Supporting Information for:

Transcription

1 Supporting Information for: Paired Electrolysis in the Simultaneous Production of Synthetic Intermediates and Substrates Mark J Llorente, a Bichlien H Nguyen, b Clifford P Kubiak,*,a Kevin D Moeller b a Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Drive, Mail Code 0358, La Jolla, California , United States b Department of Chemistry, Washington University in St. Louis, St. Louis, Missouri 63130, United StatesElectrochemical experiments investigating both the paired electrolyses and the individual half-reactions were performed using a Gamry Reference 600 Potentiostat which can record a secondary input voltage during standard experiments. This signal input was used to record the voltage of the auxiliary electrode vs a reference. A ~30 ml total volume glass H-Cell with fine glass frit and shared head space was used for all paired electrolysis studies as shown in Figure 1. The cells were sealed with septa with thin wire leads and taped. Cells were tested to ensure they were gas tight by filling cells with forming gas and measuring H 2 concentration over time. Both compartments used their own 5 mm diameter glassy carbon rod electrode purchased from Tokai Carbon, Ltd. and Ag/AgCl electrodes which were made by anodizing of % pure silver wire in 3 M HCl solutions. Each reference electrode was separated from solution via glass and vycor tip. Ferrocene was added to solution after experiments were completed to confirm the voltages of reference electrodes. Resistance measurements between the glassy carbon electrodes and their respective reference electrodes ranged between 20 to 40 Ω between experiments which corresponds to upwards of 120 mv of resistive voltage at 3 ma. Post-experiment voltage S1

2 compensation on the as-measured voltage auxiliary electrode was performed for slow scan rate ( 20 mv/s) and steady state experiments. Figure S1. Electrochemical H-cell sealed with septa. Each compartment contains its own Ag/AgCl reference electrode separated from solution by a glass tube terminated by a vycor tip and its own carbon rod electrode. Both compartments contain a mixture of 8:3:1 acetonitrile:tetrahydrofuran:methanol and 0.8 M Et 4 NBF 4. As shown, the left compartment contains 5 mm ceric ammonium nitrate, 40 mm o-phenylenediamine, and 44 mm syringaldehyde. The right compartment contains 5 mm Re(4,4 -di-tert-butyl-2,2 - bipyridine)(co) 3 Cl. Anode and cathode electrodes were placed 1.5 cm into solution giving 2.6 cm 2 of total working area. Chemicals were used as purchased unless otherwise noted. As described in the paper, an optimal mixed solvent system of 8:3:1 CH 3 CN:THF:MeOH and 0.8 M Et 4 NBF 4 was used to ensure solvation of substrates and products. Methanol served two primary functions, solvating the condensation product and maintaining low but constant proton concentrations required for the CO 2 reduction reaction.(1) An unintentional consequence of using methanol together with the small cation tetraethylammonium electrolyte is the combined facilitation of Cl - loss from the Re(4,4 -di-tert-butyl-2,2 -bipyridine)(co) 3 Cl (Re(bipy-tBu)(CO) 3 Cl) electrocatalyst necessary for catalyst activation and stabilization of the active oxidation state of the catalyst as shown by the early onset of catalytic current in Figures 3 and 4.(2, 3) The cathode S2

3 Current (ma) Current (ma) compartment contained 5 mm of the Re(bipy-tBu)(CO) 3 Cl which was synthesized as previously published.(4) The anode compartment contained 5 mm of the electrocatalytic mediator ceric ammonium nitrate (CAN), 40 mm o-phenylenediamine, and 44 mm syringaldehyde unless otherwise noted. Bone dry N 2 or bone dry CO 2 was sparged into stirred solution from both compartments for five minutes. Solution was magnetically stirred using Teflon bars during steady state experiments including chronopotentiometry, chronoamperometry, and bulk electrolyses N 2 CO N 2 CO Voltage (V vs Ag/AgCl) Voltage (V vs. Ag/AgCl) Figure S2. (Left) Cyclic voltammograms of 5 mm Re(bipy-tBu)(CO) 3 Cl catalyst in 8:3:1 CH 3 CN:THF:MeOH 0.8 M Et 4 NBF 4 solution using a Tokai glassy carbon rod (2.55 cm 2 working area) after sparging with N 2 and CO 2. Scan rate is 100 mv/s. Due to high series resistances when using a separator, currents exceeding ±5 ma exceeded standard potentiostat voltage limits. (Right) Cyclic voltammogram data for the same conditions using an analytical BASi 3 mm diameter glassy carbon working electrode in a custom Teflon cell. Ag/AgCl wire was used as the reference and Platinum wire was used as the counter electrode. S3

