J. Phys. Chem.

C 2007, 111, 1123-1130 1123

Shape-Controlled Syntheses of Gold Nanoprisms and Nanorods Influenced by Specific

Adsorption of Halide Ions

Tai Hwan Ha,* Hee-Joon Koo, and Bong Hyun Chung*

BioNanotechnology Research Center, Korea Research Institute of Bioscience and Biotechnolgoy (KRIBB),
Daejeon 305-806, Korea
ReceiVed: October 2, 2006; In Final Form: NoVember 14, 2006

This paper describes the effect of halide ions during the seed-mediated growth of gold nanoparticles employing
cetyltrimethylammonium bromide (CTAB) as a cationic surfactant system. With the addition of a small amount
of iodide ion (∼20 µM) in a growth solution, the major product of the gold nanostructures formed were
notably changed into triangular nanoprisms in the presence of excessive bromide ion (∼0.1 M); otherwise,
in its absence, nanorods with an aspect ratio of ∼11 were the main products. The major role of the iodide ion
was in retarding the overall rate of crystal growth, and the iodide adsorption appeared to repress the crystal
growth along Au(111) direction, resulting in Au(111)-faced triangular nanoprisms. When the counteranions
of the surfactant were replaced with chloride ions, a novel nanostructure (i.e., nanorice) was manufactured,
which demonstrates the effectiveness of the adsorption of halide ions. However, this finding is quite contrasted
with the work of the Sastry group (J. Nanosci. Nanotechnol. 2005, 5, 1721-1727), wherein iodide ions
strongly suppress the formation of nanoprism. The distinctive results are attributed to different experimental
conditions for reducing gold precursors. Nonetheless, overall these observations suggest that the specific
adsorption of halide ions is an important factor for a complete control over the shape developments in the
seed-mediated growth of gold crystals.

1. Introduction trochemical methods.30 Murphy et al. and other groups have

The variety of the optoelectronic properties of metal nano- developed a seed-mediated growth procedure to create gold
materials mainly originates from the diversity of their shapes nanorods in which cetyltrimethylammonium bromide (CTAB)
and sizes in addition to their inherent electronic properties. For is used as a major shape-directing organic template, with
a decade, the main stream of research efforts has focused on ascorbic acid used as a major reductant.15,16 Judicious control
novel routes for isotropic nanoparticles or on the deliberate over the composition of the surfactant,16,31 the preparation of
control of a high production yield and improved homogeneity.1-3 the seed solution,32 and the ratios of additives33 in the growth
Isotropic alloy nanoparticles composed of two or more metal solution rendered the creation of the gold nanorods with diverse
ingredients were also exploited to tune their optical properties.4-7 aspect ratios, higher production yields, and specific morphol-
In addition to ramifying the composition, shape-controlled (thus ogies. However, the manufacture of gold nanomaterials with
anisotropic) metal nanostructures have been highlighted recently, different shapes appears more of an art than a science, as batch-
mainly due to their unusual optical properties, i.e., Raman to-batch experiment yielded the desired nanocrystal only with
scattering at the surface of the nanostructures or diverse apparent varying degrees of reproducibility. These previous investigations
colors that stem from their distinct sizes and shapes.8-16 The largely suggest that the reaction conditions or mechanisms for
manufacture of these nanomaterials has been a starting point the modulation of crystal shapes are not well understood to the
for the development of innovative biosensors,17-20 optoelectronic extent necessary for the mass production of nanocrystals with
devices,21 bioimaging,22 and biochip applications.23 Concomi- desired morphologies.
tantly, more advanced metal nanostructures that were compos- Herein, in the seed-mediated crystal growth, a description of
ited with other inorganic nanoparticles or aligned superstructures the manufacture of mainly two types of gold nanostructures such
of different nanoparticles have been attempted to tackle future as triangular nanoprism or linear nanorod is given, along with
challenges in nanoelectronics and in biosensor applications.24-27 that of spherical faceted nanoparticles as a ubiquitously present
In an effort to control the shape of nanocrystals, diverse byproduct. The presence of distinct halide ions and their molar
synthetic routes has been explored for nanocubes,13,14 hollow ratios in this reaction condition resulted in the formation of
spheres,20 nanorods (or nanowires),15,16 and triangular prisms9-11 diverse morphologies, i.e., spheroid, triangular (or irregularly
from inorganic precursors of silver or gold. Their manufacturing circular) nanoplates, or nanorods. In particular, a minuscule
protocols have been continually improved by elaborate controls
amount of iodide ion was crucial for the formation of a triangular
over reaction parameters related to the production yield,
disk (and a circular disk at higher concentration) instead of a
homogeneity, cost, and environmental friendliness. For instance,
nanorod. In spite of the minimal amount of iodide ion, the
a gold nanorod (or nanowire) can be prepared by employing
dramatic change of the nanocrystal morphology is largely
conventional wet chemistry,15,16,28 photochemical,29 and elec-
attributed to the adsorption of iodide on low-indexed gold
* Corresponding authors. E-mail: [email protected] (T.H.H.); surfaces, allowing for a distinct growth rate of the present redox
[email protected] (B.H.C.). pair (HAuCl4-CTAB/ascorbic acid) on each facet.
10.1021/jp066454l CCC: $37.00 © 2007 American Chemical Society
Published on Web 12/29/2006
1124 J. Phys. Chem. C, Vol. 111, No. 3, 2007 Ha et al.

