Influence of Iodide Ions on the Growth of Gold Nanorods: Tuning Tip Curvature and Surface Plasmon Resonance

FULLPAPER

3780

DOI: 10.1002/adfm.200800706

Influence of Iodide Ions on the Growth of Gold Nanorods: TuningTip Curvature and Surface Plasmon Resonance**

By Marek Grzelczak, Ana Sanchez-Iglesias, Benito Rodrıguez-Gonzalez, Ramon Alvarez-Puebla,Jorge Perez-Juste,* and Luis M. Liz-Marzan*

This paper describes morphological and optical changes induced by seed-mediated growth of gold nanorods in the presence of

iodide ions. Addition of small amounts of iodide to the growth solution results in the growth of nanoparticles with dumbbell-like

structure, meaning that gold salt reduction takes place preferentially at the rod tips. However, when excess iodide is added,

homogeneous rod growth is observed, and therefore the original shape is retained. By controlling the experimental conditions,

the position of the longitudinal plasmon band of grown nanorods can be shifted up to as much as 250 nm. These optical effects

were also simulated by means of the boundary element method (BEM), achieving an excellent agreement with the experimental

spectra. X-ray photoelectron spectroscopy (XPS) and surface enhanced Raman spectroscopy (SERS) analysis of the gold

nanorods before and after iodide addition revealed the presence of AuI and AgI at the particles surface. A growth mechanism is

proposed on the basis of preferential iodide adsorption at the tips {111} facets, leading to the formation of AgI, followed by

reduction of gold salt precursor due to a decrease in the surface redox potential.

1. Introduction

Gold nanorods attract enormous attention because of their

extremely interesting, surface-plasmon based optical proper-

ties,[1] which render them ideal candidates for a large number

of applications, in particular for medical uses in diagnosis and

therapy.[2–4] The optical response of nanorods typically

involves two surface plasmon modes, associated with the

oscillation of conduction electrons, either parallel (long-

itudinal surface plasmon, LSP) or perpendicular (transverse

surface plasmon, TSP) to the rod long axis. One of the most

attractive features of nanorods is the exquisite dependence of

the LSP frequency with small variations in the aspect ratio,

with no need to significantly vary the overall dimensions.

However, in order to achieve a narrow plasmon band in the

near-IR (the water optical window, suited for medical

applications),[5] considerably longer rods are required, which

are often more polydisperse.[6] Therefore, synthetic methods

are still required for tuning the morphology of gold rods in such

[*] Prof. L. M. Liz-Marzan, Dr. J. Perez-Juste, Dr. M. Grzelczak,A. Sanchez-Iglesias, Dr. B. Rodrıguez-Gonzalez, Dr. R. Alvarez-PueblaDepartamento de Quımica Fısica and Unidad Asociada CSICUniversidade de VigoVigo 36310 (Spain)E-mail: [email protected]; [email protected]

[**] The authors thank Prof. F. Javier Garcıa de Abajo (CSIC) for providingthe BEM software and assisting with modeling. This work wassupported by the Spanish MEC (MAT2007-62696; Consolider-IngenioNanobiomed) and Xunta de Galicia (PGIDIT06TMT31402PR). Sup-porting Information is available online from Wiley InterScience orfrom the author.

� 2008 WILEY-VCH Verlag GmbH &

a way that the LSP can be tailored with no need to make very

long rods.

Although various methods have been proposed for the

reproducible and controlled synthesis of gold nanorods,

nowadays almost exclusively the seeded growth method is

used for both scientific studies and practical applications. This

method, based on the reduction of a gold salt on pre-made

small seeds by a weak reducing agent (ascorbic acid), in the

presence of a cationic surfactant (most frequently cetyl-

trimethylammonium bromide, CTAB), has been optimized

along the present decade to produce nanorods with a wide

range of aspect ratios and dimensions.[7–9] Additionally,

variations of the method have been found to result in a

variety of other morphologies, such as cubes,[10] plates,[11]

stars,[10] or dog bone-like nanoparticles.[12,13] While the precise

mechanism involved in these morphological changes is not

completely understood, and several possibilities have been

proposed, including electric field-directed preferential reduc-

tion at the tips,[14] preferential adsorption of CTAB on certain

crystallographic faces,[15] or underpotential deposition of a

small amount of silver,[8] we still see every now and then new

reports which do not really fit with these models, and thus a

further insight in the growth mechanism is still required.

Recently,[11] the presence of iodide ions during seeded

growth on small seeds has been shown to lead to the formation

of uniform gold nanoplates. The authors claimed that for-

mation of the planar structure was due to growth inhibition of

the {111} facets by strongly bound iodide ions, accompanied by

preferential gold reduction on the more curved edges, which

are less protected by CTAB. In this work, we explored the role

of iodide ions on rod growth, when using pre-grown gold

Co. KGaA, Weinheim Adv. Funct. Mater. 2008, 18, 3780–3786

FULLPAPER

M. Grzelczak et al. / Tuning Tip Curvature in Gold Nanorods

nanorods as seeds, on which additional gold salt was reduced,

in the presence of different amounts of iodide. The results

demonstrate that by changing iodide concentration different

morphologies can be obtained, ranging from thicker nanorods

(uniform growth) to well-defined, dumbbell-like particles,

which was confirmed by electron tomography. As expected, all

these morphological changes result in notable variations of the

LSP band position, providing us with a simple tool to tune the

optical properties, through controlled morphological changes,

but with no need to strongly affect the global dimensions of

the original rods. Detailed X-ray photoelectron spectroscopy

(XPS) and surface enhanced Raman spectroscopy (SERS)

studies provide important information related to the role of

iodide during growth.

