As a chemical raw material, the annual worldwide output of hydrogen peroxide (H₂O₂) is about 4.5 million tons, which fully demonstrates the importance of this chemical [1]. As an efficient green oxidant, hydrogen peroxide is widely used in pulp bleaching, chemical synthesis, energy conversion, sewage treatment, and other applications [2-7]. At present, the industrial production of H₂O₂ mainly relies on the anthraquinone method, which consumes a considerable amount of energy and produces a large number of chemical pollutants due to the multistep hydrogenation and oxidation reactions [8, 9]. Furthermore, there are various safety risks associated with the storage and transportation of concentrated H₂O₂ [10]. H₂O₂ can also be directly synthesized from hydrogen and oxygen in the presence of noble metal catalysts, in acidic aqueous solution [11, 12] or by plasma methods [13]. However, the direct synthesis of H₂O₂ also has some disadvantages, such as the risk of explosion of the gas mixture and the high cost of the noble metal catalysts.

An attractive alternative method for the in situ synthesis of H₂O₂ is the direct electroreduction of O₂ dissolved in alkaline media (Eq. (1)) or in acid media (Eq. (2)) to H₂O₂, where the oxygen reduction reaction (ORR) undergoes a two-electron pathway [14-16]:

(1)

(2)

One of the key steps to promote the ORR to synthesize H₂O₂ is to develop catalysts with high activity and selectivity. Some noble metals such as Au, Pt, Pd, and Ag have been reported to show good catalytic activity for H₂O₂ generation from the ORR [17-22]. However, the high cost of precious metals could limit the bulk use of these catalysts. As a matter of fact, it is widely believed that carbonaceous materials are the most suitable alternatives as cathodic catalysts to produce H₂O₂ from the ORR, due to their advantages in terms of catalytic selectivity, cost, and environmental compatibility [23-25]. Recent papers have reported the use of N-doped non-metallic carbon materials [26, 27] or non-noble metal and carbon gel complexes [28] for the electrochemical generation of H₂O₂. These carbon materials exhibit good catalytic activity, but their structure-activity relationships need to be further studied. As types of carbon materials, multi-walled carbon nanotubes (MWCNTs) and zero-valent metal compounds have recently been used to catalyze the ORR for H₂O₂ production [29-31]. Because the metal dissolves gradually, there is still room to improve the stability of these carbon nanotube-based materials. Recently, Lu et al. [32] reported the direct use of MWCNTs as non-metallic materials to catalyze the ORR-based production of H₂O₂, after the oxidation of commercially available carbon nanotubes. On the other hand, the authors did not investigate systematic changes in the structure and catalytic properties of the MWCNTs at different oxidation levels.

In this study, we have prepared partially oxidized MWCNTs, whose outer layers were destroyed and oxidized while the inner layers remained intact, by carefully controlling the conditions of the oxidation treatment. The as-prepared samples were characterized by various physicochemical methods, such as X-ray diffraction (XRD), transmission electron microscopy (TEM), and Fourier transform infrared spectroscopy (FT-IR). Compared to commercial raw MWCNTs, the partially oxidized MWCNTs possess a greatly improved selectivity toward H₂O₂. When the oxidation level of the oxidized MWCNTs is too low or too high, the improvement in catalytic performance over that of raw MWCNTs is limited, due to the compositional and structural changes being too small or too large. Only with an appropriate degree of oxidation, leading to limited structural changes, the oxidized MWCNTs possess good conductivity as well as a sufficient number of catalytic activity sites, and their ORR catalytic performance to produce H₂O₂ can be improved to the greatest extent. This conclusion was further verified by various electrochemical measurements and by assessing the performance of in situ H₂O₂ generation from the ORR after the oxidation of the MWCNTs.

Raw MWCNTs (diameter = 30-50 nm, length = 0.5-2 μm) were purchased from Nanjing XFNANO Materials Tech Co., Ltd. K₂TiO(C₂O₄)₂·2H₂O, KOH, and polytetrafluoroethylene (PTFE, 60 wt%, emulsion) were purchased from Shanghai Aladdin Reagent Co., Ltd. All other reagents, including H₂SO₄, KMnO₄, H₂O₂ (30 wt% in H₂O), HCl, and C₂H₅OH were purchased from Guangzhou Chemical Reagent Factory. All reagents were of analytical grade and used without further purification. Deionized (DI) water was used in all experiments.

