Advanced bimetallic In–Cu/Ag/Au nanostructures via microemulsion-based reaction

Abstract

Bimetallic nanomaterials and nanostructures constituted of the coinage metals (Cu, Ag, Au) and indium with elaborate compositions and structures are realized via a microemulsion-based approach. In detail, this comprises Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles, In-Ag Janushead-like nanoparticles, Ag⁰ hollow spheres, Ag₃In@In core@shell nanoparticles, Au@AuIn₂@In core@shell-A@shell-B nanoparticles and AuIn₂ nanoparticles. To obtain these advanced architectures, two approaches are applied: (1) In⁰ nanoparticles—pre-synthesized in a microemulsion—were reacted in a follow-up reaction with CuCl₂·2H₂O, AgNO₃ or KAuCl₄; (2) simultaneous co-reduction of InCl₃·4H₂O together with CuCl₂·2H₂O, AgNO₃ or KAuCl₄ in a microemulsion. Characterization of the resulting advanced structures and compositions requires elaborate electron microscopy techniques, combined with energy dispersive X-ray spectroscopy for chemical analyses of single nanoparticles as well as X-ray powder diffraction and optical spectroscopy. The versatility of the experimental approach toward complex nanoparticle architectures is related to a precise control and fine-tuning of the experimental conditions. The resulting tool kit of In–Cu/In–Ag/In–Au-based bimetallic and intermetallic nanomaterials and, in general, of nanostructured metal architectures with such variability and complexity have not yet been described.

---

1. Introduction

Bimetallic and intermetallic nanomaterials with elaborate composition, structure and morphology are of general interest to fundamental science and of rapidly increasing relevance to technical application.¹ In general, this is related to the unusual properties of nanoscaled metals, which can differ significantly from the bulk (e.g. due to small size, large surface, quantum-size effects).² Thus, the chemical reactivity of nanoscaled metals can be dramatically increased, too. While a high reactivity can be disadvantageous (i.e. fast oxidation/hydrolysis, low temperature stability, rapid agglomeration/aggregation), reactive nanoparticles can be highly beneficial with regard to chemical synthesis, too.³ In particular, the high reactivity of nanoscaled metals can be used to initiate chemical reactions at ambient conditions and under kinetic control, which may allow the realization of new compositions, structures and morphologies that are far beyond the thermodynamically most stable phases, as they are typically given by the relevant phase diagrams.

In particular, nanoscaled less-noble metals can represent attractive candidates for use as reactive precursors, as they could serve as pure element sources for various chemical conversions to bimetallic/intermetallic compounds. This is even more interesting since a transfer of standard approaches known from bulk metals (e.g. ball milling,⁴ melting techniques⁵) to the nanoscale is not straightforward. Dedicated strategies for the realization of nanoscaled bimetals and intermetallics, on the other hand, are currently limited.⁶ Shape- and composition-controlled synthesis of multi-metal nanoparticles has been largely limited to those metals that have similar electrochemical potential and reactivity.^6,7 The systems Fe–Pt,⁸ Sn–Pt/Pb–Pt,⁹ or Ni–Au¹⁰ represent selected examples of metals exhibiting a significantly different electrochemical potential. Intermetallic nanoparticles containing In⁰—to the best of our knowledge—have only been reported for InPt₃.¹¹

Most intriguingly, Schaak et al. have prepared Sn⁰ nanoparticles via a polyol-mediated synthesis and reacted these as a sacrificial template (i.e. reduction of a second metal under partial oxidation of the pre-synthesized Sn⁰) with Co²⁺, Ni²⁺ or Pd²⁺, to gain intermetallic compounds.¹² The authors also obtained AuCu and AuCu₃ nanoparticles from thermal annealing of bimetallic Au⁰/Cu⁰ aggregates.¹³ AuCu and AuCu₃ have been alternatively prepared via seed-based diffusion of reactive copper species into pre-synthesised Au⁰ nanocrystals.¹⁴ Recently, Armbrüster et al. synthesized intermetallic Ga–Pd nanoparticles based on pre-synthesized Pd⁰ nanoparticles.¹⁵ Reacting pre-synthesized metal nanoparticles can also result in core@shell structures, such as Ag@Au/Au@Ag,¹⁶ Pd@Au@FePt,¹⁷ or Au@Pb@Pt.¹⁸ The latter, more elaborate structures are however limited to a few examples. Note that most of the above reactions require an inert-gas atmosphere (N₂, Ar) and elevated temperatures (>100 °C). In general, the synthetic strategies summarized above are most often limited to a single composition, structure and/or morphology.

To explore the potential of reactive, less-noble metals for conversion reactions and the formation of nanoscaled intermetallics, we have selected indium (In⁰) nanoparticles as a precursor. With its electrochemical potential (E₀(bulk In⁰) = −0.34 V)¹⁹ and its further increased reactivity as a nanoparticle, In⁰ may serve as a promising precursor for various conversion reactions, without being too reactive to control these reactions. As a case study, we have reacted indium with the coinage metal salts (i.e. CuCl₂·2H₂O, AgNO₃, KAuCl₄) based on simple microemulsion techniques.²⁰ The underlying mild, ambient conditions (i.e. water as the solvent; temperature ≤50 °C) enabled us to fine-tune the relevant experimental conditions far beyond the thermodynamically most stable phases. This strategy—namely follow-up and co-reduction reactions—allows us to precisely adjust the composition, structure and morphology of the resulting bimetallic and intermetallic nanomaterials, including advanced core–shell nanostructures, nanoscale hollow spheres and bimetallic Janushead-like nanoparticles. A synthetic strategy, as well as the resulting advanced architectures, are firstly reported and lead to a tool kit of bimetallic and intermetallic nanoparticles that may also be relevant when aiming at new materials for use in electronics,²¹ magnetism,⁸ optics²² and catalysis.²³

2. Materials and methods

2.1 Pre-synthesized In⁰ nanoparticles

Microemulsion-based synthesis. A w/o (water-in-oil) microemulsion was established by mixing 25.0 mL n-dodecane as the non-polar phase, 5.0 mL n-hexanol as the co-surfactant and 0.91 g CTAB as the surfactant, with a solution of 29.3 mg (0.1 mmol) InCl₃·4H₂O in 1.0 mL demineralized water as the polar phase. This micellar system was stirred at ambient temperature under dynamic nitrogen purging to obtain a transparent, colorless microemulsion. The reduction of In³⁺ was induced by addition of 18.9 mg (0.5 mmol) NaBH₄ dissolved in 1.0 mL demineralized water under continuous stirring. Thereafter, the micellar system turned grey within a few seconds. After an additional 10 min, the microemulsion was destabilized by the addition of 10.0 mL diethylene glycol (DEG). This measure initiated a phase separation with accumulation of the In⁰ nanoparticles in the bottom DEG phase. Thereafter, the nanoparticles were separated by removing the upper non-polar phase containing most of the surfactant, and collected by centrifugation of the DEG phase. The solid remnant was washed twice by sequential redispersion/centrifugation in/from ethanol. Colloidally stable suspensions were obtained by redispersion of the washed In⁰ nanoparticles in ethanol; powder samples were yielded by drying the remnant for 12 h at ambient temperature in air.

Phase-transfer reaction. In order to avoid misinterpretation due to eventual agglomeration of the nanoparticles during the centrifugation/redispersion procedure, all nanoparticles (except for In⁰ precursor nanoparticles and nanoscale Ag⁰ hollow spheres) were alternatively purified via a two-step phase transfer. First, the nanoparticles were separated from the micellar system as described above via the addition of DEG. After removing the upper non-polar phase, the DEG bottom-phase containing the nanoparticles was mixed with 1 ml oleylamine. The latter is known as an efficient non-polar surface capping and has been already described for various types of nanomaterials. Subsequent to ultrasound treatment, 5 mL dodecane was added to the DEG–oleylamine suspension under vigorous stirring. This measure resulted in a second phase transfer, with the nanoparticles now located in the upper dodecane phase. This dodecane phase was used for all sample characterization of the nanoscaled intermetallic compounds via electron microscopy and UV-Vis spectroscopy.

2.2 Follow-up approach

Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles. These nanoparticles were prepared via a follow-up reaction of the pre-synthesized In⁰ precursor nanoparticles. Thus, a w/o-microemulsion was established as described above and heated to 50 °C. After the synthesis of nanoscaled In⁰, a solution of 17.0 mg (0.1 mmol) CuCl₂·2H₂O in 0.5 mL demineralized water was added to the micellar system. The color of the system instantaneously changed from grey to anthracite. The reaction was terminated thereafter upon addition of 10.0 mL DEG. During phase separation the nanoparticles accumulated in the bottom DEG-phase and separated by centrifugation. The solid remnant was washed twice by sequential redispersion/centrifugation in/from ethanol. Colloidally stable suspensions and powders of In@CuIn@Cu₁₁In₉ were obtained as described for the In⁰ nanoparticles.