2. Results and Discussion

The results reported here are related to the seeded growth

on pre-formed gold nanorods. All experiments were carried

out using the same amount of rods as seeds, as well as the same

concentrations of ascorbic acid and CTAB, as described in

Section 4. Since we were interested in studying the effect of

iodide ions on the growth of Au nanorods, we first carried out a

systematic variation of the amount of KI present during growth

(measured as [KI]/[Au0] molar ratio, [Au0] being the molar

concentration of Au metal in the nanorod seed solution), for a

constant [Au0]/[Au3þ] (¼ 0.6) molar ratio, so that the same

total volume increase were achieved. The UV–Vis–NIR

spectra of the corresponding final colloids are displayed in

Figure 1. It is immediately apparent from this figure that the

presence of low ([KI]/[Au0]< 1) and high ([KI]/[Au0]> 1)

amounts of KI lead to gradual red-shift and blue-shift of the

longitudinal plasmon band, respectively, as compared to the

spectrum of the starting rod (seed) solution, while intermediate

concentrations resulted in intermediate band positions. A

400 600 800 1000 12000.0

0.2

0.4

0.6

0.8

1.0

b Nor

m. A

bsor

banc

e

Wavelength (nm)

a seed solution

[KI]/[Au0]b 0.0c 3.66d 2e 0.8f 0.33g 0.038

=223 nm

a

Figure 1. UV–Vis–NIR spectra of gold nanorods before (a) and aftergrowth in the presence of different amounts of KI (see labels), keepingthe same Au0/Au3þ and ascorbic acid/Au3þ molar ratios.

Adv. Funct. Mater. 2008, 18, 3780–3786 � 2008 WILEY-VCH Verl

closer look at the transverse plasmon band (see Figs. S1 and S2,

Supporting Information) also reveals an increase in the

intensity of the band with increasing the size of the tips. For

both the longitudinal and transverse plasmon bands, the largest

red-shift was observed at a ratio as low as 0.04, producing an

LSP shift of 122 nm.

According to the well-known correlation between aspect

ratio and plasmon band position, the observed shifts seem to

indicate an increase in aspect ratio for low [KI] and a decrease

for high [KI]. However, TEM observation of the corresponding

nanoparticles (see Fig. 2) does not really confirm this

prediction. While in the presence of high iodide concentration,

the rods were indeed observed to grow uniformly, with the

expected decrease in aspect ratio, smaller amounts of added KI

induced preferential growth at the rod tips, while the central

part of the rods remained almost unaltered (at least for the

lowest KI concentration). The dumbbell (or peanut-like)

morphology of the final particles was confirmed by electron

tomography, which allowed us to obtain 3D representations of

the particles, as that shown in Figure 2h, and even rendering

movies that show rotation of the particle (see example in the

Supporting Information).

A summary of the dimensions and aspect ratios of the initial

rods and those obtained after growth in the presence of

different [KI]/[Au] ratios, is provided in the plots shown in

Figure 3 (see also Fig. S4 and Table S1 in the Supporting

Information). While the total length of all grown rods is of the

same order, the thickness values strongly depend on whether

they have been measured at the center of the rod or near the

tips. Accordingly, the values of aspect ratio can become very

different, depending on which thickness is considered.

Because of the morphological changes observed in the TEM

images, modeling of the optical response of these rods required

the use of a numerical method that allowed us to design different

shapes resembling the experimentally measured nanoparticles.

The boundary element method (BEM)[16,17] is well suited for

this, since it allows fast computation of the optical spectra when

the particles have a rotation axis, which can be assumed for the

differentmorphologiesobtainedinourexperiments, since inthis

way, only the outer edge needs to be parametrized. The models

for the different geometries of the particles grown in the

presence of various iodide concentrations are shown in the inset

of Figure 4, and should be compared with the TEM images in

Figure 2. Using these models, complete UV–Vis–NIR spectra

couldbecalculated,whichareplottedinFigure4.Comparisonof

these plots with the experimental spectra shown in Figure 1

indicates that the obtained agreement is very good, even for

these simple geometrical models. Apart from small deviations

between experimental and calculated band positions (see

Supporting Information, Fig. S2), the agreement is excellent

and confirms that the plasmon resonance can be easily

manipulated through iodide-mediated growth. It should be

stressed that, even though the three spectra corresponding to

dumbbell-like nanoparticles that have the same nominal aspect

ratio, small changes in the extent and curvature of the tips lead to

LPB shifts of up to 130 nm.

ag GmbH & Co. KGaA, Weinheim www.afm-journal.de 3781

FULLPAPER


Figure 2. TEM micrographs presenting the initial gold rods (a), and rods grown in the absence (b), and in thepresence (c–g) of KI, with the amount of KI decreasing from (c) to (g). Electron tomography 3D-reconstruction(h) of a single Au dumbbell from the sample shown in (g).

