Introduction

Electro-Fenton (EF) is one of the powerful and widely used advanced oxidation processes (AOPs) for the treatment of industrial wastewater containing non-biodegradable organic pollutants (Brillas and Martinez-Huitle 2015; Martinez-Huitle et al. 2015; Sirés et al. 2014). The EF system can continuously supply H2O2 through the two-electron reduction of oxygen on the cathode (Eq. 1). Meanwhile, Fe2+ is added to react with H2O2 for producing ·OH following Eq. (2) with the Fe2+ regeneration by a direct cathodic reaction (Eq. 3) (Brillas et al. 2009).

$$ {\mathrm{O}}_2+2{\mathrm{H}}^{+}+2{\mathrm{e}}^{\bar{\mkern6mu}}\to {\mathrm{H}}_2{\mathrm{O}}_2 $$
(1)
$$ {Fe}^{2+}+{\mathrm{H}}_2{\mathrm{O}}_2\to {Fe}^{3+}{+}^{\cdot }{OH+OH}^{-} $$
(2)
$$ {Fe}^{3+}+\mathrm{e}\bar{\mkern6mu}\to {Fe}^{2+} $$
(3)

However, the homogeneous EF requires an optimum of pH = 3 and a certain amounts of Fe2+ and the effluent must be neutralized, which results in the generation of abundant sludge (Garcia-Segura et al. 2012; Wang et al. 2013b). So, heterogeneous EF oxidation has become prevalent for wastewater treatment, where soluble Fe2+ is replaced by Fe containing solids without the need of low pH and iron sludge treatment (Ai et al. 2007b; Ammar et al. 2015; Dhakshinamoorthy et al. 2012; Feng et al. 2011; Navalon et al. 2011; Zhang et al. 2012).

In the past decades, the potential applications of heterogeneous catalysts have been investigated widely. Various kinds of iron oxides and iron hydroxides, such as Fe3O4, α-Fe2O3, and α-FeOOH have already been used to activate H2O2. However, many of them show lower catalytic activity than soluble Fe2+ and need the aid of ultrasound (Huang et al. 2012; Segura et al. 2012) and UV/visible light irradiation (Liang et al. 2012; Zhao et al. 2013), increasing the cost for wastewater treatment.

Recently, other ways have been developed to enhance the activity of heterogeneous catalysts, e.g., reducing the size of catalysts to nano-scale to increase the surface energy (Valdes-Solis et al. 2007), loading the catalysts on carriers with high surface area to improve their dispersion (Zhao et al. 2012). Many researchers also have developed new methods to obtain heterogeneous catalysts. For example, Wang et al. (2013a) used ferrite-carbon aerogel (FCA) monoliths with different iron/carbon ratios as heterogeneous catalyst for metalaxyl degradation. Wang et al. (2015b) used mesoporous copper ferrite (meso-CuFe2O4) as a heterogeneous EF catalyst for the degradation of organic contaminants.

In EF process, it is essential to choose an appropriate cathode material for effective production of H2O2. Carbonaceous materials are the most familiar materials used as cathode, such as graphite (Scialdone et al. 2013; Zhang et al. 2008a), carbon or graphite felt (Oturan et al. 2013; Panizza and Oturan 2011; Pimentel et al. 2008; Zhou et al. 2012), carbon sponge (Ӧzcan et al. 2008, 2009), activated carbon fiber (Wang et al. 2005), and carbon-PTFE air-diffusion electrode (Brillas et al. 2000). Among them, graphite felt (GF) has been regarded as one of the most widely used cathode materials due to their large 3D active surface, mechanical integrity, commercial availability, easy acquisition, and efficient cathodic regeneration of Fe2+ (Zhou et al. 2012). To further improve the electro-catalytic activity of these carbonaceous materials, various attempts have been made for a more efficient H2O2 production (Zhang et al. 2008b, c, 2009). Zhou et al. (2014) used ethanol and hydrazine hydrate to modify GF as cathode in EF system and showed the hydrogen peroxide accumulation increased by 160 % after modification. Bonakdarpour et al. (2011) reported metal-carbon composite catalysts for electro-synthesis of H2O2 and showed Co-carbon had the best catalytic activity. In our previous work, a simple method using carbon black and PTFE was firstly developed to modify graphite felt and after modification, the yield of H2O2 could be significantly increased by about 10.7 times (Yu et al. 2015a).

In the present work, we prepared metal–carbon graphite felt as cathode for EF process. The structure, morphology, and formation mechanism of the composite cathode were investigated. To evaluate its catalytic activity in the heterogeneous EF process, a model azo dye, methyl orange (MO), was chosen as the target pollutant.