2. Experimental Section surfactant with a longer aliphatic surfactant and smaller seed
solution yielded longer nanorods. The presence of silver ions
2.1 Chemicals. Hydrogen tetrachloroaurate(III) trihydrate
dramatically improved the yield of the formation of a short gold
(HAuCl4‚3H2O, 99.9%), cetyltrimethylammounium bromide
nanorod to an extent greater than 95%,16,34 and a further
(CTAB, 99%), sodium borohydride (NaBH4, 99%), L-ascorbic
modulation of the relative ratio yielded nanoprism, hexagon,
acid (99%), potassium bromide (KBr, 99%), potassium iodide
cube, and even star-shaped branched nanoparticles at a lower
(KI, 99%), potassium cyanide (KCN, 97%), potassium thio-
cyanide (KSCN, 99+%), and cetyltrimethylammounium chlo- concentration of CTAB.33
ride (CTAC, 25 wt % solution in water) were purchased from Among the chemicals present in the seed-mediated growth
Aldrich (Milwaukee, WI) and used as received. Potassium of gold nanocrystals, the influences of ascorbic acid, CTAB,
chloride (KCl, 99.5%) and trisodium citrate dehydrate (99%) chloroaurate ion, and their ratios have been extensively
were purchased from Junsei (Japan) and Kanto (Japan). Deion- studied.15,16,31-34 Since several reports also treated the content
ized water was used throughout during the preparations of of halide ions (mainly chloride, bromide, and iodide ions) as a
aqueous solutions. systematic variable for the shape control of metal nanostructures,
2.2. Methods. 2.2.1. Preparation of the Gold Seeds. A 20 an investigation on the halide ion effect during the seed-mediated
mL volume of aqueous solution containing 2.5 × 10-4 M growth attracted our attention. These specific adsorptions are
HAuCl4 and 2.5 × 10-4 M trisodium citrate was prepared in a well-documented in the field of surface chemistry and electro-
conical flask. To this solution, 0.6 mL of ice-cold 0.1 M NaBH4 chemistry;35 all the halide ions are specifically adsorbed onto
solution was quickly added with vigorous stirring. The solution low-indexed gold surfaces (i.e., Au(111), Au(110), and Au(100))
became pink in color immediately after the addition of NaBH4 except for the fluoride ion. Moreover, it is well-known in
solution, and the solution was sustained for 2 min with vigorous colloidal chemistry that a counterion affects the process of
stirring. The seed particles in this solution were used as seeds micellar adsorption and/or the surface morphology of cationic
within 2-5 h after preparation. The average particle size surfactants residing on a solid surface.36,37 Assuming that the
measured from the transmission electron micrograph was 2-4 growth from a seed crystal would be governed by halide adlayers
nm.32 and/or CTAB bilayer structures, a halide ion in a reaction vessel
2.2.2. Growth of Gold Nanocrystals from the Seed Solutions. may be a more effective variable in the area of anisotropic
For the growth of nanorods, 10 mL of growth solution crystal growth. Inspired by this background information, the
containing 2.5 × 10-4 M HAuCl4 and 0.1 M cetyltrimethylam- influences of various halide ions upon the formation of
monium bromide (CTAB) was mixed with 0.2 mL of freshly anisotropic gold nanostructures were systematically investigated.
prepared ascorbic acid solution (0.1M) in a clean test tube, which In this study, only the first step of the seed-mediated growth
resulted in a colorless solution; the concentration of ascorbic procedure was examined, as the second and third steps are only
acid is four times as large as the value usually applied in the employed for the elongation of the nanocrystals formed in the
formation of nanorod and is adopted as an optimum for the previous step.
growth of nanoprisms. To this solution, 0.025 mL of the seed In the “original” crystal growth (as shown in Figure 1a), a
solution was added and gently mixed. No further stirring or gold nanorod was successfully synthesized, as predicted from
agitation was done during crystal growth. Within 5-10 min, the previous reports. The average aspect ratio was determined
the solution color became reddish in color. For the growth of to be 11 ( 3, and a comparable number of faceted nanoparticles
nanoprisms, the growth solutions were prepared by adding 10 and a relatively small number of triangular nanoprisms were
mL of a mixture of a solution identical with that used to simultaneously observed. As counted from TEM images, the
synthesize gold nanorods in addition to small concentrations yield of nanorod was about 42 ( 3% in number density and
of KI. To this solution, 0.2 mL of 0.1 M ascorbic acid was the yield of nanoprism was 10 ( 3% (see the Supporting
added, which resulted in a colorless solution. Then, 0.025 mL Information). As shown in the HRTEM image (Figure 1c), the
of seed was mixed with the solution. As a control experiment, observed nanorod was a twinned pentahedron rod along the 011
a comparable amount of KSCN or KCN (100 µM) was added direction with five or more {100}/{110} faces.9,38 The power
into the growth solution instead of KI. For the synthesis of spectrum of a corresponding image shows a pattern similar to
nanorices, the same solution as that used to synthesize gold that of the previously reported results; the diffraction spots in
nanorods was prepared containing cetyltrimethylammonium the power spectrum indicated typical two zone axes (i.e., [11h1]
chloride (CTAC) instead of CTAB. and [11h0]) due to the twinning.39-41 In its UV/vis spectra, the
2.3. Instrumentation. Absorption spectra of the solutions corresponding band for the nanorod could not be definitely
were taken with a DU800 UV/vis spectrophotometer (Beckman assigned in the present dynamic range (data not shown), but its
Coulter) in a wavelength range from 400 to 1100 nm. Transmis- presence could be recognized by a floating and rising baseline
sion electron microscopy (TEM) images were acquired with a over ∼1000 nm.
CM20 (Philips) electron microscope at 120 kV using a carbon- In contrast, the addition of a trace of iodide ion (∼20 µM in
coated copper grid. The high-resolution image samples were a final concentration) into the reaction vessel dramatically
examined under a Tecnai F30 super-twin (FEI) transmission changed the shape of the gold crystal, as evidenced in UV/vis
electron microscope at 300 kV. spectra and the corresponding TEM image (Figure 1b). In UV/
vis spectroscopic studies, as discussed in the later section, a
3. Results and Discussion new and broad band near ∼1000 nm can be assigned as the
Murphy et al. developed a three-step seed-mediated growth nanoprism band. In Figure 1b, a substantial number of nano-
procedure to manufacture gold nanorods with controllable aspect prisms with an edge length of 98 (17 nm were clearly
ratios.15,31-33 In this protocol, it was presumed that a cationic observable with a comparable number of faceted symmetric
surfactant, CTAB, acts as a soft template to direct the nano- nanoparticles (i.e., hexagons and pentagons); the production
structure. The crystal growth is controlled by the interplay of yield of nanoprism was estimated to be 45 ( 5% (see the
many kinetic and thermodynamic parameters to yield nanoc- Supporting Information). The observed faceted pentagons or
rystals with a desired morphology. For instance, employing a hexagons all had only Au(111) faces.39,42,43 Intriguingly, a gold
Shape-Controlled Syntheses of Gold Nanoprisms J. Phys. Chem. C, Vol. 111, No. 3, 2007 1125

