Hansen Solubility Parameters: A Tool for Solvent Selection for

Jul 1, 2019 - The high polysaccharide content of sugar cane bagasse associated with its availability makes it a promising source of fermentable sugars...
1 downloads 0 Views 1MB Size
Article Cite This: Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/IECR

Hansen Solubility Parameters: A Tool for Solvent Selection for Organosolv Delignification Lísias P. Novo*,† and Antonio A. S. Curvelo‡ †

Downloaded via BUFFALO STATE on July 25, 2019 at 12:42:28 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Departamento de Engenharia de Materiais, Escola de Engenharia de São Carlos, Universidade de São Paulo, Avenida João Dagnone 1100,São Carlos, São Paulo 13563-120, Brazil ‡ Instituto de Química de São Carlos, Universidade de São Paulo, Avenida Trabalhador São-carlense, 400-Caixa Postal 780, São Carlos, São Paulo 13560-970, Brasil S Supporting Information *

ABSTRACT: The high polysaccharide content of sugar cane bagasse associated with its availability makes it a promising source of fermentable sugars for the production of secondgeneration ethanol. In this context, the separation the macro components in the cell walls of sugar cane bagasse can be performed by organosolv processes. This research was carried out to study the effect of the solvents on the organosolv delignification of sugar cane bagasse by using Hansen’s solubility parameters. Thus, a new and optimized sphere of solubility for sugar cane bagasse lignin was determined, with δD = 21.42 MPa1/2, δP = 8.57 MPa1/2, δH = 21.80 MPa1/2 and Ro = 13.56. It was observed that this new lignin sphere of solubility allowed a linear correlation between the solvent/ solute distance and the delignification extent data (linear regression coefficient of 0.93856). These results indicate that the Hansen solubility parameters can be used as a tool for choosing the solvent for an effective delignification of lignocellulosic raw materials.

1. INTRODUCTION The sugar cane harvest in Brazil reached an average of 650 × 106 tons over four seasons (2013/2014−2016/2017), leading to a production of 38.7 × 106 tons of sugar and 27.3 × 109 liters of alcohol.1 This high production is a consequence of two main factors: the cost of sugar in the international market and the need for production/use of fuels from renewable sources. Since 1975, when the Brazilian Alcohol Program (Proálcool) was created, gasoline was partially (anhydrous ethanol/ gasoline mixtures) or totally (92−94% ethanol) replaced by ethanol as a fuel in light vehicles.2 As a consequence, Brazil became the only country currently deploying the large-scale use of ethanol as an alternative to petroleum-based fuels3,4 and has now become the world’s second largest producer of this chemical.5 The main byproduct of the sugar and alcohol industry is sugar cane bagasse, corresponding in the same four seasons to 162.5 × 106 tons/year (for each ton of sugar cane processed, about 250 kg of bagasse are produced, with humidity of about 50%).3 Most sugar cane bagasse is employed as a solid fuel in boilers to produce heat and steam for the process. In the large mills, part of the heat and steam is transformed into electricity for sale to the public grid.3,6,7 Besides the utilization as solid fuel, sugar cane bagasse can be used as feedstock in the context of a biorefinery, capable of producing second-generation ethanol (by fermentation of © XXXX American Chemical Society

sugars released from cellulose and hemicelluloses hydrolysis) or other chemicals with higher added value.5,8 Considering the second-generation ethanol production as the primary goal, the separation of the different fractions of the macromolecules present in the biomass is always necessary.9 In this context, we highlight the delignification processes that consist of breaking ether linkages in lignin to produce smaller fragments able to be dissolved in solvent media.6,10 Among the delignification processes, the organosolv process is particularly interesting because of the possibility of recovery and recycling the solvents. In organosolv processes it is still possible to add catalyst (acidic or basic) and apply high pressures (subcritical and/or supercritical fluid conditions) to optimize the reaction conditions.9,11 Therefore, for organosolv processes, the dissolution of the lignin fragments plays an essential role to the delignification efficiency. In this context, Solubility Parameters can provide valuable information on the mechanisms underlining the solubility of lignin, thus enabling the rational choice of the best solvent composition for the extraction of lignin. Received: Revised: Accepted: Published: A

February 14, 2019 May 22, 2019 July 1, 2019 July 1, 2019 DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research The solubility theories are based on the condition that, for the occurrence of solubility of two liquids, it is necessary that the intermolecular interactions between the molecules of the one component (A−A) and the molecules of the second component (B−B) are of the same order of magnitude and, therefore, can be broken to form A−B interactions.12 Hildebrand established the first theory of solubility parameters and described it as being the square root of the cohesive energy density (CED), as shown in eq 1.13 δi =

