Magnetic carnosine-based metal-organic framework nanoparticles: fabrication, characterization and application as arsenic adsorbent

Graphical abstract

This study centers on the controllable synthesis, characterization, and application of a novel magnetic bio-metal-organic framework (Bio-MOF) for the adsorption and subsequent removal of arsenic from aqueous solutions. Zinc ions and carnosine (Car) were exploited to construct the Car-based MOF on the surface of magnetite (Fe₃O₄ NPs). The Magnetite precoating with Car led to an increase in the yield and the uniform formation of the magnetic MOF. The prepared magnetic Bio-MOF nanoparticles (Fe₃O₄-Car-MOF NPs) had semi-spherical shape with the size in the range of 35–77 nm, and the crystalline pattern of both magnetite and Car-based MOF. The NPs were employed as an adsorbent for arsenic (As) removal. The adsorption analyses revealed that all studied independent variables including pH, adsorbent dose, and initial arsenic concentration had a significant effect on the arsenic adsorption, and the adsorption data were well matched to the quadratic model. The predicted adsorption values were close to the experimental values confirming the validity of the suggested model. Furthermore, adsorbent dose and pH had a positive effect on arsenic removal, whereas arsenic concentration had a negative effect. The adsorption isotherm and kinetic studies both revealed that As adsorption fitted best to the Freundlich isotherm model. The maximum monolayer adsorption capacity (94.33 mg/g) was achieved at room temperature, pH of 8.5 and adsorbent dose of 0.4 g/L. Finally, the results demonstrated that the adsorbent could be efficiently applied for arsenic removal from aqueous environment.

Introduction

With the rapid development of industrialization, inorganic pollutants such as heavy metal ions from wastewater, even at a small scale, have been a serious threat to human health and the ecological systems [43]. One of the most deleterious and toxic heavy metals found in water is Arsenic (As). It is a poisoning element that is introduced as class I carcinogen for human by the international agency for research on cancer (IARC) [13, 18]. Naturally occurring arsenic in rocks and industrial arsenic contained in pesticides are the main sources of water contamination. Because of its carcinogenic character and grave health issues for humans, the World Health Organization (WHO) has recommended a standard value for drinking water of 10 μg/L [40].

Not only is arsenic naturally present in the soil and can contaminate surface or underground aquifers, but also anthropological activities such as draining arsenic-containing effluents may increase arsenic concentrations [6, 34]. For this purpose, several removal processes such as chemical precipitation, membrane techniques, ion exchange, adsorption, and electrodialysis (ED) have been applied for As removal from contaminated resources. Among various techniques, adsorption is considered as one of the most efficient processes for drinking water treatment, having advantages such as low cost and ease in operation. Several studies have been conducted to develop numerous natural or synthetic adsorbents for arsenic removal, and some of them are available commercially [7].

As a new class of crystalline inorganic-organic porous hybrid materials, metal-organic frameworks (MOFs) have emerged as a type of promising material for the development of novel adsorbents [20, 24]. These nanomaterials are constructed by multidentate organic ligands and transition metal ions or clusters [44]. Due to the exceptional internal surface area, tailored structure, tunable pore size, and shape of MOFs, the nanomaterials can be easily controlled to facilitate the uptake of a targeted guest molecule [22]. These valuable features of MOF inspired researchers to employ them as catalyst [15], support for gas storage and separation [36], carrier for drug delivery [14, 23], and adsorbent for contaminant removal [25]. However, MOFs suffer from low recyclability and reusability limiting their aplication as an adsorbent in contaminant removal.

Magnetic MOFs overcome the problem and offer practical and versatile process for separating and recycling the adsorbents using exterior magnetic fields [43]. Different strategies have been reported for the preparation of magnetic MOF including Layer-by-Layer (LbL) growth method, embedding method, encapsulation method, and some recently reported method like dry gel conversion strategy, self-template method, magnetic macroporous polymer substrate strategy, seed crystal method, and solvent-free mechanochemical approach [26]. Among different magnetic NPs (MNPs), magnetite (Fe₃O₄) is one of the well-known NPs widely employed in the design and preparation of various adsorbents for chemical pollutants removal [3, 4, 27, 8].

Up to now, a few pioneering studies have been reported related to the removal of heavy metal ions from water by magnetic MOF. Fei Ke et al. reported fabrication of thiol-functionalized porous magnetic MOF as an adsorbent for selective removal of Hg²⁺ and Pb²⁺ from wastewater [30]. In another study, R. Ricco et al. reported aluminum-based magnetic MOF for uptaking Lead (II) from groundwater [21].

