3.1. Characterization of Fe3O4 NPs/CA
The FTIR spectra of Fe3O4/CA, and Fe3O4/CA after the adsorption of MB are presented in Fig. 2 (a, and b). The FTIR spectra of Fe3O4/CA exhibited intense peaks between 580 cm− 1 and 630 cm− 1, which are attributed to the stretching vibration mode associated with the Fe–O bonds in the crystal lattice of Fe3O4 (Waldron 1955). A band at 1629 cm− 1 and the broad band centered at 3435 cm− 1 are related to the presence of hydroxyl groups and assigned to OH–bending and OH–stretching, respectively (Patrikiadou et al. 2016). Other bands were observed in the spectra of Fe3O4/CA. The peak in Fig. 2a at 1562.34 cm− 1 refers to the C = C stretching vibration in the aromatic C–C bond. The peak at 1562.34 cm− 1 indicates that impregnation with Fe3O4 NPs led to an increase in the aromatic properties. The peaks at 1415.75 cm− 1 and 1327.03 cm− 1 are attributed to the bending and stretching vibrations of CH2 and C–N bonds of the Fe3O4 NPs (Yan et al. 2015 ; Dovbeshko et al. 2000)[29,30], while the peak between 1157.29 cm− 1 and 1114.86 cm− 1 is attributed to the stretching between C–O and O–H for –COOH (Ragab and El Nemr 2019). In addition, effective functional groups appeared at 1041.56 cm− 1 as a sulfoxide (S = O). From the relevant results in Fig. 2b, we investigated the shifts in the absorbance peaks after MB dye adsorption due to the electrostatic interactions between MB and adsorbent.
The vibrating sample magnetometry (VSM) hysteresis loop magnetization curve was produced using a magnetometer at 298°K. Figure 3 shows the Fe3O4/CA magnetization curve, confirming its superparamagnetic behavior. The obtained saturation magnetization value of 24.5 emu/g is in good agreement with the known values for magnetic nanoparticles (He and Gao 2010).
The XRD pattern of the Fe3O4/CA is illustrated in Fig. 4. The magnetite characteristic peak can be seen at 2Ɵ = 30.3°, 35.8°, 43.01°, 54.14°, 57.5°, and 62.6°, while the high intensity corresponds to the 2 Ɵ of 35.8°. These diffraction peaks are related to the corresponding indices of (220), (311), (400), (422), (511), and (440), and a comparison with JCPDS card No. 79–0418 indicates that the prepared Fe3O4/CA particles could be easily matched to Fe3O4. A Scherrer equation was applied to calculate the average crystallite size, as follows (Smilgies 2009):
where Dc is the crystalline diameter in nm; K is the Scherrer constant; \(\lambda\) is the wavelength of the X-ray radiation; 𝜷 is the FWHM of the diffraction peak; and Ɵ is the peak diffraction angle.
The average crystallite size was computed to be 12.5 nm.
Figure 5 shows the surface morphology of the Fe3O4 NPs and Fe3O4/CA. These indicate that the Fe3O4 NPs were predominantly spherical nanoparticles of approximately similar diameters (Fig. 5a). In contrast, Fig. 5b shows that Fe3O4/CA particles were exposed to some aggregation. In most cases, this agglomeration is caused by the lack of a capping agent in the plant extract used in the nanoparticle manufacturing (Mascolo et al. 2013) [34]; however, these agglomerations are commonly recognized as overlapping pieces held together by weak Van der Waals forces (Altman et al. 2005).
From the literature concerning manufacturing Fe3O4 using Rhus coriaria extract under comparable circumstances, (Chen et al. 2013) reported that the majority of Fe3O4 particles were large (5–50 m) and asymmetrical.
The chemical compositions of the samples were analyzed using the EDX technique. Figure 6 shows the EDX spectrum of Fe3O4 NPs after juice loading, which confirms the presence of iron elements in the compounds. The iron percent (72.1) and oxygen percent (19.9) were recorded in the total weight of the Fe3O4 NPs, while additional elements occurred in minor amounts.