CED =

ΔUi Vi̅

For practical applications, when RED is less than 1, the affinity between the solvent and the polymer is high. If RED is greater than 1, the affinity between the solvent and the polymer is low and as it grows, the affinity between the two decreases progressively. When RED is equal to 0, there is no difference between the solvent and polymer interaction energies. Therefore, the affinity between the solvent and polymer reaches its maximum. The boundary condition of polymer dissolution occurs when the value is equal or close to 1.12,14−16 The Hansen solubility parameters (HSP) for lignin as determined by Hansen and Björkman15 are δD= 21.9 MPa1/2, δP = 14.1 MPa1/2, δD), δH = 16.9 MPa1/2, and R0 = 13.7. These values were confirmed by Vebber et al.17 in a determination performed applying genetic algorithms. The application of solubility parameters in the study of dissolution of the lignins begun with the seminal paper by Conrad Schuerch published in 1952.18 In that work, the solubility of (previously) extracted lignins was determined in several solvents at room temperature and correlated with the solubility parameter of the solvents. From the obtained results Schuerch proposed a Hildebrand solubility parameter for lignin as (around) 22.5 MPa1/2 and that the solubility increases with the hydrogen bonding capability of the solvent.18 The Hildebrand solubility parameters were also determined for lignins extracted from different processes and raw materials. Ni et al. reported a value of 28.1 MPa1/2 for hardwood Alcell,19 whereas Wang et al.20 calculated a value of 28.7 MPa1/2 for enzymatic hydrolysis/mild acidolysis lignin from sugar cane bagasse. The fractionation of isolated lignins was also studied by applying the solubility parameters approach aiming at the production of more homogeneous lignin samples. In this context, Rapponen et al.21 reported the fractionation of hardand softwood lignins isolated by means of organosolv and Kraft processes. Li et al.22 described the fractionation of Bamboo organosolv lignin in a work that supports the claim by Schuerch18 concerning the role of the solvent in the dissolution of lignins. Boeriu et al.23 and Duval et al.24 also applied the solubility parameters concept to the study of lignin fractionation, by either using the Hildebrand approach as described by Boeriu et al.23 or following the Hansen threedimensional theory as reported by Duval et al.24 A different utilization of the solubility parameters concepts can be found in the study of delignification of lignocellulosic substrates. Thus, Balogh et al.25 studied the organosolv delignification of Pinus caribaea hondurensis with nine different organic solvent/water mixtures. The obtained relationship between the delignification extent and Hildebrand solubility parameter exhibited a “Bell shape” pattern centered at δ = 23 MPa1/2. Other study assessed the effect of water content on the organosolv solvent mixture was studied in the delignification of hydrolyzed almond shells by Quesada-Medina et al.26 The highest extraction yield was obtained with 75% organic solvent (ethanol, dioxane or acetone) and the corresponding solubility parameter for the hydrolyzed lignin from almond shells was equal to 30 MPa1/2. Ye et al.27 also applied the Hildebrand solubility parameter to explain the effect of the solvent in the delignification of enzymatically hydrolyzed cornstalks. Employing ethanol, dioxane and tetrahydrofuran in mixtures with water, they found delignification extent in the range from 34.5 to 53.7% and δ = 28 MPa1/2 for cornstalk lignin. Recently, Cheng et al.28 described the utilization of HSP to correlate delignification of poplar and rice straw with the

(1)

Where ΔUi and V̅ i are the internal energy and the molar volume of the pure “i” component, respectively. The same way, for polymer dissolution, it is necessary that the solubility parameters of the solvent and solute are close to each other. However, this solubility parameter does not discriminate the kind of intermolecular interactions and in some cases predicted incorrect results. To improve the Hildebrand’s methodology, Hansen proposed a new treatment based on the three different types of intermolecular interactions: dispersive interactions (non polar interactions, δD), polar interactions (between permanent dipoles, δP), and hydrogen bonds (δH). Therefore, δ can be described as the square root of the sum of the quadratic of the three different kinds of interactions.12,14 Thus, for Hansen, a solvent could be represented as been a point in a threedimensional space, with coordinates δD, δP and δH. Similarly, a solvent mixture can also be represented as a point in threem dimensional space, but in this case with coordinates δm D , δP , and δm , as expressed in eq 2: H ij ∑ ϕδ D, i ∑ ϕδ P, i ∑ ϕδ H, i yz zz (δ Dm , δ Pm , δ Hm) ≡ jjjj i i , i i , i i zz ∑ ∑ ∑ ϕ ϕ ϕ i i i i i i { k

(2)

Where ϕi is the volume fraction of the solvent in the mixture. As consequence, if two solvents do not solubilize a solute and these two solvents have their coordinates in opposite sides (from the solubility parameter) of the solute, it is possible to produce a mixture of these two solvents that will solubilize the solute12,14,15 Likewise, a polymer can be represented, not as a point, but as a volume in space, and for the polymer dissolution the point that represents the solubility parameter of the solvent must be inserted within the volume corresponding to the parameter solubility of the polymer.12,15 Hansen states that the volume corresponding to the solubility parameter of a polymer can be described having a pol pol radius R0 from the center with δpol D , δP and δH . To establish a parameter that correlates the interaction between a polymer and a solvent (or solvent mixture), Hansen proposed a parameter called relative energy difference (RED), that is correlated with the affinity solvent/solute, and defined as the ratio between the radius of interaction (Ra), eq 3, and the experimental sphere radius for the polymer (R0), as shown in eq 4.12,14−16 Ra =

4(δDpol − δDsolv)2 + (δPpol − δPsolv)2 + (δ Hpol − δ Hsolv)2 (3)

