Nanomaterials engineering and applications in catalysis

Formation of Ag nanoplates promoted by H₂O₂

Through a systematic study on a typical chemical reduction reaction, we confirmed that the formation of Ag nanoplates could be promoted by the addition of a mild oxidant, H₂O₂. In the direct chemical reduction route, Ag nanoplates are obtained by reducing an aqueous solution of AgNO₃ with NaBH₄ in the presence of trisodium citrate (TSC), polyvinyl pyrrolidone (PVP), and H₂O₂ [35]. In the absence of H₂O₂, when PVP or citrate was used as the surfactant, only quasi-spherical Ag nanoparticles could be obtained, as evidenced by the characteristic absorption peak around 400 nm (Fig. 1a) as well as the TEM characterization (Fig. 1b). When H₂O₂ is introduced into the system, some anisotropic structures formed, suggested by the appearance of a peak in the longer wavelength range. As shown in Fig. 1a, a sharp peak around 450 nm appeared along with a shoulder at ~400 nm when the concentration of H₂O₂ is 5 mM. The sharp characteristic peak around 400 nm disappeared as the concentration of H₂O₂ was increased to 10 mM, implying the disappearance of spherical Ag nanoparticles. The weak shoulder around 380–420 nm could be attributed to the in-plane quadrupole resonance of Ag nanoplates. When the concentration of H₂O₂ was increased to 20 mM, the quadrupole resonance became more pronounced and the dipole resonance red-shifted to ~600 nm. The product is triangular Ag nanoplates with almost 100 % yield, as shown in Fig. 1d. The thickness of the Ag nanoplates could be varied from ~3 to ~6 nm by varying the concentration of sodium borohydride. The reaction process has been monitored using a UV/vis spectrometer to investigate the detailed formation process of the Ag nanoplates. As shown in Fig. 1c, only the characteristic peak of spherical Ag nanoparticles (~400 nm) could be observed in the early stage. The peak intensity increased quickly and reached the maximum within 5 s after the beginning of the formation of nanoparticles. The peak intensity at 400 nm then gradually dropped, indicating the consumption of spherical particles, while another peak at ~500 nm emerged and red-shifted to longer wavelengths, suggesting the formation and growth of anisotropic structures.

Fig. 1

(a) UV/vis spectra of Ag nanoparticles obtained by controlling the concentrations of H₂O₂; (b) TEM image showing the morphology of products prepared with only PVP as the surfactant and in the absence of H₂O₂; (c) in situ measurement of the formation of Ag nanoplates with a time interval of 7.5 s. The inset in (c) shows the UV/vis spectra of the initiation stage during the synthesis of Ag nanoplates with a time interval of 0.75 s; (d) TEM image showing the typical morphology of Ag nanoplates in the presence of H₂O₂. The inset in (d) shows a TEM image in which Ag nanoplates stand vertically upon their edges. Adapted with permission from ref. [7].

Based on the careful and systematic study, a plausible mechanism for the synthesis of Ag nanoplates has been proposed: Ag nanoparticles were first formed through the reduction of Ag⁺ ions by NaBH₄. Thanks to the Ag-citrate coordinating interaction and the presence of the powerful etchant, H₂O₂, the newly formed Ag nanoparticles contain many defects, including the twinned defects that favor the planar growth into plate shapes. In the previous paper, we have suspected that H₂O₂ can remove the relatively unstable nanoparticles at this stage, leaving only the most stable ones [7]. Recently, our further research showed that Ag nanoparticles can trigger the decomposition of H₂O₂ to release active oxygen, which favors the formation of twinned defects.

This improved understanding not only provides us with a highly reproducible synthetic tool but also allows us to study the role of each reagent. For example, contrary to previous reports in which citrate has been considered to be a “magic” component critically required for the formation of Ag nanoplates, our investigation clearly points out, for the first time, that citrate is not an irreplaceable component: it is required to stabilize the formed Ag nanoplate nuclei through preferential binding to the (111) facets but can be completely substituted by many other carboxyl compounds containing two carboxylate groups. We found that PVP is not required for obtaining high-quality Ag nanoplates and the exclusion of PVP could improve the reproducibility of the synthesis. In our further research, H₂O₂ has been successfully used to improve the yield of traditional seed-mediated growth process, further confirming the key role of H₂O₂ in the synthesis of Ag nanoplates [36].