Bio-MOFs, which are composed of biomolecules and biocompatible transition metals like zinc or iron, have attracted increasing attention as environmentally benign materials for the application in various branches of science [9, 31]. Peptides are safe and naturally-occurring compounds for the preparation of MOFs endowing the target adsorbent with appropriate biocompatibility, chemical diversity, high flexibility, and excellent adsorption ability for heavy metal removal [16, 28, 32]. L-Carnosine (Car) is a naturally occurring dipeptide located at a high level in nerve tissues and skeletal muscle as a neuropeptide bearing valuable functional groups with the high ability to bind with various metals [5, 11, 19]. These valuable features make Car as a promising multidentate organic bridge for the MOFs preparation and metal removal.

In the light of the facts mentioned above and in continuation of our previous studies on the contaminants removal [1, 37], in this work, a novel magnetic Bio-MOF was introduced for As removal. Fe₃O₄ NPs were employed as a core endowing the target adsorbent with easily separation and recycling. Car-based MOF was exploited as a shell to offer high surface area, appropriate functionality, and porosity to the adsorbent improving As removal. Car coated Fe₃O₄ NPs (Fe₃O₄-Car), and novel magnetic Bio-MOF containing Car and Zinc ions (Fe₃O₄-Car-MOF) were separately prepared and fully characterized. The performance of the designed adsorbents was also evaluated and compared in terms of As removal. To the best of our knowledge, this is the first report on the preparation of magnetic Car-based MOFs and their application for the removal of any contaminants.

Materials and methods

Materials

Iron (III) chloride dihydrate (FeCl₃.2H₂O), FeSO₄.4H₂O, L-carnosine (Car), zinc nitrate hexahydrate, sodium arsenate, and other materials and reagents were purchased from the Merck Company (Darmstadt, Germany) and used without any further purification.

Synthesis procedure

Car coated Fe₃O₄ MNPs were prepared according to the previously reported method [10]. An aqueous solution of FeSO₄.4H₂O (1.2 g) and FeCl₃.2H₂O (2.34 g) in 100 mL of water was prepared and kept at constant temperature of 80 °C for 30 min under vigorous stirring. 4 g of Car was added to the aqueous solution during 1 h. Then, aqueous ammonia solution was added dropwise to reach pH 11. A black suspension was formed and it was allowed to stir for 1.5 h. All steps were carried out under inert atmosphere and vigorous stirring. Car coated magnetite NPs (Fe₃O₄-Car NPs) were separated from the aqueous solution by magnetic decantation washed with distilled water several times and dried in an oven overnight. Non-magnetic Car-based MOF was prepared according to the reported method [19].

For the preparation of Fe₃O₄-Car-MOF NPs, aqueous solution of Zn (NO₃)₂.6H₂O (5.26 mL, 0.336 M), aqueous solution of Car (2 mL, 0.442 M), 0.4 g of well-dispersed Fe₃O₄-Car in 2 mL of water and 20 mL of DMF were all combined while stirring. The suspension was heated at 100 °C for 12 h with a ramping rate of 1 °C/min. It was then cooled down to room temperature with a ramping rate of 1 °C/min. The final product was collected by an external magnetic field, washed thoroughly with methanol, and finally dried. The schematic procedure for the preparation of Fe₃O₄-Car-MOF NPs is seen in Fig. 1.

Fig. 1.

Schematic illustration for the preparation of Fe₃O₄-Car-MOF NPs

Characterization

The morphology and size of the NPs were determined by field emission scanning electron microscopy (FE-SEM, MIRAIII Tescan). Sample elemental analysis was performed by energy dispersive spectrometry, EDAX (SAMX). The crystal structure of Fe₃O₄-Car-MOF NPs was determined by X-ray powder diffraction, XRD (PHILIPS PW1730 diffractometer using Cu Kα radiation). The FTIR spectra were recorded by a Fourier transform infrared spectrometer, FTIR (Perkin-Elmer) at the wavelength range of 4000–400 cm⁻¹ using the KBr pellet technique. Magnetic properties were analyzed using Vibrating Sample Magnetometer, VSM (Meghnatis Daghigh Kavir Co, Iran) with a maximum magnetic field of 10 kOe at 25 °C.