4 Current (ma) N 2 CO 2 CO 2 + 2mL MeOH CO 2 + 2mL MeOH + 6mL THF Voltage (V vs Ag/AgCl) Figure S3. A series of cyclic voltammograms with a scan rate of 100 mv/s showing the effects of sequentially adding the different components of the solvent mixture to an initial concentration of 0.12 M Et 4 NBF 4 and 7.5 mm Re(bipy-tBu)(CO) 3 Cl catalyst in 16 ml CH 3 CN. In black is the voltammogram corresponding to an N 2 sparged solution. In red is a CO 2 sparged solution with trace water acting as the weak Brønsted acid for the CO 2 reduction. In blue, 2 ml of methanol is added shifting the onset of catalytic current to less negative potentials. In green, 6 ml of THF is added with the additional effect of lowering the onset of faradaic current. Peak current is lowered to 75% of the previous CV due to dilution of catalyst. S4

5 Current (ma) M MeOH in 1.2M TEABF M MeOH in 1.1M TEABF Voltage (V vs Ag/AgCl) Figure S4. depicts the cyclic voltammagrams of Re(bipy-tBu)(CO) 3 Cl at a scan rate of 100 mv/s in N 2 sparged solutions with and without added methanol. The initial increase in current and lack of a reverse oxidation peak near -0.2 V is attributed to methanol facilitating Cl - loss from the catalyst. S5

6 A custom Teflon cell was used in preliminary cyclic voltammetry (CV) experiments to characterize the mechanism by which the condensation half-reaction takes place. In the absence of the CAN mediator, a cloudy violet-black mixture of oxidized phenolic oligomers and stable organic radicals are produced. CV experiments were performed using a BASi glassy carbon electrode 3 mm diameter working electrode. An Ag/AgCl wire was used as the reference electrode and was separated from solution by a glass compartment and Vycor tip. Platinum wire was used as the counter electrode. Solvent was an 8:3:1 CH 3 CN:THF:MeOH mixture with 0.8 M Et 4 NBF 4. Concentrations of mediator and substrates were varied as follows CAN (mm) o-phenylenediamine (mm) Syingaldehyde (mm) Table S1. shows the varying concentrations of catalyst and substrates for the anodic condensation reaction as performed in 0.8 M Et 4 NBF 4 and 8:3:1 CH 3 CN:THF:MeOH. Cyclic voltammograms at each of these concentrations were performed at various scan rates. S6

7 Current (ma) 6 Saturated CAN Saturated CAN x20 Saturated CAN + 20mM Substrates Voltage (V vs Ag/AgCl) Figure S5. depicts cyclic voltammograms for the electrochemical behavior of the 5mM cerium ammonium nitrate electrocatalyst in 8:3:1 CH 3 CN:THF:MeOH 0.8 M TEABF 4 solution using a Tokai glassy carbon rod (2.55 cm 2 working area) after sparging the H-cell with N 2. A redox feature corresponding to the Ce(III/IV) couple can be seen at 1.0 V vs Ag/AgCl. Scan rate is 100 mv/s. Currents exceeding 5 ma in the electrochemical H-cell exceeded the potentiostat s voltage limit and could not be acquired. Current behavior in red corresponds to 20 mm o- phenylendiamine and 20 mm syringaldehyde which matches well with faradaic current behavior for the condensation reaction shown in Figure 5b. S7

