Adsorption of ionic liquid from aqueous solutions using functional corncob-cellulose nanocrystals

Abstract

Here we report novel adsorbents, based on corncob-cellulose nanocrystals, for the efficient absorption of ionic liquid (1-butyl-3-methylimidazolium chloride, [Bmim]Cl) from aqueous solutions. Diethylenetriaminepentaacetic acid (DTPA) and sulfosalicylic acid (SSA) were introduced onto the cellulose nanocrystals to improve their adsorption efficiencies. The maximum [Bmim]Cl adsorption capacities of the corncob cellulose nanocrystal–GMA–DTPA adsorbent (CNGD) and the corncob cellulose nanocrystal–GMA–SSA adsorbent (CNGS) reached 0.473 mmol g⁻¹ and 0.499 mmol g⁻¹, respectively, which are higher than those of many other reported adsorbents. Moreover, the monolayer adsorption of [Bmim]Cl onto the functional cellulose nanocrystal adsorbents was revealed by a better fitting of the Langmuir model. Adsorption kinetics studies indicated that the adsorption behavior of [Bmim]Cl followed a pseudo-second-order kinetics model. Furthermore, the reusability performances and cycling behaviors confirmed the promising potential of the novel biosorbents for the removal of [Bmim]Cl from aqueous environments.

---

1. Introduction

Within the last few decades, ionic liquids (ILs) have been in the spotlight of the scientific and industrial community, due to their exceptional physicochemical properties, such as low vapor pressure, high chemical and thermal stability, wide liquid window and high solvent capacity.^1,2 Another remarkable advantage of ILs is that their scaffolds can be conveniently tailored by adequate manipulation of the cation and anion pairs, which expands their possible applications in areas including organic, physical, and analytical chemistry, biology, and biomass processing.^3–5 Although they are characterized as “green solvents” due to their negligible vapor pressures, extensive evidence has indicated that they have harmful effects on living organisms such as duckweed, algae, daphnia, and zebrafish.^6,7 In addition, when released into aquatic environments, it is hard to degrade them under the natural environmental conditions, due to their relatively high stabilities. Therefore, the application of ILs on an industrial scale may lead to contaminated wastewater streams, therefore the removal or recovery of ILs has to be considered as a significant topic, in order to prevent their release into the aquatic environment.

To date, several physicochemical and biological strategies, including thermal degradation,⁸ chemical oxidation,⁹ photocatalysis,¹⁰ biodegradation¹¹ and adsorption¹² have been used for eliminating ILs from wastewater.

Among these techniques, adsorption is favored due to its eco-friendly, cost-effective and non-destructive characteristics. Thus, developing promising adsorbents with high adsorption capacities is of great significance, as this would help deal with the challenge of the treatment of IL pollution in the environment. Recently, cellulose nanocrystals (CNCs), a one-dimensional nanosized material that can be inexhaustibly obtained from a variety of highly available and renewable cellulose-rich sources, have drawn a tremendous level of attention from the materials community, owing to their high aspect ratio, high crystallinity, low density, nanoscale dimension and unique morphology.¹³ With high specific surface areas and abundant surface hydroxyl groups, CNCs have also emerged as promising adsorbents for water remediation.^14,15 Moreover, CNCs display high reactivity and are well suited for scaffold fabrication of functional adsorbents, which expands their potential applications in water purification and removal of pollutants.¹⁶ Furthermore, in many studies, CNCs have been proven to possess excellent biodegradability and a low ecotoxicological risk.^17,18 Due to these characteristics, CNC-based adsorbents have been found to work effectively in the removal of hazardous materials, such as heavy metals,¹⁹ dyes^20,21 and pesticides.²² Nevertheless, the applications of functional cellulose nanocrystals for the treatment of ILs have not been investigated previously.

The objective of this work is to evaluate the efficiency of functional corncob-cellulose nanocrystals in the removal of ionic liquids from aqueous solutions. Different types of functional adsorbents based on corncob cellulose nanocrystals were prepared via modification with diethylenetriaminepentaacetic acid and sulfosalicylic acid, respectively. Their adsorptive behaviors towards [Bmim]Cl, a typical hydrophilic ionic liquid, were investigated systematically. Sorption kinetics and isotherms were measured to obtain in-depth information associated with the adsorption processes. Finally, comparisons of the maximum adsorption capabilities of different adsorbents were also performed.

