Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Supplementary Information for A Mechanistic Study of Proton Reduction by a Pentapyridine Cobalt Complex: Evidence for the Involvement of an Anation-Based Pathway Amanda E. King1, Yogesh Surendranath1, Nicholas A. Piro1,5, Julian P. Bigi1,5, Jeffrey R. Long1,4* and Christopher J. Chang1,2,3,5* Departments of 1Chemistry and 2Molecular and Cell Biology and the 3Howard Hughes Medical Institute, University of California, Berkeley, California 94720, and 4Materials Sciences Division and 5Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720 S1 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Materials The ligands PY5Me , CF PY5Me , and NMe PY5Me were synthesized as previously described, as were 2 3 2 2 2 [Co(PY5Me )(MeCN)](OTf) , [Co(CF PY5Me )(MeCN)](OTf) (5) and 2 2 3 2 2 [Co(NMe PY5Me )(MeCN)](OTf) (6) were also synthesized as previously described. All other 2 2 2 compounds purchased from chemical vendors were used as received. Acetonitrile, toluene, and tetrahydrofuran were dried over activated 4 Å molecular sieves, passed over a column of activated alumina, and stored over 3 Å molecular sieves under an N atmosphere. Unless otherwise noted, all 2 manipulations were performed at room temperature under an N atmosphere using either Schlenk or glove 2 box techniques. Physical Methods NMR spectra were recorded on Bruker spectrometers operating at 300 or 400 MHz as noted. Chemical shifts are reported in ppm relative to residual protiated solvent; coupling constants are reported in Hz. UV-visible spectra were recorded on a Hewlet-Packard 8452A photodiode array spectrometer. Spectroelectrochemical experiments were obtained using a reticulated vitreous carbon working electrode and platinum electrodes as the reference and auxiliary electrodes. Contact to the vitreous carbon electrode was achieved using a Pt wire. The auxiliary electrode was separated from the solution using a Vycor tip.An Agilent 490-GC micro-gas chromatograph with a mol sieve column and heated syringe injector was used for H quantification. Carbon, hydrogen, and nitrogen analyses were obtained from the 2 Microanalytical Laboratory of the University of California, Berkeley. Mass spectra were determined at the University of California, Berkeley Mass Spectrometry Facility. Cyclic voltammetry experiments were carried out using BASI’s Epsilon potentiostat. A glassy carbon working electrode and silver (reference) and platinum (auxiliary) wires were used for cyclic voltammetry experiments in CH CN with Bu NPF as 3 4 6 a supporting electrolyte. Ferrocene (E = 0.64 V vs SHE) was added during each experiment as an Fc+/0 internal reference. Single-crystal X-Ray diffraction was conducted at University of California, Berkeley, College of Chemistry, X-Ray Crystallography Facility. Crystals were mounted on nylon loops in Paratone-N hydrocarbon oil. Air-sensitive samples were transferred from the glove box to Paratone-N and mounted quickly to avoid decomposition. All data collections were performed on either a Bruker SMART, or Bruker APEX diffractometer equipped with a CCD area detector and a low temperature apparatus. Data integration was performed using SAINT. Preliminary data analysis and absorption correction were performed using XPREP and SADABS. Structure solution and refinement was performed using SHELX software package. All DFT calculations on 4 were carried out using the Gaussian09 program suite at the unrestricted B3LYP/6-31G(d) level of theory. Starting geometries were derived from the crystal structure of 4, where C–H bond distances were first normalized to 1.083 Å. The geometries were first run for ca. 200 cycles using default settings, and