Supporting Information High-Performance Oxygen Redox Catalysis ...
Supporting Information
High-Performance Oxygen Redox Catalysis with Multifunctional Cobalt Oxide Nanochains: Morphology Dependent Activity Prashanth W. Menezes,† Arindam Indra,† Diego González-Flores,‡ Nastaran Ranjbar Sahraie,§ Ivelina Zaharieva,‡ Michael Schwarze,† Peter Strasser,§* Holger Dau,‡* and Matthias Driess†* †
Department of Chemistry: Metalorganics and Inorganic Materials, Technische Universität Berlin, Straße des 17 Juni 135, Sekr. C2, 10623 Berlin, Germany ‡
Fachbereich Physik, Freie Universität Berlin, Arnimallee 14, 14195 Berlin, Germany
§
Department of Chemistry, The Electrochemical Energy, Catalysis, and Materials Science Group, Technische Universität Berlin, Straße des 17 Juni 124, Sekr. TC3, 10623 Berlin, Germany *Corresponding
Authors:
Email: [email protected] [email protected] [email protected]
Chemicals All chemical reagents (analytical grade) were used as received without any further purification. Deionized water was used throughout the experiment. Commercially available cobalt(II) acetate tetrahydrate (Co(CH3COO)2·4H2O), ammonium oxalate dihydrate ((NH4)2C2O4·2H2O), cobalt oxide (Co3O4), cobalt monoxide (CoO), ceric ammonium nitrate ((NH4)2[Ce(NO3)6]), sodium peroxodisulfate (Na2S2O8) and tris(bipyridine)ruthenium(II) chloride hexahydrate ([Ru(bpy)3]Cl2·6H2O) were obtained from Sigma Aldrich whereas cetyltrimethylammonium bromide (CTAB) was purchased from Alfa Aesar. Instrumental Phase identification of the samples were determined using PXRD on a Bruker AXS D8 advanced automatic diffractometer equipped with a position sensitive detector (PSD) and a curved germanium (111) primary monochromator. The radiation used was Cu-Kα (λ = 1.5418 Å). The XRD profiles recorded were in the range of 5° < 2θ < 80° and the diffraction pattern fitting were carried out using the program WinxPow. Similarly, the structural models were drawn with the DIAMOND program version 3.0. The chemical composition of the precursor was confirmed by ICP-AES on a Thermo Jarrell Ash Trace Scan analyzer. The samples were dissolved in acid solutions (aqua regia) and the results of three independent measurements were averaged which showed good agreement with the chemical formulae. The quantification of the precursor and oxide was also estimated by the elemental analyses that were performed on a Flash EA 112 Thermo Finnigan elemental analyzer. SEM was used to evaluate size and morphology and EDX analyses were used to semi-quantitatively determine the cobalt present on the sample surfaces. The samples were placed on a silicon wafer and
1
the measurements were carried on a LEO DSM 982 microscope integrated with EDX (EDAX, Appollo XPP). Data handling and analysis were carried out with the software package EDAX. The microstructure of the samples was investigated by TEM analysis. A small amount of the sample powder was placed on a TEM-grid (carbon film on 300 mesh Cu-grid, Plano GmbH, Wetzlar, Germany). The microstructure (morphology, particle size, phase composition, crystallinity) of the samples was studied by a FEI Tecnai G2 20 S-TWIN transmission electron microscope (FEI Company, Eindhoven, Netherlands) equipped with a LaB6-source at 200 kV acceleration voltage. EDX analysis were carried out with an EDAX r-TEM SUTW Detector (Si (Li)-detector). Images were recorded with a GATAN MS794 P CCD-camera. Both SEM and TEM experiments were carried out at the Zentrum für Elektronenmikroskopie (ZELMI) of the TU Berlin. The surface area and the pore size distributions were carried out on a Quantachrome Autosorb-1 apparatus. Nitrogen adsorption/desorption isotherms were determined at -196 ˚C after degassing the sample at 150 ˚C overnight. The BET surface areas (SBET) were estimated by adsorption data in a relative pressure range from 0.01 to 0.1 and the pore size distribution (PSD) was calculated by analyzing the adsorption data of the N2 isotherm using the Barret-Joyner-Halenda (BJH) method. Simultaneous constant rate TGA analysis of the well ground oxalate precursor was performed on a Rubotherm set up. The samples dried at 80 ˚C were placed in an open alumina crucible and heated at 5 ˚C/min to 600 ˚C in a continuous nitrogen gas flow and cooled down to the room temperature. The TG curves were corrected by subtraction of a blank run and the solid product obtained was examined by PXRD. The presence of different modes of vibrations of the precursor and the metal oxides were studied using a BIORAD FTS 6000 FTIR spectrometer under attenuated total reflection (ATR) conditions. The data were recorded in the range of 400–4000 cm-1 with the average of thirty two scans at 4 cm-1 resolution. Laser-Raman spectra were collected on a Renishaw Ramanscope in the backscattering configuration using 514 nm (2.41 eV) wavelength laser under ambient condition. TPR experiments were performed by Thermo scientific TPD/R/O 1110 instrument under an argon flow of 20 ml min-1 as a carrier gas and a 5 vol.% H 2/Ar mixture (20 ml min-1) as the probe gas. A small amount of catalyst (15 mg) was mounted in a quartz reactor tube assembling in a furnace whose temperature was continuously recorded by thermocouple inserted in the catalyst reactor. The furnace temperature increased using a constant heat rate of 10 °C min-1 in the temperature range from room temperature to final temperature of 800 °C and held at this temperature for 15 minutes. Subsequently the reactor was cooled to room temperature using nitrogen flow. The XPS measurements were performed using a Kratos Axis Ultra X-ray photoelectron spectrometer (Karatos Analytical Ltd., Manchester, UK) using an Al Kα monochromatic radiation source (1486.7 eV) with 90° takeoff angle (normal to analyzer). The vacuum pressure in the analyzing chamber was maintained at 2 x 10-9 Torr. The high-resolution XPS spectra were collected for C1s, O1s and Co2p levels with pass energy 20 eV and step 0.1 eV. The binding energies were calibrated relative to C1s peak energy position as 285.0 eV. Data analyses were performed using Casa XPS (Casa Software Ltd.) and Vision data processing program (Kratos Analytical Ltd.). The evolved oxygen gas in photochemical water oxidation experiments were quantified by a gas chromatograph (GC). An Agilent 7890A gas chromatograph was used to determine the oxygen content in a headspace. The GC was equipped with a carboxen-1000 column and a thermal conductivity detector (TCD). The carrier gas was argon with a flow rate of 30 mL/min. For analysis of H2O18 labelled experiments, a GC-MS system equipped with a tungsten filament from Agilent (5975C GC/MSD) was used. The voltage of the secondary electron multiplier (SEM) was 960 V.
