Inconsistency of BDD reactivity assessed by ferri/ferro-cyanide redox system and electrocatalytic degradation capability

Abstract Ferri/ferro-cyanide ([Fe(CN)6]3−/4−) redox couple has been extensively used as a benchmark to assess electrocatalytic degradation capability of boron-doped diamond (BDD) electrodes. However, the [Fe(CN)6]3−/4− is far more sensitive to the surface terminal groups of BDD surface than the other factors (e.g. surface morphology and electrode configuration) that are closely related to electrocatalytic degradation properties. Thus, inconsistency exists while correlating the degradation properties of BDD with electrochemical properties determined from ferri/ferro-cyanide redox couple. Herein, an exemplar pollutant, reactive blue 19 (RB-19), was electrochemically degraded using various terminated BDD electrodes, including hydrogen-terminated (H-BDD), oxygen-terminated (O-BDD) and porous oxygen-terminated BDDs (OE-BDD), obtained via cathodic and anodic polarization as well as oxygen plasma etching, respectively. Surprisingly, OE-BDD with the lower heterogeneous electron transfer rate constant for [Fe(CN)6]3−/4− showed a better electrocatalytic degradation capability toward RB-19, indicating the inconsistency for qualitatively evaluating degradation properties according to the kinetic parameters extracted from [Fe(CN)6]3−/4− redox system.


Introduction
Conductive diamond showed diverse distinctive properties such as a wide potential window (4.6 V in organic solutions [1]), high biocompatibility [2], and exceptional electrochemical stability, due to the existence of saturated sp 3 -hybridized C-C bond in the diamond crystal lattice, which has been extensively used in electro-analysis, electro-catalytic oxidative degradation, and electrical energy storage [1,[3][4][5][6]. The electrocatalytic degradation capability of diamond depend on many factors, such as the boron-doping concentration [7,8], the ratio of sp 2 /sp 3 [9], surface termination [10], and micro/nanometer-sized surface structure of borondoped diamond (BDD). The highly doped BDD thin films usually show better electrochemical activity than the lowly doped ones due to the increased electrical conductivity [11], and so does the addition of the impurity sp 2 phase (i.e. graphite phase). In addition, constructing porous morphology on the surface of BDD is beneficial to increasing its specific surface area, thereby increasing the number of reaction sites [12].
The ferri/ferro-cyanide redox couple ([Fe(CN) 6 ] 3−/4− ) has been widely used to assess electrochemical properties of BDD electrodes, including heterogeneous kinetic rate constant (k 0 ), the reversibility/irreversibility behavior, the charge transfer resistance, and electrochemically active surface area (ECSA) if completely or nearly reversible [13]. For instance, k 0 can be extracted from the relation of the peak currents, the peak potential separation (ΔE p ) and the scan rate according to the Nicholson method [14]. Reversible redox reactions are characterized by high rates of charge transfer (i.e. k 0 >10 −1 cm s −1 ), and k 0 for quasi-reversible reactions ranges between 10 −1 and 10 −5 cm s −1 , and <10 −5 cm s −1 for irreversible systems. Alternatively, for one-electron transfer reaction, the relative reversibility behavior can also be simply assessed by the ΔE p values in which the reaction is considered as reversible if ΔE p is around 59 mV at 25 °C, irreversible if larger than 200 mV, and quasi-reversible if between them [13].
Owing to its easy access to abundant information with respect to electrochemical properties of the underlying electrode, the ferri/ferro-cyanide redox system was widely reported to characterize the electrochemical activity of BDD electrodes and tentatively establish the relation between activities and electrochemical degradation properties. A higher k 0 usually corresponds to the faster electron transfer rate and a higher ECSA leads to the larger electroactive sites, thereby resulting in the better electrocatalytic performance, which is easy for granted to observe the better electrocatalytic degradation performance for the target organic compound of an interest. For instance, previous reports extracted the key electrochemical properties of different BDD electrodes from the ferri/ferro-cyanide redox system according to Randles-Sevcik equation in which the slope of a linear relation of peak currents versus scan rates yields the resulting ECSA. They attributed the improved electrocatalytic degradation performance to the increased ECSA of those BDD electrodes which could provide more electroactive sites to degrade the organic compound (e.g. reactive blue-19 and tetracycline) [15,16]. Recently, Li et al. utilized the similar redox system to pick up the optimal BDD electrodes grown on the Ta substrate to carry out electrocatalytic degradation of organic compounds.
However, the electrochemical behavior of [Fe(CN) 6 ] 3−/4− is far more sensitive to the surface functional groups of BDD surface than the other factors such as surface morphology, electrode configuration and charge transfer resistance, which are closely related to electrocatalytic degradation performance. Compared to the oxygen-terminated BDD (O-BDD), the hydrogen-terminated BDD (H-BDD) presents faster electron transfer kinetics of Fe(CN) 6 4−/3− than the former [17]. Indeed, when the O-BDD transferred to H-BDD via a cathodic pretreatment by applying a bias potential of −3.0 V for 30 min, the H-BDD showed the increased intensity of peak current accompanied by a narrowed peak potential separation [18]. This change of BDD reactivity is closely related to surface functional groups since the oxygen-containing groups would repel negatively charged substances (i.e. Fe(CN) 6 3−/4− ) due to electrostatic repulsion and simultaneously, partially block adsorption sites and inhibit the transfer pathway of electrons, thereby reducing electrocatalytic properties [19,20]. Thus, a question arises whether or not the redox couple [Fe(CN) 6 ] 3−/4− is reliable to act as a standard to select the optimal electrodes which are anticipated to exhibit the optimal electrocatalytic degradation performance, owing to the fact that the redox couple [Fe(CN) 6 ] 3−/4− is sensitive to the surface functional groups. Herein, we designed three BDD films with equivalent geometric surface area and various surface functional groups via polarization treatment and oxygen plasma etching and investigated their electrochemical behaviors for the Fe(CN) 6 3−/4− redox system as well as respective electrocatalytic degradation performance toward a typical dye pollutant (i.e. reactive blue 19 (RB-19)).