then the integration grid (Ultrafine) and maximum step size were each made smaller to finish the minimization on the apparently very flat potential energy surface. Final structures each contained no imaginary frequencies. Molecular orbitals were orthogonalized for maximal α/β overlap prior to visualization. Crystallographic Structure Determinations The X-ray crystallographic data collection was carried out on a Bruker three-circle diffractometer mounted with an SMART 1000 detector using monochromated Mo Kα radiation (0.71073 Å) outfitted with a low-temperature, nitrogen-stream aperture, an APEXII CCD detector, and equipped with an Oxford Cryostream 700. The structure was solved using direct methods in conjunction with standard difference Fourier techniques and refined by full-matrix least-squares procedures.4 A semi-empirical absorption correction (SADABS) was applied to the diffraction data for all structures. All non-hydrogen atoms were refined anisotropically, and hydrogen atoms were treated as idealized contributions and refined isotropically. A summary of crystallographic data is given in Table S1. All software used for diffraction data processing and crystal-structure solution and refinement are contained in the APEX2 program suite (Bruker AXS, Madison, WI). S2 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Syntheses Synthesis of [Co(PY5Me )(MeCN)](PF ) (1⋅(PF )). 1 equivalent of CoCl (29.3 mg, 225 μmol) was 2 6 2 6 2 added to 1 equivalent of PY5Me (100.0 mg, 225 μmol) in 10 mL of acetonitrile. After 20 minutes, 2 2 equivalents of AgPF (56.9 mg, 450 μmol) were added to the blue-green slurry. The addition of AgPF 6 6 resulted in the immediate precipitation of silver chloride and the formation of a light orange solution. The reaction was allowed to stir for an additional 2 hours before being filtered and concentrated to 1 mL. X- ray quality crystals were obtained by slow diffusion of diethyl ether into a concentrated acetonitrile solution of 3⋅(PF ) . Yield: 150.0 mg, 84% yield. Anal. Calcd. for C H CoF P N : C, 44.67%; H, 6 2 31 28 12 2 6 3.39%; N, 10.08%. Found: C, 44.20%; H, 3.50%; N, 9.91%. Synthesis of [Co(PY5Me )](BPh )(3⋅BPh ). 1 equivalent of Co(PPh ) Cl (197.8 mg, 225 μmol) was 2 4 4 3 3 added to 1 equivalent of PY5Me (100.0 mg, 225 μmol) in 10 mL of toluene; after stirring overnight the 2 reaction color had changed from lime-green to dark blue. The reaction was filtered, and the remaining blue solid was washed with 20 mL of toluene and 20 mL of diethyl ether to remove residual PPh and 3 dried under vacuum. The solid was suspended in THF, and 1 equivalent of NaBPh (77.0 mg, 225 μmol) 4 was added; the resulting dark blue solution was allowed to stir overnight. Removal of the resulting NaCl via filtration, removal of the bulk THF solvent, and subsequent recrystallization via slow diffusion of diethyl ether into a difluorobenzene solution produced X-ray quality crystals. Yield: 65.0 mg, 79% yield. Anal. Calcd. for C H CoBN OF : C, 75.07%; H, 5.70%; N, 6.95%. Found: C, 74.31%; H, 5.54%; N, 63 57 5 2 6.91%. Synthesis of [Co(PY5Me )(MeCN)](OTf) (BF ) (2⋅(OTf) (BF )). 1 equivalent of NOBF (13.9 μmol) 2 2 4 2 4 4 was added to 1 equivalent of 3⋅(OTf) (100.0 mg, 119 μmol) in 10 mL of acetonitrile. The light orange 2 solution intensified in color. The reaction was allowed to stir for an additional 2 hours before being filtered and concentrated to 1 mL. X-ray quality crystals were obtained by slow diffusion of diethyl ether into a concentrated acetonitrile solution of 3⋅(OTf) . Yield: 80.0 mg, 72% yield. Anal. Calcd. for 2 C H BCoF N O S : C, 41.96%; H, 2.84%; N, 7.89%. Found: C, 41.44%; H, 2.86%; N, 8.72%. 31 25 10 5 5 2 Synthesis of [Co(PY5Me )(OAc)](MeCN)(OTf) (4). 