Partially oxidized MWCNTs were prepared via a very simple surface oxidation treatment. Typically, 30.0 mg raw MWCNTs, 30.0 mL concentrated sulfuric acid, and 150.0 mg KMnO₄ powder were added into a three-necked, round-bottomed glass flask, which was placed in a water bath. Then, the flask was equipped with a reflux condenser, magnetic stirrer, and thermometer. After stirring the mixture for 1 h at different temperatures (20, 30, 40, 50, and 60 ℃), excess H₂O₂ (15.0 mL) was slowly added into the flask until no bubbles formed, followed by the addition of 150.0 mL DI water. Next, the slurry was removed, cooled down naturally to room temperature, filtered through a millipore membrane (0.45 μm), and washed with plenty of 1.0 mol L^-1 aqueous HCl solution and deionized water until the washings became neutral, with a pH value of 7.0. Finally, the black residue was dispersed in an appropriate quantity of water and then dried by the freeze-drying method.

Moreover, when the oxidation temperature was kept at 40 ℃, the reaction time was changed to 1, 2, and 3 h. The corresponding partially oxidized MWCNT products were denoted as O-CNTs-X-Y, in which X and Y represent the reaction temperature and time, respectively.

The morphology and structure of the products were characterized by TEM (JEM-2010HR). Wide-angle XRD patterns were recorded on a Bruker D8 X-ray diffractometer using Cu K radiation (40 kV, 26 mA). The surface functional groups of the samples were investigated by FT-IR spectroscopy (Bruker EQUINOX 55). The Raman spectra of the samples were recorded on a HORIBA Jobin Yvon, LabRam HR800 laser confocal microspectrometer. X-ray photoelectron spectroscopy (XPS) measurements were performed on an ESCALAB 250 spectrometer.

All electrochemical tests were performed on a computer-controlled standard three-electrode system connected to an Auto84480 instrument at room temperature. Platinum foil (1.0 cm × 1.0 cm) and Ag/AgCl (saturated KCl) were used as counter and reference electrodes, respectively. Without specification, all potentials were referenced to the reversible hydrogen electrode (RHE) scale, using Eq. (3) [33]:

(3)

where 0.197 V is the standard potential of the Ag/AgCl reference electrode in a saturated aqueous solution of KCl at 25 ℃.

A glassy carbon disk (of 0.2475 cm² area) was used as the working electrode. The catalyst ink was prepared by mixing the as-prepared oxidized MWCNTs catalysts (2.0 mg), a Nafion^® solution (5.0 wt%, 20.0 μL), and ethanol (500.0 μL), followed by ultrasonication for 1 h. Then, 10.0 μL of the ink suspension was dropped onto the glassy carbon disk and dried naturally. All measurements were conducted in O₂-saturated 0.1 M KOH (pH ~ 13) with a scan rate of 10 mV s^-1; the corresponding electrochemical polarization curves in N₂-saturated electrolytes were also recorded as a reference.

The electrocatalytic activity and selectivity toward H₂O₂ of the catalysts were investigated by polarization and rotating ring disk electrode (RRDE) measurements. For the latter measurements, the Pt ring electrode was polarized at 0.3 V (vs. Ag/AgCl) in the electrolytes. The electron transfer number (n) and the HO₂^- intermediate production percentage (%HO₂^-) were calculated based on Eqs. (4) and (5), respectively [34, 35]:

(4)

(5)

where I_D is the disk current, I_R is the ring current, and N (= 0.37) is the H₂O₂ collection efficiency of the Pt ring.

Electrochemical impedance spectroscopy (EIS) measurements were carried out in O₂-saturated 0.1 mol L^-1 KOH at 0.75 V, in the frequency range from 100 kHz to 0.01 Hz, at a 5 mV amplitude and with a rotation rate of 1600 rpm.

The stability was evaluated at a constant potential of 0.6 V, in O₂-saturated 0.1 mol L^-1 KOH at room temperature (rotation rate 1600 rpm).

