Boosting Oxygen Reduction through Microenvironment Modulation to Enhance Mass Transportation

Electroreduction of oxygen driven by renewable electricity holds significant promise for the sustainable production of value‐added hydrogen peroxide (H2O2). While water is a desirable source of protons and electrons for this reaction, its low gas solubility often limits the transportation of gas molecules and consequently leads to large concentration overpotential, resulting in unsatisfactory energy efficiencies. Herein, a facile and effective strategy to promote the 2e− oxygen reduction reaction (ORR) through microenvironment modulation is presented. In this work, it is specifically aimed to facilitate oxygen transportation at the reaction interface, particularly at high rates. To achieve this, hydrophobic polytetrafluoroethylene (PTFE) particles are introduced into a catalyst layer containing amino‐group‐functionalized carbon nanotube (CNT‐NH2) as the ORR catalyst for H2O2 production. As a result, the PTFE‐modified CNT‐NH2‐based gas‐diffusion electrode (GDE) substantially improves ORR activity. At 100 mA cm−2, the PTFE‐modified CNT‐NH2 achieves a high cathodic energy efficiency of 92%, 1.5 times higher than the pristine CNT‐NH2‐based GDE (63%). Detailed kinetic analysis reveals that this enhanced ORR performance is indeed due to the enhanced oxygen transportation induced by the persistent hydrophobic microenvironment created by the PTFE‐modified catalyst layer, reducing the concentration overpotential during ORR.


Introduction
Hydrogen peroxide (H 2 O 2 ), recognized as an environmentally benign chemical oxidant, holds considerable significance as a value-added chemical within the contemporary chemical industry.
Its versatile applications encompass sewage treatment, [1,2] propylene epoxidation, [3] paper disinfection, and bleaching. [4]Additionally, H 2 O 2 shows great potential as a green energy carrier owing to its high energy density. [5]owever, the existing large-scale production of H 2 O 2 , accounting for 95%, relies on the energy-intensive anthraquinone oxidation process.[8] In the light of these limitations, the 2e À electrochemical oxygen reduction reaction (ORR), involving the directly reduction of O 2 to H 2 O 2 using renewable electrical energy emerges as a promising and sustainable alternative for H 2 O 2 production under mild operating conditions.
In electrochemical reactions, the measured overpotentials often consist of three primary compositions: activation overpotentials, concentration overpotential, and Ohmic drop.Previous studies have primarily focused on mitigating the activation overpotential of ORR by employing strategies such as heteroatom doping [30,31] and defect engineering [18] to enhance the intrinsic activity of catalysts.In contrast, the limited solubility of O 2 in H 2 O gives rise to the concentration overpotential caused by the insufficient O 2 mass transfer, which is the primary contributor restricting the ORR performance under practical relevant operating conditions. [32,33]To address this challenge, several approaches have been developed to enhance the O 2 mass DOI: 10.1002/aesr.202300143Electroreduction of oxygen driven by renewable electricity holds significant promise for the sustainable production of value-added hydrogen peroxide (H 2 O 2 ).While water is a desirable source of protons and electrons for this reaction, its low gas solubility often limits the transportation of gas molecules and consequently leads to large concentration overpotential, resulting in unsatisfactory energy efficiencies.Herein, a facile and effective strategy to promote the 2e À oxygen reduction reaction (ORR) through microenvironment modulation is presented.In this work, it is specifically aimed to facilitate oxygen transportation at the reaction interface, particularly at high rates.To achieve this, hydrophobic polytetrafluoroethylene (PTFE) particles are introduced into a catalyst layer containing amino-group-functionalized carbon nanotube (CNT-NH 2 ) as the ORR catalyst for H 2 O 2 production.As a result, the PTFE-modified CNT-NH 2 -based gas-diffusion electrode (GDE) substantially improves ORR activity.At 100 mA cm À2 , the PTFE-modified CNT-NH 2 achieves a high cathodic energy efficiency of 92%, 1.5 times higher than the pristine CNT-NH 2 -based GDE (63%).Detailed kinetic analysis reveals that this enhanced ORR performance is indeed due to the enhanced oxygen transportation induced by the persistent hydrophobic microenvironment created by the PTFE-modified catalyst layer, reducing the concentration overpotential during ORR.
