Wastewater treatment from the biodiesel production using waste cooking oil by electrocoagulation: a multivariate approach

Biodiesel, one of the most attractive and promising alternative fuels, has been categorized among renewable resources and is currently being developed all over the world (Yaakob et al. 2013; Daud et al. 2015). Biodiesel has been preferred as an alternative energy resource because of its desirable chemical properties such as biodegradability, non-toxicity and carbon neutrality (Omar & Amin 2011; Tan et al. 2011). Carbon neutral fuels absorb carbon dioxide as they grow, and when burned, the same carbon is released back into the atmosphere (Mathews 2008). From an environmental perspective, biodiesel is considered ‘carbon neutral’ because all CO₂ released during consumption is removed from the atmosphere for the growth of vegetable oil plants (Barnwal & Sharma 2005). It also produces low nitrogen, carbon dioxide, sulfur, lead and hydrocarbon emissions in comparison to other non-renewable fuels such as petroleum or diesel, and this ensures a positive impact on the environment (Demirbas 2009).

Four main techniques may be used in biodiesel production: direct use and raw oil blending, pyrolysis, micro-emulsions and transesterification using edible oils, non-edible oils and reusable wastes such as feedstock (Vyas et al. 2010; Daud et al. 2015). Nowadays, transesterification is a common method for producing biodiesel (Siddiquee & Rohani 2011; Abbaszaadeh et al. 2012). The use of edible oils in biodiesel production can cause a food crisis and unnecessary clearing of forests for plantation (Chen et al. 2009; Talebian-Kiakalaieh et al. 2013). Therefore, many researchers have focused on waste cooking oils (WCO) and non-edible oils such as algae, microalgae and jatropha oils (Ganapathy et al. 2009; Abou-Shanab et al. 2011; Karatay & Dönmez 2011). WCO from homes and restaurants can be used in production of biodiesel to reduce water pollution and prevent blockages in water drainage systems (Yaakob et al. 2013). The use of WCO in biodiesel production can also reduce the costs of biodiesel production by 60–90% (Marchetti et al. 2008; Daud et al. 2015).

In the biodiesel production process, wastewater is mainly generated by the washing stage. The amount of wastewater produced by the washing process is high (approximately 20–120 L of wastewater/100 L biodiesel) (Suehara et al. 2005; Srirangsan et al. 2009). Biodiesel wastewater is characterized by high chemical oxygen demand (COD), suspended solids (SS) and oil & grease. It is viscous wastewater with an opaque white color (Jaruwat et al. 2010; Rattanapan et al. 2011; Ramírez et al. 2012). Discharge of untreated biodiesel wastewater into public sewage systems leads to plugging of the pipe systems due to the high concentration of oil & grease. It might also affect the microbial activity of activated sludge in biological wastewater treatment plants (Daud et al. 2015). Individual biodiesel wastewater treatment methods have been employed for the removal of COD, SS and oil & grease, including coagulation (Kumjadpai et al. 2011; Ngamlerdpokin et al. 2011; Xie et al. 2011), electrocoagulation (Chavalparit & Ongwandee 2009; Srirangsan et al. 2009; Ngamlerdpokin et al. 2011; Sandhwar & Prasad 2018), adsorption (Pitakpoolsil & Hunsom 2013), biological processes (Suehara et al. 2005; Ramírez et al. 2012), and microbial fuel cell systems (Sukkasem et al. 2011).

As shown in Equations (9)–(16), as Al⁺³ is generated it reacts with hydroxyl to form Al(OH)₂⁺ at low pH. Then, as pH increases, Al(OH)₃ is polymerized to form Al_n(OH)_3n (Al-Qodah & Al-Shannag 2017).

The EC process is affected by several parameters such as initial current density, pH, reaction time and types of sacrificial electrodes. Treatment efficiency might also affect wastewater type (Butler et al. 2011). The EC process has been successfully applied to different types of wastewater. However, there are only a few studies that have investigated biodiesel wastewater treatment using the EC process so far.

This study investigated the treatability of biodiesel wastewater by the EC process using aluminum and iron electrodes with the addition of NaCl as an electrolyte solution. The effectiveness of the EC process using each electrode material for the treatment of biodiesel wastewater was evaluated and optimized in terms of removal efficiency. The approach of response surface method (RSM) using the central composite design (CCD) was chosen to improve a model and determine the process parameters' interactive effects. Cost analyses of the EC process and characteristics of waste sludge were also determined at the optimum operating conditions.

In this study, the biodiesel wastewater was taken from a small-scale biodiesel production plant in the province of Istanbul in Turkey. In this plant, waste cooking oil is used as feedstock, and the transesterification process is employed for biodiesel production. The main source of the wastewater was from the washing process. The biodiesel wastewater characteristics are shown in Table 1. The wastewater had 29,550 mg/L COD and 3.72 mS/cm conductivity at 20 °C. The pH of the raw biodiesel wastewater was 4.82.