Adsorption test

Sodium arsenate was used for the preparation of arsenic solutions with different concentrations. In each run, a predetermined dose of adsorbent was poured in an Erlenmeyer flask containing a specific concentration of arsenic, and the solution was agitated on a shaker at a speed of 200 rpm. In all adsorption experiments, the contact time was 180 min, which was the optimum time for the adsorption according to the pretest. At the end of each run, the adsorbent was separated from the solution by an external magnet, and the samples were analyzed by an ICP for the detection of the residual arsenic. All runs were conducted at room temperature.

Isotherm experiments

Langmuir and Freundlich isotherm

The relation between the concentration of adsorbent in the solution and the removed As values were determined by the Langmuir and Freundlich isotherm.

The Langmuir isotherm model is centered on the monolayer distribution of solute molecules onto the surface of the adsorbent. The linear form of the Langmuir model is presented via Eq. (1) [35].

$$\frac{C_e}{q_e}=\frac{1}{K_a{q}_m}+\frac{C_e}{q_m}$$ 1

q_m, K_a and C_e (the Langmuir constants) representing the maximum adsorption capacity (mg/g), the adsorption energy (L/mg), and equilibrium concentration of As in solution (L/mg) were all calculated from the plot C_e/q_e versus C_e.

The R_L parameter is referred to the non-dimensional constant factor expressing the dependence of the adsorbed substance on the adsorbent and it was calculated using Eq. (2) [35].

$$R_L=\frac{1}{1+K_a{C}_0}$$ 2

Where, R_L is dimensionless constant, and C₀ is the initial concentration of As (mg/L).

Freundlich isotherm model evaluates the multilayer adsorption of adsorbate on the surface of adsorbent. It also assumes that the adsorption occurs on heterogeneous surfaces expressed via Eq. (3) [29].

$$log{q}_e=log{K}_f+\frac{1}{n}\left(log{C}_e\right)$$ 3

Where, K_f and n are Freundlich constants and q_e is adsorption capacity (mg/g). The adsorption intensity of the system was derived from the plot of lnq_e versus lnC_e.

Kinetic experiments

Adsorption kinetics of As was assessed by the experimental data [29]. To investigate the mechanism of the adsorption, the transient behavior of the As adsorption process was analyzed using the pseudo-first-order and pseudo-second-order kinetics, explained below:

Pseudo-first-order kinetic

The linear form of the pseudo-first-order kinetic model is expressed by Eq. (4) [29].

$$log\left(q_e-q_t\right)=log\left(q_e\right)-\left(\frac{k_1}{2.303}\right)t$$ 4

Where, q_e, q_t, and k₁ are the amount of adsorbed As at the equilibrium condition (mg/g), the amount of adsorbed As at a given time (t) and the equilibrium rate constant of the pseudo-first-order kinetic model (min⁻¹), respectively. k₁ was obtained from drawing log (q_e – q_t) versus (t).

Pseudo-second-order kinetic

The linear form of pseudo-second-order kinetic was obtained via Eq. (5).

$$\frac{t}{q_t}=\frac{1}{k_2{q_e}^2}+\frac{1}{q_e}t$$ 5

From the plot of (t/q_t) versus (t), the values of the pseudo-second-order rate constants (k₂) (g mg⁻¹ min⁻¹) and q_e (mg g⁻¹) was calculated.

The initial adsorption rate (h) was calculated at zero time, via Eq. (6):

$$h=k_2{q_e}^2$$ 6

Results and discussions

Fe₃O₄-Car and Fe₃O₄-Car-MOF NPs were separately synthesized, and their capacity for the As removal was compared. In our initial attempts to prepare magnetic Car-based MOF, naked Fe₃O₄ MNPs were treated with zinc ions and Car. However, the non-magnetic Bio-MOF was also formed during the synthesis, decreasing the yield of the target adsorbent. This could be due to the inappropriate affinity of zinc ions to the surface of the Fe₃O₄ MNPs resulting in the non-uniform formation of Bio-MOF. Therefore, in our next attempts, Fe₃O₄ MNPs were initially functionalized with Car to provide an ideal condition for zinc ions immobilization and Bio-MOF formation. To achieve this goal, Fe₃O₄ MNPs were prepared in the presence of Car to endow the surface of the MNPs with appropriate amine and imidazole groups as Zinc chelating agents. Finally, Fe₃O₄-Car NPs were converted to Fe₃O₄-Car-MOF through the hydrothermal formation of MOF by Car and Zinc ions as cationic metal and multidentate organic bridge, respectively (Fig. 1). The prepared NPs were characterized by various methods including FTIR, VSM, XRD, FESEM, EDAX, elemental mapping and TEM analyses. The efficacy of the system was investigated for the As removal from an aqueous environment.

