The electrosynthesis was accomplished according to the 2⁴ factorial shown in table I. The variables of temperature, current density, and the type and concentration of the electrolyte were analyzed and contrasted using the Yates algorithm and the “F” test to determine which had the main effect on the electrosynthesis of ferrate. The experiments were carried out in an electrochemical batch cell using iron as the anode and BDD as the cathode; the anodic surface area was 12 cm². 30 mL of the electrolyte was placed in the electrosynthesis reactor, and the current intensity was controlled using a power supply and a digital multimeter. The concentration of ferrate was determined by measuring the absorbance at 505 nm using a Varian Cary E1 spectrophotometer, and applying a molar absorption coefficient of 1150 1/M/cm (Lee et al. 2005, Ramseier Peter et al. 2011, Sharma 2013).

TABLE I

Factors	Symbol	Level 1 (-)	Level 2 (+)
Electrolyte type	A	NaOH	KOH
Electrolyte concentration (M)	B	14	20
Current density (mA/cm2)	C	30	100
Temperature (ºC)	D	4	RT

(-) low level, (+) high level, (RT) room temperature

Ferrate(VI) was characterized by three analytical techniques: a) infrared spectroscopy (IR) to observe the characteristic functional groups at 324 and 800 1/cm; b) cyclic voltammetry (CV), an electroanalytical technique used for the determination of compounds that are oxidized or reduced electrochemically, and c) ultraviolet-visible spectroscopy (UV-Vis), which provides information via the characteristic absorption peak at 505 nm.

Indigo blue dye was obtained from an industrial dyeing operation located in the State of Mexico, and was characterized using IR spectroscopy and UV-Vis spectroscopy in a previous report (Almazán-Sánchez et al. 2016). For this, 10, 30, 40, 50, and 100 mg/L aqueous solutions of the dye were prepared (pH = 8.0), and the concentration was determined by UV-Vis spectrometry at 591 nm, based on the previous results.

A wastewater sample from the effluent of an industrial textile operation located in the State of Mexico was collected in plastic containers. The following parameters of the sample were analyzed: temperature, biochemical oxygen demand (BOD₅), chemical oxygen demand (COD), total organic carbon (TOC), color (Pt-Co units), turbidity, pH, sulfates (SO₄ ^2-), ammonia nitrogen (NH₄ ⁺), total suspended solids (TSS), methylene blue active substances (MBAS), nitrates (NO₃ ^-), nitrites (NO₂ ^-), total phosphorus, total solids (TS), total dissolved solids (TDS), oil and grease, residual chlorine, magnesium, sodium, and calcium, according to standard methods and procedures (APHA 2012). The sample was also analyzed using IR spectroscopy and UV-Vis to determine the mechanism of dye removal. The same parameters were analyzed in the treated wastewater in order to evaluate the efficiency of ferrate.

A jar test was used to determine the optimum time and ferrate dose for the treatment of the aqueous solutions and denim wastewater. Batch experiments were carried out by combining 8 mL of the indigo blue aqueous solutions with different initial concentrations (10, 30, 40, 50, or 100 mg/L) with 25, 50, 100, 200, or 300 mg/L of previously synthesized potassium and sodium ferrate. The solutions were mixed rapidly (250 rpm) for 1 min, and then at a lower rate (50 rpm) for 26 h. Aliquots were collected at different times (1, 2, 3, 4, 5, 6, 7, and 26 h) and the concentration of indigo blue was evaluated in order to calculate the removal percentage at the end of the treatment. For the denim wastewater, the treatments were carried out under the same conditions, and the color removal efficiency was evaluated after 1 h. All experiments were carried out in triplicate.

The effect of different variables in the electrosynthesis of ferrate was evaluated using a 2⁴ factorial design. The effect of each factor was analyzed separately in order to determine its influence on the process. Table II shows the optimal operational conditions for the electrochemical processes. According to the results, the best ferrate yield was obtained in experiment 15, using 20 M NaOH, and applying a current density of 100 mA/cm² at room temperature (25 ºC).

