers have been working on methods to efficiently extract lignin and hemicellulose from black liquor [18][19][20], or to create more atom-efficient pulping processes like organosolv pulping [21], from which sulfur-free lignin can be extracted as a byproduct with useful application potential.
Deep eutectic solvents (DESs) are a promising class of solvents that have emerged over the past two decades [22].They share plenty of characteristics with ionic liquids, but deep eutectic solvents have several important advantages.Depending on the starting materials, DESs can consist of two relatively inexpensive, biocompatible, biodegradable and non-toxic components which are mixed to prepare the desired solvent [23,24], conceptually such solvents are composite solvents [25].DESs consist of a hydrogen bond donor (HBD) and a hydrogen bond acceptor (HDA) component.The highly non-ideal interaction between the HBD and HDA caused by the hydrogen bonding results in a charge delocalization which drastically lowers the melting point of the mixture (>50 • C) relative to the ideal mixture [23].DESs are proposed as a possible alternative to the Kraft pulping process for the delignification of biomass [14,26,27].Research has shown that the combination of lactic acid and choline chloride is an effective DES for delignifying lignocellulose [7,[28][29][30].Previous research conducted by Smink et al. showed in a DES-based delignification process using lactic acid and choline chloride, the chloride anion is responsible for a faster delignification [7] and studies on the reaction mechanism have shown that chloride is catalyzing the cleavage of the β-O-4 bond of the lignin polymer [27].Hemicellulose, along with lignin, dissolves in DES during the pulping process and can be transformed into hemicellulose derivatives.Furanics, including furfural and 5-hydroxymethyl-2-furfural (5-HMF), are hemicellulose derivatives that result from the dehydration of C 6 and C 5 sugars.Therefore, the mixture leaving the pulping reactor consists of the liberated cellulose fibers and DES containing hemicellulose and hemicellulose derivatives and lignin originating from the wood biomass [28].In the conceptual process, the cellulose fibers are removed from the mixture in a pressing and water washing step [28].The remaining mixture is called DES-black liquor.Valorization of the different lignin fractions with different molecular weights and the furanics from this stream is essential for development of the DES-pulping process [28,31,32].For this reason, economically viable and environmentally friendly technologies must be used to sustainably fractionate lignin and furanics from the stream [33].
In addition to ultrafiltration [33] and cold water precipitation [34], liquid-liquid extraction (LLX) is a potential separation technique to recover lignin and furanics from DES-black liquor if a suitable solvent is used [32,34,35].A combination of several separation techniques would be preferable, because that will allow forboth recovery and fractionation of lignin with varying molecular weights and furanics from DES-black liquor [34].Previous research suggested a two-step DES-regeneration process involving cold-water precipitation and LLX for complete lignin fraction and furanic recovery from the DES.When water is added to a DES-black liquor with ratio of 3.5:1 [g/g] water to DES-black liquor, the lignin fractions with a higher molecular weight (>2500 g/mol) are precipitated [32,34].The water required for precipitation is already available in the process by re-purposing the exiting stream of the pulp washing step [28].The smaller molecular weight lignin fraction which remained is extracted in a subsequent LLX step which is improved by the presence of water in the DES mixture [28,34], while also the hydrodynamics in the operation improves due to the precipitation of the largest lignins.Essentially, when no water precipitation is done to remove high molecular weight lignins, emulsification is regularly observed in LLX.The furanics (e.g.furfural and 5-hydroxymethyl-2-furfural (5-HMF)) and other hemicellulose byproducts can potentially be extracted in this step.Therefore, both techniques will complement each other to recover the lignin and furanics.Based on the previous results [32,35], guaiacol and 2-methyltetrahydrofuran (2-MTHF) are two solvent options that may be considered for use in this process.The adoption of extraction systems utilizing these bio-based solvents offers for future industrial processing the advantage that the solvents utilized may be obtained from the biorefinery itself, rather than from oil refineries, which can help addressing environmental concerns and reduce dependency on fossil sources.The solvents in the current study applied, 2-MTHF as well as guaiacol, are derived from renewable sources.
The earlier research [32] focused on the single-stage batch LLX approach to identify suitable solvents.However, there remained a notable gap in existing knowledge regarding the continuous processing of DES-black liquor for the recovery of components and the regeneration of DES.Given the complex nature of this multi-component stream after DES-pulping, a comprehensive investigation into the efficiency and behavior of the solvent extraction process within continuous extraction equipment is crucial for further development of processes, and studied in the work described in this paper.Continuous LLX is a widely used separation technique in industries such as pharmaceuticals, chemicals, and biotechnology [36][37][38][39][40]. Continuous operation is typical for large scale industrial processes, offering improved efficiency and better process control, enhancing the productivity and quality of LLX processes [36,41].In this work, the potential of two typical types of continuous equipment have been investigated, being the continuous centrifugal contactor separator (CCCS) and the Karr column.The LLX has been applied after cold water precipitation, the intended combination of unit operations in the DES-regeneration.
The first approach towards continuous LLX in this study involves the utilization of CCCS [42].A continuous centrifuge contactor separator is a specialized apparatus designed to separate immiscible liquids based on their density differences using centrifugal force.The feed mixture and the extracting solvent are continuously fed into the contactor separator, which induces phase separation due to the difference in their densities.The high centrifugal force generated within the system enhances phase separation and promotes faster and more effective mass transfer, resulting in improved extraction efficiency and higher product yields [39,40,43].Additionally, the continuous nature of the process allows for better control over extraction parameters such as residence time, flow rates, and solvent-to-feed ratios.This control facilitates optimization of the extraction process and ensures consistent and reproducible results [38,42,44].The reason for applying this equipment is that it allows large flows at a small equipment size, which is interesting for scale-up at pilot scale.The same scale device may be used on the lab scale with only a few liters solvent, up to hundreds of liters in the pilot scale.
A Karr reciprocating plate column operates on the principle of countercurrent extraction [45].The liquid phases flow in opposite directions, allowing for efficient mass transfer between the phases.The extraction occurs as the feed mixture and solvent move through the column, and the desired components are selectively transferred from one phase to the other [46,47].The hydrodynamic and mass transfer performance of Karr column has been subject for many studies for different LLX systems including ionic liquid systems [47][48][49][50].Although scaling up to pilot scale is more difficult with this type of equipment, and a new column would be necessary, the option to investigate multistage extraction is highly valuable.
In the current work, we thus report on the use of both the CCCS and the Karr column for performing LLX in the DES regeneration process.The feed mixture that was considered was an aqueous dilution of DESblack liquor obtained by precipitation of larger lignin by adding water to DES-black liquor at a ratio of 3.5 [g/g].Because the larger lignins precipitate fully at this water to DES ratio [32], only lignin with a molecular weight of lower than 2500 Da was present.Having two CCCS devices available, both single-stage and two-stage countercurrent LLX in CCCS were carried out for the solvents 2-MTHF and guaiacol.Furthermore, countercurrent extraction for both solvents was performed in a Karr extraction column where the number of theoretical stages was determined based on equilibrium data using the analytical Kremser method.For both equipment, the extraction yields for smaller molecular weight lignin (<2500 Da), furfural and 5-HMF were evaluated.

