Synthesis of a Lignin-Fe/Mn Binary Oxide Blend Nanocomposite and Its Adsorption Capacity for Methylene Blue.

A high-performance modified lignin adsorbent was prepared through coprecipitation of ferrous, ferric, and permanganate with lignin in sodium hydroxide solution. The structural characteristics of the synthesized lignin-Fe/Mn binary oxide blend nanocomposite (L-F/M) and its performance on the methylene blue (MB) removal from aqueous solution were evaluated. Influence factors of adsorption effects were analyzed including pH, contact time, dye concentration, temperature, and thermodynamics. The pseudo-second-order kinetic model well described the adsorption kinetics, and the adsorption isotherms best fitted the Langmuir model with a maximum adsorption capacity of 252.05 mg g-1 at 298 K. The adsorption mechanism showed that the L-F/M introduced the metallic element and negative charges to the lignin surface, which improved the adherence of MB via hydrogen bonding, electrostatic interaction, and coordination. Moreover, the removal ratio of MB maintained 81.2% after being used in five adsorption-desorption cycles. Results indicated that the L-F/M obtained was an efficient candidate for dye wastewater treatment.

Introduction

Dyes have been widely used in various industries and pose a great challenge to the environment. Because most of them present favorable stability, they are hard to be naturally decomposed. This results in enhancement of the chemical oxygen demand and brings potential toxicity due to their physicochemical properties. In addition, they may further influence the regular growth of aquatic organisms when discharged into a natural body of water.[1−3] The wastewater produced in the textile industry contains large amounts of dyes. In addition, the salt, alkali, and hydroxides used to promote the combination between dyes and fibers in the process are also presented.[4] It is reported that the total emissions of textile wastewater accounted for approximately 20% of the total amount of sewage worldwide.[5] In particular, methylene blue (MB), one of the most widely used cationic dyes in the textile industry, is known to cause organic pollution in water and severe threat to public health.[6−9] A small amount of MB (>7.0 mg kg–1) can result in nausea, abdominal pain, hypertension, respiratory distress, mental disorder, and even blindness.[10,11] Therefore, the removal of MB from textile wastewater is urgently necessary. Various technologies are available for MB removal to reduce the threats to humans and ecosystems, including photocatalytic degradation, ozonation, ion exchange, ultrafiltration, and adsorption. With respect to the ease of operation, fast reaction rate, low cost, and the good chemical stability of MB, adsorption is particularly appealing and promising among these treatments.[12−14] Many materials such as porous carbon and metal–organic frameworks are used as advanced adsorbents. They perform excellently but still have limitations such as the high cost of large-scale production and the secondary pollution in synthesis.[15−18] Consequently, it is of great importance to develop a cost-effective, green, and efficient alternative adsorbent to accelerate the treatment of MB effluent pollution.

