Fe3O4 nanoparticle-coated mushroom source biomaterial for Cr(VI) polluted liquid treatment and mechanism research

Agrocybe cylindracea substrate–Fe3O4 (ACS–Fe3O4), a Fe3O4 nanoparticle-coated biomaterial derived from agriculture waste from mushroom cultivation, was developed to remove hexavalent chromium (Cr(VI)) from liquid. After modification, material surface became uneven with polyporous and crinkly structure which improved Cr-accommodation ability in a sound manner. Optimized by the Taguchi method, Cr(VI) removal percentage was up to 73.88 at 240 min, 40°C, pH 3, Cr(VI) concentration 200 mg l−1, dosage 12 g l−1, rpm 200. The efficient Cr(VI) removal was due to the combined effect of adsorption and redox. In addition, verification test using tannery wastewater, with removal percentage of Cr(VI) and total Cr reaching 98.35 and 95.6, provided further evidence for the efficiency and feasibility of ACS–Fe3O4. The effect of storage time of the material on Cr(VI) removal was small, which enhanced its value in practical application. Results indicated that metal removal was mainly influenced by solution concentration, adsorbent dosage and treatment time. The experimental data obtained were successfully fitted with the Langmuir isotherm model. Thermodynamic study indicated the endothermic nature of the process. The results confirmed that ACS–Fe3O4 as novel material derived from waste, with long-term stability, could be applied for heavy metal removal from wastewater and waste cycling.


Preparation of ACS-Fe 3 O 4
ACS-Fe 3 O 4 materials were prepared according to a published method [23]. FeCl 2 ·4H 2 O (3.0 g) and FeCl 3 ·6H 2 O (6.1 g) were dissolved in 100 ml of water. The mixture was heated to 90°C in a 250 ml round-bottom flask equipped with a reflux condenser. The reaction solution was magnetically stirred throughout the process. To the mixture, 10 ml of 25% ammonium hydroxide and 0.50 g ACS particle (for ACS-Fe 3 O 4 synthesis) were added rapidly and sequentially. The mixture was aged at 90 ± 5°C for an additional 30 min. The solid products were washed with water and dried to constant weight in a vacuum oven at approximately 40°C. The particles were stored in a vacuum desiccator. Pure Fe 3 O 4 was also prepared with the same method without ACS. The ability of pure Fe 3 O 4 and ACS-Fe 3 O 4 for Cr(VI) removal was also tested with the same condition mentioned in §2.1.

Chemicals and equipment
All the chemicals and reagents used were of analytical grade (Kelong Chemical Reagent Factory, Chengdu, China). Potassium dichromate was used as adsorbate. The stock adsorbate solutions (1000 mg l −1 and 5000 mg l −1 ) were prepared by dissolving 2.828 and 14.140 g of potassium dichromate in 1 l ultrapure water, respectively. All working solutions were obtained by dilution. Cr(VI) concentration was determined by a spectrophotometer according to Chinese National Standard GB/T 7467-87. The total Cr was determined by a flame atomic absorption spectrometer (AA700, Perkin-Elmer, USA). The Cr(III) content was the difference between total Cr and Cr(VI) in the solution.

Characterization of ACS-Fe 3 O 4
SEM (JSM-5900LV, Japan) was used to identify the surface morphology features of raw ACS and ACS-Fe 3 O 4 . The chemical elements on the surface of biomaterial and the main functional groups were analysed by EDX and FTIR spectroscopy (NEXUS-650, USA).

Taguchi experiment design
The Taguchi method had been widely used as a systematic approach to optimize the design parameters, which can minimize the overall testing time and the experimental costs [24,25]. Using the specially designed orthogonal array, the optimum experiment conditions can be determined. Accordingly, an analysis of the signal-to-noise (S/N) ratio was applied to evaluate the experimental results. This study was designed to acquire the optimized operational conditions for the maximum Cr(VI) removal percentage; therefore, the HB (higher is best) S/N ratio analysis was adopted.
The analysis of mean statistical approach was conducted to evaluate the optimal conditions. The details and equations for Taguchi experiment design and optimization study are shown in electronic supplementary material, table S1. 2.8. 1

. Sorption isotherms study
A series of Cr(VI) solutions with different initial concentrations were prepared (50, 100, 150, 200 and 300 mg l −1 ). The experiment conditions were: the dose 5 g l −1 , 24 h, 180 rpm. After equilibrium, q        (mg g −1 ), weight of Cr removed per unit of dry adsorbent weight, was calculated by where C 0 and C e (mg g −1 ) are the initial and equilibrium concentrations of Cr(VI), respectively. V (l) is the volume of the solution, W (g) is the mass of dry adsorbent [26]. Langmuir and Freundlich isotherm models were used to analyse the sorption equilibrium data [27].

