Surface Processes Control the Fate of Reactive Oxidants Generated by Electrochemical Activation of Hydrogen Peroxide on Stainless-Steel Electrodes

Low-cost stainless-steel electrodes can activate hydrogen peroxide (H2O2) by converting it into a hydroxyl radical (•OH) and other reactive oxidants. At an applied potential of +0.020 V, the stainless-steel electrode produced •OH with a yield that was over an order of magnitude higher than that reported for other systems that employ iron oxides as catalysts under circumneutral pH conditions. Decreasing the applied potential at pH 8 and 9 enhanced the rate of H2O2 loss by shifting the process to a reaction mechanism that resulted in the formation of an Fe(IV) species. Significant metal leaching was only observed under acidic pH conditions (i.e., at pH <6), with the release of dissolved Fe and Cr occurring as the thickness of the passivation layer decreased. Despite the relatively high yield of •OH production under circumneutral pH conditions, most of the oxidants were scavenged by the electrode surface when contaminant concentrations comparable to those expected in drinking water sources were tested. The stainless-steel electrode efficiently removed trace organic contaminants from an authentic surface water sample without contaminating the water with Fe and Cr. With further development, stainless-steel electrodes could provide a cost-effective alternative to other H2O2 activation processes, such as those by ultraviolet light.


Text S1. Pre-condition of the stainless-steel electrode
To provide a stable performance, a new Scotch-Brite 20 g stainless-steel scrubber (catalogue number 214C, 3M Company, St. Paul, MN, USA; 80 cm 2 /g specific surface area) was cut to around 4.5 g and pre-conditioned for each operating condition tested (i.e., pH and potential).The same electrode was used for all experiments under that operating condition.No significant decrease in the yield of oxidants was observed over time except in experiments conducted at pH 6 and +0.020 V.In these experiments, decreases in performance were observed after pre-conditioning and after completion of 15 experiments.After the decreases in performance were observed, another electrode was pre-conditioned following the same protocol and used for the remaining experiments.Before and after each experiment, the stainless-steel electrodes were rinsed thoroughly with ultrapure water and air-dried for at least 30 minutes before being reused.
The pre-conditioning process was performed with the working electrode chamber operated in continuous stirred-tank reactor mode to avoid any metal accumulation.The working electrode chamber was fed with a 0.2 M Na2SO4 electrolyte that containing methanol (100 mM), During the pre-conditioning process, new stainless-steel electrodes underwent a period of rapid corrosion which was accompanied by the release of metals and variation in performance.
After approximately one hour, the leaching of metals substantially decreased and the kinetics of H2O2 activation became more reproducible.When operated at pH values below 7 and a potential of +0.020 V in continuous flow mode, elevated metal concentrations were observed within 20 minutes (Figure S2C-E).At higher pH values, substantial metal leaching was never observed.
Therefore, pre-conditioning (i.e., operating the electrode for one hour) followed by disposal of leached metals may be needed prior to treating water with an initial pH value below 7.Under all tested conditions, the electrode performance stabilized within one hour in terms of H2O2 activation, reactive oxidant production, metal leaching, and current density (Figure S2).Surface water was collected during a storm event from Strawberry Creek on the University of California, Berkeley campus (37°52'26.9"N122°15'41.8"W)on December 22, 2021, around 1:30 PM, Pacific Time.The surface water runoff was filtered through 0.7 µm glass fiber filter (MilliporeSigma, Burlington, MA) followed by 0.45 µm Supor® 450 Membrane (Gelman Sciences, Ann Arbor, MI) prior to experiment.A summary of the water quality parameters is provided in Table S1.

Ag/AgCl
Pt mesh The collected surface water was amended with 200 mM of Na2SO4 to avoid overloading of the potentiostat.To test the performance of the stainless-steel electrode under circumneutral pH conditions, the initial pH of the surface water was adjusted from 7.8 to 6.0 with diluted H2SO4 prior to the electrolysis.(Solution acidification on this magnitude might also occur if the electrode had been part of a three-electrode system in which the cathodes for H2O2 production and activation were preceded by an anode.)The pH of the solution increased from 6.0 to around 7.3 gradually during electrolysis, potentially caused by CO2 partitioning between the solution and atmosphere.Based on its alkalinity, the equilibrium pH was estimated to be 8.0 when the water was equilibrated with the atmosphere.
Control experiments indicated that H2O2 did not react with carbamazepine or atrazine in the absence of stainless-steel electrode (Figure S6).Control experiments were also conducted in the absence of H2O2 to assess contaminant removal by processes taking place on the electrodes (e.g., direct electron transfer).Results indicated less than 10% loss of atrazine and about 25% loss of carbamazepine over the two-hour experiment (Figure S7).This was substantially lower than the removal observed in the presence of H2O2 (Figure 6A).

