Characterization of Organosolv Lignin Particles and Their Affinity to Sulfide Mineral Surfaces

: Organosolv lignin nanoparticles have been recently evaluated for their use in mineral froth flotation as a flotation reagent, and as a result, the recovery of the target minerals was improved and the selectivity of the process was increased. However, the mechanism of lignin activity in mineral froth flotation is not known. Therefore, this study is the first step in understanding the interaction of organosolv lignin with the mineral surface. As such, the organosolv lignin was characterized by GPC and 31 P NMR, where the structural differences between the birch and spruce lignins were determined. The molecular size and lignol unit composition were evaluated. Subsequently, the morphology and size of the organosolv lignin particles were examined for all 4 produced types: BN, BM, SN, and SM. The ζ potential was measured in the pH range of 2 − 11. All particles had a high negative charge, which indicated good stability of the dispersion in the alkali range. The stability of their colloidal dispersion was observed under increasing concentrations of mono-and divalent cations, and electrostatic repulsion was identified as the main stabilization mechanism. Finally, QCM-D was used to study the interaction of the lignin particles with the mineral surfaces of chalcopyrite, pyrite, and galena, which gave insight into the possible mechanism during the flotation process.


INTRODUCTION
Lignin is the second most abundant natural polymer, which can be extracted for lignocellulosic biomass, and used for many industrial applications. 1,2Lignin content in lignocellulosic biomass is about 5−30% depending on the source (hardwood, softwood, grass).In 2015, the lignin production reached 100 million tons, but only 2% of that volume was sold for industrial applications, while the rest was used for energy production through burning. 3Chemically, lignin is a complex phenolic heteropolymer based on the three main monomeric units, pcoumaryl, coniferyl, and sinapyl, called lignols.The radical forms of the lignols, generated upon polymerization, are phydroxyphenyl (H), guaiacyl (G), and syringyl (S). 4 The ratio of monomeric units in the polyphenolic structure is affected by the source of lignin.The hardwood, such as birch, is expected to have a high content of condensed syringyl groups, which is characteristic of angiosperm lignins.The softwood is expected to have a high content of guaiacyl (G groups), which is common for gymnosperm plants (conifer), such as spruce. 4,5here is an intrinsic amphiphilicity of the lignin molecule due to the presence of phenolic and aliphatic hydroxyl groups, carboxylic groups, and aromatic rings, which are present in various quantities along the polymeric backbone, depending on the chosen type of lignin. 6The lignin isolation method controls the degree of polymerization, where the used temperatures, solvent content, and presence of a catalyst will affect the final size of the lignin molecules.Thus, the combination of different lignin sources and different pretreatments can lead to fundamentally different lignin molecules. 4,7herefore, lignin can be considered as a source for a range of valuable compounds, with different compositions and properties, which can be utilized in a variety of industrial applications. 8,9he organosolv pretreatment is a fractionation method, which allows sufficient separation of lignocellulosic biomass into the three main fractions of cellulose, hemicellulose, and lignin. 7The lignin recovered in such fractionation is of high purity, with low content of ash and sulfur, and is not highly modified by the pretreatment. 4,7These characteristics are advantageous for the further applications of lignin: (1) in agriculture and pharmacy as active compound controlled release carriers; (2) for the production of bioplastics, nanocomposites, and nanoparticles; and (3) as an adsorbent of heavy metals and UV-light. 2he application of lignin nanoparticles is currently extensively studied. 1,10,11In our previous study, the organosolv lignin nano-and microparticles (OLPs) were evaluated for their use as reagents in mineral froth flotation. 12Mineral froth flotation is a mineral processing technique that allows the selective separation of minerals based on the hydrophobicity of their surface.The hydrophobic minerals are lifted by bubbles to the surface, where a froth is formed and subsequently removed.In general, the process requires a whole spectrum of reagents for it to be efficient, selective, and economically feasible.Among the reagents, a frother is normally required to facilitate the creation of froth with specific stability to allow froth separation.A collector is necessary to facilitate the connection between the mineral and bubble surfaces to enable the transport of mineral pulp into the froth.Finally, modifiers are reagents that are used to improve the selectivity of mineral separation, as they selectively modify the mineral surfaces to either facilitate or eliminate the attachment of the collector to the mineral surface.The results of our previous study have shown that lignin nano-and microparticles can be used as a flotation reagent for copper recovery from sulfide ores. 12The copper recovery from the sulfide ore exceeded 80% with a grade of 8.6% in the rougher flotation tests, and the selectivity toward copper was improved as less iron was recovered.In addition, the OLP in the reagent mixture with xanthate improved the recovery of copper (91%), lead (85%), and zinc (98%) from a sulfide ore sample, compared to the xanthate alone, where the recovery was 84, 70, and 96%, respectively.The grade and selectivity increased as fewer iron-bearing minerals were recovered in the process.Finally, the difference in the performance of the OLP was evaluated, which showed that there was a significant difference in recovery and grade based on the type and size of the lignin particles.Using SM, BM, and BN in the Cu−Ni ore sample resulted in a Cu recovery of 87, 70, and 77% with a grade of 8.9, 8.6, and 7.9%, respectively.Thus, the size and lignin source of the particle affect the flotation results. 12However, the mechanism was unknown.Other authors reported the use of lignin, mainly lignosulfonates, in mineral froth flotation, 13−15 but very little was reported on the mechanism and interaction of lignin and mineral surfaces. 16his paper aims to make the first step in understanding the interaction of organosolv lignin and mineral surface and further understanding the mechanism by which the lignin improves the selectivity of the mineral froth flotation.Therefore, knowing the surface chemistry of the OLP will be crucial in understanding the process.This paper discusses the characterization of lignin isolated from spruce and birch, preparation and morphology of the OLP, and surface characterization of the OLP, such as ζ potential, size, and stability of the OLP dispersion.Finally, the interaction between the OLP and mineral surfaces was studied using a quartz crystal microbalance with dissipation monitoring.