2
Figure S1. PXRD and Miller indices (hkl) of as-prepared cobalt oxalate precursor (CoC2O4∙2H2O, JCPDS 25-250) via inverse micelle approach.
Figure S2. PXRD pattern and Miller indices (hkl) of as-obtained cobalt oxide nanochains (Co3O4, JCPDS 42-1467) from cobalt oxalate precursor.
3
Figure S3. The crystal structure of (a) the cobalt oxalate dihydrate precursor, which comprises one dimensional chains with each cobalt atom coordinated by two bidentate oxalate ligands and two water molecules,1 and (b) cobalt oxide, which crystallizes in the normal spinel structure with Co2+ atoms in tetrahedral and Co3+ atoms in the octahedral sites. The unit cell is shown on the bottom left and the central structural motif is on the bottom right. 2
Figure S4. Nanorods of the as-prepared cobalt oxalate precursor are shown in SEM (left) and TEM (right) micrographs. The nanorods had a mean diameter of around 400 nm with a length of several hundred nanometers.
4
Table S1. Quantification of cobalt from cobalt oxalate precursor obtained by ICP-AES. Three independent measurements were performed for the reliability of the experiments. Cobalt (mass %) Calculated
32.21
Measurement 1
32.17
Measurement 2
32.27
Measurement 3
32.22
Averaged
32.22
Figure S5. The presence of cobalt in the cobalt oxalate precursor (top) and the cobalt oxide (bottom) were determined by the EDX. Appearance of peaks for carbon and copper are due to TEM grid (carbon film on 300 mesh Cu-grid).
5
Table S2. Results from the elemental analysis (C, H, N) of as-prepared cobalt oxalate dihydrate that is in accordance with the calculated chemical formula. Three independent measurements were performed to ensure the reliability of the experiments.
Calculated Measurement 1 Measurement 2 Measurement 3 Averaged
Carbon (mass%) 13.13 13.28 13.24 13.18 13.23
Hydrogen (mass%) 2.20 2.26 2.23 2.23 2.24
Nitrogen (mass%) 0 0 0 0 0
Figure S6. TGA (blue solid line) and its differential (DTG, red dotted line) plot of the CoC 2O4∙2H2O precursor thermally treated from room temperature to 700 ˚C in nitrogen atmosphere at the rate of 5 ˚C/min. Two distinct mass loss steps were observed. The first mass loss occurs between 130 to 180 ˚C, which corresponds to the release of the structural water upon formation of the anhydrous CoC 2O4. This mass loss corresponds to a broad DTG peak at 149 ˚C. The experimental mass loss (19.4%) obtained is consistent with the calculated weight (19.6%) of two water molecules. The second mass loss occurred at 200 to 360 ˚C with a distinct DTG peak at 351 ˚C, transforming the anhydrous oxalate phase into the oxide phase. The mass loss within this step was 47.6 %, which is very close to the calculated value (48.1%) for two molecules of carbon dioxide.3-5
6
Figure S7. PXRD pattern and Miller indices (hkl) of the decomposition product of cobalt oxalate precursor subjected to thermogravimetric analysis (TGA). The phase was identified as the mixture of metallic Co (red, JCPDS 15-806) and CoO (black, JCPDS 74-2392).
Figure S8. FT-IR transmission spectrum of as-prepared cobalt oxalate nanorods.3,4
7
Table S3. IR absorption maxima (cm-1) of CoC2O4∙2H2O corresponding to Figure S8 that match well with the maxima previously reported for the metal oxalate precursor.3,4 IR maxima / cm-1 Assignments IR maxima / cm-1 3365 γ(OH)(H2O) 825 1621 γas(C–O) 741 1362 γs(C–O) 495 1316 δ(OCO) as: asymetric, s:symmetric, γ:stretching, δ: bending, ρ:scissoring
Assignments γs(C–C) + δ(OCO) ρ(H2O) δ ring
Figure S9. FT-IR transmission spectrum of cobalt oxide in the region 400-1200 cm-1 showing symmetric Co―O stretching vibrations. The shown spectrum is in accordance with the previously reported spectra of cobalt oxides.4
8
Figure S10. Raman spectra of the cobalt oxalate precursor in the region 400-1600 cm-1 (left) and the cobalt oxide in the region of 300-1000 cm-1 (right). For cobalt oxalate dihydrate, the band identified at 1472 cm-1 could be assigned to the ν(C-O) stretching mode whereas at 914 cm-1 can be assigned to ν(C-C) stretching. The band appearing at 584 cm-1 of low intensity corresponds to water liberation modes and the band at 525 cm-1 belongs to the symmetric OCO bending.6,7 Four Raman bands at 661 cm-1, 598 cm-1, 504 cm-1 and 460 cm-1 could be assigned for the literature reported Co3O4 phase.8,9
Figure S11. The H2-temperature programmed reduction (H2-TPR) profiles of Co3O4 nanochains, Co3O4/C and C. Both Co3O4 and Co3O4/C show two step reduction process between 300 ˚C to 600 ˚C. For Co3O4, the first reduction peak appears around 320 ˚C which corresponds to the reduction of Co 3O4 to CoO. The second broad peaks at higher temperature originate from the reduction of CoO to Co. This reduction step is strongly dependent on the morphology of the Co 3O4 materials and often broad signals have been observed depending on the particle sizes and morphology. Co 3O4/C shows a clear two step reduction due to its well separated and distributed particles over carbon support. However, the carbon support C does not undergo any reduction.10
9
Figure S12. Typical high-resolution XPS spectra of the regions containing the Co2p, O1s, and C1s of CoC2O4∙2H2O precursor. The Co(II) and Co(III) have almost similar 2p binding energies but they can be differentiated by the Co2p1/2−2p3/2 spin−orbit level energy spacing. The binding energies of cobalt in this case were found to be 797.3 eV for Co2p1/2 and 781.3 for Co2p3/2 and are consistent with Co(II).11 The O1s binding energy peaks can be deconvoluted into two different carbon-oxygen functional groups. The peak at 533.9 eV corresponds to the binding energy of C–O bond or water ligands and the peak at 532.2 eV could be assigned to the C=O bond. The deconvoluted C 1s spectrum comprises of three binding energy peaks at 284.9 eV, 286.4 eV and 288.8 eV. The peaks at 284.9 eV and 286.4 eV are attributed to the carbon in C–C and C–O environments. The peak at 288.8 eV could be assigned to the C=O from the oxalate anion.12,13
10
Figure S13. SEM image of the cross-section of the thin film of Co3O4 (3 mg on FTO) is shown on the left. The thickness of the film was determined to be around 800 nm. The higher resolution SEM image of the nanochains deposited on the surface is shown on the right. The photograph of the film taken after the deposition is shown in the inset.
