Specific ion effects directed noble metal aerogels: Versatile manipulation for electrocatalysis and beyond

Specific ion effects are demonstrated to create and flexibly manipulate noble metal aerogels for versatile applications.


The PDF file includes:
Supplementary Materials and Methods Fig. S1. Characterizations of NP precursors. Fig. S2. Self-healing behavior of NH 4 F-induced gold hydrogels.  Fig. S17. Scanning TEM-EDX analysis of different alloy gels prepared by one-step method. Fig. S18. High-angle annular dark-field scanning transmission electron microscopy imaging and EDX analysis of core-shell structured alloy gels. Fig. S19. SEM images of uncompressed aerogels. Fig. S20. Cross-sectional SEM images of compressed aerogels. Fig. S21. Electrocatalytic performance of different commercial and gel catalysts. Table S1. Summary of the gelation behavior of gold induced by different salts. Table S2. Summary of nitrogen adsorption data and ligament sizes of as-prepared aerogels. Table S3. Elemental analysis of different alloy aerogels. Table S4. Comparison of parameters of NMFs in literature.

Preparation of Gold Hydrogels
Aqueous solution of NH 4 F (1 M, 555 μL) was added in as-prepared gold NP solution (5 mL), stirring for 10~20 s, followed by grounding 4~12 h to get free-standing gold hydrogel. The as-prepared gel is washed by a large amount water for 4~5 times with a total duration of 3 days to remove possible residues. For NP precursor solutions with different c M , the final concentration of NH 4 F (c salt ) was fixed to 100 mM. For hydrogels prepared by using different c salt , the c M was fixed to 0.2 mM.
For scale-up production, aqueous solution of trisodium citrate dehydrate (400 mM, 4 mL) and HAuCl 4 ·3H 2 O (32.5 mM, 4.92 mL) were added successively in 788 mL water and stirred for ~15 min. Then freshly prepared NaBH 4 aqueous solution (200 mM, 3.2 mL) was rapidly injected in the above mixture, followed by stirring for 30~60 min. Afterwards, aqueous solution of NH 4 F (1 M, 88.9 mL) was quickly added, stirring for ~20 s, followed by grounding ~12 h to allow complete reaction.

Fabrication of Alloy NMHs
Alloy hydrogels were synthesized by either one-pot or dynamic shelling approach (DSA). The latter strategy could be used to controlled fabricate core-shell structured gels.
For one-pot method, the procedure is the same to that of single-metal hydrogels, except that single metal precursor salts were replaced by two or more metal precursor salts. The molar ratio of different metal precursor salts is 1:1 for bi-metallic system and 1:1:1 for tri-metallic system, and the total concentration of metal salts in final solution is fixed to 0.2 mM.

Preparation of Noble Metal Aerogels
After purification by water, hydrogels were solvent-exchanged with tert-butanol for 2~3 times. Afterwards, wet gels were flash freezed by liquid nitrogen and remained at -196 ºC for ~10 min to enable complete freezing. The frozen samples were put into the chamber of freeze drier (TOPTI-12S-80) and dried for 12~24 h at ~1Pa. The temperature of the cold trap was set to -80 ºC.

Compression of Aerogels
To make a lustrous and solid aerogel pellet, the as-prepared aerogel was pressed manually by using a polished stainless-steel cylinder for a few seconds. To make the hetero-structured compressed Au-Ag aerogel, the original Au and Ag aerogel were placed together with a small overlapping fraction, followed by compressing as described above.

Macroscopic Force Analysis during Gelation Process
In solution, the aggregate is imposed by gravity (G), buoyancy (f), and viscous drag (F).(39) When the aggregate falls down, the direction of F is the same with that of f, while opposite with that of G. To facilitate estimation, given that all aggregates during gelation have the similar shape (quasi sphere) and density, only differ in sizes. G, F, f, and the volume of aggregate (V) could be expressed as follow and v denote the density of aggregate, density of solution, viscosity of solution, equivalent diameter of aggregate, and velocity of the aggregate.
For F that is expressed by Stock's law, the laminar flow need to be met, where the Reynolds number (Re) should be very small, e.g. less than 10 for the system of spheres. In our system, assume the viscosity and density of solution is close to the water (~10 -3 Pa s and 10 3 kg m -3 ), the maximum velocity (v max ) of the aggregate is 10 times of average settling down velocity (v s ), and the largest size of aggregate in solution is 100 μm. Since the precipitate appeared usually after 1 hours upon reaction for a system with a solution height of 5 cm, so the v s should be less than 5 cm h -1 , thus a v max of 50 cm h -1 was adopted. From above equation, it's clear that acceleration a is determined by the viscous drag term, which is positively related to the mass of aggregate. Hence, at the same velocity, a of big aggregate is always larger than that of the smaller one. So given that the initial velocity is zero, the velocity of larger aggregate should be always larger than that of the smaller one. At the final equilibrium state, the equilibrium speed is also positively related to the size, i.e. the mass of aggregate. Therefore, during whole gelation process, large aggregates would fall down all the way quicker than smaller ones. In this way, during gelation process, the gradually formed large aggregates will continuously fall down and eventually precipitate and gelate at the bottom of the vial.