The schematic presentation of the experimental set-up is shown in Figure 1. A laboratory-scale plexiglass EC reactor was used, with a 9-cm diameter and 13-cm height. The aluminum and iron electrodes (one anode and one cathode; Al-Al and Fe-Fe) were used as the sacrificial electrodes with dimensions of 11.5 cm height, 6 cm width and 0.1 cm thickness. In the optimization experiments, the effective electrode area and distance between the electrodes were 62.2 cm² and 2 cm, respectively. The electrodes were connected to a digital DC power supply (ADC-3306D) to impose the current in the laboratory experiments.

The test volume of biodiesel wastewater for each run was 400 mL. The laboratory experiments were performed under the operating conditions of 16 to 80 mA/cm² current density, pH 5–9 and 10–30 reaction time. The initial pH was adjusted using H₂SO₄ (6 N) and NaOH (6 N) solutions to reach the desired values. An electrolyte solution (NaCl: 4.5 g/L wastewater) was used, as the electrical conductivity of wastewater was not found to be sufficient for the EC process. As is known, the current efficiency will decrease when the electrolytic conductivity is low and generally NaCl is added in order to increase the electrolytic conductivity (Liu et al. 2010). Analyses were conducted by taking the supernatant from the reactor at the end of the settling process (180 min) after each run.

The analyses of COD, TSS, orthophosphate, oil & grease and chloride were carried out in accordance with Standard Methods (APHA 2005), and all the chemicals used in this study were of analytical-reagent grade. The sludge scanning electron microscopy (SEM) and Fourier transform infrared spectroscopy (FT-IR) analyses were performed to identify the sludge characterization using the Zeiss EVO^® LS 10 and Perkin Elmer Spectrum 100 Fourier Transform Infrared Spectrophotometer, respectively.

The analysis of variance (ANOVA) and the statistical analysis of the experimental data were conducted with the Statgraphics Centurion XVII software. The interaction between the operating factors and the COD removal was obtained by ANOVA. The quality of the fit of the polynomial model was expressed by the coefficient of determination (R²) and adjusted R². Fisher's F-test was also used to check the statistical significance in the same software; a large F-value indicates high significance for the corresponding term. Based on the p-value, the model terms were selected or rejected (p-value ≪ 0.05 shows that the model terms are significant at 95% confidence level).

The results of ANOVA for the COD removals using Al and Fe electrodes are presented in Table 4. The very high correlation coefficient (R²) values of 95.20% and 91.43% confirm the high accuracy of the model equations for all operating parameters in the EC process using Al and Fe electrodes, respectively. On the other hand, relatively high R² between the experimental results and model-predicted values verified the fit of the quadratic equations. A closer value of R² and Adj.R² confirms an appreciable interaction between the independent variables (the differences between R² and Adj.R² are lower than 20%). The p-values were applied for the significance of each model. It can be stated that the model terms are significant or highly significant if the p-values are found to be smaller than 0.05 or 0.0001, respectively. The sum of square (SS) of residual error was found to be very low in comparison to the total SS for the COD responses for the Al and Fe electrodes, which confirms the high reliability of the fitted second order polynomial models with a 95% confidence interval. The F-values of the model 11.032 and 5.924 for the Al and Fe electrodes respectively indicate that the model was significant (Table 4).

Table 5 presents the ANOVA results of the operating parameters of the predicted response surface quadratic model and other statistical parameters of the COD removal values for the Al and Fe electrodes. The linear coefficients of current density (X₁), initial pH (X₂) and electrolysis time (X₃) had significant effects on COD removal efficiency for Al electrodes (p-value <0.05). The current density's second order terms (X₃²) and the quadratic terms of X₁X₂ and X₂X₃ also had significant effects on the removal of COD for the Al electrodes (p-value: 0.0096, 0.0214 and 0.0356, respectively). In terms of the Fe electrodes, the results of ANOVA indicated that only current density (X₁), electrolysis time (X₃), the second order term of current density (X₁²) and the quadratic term of X₁X₃ had significant effects on the removal of COD.

