Adsorption of acid and basic dye from the simulated wastewater using carbonized microplastic particles synthesized from recycled polyethylene terephthalate plastic waste bottles: an integrated approach for experimental and practical applications

In recent decades, polymer-made products (plastic) have become an essential part of our daily lifestyle. Globally, polyethylene terephthalate (PET) is deliberated as the massive consumable polymer in industries that produce bottles and containers for food, medicine, and soft drinks due to its exclusive features, including light weight, low cost, transparency, good insulation, excellent tensile properties, chemical resistance, ease of handling, and flexibility (Rahmawati et al. 2019; Din et al. 2020), consequently, every year, a larger volume of PET solid waste is generated, which creates serious environmental problems (El Essawy et al. 2017a). Polyethylene terephthalate is the most plentiful municipal and industrial waste, but it has no returned valuable application due its to non-biodegradability, unsuitable mechanical properties, thermal stability, and lower electrical conductivity (Mallakpour & Behranvand 2016). As a result, disposing of PET waste is a major challenge for everyone around the globe, particularly in developing countries. The common approach to the management of PET waste is incineration and landfilling (Djahed et al. 2015), in a developing country, but these methods may adversely affect the environment by releasing heat and volatile compounds and contamination of landfilling sites, respectively (El Essawy et al. 2017a). On the other hand, recycling is one of the appropriate potential methods for reducing the landfilling and incarnation of PET waste; however, it has a lower application due to the low cost–benefit ratio of its associated technology. Activating carbon production from PET plastic is an effective alternative approach to PET waste management (Djahed et al. 2015). Providentially, PET bottles are composed of high carbon and very few mineral contaminants in their chemical components, which are possibly useful as black carbon materials for pollutants adsorption (Mendoza-Carrasco et al. 2016). Dyes are hazardous toxic pollutants that are widely used as a colouring agent in a variety of industries, including tanneries, paints, pulp, food, paper, varnishes, ink, plastics, and textiles (Chakraborty et al. 2020). Globally, about 10,000 different dyes and pigments are available with a production of 7 × 10⁵ tonnes/year. About 50% of applied dyes are released as coloured effluent from textile industries (Ghosh et al. 2020). The occurrence of coloured pollutants in water bodies is harmful to the aquatic ecosystem, as they hinder the dispersion of the sunlight vital for the photosynthesis of aquatic flora and produce mycotoxins for fish and other aquatic organisms (Chakraborty et al. 2021). Moreover, exposure to elevated levels of textile dyes and their by-products can cause severe damage to human health, including kidneys, liver, central nervous system, brain, and reproductive organs (Al-Zawahreh et al. 2021; Zaman et al. 2021). Consequently, significant attention must be paid to the treatment of coloured effluent before discharging into the receiving for the protection of environmental quality and the sustainability of textile industries (Chakraborty et al. 2021). Due to their ease of application, vibrant colour, excellent binding ability, and rapid colour change properties, methyl orange (MO) and methylene blue (MB) are frequently utilized in a variety of industrial sectors (Ghosh et al. 2020). However, their removal is challenging because of their intricate aromatic ring structure, inability to degrade, xenobiotic characteristics, strong light visibility, stability, and oxidation reaction (Dehghani et al. 2021). Numerous methods have been applied for the treatment of coloured effluents, such as chemical precipitation, membrane filtration, coagulation, ion exchange, oxidation, photo-catalytic degradation, ozonation, electrochemical technologies, solid-phase extraction, and adsorption (Ghosh et al. 2018; Al-Ghouti & Al-Absi 2020; Zaman et al. 2022). Conversely, several azo dyes are resistant to chemicals and resistive to photo- and bio-degradation (Al-Zawahreh et al. 2021). Adsorption is the best method among them due to its simplicity of operation and design, low price, great efficiency, richness of the source, lack of sludge, adaptability, and insensitivity to harmful substances. Commercial activated carbon is also used as an adsorbent, but it is highly expensive. Diverse bio-adsorbents have been utilized to remove the colour from wastewater, as broadly reviewed by Yagub et al. (2014) and Chakraborty et al. (2020). Therefore, many researchers are becoming increasingly interested in the synthesis of new, inexpensive, and environmentally friendly adsorbents with an excellent performance from discarded materials for the removal of coloured effluent. So, this study aims to prepare activated carbon from PET bottle waste through an innovative and inexpensive method. To remove cationic and anionic dyes from an aqueous environment, the produced activated carbon's adsorption behaviour is evaluated under a variety of experimental settings, including initial dye concentration, contact time, temperature, solution pH, and adsorbent dose. Additionally, investigations of the adsorption kinetics, equilibrium, and thermodynamics have been conducted. Finally, the application of carbonized microplastic particles (CMPs) as an adsorbent to actual wastewater was also studied. Wastewater treated by waste materials is a challenging area, but it provides dual profits for water treatment and waste management.