Experimental

Transition metal-based catalysts and electrodes preparation

All chemicals used in this study were of analytical grade and used as received without further purification. The GF (Shanghai Qijie Carbon Material Co., Ltd.) were degreased in an ultrasonic bath with acetone and deionized water in sequence, dried at 80 °C for 24 h. Appropriate amounts of carbon black and metal nitrate salts were mixed in an ultrasonic bath for 30 min and dried overnight at 70 °C in an oven. The ferrous precursor consisted of Fe(NO3)2·6H2O with a 99.99 % purity. Similar nitrate salts were used for Co, Ce, and Cu as well. The mixture was heat treated in a ceramic tube furnace at 900 °C under flowing N2 for 2 h. The transition metal-based catalysts were marked as C-metal. Appropriate amounts of C-metal (0.3 g), PTFE, distilled water (30 mL), and n-butanol (3 %) were mixed in an ultrasonic bath for 10 min to create a highly dispersed mixture. The pretreated GF were immersed into the mixture and sonicated for 30 min and then dried at 80 °C for 24 h. At last, the samples were annealed at 360 °C for 30 min. The composite electrodes were marked as GF-C, GF-Fe, GF-Co, GF-Ce, and GF-Cu, respectively.

Electro-generation of H2O2 and MO degradation

The H2O2 electro-generation experiments were performed in an undivided cell (0.1 L) at the current density of 50 A/m2 with 0.05 M Na2SO4 as the supporting electrolyte at room temperature, stirring at 300 rpm with a magnetic bar. Ti/IrO2-RuO2 (4 × 2 cm) was used as anode and the modified GF (4 × 5 cm, thickness 5 mm) was selected as cathode. The distance between the anode and cathode was 2.5 cm. At time intervals, 1 mL samples were taken for analyzing the concentration of H2O2. The degradation of MO by EF process was carried out in the same apparatus at initial MO concentration of 50 mg/L at the current density of 50 A/m2. The solution pH was adjusted to 3 after 0.2 mM Fe2+ was added into the solution as the catalyst. The MO samples were taken to determine the degradation efficiency.

Analysis

Linear sweep voltammetry (LSV) was carried out to compare the electro-chemical behavior during H2O2 production. The performance was recorded by the CHI660D workstation (CH Instruments, Chenhua, Shanghai, China) at a scan rate of 10 mV/s in a three-electrode system (0.1 L), using the modified GF as the working electrode, Ti/IrO2-RuO2 as counter electrode, and saturated calomel electrode (SCE) as the reference electrode at ambient temperature. The surface structure of the electrode was studied by scanning electron microscopy (SEM; LEO-1530VP, Germany). The surface elemental composition of the as-prepared electrodes were determined by XPS (Perkin-Elmer, a RBD upgraded PHI-5000 C ESCA system, Physical Electronics, USA) with Mg K radiation.

The concentration of H2O2 was monitored by UV-Vis spectrophotometer (UV-2600, Shimadzu) at λ max = 400 nm using the potassium titanium (IV) oxalate method (Yu et al. 2015b). The current efficiency for H2O2 production was calculated according to the formula (4) (Brillas et al. 2009):

$$ CE\ \left(\%\right)=\frac{nFCV}{{\displaystyle {\int}_0^tI\mathrm{d}t}}\times 100 $$
(4)

where n is the number of electrons transferred for oxygen reduction for H2O2, F is the Faraday constant (96,485 C/mol), C is the concentration of H2O2 (mol/L), V is the bulk volume (L), I is the current (A), and t is the electrolysis time (s).

The absorbance (A) of MO was determined by the same spectrophotometer at a wavelength of 464 nm. The MO removal efficiency (η) was calculated using the following formula:

$$ \eta \left(\%\right)=\frac{A_0-{A}_t}{A_0}\times 100 $$
(5)

where A 0 and A t denote the absorbance of the MO solution at initial and given time, respectively.

The total organic carbon (TOC) of the final sample in the EF process was determined by a TOC analyzer (Analytikjena multi N/C 3100, Germany).

Energy-related parameters are essential figures of merit for comparing the viability of EF and related methods between them and against other electro-chemical technologies. Operating at constant I, energy consumption per unit volume is obtained from the formula (6) (Brillas et al. 2009):

$$ \mathrm{Energy}\;\mathrm{consumption}\;\left({kWh\ \mathrm{m}}^{-3}\right)=\frac{UIt}{V} $$
(6)

Where U is the average cell voltage (V), I is the current (A), t is the electrolysis time (h), and V is the bulk volume (L).