Chemicals used
2-MTHF (>99 wt%, 250 ppm BHT), and Guaiacol (natural, >99 wt %) were supplied from Sigma-Aldrich.For feed and sample preparation and analysis, Milli-Q water was used.DES-black liquor has been provided by Centre Technique du papier (CTP), where the delignification of biomass was performed at pilot scale at 130 • C for 4 h with a DES to wood (Sappi Austria Produktions-GmbH & Co. KG.spruce chips) ratio of 8.6 [g/g] and DES composed of lactic acid and choline chloride with a 10 to 1 [g/g] weight ratio.A Karl-Fischer titration showed a weight percentage of 21 ± 1 % water in the DES-black liquor.

Chemical analysis 2.2.1. GPC
The molar weight distribution of the samples was calculated using gel permeation chromatography (GPC).This was carried out using an Agilent 1260 Infinity series with a refractive index detector and a UV detector operating at 254 nm with a set of three GPC PLgel 3 m MIXED-E column in series.To facilitate complete dissolution of the lignin, tetrahydrofuran (THF) and water were used as the mobile phase of the column at a volume ratio of 95:5 (v:v).40 mg of sample was dissolved in 4 mL of the mobile phase.The operating conditions of the GPC column were a flow rate of 1 mL/min at 40 • C. The polystyrene solutions with molecular weights ranging from 162 to 27,810 Da were used to calibrate molar weight distributions.To determine the mass ratio of lignin in each sample, the UV absorbance corrected by dilution factor and sample density.