Lignin Production and Characterization.
The lignin used for the preparation of lignin particles was extracted by organosolv pretreatment of birch and spruce wood chips as previously described. 17Specifically, the birch wood chips were pretreated at 183 °C for 1 h in a 50% v/v ethanol in water solution, while the spruce wood chips were pretreated in 60% v/v ethanol in water solution at 183 °C for 1 h.The organosolv lignin produced was of high purity and stored as a freeze-dried powder (Telstar lyoquest, Terrassa, Spain).Specifically, birch lignin contained less than 5 and 1% wt xylan and ash, respectively, while spruce lignin contained less than 6% wt glucan and xylan with less than 1% wt ash. 17he molecular weight of organosolv lignin was determined by gel permeation chromatography (GPC).Lignin was derivatized by adding 0.9 mL of glacial acetic acid and 0.1 mL of acetyl bromide to 5 mg of lignin powder.Subsequently, the sample was stirred for 2 h at room temperature in closed vials.To remove the solvents, the solution was transferred to a round flask and evaporated in a rotary evaporator (Heidolph, Schwabach, Germany) at 50 °C and 50 mbar.The sample was washed twice with 1 mL of tetrahydrofuran (THF), followed by solvent evaporation.Finally, the sample was dissolved in 1 mL of THF and filtered through 0.22 μm hydrophobic filters (Sartorius, Goẗtingen, Germany).The samples were analyzed by HPLC using a UV detector (set at 280 nm) and a Styragel HR 4E column (Waters, Milford, MA), operated at 40 °C, with THF as the mobile phase, and a flow rate of 0.6 mL/min.One measurement was done for each lignin sample.The calibration was done by using polystyrene (Sigma-Aldrich, St. Louis, MO).The numbers were rounded up at 100 s due to the resolution of the method.
For quantitative 31 P NMR analysis, 18 approximately 120 mg of lignin was dissolved in 1.6 mL of anhydrous CDCl 3 /pyridine solution (1:1.6 (v/v)).Subsequently, 400 μL of a standard solution (0.1 M cholesterol in anhydrous CDCl 3 /pyridine solution) containing Cr 3+ acetylacetonate as the relaxation agent was added to this solution, followed by the addition of 400 μL of 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxa-phospholane (Cl-TMDP, 95%, Sigma-Aldrich).The mixture was stirred at room temperature for 2 h and then transferred to 10 mm NMR tubes.The spectra were recorded using a Bruker Ascend Aeon WB 400 (Bruker BioSpin AG, Fallanden, Switzerland) NMR spectrometer (72 scans at 25 °C).The chemical shifts reported are relative to the reaction product of water with Cl-TMDP, which gives a sharp signal in pyridine/CDCl 3 at 132.2 ppm.The signals for aliphatic, condensed, guaiacyl, p−OH phenol, and carboxylic groups were at 149.0−146.0,144.27−140.27,140.24−138.8,138.8−137.4,and 135.5−134.0,respectively.Quantitative analysis was performed based on previous literature reports. 19.2.Organosolv Lignin Particle (OLP) Preparation.The nanoparticle production started by dissolving 5% (w/v) of the freeze-Scheme 1. Illustration of the Nanoparticle and Microparticle Production Process dried organosolv in 75% (v/v) ethanol/water solution (Scheme 1).Subsequently, the lignin solution was homogenized at 750 bar using a pressure homogenizer (APV-2000, SPX FLOW, Charlotte, NC).The homogenized lignin solution was diluted by deionized water (1:6), resulting in the formation of nanoparticles.The preparation process yielded spherical nanoparticles that were smaller than 500 nm.Finally, the nanoparticle dispersion was freeze-dried (Telstar lyoquest, Terrassa, Spain) to obtain the nanoparticles in powder form.20 The production of microparticles started by dissolving 1% (w/v) of the freeze-dried organosolv lignin in 75% (v/v) ethanol/water solution.Subsequently, the lignin solution was homogenized at 750 bar using a pressure homogenizer (APV-2000, SPX FLOW, Charlotte, NC).The homogenized lignin liquid was placed on a rotary evaporator (Heidolph, Schwabach, Germany), and the ethanol content was reduced.The formation of microparticles was indicated by a color change, which occurred when about 70−80% of the solution was removed.The process yielded spherical microparticles that were about 1 μm in diameter.To obtain the microparticles as a dry powder, we freeze-dried the sample (Telstar lyoquest, Terrassa, Spain).