Figure S14. Chronopotentiometric measurements used to determine the overpotential for a current density of 0.5 mAcm-2 in 0.1 M KOH (pH 13) and in 0.1 M phosphate buffer (pH 7) at room temperature.
11
Figure S15. SEM images of the surface of the thin film of Co3O4 after the chronopotentiometric measurements in 0.1 M KOH solution (pH 13). The surface view is shown on the left and the higher resolution SEM image of the nanochains deposited on the surface is shown on the right. No significant change in surface morphology was observed. The photograph of the film after the electrochemical measurements is shown in the inset.
Figure S16. SEM images of the surface of the thin film of Co3O4 after the chronopotentiometric measurements in 0.1 M phosphate buffer solution (pH 7). The surface view is shown on the left and the higher resolution SEM image of the nanochains deposited on the surface is shown on the right. No significant change in surface morphology was observed. The photograph of the film after the electrochemical measurements is shown in the inset.
12
.
Figure S17. High-resolution XPS spectra of the nanochains of Co3O4 thin film after the chronopotentiometric measurements in 0.1 M phosphate buffer (pH 7) solution. Two peaks located at binding energies of about 780.2 eV for Co2p3/2 and 795.3 eV for Co2p1/2 are consistent with the presence of Co2+ and Co3+ and agrees well with the literature reported examples.11,14 However, a significant increase in the Co3+ character from the deconvolution area shows that large amount of surface Co2+ were oxidized to Co3+. The O1s spectrum is deconvoluted into three peaks. The peaks at 530 eV could be assigned to the oxygen atoms of Co3O4 whereas a large dominance of –OH species absorbed on the surface were observed at the binding energy of 531.8 eV. A small peak at around 535.3 eV was also attained and is correlated to the absorbed water molecules.15 The Co2p and O1s spectra obtained here are consistent with the one presented for 0.1 M KOH solution confirming similar potential-induced effects in both pH 7 and pH 13 solutions (see Figure S18).
13
Figure S18. TEM (a, c, e) and HRTEM (b, d, f) images of solvothermally prepared Co3O4, commercial Co3O4 and CoO. Electron diffraction pattern is shown in the inset.
14
Figure S19. PXRD pattern and Miller indices (hkl) of solvothermally prepared and commercial cobalt oxide (Co3O4, JCPDS 42-1467) and cobalt monoxide (CoO, JCPDS 34-1004).
15
Scheme S1. Photocatalytic cycle of water oxidation with Na2S2O8 and [Ru(bpy)3]2+ system. In the Ru(bpy)3]2+-S2O82-system, the [Ru(bpy)3]2+ absorbs visible light and generates electron-hole pairs on the surface of the catalyst. The electrons produced were expelled by the sacrificial electron acceptor S2O82by further oxidizing the [Ru(bpy)3]2+ to [Ru(bpy)3]3+ and reducing S2O82- to SO42- and a sulphate radical (SO4-∙). Thus the formed radical can subsequently further oxidize [Ru(bpy)3]2+ to yield [Ru(bpy)3]3+. Hence the [Ru(bpy)3]3+ molecule donates its holes to the catalyst and reverts back to the [Ru(bpy) 3]2+ where two water molecules are oxidized to form one oxygen molecule. The dissolved O2 content was analyzed by a Clark oxygen electrode system.16
Figure S20. Dissolved oxygen concentrations of Co3O4 catalyst measured by a Clark electrode in buffers of borate (pH 9), phosphate (pH 7), carbonate (pH 4.7) and acetate (pH 5.8) using Na2S2O8 as a two electron acceptor and Ru(bpy)32+ as a photosensitizer (300 W Xe lamp with 395 cutoff filter).
16
Figure S21. Dissolved oxygen concentration of Co3O4 catalyst measured by a Clark electrode in borate buffer (pH = 9) using Na2S2O8 as a two electron acceptor and Ru(bpy)32+ as a photosensitizer. The borate buffer itself shows the oxygen evolution of 90 µ mol/L without addition of the catalyst (but with electron acceptor and photosensitizer) likely explainable by photosensitizer degradation and formation of catalytic RuO2 due to the high oxidation potential of [Ru(bpy)3]3+ at pH 9.
Figure S22. The surface-area normalized plot of photochemical water oxidation with nanochains of Co3O4, solvothermal Co3O4, commercial Co3O4 and CoO.
17
Table S4. Values of GC detection of oxygen gas in the head space of the reactor containing Co3O4 (see SI, Experimental section for details). The gas was collected after the photochemical experiments irradiated by xenon lamp for two hours. No hydrogen was detected.
Vol% O2
Gas volume (mL)
O2 (mL/h)
O2 (µmol/h)
64.02
0.661
15
0.049
2.00
49.70
0.724
15
0.054
2.19
0.692
15
0.051
2.09
Catalyst
Vol. % O2 (with air)
Area O2
Area N2
Co3O4
0.94
65.07
0.94
65.27
averaged
Figure S23. GC-MS detection of oxygen evolved in the headspace of reactions containing Co3O4 catalyst in phosphate buffer (pH = 7) using Na2S2O8 as a two electron acceptor and Ru(bpy)32+ as a photosensitizer. The resulting ratios (area) of m/z 28 (N14N14), m/z 36 (O18O18), m/z 34 (O16O18), and m/z 32 (O16O16) are shown before and after irradiation with a xenon lamp (395 nm cutoff filter) for two hours.