2. Materials and methods

2.1. Materials

The corncob used in this study was obtained from a farm in Kaifeng City, Henan Province, China. The corncob feedstock was washed thoroughly with deionized water for the removal of dust and surface impurities, then dried and milled to pass through a 60-mesh sieve. 1-Butyl-3-methyl imidazolium chloride ([Bmim]Cl (99%)) was purchased from Henan Lihua pharmaceutical Co., Ltd. (Xinxiang). Glycidyl methacrylate (>99%, GMA) was obtained from Shangqiu Shengyuan Industrial Assistant Co., Ltd. (Henan, China). SSA and DTPA were both provided by the Enterprise Group of Chemical Reagent Co., Ltd. (Henan, China). The chemicals used for the syntheses were of reagent grade and were commercially available. Deionized water was used throughout the experiments.

2.2. Synthesis of the biosorbents

2.2.1. Preparation of CNCs from corncob. CNCs were isolated from agricultural waste corncob via the following procedure: 20.0 g of corncob power was extracted with toluene/ethanol (2 : 1, v/v) using a soxhlet extractor for 20 h to remove wax, as well as phenolics, pigments and oils. The treated powder (15.0 g) was then delignified with 1.4% NaClO₂ solution (300 mL, pH = 3.5) by heating at 70 °C for 5 h, and then washed with deionized water, ethanol and acetone. The delignified sample was added into 200 mL of 5% KOH solution at room temperature for 24 h and then heated at 90 °C for 2 h. The pure corncob cellulose was obtained, after a freeze-drying process under vacuum, for further synthesis. After that, 10.0 g of the pure corncob cellulose and 85 mL of 64 wt% sulfuric acid were mixed at 45 °C for 30 min with vigorous mechanical stirring. 300 mL of deionized water was then added into the above suspension to stop the previous reaction. Excess sulphuric acid was removed via repetitive centrifugation with copious amounts of water (3000 rpm, 20 min per cycle) until the supernatant became turbid. The CNCs were obtained by repeated dialysis and freeze-dried for further usage.

2.2.2. Epoxidation of the CNCs. A CNC suspension (0.03 g mL⁻¹, 100 mL) was stirred and sparged with a slow stream of nitrogen gas in a round bottom flask for 30 min to remove traces of air trapped in the CNCs. 0.1 mol L⁻¹ nitric acid was added until the pH of the mixture was adjusted to 1.4. Then 30 mL of GMA was dropped into the above solution in batches over 30 min. The suspension was stirred continuously for an additional hour at room temperature, followed by centrifuging and washing to neutral pH. Finally, the polymer was extracted with acetone using a soxhlet extractor and dried under reduced pressure at 30 °C for 5 h to obtain epoxy cellulose nanocrystals (CNG).²³

2.2.3. Surface modification of CNG with DTPA. CNGD was synthesized by grafting DTPA onto CNG via a ring-opening reaction. A one-pot synthesis of CNGD was conducted. Briefly, CNG (1.6 g), DTPA (2.5 g), NaOH (3.0 g) and H₂O (100 mL) were mixed in a flask. The mixture was kept at 110 °C for 24 h with constant magnetic stirring in a nitrogen gas atmosphere. The obtained CNGD was washed with an excess of deionized water via centrifugation to remove any residual diethylenetriamine pentaacetic acid and NaOH, then dried in an oven at 80 °C for 12 h.

2.2.4. Functionalization of CNG with SSA. CNGS was prepared according to the following modified method, which has been reported in our previous work.²³ Briefly, CNG (2.0 g) was suspended in deionized water (100 mL) with constant magnetic stirring for 30 min at room temperature under a nitrogen gas atmosphere. After that, 2.6 g of SSA and 0.6 g of NaOH were added to the suspension and allowed to react with CNG for 6 h, with the temperature rising to 85 °C. The precipitates were collected and thoroughly washed with deionized water via repeat centrifugation. Finally, the CNGS was dried overnight in an oven at 80 °C for further usage.

The structures of the two functional cellulose nanocrystals are shown in Fig. 1. The detailed information of the characterization methods used is given in the ESI.†

Fig. 1 The structures of CNGD and CNGS.