transfer at the electrode/electrolyte interface, such as the construction of porous liquids [34] and ionic liquids layer on catalyst surface. [35]t the reactor level, gas-diffusion electrode (GDE) has emerged as a viable solution to circumvent the mass transfer limitations arising from the inadequate solubility of O 2 in the electrolyte.The GDE comprises a porous gas-diffusion layer (GDL) and a catalyst layer.The catalyst layer contacts the liquid, and the gas directly diffuses from GDL to the catalyst layer and participates in the reaction without requiring dissolution into the electrolyte. [36]Therefore, the maintenance of the hydrophobicity within the catalyst layer and the prevention of the catalyst layer flooded by the electrolyte have become keys.[39][40][41] For example, hydrophobic polytetrafluoroethylene (PTFE) was employed to modify the catalyst layer directly, wherein the carbon black catalyst was wrapped by PTFE through calcination. [38]In another work, PTFE was introduced on the carbon fiber of the GDL by calcination, [39] followed by the loading of carbon black catalyst.This modification substantially enhanced the utilization rate of O 2 and the yield of H 2 O 2 .Although these studies appear to improve the performance of O 2 reduction to H 2 O 2 , it remains to be understood that how PTFE influences O 2 mass transfer during the ORR.The debate surrounding whether electrochemical reactions in GDEs occur through improved mass transfer of gaseous gas reactants at the solid-liquid-gas interface or through enhanced mass transfer of dissolved gas reactants at the conventional electrodeelectrolyte interface remains controversial. [38,39,42,43][40][41] However, the pore structure of GDE will be altered due to PTFE shrinkage and agglomeration during the calcination process, this will block the catalytic active sites and lead to increased activation overpotential and further reduction in the ORR performance [39,44] In addition, the melting of the nonconductive PTFE during calcination on the catalyst and carbon fiber in GDL will reduce the electronic conductivity of the GDE, [39,45,46] thereby increasing the Ohmic overpotential during ORR.Moreover, the final state of the PTFE is affected by many factors such as the operation time, temperature, etc., during calcination, making accurate control of these modification challenging.This lack of control may explain why the mass ratio of PTFE to the catalyst exhibits no discernible correlations to the ORR performance. [46]Overall, due to that above the inherent limitations, the current understanding of how PTFE modification enhances O 2 mass transport and the subsequent impact of this enhancement on ORR performance remains somewhat ambiguous and inconclusive.Hence, there is a need to conduct detailed mechanistic studies on the aforementioned interfacial modifications to further optimize the O 2 mass transfer without sacrificing desirable properties such as high active-site density, electronic conductivity, etc.
In our previous work, we found that the amino-groupfunctionalized carbon nanotubes (CNT-NH 2 ) exhibited improved performance for 2e À ORR compared to oxidized CNT, especially at high current densities.This improvement was attributed to its capacity in preserving the hydrophobicity of the catalyst layer. [47]ence, we are intrigued by the potential influence of microenvironment engineering on ORR performance and whether the performance of CNT-NH 2 could be further improved.
In this study, taking CNT-NH 2 as a model catalyst, we developed a straightforward microenvironmental modulation strategy to enhance the 2e À ORR performance by physically mixing PTFE particles into the catalyst layer directly.Specifically, PTFE was dispersed into the catalyst ink for preparing the catalyst layer of the GDE.With an optimized PTFE loading, we observed a significantly increase in the cathodic energy efficiency for H 2 O 2 formation at practical relevant rate, from 62% to 92% in a typical flow cell configuration while maintaining a high H 2 O 2 selectivity (%80%).To elucidate the impact of PTFE on ORR, we carefully investigated the ORR kinetics using the rotating ring-disk electrode (RRDE) system in the presence of PTFE, and performed a COMSOL simulation to understand the solid-liquid-gas contact state of CNT-NH 2 with different hydrophobicity.It is concluded that the promotion of H 2 O 2 production by PTFE can be attributed to the fine balance of liquid/gas phase achieved at the triple-phase interface, which enhanced the mass transfer of gas-phase O 2 , thereby significantly reducing the concentration overpotential, while also the maintaining of the hydrophobicity of the catalyst layer during the ORR process.

Characterization of the GDEs
Two GDEs were prepared for comparison.One used the pristine catalyst ink consisting only of CNT-NH 2 as the catalyst, while the other used the same CNT-NH 2 ink however with the addition of PTFE particles (CNT-NH 2 /PTFE).First, the morphology of the CNT-NH 2 was characterized using scanning electron microscope (SEM).As shown in Figure 1a and S1, Supporting Information, the typical morphology of CNT-NH 2 can be clearly observed on the GDE surface.Fourier-transform infrared (FTIR) spectroscopy was employed to identify the characteristic chemical groups within the CNT-NH 2 .As shown in the FTIR spectra in Figure 1b, the characteristic adsorption bands at 3450 and 1633 cm À1 can be attributed to the amine-functional groups on the CNTs.As for CNT-NH 2 /PTFE shown in Figure 1c and S1, Supporting Information, similar CNT morphology was observed with PTFE particles dispersed among the CNTs.The powder X-Ray diffraction (PXRD) patterns of these two samples are presented in Figure 1d.CNT-NH 2 /PTFE had the similar PXRD pattern to pure CNT-NH 2 , in which characteristic peaks of CNTs are at 2θ = 26°, 41°, and 54°referring to the (002), (100), and (004) reflections, respectively, indicating an unaltered crystal structure of CNT upon the addition of PTFE.Energy-dispersive X-Ray spectroscopy (EDS) element mapping images of CNT-NH 2 / PTFE in Figure 1e indicate the presence of carbon, nitrogen, and fluorine elements, and fluorine element distributed evenly throughout the sample.Taken together, the aforementioned physical characterization results confirm the successfully dispersion of PTFE particles into the catalyst layer of the CNT-NH 2 / PTFE without introducing any morphological or structural changes to the CNT-NH 2 catalyst.