FE-SEM analysis

The morphology and size of the NPs were investigated by FE-SEM and TEM analyses. Figure 2 shows FE-SEM images of the Fe₃O₄-Car and Fe₃O₄-Car-MOF NPs having a semi-spherical shape and almost monodisperse-size distribution. Fe₃O₄-Car NPs were in the range of 25–45 nm with some agglomeration. The increased size after coating with MOF (about 35–75 nm) could be due to the formation of the layer of Car-based MOF on the surface of Fe₃O₄ MNPs. As can be seen in Fig. 2, the separate particles with almost smooth surfaces confirm that the formation of MOF was accomplished. The TEM images of the NPs revealed agglomerated NPs which could be due to the high tendency of the magnetic NPs to each other due to the dipole-dipole interaction. However, separated spherical NPs with the size of less than 30 nm could be seen in the images. No significant change was observed after formation of MOF on the surface of Fe₃O₄-Car NPs (Fig. 2).

Fig. 2.

(a, c) FE-SEM images size distribution histograms determined by Image J 1.44p software (b, d) and TEM images (e, f) of Fe₃O₄-Car and Fe₃O₄-Car-MOF NPs, respectively

EDAX analysis

Figure 3 depicts FESEM-EDX elemental mapping of Fe₃O₄-Car-MOF NPs. There was no evidence indicating the presence of any impurity within the structure of Fe₃O₄-Car-MOF NPs. The pattern revealed that the NPs were composed of Fe, Zn, C, N, and O elements with the weight ratio of 34.5, 4.5, 17.9, 11.0, and 31.9%, respectively. Appropriate distribution of the elements confirmed successfully preparation of the NPs.

Fig. 3.

FESEM-EDX elemental mapping of Fe₃O₄-Car-MOF NPs

XRD analysis

The crystalline structure and the composition of Fe₃O₄-Car-MOF were also evaluated with single-crystal X-ray diffraction patterns (Fig. 4). The obtained patterns for Fe₃O₄-Car and Car-based MOF in bulk form (non-magnetic Car-based MOF) were in good agreement with the characterization previously reported elsewhere [19]. The X-ray diffraction pattern of Fe₃O₄-Car-MOF highlights high crystallinity and purity of the product with the specific peaks related to both cubic inverse spinel form of Fe₃O₄ and the MOF. The intensity of the peaks related to the cubic inverse spinel form of Fe₃O₄ NPs was somewhat decreased and broadening was also happened after MOF coating indicating the amorphous structure of the target NPs. The Debye–Scherrer equation was applied for calculating the average crystallite size of Fe₃O₄-Car-MOF:

$$D=\frac{0.9\lambda }{\mathit{\text{βCosθ}}}$$ 7

Fig. 4.

XRD patterns of the synthesized NPs and PXRD of the naked Fe₃O₄ and free MOF. Stars and circles show the essential peaks of the naked Fe₃O₄ and free MOF, respectively

In this formula, X-ray radiation wavelength (λ) is 0.154059 nm, β represents the corrected band broadening, and θ is referred to the Bragg angle [19]. The average crystalline size of Fe₃O₄-Car-MOF particles was calculated to be about 28 nm using the Debye–Scherrer equation.

The X-ray diffraction patterns of Fe₃O₄-Car demonstrated that the distinct peaks shown at 2θ of 30°, 36°, 43°, 54°, 57°, 63° are attributed to the (220), (311), (400) (422), (511), (440) planes of Fe₃O₄ which was also observed in the pattern of Fe₃O₄-Car-MOF. The essential crystalline plates of the Car-based MOF were also identified with a slight shift in the diffraction pattern of Fe₃O₄-Car-MOF reaffirming the successful formation of the MOF on the surface of Fe₃O₄ MNPs.