Surface features (i.e., specific surface area and pore volume) are crucial elements that must be studied because they have a significant impact on the adsorption capacity. Table 3 provides the BET surface area and pore volume of the Fe3O4 and Fe3O4/CA. The tabulated results show that the surface area and pore volume of Fe3O4/CA were greater than those of Fe3O4. The increase in the surface area and pore volume can enhance the adsorbent dye removal efficiency (Moosavi et al. 2020).
Table 3
Fe3O4 and Fe3O4/CA textural properties
|
Fe3O4
|
Fe3O4/CA
|
Surface area (m2/g)
|
6.67
|
32.656
|
Pore volume (cm3/g)
|
0.00472
|
0.02658
|
3.2 Regression model equation
The experimental results of the % dye removal were modeled using RSM combined with a backward regression method, which was performed automatically in Design-Expert 7.0.0. The developed model for % dye removal used a second-order polynomial quadratic equation in terms of the coded factors, as follows:
% dye removal = + 59.63–4.22 * A + 18.66 * B + 11.60 * C + 6.13 * D -5.76 * A * B + 2.01 * A * D − 8.53 * B * C − 1.88 * B * D + 3.16 * C^2 (3)
where A is the initial dye concontration (mg/L), B pertains to the solution pH, C is the adsorbent dose (mg/L), and D denotes the contact time (min). In the above model, positive signs point to these variables exerting a positive effect on the % dye removal, whereas negative signs point to an antagonistic effect on the % dye removal. The positive value relating to the adsorbent dose (B), contact time (C), and solution pH had a positive effect, as these terms increased the % dye removal. Conversely, the negative value relating to the initial dye concontration (A) indicates that it exerted an antagonistic effect on the % dye removal of Fe3O4/CA.
To study the adequacy and significance of the current model, an ANOVA test was performed, with the results presented in Table 4. The p value was set at < 0.05, and the F-value was high, at 116.16. These results indicate that the second-order quadratic model was adequate for describing the MB removal process using Fe3O4/CA. Values of “Prob > F” less than 0.0500 indicate that the model terms are significant. In this case A, B, C, D, AB, AD, BC, BD, CD, and C2 are significant model terms. The accuracy of the above model was evaluated in terms of the regression coefficients (R2), adjusted R2, and predicted R2, which were 0.9812, 0.9728, and 0.9346, respectively.
Table 4
ANOVA results for the MB dye removal model
Source
|
Sum of
Squares
|
df
|
Mean
Square
|
F-value
|
p-value
Prob > F
|
|
Model
|
9668.99
|
9
|
1074.33
|
116.16
|
< 0.0001
|
significant
|
A: Initial con. (mg/L)
|
321.31
|
1
|
321.31
|
34.74
|
< 0.0001
|
|
B: pH
|
5928.76
|
1
|
5928.76
|
641.01
|
< 0.0001
|
|
C: Adsorbent dose (mg/L)
|
1071.29
|
1
|
1071.29
|
115.83
|
< 0.0001
|
|
D: Time (min)
|
676.02
|
1
|
676.02
|
73.09
|
< 0.0001
|
|
AB
|
531.65
|
1
|
531.65
|
57.48
|
< 0.0001
|
|
AD
|
64.60
|
1
|
64.60
|
6.98
|
0.0156
|
|
BC
|
775.76
|
1
|
775.76
|
83.87
|
< 0.0001
|
|
BD
|
56.29
|
1
|
56.29
|
6.09
|
0.0228
|
|
C^2
|
31.99
|
1
|
31.99
|
3.46
|
0.0777
|
|
Residual
|
184.98
|
20
|
9.25
|
|
|
|
Lack of Fit
|
184.98
|
15
|
12.33
|
|
|
|
Pure Error
|
0.000
|
5
|
0.000
|
|
|
|
Cor. Total
|
9853.97
|
29
|
|
|
|
|
Std. Dev.
|
3.04
|
R-Squared
|
0.9812
|
Mean
|
63.13
|
Adj. R-Squared
|
0.9728
|
C.V. %
|
4.82
|
Pred. R-Squared
|
0.9346
|
PRESS
|
644.30
|
Adeq. Precision
|
43.126
|
The actual verses the predicted values of the % dye removal are plotted in Fig. 7, supporting a good agreement between the observed and predicted values, with similar findings by (AbdulRazak et al. 2018).