Figure 2. (a) UV/vis spectra demonstrating the dependence of the

iodide concentration on the formation of nanoprism. (b) Kinetic UV/
vis spectra plotted at an interval of 150 s, showing the developments
of the nanoprism band in the presence of iodide ion (200 µM). In the
inset, the band intensities of faceted nanoparticles (at ∼530 nm) in the
presence of iodide ions (200 µM and 10 mM) were plotted versus time
Figure 1. TEM images of manufactured gold nanostructures (a) in and compared with the result in its absence. (c) UV/vis spectra showing
the absence of iodide ions and (b) in the presence of a trace amount of the pH dependence of nanoprism formation in a short pH range (2.2-
iodide ion (∼10 µM). The as-prepared gold nanocrystals were mildly 3.5). (d) TEM images showing typical morphology changes with the
centrifuged at ∼1000 rpm for 5 min to separate isotropic and faceted addition iodide ion (top images) and with pH variations (bottom images)
nanoparticles and were loaded on TEM grids. Also shown are high- during the formation of nanoprisms.
resolution TEM images of (c) a single nanorod and (d) a triangular
nanoprism obtained from (a) and (b), respectively, along with the slope decreased into a third of the value observed in the absence
corresponding power spectra (insets). The power spectrum of a nanorod
of iodide ions. Furthermore, the addition of iodide ions at a
(inset of (c)) shows typical diffraction patterns of pentatwinned nanorod
(two zone axes: [11h1] and [11h0]), while that of the triangular nanoprism high concentration altered the shape of the nanoprisms and
(inset of (d)) shows hexagonal diffraction patterns from the {220} planes resulted in rough edge characteristics demonstrated in Figure
(single zone axis: [11h1]). 2c (top images). The edge of nanoprism became somewhat
blunted or irregular, and even circular nanoplates were often
nanorod, which would have been a major product in the absence observed above an iodide concentration of ∼5 mM. This
of iodide ions, was barely observable in the presence of iodide observation is attributed to a possible multilayered solid iodine
ion. The normal orientation of the observed flat nanoprism was or similarly dissolution of metallic gold into an AuI salt
assigned as the 111 direction, as a hexagonal pattern on the formation.35 Overall these observations indicate that even a small
power spectrum was definitely observed, which corresponds to amount of iodide ion drastically suppresses the crystal growth
diffractions from {220} planes in the reciprocal space according of Au(111)-faced nanoprisms as well as faceted penta- or
to the basic fcc structure of metallic gold (a ) 0.4078 nm).39,42,43 hexatwinned nanoparticles. Although not shown here, addition
To investigate the effect of iodide ion more deeply, the of other anions such as CN- and SCN- (∼100 µM), which are
concentration dependence and the growth kinetics were exam- known to behave as pseudohalide ions, had no effect on the
ined. As shown in Figure 2a, upon the increase in the crystal growth when CTAB was employed as a main surfactant.
concentration of the iodide ions from 10 µM to 10 mM, the The formation of a nanoprism in the presence of iodide ions
intensity and position of the nanoprism band monotonically was dependent on the solution pH and the temperature. As the
decreased and blue-shifted, respectively, indicating that iodide pH of solution increased from ∼2.2 to ∼3.5, the nanoprism
ions over a critical concentration (∼5 µM) relatively suppress band became intensified with its position red-shifted (Figure
the growth of nanoprisms compared to faceted nanoparticles; 2d), indicating a facile formation of a nanoprism at a higher
at concentrations lower than ∼2 µM, the formation of nanorods pH condition. Unlikely with the concentrated iodide ion case,
was again observable. Moreover, the addition of iodide ion a larger nanoprism without a notable variation of its shape was
retarded the growth rate of all the nanoparticles, as can be seen synthesized, as shown in the lower images of Figure 2c.
in the following kinetic experiment (Figure 2b). In the kinetic However, further increase in the pH to higher than ∼4.0
UV/vis observations recorded at an interval of 150 s, the band gradually exacerbated the reproducibility of the crystal growth
at 530 nm was intensified without any movement of the position and even totally hindered nanoprism formation above pH ∼4.5.
during the crystal growth. In contrast, the band of nanoprism The feasible formation of a nanoprism with an addition of a
centered initially at ∼700 nm gradually red-shifted to ∼1000 small amount of NaOH has been observed in a previous report
nm, indicative of the growing dimension of a single nanoprism. by Mirkin et al.,9 but the growth mechanism was not presented
When the absorbance at 530 nm versus time evolution was in their investigation. Presumably, the facilitated formation of
recorded, it was also revealed that an increase in the concentra- a nanoprism at higher pH values should be understood on the
tion of iodide ions substantially slowed the kinetics of the faceted grounds that the reduction power of ascorbic acid can be
nanoparticles, as denoted in the inset of Figure 2b; the growth strengthened due to the elimination of protons produced by the
rate (at the concentration of 200 µM) estimated by the initial reduction.44 In other words, the chemical potential of ascorbic
1126 J. Phys. Chem. C, Vol. 111, No. 3, 2007 Ha et al.