2.3.Morphology.The morphology of the OLP was observed by scanning electron microscopy (SEM, FEI Magellan 400 field-emission XHR-SEM).The samples were placed on conductive carbon tape prior to the analysis, and the images were taken at a low accelerating voltage of 3 kV and a beam current of 6.3 pA.The primary size of the particles was measured using ImageJ software (LOCI, University of Wisconsin).For each measurement, at least 100 particles were measured from at least 3 different SEM images.

ζ Potential.
The concentration of 0.05 mg/mL spherical OLPs in a 10 −3 M KCl solution was analyzed in a Zetasizer Nano Series (Malvern Analytical, Malvern, UK) instrument to determine their ζ-potential.All measurements were performed in triplicate of the same sample.The ζ potential measurement was done in a pH range of 2−11, which was adjusted by HCl and NaOH.
To observe the effect of Na + and Ca 2+ ions on the ζ potential of the OLP, the concentration of 0.05 mg/mL spherical OLPs in 2×, 4×, 6×, 8×, and 10× 10 −4 M NaCl and 2×, 4×, 6×, 8×, and 10 × 10 −5 M CaCl 2 solutions was analyzed.As a result, the ratio between the lignin and cations was equivalent to the ratio of OLP and cations used in the dispersion stability measurement.All measurements were performed in triplicate of the same sample.
2.5.Lignin Dispersion Preparation.For particle dispersion preparation, two separate approaches were selected.Thus, two types of OLP dispersions were prepared: pH adjusted and pH nonadjusted.The pH nonadjusted was 1 wt % OLP suspension in distilled water that was sonicated for 10 min in a sonication bath (Branson CPX3800H) at the sonication power level of 110 W. The pH value of the first solution was 4.4−4.7 depending on the OLP type.For the pH adjusted dispersions, 0.5 wt % OLP suspension in distilled water was homogenized for 5 min at 13,500 rpm by an Ultra-Turrax T25 homogenizer (IKA, Germany).After that, the homogenized dispersion was adjusted to pH 9.5 by adding NaOH and subsequently homogenized again for 2 min at 13,500 rpm.
The dispersions prepared according to the above-described procedure were used directly for size distribution measurement and Turbiscan measurement.For the QCM-D measurement, the dispersions were allowed to stand on the bench overnight.The sediments were separated from the stable suspension, which was used for the QCM-D measurement to avoid clogging of the tubing by bigger particles.Out of the pH nonadjusted 1% wt OLP dispersion, about 0.1% wt dispersion of nanoparticles were obtained, and about 0.2−0.3wt % dispersion of microparticles was obtained.In the case of the pH adjusted 0.5% wt OLP dispersion at pH 9.5, insignificant amounts of sediments occurred in the dispersion.
2.6.Laser Diffraction Measurements.The hydrodynamic particle size of the OLP in the dispersions of the two main dispersions, freeze-dried, and resuspended with and without pH adjustment, was measured using the laser diffraction technique.In addition, freshly prepared particles before the freeze-drying process and decanted fraction of the dispersions were measured for SN and BM.A Malvern Mastersizer MicroPlus (Malvern Analytical Ltd., Malvern, UK) was used for the measurement.First, 500 mL of degassed water was placed in a 750 mL beaker, and the pump speed was set to 2000 rpm to create a flow of water through a flow cell.After the background measurement, a few milliliters of stock dispersions was added to the circulating water until around 15% obstruction was achieved.The measurements consisted of 15,000 measurement sweeps, and a volume-based size distribution was obtained, as it is a more suitable parameter in the case of very polydisperse systems compared to the intensity-based size distribution. 21The distribution is expressed in terms of the volumes of the equivalent spheres.

Stability of Dispersions.
The stability of the OLP dispersions was measured using a Turbiscan Lab Expert device (Formulaction, Toulouse, France). 2220 mL of OLP dispersion was placed in a 25 mL glass vial and closed with a plastic cap.The sample was shaken before measurement.The adjustment of the Na + and Ca 2+ ion strength was done by the addition of 1 mL of the ion solution into the glass vial before further measurement, and the sample was shaken and measured again.For the Na + and Ca 2+ ion adjustment, a 1 and 0.1 M solution of NaCl and CaCl 2 were used, respectively.The stability was measured as a value of backscattering at the height range of 20− 25 mm of the glass vial every 2 min within a period of 30 min.The data are expressed as a difference in the backscattering of dispersion in the function of time.