From the ANOVA results, the percentage contributions (PCs) of each independent variable for the % dye removal were calculated using Eq. 4 (Shakor et al. 2022), with the results presented in Table 5. The solution pH (B) showed the highest influence on the % dye removal as its contribution was 62.7%, followed by the adsorbent dose (C) at 11.3%, interaction (BC) at 8.2%, contact time (D) at 7.1%, interaction (AB) at 5.6%, and initial dye concontration (A) at 3.4%. The interaction terms (AD), (BD), and (C2) had the lowest effects on the % dye removal.
$$PC\%=\frac{SS}{\sum SS}x 100$$
4
where SS is the sum of the squares of these terms.
Table 5
PC% of the terms in the current model
Terms
|
A
|
B
|
C
|
D
|
AB
|
AD
|
BC
|
BD
|
C2
|
PC%
|
3.4
|
62.7
|
11.3
|
7.1
|
5.6
|
0.7
|
8.2
|
0.6
|
0.3
|
3.3 One factor plot
Design-Expert 7.0.0 software was used to evaluate the influence of each operation variable on the MB dye removal, as shown in Figs. 8–11.
Figure 8 shows the influence of the initial dye concontration on the % dye removal at a constant pH (B) of 9, adsorbent dose (C) of 200 mg/L, and time (D) of 89.75 min. The % dye removal decreased from 89.8 to 73.8% as the initial dye concontration increased from 10 to 50 mg/L. This is because at lower initial dye concentrations, a higher surface area was obtained due to the smaller number of MB molecules. Conversely, at higher dye concentrations, a large number of dye molecules interacted with the accessible adsorption sites. This is in keeping with results from the literature concerning MB removal (Majid et al. 2019).
Figure 9 shows the effect of the solution pH on the % dye removal, with a constant initial dye concentration (A) of 38.19 mg/L, adsorbent dose (C) of 200 mg/L, and time (D) of 89.75 min. The % dye removal increased from 38.9 to 78.55% as the solution pH increased from 3 to 9. This occurred because of the influence of the solution’s pH on the Fe3O4/CA surfaces. At a pH of 9, the negatively charged surface of the Fe3O4/CA adsorbed the most cationic dye (MB) due to an increase in the electrostatic attraction between the negative surface charge and the MB dye (Majid et al. 2019), while, at a pH of 3, the protons vied with the MB dyes to adsorb onto the Fe3O4/CA surfaces; as a result, the % dye removal decreased in this acidic medium.
Figure 10 displays the influence of the adsorbent dose on the % dye removal at an initial dye concentration (A) of 38.19 mg/L, pH (B) of 9, and time (D) of 89.75 min. The % dye removal increased from 78.55 to 85.46% as the adsorbent dose increased from 200 to 1000 mg/L. This was because as the adsorbent dose increased, the surface area expanded, generating a greater number of active sites, thereby increasing the probability that the adsorbate would find active sites (Al-Dahri et al. 2022). The slope being close to zero confirmed that the adsorbent dose had little influence on the responses.
Figure 11 illustrates the influence of the contact time on the % dye removal at a constant initial dye concentration (A) of 38.19 mg/L, pH (B) of 9, and adsorbent dose (C) of 200 mg/L. The % dye removal increased from 68.44 to 78.6% as the time increased from 30 to 90 min. The results indicate the positive effect of the contact time on the % dye removal.
3.4 Combined influence of the variables on the MB dye removal
This study investigated the influence of the initial dye concontration, pH of the solution, adsorbent dose, and contact time of the Fe3O4/CA on the MB dye removal as well as their joint influences.
Figure 12 illustrates the interaction between the initial dye concontration and pH at a constant contact time of 89.75 min and adsorbent dose of 200 mg/L. The MB removal from the aqueous solution increased as the pH of the solution increased from 3 to 9 across the entire initial MB concentration range (up to 50 mg/L). This can be explained by the influence of the solution’s pH on the adsorbent surface. Figure 13 shows the zeta potential of Fe3O4 at varying pH levels. The positive values of the Fe3O4 zeta potential changed to negative values = 7.6 (PHpzc). At a pH of 9, the negatively charged surface of the adsorbent most significantly affected the adsorption of the MB dyes due to the electrostatic attraction between the negative surface charge and the cationic dye (Borah et al. 2015). In contrast, at a pH of 3, the protons competed with the MB dyes in trying to adsorb onto the available surfaces; as a result, the percentage dye removal decreased under these acidic conditions. These results confirmed those reported in earlier studies (Anirudhan and Ramachandran 2015; Hamdy et al. 2018), which showed an increase in cationic dye adsorption with a rise in the pH level.