TABLE II

Experiment	Interaction	Factors			(FeO4)2- (g/L)		Energy consumption (kWh/kg)
		A	B (M)	C (mA/cm2)	D (ºC)
1	M*	NaOH	14	30	4	0.20	10.86
2	A	KOH	14	30	4	0.04	44.86
3	B	NaOH	20	30	4	1.42	1.53
4	AB	KOH	20	30	4	0.05	37.81
5	C	NaOH	14	100	4	0.81	17.03
6	AC	KOH	14	100	4	0.15	90.43
7	BC	NaOH	20	100	4	0.58	12.50
8	ABC	KOH	20	100	4	0.33	27.24
9	D	NaOH	14	30	RT	1.32	1.65
10	AD	KOH	14	30	RT	0.25	5.69
11	BD	NaOH	20	30	RT	2.31	0.63
12	ABD	KOH	20	30	RT	0.20	6.89
13	CD	NaOH	14	100	RT	1.38	4.43
14	ACD	KOH	14	100	RT	0.42	21.94
15	BCD	NaOH	20	100	RT	6.89	2.01
16	ABCD	KOH	20	100	RT	0.31	44.49

* M is the standard order. It refers to the experiment in which all the variables are in the low level, RT = room temperature

Figure 1 shows that the highest production of ferrate was achieved by applying a current density of 100 mA/cm², with the yield of ferrate reaching 6.89 g/L (41.54 mM). At a current density of 30 mA/cm², only 2.31 g/L (13.92 mM) of ferrate was synthesized using NaOH as the electrolyte, as shown in table II. Many studies have reported the synthesis of ferrate under different experimental conditions, some of which obtained 2.03, 1.91, and 1.74 mM of ferrate in 20 M NaOH media with current densities of 4.55, 3.08, and 1.47 mA/cm², respectively (Barışçı et al. 2014). Alsheyab et al. (2010) achieved concentrations in the range 0.54-3.25 mM by applying current densities from 1.8 to 18.1 mA/cm² in 14 M NaOH. Other authors achieved yields of 7.3 and 1.5 mM at 3.3 and 0.5 mA/cm² in 16 M NaOH (Ding et al. 2004). Finally, a concentration of approximately 0.23 mM was achieved at 125 mA/cm² using 14 M KOH (Sánchez-Carretero et al. 2010).

Fig 1

Different mechanisms have been proposed to explain the process of ferrate production in the electrochemical cell; equations 1-7 (Mácová et al. 2009) describe anodic dissolution and involve the formation of a passivating layer over the electrode, reducing the ferrate yield. This film consists of FeOOH and Fe(OH)₂ (Eq. 2-3), which block the flow of electrons. Additionally, Fe₃O₄ (Eq. 1) can act as a barrier against further iron dissolution and prevent the formation of ferrate; it also competes with oxygen evolution (Eq. 7). The hydroxyl ions promote the reaction by dissolving the oxide/hydroxide layer, as shown in equations 4-6. Hydrogen gas is produced at the cathode in the global reaction (Barışçı et al. 2014). The electrochemical reactions generated at the anode and cathode are shown in figure 2 (Zou and Chin 1988, Jiang and Lloyd 2002, Alsheyab et al. 2010). The sacrificial iron anode is oxidized from iron (0) to iron (VI) in the presence of hydroxyl ions, and simultaneous oxygen evolution and gas production take place. The energy consumption of the reaction was 2.01 KWh/kg, which was in agreement with the report of Barışçı et al. (2014).

3 F e ° +   8 O H -   →   F e 3 O 4 +   4 H 2 O +   8 e - (1)

F e ° +   2 O H -   →   F e ( O H ) 2 +   2 e - (2)

F e ( O H ) 2 +   O H -   → F e O O H +   H 2 O +   e - (3)

F e O O H + O H -   →   F e O 2 -   +   H 2 O (4)

F e O 2 -   +   2 O H -   →   F e O 3 2 -   +   H 2 O   +   e - (5)

3 F e O 3 2 -   +   H 2 O   →   2 F e O 2 -   +   F e O 4 2 -   +   2 O H - (6)

2 O H -   →   H 2 O   +   1 2 O 2   +   2 e - (7)

Fig 2

The type of electrolyte used affects iron dissolution, the formation of the oxide/hydroxide layer, and the solubility of the dissolution products. Some researchers have suggested that NaOH is the best electrolyte for the electrochemical preparation of ferrate (Bouzek et al. 1999, Bouzek and Bergmann 1999, Ding et al. 2004, Xu et al. 2009, Barışçı et al. 2014), while others have reported that better electrosynthesis results were obtained using KOH (He et al. 2005). Na₂FeO₄ remains soluble in saturated aqueous NaOH solutions, but K₂FeO₄ is insoluble in a saturated KOH solution (Jiang and Lloyd 2002). In this study, NaOH was found to be the best alkaline medium, as shown in figure 3a and 3b. When KOH was used, only 0.42 g/L of ferrate was produced.

