A kinetic, isotherm adsorption, and thermodynamic study of Congo red coagulation using Leucaena crude extract as natural coagulant

Coagulation and flocculation are commonly employed technologies in water and wastewater treatment. There are some well-known advantages of coagulation and flocculation such as the simplicity of the process, commercially available chemicals, low equipment cost, and its ability to reduce contaminants (chemical oxygen demand, biochemical oxygen demand, organic content, and the like) (Crini & Lichtfouse 2019). The chemicals that are used as coagulant and flocculant can be classified as inorganic salts, such as aluminium sulphate, ferric chloride, ferric sulphate, ferrous sulphate; inorganic polymers, such as polyaluminium chloride, polyferric chloride, polyaluminium sulphate, and polyaluminium ferric chloride; and organic polymers, for example amino methyl polyacrylamide, polyamine, and polyethylenimine (Verma et al. 2012; Lal & Garg 2017, 2019). However, application of these coagulants poses some drawbacks such as their negative impact on human health, production of non-biodegradable sludge in high volume, and decrease of the water's pH after treatment (Yin 2010). Thus, alternative coagulants are needed to cope with these specific drawbacks.

In recent years, there has been an increasing trend in the investigation of utilizing natural-based coagulants to treat various water and wastewater. Generally, natural coagulants could be classified by their active coagulating agent, namely protein, polysaccharides, and polyphenols. These active coagulating agents are isolated from plants (seeds, fruits, leaves, roots) and animals, as well as microorganisms. Several studies recently reported on Azadirachta indica (Ahmad et al. 2021), Hibiscus esulentus L. (Okolo et al. 2021), Azadirachta indica A. Juss. (Thirugnanasambandham & Karri 2021), and Moringa oleifera (Yamaguchi et al. 2021). The application of natural coagulants in water and wastewater treatment brings some advantages, namely being relatively cost-effective, accessible, and producing less sludge with biodegradable characteristics compared to chemical coagulants (Ang & Mohammad 2020). All these positive traits of natural coagulants reflect their potential as a promising green alternative to substitute chemical coagulants.

Previously, our research team applied leucaena seed (Leucaena leucocephala) as a natural coagulant to treat turbidity (Kristianto et al. 2018) and remove colour (Kristianto et al. 2020a, 2020b) from synthetic wastewater. Based on those studies, it was found that charge neutralization is the coagulation mechanism of leucaena protein. Furthermore, according to Beltran-Heredia, Sanchez-Martın (Beltran-Heredia et al. 2009) after charge neutralization, a complex between dye and coagulant is formed and followed by flocs formation via sorption mechanisms. The adsorption process is the rate controlling step; thus, the coagulation process could be modelled using adsorption models. Considering this coagulation mechanism, a number of researchers have reported the application of adsorption isotherm and kinetic models to describe the coagulation process by both natural (Rani & Talikoti 2013; Menkiti et al. 2018; Ohale et al. 2020; Ejimofor et al. 2022; Kristianto et al. 2022) and inorganic coagulant (Hussain et al. 2013; Hossain et al. 2019; Ngteni et al. 2020; Wu et al. 2020). In recent research by Vishali, Mullai (Vishali et al. 2022), several isotherm adsorption models and kinetics were employed to describe coagulation process of synthetic paint wastewater by using saline extract of Strychnos potatorum, Cactus opuntia and Portunus sanguinolentus. It was found that the coagulant was heterogenous in nature with the monolayer adsorption process, indicated by the suitability of the data and model (Vishali et al. 2022). Another study by Nnaji, Anadebe (Nnaji et al. 2022) investigated the coagulation of Cibacron blue and dispersed red dye wastewater using Luffa cylindrica seed extract. It was discovered that the coagulation process followed the Freundlich isotherm and pseudo-second order kinetic model (Nnaji et al. 2022). The study of isotherm models is useful to provide insight into colloid and coagulant amalgamation, while kinetic and thermodynamic models are utilized to describe the coagulation behaviour (Yasin et al. 2020).

This study has investigated the effect of Congo red concentration and coagulation temperature on the Congo red removal. Congo red, an anionic azo dye, is used as a model substance of organic pollutants. Furthermore, for the first time, more insights into the leucaena extract coagulation mechanism were observed by various isotherm adsorption models, namely Langmuir, Freundlich, Temkin, and Brunauer-Emmett-Teller (BET). The sedimentation kinetic was also investigated by several adsorption kinetic models, namely pseudo-first and second order, Elovich, and intraparticle diffusion models. Finally, the thermodynamic properties of the Congo red coagulation using leucaena crude extract were determined as well.