8 Current (ma) Current (ma) mM CAN 10mM Syr=O, 10mM o-phen 10mM Syr=O, 20mM o-phen 10mM Syr=O, 40mM o-phen mM Can 10 Syr=O, 40 o-phen 20 Syr=O, 40 o-phen 40 Syr=O, 40 o-phen Voltage (V vs Ag/AgCl) Voltage (V vs Ag/AgCl) Figure S6. depicts increasing substrate concentrations in cyclic voltammograms of 5mM CAN in 8:3:1 CH 3 CN:THF:MeOH and 0.8 M TEABF 4 using a BASi glassy carbon working electrode 3 mm in diameter after sparging with N 2. Scan rate is 100 mv/s. Figure 6a (Left) shows current behavior with increasing o-phenylenediamine. Currents at higher potentials begin to decrease with increased o-phenylenediamine, possibly indicating the formation of current hindering intermediates such as stable aromatic radicals or oligomers. Lower potential current begins to increase with increased concentration indicating the potentials for the initial oxidation of o- phenylenediamine is less positive than the potential required for activation of the catalyst without substrate available. Figure 6b (Right) shows current behavior with increasing syringaldehyde. At 40 mm, current dramatically increases and current-voltage behavior begins to match well with that shown in Figure 5. In order to verify that catalysis was being performed in both compartments and at the chosen operating conditions, gas was sampled from the shared headspace and gaseous products were quantified via gas chromatography on a Hewlett-Packard 7890A series gas chromatograph equipped with two molsieve columns (30 m 0.53 mm i.d. 25 μm film). A 1 ml injection was split between the two columns, one with N 2 carrier gas and one with He carrier gas, in order to quantify both CO and H 2 simultaneously. No H 2 was detected. Gas chromatography calibration curves were made by sampling known volumes of CO and H 2 gas. After bulk electrolyses, condensation products were first extracted in DCM from a ph 7 buffer then refluxed. Product was dissolved in deuterated methanol in order to take H-NMR spectra. Figure 7 and 8 show quantification and characterization of the two products. S8

9 TCD Signal min FE=~100% 64min FE=~100% 120min FE=~90% Time (minutes) Figure S7. shows gas chromatographs corresponding to samples from different times during a two hour bulk electrolysis with constant current of 2.75 ma as performed in an H-cell using glassy carbon rods. Cathode compartment contained 5mM Re(bipy-tBu)(CO) 3 Cl. Anode compartment contained 5 mm CAN, 40 mm o-phenylenediamine, and 44 mm syringaldehyde. Both are sparged with CO 2 until saturation. S9

10 [M+H] + OPSAC Agilent 6230 ESI-TOFMS Positive Ion Mode Figure S8. (Top) shows an H-NMR spectrum of the extracted product from the anode compartment using the same conditions as Figure S7. (Bottom) shows the high resolution mass spectrum of the isolated product post H-NMR with the majority peak corresponding to the protonated form of the benzimidazole product. 4-(1H-benzo[d]imidazol-2-yl)-2,6-dimethoxyphenol: The anodic chamber of a H-cell with a fine glass frit was charged with o-phenylenediamine (40 mg, 0.37mmol), syringealdehyde (74 S10

