Direct synthesis of hydrogen peroxide from plasma-water interactions

Hydrogen peroxide (H2O2) is usually considered to be an important reagent in green chemistry since water is the only by-product in H2O2 involved oxidation reactions. Early studies show that direct synthesis of H2O2 by plasma-water interactions is possible, while the factors affecting the H2O2 production in this method remain unclear. Herein, we present a study on the H2O2 synthesis by atmospheric pressure plasma-water interactions. The results indicate that the most important factors for the H2O2 production are the processes taking place at the plasma-water interface, including sputtering, electric field induced hydrated ion emission, and evaporation. The H2O2 production rate reaches ~1200 μmol/h when the liquid cathode is purified water or an aqueous solution of NaCl with an initial conductivity of 10500 μS cm−1.


Experimental setup
shows the experimental setup and a photograph of the plasma-liquid interactions. The cylinder-like cell was made from polytetrafluoroethene and its geometry parameters are also indicated in Fig. S1(a). A home-made direct current power source was used to ignite the Ar atmospheric pressure discharge plasma between a tungsten steel tube and an aquesous solution surface. A 10-k resistor was connected in series with the tungsten steel electrode to avoid the plasma transfer from glow-like discharge to arc. The Ar flow rate was 20 sccm. A graphite rod (5 mm in diameter) was placed at the bottom of the solution to act as an inert electrode. The initial conductivity of the solution was adjusted by dissolving NaCl or NaOH in purified water. The purified water has a conductivity of 1.60 S cm -1 . A peristaltic pump was used to circulate the 400 ml solution at a rate of 200 ml/min. Unless stated otherwise, the solution acts as cathode (positive voltage applied to the tungsten steel electrode), the discharge gap was 3 mm, and the discharge current was 30 mA.
The voltage between the tungsten steel and the graphite electrodes was measured by a high voltage (H.V.) probe (Tektronix P6015A) and the current was achieved from dividing the voltage across a 10- resistor which was in series connected with the graphite electrode. The pH value and temperature of the solution were measured by a pH detector with a temperature sensor (Yesmylab SX620), and the solution conductivity was measured by a conductivity detector (Yesmylab SX650). Figure S2 presents the discharge voltage evolution when the discharge current is 30 mA. The results demonstrate that the voltage for the purified water changes very quickly before 20 min plasma operation, while it almost keeps constant for liquids with high initial conductivity. The variation of the solution conductivity shown in Fig. S3 might account for this phenomenon. Except for NaOH (its conductivity being almost constant), the increase of the solution conductivity is about 600 S cm -1 after 60 min plasma treatment. For the solutions with initial conductivities of 1440 S cm -1 , 4800

Electrical characterziaiton of the discharge plasma
S cm -1 , and 10500 S cm -1 , these changes are not enough to change the discharge voltage so much, but for the purified water which has an initial conductivity of 1.60 S cm -1 , the change of conducivity is relatively large (258 S cm -1 after plamsa treatment of 20 min, almost 160 times of the initial conductivity of 1.60 S cm -1 ), and thus the discharge voltage changes very quickly in the first 20 min plamsa treatment in the case of purified water. For all the cases, the discharge plasma is a glow-like discharge, and therefore we only show the dependence of the discharge voltage on the discharge current for one case in Fig. S4 which demonstrates a glow-like characteristic of the plasma, i.e., the voltage is almost independent on the current.

Investigation of the cathode voltage fall
As shown in

The pH value and temperature change of the liquid during plasma treatment
During the plasma treatment, as shown in Fig. S6(a)

H 2 O 2 yield measurement
Because H2O2 can react with titanium sulfate in strong acid to form H2TiO4 (Ti 4+ + H2O2 + 2H2O → H2TiO4 + 4H + ) and the absorption intensity of the yellow-coloured H2TiO4 in 410 nm is proportional to the reacted H2O2 concentration 6-9 . We can use it to determine the synthesized H2O2 concentration. 7.5 ml [Ti(SO4)2,120 g/l) was added to 250 ml H2SO4 (1.5 M) to obtain the test solution of titanium sulfate. We used H2O2 with standard concentrations to obtain the proportionality between the absorption intensity of H2O2 at 410 nm, and the results are presented in Fig. S9(a). Once the proportionality is obtained, the H2O2 yield is estimated by the following equation: where k is the proportionality obtained by linearly fitting Fig. S9(b), I is the absorption intensity of synthesized H2O2 at 410 nm, and V is the solution volume (in our case, 400 ml).

Spectroscopic diagnostic of the discharge plasma
We also compared the optical emission spectra for plasmas using liquid cathode and anode, optical emission spectroscopy (Ocean Optics USB2000+) was used to detect the optical emission spectra and the optical fiber was located near the liquid surface and 20 mm away from the plasma. The results are shown in Fig. S10. Evidently, the emission intensity of OH line (309 nm) in the case with a solution cathode is much higher than that with a solution anode. Also, Na line (589 nm) is hardly detected in the case of solution andoe, while its emission intensity is very strong in the plasma using a solution cathode. In the case of liquid cathode, Na is drawed out from the solution phase by positive ion sputtering and excited in the discharge zone which finally emits 589 nm light. These results indirectly support our conclusion of the water constituents transfer at the plasma-liquid interface.  Figure S9. Optical emission spectra of plasmas using a solution cathode and anode.