Lignin, an aromatic polymer available in nature, is generated as a by-product from the pulping industry with tens of million tons annually.[19] Most lignin is burned to generate heat, which not only wastes the resources but also decreases economic benefits of biorefinery. However, lignin is rich in functional groups such as benzene ring, aliphatic/phenolic hydroxyl, and methoxy groups, which make it a promising raw material for preparing low-cost adsorbents especially in MB removal, since it can strengthen the combination between lignin and MB through many interacting forces including the chemical or physical action, π–π stacking, and hydrogen bonds.[20,21] However, it has been proven that it is hard to obtain a high MB removal efficiency when using pure lignin as an adsorbent directly. It was found that the MB adsorption capacity of organosolv lignin from rice straw was 40.02 mg g–1.[22] Similarly, Cemin et al. extracted lignin from Eucalyptus grandis sawdust, in which the maximum adsorption capacity for MB was 32 mg g–1.[23] Several researchers have developed a series of lignin-based advanced materials for further enhancing MB removal. A porous graphene oxide/alkali lignin aerogel composite exhibited an excellent MB removal capacity of 1185.98 mg g–1.[24] Alginate/carboxylated lignin hybrid beads had the adsorption capacity of MB as high as 613.0 mg g–1, and manganese-modified lignin biochar showed a high MB adsorption ability of 248.96 mg g–1.[25,26] Taleb et al. modified sulfuric acid lignin isolated from spent coffee grounds through phenolation and acetylation and found that the MB removal rate increased from 64.69 mg g–1 to 91.34 and 71.45 mg g–1, respectively, which might be attributed to the π–π interaction and the increased electrostatic interaction.[27] However, such complex synthesis steps, high-cost reagents, and energy consumption limit their possibility of high-volume production; therefore, exploring lignin-based adsorbents with characteristics of cost-effectiveness and ease of synthesis is urgent. Lignin-derived metal/metal oxide nanomaterials are expected to well balance the cost and performance in developing a novel adsorbent.[28] The introduction of metal can provide a high specific surface area and highly negative charges of raw lignin, resulting in the excellent combination capacity with cationic dyes.[29] It has been reported that the doping of Mn enhanced the MB removal efficiency due to the improvement of surface electron transfer and exchange. Liu et al. prepared manganese-modified lignin biochar for the adsorption of MB and found that the maximum adsorption capacity increased to 248.96 mg g–1, as compared to 234.65 mg g–1 for unmodified biochar.[26] Zhai et al. synthesized lignin nanoparticles by doping MnO2 based on oxidative reaction, which exhibited an excellent MB adsorption capacity of 806 mg g–1.[29] Magnetic adsorbents with the advantages of easiness to recover from wastewater after purification and avoidance of secondary pollution are of great research interest.[30] Li et al. prepared lignin-grafted magnetic nanoparticles under a nitrogen atmosphere, and the adsorbent had a maximum amount of 211.42 mg g–1 for MB with a quick magnetic separation.[31] Dai et al. synthesized a lignin-sodium citrate/Fe3O4 magnetic adsorbent through a one-step hydrothermal method at 120 °C, and the adsorbent exhibited a supersorption capacity of 281.4 mg g–1 for MB.[32]

Despite the fact that the lignin-derived metal/metal oxide nanomaterials provide a promising way for developing novel and available adsorbents, the extra chemical pollution of synthesis remained in the process.[33] Coprecipitation is a novel synthesis method for synthesizing metal/metal oxide nanomaterials with the advantages of high atomic efficiency, easiness to implement, low cost, and energy saving, which offers great convenience in synthesizing a lignin-based adsorbent.[34,35] Moreover, lignin can be used as the polymer matrix to prevent the agglomeration of nanoparticles, which may cause the adsorption decrease.[36,37] Since research on synthesizing a lignin-based adsorbent for simultaneously doping iron and manganese elements was rarely reported, we provided a novel low-cost and energy-saving coprecipitation method. Fe2O3 and MnO2 binary metal oxide was utilized to prepare a magnetic adsorbent and its performance in MB adsorption was evaluated in the present study. The adsorption behavior including adsorption kinetics, isotherms, and thermodynamics was investigated and the key factors influencing the performance of the lignin-Fe/Mn binary oxide blend nanocomposite (L-F/M) were identified. The results of the present work will provide a new light in developing value-added materials of lignin.

Results and Discussion

SEM Analysis

The SEM images of raw alkali lignin, L-F/M, and the L-F/M after the MB adsorption are depicted in Figure . As compared with the smooth surface of the raw alkali lignin, the surface of L-F/M became rough and wrinkled, indicating that the vast surface area may result in more adsorption sites for boosting MB sticking to the L-F/M. In the meantime, many uniform nanoparticle spheres of the cluster structure were homogeneously deposited onto the L-F/M surface, supporting the formation and attachment of binary metal oxide nanoparticles and demonstrating that the aggregation of nanoparticles was inhibited effectively. After dye adsorption, the L-F/M still presented a typical sphere morphology like the original adsorbent. However, the density of distribution decreased, indicating its better stability, and the surface played an important role in adsorption.

Figure 1

SEM images of (a) raw alkali lignin, (b) L-F/M, and (c) L-F/M with MB.