acid is maintained at a higher state with the addition of NaOH. interest. As such, a plausible rationale is given for the present
However, the suppressing effect upon the further pH increase observation of mainly two different crystal growths under the
cannot be easily understood at this stage, though a plausible current reaction condition, thereby providing a more systematic
scenario is presented in a later section (vide infra). picture for the growth of gold nanocrystals.
Meanwhile, as the reaction temperature increases to 30 °C, At this stage, the gold reduction by iodide ions (3I- f I3-
spherical nanoparticles became dominant (data not shown). In + 2e-) is excluded since no noticeable change in color was
contrast, the band of nanoprisms became intensified and red- observed before the addition of the seed solution as noted in
shifted as the temperature was lowered to 10 °C, although with the Supporting Information; red pink from gold nanocrystals
some precipitation of CTAB during or before the crystal growth. or deep purple from iodine is expected from the iodide oxidation.
This observation indicates that a lower temperature is advanta- It is also noteworthy that chloride ions of AuCl4- appear to be
geous for the formation of nanoprisms, albeit with a substantial replaced by bromide ions excessively present in a CTAB
broadening of its size distribution. Thus, a water bath maintained solution (0.1 M). Similarly, although the ligand exchange was
at 15 °C was methodically used during the initial growth stage also observed in the iodide concentration higher than 0.2 mM,
(approximately for ∼1 h) and, subsequently, the reaction the growth solution demonstrated very similar spectral features
solution kept at an ambient temperature (20-25 °C) for further upon the addition of ascorbic acid; the solution turned transpar-
analyses. ent and exhibited no further color changes at least for 24 h as
Recently, the suppression of crystal growth by iodide ions observed in the growth solution without iodide ions. The ligand
was observed in the reduction of gold precursors by citric acid exchange appears not to play a major role in the nanoprism
and by the leaf extract of lemongrass.45,46 In these investigations, formation as the nanocrystals were successfully formed in a
the addition of iodide ions yielded a larger number of spherical very low concentration of iodide ions (10-200 µM), in which
nanoparticles while drastically decreasing the proportion of range minimal spectral changes (i.e., the ligand exchanges) were
nanoprisms. It also distorted the shapes of nanoprisms into observed (see the Supporting Informtion).
circular nanodisks (with fairly irregular edges). These observa- Upon the addition of a small amount of iodide ion, it is less
tions were in part repeated in the present experiment, as seen plausible to assume that the microstructures of the surfactant
in the upper images of Figure 2c; the proportion of spherical in a bulk solution will change, since the added chemical was
particles increased and the rate of crystal formation was of negligible amount compared to the concentrations of CTAB
drastically slowed down upon an increase in the concentration and HAuCl4. In general, the cationic surfactant forms a globular
of iodide ions. In these studies, it was concluded that bromide micelle structure in a concentration over its critical micelle
ions in the leaf extract of lemongrass had promoted the concentration (cmc, ∼1 mM for CTAB).36,37 As the concentra-
formation of nanoprisms; presumably, oligomeric sugar groups tion of the surfactant increases in aqueous solution, the
present in the leaf extract appear to reduce the gold ions. On microstructure experiences several phase transitions, becoming,
the other hand, chloride ions in the reduction via the Turkevich for instance, hexagonal, cubic, or lamellar structures depending
approach much stimulated the formation of nanoprisms. At any on the aliphatic chain length, cationic head group, salt concen-
rate, the addition of iodide ions strongly suppressed the trations, and/or added chemicals.47-49 Considering the working
formation of nanoprisms in both conditions. In contrast, for the concentration of the present system (0.10 M), spherical or short
growth in this study, bromide ions promoted the formation of rod-shaped micelle structures are dominant and appear to be
nanorods as a major product while a minuscule of iodide ion invariant with the addition of a minute amount of iodide ion.48,49
dramatically promoted the formation of nanoprisms. Additionally, it is notable that the reactive gold complexes (Au-
These controversial observations and subsequently distinct (III) or Au(I)) are in a tightly bounded form with the CTAB
conjectures for the growth mechanism should be attributed to micelle as revealed in a previous investigation.50
experimental differences on several points from the present Regardless of the bulk structure of CTAB, the bilayer
reaction parameters. First, the seeds both in the Turkevich formation (or micellar adsorption) on growing gold nanostruc-
approach and in the biogenic reduction are formed intact during tures is an additional possible explanation for the morphology
the reduction process, where seed formation is not virtually change, as the influx of chemicals into the gold surface appears
separated from a growth stage. However, in this study, the seed to be more tightly controlled through the bilayer structure.51
was first reduced with NaBH4 and added to a separated growth This picture can be further enforced by the observation that,
solution. Second, in the present work (employing CTAB), the irrespective of the presence of iodide ions, neither a well-defined
reduction reaction on a gold surface is quite different from that nanostructure was found nor were once-formed gold crystals
either in the Turkevich method or in the method using the leaf of random shapes stabilized in the absence of CTAB. It has
extract. The reducing agent (ascorbic acid vs citric acid or sugar) been reported that the cationic surfactant (CTAB) was adsorbed
as well as the oxidation state of gold ion (Au(I) vs Au(III)) onto negatively charged glass or mica surfaces, wherein the
might be a factor yielding the contradicting observations. As surface morphology of CTAB deviated from spherical micelle
will be mentioned later, a preferred face in a specific surface to rodlike (or wormlike tubular) micelles or planar bilayer
reaction is strongly dependent on a reaction condition. Finally, structures.36,51,52 Their adsorption kinetics and surface morphol-
the reaction time for the Turkevich method was typically ∼24 ogies were dependent on the bulk concentration, the ionic
h (the time being ∼12 h for the biogenic reduction), while strength, the counterion, and the chain length of a surfactant in
ascorbic acid in the seed-mediated reaction accomplished its question. At present, it is noteworthy that the adsorption of
reduction within ∼1 h. In addition, the formation of nanorods CTAB onto a gold surface was mediated by the precedent
was hardly observable in the reductions using citric acid or the specific adsorption of halide ions, rendering the surface to bear
leaf extract of lemongrass, suggesting quite different reaction net negative charges. In other words, under the CTAB template,
mechanisms and subsequently different results. On these an adsorption of bromide ions is followed by a bilayer formation
grounds, a picture should be avoided that Au(111) is the most on gold nanostructures; the adsorption of a CTA+ ion onto gold
hindered face at any gold reduction in the presence of iodide surface as expressed in a small number of papers appears to
ions; it would rather be dependent on a specific reaction of describe this picture implicitly.12,33 A series of adsorption
Shape-Controlled Syntheses of Gold Nanoprisms J. Phys. Chem. C, Vol. 111, No. 3, 2007 1127