Shape conversion from metallic Ag particles to Ag nanoplates

The standard potential in the peroxide-water couple is dependent on the pH value of the solution [37, 38]. In acidic solutions

H₂O₂+2H⁺+2e⁻→2H₂O (E°=1.763 V)

And in alkaline solutions

H₂O₂+2e⁻→2OH^–(E°=0.867 V)

Since the potentials under both acidic solution (E°=1.763 V) and alkaline solution (E°=0.867 V) are higher than that of Ag⁺/Ag (E°=0.7996 V), H₂O₂ can be used as an etchant to dissolve metallic Ag and produce Ag⁺ ions. As a result, it can be used to convert large metallic Ag materials to Ag nanoplates, which has not been reported in the literature [24]. Figure 2 shows two examples of the conversion from pre-formed Ag nanowires and nanoparticles to Ag nanoplates, respectively. When H₂O₂ was added to the pre-formed colloidal nanostructures, the solution gradually became colorless with the appearance of small bubbles due to the decomposition of H₂O₂. When NaBH₄ was then added to reduce Ag⁺ back to Ag⁰, the solution gradually became yellow, red, and blue, indicating the formation of Ag nanoplates. The successful conversion has been confirmed by both UV/vis spectra and TEM images (Fig. 2). From the point of view of synthesis, the use of metallic Ag as the source ensures a higher degree of reproducibility as there is less possibility of introducing disturbance such as the anions of the Ag salt to the reaction.

Fig. 2

UV/vis spectra and TEM images showing both (a–c) Ag nanowires with lengths of up to 10 μm and (d–f) Ag nanoparticles with irregular shapes can be converted to Ag nanoplates in the presence of H₂O₂. Adapted with permission from ref. [7].

Recently, Ekgasit and co-workers have found that H₂O₂ can not only act as an oxidant but also a reducing agent in the conversion of Ag nanoplates [39]. Under alkaline condition

H₂O₂+2Ag⁺+2HO⁻→2Ag+2H₂O+O₂ (E°=0.947 V)

As a result, starch-stabilized Ag nanospheres could be converted to Ag nanoplates by H₂O₂ without the addition of any other reducing agent. This result, again, confirmed the critical role of H₂O₂ in the synthesis of Ag nanoplates.

Precise tuning of Ag nanoplates within a wide range

Since the physiochemical properties of nanostructures are highly dependent on their morphologies, it is therefore important to find a way to precisely control the shape and size of nanostructures. We recently reported a two-step procedure for the controlled growth of Ag nanoplates with extremely high aspect ratios and consequently widely tunable plasmonic bands [40]. With the advantages of precise control and high reproducibility, this two-step procedure can conveniently produce nanoplates with high aspect ratios that have not been realized previously by using one-step reactions.

Ag nanoplates with size around 20–30 nm were first synthesized based on the previous mentioned method [7, 35], in which PVP and citrate were used as the surfactants to ensure the high-yield production of plate-like seeds. It is believed that PVP and citrate can selectively bind to the (100) and (111) facets of Ag nanoplates, respectively. The as-prepared small Ag nanoplates were then washed to remove PVP to expose the side facets and served as seeds for further growth. When excess citrate was added, the growth along the vertical axis was blocked, allowing only extensive growth along the lateral axis of the plates (Fig. 3a). Generally, both self-nucleation process and seed-mediated epitaxial growth can occur in the seeded-growth process. On the basis of the Gibbs–Thomson equation, a slow reaction rate as well as a low concentration of monomer is normally favorable for the seed-mediated epitaxial growth, while a fast reaction rate is favorable for the self-nucleation process [41–44]. To eliminate the self-nucleation process, the reaction rate was slowed down by using a mild reducing agent, ascorbic acid, and pre-treating Ag ions with citrate ions to form Ag-citrate complexes [45]. The as-prepared complex can be reduced by ascorbic acid and release more citrate ions, which can help maintain anisotropic growth without introducing any new ligand.