FT-IR Spectrum Analysis

FT-IR spectra within the wavenumber range of 400 to 4000 cm–1 are presented in Figure . As for the spectrum of alkali lignin, the absorption peak at 1100 cm–1 was ascribed to the CH in-plane deformation from the guaiacyl unit of lignin and the peaks from 1300 to 1600 cm–1 were assigned to the aromatic skeleton vibrations and C–H deformation, respectively.[38] The observed peaks from 3000 to 3600 cm–1 were attributed to the hydroxyl groups in lignin. With the addition of binary metal oxide, new adsorption peaks appeared at 500 to 650 cm–1 and around 1000 to 1400 cm–1 and the peaks attributed to Fe–O, Fe–O–H stretching vibration, and Mn–O–H bending vibrations became sharp, illustrating the successful modification of lignin.[39−41] The extra absorption peaks at 1300 cm–1 demonstrated that MB had been adsorbed already.[24] For the L-F/M after adsorption, the peak of Fe–O shifted to the left and widened, implying the existence of coordination interaction between the lone pair electrons from the nitrogen atom of MB and the empty orbitals of metal oxide.[42] It was also obvious that the peak of benzene skeleton vibration ranged from 1300 to 1400 cm–1 was splitting, revealing the formation of π–π interaction between the aromatic rings both in lignin and MB. In addition, after combination with MB, the slight decrement and broadening could be noticed in the intensity and wavelength of the peaks of OH stretching vibration (3400 cm–1) for the L-F/M, indicating the role of hydrogen bond interaction in the adsorption process.

Figure 2

FT-IR spectra of alkali lignin and the absorbance changes of the L-F/M before and after adsorption (from top to bottom).

Specific Surface Area and Pore Size of the L-F/M

Figure demonstrates the N2 adsorption curves and pore properties of the L-F/M. It exhibited reversible type IV isotherms, which suggested that the capacity of adsorption mainly depended on the mono- and multilayer adsorption with capillary condensation.[43] The specific surface area of the L-F/M was further determined to be 82 m2 g–1 through the Brunauer–Emmett–Teller (BET) model, which was higher than that of pure lignin (82 m2 g–1 vs 20 m2 g–1 from E. grandis lignin and 0.1 m2 g–1 from straw lignin) and was likely due to the introduction of Fe/Mn binary oxide.[23,44] However, the total pore volume was only 0.097 cm3 g–1, and the Barrett–Joyner–Halenda (BJH) method was further used to analyze the pore distribution of the L-F/M. It can be seen that most radii of pores distributed in the mesoporous range from 1 to 3 nm with an average pore size of 3.6 nm, as previously fitted type IV isotherms.

Figure 3

BET analysis of the L-F/M: (a) N2 adsorption–desorption isotherm and (b) pore size distribution.

Thermal Stability

Figure shows the thermal decomposition process of the L-F/M, and the weight loss could be divided into three stages. The initial weight loss below 200 °C was attributed to the evaporation of the residual solvent for the sample. Subsequently, the degradation of lignin caused the second mass decline in the range of 200–500 °C. The final falling after 500 °C was due to the decomposition of both lignin and Fe/Mn binary oxide. The introduction of Fe/Mn binary oxide significantly enhanced the thermal stability of lignin with the total mass losing less than 25% of the initial adsorbent sample at 800 °C, which was higher than pure lignin (25% vs 60% at 800 °C).[45,46]

Figure 4

TG-DTA curve of the L-F/M.

Adsorption Experiments and Magnetic Properties

The preliminary results of adsorption experiments are presented in Figure . The L-F/M was added at a certain concentration of MB solution and after a certain period of adsorption. The color of MB solution virtually disappeared and the adsorbent could be recovered using a magnet further. Therefore, the advantages of the L-F/M included both high MB removal capacity and rapid magnetic separation ability, indicating broad application prospects in purifying dye wastewater. As further evidence from the magnetization curve in Figure d, the saturated magnetization of the L-F/M was about 14 emu g–1, whereas pure lignin was nonmagnetic, showing its superparamagnetic characteristic.[47]

Figure 5

Preliminary results of adsorption experiments of (a) MB solution, (b) MB solution after adding the L-F/M for 6 h, (c) the recovery of the L-F/M, and (d) the magnetic properties of the L-F/M.