The in situ generation of H₂O₂ was investigated using the three-electrode system described above. Teflon-treated carbon paper (CP, 2.0 cm × 2.0 cm) loaded with the oxidized MWCNTs (~ 2.0 mg cm^-2) was used as the working electrode, which was fabricated in three steps. First, the carbon paper was soaked in PTFE emulsion (2 wt%) for several minutes, and then thermally treated at 350 ℃ for 30 min, resulting in a PTFE content of 30 wt%. The oxidized MWCNTs (8.0 mg) were dispersed into a mixture containing 2.0 mL ethanol and 80.0 μL Nafion^® solution (5 wt%). After ultrasonicating the mixture for 30 min, the catalyst was finally loaded on the treated CP by drop-casting it on a heating plate at 80 ℃. The measurements were carried out in a two-compartment cell with a Nafion^®-117 membrane as the separator. The two compartments were filled with a aqueous solution (0.1 mol L^-1) of KOH. The electrolyte was saturated with oxygen by purging O₂ for at least 30 min before the tests. Moreover, during the H₂O₂ accumulation process, O₂ was bubbled into the solution at a flow rate of 600 mL min^-1. Sample aliquots (1.00 mL) in the cathode compartment were withdrawn from the solution at specific time intervals for further quantitative analysis. The concentration of H₂O₂ in the samples was determined by ultraviolet-visible spectrophotometry using K₂TiO(C₂O₄)₂ as color indicator at 400 nm [36].

The current efficiency of the H₂O₂ electrosynthesis was calculated using Eq. (6) [37]:

(6)

where n is the number of transferred electrons in the reduction of O₂ to H₂O₂ (n = 2), F is the Faraday constant (96485 C mol^-1), c is the H₂O₂ concentration (mol L^-1), V is the volume of electrolyte (L), I is the current intensity (A), and t is the electrocatalysis time (s).

The oxidized MWCNTs were obtained through a facile oxidation process, by stirring a mixture of raw MWCNTs, concentrated sulfuric acid, and KMnO₄ at different temperatures for a given time. The crystal structures of the O-CNTs-X-Y prepared under various conditions were first investigated and analyzed by XRD (Fig. 1). All patterns show three graphitic peaks. The broad peak at 26.0° is attributed to the (002) plane of the graphitic structure of the CNTs, while the other two weak peaks at 42.4° and 44.7° correspond to the (100) and (101) diffractions of the CNTs, respectively [38, 39]. No other peaks are observed in the patterns, indicating that no other impurities are introduced into the samples, or that the impurity content is negligible. By comparing the XRD patterns of the raw and oxidized MWCNTs, it can be easily found that the oxidation treatment reduces the intensity of the peaks. Moreover, increasing the reaction temperature (Fig. 1(a)) or the reaction time (Fig. 1(b)) results in a weaker XRD peak intensity of the oxidized MWCNTs. This behavior may be due to the increase in oxidation temperature or oxidation time, which could result in increased damage to the graphitic structure of the MWCNTs.

The damage to the structure of the MWCNTs caused by the oxidation treatment under different conditions can be confirmed by inspecting the morphologies of the samples, which were investigated by TEM. Typical images of the OCNTs-X-Y are shown in Figs. 2 and 3. Fig. 2(a) and (b) shows that, when the reaction time is kept constant (1 h), the MWCNTs maintain their tubular morphology after the oxidation treatment at the lower reaction temperature (20 ℃), and the degree of damage to the MWCNT walls is relatively small. Upon increasing the reaction temperature to 40 ℃, the tubular structure of the MWCNTs is generally retained, while the walls of the nanotubes begin to break down, as shown by the defects clearly visible in the MWCNTs (Fig. 2(c)) after the oxidation treatment. As the reaction temperature is further increased to 60 ℃, the nanotube walls are thoroughly etched and the oxidized products completely lose their tubular shape (Fig. 2(d)). When the reaction temperature is maintained constant, the time of the oxidation treatment plays a similar role to the treatment temperature in the destruction of the structure of the MWCNTs, that is, the etching degree of the MWCNT walls increases with an increase in reaction time. After a certain oxidation time, the walls become partially corroded. On the other hand, when the oxidation time is long enough, the walls are completely corroded and the as-prepared O-CNTs entirely lose their tube-like structure (Fig. 2(e) and (f)).