Analytical-grade chemicals were utilized throughout all experiments without further purification. MB, sodium hydroxide (NaOH), MO, and hydrochloric acid (HCl) were purchased from Sigma-Aldrich (USA). All experiments were conducted with double-distilled water. In this study, MB and MO were selected as basic and acid dyes, respectively.

The adsorption isotherm tests were conducted using 250 mL of MO and MB solutions in several 500 mL beakers with initial dye concentrations varying from 5 to 70 mg/L. Each beaker received the same fixed dosage of adsorbent (10 g/L), and the rotation speed was kept at 200 rpm with the optimum pH values of 3 and 11 at room temperature (25 ± 2 °C) and equilibrated contact times of 30 min, respectively. Adsorption isotherms provide a complete knowledge of the nature of association by demonstrating how the adsorbate behaves with sorbent materials. In this study, equilibrium adsorption isotherm models of Langmuir and Freundlich have been used, as presented in Table S1.

The adsorption kinetic studies involved adding 10 g/L CMP into 350 mL of MO solution, which had 20 mg/L of MO and MB dye concentration, at optimal pH 3 and pH 11, and an agitation speed of 200 rpm at ambient temperature (25 °C), respectively. Following specific time intervals (1, 2, 3, 5, 7, 10, 30, 60, 90, 120, 150, 180, and 210 min), samples from the solution were taken, filtered, and analysed. Adsorption kinetics includes details on the rate of MO and MB adsorption by the adsorbents, the amount of time needed to complete the adsorption process, and the reaction's mechanism. The kinetics behaviour was identified using Ho and McKay's pseudo-second-order model and Lagergren's pseudo-first-order model, whereas the intraparticle diffusion model was employed to examine the adsorption process's possible rate-controlling step and diffusion mechanism, as presented in Table S1.

The values of ΔH and ΔS were determined from the slope and intercept of the plot of ln K_d versus 1/T, respectively, and the values of ΔG from the K_d value at each temperature (Abd Elhafez et al. 2017).

The Box–Behnken (BB) design is a frequently used standard statistical method for process optimization with the least amount of tests, evaluating the likely correlations between investigated parameters and their effects on the adsorption of dyes (Gadekar & Ahammed 2019). The response variable was the quantity of MO and MB adsorbed per unit mass of the adsorbent and the experimental factors were the solution pH (X₁), initial MO dye concentration (X₂), and dose of absorbent (X₃). The factor levels were denoted by Codes −1 (lower limit), 0 (central point), and 1 (upper limit). Table 1 shows the BB design levels for experiments.

In the present study, a back-propagation algorithm with a log sigmoidal (logsig) or tan-sigmoidal (tansig) function was applied to a feed-forward network with three layers (input, hidden, and output layers), as illustrated in Figure 3. The subgroups of training (60%), validation (20%), and test (20%) networks were randomly selected. Experimental conditions were used as input parameters for the artificial neutral network (ANN). The colour removal percentage was the output layer. The neurons mandatory in the hidden layer were assessed by trial and error to gain the highest regression with the lower errors (Gadekar & Ahammed 2019). In this study, four neurons were used in an ANN model for projection that had been validated and tested. For ANN prediction, MATLAB (R2020a) was applied.

In the adsorption study, the initial solution pH value is highly influenced by stimulating the adsorbent's external charge, the exterior binding sites, and the rate of dye ionization (Chakraborty et al. 2021). There may be a huge number of active sites in CMP and the uptake of dye molecules can be associated with both the chemistry of the dye solution and the active sites. Over a pH range of 3–11, the influence of solution pH values on dye removal is investigated. The optimal pH for MB adsorption is pH 11 and it has been observed that with rising solution pH (3–11), the elimination of MB increased (95–99%) (Figure 2(b)). This effect can be explained by the cationic dye being positively charged as it dissolves in water. Therefore, the CMP's positively charged surface prevents the cationic MB from adhering to it in an acidic environment (Chakraborty et al. 2020). The rate of MB adsorption increased with increasing solution pH as a result of an increase in the electrostatic attraction between the positive dye molecules and the negatively charged CMP surface. On the other hand, oxygen-containing functional groups caused the CMP surface to become positively charged at low pH levels. Low pH values also point to a relatively high proton concentration, which works closely with MB cationic molecules to create the adsorption sites on the adsorbent surface. Therefore, at low pH values, the removal of MB dye reduces. On the other hand, the study's findings showed that MO elimination varied between pH 3 and pH 11. With a higher solution pH, it can be noted that the percentage of MO removed reduced from 99 to 21% (pH 3–11). The highest percentage removal of MO detected at pH 3 is 99%. The high percentage removal of the MO solution in an acidic environment is probably due to the electrostatic interactions between the positively charged adsorbent and the negatively charged MO dye anions (Zaman et al. 2021), represented by the following equations:

As a result, anionic dye adsorption rises at low pH levels and falls at high pH levels.

On the other hand, pHpzc is a key element that provides information on the level of ionization on the adsorbent surface and its interactions with the adsorbate. Figure S1(b) displays the curve of (pH_i – pH_f) vs. pH_i. This figure shows that the pHpzc value of the CMP adsorbent was found to be 2.99. The CMP adsorbent's pHpzc value reveals that the surface of the material is positively charged at pH levels below pHpzc and negatively charged at pH levels above pHpzc. So, pHpzc also confirms that a low pH of 3 is highly significant for acid dye removal and a high pH of 11 for basic dye removal from wastewater. This finding is in line with the findings of the studies conducted by Laszlo & Szűcs (2001) and Santos et al. (2020).

Figure 2(d) indicates that the adsorption is significantly regulated by the initial dye concentration. With an increase in initial dye concentration (5–70 mg/L), the removal rates of MO (99–80%) and MB (96–84%) decrease because, at the constant dose, the adsorbent's external surface is saturated and dye molecules are blocking the pores. Therefore, the initial dye concentration and contact time have a considerable impact on the removal efficiency. While the adsorption capacity of CMP reduced from 0.49 to 5.64 mg/g with raising MO dye concentration from 5 to 70 mg/L at a fixed CMP dose of 10 g/L, this may be because of the strong interaction between dye molecules and adsorbent surface, which increases the significant driving force to transfer a high mass of MO and MB from the liquid to the solid-phase in aqueous solution (Zaman et al. 2021).

In this study, to evaluate the adsorbate's adsorption capacity, the adsorbent dose is regarded as a crucial technical parameter that primarily influences the removal of contaminants from wastewater (Chakraborty et al. 2020). Under the optimum experimental conditions, a variety of adsorbent dosages (1–20 g/L) were used in the adsorption experiments. Figure 2(c) shows the impact of CMP dose on the adsorption process. It demonstrates that dye removal efficiency increased with increasing CMP doses (1–20 g/L) for MO and MB, respectively, due to larger surface areas and more interchangeable sites on the adsorbent surface, resulting in higher adsorption performance (Chakraborty et al. 2021). However, a higher adsorbent dose (1–20 g/L) revealed a reduced adsorption rate (5.31–0.99 mg/g for MO and 13.24–0.99 mg/g for MB), which may be due to dye molecules competing for space on the CMP or overlapping with one another (e.g. aggregation). A similar explanation was found by Kumari et al. (2022).

The Langmuir and Freundlich isotherms are frequently used in the solid/liquid system. In this study, these models are used to examine the equilibrium experimental data of MO and MB adsorption onto CMP. The correlation coefficients (R²), chi-square (χ²), residual sum square (RSS), and root mean square errors (RMSE) value (Table 2) determine the best-fitted result for isotherms. The results demonstrated that the Langmuir isotherm for the adsorption of MO and MB onto CMP fitted better than the Freundlich isotherm (R² = 0.937, RSS = 4.610; RMSE = 0.679, and χ² = 0.313) in terms of a higher correlation coefficient value (R² = 0.972), reduced error (RSS = 0.689; RMSE = 0.262), and chi-square value (χ² = 0.044) (Table 2). This demonstrates that dye molecules cover CMP in a monolayer with uniform distribution. The monolayer maximum adsorption capacities of the dye molecules onto CMP were 5.68 mg/g for MO dye and 6.651 mg/g for MB dye, respectively (Table 2). The Langmuir isotherm calculation showed that the conditions for MO and MB dye adsorption onto CMP were favourable since the R_L values of the dyes lie between 0 and 1 (Table 2). However, the value of adsorption intensity (n) for CMP determined from the Freundlich isotherm was more than 1, suggesting that the adsorption mechanism was favourable for CMP as an adsorbent for the removal of MO and MB dye from aqueous solutions. (Table 2). Shen et al. (2022) found that the Langmuir isotherm was the best-fitted model for the adsorption process using polypyrrole-modified plastic-carbon. The adsorption capacities of CMP are comparable with other adsorbents for acid and basic dye adsorption (Table S2).