Adsorption kinetics study
To investigate the mechanism and characteristics of the adsorption of Cr(VI), pseudo-first-order and pseudo-second-order models were used to test the data [28].

Scanning electron microscopy results
As shown in figure 2a, before modification, the material was bulky, and its surface was uneven and irregular. However, after modification, the particle was further broken, particle size shrank, and the surface became smoother because of the coating. The modification also provided ACS with nanostructure which was helpful for adsorption. The irregularities of ACS enhanced its ability to adsorb metal ions and also decreased the mass transfer resistance. After modification, the decrease of particle size increased its specific surface area which enhanced the biosorption efficiency.

Energy-dispersive X-ray analysis
The surface of ACS was composed of carbon (22.10 wt%) and oxygen (61.91 wt%) as well as a small amount of Mg, Ca and K before modification (figure 2b). After coating, Fe (76.18 wt%) became its main component, verifying the availability of modification. After the treatment, Cr was identified on the surface of ACS-Fe 3 O 4 , confirming the Cr adherence onto absorbent.

Fourier transform infrared analysis
The infrared spectra of ACS, ACS-Fe 3 O 4 and ACS-Fe 3 O 4 -Cr are demonstrated in figure 2c. The broad band observed between 3000 and 3700 cm −1 indicated the existence of -OH and -NH groups on both unloaded and Cr-loaded biomaterial. It has been reported [29] that biosorbents normally have intense absorption bands around 3200-3500 cm −1 . The spectra of Cr-loaded material also displayed absorption peaks at 3126 and 2925 cm −1 , corresponding to the stretching of C-H bonds of methyne and methylene groups [29,30]. The region between 1690 and 1500 cm −1 represented the C ≡ C stretching in aromatic rings [31]. The peaks observed at 1635 and 1501 cm −1 could be attributed to this vibration. The peak observed at 1033 cm −1 could be related to the vibration of C-OH in alcohol group and carboxyl [32].
The peak observed at 1395 cm −1 represented the vibration of -CH-(CH 3 ) [33]. The band at 522 cm −1 represents C-N-C scissoring that was only found in protein structures [29],