ORR in GDE Cell
Subsequently, we employed a typical flow cell to evaluate the ORR performance after adding PTFE into the catalyst layer.As shown in Figure 2a (Figure S2 and S3, Supporting Information), at various current densities, ranging from 20 to 200 mA cm À2 , the electrode potentials showed significant reduction (over 100 mV at each current density) on CNT-NH 2 /PTFE compared to that of CNT-NH 2 .For instance, a potential of 0.80 V versus reversible hydrogen electrode (RHE) was obtained on CNT-NH 2 /PTFE electrode at a current density of 200 mA cm À2 , which is comparable to that of CNT-NH 2 at 20 mA cm À2 .This indicates the superior ORR activity on the CNT-NH 2 /PTFE electrode compared to on the CNT-NH 2 electrode.In addition, the selectivity of O 2 reduced to H 2 O 2 of these two electrodes was also analyzed.Figure 2b presents the Faradic efficiency (FE) of H 2 O 2 at 100 mA cm À2 , revealing a higher FE for H 2 O 2 on the CNT-NH 2 /PTFE electrode (77%) compared to that on the CNT-NH 2 electrode (73%).This represents another improvement for ORR upon the introduction of PTFE particles.With these data, we can estimate the cathodic energy efficiency of CNT-NH 2 /PTFE electrode to be 92%, substantially higher than that obtained on the CNT-NH 2 electrode (63%) at 100 mA cm À2 .Overall, these results demonstrate that the 2e À ORR of O 2 to H 2 O 2 on the CNT-NH 2 electrode can be promoted significantly by simply introducing hydrophobic PTFE particles into the catalyst layer.
To gain insight into the potential enhancement of O 2 transport within the catalyst layer, the ORR activities of these two electrodes were compared under various O 2 gas flow rates, as shown in Figure 2c (Figure S4 and S5, Supporting Information).The pure CNT-NH 2 electrode shows a weak dependence of the applied potential on the O 2 flow rate, with a slight increase from 0.69 to 0.72 V versus RHE as the flow rate increased from 1 to 5 sccm at a fixed current density of 100 mA cm À2 .In contrast, the potential on the CNT-NH 2 /PTFE electrode exhibits a distinct trend, rapidly increasing from 0.71 to 0.84 V versus RHE as the O 2 flow rate reached 5 sccm, after which it plateaus.As suggested previously, a higher O 2 flow rate results in more dissolved oxygen at the reaction interface. [36]Hence, if the ORR in GDE primarily occurred through reactions with dissolved O 2 at the traditional electrolyte-electrode interface, a notable dependence between the O 2 flow rate and electrode potential would be observed for both the CNT-NH 2 /PTFE and CNT-NH 2 electrodes.polarization curves and electrochemical impedance spectroscopy (EIS).The Tafel slopes are similar on CNT-NH 2 /PTFE and CNT-NH 2 catalysts, suggesting that ORR mechanism was not changed after adding PTFE into catalyst layer (Figure S6b, Supporting Information).EIS studies show a lower O 2 mass-transport resistance (R mt ) on the CNT-NH 2 /PTFE catalyst (Figure S7, Supporting Information).The correlation of frequency and imaginary part of complex capacitance (C 00 (ω)) was calculated to compare the resistance of ion transport. [48]In Figure S8, Supporting Information, the mild difference between both CNT-NH 2 /PTFE and CNT-NH 2 at various potentials is an indicative of the similar ion-transport resistance.Taken together, we show that the introduction of hydrophobic PTFE particles in the catalyst layer enhances the ORR performance by facilitating the accelerated gaseous O 2 mass transport at the triple-phase interface.
The electrochemical active surface areas of CNT-NH 2 /PTFE and CNT-NH 2 electrodes were estimated by measuring their corresponding electrochemical double-layer capacitance (C dl ).As shown in Figure 2d, CNT-NH 2 /PTFE exhibits an even smaller C dl (3.88 mF cm À2 ) than CNT-NH 2 (4.71 mF cm À2 ).The electric C dl arises from the interactions between the GDE surface charges and the counter ions within the electrolyte.Therefore, the smaller C dl observed for the CNH-NH 2 /PTFE electrode indicates the less solid-liquid interfaces and, in turn, more triple-phase interfaces.These results demonstrate that the presence of PTFE dispersed in the catalyst layer promotes the O 2 masstransport process by creating a hydrophobic microenvironment, resulting in dominant solid-liquid-gas interfaces, which is desirable for efficient ORR.
