Estimating the percentage effects of Bemacid red dye adsorption dynamic parameters using a full factorial design approach

Textile industries have used increasingly large quantities of industrial dyes, leading to a corresponding increase in the quantity of unloaded wastewater. The presence of these dyes in wastewater gives it an unattractive color and has negative effects on aquatic ecosystems, including the reduction of the permeation of light and photosynthesis (Zümriye 2005; Yagub et al. 2014).

To address the issue of water pollution, various treatment processes have been utilized to remove dyes from wastewater. These methods include oxidation, coagulation and flocculation, biological treatment, membrane filtration, and others. Of these, the adsorption process has emerged as the simplest and most effective alternative treatment for removing dyes from wastewater (Wong et al. 2004; Crini & Badot 2008; Almeida & Corso 2014). Activated charcoal is the most commonly used adsorbent in the adsorption process, but its relatively high cost often limits its usage (Gupta & Suhas 2009; Mobasherpour et al. 2014). As an alternative, it is recommended to utilize low-cost adsorbents, such as industrial waste materials like those from the steel and fertilizer industries (Jain et al. 2003), brewery waste (Ouazani et al. 2017), natural materials (Bulgariu et al. 2019), or agricultural by-products (Bharathi & Ramesh 2013). The adsorption technique can be performed using either batch or continuous modes (Bharathi & Ramesh 2013). The batch mode is typically used in laboratory-scale experiments but is not suitable for industrial applications due to the continuous flow of pollutants and the large volumes involved. In contrast, the continuous mode (or dynamic mode) is suitable for large-scale applications in wastewater treatment due to its flexibility and ease of operation (García-Mateos et al. 2015).

The breakthrough curve (BTC) obtained from the dynamic study provides significant insights into the dynamic behavior of the adsorption column and is crucial for designing, optimizing, and operating the separation process. Several dynamic models, such as the Bed Depth Service Time (BDST) model, Clark model, Yoon-Landon model, Adams-Bohart model, and Thomas model, are utilized to evaluate and predict the efficiency of the column (Kyzas et al. 2012; Sicupira et al. 2015; Yang et al. 2015).

In this study, the feasibility of using brewery waste as a low-cost adsorbent for the removal of ETL dye in a packed bed column adsorption process was investigated. The effects of various dynamic factors such as bed height, inlet ETL dye concentration, and flow rate were evaluated in a continuous column adsorption system. The kinetic constants of the BTCs were determined and optimized using the different dynamic models. The percentage of dependence between dynamic factors and adsorption capacity is estimated by using the full factorial design.

In our previous study (Ouazani et al. 2017), we described of brewery waste. Based on the results of Fourier Transform Infra Red (FTIR) analysis, it was found that the waste contains lignin and cellulose, classifying it as a lignocellulosic adsorbent. The ETL dye, which is used in this study, is an industrial dye commonly used in the textile industry. The dye was obtained from the SOITEX plant in Tlemcen, which is located in western Algeria. The structure of the dye is unknown, and the solution used in this study was prepared by dissolving 1 g of the ETL dye in 1 L of distilled water.

The spectrophotometer analysis UV-Visible (HACH DR 2000) is utilized to measure the residual concentration of samples at 500 nm.

The study on fixed-bed column adsorption involved packing a specific amount of brewery waste into a glass column with an internal diameter of 1.1 cm and a length of 20 cm. The column was then subjected to a controlled flow rate of ETL dye solution at a predetermined concentration. At predetermined time intervals, samples were collected and analyzed to determine the remaining concentration of ETL dye. The impact of various operational parameters, including bed height (2.5–5 cm), feed flow rate (1.5–5.1 mL/min), and inlet ETL dye concentration (25–100 mg/L), was investigated. The effectiveness of the packed bed was determined by analyzing the BTC, which plots the concentration of adsorbed dye against time. The efficiency of the packed bed can also be represented by the normalized concentration, which is the ratio of the residual dye concentration to the inlet dye concentration (C/C₀), plotted against either flow time (t) or effluent volume (V_eff = Q*t_total). From this formula, the t_total and Q are the total process time (min) and volumetric flow rate (mL/min), respectively (Kannan et al. 2012; Pei-Jen et al. 2014).

To assess the experimental results and make the packed bed suitable for industrial applications, various models such as Adams-Bohart, Thomas, Clark, and BDST were employed to analyze the performance of the packed bed.

Equation (1) can be used to determine the rate constant (K_T) and the adsorption capacity (q₀) of the packaged bed. M is the weight of the adsorbent (g), Q is the flow (L/min), and C is the residual concentration of the Adsorbat in time t. The K_T and q₀ coefficients can be calculated by plotting Ln (C₀/C − 1) against time (t) (Hameed 2009).

The quantity of dye captured by the fixed bed is denoted by the area under the BTC (C versus t) and can be determined through numerical integration (Sicupira et al. 2015). The ratio of the dye mass (m_ad) and the adsorbent mass (M) can be used to determine the capacity (q) of the packed bed.

The performance of the continuous fixed-bed column process was depicted by plotting the BTC, which shows the relationship between the residual dye concentration and the dye inlet concentration (C_res/C₀) as a function of operating time (t).