Fig. 3

(a) Schematic illustration of the anisotropic seeded growth of Ag nanoplates based on selective ligand adhesion; TEM image (b) and SEM images (c–g) showing the evolution of Ag nanoplates during the stepwise growth process: from original seeds (b) after removal of PVP by centrifuging and washing to larger Ag nanoplates after (c) one, (d) two, (e) four, (f) five, and (g) seven cycles of seeded growth. Adapted with permission from ref. [40].

As shown in Fig. 3b, the original Ag seeds were triangular nanoplates with a mean size around 25 nm. The edge length of the Ag nanoplate can gradually increase to ~130 nm after one cycle of growth (Fig. 3c). The growth cycle could be repeated many times and the edge length of Ag nanoplates could be readily increased to ~4 μm (Fig. 3d–g). An interesting “size distribution focusing” effect has been observed during the process. The initial Ag seeds have a broad size distribution, which become highly uniform after extended growth. The size distribution focusing effect could be attributed to the fact that the smaller nanoparticles grow more rapidly than the larger ones as driven by the higher surface energy. In addition, larger particles have a slower growth rate even when the deposition rates of Ag atoms are the same due to the greater ratio of volume versus size for large particles, meaning more Ag is needed to produce a net edge length increase. Thanks to the effective blocking effect of citrate ions and the formation of Ag-citrate complexes, only small change in the thickness of Ag nanoplates has been observed. As a result, Ag nanoplates with extremely high aspect ratio (edge length/thickness >400) have been obtained. And the plasmon band has been readily tuned from the visible to the infrared (IR) spectrum by controlling the aspect ratio of the nanoplates. Due to the fact that blood and tissue are most transparent in the region of near-IR (NIR) range, the ability to cover the entire near-IR (NIR) spectrum makes the as-prepared Ag nanoplates ideal candidates for biomedical applications such as photothermal cancer therapy [46–49].

Since the complexation can significantly suppress the self-nucleation process, this idea has been further extended to the Au nanoparticle system. A one-step seeded growth has been developed by the Yin group to obtain Au nanoparticles with widely tuned size [43]. To effectively suppress the self-nucleation process, a growth solution was first prepared by mixing HAuCl₄, KI solution, PVP, and ascorbic acid solution. In this solution, HAuCl₄ is the Au source and ascorbic acid is the reducing agent. KI was used as a strong coordinating ligand to Au ions to decrease the reduction potential of Au salt, eliminating the burst self-nucleation process [42, 50]. At the same time, PVP has been employed as a bi-surfactant to stabilize the atomic Au monomers and further avoid the self-nucleation process [51]. PVP can also serve as the capping agent to prevent inter-particle agglomeration. The as-prepared growth solution can stay stable even at considerably high concentration. Upon injecting small Au nanoparticles as the seeds, the growth process occurs immediately. Through this seeded growth method, monodispersed Au nanoparticles with tunable size from ~10 to ~200 nm could be easily obtained. This new approach shines some light on the effective suppression of self-nucleation regardless of the seed/Au precursor ratio, which might be able to inspire some new ideas in other systems.

