Removal of cationic dyes from a synthetic effluent using a calcium surfactant

Modern industrial operations use different types of dyes in routine processes, and industrial effluents containing excess dyes are harmful to the environment and human health.

In the environment, dyes, which are difficult to degrade, can interfere with natural processes by blocking the passage of light through the water, inhibiting photosynthesis, and affecting the growth of microorganisms and aquatic biota (Crini & Badot 2008; Anirudhan & Ramachandran 2015). In the case of human health, some dyes may have mutagenic and/or carcinogenic properties (Li et al. 2017; Mayani et al. 2017).

Methylene blue (MB) is a heterocyclic, cationic dye of the phenothiazine class used in the pharmaceutical, textile, plastic, cosmetics, leather, paper, and food industries (Peydayesh & Rahbar-Kelishami 2015). MB can cause increased pulse rate, shock, emesis, jaundice, cyanosis, and tissue necrosis (Zhou et al. 2015; Zhu et al. 2015).

Malachite green (MG) is a cationic dye of the triphenylmethane class, commonly used for dyeing wool, leather and paper, and as a fungicide, parasiticide and bactericide in aquaculture (Oyelude et al. 2018). MG is toxic and can accumulate in the tissues of animals that ingest water contaminated with it. It can cause the destruction of the liver, kidneys and intestine and is carcinogenic (Zhang et al. 2017).

Therefore, the efficient removal of dyes from effluents has been a hot research topic. Several types of processes are being studied for the treatment of effluents contaminated by these cationic dyes, namely precipitation (Lee et al. 2011), coagulation–flocculation (Yang et al. 2016), ion exchange (Sharma et al. 2016), photocatalysis (Khatri & Rana 2020), biological degradation (Sun et al. 2021), membrane filtration (Benosmane et al. 2022), and adsorption.

Adsorption is one of the most promising methods due to its simplicity, flexibility, efficiency and ability to sequester many compounds (Peydayesh & Rahbar-Kelishami 2015). Several types of adsorbents have been reported to be effective in removing MB and MG, such as kaolin (Mouni et al. 2018), biomass ash (Novais et al. 2018), nanochitosan (Salamat et al. 2019), composite hydrogel (Verma et al. 2020), activated carbon (Mariana et al. 2021), and biochar (Giri et al. 2022).

Surfactants are amphiphilic substances, widely used in effluent treatment (Petcu et al. 2016; Mortada 2020; Melo et al. 2021), and anionic surfactants the most widely used. The most common anionic surfactant is obtained by saponification of oils and/or fats and a strong base, such as sodium hydroxide (NaOH) or potassium hydroxide (KOH). This type of surfactant precipitates in the presence of calcium ions, forming calcium surfactant (CaSF). CaSF has a high affinity for organic compounds, being able to adsorb dyes.

CaSFs applied in effluent treatment have already been studied in coagulation–flocculation (Yang et al. 2016), micellar solubilization, and ionic flocculation (Melo et al. 2021; Teixeira et al. 2022a, 2022b). However, CaSF use as an adsorbent is novel; therefore, there are many aspects to investigate.

This study proposed to treat two synthetic effluents containing cationic dyes of different classes (MB and MG) as model pollutants. The study evaluated the dye removal efficiency using a CaSF obtained by means of an anionic surfactant originated from frying soybean oil, measuring the process kinetics, mechanism, and adsorption balance, as well as the morphology, functional groups and elemental composition of the CaSF. An attractive factor for the use of this CaSF, in addition to being easily obtained at a low cost, is its green and sustainable footprint.

The CaSF used in this study was synthesized from an anionic surfactant and Ca²⁺ ions in an aqueous medium. The anionic surfactant was produced by saponification between frying soybean oil and NaOH (85.14 and 14.86% by mass, respectively). This ratio was based on the saponification index of the oil (154 g of oil and 26.88 g of NaOH). A 0.0625 M calcium chloride (CaCl₂) solution was used to provide the Ca²⁺ ions needed.

In a 5 L beaker, 20 g of the anionic surfactant was dissolved in 3.20 L of deionized water using a magnetic stirrer (100 rpm). After complete dissolution, 0.8 L of the 0.065 M CaCl₂ solution was added, and the stirring was reduced to 20 rpm for the formation of CaSF. CaSF flocs were filtered and oven-dried (50 °C, 6 h). After drying, the CaSF was macerated and sieved.

The CaSF surface characteristics, morphology and composition were analyzed by scanning electron microscopy (SEM) and energy-dispersive spectroscopy (EDS) using a scanning electron microscope (FEI, Inspect S50).

