Supplementary Information. Carolyn Richmonds, Megan Witzke, Brandon Bartling, Seung Whan Lee, Jesse Wainright,

Transcription

1 Supplementary Information Electron transfer reactions at the plasma-liquid interface Carolyn Richmonds, Megan Witzke, Brandon Bartling, Seung Whan Lee, Jesse Wainright, Chung-Chiun Liu, and R. Mohan Sankaran*, Department of Chemical Engineering, Case Western Reserve University, Cleveland, OH *To whom correspondence should be addressed: Plasma-assisted electrochemical reduction of ferricyanide to ferrocyanide Electrochemical experiments were performed in a two compartment glass cell separated by a glass frit. The catholyte and anolyte solutions consisted of 0.2 mm potassium ferricyanide [K 3 Fe(CN) 6 ] and 0.1 M KCl in deionized water, and 0.1 M KCl in deionized water, respectively. A stainless steel capillary tube (180 μm inside diameter, 5 cm length, Varian, Inc) was positioned 1-2 millimeters above the surface of the catholyte bath to ignite a plasma and serve as a gaseous cathode, and an Ag/AgCl mesh (12.5 x 12.5 mm Ag mesh, mm dia. wire, Alfa Aesar) chlorided in 3 M KCl at 1 ma/cm 2 was immersed in the anolyte to serve as the anode. Pressurized argon gas was coupled to the metal capillary tube and controlled by a mass flow controller (Model UFC 1000, Unit Instruments, Inc.) to provide a constant gas flow rate of 30 sccm (standard cubic centimeters per minute) which was low enough to prevent the solution surface from being disturbed. A negative DC power supply was used to ignite and sustain an atmospheric-pressure plasma at currents of either 3 or 6 ma at the surface of the catholyte. Experiments were run for time increments ranging from 2 to 15 min. The catholyte solution was then collected for UV-vis absorbance spectroscopy, cyclic voltammetry, and ph measurements. S1

2 The metal capillary tube was cleaned between trials by pushing a tungsten wire through the capillary tube, polishing the outside surface with sand paper, and sonicating in acetone. Voltage and current measurements The electrochemical cell consisting of a plasma cathode and an Ag/AgCl anode was operated with an electrical circuit shown in Fig. S1. A large ballast resistor on the cathode side, R cathode =325.2 kω, was used to control the discharge current. A smaller resistor on the ground side, R anode =500 Ω, was also used. We measured the voltage drops across these resistors to obtain and compare the discharge currents on the cathode and anode sides (discharge current and cell current are synonymous in our case). Voltages greater than 1000 V were measured with a high voltage probe (Oscilloprobe P4100, 100:1 V). Table S1 shows typical values of the voltage drops across various parts of the electrical circuit we note that the voltage from the power supply is distributed across R cathode, the gas discharge, the cell, and R anode. The voltage drop across the gas discharge and the cell are combined and shown as V cathode-anode. As the power supply voltage is raised, the voltage drops increase and, therefore, the current through the cell (or discharge) is increased. The current was obtained from either the voltage drop across R cathode or R anode. The two currents (specified as i cathode and i anode ) were found to differ by less than 2% (Table S2), confirming that there were negligible current losses through the electrochemical cell. Solution potential measurement In order to understand the electrochemical reaction at the plasma-liquid interface, we made several attempts to measure the potential in this region. Note that electrochemical potentials are normally obtained directly from the metal electrodes, either as the voltage S2

3 difference between the cathode and anode electrodes or as the voltage difference between the cathode or anode and a reference electrode this is not straightforward in our case because the electrode of interest, the cathode, is at a very high voltage and there is a voltage drop from the metal (the stainless steel capillary tube), across the gas discharge, to the surface of the electrolyte. To obtain the solution potential at the plasma-liquid interface (which is the potential that actually drives the electrochemical reaction), a Pt wire was platinized with a measured surface area of 44 cm 2 and placed in contact with the solution as close as possible to the point where the plasma impinged on the solution surface. We then measured the voltage between the Pt probe and the Ag/AgCl anode at different discharge (i.e. cell) currents as shown in Fig. S2. However, we found that these measurements corresponded to the overall impedance of the cell (due to the conductivity of the electrolyte and the glass frit) which is ~750 kω (given by the slope of the VI curve in Fig. S2) and did not allow a meaningful surface potential to be obtained. Alternatively, we measured the voltage between the Pt probe and an Ag/AgCl reference electrode placed in the same cathode bath. We measured an initial solution potential (before igniting the plasma) of 0.43 V. We independently calculated the solution potentials by using the Nernst equation which relates the electrochemical potential for a redox couple at different activities (or concentrations for dilute solutions): ln where is the standard state reduction potential for the ferricyanide/ferrocyanide couple vs. Ag/AgCl (for the ferricyanide/ferrocyanide couple, =0.36 V vs. SHE and =0.197 V for Ag/AgCl vs. SHE; therefore =0.163 V vs. Ag/AgCl), R is the gas constant, T is the S3