Preparation of BDD electrodes with different surface functional groups
Hot-filament chemical vapor deposition (HF-CVD) was used to grow BDD films on Si substrate (20 × 25 × 1 mm 3 ). The growth temperature was set as 800 °C and the pressure was maintained at 3 KPa with mixed gaseous flow ratios of H 2 :CH 4 :B 2 H 6 being 98:3:0.6, similar to previous works [21]. H-BDD electrode was obtained by cathodic polarization at a bias voltage of −3 V vs Ag/AgCl for 900 s in 0.5 M H 2 SO 4 solution, while O-BDD electrode by anodic polarization at a bias voltage of +3 V Ag/AgCl. Oxygen plasma etching was used to etch the O-BDD to form the porous morphology, denoted as OE-BDD. The etching temperature levelled off at 800 °C and the time was set as 2 min, similar to our recent work [21].

Water degradation experiment
Electrocatalytic degradation of 100 mg L −1 reactive blue 19 was conducted in a 500 mL container with 0.5 A direct-current power, using BDD electrodes (size of 20 × 25 × 1 mm 3 ) as the anode and a stainless steel sheet (20 × 25 × 1 mm 3 ) as the cathode. The distance between the two electrodes during the treatment was 1 cm. The supporting electrolyte was 0.1 mol/L Na 2 SO 4 and the total degradation time was 120 min. The experiment was under constant stirring with a speed of 300 rpm/min.

Chromaticity removal rate
The absorbance of RB-19 in the visible region at a maximum absorption wavelength of 595.5 nm was measured with an ultraviolet-visible spectrophotometer (UV-8000S, Metash), and the color removal of the solution (color removal) was calculated using the following equation [22]: Where t is the degradation time (s), A 0 and A t are the initial absorbance of RB-19 at the maximum wavelength in the visible region and the absorbance at the moment of degradation t (a.u.).

COD removal rate
The test solution was mixed with COD reference reagent and placed in DRB200 digestion reactor (HACH, USA) for high temperature digestion, and then the DR3900 (HACH) water quality detector/COD heavy metal tester was used to detect COD in the solution, and the following equation was used to calculate the removal rate of COD and the current efficiency [22].
where COD 0 is the initial COD value of organic matter (mg L −1 ), COD t is the COD value of organic matter at degradation t (mg L −1 ); ΔCOD is the change in COD of active blue 19 during degradation (mg L −1 ) [23]; where F is Faraday's constant (96,487C mol −1 ), V s is the volume of solution (L), I is the applied current (A); 8 is the equivalent mass of oxygen (g/eq −1 ), and t is the electrolysis time (s).

Energy consumption of the degradation process EC
where E C is the energy consumption of the degradation process (kW·h·m −3 ), E cell is the average value of the degradation voltage (V), I is the current value (A), and t is the degradation time (h) [15].