18
Figure S24. A representative PXRD pattern of cobalt oxide (Co3O4, JCPDS 42-1467) after photochemical water oxidation.
Figure S25. HR-TEM image of Co3O4 after the chemical water oxidation experiments with CAN. The morphology (Figure top, left) and the crystallinity (Figure top, right) of the nanochains are preserved after the experiment. The inset shows the Fast Fourier Transformed (FFT) pattern of the nanochains. EDX confirms the presence of only cobalt and no cerium was detected. Appearance of peaks for carbon and copper are due to TEM grid (carbon film on 300 mesh Cu-grid).
19
Figure S26. HR-TEM image of Co3O4 after the photochemical water oxidation experiments in S2O82–/ Ru(bpy)32+ system in phosphate buffer. The morphology (Figure top, left) and the crystallinity (Figure top, right) of the nanochains are preserved after the photochemical experiment. The inset shows the FFT pattern of the nanochains. EDX confirms the presence of only cobalt without any additional elements. Appearance of peaks for carbon and copper are due to TEM grid (carbon film on 300 mesh Cu-grid).
20
Figure S27. High-resolution XPS spectra of the nanochains of Co3O4 after the chemical water oxidation measurements in CAN solution. Two peaks located at binding energies of about 780.5 eV for Co2p3/2 and 795.5 eV for Co2p1/2 are consistent with the presence of Co2+ and Co3+ and agrees well with the standard literature reported results.11,14 However, a significant increase in the Co3+ character from the deconvolution area showed that large amount of surface Co2+ were oxidized to Co3+. The deconvoluted O 1s spectrum shows three peaks related to oxygen atoms of Co3O4 (530.04 eV), –OH species absorbed on the surface (532.5 eV) and chemisorbed oxygen or adsorbed water molecules (534.2 eV).15 The increase in the –OH peak in O 1s spectrum could possibly due to the changes in the surface groups as well as of the bulk material.
Figure S28. High-resolution XPS spectra of the nanochains of Co3O4 after the photochemical water oxidation measurements in S2O82––Ru(bpy)32+ system in phosphate buffer. Both, the deconvoluted Co2p and O1s spectra show similar effects as observed in the case of chemical water oxidation. The dominant increase in the –OH and H2O peak in O1s not only suggests the changes in the surface group but also possible contribution from the bulk material after two hours of photochemical experiments.
21
Figure S29. TEM (left) and HRTEM (right) images of Co3O4 spherical particles synthesized without micelle. The electron diffraction pattern is shown in the inset.
Figure S30. Dissolved oxygen concentrations measured by a Clark electrode in deoxygenated aqueous solutions containing cobalt oxide catalysts (synthesized with and without micelle route) and 0.5 M ceric ammonium nitrate (CAN) as an oxidant (catalyst amount is 1 mg).
22
Figure S31. Dissolved oxygen concentration of cobalt oxides (0.5 mg) catalysts (synthesized with and without micelle route) measured by a Clark electrode in S2O82-–Ru(bpy)32+ system using phosphate (pH 7) buffer (300 W Xe lamp with 395 cutoff filter).
Figure S32. Chronoamperometric curves (percentage of retained current as a function of operation time) of Co3O4/C and a benchmark Pt/C electrodes measured at 0.8 V versus RHE in oxygen saturated 0.1 M KOH electrolyte for ORR.
23
Figure S33. Comparison of the LSVs Co3O4/C after the chronoamperometry experiments with Pt/C and Co3O4/C in oxygen saturated 0.1 M KOH with the scan rate of 10 mV/s (at 1500 rpm).
Koutecky–Levich equation:
ik = Kinetic current density ω = Angular velocity of the disk (ω = 2πN, N is the rotation speed) n = Overall transferred electron number per molecular oxygen in the ORR F = Faraday constant A = Geometric electrode area k = Rate constant of the reaction C0 = Saturated O2 concentration in the electrolyte DO2 = Diffusion coefficient of O2 ν = Kinematic viscosity of the electrolyte
24
References (1) Deyrieux, R.; Berro, C.; Peneloux, A. Bull. Soc. Chim. Fr. 1973, 25-34. (2) Roth, W. L. J. Phys. Chem. Solids 1964, 25, 1-10. (3) Wang, D.; Wang, Q.; Wang, T. Inorg. Chem. 2011, 50, 6482-6492. (4) Salavati-Niasari, M.; Mir, N.; Davar, F. J. Phys. Chem. Solids 2009, 70, 847-852. (5) Liu, Z.; Liu, Z.; Li, Q.; Yang, T.; Zhang, D. Mater. Chem. Phys. 2011, 131, 102-107. (6) Frost, R. L.; Yang, J.; Ding, Z. Chin. Sci. Bull. 2003, 48, 1844-1852. (7) Mancilla, N.; Caliva, V.; D'Antonio, M. C.; Gonzalez-Baro, A. C.; Baran, E. J. J. Raman Spectroscopy 2009, 40, 915-920. (8) Rashad, M.; Ruesing, M.; Berth, G.; Lischka, K.; Pawlis, A. J. Nanomater. 2013, 724853. (9) Hadjiev, V. G.; Iliev, M. N.; Vergilov, I. V. J. Phys. C 1988, 21, L199-L201. (10) Xie, X.; Shen, W. Nanoscale 2009, 1, 50-60. (11) Oku, M.; Hirokawa, K. J. Electron Spectros. Relat. Phenom. 1976, 8, 475-481. (12) Yuan, X.; Zhang, M.; Chen, X.; An, N.; Liu, G.; Liu, Y.; Zhang, W.; Yan, W.; Jia, M. Appl.Catal. A 2012, 439, 149-155. (13) Shinde, V.; Mandale, A. B.; Patil, K. R.; Gaikwad, A. B.; Patil, P. P. Surf. Coat. Technology 2006, 200, 5094-5101. (14) Oku, M.; Sato, Y. Appl. Surf. Sci. 1992, 55, 37-41. (15) Alrehaily, L. M.; Joseph, J. M.; Biesinger, M. C.; Guzonas, D. A.; Wren, J. C. Phys. Chem. Chem. Phy. 2013, 15, 1014-1024. (16) Harriman, A.; Pickering, I. J.; Thomas, J. M.; Christensen, P. A. J. Chem. Soc. Farad. Trans. 1988, 84, 2795-2806.
