Copyright WILEY‐VCH Verlag GmbH & Co. KGaA, 69469 Weinheim, Germany, 2011.

Supporting Information

for Small, DOI: 10.1002/smll.201101882

Metallic Nanobowls by Galvanic Replacement Reaction on Heterodimeric Nanoparticles Yonatan Ridelman, Gurvinder Singh, Ronit Popovitz-Biro, Sharon G. Wolf, Sanjib Das, and Rafal Klajn*

- 1 -

SUPPORTING INFORMATION

Metallic Nanobowls by Galvanic Replacement Reaction on Heterodimeric Nanoparticles** Yonatan Ridelman, Gurvinder Singh, Ronit Popovitz-Biro, Sharon G. Wolf, Sanjib Das, and Rafal Klajn* [*] Y. Ridelman, Dr. G. Singh, Dr. S. Das, Dr. R. Klajn Department of Organic Chemistry Weizmann Institute of Science Rehovot, 76100 (Israel) E-mail: rafal.klajn@weizmann.ac.il Dr. R. Popovitz-Biro, Dr. S. G. Wolf Department of Chemical Research Support Weizmann Institute of Science Rehovot, 76100 (Israel) This work was supported by the European Union Marie Curie Reintegration Grant, the G. M. J. Schmidt-Minerva Center for Supramolecular Architectures, the Helen and Martin Kimmel Center for Molecular Design, and the Minerva Foundation with funding from the Federal German Ministry for Education and Research. The EM studies were conducted at the Irving and Cherna Moskowitz Center for Nano and Bio-Nano Imaging at the Weizmann Institute. We thank Dr. Arye Tishbee for performing ICP-MS measurements and Dr. Inna Popov (Harvey M. Krueger Center for Nanoscience and Nanotechnology, Hebrew University of Jerusalem) for performing STEM-EDX studies. R. K. is the incumbent of the Robert Edward and Roselyn Rich Manson Career Development Chair. 1. Preparation of Fe3O4 nanoparticles

Monodisperse Fe3O4 nanoparticles of various sizes were prepared by thermal decomposition of iron

(III) oleate in the presence of oleic acid, according to a modified literature procedure.[31] First, iron (III)

oleate was synthesized by reacting sodium oleate with iron (III) chloride. Briefly, sodium oleate (36.5 g,

120 mmol) (TCI, >97%; we found that the high purity of sodium oleate was critical for reproducible

synthesis of high-quality, monodisperse Fe3O4 NPs) and iron (III) chloride hexahydrate (10.8 g, 40

mmol) (Alfa Aesar, 98%) were dissolved in a mixture of solvents composed of 60 mL of distilled

water, 80 mL of ethanol and 140 mL of hexane. The resulting solution was refluxed (bath temperature

was set up to ο70 C ) with vigorous stirring and under a gentle flow of nitrogen. The reaction was

discontinued after four hours and the upper (hexane) layer containing the product was collected,

washed three times with 30 mL of distilled water in a separatory funnel and dried over magnesium

sulfate. Removal of the solvent in vacuo gave the product in the form of a waxy solid in a near

quantitative yield.

- 2 -

We found that the diameter of monodisperse Fe3O4 NPs could conveniently be regulated by the

amount of oleic acid added during the synthesis (noleic acid / niron oleate = χ). In a typical synthesis of Fe3O4

NPs, iron (III) oleate (1.600 g, 1.78 mmol) was dissolved in 25 mL of octadecene (Aldrich, 90%) and a

given amount of oleic acid (Alfa Aesar, 90%) was added. The solution was then heated with a constant

heating rate of 3○C min–1 to T = 310○C, and was left at this temperature for 30 min. After cooling

down to room temperature, the NPs were precipitated with a mixture of solvents composed of n-hexane,

isopropanol and acetone (v / v / v = 1 : 2 : 2). The transparent supernatant was discarded and the solids

were washed with a mixture of n-hexane and acetone (v / v = 1 : 2). Evacuation of the solvent in vacuo

afforded monodisperse Fe3O4 nanoparticles (120 mg, 0.518 mmol, yield = 87%). The diameter of the

Fe3O4 NPs was inversely proportional to χ – for example, for χ = 1.42 we obtained Fe3O4 NPs of

diameter, d ≈ 7.6 nm, χ = 1.07 resulted in d ≈ 12 nm, χ = 1.01 gave NPs of d ≈ 17 nm, χ = 0.979

afforded NPs of d ≈ 23 nm, whereas χ = 0.890 yielded NPs of d ≈ 26 nm.