RED =

Ra R0

(4) B

DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

hydroxide solution to remove adsorbed lignin and to liberate the cellulosic fibers.10 This procedure was previously determined as been indistinguishable to the extraction of the lignin fragments adsorbed on the fibers of bagasse with a solution of water and the organic solvent used in delignification. After this procedure, the pulps were filtered under reduced pressure and then washed with water until a neutral pH. The pulps were dried at room temperature to moisture below 10%. 2.3. Determination of Pulp Yield and Residual Lignin Contents. the delignification experiments were characterized by the yield of the reactions by a simple gravimetric method (eq 5). Furthermore, by characterizing the lignin content in the different pulps and liquors and comparing it with the lignin content in the raw bagasse, it was possible to obtain the delignification extents.6

solubility parameter of the solvent used in organosolv delignifications. In that study the value of HSP for lignin was obtained from isolated lignin (milled wood lignin) as reported in the literature.12 The utilization of solubility parameter in the context of lignocellulosic biomass deconstruction and cellulose dissolution was recently reviewed Rinaldi and co-workers.29−33 Considering that the most of these delignification studies applied the Hildebrand approach and values of HSP for lignin from previously isolated samples, this study aimed to apply the Hansen solubility parameter theory to the delignification extent (or delignification yield) in organosolv treatment of sugar cane bagasse. In this contex, the delignification extent and the corresponding solubility parameter of the solvent/ water mixtures were employed to determine the Hansen solubility parameters and the solubility sphere radius specific for sugar cane bagasse lignin in the reaction conditions used. In this manner, the analysis of the solubility parameter of the solvent with the experimental solubility sphere for sugar cane bagasse lignin can provide a guide to select the organic solvent−water mixture for an effective organosolv delignification.

delig. yield =

mdelignifed material minitial sample

(5)

The lignin content of the pulp and raw bagasse samples were obtained by a modified Klason method. First, the solid samples were ground in a knife mill “Solab” SL31, with a 40-mesh screen. Then, the grounded samples were submitted to hydrolysis with a 72% sulfuric acid solution at 25 °C for 2 h, with constant stirring. After that, the solutions were diluted to a final concentration of 3% of sulfuric acid, and heated in an autoclave at 120 °C for 1 h. After the reaction, the solid and liquid phases were separated using filtered crucible ASTM 10− 15 previously dried and weighed. The solid phase corresponds to the acid-insoluble lignin while the solution contains the sugars and other degradation products from the hydrolysis of polysaccharides and a small portion of acid-soluble lignin. Furthermore, the acid-soluble lignin was determined by ultraviolet absorbance at 215 and 280 nm.34 The delignification extent was determined using eq 6. ÅÄÅ bagasse ÑÉÑ pulp pulp Y ÅÅ m Ñ ÅÅ lignin − mlignin 100 ÑÑÑ ÅÅ ÑÑ delig.extent(%) = ÅÅ ÑÑ100 bagasse ÅÅ ÑÑ mlignan ÅÅ ÑÑ ÅÅÇ ÑÑÖ (6)

2. EXPERIMENTAL SECTION 2.1. Sugar Cane Bagasse Preparation. The industrial sugar cane bagasse from Iacanga Sugar and Alcohol Mill (Iacanga − São Paulo state) was submitted to a washing process at 70 °C for 1 h under constant stirring to remove residual sugars. It was then dried under open air until 10% humidity and stored. 2.2. Organosolv Delignification. The organosolv delignification reactions were carried out using a constant reaction condition for the different solvent systems: temperature of 125 °C, 60 min of reaction time, hydrochloric acid concentration of 0.05 mol L−1 (hydrochloric acid 37% purchased from Quemis), solvent/0.05 mol L−1 HCl aqueous solution ratio of 9/1 and solid/liquid ratio of 1/10 (10g of bagasse). Therefore, the only variable was the solvent system used to dissolve the lignin fragments. The following solvents were used for the organosolv delignification: 1,4-dioxane (Tedia), 1-butanol (Mallinckrodt), 2-butanol (Oxiteno), acetone (Quemis), acetonitrile (J. T. Baker), cyclohexane (Quemis), ethanol (J. T. Baker), glycerol (Quemis), isopropanol (Synth), methanol (Quemis), methyl ethyl ketone (Oxiteno), tetrahydrofuran (J. T. Baker), and toluene (Quemis). An experiment was also carried out using only water as a solvent. HSP for the solvents and solvent mixtures containing or not containing water are shown in Table S1. The use of aqueous solution of HCl in the solvent mixtures promotes the catalysis of ether bond cleavage in the lignin macromolecules and guarantees similar removal of hemicelluloses in all organosolv treatments. The delignification experiments were performed in 195 cm3 stainless steel reactors (5.0 OD × 4.0 ID × 15.5 cm length) and cap screws. It was equipped with a self-sealed Teflon Oring closure. The heating was provided by a thermostatic bath containing glycerol as the heating fluid and stirred mechanically. The reactors were submerged into the glycerol bath at the desired reaction temperature and after the reaction time the reactors were quenched in an ice bath to stop the reaction.6 The bagasse pulps were separated from the pulping liquors by filtration under reduced pressure. The pulps were submitted to a defibration process with a 1% (by weight) sodium

(

)