Monitoring the shape evolution of Ag nanoplates using a Au marker

One interesting phenomenon we observed in the seeded-growth process is the shape change of Ag nanoplates. As shown in Fig. 3, at the earlier stage of seeded growth, the corners of Ag nanoplates are extensively rounded so that the plates appear circular in SEM images. As the size increases, sharp corners are developed, then gradually truncated, and finally producing a mixture of triangular and hexagonal nanoplates. The similar shape transition phenomena have also been observed recently [52, 53]. Xia et al. have attributed it to the possibility that a truncated structure might be more thermodynamically stable, but that the growth was kinetically controlled and thus ultimately the plates returned to their original triangular shapes. However, no solid evidence has been provided. Although the crystal structure of the 2D Ag nanoplates and Au nanoplates seems to be simple, many different model crystal structures have been proposed. For example, the 1/3{422} reflections have been observed in the electron diffraction (ED) pattern of Ag nanoplates and Au nanoplates. However, for a perfect fcc structure, the 1/3{422} reflection should be forbidden. Pileni and co-workers have attributed it to parallel stacking faults in the <111> direction [54] and further pointed out that these stacking faults should be responsible for the formation of plate structures by providing low-energy reentrant grooves which energetically favor lateral crystal growth. Later, Rocha and Zanchet conducted a careful study by using both high-resolution transmission electron microscopy (HRTEM) and X-ray diffraction (XRD). They pointed out that the internal structure of Ag nanoplates is indeed very complex, containing many twins and stacking faults [55]. They found that local hcp regions, rather than fcc structure, could form due to the existence of the planar defects in the <111> direction. The formation of hcp regions could explain the existence of forbidden 2.50 Å fringes that are observed in <111> orientated nanoplates. The formation mechanism of the hcp layer has been further elucidated by Kelly et al., and it has been used to explain the evolution of Ag nanoplates [56]. It is ascertained that Ag nanoplates are comprised of two fcc regions of different thicknesses, sandwiching an hcp layer originating from a series of internal stacking faults. In their devised crystal structure model, each nanoplate edge consists of a {111} facet and a {100} facet of different sizes. As a result, there are six edges in a truncated triangular nanoplates, half featuring a larger {111} facet and the other half with a dominant {100} facet. The formation of a triangular structure bonded by {111} dominated edges could be explained by the preferential deposition on the less thermodynamically stable {100} dominated edges.

To further elucidate the seeded-growth process and identify the crystal structure of Ag nanoplates, we designed a “marker” experiment to monitor the shape change of Ag nanoplates during the process [57]. In the experiment, Ag nanoplates were first synthesized based on the direct chemical reduction method [7], in which PVP was eliminated to avoid any disturbance. A galvanic replacement process was then applied to the Ag nanoplates by using Au cations [58]. As shown in Fig. 4b, the as-prepared seeds are nanoframes. Because the lattice mismatch between Au and Ag is negligible while the difference in atomic number is so significant, Au can be a good candidate for both seeding and serving as the marker to outline the boundaries of the Ag nanoplates.

Fig. 4

(a) Schematic illustration of the structural change of Ag nanoplates during seeded growth. Preferential Ag deposition on the {100} dominated edges of the seeds leads to a rounded or hexagonal intermediate with a mix of {100} and {111} dominated edges, after which further Ag deposition produces a larger triangular plate with {111} dominated edges; (b–g) step-by-step growth of Ag nanoplates using Au nanoframes as seeds. (b) Au nanoframe seeds and frame/plates after the addition of (c) 0.3 μmol, (d) 3 μmol, (e) 4.1 μmol, (f) 7.5 μmol, and (g) 10.75 μmol of Ag nitrate. Adapted with permission from ref. [57].

As shown in Fig. 4c, the nanoframes can serve as seeds for the seeded growth process. A backfilling process was first observed when the seeded growth process started, followed by the normal growth into large nanoplates (Fig. 4d–g). Mirkin et al. have pointed out that the backfilling process could be attributed to the high-energy facets in the nanoframe interior [58]. The outward growth happened in the same manner as Ag nanoplates: It has a similar shape transition and ultimately returns to a triangular morphology (Fig. 4d–g). One interesting phenomenon is that there is a preference for deposition on the edges of the nanoframe/nanoplate with little to none occurring at the corners throughout the various stages of outward growth, causing the triangular to circular/hexagonal transition. The shape returns to triangular when more Ag is deposited, resulting in the formation of new corners at the deposition sites (Fig. 4f). The majority of the obtained product contains an embedded frame pointed 180° relative to the original orientation of the Ag nanoplate (Fig. 4f). No additional shape change could be observed even when more Ag atoms were deposited to make larger nanoplates (Fig. 4g). It can be concluded from the observation that the exclusive deposition of Ag atoms on the nanoplate edges, rather than the reaction kinetics, induced the shape transition.

This transition process could be well explained by using Kelly et al.’s crystal structure model, as illustrated in Fig. 4a. The edges of the initial Ag nanoplates are dominated by {100} facets, formed during the rapid reduction process. Because a weaker reducing agent, ascorbic acid, is used for the seeded growth process and the concentration of Ag ions is deliberately low, the thermodynamic consideration becomes more important, leading to selective deposition on {100} dominated sides and causing the observed shape transition to a structure with {111} dominated edges.