4 temperature, F is Faraday s constant, z is the number of electrons transferred, is the activity of the reduced species, and is the activity of the oxidized species. For an initial solution of nearly pure ferricyanide, the Nernst equation is used to obtain an ~0.43 V (assuming > and <0.0001), in agreement with our measurements. After the plasma was ignited, we found that this solution potential (with respect to the Ag/AgCl reference electrode) decreased, confirming the reduction of ferricyanide as predicted by the Nernst equation. UV-vis absorbance spectroscopy The ferricyanide concentration in solution (after plasma exposure) was monitored by ultraviolet-visible (UV-vis) absorbance spectroscopy (Shimadzu UV-1800). Solutions of ferricyanide have a well known absorbance spectrum (ferricyanide is yellow) while ferrocyanide does not absorb (ferrocyanide is colorless) see Fig. S3. A calibration curve was obtained by preparing solutions of potassium ferriycanide [K 3 Fe(CN) 6 ] and potassium ferrocyanide [K 4 Fe(CN) 6 ] with varying relative concentrations (total concentration=0.2 mm) in 0.1 M KCl (in deionized water) and relating the absorbance intensity at 420 nm. As expected from Beer s law (A=εCl), the relationship between absorbance intensity and ferricyanide concentration is linear (Fig. S4). Using this calibration curve for ferricyanide absorbance, we obtained the amount of ferricyanide reduced by the plasma at different discharge currents and times by performing UVvis absorbance spectroscopy on the catholyte solutions after plasma exposure and relating the absorbance intensity to the relative amount of ferricyanide by the following equation: % S4

5 To find the % ferricyanide reduced, the above equation was simply subtracted from 100: % The number of molecules that were reduced was then calculated by multiplying the above % ferricyanide reduced in solution by the catholyte volume (15 ml), initial ferricyanide concentration (0.2 mm), and Avogadro s number. As given by the following reaction, K 3 Fe(CN) 6 + e - K 4 Fe(CN) 6 the number of molecules of ferricyande reduced is equal to the number of electrons transferred from the plasma to ferricyanide. During plasma exposure, the ph of the solutions increased (as discussed in a following section). To ensure that the ph change did not influence the absorbance of ferricyanide/ferrocyanide, we also collected absorbance spectra and obtained a calibration curve for the absorbance of ferricyanide at different concentrations and phs. The ph was controlled by adding concentrated NaOH to our solutions. As shown in Fig. S5, the absorbance intensity of ferricyanide did not vary with ph over the range that was studied here and the calibration curve at ph=7 was valid. Cyclic voltammetry To further assess the reduction of ferricyanide, an independent analysis by cyclic voltammetry (CV) was performed. CV scans were obtained at 2000 rpm from -0.5 V to 0.6 V S5

6 with a CH Instruments potentiostat and workstation software and a Pine Instruments MSRX motor-speed controller. The three electrode system consisted of a Pt mesh counter electrode, an Ag/AgCl reference electrode, and a glassy carbon rotating disc working electrode from Pine Research Instruments. For each sample, we ran four cycles and averaged the last two scans representative scans are shown in Fig. S6. Solutions of potassium ferriycanide [K 3 Fe(CN) 6 ] and potassium ferrocyanide (K 4 Fe(CN) 6 ] in 0.1 M KCl with varying relative concentration (total concentration=0.2 mm) were prepared. A calibration curve was obtained by correlating the currents at 0.6 V in the CV scans to the relative amount (i.e. percent) of ferricyanide in the solutions (Fig. S7). The following linear equation was fit to the calibration curve and used to estimate the amount of ferricyanide reduced by the plasma: % To find the % ferricyanide reduced, the above equation was simply subtracted from 100: % We compared the values for % ferricyanide reduction obtained from our analysis of the CV data with the absorbance results and found excellent correspondence (Fig. S8). However, the absorbance results were considered to be more quantitative since CV measurements can be affected by diffusion and other transport limitations. ph measurements S6

7 In order to assess water electrolysis by the plasma, we measured the ph change in the catholyte bath. The half-cell reaction for water electrolysis that occurs at the cathode electrode (under acidic conditions) is: H + + e - ½ H 2 (g) and the relation between H+ concentration and solution ph is: log Thus, the ph change in the catholyte bath should provide an indication of H + reduction. The ph was obtained after plasma exposure using a digital ph meter (Accumet Basic Model AB15). The ph measurements are shown in Fig. S9 and the calculations for H+ reduction are shown in Table S2. Anodic dissolution Anodic dissolution experiments were carried out to understand the transfer of current through the plasma-electrochemical cell, from the cathode to the anode. The Ag/AgCl anode was replaced with a metal (Ag or Cu) that could undergo a measurable anodic dissolution (i.e. oxidation) as given by Faraday s law: S7