2.8.Quartz Crystal Microbalance with Dissipation Monitoring.The quartz crystal microbalance with dissipation monitoring (QCM-D) 23 measurement was used to study the interaction between the OLP and mineral surfaces of chalcopyrite, pyrite, and galena, which were custom-made by Q-Sense Ab (Gothenburg, Sweden).The measurements were performed on Q-Sense E4 equipment (Q-Sense Ab, Gothenburg, Sweden).Before use, the mineral-coated QCM sensors were cleaned with 10 min of UV treatment, followed by 10 min of sonication in pure ethanol.Further, the sensors were rinsed with deionized water and dried by nitrogen.During the initial measurement setup, the fundamental frequency and all of the overtones were found easily, and the shape of the peaks was also normal, which indicated that the crystal coating was stable.In total, 8 different OLP dispersions were examined.For the examination of the nonadjusted pH of the OLP, at the lower pH (4.4−4.7), the measurement was performed with deionized water.For the examination of the OLP dispersion with the pH adjusted to 9.5, the measurement was performed with the NaOH solution in deionized water with a pH of 9.5.The stable dispersions were diluted 10 times in distilled water before the QCM-D measurements, which were performed at a flow rate of 50 μL/min.First, deionized water was pumped through the system until a stable baseline was established.The OLP dispersion (0.01−0.05 wt %) was infused onto the crystal for 60 min, followed by a 60 min rinsing with water.Time evolutions about changes in resonance frequency (Δf) and energy dissipation (ΔD) were acquired.Each measurement was performed in duplicate, where two different crystals were used with the same OLP dispersion.In the case of Galena, the measurement was repeated four times in total to verify the frequency drift.

Lignin Characterization.
The chemical structure of lignin varies significantly depending on the lignin source e.g., hardwood plants, softwood plants, and grass.The degree of polymerization is affected by the isolation methods. 4Therefore, different types of lignin can be isolated based on the plant source and specific conditions of the isolation.The organosolv lignins, obtained from spruce and birch, were characterized by gel permeation chromatography (GPC) for their molecular weight and by quantitative 31 P NMR analysis to determine the content of aliphatic, phenolic, and carboxylic groups (Table 1).
GPC was used to determine the weight-average molecular weight (M w ), number-average molecular weight (M n ), and polydispersity index (PDI).The M w of the organosolv lignin is typically in the range of 0.5−10.8kDa compared to 1.5−5.0kDa for the kraft lignin and 5.0−50.0−26 In addition, the polydispersity for lignosulfonate and kraft lignin is high compared to the organosolv lignin. 2 For the organosolv spruce lignin, the M w and M n values were 4600 and 1200, respectively, with a PDI of 3.9.Similar data were reported by Gordobil et al., as the M w and M n values were 3081 and 1065, respectively, with a PDI of 2.89 for spruce lignin produced by organosolv pretreatment with 50% ethanol solution at 180 °C for 60 min with an acid catalyst. 25The conditions were similar but the use of an acid catalyst led to lower M w and M n values compared to the ones obtained in this study. 17The M w and M n values were 4200 and 1200, respectively, with a PDI of 3.5 for the birch organosolv lignin.Similar results were obtained after organosolv pretreatment of birch sawdust at the same organosolv pretreatment conditions, where the M w and M n values were 4600 and 1200, respectively, with a PDI of 3.88. 24The chromatograms of the GPC measurements are presented in Figure S1.
The content of aliphatic OH groups and carboxylic groups was 3.47 and 0.19 mmol/g in the spruce lignin.The total phenolics content was 2.73 mmol/g with a high guaiacyl (G) content of 1.53 mmol/g, which was expected as spruce is a gymnosperm plant (conifer). 27The content of aliphatic OH groups in birch lignin was slightly higher and reached up to 3.78 mmol/g, while the content of carboxylic groups was as low as 0.07 mmol/g.The total phenolics content was 2.29 mmol/g with a high condensed (S) content of 1.58 mmol/g, which is a common characteristic of angiosperm lignins. 27.2.Organosolv Lignin Particle Preparation and Morphology.The SEM imaging showed that the lignin particles have a well-defined spherical shape with a narrow size range.The primary size of the nanoparticles was 96.1 ± 20.1 and 108.6 ± 21.5 nm for BN and SN, respectively, with very few misshaped artifacts (Figure 1A,B).The primary size of the microparticles was 1.15 ± 0.21 and 1.00 ± 0.21 μm for BM and SM (Figure 1C,D), respectively.The size of the produced particles was affected by the production method. 20The nanoparticles were formed instantly as the homogenized liquid was mixed with deionized water by pouring, while the microparticles were formed over a period of few minutes when ethanol was evaporated from the solution.The general mechanism of the particle formation is the same for both particle types.The bigger and more hydrophobic molecules create the core of the particle, while the smaller and more  hydrophilic molecules create the outer layer of the particle.The mechanism of lignin self-assembly from 70% v/v ethanol solution was described by Sipponen et al. ( 2018) as a molecular-weight-dependent precipitation. 6When the ethanol content was decreased from 62 to 36%, the high-molecularweight lignin created crumpled particles with incompletely fused small particles at their surfaces.When the ethanol content dropped below 26%, the adsorption of smaller molecules rendered the particles gradually smoother and spherical lignin nanoparticles were obtained. 6The surface-areato-volume ratio plays a crucial role in the morphology of the particles.By increasing the surface area to volume ratio, the percentage of atoms at the surface is increasing as well.The smaller the particle, the more atoms are on the surface of the particle. 28Therefore, the hydrophilic small molecules will be distributed more thinly in the case of nanoparticles compared with the microparticles, where a thicker layer would be expected.