The high-resolution transmission electron microscopy (HRTEM) images illustrate more clearly the changes affecting the MWCNT walls during the oxidation process. For the commercial MWCNTs, the images show very clear stripes due to the high crystallinity, the graphite layers are evenly arranged, and the tube walls are intact (Fig. 3(a)). With an increase in oxidation degree, the structure of the MWCNT walls is gradually destroyed. For partially oxidized MWCNTs samples such as O-CNTs-40-1, which was treated at 40 ℃ for 1 h, the HRTEM image (Fig. 3(b)) shows that the outer layers of the MWCNT walls suffer significant damage. Noticeable amounts of defects and ruptures are present on the outer layers of the nanotube wall (corresponding to the outer side of the red dotted line in Fig. 3(b)), while the stripes corresponding to the graphite layers are still visible on the inner wall, indicating that the oxidation treatment has little impact. The thinned graphitized MWCNT walls appear to be covered with an amorphous carbonaceous coating. In this case, if the MWCNTs are further oxidized by increasing the oxidation temperature or extending the treatment time, the damage to the MWCNT walls will propagate from the outer to the inner layers of the nanotube walls, the amorphous carbonaceous layer will gradually thicken, and finally the MWCNT walls will be completely destroyed. Only a very small number of rather short stripes representing the graphite structure can be seen in Fig. 3(c). Incidentally, a small number of darker areas are observed in the figure, where the MWCNTs stack together or have thicker micro-ends.

To further understand the effects of the oxidation treatment on the number of defects and on the structure and crystallization of the O-CNTs, Raman spectroscopy was employed to characterize the obtained MWCNTs samples. The spectra of all materials are presented in Fig. 4. All spectra show a strong D band at ~ 1330 cm^-1, which is related to structural defects, disorder, and vacancies in the graphene lattice. A G band at about 1580 cm^-1, associated with the crystalline graphitic structure of the CNTs [40], can also be observed. The intensity ratio of the D and G peaks (I_D/I_G) was used to inspect the structural damage in the O-CNTs caused by the oxidation treatment. In general, the I_D/I_G ratios of the O-CNTs are higher than those of the raw MWCNTs, which indicates that the oxidation process destroys the MWCNT structure and creates crystal defects in the graphitic structure [41]. This result is consistent with the XRD and TEM results discussed above. Fig. 4(a) and (b) also shows that, at low oxidation levels, the I_D/I_G ratio increases with an increase in oxidation degree, until it reaches a maximum at higher oxidation degrees. Further increasing the oxidation level leads to a decrease in the I_D/I_G ratio, due to the fact that the oxygen-bonded saturated sp³ configuration C, which forms the defects, could be desorbed in the form of C=O and O-C=O species, thus resulting in the decrease in intensity of the D band [42].

In order to investigate the changes in chemical composition of the oxidized MWCNTs, the functional groups of the as-prepared samples were characterized by FT-IR. As shown in Fig. 5, the spectra of the O-CNTs show noticeable changes in the intensity of the characteristic peaks of several functional groups, compared to those of the raw MWCNTs. The wide absorption peak at ~ 3438 cm^-1 is the characteristic vibration peak of the hydroxyl (-OH) in the acidic functional groups on the MWCNTs [43]. The bands at ~ 2920, 1722, and 1655 cm^-1 are attributed to the stretching vibrations of the C-H [44], C=O [45], and C=C [30] groups of the MWCNTs, respectively. In addition, with a further increase in the oxidation degree, C-O-C (~ 1229 cm^-1) and C-O (~ 1050 cm^-1) [46] signals appear in the FT-IR spectra. Overall, these typical bands reveal that oxygen-containing functional groups, some of which (-COOH and C-O-C) were reported to favor the electrosynthesis of H₂O₂ [32], have been introduced on the surface of the O-CNTs.

In order to better understand the effect of the oxidation treatment on the elemental contents on the MWCNT surface, the chemical states of all oxidized MWCNTs samples were studied by XPS. The corresponding results are given in Fig. 6. The oxygen content of the MWCNTs before oxidation is negligible, as only one C1s peak is visible in the recorded spectra of the raw samples, indicating the absence of any oxygenated functional groups. The O1s peak appears in the spectra of the oxidized MWCNTs and its intensity gradually increases with an increase in oxidation level, denoting a corresponding increase in the oxygen content (Fig. 6(a) and (d)). The high-resolution C1s spectra (Fig. 6(b) and (e)) could be deconvoluted into six specific bands including those of C=C, C-C, C-O, C=O, and O=C-O groups at 284.7, 285.1, 286.1, 287.0, and 288.9 eV, respectively, as well as the π-π^* transition at 290.5 eV [47, 48]. Moreover, the deconvolution of the high-resolution O1s spectra (Fig. 6(c) and (f)) suggests the presence of two oxygenated species on the surface of the O-CNTs: C=O at 531.6 eV and C-O at 533.2 eV [47]. The high-resolution C1s and O1s spectra thus support the conclusions of the FT-IR analysis discussed above. The relative contents of oxygen-containing groups in the samples were obtained by measuring the relative peak areas in Fig. 6(e) and (f), and the results are summarized in Table 1. The data in the table clearly show that the oxidation treatment resulted in a sustained increase in the relative content of oxygen-containing groups on the surface of the MWCNTs.