Preparation of leucaena crude extract as natural coagulant was conducted by following previous studies (Kristianto et al. 2019; Kristanda et al. 2021). To summarize, dried leucaena seeds from a local market in Probolinggo, East Java were washed, oven dried (60 °C, 4 h), crushed, and sieved to obtain leucaena seed kernel powder. The seed kernel powder was subsequently extracted using a 1 M NaCl saline solution at pH 9.0 for 1 h. The seed cake was then separated from the extract and the filtrate was used as natural coagulant. The protein concentration in the extract was determined by using the Bradford method with bovine serum albumin (BSA) as the standard (Bradford 1976), thus the protein concentration was expressed as mg equivalent BSA per mL (mg eq BSA/mL). The extract was freshly made prior to each experiment to prevent any degradation during storage.

The coagulation study of Congo red was made by using a jar test apparatus equipped with a water bath and thermostat. Congo red solutions at various initial concentrations (50–100 mg/L) were prepared, and the pH was adjusted by using a 0.1 M HCl or NaOH solution. The coagulation pH was fixed at 3.0, which was found to be the best pH for coagulation of Congo red by using leucaena extract (Kristianto et al. 2019, 2020a, 2020b). The temperature of Congo red solutions was also adjusted according to the variations (30–50 °C), prior to coagulant addition. A total amount of 3 mL of leucaena crude extract was added to the Congo red solution, followed by rapid (200 rpm, 3 min) and slow mixing (60 rpm, 30 min). The mixture was subsequently left to settle for 60 min, during which a sample was taken and measured every 5 min. Determination of Congo red concentration was accomplished by using spectrophotometer visible (Thermoscientific Genesys 150) at Congo red maximum wavelength. All experiments were performed in duplicate.

The kinetic and isotherm models used in this study are presented in Tables 1 and 2, respectively.

The effect of Congo red concentration and temperature on the Congo red is presented in Figure 1. Furthermore, the result of analysis of variance (ANOVA) in this study is presented in Table 3. It can be observed that Congo red concentration, coagulation temperature, and their interaction are significant to the Congo red removal with a confidence level of 95%. It can also be discerned that with the increase of Congo red concentration, the percentage removal was decreased with the same coagulant dosage. As the Congo red concentration increased, more colloid particles existed in the system. Thus, the coagulant became limited at higher colloid concentration, resulting in the decrease of percentage removal. A similar trend has been reported by previous researchers (Vishali & Karthikeyan 2018; Luo et al. 2019; Vishali et al. 2020). On the other hand, with the increase of operating temperature, the Congo red removal followed suit and increased as well. It is generally known that temperature can affect the viscosity and surface tension of the aqueous system. Furthermore, these physical parameters have a positive influence to the Brownian motion of the particles that lead to aggregation and floc formation (Fitzpatrick et al. 2004; He & Xie 2021). A similar trend has also been reported by previous researchers (Patel & Vashi 2012; Rodrigues et al. 2013; Mahmoudabadi et al. 2019).

The profile of percentage removal at various initial concentrations and various temperatures is presented in Figure 2. It shows that similar percentage removal vs. time profile was observed from both figures, where a rapid increase of removal occurred during the initial 5 min, a further increase is observed until approximately 40 min, and it becomes relatively constant until 60 min. The rapid removal increase at initial sedimentation time is due to the charge neutralization mechanism that is still dominant in this phase (Obiora-Okafo et al. 2020). After that, adsorption and flocs formation are the dominant steps that require more time to complete. The data obtained were then fitted to various kinetic models, namely pseudo-first order, pseudo-second order, Elovich, and intraparticle diffusion.

The pseudo-first order kinetic model assumes that the rate of adsorbate uptake per time is proportional to the difference of adsorption capacity that is usually suitable to describe the initial step of an adsorption process. Furthermore, the pseudo-first order kinetic model is usually suitable in a system with high initial adsorbate concentration and only a few active sites available on the adsorbate (Wang & Guo 2020). The parameters of the pseudo-first order kinetic model obtained in this study are presented in Table 4. Based on the evaluation of the model's suitability, it can be observed that this model generally has a good R² value, ranging from 0.74 to 0.97. However, the low suitability of the qt model compared to the experimental data is shown by the relatively high χ² and SSE value.

The pseudo-second order kinetic model is suitable to describe the adsorption process at low adsorbate concentration and adsorbent with abundant active sites on its surface (Wang & Guo 2020). The adsorption process is assumed to be second order with chemical sorption as the rate determining step. The chemical sorption happens due to the high content of active sites and their interaction with adsorbate via covalent forces and electron exchange (Qiu et al. 2009). The parameters of the pseudo-second order kinetic model are presented in Table 5. It can be seen that this model is highly suitable for the experimental data, as shown in high R² (0.99) and low χ² and SSE values.