2.3. Adsorption/desorption experiments

All adsorption experiments were carried out by adding 20.0 mg of adsorbent to 50 mL of polypropylene in centrifuge tubes containing 20 mL of [Bmim]Cl aqueous solution, and were all performed in triplicate. The effect of pH on adsorption was evaluated by changing the initial pH value from 3.0–9.0 (0.1 mol L⁻¹ NaOH and/or HCl solution was used to adjust the pH). Adsorption kinetics studies were conducted with an initial [Bmim]Cl concentration of 0.2 mmol g⁻¹ at 25 °C and the independent samples were measured at time intervals ranging from 0.5–25 h. The adsorption isotherm experiments were performed with various initial [Bmim]Cl concentrations (C₀ = 0.05–5 mmol L⁻¹) at different temperatures (25, 45 and 65 °C) for 24 h. After each adsorption experiment, the precipitate was separated by centrifugation and the [Bmim]Cl concentration in the filtrate was determined with a UV-Vis spectrophotometer (TU-1900, Beijing Purkinje General Instrument. Co. Ltd. Beijing, China) at 211 nm. The equilibrium adsorption capacity (q_e, mmol g⁻¹) of [Bmim]Cl was calculated using eqn (S1) in ESI.†

In the desorption experiments, HCl aqueous solution (0.1 mol L⁻¹) and ethanol were selected as the eluents. After the adsorption experiments, the spent adsorbents were recovered by centrifugation, and dried. Then, 20 mL of eluent was added and ultrasonicated for 2 h. The final precipitant was dried for re-use in the above adsorption experiments. This same step was repeated five times.

3. Results and discussion

3.1. Characterization of CNGD and CNGS

The surface functionalization results were supported by FT-IR spectroscopy analyses (Fig. 2a). The FT-IR peaks of the CNCs prepared from corncob were completely consistent with the characteristic peaks of pure cellulose.¹⁵ A broad peak between 3200–3600 cm⁻¹ was attributed to –OH stretching, the peak at 2905 cm⁻¹ was assigned to the C–H vibration of the –CH₂ groups, and the peak for the C–O antisymmetric bridge stretching appeared at 1163 cm⁻¹.¹⁹ In the FT-IR spectrum of CNG, the appearance of the characteristic peaks of an epoxy group (756 cm⁻¹, 842 cm⁻¹, 910 cm⁻¹), and a carbonyl group of an ester group (1742 cm⁻¹) suggested that GMA was grafted onto the CNCs successfully.²³ In the spectra of CNGD and CNGS, the characteristic peaks of the epoxy group disappeared. The FTIR spectrum of CNGS shows a characteristic peak at 1170 cm⁻¹, corresponding to the S–O stretching vibration of the –SO₃H groups. These observations provide evidence that the adsorbents CNGD and CNGS were successfully prepared.

Fig. 2 (a) FTIR spectra and (b) X-ray diffraction patterns of CNCs, CNG, CNGD and CNGS.

The surface chemical modifications were further substantiated by employing X-ray diffraction analysis (Fig. 2b). The CNCs’ diffractogram displays three distinct peaks at around 14.7°, 22.5°, and 34.4°, which are characteristic of cellulose I. These results suggest that the corn cellulose does not undergo any substantial structural changes during the sulfuric acid hydrolysis process.^24,25 In comparison with the CNCs, the XRD diffraction pattern of CNG exhibits a broad peak at 2θ = 18.4°, and the intensity of the peak at 2θ = 22.5° diminishes noticeably. CNGD and CNGS present similar diffraction profiles to CNG. Then, crystallinity indexes (CrI) were calculated to reveal the changes of the crystalline structure, using the Segal method, as shown in the following equation:

where I₂₀₀ is the overall intensity for the crystalline portion near 2θ = 22.5°, and I_am is the intensity of the baseline for the amorphous portion at about 2θ = 18°. The crystallinities of CNGD and CNGS are 24.7% and 23.2%, respectively. These show a significant decline in comparison with CNG (CrI = 57.7%), which suggests that the crystalline transformation occurred in the modification process.

