Catalytic Properties of Cryogelated Noble Metal Aerogels

2.1 Nanocrystals synthesis

For the synthesis of the noble metal NCs, we use the modified citrate reduction of metal salts from Enustun and Turkevich [22]. In a 1 L flask 36.2 mL of an aqueous 0.2 wt% dihydrogen tetrachloroplatinate (IV) hexahydrate (ACS, 99.95%) solution was diluted with deionized (DI) water to a total volume of 500 mL. The solution was stirred and heated until boiling. Then 11.6 mL of a 1 wt% trisodium citrate solution (ABCR, 99%) were added and stirred for 30 s. Subsequently 5.5 mL of an ice cold sodium borohydride solution (Fluka ≥99%) with 0.076 wt% were added. The solution immediately changed the color from yellow to brown. Au NCs were prepared with the same method using 12 mL of an aqueous 0.2 wt% silver nitrate (Ag NO₃) (Alfa Aesar, 99%) solution instead of dihydrogen tetrachloroplatinate (IV) hexahydrate. When sodium borohydride is added the solution changes from colorless to yellow.

For Pd NCs 0.051 g of Pd(II)chloride (0.288 mmol, Sigma Aldrich, 99.999%) were dissolved in 10 mL DI water and 37% hydrochloric acid (Sigma Aldrich, reagent grade) with a molar ratio of 1:2 and heated until the solution was clear. Afterwards it was diluted with distilled water to a total volume of 500 mL. The temperature of the solution before sodium borohydride addition was 80°C. The addition of trisodium citrate and sodium borohydride was the same as for Pt and Ag.

Gold nanocrystals (Au NCs) can be synthesized at room temperature using 29 mL of an aqueous 0.2 wt% hydrogen tetrachloroaurate (III) trihydrate solution (ABCR, 99.99%) diluted to 500 mL total volume with DI water. Thirty seconds after the addition of 11.6 mL of a 1 wt% trisodium citrate solution, 5.8 mL of an aqueous 0.07 wt% sodium borohydride solution were added. The solution immediately changed from yellow to red.

2.2 Aerogel fabrication

According to our previous results [21], beside employing an aqueous solution of the NCs, the requirement for aerogel fabrication via the cryogelation method is a volume fraction of NCs of at least 0.1%, in order to avoid shrinkage of the resulting monoliths. After the synthesis the particles were washed and concentrated by factor 100 over an ultrafiltration cell with a 30 kDa regenerated cellulose membrane (Satorius Stedim) and a pressure of 5.5 bar. This solution can be stored for several weeks. To prepare them for aerogel fabrication the colloidal solution was concentrated again by factor 10 by means of centrifuge filters (Amicon Ultra 15, 10 kDa, Merck Milipore). These solutions were then injected into liquid nitrogen, cooled for 10 min to ensure complete freezing and subsequently freeze dried at a pressure lower than 0.05 mbar for at least 24 h (see Figure 1). If the structure is not completely dry, the remaining ice would melt and the structure would collapse.

Fig. 1:

Freezing mechanism of the aqueous NCs solution (left) in liquid nitrogen. Ice crystals form all over the liquid drop, pushing the NCs into a filigree network superstructure. After removing the ice crystals through freeze drying the self-supporting highly-porous NCs superstructure remains (right).

2.3 Catalytic experimental setup

Catalytic activity measurements were carried out in a custom made flow-microreactor system (i.e. a quartz tube with internal diameter=4 mm inserted in an electric furnace) where aerogel masses in the range of 15–40 mg were loaded between two quartz wool layers. For each experiment, the reactor was flushed for at least 30 min with a inert gas mixture of 6% O₂ balanced with He and afterwards heated from 25 to 200°C at 5°C/min while flowing 40–80 Ncc/min (norm cubic centimeters per minute: at standard pressure and room temperature) of reaction mixture (1.3% v/v CO, 10% v/v O₂, balance He), and then cooled back to room temperature. The CO and CO₂ outlet concentrations were continuously measured by means of a nondispersive infrared (NDIR) photometer (ABB Uras 26).

3.1 Electron microscopy characterization

The macroscopic structure of the resulting aerogel monoliths can be described as highly porous, voluminous and therefore with low density. Weighing the monoliths gives masses ranging from 20 mg for Au and Pd NC aerogels to 60 mg for Au and Pt NC aerogels. With an applied starting solution volume of 1 cm³ and the visual confirmation of no shrinkage during cryogelation a density ranging from 20 to 60 mg cm⁻³ can be estimated. This corresponds to 0.2% of the respective bulk material density. Using electron microscopy the micro- and nanostructure of the monolith is characterized (see Figure 2).

