1 Introduction

The Textile industry plays a major role in polluting natural water resources and depletion of potable water (Akhtar et al. 2020; Gökkuș and Yıldız 2014). It is estimated that about 5 × 105 tons of dyes are consumed and 250–350 m3 of colored wastewater is generated for every ton of finished Textile product (Arora 2014). Dyes are classified into three major groups: (1) cationic i.e. basic dyes; (2) anionic i.e. direct, acid, reactive dyes; and (3) non-ionic. As cotton fabric makes up about 50% of the total fiber consumption in the world, cotton dyeing with reactive dyes has led to an increased focus on studies of the decolorization of reactive dye-bearing Textile effluents. Furthermore, alkaline dyebath hydrolysis and lower fixation rates caused an additional waste of the reactive dyes (Sharma and Rashmi 2013). Studies have reported that during the dying process, significant amounts of dyes (≈ 16–52%) fail to adhere to fabric, and are released as wastewater (Rehman et al. 2018). The reactive dyeing process has a much higher water utilization rate compared to dyeing with other types of dyes resulting generation of large volumes of highly colored spent dye baths (Sharma and Rashmi 2013). Dyes are highly toxic, carcinogenic, and mutagenic in nature (Zhang et al. 2015). Their relatively longer half-lives and poor biodegradability cause damage to the aquatic environment and human health (Lellis et al. 2019; Hassan and Christopher 2018). Dye-bearing wastewaters, discharged to a water body, can lead to reduced sunlight penetration; a lower rate of photosynthesis and dissolved oxygen levels, and an increase in turbidity (Hassan and Christopher 2018). Therefore, stringent regulations have been imposed by the statutory bodies in India and abroad, to prevent the discharge of colored effluents and to promote the installation of the on-site treatment units.

It is extremely difficult to treat reactive dyes bearing effluents by adopting conventional treatment approaches. The Reactive Orange 16 (RO16) is one such anionic dye, belonging to the azo group and soluble in the aqueous phase. These dyes impart color owing to the chromophoric groups in their molecular structures and fix the color to the fabric due to the auxotrophic groups which react with cellulose hydroxyl or amino acid functional group of wool and similar synthetic fabric (Shen and Gondal 2017; Wang et al. 2018). The azo (N=N) group in the RO16 dye structure results in poor biodegradability leading to a lower removal efficiency of these dyes in the engineered treatment system. Failure to rely on sorption and aerobic or anaerobic biodegradation makes it necessary to identify sustainable treatment strategies for the removal of these dyes from wastewater. This indeed would be an important step undertaken to prevent the release of such recalcitrant and xenobiotic pollutants into the ecosystem. The various conventional methods that are used for dye effluent treatment include adsorption (Jesionowski et al. 2012), coagulation (Joo et al. 2007), electro-coagulation (Mook et al. 2017), reverse osmosis (Wang et al. 2018), and bio-degradation (Bouraie and Din 2016), etc. Due to the high stability and low biodegradability of dyes, these methods have several shortcomings such as incomplete degradation, high operational and chemical costs, limited applicability, and handling issues of sludge production as secondary pollutants. Table 1 summarises the comparative study of some most commonly used methods of dye treatment.

Table 1 Comparison of commonly used treatment technology for dye wastewater treatment
Full size table

In the last few decades, advanced oxidation processes (AOPs) have been significantly developed and extensively used to treat persistent organic pollutants (POPs) present in complex industrial effluents. These processes are more efficient and effective than conventional physicochemical methods (Brillas and Martínez-Huitle 2015). In AOPs, the two methods; anodic oxidation (AO); and electro-Fenton (EF) oxidation are frequently used due to their ample in-situ hydroxyl radical (HO·) production. These oxidation methods have a greater redox potential (Eo = 2.80 V vs. SHE) and a higher capacity to oxidize pollutants (Sirés et al. 2014).