HPLC
The components other than lignin in the samples were measured using high performance liquid chromatography (HPLC).In order to do this, a HPLC machine from the Agilent 1200 series was used with a UVdetector operating at 285 nm, and a refractive index detector kept at 55 • C. The applied column was a Hi-Plex-H column, operated at 60 • C. Sulfuric acid, at a flow rate of 0.6 mL/min, made up the mobile phase.

Experimental procedure 2.3.1. Feed preparation and solvent choice
Milli-Q water was added to the DES-black liquor that was generated in a pilot-scaled pulping experiment at CTP (see section 2.1) in a ratio of 3.5:1 [g/g].After stirring the solution for an hour at room temperature with a magnetic stirrer, it was allowed to settle for at least 24 h to allow the lignin with a higher molecular weight to precipitate [32,34].Then, the precipitated lignin was removed from the liquid by vacuum filtration.The filtrate was used as feed for the LLX study in this work.The solvents that have been used in this study were based on earlier work in batch extraction studies [32,34,35], and the methodology to first precipitate the heavy lignins was also based on earlier work [34], in particular, initial measurements without this precipitation have shown that LLX cannot be operated stably due to persistent emulsification.All experiments described in the next subsections are therefore based on DES-black liquor that was pretreated by precipitation of the heaviest lignins.

LLX in CCCS (Single-stage and Two-stage countercurrent)
Centrifuge Model HC-2, made by Centri Tech Separation Twente (CTST) is used to conduct CCCS experiments in continuous mode.Fig. 1 (a) shows the schematic for the single-stage CCCS setup.As shown, the CCCS itself consists of several subsections.The two (partially) immiscible phases are individually introduced into the annular zone surrounding the centrifuge at either the light phase or heavy phase inlet, where they are efficiently mixed.The mixture then enters the rotor chamber from an opening at the bottom.The centrifugal forces will accelerate the settling, and therefore separation, of the two phases.The two phases are separately collected in two outlets by utilizing a light phase weir and heavy phase weir with different sizes which are mixture specific.Fig. 1(b) displays a two-stage countercurrent extraction configuration consisting of two CCCSs.The experiments were operated by starting the engine of the CCCSs and following that the heavy phase pump was started at desired flowrate.The density of phases was used to calculate the corresponding volume flowrate for desired mass flow rate.After filling up the CCCS with the heavy phase and establishment of the heavy phase outlet, the light phase pump was started.The samples were taken over time from outlet of both phases and analyzed with GPC and HPLC.
The operating conditions (flowrates and rotor speed) of the centrifuge for each solvent were chosen in order to minimize the contamination of each phase, obtaining almost full separation of both phases at the outlets.The size of the light phase weir in the centrifuge was fixed and does not change between experiments using different solvents.For 2-MTHF system, the heavy phase weir with a diameter of 24.77 mm (0.975 in.) was used.The light phase is 2-MTHF and the heavy phase is the aqueous DES-black liquor.For guaiacol as the solvent, the heavy phase weir with a diameter of 22.86 mm (0.90 in.) was used.For both solvents each phase was continuously fed to the device at 30 g/min and the rotor speed of the centrifuge was set at 3600 RPM.Under the mentioned operating conditions, the liquid hold-up of the centrifuge was determined to be 170 mL.The extraction experiments using guaiacol were done by pre-heating the aqueous DES-black liquor and solvent up to 30-35 • C to prevent the crystallization of guaiacol.For the experiments with 2-MTHF as solvent, the two phases were entered the setup at room temperatures.Each experiment was carried out twice.

LLX in Karr column
The Karr column, referred to as bench top unit (BTU), made by Koch Modular Process Systems, is depicted in Fig. 2. The set-up consists of an agitated column with an internal diameter of 5/8″ containing reciprocating internals.The stainless steel plate stack reciprocates up and down which is a typical characteristic for the Karr column.For both the solvents 2-MTHF and guaiacol, the solvent formed the dispersed phase, and the DES-black liquor formed the continuous phase.The setup is shown in Fig. 2 when the solvent is in the light phase.If the solvent is the heavy phase (i.e.Guaiacol), the interface in the bottom chamber of the column is regulated.In Table 1, column specifications have been provided.The agitation of the column is realized using air, and the air flows needed to realize a certain frequency were calibrated as a function of the agitation speed expressed in strokes per minute (SPM).The solvent phase pump was started when the continuous phase outflow established.The flowrates of the aqueous DES-black liquor and solvent were fixed at 30 [g/ min].After an interface was established in the column, the agitator mechanism was switched on at 80 SPM.The motor of the agitator mechanism was powered by instrument air.The flow to the motor was regulated with a manual needle valve.The system at start-up was in steady state after five column turnovers.The turnover was calculated by dividing the volume of the column by the sum of the flowrates of both the feed and solvent.Three column turnovers were experimentally determined to be sufficient for each successive experiment to reach steady state.To avoid crystallization during the guaiacol extraction experiments, the aqueous DES-black liquor and solvent were heated to 30-35 • C before use, while the two phases were introduced into the apparatus at room temperature for 2-MTHF system experiments.Samples were taken over time from the heavy phase and the light phase outlets and analyzed by GPC and HPLC.All experiments were repeated at least twice.