The adsorption behaviour of MO and MB onto CMP was examined using three kinetic models (Lagergren pseudo-first-order, Ho's pseudo-second-order, and intraparticle diffusion), and the values of model parameters are listed in Table 3. The calculated (q_e,cal) values from the pseudo-second-order kinetic model also correlated with the experimental (q_e,exp) values for CMP (Table 3), indicating that the adsorption process follows the pseudo-second-order kinetic model. Therefore, it appears that chemisorption or the electrostatic interactions between the molecules of the adsorbent and the dye regulate the entire adsorption process. Intraparticle diffusion was applied to investigate the diffusion mechanisms. The intraparticle diffusion shows multi-linearity and did not pass through the origin (Figure S2) indicating that two or more mechanisms such as intraparticle diffusion, bulk diffusion, and film diffusion controlled these dye adsorption process, which was the rate-limiting step of MO and MB dye adsorption onto CMP. These two processes may occur simultaneously. Santos et al. (2020), Shen et al. (2022) and Kumari et al. (2022) found similar observations in their adsorption studies.

The values of the thermodynamic study parameters are shown in Table 4 and Figure S3(a). A Van't Hoff plot of lnk_d vs. 1/T produced a straight line with R² values of 0.993 and 0.990 for the MB and MO, respectively. The observation that both dyes have negative ΔG⁰ values demonstrates that the adsorption process is spontaneous and practical (El Essawy et al. 2017b).

Table 4

Thermodynamic modelling of CMP on MO removal

.	MB .			.	MO .			.
Temperature (K) .	(kJ/mol) .	(kJ/mol) .	(J/mol/K) .	R2 .	(kJ/mol) .	(kJ/mol) .	(J/mol/K) .	R2 .
298	− 4.301	−4.498	0.703	0.993	− 8.999	−46.163	125.109	0.990
313	− 4.261				− 6.888
323	− 4.256				− 5.509
333	− 4.284				− 4.742

.	MB .			.	MO .			.
Temperature (K) .	(kJ/mol) .	(kJ/mol) .	(J/mol/K) .	R2 .	(kJ/mol) .	(kJ/mol) .	(J/mol/K) .	R2 .
298	− 4.301	−4.498	0.703	0.993	− 8.999	−46.163	125.109	0.990
313	− 4.261				− 6.888
323	− 4.256				− 5.509
333	− 4.284				− 4.742

View Large

The negative value of ΔH⁰ validates the exothermic nature of the adsorption, indicating that dye uptake reduces with increasing temperature, supported by Figure S3(a), but no significant changes were observed after 298 K (25 °C). So this temperature was chosen as the optimum for further adsorption process using CMP. The positive ΔS⁰ also recommends a reduction in randomness between the CMP and dye molecule interfaces. Additionally, the positive value of ΔS⁰ demonstrates the adsorbent's attraction to MB and MO dye adsorption. Santos et al. (2020) found the exothermic nature of adsorption for dye adsorption by PET waste bottles.

The ANN network was established to assess colour removal as presented in Figure 3, which has topologies of 3:4:1 and 3:3:1 for MO and MB, respectively. In this study tan and log sigmoidal transfer functions, as well as hidden layer neurons from 3 to 7, were investigated. In many cases, about 1–20 tested networks of neurons were used in the hidden layer. However, ANN presentation is directly influenced by the number of hidden neurons while larger and lower numbers of hidden neurons hamper the estimation accuracy. Consequently, the optimal numbers of hidden neurons selection help to avoid over- and underestimation (Gadekar & Ahammed 2019). Tables S3, S4 display the effectiveness of the ANN network topology and transfer function. The performance of the network is increased with increasing neuron numbers (Tables S3, S4), while R-value did not show similar results in the training phase. The network of tan-sigmoidal of topology 3:3:1 was selected as the best model for MB due to lower MSE (0.710) and R-value close to 1 for all training, validation, and testing phases, the network of tan-sigmoidal of topology 3:4:1 was nominated as the model for MO due to its lower MSE (0.707) and higher R values for all aspects (all training, validation, and testing phases) (Tables S3, S4). Figure S4 shows the ANN-predicted and experimental values of colour removal in all aspects of network validation, where linear regression analysis was conducted between experimental and ANN-predicted values to validate the network performance. Based on the regression coefficient (R²), the trained artificial neural network model was satisfactory. Gadekar & Ahammed (2019) found a good association between dye adsorption modelling and artificial neural network models.