The morphology of the CNCs used for modification was investigated by using a field emission scanning electron microscope (SEM) and an atomic force microscope (AFM). Fig. 3a and b show the AFM and SEM images of the CNCs, respectively, revealing the rod-like shapes. The AFM image indicates that the CNCs suspension is uniformly dispersed, and the thicknesses of these crystals were measured to be about 25 nm. From the SEM image (Fig. 3b), it was easily found that the CNCs undergo self-assembly during the freeze-drying process, and the homogeneous structure of the CNCs is about 500 nm in length. When modified with DTPA and SSA, the obtained CNGD and CNGS display similar morphologies (Fig. 3c and d). The functional cellulose nanocrystal particulates are closely packed with a block structure, which might be attributed to the intermolecular force of hydrogen bonding interactions and the introduction of the –COO and –SO₃H groups. In addition, EDS was coupled with SEM to examine the surface elemental compositions of the functional cellulose nanocrystals. The EDS analyses are shown in Fig. S1.† The elemental composition reveals that only the elements carbon and oxygen are present in the surface of the CNCs but the CNGD and CNGS present a small amount of Na and S, respectively. These findings indicate the DTPA and SSA were introduced to the cellulose nanocrystals successfully, and are consistent with the findings of the FT-IR analysis.

Fig. 3 (a) AFM image of CNCs aqueous suspensions; SEM images of (b) CNCs, (c) CNGD and (d) CNGS powders.

3.2. Determination of oxygen-containing functional group content

The oxygen-containing functional group content of CNGD and CNGS was determined by the reported method (eqn (S2)).²³ As shown in Table 1, the sulfonic acid groups and carboxyl groups of CNGS possess the same value (1.99 mmol g⁻¹) and the content of carboxyl groups in CNGD is 6.20 mmol g⁻¹. Moreover, it can be seen that the total content of oxygen-containing functional groups in CNGD (6.20 mmol g⁻¹) is much higher than that of CNGS (3.98 mmol g⁻¹).

Table 1 Oxygen-containing functional group content of CNGD and CNGS

Adsorbent	Total content (mmol g^−1)	–COOH (mmol g^−1)	–SO3H (mmol g^−1)
CNGD	6.20	6.20	—
CNGS	3.98	1.99	1.99

---

3.3. Effect of pH on [Bmim]Cl adsorption

Solution pH is an important factor which influences the adsorption properties of adsorbents during the adsorption process.²⁶ The effects of solution pH on [Bmim]Cl removal by CNGD and CNGS were studied and the results are presented in Fig. 4a. The removal rate of [Bmim]Cl by CNGD and CNGS increased rapidly in the pH range of 3.0–6.0 and the maximum [Bmim]Cl removal rate was achieved at pH 6.0, which indicates that the electrostatic attraction between [Bmim]Cl cations and the negatively charged sites of the adsorbents is the major driving force in the adsorption process. At a lower pH, the competitive adsorption between the protons and [Bmim]Cl decreases the [Bmim]Cl adsorption onto the adsorbents. Furthermore, as the solution pH increases, successive deprotonation of the adsorbents greatly enhances the electrostatic attraction between the positively charged [Bmim]Cl cations and the negatively charged adsorbent surfaces, resulting in a proportional increase in adsorption.^12,27 When the pH was increased from 6.0 to 10.0, the adsorption capacity of the two adsorbents was kept almost constant. These results show that CNGD and CNGS work effectively over a wide range of pH.

Fig. 4 (a) Effect of pH on [Bmim]Cl adsorption; (b) pseudo-second-order kinetic linear fit for [Bmim]Cl onto CNGD and CNGS; (c) adsorption isotherms of [Bmim]Cl onto CNGD at different temperatures; (d) adsorption isotherms of [Bmim]Cl onto CNGS at different temperatures.

3.4. Adsorption kinetics

The adsorption of [Bmim]Cl from aqueous solutions onto the surfaces of CNGD and CNGS, as a function of contact time, is shown in Fig. 4b (the inset picture). It is apparent that the adsorption process can be divided into two stages for CNGD and CNGS. The [Bmim]Cl adsorption rates were rapid during the initial reaction stage (0.5–5 h) and then gradually decreased, reaching adsorption equilibrium at approximately 17 h for CNGD and 21 h for CNGS. The fast adsorption stage of [Bmim]Cl could be attributed to the fact that an abundance of vacant adsorption sites are available initially. With the advancement of the adsorption process, many adsorption sites were occupied by [Bmim]Cl cations and the concentration of available adsorption sites was lower, resulting in a gradual decrease in adsorption rate at higher conversion. To ensure full equilibrium, we conducted the following adsorption experiments with a contact time of 24 h. Two kinetics models (i.e., pseudo-first-order and pseudo-second-order models) were employed to evaluate the [Bmim]Cl adsorption mechanisms of CNGD and CNGS. Their linearized rate equations are expressed as eqn (S3) and (S4).