Morphology and structure characterization
The BDD electrodes were subjected to polarization and oxidation etching treatments, respectively, and the SEM images are shown in Figure 1. The H-BDD and O-BDD electrodes present a uniform and compact polycrystalline structure with the size of the crystal grains of around 2-4 µm with clear grain boundaries, and the grain shape of the diamond almost remains intact (Figure 1(a,b)). The surface morphology of the O-BDD and H-BDD was similar to that of the as-deposited BDD film sample (see Figure S1), indicating that the polarization treatment did not change the electrode morphology. In contrast, the surface morphology of OE-BDD changed significantly with dense etched pits while the edges of BDD remain clear (Figure 1(c)), formed by the reaction between oxygen and C-C bonding during oxygen plasma etching [24,25]. The porous surface morphology of BDD is consistent with this work by Liu et al. [21], during which the BDD films after oxidation etched presented porous micrography and a larger specific surface area than the unetched sample. Figure 1(e) shows the XRD patterns of the three BDD electrodes. There is an apparent diffraction peak of diamond (111) and weak peaks of (220) crystal planes at 2θ = 43.8° and 76°, indicating the BDD samples are mainly (111) oriented. Considering the important effect of B-doping concentration on the catalytic properties of diamond, we characterized the Raman spectra of OE-BDD, O-BDD, and H-BDD. The characteristic peak of diamond appearing around 1332 cm −1 was observed on the three BDD thin films (Figure 1(f)), and there are no obvious graphite peaks around 1350 and 1580 cm −1 [26], reflecting that all electrodes have good diamond quality. This is because oxygen plasma etching rate is faster for sp 2 carbon than sp 3 phase and therefore the former is removed during the oxygen plasma process. We calculated the density of doped boron ([B]) of BDD films from the peak wavenumber at around 400-600 cm −1 using the formula [27], To obtain information about the abundance of C-O bonds on BDD surface, XPS measurements were performed on BDD electrodes. Figure 2(a-c) show the O atom% of H-BDD (3.53%) is far less than O-BDD (8.7%). That is because the abundance of C-H bonds would increase and that of C-O bonds decreases with greater negative potential (−3 V) [28]. The mechanism for the hydrogenation of BDD thin films by cathodic polarization treatment was proposed by Einaga [28]. The process of cathodic polarization could generate hydrogen atoms to abstracted the C-O bonds, resulting in the creation of dangling bonds on the BDD surface, and the dangling bonds would react with the hydrogen to generate new C-H bonds [28]. The O atom% of OE-BDD (8.4%) is similar to that of O-BDD (8.7%), that because the as-deposited BDD surface covered with C-H bonds can be oxygenated by atomic oxygen during the oxygen plasma etching process. Figure 2(d-f) illustrates the XPS spectra of C1s peak of three BDD samples. Two main peak components, C1 peak at 284.7 eV, and C2 peak at 285.5 eV are observed, which are attributed to sp 3 C-C bond from BDD [29,30], C-O bonds at the grain boundaries on the electrode surface  [31], respectively. As show in the Figure 2(d-f), there are more C-O bonds on OE-BDD and O-BDD than on H-BDD.
The O-terminal can interact with the hydroxide bond in the water molecule to increase the surface energy, thus making the surface tends to be hydrophilic [32,33]. The wettability of the BDD films was compared to further characterize surface C-O coverage. It was found that the H-BDD was hydrophilic and had a wetting Angle (WA) of 91.03°, while WA of the BDD films after anodization and oxygen plasma etching were 44.57° and 39.22°, respectively (Figure 3). The results indirectly demonstrated the abundance of C − O bonds increases under high anodic polarization current density and oxygen plasma treatment.

Electrochemical characterization
The cyclic voltammogram (CV) curves are commonly used to investigate the electrochemical performance of the electrodes in the [Fe(CN) 6 ] 3−/4− redox couple [34]. Compare the CV of three BDD electrodes at 20 mV s −1 in a solution containing 1 mM K 3 Fe(CN) 6 (Figure 4(a)), the anodic peak potential of H-BDD is more negative than O-BDD and OE-BDD, which indicates that the electron transfer of K 3 Fe(CN) 6 at O-terminated BDD electrodes required larger driving force. Besides, the H-BDD has the largest oxidation peak intensity of 0.74 mA, 1.53 times the OE-BDD (0.48 mA) and 1.26 times the O-BDD (0.59 mA) (Figure 4(b)), showing that the H-BDD films had higher electrochemical activity than the other two in the ferri/ferro-cyanide system. The peak potential separation (ΔE p ) of redox reactions observed at the OE-BDD is 41 mV larger than that H-BDD (Figure 4(b)), indicating the poor reversibility of the O-terminated BDD electrode [35]. Figure 5(a-c) shows the CV from three modified BDD electrodes in 1 mM K 3 Fe(CN) 6 with 0.1 M KCl solution as the supporting electrolyte at various scan rates. As the scan rate increases, the oxidation peak potential of three electrodes moves in the positive potential ( Figure 5(d)). Meanwhile, the peak currents are linearly proportional to the square root of scan rate.
where Ψ is Nicholson's kinetic parameter dependent on ΔE p in mV (Equation (6)) [36]. n is the number of electrons transferred in the oxidation/reduction process (here, n = 1), R is the universal gas constant and T is the temperature. According to the plot of Ψ and [ ] / π DnvF RT 1 2 , the H-BDD had the fastest electron transfer among these electrodes due to its maximum slope (see Figure 6). According to the linear calibration relations (Ψ = 0.0041  (R 2 = 0.998)) has lower values of 2.6 × 10 −3 and 2.7 × 10 −3 cm s −1 , respectively (Figure 3(b)). The decreased k 0 found at O-BDD and OE-BDD might be because oxygenated surface exhibited thicker depletion layer than hydrogenated surface [37] and O sites had electrostatic repulsion to the Fe(CN) 6 3−/4− , preventing it from adsorbing on the electrode surface [20], and thus inhibiting the electron transfer between the redox group and BDD surface.