25
High-Performance Oxygen Redox Catalysis with Multifunctional Cobalt Oxide Nanochains: Morphology Dependent Activity Prashanth W. Menezes,† Arindam Indra,† Diego González-Flores,‡ Nastaran Ranjbar Sahraie,§ Ivelina Zaharieva,‡ Michael Schwarze,† Peter Strasser,§* Holger Dau,‡* and Matthias Driess†* †
Department of Chemistry: Metalorganics and Inorganic Materials, Technische Universität Berlin, Straße des 17 Juni 135, Sekr. C2, 10623 Berlin, Germany ‡
Fachbereich Physik, Freie Universität Berlin, Arnimallee 14, 14195 Berlin, Germany
§
Department of Chemistry, The Electrochemical Energy, Catalysis, and Materials Science Group, Technische Universität Berlin, Straße des 17 Juni 124, Sekr. TC3, 10623 Berlin, Germany *Corresponding
Authors:
Email: [email protected] [email protected] [email protected]
Chemicals All chemical reagents (analytical grade) were used as received without any further purification. Deionized water was used throughout the experiment. Commercially available cobalt(II) acetate tetrahydrate (Co(CH3COO)2·4H2O), ammonium oxalate dihydrate ((NH4)2C2O4·2H2O), cobalt oxide (Co3O4), cobalt monoxide (CoO), ceric ammonium nitrate ((NH4)2[Ce(NO3)6]), sodium peroxodisulfate (Na2S2O8) and tris(bipyridine)ruthenium(II) chloride hexahydrate ([Ru(bpy)3]Cl2·6H2O) were obtained from Sigma Aldrich whereas cetyltrimethylammonium bromide (CTAB) was purchased from Alfa Aesar. Instrumental Phase identification of the samples were determined using PXRD on a Bruker AXS D8 advanced automatic diffractometer equipped with a position sensitive detector (PSD) and a curved germanium (111) primary monochromator. The radiation used was Cu-Kα (λ = 1.5418 Å). The XRD profiles recorded were in the range of 5° < 2θ < 80° and the diffraction pattern fitting were carried out using the program WinxPow. Similarly, the structural models were drawn with the DIAMOND program version 3.0. The chemical composition of the precursor was confirmed by ICP-AES on a Thermo Jarrell Ash Trace Scan analyzer. The samples were dissolved in acid solutions (aqua regia) and the results of three independent measurements were averaged which showed good agreement with the chemical formulae. The quantification of the precursor and oxide was also estimated by the elemental analyses that were performed on a Flash EA 112 Thermo Finnigan elemental analyzer. SEM was used to evaluate size and morphology and EDX analyses were used to semi-quantitatively determine the cobalt present on the sample surfaces. The samples were placed on a silicon wafer and
1
the measurements were carried on a LEO DSM 982 microscope integrated with EDX (EDAX, Appollo XPP). Data handling and analysis were carried out with the software package EDAX. The microstructure of the samples was investigated by TEM analysis. A small amount of the sample powder was placed on a TEM-grid (carbon film on 300 mesh Cu-grid, Plano GmbH, Wetzlar, Germany). The microstructure (morphology, particle size, phase composition, crystallinity) of the samples was studied by a FEI Tecnai G2 20 S-TWIN transmission electron microscope (FEI Company, Eindhoven, Netherlands) equipped with a LaB6-source at 200 kV acceleration voltage. EDX analysis were carried out with an EDAX r-TEM SUTW Detector (Si (Li)-detector). Images were recorded with a GATAN MS794 P CCD-camera. Both SEM and TEM experiments were carried out at the Zentrum für Elektronenmikroskopie (ZELMI) of the TU Berlin. The surface area and the pore size distributions were carried out on a Quantachrome Autosorb-1 apparatus. Nitrogen adsorption/desorption isotherms were determined at -196 ˚C after degassing the sample at 150 ˚C overnight. The BET surface areas (SBET) were estimated by adsorption data in a relative pressure range from 0.01 to 0.1 and the pore size distribution (PSD) was calculated by analyzing the adsorption data of the N2 isotherm using the Barret-Joyner-Halenda (BJH) method. Simultaneous constant rate TGA analysis of the well ground oxalate precursor was performed on a Rubotherm set up. The samples dried at 80 ˚C were placed in an open alumina crucible and heated at 5 ˚C/min to 600 ˚C in a continuous nitrogen gas flow and cooled down to the room temperature. The TG curves were corrected by subtraction of a blank run and the solid product obtained was examined by PXRD. The presence of different modes of vibrations of the precursor and the metal oxides were studied using a BIORAD FTS 6000 FTIR spectrometer under attenuated total reflection (ATR) conditions. The data were recorded in the range of 400–4000 cm-1 with the average of thirty two scans at 4 cm-1 resolution. Laser-Raman spectra were collected on a Renishaw Ramanscope in the backscattering configuration using 514 nm (2.41 eV) wavelength laser under ambient condition. TPR experiments were performed by Thermo scientific TPD/R/O 1110 instrument under an argon flow of 20 ml min-1 as a carrier gas and a 5 vol.% H 2/Ar mixture (20 ml min-1) as the probe gas. A small amount of catalyst (15 mg) was mounted in a quartz reactor tube assembling in a furnace whose temperature was continuously recorded by thermocouple inserted in the catalyst reactor. The furnace temperature increased using a constant heat rate of 10 °C min-1 in the temperature range from room temperature to final temperature of 800 °C and held at this temperature for 15 minutes. Subsequently the reactor was cooled to room temperature using nitrogen flow. The XPS measurements were performed using a Kratos Axis Ultra X-ray photoelectron spectrometer (Karatos Analytical Ltd., Manchester, UK) using an Al Kα monochromatic radiation source (1486.7 eV) with 90° takeoff angle (normal to analyzer). The vacuum pressure in the analyzing chamber was maintained at 2 x 10-9 Torr. The high-resolution XPS spectra were collected for C1s, O1s and Co2p levels with pass energy 20 eV and step 0.1 eV. The binding energies were calibrated relative to C1s peak energy position as 285.0 eV. Data analyses were performed using Casa XPS (Casa Software Ltd.) and Vision data processing program (Kratos Analytical Ltd.). The evolved oxygen gas in photochemical water oxidation experiments were quantified by a gas chromatograph (GC). An Agilent 7890A gas chromatograph was used to determine the oxygen content in a headspace. The GC was equipped with a carboxen-1000 column and a thermal conductivity detector (TCD). The carrier gas was argon with a flow rate of 30 mL/min. For analysis of H2O18 labelled experiments, a GC-MS system equipped with a tungsten filament from Agilent (5975C GC/MSD) was used. The voltage of the secondary electron multiplier (SEM) was 960 V.