Stabilizing Ag nanoplates through the Ag@Au core/shell nanostructure

As we mentioned above, the poor chemical and structural stability of Ag nanostructure has been the major concern in their practical applications. The Yin group recently developed a Au coating strategy that can not only enhance the stability of Ag nanoplates but also keep their excellent plasmonic property [50]. It is well known that galvanic replacement would happen when Ag nanoparticles are exposed to Au salt due to the different oxidative potentials. As a result, the key to uniformly depositing a thin layer of Au on the Ag nanoplates is to minimize the galvanic replacement process.

Similar to the above-mentioned one-step seeded growth of Au nanoparticles, an aqueous solution containing HAuCl₄, KI, and PVP was first prepared as a growth solution, which was then slowly added into a mixture containing PVP, diethylamine, ascorbic acid, and Ag nanoplates. In the growth solution, thanks to the complexation between Au ions and I^– ions, the reduction potential of Au salt has been significantly decreased to +0.56 V (vs. standard hydrogen electrode, SHE), which is much lower than those for Au ions (E°(Au³⁺)=+1.52, E°(AuCl₄⁻)=+0.93 V, vs. SHE). The addition of diethylamine can ensure homogeneous Au deposition on the surface of Ag nanoplates, due to the unselectively binding of diethylamine on the surface of Ag nanoplates. Figure 5a shows a typical TEM image of the obtained Ag@Au core/shell nanoplates. Although some nanoplates contain small voids on the surface that are caused by galvanic replacement, the original triangular/hexagonal shapes are well maintained, suggesting that the galvanic replacement has been significantly suppressed. Notably, the plasmonic property has been well maintained, as shown in Fig. 5b. The intensity of the plasmon band of the final Ag@Au core/shell nanoplates is very close to that of the original Ag nanoplates. A slight blue-shift of the plasmonic peak has been observed, which can be attributed to the decrease of the aspect ratio. The core/shell structure has been further confirmed by the high-resolution energy-dispersive X-ray spectroscopy (EDS). Figure 5c shows a typical EDS line scan across the surface of a single nanoplate. Both Ag and Au have been detected in the bulk area of the core/shell nanoplates. About 0.5 nm from the edge of the nanoplates, the atomic percentage of Ag drops to zero and that of Au rises to 100 %, which is consistent with the fact that the edge Au atoms are deposited through reduction of Au cations instead of partial replacement of Ag.

Fig. 5

(a) TEM image of the Ag@Au core/shell nanoplates; (b) UV/vis/NIR spectra of the original Ag nanoplates and Ag@Au core/shell nanostructures prepared by adding various amounts of Au growth solution; (c) atomic percentage profile of a Ag@Au nanoplate calculated from the relative counts in the EDS line scan indicated by the arrow in the STEM image below; (d–i) stability of the Ag@Au nanoplates in (d) phosphate buffer solution (10 mm, pH 7.4, PVP 0.5 %), (e) NaCl solution (20 mm, PVP 0.5 %), (f) PBS (10 mm, NaCl 150 mm, pH 7.4, PVP 0.5 %), and (g) H₂O₂ solution (2.1 %), monitored by UV/vis/NIR spectrophotometry. A TEM image of the Ag@Au nanoplates after treatment in H₂O₂ for 2 h is shown in the inset to (g). Stability of (h) pristine Ag nanoplates and (i) MHA-stabilized Ag nanoplates in H₂O₂ (2.1 %). Adapted with permission from ref. [50].