The EF oxidation method involves the coupling of the Fenton reaction with electrochemistry. The technology has been reported as promising and efficient for various POP treatments such as dyes, leachate, oily waste, medical waste, and pesticides (Gökkuş and Yıldız 2015; Nidheesh and Gandhimathi 2012). It attracted attention owing to several reasons like lower operational cost, high HO· generation, wide pollutants treatment range, efficiency adjustment (with additional HO· generation), and minimum sludge production (reproduction of Fe2+ from cathodic Fe+3 reduction) (Eq. 1):

$$Fe^{3 + } + e^{ - } \to Fe^{2 + }$$
(1)

The reaction mechanism of the EF process depends on the selected pollutants and used electrode materials. For the most common mechanism, hydrogen peroxide (H2O2) is electrochemically generated on the cathode in continuous chain mode by electrochemical reduction of atmospheric O2 under acidic condition (Eq. 2):

$$O_{2} + 2H^{ + } + 2e^{ - } \to H_{2} O_{2}$$
(2)

Further, with small amount of iron catalyst (Fe2+) present, H2O2 generates HO·, which oxidizes the complex POP into simple molecular products (Eqs. 3 and 4).

$$H_{2} O_{2} + \, Fe^{2 + } \to Fe^{3 + } + OH^{ - } + HO \cdot$$
(3)
$$POP + HO \cdot \to Simple\;\;Molecular\;\;Products + CO_{2} + H_{2} O$$
(4)

However, in the case of electrochemical reactions, the most critical aspect is the possibility of higher energy consumption due to the occurrence of parallel competitive reactions in solvent. For example, the simultaneous occurrence of oxygen evolution reaction in the solvent retards the desired oxidative reaction and causes abatement in current efficiency. Therefore, it is very important to select a suitable electrode material, for a higher oxygen evolution reaction, that can result in a higher Faradic yield. To cope with the problem, carbonaceous electrode materials are most commonly used because of their low catalytic activity for H2O2 decomposition (Daneshvar et al. 2008). There are various carbonaceous materials tested for H2O2 generation like carbon felts (Çelebi et al. 2015), carbon sponge (Özcan et al. 2008), carbon nano-tubes (Gao et al. 2015), graphite felt (Daneshvar et al. 2008), graphite paper (Fernández de Dios et al. 2015), and graphite plates (Nidheesh et al. 2014). On the flip side, another method that is also used to cut down energy utilization is statistical process optimization; which reduces the number of experimental runs and operational costs (Gökkuş and Yıldız 2015).

In the study, we run the batch-mode electro-Fenton process using RO16 as the model pollutant dye. Carbon sponge anode and graphite plate cathode were used as carbonaceous electrodes because of their high electrocatalytic activity. In literature, various studies have been reported for dye treatment, but mostly they focused on parametric analysis. Here, we aimed to study the process from the fundamental operational parameter to field operational suitability by optimizing the process with the Response Surface Methodology (RSM) statistical tool. The EF process was first studied for its key operational factors effect (pH, initial RO16 concentration, total electrolysis time, and current density) on decolorization and COD removal. Further, with RSM, the multi-parameter optimization was performed. The method significantly reduces the number of experiments with a subsequent reduction in experimental cost; with the additional benefit of specifying the most significant and least/adversely affecting operational factors. The study also presents the RO16 decay kinetic and the EF process total operational costs analysis in the last section.

2 Materials and Methods

2.1 Chemicals

RO16 dye was purchased from Sigma Aldrich, USA. Sodium hydroxide, NaOH (Pellets, 98% purity), Sulfuric acid, H2SO4 (98% purity), and Sodium chloride, NaCl (Pellets, 99% purity) were arranged from Merck Co., Germany. Acetone, (CH3)2CO (purity > 99%) and Hydrogen chloride, HCl (purity > 55%) were used for electrode cleaning. All the chemicals used were of analytical grade (AR).

2.2 Experimental Arrangement

The batch mode experiments were performed in a 500 mL volume undivided Pyrex glass beaker at room temperature. Porous carbon sponge anode and solid graphite plate cathode (surface area = 20 cm2; 5 cm × 4 cm) electrodes were used with the fixed mutual spacing of 2.0 cm (Fig. 1).

Fig. 1
figure 1

Lab-scale experimental setup of the electro-Fenton process

Full size image

For current density, DC power supply (0–30 V, 2.5 A) was used. NaCl (0.4 M) and FeSO4.7H2O (0.5 mM) were used as supporting electrolytes and iron catalysts (Fe2+), respectively. For pH value adjustment solutions of 0.1 M NaOH and 0.1 M HCl were applied. Magnetic stirrer (rpm = 300; Witeg MSH-30A 230 V Germany) and Aerator (flow rate = 1.2 L/min Sobo, Model: SB333A, China) were used for homogenous mixing and oxygen supply, respectively.