2
Figure S1. PXRD and Miller indices (hkl) of as-prepared cobalt oxalate precursor (CoC2O4∙2H2O, JCPDS 25-250) via inverse micelle approach.
Figure S2. PXRD pattern and Miller indices (hkl) of as-obtained cobalt oxide nanochains (Co3O4, JCPDS 42-1467) from cobalt oxalate precursor.
3
Figure S3. The crystal structure of (a) the cobalt oxalate dihydrate precursor, which comprises one dimensional chains with each cobalt atom coordinated by two bidentate oxalate ligands and two water molecules,1 and (b) cobalt oxide, which crystallizes in the normal spinel structure with Co2+ atoms in tetrahedral and Co3+ atoms in the octahedral sites. The unit cell is shown on the bottom left and the central structural motif is on the bottom right. 2
Figure S4. Nanorods of the as-prepared cobalt oxalate precursor are shown in SEM (left) and TEM (right) micrographs. The nanorods had a mean diameter of around 400 nm with a length of several hundred nanometers.
4
Table S1. Quantification of cobalt from cobalt oxalate precursor obtained by ICP-AES. Three independent measurements were performed for the reliability of the experiments. Cobalt (mass %) Calculated
32.21
Measurement 1
32.17
Measurement 2
32.27
Measurement 3
32.22
Averaged
32.22
Figure S5. The presence of cobalt in the cobalt oxalate precursor (top) and the cobalt oxide (bottom) were determined by the EDX. Appearance of peaks for carbon and copper are due to TEM grid (carbon film on 300 mesh Cu-grid).
5
Table S2. Results from the elemental analysis (C, H, N) of as-prepared cobalt oxalate dihydrate that is in accordance with the calculated chemical formula. Three independent measurements were performed to ensure the reliability of the experiments.
Calculated Measurement 1 Measurement 2 Measurement 3 Averaged
Carbon (mass%) 13.13 13.28 13.24 13.18 13.23
Hydrogen (mass%) 2.20 2.26 2.23 2.23 2.24
Nitrogen (mass%) 0 0 0 0 0
Figure S6. TGA (blue solid line) and its differential (DTG, red dotted line) plot of the CoC 2O4∙2H2O precursor thermally treated from room temperature to 700 ˚C in nitrogen atmosphere at the rate of 5 ˚C/min. Two distinct mass loss steps were observed. The first mass loss occurs between 130 to 180 ˚C, which corresponds to the release of the structural water upon formation of the anhydrous CoC 2O4. This mass loss corresponds to a broad DTG peak at 149 ˚C. The experimental mass loss (19.4%) obtained is consistent with the calculated weight (19.6%) of two water molecules. The second mass loss occurred at 200 to 360 ˚C with a distinct DTG peak at 351 ˚C, transforming the anhydrous oxalate phase into the oxide phase. The mass loss within this step was 47.6 %, which is very close to the calculated value (48.1%) for two molecules of carbon dioxide.3-5
6
Figure S7. PXRD pattern and Miller indices (hkl) of the decomposition product of cobalt oxalate precursor subjected to thermogravimetric analysis (TGA). The phase was identified as the mixture of metallic Co (red, JCPDS 15-806) and CoO (black, JCPDS 74-2392).
Figure S8. FT-IR transmission spectrum of as-prepared cobalt oxalate nanorods.3,4
7
Table S3. IR absorption maxima (cm-1) of CoC2O4∙2H2O corresponding to Figure S8 that match well with the maxima previously reported for the metal oxalate precursor.3,4 IR maxima / cm-1 Assignments IR maxima / cm-1 3365 γ(OH)(H2O) 825 1621 γas(C–O) 741 1362 γs(C–O) 495 1316 δ(OCO) as: asymetric, s:symmetric, γ:stretching, δ: bending, ρ:scissoring
Assignments γs(C–C) + δ(OCO) ρ(H2O) δ ring
Figure S9. FT-IR transmission spectrum of cobalt oxide in the region 400-1200 cm-1 showing symmetric Co―O stretching vibrations. The shown spectrum is in accordance with the previously reported spectra of cobalt oxides.4
8
Figure S10. Raman spectra of the cobalt oxalate precursor in the region 400-1600 cm-1 (left) and the cobalt oxide in the region of 300-1000 cm-1 (right). For cobalt oxalate dihydrate, the band identified at 1472 cm-1 could be assigned to the ν(C-O) stretching mode whereas at 914 cm-1 can be assigned to ν(C-C) stretching. The band appearing at 584 cm-1 of low intensity corresponds to water liberation modes and the band at 525 cm-1 belongs to the symmetric OCO bending.6,7 Four Raman bands at 661 cm-1, 598 cm-1, 504 cm-1 and 460 cm-1 could be assigned for the literature reported Co3O4 phase.8,9
Figure S11. The H2-temperature programmed reduction (H2-TPR) profiles of Co3O4 nanochains, Co3O4/C and C. Both Co3O4 and Co3O4/C show two step reduction process between 300 ˚C to 600 ˚C. For Co3O4, the first reduction peak appears around 320 ˚C which corresponds to the reduction of Co 3O4 to CoO. The second broad peaks at higher temperature originate from the reduction of CoO to Co. This reduction step is strongly dependent on the morphology of the Co 3O4 materials and often broad signals have been observed depending on the particle sizes and morphology. Co 3O4/C shows a clear two step reduction due to its well separated and distributed particles over carbon support. However, the carbon support C does not undergo any reduction.10
9
Figure S12. Typical high-resolution XPS spectra of the regions containing the Co2p, O1s, and C1s of CoC2O4∙2H2O precursor. The Co(II) and Co(III) have almost similar 2p binding energies but they can be differentiated by the Co2p1/2−2p3/2 spin−orbit level energy spacing. The binding energies of cobalt in this case were found to be 797.3 eV for Co2p1/2 and 781.3 for Co2p3/2 and are consistent with Co(II).11 The O1s binding energy peaks can be deconvoluted into two different carbon-oxygen functional groups. The peak at 533.9 eV corresponds to the binding energy of C–O bond or water ligands and the peak at 532.2 eV could be assigned to the C=O bond. The deconvoluted C 1s spectrum comprises of three binding energy peaks at 284.9 eV, 286.4 eV and 288.8 eV. The peaks at 284.9 eV and 286.4 eV are attributed to the carbon in C–C and C–O environments. The peak at 288.8 eV could be assigned to the C=O from the oxalate anion.12,13
10
Figure S13. SEM image of the cross-section of the thin film of Co3O4 (3 mg on FTO) is shown on the left. The thickness of the film was determined to be around 800 nm. The higher resolution SEM image of the nanochains deposited on the surface is shown on the right. The photograph of the film taken after the deposition is shown in the inset.
Figure S14. Chronopotentiometric measurements used to determine the overpotential for a current density of 0.5 mAcm-2 in 0.1 M KOH (pH 13) and in 0.1 M phosphate buffer (pH 7) at room temperature.
11
Figure S15. SEM images of the surface of the thin film of Co3O4 after the chronopotentiometric measurements in 0.1 M KOH solution (pH 13). The surface view is shown on the left and the higher resolution SEM image of the nanochains deposited on the surface is shown on the right. No significant change in surface morphology was observed. The photograph of the film after the electrochemical measurements is shown in the inset.
Figure S16. SEM images of the surface of the thin film of Co3O4 after the chronopotentiometric measurements in 0.1 M phosphate buffer solution (pH 7). The surface view is shown on the left and the higher resolution SEM image of the nanochains deposited on the surface is shown on the right. No significant change in surface morphology was observed. The photograph of the film after the electrochemical measurements is shown in the inset.
12
.
Figure S17. High-resolution XPS spectra of the nanochains of Co3O4 thin film after the chronopotentiometric measurements in 0.1 M phosphate buffer (pH 7) solution. Two peaks located at binding energies of about 780.2 eV for Co2p3/2 and 795.3 eV for Co2p1/2 are consistent with the presence of Co2+ and Co3+ and agrees well with the literature reported examples.11,14 However, a significant increase in the Co3+ character from the deconvolution area shows that large amount of surface Co2+ were oxidized to Co3+. The O1s spectrum is deconvoluted into three peaks. The peaks at 530 eV could be assigned to the oxygen atoms of Co3O4 whereas a large dominance of –OH species absorbed on the surface were observed at the binding energy of 531.8 eV. A small peak at around 535.3 eV was also attained and is correlated to the absorbed water molecules.15 The Co2p and O1s spectra obtained here are consistent with the one presented for 0.1 M KOH solution confirming similar potential-induced effects in both pH 7 and pH 13 solutions (see Figure S18).
13
Figure S18. TEM (a, c, e) and HRTEM (b, d, f) images of solvothermally prepared Co3O4, commercial Co3O4 and CoO. Electron diffraction pattern is shown in the inset.
14
Figure S19. PXRD pattern and Miller indices (hkl) of solvothermally prepared and commercial cobalt oxide (Co3O4, JCPDS 42-1467) and cobalt monoxide (CoO, JCPDS 34-1004).
15
Scheme S1. Photocatalytic cycle of water oxidation with Na2S2O8 and [Ru(bpy)3]2+ system. In the Ru(bpy)3]2+-S2O82-system, the [Ru(bpy)3]2+ absorbs visible light and generates electron-hole pairs on the surface of the catalyst. The electrons produced were expelled by the sacrificial electron acceptor S2O82by further oxidizing the [Ru(bpy)3]2+ to [Ru(bpy)3]3+ and reducing S2O82- to SO42- and a sulphate radical (SO4-∙). Thus the formed radical can subsequently further oxidize [Ru(bpy)3]2+ to yield [Ru(bpy)3]3+. Hence the [Ru(bpy)3]3+ molecule donates its holes to the catalyst and reverts back to the [Ru(bpy) 3]2+ where two water molecules are oxidized to form one oxygen molecule. The dissolved O2 content was analyzed by a Clark oxygen electrode system.16
Figure S20. Dissolved oxygen concentrations of Co3O4 catalyst measured by a Clark electrode in buffers of borate (pH 9), phosphate (pH 7), carbonate (pH 4.7) and acetate (pH 5.8) using Na2S2O8 as a two electron acceptor and Ru(bpy)32+ as a photosensitizer (300 W Xe lamp with 395 cutoff filter).
16
Figure S21. Dissolved oxygen concentration of Co3O4 catalyst measured by a Clark electrode in borate buffer (pH = 9) using Na2S2O8 as a two electron acceptor and Ru(bpy)32+ as a photosensitizer. The borate buffer itself shows the oxygen evolution of 90 µ mol/L without addition of the catalyst (but with electron acceptor and photosensitizer) likely explainable by photosensitizer degradation and formation of catalytic RuO2 due to the high oxidation potential of [Ru(bpy)3]3+ at pH 9.
Figure S22. The surface-area normalized plot of photochemical water oxidation with nanochains of Co3O4, solvothermal Co3O4, commercial Co3O4 and CoO.
17
Table S4. Values of GC detection of oxygen gas in the head space of the reactor containing Co3O4 (see SI, Experimental section for details). The gas was collected after the photochemical experiments irradiated by xenon lamp for two hours. No hydrogen was detected.
Vol% O2
Gas volume (mL)
O2 (mL/h)
O2 (µmol/h)
64.02
0.661
15
0.049
2.00
49.70
0.724
15
0.054
2.19
0.692
15
0.051
2.09
Catalyst
Vol. % O2 (with air)
Area O2
Area N2
Co3O4
0.94
65.07
0.94
65.27
averaged
Figure S23. GC-MS detection of oxygen evolved in the headspace of reactions containing Co3O4 catalyst in phosphate buffer (pH = 7) using Na2S2O8 as a two electron acceptor and Ru(bpy)32+ as a photosensitizer. The resulting ratios (area) of m/z 28 (N14N14), m/z 36 (O18O18), m/z 34 (O16O18), and m/z 32 (O16O16) are shown before and after irradiation with a xenon lamp (395 nm cutoff filter) for two hours.