11 mg, 0.40 mmol), ceric ammonium nitrate (41 mg, 0.07 mmol), and 0.1 M LiClO 4 (210 mg) in 5:1 THF:MeOH (12 ml). The cathodic chamber was charged with 1 mm Re(bpy-tBu)(CO) 3 Cl (12 mg) and 0.1 M LiClO 4 (210 mg). A RVC anode and carbon cathode were inserted into the flask, and 2.5 F/mole of charge was passed at 4 ma of current. The solution was diluted with buffered solution (ph=7), extracted with EtOAc, and washed with brine to give a 65% yield by NMR. 1 HNMR (300 MHz, CD 3 OD): 7.57 (dd, 5.9, 2.9 Hz, 2H), 7.45 (s, 2H), 7.23 (dd, 5.9, 2.9 Hz, 2H), 3.97 (s, 6H); 13 CNMR (75 MHz, CD 3 OD): 153.9, 149.7, 140.1, 139.1, 123.7, 121.3, 115.5, 105.3, 56.9 IR (neat, KBr) 3310, 2957, 2929, 1607, 1503 cm -1 ESI HRMS m/z (M+H) + : cald , obsd Photovoltaic devices were created from individual cells that were rated to perform with a minimum short circuit current of 10 ma and minimum open circuit potential of 0.5 V with dimensions of 0.2cm x 2.2cm under standard Air Mass 1.5 solar irradiance, though individual cells showed variance in short circuit current of 12±2 ma and open circuit voltage of 0.50 ±0.05 mv. When placed in series, all photovoltaic cells pass the same photocurrent but their voltages add. According to manufacturer specifications, when eleven are placed in series, the new device would give a short circuit current of 10mA and a larger open circuit voltage of 5.5V. The as built device performed at a calculated 11% efficiency when irradiated by the ~98 mw/cm 2 total solar flux reported for the day and hour of measurement in La Jolla, CA. PV efficiency is calculated by measuring power output at the maximum power point and dividing it by the impinging solar flux power. When compared to the current-voltage requirements of the electrolysis cell, a smaller photocurrent is desired so as to minimize energy loss due to poor current and voltage matching, so a new device made from smaller cells was constructed. Figure 9 shows the current-voltage behavior for both of these devices compared to the characteristic current-voltage behavior of the electrolysis cell. Figure 10 shows images of the assembled photovoltaic devices. S11

12 Current (ma) Photocurrent (Full Cells) Photocurrent (Chipped Cells) Electrolysis Current Voltage (V) Figure S9. Current and voltage behavior of the created photovoltaics under ~98 mw/cm 2 illumination. The device made from the chipped cells which has an optimal power point of 4.28 V at 3.12 ma is nearly power-matched to the electrolysis cell. The operating point when coupling the two lies at the intersection of the orange and blue curves, which under these conditions would be 4.4 V at 3.0 ma. S12

13 Figure S10. (Left) An image of the soldered individual Solar Made Supercell 10 ma 0.5 V photovoltaic cells. (Right) An image of the ¼ size chipped and soldered cells. Paired electrolyses driven by the second photovoltaic device were performed under direct solar flux. Non-ideal weather conditions lead to inconsistent photocurrents and made it difficult to quantify yields vs total charge passed. Figure 11 shows the current-voltage behavior as measured using an ammeter and voltmeter for the photovoltaic-driven electrolysis during an overcast day which had the least solar flux variance. Even under lower current conditions, the characteristic increase in voltage due to depletion of substrate at the anode arises. S13

14 Voltage and Current vs Time Voltage (V) Current (ma) Figure S11. records the current and voltage demand as a function of time for the electrolysis device as powered by the second photovoltaic. Both the lower operating current and voltage as compared to Figure S9. and signal fluctuations are due to a mild but persistent cloud layer on the day of this experiment. Note that around 33 minutes, an increase in voltage demand occurs with little increase in current, which we attribute primarily to the increased anodic voltage demand for the condensation reaction as described in the main paper. While the CAN mediator was sufficient for selectively forming the condensation product, a catalyst with higher turnover rates and that performs at a lower overpotentials would be preferable in future pairings to improve product yield and power demand. 1. Smieja J, Benson E, Kumar B, Grice K, Seu C, Miller A et al. Kinetic and structural studies, origins of selectivity, and interfacial charge transfer in the artificial photosynthesis of CO. Proceedings of the National Academy of Sciences. 2012;109(39): Grice K, Kubiak C. Recent Studies of Rhenium and Manganese Bipyridine Carbonyl Catalysts for the Electrochemical Reduction of CO2. CO2 Chemistry. 2014;: Saeki T, Hashimoto K, Kimura N, Omata K, Fujishima A. Electrochemical reduction of CO2 with high current density in a CO2 + methanol medium II. CO formation promoted by tetrabutylammonium cation. Journal of Electroanalytical Chemistry. 1995;390(1-2): Smieja J, Kubiak C. Re(bipy-tBu)(CO) 3 Cl improved Catalytic Activity for Reduction of Carbon Dioxide: IR-Spectroelectrochemical and Mechanistic Studies. Inorganic Chemistry. 2010;49(20): S14