3782

A second set of experiments was carried out, in which the

added amount of seeds was varied, while maintaining a

constant [KI]/[Au] ratio. In order to perform a more

exhaustive analysis, two series of samples were prepared, with

the higher and lower [KI] defined in Figures 1–4, which

produced very different morphologies. Representative TEM

images of the resulting nanoparticles are shown in Figure 5.

60

70

80

90

X

Asp

ect r

atio

nm

15

20

25

X

tip middle

Width

Length

0 1 2 3 4

3

4

5

6

7

X

[KI]/[Au0]

Figure 3. Values of length, width, and aspect ratio for grown gold nanor-ods, as a function of [KI]/[Au] molar ratio. These measurements are basedon the TEM images shown in Figure 2. The crosses correspond to the initialgold nanorods.

Figure 4. Calculated (BEM) extinctwith similar morphologies to thoseSchematic drawing of the geometr

www.afm-journal.de � 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Again, dumbbell-like particles

were only obtained for low [KI],

regardless of the number of

seeds, while for high [KI], uni-

form growth is consistently

observed. In both cases, when

lower concentrations of seeds

were used, bigger particles were

obtained. It is clear thus that the

dominating parameter defining

the final shape is the amount of

iodide related to the number of

seed nanorods, which is likely to

be related to the preferential

adsorption of iodide ions on

{111} faces, as compared to

{110} and {100}. In this way,

low amounts of iodide would

inhibit growth mainly at the

lateral sides and the very tip

ends (top and bottom faces of the rod), while growth would

take place at the rounding part of the tips, resulting in

dumbbell formation. When the amount of iodide is increased,

adsorption must necessarily become more homogeneous and

thereby growth takes place progressively closer to homo-

geneous growth, as observed both in the absence of KI and in

the presence of a large KI concentration. It should be clear

that, inhibition of Au growth on iodide-containing surfaces

does not mean that there is no growth, but rather that it takes

place at a slower rate.

The UV–Vis–NIR spectra of these series of samples (Fig. S3,

Supporting Information) reveals a gradual blue-shift of the

longitudinal plasmon band for high iodide content, which is

consistent with a larger increase of the rod thickness as

compared to length (decrease of aspect ratio), as the amount of

seeds is decreased, while for low iodide content, the dumbbell

morphology is retained but the initial red-shift observed for the

ion spectra of Au rods and dumbbellsobtained experimentally (Fig. 2). Inset:ies used for BEM spectral simulation.

Adv. Funct. Mater. 2008, 18, 3780–3786

FULLPAPER


Figure 5. TEM images of nanoparticles resulting from the growth of Aunanorods for two [KI]/[Au] ratios: 3.6 (left) and 0.083 (right). The relativeamount of seeds employed for the growth decreases from top to bottom,so that larger particles are obtained.

selective growth at the tips, is reverted when the particles get

thicker and the aspect ratio is globally reduced.

Following on the effort to propose a mechanism that accounts

for the iodide-induced change in tip curvature of gold nanorods,

we need to focus on the crystalline structure and the surface

chemistry of the rods, both before and after iodide addition. The

initial rods are enclosed within eight {110} and {100} alternating

lateral facets, their tips being terminated by {100}, {110}, and

{111} facets.[18] For similar nanorods, elemental analysis carried

out by Orendorff and Murphy[19] showed that the gold nanorods

indeed contain silver, in a percentage that can vary from 2.5 up to

4.3%, depending on the experimental conditions. These authors

suggested that silver metal is mostly distributed on the surface of

the particles, in accordance with the underpotential deposition

model proposed by Liu and Guyot-Sionnest[8] for gold nanorod

growth. We carried out surface characterization of the nanorods

by means of XPS and SERS. The XPS survey spectrum of the

initial nanorods (Fig. S5, Supporting Information) shows the

presence of bromide and clearly demonstrates the presence of

silver. Quantification of the percentage of silver compared to

gold reveals a higher percentage of silver of around 9%, as

compared to the values previously reported from elemental

analysis. This confirms that a higher percentage of silver atoms is

present at or close to the surface of the particles. Additionally, as

will be shown below, the silver content remains constant after KI


addition, while bromide ions are completely removed and

replaced by iodide. For the initial rods (Fig. S5B, Supporting

Information) the characteristic energy for gold (Au-4f7/2),

usually at 83.8 eV,[20] was found to be slightly shifted to 83.48 eV.