2.4. Lignin Solubility Sphere. Hansen and Björkman15 reported the HSP for extracted wood lignin as the sphere with δD, δP, δH, and Ro values equal to 21.9, 14.1, 16.9, and 13.7, respectively. However, these values cannot be assumed universally for all lignin samples and organosolv delignification processes. The great chemical heterogeneity between lignins from different sources in addition to the method of extraction of the lignin (that fragments it differently) should provide other values for the solubility sphere of sugar cane lignin. Thus, the organosolv delignification data was used to obtain HSP for organosolv sugar cane bagasse lignin. For that, a program developed by Gharagheizi29 for MatLab was used. This program utilizes the same algorithm to calculate the solubility parameters that is used in the program developed by Hansen;12 however, by applying modifications, the algorithm results in a more accurate response.35 In these algorithms, it is necessary to input a database, containing the number of solvents used, the solubility parameters for this set of solvents, and the solubility of the substrate in the solvent, for the calculation of the HSP. The solubility of the substrate in each solvent is expressed as a value C

DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research of 1 or 0 for the miscibility and immiscibility of the solute in the solvent, respectively.12,35 However, the solubility concept can be often stated as different phenomena. For example, solubility can be determined as (i) the formation of a unique phase; (ii) a given percentage of swelling; (iii) the absorption of the solvent by a polymer; (iv) the rupture of the substrate in a certain time; (v) by the suspension of a pigment for a long period.12,35 In this work, it was assumed that the solubility of the sugar cane lignin in each solvent mixture is directly related to the delignification extent determined in the organosolv treatment with this solvent mixture. The limits for the choice of solubility (1) or insolubility (0) values were tested from (higher than) 50% to 70%.

3. RESULTS AND DISCUSSION Table S2 shows the yields for the organosolv delignification experiments with the corresponding residual lignin content of the pulps (insoluble and soluble lignins) and the delignification extents. By the same characterization methods, raw sugar cane was characterized as having on a weight dry basis (wt %) 21.14 ± 0.39% of acid insoluble lignin and 1.08 ± 0.17% of acid soluble lignin, resulting in a total of 22.22 ± 0.39% of lignin. These values are in good agreement with that one reported by Vallejos et al.,7 that employed the standardized TAPPI procedure of acid insoluble lignin analysis for sugar cane bagasse (21.7 wt % of lignin, been 19.5 wt % of acid insoluble lignin and 2.2 wt % of acid soluble lignin). Sugar cane bagasse chemical composition from different locations also presented good similarity to the lignin contents obtained in this study.36,37 Additionally, these results are consistent with the selective removal of hemicelluloses and dissolution of lowmolecular-weight lignin fragments. Balogh et al.,25 with similar delignification conditions (125 °C, 360 min, 0.2 mol L−1 of hydrochloric acid, solvent/water ratio of 9/1 and solid/liquid ratio of 1/8), achieved smaller delignification extents for sawdust Pinus caribaea hondurensis than in this study, despite the use of a higher acid concentration and longer extraction time. This comparison indicates that removing lignin from sugar cane bagasse is easier than from Pinus, because the bagasse is less lignified and absorbs liquids easily,38 allowing a more efficient impregnation of the pulping liquor and promoting a more significant delignification, as already described in the literature.39 The direct relationship between the delignification extent data (Table S2) and the correspondent Hildebrand solubility parameters did not result in a good fit, as shown in Figure 1. The position of the data for the delignifications performed with 2-butanol, acetonitrile, 1-butanol, and isopropanol shows a deficiency of the Hildebrand solubility parameter model. These four solvents have practically the same value of solubility parameter. However, they yield delignification extents with more than 20 percentage points of difference. On the other hand, isopropanol, ethanol and methanol present different solubility parameter and render almost the same delignification extent. The Hildebrand solubility parameter cannot differentiate the several types of intermolecular interactions, which result in similar results for solvents that interact differently. In an attempt to get a better correlation between delignification extent data and solubility parameters, it was applied the HSP with lignin parameters (δD, δP, δH, and Ro) determined by Hansen and Björkman.15 In this approach, it is assumed that the Relative Energy Difference (Ro) should

Figure 1. Delignification extent of organosolv treatments of sugar cane bagasse versus Hildebrand solubility parameter of the solvent mixtures.

provide a linear relationship with the delignification extents, as both are expressions of lignin solubility. In other words, lower values of RED (especially for RED < 1) for a solvent/lignin system should provide greater delignification extent. Presented in Figure 2 is the plot of delignification extent data with the RED values in relation to Hansen and Björkman lignin HSP,15 determined for a solvent mixtures containing 10% of water (v:v).

Figure 2. Relative energy difference (RED) from Hansen and Björkman lignin solubility sphere versus delignification extents of organosolv treatments of sugar cane bagasse.