The hydrophobicity of the GDE surface induced by PTFE particles was further evaluated by contact-angle measurement.As shown in Figure 2e, it can be observed that CNT-NH 2 /PTFE electrode exhibits a contact angle of 147.6°slightly larger than CNT-NH 2 electrode (145.6°).At first glance, this indicates the apparent hydrophobicity of the GDE is not increased by the introduction of PTFE into the catalyst layer.However, it should be noted that the hydrophobicity of the catalyst layer would gradually change during the electrochemical reaction.In fact, GDE surfaces tend to become more hydrophilic as the electrochemical reaction proceeding, which causes the electrolyte to flood the catalyst layer, thereby inhibiting the mass transfer of gas-phase reactants and reducing the reaction rate.To assess whether the high hydrophobicity of these GDEs retains after catalysis, we remeasured the contact angles of the CNT-NH 2 /PTFE and CNT-NH 2 electrodes after ORR at 50 mA cm À2 for 1 h.As shown in Figure 2e, the contact angle decreased slightly from 147.6°to 140.0°after reaction for CNT-NH 2 /PTFE electrode.In contrast, CNT-NH 2 electrode exhibited a significantly decrease in contact angle, which dropped from 145.6°to 85.0°under the same conditions.The stability of catalyst was further evaluated by chronopotentiometry at a fixed current density of 100 mA cm À2 .As shown in Figure S9, Supporting Information, the cathode electrode potential of CNT-NH 2 /PTFE was consistently maintained around 0.75 V throughout >8 h of operation, with an approximate 75% FE observed for the entire duration of the test.In contrast, a decrease in H 2 O 2 FE was observed for bare CNT-NH 2 after 6 h.In contrast, as shown in Figure S10-S12, Supporting Information, negligible changes are observed in terms of morphologies for both CNT-NH 2 /PTFE and CNT-NH 2 after reactions, indicating no obvious changes occurred to the catalyst itself.Therefore, we believe that the hydrophobic PTFE acts as a protective element, preserving the hydrophobic nature of the catalyst layer during the reaction.This preservation helps maintain the solid-liquid-gas interface, facilitating the mass transfer of O 2 during ORR and reducing the concentration overpotential.
Furthermore, to evaluate the applicability of the PTFE effect, we measured the ORR performance of the PTFE-modified GDE based on commercial Pt/C catalyst (the state-of-the-art 4e À ORR electrocatalyst).Similarly, two types of Pt/C electrodes were prepared, Pt/C/PTFE electrode with the introduction of PTFE particles in the catalyst ink, and the pristine Pt/C electrode.These two GDEs also showed nearly identical morphology (Figure S13, Supporting Information).As shown in Figure 2f (Figure S14 and S15, Supporting Information), once again, Pt/C/PTFE exhibits substantially lower overpotentials to drive the ORR across a broad range of current densities, from 20 to 300 mA cm À2 .The difference is even more significant at large current densities, suggesting the dominate effect of O 2 mass transport within the conventional design.This result confirms that our strategy can be extended into systems with different catalysts, and potentially different gas involving electrochemical processes.

Investigating the PTFE Effect using RRDE
Although the catalytic interface in the RRDE configuration is not exactly identical to that in a flow cell, RRDE enables robust kinetic studies by precisely control the convection flow for efficient mass transportation.This is beneficial for studying the influence of the hydrophobic microenvironment created by PTFE on the kinetics of ORR.We first performed RRDE voltammetry measurements on catalysts with different loadings of PTFE. Figure 3a presents the recorded ORR polarization curves on CNT-NH 2 with an increasing loading of PTFE from 10% to 60%, using 0.1 M KOH as the electrolyte and a rotation rate of 1600 rpm.In the absence of PTFE, a normal S-shaped profile is obtained with a plateau current density of %3.3 mA cm À2 due to O 2 mass-transport limitations.However, the plateau current density for ORR exhibits an increase when PTFE is added into the catalyst layer.Specifically, the plateau current density reaches 5.4 mA cm À2 for a PTFE loading of 10% and 4.6 mA cm À2 for a higher PTFE loading.As shown in Figure S16a, Supporting Information, the selectivity of the 2e À ORR on CNT-NH 2 with PTFE are lower than on pure CNT-NH 2 .This result seems to be inconsistent with the results measured in the flow cell.However, it should be noted that there is not much difference in the ring current (Figure S16b, Supporting Information), i.e., the amount of H 2 O 2 detected by RRDE does not differ significantly, but the disk currents on PTFE-modified CNT-NH 2 are even higher.This implies that on PTFE-modified CNT-NH 2 , the generated H 2 O 2 on the disk of RRDE might be immediately further reduced before it could be transferred to the ring, resulting in a higher current.This indirectly proves that the CNT-NH 2 catalyst modified with PTFE has higher activity for 2e À ORR.As shown in Figure 3b, CNT-NH 2 catalysts with PTFE exhibit more positive half-wave potential (E 1/2 ) compared to that without PTFE.Furthermore, the E 1/2 increases from 0.63 to 0.65 V vs. RHE with the increase in PTFE mass ratio.This observation is further supported by the comparison of kinetic current density for ORR on CNT-NH 2 with different PTFE loadings (Figure S17, Supporting Information).These results align with the improved performance observed in the flow cell experiments.