8 where i is the discharge (i.e. cell) current, t is the total time current is passed, M is the molar mass of the metal species oxidized, z is the valency of the metal ion, and F is Faraday s constant. The oxidation of metals proceeds as: M 0 z e- +M z+ where z is 1 for Ag and 2 for Cu. Ag foil (25 mm x 76 mm, ESPI Metals) or Cu wire (1 mm outside diameter) was used with 1 mm HCl or 1 mm HNO 3, respectively. The metal anodes were weighed using an analytical balance (Sartorius Model BP210D) prior to and after their oxidation to determine the weight lost for a given current and time. S8

9 Table S1. Voltage drops measured at various points in our electrical circuit shown in Fig. S1. The power supply voltage is distributed through a ballast resistor (R cathode ), the gas discharge, the cell, and a current measuring resistor (R anode ) [we note that the power supply was a negative high voltage source]. V cathode refers to the voltage drop across the ballast resistor (325.2 kω), V cathodeanode refers to the voltage drop between the cathode electrode (i.e. metal capillary tube) and the anode electrode (i.e. Ag/AgCl), and V anode refers to the voltage drop across the current measuring resistor (500 Ω) resistor. As the power supply voltage is increased, the voltage drops increase and the discharge (or cell) current increases. Power supply voltage (Volts) V cathode (Volts) V cathode-anode (Volts) V anode (Volts) S9

10 Table S2. Discharge (or cell) current as measured at the cathode (i cathode ) and anode (i anode ) side of the electrical circuit shown in Fig. S1. i cathode (ma) i anode (ma) Current difference (%) S10

11 Table S3. ph change and H + reduction as a result of plasma exposure. The catholyte solution initially contained 0.2 mm K 3 Fe(CN) 6 and 0.1 M KCl in deionized water (ph~7.34). Current (ma) Exposure Time (min) Initial ph Final ph # H + inital # H + final # H + reduced E E E E E E E E E E E E E E E E E E E E E E E E+14 S11

12 Figure captions Figure S1. Schematic of electrical circuit and voltage drops measured to obtain discharge current (i.e. cell current) at the anode and cathode sides. Figure S2. Cell potential versus discharge (i.e. cell) current for the plasma electrochemical cell. The cell potential was obtained by placing a Pt wire in contact with the solution surface at the plasma-liquid interface, and measuring the voltage with respect to the Ag/AgCl anode electrode. The linear curve indicates that we are measuring the ohmic resistance of the cell which is ~750 kω. Figure S3. Absorbance spectra of solutions of K 3 Fe(CN) 6 and K 4 Fe(CN) 6 at varying relative concentrations (total concentration=0.2 mm) with 0.1 M KCl in deionized water. Figure S4. Calibration curve for UV-vis absorbance measurements showing % ferricyanide in solution vs. absorbance intensity at 420 nm. The solutions contained K 3 Fe(CN) 6 and K 4 Fe(CN) 6 at varying relative concentrations (total concentration=0.2 mm) with 0.1 M KCl in deionized water. Figure S5. UV-vis absorbance measurements showing % ferricyanide in solution vs. absorbance intensity at 420 nm as a function of ph. The solutions contained K 3 Fe(CN) 6 and K 4 Fe(CN) 6 at varying relative concentrations (total concentration=0.2 mm) with 0.1 M KCl in deionized water. The ph in solution was controlled by adding concentrated NaOH. S12

13 Figure S6. Cyclic voltammogram (CV) scans of solutions of K 3 Fe(CN) 6 and K 4 Fe(CN) 6 at varying relative concentrations (total concentration=0.2 mm) with 0.1 M KCl in deionized water. Figure S7. Calibration curve for cyclic voltammetry (CV) measurements comparing % ferricyanide in solution to current at 0.6 V. The solutions contained K 3 Fe(CN) 6 and K 3 Fe(CN) 6 at varying relative concentrations (total concentration=0.2 mm) with 0.1 M KCl in deionized water. Figure S8. Comparison of percent ferricyanide reduced, as estimated from UV-vis absorbance spectrometry and cyclic voltammetry, at different discharge currents and times. Figure S9. ph of the catholyte initially containing 0.2 mm ferricyanide and 0.1 M KCl in deionized water after plasma exposure for different times. S13

14 Figure S1 S14

15 -1 Potential (V) vs. Ag/AgCl anode y=-757x Discharge current, i D (ma) Figure S2 S15

16 Absorbance % 25% 50% 75% 100% Wavelength (nm) Figure S3 S16

17 100 Percent Ferricyanide (%) Absorbance Figure S4 S17

18 Percent Ferricyanide (%) ph 7 ph 9 ph 10 ph Absorbance Figure S5 S18

19 2x10-5 0% 50% 1x % Current (A) 0-1x x x Potential (V) vs. Ag/AgCl Figure S6 S19

20 100 Percent Ferricyanide (%) x x10-5 Current at 0.6 V (A) Figure S7 S20

21 Percent Ferricyanide Reduced (%) ma (CV results) 6 ma (CV results) 3 ma (UV-vis absorbance results) 6 ma (UV-vis absorbance results) Time (min) Figure S8 S21

22 ma 6 ma 10 ph Time (min) Figure S9 S22