Effect of Adsorption Time and the Kinetics

The contact time between the L-F/M and MB influenced the adsorption efficiency. As clearly seen from Figure a, the adsorption capacity for MB increased rapidly within the first 240 min at all different temperatures (298, 303, and 313 K) and then increased slowly until the equilibrium was reached after 330 min with the maximum MB removal to 99.8 mg g–1 at 313 K. The high adsorption rate in the initial stage could be attributed to the existence of many available active adsorption sites on the surface of the L-F/M, whereas at a later stage due to the repulsive forces between dye molecules adsorbed, the mass transfer resistance reduced the adsorption rate until its equilibrium was reached. Moreover, the adsorption rate at 313 K was higher than those at 303 and 298 K, which was ascribed to molecular motion facilitated by heat.

Figure 6

(a) Effect of contact time on the MB removal at different temperatures fitted with the pseudo-first-order model, (b) pseudo-second-order kinetic model for adsorption of MB on the L-F/M, and (c) intraparticle diffusion model for MB removal onto the L-F/M at various initial temperatures.

Several physical models can be used to describe adsorption kinetics and the most common forms are pseudo-first-order (PFO) and pseudo-second-order (PSO) models (eqs and 2).[48,49]where q (mg g–1) is the amount adsorbed at a certain time t (min) and qe (mg g–1) is the amount adsorbed at equilibrium; k1 (min–1) and k2 (g mg min–1) are the PFO and PSO kinetic rate constants, respectively.

Figure b presents the fitting result of PSO models, and the parameters of two models calculated are listed in Table . By comparing R2, the PSO model was the most suitable model to simulate the adsorption process and the residual analysis further confirmed the result (reduced chi-sqr time degree of freedom lower than residual sum of squares of the PFO model). Therefore, the rate control of MB adsorption on the L-F/M surface was mainly chemisorption.

Table 1

Parameters of Adsorption Kinetics

T (K)	model	qe (mg g–1)	K1 (min–1)	K2 (g mg–1 min–1)	R2	residual
298	PFO	83.93 ± 1.12	0.01609 ± 0.00087		0.97	5.13
	PSO	99.97 ± 0.59		0.00018 ± 0.00001	0.99	0.0032
303	PFO	89.49 ± 0.92	0.01633 ± 0.00069		0.98	3.54
	PSO	106.38 ± 1.15		0.00018 ± 0.00001	0.99	0.0048
313	PFO	98.07 ± 1.13	0.01926 ± 0.00102		0.97	6.58
	PSO	112.35 ± 0.87		0.00023 ± 0.00001	0.99	0.0043

In addition, internal mass transfer is generally considered as the determinant factor of the dye adsorption rate; therefore, the intraparticle diffusion model is used to test experimental data by means of (eq )[50,51]where Kid is the constant of intraparticle diffusion (mg g–1 min1/2), which represents the rate of the diffusion process for each stage. C is a constant related to the thickness of the boundary layer.

As displayed in Figure c, the fitting lines were not straight and not through the origin, indicating that the adsorption rate was not only constrained by intraparticle diffusion.[24] According to the model, the MB removal for the L-F/M could be divided into three parts. During the first phase, MB migrated to the external surface of the L-F/M. Subsequently, MB diffused into the pores and/or walls of the L-F/M, and in the third phase, MB deposited on the inside of the L-F/M and the adsorption process settled into equilibrium.[52,53]Table lists the parameters of the intraparticle diffusion model. The Kid slightly increased with the increase in temperature and decreased as the adsorption process progressed, resulting from the thicker boundary layer.

Table 2

Parameters of the Intraparticle Diffusion Model

intraparticle diffusion	298 K	303 K	313 K
Kid1 (mg g–1 min1/2)	7.28 ± 0.74	7.81 ± 0.99	8.06 ± 1.11
C1 (mg g–1)	–4.18 ± 5.75	–4.76 ± 7.67	3.94 ± 8.63
R12	0.98	0.98	0.98
residual	4.45	7.93	10.05
Kid2 (mg g–1 min1/2)	2.88 ± 0.15	2.82 ± 0.32	3.42 ± 0.44
C2 (mg g–1)	38.19 ± 1.94	44.71 ± 4.17	47.29 ± 5.61
R22	0.99	0.97	0.97
residual	0.32	1.47	2.65
Kid3 (mg g–1 min1/2)	1.52 ± 0.24	1.24 ± 0.29	0.69 ± 0.15
C3 (mg g–1)	58.71 ± 4.12	68.86 ± 4.82	87.51 ± 2.53
R32	0.95	0.90	0.91
residual	0.47	0.65	0.18