To verify the aforementioned improved performance in ORR is primarily attributed to the optimized microenvironment instead of other factors, such as reaction pathway, morphology, conductivity, etc., we first conducted Tafel analysis for ORR on the CNT-NH 2 catalysts with different PTFE loadings (Figure S18, Supporting Information).It turns out that similar Tafel slopes were obtained for ORR on the catalyst with and without PTFE, about 60 mV dec À1 in each case.This Tafel slope indicates that the rate-determining step of ORR is likely to be the initial protonation of the surface-adsorbed O . Furthermore, we compared the electrochemical C dl and the morphology of CNT-NH 2 with and without PTFE.As expect, similar C dl were observed on all CNT-NH 2 catalysts (Figure S19, Supporting Information).Taken together, we believe that negligible changes occurred upon the addition of PTFE into catalyst layer regarding the catalyst morphology/structure and ORR reaction pathway.In addition, EIS was employed to analyze the conductivity, charge transfer, and O 2 masstransport kinetics for the aforementioned catalysts.As shown in Figure 3c, the EIS spectra of CNT-NH 2 under ORR conditions consist of two semicircles in high-and low-frequency regions, respectively.The semicircle locates at high-frequency region is likely associated to the ORR charge-transfer resistance (R ct ), while the one at low-frequency region can be attributed to the O 2 mass transport. [49]Consequently, the simulated solution resistance (R s ), R ct , and O 2 R mt of CNT-NH 2 samples are compared in Figure 3d.At a given potential, nearly identical solution resistances (around 43 Ω) were obtained for different samples, indicating negligible difference existed in electrode conductivity/solution resistance.Similarly, the estimated R ct of ORR on pure CNT-NH 2 is similar with those with PTFE, suggesting that the introduction of PTFE into the catalyst layer does not affect the electron transfer rate during ORR on CNT-NH 2 .In contrast, the R mt decreases dramatically upon the addition of PTFE, i.e., R mt decreased from 111.5 to 19.3 Ω with only 10% PTFE added into the CNT-NH 2 catalyst layer.At larger overpotentials (O 2 -diffusion-limiting regions), notably smaller R mt was observed on the PTFE-modified CNT-NH 2 compared to the pristine CNT-NH 2 electrode (Figure S20, Supporting Information).Additionally, we analyzed the ion-transport resistance by plotting frequency against C 00 (ω), revealing minor differences between CNT-NH 2 /PTFE with various PTFE mass ratios and pristine CNT-NH 2 at different potentials, as depicted in Figure S21, Supporting Information.These EIS results further confirm that the enhanced ORR performance of CNT-NH 2 /PTFE can be attributed to the improved O 2 mass transport facilitated by the hydrophobic microenvironment induced by PTFE, with no significant impact on resistance related to ion transport, charge transport, or conductivity, consistent with EIS results in the flow cell at relatively high current densities.Moreover, this is supported by a comparison of the O 2 uptake rates (υ O 2 ) on CNT-NH 2 with varying PTFE loading.As shown in Figure 3e, CNT-NH 2 catalyst exhibits a moderate υ O 2 of 1.79 Â 10 À4 mmol cm À2 s À1 in absence of PTFE.Upon the addition of PTFE, υ O 2 increases with the loading of PTFE, reaching the highest υ O 2 of 2.46 Â 10 À4 mmol cm À2 s À1 at 60% PTFE.
Overall, based on the aforementioned RRDE data, we believe the introduction of PTFE particles into the catalyst layer does not alter the catalytic mechanism, the intrinsic activity of the catalyst, and the conductivity of the electrode.Hence, the enhanced ORR performance can be attributed to the facilitated O 2 mass transfer at the reaction interface.As shown in Figure 3f, the presence of PTFE could likely mitigate the concentration gradient of O 2 between electrode surface (x = 0) and bulk electrolyte (x = δ), and further reduce the concentration overpotential of ORR.However, the mass transport of O 2 has been optimized in the RRDE configuration.In addition, the ORR at the electrode/electrolyte interface in RRDE relies on dissolved O 2 .Hence, the PTFE effect on ORR performance in RRDE system is not as significant as that in the flow cell.In particular, the reduction in the apparent E 1/2 is only 22 mV on CNT-NH 2 /PTFE compared with CNT-NH 2 , while the reduction in ORR overpotential is over 130 mV at 100 mA cm À2 in the flow cell after the introduction of PTFE to the same catalyst.Once again, this implies that the improved ORR performance on CNT-NH 2 /PTFE in the flow cell configuration is primarily attributed to the enhanced gas-phase O 2 mass transport rather than dissolved O 2 .

