Regeneration of Deep Eutectic Solvent Post Biomass Delignification
Mahsa Gholami is a PhD student in the Department Sustainable Process Technology. Promotors are prof.dr.ir. B. Schuur, prof.dr. S.R.A. Kersten and prof.dr.ing. M.B. Franke from the Faculty of Science & Technology.
Recent years have seen a surge in interest in lignocellulosic biomass as a renewable resource across various industries, owing to its potential to reduce the reliance on fossil fuels and mitigate environmental impacts. Lignocellulosic biomass, comprised mainly of cellulose, hemicellulose, and lignin, poses a complex matrix requiring efficient separation techniques for value-added fractionation. Lignin, in particular, has garnered attention for its potential valorization into high-value chemicals and materials, with the paper and pulp industry serving as a primary source of lignin as a by-product. While the Kraft pulping process effectively extracts cellulose, isolating lignin as byproduct is hindered by the sulfur and high repolymerization degree, reducing the value, which leads to exploration of novel solvent systems like deep eutectic solvents (DESs) that may obtain higher value lignin next to the cellulose.
Chapter 1 reviews delignification processes utilizing four categories of green solvents, starting with alcohols and organic acids (collectively termed organosolv processes), followed by ionic liquids (IL)s and DESs. Alcohols and some acids benefit from their vapor pressure, enabling regeneration through evaporation and distillation. In contrast, the low vapor pressure of ILs and DESs offers advantages during pulping, such as less severe operational conditions, reduced energy usage, and opportunities for multiproduct biorefineries. However, low-volatility solvents present challenges in solvent regeneration. The chapter emphasizes the recyclability and reusability of these solvents alongside the delignification process.
In DES-based pulping, next to the cellulose fibers used for paper making, the resulting liquor phase leaving the reactor is termed DES-black liquor, and contains lignin, hemicellulose, and derivatives like furfural and 5-HMF. Among DES formulations, lactic acid-choline chloride emerges as an effective mixture for lignin and hemicellulose removal from the wood matrix. Key for the success of this process is the regeneration of DES, crucial for both solvent retrieval and the extraction of dissolved lignin and hemicellulose post-biomass delignification, forming the core focus of this thesis. The strategies for DES regeneration as studied are divided into two distinct methods. The first method, referred to as the LLX-based approach, encompasses chapters 2 through 4. These chapters delve into the experimental investigation associated with DES regeneration through liquid-liquid extraction (LLX). The second method, termed the membrane-based approach, is covered in chapters 5 and 6. This section focuses on the experimental investigation related to DES regeneration utilizing membrane-based techniques. Lastly, Chapter 7 provides detailed energy usage calculations for lignin recovery using DES and various approaches, outlining the potential heat requirements and the processes for extracting and purifying valuable lignin and furanic components.
In Chapter 2, a method is studied in which water precipitation followed by LLX is suggested for extracting lignin and furanics (such as furfural and 5-HMF) from DES composed of choline chloride and lactic acid. The addition of water prior to LLX expands the range of solvents usable in the process compared to DES phases without water addition. A solvent screening study was conducted with six different solvents, including 2-methyltetrahydrofuran (2-MTHF), guaiacol, 2,2,5,5-tetramethyloxolane (TMO), eugenol, m-xylene, and toluene. Among these, 2-MTHF, guaiacol, and TMO were identified as suitable solvents for lignin and furanics recovery. The distribution coefficients of lignin and furanics were compared using these solvents with varying amounts of added water to DES-black liquor in the precipitation step. The results indicated that increasing the amount of water in DES-black liquor improved the precipitation yield of lignin, but beyond a certain ratio, further improvement was not observed. LLX facilitated the extraction of water-soluble lignin, with extraction yields exceeding 45% for smaller lignin fractions and over 70% for larger lignin fractions when water was added to DES-black liquor at a ratio of 1:1 g/g for precipitation followed by a single-stage extraction. The distribution coefficients for smaller lignin fractions were higher in 2-MTHF and guaiacol compared to TMO. While all three solvents were able to extract furfural with distribution coefficients higher than 1, guaiacol exhibited higher distributions of furfural and 5-HMF compared to 2-MTHF and TMO. The addition of water to DES-black liquor during the precipitation step improved the recovery of furanics and smaller lignin fractions in LLX while reducing the leaching of lactic acid and water into the solvents, but also increases the energy costs for regeneration. The optimal amount of water added during the precipitation step is important for industrial applications.