In–Ag Janushead-like nanoparticles. These nanoparticles were also prepared via a follow-up reaction of the pre-synthesized In⁰ precursor nanoparticles. In general, the In⁰ nanoparticles were prepared as described above, but using cetyltrimethylammonium tosylate (CTATos) as a surfactant instead of CTAB, to avoid any precipitation of AgBr. After the nucleation of nanoscaled In⁰—as indicated by the grey appearance of the microemulsion—a solution of 17.0 mg (0.1 mmol) AgNO₃ in 0.5 mL demineralized water was rapidly added. Thereafter, the color of the micellar system turned black. After an additional 10 min, the In–Ag Janushead-like nanoparticles were separated from the micellar system by the addition of 10.0 mL DEG and purified as described above.

Ag⁰ hollow spheres. To obtain nanoscale Ag⁰ hollow spheres, the In⁰ precursor nanoparticles were prepared and purified as described above, followed by resuspending In⁰ in 20 mL ethanol. 5.0 mL of this suspension were mixed with a solution of 68.0 mg (0.4 mmol, fourfold excess) AgNO₃ in 5.0 mL ethanol and stirred for 6 days at ambient temperature. Already after the first day, the grayish suspension exhibited a slight yellow appearance. After 6 days, the suspension had a clear yellow color. After centrifugation, the resulting Ag⁰ hollow spheres were redispersed in ethanol with 0.05 wt% polyvinylpyrrolidone (PVP) and centrifuged again. The solid remnant was purified as described above.

Au@AuIn₂@In core@shell-A@shell-B nanoparticles. These nanoparticles were prepared via a follow-up reaction of the In⁰ precursor nanoparticles as well. A w/o-microemulsion was established as described above and heated to 50 °C. The In⁰ nanoparticles were nucleated and stirred for one minute. Thereafter, a solution of 38.0 mg (0.1 mmol) KAuCl₄ in 1.0 mL demineralized water was added to the micellar system. The color of the microemulsion rapidly turned from grey to black. After cooling to room temperature, the reaction was terminated upon the addition of 10.0 mL DEG. During phase separation, the nanoparticles accumulated in the bottom DEG-phase. The Au@AuIn₂@In core@shell-A@shell-B nanoparticles were separated by centrifugation and redispersed in ethanol with 0.05 wt% PVP in order to stabilize the nanoparticles. The solid remnant was purified as described above.

2.3 Co-reduction approach

Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles. These nanoparticles were prepared via co-reduction of In³⁺ and Cu²⁺ in a w/o-microemulsion. Therefore, 25.0 mL n-dodecane as the non-polar phase, 5.0 mL n-hexanol as the co-surfactant and 0.91 g cetyltrimethylammonium bromide (CTAB) as the surfactant were mixed with a solution of 29.3 mg (0.1 mmol) InCl₃·4H₂O and 17.0 mg (0.1 mmol) CuCl₂ in 1.0 mL demineralized water as the polar phase. The micellar system was stirred at room temperature under dynamic nitrogen purging to obtain a transparent, colorless microemulsion. Simultaneously, 20.0 mg (0.5 mmol) NaBH₄ was dissolved in 0.5 mL of a 0.025 M NaOH solution and rapidly added to the micellar system. After 10 min the Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles were separated from the micellar system by adding 10.0 mL DEG and purified as described above.

Ag₃In@In core–shell nanoparticles. Ag₃In@In nanoparticles were synthesized via co-reduction of In³⁺ and Ag⁺ in a w/o-microemulsion, which was established as described above. A solution of 30.0 mg (0.1 mmol) In(NO₃)₃·xH₂O and 17.0 mg (0.1 mmol) AgNO₃ in 1.0 mL demineralized water was used as the polar phase. The micellar system was stirred at 50 °C under dynamic nitrogen purging to obtain a transparent, colorless microemulsion. Simultaneously, 20.0 mg (0.5 mmol) NaBH₄ was dissolved in 0.5 mL of a 0.025 M NaOH solution and rapidly added to the micellar system. After 10 min the Ag₃In@In core–shell nanoparticles were separated from the micellar system by adding 10.0 mL DEG and purified as described above.

AuIn₂ nanoparticles. These nanoparticles were also synthesized via co-reduction of In³⁺ and Au³⁺ in a w/o-microemulsion, which was established as described above. A solution of 29.3 mg (0.1 mmol) InCl₃·4H₂O and 18.9 mg (0.05 mmol) KAuCl₄ in 1.0 mL demineralized water was used as the polar phase. This micellar system was stirred at 50 °C under dynamic nitrogen purging to obtain a transparent, orange-red microemulsion. Simultaneously, 20.0 mg (0.5 mmol) NaBH₄ was dissolved in 0.5 mL of a 0.025 M NaOH solution and injected into the microemulsion. After 10 min, the AuIn₂ nanoparticles were separated from the micellar system by adding 10.0 mL DEG and purified as described above.

2.4 Energy-dispersive X-ray spectroscopy (EDXS) and analysis of concentration profiles

Although advanced electron microscopy (including STEM, HAADF-STEM, HRTEM, SAED) is essential to characterize the as-prepared bimetallic nanostructures described here, the underlying equipment and techniques can be nevertheless denoted as standard techniques—and are therefore summarized in the ESI.† EDXS profile analysis, as used in the following to elucidate the complex composition, structure and morphology of the as-prepared elaborate nanoparticle architectures, is however new and far beyond standard characterization. Therefore, the underlying procedure is explained here in detail.

EDXS line profiles were recorded with an FEI Titan³ 80–300 microscope by applying a drift-correction routine via cross correlation of several images, which yields a local precision better than 1.0 nm. The drift-corrected EDXS line profiles were taken with a probe diameter of 0.5 nm and a distance of about 1.0 nm between two measuring points. With respect to a quantification of the EDX spectra from core–shell nanoparticles, one has to keep in mind that only compositions are obtained that are averaged along the electron-beam direction. In concrete, the whole volume along the electron trajectory contributes to the detected X-ray signal. To evaluate the composition of different shells of a core–shell nanoparticle, a procedure was developed which is outlined in the following.

The analysis is based on the concentration profiles of different chemical elements within a nanoparticle determined from EDX spectra measured along a line scan that passes through its centre. The distance of 1 nm between two successive measured EDX spectra in the concentration profile suggests a division of the nanoparticle into sub-shells (SS_n with n = 1,2,3…) with an equal thickness of 1 nm. Fig. 1 shows the 2-dimensional cross-section of an ideal spherical nanoparticle of radius R and the first two sub-shells represented by different colors. Here, the z-axis is parallel to the electron-beam direction, the x-axis represents the line-scan direction, and O is the nanoparticle centre. If c_Aⁱ is the concentration of the chemical element i measured at point A, one can consider that it also represents the average concentration of element i within the first sub-shell c_Aⁱ = cⁱ_SS1. Then, the concentration c_Bⁱ measured at the next point B can be described as the weighted concentration of element i within the first and the second sub-shell. The weights w₁ and w₂ are proportional to the corresponding thickness of each sub-shell, measured along the vertical line that passes through B parallel to the z-axis (Fig. 1), i.e. to the sub-shell thicknesses passed by the electron beam. Thus, the average concentration of element i in the second sub-shell cⁱ_SS2 can be calculated using the cⁱ_SS1 value and the following relation:

with d₁ and d₂ given by:

Fig. 1 Cross-section of an ideal spherical nanoparticle (R: radius; z-axis parallel to electron-beam; x-axis corresponds to line-scan direction; with two sub-shells SS1 and SS2, both 1 nm in diameter, indicated by red/blue color.

Similarly, one can calculate the average concentrations of all chemical elements in different sub-shells that are contained in the nanoparticle. As a result, the average composition of the nanoparticle shell/core can be calculated as the normalized concentration of all chemical elements found within different sub-shells that form the core/shell of the nanoparticle.

3. Results and discussion

3.1 Strategy of synthesis: follow-up and co-reduction approach

All the following bimetallic nanoparticles in the systems In–Cu, In–Ag and In–Au are realized based on microemulsion techniques. To obtain the resulting elaborate architectures, two slightly different approaches were applied (Table 1, Fig. 2):

Fig. 2 Scheme displaying the follow-up and co-reduction approach for the microemulsion-based reaction of In⁰ and the coinage metals as well as photos of the resulting nanoparticles: (left to right) In⁰ precursor; Cu₁₁In₉@CuIn@In; Ag hollow spheres; Ag₃In@In; Au@AuIn₂@In; AuIn₂ (all suspensions in ethanol).