The CaSF was analyzed before and after dye adsorption by Fourier Transform Infrared Spectroscopy (FTIR) (PerkinElmer, Spectrum Two) to identify surface functional groups. The recorded spectra were in the range of 4,000–450 cm⁻¹ at 4 cm⁻¹ resolution using the potassium bromide (KBr) pellet method.

The point of zero charge (pH_PZC) was determined using the eleven-point methodology (Robles & Regalbuto 2004). A volume of 20 mL of these solutions was transferred to Erlenmeyer flasks containing approximately 20 mg of CaSF. The samples were shaken for a period of 6 h on a vibrating table (Nova Ethics, model 109). The pH of the solutions was measured and a graph of the final pH vs. initial pH was plotted to determine the pH_PZC.

The optimal values to promote dyes removal are shown in Table 1.

The dosing effect was studied using varying concentrations of CaSF (1, 2, 4, 5, 8, 10, 15, 20 and 25 g·L⁻¹) interacting with a 50 mg·L⁻¹ MB or MG solution under optimal conditions (Table 1).

The stirring effect was studied at 0, 20, 50, 80 and 100 rpm under optimal conditions (Table 1). MB and MG initial concentrations were both 50 mg·L⁻¹.

The adsorption kinetics were analyzed using 0–300 min intervals under optimal conditions (Table 1). MB and MG initial concentrations were both 50 mg·L⁻¹.

In the equilibrium study, samples containing dye concentrations of 10, 50, 100, 200, 300, 400 and 500 mg·L⁻¹ were treated under optimal conditions (Table 1). Equilibrium was evaluated at temperatures of 20, 30 and 40 °C.

The value of E enables the assessment of the nature of the process, whether it is physical adsorption (E < 8 kJ·mol⁻¹), ion exchange (8 < E < 16 kJ·mol⁻¹), or chemical adsorption (E > 16 kJ·mol⁻¹) (Kaveeshwar et al. 2018).

Figure 1 shows that CaSF is amorphous. The anionic surfactant inhibits the formation of the crystalline phase of CaSF, while the nature of such inhibition depends on the concentration of the anionic surfactant in the solution (Bujan et al. 2001; Wei et al. 2004). As the concentration exceeded the critical micellar concentration, the anionic surfactant–Ca²⁺ interaction promoted miniature crystallization of CaSF, resulting in the formation of a few trigonal-shaped crystals (Figure 1).

According to the EDS results, CaSF is mainly composed of C (70.8%), Ca (17.1%) and O (10%) (Figure 2(a)). Figure 2(b) and 2(c) show that CaSF is able to adsorb both MB and MG, as there was an increase in C after adsorption (to 75.20 and 80.80%, respectively). The amount of Ca dropped in both cases, strongly indicating that the interaction between CaSF and the dyes results in the redissolution of Ca. Similar results were obtained in the removal of MB by flocculation–precipitation (Yang et al. 2016).

The FTIR spectra (Figure 3) show some absorption peaks belonging to different functional groups, which was expected as CaSF is a complex compound. The peak at 3,417 cm⁻¹ is typical of the elongation of the OH⁻ functional group and indicates the presence of hydrogen bonds. The peak at 3,011 cm⁻¹ is typical of simple unsaturated olefinic compounds (C = C). The peaks at 2,921–2,850 cm⁻¹ correspond to the stretching of linear long-chain aliphatic compounds (C–H). The relative high intensity of these peaks points to the fact that CaSF has no ramifications in its molecular structure (Mirghani et al. 2002; Nandiyanto et al. 2019).

The peak at 1,693 cm⁻¹ indicates alkenyl stretch (C = C), while the sequence of peaks at 1,578–1,541 and 1,432–1,420 cm⁻¹ suggests the presence of carbonyl (C = O) groups in the form of carboxylate or carboxylic acid salt. The peak at 1,113 cm⁻¹ indicates the presence of ester (COO). Finally, the peaks at 722–675 cm⁻¹ indicate the deformation of olefins group (C–H) (Mirghani et al. 2002; Nandiyanto et al. 2019).

These functional groups can form surface complexes or bonds with the dye molecules and, consequently, promote their adsorption onto the surface of CaSF. The comparison between the CaSF FTIR spectra before and after dye adsorption shows that, after the adsorption, there were changes in the intensities of the absorption peaks mentioned above. Therefore, the functional groups concerned influence MB and MG adsorption on CaSF. Similar results were found for the removal of methyl red from aqueous matrices using CaSF (Teixeira et al. 2023).