Mathematical formulation for data analysis
The extraction yields, Y i (%), for furanics and lignin were calculated using Eq.(1): where w i,F and w i,S are the measured mass fraction of component i in the feed and the solvent phase in [g/g], respectively.S is the mass flowrate of the solvent phase and F is the mass flowrate of the feed in [g/min].
Similarly, the leaching of lactic acid in the solvent phase over time was calculated using Eq.(2).
It should be noted that in the above equations, when the concentrations of components in the solvent phase were too low to be determined by analysis, the concentrations of the raffinate phase were used only, and solvent phase concentrations were determined by mass balance.HPLC was employed to quantify the concentrations of furanics and lactic acid in the samples, while GPC was utilized for lignin analysis.
The number of theoretical stages in the Karr column was determined using the Kremser equation for a dilute system [51]: In Eq. ( 3) w i,R,N is mass fraction of component i coming out from stage N of a multistage countercurrent extraction.E i is extraction factor determined by employing Eq. ( 4), where D i represents distribution coefficient and S/F denotes the solvent-to-feed ratio.
Table 2 provides distribution coefficients for components as measured in recent work [32].

LLX in CCCS -Stability of operation
The experiments in the CCCS, as described in section 3.1, were continued for 30 min, for both single-stage extraction and two-stage countercurrent extraction.Under operational conditions (feed and solvent flowrate of 30 g/min and rotor speed of 3600 RPM), the 2-MTHF system showed a raffinate phase that was contaminated with up to 10 vol% of extract phase at the end of the extraction.For the guaiacol system, the contamination of the extract phase in the raffinate phase was below 5 % of total volume under the same operational conditions.For both solvents, contamination of the extract phase by raffinate was less than 2 % of total volume.Using Eq. ( 1), the extraction yields (Y i ) for furfural, 5-HMF, and lignin (<2500 Da) were calculated.Leaching of lactic acid into solvents during extraction was calculated using Eq. ( 2).The molar mass distributions of feed and raffinate phase after singlestage countercurrent extraction with 2-MTHF and guaiacol are shown in supplementary information.S.

Single-stage
The single-stage extraction yields in CCCS for lignin (2500 Da), furfural, and 5-HMF, as well as the leaching of lactic acid into the  According to previous research [32], and as demonstrated in Table 2, the distribution coefficients for furanics in guaiacol are higher than those for 2-MTHF, which results in a higher extraction yield for these components when guaiacol is used.Also, lactic acid leaching in guaiacol is less than that of 2-MTHF.In general, lignin and furfural have higher extraction yields than 5-HMF for both solvents.These results are also evident in the continuous extraction results displayed in Fig. 3.