The kinetic parameters and correlation coefficients are listed in Table 2. The data show an excellent fit to the pseudo-second-order model with a high R² (>0.99) value. A good fit can be further confirmed by the fact that the calculated q_e values of CNGD (0.101) and CNGS (0.089) are very close to the measured ones (0.105, and 0.091, respectively). The fitting curves of the pseudo-second-order rate model (Fig. 4b) indicate that chemical interactions were possibly involved in the adsorption process. Similar kinetics study results were also observed for the [Bmim]Cl adsorption for many adsorbents, e.g. carbonaceous materials.^27,28

Table 2 The kinetic parameters of the pseudo-first-order and pseudo-second-order models for [Bimim]Cl adsorption onto CNGD and CNGS

Adsorbent	qe (exp) (mmol g^−1)	Pseudo-second-order model			Pseudo-first-order model
		k2 (g mmol^−1 h^−1)	qe (cacl) (mmol g^−1)	R^2	k1 (h^−1)	qe (cacl) (mmol g^−1)	R^2
CNGD	0.101	9.56	0.105	0.997	0.013	0.873	0.905
CNGS	0.089	5.33	0.091	0.999	0.005	0.072	0.827

---

3.5. Adsorption isotherms for [Bmim]Cl removal

The adsorbed amounts of [Bmim]Cl on CNGD and CNGS at different temperatures (25 °C, 45 °C and 65 °C) were investigated. As shown in Fig. 4c and d, when increasing the initial [Bmim]Cl concentration, the q_e values accordingly rise up and then approach maximal values for both adsorbents. This is due to the enhancement of the driving force between the adsorbents and [Bmim]Cl when the initial concentration increases. For CNGD, the maximal values increase with the temperature changing from 25 °C to 65 °C. However, a reverse tendency can be seen for CNGS, where the highest maximal adsorption capacity is achieved at 25 °C. These results indicate that temperature is an important factor influencing the adsorption capacities of CNGD and CNGS for removal of [Bmim]Cl.

In order to determine the mechanism of [Bmim]Cl adsorption onto the adsorbents, adsorption isotherms of [Bmim]Cl onto CNGD and CNGS were assessed at 25 °C, 45 °C and 65 °C. The two most commonly used isotherm models (i.e., Langmuir and Freundlich) were employed to fit the [Bmim]Cl removal process by CNGD and CNGS, and are described as eqn (S5) and (S6) in ESI.†

The fitting results of the Langmuir and Freundlich isotherms are shown in Table 3. According to the correlation coefficient values (R²), it is confirmed that the [Bmim]Cl adsorption isotherm by CNGD and CNGS fits better with the Langmuir model, which suggests that [Bmim]Cl adsorption by functional cellulose nanocrystals is limited by the monolayer coverage, and that the adsorption sites on the surfaces of CNGD and CNGS are evenly distributed.¹⁵ The calculated maximum [Bmim]Cl adsorption capacities of CNGD and CNGS, as derived from the Langmuir model, are 0.473 mmol g⁻¹ and 0.499 mmol g⁻¹, respectively. Compared with the adsorbed amount of CNGD, CNGS exhibits a more favorable adsorption performance with a lower total of oxygen-containing functional groups, which could be attributed to the fact that the surface functional groups of the adsorbents are the pivotal driving force for high [Bmim]Cl removal, and that the π–π stacking interactions between CNGS and [Bmim]Cl²⁹ contribute to the whole binding strength, thus greatly improving the adsorption capacity of CNGS. With the temperature rising from 25 °C to 65 °C, the adsorption capacity of [Bmim]Cl on CNGD increased constantly, while the adsorption capacity of CNGS decreased. Furthermore, in order to evaluate the adsorption capacities of the functional cellulose nanocrystal adsorbents for the treatment of [Bmim]Cl, the adsorption capacities of many other reported adsorbents are listed in Table 4. It can be seen that CNGD and CNGS show greater adsorption capabilities in comparison with other adsorbents, including exchange resins and carbonaceous materials, which reveals the advantage of CNGD and CNGS as potential adsorbents for ionic liquid cleanup.