Table 1 Summary of experimental data and composition for microemulsion-based In–Cu/Ag/Au nanostructures

Type of Precursor		Type of approach	Molar ratio In : M : BH4^−	Resulting composition	Size^a/nm
In	M
a Average particle diameter according to BF-STEM analysis.
InCl3·4H2O	—	—	1 : 0 : 1	In precursor	12(2)
In^0-NP	CuCl2	follow-up	1 : 1 : 5	Cu11In9@CuIn@In core@shell	14(6)
InCl3·4H2O	CuCl2	co-reduction	1 : 1 : 5	Cu11In9@CuIn@In core@shell	6(2)
In^0-NP	AgNO3	follow-up	1 : 1 : 5	Janushead-like In–Ag	13(3)
In^0-NP	AgNO3	follow-up	1 : 4 : 5	Ag hollow spheres	11(2)
InCl3·4H2O	AgNO3	co-reduction	1 : 1 : 5	Ag3In@In core@shell	9(3)
In^0-NP	KAuCl4	follow-up	1 : 1 : 5	Au@AuIn2@In core@shell	14(5)
InCl3·4H2O	KAuCl4	co-reduction	1 : 0.5 : 5	AuIn2	4(2)

---

1. In⁰ precursor nanoparticles—pre-synthesized in a microemulsion—were reacted in a follow-up reaction with the coinage metal salts (i.e. CuCl₂·2H₂O, AgNO₃, KAuCl₄).

2. Co-reduction of InCl₃·4H₂O simultaneously with the coinage metal salt (i.e. CuCl₂·2H₂O, AgNO₃, KAuCl₄) in a microemulsion.

In the follow-up approach, In⁰ nanoparticles were first prepared by reduction of InCl₃·4H₂O with NaBH₄ in a microemulsion (see section 3.2). Subsequently, these pre-synthesized “precursor In⁰ nanoparticles” were reacted with the much more noble coinage metals (i.e. CuCl₂·2H₂O, AgNO₃, KAuCl₄) to obtain bimetallic In–Cu, In–Ag and In–Au nanoparticles. The relevant coinage metal salt was added to the micellar system subsequent to the completed nucleation of In⁰. This point in time is qualitatively indicated by an ebbing of the hydrogen gas formation and the occurrence of a stable grey color of the suspension. Now, the coinage metal salt solution attains micelles that already contain In⁰ nanoparticles as well as excess NaBH₄. Therefore, the coinage metal cations were instantaneously reduced to the elemental metal as well. Due to the spatially limited volume of the micelles, the In⁰ nanoparticles serve as a seed for the nucleation of the coinage metal. Depending on the interdiffusion of metals and the thermodynamics of alloy formation, nanoscaled intermetallic compounds or non-alloyed bimetallic nanoparticles—including core–shell and Janushead-like architectures—are obtained.

For this follow-up approach, the procedure of adding the coinage metal precursor to the micellar system turned out to be a key factor in controlling the reaction. If the coinage metal precursor was added too early, reduction and nucleation of indium might not be complete, resulting in quite a high excess of non-consumed NaBH₄. As a consequence, non-controlled simultaneous reduction and nucleation of indium, as well as of the coinage metal, occurred and led to nanoparticles with inhomogeneous size and composition. If—by contrast—the coinage metal precursor was added too late, all NaBH₄ was consumed by the aqueous medium due to hydrogen formation. This resulted in a very slow and incomplete reduction of the coinage metal. In such a case, reduction of the noble metal is driven by slow re-oxidation of the initial In⁰ nanoparticles—leading to formation of the elemental coinage metal only. Altogether, the course of the reaction has to be adjusted with regard to each relevant coinage metal. In detail, bimetallic nanoparticles containing copper and silver require a fast sequential injection of the coinage metal precursor as soon as the nucleation of indium is finished. This guarantees an appropriate amount of NaBH₄ to remain for the reduction of the coinage metal. In contrast, a certain delay between the nucleation of In⁰ and the injection of the gold precursor is required to gain bimetallic indium–gold phases. Otherwise, the reduction proceeded too fast and initiated a separate nucleation of nanosized gold. Since the pure coinage metals and the bimetallic systems exhibit different surface-plasmon resonance absorptions, the course of the reaction can be qualitatively followed and visualized via the particular color (Fig. 2) and quantified by optical spectroscopy (Fig. S2, ESI†).

For the co-reduction approach, InCl₃·4H₂O and the relevant coinage metal salt (i.e. CuCl₂·2H₂O, AgNO₃, KAuCl₄) were both introduced in the microemulsion and simultaneously reduced by subsequent addition of excess NaBH₄. Note that the co-reductions shown here were carried out at comparably low temperatures (e.g. 20–50 °C).

3.2 In⁰ precursor nanoparticles

Microemulsion techniques are well known for the synthesis of various inorganic nanomaterials.²⁴ Although this already comprises several metals (e.g. Bi, Co, Ni, Pd, Pt, Cu, Ag, Au),²⁵ nanoscale indium—to the best of our knowledge—has not yet been prepared via microemulsion techniques. Here, InCl₃·4H₂O was introduced into a standard w/o-(water-in-oil) microemulsion. Subsequent to addition of excess NaBH₄ as the reducing agent, nucleation of In⁰ is indicated by a deep grey color occurring within a few seconds. The course of the reaction can be clearly monitored by this color change and the simultaneous gas formation due to hydrogen release from NaBH₄. Finally, the microemulsion was destabilized by the addition of diethylene glycol (DEG).^20b At the same time, the nanoparticles accumulate in the DEG bottom phase. After centrifugation and careful washing, the In⁰ precursor nanoparticles were resuspended in ethanol to form a grayish dispersion, which is stable against reoxidation and agglomeration for months (Fig. 2). Previous studies have evidenced that these suspensions can be stored in air without the need for any inert gas atmosphere and without formation of In(OH)₃ or In₂O₃.²⁶

Size and shape of the In⁰ precursor nanoparticles are elucidated by electron microscopy (Fig. 3). Low-energy BF-STEM and TEM overview images show a manifold of spherical and nearly monodisperse particles, 12(2) nm in diameter (Fig. 3a,b,h). According to HRTEM images, the In⁰ nanoparticles show small crystalline areas with an average lattice fringe distance of 276(10) pm (Fig. 3c), which corresponds well with tetragonal bulk-In⁰ (d₁₀₁: 272 pm). The small dimensions of the crystalline areas points to an amorphous or quasi-molten state of the nanoparticles due to high-energy electron bombardment. Taking the low melting point of bulk-In⁰ (156 °C) into account, such behavior is expected and has been frequently observed for nanostructured indium samples.^26,27

Fig. 3 Analysis of In⁰ precursor nanoparticles: (a) low-energy BF-STEM overview image; ( b) TEM overview image; (c) typical HRTEM image; (d) HAADF-STEM image; (e) area-EDX spectrum (of dashed box in 3d); (f) Debye-SAED pattern; (g) XRD pattern; (h) size distribution according to statistical evaluation of STEM images.

SAED patterns show an intense diffraction ring corresponding to a distance of 272(3) pm (Fig. 3f), which is again consistent with bulk-In⁰ (d₁₀₁: 272 pm). The broadening of the ring and the absence of further diffraction rings can be ascribed to structural disorder. Some discrete (101) Bragg reflections are nevertheless visible and evidence the crystallinity of at least some nanoparticles. Since structural disorder might be an artifact of high-energy electron illumination, X-ray diffraction was involved to acquire a statistically relevant number of nanoparticles as present in a powder sample (Fig. 3g). The lattice parameters of the nanoparticles derived from XRD line positions (a = 325.5(5), c = 493.4(7) pm) are in good agreement with tetragonal bulk-In⁰ (a = 324.8, c = 494.8 pm). This demonstrates the crystallinity and purity of the pre-synthesized In⁰ precursor nanoparticles.

Considering the electrochemical potential of bulk-In⁰ (E₀: −0.34 V),¹⁹ fast re-oxidation—especially of high-surface nanoparticles—could have been expected. However, the In⁰ precursor nanoparticles turned out to be stable even in the absence of inert conditions. With regard to this finding, the surface capping of the purified nanoparticles was analyzed by FT-IR spectroscopy (Fig. S1, ESI†). Besides a broad vibration at 3450–3250 cm⁻¹ (ν(OH)) with a shoulder at 2950–2850 cm⁻¹ (ν(CH)), several vibrations in the fingerprint area (1300–800 cm⁻¹) indicate the presence of DEG as a capping agent. Hence, the unexpected stability of the In⁰ nanoparticles is ascribed to effective protection of its surface by DEG. This view is consistent with the role of DEG as reducing agent and colloidal stabilizer in the polyol-mediated synthesis of, for instance, Co⁰, Ni⁰ or Bi⁰ nanoparticles.²⁸ The overall good stability of pre-synthesized In⁰ is also beneficial in view of a controlled reaction to intermetallic compounds.