Two-stage countercurrent
A two-stage countercurrent extraction process was conducted for both the solvents 2-MTHF and guaiacol with a solvent-to-feed massbased ratio of 1. Fig. 4 illustrates the extraction yield of lignin and furanics, as well as the leaching of lactic acid into the solvent.In all situation the steady-state condition reaches within 15 min.As can be seen for lignin (<2500 Da), the extraction yield does not exhibit a substantial increase when compared to the results of single-stage extraction (Fig. 3 (a)).A possible explanation is that the residual lignin present in the feed exhibits notable distinctions in its properties compared to the lignin that has already been extracted.The different characteristics will be influenced by a different molecular weight and the number of functional groups that can influence the distribution coefficient [32].This observation suggests that the lignin present in the raffinate after a first stage of the extraction (Supplementary information, Figure S), may exhibit a higher degree of hydrophilicity, making its extractability using a nonpolar solvent much lower.Therefore, in this situation, unless significantly higher solvent to feed ratios would be applied, additional stages would not further improve the extraction yield of lignin.
Fig. 4(b) shows the extraction yield of furfural.The utilization of guaiacol as a solvent has been observed to result in an approximate 5 % Fig. 3.The extraction yield of (a) lignin, (b) furfural, (c) 5-HMF and (d) leaching of lactic acid, in the single-stage CCCS using 2-MTHF and guaiacol as solvent.For both solvents each phase was continuously fed to the device at 30 g/min and the rotor speed of the centrifuge was set at 3600 RPM.The lines are batch equilibrium extraction yields and leaning of lactic acid.
increase in the extraction yield of furfural.This improvement allows for nearly complete recovery of furfural within two stages.The extraction yield for 2-MTHF is approximately improved by 15 % compared to the single-stage extraction, resulting in an enhancement of up to 90 %.The utilization of a two-stage countercurrent extraction method with guaiacol does not appear to enhance the extraction efficiency of 5-HMF, as the extraction yield only exhibits a marginal increase of 2.5 %.However, it appears that increasing the number of stages for 2-MTHF has proven to be advantageous, as it has resulted in an enhancement of the extraction yield by around 15 % (Fig. 4(c)).As depicted in Fig. 4 (d), the lactic acid leaching into guaiacol for the two-stage countercurrent process exhibits an increase of 2-3 wt%, in stark contrast to the behavior observed in the 2-MTHF scenario.In the latter case, increasing the number of stages from a single stage to two stages results in a further increase (around 20 %) in the amount of lactic acid leached to 2-MTHF.The maximum solubility limit of lactic acid in guaiacol seems to have been reached in a single step.On the other hand, lactic acid is much more soluble in 2-MTHF and has not yet reached its maximum solubility limit in the solvent.

LLX in Karr column
Countercurrent LLX was conducted in a Karr column using solvent-to-feed ratios of 1 and 0.5.Fig. 5 illustrates the extraction yield for each component, as well as the leaching of lactic acid into solvents.Both solvent systems used in the Karr column exhibit an extraction yield of roughly 99 % for furfural for solvent-to-feed ratios of 0.5 and 1.In contrast to guaiacol, where extraction yields of 5-HMF were attained 90 % at solvent-to-feed ratios of 0.5 and 1, 5-HMF displayed a high extraction yield of 99 % when the solvent-to-feed ratio for 2-MTHF was 1.With a solvent-to-feed ratio of 0.5, the extraction yield of 5-HMF with 2-MTHF is around 80 %.The extraction yield for lignin is around 99 % for 2-MTHF and around 90 % for guaiacol at both solvent-to-feed ratios.Leaching of lactic acid by increasing the number of stages of countercurrent extraction employing a Karr column increases using 2-MTHF while it is same as single stage for guaiacol.This also has been seen in two-stage countercurrent LLX in CCCS (Fig. 4(d)).It can be attributed to the solubility of lactic acid in guaiacol, which has already reached its maximum in a single stage.However, in the case of 2-MTHF, the solubility of lactic acid appears to be higher, and increasing the number of stages enhances the leaching of lactic acid in 2-MTHF.
The Kremser equation was employed to calculate the number of theoretical stages for countercurrent liquid-liquid extraction (LLX) using 2-MTHF and guaiacol as solvents, with solvent-to-feed ratios (S/F) of 0.5 and 1 (Table 3).The objective is to achieve a 99 % extraction yield for every individual component.The validity of the Kremser equation is Fig. 4. The extraction yield of (a) lignin, (b) furfural, (c) 5-HMF and (d) leaching of lactic acid, in the two-stage countercurrent CCCS using 2-MTHF and guaiacol as solvent.Both the solvents and the DES-based phase were continuously fed to the device at 30 g/min and the rotor speed of the centrifuge was set at 3600 RPM.conditioned upon the immiscibility of the phases, a low solute concentration, and the constancy of the distribution coefficient.In this study, it is assumed that both solvents are immiscible with the feed [52,53].
However, it should be noted that in reality, 2-MTHF exhibits a leaching rate of approximately 8.5-9.3 % into the feed, while guaiacol leaches at a rate of 3.2-3.4% for both studied solvent-to-feed ratios.Due to the relatively high concentration of lactic acid in the feed, the Kremser methods may not provide accurate results for lactic acid leaching, particularly in the case of 2-MTHF, as the solvent leaching into the feed is also higher than that of guaiacol.However, the required data is unavailable for creating relevant (pseudo) ternary diagrams.Therefore, despite the fact that the boundary conditions are not met, it is still used as an approximation for comparison with guaiacol results.
The results indicate that in order to achieve a 99 % furfural recovery using a solvent-to-feed ratio of 1, it is necessary to employ 4-5 equilibrium stages for 2-MTHF.Similarly, for guaiacol, 2-3 equilibrium stages are required to achieve the same level of recovery.When comparing the experimental furfural extraction yield (Fig. 4), it is observed that both cases in the experimental Karr column exhibit a yield of approximately 99 %.This suggests that the minimum required stages for furfural extraction have already been achieved.Based on the Kremser equation, achieving the necessary number of stages for 5-HMF with a solvent-to-feed ratio of 0.5 presents a significant challenge for 2-MTHF.In this case, the experimental Karr column did not achieve a 99 % recovery at its outlet.In order to achieve a 99 % recovery of 5-HMF, it is necessary to employ 8-9 stages for the guaiacol process, with a solvent-