The Elovich kinetic model is based on assumptions that the activation energy increases with adsorption time and adsorption occurs on heterogeneous adsorbents (Wang & Guo 2020). The Elovich model is usually useful to describe the phenomenon of adsorption. The parameters of the Elovich model are presented in Table 6. The small a values (a < 1) indicated a homogenous adsorbent of the natural coagulant, while large values of b (b > 1) suggested the occurrence of chemisorption in this study (Vishali & Karthikeyan 2018).

The intraparticle diffusion model, also known as the Weber-Morris model, is commonly used to describe the internal diffusion of adsorbates within the adsorbent as the rate determining step (Wang & Guo 2020). The typical fitting of the intraparticle model is presented in Figure 3. It can be observed that the plot consists of two linear lines. The parameters for the intraparticle diffusion model in this study are presented in Tables 7 and 8 for the first and second linear section of the model, respectively. It can be discerned that in all experiment conditions the C values are nonzero. The presence of two linear lines and nonzero C values mean that the rate determining steps in this study are both intraparticle and film diffusion (An 2020). Further comparison of the intraparticle diffusion constant in the first (k_d1) and second (k_d2) linear section shows that the k_d1 values are bigger than k_d2. This indicates that the film diffusion during the adsorption process was more rapid compared with the intraparticle diffusion (Ohale et al. 2020). A similar trend in coagulation studies has also been reported previously (Okey-Onyesolu et al. 2020).

The agreement of experimental and model data was evaluated by the R², χ², and SSE values. It can be observed that all kinetic models gave relatively high R² values around 0.8–0.9, except for the fitting of the second section intra-particle model. Among all kinetic models, it can also be seen that the pseudo-second order model showed R² > 0.99 with very low χ² and SSE values and good fit of model and experimental data as presented in Figure 4. This indicates the pseudo-second order model was the most suitable model to describe the coagulation kinetic in this study. The use of pseudo-second order kinetic model to describe the coagulation process has also been reported by previous researchers (Obiora-Okafo et al. 2019; Ngteni et al. 2020; Okey-Onyesolu et al. 2020). The suitability to the pseudo-second order kinetic model indicates that the coagulation process is chemisorption in nature, due to the interaction between coagulant and Congo red molecules via hydrogen or dipole-dipole bonds (Jadhav & Mahajan 2014).

The parameters of various isotherm adsorption models, namely the Langmuir, Freundlich, Temkin, and BET models, are presented in Table 9. The Langmuir model assumes a homogeneous adsorption phenomenon with constant enthalpy of each molecule. The Langmuir constant (Kl) illustrates the adsorption capacity of the coagulant. It can be seen that with the increase of temperature, the Kl values also increased. Furthermore, the calculated separation factor (Rl) ranged from 0.169 to 0.289, 0.047 to 0.090, and 0.006 to 0.012 for coagulation temperature of 30°, 40°, and 50 °C, respectively. The Rl values ranging between 0 < Rl < 1 indicate a favourable adsorption, while the decreasing trend of Rl indicates an increased affinity between Congo red molecules and leucaena protein during the coagulation (Lin et al. 2011).

The Freundlich model is based on the assumption that adsorption of multi-layered adsorbates on heterogeneous adsorbent surfaces with different energy levels. The Freundlich constant, Kf (L/mg), illustrates the relation between adsorbate and adsorbent as well as the adsorption capacity. It can be inferred from Table 8 that KF values increased with the increase of coagulation temperature, indicating the increased adsorption capacity with the increase of temperature and the endothermic nature of the coagulation adsorption in this study. Furthermore, the 1/n values fall within the range of 0 < 1/n < 1 indicating a favourable coagulation adsorption of Congo red dye using leucaena protein.

The Temkin model assumes that heat of adsorption is a function of temperature and linearly decreases due to interaction of adsorbate and adsorbent during the adsorption process (Al-Ghouti & Da'ana 2020). The A value of the Temkin model indicates the binding energy. The increase of A values with the increase of coagulation temperature confirmed the endothermic nature of Congo red coagulation adsorption in this study (Aljeboree et al. 2017).

The BET adsorption is known as a modification of the Langmuir model. In the BET model, adsorption of dye molecules is randomly distributed and a multilayer adsorbate is formed. There are two parameters in the BET model, namely A, which is the BET constant, and Xm (mg/mg), the amount of adsorbed Congo red molecule by the coagulant to form a monolayer (Hossain et al. 2019). The increase of xm values demonstrates the increase of Congo red molecule that was adsorbed in the form of monolayer on the coagulation adsorption process. Furthermore, the negative A values implied that the surface of the coagulant was saturated with the Congo red molecules (Hossain et al. 2019).