experiments of CTAB with different counterions on a silicon

surface showed that a bromide ion has 5-fold greater binding
tendency for a CTA+ cation than a chloride ion on the micelle
surface.36,37 Simultaneously, a lower cmc in bulk solutions and
less curved micelle formation were facilitated in CTAB
compared to cetyltrimethylammonium chloride (CTAC); this
observation was attributed to the fact that the bromide ion is
much more effective in screening accumulated charges on the
micelle surface.36,37,47 In these studies, an iodide ion was reported
to bind even stronger with the cationic micelle surface compared
to a bromide ion.36,47 On these grounds, in the presence of an
iodide ion, a nanoprism seems to be favored over the formation
of a nanorod, which requires more of a curved surface geometry.
However, the stabilized bilayer formation appeared to make
rather a limited contribution to the morphology change, as the
concentration of the iodide ions was ∼5000 times lower than
that of bromide ion (10 µM vs 0.1 M) and the iodide ions were
equally distributed not only to the adsorbed micelle (or bilayer)
structure on the gold nanostructures but to globular micelles
present in the bulk solution.
Finally considered is whether the specific adsorption layer
underlying the CTAB bilayer structures was indeed affected by
the addition of iodide ions. The strong adsorption of halide ions
on low-indexed gold surfaces has been reported by a myriad of
studies.35 Many electrochemical investigations combined with
STM or surface X-ray scattering have revealed that the
adsorption power on an Au(111) surface is in the order of iodide,
bromide, and chloride ions.35,53,54 For instance, the Gibbs free
energy change during the specific adsorption of iodide ions on
an Au(111) surface was estimated to be much larger than that
of bromide ion by 30-70 kJ/mol, yet the precise values were
dependent on the surface potentials and on specific crystallin-
ity.35,54 As such, the added iodide ions appear to displace
bromide ions preadsorbed on the low-indexed facets of gold Figure 3. UV/vis spectra and corresponding typical TEM images of
seeds or as-grown gold crystals. The interaction of iodide ions diverse gold nanostructures formed under the CTAC template (a) with
with metal atoms was unique compared to chloride ions in that dilute concentrations of bromide ions (<1 mM), (b) with more
metal-halide bonding is dominated by covalent characteristics, concentrated solutions (>1 mM), and (c) in the presence of iodide ions
as evidenced in a SERS investigation showing potential- (50-100 µM) instead of bromide ions.
independent vibrational frequencies of the metal-halide bond.53,55 by reaction. Nevertheless, it seems quite plausible that the
This is further manifested in potential excursion experiments exclusive observation of Au(111) as a planar face in the
demonstrating the onset of iodide ion adsorption at extremely nanoprism is attributed to its markedly reduced reactivity caused
negative voltages (e.g., -1.0 V vs SCE reference). by iodide adlayer compared to other facets. However, this
Although the stronger adsorption tendency of iodide ions over conjecture should be confined in the present reaction species
other halide ions has been firmly supported through various (i.e., ascorbic acid and Au(I)-CTA+ system) since an adverse
experimental evidence, it has remained elusive which facet is trend was observed in the gold reduction by citric acid without
more reactive toward a gold reduction by ascorbic acid in the a noticeable surfactant; the gold nanoprism is one of the
presence of these halide ions. If one infers from facets observed preferred nanostructures with the addition of chloride ions.46
in the nanoprism formation, Au(111) faces associated with an To fortify the plausible scenario for the effect of halide ions,
iodide adlayer inhibit gold reduction more severely than other further control experiments were carried out by monitoring the
facets (i.e., Au(110) or Au(100)); in the presence of a bromide crystal shapes formed under various reaction constraints. As
adlayer, by contrast, Au(100) and/or Au(110) facets are less bromide ions remain a stronger adsorbate on low-indexed gold
reactive to the same reaction. According to investigations of surfaces compared to chloride ions,35 a new template system
several surface reactions, distinct surface activities for different employing CTAC appeared attractive as an alternative directing
facets have been reported. For instance, it was demonstrated agent; the length of aliphatic chain is identical for CTAC but
that an Au(100) face in an alkaline solution had the highest has a different counterion (i.e., chloride ion). This control system
activity in terms of an oxygen/hydrogen peroxide reduction was expected to be fruitful for outlining the overall role of halide
compared to other low-indexed surfaces, by nearly 4 orders of ions on crystal growth under a cationic surfactant system and
magnitude.56 Conversely, the Au(111) face was more effective was expected to provide additional evidence, albeit circumstan-
than Au(100) or Au(110) faces for the reduction of hexaami- tial, for the mechanistic picture described above. Currently,
necobalt ions.57,58 In spite of these examples, it is thus far CTAC is assumed to have a micelle structure nearly identical
impracticable to surmise which facet is less preferred during with that of CTAB due to the very similar chemical structures
the growth of a gold crystal, as the detailed adlayer structure of these two surfactants.36,37
and/or surface reconstruction by the adlayer are largely un- As shown in Figure 3a, only isotropic and nanorice-shaped
known;59 a preferred facet in a reduction seems to vary reaction nanoparticles, instead of nanorods or nanoprisms, were exclu-
1128 J. Phys. Chem. C, Vol. 111, No. 3, 2007 Ha et al.