Table 3
Number of theoretical stages (Nts) for countercurrent LLX for 2-MTHF and guaiacol as solvents with solvent-to-feed ratios (S/F) of 0.5 and 1, calculated using Kremser equation (Eq.( 3)).The objective is to achieve a 99% extraction yield for each component.to-feed ratio of 0.5.The experimental results indicate a recovery rate of approximately 90 % for the outlet flow, suggesting that the Karr column was operating with a lower number of stages in this particular case.However, when it comes to lignin, it is observed that even though the number of stages required is less than 2 for guaiacol at a solvent-to-feed ratio of 0.5 or 1, experimental results show that a 99 % recovery is not attained.As explained in section 3.1.2,this behavior can be linked to the inherent characteristic of the residual lignin present in the raffinate, which exhibits a notable resistance towards extraction by non-polar solvents.However in the Karr column, for the case of 2-MTHF, the extraction of smaller lignin can be enhanced by employing more than two stages.Conversely, for guaiacol, the maximum achievable recovery, which is around 90 %, is already attained with a two-stage system.According to our previous study [32], hydrogen bonds between lignin and solvent are responsible for the extraction of lignin into both solvents.During each stage of the extraction process, there is an increase in the leaching of lactic acid into 2-MTHF (Fig. 5(b)).This results in an enhanced hydrophobic nature of the 2-MTHF [32], enabling it to effectively extract smaller lignin that possess hydrophobic characteristics.The extraction yield achieved in this process is 99 %.Since the lactic acid content remains constant for guaiacol (Fig. 5(b)), it can be concluded that the extraction of lignin does not improve after the first stage.

Evaluation of the investigated extraction equipment
Two types of continuous extraction equipment have been investigated in this study, each of them with advantages and in this subsection, some characteristics of the two types of equipment are discussed.The selection of extraction equipment for large scale industrial processes may be done on the basis of decision schemes, and among the critical parameters for such decision making is the emulsification tendency.Karr columns are a class of extraction columns that offer multistage countercurrent operation and provide beneficial behavior for systems prone to emulsion formation.This enables enhanced operational adaptability while upholding the efficacy of the extraction procedure [54].Since it was experienced in these studies that lignin-rich streams of DES can be emulsifying, this type of column that was readily available in the lab was studied in addition to the CCCS that was originally the type of equipment that was selected.
The CCCS had been selected on the basis of the wide flow rate range at which the laboratory scaled device can work (typically around 10 mL/ min up to 2 L/min), which offered opportunities to separate effluents from small scale laboratory experiments as well as working up the DESblack liquor from the pilot scaled trials at one of the consortium partners.Further scalability towards full scale once selected for scaling up would also be straightforward.On the downside, each CCCS device is intrinsically maximal one equilibrium stage, which might become capital expensive when many stages are necessary.Furthermore, studies on CCCS have shown that it faces specific challenges, especially when operating at low rotational speed or high feed flowrate, which contributes to the light phase being contaminated by the heavy phase in outlet.This contamination necessitates the use of higher rotational speed, which results in higher energy usage that adds to the overall cost of the process [55,56].
Both CCCS and Karr columns are ultimately regarded as viable options for countercurrent extraction.The decision between the two alternatives is determined by various factors, including specific process requirements, the composition of the mixture being processed, desired selectivity, scalability considerations, and the overall efficiency of the entire DES regeneration process.Especially when mobility is important, the CCCS is a strong choice, while for larger numbers of stages, a column is a better choice, at limited maximal flows with comparable dimensions of the equipment.