This shift may be attributed to the formation of Au-Br species at

the surface of the particles, which is consistent with the

observation of a smaller shift in the Au-4f7/2 band after the

addition of KI (83.54 eV), because of the increase in the covalent

character of the gold halide bond.[21] The characteristic silver

peaks (Ag-3d5/2) were also shifted toward lower energies, from

their normal energy of 368 eV[22] to 367.34 and 367.49 eV for Ag-

Brand Ag-I, respectively (Fig. S5C,Supporting Information). In

this case, deconvolution of the complex signal is safer than for

gold, since silver remains as a minority component, around 9%

(related to gold), in both materials. Figures S5D and E in the

Supporting Information show that Ag-3d5/2 bands comprise two

contributions: one at 368 eV, with lower intensity and a binding

energy common to both samples, which is ascribed to metallic

silver; whereas the second, with higher intensity, is located at

367.34 eV for Ag-Br and 367.54 eV for Ag-I components,

respectively. Conclusive evidence of the presence of bromide

and iodide ions bonded to the metallic surface of the rods was

obtained by careful deconvolution of their Br-3d5/2 and I-3d5/2

characteristic bands, as well as by SERS (Fig. 6). Deconvolution

of Br-3d5/2 bands (Fig. 6A) shows the presence of two

contributions at 67.7 and 67.3 eV for gold and silver bromides,

respectively.[23] In the case of iodide (Fig. 6B), the I-3d5/2 peak

also shows two contributions at 620.1 and 619.4 eV, which can be

unambiguously assigned to gold and silver iodide, respec-

tively.[24,25] Notably, the silver halide signal is in both cases much

larger than that of gold, suggesting that Ag ions are concentrated

at the surface.

Additional evidence of the presence of Ag and Au halides

on the surface of the rods was gained through SERS

spectroscopy. SERS measurements were performed on the

different rod samples and on citrate reduced gold and silver

colloids, which were previously exposed to 10�3 M solutions of

KBr or KI (Fig. 6C). SERS spectra of Au nanoparticles display

broad and intense peaks at 196 and 160 cm�1, which are

attributed to Au-Br and Au-I stretchings. In the case of silver,

both SERS spectra show complex peaks at 152 and 121 cm�1,

assigned to Ag-Br and Ag-I. In both cases, the obtained SERS

results agree well with published data[26,27] and show that the

halides are coordinated to the metallic surface. Spectra for

both nanorod samples (before and after KI addition) show

patterns corresponding to the combination of the correspond-

ing halides. Additionally, after KI addition, no Br signals

remain. This absence of Br ions on the sample treated with

KI is very likely related to the higher stability of gold or

silver iodides over their respective bromides, in full agree-

ment with their solubility products, Ks(AgI)¼ 3.90� 10�17,

Ks(AgBr)¼ 2.71� 10�13,[25] Ks(AuI)¼ 10�16, and Ks(AuBr)¼10�12.[28]

Based on this characterization, and assuming the single

crystalline structure of the nanorods we can propose a growth

mechanism based on the following points:


FULLPAPER


Figure 6. High resolution XPS characteristic bands and their deconvoluted gold and silvercontributions for (A) Br 3d (in the initial rods) and (B) I 3d (after KI addition). (C) SERS spectraof the initial rods, and after treatment with KI. For comparison, an analytical blank was preparedby adding either KI or KBr solutions to regular gold or silver citrate reduced nanospheres. Notethat in both cases the rods show a complex signal that is composed of contributions from bothsilver and gold halides.

3784

– Silver is present mostly on the surface of the gold nanorods,

as demonstrated by elemental analysis,[19] XPS and SERS

(Fig. 6).

– Iodide ions have higher binding affinity to gold than bromide

ions. This was also confirmed by XPS and SERS analysis.

– Iodide ions have different binding affinity to different gold

crystalline facets. Iodide ions adsorb preferentially onto

{111} facets.[11]

– The surface redox potential of gold nanorods is affected by

adsorption of AgI(UPD) or AgBr(UPD). Chumanov and

coworkers[29] have reported a decrease in the surface redox

potential of silver nanoparticles when iodide ions were

450 600 750 900 10500.0

0.2

0.4

0.6

0.8

C-With KI

B-No KI

Ab

sorb

ance

Wavelength [nm]

A-Rod Initial

Figure 7. UV–Vis–NIR spectra of gold nanorods prepared without silvernitrate (A) and further overgrowth by gold shell in the absence (B) andpresence (C) of KI.

www.afm-journal.de � 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

adsorbed, replacing bromide, which in

their case would favor the reduction of

cytochrome C.

Based on these findings we suggest that

when a small amount of KI is present in

solution, the exchange of bromide by iodide

takes place at the tips, where {111} facets are

presentduetothehigheraffinityof iodideions

to these facets, as compared to {110} and {100}.

Therefore, the tips enclosed by {111} facets

contain AgI and AuI on their surface, while in

the lateral faces bromide ions remain

adsorbed. Subsequently, when ascorbic acid

is added, the nanorods are cathodically

polarized and gold ions get catalytically

reduced at the surfaces where the surface

redox potential is lower,[29] that is, AgI coated

surfaces at the tips, leading to dumbbell

formation. Conversely, from the results

obtained with excess KI present in solution,

full overcoating would result from complete

adsorption of iodide on the whole surface of

the nanorods, i.e., equal surface redox poten-

tial at the tips and the sides of the nanorods.

Although this seems to contradict the

results presented by Ha et al.,[11] where preferential adsorption

of iodide on {111} gold facets was claimed to block growth in this

direction and end up with triangular shapes, we need to point out

that the seeds used in that case were Ag-free, while there is Ag

presentat thesurfaceofourgoldnanorodseeds.The influenceof

silver on tip growth was confirmed through a control experiment

in which penta-twinned gold nanorods (grown in the absence of

silver ions[14])wereusedasseeds.Theserodshave{111} facetson

the tips and {100} facets on the sides.[30] Growth experiments on

these new seeds were carried out, both in the presence and in the

absence of KI, observing in both cases a blue-shift of the LSP

band (see Fig. 7), which suggests a decrease of the aspect ratio.