The study pH_PZC makes it possible to find the ideal pH for the adsorption process of dyes on CaSF to occur. Figure 4 shows that CaSF has a buffer zone, or the point of zero charge, at pH values between 3 and 8 (circled in red). Therefore, the pH_PZC of CaSF is 6. Anion adsorption is favored at pHs below the pH_PZC, while cation adsorption is favored at pHs above it (Kumar et al. 2010). MB and MG are both cationic, so their adsorption is favored when the pH exceeds the pH_PZC. The ideal pH defined for the process was 7, thus eliminating the need for sample buffering or pH adjustment.

Figure 5 shows that CaSF is more efficient in removing MG than MB and that the dosage influences the process. The observed trend was that removal efficiency increases with increasing CaSF dosage for both dyes until the optimal dose is attained. Dosages above optimal do not significantly affect removal efficiency.

The lowest MB removal efficiency (53%) was achieved at the lowest CaSF dosage (1 g·L⁻¹), while the highest (84%) was achieved at 10 g·L⁻¹, as above the dosage removal efficiency does not change. Therefore, the optimal CaSF dosage for MB is 10 g·L⁻¹ and was used to evaluate the other test parameters.

The lowest MG removal efficiency (86%) was achieved at the lowest CaSF dosage (1 g·L⁻¹), while the highest efficiency (100%) was achieved at 8 g·L⁻¹, as above the removal efficiency does not change. Therefore, the optimal CaSF dosage for MG is 8 g·L⁻¹, and it was the one used for the evaluation of the other tested parameters.

As the CaSF dosage increases, the availability of active adsorption sites also increases and, consequently, dye removal efficiency also increases. Dosages above established optimal values, in addition to not increasing efficiency, promote unsaturation of active adsorption sites, causing waste of CaSF.

Stirring influences the adsorption of both dyes (Figure 6) as the removal efficiency varied with the variation of the stirring rate until an equilibrium was reached. For MB, the process carried out without stirring (0 rpm) was able to remove 69%, while for MG there was a removal of 62%. When stirring was increased to 20 rpm, the MB and MG removal rates increased to 82 and 97%, respectively, indicating that the MG adsorption process is more dependent on stirring than that of MB.

The optimal stirring rate was 80 rpm for both dyes, showing removal rates of 84 and 100% for MB and MG, respectively. Stirring rates above 80 rpm showed no increase in the removal efficiency.

As the stirring rate is increased, the CaSF adsorption sites become more accessible due to the increased turbulence. Consequently, there is a decrease in the thickness of the liquid film around the CaSF particles and an increase in the diffusion of dye molecules through it, generating an increase in the mass transfer of free MB and MG in solution to the CaSF surface. After reaching the optimal condition, more aggressive stirring becomes unnecessary and can be avoided in order to save energy and material costs.

Figure 7 shows that the adsorption of both dyes occurs very quickly in the first 20 min and then decelerates until reaching equilibrium at 90 min. The high early adsorption rate is due to the high availability of free sites on the CaSF surface. After 20 min, adsorption is slower because CaSF site availability has decreased considerably and those dye molecules already adsorbed tend to repel free dye molecules in solution (Oyelude et al. 2018).

Table 2 shows that for MB, the PSO model best fitted the experimental data, presenting the highest determination coefficient (R_adj² = 0.99), the lowest error (RSS = 0.10), and q_e closest to the experimental equilibrium adsorption capacity (q_{e_exp}).

For MG, the PFO model best fitted the experimental data (R_adj² = 0.99 and RSS = 0.70). The PSO model also fitted the MG data but was not the best fit. The Elovich model did not fit either dataset (R_adj² < 0.95). The fact that the PSO model fits both dyes is a strong indication that chemisorption governs the process. It also indicates that physisorption may occur during in the early stages, which explains that the PFO model also fits the process. Zhang et al. (2017) obtained similar results using hydrochar derived from algal residues.

The study of the equilibrium at different temperatures through adsorption isotherms can provide parameters such as CaSF's maximal adsorption capacity and the type of adsorption involved (physisorption or chemisorption). It can also indicate whether it is the endothermic or exothermic proces (Table 3).

As can be seen, the temperature does not influence CaSF's adsorption of MB, as it has little effect on the experimental adsorption capacity (q_{max_exp}). The increase in q_{max_exp} and q_max with temperature increase from 20 to 30 °C indicates that the process is endothermic. The Langmuir model best fitted the experimental data, with R_adj² = 0.99 and lower error values (RSS), showing a q_max of 199 mg·g⁻¹ at 30 °C. Therefore, the adsorption of MB occurs on a monolayer with a defined number of sites of equivalent energies (Nascimento et al. 2020).