2.3 Analytical Procedures

The pH of the samples was measured by pH meter (Hack, Germany). Percentage decolorization was evaluated through a spectrophotometer (UV-3000 LAB INDIA, Germany) at λmax by measuring the absorbance. Percentage COD removal was calculated as per standard methods (APHA 2005) before and after the treatment. The process performance efficiency and electrical energy consumption (EE, kWh/m3) were calculated according to Eqs. (5) and (6) (Ahmad and Basu 2022a):

$$\% Degradation = \frac{{C_{o} - C_{t} }}{{C_{o} }}$$
(5)
$$EE({\text{kWh}}/{\text{m}}^{3} ) = \frac{VIt}{{V_{s} }}$$
(6)

where Co and Ct are concentrations of RO16 (mg/L) before and after the treatment, respectively; V is the cell voltage (V); I is the applied current; t is the electrolysis time (min); Vs (m3) is the treated dye volume.

2.4 Design of Experiment (DOE)

Design-Expert software (version 11.1.2.0) was used for the design of the experiment (DOE). response surface methodology (RSM) was used to optimize the process parameters. The performance of the EF process was evaluated in terms of percentage removal of decolorization and COD. Four operational parameters (pH, electrolysis time, pollutant concentration, and current density) were investigated by using a central composite design (CCD). A five-level coded (− α, − 1, 0, 1, + α) experimental design was formulated with the selected ranges of design as given in Table 2.

Table 2 Experimental variables with their range and levels
Full size table

In CCD, three categories of experimental runs were involved i.e., center, axial, and factorial. The center runs are used to determine the experimental error; axial points ensure the model reliability; and the factorial points are the constant equidistant points from the center of the design. The total number of experimental runs was computed by using Eq. (7) (Owolabi et al. 2015):

$$\begin{aligned} Total\;\;numbers\;\;of\;\;runs \, \left( N \right) \, & = \, Factorial\;\;run \, + \, axial\;\;run \, + \, center\;\;run \\ & = \, 2^{n} + 2n + n_{c} \\ & = \, 2^{4} + 2\left( 4 \right) + 6 = 30 \\ \end{aligned}$$
(7)

where n, nc are the total number of independent factors and centre points, respectively. The model suggested a total of 30 experimental run requirements for the study (16 factorials, 6 axials, and 6 centers).

2.5 Analysis of Variance (ANOVA)

With ANOVA, the validations of the model and mutual effects of operational parameters were determined. In ANOVA, the statistical parameter p < 0.05 determined the significant level of the polynomial equation; whereas the F-test verified the model fitting. With correlation factor (predicted and adj. R2) the validity of model fitting was checked (Ahmad and Basu 2022b). For the study, the quadratic equation was found as the best-fitted equation for the experimental data (Table 3).

Table 3 ANOVA results for decolorization and COD removal
Full size table

2.6 Pareto Analysis

The Pareto graph is used for the interpretation of significant findings of results. The graph shows the percentage effect of each operational factor on the response variables and calculated as Eq. (8) (Ahmadzadeh et al. 2017):

$$P_{i} = \left(\frac{{b_{i}^{2} }}{{\Sigma b_{i}^{2} }}\right) \times 100\quad \left( {{\text{for}}\;\;{\text{all}}\;\;{\text{values}}\;\;{\text{of}}\;\;i \ne \, 0} \right)$$
(8)

where \(b_{i}^{2}\) is the estimation of the square root of the operational factor and \(\Sigma b_{i}^{2}\) is the sum of the square root of all the operational factors.

2.7 Oxidative Degradation Kinetics

In the EF oxidation process, the rate of oxidation by free HO· radical can be given as Eq. (9):

$$r = - \frac{dq}{{dt}} = k_{OH} q_{OH} q_{POP} + \sum\limits_{i}^{m} {k_{Oth} q_{Oth} q_{Oth} }$$
(9)

where q is the concentration of particular reactants and Oth represents the possible availability of other oxidants such as hydro-peroxyl radical (HOO·). A modified Eq. (10) after including the apparent rate constant (kapp) with HO· concentration can be re-written as (Duc et al. 2021):

$$r = - \frac{dq}{{dt}} = k_{app}^{\prime} q_{POP}$$
(10)

Integrating the Eq. (10), we got the pseudo 1st order equation as:

$$ln \, \frac{{q_{o} }}{q}{ = }k_{app}^{\prime} \times t$$
(11)

where q and qo represent the concentration at initial and after time interval t, respectively; and \(k_{app}^{\prime }\) is apparent pseudo 1st order rate constant.