Results of single-factor experiments
The results of single-factor experiments are shown in figure 3. Figure 3a shows that when 0.2 g ACS-Fe 3 O 4 was suspended in a 50 ml Cr(VI) solution (200 mg l −1 , pH 7) at 180 rpm, 25°C, the adsorption capacity of ACS-Fe 3 O 4 at different initial concentrations increased fast at the beginning, then slowed down until reaching equilibrium. The adsorption process mainly happening at the start could be attributed to the instantaneous utilization of the most readily available active sites on the material surface. The saturation of the available binding sites slowed down the adsorption speed later. The increase of Cr(VI) concentration led to an increase in the biosorption uptake, so was the equilibrium time. The competition of amount of ions during the adsorption on biosorbent contributed to this increase. Unlike other experiments [28], in the current study, chemical reaction (Cr 6+ + 3Fe 2+ → Cr 3+ + 3Fe 3+ ) happened along with adsorption process which extended the equilibrium time. After the saturation of the available binding sites, chemical reaction of Cr(VI) and Fe(II) made the main contribution to Cr(VI) removal. The turning points, 120, 180 and 240 min, were chosen as factor levels in optimal experiment.
Profile for Cr(VI) removal when 0.2 g ACS-Fe 3 O 4 was suspended in a series of 50 ml Cr(VI) solutions (pH 7) at 25°C, 180 rpm for 24 h is shown in figure 3b. With the increase of Cr(VI) concentration from 20 to 1000 mg l −1 , biosorption capacity of ACS-Fe 3 O 4 increased from 2.46 to 40.08 mg g −1 and the removal percentage of Cr(VI) decreased from 98.28% to 7.94%. The interaction between Cr(VI) and biosorbent was improved because of the increase of Cr(VI) concentration, resulting in the improvement of adsorption capacity. However, limited active sites and Fe 3 O 4 amount on biosorbent surface restricted the biosorption capacity of biosorbent. And at higher ion concentration, the active sites and Fe 3 O 4 were saturated or exhausted which ended the process. The turning points 200, 400 and 600 mg l −1 were chosen as factor levels in optimal experiment. Figure 3c presents the effect of adsorbent dose with ACS-Fe 3 O 4 suspended in the 50 ml 200 mg l −1 Cr(VI) solution (pH 7) for 24 h at 180 rpm, 25°C. With the increase of dosage (2-24 g l −1 ), the removal percentage increased gradually with a slowing trend (from 20.37 to 95.07%). Similar to other reports [32,34], the increase of Cr(VI) removal was due to the increase of active sites and reaction substrate in biosorbent. Further increase of dose concentration did not contribute to the removal, indicating the excessive addition of biosorbent is uneconomical, which was why 8, 10, 12 g l −1 were chosen as factor levels in optimal experiment.
The effects of pH are shown in figure 3d with 0.2 g ACS-Fe 3 O 4 suspended in the 50 ml Cr(VI) solution (200 mg l −1 ) for 24 h at 180 rpm, 25°C. With the increase of pH (from 0 to 8), the removal percentage decreased (from 97.43 to 29.33). It has been confirmed that Cr(VI) removal decreased with pH increase [35,36]. The maximal adsorption efficiency happened when the pH was pretty low because solution pH changed the form of the chromium ion, protonation level and the surface charge of the adsorbent [37]. With neutral or alkaline pH, the main form of chromium ion was CrO 4 2− . However, Cr(VI) gradually became the main form with the decrease of pH (CrO 4 2− + 8H + → Cr 6+ + 4H 2 O), which increased the competition ability of Cr(VI) with dichromate ion or its hydroxide form, enhancing adsorption efficiency. However, extreme acid condition was not practical, hence acid (pH at 3), neutral (pH at 7) and alkaline (pH at 11) were chosen as factor levels in optimal experiment. It can be seen in figure 3e that removal percentage of Cr(VI) was enhanced (from 36.22 to 40.59) when rpm increased from 0 to 100. However, no significant difference was observed with further increase of rpm. The strengthening of vibration contributed to the contact of biosorbent and ions. And when this vibration reached a certain level, higher stirring rate contributed less to the adsorption. For better investigation of the effect of stirring rate on the adsorption process, three factor levels of 0, 100 and 200 rpm were introduced in optimal experiment.
The temperature presented similar effect as rpm (figure 3f ). Cr(VI) removal percentage was enhanced from 27.59 to 33.57 when temperature increased from 20 to 30°C, while no significant change happened when temperature increased from 30 to 40°C. The increase of temperature accelerated the molecular movement and contributed to the interaction of adsorbents and solution. However, after reaching a certain level, this acceleration weakened. For better investigation of temperature influence, three factor levels of 20, 30, 40°C were introduced in optimal experiment. 3.4