Fig 3

According to the previous equations, ferrate production increases with increasing HO^- ion concentration. A concentrated medium increases the solubility of lower-valence iron species and promotes ferrate stability (Zou and Chin 1988, He et al. 2006, Mácová et al. 2009). When KOH is used at high concentrations (over 500 g/L), the oxides and hydroxides formed crystalize on the anode surface (Bouzek et al. 1999). In this research, concentrations of 14 M and 20 M were evaluated. Using 20 M NaOH, the maximum ferrate concentration of 6.89 g/L was obtained (Fig. 3c), while a concentration of 1.38 g/L was obtained using 14 M NaOH (Fig. 3d).

Changing the temperature could induce the depassivation of the anode surface. The influence of temperature on electrochemical ferrate production has previously been reported by several researchers (Beck et al. 1985). For instance, Barışçı et al. (2014) studied the effect of using temperatures of 20, 30, 50, and 65 ºC with high-purity iron anodes. The highest production rate after 75 min was obtained at 65 ºC, and the lowest production rate was observed at 20 ºC. They concluded that higher temperatures led to better production in ferrate synthesis using high-purity iron anodes. In this work, similar behavior was observed; the maximum ferrate concentration was achieved at room temperature (25 ºC), whereas at 4 ºC, only 1.42 g/L was obtained (Fig. 4).

Fig 4

The results were validated using the Yates algorithm to explain the synergic effect of all the operational variables. The Yates equation obtained is as follows:

Y   =   1.03   -   1.63   A   +   0.98   B   +   0.54   C   +   1.09   D         (8)

Where:

A (type of electrolyte) = -1.63 g/L, B (concentration of electrolyte) = 0.98 g/L, C (current density) = 0.54 g/L, and D (temperature) = 1.09 g/L.

Based on the electrolyte type parameter (A), changing the electrolyte from NaOH electrolyte to KOH reduced the ferrate concentration. When the concentration of electrolyte (B) was increased from 14 to 20 M, the synthesis of ferrate was enhanced. Increasing the current density (C) from 30 to 100 mA/cm² had a positive effect on the electrochemical process, improving the production of ferrate. Increasing the temperature (D) from 4 ºC to room temperature (25 ºC) also enhanced the concentration of ferrate.

Table III shows the effects calculated by the Yates algorithm; through these data, the sum of squares, estimated coefficients, standard deviation and the “F” test were determined. Variables B, C and D have effects on both levels in the electrochemical synthesis of ferrate (VI); while the variable A (type of electrolyte) is optimal with NaOH at the low level.

TABLE III

Interaction	Effect	Sum of squares	Estimated coefficients	Standard deviation	“F” Test
M	1.03			0.44	2.96
A	-1.64	10.73	-0.82	0.44	-1.86
B	0.93	3.47	0.47	0.44	1.06
AB	-0.97	3.76	-0.48	0.44	-1.10
C	0.60	1.43	0.30	0.44	0.68
AC	-0.48	0.90	-0.24	0.44	-0.54
BC	0.41	0.66	0.20	0.44	0.46
ABC	-0.33	0.44	-0.17	0.44	-0.38
D	1.15	5.29	0.58	0.44	1.31
AD	-1.04	4.35	-0.52	0.44	-1.19
BD	0.65	1.71	0.33	0.44	0.74
ABD	-0.70	1.94	-0.35	0.44	-0.79
CD	0.63	1.60	0.32	0.44	0.72
ACD	-0.61	1.51	-0.31	0.44	-0.70
BCD	0.91	3.32	0.46	0.44	1.04
ABCD	-0.81	2.63	-0.41	0.44	-0.92

In conclusion, the maximum experimental concentration of ferrate produced in 180 min was 6.89 g/L, using 20 M NaOH, a current density of 100 mA/cm², and room temperature (25 ºC).

Ferrate (VI) has a characteristic purple-violet color, which corresponds to a visible and near-infrared absorption spectrum from about 450 to 600 nm. Figure 5a shows the UV/Vis spectra of electrogenerated Na₂FeO₄ and K₂FeO₄ in 20 M media. The spectra of both ferrates showed maximum absorption at 505 nm. These spectra were compared with that of a commercial ferrate, and were found to contain the same features. These results agree with those of Licht et al. (2001), Jiang and Lloyd (2002) and Barışçı et al. (2014), who all observed the same wavelength for ferrate.

Fig 5

In addition, ferrates have also been characterized using Fourier transform infrared (FTIR) spectra, in which they display characteristic absorption bands at 925 and 523 cm^-1 in the infrared region with triplet splitting (Fig. 5b; Sharma 2013).