Likewise, the temperature does not influence MG adsorption by CaSF, as the q_{max_exp} did varied little with increasing temperature. The Langmuir model showed the highest q_max value at 40 °C (123 mg·g⁻¹), indicating that the process is endothermic. However, the DR model, which best fits the data, indicates that CaSF's MG adsorption occurs in a multilayered and heterogeneous way (Ruthven 1984; Bonilla-Petriciolet et al. 2017). The values of E (>16 kJ·mol⁻¹) presented in the DR model indicate that for both dyes, the process is one of chemisorption (Kaveeshwar et al. 2018).

The fact that temperature does not influence the process indicates that CaSF is stable with respect to temperature variation. Similar results were obtained with the removal of Disperse Blue 56 (Melo et al. 2021) and Yellow 27 (Melo et al. 2017) using ionic flocculation.

Figure 8 shows that no point touches the origin of the graph for either dye, indicating that their adsorption by CaSF involves two or more steps. The increase of the constant C at each step indicates an increase of the boundary layer effect in the process and also suggests that more than one mechanism is involved as a limiting factor in the adsorption of dye (Teixeira et al. 2022a, 2022b).

For both dyes (Table 4), k_di has the highest value in the first step. The third step, equilibrium, has the lowest k_di values, yet without representing an adsorption limiting factor (Melo et al. 2017). The second step, corresponding to intraparticle diffusion, presents the lowest values of k_di, a strong indication that this step controls the process. Similar results were found in MG removal by ionic flocculation (Teixeira et al. 2022a, 2022b).

Figure 9 shows that, after three cycles, MB removal efficiency decreased from 84 to 61%. The same happened to the desorption efficiency, which dropped from 52 to 22%, and to the adsorption capacity of CaSF, which also dropped from 4.6 to 3.1 mg·g⁻¹. Likewise, Figure 10 shows that, for MG, in the first two cycles CaSF maintained its dye and desorption efficiencies and its adsorption capacity (100%, 8% and 6.2 mg·g⁻¹, respectively). In the third cycle, MG removal efficiency dropped to 70%, desorption efficiency to 6%, and adsorption capacity to 4.4 mg·g⁻¹. The low desorption efficiency of both dyes is a strong indication that chemisorption is the process involved, confirming the DR isotherm's E values.

Regeneration becomes more difficult with each cycle because the dye–CaSF interaction increases, that is, the strength of the bonds involved increases, making dye desorption increasingly difficult, eventually leading to saturation of the CaSF and to decreasing adsorption capacity. Even after three cycles of CaSF, however, it showed good removal efficiency (>60%), indicating that it can be used as an adsorbent for both MB and MG.

CaSF obtained from frying soybean oil is amorphous and contains functional groups such as OH⁻ and C = O, that favor cationic dye adsorption. The pH_PZC CaSF is 6, indicating that a pH 7 (neutral) can be used as the optimal for the process, thus avoiding the need to adjust the effluent's pH. The optimal CaSF dosage removal efficiency was 10 g·L⁻¹ for MB and 8 g·L⁻¹ for MG. Stirring affects the process and MG is more sensitive to this than MB. The equilibrium time for both dyes was 90 min. Adsorption capacity varied little with temperature and dye adsorption appears to be endothermic.

The PSO model fitted the MB kinetics data best, while the PFO model was best for MG. The Langmuir and the DR isotherm models were the best fit for MB and MG adsorption, respectively. The adsorption mechanism study indicated that the process has more than one step and that intraparticle diffusion controls adsorption for both dyes. The adsorption–desorption cycle study showed dye removal efficiency above 60% for up to three cycles.

The authors would like to acknowledge the financial support for this work of Cearense Foundation for Scientific and Technological Development Support – FUNCAP, BPI (Grant N◦ BP3-0139-00276.01.00/18 and N◦ BP4-0172-00080.01.00/20) and Central Analítica of the Federal University of Ceará (UFC).

Y.N.T. investigated and validated the article, developed the methodology, and wrote the original draft. F.J.d.P.F. conceptualized the whole article, conducted funding acquisition and formal analysis, brought resources, wrote the review and edited the article, and administered the project. V.P.B. developed the methodology and conducted formal analysis. J.M.C.M. did formal analysis and wrote the review and edited the article. D.B.S. and J.V.S.N. visualized the article and conducted methodology and formal analysis. J.d.Q.d.S. and R.N.P.T. visualized the article and did formal analysis.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.