Figure S1. Representative TEM images of Fe3O4 NPs of obtained at χ = 1.42 (left; d ≈ 7.6 nm), χ = 0.979 (center; d ≈ 23 nm) and χ = 0.890 (right; d ≈ 26 nm). 2. Preparation of Ag-Fe3O4 heterodimeric nanoparticles

Ag-Fe3O4 heterodimeric nanoparticles were prepared based on a modified literature procedure.[32] In a

typical synthesis, silver acetate (25 mg, 0.150 mmol) (Strem, 99%) was dissolved in toluene (20 mL)

containing oleylamine (1.219 g, 1.499 mL, 4.56 mmol) (Aldrich, 70%), and then Fe3O4 NPs (d ≤ 26 nm;

20 mg) were added. The solution was heated to T = 70○C under nitrogen atmosphere, and kept at this

temperature for 12 h. The solution was cooled down to room temperature and the resulting

heterodimeric NPs were purified by the addition of ethanol (1.0 volume equivalents with respect to the

original volume) followed by centrifugation. The solids were carefully separated from the supernatant and

- 3 -

redispersed in pure toluene. The concentration of the solution was determined by means of inductively

coupled plasma mass spectrometry (ICP-MS) (ELAN 9000, PerkinElmer SCIEX). Representative high

resolution TEM images of our heterodimers are shown in Figure S2. The diameter of the silver domain

could be controlled by the amount of silver acetate used for the reaction, or by discontinuing the

reaction after different times (see Figure S3a, b). Figure S4 shows representative TEM images of HDs

obtained from Fe3O4 NPs of d ≈ 12 nm (left) and d ≈ 23 nm (center), as well as images of Fe3O4 NPs

hosting, on average, more than one Ag domain each (obtained from d ≈ 26 nm Fe3O4; Figure S4,

right).

Figure S2. HR TEM images of individual Ag-Fe3O4 HDs (obtained from 12 nm Fe3O4 NPs) revealing mono- and polycrystalline nature of the Fe3O4 (larger) and Ag (smaller) domains, respectively.

Figure S3. Preparation of Ag-Fe3O4 heterodimeric NPs. a, b) Size evolution of the silver domain with time. Dashed line in b) indicates the theoretical maximum diameter of Ag calculated based on the amount of silver acetate used for the synthesis, and assuming that Ag+ is quantitatively reduced to Ag

- 4 -

comprising the HDs. One can see that it takes ~48 hrs to achieve a complete reduction of Ag+. c) Linear dependence of the size of Ag on the diameter of Fe3O4. We attribute this observation to the fact that larger Fe3O4 seeds can provide a relatively large interfacial surface area of contact with the resulting Ag. The growth of Ag on the smaller Fe3O4, on the other hand, is arrested relatively quickly due to the larger curvature of Fe3O4 NPs.

Figure S4. TEM images of Ag-Fe3O4 nanoparticles obtained under the same conditions from Fe3O4 NPs of different diameters (left: d ≈ 12 nm; center: d ≈ 23 nm; right: d ≈ 26 nm) (note that the scale in all images is the same). 3. Preparation of Ag/Au nanobowls

Preparation of Ag/Au nanobowls: A stock solution of gold (I) was first prepared by sonicating solid

AuCl (2.0 mg, 8.6 μmol) (Acros, >99%) in toluene (3.98 mL) containing oleylamine[S1] (16.26 mg,

20.0 μL, 60.8 μmol) (Aldrich, 70%) for 3 min. The stock solution was then diluted with pure toluene to

give a gold (I) solution of cAuCl = 0.15 mM. A solution of Ag-Fe3O4 HDs (nAg = 1.5 μmol; different

diameters of Ag and Fe3O4 domains) in 19.5 mL of toluene containing excess of oleylamine (615 mg,

756 μL, 2.30 mmol) was placed in a two-neck round-bottom flask immersed in an oil bath set to Τ =

50○C. To this HD solution, the freshly prepared AuCl was injected under a nitrogen atmosphere (with

a syringe pump; injection rate of 0.05 μmol min–1. After a given amount of AuCl was added, the

reaction mixture was cooled down to room temperature, and the Ag/Au nanobowls were purified by the

- 5 -

addition of 1.5 volume equivalents of EtOH (with respect to the volume of the reaction mixture)

followed by centrifugation. The resulting precipitates – a mixture of nanobowls and free Fe3O4 NPs –

were then redispersed in toluene. The magnetic NPs were removed with the help of a magnet (see Figure

3a in the main text) to give a solution of pure nanobowls.

To verify that the elemental composition of the nanobowls matches that expected from the

amount of AuCl added, we used energy dispersive X-ray spectroscopy (EDX) to quantitatively analyze the

Ag-to-Au ratio within our NPs. For example, analysis of multiple individual nanobowls obtained at θ = 0.20

reveals that they are composed of 20.9 ± 1.7% Au and 79.1 ± 1.7% Ag, thus indicating that AuCl is

quantitatively converted to metallic Au comprising the nanobowls. The TEM pictures below show

bowls of two different sizes obtained from differently sized precursor HDs.