The surface conditioning of the In⁰ precursor nanoparticles was further investigated by optical spectroscopy (Fig. S2, ESI†). Here, the surface-plasmon resonance absorption of indium (240–280 nm)^26,29 is overlapped by the strong and broad absorption of DEG in the UV range, and hence not clearly visible. While a characteristic plasmon-resonance absorption of In⁰ is, in fact, not observed, UV-Vis spectroscopy is nevertheless a useful tool to analyze the plasmon-resonance absorption of the coinage metals when next performing the follow-up and co-reduction approach.

Finally, EDX spectra from single In⁰ precursor nanoparticles were taken at a primary electron energy of 300 keV. Due to the sensitivity of the In⁰ nanoparticles to the high-energy electron beam, EDXS line scans could not be performed. However, EDX spectra were obtained during scanning of a rectangular area within a single particle for about 120 s (Fig. 3d). The resulting EDX spectrum contains the characteristic L-lines of indium as well as of carbon (Kα1), oxygen (Kα1) and the L-lines of copper (Fig. 3e). While considering DEG as the surface capping, the presence of carbon and oxygen is not a surprise. Note that the copper signal, as well as a certain amount of carbon and oxygen, also stem from the carbon-film copper grid used as the sample holder. In summary, the presence and composition of DEG-capped In⁰ precursor nanoparticles is reliably evidenced based on the above discussed analytical tools.

3.3 The In–Cu system

Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles (follow-up approach). To characterize the nanomaterial obtained via a follow-up reaction of In⁰ precursor nanoparticles and equimolar amounts of CuCl₂·2H₂O (Table 1, Fig. 2), first of all, X-ray powder diffraction analysis was involved (Fig. 4h). The XRD patterns show the diffraction lines of the intermetallic compounds Cu₁₁In₉ and CuIn, as well as of In⁰. The lattice parameters of all constituents were calculated from their XRD peak positions and result in: monoclinic Cu₁₁In₉ (a = 1345(27), b = 435(9), c = 701(14) pm, β = 55.6(1.1)°), monoclinic CuIn (a = 736(17), b = 1515(36), c = 1018(24) pm, β = 133.7(3.2)°) and tetragonal In⁰ (a = 326(1), c = 494(1) pm). These values are in good agreement with the corresponding reference data.^30,31 While elemental Cu is not observed and while Cu is over-stoichiometric in the intermetallic compound Cu₁₁In₉, the presence of elemental In is in accordance with an overall In : Cu ratio of 1 : 1, as introduced in the synthesis. The absence of Cu⁰ is also validated by comparing UV-Vis spectra of the sample to reference spectra of Cu⁰ nanoparticles (Fig. S2, ESI†).³² The observed strong absorption at 230 nm originates from oleylamine as a surface capping. Based on the above findings, now, the question arises as to whether Cu₁₁In₉, CuIn and In⁰ are obtained as separate, individual nanoparticles or whether a certain gradient or core–shell structure emerges for each single particle.

Fig. 4 Analysis of Cu₁₁In₉@CuIn@In nanoparticles (follow-up approach): (a) low-energy BF-STEM overview image; (b) typical HRTEM image; (c) experimental diffractogram of inner particle region (of dashed area in 4b) and calculated diffraction pattern of monoclinic bulk-Cu₁₁In₉ ([001] zone axis, white circles); (d) HAADF-STEM image; (e) EDXS line scan (of yellow line in 4d); (f) 3D reconstruction of nanoparticle from composition profile; (g) indexed SAED pattern of nanoparticle ensemble; (h) XRD pattern; (i) size distribution according to statistical evaluation of STEM images.

BF-STEM images allow verification of the size and morphology of the obtained nanoparticles (Fig. 4a). With a mean size of 14(6) nm, particle diameter and size distribution are slightly broadened as compared to the initial In⁰ nanoparticles (Fig. 4i). HRTEM images of a single In–Cu nanoparticle show the crystalline, single phase Cu₁₁In₉ as the central core (Fig. 4b). This is also evidenced by good agreement between the experimental diffractogram and the calculated diffraction pattern of monoclinic bulk-Cu₁₁In₉ oriented along the [001] zone axis (Fig. 4c). In contrast, no lattice fringes were observed for the outer nanoparticle shell, which might be due to its small thickness, as well as due to an amorphous or quasi-molten state of the shell (see sections 3.2, 3.4 and 3.5). This view is supported by a distance of 275 pm, deduced from the diffuse diffraction ring in Fig. 4c that can be correlated to bulk-In⁰ (d₁₀₁: 272 pm). SAED analysis of single nanoparticles, however, confirms the results stemming from XRD (Fig. 4g). All observed Bragg peaks are again in accordance with Cu₁₁In₉, CuIn and In⁰.

HAADF-STEM images and EDXS line profiles measured across a single In–Cu nanoparticle with a diameter of 39 nm further elucidate the nanoparticle architecture (Fig. 4d). Quantification of In and Cu concentrations from EDX spectra was performed using the Cu–K and In–L series (Fig. 4e). Thus, a ∼2 nm thin outer shell of In⁰ can be deduced from the concentration profile. A more detailed composition analysis (see section 2.4) indicates the presence of a second shell, 3–4 nm in diameter, with an average composition Cu_0.53(9)In_0.47(9) that corresponds approximately to the CuIn phase. These two shells surround a nanoparticle core with a diameter of 27–29 nm. This particle core exhibits the composition Cu_0.64(15)In_0.36(15), which can be assigned to Cu₁₁In₉, as already suggested by the diffractogram (Fig. 4c).

Based on the analysis of single nanoparticles (HRTEM, HAADF-STEM, EDXS composition profiles) and based on the analysis of a statistically relevant number of particles in a powder sample (STEM, XRD, SAED), the complex composition and structure of Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles is reliably evidenced. The formation of this advanced architecture can be ascribed to a diffusion of copper into the initially prepared In⁰ precursor nanoparticles. While the synthesis of Cu₁₁In₉@CuIn@In was performed at low temperature (50 °C), an equilibration of the system to form CuIn as a single phase is obviously suppressed. In contrast, Cu₁₁In₉ was formed as the most stable intermetallic compound under these conditions. This is even more surprising since Cu₁₁In₉ as the bulk phase is typically formed at temperatures above 150 °C.³³ After a certain reduction of the copper concentration, CuIn becomes the more stable compound and is formed next. After the reduction of all the copper—with an overall In : Cu ratio of 1 : 1—excess In⁰ remains. While In⁰ represents the most mobile and the least stable compound, it forms the outer shell of the nanoparticle architecture. The obtained elaborate composition and structure is first observed and can be highly relevant to catalysis as well as a precursor for CIS solar cells.³⁴

Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles (co-reduction approach). For the In–Cu system only, the co-reduction approach led to similar results to the above follow-up approach. Thus, by simultaneous reduction of In³⁺ and Cu²⁺ in a microemulsion, Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles are again formed and exhibit a slightly reduced mean diameter of 6(2) nm (Fig. S3, ESI†). Here, addition of NaBH₄ instantaneously changed the color to black, indicating the formation of metal nanoparticles (Fig. 2).

Similar to the Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles that were yielded via the follow-up approach (Fig. 4), the composition, structure and morphology of the here obtained particle architecture were validated based on advanced electron microscopy (BF-STEM, HAADF-STEM, HRTEM), detailed analysis of X-ray and electron diffraction data and concentration profiles measured via EDXS line scans (Fig. S3, ESI†). Again, a complex Cu₁₁In₉@CuIn@In core@shell-A@shell-B structure was validated. While being similar to the results of the follow-up approach, these data are discussed in detail in the ESI.†

3.4 The system In–Ag

In–Ag Janushead-like nanoparticles (follow-up approach). Reacting In⁰ precursor nanoparticles with equimolar amounts of AgNO₃ and excess NaBH₄ resulted in dark-brown suspensions (Fig. 2). BF-STEM images show non-agglomerated nanoparticles with a narrow size distribution (Fig. 5a,i). Although of similar size, the shape of these In–Ag nanoparticles is, however, slightly more lens-shaped (shorter axis: 11(2) nm; longer axis: 13(3) nm), as compared to the spherical precursor In⁰. The detailed particle architecture was determined by HRTEM analysis (Fig. 5b). Here, the slightly elliptical shape is confirmed. Moreover, a distinct structural difference becomes obvious when comparing the left-hand side (about 6–8 nm in size) and the right-hand side (about 3–5 nm in size) of the obtained nanoparticles (Fig. 5b). Diffractograms of the two particle regions are consistent with cubic bulk-Ag⁰ oriented along the [110] zone axis on the left-hand side (Fig. 5c), and tetragonal bulk-In⁰ oriented along the [102] zone axis on the right-hand side (Fig. 5d).