The response curve graphs of the quadratic models were presented for Al and Fe electrodes in Figure 2. Figure 2(a)–2(c) show the effects of current density, initial pH and electrolysis time on the removal efficiency of COD for the Al electrode. As can be seen from Figure 2(a), a slight increase in COD removal efficiency was observed with an increase in the current density. It can be shown from Figure 2(b) that at first the COD removal efficiency increases with increased current density and electrolysis time for the Al electrode. However, as the current density and electrolysis time reached maximum values, removal efficiency of COD no longer improved. As can be seen from Figure 2(c), the initial pH has an adverse effect on COD removal efficiencies and the increase in initial pH led to a decrease in COD removal. Figure 2(d)–2(f) show the effect of operating parameters on COD removal efficiency for the Fe electrode. Figure 2(d) indicates the effect of current density and initial pH on COD removal, and it was observed that high COD removals were obtained at high current density values. As can be seen from Figure 2(d) and 2(f), initial pH significant effect on COD removal. Figure 2(e) shows the interaction of current density and electrolysis time. It can be said that the electrolysis time has a positive significant effect on COD removal efficiencies and the COD removal efficiency increased slightly with an increase in electrolysis time.

In order to accurately determine the optimum operating conditions for maximum COD removal, response surface and desirability functions were used. The optimized conditions that were attained from the optimization of COD removal values for the Al and Fe electrodes are given in Table 6. The optimum operating conditions were 43 mA/cm², 5, and 21 min for current density, initial pH and electrolysis time respectively for Al electrodes. Under the optimized conditions, the optimum model prediction value of COD removal was 62.9%. In order to control the accuracy of the model-predicted value, additional EC experiments were performed; 62.7% COD removal efficiency was achieved and COD concentration was determined to be 11,025 mg/L after the EC process using Al electrodes. On the other hand, the optimum conditions for the Fe electrode were 47 mA/cm², 7.7 and 30 min for current density, initial time and electrolysis time, respectively. The optimum COD removals were determined as 67.1% and 63.4% by model prediction and additional experiments, respectively. It was noted that COD concentration decreased to 10,815 mg/L from 29,550 mg/L after the EC treatment with Fe electrodes.

The oil & grease, total phosphorus, orthophosphate and TSS removal efficiencies were found to be above 86.7% under the optimum conditions determined for the removal of COD for both electrode types. Direct discharge limits to a receiving water body for biodiesel wastewater in Turkey are 400 mg/L COD, 200 mg/L TSS, 20 mg/L oil & grease and 2 mg/L orthophosphate (Ministry of Environment and Urbanization, MoEU). According to the removal results obtained at the optimized conditions, the biodiesel wastewater direct discharge limits (COD must be <400 mg/L) or discharge limit standards for municipal sewage systems (COD must be <4,000 mg/L) have not been met based on the standards in the Water Pollution Control Regulation of Turkey, as the COD concentrations of treated wastewater were found to be above these discharge standards. Therefore, further advanced treatment processes should be investigated in order to meet the discharge limits for biodiesel wastewater.

Energy consumptions and mass loss of electrodes are the main costs of the EC process in wastewater treatment. Chemical consumption only results from the pH adjustment of wastewater before the EC process. The operational costs were found to be 6.43 €/m³ and 7.01 €/m³ for aluminum and iron electrodes, respectively, at optimum operating conditions.

Under the optimum conditions for Al and Fe electrodes, the sludge SEM and FT-IR analyses were performed to identify the sludge characterization. Figure 3 shows the SEM images for aluminum and iron sludge. As can be seen in Figure 3(a)–3(c), aluminum sludge is more porous as compared to iron sludge (Figure 3(d)–3(f)) and the particle size of iron sludge remains smaller than aluminum sludge.

From the FT-IR spectra of produced sludges (Figure 4), the band in the region of 2,925–2,855 cm⁻¹ represented CH₃ alkane stretching for both sludge samples, which was attributed to fatty acid methyl esters from the transesterification of biodiesel wastewater used oil. Peaks at the 1,742.99 and 1,741.52 cm⁻¹ band of both Al and Fe sludge samples are likely to be attributed to methyl ester (COOMe) and free faty acids (Ngamlerdpokin et al. 2011). On the other hand, the band in the region of 2,924–3,378 cm⁻¹ can be associated with the O-H stretching which is associated with adsorbed water (Guzman et al. 2016). The peaks between the 1,631.63 and 1,547.47 cm⁻¹ bands of both sludge samples correspond to γ(OH) water bending vibration and hydroxyl bending (Parga et al. 2005). The bands at 1,374.77 and 1,467.40 cm^–1 can be attributed to Al-O-H bending. On the other hand, 1,020.24 and 1,242.04 cm⁻¹ bands were associated with carbonyl groups or carbohydrate bending vibrations (Anand et al. 2014). The FT-IR results for sludges indicate that pollutants in the biodiesel wastewater including oil & grease and other pollutants were removed successfully, as the peaks at 2,925–2,855 cm⁻¹, 1,742.99–1,741.52 cm⁻¹ and 1,020.24–1,242.04 cm⁻¹ were associated with fatty acid methyl esters, methyl ester (COOMe) and free fatty acids and carbonyl groups or carbohydrate bending vibrations, respectively.