Effect of PTFE Loading on the Microenvironment
To achieve an optimal gas/solid/liquid triple-phase microenvironment inside the catalyst layer, the loading of the PTFE particle was optimized.Figure 4a, S22, Supporting Information, show the electrode potential on CNT-NH 2 /PTFE with various PTFE mass ratios at 100 mA cm À2 .The electrode potential exhibited a volcanic relationship with PTFE mass ratio, where the potential first increased from 0.69 to 0.83 V versus RHE as PTFE mass ratio increased from 0% to 40% and dropped to 0.66 V versus RHE when the PTFE mass ratio was further increased to 60%.Similar phenomenon occurred with regarding the ORR selectivity (Figure 4b), and the FE of H 2 O 2 also first increased along with the PTFE loading, however, declined at high PTFE mass ratio of 60%.Consequently, as shown in Figure 4a, the cathodic energy efficiency also shows a strong dependence on the PTFE mass ratio.As expected, a maximum cathodic energy efficiency of 92% is obtained with 40% PTFE loaded into the catalyst layer.These results suggest that the ORR energy efficiency can be improved substantially by reducing the concentration overpotential, and one effective strategy is to introduce hydrophobic (i.e., PTFE) particles directly to the catalyst layer.
We also measured the contact angles of the CNT-NH 2 electrodes with different PTFE mass ratios (Figure S23, Supporting Information).As the previous results, the contact angles measured on the GDEs are not altered before the ORR test, at least not significantly.However, different contact angles were observed for CNT-NH 2 electrodes with different PTFE loadings after the ORR test.As shown in Figure 4c, with all the PTFE loading studied, the hydrophobicity of electrode layer was largely retained after ORR.
To understand the interplays among electrode hydrophobicity, gas/liquid-phase balance, and ORR performance, we plotted an illustration based on our system and results, as shown in Figure 4d.First, the hydrophobicity of electrode surface during reaction can be retained by adding PTFE particles to the catalyst layer, as verified by the aforementioned contact-angle measurements.In addition, the gas/liquid-phase balance could be affected by the hydrophobicity of the electrode.In absence of PTFE, the electrode tends to become more hydrophilic and the ORR occurs primarily at the electrolyte-electrode interface, predominantly involving dissolved O 2 in the electrolyte.In this scenario, the ORR performance is largely constrained by the low solubility of O 2 , leading to significantly concentration overpotentials particularly at high reaction rates.Conversely, the hydrophobicity of electrode surface can be well maintained in presence of PTFE, ensures ORR predominately occurs at the solid-liquidgas interface.In this scenario, gaseous O 2 could easily participate in ORR at the triple-phase interface, avoiding the limitation of low solubility of O 2 , and reduce the concentration overpotential.However, when access amount of PTFE is loaded to the catalyst surface, the hydrophobic nature of the PTFE will repel the liquid electrolyte and sacrifice the optimized triple-phase boundaries for ORR, and further impede the ORR performance.Consequently, the cathodic energy efficiency exhibits a volcano-type correlation on the PTFE loading (Figure 4e).To probe this hypothesis, computational fluid dynamic (CFD) simulations were conducted to analyze the solid-liquid-gas contact state for CNT-NH 2 with different hydrophobicity.Figure 4f,g shows the formation of continuous gas layer on the surface of CNT-NH 2 /PTFE with high hydrophobicity, while an incontiguous gas layer is formed above the surface of CNT-NH 2 with low hydrophobicity.On the one hand, although the O 2 /H 2 O volume fraction at the oxygen inlet is close to 1 in both scenarios, the O 2 /H 2 O volume fraction farther from the O 2 inlet on the catalyst surface is higher when the surface possesses a higher hydrophobicity.O 2 transported to the more hydrophobic surface catalyst tends to form a dense gas layer instead of diffusing into the electrolyte, resulting in a higher O 2 concentration on the catalyst surface and reduction in the concentration overpotential.On the other hand, when the catalyst surface is less hydrophobic, the catalyst contacts the electrolyte directly.Thus, the CNT-NH 2 without PTFE modification will be flooded more easily than that modified by PTFE, which rationalizes its smaller contact-angle post reaction.In addition, the potential influence of surface pressure stemming from hydrophobicity was investigated by CFD.As shown in Figure S24, Supporting Information, the surface pressure on the CNT-NH 2 /PTFE electrode is approximately 3 Â 10 6 Pa, which is slightly smaller to that observed on the pristine CNT-NH 2 electrode (%4 Â 10 6 Pa).This result indicates that the impact of surface pressure stemming from hydrophobicity is likely minimal.Taken these together, we have observed improved 2e À ORR performance upon the introduction of PTFE using our approach compared to the previous works (Figure S25, Supporting Information).Overall, an appropriate amount of PTFE particles is critical in maintaining the hydrophobicity of the catalyst layer and ensure the optimal balance of the triple-phase interfaces for efficient ORR.