Fig. 2:

Electron microscopic characterization of a Pt aerogel. SEM images showing the microscopic structure (A and B), which can be described as thin, interconnected sheets and wires. The inset in A shows the macroscopic Pt aerogel made from 1 mL of a 100 times concentrated NC solution (volume fraction of Pt NC ~0.025%/black bar is equal to 1 cm). Higher magnifications reveal the assembly of single NCs to form the just described sheets. Transmission electron microscope images (C and D) show that the NCs within the sheets are randomly oriented and retain the shape and size of the starting solution as already reported in our earlier work [21].

On a microscopic scale the monoliths can be described as thin, interconnected sheets with random orientation forming the highly porous structure of the monoliths. The sheets show, in many cases, the tendency to bend or start to enroll. Within this structure wires can also be found which might be completely enrolled sheets. Higher magnifications and switching to transmission electron microscopy (TEM) reveals that the sheets are built up from the single NCs with random orientation. Through various scanning electron microscopy (SEM) and TEM images the thickness of the sheets can be estimated from perfectly perpendicular oriented sheets and is in the range of 5–50 nm corresponding to one to 10 particle layers for the given particle size. The shape and the size of the single NCs (see Supporting Information) do not change during the formation of the Pt aerogel. The structures shown are also representative for Pd aerogels. A more detailed characterization of the structure of the noble metal aerogels and their starting colloids can be found in our previous report [21]. However, softer (i.e. in terms of Young’s modulus where Au and Ag have 79 and 83 GPa, respectively, while Pt has 168 GPa) noble metals show deformations of the NC. For example, in Au aerogels we observe ~10 nm platelets in lateral dimension instead of 4 nm spheres as already shown in our previous work [21].

3.2 Catalytic activity

For a full examination of the catalytic potential of the aerogels, the conversion of CO to CO₂ was chosen. This model reaction is frequently used to characterize new materials, ^[1] and the results obtained can be transferred to other heterogeneous reactions, and they have also relevance from an industrial point of view (e.g. diesel oxidation catalysts, hydrogen production from fossil fuels) [23], [24], [25].

Figure 3A shows the first two test runs of a Pd aerogel with a total addition of 1.3% v/v CO. The results indicate clearly an excess of CO_x species around 3000 s elapsed time in the first run. We can observe this behavior for all aerogels (Au, Ag, Pt and Pd) and we attribute it to the removal of residual citrate ligands from the NCs surface. This is confirmed by the second reaction cycle, where the C balance is respected throughout the whole reaction cycle. In the case of the Pt aerogel, the removal of the ligands results in a violent reaction. By tilting the vial of the aerogel, the monolith starts a spontaneous oxidation as can be seen in Figures 3B and C. This “explosion” is most likely related to the high specific surface and catalytic properties of Pt. The oxidation is also observed under inert atmosphere (while being in the freeze dryer flushed with nitrogen after drying) but with a pale colorless flame, which is attributed to the lack of oxygen. Additionally, this spontaneous oxidation can be inhibited, e.g. by changing the ligands of the Pt NCs from citrate ligands to thiol ligands (such as mercaptopropionic acid or mercaptosuccinic acid) before freezing. Alternatively, when lowering the NCs volume fraction below 0.025%, spontaneous oxidations was not observed.

Fig. 3:

(A) The CO_x species composition of a Pd aerogels over the first two test cycles showing additional C in the first run at around 3000 s due to ligand oxidation. (B) Image series of the “exploding” Pt aerogel. (C) Image of the Pt aerogel after the “explosion”.

The catalytic performances of the gels towards CO oxidation after this first “activation” cycle, can be seen in Table 1. Pt performs best and shows full conversion, as well as Pd, while Ag showed lower catalytic activity. Surprisingly, Au shows up to 40% conversion although the crystal size with ~10nm is much bigger than 5 nm, which was reported as the minimum size before [26]. We also checked the influence of the NCs size and found that the catalytic performance decreased with increasing NCs size, which is most likely due to the decreasing specific surface area for bigger particles. To benchmark the aerogel activity, we prepared Pt NCs (5wt%) on alumina by colloidal deposition with the very same particles. This technique is commonly employed to prepare state-of-the-art model catalysts based on metallic NCs [27]. While the ignition temperature (temperature at which 50% conversion is reached during a heating transient) of the aerogel and of the supported NCs is similar in both cases, the extinction temperature (temperature at which 50% conversion is reached during a cooling transient) is lower in case of the aerogel (i.e. the gel outperforms the supported catalyst) as can be seen in Figure 4A. The hysteresis phenomena have been extensively studied over Pt/Al₂O₃ catalysts and has often been ascribed to the local overheating of the NCs due to the exothermicity of the CO oxidation reaction [28], [29], [30], [31], [32]. Our monolith is characterized by a fully active catalytic surface and by a better heat transfer in comparison with the Pt NCs deposited on alumina, where the support (i.e. Al₂O₃) acts as a heat sink. This property might justify the better performances of the aerogel in comparison with the supported catalyst.