Figure S5. Representative TEM images of θ = 0.25 nanobowls obtained from differently sized heterodimeric nanoparticles: a) ~12 nm/~7 nm, and b) ~23 nm/~11 nm Ag-Fe3O4 HDs. Note that the nanobowl in (a) is loosely connected to a Fe3O4 NP, perhaps its parent NP from which it is dissociating. 4. Br–/O2-induced etching of Ag/Au nanobowls followed by UV-Vis spectroscopy

Partial etching of Ag/Au nanobowls was induced by the addition of 50 eq of didodecyldimethyl-

ammonium bromide (DDAB) to a toluene solution of the nanobowls. The solution was left exposed to

air. The nanobowls referred to in the main text (Figure 2b) were analyzed after 4 hours of etching (red

line in Figure S6). During the initial ~8 hours the only noticeable change in the structure of the

nanobowls was the expansion of the cavity – longer times led to the destruction of the nanobowls and

their transformation into a mixture of O-shaped, C-shaped, and other irregular nanoparticles.

- 6 -

Figure S6. Changes in optical spectra of Ag/Au nanobowls accompanying partial etching with DDAB. 5. Control experiments on free Ag NPs (NPs missing the Fe3O4 "protecting group")

When the GRR was carried out under the same conditions as those employed for the preparation of

nanobowls (see Supporting Information, Section 3), but on free Ag NPs (of the same size as the Ag domains

of the HDs), the reaction product was composed of C-shaped nanoparticles, rings and other ill-defined NPs

(Figure S7, top). These results indicate that the presence of "protecting group" not only enables the

preparation of nanobowls, but also directs the reaction to produce only one type of NP morphology.

Figure S7. Results of the galvanic replacement reaction between AuCl and (top) Ag NPs, and (bottom) Ag-Fe3O4 HD NPs. Scale bars = 50 nm.

- 7 -

6. Further mechanistic studies support the role of Fe3O4 as the "protecting group"

To further confirm that the Fe3O4 domain directs the spatial deposition of Au (and the concomitant

formation of the cavity), we attempted to tune the reaction conditions of the GRR such that the metallic

domains remain in intimate contact with their parent Fe3O4 NPs. We reasoned the two domains would

stay connected – as a consequence of the tendency to decrease unfavorable metal / solvent interface area –

when the reaction took place in a more polar solvent. In fact, when the reaction was carried out in

chloroform at T = 0○C, the two domains not only stayed in touch throughout the reaction, but the

metallic parts assumed a shape of a letter "C" to further decrease the contact with the solvent (Figure S8;

compare with the highly branched structures in Figure 3f obtained under the same conditions but in

toluene as the solvent). While these modified conditions can provide a route to C-shaped metallic NPs,

the point we would like to make here is that in all the particles observed (Figure S8), the concave side

of the metallic domain faced Fe3O4, thus supporting the view that Fe3O4 acts as the "protecting group".

Figure S8. TEM images of Fe3O4-(C-shaped Au) heterodimers prepared in CHCl3 at T = 0○C.

- 8 -

7. HR TEM images confirming the single crystalline nature of the Au domain (see Figure 3a, b)

Figure S9. HR TEM images and the corresponding FFT patterns of gold domains of the Fe3O4-Au heterodimers prepared at T = 100○C. The particles shown in a)-c) are single-crystalline, whereas the one in d) is a twinned crystal. All scale bars correspond to 5 nm. 8. Electron microscopy studies

Transmission electron microscopy (TEM) studies were performed on a Philips CM120 Super Twin

microscope operating at 120 kV. High-resolution (HR) TEM was performed on an FEI Tecnai F30 UT

microscope operating at 300 kV and micrographs were acquired by a Gatan Ultrascan 1000 CCD camera.

STEM-EDX measurements were performed on an FEI Tecnai F20 G2 microscope equipped with an EDX

detector (EDAX-TSL). Scanning electron microscopy (SEM) studies were performed on a Supra 55VP

LEO microscope operated at 15 kV. STEM tomography was performed on an FEI Tecnai F20 G2 microscope

in STEM mode. Tomograms were acquired at 110k magnification, with 2 degrees tilt separation, using

the Saxton collection scheme.[S1] Tomograms were collected automatically, using FEI Explore-3D software,

and the reconstructions performed with FEI Inspect-3D software. The contoured images of individual

particles were obtained using isosurface rendering with Avizo software (VSG, Burlington MA, USA).

- 9 -

9. Additional tomograms of nanobowls

Figure S10. Additional three-dimensional reconstructions of θ = 0.25 nanobowls. The tomograms were obtained from a tilt series of STEM images. 10. Supporting references

S1. Oleylamine (used in a large excess) played at least three roles during the GRR: first, it enabled

solubilization of Au+ in toluene; second, it slowed down the reaction thus enabling the preparation

of uniform nanobowls; third, it kept the product of the reaction (AgCl) soluble in the reaction

medium, thus preventing it from depositing on the NPs – see Refs. 21, 31 in the main text.

S2. W.O. Saxton, W. Baumeister, M. Hahn, Ultramicroscopy 1984, 13, 57.