The suitability of the isotherm adsorption models was examined by observing the R², χ² and SSE values. It can be seen in Table 8 that in terms of R², the BET model gave the highest R² compared to other models. However, based on the fitting of experimental and model data as seen in Figure 5, the BET model turned out to be the least suitable of the four models. Furthermore, it can be observed that the Langmuir model gave the smallest χ² and SSE values, confirming the suitability of the Langmuir model in this study. Thus, it is safe to conclude that the coagulation adsorption of Congo red by leucaena protein occurred in a form of monolayer and on a homogenous surface. The suitability of the Langmuir model to describe the coagulation using natural coagulant has also been reported by previous researchers (Antov et al. 2010; Kebaili et al. 2018; Menkiti et al. 2018).

The thermodynamic parameters of Congo red coagulation adsorption in this study are presented in Table 10. It shows that at all coagulation temperatures the ΔG° values were negative, indicating the spontaneous nature of the coagulation process. The increase of ΔG° negativity with the increase of coagulation temperature indicated the increase of spontaneity at higher temperature. Furthermore, it can be discerned that the positive ΔS° indicated the randomness of the interaction of Congo red and leucaena protein during the coagulation process. The magnitude of ΔH° value could be used to infer whether the coagulation adsorption process was physisorption (ΔH° < 40 kJ/mol) or chemisorption (ΔH° > 40 kJ/mol) (Egbosiuba et al. 2020). The ΔH° value of 142.96 kJ/mol suggested that the Congo red coagulation was more a matter of chemisorption than physisorption. Furthermore, the positive value of ΔH° confirmed the endothermic characteristic of Congo red coagulation adsorption.

To evaluate the possible application of leucaena crude extract as a natural coagulant to substitute for inorganic chemical coagulant, its economic aspects and feasibility were analysed. Calculation of coagulant cost was based on the ingredients that were needed to obtain the coagulant, disregarding other costs such as energy, labor, depreciation of capital cost, and the like. In the preparation of leucaena extract, the contributing cost consists of leucaena seeds ($3.45/kg), NaCl, and distilled water. Calculation was based on the optimum coagulation dosage for 50 mg/L Congo red removal was found to be 10 mL/L using leucaena extract (Kristianto et al. 2019), compared with ferrous sulphate (159.64 mg/L) (Luo et al. 2019) and aluminium sulphate (40 mg/L) (Goudjil et al. 2021). The cost was estimated taking into consideration that the leucaena seed kernel amounted to 50%-w of the total seeds weight (Sethi & Kulkarni 1995), and chemical cost was derived from a marketplace in Indonesia. Based on the estimated cost presented in Table 11, it could be observed that utilization of leucaena extract as natural coagulant was the most expensive, compared to ferrous sulphate and aluminium sulphate. The high cost comes from utilization of NaCl and distilled water for protein extraction that accounts for up to 85% of the cost. A similar observation was reported by Ang, Kiatkittipong (Ang et al. 2020), namely that the extraction process would contribute greatly to the natural coagulant cost. The higher natural coagulant cost compared to inorganic coagulant has also been reported by other researchers (Zheng et al. 2021). However, despite the high cost, the use of the natural coagulant was worthwhile because of the environmental and social impact of the natural coagulant's life cycle analysis, which has been reported to be better than inorganic coagulants' (Niquette et al. 2004; Zheng et al. 2021). It is generally known that environmental and social impact, as well as economic aspects, are important factors in achieving the actual sustainability of a process. Further optimization of the extraction process is needed to minimize the cost of natural coagulant from leucaena seeds.

This study has investigated the effect of Congo red concentration and temperature on the coagulation performance of leucaena crude extract. It was found that both variables and their interaction were significant to the removal of Congo red. Furthermore, the coagulation process was investigated by using several adsorption kinetic and isotherm models. It was found that the sedimentation process was well described by the pseudo-second order kinetic model, indicating chemisorption of Congo red on the coagulant. From the intraparticle model, it could be inferred that both intraparticle and film diffusion were the determining steps, where the latter was more dominant in the diffusion process. The experimental data using isotherm models show that the Langmuir model is the best model to describe the coagulation process, inferring the monolayer adsorption process in this study. Furthermore, based on the thermodynamic parameters, it was found that the coagulation was spontaneous and endothermic in nature. The results of this study show that leucaena extract could be used as alternative coagulant to treat colour wastewater. Further study to optimize the extraction process to lower the coagulant cost and further application in more complex wastewater are needed, prior to its commercial application.

This study is funded by Parahyangan Catholic University's Centre of Research and Community Service under contract No III/LPPM/2022-02/34-P. The authors are grateful for the financial support provided.

The authors declare they have no competing interests.

Data cannot be made publicly available; readers should contact the corresponding author for details.