Chapter 3 describes an investigation on the use of continuous LLX equipment for the regeneration of DES-black liquor diluted with water, based on lactic acid-choline chloride, obtained after lignin precipitation. Continuous countercurrent extraction processes were performed using a continuous countercurrent centrifugal separator (CCCS) and a Karr column. Biomass-derived solvents, 2-MTHF, and guaiacol, were employed in single-stage and multistage extractions. CCCS demonstrated efficient single-stage extraction, with extraction yields close to batch yields for both solvents. Applying a two-stage countercurrent processes with two CCCS devices improved furfural and 5-HMF extraction efficiencies, while lignin (< 2500 g/mol) extraction remained consistently around 90%. In the experimental Karr column, which had a diameter of 15.9 mm and an effective height of 601 mm, both solvent systems achieved an extraction yield of approximately 99% for furfural. The Kremser equation shows that guaiacol requires less than two steps to remove 99% of the lignin. Experimental findings, however, indicate that 99% recovery is not achievable because the residual lignin is inherently resistant to non-polar solvents. Compared with extraction by guaiacol, the multi-stage extraction using 2-MTHF enhances the extraction of smaller lignin by increasing the leaching of lactic acid, which makes it more hydrophilic and allows for a 99% yield. 5-HMF showed a high extraction yield of 99% when the ratio of solvent to feed was 1. On the other hand, for guaiacol, extraction yields of around 90% were obtained at solvent-to-feed ratios of 0.5 and 1. Thus for both solvents, sufficient extraction was measured, and process design information has been made available for further process development.
Chapter 4 describes an experimental investigation on the vapor-liquid equilibrium (VLE) and solid-liquid equilibrium (SLE) data associated with the regeneration process of guaiacol and TMO after LLX. These solvents are potential options in the regeneration process of DES composed of lactic acid and choline chloride. The study revealed the formation of a solid solution between guaiacol and lactic acid. The vapor pressures of TMO and guaiacol were measured within a specific temperature range, allowing for simulations to assess effective approaches for solvent recovery in biorefinery processes using these solvents.
Chapter 5 describes an investigation on an ultrafiltration-diafiltration (UF-DF) process to recover and purify lignin dissolved in a DES made of lactic acid and choline chloride after delignification. Experiments were conducted using a stirred dead-end setup to investigate the rejection of lignin by a polyether sulfone ultrafiltration (UF) membrane with a molecular weight cut-off of 5000. Kraft lignin was employed as the model compound representing lignin extracted from a DES-based pulping process. The lignin was dissolved in the deep eutectic solvent (DES), composed of lactic acid and choline chloride, obtained after delignification, was diluted with a 70:30 (v/v) acetone-water mixture. This dilution was chosen because of the high solubility of lignin in this particular mixture. The results revealed that lignin molecules with a molecular weight of 200 g/mol or higher could be effectively rejected, achieving rejection values approaching 0.85 for feed solutions with lignin concentrations ranging from 4.5 to 30 g/L. Furthermore, it was observed that membrane fouling and flux decline were influenced by the concentration of DES. A modeling approach was developed to predict the system size and operating conditions required for conducting the diafiltration process on a large scale. This model incorporated experimental data on lignin fouling at different feed concentrations and hydrodynamic flow characteristics, enabling the optimization of the diafiltration process. The modeling findings indicated that achieving a lignin concentration of 30 g/L, which was considered the maximum solubility in acetone-water (70:30 v/v), requires a pre-concentration step with a membrane area of 13 m² and a recovery ratio of 0.9. Subsequently, 3 diafiltration steps are necessary, each with membrane areas of less than 2 m² and a recovery ratio of 0.85. However, once a fouling layer develops, the pre-concentration stage encounters increased filtration resistance and lower flux, resulting in a decreased recovery ratio (dropping to 0.54, 0.22, and <0.1 for diafiltration steps 1, 2, and 3, respectively) for the same membrane area. This necessitates a larger system size. To refine predictions of system size at an industrial scale, it was crucial to simulate industrial crossflow operations, which became the primary focus of the subsequent chapter.