Table 3 Parameters of the adsorption isotherm models

Adsorbent	Temp. (°C)	Langmuir			Freundlich
		qm (mmol g^−1)	KL (L mmol^−1)	R^2	KF	n	R^2
CNGD	25	0.341	5.77	0.995	0.245	0.260	0.738
	45	0.370	6.55	0.999	0.276	0.240	0.813
	65	0.473	4.08	0.998	0.288	0.288	0.924
CNGS	25	0.499	2.35	0.998	0.294	0.322	0.819
	45	0.375	1.78	0.993	0.214	0.335	0.911
	65	0.258	1.04	0.999	0.115	0.441	0.926

---

Table 4 Comparison of maximum adsorption capacities of various sorbents for [Bmim]Cl

Adsorbents	qmax (mmol g^−1)	Raw materials	Adsorption time	Equilibrium time	Cycle number (desorption rate)	Reference
Forest soil	0.108	Soil	24 h	None	None	30
CSCM	0.520	Corn stalk	2 h	1 min	None	27
FCM	0.171	Cellulose	24 h	120 min	Three (73–75%)	28
D113	0.20	Commercial resin	5 h	90 min	None	31
Chinese AC	0.230	Activated carbon	24 h	None	None	32
a-CM	0.322	Cellulose	24 h	None	None	33
CNGD	0.473	Corncob	24 h	17 h	Five (90.4–94.3%)	Present study
CNGS	0.499	Corncob	24 h	21 h	Five (66.9–70.8%)	Present study

---

3.6. Desorption and regeneration

Desorption studies are some of the most fundamental tools used to evaluate the reusability of adsorbents. The desorption abilities of HCl aqueous solution and ethanol were investigated. It can be seen from Fig. S2† that ethanol is a more efficient eluent for the desorption of [Bmim]Cl from CNGD and CNGS. Therefore ethanol was chosen as the eluent for the cycling adsorption–desorption study which was conducted. As shown in Fig. 5, CNGD and CNGS demonstrate encouraging cycling behaviors. After five adsorption/desorption cycles, the adsorption capacity of [Bmim]Cl onto the recycled CNGD still remained at 0.091 mmol g⁻¹, which was only reduced by 3.19% compared with that of the first cycle (0.105 mmol g⁻¹). The slight decrease in the adsorption efficiencies of CNGD and CNGS may be attributed to a progressive saturation of the active sites of the adsorbents. These results indicate that CNGD and CNGS could be regenerated effectively and reused for adsorption of [Bmim]Cl, with good adsorption performance, at least five times.

Fig. 5 Adsorption–desorption cycles of CNGD and CNGS for [Bmim]Cl.

4. Conclusions

In this study, two novel absorbents based on corncob-cellulose nanocrystals were designed and prepared via modification of diethylenetriaminepentaacetic acid and sulfosalicylic acid, and exhibited excellent absorption performances for [Bmim]Cl from aqueous solutions. SEM-EDS, FT-IR and XRD analyses verified the effective surface functionalization of the cellulose nanocrystals. The adsorption capacities increased with increasing initial [Bmim]Cl concentration and contact time. The kinetics of adsorption followed the pseudo-second-order model, and the adsorption isotherm studies indicated that the Langmuir model could be used to describe the experimental data, with high adsorption capacities of 0.473 mmol g⁻¹ for CNGD and 0.499 mmol g⁻¹ for CNGS, respectively. CNGD and CNGS demonstrated encouraging reusability performances and cycling behaviors using ethanol as the eluent. These findings suggest that functional cellulose nanocrystals are promising adsorbents for the removal of ionic liquids, thus reducing the risk of their persistency in water environments.

Acknowledgements

The authors are grateful to the National Natural Science Foundation of China (21576071), and the International Science and Technology Cooperation Project of Henan Province (152102410023).