According to the above analysis, the oxidation treatment of the MWCNTs gradually destroys their structure and introduces defects and oxygen-containing functional groups on their surface. Defects and oxygen-containing functional groups can alter the electronic structure of carbon materials; this leads to different binding energies of the intermediates generated during the ORR process to the active sites of the carbon-based materials, which in turn affect the ORR catalytic activity of the catalysts [49, 50]. According to density functional theory (DFT) calculation results reported elsewhere, the defects introduced in the MWCNTs have weak cohesion with oxygenated species, which ensures that the O-O bonds are preserved, thus improving the selectivity toward H₂O₂ formation [51]. Moreover, the C-O-C and -COOH functional groups bind OOH^*, the only intermediate involved in the two-electron reduction of O₂, at very favorable energies for the two-electron oxygen reduction to H₂O₂, indicating high reactivity in the electrochemical synthesis of H₂O₂ [32]. Defects and oxygen-containing functional groups such as C-O-C and -COOH thus result in the ORR to produce H₂O₂. Therefore, there is no doubt that the oxidation treatment of the MWCNTs can alter their corresponding catalytic activity. However, the catalytic activity of MWCNTs with different oxidation degrees can also vary depending on their different structures and oxygen contents. Therefore, RRDE measurements, where the amount of H₂O₂ generated at the disk electrode could be accurately determined by the ring current, were conducted in O₂-saturated 0.1 M KOH solution (pH = 13) to evaluate the ORR pathway for all oxidized MWCNT samples during the ORR process. The experimental results shown in Fig. 7 allow identifying the O-CNTs catalysts with the most suitable oxidation degree.

As clearly shown in the RRDE voltammograms of Fig. 7(a), the disk and ring currents corresponding to O-CNTs-60-1 are the smallest among the O-CNTs subjected to different treatment temperatures. This may be due to the very high level of oxidation of the O-CNTs-60-1 samples, whose structure changed to an amorphous carbonaceous state with low crystallinity (Figs. 2(d) and 3(c)), which does not promote electron transfer and transmission. Figs. 7(b) and 7(c) show an improvement in both the yield of H₂O₂ and the number of transferred electrons in the ORR-based synthesis of H₂O₂ when the MWCNT catalysts are oxidized. The same conclusion can be drawn from Figs. 7(e) and 7(f). Based on the previous conclusions, this could be due to the introduction of defects and catalytically active oxygen-containing functional groups on the surface of the MWCNTs in the oxidation process. More specifically, the sample subjected to an oxidation treatment at 40 ℃ for 1 h shows the best catalytic performance. Compared to those of the unoxidized MWCNTs and of the products obtained under other oxidation treatment conditions, the O-CNTs-40-1 sample exhibits a higher current and a lower overpotential, indicating a higher reactivity. In particular, compared to that of the commercial raw MWCNTs, the selectivity toward H₂O₂ of O-CNTs-40-1 shows a significant increase, from less than 30% to about 50%, and the corresponding number of transferred electrons (n) decreases from more than 3.4 to less than 3.0, which indicates that the ORR has a higher tendency to undergo a 2e^- pathway. It is worth pointing out that Figs. 7(b), (c), (e), and (f) also show that the selectivity toward H₂O₂ and the n value change significantly at negative potentials, which may be due to the side hydrogen evolution reaction.

In addition to the RRDE results, the results of the EIS measurements (Fig. 8) also reveal that O-CNTs-40-1h has the best catalytic activity, because its charge-transfer and intrinsic resistances are lower than those of the raw MWCNTs or of the other oxidized MWCNTs. Overall, these results support the preliminary conclusion that the partially oxidized O-CNTs-40-1 sample exhibits higher catalytic activity and selectivity in the ORR-based production of H₂O₂, compared to those of raw MWCNTs and of other oxidized MWCNTs with lower or higher oxidation degree.