Finally, the cyclic voltammetry (CV) technique was used to determine the redox behavior of the different compounds. Figure 5c shows the CV obtained at a scan rate of 100 mV/s using Ag/AgCl as the reference electrode, iron as the working electrode, and carbon paste as the counter electrode. Curve “a” shows only the 20 M NaOH. In curves “b,” “c,” and “d,” different concentrations of ferrate were added, and four peaks were obtained; the peaks numbered 1 and 2 correspond to the redox pair Fe²⁺/Fe³⁺. This pair exhibits a quasi-reversible one-electron charge transfer, while peak 3 is associated with the ferrate ion formation at -1.36 V (Licht et al. 2001, de Koninck and Bélanger 2003, Barışçı et al. 2014).

The dye removal percentages for the aqueous indigo blue solutions using Na₂FeO₄ and K₂FeO₄ are presented in figure 6. The maximum dye removal percentages (over 77 %) were achieved using Na₂FeO₄ (300 mg/L dose) for all dye concentrations, while only 94 and 96 % dye removal was achieved for 20 and 30 mg/L dye concentrations at a dose of 200 mg/L. Na₂FeO₄ showed the better results at high doses only (Fig. 6a). Based on these results, the removal percentage strongly depends on the dose of ferrate. Very high doses of ferrate proved to be the most effective in reducing the organic concentration, because for a given volume of solution, proportionally increasing the quantity of the reactants could promote the probability of collision between the target pollutant and the oxidant, leading to an increase in the degradation efficiency (Jiang 2007, Han et al. 2013).

Fig 6

As shown in figure 6b, K₂FeO₄ was superior at low dye concentrations (10, 20, and 30 mg/L) using intermediate doses. When the dye concentration was 10 mg/L, doses of 25, 50, and 100 mg/L of K₂FeO₄ gave the best dye removal results (76.9, 79.5, and 87.5 % respectively). Using dye concentrations of 20 and 30 mg/L and a ferrate dose of 200 mg/L, around 70 % color removal was achieved. When the ferrate dose was increased to 300 mg/L, 90 % color removal was achieved. Li et al. (2009) treated a 50 mg/L aqueous solution of an azo dye (orange II) with 10 mg/L K₂FeO₄ and 62 % color removal was obtained. Xu et al. (2009) treated the azo dye X-3B red in aqueous solution (initial concentration 0.08 mM) with 25 mg/L K₂FeO₄, achieving 95 % color removal at pH 10.

Ferrate ion is a strong oxidant over the entire pH range: Its redox potentials are 2.20 and 0.72 V in acidic and basic media (Eqs. 9-10), respectively.

F e O 4 2 - +   8 H + +   3 e - →   F e 3 + +   4 H 2 O     E o = 2.2   V   (9)

F e O 4 2 - +   4 H 2 O +   3 e - →   F e ( O H ) 3 +   5 O H -   E o = 0.7   V   (10)

The pH has a great effect on the decolorization of indigo blue dye by ferrate (Li et al. 2009, Xu et al. 2009) due to its effect on the redox potential and ferrate stability. The oxidation potential of ferrate decreases with increasing pH, because although ferrate is a more powerful oxidant at pH < 6, it is also highly unstable under these acidic conditions. Conversely, it becomes more chemically stable but has a weaker oxidizing ability at pH > 9. Thus, the optimum pH range is 6-9; the removal efficiency is low above and below this pH. In the experiments, the pH was increased to 11-12, as increasing the pH can cause the hydrolysis of Fe compounds to Fe(OH)_{3 (aq)} or colloids, which are capable of adsorbing dye molecules and removing them from the solution.

In the current research, testing was carried out at room temperature. However, in accordance with the report of Han et al. (2013), increasing the temperature of the reaction medium would enhance the probability of collision between the target pollutant and oxidizing agent, and thus accelerate the reaction rate.

Figure 7 shows the possible mechanisms for the oxidation of denim blue by ferrate. In the first step, the destabilization of the first N-H bond occurs in the presence of ferrate, free radicals are formed, and ferrate is reduced from Fe(VI) to Fe(V). In the next step, the destabilization of the second N-H bond occurs through FeO₄ ^-, and a conjugated double bond is formed. Then, the decomposition of FeO₄ ^- to Fe(OH)₃ and O₂ in the basic media (OH^-) occurs, and the interaction between oxygen and water breaks the conjugated bond, regenerating the basic media. Finally, organic by-products are formed. Oxidation and coagulation are thus the main mechanisms of the dye removal process.

Fig 7