In the case of a pseudo 2nd-order equation reaction, the equation will be:

$$r = - \frac{dq}{{dt}} = k_{app}^{\prime \prime } q_{POP}^{2}$$
(12)

Again, integration of Eq. (12) gives the final equation for pseudo 2nd-order as:

$$\frac{t}{q} = \frac{1}{{k_{app}^{\prime \prime } q_{e}^{2} }} + \frac{t}{{q_{e} }}$$
(13)

where qe is the equilibrium concentration and \(k_{app}^{\prime \prime }\) is the apparent pseudo 2nd order rate constant. The rate constants \(k_{app}^{\prime }\) and \(k_{app}^{\prime \prime }\) can be obtained by plotting ln (qo − q) versus t and t/q versus t graphs.

3 Results and Discussion

3.1 Model Fitting and Statistical Analysis

The experimental results for responses were modelled with Design-Expert software (version 11.1.2.0) and second-order polynomial equations were obtained (Eqs. 14 and 15).

$$\begin{aligned} Decolorization\;\;rate\left( \% \right) & = 71.50 + 2.51A + 7.08B - 4.42C \\ & \quad+ 11.25D - 8.23A^{2} - 3.98B^{2} \\ & \quad + 4.02C^{2} - 0.8542D^{2} + 0.3750AB \\ & \quad- 1.62AC - 0.2500AD + 1.50BC + 0.3750BD - 2.87CD \\ \end{aligned}$$
(14)
$$\begin{aligned} COD\;\; \, Removal\left( \% \right) & = 60.50 + 1.59A + 7.13B - 4.88C + 12.04D\\ & \quad - 6.97A^{2} - 2.97B^{2} + 5.28C^{2} \\ & \quad - 1.22D^{2} + 0.1875AB - 1.56AC - 0.9375AD \\ & \quad+ 0.1875BC + 0.3125BD - 3.44CD \\ \end{aligned}$$
(15)

These polynomial model equations explain the interaction of influencing parameters on decolorization and COD removal for the EF process. In the model equation, the parameters having favourable effects on the responses are represented with positive terms; while negative terms showed adversely affecting parameters. The complete CCD matrix, with the experimental results of response values of decolorization (%) and COD removal (%), is given in Table 4.

Table 4 The design matrix, predicted and observed values for the response variables
Full size table

For validation of the selected 2nd-order polynomial model, ANOVA was conducted with five parameters to check the variability of data between experimental and predicted results: (1) the Prob. > F to check the significant level of the model that should be < 0.0500; (2) the lack of fit test, to check the model fitting that should be > 0.0500; (3) adequate precision test to check signal-to-noise (S/N) ratio that should be ≥ 4; (4) adjusted regression coefficient (Adj. R2); and (v) regression coefficient (R2) that should be closer to 1 (Ahmad and Basu 2022a, b). The results showed that the response values met the model validation and accuracy criteria and can be used to anticipate the EF process performance (Table 5).

Table 5 ANOVA result of the EF process
Full size table

3.2 Significance of Operational Parameters

3.2.1 Pareto Graph

In the Pareto graph, the current density (D) was found to have the highest percentage contribution for the decolorization (%) and COD removal (%) with a total of 40.8% and 45.06%, respectively. The interaction effect of pH (A2) and electrolysis time (B) has a significant contribution to decolorization (38%) and COD removal (30.9%) showing a contributory effect on the process performance. For both responses, the total substantial contributions for the above three factors were observed as 65.76% and 69.85%, respectively. Factors A, C, CD, and B2 have shown less significant contributors whereas AB, AC, AD, BC, BD, and D2 were observed not to contribute any significant effect on the process performance (Fig. 2). The observed result for current density was supported by Daneshvar et al. (2007) work on Acid Yellow 23 removals; whereas the pH and electrolysis time contribution effects were also reported by Bashir et al. (2019) and GilPavas et al. (2019) work on palm oil mill effluent and Textile wastewater treatment.

Fig. 2
figure 2

Pareto graph for operational parameters

Full size image

3.2.2 Perturbation Plots

For comparing the effect of the different input factors on the response values at any particular point of design space, perturbation plots were drawn. These plots interpreted the process performance for variation of an operational factor within the operational range, with other factors as constant. The sensitivity of response for any operational factor can be judged by the steepness of the curvature of the plot. In our study, two perturbation charts were plotted at the optimal point of design space (A = 3.5; B = 42.5 min; C = 135 mg/L; and D = 17.5 mA/cm2) for the response values of decolorization and COD removal (Fig. 3). The steepness of plots showed that the process was most significantly affected by current density (D), and least by dye concentration (C) for both observed responses of decolorization and COD removal. The results were supported by Akhtar et al. (2020) work on Congo Red dye removal by the electro-coagulation process.