ζ Potential Measurement of Lignin.
The stability of the dispersion can be predicted by measuring the ζ potential of the dispersion.For values close to zero, maximal agglomeration and precipitation are expected.The stability of the dispersion improves when the value is further below or above zero and good stability is expected from the value of ±40 mV. 20,29−32 The results of the ζ-potential measurements are presented in Figure 2. From the pH value of 2, all the OLPs follow the same trend, with increasing pH value, the ζ potential value becomes more negative.The change can be attributed to the deprotonation of the carboxylic and phenolic groups that are present in the lignin particles.For BM, the ζ potential values reached −25.7 mV at pH 6, followed by a slow increase to −19.5 mV at pH 8. Thus, BM would be expected to have the lowest stability compared to the other OLPs.The obtained values for BN and SN were similar, with the lowest values −41.8 and −44.6 mV at pH 10, respectively, while SM had the lowest value, −55.1 mV, of all the OLPs measured at pH 9. The difference in the charge of the OLPs could be explained by a difference in the lignin source.Spruce lignin contains significantly more carboxylic groups (0.19 mmol/g) compared to birch lignin (0.07 mmol/g), which would explain the difference in ζ potential under acidic conditions (pH 3−4).Under alkali conditions (9.5−10.5) the biggest contribution to the change of ζ potential would be caused by the phenolic group, where the birch lignin (2.29 mmol/g) is less of the total phenolic groups than spruce (2.73 mmol/g).The significant difference in the ζ potential value under alkali conditions below pH 9 would be caused by the total number of charged groups in the particles as birch lignin has lower amounts of the charged carboxylic and phenolic groups compared to the spruce lignin.The steep change in the negative value of SM could be caused by a high negative net charge, which increases the intramolecular electrostatic repulsion and leads to destabilization of the microparticles and its swelling. 33This would explain the change in the ζ potential value as swelling particles would be bigger and thus slower in the electrophoretic mobility measurement, which would then appear as an increase in the ζ potential. 29This would also explain the small increase for both nanoparticles, BN and SN, which would swell slightly and increase in size.Overall, the measurement supports the observation that the OLP dispersions are not very stable at the nonadjusted pH (4.4−4.7) and that the stability increases with the increased pH value, which applies to BM, as well, even if the ζ potential suggests it should not (see Section 3.5).This phenomenon can be explained by the limitation of the instrumentation.While ζ potential measurement measures the electrostatic repulsive forces, it does not provide any insight into the attractive van der Waals forces, which contribute to the colloidal stability.
−37 3.4.Size Distribution of Lignin Particle Dispersions.The SEM imaging suggested that the OLPs are universally spherical with very few misshaped artifacts.The size of the spherical OLP was around 100 nm for nanoparticles and 1 μm for microparticles.However, after resuspending the freezedried powder of OLPs in distilled water, aggregation and precipitation occurred.Thus, the size of the OLP was measured by laser diffraction in addition to SEM imaging.For a deeper understanding of the OLP, the measurement was first performed with freshly prepared OLPs, not subjected to freeze-drying.The second measurement was carried out with freeze-dried resuspended distilled water.Table S2 summarizes the D10, D50, and D90 values obtained from the measurement.
After the preparation of the OLP, there was about 10−15% of ethanol still present in the solution, and no aggregation and precipitation were observed.However, the laser diffraction measurement suggested that the SN already formed aggregates in the size of 10−100 μm (Figure 3).BN was much less aggregated, with the majority of the particles below 1 μm, but some aggregates of bigger sizes were present.