Figure 2 clearly shows that all the RED values for the solvent mixtures fall off the lignin solubility sphere, since they are greater than 1. Moreover, the adjustment of the points to the delignification extent data is not entirely satisfactory, since there is a low R2 value. This outcome contradicts the high delignification extents (higher than 70%) obtained for several of the solvent mixtures. This problem appears to be related to the inadequacy of the reported HSP values15 for the sugar cane bagasse lignin. Therefore, new solubility parameters suitable for sugar cane lignin should be determined. For that, the Gharagheizi program35 was used, considering different minimum delignification extents as indicative for the solubility D

DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Table 1. Delignification Extent Minimum, Generated Sphere of Solubility, and Linear Regression Coefficient of Determination Data

a

Lit.: Hansen and Björkman15 lignin solubility sphere.

of sugar cane lignin. Considering the range of delignification obtained in the organosolv treatment of sugar cane bagasse, five delignification values (50, 60, 70, 72, and 75%) were chosen as the minimum to characterize the solubility of sugar cane lignin. The solubility parameters (δD, δP, δH, and Ro) and the correspondent linear regression coefficient (R2) for the five sets are shown in Table 1. Excepting the first set, all the remaining sets showed a similar or best fit when comparing with the reported HSP values for lignin. The best linear fit corresponds to a delignification extent minimum of 70% (3rd set). Despite progress, the coefficient of determination of 0.84662 is still lower than the expected for a good linear regression. The values of δP, δH, and Ro showed important changes when compared to reported values for HSP, with a decrease in 6 units for δP and increase of 5.6 units for δH. The value of Ro for this third set also increases (c.a. 2.3 units). The plot of Relative Energy Difference (RED) from the data obtained in the third set versus delignification extents is showed in Figure 3. Besides the improvement in the correlation between RED and delignification extents, the five best solvent mixtures (delignification extent >70%) presented RED lower than 1. The RED values for the remaining solvents are well dispersed in the entire range of RED. However, from the plot of the delignification extent, considering the third set of solubility sphere showed in Figure 3, the best solvents (1,4-dioxane and 1-butanol) did not have the lowest values of RED and that 1,4dioxane is slightly far from the fitted regression line (although in 95% confidence limit). The excellent performance of 1,4-dioxane/water mixtures to extract lignin has been known since the early works of Björkman40 and Pepper et al.41 In particular, the delignification with acidic 1,4-dioxane/water mixtures employed in the present work is very similar to the acidolysis methodology proposed by Pepper et al.41 The application of solubility parameters in the delignification processes should also consider

Figure 3. Relative energy difference (RED) from 3rd set data versus delignification extents of organosolv treatments of sugar cane bagasse.

the role of the solvent mixture in the hydrolysis of ether bonds in “protolignin” structure.25 The higher efficiency in generating low molar mass lignins fragments will increase the delignification/extraction yield in solvent mixtures with HSP close to that of lignins. Thus, the deviation observed in Figure 3 for 1,4-dioxane/water mixture could be a consequence of a synergistic association of the efficiency to produce lignins fragments with the quality of the solvent to dissolve them. To obtain HSP values for sugar cane bagasse lignin from the best correlation between RED and delignification extents the linear fitting was recalculated without the contribution of the 1,4-dioxane/water and the results are displayed in Figure 4. This removal does not indicate that there was an error in the delignification reaction but that the lignin extraction with 1,4dioxane/water is improved by the efficiency of lignin fragmentation. Moreover, to obtain the lignin solubility E

DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

butanol, and water were made, considering the theoretical RED values (in relation to the unoptimized solubility sphere), keeping the volume percentage of water equal to 10%. This new set of organic solvent mixtures is an application of HSP to the pretreatment of sugar cane bagasse. At the same time, it is still possible to improve the fit of the data by adding the results of the delignifications performed with these solvents mixtures, as they provide solubility parameters in other regions without previous points. With that, it is possible to optimize the position of the center and radius of the solubility sphere. This new set of organic solvent mixtures consists of three ternary and one quaternary solutions. Always with 10% of water, the ratio of organic solvents (by volume) in the ternary mixtures was glycerol:ethanol (1:4); ethanol:2-butanol (2:3) and glycerol:2-butanol (3:7). The ratio of organic solvents and water in the quaternary solutions was glycerol:ethanol:2butanol:water (3:3:3:1). All the solutions exhibited RED values lower than 1.0. The delignification extent results of the reactions performed with this new set of solvent mixtures are shown in Supplementary Table 3. In line with the RED values, the organosolv treatment with the new set of solvent mixtures yielded delignification extents in the range from 76 to 85%. These results confirm the feasibility to use renewable solvents in the pretreatment of lignocellulosic raw materials. The inclusion of these results to recalculate the HSP values for sugar cane bagasse lignin is displayed in Table 2. The optimization resulted in a small change in the Hansen solubility parameters values (0.5−1.2 MPa1/2). However, the radius of the solubility sphere got smaller, decreasing 2.5 MPa1/2. This decrease is a good indication that the optimization was effective, because Hansen described a polymer solubility parameter as the center of the smaller sphere that contains the solvents were this polymer is soluble.12,15 When comparing with the results obtained from set 3, the inclusion of the new set of solvent mixtures improves the correlation of RED vs delignification extents (Figure 5), resulting in a better linear regression coefficient (0.93856). Excluding 1,4-dioxane, there is a low dispersion of data along RED and delignification extent axes (see Figure 5). Observing the Hansen solubility parameters for 1,4-dioxane,42 it can be seen that this solvent has a high capacity to perform dispersive interactions, while having an intermediate ability to interact via

Figure 4. Relative energy difference (RED) from 3rd set data versus delignification extents of organosolv treatments of sugar cane bagasse (without 1,4-dioxane).