Fig. 4:

(A) CO conversion in dependence on the temperature for different particle sizes. The best performance was achieved for gels built from 4 nm particles. The dotted line shows the benchmark experiment with the 4 nm particles deposited on alumina as it is state of the art for catalyst preparation. (B) Time evolution of the COx concentration at room temperature showing immediate conversion and self-poisoning for Pt and Pd aerogels.

The Pt and Pd monoliths show CO oxidation catalytic activity already at room temperature, proving the accessibility of the NCs surface. The activity drops then to zero within few 100 s of exposure to the CO/O₂ mixture (Figure 4B), in line with the well known poisoning effect of CO on the surface of Pt and Pd catalysts [33], [34], [35]: when carbon monoxide is fed at room temperature, the molecules adsorb onto the catalyst surface and convert to CO₂ (as can be seen in Figure 4B). So at first, all sites on the surface are available for reactions. However, at this temperature the reaction sites are blocked over time from the very same CO, which strongly binds to the metal surface hence leading to a decrease of activity. Thus, to enable fast kinetics for a full and continuous carbon monoxide conversion in our Pt or Pd aerogel samples, temperatures up to 200°C are necessary.

As shown in an earlier report [21], when varying the volume fraction of the NCs we observe a threshold at 0.25 vol% for yielding monoliths with the same volume as the employed NCs colloid. Below this value the aerogel monoliths shrink due to insufficient building blocks and therefore a partly or even complete collapse of the network structure. However, the particle layer of the sheets become also thinner up to a point were only sheets with a particle monolayer could be observed. While concentrations of 0.0025 vol% and below yielded no usable monolith, concentrations of around 0.025 vol% do (although they suffer strong shrinkage of around 70% during drying). Figure 5 shows the comparison of Pd and Pt, which was the most active material and therefore the best choice to observe dependencies of morphology and catalytic performance. It turns out that lower volume fractions in the starting solutions have a negative impact on the catalytic performance of the aerogels. While the T₅₀ ignition is similar, the T₅₀ extinction temperature is around 20–30 K higher. The temperature decrease is related to the heat transfer within the aerogel and could be already observed in the system with Pt/Al₂O₃. Thinner sheets have a lower heat transfer and therefore a faster decrease in catalytic performances. The CO conversion, which can be seen for Pd below 125°C can be attributed to the oxidation of the surface ligands. Because of the low ligand amount, this can not be observed for Pt aerogels.

Fig. 5:

Comparison of the CO conversion of aerogels made from Pd and Pt NCs. The solid lines show aerogels made from a start solution with 0.25% vol% NCs volume fraction while the dashed lines show those made from 0.025%.

3.3 Thermal and catalytic stability

The thermal stability of the noble metal aerogels was investigated in the range within 25°C–200°C (as can be derived from Figure 6). All monoliths show shrinkage during the catalytic measurements. While Au and Ag aerogels show extremely low thermal stability and decreasing catalytic activity within the first runs, the catalytic activity of Pt is more stable but also decreases over five runs with an approximately 15 K higher T₅₀ (see Figure 6A). Instead, Pd shows a completely stable performance in the entire temperature range (r.t. −200°C) for several runs (Figure 6B). TEM images of cycled Pt samples (Figure 7) were measured, showing that the NCs in the aerogel-building sheets are sintered together. However, the aerogel samples have still their porous and polycrystalline nature with an increased domain size of ≤20 nm.

Fig. 6:

(A) CO conversion of Pt aerogels over six cycles (each cycle takes 120 min). The T₅₀ ignition temperature increases with each cycle showing a decreasing performance. (B) CO conversion of Pd aerogels over six cycles (each cycle takes 120 min). The T₅₀ ignition temperature does not change at all.

Fig. 7:

TEM characterization of a Pt aerogel (volume fraction 0.025%) before (A and C) and after catalytic measurements (B and D) in two different magnifications.