The aggregation of SN compared to that of BN could be explained by the different lignin source.Spruce lignin contains significantly more carboxylic groups (0.19 mmol/g) compared to that of birch lignin (0.07 mmol/g).The carboxylic groups at lower pH should have enough of a negative charge for electrostatic repulsion to stabilize the dispersion.However, the total amount of aliphatic, phenolic, and carboxylic groups is not the only difference in the lignin structure.There is a crucial difference in the monolignol content in the spruce and birch lignin.While softwood lignin consists mainly of guaiacyl units, hardwood lignin contains guaiacyl and syringyl units.Spruce lignin contains 1.53 mmol/g guaiacyl units, while birch lignin contains only 0.57 mmol/g guaiacyl units and 1.58 mmol/g syringyl units.The increased amount of guaiacyl units increases the branching of the lignin molecule and is known to encourage lignin condensation. 38In addition, based on the mechanism of the particle formation, a thicker layer of smaller hydrophilic molecules would be expected on the surface of microparticles, which would make them more stable in the colloidal suspension. 6Thus, the microparticles were in the size range (D50 = 1.25−1.30μm; D90 = 1.88−2.03)that was observed by the SEM imaging (1.00−1.15μm), which was around 1 μm.Finally, a similar amount of lignin was used for the particle preparation, but when nanoparticles are formed, their total number is significantly higher than the total number of produced microparticles.The particles have the tendency to agglomerate when their concentration is high. 6fter the freeze-drying process and subsequent resuspension in distilled water, all OLPs formed aggregates in the dispersion.The nanoparticles had almost identical size profiles with the two main peak values around 7 and 160 μm (Figure 4).The BM range increased from around 1 μm up to 100 μm with the two main peak values at roughly 4 and 30 μm.The least aggregation was observed for the SM, where the majority of particles was in the range of 1−10 μm with a small number of particles as big as 100 μm.The value of the high peak was 3 μm.There was a significant difference in the size of the OLP observed by SEM and measured by laser diffraction after the freeze-drying process.While the effect of drying on the properties of lignin was not well studied yet, a similar phenomenon of aggregation after drying was observed in cellulose.During the drying process, the surface properties of cellulose change.Thus, the fibrils can irreversibly aggregate by hydrogen bonds and create fixed domains, which cannot be easily accessible by water. 39Agustin et al. suggested that the wet lignin, 40 i.e., before a drying process, maintains interaction with water apart from the lignin−lignin interactions (Hbonding, van der Waals, π−π), 41 which then contributes to possible disintegration of the aggregates.When the water is removed by drying, the water−lignin interaction is no longer in effect, which could lead to the collapse of the aggregates.As a result, rigid and compact particles could be formed. 40The SM, with the most negative ζ potential and the most charged molecules on the surface, exhibited the lowest aggregation. 6he pH value is also crucial for the stability of the OLP. 6,34,37The pH value of the measured OLP was in a range of the native pH 4.4−4.7,where a certain degree of aggregation was expected based on the ζ potential measurement results.The surface charge of the particles at a lower pH is not high enough to prevent aggregation.Therefore, additional measurement was done with the OLP that was dispersed in distilled water, and the pH of the dispersion was adjusted to 9.5 with NaOH and homogenized.As shown in Figure 4, the size range decreased significantly.In the case of SN, the size profile was slightly smaller than freshly prepared SN, so the deprotonation of the charged groups leads to less aggregation in the system. 34he size of BM at pH 9.5 also decreased compared to the size of the OLP dispersion at the lower pH, but it did not reach the size range of the freshly prepared BM, which would support the theory of the formation of rigid and compact particles during the drying process.In addition, the low pH dispersion of BM was decanted, and the stable dispersion was measured  by laser diffraction.All of the aggregates were removed, and the size range was very similar to the freshly prepared BM.
3.5.Stability of Lignin Particle Dispersions.The stability of the OLP dispersion in salt solutions was evaluated by a Turbiscan Lab Expert device, which can detect changes in emulsions and dispersions, such as creaming, sedimentation, agglomeration, aggregation, and coalescence.Polymers can be affected by salts in many aspects, such as shifts in hydrophobicity/hydrophilicity, cloud point, and lower critical solution temperature.In addition, changes in the molecule structure conformation of the polymer can occur with the salt introduction into the solution. 42Therefore, all four OLP types were tested under two different pH conditions with increasing concentrations of Na + and Ca 2+ ions to examine the effect of the cation valency on dispersion stability.
For the control measurement, with no added ions at the lower pH (Figure S3), only SM was stable, while BM, SN, and BN started to slowly sediment.The higher stability of SM compared to BM could be explained by ζ potential values of −36.1 and −24.5 mV, respectively, at pH 5. The pH adjustment to pH of 9.5 with the following homogenization significantly increased the stability of the system, which is supported by the more negative surface charge measured by ζ potential measurement.All of the OLP controls, with no added ions, were stable at the higher pH value (Figure S4).The microparticle dispersions were more stable than the nanoparticle dispersions, which could be a result of their size as measured by the laser diffraction.The real size of the aggregates in the nanoparticle dispersion was bigger than the size of aggregates in the microparticle dispersions.The homogenization process produced smaller aggregates in the microparticle dispersions compared to the nanoparticle ones.