parameter it is needed to keep the delignification extent of 1,4dioxane in the database as a soluble system. Figure 4 presents that the exclusion of 1,4-dioxane resulted in a significant increase in the fitting of the linear regression, resulting in R2 value of 0.90546. This value is acceptable, considering that variables such as temperature dependence of the solubility parameters were not considered in the model. Despite this approximation, the HSP values proposed in this work for sugar cane bagasse lignin should be applied to the choice of organic solvent for delignification processes, while the values obtained directly from lignins should be considered in applications for isolated lignin samples. The industrial use of organosolv processes in the pretreatment of sugar cane bagasse will only be possible with the use of cheap and abundant solvents, preferably the green ones and obtained from renewable resources. Among the solvents that fit these requirements it was decided to perform the delignification of sugar cane bagasse with glycerol, in mixtures with ethanol and 2-butanol, always with 10% of water (by volume). Glycerol is a byproduct of biodiesel industry, whereas ethanol and 2-butanol are produced from fermentation of sugars. Tertiary and quaternary mixtures of glycerol, ethanol, 2-

Table 2. Optimized Solubility Sphere and Linear Regression Coefficient of Determination Data

a

Without the inclusion of 1,4-dioxane. bLit.: Hansen and Björkman15 lignin solubility sphere. F

DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

for the sugar cane bagasse lignin extracted under the specific conditions. The optimized solubility sphere allowed a better data fit with a linear regression determination coefficient of 0.93856, excluding the data from 1,4-dioxane. A specific set of values of HSP for sugar cane bagasse lignin was determined as δD = 21.42 MPa1/2, δP = 8.57 MPa1/2, δH = 21.80 MPa1/2, and Ro = 13.56. Finally, it was found that because of the difficulties in checking the lignin solubility parameter (it does not exist in native form as an isolated structure), the solubility parameter for the actual lignin cannot be stated with certainty, despite the good fit of the data. However, under the conditions studied, this sphere of solubility allows a better prediction of a delignification extent and the selection of a solvent, considering the desired delignification.



Figure 5. Relative energy difference (RED) for optimized lignin solubility sphere versus delignification extents of organosolv treatments of sugar cane bagasse (linear regression without 1,4-dioxane).

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.iecr.9b00875.

hydrogen bonds. On the other hand, the other organic solvent mixtures that lead to high delignification extents are mainly formed by alcohols with high capacity for hydrogen bonds associated with intermediate values for polar interactions.23 The mixture with 10% water affects the final properties of the organic solvent:water (9:1) solutions with particular increase of δP and δH for dioxane:water solutions and minor alterations for alcohol:water solutions. The optimized HSP values for sugar cane bagasse lignin obtained in this work differs from that one published by Hansen and Bjö rkman.15 Although both determinations present similar results for dispersive interactions (δD) and the sphere radius of solubility (R0), the polar and hydrogen bond interactions are different in each determination. Even though not completely understood, these differences could be attributed to intrinsic characteristics of sugar cane bagasse lignin. Lignins from sugar cane (as well as lignins from grasses) are classified as HGS lignin and contain higher amount of phydroxy phenyl moieties, including p-coumarates and ferulates.43 Thus, it is important to assume specific values for HSP in function of the source of lignins.



Tables S1−S3 (PDF)

AUTHOR INFORMATION

Corresponding Author

*Email: [email protected]. ORCID

Lísias P. Novo: 0000-0003-1747-1733 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors are thankful to the Institute of Chemistry of São Carlos and School of Engineering of São Carlos from the University of São Paulo that provided the structure to develop the research. The authors are also grateful to Fundaçaõ de Amparo à Pesquisa do Estado de São Paulo (FAPESP - Process 2016/07370-6) for the scholarship and to Conselho Nacional de Pesquisa e Desenvolvimento (CNPq -Process 305310/ 2015-1) and Coordenaçaõ de Aperfeicoamento de Pessoal de ́ Superior (CAPES) for funding this research. Nivel

4. CONCLUSIONS By adjusting the delignification data to the Hildebrand solubility parameter of the different solvents, it was concluded that this parameter is not the ideal parameter to relate solubility with delignification extents. The Hildebrand parameter does not differentiate the type of interactions that each solvent can perform with lignin. Thus, the Hansen parameters were found as a better way to describe the relationship between the organosolv delignification under the conditions used and the relative energy difference of the center of the lignin solubility sphere to the solvent solubility parameter. It was also concluded that the Hansen solubility parameters verified in the literature for lignin do not allow a linear fit of the delignification data. Therefore, a new lignin solubility sphere was calculated from delignification yields. With an unoptimized sphere, four solvent mixtures of green and renewable industrial solvents were applied, indicating the possibility of the use of HSP for predicting/choosing solvents for the organossolv delignification of lignocellulosic raw materials. Moreover, these mixtures could also be used to optimize the solubility sphere