Due to the high amount of wastewater (containing high COD) produced during biodiesel production, biodiesel wastewater should be treated effectively. Although the EC process is one of the attractive and efficient treatment methods, there are only a few studies that have investigated the treatment of biodiesel wastewater by the EC process so far. Chavalparit & Ongwandee (2009) investigated the EC process using aluminum and graphite electrodes for the biodiesel wastewater treatment. The authors optimized the operating conditions using the Box-Behnken design and RSM. The optimum operating conditions were determined as 6.06 of pH, 18.2 V of applied voltage and 23.5 min of reaction time. In these operating conditions, COD, oil & grease and suspended solids removal efficiencies were 55.43%, 98.42% and 96.59%, respectively. The operating cost was also investigated, and it was found that 5.57 kWh/m³ of power would be required for the treatment of biodiesel wastewater at the optimum operating conditions. On the other hand, Srirangsan et al. (2009) examined the treatment of biodiesel wastewater by the EC process using different types of electrode pairs including aluminum, iron and carbon (Fe-Fe, Fe-C, Al-Al, Al-C and C-C). At the end of the optimization study, the process was efficient at 6, 8.32 mA/cm² and 25 min for the pH, current density and reaction time for the pair of Al-C. COD, oil & grease and suspended solids removal efficiencies were found as 55.4%, 97.8% and 96.9%, respectively. Ngamlerdpokin et al. (2011) also compared the chemical coagulation and electrocoagulation processes for remediation of biodiesel wastewater. In their study, the free fatty acids in biodiesel wastewater were chemically removed from the wastewater using mineral acids, H₂SO₄ and HCl at 2.5 pH and 7 min, and 38.94% COD, 76.32% biochemical oxygen demand (BOD₅) and 99.36% of oil & grease was removed. After the pretreatment, chemical coagulation and EC were investigated, and both treatment methods were found to be effective. In the chemical coagulation, 97.5% COD, 97.2% BOD₅ and 98.2% oil & grease removal efficiencies were obtained at the optimum conditions (6 pH and 2 g/L of Al₂(SO₄)₃). The optimum conditions were found to be 7.4 pH, 13 mA/cm² current density and 4 h reaction time in the EC process. The COD, BOD₅ and oil & grease removal efficiencies were 99.6%, 91.5% and 98.7% in the EC process using Fe electrodes. Although a lower cost was obtained by the chemical coagulation process, the EC process provided a better effluent quality in terms of the investigated parameters. In this study, 62.7% and 63.4% COD removal efficiencies were found at the optimum operating conditions for the Al-Al and Fe-Fe electrodes (43 mA/cm² for current density, 5 for initial pH, and 21 min for electrolysis time for Al electrode and 47 mA/cm², 7.7 and 30 min for current density, initial pH and electrolysis time for Fe electrode), respectively. The operating cost was determined as 6.43 €/m³ and 7.01 €/m³ for the Al and Fe electrodes, respectively. It may be stated that the results of this study and the previous studies show that the treatment of biodiesel wastewater by the EC process is suitable, and the EC process is a cost-effective treatment process. However, further treatment methods such as electro-oxidation using different electrodes and peroxy-electrocoagulation should be investigated to achieve the local discharge standards.

This study focused on the treatment of biodiesel wastewater by the EC process using Al and Fe electrodes to remove COD, oil & grease, total phosphorus, orthophosphate and TSS from wastewater. For this purpose, laboratory experiments were performed using the CCD for the optimization of process parameters including current density, initial pH and electrolysis time, and RSM was used for the process optimization of EC. It was noted that the optimization studies were only performed for COD removal. The ANOVA results show that the high R² (>91.43) ensures a satisfactory adjustment of the second-order regression model with the experimental data. The optimization of the model provides the optimum conditions for the removal of COD using Al electrodes at a current density of 43 mA/cm², initial pH of 5 and 21 minutes of electrolysis time. The removal efficiencies of COD, oil & grease, total phosphorus, orthophosphate and TSS were found to be 62.7%, 89.9%, 98.9%, 99.5% and 86.7% at the optimized conditions, respectively. On the other hand, under the optimum operating conditions for the removal of COD using the Fe electrodes (current density: 47 mA/cm², initial pH: 7.7 and electrolysis time: 30 minutes), the removal efficiencies were found to be almost equal to the results that were obtained using the Al electrodes. The total operational cost was found to be 6.43 €/m³ and 7.01 €/m³ for the aluminum and iron electrodes, respectively. Consequently, the obtained results indicate that excellent removal for oil & grease, total phosphorus, orthophosphorus and TSS can be achieved by the EC process and RSM is a powerful technique to optimize the operating parameters.