At a lower pH, the dispersions of spruce nano-and microparticles were strongly affected by the addition of Na + and Ca 2+ ions, while the dispersions of birch particles were significantly more stable.This is in correlation with the results of the ζ potential measurements, which showed that spruce particles, both SN and SM, are more charged than birch particles, BN and BM.Based on the NMR results, spruce lignin contains almost three times more carboxylic groups than birch lignin, which could explain the difference in surface charge.The addition of cations to the negatively charged lignin dispersion led to a decrease in the negative surface charge.Thus, the electrostatic repulsion, which is the main stabilization mechanism, is significantly lowered. 43This was confirmed during the additional ζ potential measurement (Figures S5 and S6).The effect is visible mainly in the case of the microparticles after the interaction with both cations.However, the addition of cations in the dispersion may also enhance the hydrophobic interactions by neutralizing the surface charge, which could explain the aggregation. 44Finally, the Ca 2+ ions are divalent and thus can facilitate the bridging of the negatively charged particles.As a result, quick sedimenta-tion can be observed, even at very low concentrations of Ca 2+ ions.
At a higher pH, the particle dispersions were homogenized and the general particle size was significantly lowered, which improved the stability of the dispersion.In addition, the surface charge significantly decreased based on the ζ potential measurement, which indicated improved stability of the dispersion and more available negatively charged groups.The microparticle dispersion was so stable that even the highest concentration of the salts did not cause a significant change to the stability within 30 min of the measurement.Birch nanoparticles were affected more by a higher concentration of Na + ions than the spruce nanoparticles, which could be explained by the higher content of deprotonated carboxylic groups in the spruce lignin.As a result, the negative charge of the particle was lowered and the stability of the dispersion decreased.In all cases, aggregation followed by sedimentation was observed during the measurements, except for the nanoparticles adjusted to higher pH in the presence of Ca 2+ ions, where coagulation followed by sedimentation was observed (peak in the graphs).This phenomenon could be explained by the bridging ability of divalent ions, such as Ca 2+ .
In addition, Zongo et al. investigated the effect of the Hofmeister series on polyphenol-based nanostructures (lignin microcapsules).When comparing the Na + and Ca 2+ ions, the Ca 2+ ions increased the instability of the lignin microcapsules due to their ability to adsorb on the hydrophobic parts and effectively disintegrate the microcapsules, while Na + ions caused aggregation of the microcapsules by neutralizing the surface charge. 33.6.QCM-D Measurement.Quartz crystal microbalance with dissipation monitoring is a real-time technique for analyzing surface interaction phenomena. 23Increasing the mass attached to the surface of the crystal causes a drop in frequency (f), while an increasing frequency would indicate a loss in the crystal's mass.The change in the dissipation factor (D) depends on the rigidity of the attached layer.A rigid layer does not cause a significant change in the dissipation factor, while a soft layer of the attached material on the crystal would increase the dissipation significantly. 23halcopyrite, pyrite, and galena are sulfide minerals with a hydrophobic surface, and their ζ potential trends are similar as the values transit from slightly positive at low pH to negative at higher pH values. 45,46At slightly acidic to neutral pH, the ζ potential value is close to zero as the mineral surface consists of metal cations and oxidized sulfur anions, whereas at higher pH, the cations are no longer the main part of the mineral surface and the oxidized sulfur groups dominate. 47he interaction of the chalcopyrite surface with the OLP showed that the attachment of the lignin particles to the mineral surface is very rigid, as the dissipation value did not change during the measurement (Figure S7).Under both pH values, the birch particles attached in a higher quantity than the spruce particles (Table 2).This could be a result of the difference in the ζ potential of the OLP and the number of carboxylic groups in the lignin sources.The size of the particles did not have a significant effect on the OLP interaction with the mineral surface of chalcopyrite.However, the overall attached mass quantity was small as the change in frequency was rather small (ΔF = 0−10), and the majority of the mass was washed off by water in the second step of the measurement.Only the birch particles at the higher pH value showed a stronger attachment in higher quantities (ΔF = 20).The result of the interaction of the OLP with pyrite was almost identical to the interaction with the chalcopyrite surface (Figure S8).The attachment was very rigid as the dissipation change was extremely low.The birch particles attached in higher quantities than spruce, but the general number of particles attached to the pyrite surface was higher compared to the chalcopyrite surface as the change in frequency was bigger.While there was almost no interaction between nanoparticles and chalcopyrite at the lower pH, there was a very strong interaction between pyrite and nanoparticles under those conditions.The change in frequency in the pyrite crystal was 20 times bigger than the change in the chalcopyrite crystal.In addition, the attachment of the particles was very soft as there was a significant change in dissipation.This might be caused by aggregation of the nanoparticles on the surface of the mineral during the measurement at the lower pH.