Fig. 5 In–Ag Janushead-like nanoparticles (follow-up approach): (a) BF-STEM overview image; (b) typical HRTEM image; (c and d) experimental diffractograms of left-hand/right-hand particle side (of dashed area in 5b) as well as calculated diffraction pattern of cubic bulk-Ag⁰ ([110] zone axis, white circles) and of tetragonal bulk-In⁰ ([102] zone axis, white circles); (e) HAADF-STEM image; (f) spot EDX spectra (of blue/orange spots in 5e); (g) indexed SAED pattern of nanoparticle ensemble; (h) XRD pattern; (i) size distribution according to statistical evaluation of STEM images.

The composition of the two separate regions was further analyzed by EDXS of a single particle with a probe size of 1 nm at the positions indicated by the orange and blue dot in the HAADF-STEM image (Fig. 5e). Here, the EDX spectra taken in the Ag-rich region (right-hand side, orange dot) indeed show Ag–L lines with high intensity, whereas In–L lines only had low intensity (Fig. 5f). Quantification yielded 92(2) at% Ag and 8(1) at% In for the particle side showing a bright contrast (Fig. 5e). EDX spectra recorded in the In-rich side (left-hand side, blue dot), in contrast, show Ag–L lines with low intensity, and instead the In–L series with high intensity (Fig. 5f), resulting in 4(1) at% Ag and 96(2) at% In (Fig. 5e). This again evidences a “Janushead-like” architecture with silver and indium as opposite “particle faces”. The bimetallic structure of the In–Ag nanoparticles is further confirmed by XRD and SAED. Thus, XRD patterns indicate the presence of Ag⁰ (a = 405.9(5) pm) and In⁰ (a = 325.3(1), c = 493.7(1) pm) with calculated lattice parameters that are well in agreement with the reference data for cubic bulk-Ag⁰ and tetragonal bulk-In⁰ (Fig. 5h). Similarly, the Bragg reflections in SAED patterns can be attributed to the presence of Ag⁰ and In⁰ (Fig. 5g).

Qualitatively, the formation of the Janushead-like In–Ag nanoparticles can already be followed by the naked eye in the course of the reaction. Hence, the pristine grey suspension of In⁰ precursor nanoparticles turned to a characteristic dark-brown subsequent to the addition of AgNO₃. Quantification via UV-Vis spectroscopy evidences the characteristic surface-plasmon resonance of silver (Fig. S2, ESI†). Thus, the absorption maximum appears at 389 nm, which is in good agreement with literature data referring to values of around 400 nm for nanoscaled silver of similar size.³⁵ A certain blue-shift of the plasmon resonance can be attributed to the electronic structure of Ag⁰, which is influenced here by intimate contact with In⁰.³⁶ While the less-noble indium will naturally transfer electrons to silver, such blue-shift is to be expected for bimetallic Janushead-like architecture.^23c In addition, the oleylamine capping causes a strong absorption at around 230 nm that overlaps the surface-plasmon resonance of In⁰ (see section 3.2).

In summary, a Janushead-like architecture with In⁰ and Ag⁰ as opposite “faces” is obtained. The asymmetric structure of the nanoparticles can be explained by formation of a first silver seed crystal on the surface of the In⁰ precursor nanoparticles, right at the beginning of the reduction of Ag⁺. While the reduction of AgNO₃ proceeds, the growth of this first Ag⁰ seed crystal is obviously favored, as compared to the formation of additional seeds on the indium surface. This finding is in accordance with the fact that alloying of pre-formed indium and silver normally requires elevated temperatures (e.g. furnace, arc-melting techniques, microwave synthesis).³⁷ As a consequence of the proceeding growth of the once formed silver seed, bimetallic Janushead-like nanoparticles are finally obtained. When comparing the diameter of the initial In⁰ (12 nm) and the diameter of In⁰ remaining in the bimetallic In–Ag nanoparticles (5 nm), at least a partial dissolution of In⁰ becomes obvious. This finding can be attributed to the nanosized bimetal, which also represents an electrochemical local cell. Altogether, heterostructured bimetallic nanoparticles with diameters <50 nm are rare and limited to, for instance, Au–Pt heterodimers³⁸ or Au–Ag heterometallic nanorods and icosahedra.³⁹ A Janushead-like In–Ag architecture is first described.

Ag hollow spheres (follow-up approach). When reacting the In⁰ precursor nanoparticles with a fourfold excess of AgNO₃, the course of the reaction is different from the above described Janushead-like In–Ag nanoparticles (Table 1, Fig. 2). The large excess of the more noble silver results in the re-oxidation and more-or-less complete dissolution of indium. Hence, the reaction corresponds to the so-called sacrificial-template method as introduced by Xia et al., and is driven by the different electrochemical potentials of indium and silver.⁴⁰ Under these conditions silver forms a shell around the In⁰ nanoparticles. While the redox-reaction proceeds, the indium core is consumed more and more, with a shell of Ag⁰ finally remaining. As a consequence, nanoscale silver hollow spheres are formed in due course that exhibit almost the same size as purported by the initial In⁰ template.

According to BF-STEM and TEM overview images the nanoparticles exhibit an average diameter of 11(2) nm, which is very similar to the initial In⁰ nanoparticles (12 nm) (Fig. 6a,b,h). The nanoparticles exhibit a ring-type contrast, indicating the presence of an inner cavity (Fig. 6a,b). HRTEM images evidence the formation of nanoscale hollow spheres with an outer diameter of 18 nm and a wall thickness of 7 nm (Fig. 6c). The crystallinity and composition of the sphere wall are confirmed as elemental silver with lattice distances of 237(14) pm (cubic bulk-Ag⁰ d₁₁₁: 236 pm).⁴¹ Some hollow spheres, however, seem to be not completely empty, but to exhibit a certain inner structure. Note furthermore that the XRD pattern indicates the presence of In⁰ in addition to Ag⁰ (Fig. 6g). Here, the calculated lattice parameters with a = 408.9(9) pm, and a = 326.9(9), c = 491.9(14) pm are in agreement with cubic bulk-Ag⁰ and tetragonal bulk-In⁰. Moreover, all Bragg reflections in SAED patterns can be also attributed to In⁰ and Ag⁰ (Fig. 6f).

Fig. 6 Ag hollow spheres (follow-up approach): (a) low-energy BF-STEM overview image; (b) BF-STEM detail image; (c) HRTEM image; (d) HAADF-STEM image; (e) area EDX spectrum (of dashed area in 6d); (f) indexed SAED pattern of nanoparticle ensemble; (g) XRD pattern; (h) size distribution according to statistical evaluation of STEM images.

To verify the chemical composition, EDX spectra were recorded from single Ag⁰ hollow spheres at 30 keV primary electron energy by scanning a selected rectangular area for 120 s (Fig. 6d). Such an EDX spectrum reveals the characteristic L-lines of silver and indium as well as the lines of carbon (Kα1) and copper (L-lines) from the supporting film and the grid (Fig. 6e). Based on the observed Ag–L and In–L series, the chemical composition can be quantified to 68(7) at% Ag and 33(3) at% In. Obviously, non-consumed indium remains inside the silver shell, at least for some hollow spheres. The limited diffusibility of In³⁺ through the growing silver shell can be regarded as a kinetic reason for this residual In⁰.

Size and shape are well known to influence the surface-plasmon resonance, and thereby the optical properties of metal nanoparticles.⁴² In this regard, “hollowing” of silver or gold nanospheres typically leads to a red-shifted absorption maximum of the surface-plasmon resonance.⁴³ This effect is observed for Ag⁰ hollow spheres also (Fig. S2, ESI†). UV-Vis spectra show the absorption maximum at 423 nm, as compared to 400 nm for massive Ag⁰ nanoparticles of similar size.⁴⁴ Besides the surface-plasmon resonance, a strong absorption is observed at around 210 nm, which can be attributed to surface-bound DEG and polyvinylpyrrolidone (PVP) on the silver hollow spheres. These capping agents are also responsible for the comparatively low intensity of the plasmon-resonance absorption. Note finally that the color of the suspension of the Ag⁰ hollow spheres is significantly different from the In⁰ precursor nanoparticles, the In–Ag Janushead-like nanoparticles and the Ag₃In@In core@shell nanoparticles (Fig. 2, Fig. S2, ESI†). This again evidences the different architecture of the nanoparticles obtained in the In–Ag system.

Ag₃In@In core@shell nanoparticles (co-reduction approach). In contrast to In–Ag Janushead-like nanoparticles and nanoscale Ag⁰ hollow spheres, as obtained via the follow-up approach, the co-reduction of In³⁺ and Ag⁺ with an In : Ag ratio of 1 : 1 results in the formation of Ag₃In@In core–shell nanoparticles (Table 1, Fig. 2). Again, uniform and spherical particles are obtained with a mean size of 9(3) nm (Fig. 7a,k). While the follow-up approach did not lead to any intermetallic phase, the co-reduction of In³⁺ and Ag⁺ gives access to Ag₃In. XRD patterns confirm the presence of a hexagonal (h) Ag₃In as well as a cubic (c) Ag₃In phase (Fig. 7j). The resulting lattice parameters of a = 301.3(4), c = 465.2(6) pm (bulk h-Ag₃In: a = 299.0, c = 480.5 pm)⁴⁵ and a = 409.9(4) pm (bulk c-Ag₃In: a = 414(9) pm) are well in agreement with the literature data. Note that Bragg peaks of In⁰ are not observed.