Results and discussion

Optimization of transition metal-based catalysts

In order to obtain the optimal mass ratio of the transition metal to carbon black for GF-metal, the performance of MO degradation was tested. As shown in Fig. 1, while metal content was low, with the increasing of mass ratio of the transition metal to carbon black, MO degradation improved, and then it decreased. The degradation efficiency reached the corresponding maximum of 99.2, 94.2, 89.5 and 70.1 % with 1 wt.% Co, 1 wt.% Fe, 1 wt.% Ce, and 0.5 wt.% Cu, but for GF-C and GF the degradation efficiency were only 35.5 and 12.6 % within 120 min. It indicated that the MO removal efficiency significantly increased after transition metals were loaded. This was mainly due to more ·OH production with the existence of transition metal ions in the solution (Ai et al. 2013). However, a further increase of the transition metals content might cause an inhibition for MO degradation due to the loss of ·OH by reaction with excess transition metal ion, take Fe2+ for example (Eq. 7) (Nidheesh and Gandhimathi 2012):

$$ {Fe}^{2+}{+}^{\bullet }OH\to {Fe}^{3+}{+OH}^{\hbox{-} } $$
(7)
Fig. 1
figure 1

Effect of metal loadings on the degradation of MO: a Co, b Fe, c Ce, and d Cu. Conditions, 0.05 M Na2SO4 and 50 mg/L MO; current density, 50 A/m2 (pH = 3); air flow rate, 0.5 L/min

Full size image

So 1 wt.% Co, 1 wt.% Fe, 1 wt.% Ce, and 0.5 wt.% Cu was selected as the best dosage for MO degradation, respectively.

In order to investigate electro-catalytic activity of the as-prepared electrodes, LSV was carried out. As shown in Fig. 2, all transition metal-based electrodes exhibited higher current response for O2 reduction than the GF without loading transition metal. This result indicated that the presence of transition metal was helpful for increasing the cathode conductivity and thus reducing the cathode potential. Among the transition metal-based electrodes, the GF-Co electrode showed the highest current response while the current response of GF-Cu electrode was the lowest. This trend was in agreement with that of MO removal, which is reasonable since the electro-chemical reaction rate is determined mainly by current (Yu et al. 2015b).

Fig. 2
figure 2

Linear sweep voltammograms. Conditions: scanning rate, 50 mV/s (0.05 M Na2SO4, pH 7), oxygen saturation

Full size image

H2O2 accumulation in the heterogeneous EF system

Since H2O2 production is very important for electro-Fenton process, it is necessary to identify the H2O2 production capacity of these transition metal-based cathodes. Figure 3a, b shows the accumulation of H2O2 with different metal loadings at pH of 3 and 7, respectively. After 120 min electrolysis, the concentration of H2O2 reached 554.8, 474.7, 454.1, 440, 380.5, and 35.6 mg/L at pH = 3 using GF-Co, GF-Fe, GF-Ce, GF-Cu, GF-C, and GF, respectively. Accordingly, the concentration of H2O2 reached 516.8, 442.7, 404.1, 378.6, 315.2, and 25.5 mg/L at pH = 7. Similarly, GF-Co had the highest current efficiency for H2O2 production (Fig. 3c, d), which reached 41 and 38 % in 2 h at pH 3 and 7, respectively. These results were in agreement with the LSV results (Fig. 2). The results indicated GF-Co had the highest catalytic activity for electro-reduction of O2 to H2O2. This was in agreement with other literatures (Bonakdarpour et al. 2011). There was a slight increase of the H2O2 accumulation at pH 3 because a low pH was favorable to H2O2 production (Eq. 1). In summary, the prepared GF-metal is a very good cathode material for H2O2 production and potential to be used in electro-Fenton process.

Fig. 3
figure 3

Effect of metal loading on (a, b) H2O2 production and (c, d) current efficiency. Conditions, 0.05 M Na2SO4; current density, 50 A/m2; air flow rate, 0.5 L/min (1.0 wt.% Co, 1.0 wt.% Fe, 1.0 wt.% Ce, 0.5 wt.% Cu)

Full size image

Characterizations of the prepared electrodes

Figure 4 shows the SEM images of unmodified GF and modified GF. Before the transition metal was loaded, the GF showed a clean fiber structure composed of an entangled network of carbon microfilaments with diameters around 15 μm. After the transition metal was loaded, a large numbers of interconnected particles appeared on the fiber of GF, which would obviously change the cathode surface characteristics. These carbon particles and porous structure on the electrode surface could promote O2 electro-sorption and electro-reduction and pollutants degradation (Jin et al. 2011; Wu et al. 2010). Figure 4c shows GF-Co had a more uniform surface with particles, which might render GF-Co having the highest catalytic activity.