Effect of pH

pH is an important factor for MB removal by affecting the existing form, surface charge, and interaction of the adsorbent simultaneously. As shown in Figure a, with the increase in pH, the MB adsorption first increased from 23 mg g–1 to 88 mg g–1 and then decreased to 59 mg g–1 at 298 K, and similar trends were observed at other temperatures. The maximum quantity of MB adsorption was achieved at pH = 7, and the performance of the adsorbent in the alkali condition was better than the acid condition. It could be mainly accredited to the enhancement of electrostatic repulsion between the cationic dye MB and the positive charge on the surface of the L-F/M in the acid condition, whereas in the alkali solution, the negative charges on the surface of the L-F/M would facilitate the combination with the cation dye MB. Moreover, the stability of the L-F/M would be decreased at extreme pH conditions due to the reaction of metallic oxide in the acidic environment and the depolymerization of lignin in the alkaline environment. Therefore, the optimal condition for adsorption of MB onto the L-F/M is the neutral environment.

Figure 7

Effect of (a) pH and (b) temperature on the adsorption of MB onto the L-F/M and (c) the equilibrium adsorption amount at different temperatures and initial MB concentrations fitted with the Langmuir and Freundlich models.

Effect of Temperature

For further determining the direction of spontaneity and energy exchange in the adsorption process, the effect of temperature was examined and the thermodynamic parameters were calculated.[47] For all initial concentration solutions, the MB removal increased with increasing temperature before equilibrium (Figure b), indicating that the process was endothermic and might be interpreted as the promotion of mass transfer by heating. At 298 K, the average MB removal (qe) was below 90 mg g–1, whereas it increased to more than 100 mg g–1 at 313 K. Since industrial wastewater was generally hot, the L-F/M was applicable to purify the dye effluents in most actual cases.[54]

Further, the thermodynamic parameters including enthalpy ΔH (kJ mol–1), entropy ΔS (J mol–1 K–1), and free energy ΔG (kJ mol–1) are calculated via Van’t Hoff and Gibbs–Helmholtz equations (eqs and 5), respectively, in Table .where qe is the amount of MB adsorbed at equilibrium (mg g–1), Ce is the equilibrium concentration (mg L–1) of aqueous MB, R is the ideal gas constant (8.314 J mol–1 K–1), and T is the temperature (K). It can be seen from Table that the value of ΔH was positive, indicating that the adsorption of MB onto the L-F/M was an endothermic process. Furthermore, the positive value of ΔS suggested the certain affinity of the L-F/M for MB. There was an increase in randomness between the solid (L-F/M)/liquid (MB solution) phases, and one possible cause might be the interaction between water and MB like hydration and deaggregation.[55] The values of ΔG were negative at all temperatures and decreased from −0.63 to −1.05 kJ mol–1 with the increase in temperature from 295 to 315 K, implying that the process was spontaneous and favorable at a high temperature.

Table 3

Parameters of Thermodynamics

T (K)	ΔG (kJ mol–1)	ΔH (kJ mol–1)	ΔS (J mol–1 K–1)	R2 and residual
295	–0.63 ± 1.40	5.45 ± 0.71	20.62 ± 4.07	200 mg L–1
300	–0.74 ± 1.41			R2 = 0.99, residual = 6.76 × 10–5
305	–0.84 ± 1.43			225 mg L–1
				R2 = 0.98, residual = 9.77 × 10–5
310	–0.94 ± 1.45			250 mg L–1
315	–1.05 ± 1.47			R2 = 0.99, residual = 6.31 × 10–5

Adsorption Isotherms

Adsorption isotherms focus on establishing the equilibrium between the adsorbent and the adsorbate, which can provide insight into the mechanisms of the absorption process. Batch experiments were carried out at different concentrations ranging from 25 to 250 mg L–1 under different temperatures (298, 303, and 313 K) with 0.1 g of L-F/M to get an equilibrium concentration. The data were fitted to isotherm models by Langmuir (eq ) and Freundlich (eq ) adsorption isotherm models (Figure c).[56−58]where Ce is the equilibrium concentration of the solution (mg L–1), qmax is the maximum monolayer adsorption capacity (mg g–1), kL is a Langmuir constant (L g–1), n is the dimensionless exponent of the Freundlich model, and kF is a Freundlich constant.