Density-functional theory (DFT) Calculations
DFT calculations were carried out by using the Amsterdam Density Functional package.(40-42) The generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) functional(43) was employed together with a TZP basis set. The convergence criteria for energy and nuclear gradients were set to 10 -3 Hartree and 10 -3 Hartree/Å, respectively.
The initial structures were generated as follows: a high-temperature trajectory of a single citric-acid molecule was used to scan the conformational space with an approximate DFT method;(44) the optimized lowest-energy structure was taken as the initial structure and a single acidic proton was exchanged for the respective cations (Na + , K + , NH 4 + , Mg 2+ , Ca 2+ ). All structures generated in this way were optimized with full DFT afterwards. In case of the alkaline earth metals, the molecule was positively charged. Dispersion effects were included by using the three parameter correction developed by Grimme et al. (45) To mimic the experimental environment, the conductor-like screening model (COSMO) of solvation as implemented in ADF was used. It includes the respective molecule in a molecule-shaped cavity constructed by the atomic radii and surrounded by a dielectric medium (here: water). Taken single-charged cations as example, the binding energies (E b ) was calculated based on following equation where RCOOand X + denote deprotonated citrate and cation, respectively. E b is defined E b = E RCOOX -E RCOO--E X+ (we will display and discuss absolute values for convenience).

Microscopy Characterization
Scanning electron microscopy (SEM) analysis was performed on a Nova 200 NanoSEM scanning electron microscope. Samples were prepared by directly sticking on the conductive tape.
Transmission electron microscopy (TEM) analysis was carried out by using a FEI Tecnai G2 20 microscope operated at 200 kV. Samples were prepared by dispersing in acetone under ultrasonication (15 s to 120 s, depending on their dispersing ability), followed by dropping onto carbon-coated copper grids and drying at ambient temperature. For each formula, 2~3 TEM specimens from three different batches were analyzed and at least five different positions were surveyed for each one.
High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) imaging and spectrum imaging based on energy-dispersive X-ray spectroscopy (EDX) were performed at 200 kV with a Talos F200X microscope equipped with an X-FEG electron source and a Super-X EDX detector system (FEI). Prior to STEM analysis, the specimen mounted in a high-visibility low-background holder was placed for 2 s into a Model 1020 Plasma Cleaner (Fischione) to remove contamination.
Optical imaging was acquired by Carl Zeiss Microscopy (Colibri 7), with the magnification of 63×10.

Diffraction Characterization
X-ray powder diffraction (XRD) was carried out in reflection mode on a Siemens D5000 X-ray diffractometer operated at a voltage of 30 kV and a current of 10 mA with Cu Kα radiation (λ = 1.5406 Å). The data were collected in the 20°-90° (2θ) range with a step size of Δ2θ = 0.02°. The sample was fixed on the holder by Scotch tape. For single-metal system, the crystalline size could be estimated by the Scherrer equation applying crystallite-shape factor K=0.9.(46)

Spectroscopy Characterization
Fourier-transform infrared spectroscopy (FT-IR) spectra were recorded by Thermo Scientific Nicolet 8700 FT-IR Spectrometer configured with a Smart iTR diamond accessory.

Thermal Properties Characterization
Thermogravimetric analysis (TGA) was conducted by Diamond TG-DTA/Spectrum GX with heating rate of 10 K min -1 under nitrogen atmosphere.

Element Analysis
X-Ray photoelectron spectroscopy (XPS) was performed by an Axis Ultra spectrometer (Kratos, UK) with a high-performance Al monochromatic source operated at 15 kV. The XPS spectra were taken after all binding energies were referenced to the C 1s neutral carbon peak at 284.8 eV, and the elemental compositions were determined from peak area ratios after correction for the sensitivity factor by CasaXPS.
Inductively coupled plasma optical emission spectroscopy (ICP-OES) was performed by Perkin-Elmer Optima 7000DV optical emission spectrometer.

Nitrogen Adsorption Measurement
Nitrogen adsorption experiments were performed with Quantachrome NOVA 3000e at 77 K. The sample was outgassed at 323 K for ~24 h under vacuum before measurement. The filling rod was used to reduce the dead volume and thus improving measurement accuracy. The specific surface area was calculated by using multi-point BET equation (0.1<p/p 0 <0.3). The pore size distribution was derived by using the density functional theory method which is implanted in the software of the instrument. The total pore volume, which is calculated at p/p 0 =0.99, in consistent with the value derived by BJH (Barret-Joyner-Halender) method.