Conclusion
This study investigated the use of two different types of continuous LLX equipment that can be practically implemented to regenerate a DESblack liquor diluted with water that was based on lactic acid-choline chloride that was obtained after lignin precipitation by water precipitation.Continuous countercurrent extraction processes were done using both CCCS and a Karr column.The biomass-derived solvents, 2-MTHF and guaiacol, were employed in single-stage and two-stage extractions in the CCCS, and in multistage extractions in the Karr column.The results demonstrated for the CCCS fed at 30 g/min for both phases efficient single-stage extraction in less than 10 min for 2-MTHF and around 5 min for guaiacol, with extraction yields close to batch yields, suggesting approximately 100 % efficiency for CCCS.In two-stage countercurrent processes, improvements were observed in furfural and 5-HMF extraction efficiencies, while lignin extraction remained consistently around 90 %, similar to single-stage extraction.Lactic acid extraction increased with 2-MTHF, while it remained consistent with guaiacol.
The countercurrent extraction in the Karr column, employing two different solvent-to-feed ratios (0.5 and 1), further enhanced our understanding of the process.Notably, both solvent systems in the Karr column demonstrated an extraction yield of approximately 99 % for furfural (for both solvent-to-feed ratios of 0.5 and 1).5-HMF exhibited a high extraction yield of 99 % when the solvent-to-feed ratio was 1 for 2-MTHF, while for guaiacol, extraction yields of approximately 90 % were achieved at solvent-to-feed ratios of 0.5 and 1.The extraction of 5-HMF using 2-MTHF with a solvent-to-feed ratio of 0.5 necessitates a greater number of equilibrium stages.Consequently, the Karr column achieved a maximum extraction yield of 80 % for 2-MTHF.According to the Kremser equation, guaiacol necessitates less than two stages for achieving a 99 % lignin extraction.However, experimental results show that 99 % recovery is not possible due to residual lignin's inherent resistance to non-polar solvents.In contrast to guaiacol, 2-MTHF's multi-stage approach improves smaller lignin extraction due to increased lactic acid leaching, making it more hydrophobic and enabling 99 % yield.
In conclusion, in this first study employing continuous solvent extraction for DES regeneration after delignification, both CCCS and the Karr column are viable for countercurrent extraction, in particular after removing heavy lignins prior to the extraction and with the choice depending on specific process requirements, such as mobility and a large scalability factor at the same equipment size being beneficial for the CCCS, and the high extraction yield being beneficial for the Karr column.

Fig. 1 .
Fig. 1.The schematic representation of the CCCS setup for (a) single-stage and (b) two-stage countercurrent operation.Gray-light color indicates the light phase, patterned-green color indicates the mixing zone, and dark-green color indicates the heavy phase.

Fig. 2 .
Fig. 2. The schematic representation of the Karr column BTU setup in the situation that the solvent is the lighter phase (i.e.2-MTHF).If the solvent the interface in the column's bottom chamber is controlled.

Fig. 5 .
Fig. 5. (a)The extraction yield of lignin, furfural and 5-HMF and (b) leaching of lactic acid, in the countercurrent Karr column extraction using 2-MTHF and guaiacol as solvent with the solvent-to-feed ratio (S/F) of 1 and 0.5.For both solvents the agitation speed set at 80 SPM.

Table 1
Specification of the Karr column used in this work.

Table 2 Distribution
[32]e illustrated in Fig.3.In this figure also the equilibrium extraction yields based on earlier work[32]are displayed.As can be observed for both solvents, the extraction yields for all components will reach the batch extraction yield resulting in almost 100 % extraction efficiency.Steady-state operation for 2-MTHF was achieved after approximately 10 min, which is nearly three times the residence time.However, for guaiacol, steady state was reached in less than 10 min (around 5 min-approximately two times the residence time) for all components.Guaiacol thus achieves steady-state more rapidly than MTHF.This is probably a result of increased transfer of lactic acid into 2-MTHF compared with guaiacol, which causes a delay in mass transfer of the other components into 2-MTHF.The results clearly show the suitability of CCCS in continuous DES-regeneration process to extract furanics and lignin from DES.
[32]ficients for the components in the batch LLX extraction as reported in[32](This work is openly licensed via CC BY 4.0).solvent