The obtained adsorption isotherm parameters are shown in Table . According to the R2 and residual analysis, the Langmuir model was more suitable to describe the interactions between the L-F/M and MB compared with the Freundlich model, implying that the MB molecules were adsorbed as a homogeneous monolayer on the L-F/M. The maximum adsorption capacity (qmax) calculated by the Langmuir model for MB increased from 252.05 mg g–1 to 337.83 mg g–1 with an increase in temperature from 298 K to 313 K, demonstrating that the adsorption was endothermic reaction. In addition, the values of kL, which were intimately associated with the affinity of the active sites and energy of adsorption, were lower than 1 but higher than 0, further indicating that the adsorption process was favorable.

Table 4

Parameters of Langmuir/Freundlich Model for MB Removal on the L-F/M

	Langmuir				Freundlich
T (K)	qmax (mg g–1)	kL (L g–1)	R2	residual	kF (mg L–1)	n	R2	residual
298	252.05 ± 60.34	(2.43 ± 0.86) × 10–3	0.99	9.51	1.24 ± 0.51	1.26 ± 0.13	0.98	16.26
303	295.48 ± 93.15	(2.11 ± 0.94) × 10–3	0.98	14.78	1.19 ± 0.56	1.23 ± 0.14	0.98	22.28
313	337.83 ± 107.80	(1.83 ± 0.80) × 10–3	0.99	13.47	1.14 ± 0.50	1.21 ± 0.12	0.98	20.08

Many research studies about modified lignin adsorbents for MB removal have been published and the performance of the L-F/M was compared with several previous studies to further evaluate its performance. Table demonstrates that the capacity of the L-F/M for MB adsorption was better than the others, especially for pure lignin (Organosolv Lignin). The MB adsorption of the L-F/M increased 12 times, suggesting that the certain elemental loadings of Fe/Mn promoted the combination of MB and the adsorbent.

Table 5

Comparison of MB Adsorption Capacities in Various Modified Lignin Adsorbents

material	modification agent	capacity (mg g–1)	references
lignin from prehydrolysis liquor	horseradish peroxidase	241.1 (308 K)	Liu et al.[59]
lignin	Fe3O4, sodium citrate	149.7 (318 K)	Dai et al.[32]
sulfuric acid lignin		66.225 (298 K)	Taleb et al.[27]
lignin	5-sulfosalicylic acid	83.2 (318 K)	Jin et al.[60]
technical lignin	SiO2	60 (298 K)	Budnyak et al.[61]
organosolv lignin		20.38 (293 K)	Zhang et al.[22]
lignin	acetic acid	63.3 (313 K)	Feng et al.[62]
L-F/M	Fe/Mn binary oxide blend	252.05 (298 K)	the present work

Cycling Stability

The cycling stability plays a crucial role in evaluating the economy and feasibility of the adsorbent. The step stripping efficiency (SSE) was used to evaluate the stability between contiguous cycles. As depicted in Figure , it remained relatively stable and only a small drop of 3% occurred after the first and third cycles. The results also showed a slight decline of MB removal (recycling efficiency, RE) after five saturation/regeneration processes, which was mainly ascribed to its structural damages due to the reheating or pH change during the adsorption and desorption processes. In the final cycle, the removal efficacy of MB was still above 81% as compared with the initial adsorption capacity, indicating the excellent reusability and broad potential of the L-F/M in sewage treatment.

Figure 8

Adsorption–desorption cycle properties of the L-F/M for MB removal.