Electrocatalytic degradation of organic wastewater
From the above discussion, one can know that H-BDD exhibits better electrical activity in potassium ferricyanide system compared to O-BDD and OE-BDD. To compare the catalytic activity of the three electrodes in real wastewater degradation, we used RB-19 as a simulated dye to test the electrocatalytic performance of the three BDD electrodes. Figure 7(a-c) shows the UV-Vis test results of electrochemical oxidation of RB-19 with H-BDD, O-BDD, and OE-BDD electrodes, respectively. There are three characteristic peaks of UV-Vis spectrum from original Reactive Blue 19. The maximum  To further confirm the electrocatalytic activity of three BDD electrodes, we compared the chemical oxygen demand (COD) removal rate of the electrodes. Figure 7(e) shows a plot of energy consumption versus COD removal rate. The COD removal of H-BDD, O-BDD, and OE-BDD within 90 min were 13.75%, 20.62%, and 44.7%, respectively. The total COD removal rate of the OE-BDD electrode was increased by 3.25 times and the current efficiency was increased by 1.78 times compared to the H-BDD electrode. Besides, the OE-BDD electrode required the lowest energy consumption at the same COD removal rate (see in Figure 7(f)). The results above indicated the OE-BDD had the best electrocatalytic activity, followed by O-BDD, and finally H-BDD. OE-BDD electrodes are treated by oxygen plasma etching. The best degradation performance of OE-BDD is due to the increased specific surface area after oxygen etched. During the etching process, a large number of chemically active oxygen atoms in the oxygen plasma occupy the dangling bonds of carbon atoms and react with them to form CO or CO 2 [38], thus forming a hole structure (as shown in Figure 1(d)). Although, as shown in Figure 1(a,b), the electrochemical polarization treatment does not change the surface morphology of the BDD electrode. So the OE-BDD has a larger specific surface area compared to the O-BDD, H-BDD electrodes, increasing the contact area between wastewater and electrode, providing more sites to generate hydroxyl radicals (•OH) for RB-19 degradation. Compared to H-BDD, O-BDD had better degradation performance, possibly attributed to its hydrophilic surface. The organic degradation relies on the extent of •OH production via water discharge on the surface of the electrode [39], and the hydrophilic of surface is beneficial to the adsorption of water molecules on the electrode surface It is noteworthy that H-BDD had the worst electrocatalytic oxidation performance, while one exhibited largest electrical activity in Fe(CN) 6 3−/4− system compared to O-BDD and OE-BDD, which contradicts each other. This is because the Fe(CN) 6 3−/4− redox probe is extremely sensitive to the surface chemistry, and the O end group can inhibit its electron transfer, thus shielding the effect of the micro-nano structure on the electrochemical properties of the BDD electrode. Therefore, it is vitally important when measuring electrochemical performance to carefully assess the surface chemical property.

Conclusion
Ferri/ferro-cyanide redox couple has been widely used as a benchmark to evaluate the degradation ability of BDD electrodes. Since the [Fe(CN) 6 ] 3−/4− is far more sensitive to the surface terminal groups of BDD surface than the other factors (e.g. surface morphology and electrode configuration) relevant to electrocatalytic degradation properties, there exists inconsistency while correlating the degradation properties of BDD with electrochemical properties determined from ferri/ ferro-cyanide redox couple. The experimental results herein demonstrated that the OE-BDD with the slower heterogeneous kinetic rate constant (2.7 × 10 −3 versus 4.1 × 10 −3 cm s −1 for the H-BDD) showed the better electrocatalytic degradation capabilities toward RB-19. The total COD removal rate of the OE-BDD electrode was increased by 3.25 times that of the H-BDD electrode and the current efficiency was increased by 1.78 times. Our results indicate that ferri/ferro-cyanide redox couple could not be the only determining factor to assess the electrocatalytic degradation capabilities of electrodes in wastewater treatment. More redox systems, such as ruthenium hexamine, is suggested to be considered as the assessment criterion.

Disclosure statement
No potential conflict of interest was reported by the authors.