Thanks to the protection of Au layer, the Ag@Au core/shell structure shows excellent stability against various etchants. As shown in Fig. 5d–f, the position and the intensity of the in-plane dipole plasmon resonance bands remained essentially unchanged for 4 days or longer when the nanoplates were dispersed in (D) a phosphate buffer solution, (E) a NaCl solution, and (F) a phosphate-buffered saline (PBS) solution, respectively. Although a slight decrease in peak intensity was observed in the concentrated NaCl solution treated case, it could be primarily attributed to the aggregation of the nanoplates in a solution of high ionic strength. The as-obtained core/shell structure shows outstanding stability against H₂O₂, a strong oxidant that can easily destroy Ag nanoplates protected by conventional methods (Fig. 3g). The inset TEM image in Fig. 3g shows the morphology of Ag@Au nanoplates stored in H₂O₂ for 2 h, which is quite similar to the original Ag nanoplates, further confirming their excellent stability against strong etching agents. The protecting effect can be further proved by the control experiments, in which Ag nanoplates without any protection were quickly etched by H₂O₂ (Fig. 3h), as evidenced by a significant shift in the peak position and a dramatic decrease in the intensity over a relatively short period. The protecting effect from thiols, such as 16-mercaptohexadecanoic acid (MHA), is also very limited. As shown in Fig. 3i, the Ag–MHA nanoplates gradually degraded in the presence of H₂O₂.

Stabilizing metal nanoparticles through encapsulation within a porous shell

In the physical approach, encapsulating metal nanoparticle with a porous shell to prevent the direct contact from each other is one of the major strategies [110–114]. Figure 6b shows the typical TEM image of a model catalysts, in which Pt nanoparticles with size around 2–3 nm were supported on the surface of ~100 nm silica beads. After calcined at 1075 K, only several large crystalline particles with size ~10–20 nm can be observed (Fig. 6c), confirming the occurrence of a serious sintering process. When the SiO₂/Pt composite was coated with a thin layer of silica, great improvement in thermal stability has been achieved. As shown in Fig. 6d–f, no obvious particle size increase can be observed after being treated at 1075 K, implying successful suppressing of sintering.

Fig. 6

Illustration of the use of mesoporous silica layers for protection against sintering of dispersed metal nanoparticles [125]. As shown schematically in panel (a), the original small Pt nanoparticles supported on SiO₂ beads [panel (b)] coalesce into a few larger structures [panel (c)] after calcination at 1075 K. The corresponding TEM images from a catalyst coated with a mesoporous SiO₂ shell, shown in panels (e) and (f), prove the enhanced stability afforded by such treatment [panel (d)]. Adapted with permission from ref. [125].

Although the encapsulation of metal nanoparticles within oxide shells could significantly suppress the sintering, dense coating will block the accessibility of the active sites and make them useless. It is thus desirable to have a porous shell coating rather than a dense one. We have developed a new concept, called “surface-protected etching”, to precisely convert dense oxide shell to porous shell and explored their potential applications in catalysis [66, 115–124]. In a typical surface-protected etching process, solid oxide shells were first protected by a layer of polymeric ligands, followed by a preferential etching of material from the interior of the oxide shells. Due to the protection of polymeric ligands, the original size of the oxide shell could be well maintained, while the selective etching at the interior produces porous or even hollow structures. To transfer dense silica shells to porous ones, we used PVP as the polymeric ligand and NaOH as the etchant. By varying the etching time, the porosity of silica shell could be readily tuned, resulting in well-controlled catalytic activity. The hydrogenation of cis-2-butene using Pt nanoparticles as the catalyst has been used as the model reaction, as shown in Fig. 7. When SiO₂/Pt/SiO₂ composites without etching were used as the catalyst, a slow reaction rate, indicated by the slope of the traces at time zero, has been observed, which can be attributed to the microporous nature of sol-gel obtained silica shell. The reaction rate can be significantly higher when the samples were treated with a surface-protected etching process (by a factor of almost an order of magnitude).

Fig. 7

Illustration of the use of the surface-protected etching procedure to regain the activity of encapsulated catalysts. The kinetics of hydrogenation of cis-2-butene, in the form of the accumulation of the butane product in the batch reactor as a function of reaction time, is reported for three catalysts, namely, for 10-nm-mesoporous-silica/1-wt%-Pt/silica-bead catalysts as prepared (bottom TEM, blue) and after 40 (center TEM, green) and 60 (top TEM, red) min of etching [66]. Pt re-exposure and full catalytic activity are regained upon the 60 min etching treatment. Adapted with permission from ref. [66].