1 equivalent of NaBH(OAc) (22.8 mg, 108 μmol) 2 2 3 was added to 1 equivalent of 2⋅(OTf) (BF ) (100.0 mg, 108 μmol) in 10 mL of acetonitrile. The light 2 4 orange solution became pink in color after 1 hour. The reaction was allowed to stir for an additional 2 hours before being filtered and concentrated to 1 mL. X-ray quality crystals were obtained by slow diffusion of diethyl ether into a concentrated acetonitrile solution of 6. Yield: 70 mg, 79% yield. Anal. Calcd. for C H CoF N O S : C, 46.11%; H, 3.47%; N, 9.33%. Found: C, 46.67%; H, 2.90%; N, 9.41%. 35 31 6 6 7 2 Experimental Procedures Kinetic experiments with catalytic amounts of 1⋅(OTf) and acetic acid. A typical kinetic experiment was 2 conducted as follows. A flask was charged with 5 mL of a stock solution of Bu NPF and 3⋅(OTf) in 4 6 2 acetonitrile (100 mM and 1 mM, respectively) and subsequently purged with N . Cyclic voltammograms 2 were recorded at 50, 100, 250, and 500 mV/s. Glacial acetic acid was then added in known increments; cyclic voltammograms were recorded at 50, 100, 250, and 500 mV/s after each increment. Dependencies on both [AcOH] and [1] were determined. Kinetic experiments with stoichiometric amounts of 1⋅(OTf) and acetic acid. A typical kinetic 2 experiment was conducted as follows. A flask was charged with 5 mL of a stock solution of Bu NPF and 4 6 3⋅(OTf) in acetonitrile (500 mM and 10 mM, respectively) and subsequently purged with N . Cyclic 2 2 voltammograms were recorded at 50, 100, 250, and 500 mV/s. Glacial acetic acid was then added in known increments; cyclic voltammograms were recorded at 50, 100, 250, and 500 mV/s after each increment. Dependencies on both [AcOH] and [1] were determined. S3 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Titration of 1⋅(OTf) with tetrabutylammonium acetate. A flask was charged with 5 mL of a stock 2 solution of Bu NPF and 3⋅(OTf) in acetonitrile (500 mM and 10 mM, respectively) and subsequently 4 6 2 purged with N . Cyclic voltammograms were recorded at 50, 100, 250, and 500 mV/s. 2 Tetrabutylammonium acetate was added to give final concentrations of 5 and 10 mM; cyclic voltammograms were recorded at 50, 100, 250, and 500 mV/s after each addition. Kinetic experiments with catalytic amounts of 1⋅(OTf) and 1:1 tetrabutylammonium acetate:acetic acid 2 mixture. A typical kinetic experiment was conducted as follows. A flask was charged with 5 mL of a stock solution of Bu NPF and 3⋅(OTf) in acetonitrile (100 mM and 1 mM, respectively) and 4 6 2 subsequently purged with N . Cyclic voltammograms were recorded at 50, 100, 250, and 500 mV/s. A 1:1 2 tetrabutylammonium acetate:acetic acid solution in acetonitrile (500 mM concentration in each component) was then added in known increments; cyclic voltammograms were recorded at 50, 100, 250, and 500 mV/s after each increment. Dependencies on both [Bu NOAc:AcOH] and [1] were determined. 4 Faradaic efficiency determination with 1:1 tetrabutylammonium acetate:acetic acid mixture. Controlled-potential electrolysis was conducted using a custom-made air-tight glass double compartment cell separated by a glass frit. The working compartment was fitted with a glassy carbon rod working electrode (2.5 mm diameter, 2 cm length) and a Ag/AgNO reference electrode. The auxiliary 3 compartment was fitted with a Pt gauze electrode. The working compartment was filled with 10 mL of 35 mM acetic acid in a 0.1 M Bu NPF acetonitrile solution, while the auxiliary compartment was filled with 4 6 5 mL of 0.1 M Bu NPF acetonitrile solution, resulting in equal solution levels in both compartments. 