18
Figure S24. A representative PXRD pattern of cobalt oxide (Co3O4, JCPDS 42-1467) after photochemical water oxidation.
Figure S25. HR-TEM image of Co3O4 after the chemical water oxidation experiments with CAN. The morphology (Figure top, left) and the crystallinity (Figure top, right) of the nanochains are preserved after the experiment. The inset shows the Fast Fourier Transformed (FFT) pattern of the nanochains. EDX confirms the presence of only cobalt and no cerium was detected. Appearance of peaks for carbon and copper are due to TEM grid (carbon film on 300 mesh Cu-grid).
19
Figure S26. HR-TEM image of Co3O4 after the photochemical water oxidation experiments in S2O82–/ Ru(bpy)32+ system in phosphate buffer. The morphology (Figure top, left) and the crystallinity (Figure top, right) of the nanochains are preserved after the photochemical experiment. The inset shows the FFT pattern of the nanochains. EDX confirms the presence of only cobalt without any additional elements. Appearance of peaks for carbon and copper are due to TEM grid (carbon film on 300 mesh Cu-grid).
20
Figure S27. High-resolution XPS spectra of the nanochains of Co3O4 after the chemical water oxidation measurements in CAN solution. Two peaks located at binding energies of about 780.5 eV for Co2p3/2 and 795.5 eV for Co2p1/2 are consistent with the presence of Co2+ and Co3+ and agrees well with the standard literature reported results.11,14 However, a significant increase in the Co3+ character from the deconvolution area showed that large amount of surface Co2+ were oxidized to Co3+. The deconvoluted O 1s spectrum shows three peaks related to oxygen atoms of Co3O4 (530.04 eV), –OH species absorbed on the surface (532.5 eV) and chemisorbed oxygen or adsorbed water molecules (534.2 eV).15 The increase in the –OH peak in O 1s spectrum could possibly due to the changes in the surface groups as well as of the bulk material.
Figure S28. High-resolution XPS spectra of the nanochains of Co3O4 after the photochemical water oxidation measurements in S2O82––Ru(bpy)32+ system in phosphate buffer. Both, the deconvoluted Co2p and O1s spectra show similar effects as observed in the case of chemical water oxidation. The dominant increase in the –OH and H2O peak in O1s not only suggests the changes in the surface group but also possible contribution from the bulk material after two hours of photochemical experiments.
21
Figure S29. TEM (left) and HRTEM (right) images of Co3O4 spherical particles synthesized without micelle. The electron diffraction pattern is shown in the inset.
Figure S30. Dissolved oxygen concentrations measured by a Clark electrode in deoxygenated aqueous solutions containing cobalt oxide catalysts (synthesized with and without micelle route) and 0.5 M ceric ammonium nitrate (CAN) as an oxidant (catalyst amount is 1 mg).
22
Figure S31. Dissolved oxygen concentration of cobalt oxides (0.5 mg) catalysts (synthesized with and without micelle route) measured by a Clark electrode in S2O82-–Ru(bpy)32+ system using phosphate (pH 7) buffer (300 W Xe lamp with 395 cutoff filter).
Figure S32. Chronoamperometric curves (percentage of retained current as a function of operation time) of Co3O4/C and a benchmark Pt/C electrodes measured at 0.8 V versus RHE in oxygen saturated 0.1 M KOH electrolyte for ORR.
23
Figure S33. Comparison of the LSVs Co3O4/C after the chronoamperometry experiments with Pt/C and Co3O4/C in oxygen saturated 0.1 M KOH with the scan rate of 10 mV/s (at 1500 rpm).
Koutecky–Levich equation:
ik = Kinetic current density ω = Angular velocity of the disk (ω = 2πN, N is the rotation speed) n = Overall transferred electron number per molecular oxygen in the ORR F = Faraday constant A = Geometric electrode area k = Rate constant of the reaction C0 = Saturated O2 concentration in the electrolyte DO2 = Diffusion coefficient of O2 ν = Kinematic viscosity of the electrolyte
24
References (1) Deyrieux, R.; Berro, C.; Peneloux, A. Bull. Soc. Chim. Fr. 1973, 25-34. (2) Roth, W. L. J. Phys. Chem. Solids 1964, 25, 1-10. (3) Wang, D.; Wang, Q.; Wang, T. Inorg. Chem. 2011, 50, 6482-6492. (4) Salavati-Niasari, M.; Mir, N.; Davar, F. J. Phys. Chem. Solids 2009, 70, 847-852. (5) Liu, Z.; Liu, Z.; Li, Q.; Yang, T.; Zhang, D. Mater. Chem. Phys. 2011, 131, 102-107. (6) Frost, R. L.; Yang, J.; Ding, Z. Chin. Sci. Bull. 2003, 48, 1844-1852. (7) Mancilla, N.; Caliva, V.; D'Antonio, M. C.; Gonzalez-Baro, A. C.; Baran, E. J. J. Raman Spectroscopy 2009, 40, 915-920. (8) Rashad, M.; Ruesing, M.; Berth, G.; Lischka, K.; Pawlis, A. J. Nanomater. 2013, 724853. (9) Hadjiev, V. G.; Iliev, M. N.; Vergilov, I. V. J. Phys. C 1988, 21, L199-L201. (10) Xie, X.; Shen, W. Nanoscale 2009, 1, 50-60. (11) Oku, M.; Hirokawa, K. J. Electron Spectros. Relat. Phenom. 1976, 8, 475-481. (12) Yuan, X.; Zhang, M.; Chen, X.; An, N.; Liu, G.; Liu, Y.; Zhang, W.; Yan, W.; Jia, M. Appl.Catal. A 2012, 439, 149-155. (13) Shinde, V.; Mandale, A. B.; Patil, K. R.; Gaikwad, A. B.; Patil, P. P. Surf. Coat. Technology 2006, 200, 5094-5101. (14) Oku, M.; Sato, Y. Appl. Surf. Sci. 1992, 55, 37-41. (15) Alrehaily, L. M.; Joseph, J. M.; Biesinger, M. C.; Guzonas, D. A.; Wren, J. C. Phys. Chem. Chem. Phy. 2013, 15, 1014-1024. (16) Harriman, A.; Pickering, I. J.; Thomas, J. M.; Christensen, P. A. J. Chem. Soc. Farad. Trans. 1988, 84, 2795-2806.
25