Fig. 3
figure 3

Perturbation plots for a demineralization (%); and b COD removal (%)

Full size image

3.3 Parametric Discussion

3.3.1 Effect of pH Value

At high dye concentrations (200 mg/L), lower pH values showed enhancement in decolorization from 62.3 to 73.5%, which gradually reduced to 63% as the pH > 3.5. At lower dye concentration (70 mg/L), the decolorization initially increased from 42.3 to 75% and further decreased to 73% for pH > 3.5 (Fig. 4b). Similar trends were also obtained for COD removal (%) with a reduction in the result of approximately 10% and 1% for high (200 mg/L) and lower dye concentration (70 mg/L) for pH > 3.5, respectively (Fig. 5b). The obtained result was supported by the result reported by Wang et al. (2014) for Methylene Blue dye treatment. In general, a pH in the range of 2.5–3.5 is often reported to be more favorable for the removal of dyes in the EF process (Ahmad and Basu 2023). A higher pH leads to more ferric ions (Fe2+) precipitation in aqueous solution; whereas natural or lower pH causes the accumulation of hydroxyl ions by H+ ions (Hodaifa et al. 2013; Nieto et al. 2011). In both of the cases, an overall decrement in the oxidative power of the EF process occurred.

Fig. 4
figure 4

Operating factors interaction effect on decolorization (%)

Full size image
Fig. 5
figure 5

Operating factors interaction effect on COD (%) removal

Full size image

3.3.2 Effect of Current Density

From the 3D plots of the current density effect on decolorization and COD removal, it can be said that there was an overall increment in dye degradation with an increase in current density. The possible reason could be that at higher voltage more free electrons boost the rate of Fe+2 regeneration (Eq. 1) which subsequently increases the free HO· and H2O2 production (Moreira et al. 2017).

Figures 4c) and 5c explain the effect of current density on decolorization and COD removal, respectively. The decolorization showed an incensement with an increase in current density, independent of the pH value. However, this effect was less significant for COD removal. At a higher pH value (> 3.5), an increment in current density from 10 to 25 mA/cm2 increases the decolorization (%) from 50.53 to 73.69%; while a lower pH value (< 3.5), it was 31.61 to 73.69%. Wang et al. (2014) also reported a similar increment in color removal efficiency from 56.1 to 70.8%, for an increase in current density from 11 to 80 A/m2 in their EF process of real Textile wastewater treatment.

3.3.3 Effect of Initial Dye Concentration

The EF process effect for dye concentration has found to work better at lower values (Figs. 4d, f and 5d, f)). At higher dye concentrations, the higher formation of intermediate products reduced the amount of HO· radicals that attack directly the targeted dye pollutants, decreasing overall efficiency (García-Rodríguez et al. 2016). Also, at higher dye concentrations, the consumption of free radical HO· increases more than its production, leading to a deficiency of available free radicals (HO·) (Yousefi et al. 2018).

3.3.4 Effect of Electrolysis Time

The results on process performance for the effect of electrolysis time are shown in Figs. 4 and 5. It was found that an increment in total electrolysis time favoured the process efficiency. Figure 4e and 5e showed that for a constant current density, the decolorization increased from 70.36% (25 min) to 85.16% (60 min) and COD removal increased from 60% (25 min) to 75.85% (60 min). The possible reason could be an increment in electrolysis time caused more HO· production that caused more H2O2 decomposition subsequently (Özcan et al. 2008). For a variable pH range, this increment was up to a pH of 3.5, after that graph showed a reverse trend (Figs. 4a and 5a). This was due to the possible reason of Fe2+ precipitation in the aqueous solution, that consequently decreased HO· production and reduced the treatment efficiency as discussed earlier in Sect. 3.3.1.

3.4 Statistical Model Validation

The result showed a good agreement between predicted and experimental values Fig. 6 a, b. Further, diagnostic curves for hypotheses of mathematical models were drawn to check the model adequacy Fig. 7. In Fig. 7a and b, colored points showed the residuals, and their closeness to the bisector line confirmed the normality of residuals (Dehghan et al. 2018). In Fig. 7c–f the plots of residuals showed a random distribution without any particular trend line confirming the 2nd-order model suitability (Yousefi et al. 2018). Therefore, the plots for residuals are normal and acceptable for response values prediction.