Other Characterizations
Zeta potential and dynamic light scattering (DLS) tests were performed on ZETASIZER NANO (ZEN5600, Malvern Company). For both tests, the original solutions of gold nanoparticles and salts were purified separately with a 0.45 μm filter membrane before the tests. To reflect the real reaction conditions and to avoid the possible deviations incurred by dilution, the concentrations of all reactants (metal precursors, ligands, and salts) used here are exactly the same as described in the experimental part, i.e., 0.2 mM for metal precursors (~3.9 wt.% for gold), 2 mM for ligands, and 100 mM for salts. In this condition, although the absolute value of the hydrodynamic size derived from DLS might be partially affected due to the adopted high NP concentration (for DLS test, the typically used concentration is 0.1~1 wt.%), it is still meaningful to compare the time-lapse size evolution where the relative values count. The ligands and salts in solutions showed negligible effect on light (e.g. light transmission, absorption, scattering), which could be partially deduced from the fact that the solution is colorless and transparent with only salts and ligands. Therefore, their existence may not largely affect the accuracy of the results from the DLS tests.

Electrochemical Measurements
All electrochemical tests were performed with a three-electrode system on Autolab/PGSTAT 30 (Eco Chemie B. V. Utrecht, the Netherlands). Glassy carbon electrode (GCE, 3 mm in diameter), Ag/AgCl (saturated KCl aqueous solution) electrode, and platinum foil were used as working electrode, reference electrode, and counter electrode, respectively.
For modification of working electrode, ~1 mg catalyst was dispersed in 1 mL of 2-propanol by sonicating for ~30 min to acquire the catalyst ink. Then specific amount of ink was transferred on GCE electrode and evaporated at ambient temperature, followed by coating with 5 μL Nafion (0.5 wt % in ethanol). The concentration of Pd and Pt in ink was determined by ICP-OES, and the final loading of Pd and Pt (m Pd+Pt ) was calculated accordingly to be ~20 μg cm -2 . For commercial Pd/C (20 wt.% Pd on carbon black, Alfa) and Pt/C (20 wt.% Pt on carbon, Aldrich) catalyst, they were prepared in the same way, except that the initial concentration in 2-propanol was 4 mg mL -1 .
Cyclic voltammetry (CV) curves were conducted in nitrogen saturated 1 M KOH aqueous solution, with a voltage window between -1.0 and 0.5 V (vs. AgCl/Ag) and a scanning rate of 100 mV s -1 .
For electro-oxidation of ethanol, the test was performed under N 2 atmosphere in 1 M KOH aqueous solution containing 1 M ethanol. CV curves were recorded between -0.9 and 0.3 V (vs. AgCl/Ag) with a scanning rate of 50 mV s -1 . The peak current of the forward scanning (from negative to positive potential) and backward scanning are denoted as I f and I b , respectively. The stability test was conducted for 10000 s at the potential of forward peak current maximum. For electro-oxidation of methanol, 1 M ethanol was replaced by 1 M methanol, and all other conditions were remained the same.

Self-Propulsion Tests
For the self-propulsion test with original Ag aerogels, the solution is prepared by diluting 30 wt.% H 2 O 2 aqueous solution with deionized water to 1.5 wt.% in a plastic Petri dish. Then a piece of Ag aerogel was carefully transferred in above solution and its behavior was recorded by video.
For the test with hetero-structured compressed Au-Ag aerogel, the sample was transferred in diluted H 2 O 2 aqueous solution (2 wt.%), then the rotation behavior was observed and recorded.
The movement or rotation speed was determined by comparing the position of sample from different frames during certain period. The frame per second (FPS) of recorded video was 25, corresponding to the time resolution of 40 ms.      XPS spectra of gold aerogels directed by different salts. All results suggest few impurities (ligands, salts) exist in aerogels. The "*" in (B) denotes the peak from the noise of IR spectrometer.       S18. High-angle annular dark-field scanning transmission electron microscopy imaging and EDX analysis of core-shell structured alloy gels. (A) Au-Ag (1/1) gels prepared by one-pot method. (B-F) Au-Ag (1/1), Au-Pd (1/1), Au-Pd (9/1), Au-Pt (1/1), and Au-Pd-Pt (1/1/1) gels prepared by the dynamic shelling approach. The green arrows in the HAADF-STEM images indicate the line scanning direction.    Tables   Table S1. Summary of the gelation behavior of gold induced by different salts. The salting-out ability decrease from SO4 2to SCN -, and from NH4 + to Ca 2+ according to Hofmeister series. The symbol of "√" and "×" indicate that the product is self-supporting gel or unsupported powders. The value in bracket is the relative surface charge density of cations and anions referred to NH4 + . For calculation, the ionic radii were used according to literature  Table S2. Summary of nitrogen adsorption data and ligament sizes of as-prepared aerogels. Specific surface area SBET (m 2 g -1 ) was calculated in the partial pressure (p/p0) range of 0.1-0.3. Total pore volume Vtot is derived at the volume at p/p0=0.99. Average ligament size dave is derived from the statistic analysis of TEM measurements.