The long-term durability is an important parameter in the evaluation of electrocatalysts; hence, the stability of O-CNTs-40-1, raw MWCNTs, and oxidized MWCNTs with higher oxidation levels were studied by the chronoamperometric method. As shown in Fig. 9, O-CNTs-40-1 retains 90.5% of the initial current density after a 10000-s test, while the raw MWCNTs, O-CNT-40-3, and O-CNT-60-1 keep 83.6%, 72.2%, and 67.6% of their initial current, respectively, showing the outstanding stability of the partially oxidized O-CNTs-40-1 sample.

Based on the results of the electrochemical tests discussed above, it is clear that O-CNTs-40-1 has higher catalytic activity, selectivity, and durability in the ORR-based synthesis of H₂O₂ than the raw MWCNTs and the oxidized MWCNTs with lower or higher oxidation degree. In order to further investigate the practical applicability of O-CNTs-40-1, some samples were loaded on Teflon-treated CP to obtain a CP@O-CNTs-40-1 electrode for the in situ production of H₂O₂ through the ORR.

Under different cathodic potentials, significantly different amounts of H₂O₂ accumulate on the CP@O-CNTs-40-1 electrode during the same time. As shown in Fig. 10, with the cathode potential ranging from 0.86 to 0.46 V over 40 min, the cumulative concentration of H₂O₂ increases and reaches its maximum at 0.46 V. As the cathode potential continues to increase, the concentration of H₂O₂ decreases, which is due to the fact that both the decomposition of H₂O₂ and the evolution of H₂ at high potential begin to occur, inhibiting the 2e^+- process of the ORR. Figs. 11(a) and (b) show the cumulative concentration of H₂O₂ and the corresponding current efficiency of different cathodes at 0.46 V. Compared to those of the electrode loaded with the raw MWCNTs, the H₂O₂ concentration and current efficiency of the CP@O-CNTs-40-1 electrode show great improvement. O-CNTs-40-1 clearly shows higher potential than the unoxidized raw MWCNTs as ORR catalyst to generate H₂O₂ in practical applications. Incidentally, the current efficiency of the CP@O-CNTs-40-1 electrode in Fig. 11(b) shows a slight decrease toward the end of the experiment, which can be attributed to the increased decomposition of H₂O₂ at this potential. The measured concentration of H₂O₂ is lower than its actual generated concentration, resulting in a slight decrease in the calculated current efficiency [52].

The results of the electrochemical measurements show that O-CNTs-40-1 has enhanced activity and selectivity in the electrocatalytic reduction of oxygen to H₂O₂ compared to the raw MWCNTs. In addition, O-CNT-40-1 shows a greater performance improvement compared to those of the other oxidized samples. These findings can be understood with the help of the physicochemical characterization experiments. MWCNTs that are completely non-oxidized or have a low degree of oxidation do not possess effective catalytic active sites, while oxidized MWCNTs with high degree of oxidation have a severely damaged graphitic structure and their electron transfer performance is greatly reduced, which is also not favorable for the reaction.

In summary, oxidized MWCNTs with different oxidation degrees were used to catalyze the oxygen reduction reaction and produce H₂O₂. The experimental results show that, after the oxidation treatment, defects and oxygen-containing functional groups are introduced on the surface of the MWCNTs, which promotes H₂O₂ generation by the ORR through a two-electron process. Therefore, all oxidized MWCNTs have better catalytic activity and selectivity than the commercial non-oxidized MWCNTs. Among the oxidized MWCNTs, O-CNTs-40-1 exhibits the greatest improvement in H₂O₂ selectivity (from ~ 30% to ~ 50%) and electron transfer number (from ~ 3.4 to ~ 3.0). In terms of structure and composition, the outer walls of O-CNTs-40-1 are damaged and destroyed by the oxidation treatment, resulting in the formation of a large number of defects, which promotes the ORR. At the same time, oxygen-containing functional groups such as -COOH and C-O-C introduced on the MWCNT surface are known to be the catalytic active sites for the ORR-based production of H₂O₂. On the other hand, the inner layers of the nanotube walls are not corroded and keep their graphitic structure, which ensures that the material has good electrical conductivity. Therefore, the present partially oxidized MWCNTs have good applicability in catalyzing the reduction of oxygen to H₂O₂. The present work could provide useful information for the design of electrocatalysts capable of efficiently catalyzing the ORR for the in situ synthesis of H₂O₂; in particular, we showed that tailoring the surface oxidation degree and conductivity of carbon-based materials could significantly enhance their performance.