References

1. K. Dong, S. Zhang and J. Wang, Chem. Commun., 2016, 52, 6744–6764.
2. S. Zhang, N. Sun, X. He, X. Lu and X. Zhang, J. Phys. Chem. Ref. Data, 2006, 35, 1475–1517.
3. M. Imanari, K. Fujii, T. Mukai, N. Mizushima, H. Seki and K. Nishikawa, Phys. Chem. Chem. Phys., 2015, 17, 8750–8757.
4. D. R. MacFarlane, N. Tachikawa, M. Forsyth, J. M. Pringle, P. C. Howlett, G. D. Elliott, J. H. Davis, M. Watanabe, P. Simon and C. A. Angell, Energy Environ. Sci., 2014, 7, 232–250.
5. Y. Zhou, J. Liu, M. Xiao, Y. Meng and L. Sun, ACS Appl. Mater. Interfaces, 2016, 8, 5547–5555.
6. Y. Fan, X. Dong, L. Yan, D. Li, S. Hua, C. Hu and C. Pan, Chemosphere, 2016, 148, 241–247.
7. T. P. T. Pham, C. W. Cho and Y. S. Yun, Water Res., 2010, 44, 352–372.
8. E. Thomas, D. Thomas, K. P. Vijayalakshmi and B. K. George, RSC Adv., 2016, 6, 9421–9428.
9. E. M. Siedlecka, M. Gołębiowski, Z. Kaczyński, J. Czupryniak, T. Ossowski and P. Stepnowski, Appl. Catal., B, 2009, 91, 573–579.
10. T. Itakura, K. Hirata, M. Aoki, R. Sasai, H. Yoshida and H. Itoh, Environ. Chem. Lett., 2009, 7, 343–345.
11. J. Neumann, C.-W. Cho, S. Steudte, J. Köser, M. Uerdingen, J. Thöming and S. Stolte, Green Chem., 2012, 14, 410–418.
12. Y. Liu, L. Qiao, Y. Xiang and R. Guo, Langmuir, 2016, 32, 2582–2590.
13. C. Moreau, A. Villares, I. Capron and B. Cathala, Ind. Crops Prod., 2016, 93, 96–107.
14. R. Batmaz, N. Mohammed, M. Zaman, G. Minhas, R. M. Berry and K. C. Tam, Cellulose, 2014, 21, 1655–1665.
15. H. Qiao, Y. Zhou, F. Yu, E. Wang, Y. Min, Q. Huang, L. Pang and T. Ma, Chemosphere, 2015, 141, 297–303.
16. Y. Habibi, Chem. Soc. Rev., 2014, 43, 1519–1542.
17. L. Chen, J. Y. Zhu, C. Baez, P. Kitin and T. Elder, Green Chem., 2016, 18, 3835–3843.
18. R. M. A. Domingues, M. E. Gomes and R. L. Reis, Biomacromolecules, 2014, 15, 2327–2346.
19. P. Liu, P. F. Borrell, M. Bozic, V. Kokol, K. Oksman and A. P. Mathew, J. Hazard. Mater., 2015, 294, 177–185.
20. Z. Karim, A. P. Mathew, M. Grahn, J. Mouzon and K. Oksman, Carbohydr. Polym., 2014, 112, 668–676.
21. C. Zhou, Q. Wu, T. Lei and I. I. Negulescu, Chem. Eng. J., 2014, 251, 17–24.
22. W. Zhu, W. Ma, C. Li, J. Pan and X. Dai, Chem. Eng. J., 2015, 276, 249–260.
23. Y. Zhou, M. Zhang, X. Wang, Q. Huang, Y. Min, T. Ma and J. Niu, Ind. Eng. Chem. Res., 2014, 53, 5498–5506.
24. A. D. French, Cellulose, 2014, 21, 885–896.
25. Q. Lu, L. Tang, S. Wang, B. Huang, Y. Chen and X. Chen, Biomass Bioenergy, 2014, 70, 267–272.
26. L. Liu, J. P. Xie, Y. J. Li, Q. Zhang and J. M. Yao, Cellulose, 2016, 23, 723–736.
27. Y. Wang, F. Shen and X. Qi, RSC Adv., 2016, 6, 32505–32513.
28. X. Qi, L. Li, T. Tan, W. Chen and R. L. Smith, Environ. Sci. Technol., 2013, 47, 2792–2798.
29. K. Y. A. Lin and H. A. Chang, Chemosphere, 2015, 139, 624–631.
30. P. Stepnowski, W. Mrozik and J. Nichthauser, Environ. Sci. Technol., 2007, 41, 511–516.
31. L. Li, Y. Wang and X. Qi, RSC Adv., 2015, 5, 41352–41358.
32. A. Farooq, L. Reinert, J.-M. Levêque, N. Papaiconomou, N. Irfan and L. Duclaux, Microporous Mesoporous Mater., 2012, 158, 55–63.
33. X. Qi, L. Li, Y. Wang, N. Liu and R. L. Smith, Chem. Eng. J., 2014, 256, 407–414.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra22918b

---