REFERENCES

(1) UNICA Brazilian Sugarcane Industry Association. Unicadata. ́ (2) Bastos, V. D. Etanol, Alcoolquimica e Biorrefinarias. BNDES Setorial 2007, 25, 5−38. (3) Zanin, G. M.; Santana, C. C.; Bon, E. P. S.; Giordano, R. C. L.; De Moraes, F. F.; Andrietta, S. R.; De Carvalho Neto, C. C.; Macedo, I. C.; Fo, D. L.; Ramos, L. P. Brazilian Bioethanol Program. Appl. Biochem. Biotechnol. 2000, 84−86 (1−9), 1147−1162. (4) Gurgel, L. V. A.; Marabezi, K.; Zanbom, M. D.; da Curvelo, A. A. S. Dilute Acid Hydrolysis of Sugar Cane Bagasse at High Temperatures: A Kinetic Study of Cellulose Saccharification and Glucose Decomposition. Part I: Sulfuric Acid as the Catalyst. Ind. Eng. Chem. Res. 2012, 51 (3), 1173−1185. (5) Cherubini, F. The Biorefinery Concept: Using Biomass Instead of Oil for Producing Energy and Chemicals. Energy Convers. Manage. 2010, 51 (7), 1412−1421. (6) Novo, L. P. L. P.; Gurgel, L. V. A. L. V. A.; Marabezi, K.; Curvelo, A. A. D. S. Delignification of Sugarcane Bagasse Using Glycerol-Water Mixtures to Produce Pulps for Saccharification. Bioresour. Technol. 2011, 102 (21), 10040−10046. G

DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research (7) Vallejos, M. E.; Zambon, M. D.; Area, M. C.; da Silva Curvelo, A. A. Low Liquid−Solid Ratio (LSR) Hot Water Pretreatment of Sugarcane Bagasse. Green Chem. 2012, 14 (7), 1982. (8) FitzPatrick, M.; Champagne, P.; Cunningham, M. F.; Whitney, R. A. A Biorefinery Processing Perspective: Treatment of Lignocellulosic Materials for the Production of Value-Added Products. Bioresour. Technol. 2010, 101 (23), 8915−8922. (9) Huijgen, W. J. J.; Reith, J. H.; Den Uil, H. Pretreatment and Fractionation of Wheat Straw by an Acetone-Based Organosolv Process. Ind. Eng. Chem. Res. 2010, 49 (20), 10132−10140. (10) Pasquini, D.; Pimenta, M. T. B.; Ferreira, L. H.; Curvelo, A. A. S. Sugar Cane Bagasse Pulping Using Supercritical CO2 Associated with Co-Solvent 1-Butanol/Water. J. Supercrit. Fluids 2005, 34 (2), 125−131. (11) Ho, C. L.; Wu, K. T.; Wang, E. I. C.; Su, Y. C. Delignification Kinetics Study on the Tetrahydrofurfuryl Alcohol/HCl Pulping of Rice Straw. Ind. Eng. Chem. Res. 2011, 50 (7), 3849−3857. (12) Hansen, C. M.; Durkee, J.; Kontogeorgis, G.; Panayiotou, C.; Willliams, L.; Poulsen, T.; Priebe, H.; Redelius, P. Hansen Solubility Parameters, 2nd ed.; Hansen, C. M., Ed.; CRC Press: Boca Raton, FL, 2007. (13) Utracki, L.; Simha, R Statistical Thermodynamics Predictions of the Solubility Parameter. Polym. Int. 2004, 286 (2003), 279−286. (14) Yu, H.; Hu, J.; Chang, J. Selective Separation of Wood Components Based on Hansen ’ s Theory of Solubility. Ind. Eng. Chem. Res. 2011, 50, 7513−7519. (15) Hansen, C. M.; Björkman, a. The Ultrastructure of Wood from a Solubility Parameter Point of View. Holzforschung 1998, 52, 335− 344. (16) Hansen, C. M. The Three Dimensional Solubility Parameter − Key to Paint Component Affinities: I. Solvents, Plasticizers, Polymers, and Resins. J. Paint Technol. 1967, 39 (505), 104−117. (17) Vebber, G. C.; Pranke, P.; Pereira, C. N. Calculating Hansen Solubility Parameters of Polymers with Genetic Algorithms. J. Appl. Polym. Sci. 2014, 131 (1), 1−12. (18) Schuerch, C. J. The Solvent Properties of Liquids and Their Relation to the Solubility, Swelling, Isolation and Fractionation of Lignin. J. Am. Chem. Soc. 1952, 74 (20), 5061−5067. (19) Ni, Y.; Hu, Q. Alcell® Lignin Solubility in Ethanol−Water Mixtures. J. Appl. Polym. Sci. 1995, 57 (12), 1441−1446. (20) Wang, Q.; Chen, K.; Li, J.; Yang, G.; Liu, S.; Xu, J. The Solubility of Lignin from Bagasse in a 1,4-Butanediol/Water System. BioResources 2011, 6 (3), 3034−3043. (21) Ropponen, J.; Räsänen, L.; Rovio, S.; Ohra-aho, T.; Liitiä, T.; Mikkonen, H.; van de Pas, D.; Tamminen, T. Solvent Extraction as a Means of Preparing Homogeneous Lignin Fractions. Holzforschung 2011, 65 (4), 543−549. (22) Li, M. F.; Sun, S. N.; Xu, F.; Sun, R. C. Sequential Solvent Fractionation of Heterogeneous Bamboo Organosolv Lignin for Value-Added Application. Sep. Purif. Technol. 2012, 101, 18−25. (23) Boeriu, C. G.; Fiţigǎu, F. I.; Gosselink, R. J. A.; Frissen, A. E.; Stoutjesdijk, J.; Peter, F. Fractionation of Five Technical Lignins by Selective Extraction in Green Solvents and Characterisation of Isolated Fractions. Ind. Crops Prod. 2014, 62, 481−490. (24) Duval, A.; Vilaplana, F.; Crestini, C.; Lawoko, M. Solvent Screening for the Fractionation of Industrial Kraft Lignin. Holzforschung 2016, 70 (1), 11−20. (25) Balogh, D. T.; Curvelo, A. A. S.; De Groote, R. A. M. C. Solvent Effects on Organosolv Lignin from Pinus Caribaea Hondurensis. Holzforschung 1992, 46 (4), 343−348. (26) Quesada-Medina, J.; López-Cremades, F. J.; Olivares-Carrillo, P. Organosolv Extraction of Lignin from Hydrolyzed Almond Shells and Application of the δ-Value Theory. Bioresour. Technol. 2010, 101 (21), 8252−8260. (27) Ye, Y. Y.; Liu, Y. Q.; Chang, J. Application of Solubility Parameter Theory to Organosolv Extraction of Lignin from Enzymatically Hydrolyzed Cornstalks. BioResources 2014, 9 (2), 3417−3427.