FTIR analysis

The FTIR spectroscopy was applied to determine the type of functional groups in the structure of Fe₃O₄-Car-MOF. Figure 5 shows the FT-IR spectra of Fe₃O₄-Car and Fe₃O₄-Car-MOF NPs. In the FTIR spectra of Fe₃O₄-Car and Fe₃O₄-Car-MOF, the peak appeared at about 580 cm⁻¹ is related to the stretching vibration of the Fe-O bond of Fe₃O₄ MNPs. The broad peak appeared at 3386 cm⁻¹ could be related to the hydroxyl groups of the MNPs which was probably overlapped by the stretching vibration of NH₂ group of Car. The bands attributed to the stretching vibrations of carboxylate bonds (COO⁻), which presumably interacted with the surface of Fe₃O₄ and C=O amide bonds, are shown at 1630 and 1617 cm⁻¹, respectively. The stretching frequencies of asymmetric and symmetric CH₂ groups of Car were also observed at 2848 and 2922 cm⁻¹ [10]. In the FTIR spectrum of Fe₃O₄-Car-MOF NPs, an apparent increase in the amplitude of the NH amide bond at 1622 cm⁻¹ and C=N bond of imidazole ring at 1560 cm⁻¹ could confirm the large presence of Car as well as the successful formation of the MOF. Also, the peak at 3300 cm⁻¹ could be probably due to the stretching frequencies of the N-H bond of Car. These results along with the results of EDAX and XRD analyses confirm the successful formation of the magnetic-based MOF.

Fig. 5.

FTIR spectra of Fe₃O₄-Car and Fe₃O₄-Car-MOF NPs

VSM analysis

The magnetic property of Fe₃O₄-Car-MOF NPs was measured using VSM at room temperature and compared with the naked Fe₃O₄ and Fe₃O₄-Car NPs (Fig. 6). The applied magnetic field was varied from −9 kOe to 9 kOe. The magnetization curve did not exhibit hysteresis and coercivity. This observation indicated that NPs had superparamagnetic behavior. The saturation magnetization values for Fe₃O₄-Car and Fe₃O₄-Car-MOF were found to be 33 and 28 emu/g at room temperature, respectively. The meaningful drop in the magnetization value of Fe₃O₄-Car-MOF could be attributed to the non-magnetic character of Zinc and Car.

Fig. 6.

Magnetization curves of Fe₃O₄, Fe₃O₄-Car and Fe₃O₄-Car-MOF NPs at room temperature

As removal

Experimental design

Initial attempts to evaluate the efficacy of the NPs revealed that Fe₃O₄-Car-MOF had higher potential for As removal than Fe₃O₄-Car (about 17% more than Fe₃O₄-Car). Therefore, magnetic bio-MOF was employed for the next studies. The effect of independent variables such as adsorbent dose (0.1–0.4 g/l), pH (3–11), and initial arsenic concentration (20–40 mg/L) on the arsenic adsorption was investigated in a batch system. Central composite design (CCD) under response surface methodology was applied for the determination of the number of runs, studying the interaction of parameters, and analyzing the experimental data [2, 17].

According to the central composite design, 29 experiments were selected for investigating the effect of three different variables on the arsenic adsorption. To analyze the results, Design-Expert software (version 7, trial, Stat-Ease) was utilized [35]. Table 1 shows the central composite design matrix for the arsenic adsorption and the adsorption data. As seen in Table 1, the predicted adsorption values are close to the experimental values confirming the validity of the quadratic model.

Table 1.

Central composite design matrix for arsenic adsorption

Run order	A	B	C	As removal (%)
	(Adsorbent dose)	(pH)	(As concentration)
				Experimental	Predicted
1	0	0	0	63	59.2
2	0	0	0	62	59.2
3	−0.5	0.5	0.5	35	37.3
4	0	0	0	53	59.2
5	−0.5	−0.5	0.5	22	17.7
6	−0.5	0.5	−0.5	48	56.2
7	0.5	0.5	−0.5	72	81.6
8	0.5	−0.5	0.5	56	53.1
9	0	0	0	67	59.2
10	0.5	−0.5	−0.5	71	73.9
11	0	0	0	60	59.2
12	−0.5	−0.5	−0.5	32	37.1
13	0	0	0	54	59.2
14	0.5	0.5	0.5	61	61.2
15	0	0	0	67	59.2
16	0	0	1	23	27.9
17	0	0	0	68	59.2
18	1	0	0	89	86.7
19	0	0	0	58	59.2
20	0	0	0	52	59.2
21	0	0	0	59	59.2
22	0	0	0	60	59.2
23	0	0	−1	78	67.7
24	0	−1	0	30	32.2
25	0	1	0	67	59.4
26	0	0	0	58	59.2
27	0	0	0	62	59.2
28	0	0	0	51	59.2
29	−1	0	0	29	25.9

Analysis of variance results

According to the data presented in Table 2, all studied independent variables have a significant effect (p value <0.05) on the arsenic removal, and the quadratic model is significant (R squared 0.8868, p value <0.0001, and lack of fit 0.0509) [39]. Based on the F value, the adsorbent dose is the most important parameter in the arsenic adsorption.