Interestingly, in the presence of KI, the shift is considerably

larger (ca. 100 nm) than in the absence of KI (40 nm), while TEM

analysis shows that no dumbbell shapes were formed in either

case (Fig. S6, Supporting Information). Statistical analysis of the

dimensions show that, while the thickness increased in a similar

extent in both cases (from �24 up to �32 nm), the length

increased when no KI was present (from 109 up to 114 nm),

however it remained basically constant in the presence of KI.

This indicates again that iodide ions block {111} facets at the tips,

as expected for Ag-free rod seeds.

3. Conclusions

The growth of gold nanorods can be notably modified

through the presence of tiny amounts of iodide, in such a way

that tip growth can be greatly enhanced, resulting in the

formation of well-defined dumbbell morphologies. Our results

Adv. Funct. Mater. 2008, 18, 3780–3786

FULLPAPER


indicate that iodide can readily replace bromide ions from the

rod surface, and this replacement is more favorable at {111}

facets, which are only present at the tips. This simple procedure

becomes a solid strategy for tuning nanorod morphology, and

in turn the optical response of the system. Optical spectroscopy

characterization of the different morphologies indicates that

growth of dumbbell structures leads to strong shifts of the

longitudinal plasmon resonance (and smaller shifts of the

transverse resonance), which were found to be in good

agreement with numerical modeling based in the BEM.

Applications can be foreseen in various fields, but importantly

in LSPR biosensor design, SERS, and development of probes

for hyperthermia, since the longitudinal plasmon can be driven

to the NIR region without the need for largely increasing

particle size or aspect ratio.

4. Experimental

Gold nanorods were prepared through the well-known seededgrowth method [7,8], based on the reduction of HAuCl4 with ascorbicacid on CTAB-stabilized Au nanoparticle seeds (<3 nm), in thepresence of CTAB (0.1 M), HCl (pH 2–3), and AgNO3 (0.12 mM).Upon synthesis, the gold nanorod solution (10 mL) was centrifuged(8000 rpm, 30 min) to remove excess silver salt, ascorbic acid and HCl,and redispersed in CTAB solution (2 mL, 0.1 M), so that theconcentration of the gold nanorods in the seed solution used forfurther growth was 2.5 mM.

For the growth of gold nanorods in the presence of KI, anappropriate volume of 0.1 M CTAB (calculated for a total volume of10 mL) was mixed with HAuCl4 (0.05 mL, 0.05 M) and stored for 5 minat 27 8C to allow for complexation of gold salts, followed by addition of0.01 M KI (5.7, 50, 120, 300, 550mL) and ascorbic acid (0.04 mL, 0.1 M).Finally, a gold nanorod seed solution (0.6 mL, 2.5 mM) was addedunder stirring.

To examine the influence of gold nanorod seed concentration (atconstant [KI] / [Au] molar ratio), different growth solutions wereprepared with identical concentrations of CTAB (0.1 M) and HAuCl4(0.25 mM), but different KI concentrations chosen to fit the selected[KI]/[Au] molar ratios (3.6 and 0.083). To each of these solutions(10 mL), ascorbic acid (0.04 mL, 0.1M) was added, followed bydifferent amounts of gold nanorods (1, 0.6, 0.3, 0.1 mL; [Au] = 2.5 mM).

Optical characterization was carried out by UV–Vis–NIR spectro-scopy with a Cary 5000 spectrophotometer, using 10 mm path lengthquartz cuvettes. Transmission electron microscopy (TEM) imageswere obtained with a JEOL JEM 1010 transmission electronmicroscope operating at an acceleration voltage of 100 kV. For theelectron tomography measurements, data were acquired with a JEOLJEM 2010F microscope, operating at 200 kV, using a Gatan 912 ultrahigh tilt tomography holder. Image acquisition was carried out withmultiscan Gatan camera and the Digital Micrograph software. Thetotal tilt angle was 1258 (from �578 to þ 688), taking 125 images, oneevery 18. Image alignment was performed manually using Midassoftware, while for reconstruction the TOMOJ package was used. XPSanalysis of the samples was performed using a VG Escalab 250 iXLESCA instrument (VG Scientific), equipped with aluminum Ka1.2monochromatic radiation at 1486.92 eV X-ray source. SERS wasmeasured with a LabRam HR system (Horiba-Jobin Yvon), equippedwith a confocal optical microscope, high resolution gratings(1800 g mm�1) and a Peltier CCD detector. For the sample excitation,a near-infrared laser line (785 nm) and low power at the sample (7mW)and collection times (1 s) were used to avoid halide decomposition.Analytical blanks for SERS were prepared by adding either KBr or KI(10mL, 10�2 M) to gold or silver nanospheres (1 mL, prepared by


citrate reduction), giving rise to final halide concentrations of 10�4 M.SERS spectra of the different rod and blank samples were directlycollected from the liquid suspensions by using a macrosamplingaccessory attached directly to the microscope. Simulation of opticalspectra was based on the BEM. In the BEM, scalar and vectorpotentials f and A are used and they are expressed inside eachhomogeneous region of space (e.g., region j) as the sum of an externalfield contribution (i.e., the potentials corresponding to the incidentlight plane wave, for which the scalar potential can be chosen as zero)and surface integrals involving the noted equivalent boundary chargessj and currents hj, defined on the boundary of region j, Sj. For water, theindex of refraction was taken constant and equal to 1.333, the value at600 nm, while for gold tabulated, frequency-dependent dielectricfunctions were used [31]). The particles were described by axiallysymmetric shapes capturing the main physical aspects of their responseto external illumination. Geometrical models were devised on the basisof nanoparticles’ shape observed in the TEM images. The circularedges are rounded, in order to avoid sharp corners that can causenumerical problems. Convergence to the point where differences innumerical results cannot be resolved on the scale of the figures whenchanging the number of parametrization points N has been achievedfor N¼ 150.