Conclusion
In summary, we demonstrate a facile and effective strategy in enhancing the performance of the 2e À ORR for the production of H 2 O 2 via microenvironment modulation.Encouragingly, by simply adding hydrophobic PTFE particles into the catalyst layer, the energy efficiency of O 2 reduced to H 2 O 2 can be improved substantially, and with modest increase in product selectivity.Detailed kinetic analysis reveals that the improved ORR performance is largely attributed to the enhanced gas-phase O 2 transportation, which significantly reduces the concentration overpotential during ORR.We believe that the hydrophobic microenvironment modulation induced by the PTFE particles in the catalyst layer is the origin of this promotion effect.Furthermore, the loading of PTFE in the catalyst layer is critical in optimizing the balance of triple-phase interface to accelerate the ORR.Overall, we believe our strategy of modulating the hydrophobic microenvironment of the catalyst layer can be applied to broader research directions that rely on gaseous reactant.
Preparation of GDE for ORR in the Flow Cell: The incorporation of PTFE into catalyst layer is a physically mixing process.In this process, the catalyst suspension ink was prepared by dispersing 10 mg of catalyst (CNT-NH 2 or Pt/C), different mass ratio of PTFE powder, and 50 μL 5% Nafion in 950 μL mixture of isopropanol (610 μL) and water (340 μL).The ink was further sonicated for 15 min using.The catalyst suspension ink then was sprayed on the carbon paper by a spray gun used as GDE with a loading amount around 0.7 mg cm À2 .Then, the prepared GDE was dried at room temperature for standby.
Preparation of Electrode for ORR in the RRDE: An RRDE setup (RRDE-3A, ALS Co. Ltd) containing a glassy carbon disk electrode and a platinum ring with a collection efficiency of 0.38 was used.The same catalyst (CNT-NH 2 or Pt/C) ink as that used in GDE preparation was pipetted onto a clean glassy carbon ring disk by 7.5 μL to reach the loading amount of around 0.3 mg cm À2 , and then dried in the fume hood.
Materials Characterization: The SEM was carried out on a JEOL JEM-7610M at 5 kV equipped with EDS analyzed at 15 kV.Powder X-Ray diffraction (PXRD) patterns were recorded on a Bruker D8-advance X-Ray powder diffractometer operated at 40 kV and 30 mA with CuKα radiation (λ = 1.5406Å).FTIR spectroscopy was recorded using VERTEX 70 FTIR spectrometer.
Electrochemical Measurements: In the flow cell, the electrochemical performance was evaluated using a three-electrode system on Biologic VMP3e electrochemical workstation.The prepared GDE, a Pt sheet, and a Ag/AgCl electrode were served as working, counter, and reference electrode, respectively.The geometry surface area of the working electrode was fixed to 1 cm 2 .The catholyte and anolyte were both 1.0 M KOH, the flow rates were both 1.5 mL min À1 .The cathode side was supplied with O 2 gas in various flow rates.For the RRDE test, the prepared RRDE with catalysts, a carbon rod, and a Hg/HgO electrode were used as working, counter, and reference electrodes, respectively.The ORR performance measurement was performed in O 2 saturated 0.1 M KOH aqueous solution.LSV curves were conducted at a scan rate of 10 mV s À1 and an electrode rotation speed of 1600 rpm with a ring potential of 1.48 V versus RHE.Tafel plot was recorded using LSV method from 0.95 to 0.8 V versus RHE at a scan rate of 2 mV s À1 .
The electron transfer number (n) in the RRDE was calculated by the following equation where I d and I r are the current recorded on the glassy carbon disk and Pt ring of the RRDE, respectively, and N is the collection efficiency (0.38).
Kinetic current density ( j k ) was calculated according to the equation C dl is the electrochemical double-layer capacitance derived from the slope of linear fit of the difference between cathodic and anodic current density against the scan rate using cyclic voltammogram method at a non-Faradic potential window.
Oxygen uptake rate (υ O 2 ) was calculated based on where FE H 2 O 2 ð Þis the Faradic efficiency of H 2 O 2 and j d is the current density on the glassy carbon disk of RRDE.Solution resistance (R s ), R ct , and O 2 R mt were obtained by fitting of Nyquist plot according to the equivalent circuit using ZView software.