4 6 Solution diffusion across the glass frit was slow under static pressure. Both compartments were sparged for 15 min with N and cyclic voltammograms were recorded as controls. Catalyst 1(CF SO ) (0.1 mM) 2 3 3 2 was then added and a cyclic voltammogram was recorded. Electrolysis was conducted for 2 h and the headspace was subjected to gas chromatographic analysis. An Agilent 490-GC Micro-Gas Chromatograph with a molecular sieve column and heated syringe injector was used for product detection. The column was heated to 80 °C under Ar gas flow and an average sample volume of 200 nL was injected onto the column. Calibration. Using the double-compartment cell described above, the working compartment was filled with 50 mL of 35 mM acetic acid in a 0.1 M Bu NPF acetonitrile solution, while the auxiliary 4 6 compartment was filled with 70 mL of a 0.1 M Bu NPF acetonitrile solution. Both compartments were 4 6 sparged thoroughly with N and sealed. A 3 mL aliquot of the headspace was removed and replaced with 2 3 mL of CH . Aliquots of 0.5, 1, 2, and 3 mL of H were introduced to the headspace and the solution was 4 2 allowed to stir for at least 30 min. A sample of the headspace was injected into the gas chromatograph and the ratio of CH and H was taken as points on a calibration curve 4 2 Titration of different acids into solutions of [Co(X-PY5Me )(MeCN)]2+. The follow procedure was 2 repeated for compounds 5 and 6. A flask was charged with 3 mL of a stock solution of Bu NPF and 4 6 1⋅(OTf) in acetonitrile (100 mM and 1 mM, respectively) and subsequently purged with N . Cyclic 2 2 voltammograms were recorded at 50, 100, 250, and 500 mV/s. 50 μL of a 600 mM acid solution was added to the flask (final [acid] = 10 mM), and cyclic voltammograms were recorded at 50, 100, 250, and 500 mV/s. The potential at which the observed current equaled 8 μA was plotted versus acid pKa. Spectroelectrochemical experiments with 1⋅(OTf) . 4 mL of a solution containing Bu NPF and 1⋅(OTf) 2 4 6 2 in acetonitrile (100 mM and 1 mM, respectively) was added to a 1-cm cuvette equipped with a reticulated vitreous carbon working electrode. A starting spectrum was recorded as a baseline. A potential of -900 mV vs SHE was applied, and the current and potential were monitored during electrolysis. S4 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 UV-visible spectroscopic experiments examining protonation of 4. 4 mL of a 168 mM solution of p- toluenesulfonic acid was added to a 1-cm pathlength UV-Vis cuvette equipped with a Schlenk valve. 50 μL of a 64 mM solution of 3⋅BPh was added (0.8 mM final concentration) was added via syringe, and 4 spectra recorded at known intervals. Dead time between mixing and the attainment of the first spectrum was approximately 30 seconds. S5 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Table S1. Crystallographic data for [Co(PY5Me )(MeCN)](PF ) (1⋅(PF ) ). 2 6 2 6 2 Empirical formula C H CoF N P 31 28 12 6 2 Formula weight 833.46 Temperature 100 K Wavelength 0.71073 Å Crystal system Triclinic Space group P-1 a = 11.4618(13) Å, α = 109.795(2)° Unit cell dimensions b = 12.5008(14) Å, β = 108.383(2)° c = 13.1640(15)Å, γ = 90.824(2)° Volume 1668.6 Å3 Z 2 Density (calculated) 1.659 Absorption coefficient 0.712 F(000) 842 Crystal size 0.14 x 0.14 x 0.12 mm3 Theta range for collection 1.75° to 25.38° Index ranges -12<=h<=13, -15<=k<=13, -15<=l<=15 Reflections collected 9548 Independent reflections 5362 Completeness to max 88.6 % Absorption correction Semi-empirical from equivalents Refinement method Full-matrix least squares on F2 Data/restrains/parameters 5362 / 0 / 472 Goodness-of-fita 1.020 Final R indicesb [I > 2d(I)] R = 0.0399, wR = 0.0975 1 2 R indicesb (all data) R = 0.0523, wR = 0.1069 1 2 Largest diff. peak and hole 0.476 and -0.431 e Å-3 aGooF =[[w(F2 −F2)2]/(n −p)]1/2 o c { [ ] [ ]} bR = F − F /F , wR = w(F2 −F2)2 / w(F2)2 1/2 1 o c o 2 c o o S6 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Table S2. Crystallographic data for [Co(PY5Me )(MeCN)](OTf) BF (2⋅(OTf) BF ). 