The bed data were fitted with three different dynamic models previously cited to determine the rate constant (K_AB) and maximum adsorption capacity (q). The results for the two kinetic parameters (K_AB) and (q) and their determination coefficients are presented in Table 1. The obtained results indicate that the Clark model provided a good fit to the experimental data, with determination coefficients ranging from 0.94 to 0.97 for all operational conditions. Furthermore, Table 1 reveals that the value of the kinetic constant was influenced by the flow rate, inlet dye concentration, and bed depth. It was observed that the bed capacity (q₀) was inversely proportional to the flow rate but proportional to the inlet dye concentration.

According to the results presented in Table 1, the values of the kinetic constants obtained using the Adams-Bohart model remained constant for the operational conditions where the dye concentration was 25 mg/L and the flow rate was 1.5 mL/min, regardless of the bed altitude. However, for other conditions, a slight variation in the K_AB constant was observed. These findings suggest that the bed depth did not have a significant impact on the values of the kinetic constant.

The estimated values of the kinetic constant K_AB in this study were found to be lower than the values reported by Hamdaoui (2006) for the removal of methylene blue using sawdust of cedar, which was 34 × 10⁻³ L/mg·min at an altitude of 8 cm and a flow rate of 23 mL/h. On the contrary, the volume capacity obtained in this study is higher than the value reported by Bennani et al. (2015) when they used Moroccan clay for the removal of basic red dye 46, which was 160 mg/L under the operational conditions of a dye concentration of 160 mg/L, a bed depth of 15 cm, and a flow rate of 4 mL/min.

The results presented in Table 1 indicate that the Thomas rate constant K_T values vary significantly under different operational conditions. For instance, when the inlet dye concentration is 25 mg/L and the flow rate is 1.5 mL/min, the K_T value ranges from 2 × 10⁻³ to 22.4 × 10⁻³ L/mg·min, and a similar trend is observed for other flow rates.

The estimated values of the Thomas rate constant K_T in this study are higher than those reported in previous research. For instance, Sivakumar & Palanisamy (2009) found a rate constant of 1.4 mL/mg·min for the removal of acid blue dye 92 using activated charcoal at an initial concentration of 25 mg/L and a bed height of 5 cm. Similarly, Shadeera & Nagapadma (2015) reported a rate constant of 0.35 mL/mg·min for the removal of Brilliant diazo dye at a concentration of 20 mg/L, a bed height of 6 cm, and a flow rate of 0.8 m/min. However, the values obtained in this study are within the same range as those reported by Manase (2012) for the removal of Rhodamine B using activated charcoal.

Table 1 shows that the Clark constant r ranges from 0.11 to 0.73 min⁻¹. These values are higher than those obtained for the removal of methylene blue dye (ranging from 0.02 to 0.07 min⁻¹) as reported by Hamdaoui (2006). Notably, the highest value of the Clark constant r (0.73 min⁻¹) was observed for a bed altitude of h = 2.5 cm, a flow rate of 1.5 mL·min⁻¹, and an initial dye concentration of 100 mg/L. Furthermore, the variation in the Clark constant r was insignificant when the altitude was increased from 2.5 to 5 cm, suggesting that altitude has no influence on the constant for a consistent flow rate.

In our adsorption column study, it was found that the adsorption process of ETL dye on brewery waste was affected by inlet ETL dye dosage, bed altitude, and flow rate. A statistical study was conducted to optimize these factors and model the adsorption capacity in the experimental area.

Table 2 shows that the R² value is 0.97, indicating that 97% of the adsorption capacity value is attributed to the three studied parameters (dye concentration, inlet flow, and bed depth). This suggests that the statistical model is effective in estimating the adsorption capacity value compared to the experimental measurement. The difference between the adjusted R² and R² is less than 0.0245, implying that the statistical model is consistent with the experimental outcomes. Additionally, the software utilized in this study enables for testing noise in the studied area, with an Adeq precision greater than 4, indicating acceptable precision. In this work, the Adeq precision of 33.015 indicates that the signal is appropriate.

In Table 3 of the statistical analysis, the estimated coefficients of the model, standard error, and variance inflation factor (VIF) are presented. In this study, the VIF value is equal to 1, which indicates the orthogonality of the factors.

Figure 6 enables the deduction of additional values for the adsorption capacity from the studied area. Based on the statistical tests and diagrams, the polynomial model is considered a valid representation of the experimental design. This model allows for the prediction of responses for other values included in the studied area.

The color degradation in Figures 5 and 6 indicates that the adsorption capacity is positively influenced by the dye concentration and bed depth, whereas the inlet flow has a negative effect on dye removal. The same observations are recorded in the dynamic study.

The dynamic study confirms that brewery waste is effective in removing Bemacid red dye from aqueous solutions. Several dynamic models (BDST, Thomas, Adams-Bohart, and Clark) were used to analyze BTCs in the packed bed column. The BDST model describes the initial part of the curve, while the Thomas model fits well to the entire curve under the selected operational conditions. The results suggest that to increase the amount of dye removed, low flow rates, higher bed altitudes, and high dye concentrations should be used. The Thomas model has a high determination coefficient (R² ranging from 0.93 to 0.98) under these dynamic conditions. The full factorial design was used to model the adsorption capacity of brewery waste for the studied factors (bed altitude, dye concentration, and flow rate) using a first-degree polynomial model. This approach provides the degree of dependence between the response (adsorption capacity) and the factors, which was approximately 98% in our study.

The present research did not receive any financial support.

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

The authors declare there is no conflict.