Received: May 22, 2008Published online: September 9, 2008

[1] J. Perez-Juste, I. Pastoriza-Santos, L. M. Liz-Marzan, P. Mulvaney,

Coord. Chem. Rev. 2005, 249, 1870.

[2] X. Huang, I. H. El-Sayed, W. Qian, M. A. El-Sayed, J. Am. Chem. Soc.

2006, 128, 2115.

[3] L. Tong, Y. Zhao, T. B. Huff, M. N. Hansen, A. Wei, J.-X. Cheng, Adv.

Mat. 2007, 19, 3136.

[4] R. S. Norman, J. W. Stone, A. Gole, C. J. Murphy, T. L. Sabo-

Attwood, Nano Lett. 2008, 8, 302.

[5] L. R. Hirsch, R. J. Stafford, J. A. Bankson, S. R. Sershen, B. Rivera, R.

E. Price, J. D. Hazle, N. J. Halas, J. L. West, Proc. Natl. Acad. Sci. USA

2003, 100, 13549.

[6] B. Busbee, S. Obare, C. J. Murphy, Adv. Mater. 2003, 15, 414.

[7] B. Nikoobakht, M. A. El-Sayed, Chem. Mater. 2003, 15, 1957.

[8] M. Liu, P. Guyot-Sionnest, J. Phys. Chem. B 2005, 109, 22192.

[9] C. J. Murphy, T. K. Sau, A. M. Gole, C. J. Orendorff, J. Gao, L. Gou,

S. E. Hunyadi, T. Li, J. Phys. Chem. B 2005, 109, 13857.

[10] T. K. Sau, C. J. Murphy, J. Am. Chem. Soc. 2004, 126, 8648.

[11] T. H. Ha, H.-J. Koo, B. H. Chung, J. Phys. Chem. C 2007, 111, 1123.

[12] L. Gou, C. J. Murphy, Chem. Mater. 2005, 17, 3668.

[13] X. Xu, M. B. Cortie, Adv. Funct. Mater. 2006, 16, 2170.

[14] J. Perez-Juste, L. M. Liz-Marzan, S. Carnie, D. Y. C. Carnie,

P. Mulvaney, Adv. Funct. Mater. 2004, 14, 571.

[15] J. Gao, C. M. Bender, C. J. Murphy, Langmuir 2003, 19, 9065.

[16] F. J. Garcıa de Abajo, A. Howie, Phys. Rev. Lett. 1998, 80, 5180.

[17] F. J. Garcıa de Abajo, A. Howie, Phys. Rev. B 2002, 65, 115 418.

[18] Z. L. Wang, M. B. Mohamed, S. Link, M. A. El-Sayed, Surf. Sci. 1999,

440, 809.

[19] C. J. Orendorff, C. J. Murphy, J. Phys. Chem. B 2006, 110, 3990.

[20] D. Briggs, M. P. Seah (Eds.), Practical Surface Analysis by Auger and

X-ray Photoelectron Spectroscopy, 2nd edition, Wiley & Sons,

Chichester, UK 1992.

[21] H. Kitagawa, N. Kojima, T. Nakajima, J. Chem. Soc., Dalton Trans.

1991, 3121.

[22] J. F. Moulder, W. F. Stickle, P. E. Sobol, K. D. Bomben, Handbook of

X-ray Photoelectron Spectroscopy, Perkin-Elmer Corp., Eden Prairie,

MN, USA 1992.


FULLPAPER


3786

[23] C. D. Wagner, J. Vac. Sci. Technol. 1978, 15, 518.

[24] A. Glidle, T. Yasukawa, C. S. Hadyoon, N. Anicet, T. Matsue,

M. Nomura, J. M. Cooper, Anal. Chem. 2003, 75, 2559.

[25] M. Kawasaki, H. Ishii, Langmuir 1995, 11, 832.

[26] B. H. Loo, J. Phys. Chem. 1982, 86, 433.

[27] G. Chumanov, M. S. Sibbald, M. T. Cotton, J. Phys. Chem. B 1998, 102,

10836.

www.afm-journal.de � 2008 WILEY-VCH Verlag GmbH

[28] F. Burriel, F. Lucena, S. Arribas, J. Hernandez, Quımica Analıtica

Cualitativa, 16th edition, Paraninfo, Madrid 1998.

[29] M. S. Sibbald, G. Chumanov, T. M. Cotton, J. Phys. Chem. 1996 100,

4672.

[30] M. Hu, P. Hillyard, G. V. Hartland, T. Kosel, J. Perez-Juste,

P. Mulvaney, Nano. Lett. 2004, 4, 2493.

[31] P. B. Johnson, R. W. Christy, Phys. Rev. B 1972, 6, 4370.