.1. Optimum conditions
According to the Taguchi method, Tests 1-27 were accomplished. The Cr(VI) removal efficiency and the S/N ratio at each test condition were determined. The value in bold in table 2 represents the maximum value of S/N ratio. In electronic supplementary material, table S3, the maximum value of (M) level=i factor=I among all six factors in three levels was bolded, which indicated the optimization condition for Cr(VI) removal. The optimum conditions were: 40°C, pH 3, Cr(VI) concentration 200 mg l −1 , adsorbent 12 g l −1 , rpm 200 and 240 min. Figure 4a shows the effect of solution pH on S/N ratio. With the lowest pH at level 1 (pH 3), S/N was a peak at 30.23. And the lowest S/N of 25.51 happened at level 3. With the enhancement of temperature, S/N ratio value increased (figure 4b), indicating that higher ambient temperature influenced the adsorption more.
It is well defined in figure 4c that the S/N ratio reached the peak at 29.05, when the dosage was 12 g l −1 . While the lowest S/N ratio emerged when the dosage was 8 g l −1 . This was because the increase of adsorbent dosage led to the increase of adsorption sites, improving pollutant removal efficiency.   When it came to Cr(VI) concentration (figure 4d), the highest S/N ratio value ( 31.30) occurred at 200 mg l −1 and decreased with its increase, indicating that the increase of metal concentration played weaker role in adsorption mainly because of the limited adsorption sites on biosorbent.
According to figure 4e, the highest S/N ratio value was 29.29 when treatment time was 240 min, and the lowest value was 25.68 with treatment time at 120 min. This result indicated that although the adsorption process mainly happened in the early phase as shown in figure 3a, it continued with time increasing. And the final equilibrium state prolonged with the increase of Cr(VI) concentration. At given condition in optimal study with Cr(VI) concentration between 200 and 600 mg l −1 , the equilibrium could be around 240 min where the highest S/N ratio value was reached. Therefore, the optimum treatment time should be 240 min with Cr(VI) concentration between 200 and 600 mg l −1 .
As shown in figure 4f, the highest S/N ratio value was 29.23 at 200 rpm, which presented no significant difference from 100 rpm, indicating that beyond a certain level, the increase in rpm had a weak influence on Cr(VI) removal.

Verification test
In real industry process, the parameters of the polluted liquid could be complex, and the concentration of Cr in wastewater could be different [38]. To verify the availability of optimal conditions and feasibility of ACS-Fe 3 O 4 in practice, a verification test was conducted using tannery wastewater. The detailed experiment process is presented in the electronic supplementary material and the result is shown in

Contribution of each factor
The results of SS F andR F k are shown in table 3 and electronic supplementary material, table S5, respectively. SS T , the total sum of squares, was 19110. 170. The variance of error, V E (99.521), was also obtained. In the end, the contribution ratio of each factor was determined and shown in table 3. Initial concentration of Cr(VI) was the most influential factor in the process, whose contribution ratio was 33.310%. Solution pH had second highest significant influence with contribution ratio at 13.010%. The stirring rate, dosage and treatment time posed minor influence on the process, and the contribution ratios were 9.952%, 5.879% and 3.861%, respectively. The temperature posed adverse but small (−0.384%) effect on treatment. However, in practical application, metal concentrations and solution pH could not be efficiently controlled without high cost in practical application, hence more attention should be paid to controllable factors like rpm or dose concentration. 3. 6. Mechanism study 3.6

.1. Biosorption isotherms
Adsorption isotherms provide important information that reveals the equilibrium relationship between the adsorbate concentration in the liquid phase and the solid phase at a constant temperature. Langmuir and Freundlich models, which correspond to homogeneous and heterogeneous adsorbent surfaces, respectively [34], were chosen to describe the equilibrium characteristics of this study. The average regression coefficients (r 2 ) of the Langmuir model (0.779-0.977) were higher than those of the Freundlich model (0.716-0.922) (table 4A), indicating the Langmuir model was more suitable to describe the sorption process, and monolayer adsorption occurred on a heterogeneous adsorbent surface. Moreover, the value of b was 0.0048 (0 < b < 1), which confirmed the favourable uptake of Cr(VI).

Kinetics of Cr(VI) biosorption
Two models (pseudo-first-order and pseudo-second-order) were used to test the data (table 4B). With a higher r 2 value, the pseudo-second-order model was better fitted to the data than the pseudo-first-order model. Moreover, its calculated adsorption capacity closely fitted the experimental data. Therefore, the present adsorption system followed a predominantly pseudo-second-order kinetics model. A similar result for the treatment of waste water has also been reported in other works [39,40].

Thermodynamics of biosorption
As shown in table 4C, the values of the free energy change ( G 0 ) were negative, indicating the feasibility and spontaneous nature of the adsorption process. The positive value of S 0 indicated the increased randomness at the solid/solution interface during the adsorption, suggesting the good affinity of Cr(VI) towards the adsorbent and significant changes occurred in the internal structure of the biosorbent during biosorption. Furthermore, the positive H 0 value confirmed that the adsorption was an endothermic process.