H 2 O 2 Qualification and Calculation of Cathode Energy Efficiency: After ORR electrocatalysis, the amount of generated H 2 O 2 was determined by the standard potassium permanganate titration process according to the following reaction equation The FE of H 2 O 2 was calculated based on The cathodic energy efficiency was calculated using the following equation Cathodic energy efficiency ¼ where E Θ is the standard electrode potential for O 2 reduced to H 2 O 2 , E is the electrode potential measured experimentally, and FE is the Faradaic efficiency of H 2 O 2 .COMSOL Multiphysics Simulation: COMSOL Multiphysics 5.6 was employed to perform computational analyses in surface-tension-dominant two-phase flows. [50]We systematically adjusted the hydrophobicity of the electrode by manipulating the contact angle of the material within an oxygen-water system.The contact-angle values were determined based on contact-angle measurements.Specifically, a contact angle of 140°w as employed for CNT-NH 2 /PTFE, while a contact angle of 85°was used for pure CNT-NH 2 .The momentum equations governing the entirety of the flow field are expressed as follows [51] ρ ρ Â ∇ Â u ¼ 0 (9) where ρ (kg m À3 ) is the density, u (m s À1 ) is the velocity components in y direction, t (s) is the time, P (Pa) is the pressure, τ is the deviatoric stress tensor for a Newtonian fluid, F (N m À3 ) is the surface tension force per unit volume, g (m s À2 ) is the acceleration due to gravity, μ (N s m À2 ) is the dynamic viscosity, T (K) is the temperature (293.1 K in this work), σ (N m À1 ) is the surface tension coefficient (0.072 N m À1 in this work), k (m À1 ) is the interfacial curvature, n is the interfacial unit normal vector, and δ is the delta function centered at the interface.

Figure 1 .
Figure 1.a) Scanning electron microscope (SEM) images and b) Fourier-transform infrared (FTIR) spectra of the pristine amino-group-functionalized carbon nanotube (CNT-NH 2 ) sample.c) SEM images of CNT-NH 2 /polytetrafluoroethylene (PTFE) sample.PTFE particles are circled in yellow.d) Powder X-Ray diffraction (PXRD) pattern comparison of CNT-NH 2 /PTFE and pure CNT-NH 2 samples.e) Corresponding energy-dispersive X-ray spectrometer (EDS) element mapping images of CNT-NH 2 /PTFE sample taken from the SEM image in Figure 1c.

Figure 2 .
Figure 2. a) Potentials for oxygen reduction reaction (ORR) on CNT-NH 2 /PTFE and CNT-NH 2 electrodes at various current densities with an O 2 flow rate of 5 sccm.b) Faradic efficiency (FE) of H 2 O 2 and cathodic energy efficiency on CNT-NH 2 /PTFE and CNT-NH 2 electrodes at 100 mA cm À2 .c) Potentials for ORR at various O 2 flow rates on CNT-NH 2 /PTFE and CNT-NH 2 electrodes at 100 mA cm À2 .d) C dl fitting plots of CNT-NH 2 /PTFE and CNT-NH 2 electrodes.e) Contact-angle measurement images of CNT-NH 2 /PTFE and CNT-NH 2 electrodes before and post reaction.f ) Potentials for ORR on Pt/C/PTFE and Pt/C electrodes at various current densities with an O 2 flow rate of 5 sccm.

Figure 3 .
Figure 3. a) Rotating ring-disk electrode (RRDE) voltammograms collected on CNT-NH 2 /PTFE with different PTFE mass ratios at a scan rate of 10 mV s À1 and an electrode rotation rate of 1600 rpm in O 2 -saturated 0.1 M KOH solutions.b) E 1/2 comparison.c) Nyquist plots of these CNT-NH 2 /PTFE sample (inset is the equivalent circuit).d) Relationship between R s , R ct , and R mt with PTFE mass ratio at 0.7 V versus RHE.e) Relationship between O 2 uptake rate and PTFE mass ratio at 0.65 V vs. RHE.f ) Schematic illustration of the reduced concentration overpotential caused by the enhanced O 2 mass transport.

Figure 4 .
Figure 4. a) Potentials, FEs of H 2 O 2 , and b) cathodic energy efficiencies for ORR at 100 mA cm À2 on CNT-NH 2 /PTFE electrodes with various PTFE mass ratios added into catalyst layer.c) Difference of contact angle (Δθ) of CNT-NH 2 /PTFE electrodes with various PTFE mass ratios before and after 1 h electrolysis at 100 mA cm À2 .d) Relationships among Δθ, R mt , and cathodic energy efficiency on CNT-NH 2 /PTFE electrodes with various PTFE mass ratios.Error bars represent the standard deviation of three independent replicate experiments.Simulation results of solid-liquid-gas triple interface of f ) CNT-NH 2 /PTFE, and g) CNT-NH 2 electrodes.