Adsorption Mechanism

Given all the above results, the modification of lignin by doping Fe/Mn binary oxide facilitated the process of MB removal. The synthesis of Fe/Mn binary oxide was based on MnO4– + 3Fe2+ + 7H2O = 3Fe(OH)3 + 5H+ + MnO2.[36,63,64] The coprecipitation was driven by the change in OH– concentration. As the concentration of OH– decreased in the process of synthesis of Fe/Mn binary oxide nanocomposites, alkali lignin precipitated with them together. Lignin used in the present study originated from the pulp and paper industry, and its structure was dramatically modified. The presence of hydroxyl groups made lignin negative, and the interaction with cationic dyes was enhanced due to the functional group with a negative charge.[21] Therefore, the lignin-based adsorbent was usually used in removing cationic compounds and greatly depended on the pH of the system. Several studies reported that the introduction of iron oxides with a positive charge could not change the overall electronegativity of the lignin-based adsorbent, and as the positive charge increased, the lignin-based adsorbent tends to aggregate.[65,66] Therefore, it is expected that the removal of anionic dye would be low, and the limited interaction mainly depends on the iron and manganese oxides with a positively charge of the L-F/M instead of lignin. The pore volume of the L-F/M was far smaller than the specific surface in the BET analysis as the previous research study illustrated, indicating that the surface properties were crucial for the adsorption capacity.[59] Similar results also could be found in the fitting results from PFO, PSO, and intraparticle diffusion models, which suggested that the chemisorption process through surface functional groups played an important role in the overall adsorption rate control in addition to diffusion.[67] Four main interactions including π–π stacking, electrostatic attraction, hydrogen bonds, and coordination were put forward to reveal the possible mechanisms of MB adsorption onto the L-F/M (Figure ). Generally, the π–π stacking and hydrogen bond interactions were widespread in a bio-based adsorbent and, similar to the L-F/M, were also illustrated through the FT-IR spectra.[68−71] As discussed previously, the key to MB removal was the pH of the adsorption system, which influenced the stability of the adsorbent and interaction between the L-F/M and MB simultaneously. The structure of the adsorbent would be destroyed in either acid or basic cases, resulting in adsorption site loss. As for the interaction at different pH values, the negative charge in the oxygen-containing group on metal oxide could combine with the positive charge in the nitrogen atom from MB via electrostatic forces in neutral and alkaline conditions, whereas the only interaction in the acid condition was hydrogen bonds formed by the nitrogen atom and hydroxy in metal oxide. Moreover, the coordination bonds formed by the unoccupied electron orbitals in metal oxide and lone pair electrons in sp2 hybridization nitrogen from MB also played an important role in the adsorption process, which were inferred from the red shift of the Fe/Mn–O stretching vibration peak. However, compared with metal oxide, the empty orbitals of the hydrogen ion were preferred to combine with nitrogen in acid cases; thus, the MB removal was substantially down and lower than the case under neutral or alkaline conditions.

Figure 9

Schematic illustration of the adsorption process and interactions between the L-F/M and MB under different pH values.

Conclusions

In the current investigation, a novel lignin-based material doped with Fe and Mn elements was synthesized and its performance for MB removal was evaluated by batch experiments. Embedded binary oxide blend nanoparticles with alkali lignin had efficiently increased charge density and introduced multiple interactions, thus enhancing MB adsorption. Both the initial MB concentration and pH impacted the adsorption process. The adsorption capacity increased with increasing temperature. The negative value of the Gibbs free energy change indicated that the adsorption process was spontaneous, favorable, and exothermic. The adsorption isotherm and kinetics preferred following the Langmuir and pseudo-second-order models, respectively. The maximum equilibrium adsorption capacity of MB reached more than 100 mg g–1 at 313 K, and its efficiency only reduced approximately 19% after five desorption–adsorption cycles. In addition, the adsorbent could be easily separated by a magnet due to its paramagnetic property resulting from the doped iron oxide. The adsorbent had a high efficiency and the process was of low cost and was easy to operate, implying that the technology had promising potential in commercialization.

Experimental Section

Materials

Alkali lignin was provided by Shandong Longlive Bio-Technology Co., Ltd. (Shandong, China). FeSO4·7H2O, FeCl3·6H2O, NaOH, and KMnO4 were obtained from Beijing Chemical Works, China. Methylene blue was purchased from Sinopharm Chemical Reagent Co., Ltd., China. All chemicals employed in this experiment were of analytical grade and were used without further purification.

Synthesis of the L-F/M

Three grams of alkali lignin and 50 mL of NaOH (10 wt %) were mixed in a flask for 30 min under constant magnetic stirring (400 rpm) at room temperature. Then, 3 g of FeSO4·7H2O and 5 g of FeCl3·6H2O were both dissolved in 100 mL of deionized water and heated to 313 K, simultaneously stirring at 400 rpm in a heater/magnetic stirrer for 30 min. After that, 50 mL of lignin NaOH solution and KMnO4 aqueous solution (1.5 wt %) was slowly added into the binary iron salt solutions with stirring for 3 h without heating. Finally, the L-F/M mixture obtained was filtered through vacuum, washed with deionized water to neutral pH, and dried under 333 K.