Table 2.

ANOVA results for arsenic removal using Fe₃O₄-Car-MOF

Source	Sum of squares	df	Mean square	F value	p value
Model	6678.84	9	742.09	16.55	<0.0001
A-Adsorbent dose	3690.56	1	3690.56	82.28	<0.0001
B-pH	742.56	1	742.56	16.56	0.0007
C-As concentration	1580.06	1	1580.06	35.23	<0.0001
AB	66.13	1	66.13	1.47	0.2395
AC	1.13	1	1.13	0.025	0.8758
BC	0.13	1	0.13	2.787E-003	0.9584
A2	16.59	1	16.59	0.37	0.5502
B2	350.40	1	350.40	7.81	0.0115
C2	253.74	1	253.74	5.66	0.0280
Residual	852.20	19	44.85
Lack of fit	436.60	5	87.32	2.94	0.0509
Pure error	415.60	14	29.69
Cor total	7531.03	28

R-Squared: 0.886, Adj R-Squared: 0.833

Based on the acquired regression coefficients, the arsenic removal efficiency can be computed by Eq. (8)

$$\begin{array}{l}\mathbf{As}\mathbf{\text{removal}}\left(\%\right)=+59.24+18.06\ast \mathrm{A}+8.10\ast \mathrm{B}-11.82\ast \mathrm{C}-4.07\ast \mathrm{A}\ast \mathrm{B}-\\ 0.53\ast \mathrm{A}\ast \mathrm{C}+0.18\ast \mathrm{B}\ast \mathrm{C}-1.03\ast {\mathrm{A}}^2-4.74\ast {\mathrm{B}}^2-4.04\ast {\mathrm{C}}^2\end{array}$$ 8

Where; A, B, and C are the coded levels of adsorbent dose, pH, and arsenic concentration, respectively. The regression data showed that adsorbent dose and pH had a positive effect on the arsenic adsorption, whereas initial arsenic concentration had a negative effect.

Effect of independent variables

The effect of independent variables on the arsenic removal is seen in Fig. 7. It is clear that higher As removal could be achieved through the rise in the adsorbent dose and pH values as well as the drop in the As concentration. In other words, when the adsorbent dose increases the number of available active sites for the adsorption of arsenic ions increases, which results in the improvement of arsenic adsorption [42]. However, with increasing the concentration of arsenic the number of active sites is not enough to remove all arsenic ions, and eventually, the adsorption efficiency declines. In a related study, a similar trend was reported for the effect of pH on the arsenic removal using octanuclear Zn (II)-based polymer [41]. The maximum monolayer adsorption capacity was 94.33 mg/g and the optimum condition was obtained at room temperature, pH of 8.5 and adsorbent dose of 0.4 g/L.

Fig. 7.

Three-dimensional plots: Arsenic removal efficiency as the function of independent parameters; (a) As concentration/pH; (b) As concentration/ adsorbent dose

Langmuir and Freundlich isotherm models

Figure 8 and Table 4 show the Langmuir isotherm and the concerning kinetic parameters of Fe₃O₄-Car-MOF NPs. A good coefficient of determination was achieved for the Langmuir isotherm (R² = 0.9866). q_m, K_a and R_L (the Langmuir constants) were determined to be 94.33, 2 and 0.013 mg/g, respectively. The R_L is referred to the relative volatility in vapour–liquid equilibrium. The amount of R_L determines the type of equilibrium as follows: 0 < R_L <1 refers to the favorable equilibrium, whilst R_L > 1 is attributed to the unfavorable isotherm [12].

Fig. 8.

The plots for the Langmuir and Freundlich isotherm models regarding the adsorption of As on Fe₃O₄-Car-MOF NPs

Table 4.