Fig. 6
figure 6

Predicted versus actual values plot for a decolorization (%) and b COD removal (%)

Full size image
Fig. 7
figure 7

Diagnostics plots for decolorization and COD removal: a and b for normal probability test, c and d for residuals versus run number, and (e and f) for residuals versus predicted values

Full size image

4 Kinetic Study

Pseudo 1st and 2nd order kinetic model was used to test the best fitted equation for experimental data. The kinetic constants values for 1st and 2nd order kinetic models were obtained at different RO16 dye concentrations (Table 6). The result showed a better fitment of experimental data with the pseudo 2nd order model because of the higher regression coefficients (R2) values. Also, with higher RO16 concentration, the decreasing R2 value represents a possible increment of competitive reactions among the HO· and other oxidative by products (Mansour et al. 2015).

Table 6 Kinetic constants at different initial RO16 dye concentration (A = 3.5, C = 45 min, D = 17.5 mA/cm2)
Full size table

Figure 8 represents the relationship between rates of reaction (\(k_{app}{\prime}\) and \(k_{app}^{^{\prime\prime}}\)) with RO 16 dye concentration (qPOP). Accordingly, the kinetic models for RO16 dye degradation were given by Eqs. (16) and (17):

Fig. 8
figure 8

Pseudo 2nd order kinetic model for RO16 degradation (A = 3.5; B = 1000 mg/L; D = 17.5 mA/cm2)

Full size image

Pseudo 1st order model:

$$r = - \frac{dq}{{dt}} = 0.0094q_{POP}$$
(16)

Pseudo 2nd order model:

$$r = - \frac{dq}{{dt}} = 0.0806q_{POP}^{2}$$
(17)

Therefore, the rate of the pseudo 2nd order model was 8.57 times faster than the pseudo 1st order model.

5 Cost Analysis

The operational cost (OC) plays a major constraint on the application suitability of any treatment technology. It may include the cost of electrical energy (EE), chemical, electrode consumption (EC), etc. For the current study, the total OC at optimum condition was calculated as $0.54 m−3 for the commercial Indian electric price rate of $0.14 kWh−1. The chemical cost was $1.23 m−3 ($2.58 kg−1 for H2SO4; $2.06 kg−1 for NaCl; $3.6 kg−1 for NaOH; and $0.05 kg−1 for FeSO4.7H2O) which was about 66% of the total cost. In Zheng and Lefebvre (2019) study, they reported that the total OC for the majority of the homogeneous EF process can be distributed as electrical energy (29%), chemical cost (64%), and remaining (7%) cost for others such as electrode consumption (EC) cost. Using this study result, the other cost (EC) for the electrode materials consumption was calculated and overall results are reported in Table 7. The result concludes that the EF process can be used as a substitution method over other processes such as electro-coagulation that cost highly because of huge chemical consumption.

Table 7 Various operational costs analysis
Full size table

6 Conclusion

A batch-mode experimental study and optimization of the EF process was done with Response surface methodology (RSM). For the selected operational parameters, the optimized result was obtained as pH (A) = 3.5, electrolysis time (B) = 42.5 min, initial dye concentration (C) = 135 mg/L, and current density (D) = 17.5 mA/cm2, with decolorization and COD removal as 72% and 61% respectively. For the EF process efficiency, current density (D) affects the process most; whereas the higher RO16 dye concentration (> 135 mg/L) adversely affects the process. The ANOVA result showed 2nd order polynomial model suitability for decolorization (R2 = 0.9863) and COD removal (R2 = 0.9530) and diagnostic plots validated the developed statistical models. The Pareto graph ordered the effect of operational parameters as: current density > pH > electrolysis time > initial RO16 dye concentration. Also, the Perturbation plots support the result at optimized conditions. The kinetic study was best suited by 2nd order polynomial model (R2 = 0.9878). Also, the overall cost ($1.9 m−3) at the optimized conditions favoured an economical treatment of dye pollutants. Future research can be aimed at using other optimization tools (Taguchi design and Artificial Neural Networks) to further analyze and compare the result for the best optimization tool selection and to better understand the lab-scale EF process experimental design, particularly adapted for the economical treatment of Textile Industrial pollutants.