In the case of the galena crystal, there was a significant drift in the frequency measurement observed (Figure S9).The crystals were custom-made, and there is very little available literature about their use.Therefore, there are no references to compare our results.At the lower pH, all three mineral surfaces showed a small drift at the beginning of the measurement and in the case of chalcopyrite with SN, the drift was present for the time of the whole measurement.Generally, the increase in frequency indicates a loss of the crystal's mass, such as detachment of absorbed molecules. 23Therefore, a possible explanation for drift in this case could be a surface change, such as oxidation or dissolution of the crystal surface.This would be supported by the fact that when the surface of the mineral is covered by the OLPs, the drift is not as pronounced as the galena surface with nanoparticles under the lower pH, where the change in frequency and dissipation suggested a higher amount of the nanoparticles softly attached to the mineral surface.While in the case of microparticles, there is no attachment observed, and the drift is the highest of all of the measurements.At the higher pH, the higher the amount of attached OLP is, the smaller the drift.In addition, galena as a mineral is significantly more soluble in lower pH of 4 compared to the higher pH of 9.5, which would explain the difference in the value of the frequency drift. 48For experiments with BM at a lower pH value, the drift was almost 140 Hz, while for the higher pH value, the total drift of frequency was only about 20 Hz.Similarly, to pyrite and chalcopyrite, the birch particles attached in higher amounts compared to the spruce particles.

CONCLUSIONS
Organosolv lignin is an underutilized renewable resource that can be fractioned from residual lignocellulosic biomass from the forestry industry.The potential of lignin and lignin nanostructures in many industrial applications has been extensively studied in the past few years.Organosolv lignin nanoparticles have been recently tested as flotation reagents for mineral froth flotation processes, and the results showed improvement in the target mineral recovery and selectivity.However, the mechanism of the organosolv lignin particle activity is not known.This study was the first step toward an understanding of the mechanism of organosolv lignin particles in the mineral froth flotation process.
Organosolv lignin, isolated from spruce and birch, was characterized by GPC and 31 P NMR, which allowed the insight into the type and number of functional groups that are present in the lignin molecule of a certain size.Based on the mechanism of lignin particle self-assembly, hydrophilic small lignin molecules are expected to cover the surface of the particles; thus, it can indicate the amount and type of the functional groups covering each type of the organosolv lignin particle type that was produced.As such, spruce lignin has significantly more carboxylic groups compared to the birch lignin, which can be then observed as a difference in the measured ζ potential.Spruce particles showed notably a higher negative charge compared to birch particles.This was also important for the stability of the colloidal dispersion measured by Turbiscan, which was evaluated under increasing concentrations of mono-and divalent cations.Under the lower pH, only the SM was sufficiently stable in the dispersion, which could be explained by the ζ potential value and the higher content of carboxylic groups.While under the higher pH, all of the dispersions were sufficiently stable as the phenolic group got charged.Through this measurement, it was confirmed that the electrostatic repulsion was identified as the main stabilization mechanism.
Finally, the interactions of lignin particles with mineral surfaces of chalcopyrite, pyrite, and galena were examined.The quantity and rigidity of the particle attachment were measured for all 3 surfaces and all 4 types of lignin particles to evaluate how the attachment of the particle affects the ability of the particle to float in the mineral froth flotation.However, for a full understanding of the mechanism of organosolv lignin particles during mineral froth flotation, further studies are necessary.

Figure 3 .
Figure 3. Volume-based size distribution of the lignin particles measured by laser diffraction.The first graph shows the size of particles right after their formation by dilution or evaporation.The second graph shows the size of the particles after freeze-drying and resuspended in water.

Figure 4 .
Figure 4. Volume-based size distribution of the spruce nanoparticles and birch microparticles measured by laser diffraction.Comparison of freshly prepared (L), freeze-dried and resuspended (FD), decanted after freeze-drying (D), and homogenized at pH 9.5 (H 9.5).

Figure S1 .
Figure S1.Results of GPC for birch and spruce lignin.TableS2.Size distribution of lignin particle dispersions.Figure S3.Results of Turbiscan at pH values of 4.4−4.7. Figure S4.Results of Turbiscan at pH 9.5.Figure S5.Results of ζ potential measurement in Na + .Figure S6.Results of ζ potential measurement in Ca 2+ .Figure S7.Results of QCM-D of chalcopyrite.Figure S8.Results of QCM-D of pyrite.Figure S9.Results of QCM-D of galena (PDF)

Figure S3 .
Figure S1.Results of GPC for birch and spruce lignin.TableS2.Size distribution of lignin particle dispersions.Figure S3.Results of Turbiscan at pH values of 4.4−4.7. Figure S4.Results of Turbiscan at pH 9.5.Figure S5.Results of ζ potential measurement in Na + .Figure S6.Results of ζ potential measurement in Ca 2+ .Figure S7.Results of QCM-D of chalcopyrite.Figure S8.Results of QCM-D of pyrite.Figure S9.Results of QCM-D of galena (PDF)■ AUTHORINFORMATION

Table 1 .
Characterization Results of the Lignin Sample by GPC and NMR

Table 2 .
Overview of the QCDM Results after 1 h of Adsorption of the OLP at pH 9 for Chalcopyrite, Pyrite, and Galena