sively observed under the CTAC template. Two corresponding

bands at ∼540 nm and ∼700 nm in the UV/vis spectra were
observed, respectively (Figure 3a). Upon the further increase
in the concentration of bromide ions up to ∼1 mM (with the
addition of KBr), the nanorice band at ∼700 nm become more
intensified, as seen in both the UV/vis spectra and TEM image
(Figure 3a and the inset). Although the nanorice seemingly
resembles bipyramidal structures formed recently in the presence
of silver ions, the band in fact was an ensemble of diverse
nanostructures. Intriguingly, the nanorice band abruptly dimin-
ished on the further increase in the bromide concentration
(Figure 3b); short nanorods were instead manifested at higher
concentrations over ∼5 mM. At even higher concentrations over
∼20 mM of bromide ions, the formation of nanorods was again
pronounced with a similar aspect ratio; concomitantly, some
nanoprisms were observable as in the CTAB system without
iodide ion. Over a concentration of ∼100 mM, the nanorod
formation was nearly identical with the pure CTAB system in
terms of the aspect ratio and relative yield (see the TEM image
of Figure 3b).
The manufacture of nanoparticles with few anisotropic
characteristics in the presence of chloride ions as a major
counterion is largely consistent with the picture mentioned
above; the facet-specific crystal growth seems less important
under a chloride adlayer considering the lower adsorption
tendency of chloride ion onto gold faces. The improved
production of the nanorice band (the band of ∼700 nm in
Figure 3a) in the presence of bromide ions (∼1 mM) appears
Figure 4. (a) Schematic illustration showing the shape-controlled
to have been caused by a cooperative adlayer formation of syntheses of Au nanoprisms and nanorods according to the presence
both halide ions, but the detailed mechanism of the morph- of iodide ions under a CTAB template. In the presence of iodide ions
ology development is left for a future study. Moreover, through (10 µM-10 mM), faceted nanoparticles (separately designated in TEM
these observations, it becomes clear that bromide ions are images) were simultaneously observed with triangular nanoprisms. In
essential for a nanorod formation under the CTA+-halide ion the absence of iodide ion, nanorods along with spherical nanoparticles
surfactant system. At a glance, this observation appears to be and nanoprisms were dominantly observed. (b) Further schematic
cartoon describing the influence of the CTAB template, in conjunction
contradicted with a recent study reporting the formation of with an iodide adlayer as a passive barrier for crystal growth during
shorter nanorods, instead of the nanorice particles, with the the formation of nanoprisms. The white arrows indicate the rate of
addition of CTAC.50 However, recalling that CTAB (8 mM) growth on specific facets.
was maintained in the growth solution as an invariant even in Under the CTAB template in the absence of iodide ions, the
the addition of CTAC (up to 4 mM), the report appears not dominant observation of nanorods can be rationalized by the
to be inconsonant with the present result but to be rather assumption that the bromide adlayer causes both the Au(110)
revived here. and Au(100) facets to be more hindered than Au(111) facets.
As a small amount of iodide ion (50 µM) was added to the Nonetheless, the overall growth rates on the bromide adlayer
reaction vessel containing CTAC, nanoprisms were again were much faster than those in the presence of iodide ions, and
observable with a complete absence of nanorod. Judging from the apparent preference for a specific facet seems less conspicu-
the corresponding UV/vis spectra, the yield of nanoprisms at a ous, as both nanostructures (i.e., nanorods and nanoprisms) were
concentration of ∼100 µM was reasonably comparable with the simultaneously found along with isotropic nanocrystals. This
value obtainable in the CTAB template with a trace of iodide rationale is fairly understandable with the adsorption tendency
ion but had a relatively broader size distribution (Figure 3c). of the bromide ion inferior to that of the iodide ion, and this
As the iodide concentration increased to over ∼200 µM, the situation is summarized in Figure 4a.
nanoprism band reduced in intensity and blue-shifted as In addition, the faceted isotropic nanoparticles considered as
observed in the CTAB template. Considering the relatively a major byproduct in the nanoprism formation (featured at the
higher concentration of iodide ions for the first appearance of ∼530 nm band in the UV/vis spectra) also provide some insight
the nanoprism compared to the CTAB system (50 µM vs 5 µM), into the crystal growth. Although isotropic nanoparticles
it is implicated that bromide ions also played some role in the acquired in the presence of iodide ions were only slightly
nanoprism formation with the addition of iodide ion under the distinguishable from those obtained in the absence in the UV/
CTAB template. Overall, these observations stress (1) that a vis spectra, the surface morphologies were slightly different in
trace amount of iodide ion is essential in the formation of the HRTEM image, as demonstrated in Figure 4a. The ridges
nanoprisms regardless of the concentration of bromide or between the 5-fold twinned gold domains in the presence of
chloride ions in the solution, (2) that chloride ions solely or in iodide ions were sharper than those observed in nanocrystals
combination with bromide ions under a critical concentration obtained without iodide ions (see the TEM images in Figure
(∼1 mM), conversely, stimulate the formation of nanoparticles 4a). This observation suggests that the crystal growth was more
of irregular shapes or ill-defined nanorice crystals, and (3) that tightly controlled depending on the specific crystal facet
bromide ions over a set concentration (∼20 mM) facilitated the exposed. It is also noteworthy that these faceted nanocrystals
formation of nanorods. appear to be species “isomorphous” with a nanoprism, as the
Shape-Controlled Syntheses of Gold Nanoprisms J. Phys. Chem. C, Vol. 111, No. 3, 2007 1129