& Co. KGaA, Weinheim Adv. Funct. Mater. 2008, 18, 3780–3786

1

Influence of Iodide Ions on the Growth of Gold Nanorods: Tuning Tip

Curvature and Surface Plasmon Resonance

By Marek Grzelczak, Ana Sánchez-Iglesias, Benito Rodríguez-González, Ramón Alvarez-Puebla, Jorge Pérez-Juste*, and Luis M. Liz-Marzán*

Departamento de Química Física and Unidad Asociada CSIC-Universidade de Vigo,

Vigo 36310 (Spain)

* E-mail: [email protected]; [email protected]

Supporting Information

400 500 6000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

400 500 6000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Nor

m. A

bsor

banc

e

Wavelength / nm

KIa)

Seed Fig.1d Fig.1e Fig.1g

b)

Nor

m. C

ext

Figure S1. (a) Evolution of TPR for samples at different amount of KI as indicated in the text

(Figure 1). (b) Calculated spectra zoomed from Figure 4. In both cases a broadening of

plasmon band as the amount of KI decreases is observed due to spherical shape on the tips

(see Figure 2).

2

750

800

850

900

950

1000

1050

LS

P /

nm

[KI]/[Auo]

0 1 2 3 4

508

512

516

520

524

TS

P /

nm

Experimental Calculated

Figure S2. Peak positions of transverse (bottom) and longitudinal (top) plasmon bands of

grown nanorods, as a function of the KI/Au0 molar ratio. The black and the red points

correspond to the experimental and calculated data respectively.

3

400 600 800 1000 12000

1

2

3

4

5

[KI]/[Au0] = 0.083

Nor

m. A

bsor

banc

e

W avelength (nm)

[KI]/[Au0] = 3.6

a

600 800 1000 1200

A - Rod InitialB - Au0/Au3+=1C - Au0/Au3+=0.6D - Au0/Au3+=0.3

b

Figure S3. UV-vis-NIR spectra of gold nanorods grown using different amounts of seeds in

the solution, for two different [KI]/[Au] ratios, as indicated: 3.6 (L) and 0.083 (R). The

amount of KI was adjusted to the concentration of Au0 (seed rods) in solution. Representative

TEM images of the corresponding nanoparticles are shown in Figure 5.

4

L1- total length

L2- length on the side (overgrowing)

L3- length of spherical tip (tipgrowing)

W1- with on the tip

W2- with in the middle

C- radius of tip curvature (overgrowing)

Fig S4. Scheme showing two extreme allowed shapes. Upper: over-growth and lower: tip-growth.

Au0/Au3+=0.6 ; KI/Au0(seed) = A) Au0/Au3+ =

B); Au0/Au3+ =

Rod

Initial

No KI

0.6 A)-3.6 2 0.8 0.3 B)0.038 1 0.6 0.3

1 0.6 0.3

L1 63.8±8.6 78.9±7.6 76.1±7.3 78.0±6.3 83.3±7.1 82.8±7.5 82.0±8.9 75.8±8.2 76.4±8.0 86.8±7.7 77.6±7.7 81.4±8.3 94.2±8.2

L2 56.5±8.7 64.8± .9 60.2± .3 61.0± .1 - - - 63.8± .2 60.4± .9 66.2±8.3 - - -

L3 - - - - 33.8± .9 30.9±2.2 26.3±3.4 - - - 22.5±2.0 27.0±2.2 35.2±2.9

W1 13.7±1.6 20.4±1.6 21.6±1.6 21.4±1.5 23.3±1.7 23.9±1.4 24.9±2.3 18.0±2.0 21.5±1.7 26.3±1.8 22.7±1.9 25.2±2.1 32.6±3.0

W2 13.3±1.5 20.4±1.5 21.5±1.6 21.3±1.4 15.4±1.9 14.5±1.4 13.6±1.8 16.3±1.6 21.5±1.6 27.0±1.7 14.5±1.7 13.5±1.9 15.1±1.6

C (L1-L2)/2 3.6±0.6 7.1±0.9 8.0±1.2 8.5±1.0 - - - 6.0±1.0 8.0±1.2 10.3±1.6 - - -

AR rod L1/W2 4.8±0.9 3.9±0.5 3.5±0.4 3.7±0.4 5.5±1.1 5.8±0.8 6.1±1.0 4.7±0.7 3.6±0.4 3.2±0.4 5.4±1.0 6.2±1.2 6.3±1.1

AR tipL3/W1 - - - - 1.4±0.1 1.3±0.1 1.0±0.1 - - - 1.00±0.1 1.1±0.1 1.1±0.2

TPB 507 510 510 510 515 519 524 511 510 511 518 525 524

LPB 913 815 808 829 911 975 1031 891 807 754 999 1024 990

Table S1. Gold nanorod dimensions as indicated in Figure S4 (see text for details).