Strong binding bonds (covalent or ionic bonds) or weak binding forces (Van der Waals forces or a dipole–dipole interaction) between adsorbate molecules and the adsorbent surface influence the reversibility or desorption process (Chakraborty et al. 2021). This study examined the desorption capacities of CMP for MO and MB dyes at various pH levels (pH 5, 7, 9, and 11). Figure S3(b) shows that the desorption percentage of CMP was very low (0.5–7.19%) for MO dye and between 0.45 and 4.87% for MB dye, suggesting that strong binding forces may have existed between the adsorbent (CMP) and the adsorbed (MO and MB dye) molecules rather than rising pH levels. Finally, desorption studies suggest that this adsorbent has environmentally friendly properties because they have no probability of further environmental pollution. Similar observations were also found in other studies (Mouni et al. 2018; Al-Zawahreh et al. 2021; Chakraborty et al. 2021).

In this study, real textile washing wastewater was utilized as an adsorption medium for examining the applicability of CMP in MO and MB removal. The properties of the collected wastewater were pH (8.3), conductivity (731 μS/cm), total dissolved solids (TDS) (353 mg/L), salinity = 0.3 ppt, and absorbance at 464 and 665 nm for MO and MB, respectively. The concentration of MO and MB in the real wastewater was 237.8 and 90.6 mg/L, after the adsorption studies, these concentrations were 62.92 and 31.54 mg/L, respectively. These are corresponding to 74 and 65% removal for MO and MB, respectively (Figure S5). This shows the significant possible application of CMP in removing MO and MB from dye-containing effluent.

To minimize environmental pollution, new, inexpensive, and environmentally friendly wastewater treatment methods must be developed. For these reasons, using PET plastic waste as decolouring agents for textile effluent is very advantageous because it is considered as the most abundant municipal waste. This study tried to prepare activated carbon via the thermal dissociation of PET bottle waste. According to the findings, producing activated carbon is a good adsorbent for removing dye from aqueous solutions because it has a relatively high surface area and micropore volumes. The removal effectiveness of MB and MO reduces with increasing initial dye concentration and temperature and increases with increasing adsorbent dose. These adsorption processes reached equilibrium within 30 min for MB and MO, respectively. The adsorption of MB and MO was favoured in alkaline (pH 11) and acidic (pH 3) mediums, respectively. Significant agreement between predicted and experimental results is represented by the BB-RSM and ANN models. The entire adsorption process of MB and MO was studied kinetically and the pseudo-second-order kinetic model was more effective than the pseudo-first-order kinetic model. During the adsorption process, multiple diffusion mechanisms (including intraparticle and film diffusion) were engaged. Langmuir was the best-fitted isotherm model for this study. The negative values of ΔG, ΔH, and ΔS show that the process of dye adsorption is spontaneous, exothermic, and favourable. The performance of CMP for real wastewater is satisfactory. The whole results indicated that the prepared activated carbon from PET waste bottles is an efficient and low-cost adsorbent and could be utilized extensively in textile effluent treatment plants, where they would confirm fewer adverse effects on the environment and sustainability.

The authors would like to thank the Dr M. A. Wazed Miah Institute of Advance Studies and Research, Research Cell, Jashore University of Science and Technology, Bangladesh and the Ministry of Science and Technology, Bangladesh for the research grant.

Tapos Kumar Chakraborty, Gopal Chandra Ghosh, and Samina Zaman conceived and designed the experiments; Keya Adhikary, Md. Shahnul Islam, Ahsan Habib, Sozibur Rahman, Khandakar Rashedul Islam, Baytune Nahar Netema Khadiza Tul-Coubra, and Md. Simoon Nice performed the experiments and collected the sample. Tapos Kumar Chakraborty, Nazmul Hossain, Himel Bosu, Monisanker Halder, and Samina Zaman analysed and interpreted the data; Tapos Kumar Chakraborty, Keya Adhikary, Md. Shahnul Islam, Ahsan Habib and Samina Zaman wrote the paper.

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

The authors declare there is no conflict.