Fig. 4
figure 4

SEM image of a GF, b GF-C, c GF-Co, d GF-Fe, e GF-Ce, and f GF-Cu

Full size image

The surface element of the modified GF was studied by XPS analysis. Compared with GF-C (Fig. 5a), not only C and O elements but also iron element was observed and the ratio between O and C (O/C) increased in GF-Fe (Fig. 5b), which indicated that the number of oxygen-containing functional groups increased after Fe was loaded. The F element was also detected, which was probably due to the addition of PTFE during modification.

Fig. 5
figure 5

XPS of a GF-C and b GF-Fe; the high-resolution XPS spectrum of c C1s region, d O1s region, and e Fe2p region for GF-Fe

Full size image

For GF-Fe, peak fitting of C1s and O1s were carried out, and the results are shown in Fig. 5c, d. Based on the literatures (Wang et al. 2013a), for C1s spectra, the main peak at 284.6–284.7 eV was attributed to graphitized carbon (C=C). The other three peaks should be attributed to the defects on the GF structure(C=C, 285.1 eV), C–OH (286.0–286.3 eV), and C–O (286.8–287.0 eV). Regarding the O1s spectra, the split peaks were located at 532.2–532.7, 531.0–531.1, and 533.9–534.2 eV, which should be assigned to O–H and C–O (Wang et al. 2015a).

Figure 5e presents the high-resolution spectra of Fe 2p. The peaks centered at 713.7 and 725.1 eV were assigned to Fe(III). The peak centered at 722.0 eV was attributed to Fe0 (Ai et al. 2007a). Therefore, the iron species was mainly composed of Fe0 and Fe2O3. These oxygen-containing groups and ferrite-carbon black hybrid could be acted as the active sites capable of accelerating the electro-chemical reactions and make dissolved oxygen accessible to the cathode surface facilitating electro-generation of H2O2 (Miao et al. 2014).

The degradation of MO in the heterogeneous EF system

It is well known that pH can remarkably affect Fenton reaction (Fan et al. 2010). In acid condition, soluble Fe2+ can react with H2O2 to produce ·OH which could degrade organic pollutants even into CO2 and H2O, but in alkali condition, ferric species would precipitate as ferric hydroxides, which made it more difficult to establish a good redox system between H2O2 and Fe2+/Fe3+ (Brillas et al. 2009). The degradation efficiency of MO with the optimum metal loading under different pH values is shown in Fig. 6. In general, degradation efficiency of MO decreased with an increase of pH, which accords with EF reaction. However, degradation efficiency of MO with 1.0 wt.% Co loading were 82 % within 120 min treatment even at pH 9 while 99 % at pH 3 (Fig. 6a), which was better than the results reported in other literatures (Torres et al. 2007; Zhang et al. 2015). According to the pH values in Table 1, the pH of solution at initial pH 3 remained almost unchanged during treatment while the solutions at pH ≥5 were strongly acidified, which was attributed to the formation of carboxylic acids (El-Ghenymy et al. 2012; Isarain-Chavez et al. 2010). This could explain why degradation efficiency of MO was still high even at alkali condition.

Fig. 6
figure 6

Effect of pH on the degradation of MO: a GF-Co, b GF-Fe, c GF-Ce, and d GF-Cu. Conditions, 0.05 M Na2SO4 (50 mg/L MO); current density, 50 A/m2; air flow rate, 0.5 L/min (1.0 wt.% Co, 1.0 wt.% Fe, 1.0 wt.% Ce, 0.5 wt.% Cu)

Full size image
Table 1 Final pH of the treated solution after 120 min
Full size table