Text S4. Prediction of the reaction rate constant based on mass-transport limitation
The mass-transport limiting current density was predicted based on film theory.The mass-transfer coefficient km was predicted by: where D is the diffusion coefficient of H2O2 in water (1.5 × 10 −5 cm 2 s −1 ), 8,9 and Lf is the film thickness (about 100 μm). 10,11 rate constant limited by mass transport, klimt, was predicted by: where a is the surface area to volume ratio (m 2 /m 3 ).The reaction rate constant for the reaction between • OH and the electrode surface  surface,•OH was calculated by rearranging eq.3 into a linear form (eq.S3, Figure S17) and subtracting out contributions from H2O2, Fe 2+ and the buffer to • OH scavenging.The second-order rate constants for the reactions of buffers with • OH were measured using competition kinetics as  MES,•OH = 2.1±0.1 ×10 9 M −1 s −1 and  PIPES,•OH = 4.2±0.1×10 9M −1 s −1 , respectively (Text S7).The reaction rate constant for the reaction between the electrode surface and • OH was calculated as 6.5 × 10 5 g −1 s −1 at pH 6 and 4.9 × 10 5 g −1 s −1 at pH 7.
Text S6.Prediction of the fate of • OH.
The fraction of • OH that reacted with the electrode surface and aqueous species was estimated by: where j represents species that reacts with • OH (e.g., electrode surface, competing organic compounds, Fe 2+ , H2O2).The rate constant for the electrode surface and • OH was estimated as the average of the values observed under pH 6 and pH 7 (i.e., 5.7× 10 5 g −1 s −1 ).The competing Text S7.Reaction rate constants for • OH with organic buffer compounds.
Second-order rate constants for the reaction of • OH with MES and PIPES were measured using competition kinetics. 12Briefly, individual test compounds (MES or PIPES, 1.0 μM) was irradiated in a customized brown glass bottle (Veffective = 600 mL), using a 9 W low-pressure Hg UV-C lamp (arc length = 12.5 cm, Anyray, US).The solutions contained 0.5 μM carbamazepine as a reference compound  carbamazepine,•OH = 9.1 × 10 9 M −1 s −1 , 13 along with H2O2 as a photosensitizer (10 μM).Carbamazepine was chosen as a reference compound because it exhibits low rate of direct photolysis. 14ution pH was buffered with 100 μM phosphate buffer at 6.0 for the MES experiment and 7.0 for the PIPES experiment, respectively.The pH changed by < 0.2 units through the photolysis experiments.Details of the analytical method are described in Text S8.Because the reaction between the test compounds and H2O2 are slow 15 and no H2O2 consumption was observed when test compounds were added to H2O2-contaning solution (Figure S3), the direct oxidation of the test compounds by H2O2 was negligible over the time scale of this study.
Control experiments for direct photolysis of the test compounds and carbamazepine were conducted under similar experimental conditions in the absence of H2O2.The pseudo 1 st -order rate constants for direct photolysis of carbamazepine and test compounds in the absence of H2O2 were less than 10% of the values observed in the presence of H2O2 (Figure S18).Due to the low molar absorptivity of MES and PIPES, 16,17 the contribution of direct photolysis to the overall phototransformation in the presence of H2O2 was neglected because the presence of H2O2 was expected to further slowdown direct photolysis reactions due to competition for photons.

125 2 -3×10 - 2 pH
Figure S3.H2O2 concentrations in the presence of 100 mM of (A) methanol, (B) 2-propanol in buffer-containing Na2SO4 electrolyte.The experiments were conducted in the absence of stainless-steel electrodes.No formation of formaldehyde or acetone was observed.Error bars represent one standard deviation.

Figure S5 .
Figure S5.Schematic configuration of the undivided electrochemical reactor.

Figure S6 .
Figure S6.Concentrations of (A) H2O2 and (B) trace organic contaminants in Na2SO4-amended surface water in the absence of stainless-steel electrode.[H2O2]0 = 1.25 mg/L.Error bars represent one standard deviation; error bars not shown are smaller than symbols.

Figure S7 .0
Figure S7.Concentrations of trace organic contaminants in Na2SO4-amended surface water in the absence of H2O2.Potential = +0.020V. Error bars represent one standard deviation; error bars not shown are smaller than symbols.

Figure S13 .
Figure S11.(A) Current densities observed under different experimental conditions.(B)Relationship between the observed current densities and initial H2O2 concentration.Applied potential = +0.020V, pH = 6.Experiments conducted in deareated solution was purged with N2 for at least 20 minutes before the experiment and was continuously purged with N2 throughout the experiments in the sealed H-cell reactor.The flow rate of the N2 stream was maintained at 0.5 L/min.Error bars represent one standard deviation.

Figure S14 .
Figure S14.Atomic concentration of Fe in different oxidation states.Potential = +0.020V, The * symbol represents experiments conducted at potentials lower than +0.020V (i.e., -0.039 V at pH 8 and -0.098 V at pH9).Error bars represent one standard deviation.

Figure S15 .
Figure S15.Metal concentrations after five minutes of electrolysis under various experimental conditions.Potential = +0.020V,pH = 6.Error bars represent one standard deviation.Experiments conducted in deareated solution was purged with N2 for at least 20 minutes before the experiment and was continuously purged with N2 throughout the experiments in the sealed H-cell reactor.The flow rate of the N2 stream was maintained at 0.5 L/min.Error bars represent one standard deviation.

1 [
Figure S17.Linear relationship between inverse of formaldehyde yield and inverse of concentration of methanol.Error bars represent one standard deviation.

Table S1 .
Water quality parameters of the surface water.

Table S2 .
Reactions considered for construction of Pourbaix diagrams.