Characterization of the L-F/M

The scanning electronic microscope JSM-6700F was used to observe the change in surface morphology between raw alkali lignin and the L-F/M after gold spraying. The Fourier transform-infrared spectra (FT-IR) were obtained on a VERTEX70 Bruker spectrometer using KBr pellets to investigate the possible interaction in the adsorption process. Vibration sample magnetometry (VSM) determination was acquired on Quantum Design PPMS-9 in measuring the magnetic property of the L-F/M. The N2 adsorption–desorption isotherms, pore diameter, and specific surface area of the L-F/M were examined with a Kubo-X1000 Gas Sorption System at 353 K based on the BET analysis and BJH method. The thermal stability of the L-F/M based on TG-DTA was carried out on a Shimadzu DTG-60 DTG analyzer at the ramping rate of 10 °C/min in the temperature range of 30–500 °C under a nitrogen atmosphere (50 mL min–1).

Adsorption Experiments

All batch adsorption experiments were carried out on a DSHZ-300 constant temperature shaking water bath with a shaking speed of 250 rpm in 50 mL flasks. A MB stock solution (1000 mg L–1) was prepared by dissolving 1.0 g of MB in 1.0 L of deionized water and the desired concentrations of MB solution were further obtained through dilution. The initial pH values of MB solution were adjusted by 0.1 mol L–1 HCl or NaOH solution. After adsorption experiments, the solution was separated through vacuum filtration with 0.45 μm hydrophilic membranes. Subsequently, the absorbance of the supernatant was analyzed on a UV–vis spectrophotometer (UV 2012) at a maximum absorption wavelength of 665 nm to determine the concentration of MB before and after adsorption. The adsorption capacity of the L-F/M at equilibrium (qe, mg g–1) or at any time (qt, mg g–1) and the efficiency of MB removal were calculated by the following equations:[14]where Co, Ce, and C (mg L–1) are the initial concentration, concentration at equilibrium, and concentration at time t of MB solution, respectively. V (L) is the volume of the MB solution and m (g) is the dosage of the L-F/M.

Effect of Adsorption Time

The effect of adsorption time on MB removal was examined over the contact time range of 0–330 min under different initial temperatures (298, 303, and 313 K) with all other conditions held constant (pH = 7.0, 0.1 g of L-F/M, and 100 mL of MB solution at a concentration of 200 mg L–1).

The effect of initial solution pH on MB removal was investigated at different pH intervals (3, 5, 7, 9, and 11) under different initial temperatures (298, 303, and 313 K) with all other conditions held constant (330 min, 0.1 g of L-F/M, and 100 mL of MB solution at a concentration of 200 mg L–1).

Effect of Temperature

The effect of temperature on the removal of MB was conducted using 100 mL of MB solution at different temperatures of 298, 303, and 313 K under various concentrations (200, 225, and 250 mg L–1) with all other conditions held constant (pH = 7.0, 0.1 g of L-F/M, and 330 min).

The L-F/M (0.1 g) was added into MB solution (100 mL of MB solution at a concentration of 200 mg L–1) to finish the first adsorption process under the conditions of 303 K, pH = 6.5, and 330 min. The L-F/M was regenerated by eluting with 95% ethanol solution under 313 K for 30 min. The adsorption and regeneration cycles were repeated five times to calculate the MB removal, RE (eq ), and SSE (eq ).[72]

Statistical Analysis

All the experiments were repeated in triplicate. The data are presented as mean ± standard deviation, as indicated in the tables and figures (error bars). The best-fitting models were determined through the coefficient of determination (R2) (eq ) and residual analysis. For linearized forms, the residual analysis was performed through residual sum of squares (eq ), while for nonlinearized forms, it was performed through reduced chi-sqr (eq ). All the statistical analyses were carried out with Origin 2019b.where y is the experimental data, f is fit of the data, and y̅ is the mean of experimental data.