5

Ag-

3s

1200 1000 800 600 400 200

Binding Energy/eVI-

4s

Na-

1s

O-1s

O-2

s

O-K

LL

C-1sI-

3s/N

a-1s

I-3d

3/2

I-3d

5/2

Br-

3d3/

2/B

r-3d

5/2

Ag-

MN

N

Ag-

3d3/

2A

g-3d

5/2

Ag-

3p3/

2

Ag-

3p1/

2

Au-

4f5/

2

Au-4f7/2

Au-

4d3/

2 Au-

4d5/

2

Si-2s

Si-3

p 3/2S

i-3p 1

/2

Initial rods

Rods after the addition of KI

Ag-

3s

1200 1000 800 600 400 200

Binding Energy/eVI-

4s

Na-

1s

O-1s

O-2

s

O-K

LL

C-1sI-

3s/N

a-1s

I-3d

3/2

I-3d

5/2

Br-

3d3/

2/B

r-3d

5/2

Ag-

MN

N

Ag-

3d3/

2A

g-3d

5/2

Ag-

3p3/

2

Ag-

3p1/

2

Au-

4f5/

2

Au-4f7/2

Au-

4d3/

2 Au-

4d5/

2

Si-2s

Si-3

p 3/2S

i-3p 1

/2

Ag-

3s

1200 1000 800 600 400 200

Binding Energy/eVI-

4s

Na-

1s

O-1s

O-2

s

O-K

LL

C-1sI-

3s/N

a-1s

I-3d

3/2

I-3d

5/2

Br-

3d3/

2/B

r-3d

5/2

Ag-

MN

N

Ag-

3d3/

2A

g-3d

5/2

Ag-

3p3/

2

Ag-

3p1/

2

Au-

4f5/

2

Au-4f7/2

Au-

4d3/

2 Au-

4d5/

2

Si-2s

Si-3

p 3/2S

i-3p 1

/2

Initial rods


A

Ag-

3s

1200 1000 800 600 400 200

Binding Energy/eVI-

4s

Na-

1s

O-1s

O-2

s

O-K

LL

C-1sI-

3s/N

a-1s

I-3d

3/2

I-3d

5/2

Br-

3d3/

2/B

r-3d

5/2

Ag-

MN

N

Ag-

3d3/

2A

g-3d

5/2

Ag-

3p3/

2

Ag-

3p1/

2

Au-

4f5/

2

Au-4f7/2

Au-

4d3/

2 Au-

4d5/

2

Si-2s

Si-3

p 3/2S

i-3p 1

/2

Initial rods


Ag-

3s

1200 1000 800 600 400 200

Binding Energy/eVI-

4s

Na-

1s

O-1s

O-2

s

O-K

LL

C-1sI-

3s/N

a-1s

I-3d

3/2

I-3d

5/2

Br-

3d3/

2/B

r-3d

5/2

Ag-

MN

N

Ag-

3d3/

2A

g-3d

5/2

Ag-

3p3/

2

Ag-

3p1/

2

Au-

4f5/

2

Au-4f7/2

Au-

4d3/

2 Au-

4d5/

2

Si-2s

Si-3

p 3/2S

i-3p 1

/2

Ag-

3s

1200 1000 800 600 400 200

Binding Energy/eVI-

4s

Na-

1s

O-1s

O-2

s

O-K

LL

C-1sI-

3s/N

a-1s

I-3d

3/2

I-3d

5/2

Br-

3d3/

2/B

r-3d

5/2

Ag-

MN

N

Ag-

3d3/

2A

g-3d

5/2

Ag-

3p3/

2

Ag-

3p1/

2

Au-

4f5/

2

Au-4f7/2

Au-

4d3/

2 Au-

4d5/

2

Si-2s

Si-3

p 3/2S

i-3p 1

/2

Initial rods


A

380 375 370 365 360

94 90 86 82 78 370 368 366 364

370 368 366 364Binding Energy/eV Binding Energy/eV

Au-4f7/2Au-4f5/2

Ag-3d3/2

Ag-3d5/2

Ag-3d5/2

Ag-3d5/2

Initial rods

Initial rods

Initial rods




380 375 370 365 360 380 375 370 365 360

94 90 86 82 78 94 90 86 82 78 370 368 366 364


Au-4f7/2Au-4f5/2

Ag-3d3/2

Ag-3d5/2

Ag-3d5/2

Ag-3d5/2

Initial rods

Initial rods

Initial rods




B

C

D

E

380 375 370 365 360

94 90 86 82 78 370 368 366 364


Au-4f7/2Au-4f5/2

Ag-3d3/2

Ag-3d5/2

Ag-3d5/2

Ag-3d5/2

Initial rods

Initial rods

Initial rods




380 375 370 365 360 380 375 370 365 360

94 90 86 82 78 94 90 86 82 78 370 368 366 364


Au-4f7/2Au-4f5/2

Ag-3d3/2

Ag-3d5/2

Ag-3d5/2

Ag-3d5/2

Initial rods

Initial rods

Initial rods




B

C

D

E

Figure S5. XPS survey spectra for the initial rods and after the addition of KI (A) HR-XPS

spectra of gold (B) and silver (C) for the initial rods and after the addition of KI. Ag-3d5/2

deconvolution for the initial rods (D) and after the addition of KI (E).

6

Figure S6. Representative TEM micrographs of gold nanorods prepared without silver nitrate,

before (A) and after overgrowth in the absence (B) and in the presence (C) of KI.

Influence of Iodide Ions on the Growth of Gold Nanorods: Tuning Tip Curvature and Surface Plasmon Resonance

Documents