Fig. 7 h-/c-Ag₃In@In nanoparticles (co-reduction approach): (a) low-energy BF-STEM overview image; (b) HRTEM image of h-Ag₃In@In nanoparticle; (c) experimental diffractogram of the inner particle region (of dashed area in 7b) and calculated indexed diffraction pattern of bulk h-Ag₃In ([100] zone axis, white circles); (d) indexed SAED pattern of nanoparticle ensemble; (e) HRTEM image of c-Ag₃In@In nanoparticle; (f) experimental diffractogram of the inner particle region (of dashed area in 7e) and calculated indexed diffraction pattern of bulk c-Ag₃In ([110] zone axis, white circles); (g) typical HAADF-STEM image; (h) EDXS line scan (of yellow line in 7g); (i) 3D reconstruction of nanoparticle from composition profile; (j) XRD pattern; (k) size distribution according to statistical evaluation of STEM images.

The HRTEM and HAADF images show evidence of a core–shell structure for the obtained In–Ag nanoparticles (Fig. 7b,e,g). Here, the particle core represents either a h-Ag₃In monocrystal oriented along the [100] zone axis (Fig. 7c) or a c-Ag₃In monocrystal with a [110] zone axis (Fig. 7f).⁴⁵ In contrast, the particle shell does not show any lattice fringes. This again points to an amorphous or quasi-molten shell similar to the In⁰ precursor nanoparticles (see section 3.2) as well as to Cu₁₁In₉@CuIn@In (see section 3.3) and Au@AuIn₂@In (see section 3.5) nanoparticles. This view is confirmed by electron diffraction patterns (Fig. 7d). Here, all Bragg reflections can be rationalized based on the presence of h-/c-Ag₃In as well as of In⁰.

Silver and indium concentration profiles were measured via EDXS line scans across a single nanoparticle with a diameter of 31 nm (Fig. 7g). Quantification was performed by using the Ag–L and In–L series and confirms the presence of an In-rich shell, about 5 nm in diameter (Fig. 7h,i). The outer 1–2 nm of this shell is composed of pure In⁰, whereas the remaining 3–4 nm of the shell has an average composition of In_0.87(5)Ag_0.13(5) (see section 2.4). However, the absence of any well-defined In–Ag phases around this composition suggests that the whole nanoparticle shell is indium-rich with an average composition of In_0.92(4)Ag_0.08(4) (Fig. 7h,i). In addition, the particle core has a diameter of 21–23 nm and a chemical composition of Ag_0.71(8)In_0.29(8), which can be assigned to Ag₃In, as already suggested by the diffractograms (Fig. 7c,f). As expected, no compositional difference can be detected in view of an inner core of h- and c-Ag₃In.

In summary, all applied analytical tools (BF-STEM, HRTEM, SAED, HAADF-STEM, EDXS) suggest that a h-/c-Ag₃In@In core@shell structure is obtained. The absence of diffraction lines from In⁰ in the XRD pattern can again be attributed to an amorphous or quasi-molten outer particle shell (see sections 3.2, 3.4 and 3.5). The dark yellowish brown color of the h-/c-Ag₃In@In suspension was finally quantified by UV-Vis measurements (Fig. S2, ESI†). Besides the strong absorption of oleylamine as a capping (around 230 nm), a comparatively weak absorption is observed with a maximum at 363 nm. Taking a silver-rich intermetallic compound such as Ag₃In into account, a plasmon-resonance absorption in the region of elemental silver (around 400 nm) is within expectations, and well-known for silver-rich Au–Ag alloys.^7a,42a The obvious blue shift can be explained by the electronic influence of the less-noble indium (see section 3.4). To our knowledge Ag₃In has not yet been prepared as a nanomaterial; comparison of the optical properties to a nanoscaled reference material is therefore not possible.

3.5 The system In–Au

Au@AuIn₂@In core@shell-A@shell-B nanoparticles (follow-up approach). Nanoparticles with Au@AuIn₂@In core@shell-A@shell-B architecture were obtained after formation of In⁰ precursor nanoparticles and reaction with KAuCl₄ (Table 1, Fig. 2). According to BF-STEM images, island-like assembles of several tens of separated nanoparticles are formed (Fig. 8a). The size distribution, with mean diameters of 14(5) nm, is slightly broadened as compared to the initial In⁰ (Fig. 8i). HRTEM images display the structure of the nanoparticles and indicate their core–shell-like architecture (Fig. 8b). Thus, the inner core represents crystalline Au⁰, as demonstrated by the exact agreement between its experimental diffractogram and the calculated diffraction pattern of cubic bulk-Au oriented along the [110] zone axis (Fig. 8c).⁴⁶ The outer particle shell does not show any lattice fringes and appears amorphous or quasi-molten (see sections 3.2–3.4).

Fig. 8 Au@AuIn₂@In nanoparticles (follow-up approach): (a) low-energy BF-STEM overview image; (b) typical HRTEM image; (c) experimental diffractogram of inner particle region (of dashed area in 8b) and calculated indexed diffraction pattern of cubic bulk-Au⁰ ([110] zone axis, white circles); (d) HAADF-STEM image; (e) EDXS line scan (of yellow line in 8d); (f) 3D reconstruction of nanoparticle from composition profile; (g) indexed SAED pattern of nanoparticle ensemble; (h) XRD pattern; (i) size distribution according to statistical evaluation of STEM images.

The architecture of the In–Au nanoparticles is further confirmed by SAED and XRD. All Bragg reflections observed in SAED patterns can be attributed to AuIn₂, Au⁰ and In⁰ (Fig. 8g). While HRTEM and SAED evidence the structure and composition of single particles, the XRD patterns confirm the presence of AuIn₂ and Au⁰ for a statistically relevant number of particles in a powder sample (Fig. 8h). The calculated lattice parameters of a = 408.6(1) pm (cubic Au) and a = 650.1(2) pm (cubic AuIn₂) are in good agreement with the bulk references.^46,47 Note that Bragg reflections of In⁰ were not observed.

Typical HAADF images of the In–Au nanoparticles support their core–shell structure with a clear contrast between its shell and core (Fig. 8d). Concentration profiles from EDXS line scans across a single nanoparticle, about 24 nm in diameter, are quantified based on Au–L and In–L series (Fig. 8e). Accordingly, an outer shell of 1–2 nm is composed of In⁰, followed by a second shell of 2–3 nm with a composition Au_0.27(4)In_0.73(5) (see section 2.4). The latter corresponds well with AuIn₂. Finally, the particle core with a diameter of 14–18 nm is composed of Au⁰, as already suggested by the diffractogram. In summary, the above considerations and analytical tools confirm an Au@AuIn₂@In core@shell-A@shell-B architecture as the result of the follow-up approach (Fig. 8f).

Optically, the Au@AuIn₂@In nanoparticles exhibit a weak absorption around 539 nm (Fig. S2, ESI†) that refers to the characteristic surface-plasmon resonance of Au⁰ nanoparticles (typically positioned at around 520 nm). While considering Au⁰ as the inner core of the nanoparticles, the weak absorption might point to a small number of particles with a comparably thin shell. An appreciable number of “naked” Au⁰ nanoparticles in the sample would, in contrast, lead to a much more pronounced characteristic absorption, as was indeed found in reference experiments.

The formation of the Au@AuIn₂@In particle architecture can be ascribed to a diffusion of as-reduced Au⁰ atoms into the initial In⁰ nanoparticles. A similar observation was described previously by Mori et al., based on indium and gold nanoparticles synthesized via evaporation techniques inside of a TEM.⁴⁸ While pure gold is obviously the most stable phase under the conditions of the follow-up approach, Au⁰ represents the inner particle core. As the concentration of gold decreases over the course of the reaction, the intermetallic phase AuIn₂ is formed next as a shell. With regard to the introduced In : Au ratio of 1 : 1, residual In⁰ remains as a thin outer shell on the nanoparticles. As discussed for all other nanostructures presented here, a precise adjustment of the experimental conditions turned out to be crucial in controlling the formation of one or the other phase composition.