(28) Cheng, F.; Ouyang, T.; Sun, J.; Jiang, T.; Luo, J. Using Solubility Parameter Analysis to Understand Delignification of Poplar and Rice Straw with Catalyzed Organosolv Fractionation Processes. BioResources 2019, 14 (1), 486−499. (29) Rinaldi, R.; Jastrzebski, R.; Clough, M. T.; Ralph, J.; Kennema, M.; Bruijnincx, P. C. A.; Weckhuysen, B. M. Paving the Way for Lignin Valorisation: Recent Advances in Bioengineering, Biorefining and Catalysis. Angew. Chem., Int. Ed. 2016, 55 (29), 8164−8215. (30) de Oliveira, H. F. N.; Rinaldi, R. Understanding Cellulose Dissolution: Energetics of Interactions of Ionic Liquids and Cellobiose Revealed by Solution Microcalorimetry. ChemSusChem 2015, 8 (9), 1577−1584. (31) Rinaldi, R.; Reece, J. Chapter 13. Solution-Based Deconstruction of (Ligno)-Cellulose. In Catalysis for the Conversion of Biomass and its Derivatives; Behrens, M., Datye, A. K., Eds.; Max Planck Research Library for the History and Development of Knowledge, 2013; pp 435−462. (32) Rinaldi, R. Chapter 5. Early-Stage Conversion of Lignin over Hydrogenation Catalysts. In RSC Energy and Environment Series; 2018; Vol. 2018-Janua, pp 108−127. (33) Rinaldi, R. Chapter 4. Solvents and Solvent Effects in Biomass Conversion; 2014; pp 74−98. (34) Goldschimid, O. Ultraviolet Spectra. In Lignins: Occurrence, formation, structure and reactions; Sarkanen, K. V., Ludwig, C. H., Eds.; John Wiley & Sons, Inc.: New York, 1971; pp 241−266. (35) Gharagheizi, F. New Procedure to Calculate the Hansen Solubility Parameters of Polymers. J. Appl. Polym. Sci. 2007, 103 (1), 31−36. (36) da Silva, C. G.; Grelier, S.; Pichavant, F.; Frollini, E.; Castellan, A. Adding Value to Lignins Isolated from Sugarcane Bagasse and Miscanthus. Ind. Crops Prod. 2013, 42 (1), 87−95. (37) Zhao, X.; Wu, R.; Liu, D. Production of Pulp, Ethanol and Lignin from Sugarcane Bagasse by Alkali-Peracetic Acid Delignification. Biomass Bioenergy 2011, 35 (7), 2874−2882. (38) Nada, A.; Sefain, M. Beatbility and Water-Retention Relations of Wood and Bagasse Pulps. Cell. Chem. Technol. 1986, 20 (4), 409− 415. (39) Pasquini, D.; Pimenta, M. T. B.; Ferreira, L. H.; Curvelo, A. A. D. S. Extraction of Lignin from Sugar Cane Bagasse and Pinus Taeda Wood Chips Using Ethanol−Water Mixtures and Carbon Dioxide at High Pressures. J. Supercrit. Fluids 2005, 36 (1), 31−39. (40) Björkman, A. Studies on Finely Divided Wood. Part 1. Extraction of Lignin with Neutral Solvents. Sven. Papperstidn 1956, 59, 477−485. (41) Pepper, J. M.; Baylis, P. E. T.; Adler, E. THE ISOLATION AND PROPERTIES OF LIGNINS OBTAINED BY THE ACIDOLYSIS OF SPRUCE AND ASPEN WOODS IN DIOXANE− WATER MEDIUM. Can. J. Chem. 1959, 37 (8), 1241−1248. (42) Barton, A. Handbook of Solubility Parameters and Other Cohesion Parameters, 2nd ed.; Routledge: New York, 1991. (43) del Río, J. C.; Lino, A. G.; Colodette, J. L.; Lima, C. F.; Gutiérrez, A.; Martínez, Á . T.; Lu, F.; Ralph, J.; Rencoret, J. Differences in the Chemical Structure of the Lignins from Sugarcane Bagasse and Straw. Biomass Bioenergy 2015, 81, 322−338.

H

DOI: 10.1021/acs.iecr.9b00875 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX