Water dispersible surface-functionalized platinum/carbon nanorattles for size-selective catalysis

Surface-functionalized hollow carbon nanobubbles containing platinum in their interior perform size-selective catalysis.


Materials and General Methods
Dispersions of nanoparticles were produced with the aid of an ultrasonic bath (Bandelin Sonorex) and a high sheer mixer (T10 basic ULTRA-TURRAX (IKA)). For centrifugation a 3-30KS centrifuge (Sigma) was used. The carbon nanorattles were filtered through Amicon Ultra-15 or -4 Centrifugal Filter Units (50 kDa, Merck). The nanoparticles were analyzed by elemental microanalysis ELEMENTAR (Elementar Analysensysteme). X-ray diffraction (XRD) patterns were recorded on an X'Pert PRO-MPD diffractometer (Cu-Kα radiation, X'Celerator linear detector system; PANalytical, Netherlands) and inductively coupled plasma atomic emission spectroscopy (ICP-OES) by ULTIMA ICPOES. Transmission electron microscopy (TEM) was measured with a F30 (Tecnai F30 ST (FEI, FEG, 300 kV)), scanning transmission electron microscope (STEM) on a NovaNanoSEM 450 (FEI, operated at 30 kV) with a HAADF detector. EDX was performed on a HD2700CS. For both, the particles were deposited on a carbon/copper grid.
The analyte concentration (platinum and cobalt in nanorattles) was determined using an inductively coupled plasma time-of-flight mass spectrometer (icpTOF, TOFWERK AG). No sample treatment was conducted and samples were diluted in ultra-high purity (UHP) water (Millipore, Billerica, Massachusetts, USA). Platinum, cobalt and iridium stock solutions were prepared from single-element standard solutions (Inorganic Ventures, Christiansburg, Virginia, USA). Subsequent dilutions were prepared with a solution of 2% HNO 3 and 2% HCl (Traceselect Ultra, Fluka) in UHP water. All dilutions were carried out gravimetrically using an electronic balance (Mettler AE240, Mettler-Toledo, Grefensee, Switzerland). Final solutions of platinum and cobalt with concentrations ranging from 1 ng/g to 500 ng/g were used for external calibrations. 10 ng/g Iridium was spiked in the calibrations solutions as well as in the diluted samples as internal standard to monitor instrumental drift and analyte response. 10000 TOF extractions were integrated for one data point and 200 data points were collected for a minute. Hence, 200 data points each integrated over 300 ms were used to determine the average and standard deviation.
Mass spectrometry of the product of the platinum catalyzed oxidation of glucose and maltoheptaose was performed using an Agilent 1260 Infinity Quadrupole 6130 LC/MS. The parameters for the direct injection method (also mass spectrometer parameters) are as follows: 5 µl injection, 0.1 mL/min flow, a solvent mix of 50/50 (A: NH 4 Ac 13 mM at pH7, B: MeCN), fragmentor = 200, capillary voltage 4000 V +/-, drying gas flow = 12 l/min, nebulizer pressure = 35 psig, drying gas temperature = 250 °C. Simultaneous electrospray ionization (ESI) in positive and negative mode was performed with a fast switching dielectric capillary which allowed for injection of the sample only once. Single ion monitoring (SIM) in the negative mode was done for relative quantification of the products of the catalysis, namely gluconic acid (195 m/z) and maltoheptanonic acid (583 m/z).
The particle size distribution was characterized by optically measuring the diameter of at least 200 particles per sample in STEM/SEM images.
Production of C/CoPt-nanoparticles. Platinum acetylacetonate (0.435 g) was dissolved in a mixture of 2-ethylhexanoic acid and tetrahydrofuran (60 mL, wt/wt 1:1) to give a brownish precursor solution. The metal content of the platinum precursor was determined by adding small amounts of precursor into Erlenmeyer-flasks and burning of the organic contents by heating to 450 °C for 4 h. By measuring the mass of the residues and knowing the initial amounts of precursor which was burnt, the metal content was calculated (1.37 wt% Pt). The cobalt precursor (cobalt 2-ethylhexanoate, 65% in mineral spirits, 12% Co, ABCR) was diluted by THF (precursor / THF 1:2). The platinum and the cobalt precursors were mixed at a ratio to target 20 wt%, 10wt% and 1 wt% platinum in cobalt. The carbon-coated cobalt platinum nanoparticles were produced in a one step process by reducing flame spray pyrolysis under an oxygen-free atmosphere as described by Grass et al. 1 Synthesis of C/CoPt@initiator (2) was performed as described elsewhere. 2

Synthesis of C/CoPt@pSPM (4).
The procedure was adapted from literature. 3 All reaction steps were performed under a protective nitrogen atmosphere. The monomer solution was prepared by dissolving 3-(2-Methylprop-2-enoyloxy)propane-1-sulfonic acid potassium salt (SPM, 3) (8.6 g, 34.9 mmol) in MeOH/H 2 O (2:1, 12 mL) and consecutive degassing by nitrogen bubbling for 30 minutes. CuBr 2 (10 mg, 0.045 mmol), 2,2'-bipyridine (54 mg, 0.35 mmol), L-ascorbic acid (60 mg, 0.34 mmol) and NaCl (90 mg, 1.54 mmol) were added to the solution and it was degassed for further 5 minutes. C/Co@initiator (2) (500 mg) were placed in a Schlenk flask and degassed (3 × high vacuum pump / N 2 refill cycles). The monomer solution was then added by syringe. The reaction mixture was exposed to sonication for 10 minutes to obtain a homogeneous dispersion. It was then stirred for 18 hours at 40 °C. The poly-SPM functionalized nanoparticles (C/CoPt@pSPM, 4) were magnetically separated. After magnetic decantation, the particles were washed three times with water. Acetone (twice the volume of the washing water) was used to destabilize the particles. It was further washed with ethanol, ethyl acetate and acetone, twice each. After each washing procedure (sonication for 3 minutes in solvent) the particles were recovered by the external magnet and the washing solvent was drained. The nanoparticles were dried in a vacuum oven at 50 °C.
Synthesis of polymer coated carbon nanorattles containing Pt (5). Pestled C/CoPt@pSPM (4) (50 mg) were added to hydrochloric acid (0.1 M, 50 mL) and dispersed by ultrasonication for 15 minutes and subsequent stirring by the aid of a high sheer mixer. The dispersion was then magnetically stirred at 1000 rpm and heated to 80 °C for one week. The nanorattles and non-dissolved cobalt-platinum particles were separated from each other by magnetic separation of the latter and subsequent draining of the supernatant containing the nanorattles. The nanorattles were then collected from the reaction solvent by the aid of amicon ultra-15 centrifugal filter units and then washed five times by adding water to the concentrate, shaking and subsequent filtration.

X-ray diffraction pattern (XRD)
Fig. S1 X-ray diffraction pattern of carbon-coated cobalt-platinum-alloy nanoparticles (C/CoPt) containing 1 wt%, 10 wt% and 20 wt% platinum. Reference lines drawn in grey are taken from pdf_00-015-0806 for Co and pdf_03-065-8970 for CoPt. From the mass gain in thermogravimetric analysis the amount of platinum in the C/CoPt particles can be estimated. By knowing the weight percentage of carbon within the particles from elemental analysis and the mass gain upon oxidation of the cobalt, it is possible to calculate the amount of platinum present in the sample via the following equation: (1 − ) • ( + (1 − )) • 1.36 = 1 + ∆m ox = 2.78•(1+∆m ) −1 + 3.78 (eq. S1) where C is the weight part of the particles being carbon, Pt is the weight part of the particles being platinum and the factor 1.36 resulting from the mass gain of cobalt upon oxidation (Co  Co 3 O 4 ).
Calculating the platinum amount of the particles via equation S1 leads to the results shown in Table S3.  : calculated from the wt% of the organic part of the C/CoPt@pSPM (4) particles, determined by elemental microanalysis, with the assumption of total cobalt dissolution and that the platinum content is the one which was aimed in the particle production process (meaning 20 wt% Pt).  Fig. S4 Particle size distribution and cumulative probability of carbon-coated cobalt-platinum-alloy nanoparticles (C/CoPt) containing 1 wt%, 10 wt% and 20 wt% platinum measured by diameter analysis in STEM pictures. An average diameter of 23 ± 14 nm for C/CoPt (1 wt% Pt), 19 ± 13 nm for C/CoPt (10 wt% Pt) and 18 ± 10 nm for C/CoPt (20 wt% Pt) was found by measuring 200, 300 and 300 particles, respectively.   Table S5), c : calculated from the wt% of cobalt and platinum in the sample (determined by elemental microanalysis) and the amount of non-dissolved, magnetically removed cobalt-platinum-alloy particles 5 Characterization of platinum nanorattles

X-ray diffraction pattern (XRD)
Table S7 Particle diameters determined by peak width of XRD patterns using the Scherrer formula 4 and number-and volume-weighted mean particle diameters, determined by microscopy images, of platinum nanoparticles in the interior of the nanorattles and C/CoPt with different platinum contents (20 wt%, 10 wt% and 1 wt% platinum Determining the crystal size from the XRD patterns for the C/CoPt (20 wt% Pt) nanoparticles resulted in a smaller size (13.6 nm) than the volume-weighted mean particle diameter determined by microscopy images (17.8 nm; see Table S7 for all particle diameters). This difference could be explained by the formation of polycrystalline particles or the formation of twins as previously observed for fcc-cobalt nanoparticles. 5

Thermogravimetric analysis
Fig. S11 Relative mass loss of Pt nanorattles derived from C/CoPt (20 wt% Pt) nanoparticles upon combustion of the organic part of the particles measured by thermogravimetry in air. Pt nanorattles showed the combustion of 66.9 wt% after heating to 600 °C. : calculated from the wt% of carbon of the C/CoPt nanoparticles, determined by elemental microanalysis, with the assumption of total cobalt dissolution and that the platinum content is the one which was aimed in the particle production process (meaning 20 wt% and 1 wt% platinum).

Inductively coupled plasma time-of-flight mass spectrometry (ICP-TOFMS)
6 Size-selective catalysis 6 As further control, the reaction was also performed with hollow carbon nanobubbles not containing platinum (0.19 mg/mL, HCNB 3 ), carbon nanoparticles (0.38 mg/mL, nano activated carbon (100 nm), Nanostructured & Amorphous Materials, Inc.) and Pt nanoparticles made in the presence of hollow carbon nanobubbles (see Figure S14 for STEM images and below for synthesis procedure).
All signal of the mass spectrometry measurements were normalized by the corresponding control (case iv in the above description).

Scanning transmission electron micrographs (STEM) of Pt nanorattles after usage as catalyst
Fig. S12 Scanning transmission electron microscopy (STEM, a, b) and scanning electron microscopy (SEM, c) of platinum nanorattles made from C/CoPt (20 wt% Pt) nanoparticles after usage as catalyst for the oxidation of glucose and maltoheptaose.

Results of control reactions
Fig. S13 Results of control reactions for the platinum catalyzed oxidation of glucose and maltoheptaose with hollow carbon nanobubbles (HCNB

Formation of platinum nanoparticles in the presence of nanobubbles (Pt NP & HCNB)
Synthesis of Pt NP & HCNB. The procedure was adapted from literature 6 . An aqueous solution of H 2 PtCl 6 (0.25 mL, 0.172 mM) was added to an aqueous solution of hollow carbon nanobubbles (synthesized as described before 3 , 0.25 mL, 0.19 mg bubbles / mL). Then, NaBH 4 (0.5 mL, 5 mM) was added dropwise. The reaction was shaken for 1 h at 500 rpm and then 1 h at 1000 rpm at room temperature. The nanobubbles were washed by filtration (amicon ultra centrifugal filter 4 mL, 50 kDa).

Fig. S14
Scanning transmission electron microscopy of platinum nanoparticles made in the presence of hollow carbon nanobubbles (a-d) and platinum nanoparticles made without the presence of hollow carbon nanobubbles as reference (e, f).