Figure 14 shows the interactive effect of the initial dye concontration and contact time at a constant adsorbent dose and solution pH of Fe3O4/CA on the MB dye removal. It is obvious that the % MB dye removal decreased as the MB dye concentration increased. This indicates that the adsorption active sites were saturated on the Fe3O4/CA surfaces. Under the conditions of the same initial MB concentration and different contact times, an increase in the contact time from 30 to 90 min caused an increase in the % removal of the MB dye. Thus, it can be concluded that there were more active adsorbent sites on the Fe3O4/CA surfaces that were available and unoccupied to adsorb the MB molecules (Alsaiari et al. 2021; Yang et al. 2019). As can be seen in Fig. 14, the lines are not elliptical, which indicates that there is no interaction between the contact time and initial dye concentration.
Figure 15 represents the effect of the solution’s pH and adsorbent dosage at a fixed initial dye concontration of 38.19 mg/L and contact time of 89.75 min. It clearly shows that the adsorbent dosage had positive effects on the % dye removal, whereas the removal efficiency increased with increases in the adsorbent dosage across the entire pH range (3–9) of the solution. This may be attributed to the fact that the uptake was higher due to the availability of more active binding sites (Zhang et al. 2012).
As can be seen in Fig. 15, the lines are elliptical, which indicates that there is an interaction between the adsorbent dosage and pH of the solution.
The response surface plot for the effect of the pH and contact time on the removal efficiency at a constant initial dye concontration of 38.19 mg/L and adsorbent dose of 200 mg/L is shown in Fig. 16. As the pH of the solution increased, the % dye removal increased with the contact time from 30 to 90 min. This was due to the electrostatic interaction between the positively charged dye and the negatively charged adsorbent surface; the adsorption process worked best under these circumstances (Igwegbe et al. 2019).
3.5 Optimization and model validation
To optimize the adsorption process, target criteria were set to maximize the % MB removal, with four experimental parameters (i.e., initial dye concontration, solution pH, adsorbent dose, and contact time) chosen in the ranges considered in this study.
The optimal solution was based on a desirability value of 1. Under optimal conditions, the best maximum % dye removal of 93.14 was found to be at the initial dye concontration of 10.02 mg/L, the solution pH of 8.98, adsorbent dose of 997.99 mg/L, and contact time of 43.7 min (Fig. 17). To validate the model’s predicted results, additional experiments were conducted at the optimal experimental conditions in three replicates. The average MB removal efficiency of these three replicates was found to be 94%, which agrees with the predicted MB removal of 93.14%. It indicates the applicability of using a central composite design (CCD) to optimize and evaluate MB dye removal by Fe3O4/CA.
3.6 Mechanism of adsorption
The adsorption process is facilitated by various factors, including high surface area, high surface charge, and the presence of compatible functional groups. Of the total number of phenolic compounds, C. aurantium juice has been reported to contain 86% phenolic acids, A number of phenolic compounds are present in the Citrus aurantium juice (e.g., gallic, sinapic acid, dihydroxyphenilic acid, dihydrobenzoic acid, chlorogenic, vanillic, syringic, p-coumaric, ferulic, rosmarinic, trans-2-dihydrocinnamic, and cinnamic acids) (Tounsi et al. 2011 ). According to the FTIR spectra of Fe3O4/CA, as shown in Fig. 2b, a band at 1629 cm− 1 and the broad band centered at 3435 cm− 1 are related to the presence of hydroxyl groups and attributed to OH–bending and OH–stretching, respectively. The peaks between 1157.29 cm− 1 and 1114.86 cm− 1 are assigned to the stretching of between C–O and O–H for –COOH. Therefore, the phenolic and carboxylic groups played key roles in the adsorption mechanism.
