Toward Highly Stable Electrocatalysts via Nanoparticle Pore
Toward Highly Stable Electrocatalysts via Nanoparticle Pore Confinement Carolina Galeanoa‡ ,Josef C. Meierb,c‡, Volker Peinecked, Hans Bongarda, Ioannis Katsounarosb, Angel A. Topalovb,c, Anhui Lue, Karl J.J. Mayrhoferb*, Ferdi Schütha* a)Department of Heterogeneous Catalysis, Max‐Planck‐Institut für Kohlenforschung,Kaiser‐Wilhelm‐ Platz 1, 45470 Mülheim an der Ruhr, Germany b)Department of Interface Chemistry and Surface Engineering, Max‐Planck‐Institut für Eisenforschung GmbH, Max‐Planck‐Strasse 1, 40237 Düsseldorf, Germany c)Center for Electrochemical Sciences, Ruhr‐Universität Bochum, Universitätsstrasse 150, 44780 Bochum, Germany d)The fuel cell research center ZBT GmbH, Carl‐Benz‐Straße 201, 47057 Duisburg, Germany e)State Key Laboratory of Fine Chemicals, Dalian University of Technology, Linggong Road No.2, Dalian 116024, P.R. China
Supporting information Material preparation Synthesis of solid core‐mesoporous shell silica exotemplate (SCMS)1 A typical synthesis of 10 g of SCMS silica spheres is as follows. 32.9 mL of aqueous ammonia (28%) is mixed with 500 g of ethanol and 120 mL of deionized water. After stirring for ca. 10 min, 23.6 mL of tetraethoxysilane (TEOS, 98%) are added, and the reaction mixture is stirred for ca. 1 h. Afterwards, a mixture solution containing 14.1 mL of TEOS and 5.7 mL of octadecyltrimethoxysilane (ODTMS, 90% tech.) is drop‐wise added (for ca. 20 min) to the colloidal solution containing the silica spheres and further reacted for 5 h without stirring. The resulting SCMS spheres are separated from S1
the solution by centrifugation, dried at 75°C overnight and further calcined at 550°C under air atmosphere to produce the final uniform spherical SCMS particles. Comparison materials The comparison material Pt/Vulcan (20 wt.%) is prepared by colloidal deposition of pre‐synthesized Pt nanoparticles on commercial Vulcan support. The synthesis and characterization details of this material can be found in our previous publication.2
Material characterization High resolution transmission electron microscopy (HR‐TEM) images were obtained on a HF‐2000 microscope equipped with a cold field emitter (CFE) and operated at a maximum acceleration voltage of 200 kV. Typically, the samples were placed on a Lacey carbon film supported by a copper grid. Solid samples were deposited on the Lacey carbon film without previous dissolution. High resolution scanning electron microscopy (HR‐SEM) and scanning transmission electron microscopy (STEM) micrographs were collected on a Hitachi S‐5500 ultra‐ high resolution cold field emission scanning electron microscope. The instrument was operated at a maximum acceleration voltage of 30 kV. The samples were prepared on Lacey carbon films supported on a 400 mesh copper grid. The use of Duo‐STEM Bright Field/Dark Field detector together with the secondary electron (SE) detector geometry allows simultaneous imaging of surface morphologies in scan mode, and dark field/bright field imaging in transmission mode. The same HR‐SEM/STEM microscope was used for the identical location SEM/STEM experiments. To obtain the cross‐ sectional cuttings, the Pt@HGS material was embedded in Spurr‐resin (a low‐viscosity epoxy resin embedding medium for electron microscopy) and then subjected to the cutting procedure in an ultramicrotome (Reichert Ultracut) equipped with a diamond knife. The resulting slices of the composite present a thickness of ca. 30‐50 nm. Nitrogen sorption measurements were carried out on a Micrometrics ASAP 2010 instrument. Prior to analysis, the silica exotemplate was activated under vacuum for at S2
least 8 h and the HGS for at least 15 h at 250°C. The measurements were performed at ‐ 196°C using a static‐volumetric method. The empty volume was determined with nitrogen. The BET surface area was calculated from the adsorption data in the relative pressure interval from 0.04 to 0.2. The pore size distribution was estimated by the BJH (Barrett‐Joyner‐Halenda) method from the adsorption branch (desorption data, which are normally recommended by IUPAC for BJH analysis, may be influenced by the tensile strength effect, see Figure SI‐3). The total pore volume was estimated from the amount of nitrogen adsorbed at a relative pressure of 0.97. In‐situ X‐ray powder diffraction (XRD) data were collected in reflection geometry on a Bragg Brentano diffractometer (X’Pert PRO, PANalytical) equipped with an Anton Paar XRK900 high‐temperature reaction chamber and a CuKα1,2 radiation source (40 kV, 40 mA) with the following slit configuration: primary and secondary soller slits 0.04 rad, divergence slit 0.5°, anti‐scatter slits 1°. Instead of a monochromator, a secondary Ni filter was inserted before a position sensitive real time multi strip detector (X'Celerator, 2.12° 2 active length). The reaction chamber is equipped with a Marcor® sample holder (6‐10 mm diameter). The sample is prepared on a sieve plate (10 mm diameter, 1 mm depth) which allows the protective gas to flow through the sample and leave the chamber through an exhaust pipe. Measurements were taken under 100% nitrogen flow. The samples were heated with a heating rate of 5 °C min‐1. After reaching the appropriate temperature, the waiting time before starting data collection was set to 30 min. Data were collected in the range between 20 and 90° 2. The sample was kept for 3 h at 850°C before starting the measurement. Room temperature XRD patterns were collected with a Bragg Brentano diffractometer (STOE THETA/THETA). The instrument is equipped with a secondary graphite monochromator (CuKα1,2 radiation) and a proportional gas detector. The divergence slit was set to 0.8°, the receiving slit was set to 0.8 mm, and the width of the horizontal mask was 4 mm. The samples were prepared on a background free single crystal quartz sample holder. Thermal stability of HGS was investigated by TG‐DTA using a Netzsch STA 449C thermal analyzer by increasing the temperature from 25°C to 1000°C with a heating rate of 10 °C min‐1 in air flow of ca. 60 mL min‐1. For the determination of the Pt content, Pt@HGS and Pt/Vulcan where heated to 1000°C with a heating rate of 20 °C S3
min‐1 in air flow of ca. 60 mL min‐1. The silica content determined for the support alone is subtracted from the residual mass and the resulting mass is considered to be Pt. The Raman spectra were recorded with a HORIBA Jobin Yvon spectrometer. A laser operating at a wavelength of 632.8 nm was used as radiation source.
Electrochemical characterization Ex‐situ half cell measurements The catalyst powders were dispersed ultrasonically in ultrapure water (18 MΩ, Millipore®) for at least 45 minutes initially and again for at least 10 minutes prior to pipetting onto the glassy carbon discs (5 mm diameter, 0.196 cm² geometrical surface area). The catalysts were dried in air or under mild vacuum. The electrochemical measurements were conducted at room temperature in a 150 mL Teflon three‐ compartment electrochemical cell, using a rotating disk electrode (RDE) setup, a Gamry Reference 600 potentiostat and a Radiometer Analytical rotation controller. The potentiostat, the rotator and the gas flow were automatically regulated using an in‐house‐developed LabVIEW software.3 A graphite rod was employed as the counter electrode, and a saturated Ag/AgCl Electrode (Metrohm) served as reference. However, all potentials are given with respect to the reversible hydrogen electrode potential (RHE), which was experimentally determined for each measurement. The reference electrode compartment was separated from the main compartment with a Nafion membrane to avoid contamination with chlorides during activity and stability tests. Both activity and stability measurements were performed in 0.1 M HClO4. The electrolyte was prepared with ultrapure water and conc. HClO4 (Merck, Suprapur). Solution resistance was compensated for in all electrochemical measurements via positive feedback. The residual uncompensated resistance was less than 4Ω in all experiments. Activity measurements were performed for different amounts of catalyst for each material at the working electrode. Loadings were in the range of 5 to 30 µgPt cm‐2 at the electrode in order to obtain thin and well dispersed catalyst films. The catalyst S4
materials were subjected to cleaning cycles before activity measurements until a stable cyclovoltammogram was obtained. This procedure was extended for the Pt@HGS catalyst to typically 200 cleaning cycles (0.05 ‐ 1.35 VRHE, 0.2 V s‐1) for removal of carbon impurities, prior to determination of SA and ECSA. The general guidelines for activity measurements were followed as described previously.4 Specific activities were calculated from the anodic scan of RDE polarization curves at 0.9 VRHE, a rotation rate of 1600 rpm and a scan rate of 50 mV s‐1. In order to isolate current related to oxygen reduction, the RDE polarization curves were corrected for capacitive processes. For this purpose a cyclovoltammogram recorded with the same scan rate and potential window but in argon saturated solution was subtracted from the oxygen reduction polarization curves. The surface area was determined via electrochemical oxidation of adsorbed carbon monoxide (CO‐stripping). In each CO‐stripping experiment, carbon monoxide was adsorbed on the catalyst in a potential region (e.g. 0.05 VRHE) at which no CO oxidation occurs, until the saturation coverage was reached. Afterwards the electrolyte is purged with argon again until all remaining carbon monoxide is removed from the electrolyte, while the same potential is held. Finally, pre‐adsorbed CO is oxidized electrochemically and the charge corresponding to the CO oxidation is measured by the area of the oxidation peak. A more detailed description of general features of CO‐ stripping curves as seen in Figure 4a can be found in the literature.2 The mass activity was calculated based on the specific activity and electrochemical active surface area (ECSA), which was determined independently with several CO‐stripping experiments for at least three different catalyst loadings at the working electrode.
S5
Supporting figure 1. Thermo gravimetric analysis (TGA) in air with heating rate of 10 °C min‐1 of: amorphous hollow carbon spheres (black curve) and hollow graphitic spheres after graphitization process with Fe3+ (red curve).
Supporting figure 2. Nitrogen sorption isotherms of solid core–mesoporous shell (SCMS) silica exotemplate (green curve) and hollow graphitic spheres (HGS) replica (blue curve).
S6
Supporting figure 3. BHJ pore size distributions of SCMS silica exotemplate and HGS: A) adsorption branch and B) desorption branch. The pore size distribution obtained from the desorption data may be influenced by the tensile strength effect.
Supporting figure 4. Particle size distributions of Pt@HGS: before and after thermal treatment from 25°C to 850°C. Thermal treatment carried out under in‐situ XRD measurement conditions.
S7
Supporting figure 5. Degradation curves of Pt/Vulcan and Pt@HGS (as provided in the main text figure 4b) with error bars as resulting from the CO‐stripping surface area determination.
Supporting figure 6. The IL‐TEM images of the Pt/Vulcan reference catalyst before (A) and after (B) 3600 degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4) provides clear proof of particle growth.
S8
The degradation behavior of this catalyst was also previously studied and compared to other catalysts in the literature.2,5 Even though the catalyst shows an overlap of several degradation mechanisms (particle growth, dissolution, particle detachment and carbon corrosion were clearly observed at some catalyst locations; not shown here, for more details see ref.2) under the applied, aggressive start‐stop conditions the vast majority of all catalyst locations suffers from severe particle growth as shown in this IL‐TEM micrograph within the first 3600 degradation cycles. Especially the formation of “string‐shaped”, “L‐shaped” or “T‐shaped” clusters was observed, which is a strong indication of agglomeration.2,5 Furthermore “necking” (meaning the presence of a “bridge” with a lower diameter connecting two platinum particles) was seen for this catalyst after the electrochemical degradation test.2 Besides of this an increase in average particle size and a tail towards larger particle sizes in the particle size distribution after the degradation test are further indications that agglomeration is playing a crucial role for the Pt/Vulcan catalyst under start/stop conditions. A more quantitative analysis of the change in particle size distribution is provided in the main text in figure 6 for a typical catalyst location. The observation that agglomeration is a degradation mechanism of major importance during start/stop conditions was also demonstrated in other IL‐TEM or IL‐Tomography studies for Pt or Pt‐alloys on different types of carbon supports.6‐8
S9
Supporting figure 7. The IL‐TEM images of the Pt/Vulcan reference catalyst were recorded after 0 (A, C) and after 5000 (B, D) degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4).
The IL‐TEM images of the Pt/Vulcan reference catalyst were recorded for the pristine state (A, C) and after 5000 (B, D) degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4). Blue arrows highlight some examples of particles, which clearly become smaller during the degradation test and such provide clear proof for dissolution to occur. The red circle highlights a platinum particle, which is observed next to the carbon support on the coating of the TEM grid after the aging test, indicating particle detachment. Green circles highlight S10
regions where particle growth can clearly be observed after the degradation test. The shape of some of the formed larger platinum particles as seen after the electrochemical aging test points towards an agglomeration mechanism as parts of the original shapes of former individual particles can still be identified. A more quantitative investigation of the particle size distributions in the pristine state and after 5000 degradation cycles is depicted in supporting figure 9. Supporting figure 8. The IL‐TEM images of the Pt@HGS catalyst were recorded after 0 (A, C) and after 5000 (B, D) degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4).
S11
C and D are magnifications of the rectangular sections marked in A and B. A first qualitative evaluation of the images clearly shows that no particle growth is taking place, whereas the total number of platinum particles slightly decreases. A more quantitative investigation of the particle size distributions in the pristine state and after 5000 degradation cycles is depicted in supporting figure 9. Supporting figure 9. Corresponding particle size distributions of Pt/Vulcan and Pt@HGS for representative identical catalyst locations of both materials.
The particle size distributions were obtained from IL‐TEM of the pristine state and after 5000 degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4). Because the particles are not spherical, the shape was approximated by ellipses. A diameter corresponding to an ideal sphere was calculated for every single particle from the area obtained from the ellipse, and this was used to calculate the average spherical diameter. The dots in the graphs reflect the counts per particle diameter, while the lines are intended as a guide to the eye. The visual indications for agglomeration as observed for the Pt/Vulcan from the IL‐ TEM images (See supporting figure 7) are confirmed by the quantitative evaluation. The tail in the particle size distribution towards larger particle sizes, which was already observed after 3600 degradation cycles can also be seen after 5000 degradation cycles indicating agglomeration. While the total number of particles strongly decreases an
S12
increase in the number of small particles can be observed after 5000 degradation cycles confirming dissolution as an important degradation mechanism. In contrast to the Pt/Vulcan catalyst, Pt@HGS does not show any increase in particle size after 3600 degradation cycles (see main article figure 6) and also not after 5000 degradation cycles. This proves that no significant agglomeration is taking place for this catalyst material. The total loss in particles is also much less pronounced, which can be seen as an indication that detachment is reduced for the Pt@HGS catalyst. While for Pt/Vulcan at several locations detached particles are observed on the coating of the TEM grid, this was never observed for Pt@HGS, which is a further indication that detachment does not play a pronounced role here. However, as observable from the decrease in average particle size and from the increase of the amount of smaller particles, dissolution seems to be present also for the Pt@HGS material. Compared to the reference Pt/Vulcan catalyst, agglomeration and detachment of metal nanoparticles is strongly reduced for the Pt@HGS catalyst by pore confinement, while dissolution was found to be present for both materials.
References (1) Büchel, G.; Unger, K. K.; Matsumoto, A.; Tsutsumi, K. Adv. Mater. 1998, 10, 1036 (2) Meier, J. C.; Galeano, C.; Katsounaros, I.; Topalov, A. A.; Kostka, A.; Schüth, F.; Mayrhofer, K. J. J. ACS Catal. 2012, 832 (3) Topalov, A. A.; Katsounaros, I.; Meier, J. C.; Klemm, S. O.; Mayrhofer, K. J. J. Rev. Sci. Instrum. 2011, 82, 114103 (4) Mayrhofer, K. J. J.; Strmcnik, D.; Blizanac, B. B.; Stamenkovic, V.; Arenz, M.; Markovic, N. M. Electrochim. Acta 2008, 53, 3181 (5) Meier, J. C.; Katsounaros, I.; Galeano, C.; Bongard, H. J.; Topalov, A. A.; Kostka, A.; Karschin, A.; Schüth, F.; Mayrhofer, K. J. J. Energy Environ. Sci 2012, 5, 9319 (6) Schlögl, K.; Hanzlik, M.; Arenz, M. J. Electrochem. Soc. 2012, 159, B677 (7) Hartl, K.; Hanzlik, M.; Arenz, M. Energy Environ. Sci. 2011, 4, 234 (8) Yu, Y.; Xin, H. L.; Hovden, R.; Wang, D.; Rus, E. D.; Mundy, J. A.; Muller, D. A.; Abruña, H. D. Nano Lett. 2012, 12, 4417
S13
Supporting information Material preparation Synthesis of solid core‐mesoporous shell silica exotemplate (SCMS)1 A typical synthesis of 10 g of SCMS silica spheres is as follows. 32.9 mL of aqueous ammonia (28%) is mixed with 500 g of ethanol and 120 mL of deionized water. After stirring for ca. 10 min, 23.6 mL of tetraethoxysilane (TEOS, 98%) are added, and the reaction mixture is stirred for ca. 1 h. Afterwards, a mixture solution containing 14.1 mL of TEOS and 5.7 mL of octadecyltrimethoxysilane (ODTMS, 90% tech.) is drop‐wise added (for ca. 20 min) to the colloidal solution containing the silica spheres and further reacted for 5 h without stirring. The resulting SCMS spheres are separated from S1
the solution by centrifugation, dried at 75°C overnight and further calcined at 550°C under air atmosphere to produce the final uniform spherical SCMS particles. Comparison materials The comparison material Pt/Vulcan (20 wt.%) is prepared by colloidal deposition of pre‐synthesized Pt nanoparticles on commercial Vulcan support. The synthesis and characterization details of this material can be found in our previous publication.2
Material characterization High resolution transmission electron microscopy (HR‐TEM) images were obtained on a HF‐2000 microscope equipped with a cold field emitter (CFE) and operated at a maximum acceleration voltage of 200 kV. Typically, the samples were placed on a Lacey carbon film supported by a copper grid. Solid samples were deposited on the Lacey carbon film without previous dissolution. High resolution scanning electron microscopy (HR‐SEM) and scanning transmission electron microscopy (STEM) micrographs were collected on a Hitachi S‐5500 ultra‐ high resolution cold field emission scanning electron microscope. The instrument was operated at a maximum acceleration voltage of 30 kV. The samples were prepared on Lacey carbon films supported on a 400 mesh copper grid. The use of Duo‐STEM Bright Field/Dark Field detector together with the secondary electron (SE) detector geometry allows simultaneous imaging of surface morphologies in scan mode, and dark field/bright field imaging in transmission mode. The same HR‐SEM/STEM microscope was used for the identical location SEM/STEM experiments. To obtain the cross‐ sectional cuttings, the Pt@HGS material was embedded in Spurr‐resin (a low‐viscosity epoxy resin embedding medium for electron microscopy) and then subjected to the cutting procedure in an ultramicrotome (Reichert Ultracut) equipped with a diamond knife. The resulting slices of the composite present a thickness of ca. 30‐50 nm. Nitrogen sorption measurements were carried out on a Micrometrics ASAP 2010 instrument. Prior to analysis, the silica exotemplate was activated under vacuum for at S2
least 8 h and the HGS for at least 15 h at 250°C. The measurements were performed at ‐ 196°C using a static‐volumetric method. The empty volume was determined with nitrogen. The BET surface area was calculated from the adsorption data in the relative pressure interval from 0.04 to 0.2. The pore size distribution was estimated by the BJH (Barrett‐Joyner‐Halenda) method from the adsorption branch (desorption data, which are normally recommended by IUPAC for BJH analysis, may be influenced by the tensile strength effect, see Figure SI‐3). The total pore volume was estimated from the amount of nitrogen adsorbed at a relative pressure of 0.97. In‐situ X‐ray powder diffraction (XRD) data were collected in reflection geometry on a Bragg Brentano diffractometer (X’Pert PRO, PANalytical) equipped with an Anton Paar XRK900 high‐temperature reaction chamber and a CuKα1,2 radiation source (40 kV, 40 mA) with the following slit configuration: primary and secondary soller slits 0.04 rad, divergence slit 0.5°, anti‐scatter slits 1°. Instead of a monochromator, a secondary Ni filter was inserted before a position sensitive real time multi strip detector (X'Celerator, 2.12° 2 active length). The reaction chamber is equipped with a Marcor® sample holder (6‐10 mm diameter). The sample is prepared on a sieve plate (10 mm diameter, 1 mm depth) which allows the protective gas to flow through the sample and leave the chamber through an exhaust pipe. Measurements were taken under 100% nitrogen flow. The samples were heated with a heating rate of 5 °C min‐1. After reaching the appropriate temperature, the waiting time before starting data collection was set to 30 min. Data were collected in the range between 20 and 90° 2. The sample was kept for 3 h at 850°C before starting the measurement. Room temperature XRD patterns were collected with a Bragg Brentano diffractometer (STOE THETA/THETA). The instrument is equipped with a secondary graphite monochromator (CuKα1,2 radiation) and a proportional gas detector. The divergence slit was set to 0.8°, the receiving slit was set to 0.8 mm, and the width of the horizontal mask was 4 mm. The samples were prepared on a background free single crystal quartz sample holder. Thermal stability of HGS was investigated by TG‐DTA using a Netzsch STA 449C thermal analyzer by increasing the temperature from 25°C to 1000°C with a heating rate of 10 °C min‐1 in air flow of ca. 60 mL min‐1. For the determination of the Pt content, Pt@HGS and Pt/Vulcan where heated to 1000°C with a heating rate of 20 °C S3
min‐1 in air flow of ca. 60 mL min‐1. The silica content determined for the support alone is subtracted from the residual mass and the resulting mass is considered to be Pt. The Raman spectra were recorded with a HORIBA Jobin Yvon spectrometer. A laser operating at a wavelength of 632.8 nm was used as radiation source.
Electrochemical characterization Ex‐situ half cell measurements The catalyst powders were dispersed ultrasonically in ultrapure water (18 MΩ, Millipore®) for at least 45 minutes initially and again for at least 10 minutes prior to pipetting onto the glassy carbon discs (5 mm diameter, 0.196 cm² geometrical surface area). The catalysts were dried in air or under mild vacuum. The electrochemical measurements were conducted at room temperature in a 150 mL Teflon three‐ compartment electrochemical cell, using a rotating disk electrode (RDE) setup, a Gamry Reference 600 potentiostat and a Radiometer Analytical rotation controller. The potentiostat, the rotator and the gas flow were automatically regulated using an in‐house‐developed LabVIEW software.3 A graphite rod was employed as the counter electrode, and a saturated Ag/AgCl Electrode (Metrohm) served as reference. However, all potentials are given with respect to the reversible hydrogen electrode potential (RHE), which was experimentally determined for each measurement. The reference electrode compartment was separated from the main compartment with a Nafion membrane to avoid contamination with chlorides during activity and stability tests. Both activity and stability measurements were performed in 0.1 M HClO4. The electrolyte was prepared with ultrapure water and conc. HClO4 (Merck, Suprapur). Solution resistance was compensated for in all electrochemical measurements via positive feedback. The residual uncompensated resistance was less than 4Ω in all experiments. Activity measurements were performed for different amounts of catalyst for each material at the working electrode. Loadings were in the range of 5 to 30 µgPt cm‐2 at the electrode in order to obtain thin and well dispersed catalyst films. The catalyst S4
materials were subjected to cleaning cycles before activity measurements until a stable cyclovoltammogram was obtained. This procedure was extended for the Pt@HGS catalyst to typically 200 cleaning cycles (0.05 ‐ 1.35 VRHE, 0.2 V s‐1) for removal of carbon impurities, prior to determination of SA and ECSA. The general guidelines for activity measurements were followed as described previously.4 Specific activities were calculated from the anodic scan of RDE polarization curves at 0.9 VRHE, a rotation rate of 1600 rpm and a scan rate of 50 mV s‐1. In order to isolate current related to oxygen reduction, the RDE polarization curves were corrected for capacitive processes. For this purpose a cyclovoltammogram recorded with the same scan rate and potential window but in argon saturated solution was subtracted from the oxygen reduction polarization curves. The surface area was determined via electrochemical oxidation of adsorbed carbon monoxide (CO‐stripping). In each CO‐stripping experiment, carbon monoxide was adsorbed on the catalyst in a potential region (e.g. 0.05 VRHE) at which no CO oxidation occurs, until the saturation coverage was reached. Afterwards the electrolyte is purged with argon again until all remaining carbon monoxide is removed from the electrolyte, while the same potential is held. Finally, pre‐adsorbed CO is oxidized electrochemically and the charge corresponding to the CO oxidation is measured by the area of the oxidation peak. A more detailed description of general features of CO‐ stripping curves as seen in Figure 4a can be found in the literature.2 The mass activity was calculated based on the specific activity and electrochemical active surface area (ECSA), which was determined independently with several CO‐stripping experiments for at least three different catalyst loadings at the working electrode.
S5
Supporting figure 1. Thermo gravimetric analysis (TGA) in air with heating rate of 10 °C min‐1 of: amorphous hollow carbon spheres (black curve) and hollow graphitic spheres after graphitization process with Fe3+ (red curve).
Supporting figure 2. Nitrogen sorption isotherms of solid core–mesoporous shell (SCMS) silica exotemplate (green curve) and hollow graphitic spheres (HGS) replica (blue curve).
S6
Supporting figure 3. BHJ pore size distributions of SCMS silica exotemplate and HGS: A) adsorption branch and B) desorption branch. The pore size distribution obtained from the desorption data may be influenced by the tensile strength effect.
Supporting figure 4. Particle size distributions of Pt@HGS: before and after thermal treatment from 25°C to 850°C. Thermal treatment carried out under in‐situ XRD measurement conditions.
S7
Supporting figure 5. Degradation curves of Pt/Vulcan and Pt@HGS (as provided in the main text figure 4b) with error bars as resulting from the CO‐stripping surface area determination.
Supporting figure 6. The IL‐TEM images of the Pt/Vulcan reference catalyst before (A) and after (B) 3600 degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4) provides clear proof of particle growth.
S8
The degradation behavior of this catalyst was also previously studied and compared to other catalysts in the literature.2,5 Even though the catalyst shows an overlap of several degradation mechanisms (particle growth, dissolution, particle detachment and carbon corrosion were clearly observed at some catalyst locations; not shown here, for more details see ref.2) under the applied, aggressive start‐stop conditions the vast majority of all catalyst locations suffers from severe particle growth as shown in this IL‐TEM micrograph within the first 3600 degradation cycles. Especially the formation of “string‐shaped”, “L‐shaped” or “T‐shaped” clusters was observed, which is a strong indication of agglomeration.2,5 Furthermore “necking” (meaning the presence of a “bridge” with a lower diameter connecting two platinum particles) was seen for this catalyst after the electrochemical degradation test.2 Besides of this an increase in average particle size and a tail towards larger particle sizes in the particle size distribution after the degradation test are further indications that agglomeration is playing a crucial role for the Pt/Vulcan catalyst under start/stop conditions. A more quantitative analysis of the change in particle size distribution is provided in the main text in figure 6 for a typical catalyst location. The observation that agglomeration is a degradation mechanism of major importance during start/stop conditions was also demonstrated in other IL‐TEM or IL‐Tomography studies for Pt or Pt‐alloys on different types of carbon supports.6‐8
S9
Supporting figure 7. The IL‐TEM images of the Pt/Vulcan reference catalyst were recorded after 0 (A, C) and after 5000 (B, D) degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4).
The IL‐TEM images of the Pt/Vulcan reference catalyst were recorded for the pristine state (A, C) and after 5000 (B, D) degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4). Blue arrows highlight some examples of particles, which clearly become smaller during the degradation test and such provide clear proof for dissolution to occur. The red circle highlights a platinum particle, which is observed next to the carbon support on the coating of the TEM grid after the aging test, indicating particle detachment. Green circles highlight S10
regions where particle growth can clearly be observed after the degradation test. The shape of some of the formed larger platinum particles as seen after the electrochemical aging test points towards an agglomeration mechanism as parts of the original shapes of former individual particles can still be identified. A more quantitative investigation of the particle size distributions in the pristine state and after 5000 degradation cycles is depicted in supporting figure 9. Supporting figure 8. The IL‐TEM images of the Pt@HGS catalyst were recorded after 0 (A, C) and after 5000 (B, D) degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4).
S11
C and D are magnifications of the rectangular sections marked in A and B. A first qualitative evaluation of the images clearly shows that no particle growth is taking place, whereas the total number of platinum particles slightly decreases. A more quantitative investigation of the particle size distributions in the pristine state and after 5000 degradation cycles is depicted in supporting figure 9. Supporting figure 9. Corresponding particle size distributions of Pt/Vulcan and Pt@HGS for representative identical catalyst locations of both materials.
The particle size distributions were obtained from IL‐TEM of the pristine state and after 5000 degradation cycles between 0.4 and 1.4 VRHE (Scan rate: 1 Vs‐1; room temperature; no rotation; in 0.1 M HClO4). Because the particles are not spherical, the shape was approximated by ellipses. A diameter corresponding to an ideal sphere was calculated for every single particle from the area obtained from the ellipse, and this was used to calculate the average spherical diameter. The dots in the graphs reflect the counts per particle diameter, while the lines are intended as a guide to the eye. The visual indications for agglomeration as observed for the Pt/Vulcan from the IL‐ TEM images (See supporting figure 7) are confirmed by the quantitative evaluation. The tail in the particle size distribution towards larger particle sizes, which was already observed after 3600 degradation cycles can also be seen after 5000 degradation cycles indicating agglomeration. While the total number of particles strongly decreases an
S12
increase in the number of small particles can be observed after 5000 degradation cycles confirming dissolution as an important degradation mechanism. In contrast to the Pt/Vulcan catalyst, Pt@HGS does not show any increase in particle size after 3600 degradation cycles (see main article figure 6) and also not after 5000 degradation cycles. This proves that no significant agglomeration is taking place for this catalyst material. The total loss in particles is also much less pronounced, which can be seen as an indication that detachment is reduced for the Pt@HGS catalyst. While for Pt/Vulcan at several locations detached particles are observed on the coating of the TEM grid, this was never observed for Pt@HGS, which is a further indication that detachment does not play a pronounced role here. However, as observable from the decrease in average particle size and from the increase of the amount of smaller particles, dissolution seems to be present also for the Pt@HGS material. Compared to the reference Pt/Vulcan catalyst, agglomeration and detachment of metal nanoparticles is strongly reduced for the Pt@HGS catalyst by pore confinement, while dissolution was found to be present for both materials.
References (1) Büchel, G.; Unger, K. K.; Matsumoto, A.; Tsutsumi, K. Adv. Mater. 1998, 10, 1036 (2) Meier, J. C.; Galeano, C.; Katsounaros, I.; Topalov, A. A.; Kostka, A.; Schüth, F.; Mayrhofer, K. J. J. ACS Catal. 2012, 832 (3) Topalov, A. A.; Katsounaros, I.; Meier, J. C.; Klemm, S. O.; Mayrhofer, K. J. J. Rev. Sci. Instrum. 2011, 82, 114103 (4) Mayrhofer, K. J. J.; Strmcnik, D.; Blizanac, B. B.; Stamenkovic, V.; Arenz, M.; Markovic, N. M. Electrochim. Acta 2008, 53, 3181 (5) Meier, J. C.; Katsounaros, I.; Galeano, C.; Bongard, H. J.; Topalov, A. A.; Kostka, A.; Karschin, A.; Schüth, F.; Mayrhofer, K. J. J. Energy Environ. Sci 2012, 5, 9319 (6) Schlögl, K.; Hanzlik, M.; Arenz, M. J. Electrochem. Soc. 2012, 159, B677 (7) Hartl, K.; Hanzlik, M.; Arenz, M. Energy Environ. Sci. 2011, 4, 234 (8) Yu, Y.; Xin, H. L.; Hovden, R.; Wang, D.; Rus, E. D.; Mundy, J. A.; Muller, D. A.; Abruña, H. D. Nano Lett. 2012, 12, 4417
S13