Pseudo-first-order and pseudo-second-order kinetic parameters regarding adsorption of As

Parameters	Qe mg g−1 (Cal.)	Qe mg g−1 (Exp.)	k1 (min−1)	k2 (g mg−1 min−1)	R2
Pseudo-first-order kinetic	12.28	39.5	0.011		0.208
Pseudo-second-order kinetic	38.91	39.5		0.002	0.942

The adsorption data were also analyzed using the Freundlich isotherm, and the results are presented in Table 3. K_f and n, known as the Freundlich constants, were calculated from the plot of lnq_e versus lnC_e, which were 12.3 mg g⁻¹ and 1.91, respectively. Also, the coefficient of determination (R²) for the Freundlich isotherm was 0.9987 (Table 3). The value of n was larger than one indicating a favorable adsorption system. With comparison of coefficient of determinations (R²) of the two models involved, it could be deducted that Freundlich model fitted best to describe the adsorption process in this study. In a similar related study, the same behavior was also observed for the methylene blue adsorption onto activated carbon prepared from rattan sawdust and jute fiber carbon [12, 33].

Table 3.

Freundlich and Langmuir isotherm parameters for adsorption of As

Parameters	Langmuir	Freundlich
KL (L/mg)	0.18	18.24
Qm	94.33	–
RL	0.34	–
R2	0.9866	0.9987
n	–	1.91

Kinetics of adsorption

In this study, the applicability of the pseudo-first-order and pseudo-second-order models was tested for the adsorption of As, and the results showed that the pseudo-second-order model provided a very good coefficient of determination (R²) (Fig. 9 and Table 4). The value of the calculated equilibrium adsorption capacity (q_e) from the pseudo-first-order model was significantly lower than the experimental q_e value indicating the inapplicability of the model. For the pseudo-second-order kinetic model, the high coefficient of determination (R²), and the closeness of the calculated q_e from the model to the experimental q_e value indicate that the model is highly applicable to describe the adsorption of As on the adsorbent [38]. These results are in line with the previously reported studies such as adsorption of As onto rice husk [38] and bamboo-based activated carbon [12].

Fig. 9.

(a) Pseudo-first-order, and (b) pseudo-second-order kinetic plots for the adsorption of As

Stability of the adsorbent

Stability of the Fe₃O₄-Car-MOF was evaluated in simulated media for As removal. A predetermined dose of adsorbent (0.4 g/L) was poured in an Erlenmeyer flask containing a specific concentration of As, and the solution was agitated on a shaker at a speed of 200 rpm for 180 min, pH 8.5 at room temperature, which were the optimum parameters for the adsorption according to the pretest. At the end of each run, the adsorbent was separated from the solution by an external magnet. Four cycles of regeneration studies were carried out. FTIR analysis of the fresh Fe₃O₄-Car-MOF and the adsorbent after 4 cycles were compared with each other. The results clearly suggested stability of the adsorbent without any significant changes in the essential functional groups of the NPs (Fig. 10).

Fig. 10.

FTIR spectra of Fe₃O₄-Car-MOF NPs and Fe₃O₄-Car-MOF NPs after 4 cycles

Conclusions

In this study, a simple, rapid, and cost-effective method with no need to template, surfactant and catalyst was introduced for the synthesis of a novel magnetic Bio-MOF (Fe₃O₄-Car-MOF NPs) employed for the arsenic adsorption from aqueous solution. Fully characterization of the NPs confirmed formation of nearly spherical NPs with the size in the range of 35–77 nm bearing both character of the cubic inverse spinel form of Fe₃O₄ and Car-based MOF and appropriate elemental distribution. According to the statistical results, all studied independent parameters had a significant effect on the arsenic adsorption (p value <0.05), and the regression data indicated that adsorbent dose and pH had a positive effect on the arsenic adsorption, whereas initial arsenic concentration had a negative effect. Based on the fisher value, the adsorbent dose was the most influential parameter in the arsenic adsorption. The experimental data were well fitted to the quadratic model, and Fe₃O₄-Car-MOF NPs were able to remove arsenic up to 89%. The optimum condition was obtained at room temperature, pH of 8.5 and adsorbent dose of 0.4 g/L resulted in the maximum monolayer adsorption capacity of 94.33 mg/g. The preliminary screening exhibited appropriate stability of the NPs after four cycles without any significant changes in the functional grooups of the NPs. However, more studies are suggested to evaluate reusability and stability of the adsorbent. Finally, it is assumed that the magnetic Bio-MOF could be successfully employed for the removal of other contaminants and its application might be potentially extended to further areas of interest such as drug delivery and medical imaging.

Compliance with ethical standards

Conflict of interest

The authors declare no conflict of interest.