In addition, the mechanisms of MB removal by the Fe3O4/CA adsorbent are illustrated in Fig. 18. Hydrogen bonding between the –OH groups of the Fe3O4/CA and an aromatic heterocyclic ring of the MB was a major contributor in capturing the MB. Van der Waals forces between the phenolic ring of the adsorbent and aromatic heterocyclic ring of the MB may also have contributed significantly.
The batch study data indicated that Fe3O4/CA exhibited better MB removal efficacy with a higher solution pH. This is possibly due to electrostatic interactions because at a higher pH, the surface of the Fe3O4/CA tends to be negatively charged and to interact favorably with the positively charged MB dye. Thus, it can be said that the MB molecules were captured from water through electrostatic interaction, hydrogen bonding, and Van der Waals forces.
3.7 Influence of temperature
The influence of temperature on the MB dye removal was studied at 25, 30, and 35°C, under the following conditions: the adsorbent dose of 998 mg/L, initial MB dye concentration of 10.02 mg/ L, time of 43.7 min, and a solution pH of 8.98. The results indicated that the temperature had a small influence on the MB dye removal. For example, the dye removal efficiency increased from 93 to 95.49% by increasing the solution temperature from 20 to 35°C (Fig. 19). This result demonstrated that the adsorption of MB on Fe3O4/CA was an endothermic process because at higher temperatures the dye molecules diffused into the adsorbent. (Alguacil and López 2021) reported a similar observation, that the adsorption was an endothermic and spontaneous process.
3.8 Recycled adsorption
The cyclic adsorption performance of Fe3O4/CA was studied in this work. Upon completion of the MB removal, the Fe3O4/CA particles were separated from the reaction mixture, washed with distilled water, and reused for the removal of MB dye. As shown in Fig. 20, the % dye removal of Fe3O4/CA after five rounds of cyclic adsorption accounted for 93.14, 92, 91, 88, and 80.00% at the following conditions: pH = 8.98, adsorbent dose = 998 mg/L, initial concentration = 10.02 mg/L, and time = 43.7 min. Thus, the evidence indicates that Fe304NPs/CA is stable and reusable in removing MB dye from wastewater.
3.9 Comparative study of MB dye removal with other adsorbents
Table 6 shows the % dye removal of MB by Fe3O4/CA compared with other adsorbents through the statistical experimental design at optimum conditions. It is evident that Fe3O4/CA is an efficient adsorbent for the removal of MB dye.
Table 6
Comparison of the % dye removal of MB reported in extant studies with that of Fe3O4/CA
Adsorbents
|
Optimum conditions
|
% MB dye removal
|
Ref.
|
Activated carbon (AC) from grape leaves
|
pH (11), adsorbent dosage (12.5 g/L), MB initial concentration (100 mg/L), contact time (90 min)
|
97.4
|
(Mousavi et al. 2022)
|
Spruce sawdust (SD) coated by magnesium oxide (MgO)
|
pH (11), adsorbent dosage (3.50 g/L)
|
94.05
|
(Sharii, and Shoja 2018)
|
Cellulose dusts (CD)
|
pH (9.84), adsorbent dosage (4.38 g/L), MB concentration (75.50 mg/L), contact time (208.13 min)
|
98.05
|
(Pajaie et al. 2018)
|
Agricultural waste, cashew nut shell (CNS)
|
pH (10), adsorbent dose (2.1846 g/L), initial dye concentration (50 mg/L), contact time (62.8693 min)
|
100
|
(Subramaniam, and Ponnusamy 2015)
|
Banana leaf ash (BLA)
|
Adsorbent dose (23.9 mg/100 ml), shaking time (3 h), shaking speed (356 rpm)
|
93.75
|
(Zahanggir et al. 2022)
|
Zeolites 13X modified by magnetic nanoparticles
|
pH (8.93), adsorbent dose (1198 mg/L), temperature (53.39°C), initial MB concentration (10.05 mg/L)
|
96
|
(Majid et al. 2019)
|
Fe3O4/CA
|
pH (8.98), adsorbent dose (997.99 mg/L), initial dye concentration (10.02 mg/L), contact time (43.7 min)
|
93.14
|
This work
|
Fe3O4
|
72.2
|
This work
|