AuIn₂ nanoparticles (co-reduction approach). Similar to the In–Ag system (see section 3.4), co-reduction and follow-up approach lead to different results for the In–Au system too. Thus, co-reduction of In³⁺ and Au³⁺ in the micellar system results in the formation of intermetallic AuIn₂ nanoparticles (Table 1, Fig. 2). Again, the course of the reaction can be followed by the naked eye. When adding the yellow solution of InCl₃·4H₂O and KAuCl₄, the cetyltrimethylammonium bromide (CTAB)-based microemulsion immediately turned orange due to halogen exchange from [AuCl₄]⁻ to [AuBr₄]⁻. The addition of NaBH₄ thereafter instantaneously changed the color to black, indicating the formation of AuIn₂. The optical properties were again quantified by UV-Vis (Fig. S2, ESI†). While the strong absorption related to the oleylamine capping (about 230 nm) is clearly visible, additional absorptions—such as the characteristic surface-plasmon resonance of Au⁰ nanoparticles—are not observed and confirm the purity of the AuIn₂ sample.

BF-STEM images show spherical nanoparticles with a mean diameter of 4(2) nm (Fig. 9a,h). XRD patterns evidence the formation of the intermetallic AuIn₂ phase (Fig. 9g). From the diffraction line positions, a lattice parameter of a = 648.9(10) pm was calculated that is in agreement with cubic bulk-AuIn₂.⁴⁷ Note that no Bragg reflections of In⁰ or Au⁰ were observed. The structure of the In–Au nanoparticles is further validated by the HRTEM images (Fig. 9b). Herein, the whole particle turned out to be crystalline AuIn₂, as demonstrated by the excellent agreement between its diffractogram and the calculated diffraction pattern of cubic bulk-AuIn₂ oriented along the [111] zone axis (Fig. 9c). SAED further confirms the purity of AuIn₂ (Fig. 9f). HAADF images of the nanoparticles do not suggest the presence of any outer shell here (Fig. 9d). EDX spectra were recorded in a rectangular area of a single AuIn₂ nanoparticle (Fig. 9e) and reveal the characteristic lines of gold (M-series) and indium (L-series). Moreover, the lines of carbon (Kα1) and copper (L-series) from the supporting film and grid are observed. As a result, the In and Au content were quantified to Au₃₅₍₉₎In₆₅₍₆₎, which fits well with the intermetallic phase AuIn₂.

Fig. 9 AuIn₂ nanoparticles (co-reduction approach): (a) low-energy BF-STEM overview image; (b) typical HRTEM image; (c) experimental diffractogram of inner particle region (of dashed area in 9b) and calculated indexed diffraction pattern of cubic bulk-AuIn₂ ([111] zone axis, white circles); (d) HAADF-STEM image; (e) area EDX spectrum (of dashed area in 9d); (f) indexed SAED pattern of nanoparticle ensemble; (g) XRD pattern; (h) size distribution according to statistical evaluation of STEM images.

4. Conclusions

Follow-up approach (i.e. reaction of pre-synthesized In⁰ precursor nanoparticles with Cu²⁺, Ag⁺ or Au³⁺) and co-reduction approach (i.e. simultaneous reduction of In³⁺ together with Cu²⁺, Ag⁺ or Au³⁺) in a micellar system results in a wide range of different compositions, structures and morphologies. This includes Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles, In–Ag Janushead-like nanoparticles, Ag hollow spheres, Ag₃In@In core@shell nanoparticles, Au@AuIn₂@In core@shell-A@shell-B nanoparticles and AuIn₂ nanoparticles. Due to the spatially limited volume of the micelles, particle nucleation and growth are very much limited, and lead to nanoparticles in a narrow size range. The elaborate architecture of the nanoparticles is firstly described and requires advanced electron microscopic techniques for reliable characterization (e.g. BF-/HAADF-STEM at 300 keV and energies ≤30 keV, HRTEM, SAED, spot/line-scan/area EDXS). In addition to the analysis of selected individual nanoparticles, the results were validated based on the analysis of powder samples containing a statistically relevant number of nanoparticles (STEM, XRD, UV-Vis).

The formation of the different particle architectures can be attributed, first of all, to the low reaction temperature (≤50 °C). As a consequence, the syntheses are performed under kinetic control, allowing a fine-tuning of the experimental conditions and the resulting nanostructures. Depending on the interdiffusion of the relevant metals and the thermodynamics of alloy formation, nanoscaled intermetallic compounds or non-alloyed bimetallic nanoparticles, including core–shell and Janushead-like structures, are obtained. In general, the follow-up approach favors the formation of core–shell structures, while the co-reduction approach favors the formation of intermetallic compounds. Whereas the composition and architecture of the obtained nanoparticles are significantly different, the very similar experimental conditions (type of microemulsion, concentration of reactants, temperature, etc.) guarantee a very similar size range 5–15 nm of all obtained nanostructures.

Based on the variety of compositions, structures and morphologies that can be obtained via the follow-up and co-reduction approach for the systems In–Cu, In–Ag and In–Au, many more compounds and elaborate architectures can be expected from extending the synthesis and its flexibility toward other metals. The resulting tool kit of new architectures of bimetallic and intermetallic nanostructures may also be relevant in view of new materials for catalysis, thin-film electronics or solar cells.

Acknowledgements

The authors are grateful to the Center for Functional Nanostructures (CFN) of the Deutsche Forschungsgemeinschaft (DFG) at the Karlsruhe Institute of Technology (KIT) for financial support (analytical tools and equipment). We also acknowledge the DFG for individual grant within the project FE 911/4-1.