In Chapter 6 a lab-scale crossflow setup was utilized to comprehensively investigate flux decline, fouling, and lignin rejection under conditions resembling industrial processes. To create feed solutions with varying DES (1.8 to 229 g/L) and lignin concentrations (3.3 to 30 g/L), lignin was isolated from DES-black liquor post-biomass delignification using cold water precipitation. The obtained lignin was then mixed with DES composed of lactic acid and choline chloride and adjusted to the desired concentration. Subsequently, this mixture was diluted with a 70:30 (v:v) acetone-water mixture, chosen for its high lignin solubility. Under varying pressures, feed flowrates, and concentrations of lignin and DES, the lignin rejection consistently averaged around 0.9. Raising the flow rate from 4.36 mL/s to 17.4 mL/s and pressure to a maximum of 10 bar enhances the flux. However, when pressure increases, the compaction of the membrane may influence the pore-blocking and affect the formation of a cake on the membrane's surface. As the concentration of lignin decreases, the flow increases and the filtration resistance decreases. Additionally, it was shown that concentrations of DES at 46 g/L and below have a negligible impact on filtration resistance. Nevertheless, the greatest concentration of DES (229 g/L) leads to a significantly greater change in filtration resistance over time as a result of an increase in pore-blocking. The initial flow increases as the concentration of DES drops, resulting in a lower viscosity of the feed solution. To aid in process development aimed at minimizing fouling, the effects of membrane fouling were predicted using a modified version of a numerical simulation-based method employed in previous chapter (Chapter 5). This involved evaluating filtration flux for various feed concentrations and flow rates and fitting the data to experimental results to determine constant parameters such as maximum thickness factor, cake growth coefficient, and specific cake resistance. These parameters were then utilized to forecast flux decline and filtration resistance in a large-scale UF-D;/F process, including one pre-concentration step and 3 diafiltration modules. The modeling results indicated that over time, the required area of membrane modules increases due to the accumulation of fouling layer and filtration resistance against flux. However, for pre-concentration, a module area of less than 80 m2 appears adequate to sustain system operation for 2 days without requiring backwashing. In contrast, the required area for diafiltration modules increases because the addition of solvent increases the volume entering the system. Notably, the membrane area required for diafiltrations 2 and 3 is smaller than for diafiltration 1, as decreasing DES concentration leads to increased flux.
Chapter 7 focuses on energy usage calculations based on lignin recovery from DES using various solvents and methods involving lignin precipitation prior to LLX and membrane-based DES recovery. four scenarios were evaluated, including lignin precipitation without LLX, lignin precipitation followed by 2-MTHF-based LLX, lignin precipitation followed by guaiacol-LLX, and lignin recovery using membrane-based approach (UF-DF). Lignin precipitation without LLX has shown potential for regenerating DES with reduced lignin content without compromising delignification rates and pulp yields. Using a reasonable water mass ratio (0.42 to 1) resulted in heat usage of 12.5 to 20 GJ/ton pulp. Lignin precipitation followed by 2-MTHF-based LLX enhanced DES purity and facilitated the extraction of smaller lignin and furanics for potential valorization, but it required 2.4 times more heat due to the considered distillation recovery for furfural, presence of solvent and additional water needed for DES recovery. In contrast, lignin precipitation followed by guaiacol-LLX required less heat compared to 2-MTHF, with total heat increasing by 1.5 times without LLX, representing a 40% decrease compared to the 2-MTHF process. For membrane-based approach, excluding downstream furfural recovery costs, an optimized heat duty of 14.3 GJ/ton pulp was achieved through adjustments in water and acetone (acetone fraction of 0.9) quantities. Overall, all DES regeneration strategies, offered potential for the extraction and purification of valuable components such as lignin and furanics. While solvent extraction aids in removing smaller lignin molecules and furanics, it increases costs and complexity due to the energy required for solvent regeneration, with guaiacol resulting in lower energy costs compared to 2-MTHF.
Chapter 8 provides a comprehensive summary of the investigation into the regeneration of Deep Eutectic Solvents (DES) through two distinct approaches: LLX-based and membrane-based methods. Each approach is evaluated based on its benefits, drawbacks, perspectives, and proposed future directions.