external faces of a typical isotropic nanoparticle are mostly suppressed the overall crystal growth through specific adsorption
bounded with Au(111) faces (data not shown); a single and stimulated the formation of nanoprisms as a major product.
nanoprism has only two Au(111) facets (see the TEM images Conversely, bromide ion larger than ∼20 mM preferred to form
in Figure 4a). This isomorphic behavior suggests that the faceted nanorods with a minor fraction of nanoprisms, while chloride
nanoparticles can scarcely be excluded during the growth stage, ions solely or in combination with small amount of bromide
which is evidenced in the various control experiments described ion yielded gold nanorice (or distorted bipyramidal nanostruc-
above. When the reduction potential of gold precursors increased tures). With the addition of iodide ions, the Au(111) face was
by means of pH or temperature elevations, these faceted particles more severely hindered compared to other faces (i.e., Au(110)
were dominant products compared to triangular nanoprisms, and Au(100)); as a result, all of the flat faces of the triangular
indicating that these isotropic particles are kinetically the more nanoprisms were bounded in that direction. In all cases, faceted
favored species.60 Intriguingly, even upon the suppression of isotropic nanoparticles were cosynthesized as a byproduct that
crystal growth through the increase in the concentration of iodide could not be excluded under the present synthetic parameters.
ions (as demonstrated in Figure 2a), the faceted isotropic The most striking feature of this study is that iodide ions at an
nanoparticles were again the dominant products compared to impurity level dramatically changed the morphologies of gold
the nanoprisms. Overall these observations indicate that the nanocrystals formed under the same cationic surfactant system
formation of nanoprism in the presence of iodide ion appears (from nanorods to nanoprisms). This study not only elaborates
to occur in a narrow window of reaction parameters (vide supra). the recent controversial contention that substantial nanoprisms
Thus far, the influence of halide ions on the syntheses of are synthesized under a reaction condition designed to produce
gold nanocrystals has been demonstrated through the production nanorods but also provides a deeper insight into the seeded
of diverse crystal shapes, while the chemical backbone of a crystal growth.9 However, with recall of a recent report that
cationic surfactant was maintained as an invariant. On the basis chloride ions promote the formation of nanoprism while iodide
of the morphology deviations under several combinations of ions strongly suppressing it in the gold reduction with citric
halide ions, it appears that the specific adsorption of halide ions acid, it seems obvious that the hindered face with the addition
exerts major exquisite control over the shape evolutions by of iodide ion (i.e., Au(111)) is not applicable to all other gold
allowing distinct growth rates dependent on specific gold faces reductions; the determination of a specific face preferred during
under the current seed-mediated redox pair (Au(I)/ascorbic acid). the growth of nanostructures appears to be governed by the
Alternatively, the CTAB template that was assumed to be a physicochemical nature of the seed preparation, selection of
major geometrical constraint, in fact, made a limited contribution reducing materials, and the presence of relevant surfactant.
to the crystal growth as an agent for the passive protection of Additionally, it is also obligatory to understand and manipulate
the as-grown crystals. The limited role of the surfactant was the crystallinity, twinning, surface charges, and size of the seed
already suggested in an investigation performed by the Pileni used, to realize an “on-demand” shape control in the seed-
group upon the synthesis of copper nanocrystals; in their study, mediated growth.
AOT was used as a soft template and the effects of diverse
anions, with the exception of the iodide ion, were investi- Acknowledgment. This work has been supported by grants
gated.61,62 The authors claimed that a specific adsorption of from the Protein Chip Technology Program (MOST, Korea)
anions dominated the shape evolution; however, their assertion and KRIBB Initiative Research Program (KRIBB, Korea). We
has been mostly overlooked in the seed-mediated production are grateful to Y. C. Park (NNFC, Korea) for technical
of gold nanocrystals. Hence, the drastic influence of iodide ions comments on the TEM study.
on the formation of nanoprisms may support the crucial role of
halide adsorption on the formation of gold nanocrystals. Supporting Information Available: Experimental details
In Figure 4b, the growth of gold nanocrystals under the CTAB including TEM figures. This material is available free of charge
template system was schematically illustrated. In the presence via the Internet at http://pubs.acs.org.
of iodide ions, a stable CTAB bilayer on an iodide adlayer
preferentially formed due to its overwhelming adsorption References and Notes
tendency compared to other halide ions and controls the supply (1) Wilcoxon, J. P.; Provencio, P. P. J. Am. Chem. Soc. 2004, 126,
of added chemicals into gold surface passively. However, at 6402-6408.
the edge faces of nanoprisms (i.e., Au(110) faces), more curved (2) Kim, P. P.; Oh, S.; Crooks, R. M. Chem. Mater. 2004, 16, 167-
172.
spherical micelles should be formed as a result of geometric
(3) Pastoriza-Santos, I.; Liz-Marzan, L. M. Langmuir 2002, 18, 2888-
factors, which relatively expedites the influx of the chemicals 2894.
in the edge directions. In fact, a theoretical consideration (4) Lee, I.; Han, S. W.; Kim, K. Chem. Commun. 2001, 1782-1783.
revealed that mean free passage times for ions are shortest at (5) Zhang, J.; Worley, J.; Dénommée, S.; Kingston, C.; Jakubek, Z.
the curved surfaces.50 Associated with the differentiated reactiv- J.; Deslandes, Y.; Post, M.; Simard, B. J. Phys. Chem. B 2003, 107, 6920-
6923.
ity of Au(110) face with an iodide adlayer, the crystal growth (6) Mallin, M. P.; Murphy, C. J. Nano Lett. 2002, 2, 1235-1237.
rate in that direction seems to be more facilitated. Nonetheless, (7) Freeman, R. G.; Hommer, M. B.; Grabar, K. C.; Jackson, M. A.;
the dramatic morphology change from nanorod to nanoprism Natan, M. J. J. Phys. Chem. 1996, 100, 718-724.
can be attributed to the iodide adlayer due to the relatively low (8) Rex, M.; Hernandez, F. E.; Campiglia, A. D. Anal. Chem. 2006,
78, 445-451.
concentration of iodide ions compared to excessive bromide (9) Millstone, J. E.; Park, S.; Shuford, K. L.; Qin, L.; Schatz, G. C.;
ions. Mirkin, C. A. J. Am. Chem. Soc. 2005, 5312-5313.
(10) Metraux, G. S.; Mirkin, C. A. AdV. Mater. 2005, 17, 412-415.
3. Conclusions (11) Jin, R.; Cao, Y. C.; Hao, E.; Metraux, G. S.; Schatz, G. C.; Mirkin,
C. A. Nature 2003, 425, 487-490.
Three distinct gold nanostructures of a nanoprism, nanorod, (12) Murphy, C. J.; Sau, T. K.; Gole, A. M.; Orendorff, C. J.; Gao, J.;
Gou, L.; Hunyadi, S. E.; Li, T. J. Phys. Chem. B 2005, 109, 13857-13870.
and nanorice were manufactured by solely controlling the (13) Yu, D.; Yam, V. W. -W. J. Am. Chem. Soc. 2004, 126, 13200-
content of halide ions. A miniscule amount of iodide ion in the 13201.
presence of an excessive amount of other halide ions strongly (14) Sun, Y; Xia, Y. Science 2002, 298, 2176-2179.
1130 J. Phys. Chem. C, Vol. 111, No. 3, 2007 Ha et al.