References

1. C. J. Serpell, J. Cookson, D. Ozkaya and P. D. Beer, Nat. Chem., 2011, 3, 478.
2. (a) C. Burda, X. Chen, R. Narayanan and M. A. El-Sayed, Chem. Rev., 2005, 105, 1025; (b) N. Toshima and T. Yonezawa, New J. Chem., 1998, 1179; (c) H. Goesmann and C. Feldmann, Angew. Chem., Int. Ed., 2010, 49, 1362; (d) G. A. Ozin and A. C. Arsenault, Nanochemistry, RSC Publishing, Cambridge, 2005.
3. (a) U. Jeong, P. H. C. Camargo, Y. H. Lee and Y. Xia, J. Mater. Chem., 2006, 16, 3893; (b) Y. Yin, R. M. Rioux, C. K. Erdonmez, S. Hughes, G. A. Somorjai and A. P. Alivisatos, Science, 2004, 304, 711; (c) Y. G. Sun and Y. N. Xia, Science, 2002, 298, 2176; (d) Y. Vasquez, A. K. Sra and R. E. Schaak, J. Am. Chem. Soc., 2005, 127, 12504; (e) J. M. Nedeljković, O. I. Mićić, S. P. Ahrenkiel, A. Miedaner and A. J. Nozik, J. Am. Chem. Soc., 2004, 126, 2632; (f) Z. Bai, L. Yang, J. Zhang, L. Li, J. Lv, C. Hu and J. Zhou, Catal. Commun., 2010, 11, 919.
4. C. Suryanarayana, E. Ivanov and V. V. Boldyrev, Mater. Sci. Eng. A, 2001, 151, 304.
5. E. Zintl, A. Harder and W. Haucke, Z. Phys. Chem., 1937, 35, 354.
6. (a) R. Ferrando, J. Jellinek and R. L. Johnston, Chem. Rev., 2008, 108, 845; (b) D. Wang and Y. Li, Adv. Mater., 2011, 23, 1044; (c) X. Liu and X. Liu, Angew. Chem., 2012, 124, 3367; (d) M. Boutonnet, S. Lögdberg and E. E. Svensson, Curr. Opin. Colloid Interface Sci., 2008, 13, 270; (e) M. P. Pileni, Nat. Mater., 2003, 2, 145.
7. (a) S. Liu, G. Chen, P. N. Prasad and M. T. Swihart, Chem. Mater., 2011, 23, 4098; (b) Y. Liu, M. Chi, V. Mazumder, K. L. More, S. Soled, J. D. Henao and S. Sun, Chem. Mater., 2011, 23, 4199; (c) Y. Liu and A. R. Walker, Angew. Chem., Int. Ed., 2010, 49, 6781; (d) M. L. Wu, D. H. Chen and T. C. Huang, Langmuir, 2001, 17, 3877.
8. (a) S. Sun, C. B. Murray, D. Weller, L. Folks and A. Moser, Science, 2000, 287, 1989; (b) K. M. Hyie and I. I. Yaacob, J. Mater. Process. Technol., 2007, 191, 48.
9. (a) R. E. Cable and R. E. Schaak, J. Am. Chem. Soc., 2006, 128, 9588; (b) Z. Liu, D. Reed, G. Kwon, M. Shamsuzzoha and D. E. Nikles, J. Phys. Chem. C, 2007, 111, 14223; (c) Z. Liu, G. S. Jackson and B. W. Eichhorn, Angew. Chem., Int. Ed., 2010, 49, 3173; (d) Y. Liu, D. Li, V. R. Stamenkovic, S. Soled, J. D. Henao and S. Sun, ACS Catal., 2011, 1, 1719; (e) L. M. Magno, W. Sigle, P. van Aken, D. Angelescu and C. Stubenrauch, Phys. Chem. Chem. Phys., 2011, 13, 9134.
10. (a) B. J. Auten, B. P. Hahn, G. Vijayaraghavan, K. J. Stevenson and B. D. Chandler, J. Phys. Chem. C, 2008, 112, 5365; (b) S. Zhou, Z. Ma, H. Yin, Z. Wu, B. Eichhorn, S. H. Overbury and S. Dai, J. Phys. Chem. C, 2009, 113, 5758.
11. W. Wang, Q. Peng and Y. Li, Nano Res., 2010, 3, 574.
12. N. H. Chou and R. E. Schaak, J. Am. Chem. Soc., 2007, 129, 7339.
13. A. K. Sra and R. E. Schaak, J. Am. Chem. Soc., 2004, 126, 6667.
14. (a) W. Chen, R. Yu, L. Li, A. Wang, Q. Peng and Y. Li, Angew. Chem., Int. Ed., 2010, 49, 2917; (b) R. E. Schaak, A. K. Sra, B. M. Leonard, R. E. Cable, J. C. Bauer, Y.-F. Han, J. Means, W. Teizer, Y. Vasquez and E. S. Funck, J. Am. Chem. Soc., 2005, 127, 3506.
15. M. Armbrüster, G. Wowsnick, M. Friedrich, M. Heggen and R. Cardoso-Gil, J. Am. Chem. Soc., 2011, 133, 9112.
16. (a) Q. Zhang, J. Xie, Y. Yu and J. Y. Lee, Nanoscale, 2010, 2, 1962; (b) Y. Okuno, K. Nishioka, A. Kiya, N. Nakashima, A. Ishibashi and Y. Niidome, Nanoscale, 2010, 2, 1489.
17. V. Mazumder, M. Chi, K. L. More and S. Sun, Angew. Chem., Int. Ed., 2010, 49, 9368.
18. S. Patra and H. Yang, Bull. Korean Chem. Soc., 2009, 30, 1485.
19. S. Hakomori, J. Am. Chem. Soc., 1930, 52, 2372.
20. (a) M. P. Pileni, J. Phys. Chem., 1993, 97, 6961; (b) D. H. M. Buchold and C. Feldmann, Adv. Funct. Mater., 2008, 18, 1002.
21. X. L. Wang, W. Q. Han, J. Chen and J. Graetz, ACS Appl. Mater. Interfaces, 2010, 2, 1548.
22. M. G. Blaber, M. D. Arnold and M. J. Ford, J. Phys.: Condens. Matter, 2010, 22, 143201.
23. (a) A. Elattar, T. Takeshita, W. E. Wallace and R. S. Craig, Science, 1977, 196, 1093; (b) H. Abe, F. Matsumoto, L. R. Alden, S. C. Warren, H. D. Abruña and F. J. Di Salvo, J. Am. Chem. Soc., 2008, 130, 5452; (c) T. Yonezawa and N. Toshima, J. Mol. Catal., 1993, 83, 167.
24. (a) M. P. Pileni, Cryst. Res. Technol., 1998, 7–8, 1155; (b) C. Destrée and J. B. Nagy, Adv. Colloid Interface Sci., 2006, 123–126, 353; (c) J. Eastoe, M. J. Hollamby and L. Hudson, Adv. Colloid Interface Sci., 2006, 128–130, 5.
25. (a) B. L. Cushing, V. L. Kolesnichenko and C. J. O'Connor, Chem. Rev., 2004, 104, 3893; (b) I. Capek, Adv. Colloid Interface Sci., 2004, 110, 49.
26. C. Kind and C. Feldmann, Chem. Mater., 2011, 23, 4982.
27. (a) G. L. Allen, R. A. Bayles, W. W. Gile and W. A. Jesser, Thin Solid Films, 1986, 144, 297; (b) D. Xie, M. P. Wang, W. H. Qi and L. F. Cao, Mater. Chem. Phys., 2006, 96, 418; (c) T. H. Lim, B. Ingham, K. H. Kamarudin, P. G. Etchegoin and R. D. Tilley, Cryst. Growth Des., 2010, 10, 3854.
28. (a) F. Fievet, J. P. Lagier and M. Figlarz, Mater. Res. Bull., 1989, 29; (b) A. Luz and C. Feldmann, J. Mater. Chem., 2009, 19, 8107.
29. (a) P. Singh, S. Kumar, A. Katyal, R. Kalra and R. Chandra, Mater. Lett., 2008, 62, 4164; (b) N. H. Chou, X. Ke, P. Schiffer and R. E. Schaak, J. Am. Chem. Soc., 2008, 130, 8140; (c) P. K. Khanna, K.-W. Jun, K. B. Hong, J.-O. Baeg, R. C. Chikate and B. K. Das, Mater. Lett., 2005, 59, 1032; (d) S. Cingarapu, Z. Yang, C. M. Sorensen and K. J. Klabunde, Inorg. Chem., 2011, 50, 5000.
30. (a) T. P. Rajasekharan and K. Schubert, Z. Metallkd., 1981, 72, 275; (b) V. Simić and Ž. Marinković, J. Less Common Met., 1980, 72, 133.
31. M. E. Straumanis, P. B. Rao and W. J. James, Z. Metallkd., 1971, 62, 493.
32. (a) P. K. Khanna, S. Gaikwad, P. V. Adhyapak, N. Singh and R. Marimuthu, Mater. Lett., 2007, 61, 4711; (b) M. Singh, I. Sinha, M. Premkumar, A. K. Singh and R. K. Mandal, Colloids Surf., A, 2010, 359, 88; (c) C. Kind, A. Weber and C. Feldmann, J. Mater. Chem., 2012, 22, 987.
33. H. Okamoto, J. Phase Equilib., 2005, 26, 645.
34. C. Kind, C. Feldmann, A. Quintilla and E. Ahlswede, Chem. Mater., 2011, 23, 5269.
35. (a) C. Petit, P. Lixon and M. P. Pileni, J. Phys. Chem., 1993, 97, 12974; (b) A. Slistan-Grijalva, R. Herrera-Urbina, J. F. Rivas-Silva, M. Ávalos-Borja, F. F. Castillón-Barraza and A. Posada-Amarillas, Phys. E., 2005, 27, 104.
36. P. Mulvaney, Langmuir, 1996, 12, 788.
37. J. W. Lekse, T. J. Stagger and J. A. Aitken, Chem. Mater., 2007, 19, 3601.
38. S. I. Lim, M. Varon, I. Ojea-Jiménez, J. Arbiol and V. Puntes, J. Mater. Chem., 2011, 21, 11518.
39. M. R. Langille, J. Zhang and C. A. Mirkin, Angew. Chem., Int. Ed., 2011, 50, 3543.
40. (a) Y. Sun and Y. Xia, Anal. Chem., 2002, 74, 5297; (b) Y. Sun, B. Mayers and Y. Xia, Adv. Mater., 2003, 15, 641; (c) X. Lu, H.-Y. Tuan, J. Chen, Z.-Y. Li, B. A. Korgel and Y. Xia, J. Am. Chem. Soc., 2007, 129, 1733.
41. E. A. Owen and E. L. Yates, J. Chem. Phys., 1935, 3, 605.
42. (a) Y. Sun and Y. Xia, Analyst, 2003, 128, 686; (b) E. Hutter and J. H. Fendler, Adv. Mater., 2004, 16, 1685.
43. (a) H. P. Liang, L. J. Wan, C. L. Bai and L. Jiang, J. Phys. Chem. B, 2005, 109, 7795; (b) M. Gellner, B. Küstner and S. Schlücker, Vib. Spectrosc., 2009, 50, 43.
44. C. Kind, R. Popescu, E. Müller, D. Gerthsen and C. Feldmann, Nanoscale, 2010, 2, 2223.
45. (a) A. N. Campbell and W. F. Reynolds, Can. J. Chem., 1962, 40, 37; (b) A. N. Campbell, R. Wagemann and R. B. Ferguson, Can. J. Chem., 1970, 48, 1703.
46. J. J. Couderc, G. Garique, L. Lafourcade, Q. T. Nguyen and Z. Metallkd, Z. Metallkd., 1959, 50, 708.
47. O. Kubaschewski and Z. Elektrochem, Angew. Phys. Chem., 1938, 44, 870.
48. H. Yasuda and H. Mori, Mater. Sci. Forum, 1999, 307, 167.

---

Footnote

† Electronic supplementary information (ESI) available: Detailed information regarding the starting materials and the analytical methods are described. Moreover, the analysis of the surface capping of the In⁰ precursor nanoparticles, the complete investigation of the surface-plasmon resonance absorption of all obtained nanoparticles as well as the complete analysis of the Cu₁₁In₉@CuIn@In core@shell-A@shell-B nanoparticles gained via the co-reduction approach can be found here. See DOI: 10.1039/c2ra21659k

---