2 2 4 2 4 Empirical formula C H B Co F N O S 144 130 4 4 40 30 24 8 Formula weight 3960.22 Temperature 100 K Wavelength 0.71073 Å Crystal system Monoclinic Space group Cc a = 17.791 Å, α = 90° Unit cell dimensions b = 16.341 Å, β = 97.47° c = 27.415 Å, γ = 90° Volume 7902.4 Å3 Z 2 Density (calculated) 1.664 Absorption coefficient 0.645 F(000) 4024 Crystal size 0.20 x 0.16 x 0.12 mm3 Theta range for collection 1.50° to 25.37° Index ranges -21<=h<=21, -19<=k<=19, -33<=l<=32 Reflections collected 55077 Independent reflections 13504 Completeness to max 99.7 % Absorption correction Semi-empirical from equivalents Refinement method Full-matrix least squares on F2 Data/restrains/parameters 13504 / 8 / 1153 Goodness-of-fita 1.044 Final R indicesb [I > 2d(I)] R = 0.0707, wR = 0.1913 1 2 R indicesb (all data) R = 0.0776, wR = 0.2010 1 2 Largest diff. peak and hole 1.575 and -1.025 e Å-3 aGooF =[[w(F2 −F2)2]/(n −p)]1/2 o c { [ ] [ ]} bR = F − F /F , wR = w(F2 −F2)2 / w(F2)2 1/2 1 o c o 2 c o o S7 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Table S3. Crystallographic data for [[Co(PY5Me )]BPh (3⋅BPh ). 2 4 4 Empirical formula C H BCoN 53 45 5 Formula weight 821.70 Temperature 100 K Wavelength 1.54178 Å Crystal system Triclinic Space group P-1 a = 10.9829(8) Å, α = 102.458(5)° Unit cell dimensions b = 13.7142(9) Å, β = 107.971(5)° c = 17.0064(13) Å, γ = 101.250(5)° Volume 2282.8(3) Å3 Z 4 Density (calculated) 1.195 Absorption coefficient 3.250 F(000) 842 Crystal size 0.10 x 0.06 x 0.02 mm3 Theta range for collection 2.86° to 63.68° Index ranges -12<=h<=12, -14<=k<=15, -19<=l<=19 Reflections collected 20067 Independent reflections 7242 Completeness to max 96.5 % Absorption correction Semi-empirical from equivalents Refinement method Full-matrix least squares on F2 Data/restrains/parameters 7242 / 0 / 542 Goodness-of-fita 0.997 Final R indicesb [I > 2d(I)] R = 0.0625, wR = 0.1524 1 2 R indicesb (all data) R = 0.0957, wR = 0.1685 1 2 Largest diff. peak and hole 0.408 and -0.402 e Å-3 aGooF =[[w(F2 −F2)2]/(n −p)]1/2 o c { [ ] [ ]} bR = F − F /F , wR = w(F2 −F2)2 / w(F2)2 1/2 1 o c o 2 c o o S8 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Table S4. Crystallographic data for [Co(PY5Me )(OAc)](OTf) (4⋅(OTf) ). 2 2 2 Empirical formula C H CoF N O S 33 25 6 5 8 2 Formula weight 856.63 Temperature 100 K Wavelength 0.71073 Å Crystal system Triclinic Space group P-1 a = 11.9400(3) Å, α = 65.6960(10)° Unit cell dimensions b = 12.8220(3) Å, β = 78.9150(10)° c = 15.3280(3) Å, γ = 66.5570(10)° Volume 1960.99(8) Å3 Z 2 Density (calculated) 1.273 Absorption coefficient 0.612 F(000) 764 Crystal size 0.16 x 0.10 x 0.07 mm3 Theta range for collection 1.46° to 25.39° Index ranges -14<=h<=14, -15<=k<=15, -18<=l<=18 Reflections collected 33375 Independent reflections 7167 Completeness to max 99.5 % Absorption correction Semi-empirical from equivalents Refinement method Full-matrix least squares on F2 Data/restrains/parameters 7167 / 0 / 499 Goodness-of-fita 1.076 Final R indicesb [I > 2d(I)] R = 0.0362, wR = 0.0922 1 2 R indicesb (all data) R = 0.0412, wR = 0.0946 1 2 Largest diff. peak and hole 0.628 and -0.644 e Å-3 aGooF =[[w(F2 −F2)2]/(n −p)]1/2 o c { [ ] [ ]} bR = F − F /F , wR = w(F2 −F2)2 / w(F2)2 1/2 1 o c o 2 c o o S9 Electronic Supplementary Material (ESI) for Chemical Science This journal is © The Royal Society of Chemistry 2013 Fig. S1: UV-visible spectrum of 2 in acetonitrile (top) and 1H-NMR spectrum in CD CN (bottom). 3 S10