On the other hand, the kind of metal loading could affect the catalytic activity of the electrodes (Ai et al., 2007a, b, 2008). At pH 3, degradation efficiency of MO were 99, 93, 89, and 81 % with 1.0 wt.% Co, 1.0 wt.% Fe, 1.0 wt.% Ce, and 0.5 wt.% Cu, respectively. It indicated that GF-Co had the highest catalytic activity for H2O2 electro-generation by O2 electro-reduction and further heterogeneous EF process (Bonakdarpour et al. 2011). This result was also demonstrated in Fig. 7, in which the TOC removal efficiency was investigated with different metal loadings. At the pH of 3 and 7, the TOC removal efficiency were 93.8 and 86.2 %, 90.2 and 80.3 %, 85.1 and 70.2 %, 70 and 50.8 %, 35.2 and 14.4 %, and 12.3 and 4.5 % using GF-Co, GF-Fe, GF-Ce, GF-Cu, GF-C and GF, respectively. It proved that the electrodes had a better catalytic activity after loading metals. Compared with the degradation efficiency of MO, it also indicated that most of MO was mineralized into H2O and CO2. Table 2 shows that the degradation kinetic rate constant (k) with different metal loadings under different pH values, which was obtained by assuming a pseudo-first reaction rate and could explain the degradation trends more directly. The degradation kinetic rate constant decreased with an increase of pH and also as following order at the same pH: GF-Co, GF-Fe, GF-Ce, GF-Cu, GF-C, and GF. Particularly, when using GF at pH 3, the degradation kinetic rate constant was 0.0025 s−1, which was 25 times higher than that at pH 9, but the degradation kinetic rate constant was 0.030 s−1 using GF-Co at pH 3, which was two times higher than that at pH 9. This result demonstrated again that GF-Co had the high catalytic activity for MO degradation at a wide pH range.

Fig. 7
figure 7

TOC removal using different cathodes at pH of 3 and 7 within 120 min treatment. Conditions, 0.05 M Na2SO4 (50 mg/L MO); current density, 50 A/m2; air flow rate, 0.5 L/min (1.0 wt.% Co, 1.0 wt.% Fe, 1.0 wt.% Ce, 0.5 wt.% Cu)

Full size image
Table 2 The degradation kinetic rate constant (k) with different metal loading under different pH within 120 min treatment
Full size table

Stability test and performance comparison

Since the stability of an electrode is important for practical application, the performance of GF-Co in ten-time continuous runs was investigated for MO degradation. As shown in Fig. 8, the MO degradation efficiency was almost stable with a slight decrease during the ten-time runs. At the first time, the MO degradation efficiency was 99.5 % at pH 3 within 120 min treatment and at the tenth time, that was still 89.5 %. It indicated that the GF-Co electrode was relatively stable and reusable during the process of MO removal. The obtained results demonstrated that the GF-Co electrode had good stability, which was promising for long-term practical application.

Fig. 8
figure 8

The stability test of GF-Co in ten-time continuous runs. Conditions, 0.05 M Na2SO4 (50 mg/L MO); current density, 50 A/m2 (pH = 3, 1.0 wt.% Co)

Full size image

Table 3 shows the comparison of the results of MO degradation with other literatures. In the present work, the energy consumption for MO degradation within 2 h was only 7.0 kWh m−3, which was lower than the result of the experiment using Ti/BDD anode (Ramírez et al. 2013). And the TOC removal efficiency using GDE/Co as heterogeneous EF like was 93 %, which was higher than the result using PbO2 anode (Recio et al. 2011). These results indicated that the heterogeneous EF like using GDE/Co shows a high degradation efficiency and TOC removal for and a low-energy consumption MO degradation.

Table 3 Comparison of MO degradation with other literatures
Full size table

Confirmation of hydroxyl radical

Hydroxyl radical oxidation is a key mechanism in AOPs. To identify the function of ·OH, a hydroxyl radical scavenger, isopropanol (Richard et al. 1997), was added to the solution during degradation of MO. The effect of isopropanol (IPA) addition on the degradation efficiency of MO is shown in Fig. 9. Without IPA, the degradation efficiency of MO was 99 and 94 % at pH 3 and 7 within 120 min treatment. However, after IPA was added into the solution, the degradation efficiency of MO decreased to 21 and 27 %, respectively. This result indicated that the addition of isopropanol obviously inhibited the degradation efficiency of MO and ·OH played an important role in MO degradation.

Fig. 9
figure 9

Effect of isopropanol on degradation of MO using GF-Co. Conditions, 0.05 M Na2SO4 (50 mg/L MO); current density, 50 A/m2 (1.0 wt.% Co, 5 mL IPA)

Full size image

Conclusions

In this work, feasibility of GF-metal was demonstrated in the heterogeneous EF process for MO degradation. After loading metals, the value of O/C increased observed by XPS, the current response became higher and the degradation efficiency of MO was improved. In the H2O2 production experiment, GF-Co was proved to have the highest catalytic activity for electro-reduction of O2 to H2O2. With an increase of pH, the degradation efficiency of MO decreased slightly, but even at initial pH 9, 82 % of MO degradation efficiency was obtained using GF-Co after 120 min treatment. In the process of MO degradation at initial pH ≥5, pH of solution decreased. TOC results indicated that most of MO was mineralized into H2O and CO2. GF-Co had a high stability for MO degradation even after ten-time runs. The addition of isopropanol proved that ·OH played an important role in degradation of MO. The results show the GF-Co is very promising for wastewater treatment.