(15) Jana, N. R.; Gearheart, L.; Murphy, C. J. AdV. Mater. 2001, 13, (39) Im, S. H.; Lee, Y. T.; Wiely, B.; Xia, Y. Angew. Chem., Int. Ed.
1389-1393. 2005, 44, 2154.
(16) Nikoobakht, B.; El-Sayed, M. A. Chem. Mater. 2003, 15, 1957- (40) Wang, Z. L.; Mohamed, M. B.; Link, S.; El-Sayed, M. A. Surf.
1962. Sci. 1999, 440, L809-L814.
(17) Orendorff, C. J.; Gole, A.; Sau, T. K.; Murphy, C. J. Anal. Chem. (41) Liu, M.; Guyot-Sionnest, P. J. Phys. Chem. B. 2005, 109, 22192-
2005, 77, 3261-3266. 22200.
(18) Nikoobakht, B.; El-Sayed, M. A. J. Phys. Chem. A 2003, 107, (42) Rai, A.; Singh, A.; Ahmad, A.; Sastry, M. Langmuir 2006, 22,
3372-3372. 736-741.
(19) Raschke, G.; Brogl, S.; Susha, A. S.; Rogach, A. L.; Klar, T. A.; (43) Gai, P. L.; Harmer, M. A. Nano Lett. 2002, 2, 771-774.
Feldmann, J.; Fieres, B.; Petkov, N.; Bein, T.; Nichtl, A.; Kürzinger, K. (44) Wang, C.; Wang, T.; Ma, Z.; Su, Z. Nanotechnology 2005, 16,
Nano Lett. 2004, 4, 1853-1857. 2555-2560.
(20) Sun, Y.; Xia, Y. Anal. Chem. 2002, 74, 5297-5305. (45) Rai, A; Singh, A.; Ahmad, A.; Sastry, M. Langmuir 2006, 22, 736-
(21) Sönnichsen, C.; Alivisatos, A. P. Nano Lett. 2005, 5, 301-304. 741.
(22) Huang, X.; El-Sayed, I. H.; Qian, W.; El-Sayed, M. A. J. Am. Chem. (46) Shankar, S. S.; Bhargava, S.; Sastry, M. J. Nanosci. Nanotechnol.
Soc. 2006, 128, 2115-2120. 2005, 5, 1721-1727.
(23) Huang, W.; Qian, W.; El-Sayed, M. A. J. Phys. Chem. B 2005, (47) Blackmore, E. S.; Tiddy, G. J. T. J. Chem. Soc., Faraday Trans. 2
109, 18881-18888. 1988, 84, 1115-1127.
(24) Mokari, T.; Rothenberg, E.; Popov, I.; Costi, R.; Banin, U. Science (48) Henriksson, U.; Blackmore, E. S.; Tiddy, G. J. T.; Söderman, O.
2004, 304, 1787-1790. J. Phys. Chem. 1992, 96, 3894-3902.
(25) Shevchenko, E. V.; Talapin, D. V.; Kotov., N. A.; O’Brien, S.; (49) Krishnaswamy, R.; Ghosh, S. K.; Lakshmanan, S.; Raghunathan,
Murray, C. B. Nature 2006, 439, 55-59. V. A.; Sood, A. K. Langmuir 2005, 21, 10439-10443.
(26) Shi, W.; Zeng, H.; Sahoo, Y.; Ohulchanskyy, T. Y.; Ding, Y.; Wang, (50) Pérez-Juste, J.; Liz-Marzan, L. M.; Carnie, S.; Chan, D. Y. C.;
Z. L.; Swihart, M.; Prasad, P. N. Nano Lett. 2006, 6, 875-881. Mulvaney, P. AdV. Funct. Mater. 2004, 14, 571-579.
(27) Steiner, D.; Mokari, T.; Banin, U.; Millo, O. Phys. ReV. Lett. 2005,
(51) Nikoobakht, B.; El-Sayed, M. A. Langmuir 2001, 17, 6368-6374.
95, 056805.
(28) Jana, N. R. Chem. Commun. 2003, 1950-1951. (52) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. J. Colloid Interf.
(29) Kim, F.; Song, N. R.; Yang, P. J. Am. Chem. Soc. 2002, 124, Sci. 2003, 266, 236-244.
14316-14317. (53) Gao, P.; Weaver, M. J. J. Phys. Chem. 1986, 90, 4057-4063.
(30) Chang, S.; Shin, C.; Chen, C.; Lai, W.; Wang, C. R. Chris Langmuir (54) Chen, A.; Shi, Z.; Bizzotto, D.; Lipkowski, J.; Pettinger, B.; Bilger,
1999, 15, 701-709. C. J. Electroanal. Chem. 1999, 467, 342-353.
(31) Gao, J.; Bender, C. M.; Murphy, C. J. Langmuir 2003, 19, 9065- (55) Gao, X.; Edens, G. J.; Liu, F.; Hamelin, A.; Weaver, M. J. J. Phys.
9070. Chem. 1994, 98, 8086-8095.
(32) Gole, A.; Murphy, C. J. Chem. Mater. 2004, 16, 3633-3640. (56) Markovic, N. M.; Tidswell, I. M.; Ross, P. N. Langmuir 1994, 10,
(33) Sau, T. K.; Murphy, C. J. J. Am. Chem. Soc. 2004, 126, 8648- 1-4.
8649. (57) Hromadova, M.; Fawcett, W. R. J. Phys. Chem. A 2000, 104, 4356.
(34) Sau, T. K.; Murphy, C. J. Langmuir 2004, 20, 6414-6420. (58) Hromadova, M.; Fawcett, W. R. J. Phys. Chem. A 2001, 105, 104-
(35) Magnussen, O. M. Chem. ReV. 2002, 102, 679-725 and references 111.
therein. (59) Wang, Z. L.; Gao, R. P.; Nikoobakht, B.; El-Sayed, M. A. J. Phys.
(36) Velegol, S. B.; Fleming, B. D.; Biggs, S.; Wanless, E. J.; Tilton, Chem. B 2000, 104, 5417-5420.
R. D. Langmuir 2000, 16, 2548-2556. (60) Lofton, C.; Sigmund, W. AdV. Funct. Mater. 2005, 15, 1197-1208.
(37) Magid, L. J.; Han, Z.; Warr; G. G.; Cassidy, G. G.; Butler, P. D.; (61) Filankembo, A.; Pileni, M. P. J. Phys. Chem. B 2000, 104, 5865-
Hamilton, W. A. J. Phys. Chem. B 1997, 101, 7919-7927. 5868.
(38) Li, C.; Cai, W.; Li, Y.; Hu, J.; Liu, P. J. Phys. Chem. B 2006, 110, (62) Filankembo, A.; Giorgio, S.; Lisiecki, I.; Pileni, M. P. J. Phys.
1546-1552. Chem. B 2003, 107, 7492-7500.